Isolation and Hplc Quantitation of Kaurane-type Diterpenes and Cinnamic Acid Derivatives of Long-term Stored Leaves of Mikania Laevigata and Mikania Glomerata

Isolation and HPLC quantitation of kaurane-type diterpenes and cinnamic acid derivatives of long-term stored leaves of Mikania laevigata and Mikania glomerata ABSTRACT The leaves of Mikania laevigata and Mikania glomerata are used in Brazil to treat respiratory affections, being kaurane-type diterpenes and coumarin considered as the bioactive compounds. The present study reports an investigation on the HPLC-DAD profi les and contents of coumarin (1), trans-o-coumaric (2), kaurenoic (3), benzoylgrandifl oric (4) and cinnamoylgrandifl oric (5) acids in dried leaves of Mikania species stored in dark room under controlled conditions. Excepting 2, the constituents were isolated and purifi ed to be employed as reference compounds. The samples were analyzed at three monthly intervals up to 18 months for M. laevigata and 12 months for M. glomerata. trans-o-Coumaric was not detected in both, whereas 1 occurred only in M. laevigata. The concentrations of the assayed constituents did not vary signifi cantly within the evaluated period (p vary signifi cantly within the evaluated period (p vary signifi cantly within the evaluated period (< 0.05), for both species. In contrast, changes in the chromatographic profi les and spectral purity of peaks from 3, 4 and 5 were detected in samples of both Mikania stored for three months, while the coumarin profi le in M. laevigata modifi ed after six months of storage. The evaluation of chromatographic profi les based on spectral purity analyses of selected peaks was shown to be a more robust tool to access chemical stability of Mikania samples than the quantitation of chemical markers' contents.


INTRODUCTION
The validation of medicinal plants has only gained attention of the Brazilian scientifi c community and government in the last decade (Bertolucci et al. 2009); for that reason, the quality control of herbal drugs in the country is limited by the reduced number of chemical markers defi ned for native species, along with the lack of commercial sources of reference compounds (Braga et al. 2003).
Different chromatographic techniques are employed for the quality control of vegetal drugs and herbal products including qualitative and quantitative methods based on TLC, HPLC and GC analyses (Liang et al. 2009, Razmovski-Naumovski et al. 2010).Chromatographic methods 473-485 SUZAN K.V. BERTOLUCCI, ANA B.D. PEREIRA, JOSÉ E.B.P. PINTO, ALAÍDE B. OLIVEIRA and FERNÃO C. BRAGA are extensively used for the quality control of plant raw materials due to their ability to detect variations in chemical composition originated from intraspecifi c differences, growing conditions, harvest time, processing methods and storage period (Liang et al. 2009).The development of HPLC methods for plant analysis is not straightforward and involves exhaustive optimization of operating conditions (e.g., mobile phase composition, pH, column and temperature) in order to obtain desirable outcomes such as higher plate number, shorter analysis times, peak purity and improved peak resolution (Dharmadi and Gonzalez 2005) -this last in the case of simultaneous analysis of multiple compounds.
Mikania laevigata Schultz Bip.ex Baker Mikania laevigata Schultz Bip.ex Baker Mikania laevigata and Mikania glomerata Sprengel (Asteraceae) Mikania glomerata Sprengel (Asteraceae) Mikania glomerata are medicinal species popularly known as guaco, widely used in Brazil to treat respiratory affections (Napimoga and Yatsuda 2010).Kaurane-type diterpenes from the ent-series [kaurenoic (KA), benzoylgrandifl oric (BA) and cinnamoylgrandifl oric (CA) acids] and the derivatives of cinnamic acid, coumarin (CO) and trans-o-coumaric acid (OC) (Fig. 1) have been identifi ed as constituents of the species (Veneziani and Oliveira 1999, Vilegas et al. 1997, Oliveira et al. 1984) and may account for their alleged biological properties (Ambrosio et al. 2006, Moura et al. 2002).Previous reports have suggested similar chemical composition for the leaves of both Mikania (Veneziani and Oliveira 1999, Vilegas et al.Mikania (Veneziani and Oliveira 1999, Vilegas et al. Mikania 1997, Oliveira et al. 1984).However, we reported a marked difference in the constitution of M. laevigata and M. glomerata leaves, with the presence of KA, BA and CA in both and the lack of CO and OC in the last one (Bertolucci et al. 2009).Besides, we found higher contents of BA and CA in M. laevigata, while KA was disclosed as the most abundant compound in M. glomerata (Bertolucci et al. 2009).

