Research Article

Received: 31 October 2012, Revised: 13 February 2013, Accepted: 19 February 2013 Published online in Wiley Online Library: 17 April 2013

(wileyonlinelibrary.com) DOI 10.1002/pca.2435

574

Analysis of Cleistopholis patens Leaf and TrunkBark Oils Using Combined GC- FlameIonisation Detection, GC-Retention Index,GC–MS and 13C-NMRZana A. Ouattara,a,b Jean Brice Boti,b Antoine Coffy Ahibo,b

Joseph Casanova,a Félix Tomia and Ange Bighellia*

ABSTRACT:Introduction – Germacrenes A–C are secondary metabolites produced by various plants. They are sesquiterpene hydrocar-bons bearing the (E,E)-1,5-cyclodecadiene structure known to undergo thermal rearrangement through a [3.3]-sigmatropicreaction. Such a rearrangement was evidenced by comparing the contents of a given germacrene and the correspondingelemene calculated by GC-flame ionisation detection (FID) with the relative intensities of the signals of both molecules inthe 13C-NMR spectrum of the mixture, recorded at room temperature.Objective – To develop a protocol to identify and quantify germacrenes A, B and C and in parallel the corresponding elemenes,using a combination of GC-FID and 13C-NMR and then provide a correct analysis of Cleistopholis patens essential oils.Methods – The essential oil was submitted to GC-FID, GC-retention index, GC–MS and 13C-NMR analyses. The relativepercentages of every couple of germacrene and elemene measured by GC-FID were summed. Then, the relative ratio ofthe mean intensities of the signals of the protonated carbons of a given germacrene and the corresponding elemene wascalculated. The contents of both compounds were obtained by combining GC-FID and 13C-NMR data.Results – The true content of germacrene A/b-elemene, germacrene B/g-elemene and germacrene C/d-elemene in leaf androot oils from C. patens was evaluated by combination of GC-FID and 13C-NMR data. Correct analysis of the essential oilswas provided.Conclusion – Combined analysis of essential oil including 13C-NMR without isolation of the components, appeared reallyefficient to identify and quantify germacrene isomers and in parallel elemene isomers in essential oils. Copyright © 2013 JohnWiley & Sons, Ltd.

Keywords: 13C-NMR; germacrene; elemene; essential oil composition; Cleistopholis patens

* Correspondence to: A. Bighelli, Université de Corse-CNRS, UMR 6134 SPE,Equipe Chimie et Biomasse, Route des Sanguinaires, 20000 Ajaccio, France.E-mail: ange.bighelli@univ-corse.fr

a Université de Corse-CNRS, UMR 6134 SPE, Equipe Chimie et Biomasse,Route des Sanguinaires, 20000 Ajaccio, France

b Laboratoire de Chimie Organique Biologique, UFR-SSMT, Université FélixHouphouët-Boigny, BPV 34 Abidjan, Côte d’Ivoire

IntroductionThe genus Cleistopholis (Annonaceae) is specific to tropical Africa.The most well known of the species are C. staudii, C. glauca andC. patens; the latter two grow wild in the Ivory Coast.

Cleistopholis patens has several synonyms: C. klaineana, C.pynaertii, C. lucens, C. verschuereni, C. brevipetala and Oxymitrapatens (Le Thomas, 1969). It is present from Sierra Leone toGabon and it appreciates hydromorphic soil. In the Ivory Coast,this species grows wild in the rainforest, along river banks andin gallery forest, and it is easily cultivated. Cleistopholis patensis a tree up to 25m height, with fibrous and fragrant bark. Itsleaves are oblong elongated and can reach up to 20 cm longand 6 cm wide. Cleistopholis patens is used in Africa for variousapplications. In the Ivory Coast, the trunk bark fibres are usedby craftsmen to manufacture traditional utensils and mats. Intraditional rural housing, they are used for building of huts. InIvorian traditional medicine, the trunk bark fresh extract is usedin nasal instillation to cure headache. It is prescribed also forsleeping sickness. The decoction of the trunk bark is used as adeodorant. In addition the trunk bark of C. patens associatedwith that of Dacryodes klaineana (Burseraceae) is used as

Phytochem. Anal. 2013, 24, 574–580 Copyright © 2013 John

febrifuge. In Gabon, the very light wood is used for manufactureof rafts and pirogues (Le Thomas, 1969). The trunk bark isalso used in tuberculosis treatment or simple bronchia disease,stomach disease, diarrhoea, whitlow and oedema. In Ghana itis used for hepatic infection.

