Evaluation of morpho-anatomical and chemical differences between varieties of the medicinal plant Casearia sylvestris

Casearia sylvestris Swartz (Salicaceae) has been used in traditional medicine and its leaf extracts have been exhibited important pharmacological activities. The species presents morphological, chemical and genetic variation. Two varieties are considered due external morphological differences: C. sylvestris var. sylvestris and var. lingua. There are difficulties in definition of these varieties. The objective of this work is to evaluate chemical and morpho-anatomical differences between C. sylvestris varieties that can be applied in their distinction for pharmaceutical or botanical purposes. Transverse and paradermic sections of leaves were prepared for morphoanatomical, histochemical and quantitative microscopy (stomatal and palisade index) analyses. Diterpene profiles of the specimens were obtained by HPLC-DAD and TLC. Morpho-anatomical analyses demonstrated significant differences between the varieties only in paradermic sections: var. sylvestris polygonal epidermic cell walls and hypostomatic; var. lingua rounded epidermic cell walls and amphistomatic. No differences were observed for stomatal index; palisade index was found 2.8 for var. lingua and 3.9 for var. sylvestris. Chromatographic analyses confirmed previous results demonstrating that diterpene profile in varieties differs, with predominance of these metabolites in var. sylvestris. In conclusion, this work indicates that chromatographic analysis besides morphoanatomical analysis can be applied in distinction of C. sylvestris varieties.


INTRODUCTION
Casearia sylvestris Swartz (Salicaceae), known as "guaçatonga" or "erva-de-bugre", is a plant species distributed throughout South America that occurs in 22 states of Brazil, especially within the Atlantic and Amazon forests and the Cerrado biomes (Marquete 2001, Sleumer 1980).The plant has been widely used in traditional medicine for the treatment of snake bites and in wound healing, and also as an antiulcer and topical antiseptic (Hoehne 1939, Lorenzi andMatos 2002).Pharmacological studies on its leaf extracts have demonstrated antiulcerogenic, anti-inflammatory, antivenom, and cytotoxic (tumor cell lines) activities.Moreover, no significant toxicological effects of its ethanolic or hydroalcoholic extracts have been observed in vitro or following oral administration in animals (Basile et al. 1990, Borges et al. 2000, Ferreira et al. 2010, Maistro et al. 2004, Santos et al. 2010, Sertiè et al. 2000, Silva et al. 2004).The Brazilian Agency of Medicines (Agência Nacional de Vigilância Sanitária -ANVISA) recognizes the importance of the species that was included in a positive list of medicinal plants for sale in pharmacies (Brasil 2010).
Casearia sylvestris was first described by Swartz in Flora Indiae Occidentalis (1797) and Eichler and Martius (1871) included it in Flora Brasiliensis of Carl Friedrich Philipp von Martius, classified in Bixaceae family.Later it was classified in Flacourtiaceae family (Absy andScavone 1973, Sleumer 1980).The Angiosperm Phylogeny Group (APG) updated the classification of Angiosperms in 2003 and some genera of Flacourtiaceae (e.g.Casearia Jacq.) were classified as Salicaceae.C. sylvestris Swartz (Crateria Bentahm section) is classified in Samydeae tribe, Salicaceae family and Malpighiales order (Chase et al. 2002, The Angiosperm Phylogeny Group 2003).The macromorphology of the species is described in the work of Torres and yamamoto (1986), based on descriptions of Eichler and Martius (1871) and Sleumer (1980).The morpho-anatomical characters of the leaves were described by Absy and Scavone (1973) and Alquini and Takemori (2000).
C. sylvestris occurs in different biomes showing great morphological variability: the size, shape, texture and consistency of the leaves, the pubescence of the branches and inflorescences, the number of flowers per inflorescence, the length of peduncle, and its height -shrubs to trees (Marquete 2001, Torres andyamamoto 1986).Sleumer (1980) considered two varieties to the species due external morphological differences, connected by intermediate forms that are difficult to distinguish: C. sylvestris var.sylvestris and C. sylvestris var.lingua (Cambess.).Silva et al. (2006) reported differences observed during collection between varieties.In Cerrado biome the specimens were shrubs, with coriaceous lighter green leaves with smaller width and length, corresponding to C. sylvestris var.lingua; unlike, specimens of Atlantic Forest were trees with different heights (always higher than 2 m), green dark leaves larger in width and length, corresponding to C. sylvestris var.sylvestris.The differences between varieties were summarized by Klein and Sleumer (1984).
Besides the macromorphological differences between the varieties, metabolism variability analyses have suggested a minor expression of diterpenes DIFFERENCES BETWEEN VARIETIES OF C. sylvestris Sw. and a greater expression of flavonoids (as rutin) in C. sylvestris var.lingua.Studies of genetic variability also indicated distinct genetic profiles for the varieties (Cavallari 2008, Silva et al. 2006).
Considering the importance of C. sylvestris as medicinal plant, the chemical variability observed in previous works, mainly for the bioactive diterpenes and rutin, and the difficulties to distinguish the varieties of C. sylvestris by macromorphological analyses in the present article, we report on the evaluation of chemical and morpho-anatomical differences between C. sylvestris varieties that can be applied in their distinction for botanical purposes or pharmaceutical applications.(Cavallari 2008).