M. glomerata
Safety and effi cacy of phytopharmaceutical products are directly affected by the chemical reliability at all stages of the manufacturing processes, including stability and shelf-life of vegetal drugs (Sahoo et al. 2010).Nevertheless, there are only a few reports addressing post-harvest effects, storage and shelf-life of medicinal plants (Guimarães et al. 2011, Madan et al. 2008, Stafford et al. 2005, Fennell et al. 2004, Griggs et al. 2001).As far as we know, storage-related changes in the constituents of Mikania species have never been investigated.
Mikania species have never been investigated.Mikania Taking into account that cinnamic acid derivatives and kaurane-type diterpenes are considered the bioactive compounds of guaco, the goal of the present study was to investigate the HPLC profi les and contents of CO, OC, KA, BA and CA in stored leaves, thus requiring the isolation of some of them to be employed as reference compounds.

GENERAL EXPERIMENTAL PROCEDURES
Fractionation of the extracts and purifi cation of the isolated compounds were carried out by column chromatography on silica gel 60G (Merck 0.2-0.5 mm and Merck 0.063-0.200mm) and by preparative TLC on self-coated plates with silica gel 60G (Merck, 70-230 µm).Liebermann-Burchard, p-anisaldehyde and NP/PEG solutions were employed as spray reagents for monitoring the fractionation, along with ammonium vapor exposure, the detection was carried out under UV light at 254 and 366 nm (Wagner et al. 1984).
A Shimadzu preparative HPLC system composed of LC-8A quaternary pump, SCL-8A controller, SPD-6AV UV-VIS detector and CR4A integrator was employed for the fi nal purifi cation of the compounds.Analyses were performed on Shimpack Prep Sil (250 × 10 mm d.i.,Shimadzu) and Shim-pack Prep-ODS (250 × 10 mm d.i.) columns.Purity of the isolated compounds was checked by melting point data (without correction) determined on MQAPF-301 apparatus (Microquímica), HPLC-DAD and NMR analysis, whereas their identifi cation was accomplished by spectroscopic analysis (UV, CHEMICAL MARKERS CONTENTS IN STORED Mikania LEAVES IR, 1 H and 13 C NMR).IR spectra were recorded on a Perkin-Elmer FT-IR spectrophotometer in ATR mode with internal reference (range 4,000-600 cm −1 ). 1 H and 13  H and 13 H and C NMR spectra were obtained on Bruker Avance DRX-200 and DRX-400 equipments (Departamento de Química, UFMG), operating respectively at 200 and 400 MHz for 1 H and at 50 and 100 MHz for 13 C. TMS was employed as internal standard for both nuclei and CDCl 3 solutions were used in the analyses.Optical rotations were determined using a Bellingham Stanley ADP220 automatic recording spectropolarimeter.The analyses of chemical markers were carried out on a Waters 2695 HPLC system composed of a quaternary pump model L-6200A, autosampler, in-line degasser AF (Waters) and photo-diode array detector (Waters 2996).Waters Empower software was employed for data processing.The analyses were performed on a LiChrospher 100 RP-18 column (125 × 4 mm i.d., 5µm; Merck), in combination with a LiChrospher 100 RP-18 guard column (4 × 4 mm i.d., 5µm; Merck).

ISOLATION AND SPECTROSCOPIC CHARACTERIZATION OF CHEMICAL MARKERS
Mikania glomerata and Mikania laevigata leaves were collected from specimens cultivated at the Departamento de Agricultura, Universidade Federal de Lavras, Minas Gerais, Brazil.The leaves were collected in summer (January, 2004), from the apical, intermediate and basal regions of 16-month old plants, cultivated under solar radiation.
The species were identifi ed by Dr. Mara Rejane Ritter from the Instituto de Biociências, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil, where vouchers are deposited under the codes ICN 141992 and ICN 141990, respectively.The leaves were dried in a ventilated oven at 40°C for 72 h.The dried materials (2 kg of Mikania glomerata and 4 kg of Mikania laevigata) were ground in a knife mill, following exhaustive percolation with ethanol 96% at room temperature.Solvent was removed in a rotatory evaporator and the obtained extracts were kept in desiccators until the complete elimination of the residual solvent.The obtained crude extracts (331.6 g of M. glomerata and 564.9 g of M. laevigata) were employed for the isolation of chemical markers.

