Abstract
Duroia longiflora (Rubiaceae) is an Amazonian plant species with no previous chemical study. The aim of this work was to study the chemical composition of D. longiflora leaves and study their histochemistry. In vivo tissues of the leaves were histolocalized. Extracts were made from dried leaves with hexane, methanol, and water and the methanolic extract was fractionated. Fractionation of the dichloromethane phase of the methanolic leaf extract allowed the isolation of a biflavanoid. Via nuclear magnetic resonance (mono and bidimensional), mass spectrometer analysis, and comparison with existing literature, its chemical structure was identified as 2α,3α-epoxy-5,7,3′,4′-tetrahydroxyflavan-(4β→8)-epicatechin. This is the first report describing the histochemical and phytochemical studies of D. longiflora leaves. As far as we know, this is the first report of this substance in any Rubiaceae species.
The Amazon rainforest is considered to be one of the greatest sources of biodiversity on the planet. 1 One of the largest flowering plant families in the tropics is Rubiaceae, represented in Brazil by 126 genera and 1397 species 2 and it is characterized by the production of a wide variety of secondary metabolites including terpenes, flavonoids, alkaloids, and coumarins. 3
Duroia longiflora Ducke (Rubiaceae) belongs to the subfamily Ixoroideae, tribe Gardenieae, and are woody plants, with whole stipules, corollas with contour lobes, and rarely 1 fleshy fruit. Individuals can reach 17 m height and 15 cm in diameter 4 ; the inner bark (rhytidome) is dark brown, striated, and fissured; and the outer bark is brown with cream striations. 5 Literature search revealed only few chemical studies with Duroia species: 2 were conducted with D. hirsuta describing flavonoids and iridoid isolation 6,7 and some with D. macrophylla reported the isolation of indole alkaloids with antitumor and antituberculosis activity, along with 2 triterpenes with antituberculosis activity as well. 8 -14 So, we decided to investigate another species of the genus Duroia, D. longiflora.
Histochemistry coupled with phytochemistry can provide an excellent chemical profile. Histochemical studies detected the presence of flavonoids and flavones when stained with ferric chloride in the leaf tissue, and when stained with hydrochloric vanillin, a specific reagent for tannins, it revealed an abundance of tannins along the tissue (Figure 1a, b).

Our observations showed tannins are restricted to epidermis and wall of collenchyma, phloem, and trichomes. This information guided the phytochemical fractionation on the search of flavonoids and tannins, because of the abundance seen in the histochemical test. Phytochemical fractionation led to the isolation of a biflavanoid, identified as 2α,3α-epoxy-5,7,3′,4′-tetrahydroxyflavan-(4β→8)-epicatechin (Figure 2). This is the first report of this substance in any Rubiaceae species, as far as we know. Only 2 studies reporting the presence of this substance were found in literature, both in Sapindaceae species: in Litchi chinensis seeds 15 and in Xanthoceras sorbifolium husks. 16

Main heteronuclear multiple bond correlations (a) and correlational spectroscopy correlations (
Duroia species suffers from herbivory and the species D. hirsuta and D. saccifera developed myrmecophily association, which creates protection against herbivory. 17 Duroia macrophylla and D. longiflora lack ant symbiosis, which forced them to find another way to protect themselves from herbivory and fungal pathogens. 18,19 Studies reveal alkaloid production in D. macrophylla leaves. 10,12,13 Our study shows D. longiflora presented great concentration of tannins in their leaves. We hypothesize that the production of these compounds may be related to herbivory protection, because it is known that the presence of trichomes and flavonoids production are related to insect pest protection. 18 Also, a study reveals that production of epicatechin compounds in plants can be related to protection from fungi infections, 19 which could explain the production of this specific compound in D. longiflora. This work can be used as a basis for future chemotaxonomic and ecological studies.
