Isolation and identification of a new biflavonoid from the Amazonian tree Ceiba pentandra (L.) Gaertn. (sumaúma)

1 Introduction

Brazil has one of the largest tropical forests in the world. It is home to rich biodiversity and has great potential for discovering bioactive compounds. In this context, plant species show significant chemical and pharmacological potential (Wink, 2008; Böttger et al., 2018). However, there are few phytochemical studies on wood species. This highlights the need to increase chemical research on Amazonian timber species.

Ceiba pentandra (syn. Bombax pentandrum L.) belongs to the family Malvaceae. It is a pantropical species and one of the largest trees in Brazil. In the Amazon region, it is often called sumaúma or sapopema. This tree holds cultural importance, being regarded as a sacred tree and used by riverine and Indigenous communities as shelter and for acoustic communication. Phytochemical studies show a wide variety of specialized compounds in different parts of the plant. These include fatty acids (de Souza et al., 2026; Montcho et al., 2018; Ueda et al., 2002), steroids, triterpenes (Onuh et al., 2024; Abouelela et al., 2018; Ngounou et al., 2000; Nelly et al., 2025), sesquiterpenes (Rao et al., 1993; Kishore et al., 2003), xanthones (Taiwo et al., 2024), flavonoids (Noreen et al., 1998; Onuh et al., 2024; Aderogba et al., 2013; Ngounou et al., 2000; Ueda et al., 2002), flavanolignans (Abouelela et al., 2020), and phenolic acids (Onuh et al., 2024; Aderogba et al., 2013; Nelly et al., 2025), pointing to their significant bioactive potential. Flavonoids are the most frequently mentioned class of these compounds and are often linked to plant defense (Treutter, 2005).

Despite the importance of these findings and the diversity of phenolic compounds reported in different plant parts, there have been few studies on the phytochemistry of wood from native species of C. pentandra. This study aimed to explore the flavonoid profile of a native species of C. pentandra.

2 Materials and methods

2.1 General methods

Nuclear Magnetic Resonance (NMR) measurements were performed using Bruker Fourier 300 MHz and Bruker Ascend Evo 600 MHz spectrometers. Chemical shifts (δ) are reported in parts per million (ppm), while coupling constants (J) are expressed in Hertz (Hz). Spectrometric data were obtained using a Bruker Daltonics MicroTOF-Q-II high-resolution mass spectrometer (HRESI-MS) equipped with an electrospray ionization (ESI) source. Column chromatography (CC) utilized silica gel 60 (Merck, 70–230 and 230–400 mesh, Sigma-Aldrich, Gallen, Switzerland) and Sephadex LH-20 (Sigma-Aldrich, Schnelldorf, Germany). Thin-layer chromatography (TLC) was performed using silica gel 60 F254 pre-coated aluminum sheets (0.25 mm; Merck, Darmstadt, Germany). Visualization was achieved under UV light at 254 and 365 nm, followed by spraying with vanillin–sulfuric acid reagent. Deuterated solvents were obtained from Merck (Darmstadt, Germany).

2.2 Acquisition of wood, extraction, isolation, and purification of compounds

Wood samples were collected from the trunk of a native specimen of C. pentandra on the campus of the National Institute for Amazonian Research (INPA) in Manaus, Amazonas, Brazil (3°05′40.6″S; 59°59′15.3″W) and identified by comparison with samples available in the INPA xylotheque. This tree had fallen naturally due to termite infestation in its taproots.

The wood material (1.400 g) was pulverized and extracted by maceration in hexane for 7 days, followed by methanol for an additional 7 days, both at room temperature. After the maceration, the resulting solutions were filtered and concentrated under reduced pressure using a rotary evaporator, resulting in hexane and methanol extracts. All experimental procedures were conducted in accordance with the protocol established by the research group “Plantas da Amazônia: Química, Quimiossistemática e Atividade Biológica”.