DEVELOPMENT AND OPTIMIZATION OF HPLC CONDITIONS FOR THE ANALYSIS OF CHEMICAL MARKERS
For the analysis, 1 g of dried leaves was submitted to ultrasound-assisted extraction with ethanol (2 × 30 mL), at room temperature (20 min each cycle).The extract was fi ltered through fi lter paper to a 100 mL volumetric fl ask and fi lled up with ethanol.An aliquot (10 mL) was taken and the solvent removed in a rotatory evaporator at 50°C under reduced pressure.The obtained residue was dissolved in MeOH (1 mL) and centrifuged at 8,400 g for 10 min, being the supernatant employed for analysis.During method development, fortifi ed extracts of Mikania laevigata were prepared by adding OC to fi nal concentration of 0.1 mg/mL, whereas Mikania glomerata extracts Mikania glomerata extracts Mikania glomerata were spiked with OC and CO (fi nal concentrations of 0.1 mg/mL each) before sonication.
Method development and optimization comprised the modifi cation of parameters related to chromatographic resolution, including organic modifi er, temperature, fl ow rate and slope of gradient elution.Phosphoric acid 0.1% was added to the eluents in all assayed conditions.Peak identifi cation was based on UV data recorded on line by DAD, as well as on the co-injection of reference compounds.Chromatographic selectivity was accessed by analyzing UV spectra recorded by DAD in the ascending, upper and descending regions of the peaks, being considered pure those peaks whose spectra matched exactly.The condition was considered selective when all fi ve analyzed chemical markers showed adequate peak purity.
The effi ciency of the established chromatographic condition was also assessed by determining system suitability parameters (resolution, retention factor, tailing and number of plates) of the chemical markers.The parameters were determined using Waters Empower software according to the equations recommended by the United States Pharmacopoeia (USP 2006) and the results were the mean value of 6 replicates.Determination of t 0 was accomplished by the injection of 10 µL of sodium nitrate solution 0.01% (m/v) prepared in MeOH.The fresh leaves were conditioned in kraft paper bags and dried in a ventilated oven at 40°C until constant weight.The paper bags were transferred into polypropylene bags and stored in a dark room with controlled temperature (25 ± 5°C) and humidity (54 ± 20%).M. glomerata and M. glomerata and M. glomerata M. laevigata samples M. laevigata samples M. laevigata were stored during 12 and 18 months, respectively.
Samples of the stored material were analyzed at three-month intervals.The leaves were ground in a knife mill, sieved in 0.85 mm tamis and freezer stored (-20°C) in safelock polypropylene bags until analysis.The developed chromatographic method was applied to assess changes in the chemical markers during storage of the plant material.Analysis was based on peak purity disclosed by UV spectral curves recorded on line for CO, OC, BA, CA CHEMICAL MARKERS CONTENTS IN STORED Mikania LEAVES and KA.The contents of these compounds were quantified using a method previously described by us (Bertolucci et al. 2009).

STATISTICAL ANALYSIS
Statistical design was entirely casualized delineation, with 7 treatments for Mikania laevigata (T 0 , T 3 , T 6 , T 9 , T 12 , T 15 and T 18 , respectively to experiment start and 3-18 months of storage) and 5 treatments for Mikania glomerata (T 0 , T 3 , T 6 , T 9 and T 12 ) with 6 repetitions each.The contents of constituents were submitted to ANOVA, followed by the Scott-Knott test, employing Sisvar software, version 5.0 (Ferreira 2007).Data were considered signifi cantly different when p < 0.05.