MORPHO-ANATOMICAL ANALySES
Anatomical freehand sections (paradermic and transverse) of the middle third of fresh leaves from C. sylvestris were obtained with a blade.The sections were clarified with sodium hypochlorite commercial solution and wash with deionized water before stain with different reagents (Delafield hematoxylin, Astra blue, iodine green and toluidine blue) for selection of better cell wall staining and visualization of tissue structures.The anatomical descriptions were realized based on microscope visualization (40 x) of the sections on covered glass sheet and comparison with literature data (Absy andScavone 1973, Alquini andTakemori 2000).JOSIANE C. CLAUDINO et al.

QUANTITATIVE MICROSCOPy
Palisade index: the average number of palisade cells beneath each epidermal cell is termed the palisade index.The paradermic sections obtained and clarified as described in Morpho-Anatomical Analyses section were mounted on covered glass sheet.Five groups of four epidermal cells were selected in each section.The palisade cells lying beneath each group were counted, being included in the count wich are more than half-covered by the epidermal cells (Evans 2002, Costa 2001).
Stomatal index: the percentage proportion of the ultimate divisions of the epidermis of a leaf which has been converted into stomata is termed the stomatal index (I): Where S is the number of stomata per unit area and E is the number of ordinary epidermal cells in the same unit area.The paradermic sections (upper and lower surfaces), that were obtained and clarified as described in Morpho-Anatomical Analyses section, were stained with Astra blue and mounted on covered glass sheet (Evans 2002, Costa 2001).

CHROMATOGRAPHIC ANALySES
The following methods were developed and validated to evaluate the diterpene chromatographic profile of C. sylvestris leaf samples in the laboratories of Núcleo de Bioensaios, Biossíntese e Ecofisiologia de Produtos Naturais, Instituto de Química, UNESP, Araraquara-SP (data not published).
Extraction: Dried (40° C, 7 days) and powdered leaves (1.0 g) from the specimens of C. sylvestris were extracted twice in a test tube with 10 mL of ethyl acetate: hexane: isopropanol 91:08:01 (v/v) by sonication for 1 h.The liquid extracts were filtered, dried at room temperature and at desiccator with silica gel under reduced pressure.
Casearin-like diterpenes present two different patterns of conjugated double bond in their lateral chain (C11-C16) -C12(Z or E)/C14 or C13(16)/C14 -and their UV spectra present one band and λ max = 231-238 or 221-228 nm, respectively (Carvalho et al. 2009, Espíndola et al. 2004, Itokawa et al. 1990).Thus, we suggested that peaks with UV spectra with one band and λ max = 221-228 or 231-238 nm were relative to the casearin-like diterpenes.DIFFERENCES BETWEEN VARIETIES OF C. sylvestris Sw.Carvalho et al. (2009) analyzed different organs of C. sylvestris with similar approach and results from HPLC-DAD analyses were confirmed by 1 H NMR and TLC data.Moreover, five casearin-like diterpenes isolated from leaves of C. sylvestris were identified in its extract and presented peaks with t R in the range of 10-30 min in the same HPLC conditions (Santos 2008).Additionally, as observed by Carvalho et al. (2009), peaks with a second band in the UV spectra with λ max of c.a. 280 nm, indicate the presence of clerodane diterpenes with a aromatic ester substituent, as reported in the literature (Beutler et al. 2000).Thus, clerodane diterpenes were identified in the extract chromatograms on the basis of UV spectra of the peaks.Additionally, the comparison of retention time (t R ) and UV spectra of peaks from standard and extract chromatograms was realized.
Thin Layer Chromatography: 8.0 mg of dried extracts and 1.0 mg of caseargrewiin F and casearin B were dissolved in 1.0 mL of ethyl acetate.TLC was developed in silica gel plate aluminum backed (20 x 20 cm; 200 μm) using hexane: ethyl acetate: isopropanol 70:28:02 (v/v) as eluent and sulfuric anisaldehyde as spray reagent.