Experimental
Plant Material
Duroia longiflora leaves were collected from 3 places: (1) in Reserva Florestal A. Ducke, located at Km 26 on the Manaus-Itacoatiara Road (AM-010); (2) in ZF-2 Forest Management Base, Tropical Forestry Station, located at Km 50 on the BR-174; and (3) in Km 23 of ZF-2 branch, Amazonas-Brazil. A voucher specimen was deposited in the Herbarium of the Amazonas Federal Institute of Education, Science and Technology, no. 10 893.
Histochemical Study
Five healthy leaves from the third or fourth node of young shoots were collected from 3 individuals of D. longiflora. These were fixed in FAA70 for 48 hours and stored in 70% alcohol before being submitted to histochemical tests. The material was cut with a rotary table microtome blade adapted to a free hand cut microtome and placed in a watch glass where it was stained with ferric chloride 20 to identify phenolic substances and hydrochloric vanillin 21 to identify tannins, where brown and red colors indicated the presence of substances. Sections were placed on a glass slide and analyzed with a microscope (Zeiss Primo Star-Microimaging 37081) at 100 × 10 magnifications. The materials revealed in leaf petiolar and laminar tissues were then photographed under a microscope coupled to a digital camera (Canon PC1252).
Extracts Preparation
Leaves were dried in a forced circulation oven (30°C-40°C) and grounded in a knife mill (Tecnal, model Willye TE-650). The plant material was extracted with hexane in an ultrasonic bath (Unique) for 20 minutes, then filtered, and extracted 2 more times. Then the plant material was oven-dried and extracted with methanol and finally extracted with distilled water, by the same procedure. Hexane and methanolic extracts were concentrated in a rotary evaporator (Fisatom, model 802), whereas the aqueous extracts were lyophilized (CHRIST, model Beta 1-8 LD plus).
Chromatographic Fractionation
Methanol extract (10 g) was diluted in 1:1 methanol/H2O and partitioned with dichloromethane (DCM); this was repeated 3 times. The hydroalcoholic solution was then extracted with ethyl acetate, also 3 times. This procedure yielded 1.25 g of DCM phase, 7.9 g of ethyl acetate phase, and 1.97 g of hydroalcoholic phase. Based on prior thin layer chromatography (TLC) analysis, the DCM phase was selected for fractionation, because it showed to be rich in tannins (flavonoids) and showed the best separation condition in ethyl acetate/methanol 9:1 as eluent. The DCM phase (1 g) was fractionated in a Sephadex LH-20 column using isocratic methanol, obtaining 64 fractions which were concentrated in a rotary evaporator, and analyzed by comparative TLC with chemical reagents (sulfuric anisaldehyde and ferric chloride) and also UV lights (254 and 365 nm). After reunion, subfractions 5 to 12 (0.2844 mg) were fractionated using silica gel (60 mesh) as stationary phase and increasing gradient of the solvents, in binary combinations: DCM/ethyl acetate and ethyl acetate/methanol, resulting in 33 fractions which were mixed after TLC analysis. Subfractions 31 and 32 (18 mg) showed absorption at 254 nm when analyzed by TLC, a purple spot when stained with ceric sulfate, and a blue spot when stained with ferric chloride. These informations suggested the presence of a substance with aromatic ring together with an oxidable site on it.
Chemical identification was performed by 300 MHz nuclear magnetic resonance (NMR) analysis (Bruker Biospin AG, Forrier Model 300 UltraShield), where 1H, 13C, heteronuclear single quantum coherence, heteronuclear multiple bond correlation, correlational spectroscopy, and J-res NMR experiments were obtained. The sample was also analyzed by high-resolution mass spectrometry (electrospray ionization, using micrOTOF instrument, Bruker model), generating a quasi-molecular ion [M+1] at 577.1399 m/z, which allowed determination of the molecular formula as C30H24O12.
Footnotes
Acknowledgments
The authors would like to thank Adrian Barnett who helped with the English.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors would like to thank the Brazilian Research Agencies: CNPq (CT-Amazônia/CNPq - 405804/2013-0) and CAPES (Pro-Amazônia/CAPES - 23038.000738/2013-78) for financial support. To FAPEAM for granting the scholarship to FSC.