ISOLATION OF CHEMICAL MARKERS
The crude ethanolic extracts from the leaves of Mikania laevigata and Mikania glomerata were fractionated to isolate the major compounds in the species, to be employed as chemical markers.Fractionation of M. laevigata afforded coumarin (CO) and three ent-kaurane diterpenes, namely benzoylgrandifl oric (BA), kaurenoic (KA) and cinnamoylgrandifl oric (CA) acids.KA was also isolated from M. glomerata extract.Structure identifi cation of the isolated compounds was achieved by usual spectroscopic analysis and comparison with previously reported data (Kupriyanova 1997, Oliveira et al. 1984, Rowbotham and Schaefer 1972, Batista et al. 2005, Velandia et al. 1998, Fabbri et al. 1997, Vichnewski et al. 1977), as well as by TLC and HPLC analysis employing authentic samples.
The occurrence of the isolated compounds in the leaves of Mikania laevigata and Mikania laevigata and Mikania laevigata Mikania glomerata was investigated by HPLC analysis, along glomerata was investigated by HPLC analysis, along glomerata with trans-o-coumaric acid (OC), a biosynthetic precursor of coumarin.Identifi cation of the peaks corresponding to the isolated compounds was achieved by comparison with retention time, UV spectra and co-injection of reference compounds.Altogether, these data allowed us to propose OC, CO, KA, BA and CA as chemical markers for M. laevigata, whereas only kaurane-type diterpenes (KA, BA and CA) were found in M. glomerata CA) were found in M. glomerata CA) were found in leaves.M. glomerata leaves.M. glomerata Development and optimization of the HPLC-DAD method comprised the evaluation of 14 different conditions.Mikania laevigata was selected for method development in view of its more complex profi le than Mikania glomerata.A sample spiked with OC was employed for method development due to the low intensity of its peak found in the exploratory run.Acetonitrile and methanol were tested as organic modifi ers; ACN exhibited better resolution for the kaurane diterpenes, whereas MeOH improved resolution for the cinnamic acid derivatives.Therefore, both solvents were employed for elution.The chromatographic conditions were defi ned after exhaustive adjustment in the elution strength, gradient slope, temperature and fl ow rate.Despite several attempts to improve resolution, elution of BA and CA was partially superimposed to other peaks (Fig 1a).Spectral data recorded by DAD for BA and CA indicated maximum wavelength absorptions at 230 and 270 nm, respectively, while the partially coeluted compounds showed maximum absorption at 210 nm.Given that all previous attempts to increase resolution for the kaurane derivatives had failed, we selected different wavelengths to register the chemical markers, respectively 210 nm for OC, CO and KA, 230 nm for BA and 270 nm for CA, thus resulting in spectral purity for each chemical marker peak.The established conditions allowed the unambiguous identifi cation of the fi ve constituents in the leaves of M. laevigata.Finally, we introduced a timed wavelength program for chemical markers detection, allowing registering them at the wavelength of maximum absorbance in one single chromatogram (Fig. 1b).The established chromatographic conditions are described in Table I.
The reliability of the established conditions was checked by system suitability tests, comprising evaluation of resolution (Rs), tailing factor (T), number of plates (N), retention factor (k) and repeatability k) and repeatability k  I.The chromatograms were recorded using the timed wavelength program, except in "a", registered at 210 nm.The sample in "a", "b" and "c" were spiked with o-coumaric acid, along with coumarin in "c".Peaks: OC, trans-o-coumaric acid; CO, coumarin; BA, ent-benzoylgrandifl oric acid; CA, ent-cinnamoylgrandifl oric acid; KA, ent-kaurenoic acid.
of peak response [RSD of retention time (n = 6) for chemical markers' peaks].The results are presented in Table II, along with the limits recommended by the U.S. Food and Drug Administration (FDA 2000).All values are in accordance with FDA excepting the resolution between CO and OC.Such fi nding may not constitute a problem, since it is accepted that system suitability tests might be less rigorous for biological matrices and traces analysis (Dong et al. 2001).Moreover, some authors consider Rs ~ 1.5 adequate for quantitative analysis (Meyer 1996, Dong et al. 2001).Precision is critical for analytical routine methods; variation in retention time between runs indicates low precision.In the established method, a maximum RSD for retention time was observed for CO (0.40%), attesting the high reproducibility of the chromatographic conditions, including a suffi cient re-equilibration interval.The HPLC conditions employed to record Mikania laevigata profi le were applied to Mikania glomerata samples, disclosing kaurane-type diterpenes as major constituents (Fig. 1d).However, the occurrence of CO in the species has been previously reported (Santos et al. 2006, Oliveira et al. 1984) and therefore the chromatographic conditions were not modifi ed, since coumarin might be present in other samples of M. glomerata.
Similarly to Mikania laevigata, system suitability tests were carried out for Mikania glomerata (Table II).All analyzed parameters fell within the limits established by FDA (FDA 2000), excepting the resolution for KA.In spite of that, the obtained value (Rs = 1.73) may be considered adequate for quantitative analysis (Dong et al.