RESULTS
The macromorphological differences between the varieties of C. sylvestris reported were evident during the collection and the specimens were classified as C. sylvestris var.lingua or C. sylvestris var.sylvestris as described by Sleumer (1980) and Klein and Sleumer (1984).
In the morpho-anatomical study, the best visualization in transverse sections was achieved with dual staining: Delafield hematoxylin followed by green iodine.No significant anatomical differences were observed in transverse sections for the varieties.Morpho-anatomical characters of the leaves were in agreement with literature (Absy andScavone 1973, Alquini andTakemori 2000).In the midvein region of leaf sections we observed typical cells of xylem, phloem, collenchyma and parenchyma, in addition to fibers associated with vascular bundles (Figure 1a).Both upper and lower epidermises were uniseriate, presenting a lower number of simple nongladular trichomes.The mesophyll is asymmetrical and heterogeneous with two (rarely three) layers of palisade parenchyma and several layers of spongy parenchyma with irregular cells in shape and size.The air spaces (meatus) occur in small proportion in spongy parenchyma (Figure 1b).We also noted the presence of spiral vessels, oil glands, druses and prismatic crystals (calcium oxalate) (Figure 2).
For the paradermic sections, the best stain reagent was Astra blue.In this case we observed significant differences between the varieties.The epidermic cell walls are polygonal in C. sylvestris  Histochemical reagents did not differentiate the varieties and the results were positive for both terpenes and phenolics (all reagents tested).
The results of quantitative microscopy are in Tables I and II.The values of stomatal index for specimens of C. sylvestris var.sylvestris presented great variation: 11.9 to 20.7 % (standard deviation = 3.4).In the case of C. sylvestris var.lingua specimens these values were more homogeneous (standard deviation = 0.2).The mean value of stomatal index for both varieties was similar: 15.9 (var.sylvestris) and 16.4 % (var.lingua).On the other hand, palisade index was found 2.8 for var.lingua and 3.9 for var.sylvestris, demonstrating that this index should be useful in differentiation of C. sylvestris varieties.
TLC analyses provided information about clerodane diterpenes profile of the extracts.Caseargrewiin F (R f = 0.23) was identified by TLC (Figure 4) in the extracts of specimens AGS101, AGS102, AGS103, AGS104 (var.sylvestris), AGS44, AGS45 and AGS51 (var.lingua), while casearin B (R f = 0.26) was not identified in any extract.Moreover, TLC profiles of var.sylvestris specimens are similar, except for AGSDER.Unlike, specimens of var.lingua presented different TLC profiles, except for AGS44      UV spectra/λ max .Figure 5 shows representative chromatograms of both varieties and the structures of diterpene standards.As observed in the example of Figure 5, there is a greater total area of peaks with t R < 5 min in the chromatograms of specimens of C. sylvestris var.lingua and AGSDER (var.sylvestris) than in chromatograms of AGS101-104 (var.sylvestris).On the other hand, the total area of peaks with λ max = 221-228 or 231-238 nm is greater in chromatograms of AGS101-104 (Table III).
We classified the specimens in three groups on the basis of the observed common main peaks (peaks with larger area), identification of clerodane diterpenes (caseargrewiin F, casearin B and X) and predominance of peaks with UV spectrum with λ max = 221-228 or 231-238 nm.The diterpene standards were identified in the extracts by comparison of their t R and UV spectra.Tables IV-VI present data from chromatograms of the three groups: group 1 -AGS38, AGS44-45 (var.lingua); group 2 -AGS43 and AGS51 (var.lingua); group 3 -AGS101-104 (var.sylvestris).AGSDER demonstrated a chromatographic profile with characteristics of different groups.