The methanol extract (12 g) was fractionated on a silica gel column (70–230 mesh; height (h) × diameter (⌀) = 35.0 × 4.5 cm) and eluted with hexane, hexane:ethyl acetate (EtOAc), and EtOAc:methanol (MeOH) mixtures of increasing polarity, yielding 41 fractions. Fractions 26, 29, and 36 were further subjected to chromatographic separation. Fraction 26 (132.4 mg) was further fractionated on a silica gel column (230–400 mesh; h × ⌀ = 30 × 1.5 cm) and eluted with a mixture of dichloromethane (CH₂Cl₂) and methanol (MeOH) (95:5), yielding 17 subfractions. Subfraction 26.3 (20 mg) was processed through C18 solid-phase extraction (Strata, 55 µm), eluted with a mixture of H₂O:MeOH (80:20 and 1:1, v/v). The H₂O:MeOH (1:1) eluate was concentrated, resulting in the isolation of compound 1 (5.6 mg). Subfractions 26.7 and 26.12 afforded compounds 2 (18.1 mg) and 3 (9.9 mg) as solids, respectively. Fraction 29 (116.5 mg) was fractionated on Sephadex LH-20 (h × ⌀ = 46 × 2.8 cm), eluted with MeOH, and produced 21 subfractions. Subfraction 29.8 (21.9 mg) was subsequently fractionated over a silica gel column (230–400 mesh; h × ⌀ = 18.4 × 0.7 cm) and eluted with a mixture of CH₂Cl₂ and EtOAc (8:2), which resulted in 27 subfractions, from this, subfractions 29.8.9, 29.8.17 and 29.8.19 yielded compounds 4 (1.6 mg), 5 (2.1 mg) and 6 (3.6 mg). Fraction 36 (781.1 mg) was also fractionated on Sephadex LH-20 (h × ⌀ = 46 × 2.8 cm), eluted with MeOH, yielding 15 fractions. Subfraction 36.7 (68 mg) was separated on another Sephadex LH-20 (h × ⌀ = 24.5 × 1.1 cm) and eluted with methanol. Following this, subfraction 36.7.6 (36 mg) was processed over silica gel in a column (230–400 mesh; MeOH = 37 × 0.7 cm) and eluted with a mixture of CH₂Cl₂ and acetone, resulting in the purification of compound 7 (2.3 mg).

3 Results and discussion

Phytochemical analysis of the methanol extract resulted in the isolation and structural elucidation of new biflavonoid (7), in addition to known compounds (Figure 1), whose NMR data were compared with the literature as liquiritigenin (1) (Kavtaradze et al., 2011), garbanzol (2) (Hashida et al., 2014), 5-deoxykaempferol (3) (Yu et al., 2016), 3-methoxy-7,4′-dihydroxyflavone (4) (Kim et al., 2006), 7,4′-dihydroxyflavone (5) (Park et al., 2007) and guibourtacacidine (6) (Carvalho et al., 2008). The NMR assignments were confirmed using heteronuclear single quantum coherence (HSQC) and heteronuclear multiple bond correlation (HMBC) experiments (Tables 1, 2). The relative stereochemistry of the isolated compounds was determined based on experimental NMR data, including 1H NMR coupling constants. 2D NMR correlations (HSQC and HMBC) were also obtained in the present study. These are the first reports of the occurrence of flavonoids 1, 2 and 6 in the genus Ceiba, as well as 3, 4 and 5 in Malvaceae.

TABLE 1. ¹H and ¹³C NMR assignments for the isolated compounds 1–3.

TABLE 2. ¹H and ¹³C NMR assignments for the isolated compounds 4–6.