TABLE I Chromatographic conditions established for the analysis of chemical markers in Mikania laevigata
and Mikania glomerata leaves.
a All eluents were acidifi ed with 0.1% phosphoric acid.

TABLE II System suitability parameters determined for the analysis of chemical markers in Mikania laevigata and
Mikania glomerata leaves, employing the chromatographic conditions described in Table I.Besides, spectral homogeneity attested peak purity for the three kaurane diterpenes, indicating method selectivity for M. glomerata.

Chemical markers Parameters
Considering that the analyzed Mikania species Mikania species Mikania present different matrices and that CO and OC may occur in other samples of Mikania glomerata, evaluation of method selectivity was also mandatory for this species.A sample of M. glomerata was spiked M. glomerata was spiked M. glomerata with OC and CO and the resulting chromatogram exhibited peak purity for both compounds (Fig. 1c).System suitability tests carried out for these chemical markers in the fortifi ed sample of M. glomerata were M. glomerata were M. glomerata also in accordance with FDA recommendations (FDA 2000), apart from the resolution of OC and CO, attesting method selectivity and precision (Table II).
The results obtained so far indicate a marked difference in the analyzed samples of Mikania laevigata and Mikania glomerata, with the presence of the kaurane-type derivatives (KA, BA and CA) in both and the lack of the cinnamic acid derivatives (CO and OC) in the second species.This fi nding corroborates the absence of coumarin previously reported for specimens of M. glometara collected in distinct locations of São Paulo state, Brazil (V.L.G.Rehder et al., unpublished data).
However, we cannot assure that cinnamic acid derivatives are not produced by Mikania glomerata, since both CO and OC have been already reported for the species (Santos et al. 2006, F. Bras. IV 2005, Veneziani and Oliveira 1999, Oliveira et al. 1984).

The confl icting data may arise from misidentifi cation of Mikania species based exclusively on foliar
Mikania species based exclusively on foliar Mikania morphology, as described for the medicinal species Maytenus ilicifolia Mart.ex Reiss.and Maytenus ilicifolia Mart.ex Reiss.and Maytenus ilicifolia Maytenus aquifolium Mart.(Duarte andDebur 2005, Tiberti et al. 2007).The HPLC-DAD profi les here reported may contribute to overcome this problem and together with DNA fi ngerprints they will represent useful tools for the identifi cation of Mikania laevigata and Mikania laevigata and Mikania laevigata M. glomerata, allowing their unambiguous use for production of phytopharmaceuticals.