DISCUSSION
The morpho-anatomical comparison between the varieties revealed differences only in the paradermic sections.The shape of epidermic cell walls (polygonal in var.sylvestris and rounded in var.lingua) and stomata distribution (var.sylvestris is hypostomatic and var.lingua is amphistomatic) are characteristics that can be applied in the differentiation of the varieties as well as the palisade index.According Costa (2001) the palisade index is less affected by environmental factors than stomatal index.These results are interesting for the pharmaceutical quality control application because the analysis of the powdered drug will lead to the identification of the varieties.Obviously, the assays will be applied to more representative number of specimens from different environments or regions to be validated.Data from HPLC-DAD analyses (t R and UV spectrum) allowed to classify the specimens in three groups.In the first group -AGS38, AGS44 and AGS45 (var.lingua) -we found the following common characteristics: identification only of caseargrewiin F, predominance of peaks with λ max = 221-228 nm (22-23 peaks and total peak area of 22-55 x 10 6 ), and seven common main peaks with area > 10 6 (Tables III and IV).In TLC analyses caseargrewiin F was not identified in AGS38 probably due its minor concentration as demonstrated by HPLC analysis (small peak area).
In the second group -AGS43 and AGS51 (var.lingua) -caseargrewiin F, casearins B and X were identified.In the TLC analyses, diterpenes were not identified as discussed above.The peaks with λ max = 231-238 nm have greater area (16-17 peaks and total peak area of 21-53 x 10 6 ) than peaks with λ max = 221-228 nm (6-8 peaks and total peak area of 0.8-2.4x 10 6 ) (Table III).Finally, there are five common main peaks with area > 106 (Table V) in their chromatograms.
The third group includes AGS101-104 (var.sylvestris).In this group the peaks with λ max = 231-238 nm predominate notably in number of peaks (24-33) and total peak area (142-242 x 106) in comparison with all specimens of var.lingua (Table III).The three diterpenes were identified in these specimens and they have six common main peaks with area > 4 x 10 6 (Table VI).In the TLC analysis, casearin B was not identified as discussed above.
The specimen AGSDER presented characteristics of different groups.Its chromatogram demonstrated that peaks with λ max = 221-228 nm predominate in number (13) but no differences were observed for peak total area (8 x 10 6 ) considering the two ranges of λ max .Casearin B and caseargrewiin F were identified in HPLC-DAD analysis.
The results obtained suggest the predominance of clerodane diterpenes, especially with λ max = 231-238 nm, in C. sylvestris var.sylvestris.Moreover, there is a greater total area of peaks with t R < 5 min in the chromatograms of specimens of C. sylvestris var.lingua and AGSDER (var.sylvestris) than in chromatograms of AGS101-104 (var.sylvestris) indicating greater concentration of polar compounds (e.g.phenolics) as demonstrated in previous work (Cavallari 2008, Silva et al. 2006).AGSDER was in an urban area subjected to pollutants and it may be correlated with its different chemical pattern.
As morpho-anatomical analyses, the chemical analyses by HPLC-DAD have to be applied to greater number of specimens from different environments or regions to give more conclusive results.
In conclusion, this work indicates that chemical analyses, mainly HPLC-DAD, besides morpho-anatomical analyses of paradermic sections and palisade index can be applied in distinction of C. sylvestris varieties for botanical or pharmaceutical purposes.

Figure 1 -
Figure 1 -Tranverse section of the midvein region from the leaf of C. sylvestris var.lingua (a) and tranverse section of the region between the midvein and the margin from the leaf of C. sylvestris var.sylvestris (b).

Figure 2 -
Figure 2 -Transverse section of the region between the midvein and the margin from the leaf of C. sylvestris var.lingua.
DIFFERENCES BETWEEN VARIETIES OF C. sylvestris Sw.

TABLE II Palisade index for specimens of C. sylvestris.
1 standard deviation.1 standard deviation.

TABLE IV Data from HPLC-DAD analyses of specimens AGS38, AGS44 and AGS45. Retention time, area and λ max of selected peaks (t R > 10 min; area > 10 6 in at least one specimen of the group) are presented.
1 all spectra presented one band. 2peak identified as caseargrewiin F (t R = 16.9 min; λ max = 235 nm). 3 values in italic indicates that peak area< 10 6 for this specimen.

TABLE VI Data from HPLC analyses of specimens AGS101-104. Retention time, area and λ max of selected peaks (t R > 10 min; area >10 7 in at least one specimen of the group) are presented.
peaks identified as caseargrewiin F (t R = 16.9 min; λ max = 235 nm), casearin B (t R = 19.0min;λ max = 233 nm) or casearin X (t R = 21.4 min; λ max = 234 nm).3 peak area values in italic highlights that peak area< 10 7 for this specimen. 2