Jorge F. S. Ferreira

Floral Morphology of Artemisia annua with Special Reference to Trichomes

Floral morphology of Artemisia annua L. was described using light and scanning electron microscopy. Floral trichomes include nonglandular T-shaped filamentous trichomes, which occur on the basal bracts and pedicel of the capitulum, and 10-celled biseriate glandular trichomes on receptacle and bracts of capitula and corolla of florets. The heart-shaped biseriate glandular trichomes, composed of two columns of five cells each, are similar to those described for the leaf. During floral development, the cuticle surrounding the apical cells of biseriate glands detaches and forms a saclike reservoir for fluids excreted by the apical cells. The cuticle ruptures at anthesis, discharging its contents on the inflorescence. Artemisinin was detected by HPLC-EC from flowering branches dipped for 60 s in petroleum ether or acetonitrile, providing corroborating evidence that biseriate glandular trichomes were the likely site of artemisinin sequestration.

Roots as an enhancing factor for the production of artemisinin in shoot cultures of Artemisia annua

Artemisinin was produced in differentiated shoot cultures of Artemisia annua L. but was undetected in callus or cell cultures. The growth regulators benzyladenine, kinetin, chlormequat, and daminozide, at concentrations which severely reduced rooting, reduced artemisinin production. A highly significant correlation (1% level) was observed between shoot artemisinin content and number of roots (r=0.775**), but shoot number and artemisinin content were unrelated (r=-0.198). Benzyladenine increased shoot proliferation at 0.5 and 5.0 μM, but decreased root production at 0.5, 5.0, and 50 μM. The highest levels of artemisinin production (0.287% DW) were obtained in hormone-free medium when root production was maximized. Removal of roots from shoots cultured in hormone-free liquid medium reduced shoot artemisinin by 53% and shoot arteannuin B by 60%. Neither artemisinin, arteannuin B, or artemisinic acid were detected from roots developed in semi-solid or liquid medium.

Immunoquantitative analysis of artemisinin from Artemisia annua using polyclonal antibodies

Artemisinin was derivatized to dihydroartemisinin carboxymethylether in three steps, without disturbing the peroxide bridge, and then linked to either thyroglobulin (TGB) or bovine serum albumin (BSA). The artemisinin-TGB and -BSA conjugates were injected in female New Zealand rabbits but only the artemisinin-TGB conjugate generated polyclonal antibodies. An enzyme-linked immunosorbent assay (ELISA) was developed and the specificity of the antibodies was confirmed by comparison with pre-immune serum and by competitive assays using different dilutions of artemisinin standards. Although anti-artemisinin antibodies cross-reacted with artemisitene and dihydroartemisinin at all dilutions used, cross-reaction with deoxyartemisinin, artemisinic acid, and arteannuin B occurred only at high concentrations. ELISA successfully detected artemisinin from crude extracts in concentrations as low as 1.5 ng ml-1; and was epsilon 400-fold more sensitive than the HPLC-EC. The ELISA successfully detected and quantified artemisinin in different organs of greenhouse-grown plants and in eight clones of Artemisia annua grown in tissue culture but artemisinin was overestimated owing to cross-reactivity of the antibodies with artemisinin-related compounds present in the samples. Despite overestimation of artemisinin content, the correlations between ELISA and HPLC-EC were r = 0.92 when samples were diluted 100 times, and r = 0.90 when samples were diluted 500 times, indicating that ELISA is a potential tool for screening large A. annua populations.

Histochemical and Immunocytochemical Localization of Tropane Alkaloids in Erythroxylum coca var. coca and E. novogranatense var. novogranatense

The tropane alkaloids of Erythroxylum coca var. coca and Erythroxylum novogranatense var. novogranatense were localized using histochemical, cytochemical, and immunocytochemical probes. At the tissue level, these alkaloids were localized in leaves, stems, and fruits with Dragendorffs reagent. Alkaloids were found in the mesophyll, including palisade, spongy, and vascular parenchyma cell layers, and in some cells of the collenchyma. No alkaloids were detected in the epidermis of either Erythroxylum species. Alkaloids were also detected in the endosperm and embryos of orange and mature red fruits but not in tissues of immature green fruits. Quantitative HPLC reveals that embryos had 2.5-5 times more cocaine (w/w%) than endosperms, but 40 times less than leaves. In contrast to leaves, cocaine was a minor alkaloid present in fruits. Dragendorffs reagent was used to develop thin-layer chromatography plates and specifically stained the alkaloids extracted from leaves of both species. Complexing of alkaloids with tannins resulted in aggregates in the vacuole of E. coca leaf cells as visualized by transmission electron microscopy. Immunocytochemical localization, using anticocaine polyclonal antibodies on leaf tissues, proved that these compounds are localized in vacuoles of both photosynthetic and vascular parenchyma, as demonstrated by the use of Dragendorffs reagent. The alkaloids were associated with intravacuolar globules and appear to be aggregates with a core formed by phenolic compounds and a periphery enriched in alkaloids. The vacuolar localization of the cocaine alkaloids indicate that they may be complexed with phenols in vivo, thereby rendering them relatively immobile. The implications of the cellular and tissue localization of tropane alkaloids in Erythroxylum are discussed.