CHEMICAL MARKERS ANALYSIS DURING STORAGE
The HPLC profi les and quantitative composition of samples from Mikania leavigata and Mikania leavigata and Mikania leavigata Mikania glomerata leaves, stored under controlled conditions, glomerata leaves, stored under controlled conditions, glomerata were analyzed three-monthly.CO and OC were not detected in M. glomerata within the evaluated period M. glomerata within the evaluated period M. glomerata (Fig. 2, Table III), whereas CO was present in M. laevigata (Fig. 3, Table III).After six months of laevigata (Fig. 3, Table III).After six months of laevigata storage, peak purity analysis of CO revealed a coeluted compound, clearly detected in 12-month and 18-month stored plants (Fig. 4, CO), whose spectral data is compatible with a cinnamic acid derivative.
The CO content of 18-month stored plants (0.124 ± 0.026 %) was statistically similar to time zero samples (0.113 ± 0.033 %) (p zero samples (0.113 ± 0.033 %) (p zero samples (0.113 ± 0.033 %) ( < 0.05; Table III).Likewise, the concentration of the kaurane diterpenes did not vary signifi cantly within the evaluated storage period (p evaluated storage period (p evaluated storage period ( < 0.05), for both species (Table III).However, spectral analysis of the peaks pointed out co-eluting compounds in samples of Mikania laevigata and Mikania laevigata and Mikania laevigata Mikania glomerata stored Mikania glomerata stored Mikania glomerata for 6 and 3 months, respectively (data not shown).A marked decrease in resolution between CA and KA peaks (R S = 0.93) was detected in M. laevigata after M. laevigata after M. laevigata 9 months of storage in comparison to the experiment start (R S = 2.62).This fi nding cannot be related to the loss of column effi ciency, since the chromatograms obtained for non-stored plants showed similar resolution to T 0 samples (data not shown).
UV data recorded on line for BA peak showed co-eluting compounds in samples of Mikania glomerata stored for 3 months (Fig. 4, BA), similarly to the peak of KA in the same species (data not shown).Besides, a marked decrease in resolution was observed for both compounds.Altogether the results suggest the occurrence of chemical transformations in these derivatives.Considering that changes in the peaks of kaurane diterpenes were observed after 3 months of M. glomerata storage, the stability study of this species was discontinued after 12 months.Spectral purity evaluation of selected peaks was shown to be a robust tool to access chemical stability of both guaco species, complementary to quantitative analysis.Excepting CA and KA peaks, whose chromatograms clearly indicated loss of resolution during the storage periods (Fig. 2 and 3), chemical changes in the other compounds were only detected by peak purity analysis, since the contents of the constituents showed no signifi cant variation for both species (p variation for both species (p variation for both species ( < 0.05).
In summary, our results demonstrate that the chemical integrity of the kaurane diterpenes (BA, CA and KA), constituents of Mikania glomerata and Mikania laevigata, is affected after three months of storage, whereas modifi cations in the coumarin peak, found only in M. laevigata, are detected after six months of storage.The Brazilian Pharmacopoeia establishes a minimum content of 0.1% CO for M. laevigata (F.Bras.IV 2005), while the concentrations of kaurane diterpenes in Mikania species are not regulated by any offi cial guide.The CO contents determined for stored samples of M. laevigata ranged from 0.10% to 0.12% and therefore fulfi ll the pharmacopoeical requirement, except for the 12-month sample (Table II).Hence, based strictly on the offi cial guidelines, the analyzed samples would be considered adequate for human consumption.However, peak purity analyses of the constituents in stored samples indicate chemical changes, that may affect the biological effects of the vegetal drugs.Therefore, pharmacological analyses are required to investigate if the observed changes in chemical markers will affect the biological effi cacy and safety of guaco.
ANALYSES OF CHEMICAL MARKERS IN STORED LEAVES Leaves of Mikania laevigata and Mikania laevigata and Mikania laevigata Mikania glomerata were collected in June 2007 from the glomerata were collected in June 2007 from the glomerata apical, intermediate and basal regions of cultivated 45-month old plants (n = 6 plants for each species).

aFigure 4 -
Figure 4 -UV spectra recorded on line by DAD in the ascending and descending regions of the peaks from coumarin (CO) and benzoylgrandifl oric acid (BA) found in the chromatographic profi les of Figures 3 and 2, respectively.Storage periods: T 0 , study start; T 3 , 3 months; T 6 , 6 months; T 12 , 12 months; T 18 , 18 months.
aThe parameters were determined in the wavelength established for each chemical marker, according to the program describe in TableI.Rs: resolution; T: tailing factor; N: plate number; k: retention factor; t R : retention time; RSD: relative standard deviation.bSamples of Mikania laevigata and Mikania glomerata spiked with 0.1 mg/mL o-coumaric acid were employed in the experiments.See text for details.CO, coumarin; OC, trans-o-coumaric acid; BA, ent-benzoylgrandifl oric acid; CA, ent-cinnamoylgrandifl oric acid; KA, ent-kaurenoic acid.