Compound (7) was obtained as an amorphous colorless solid. The molecular formula was established as C₃₁H₂₈O₉ based on HRESI-MS, with the (M − H)⁻ ion observed at m/z 543.1671 (calcd. for C₃₁H₂₇O₉, error −1.9 ppm) and NMR data (Table 3). The MS/MS fragmentation pattern exhibited characteristic product ions at m/z 271.0853 and 243.0903. The ¹H NMR spectrum revealed the presence of two distinct flavonoid units, evidenced by two ABX aromatic spin systems assigned to ring A, and two AA′BB′ systems corresponding to ring B. Additionally, signals characteristic of saturated heterocyclic flavan-3,4-diol moieties (ring C) were observed. For monomer I, the A-ring displayed an ABX system at δ_H 7.26 (d, J = 8.3 Hz, H-5), 6.55 (dd, J = 8.3, 2.3 Hz, H-6), and 6.38 (d, J = 2.3 Hz, H-8). Monomer II showed a related pattern at δ_H 7.14 (d, J = 8.3 Hz, H-5″), 6.43 (dd, J = 8.3, 2.3 Hz, H-6″), and 6.31 (d, J = 2.3 Hz, H-8″), confirming analogous substitution patterns in both flavan skeletons. The B rings of both units revealed AA′BB′ systems, with signals at δ_H 7.34 (2H, H-2’/H-6′) and 6.88 (2H, H-3’/H-5′) for monomer I, and δ_H 7.29 (2H, H-2‴/H-6‴) and 6.84 (2H, H-3‴/H-5‴) for monomer II. HSQC correlations supported these assignments, linking δ_H 7.34/7.29 to δ_C 129.9 and δ_H 6.88/6.84 to δ_C 115.6. The heterocyclic C rings confirmed the presence of two flavan-3,4-diol units. For monomer I, the spin system at δ_H 4.44 (d, J = 9.8 Hz, H-2), 4.30 (dd, J = 9.8, 5.5 Hz, H-3), and 5.11 (d, J = 5.5 Hz, H-4) indicated a 2,3-trans-3,4-cis relative configuration according to Carvalho et al., 2008. HSQC correlations placed these protons at δ_C 78.3 (C-2), 76.3 (C-3), and 71.8 (C-4), consistent with oxygenated sp³ carbons. Monomer II exhibited a similar coupling pattern at δ_H 4.99 (d, J = 9.8 Hz, H-2″), 3.97 (dd, J = 9.8, 3.3 Hz, H-3″), and 4.20 (d, J = 3.3 Hz, H-4″), supporting the same relative stereochemistry according to Li et al. (2014). An additional resonance at δ_H 3.45 (s, 3H) was attributed to a methoxy group. HSQC correlations assigned H-2″, H-3″, and H-4″ to δ_C 77.8, 70.8, and 76.7, respectively. In the HMBC spectrum, the methoxy signal demonstrated a three-bond (³ J _C-H) correlation with C-4” (δ_C 76.7), confirming O-methyl substitution at this position. Key HMBC correlation between H-3 of monomer I and C-6 (or C-7 region) of monomer II supporting the C-3→C-7 interflavonoid linkage in compound 7. Long-range HMBC correlations further established the biflavonoid linkage. Notably, H-3 of monomer I (δ_H 4.30) exhibited a 4-bond correlation with C-6” (δ_C 110.2) of monomer II, supporting a C-3→C-7 interflavonoid connection (Figure 2). Comparison with literature data supported the structural assignment of compound 7, an undescribed biflavonoid, Rel-(2,3-trans,3,4-cis)-guibourtacacidine-(3→7)-(2,3-trans,3,4-cis)-4-O-methylguibourtacacidine, which we identified in this article for the first time (spectra in Supplementary Information (SI) section). It should be noted that proanthocyanidins (oligomeric flavan-3-ols) have been reported in the Malvaceae family, particularly in the genus Adansonia (Shahat, 2006; Russo et al., 2020). This indicates that the occurrence of condensed flavonoids is consistent with the known chemistry of the family. However, compound 7 differs structurally from proanthocyanidins, as it represents a dimeric biflavonoid rather than a flavan-3-ol oligomer, highlighting its relevance to the phytochemistry of Ceiba pentandra.

TABLE 3. ¹H and ¹³C NMR data for biflavonoid (7) in CD₃COCD₃.

It was interesting to observe the oxygenation pattern of all the identified compounds, as the NMR signals confirm the ABX system in ring A and AA'BB' in ring C, characterizing the low oxygenation of the flavonoids in the wood of this study, suggesting a low biosynthetic energy cost compared to the more oxygenated flavonoids in other vegetative parts of Ceiba.

The stereochemistry of compounds 1–7 was not experimentally determined in the present study. The configurations depicted were assigned based on well-established biosynthetic pathways of flavonoids and previous literature. The origin of the configuration at C-2 in flavonoids is well established to arise from the stereospecific cyclization of chalcones catalyzed by chalcone isomerase (CHI), which produces exclusively (2S)-flavanones (Jez et al., 2000). This configuration is subsequently preserved during downstream enzymatic transformations, including the conversion of flavanones to dihydroflavonols by flavanone 3-hydroxylase (F3H), which does not alter the stereochemistry at C-2 (Liu et al., 2022; Wang et al., 2020). Further along the biosynthetic pathway, dihydroflavonol 4-reductase (DFR) catalyzes a stereoselective reduction to afford leucoanthocyanidins (flavan-3,4-diols), thereby maintaining a well-defined stereochemical course throughout flavonoid biosynthesis (Winkel-Shirley, 2001).

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Lucia Pinheiro Santos Pimenta, Federal University of Minas Gerais, Brazil

Reviewed by: Abdelnaser Abdelghany Elzaawely, Tanta University, Egypt

Hashim Ibrahim, University of Wisconsin-Madison, United States