Phytochemical studies of C. patens include the isolation andstructure elucidation of terpenes (Waterman and Muhammad,1985; Ngnokam et al., 2003); alkaloids (Waterman andMuhammad,1985; Hufford et al., 1987; Liu et al., 1990; Akendengue et al., 1999)and oligorhamnosides (Seidel et al., 1999). Very little is knownabout the chemical composition of C. patens essential oils. Essen-tial oil has been isolated from various parts of Nigerian C. patens.

Wiley & Sons, Ltd.

Figure 1. Structure of germacrenes A, B and C and b-, g- andd-elemenes.

Composition of Cleistopholis patens Leaf and Trunk Bark Oils

Root oil composition was dominated by bornyl acetate and a-cadinol (content not mentioned) (Ekundayo and Oguntimein,1987). Trunk bark oil contained p-cymene (13.4%), germacrene D(12.5%) and myrcene (12.0%), whereas linalool (23.1%) was themajor component of fruit oil in addition to trans and cis linalooloxides (tetrahydrofuran) (17.7 and 17.0%). Leaf oil contained(E)-b-ocimene (31.0%), (E)-b-caryophyllene (12.8%) and linalool(8.5%) as the main components (Ekundayo et al., 1988). Thecomposition of C. patens leaf oil from Cameroun was dominatedby sesquiterpene hydrocarbons, (E)-b-caryophyllene (17.5%),germacrene D (16.1%) and germacrene B (16.0%), whereas stembark oil contained mostly d-cadinene (28.7%), a-copaene (16.9%)and germacrene B (7.4%) (Boyom et al., 2011).

In the course of our ongoing work on the characterisationof aromatic and medicinal plants growing wild in the IvoryCoast (Ouattara et al., 2012), we investigated the chemicalcomposition of the essential oils isolated from leaves and trunkbark of C. patens. In order to ensure the identification of individ-ual components and particularly sesquiterpenes, our analyseswere achieved by combination of various chromatographicand spectroscopic techniques. During the completion of thiswork, we were challenged by the identification and quantifica-tion of thermolabile compounds, precisely, germacrenes.Indeed, it appeared in various analyses that the contents of agiven germacrene and its corresponding elemene as evaluatedby GC-flame ionisation detection (FID) did not fit with therelative intensities of the signals of compounds in the 13C-NMRspectrum (spectrum recorded at room temperature). Thereforethe aim of the present work was to evaluate the true contentof every germacrene isomer in an essential oil sample, and toappreciate the degree of thermal rearrangement with respectto our GC experimental conditions. The quantification of thecorresponding elemenes was conducted in parallel. Therefore,the chemical composition of C. patens leaf and trunk bark oilswas determined taking into account the partial or total thermalrearrangement of germacrene isomers.

Experimental

Plant material

Leaves and trunk bark of C. patens were collected in May 2009, in theAhoué forest (Abidjan–Alepé route, southeastern Ivory Coast) (leaves,sample A; trunk bark, sample B), and in June 2009, in Bossé Mathié forest(near Abengourou, southeastern Ivory Coast) (trunk bark, sample C). Avoucher specimen has been deposited at the herbarium of the CentreNational Floristique (CNF, Abidjan, reference L. Aké Assi 10612).

57

Essential oil isolation and fractionation

The fresh leaves (700 g) and dried trunk bark (1000 g and 800 g) ofC. patens were subjected to hydrodistillation using a Clevenger-typeapparatus for 3 h and yielded 1.6, 1.4 and 1.7 g of essential oil and weredesignated as samples A, B and C, respectively.

An aliquot of the sample B (150mg) was fractionated into hydrocar-bons (fraction F1 eluted with pentane) and oxygenated compounds(fraction F2, eluted with diethyl ether) over silica gel (10 g, 63–200mm).Then, most of the monoterpenes contained in the hydrocarbon fractionwere concentrated under vacuum, leading to the subfraction F1’ (18mg),which was dominated by germacrene B (80% by GC, percentageobtained by summing the contents of germacrene B and g-elemene)(Fig. 1). Full sets of one- and two-dimensional NMR experiments(1H,13C, distortionless enhancement by polarization transfer (DEPT),

Phytochem. Anal. 2013, 24, 574–580 Copyright © 2013 John

heteronuclear single-quantum coherence (HSQC), heteronuclear multiple-bond correlations (HMBC)) were carried out on the semi-purified fractionF1’ to confirm the assignments for the 13C-NMR data.

Germacrene B: 1H-NMR (CDCl3; d, ppm; J, Hz): H-1 = 4.72 (broad d;12.25 Hz); H-2 = 2.30 (dd; 12.25, 4.98 Hz, 1H) and 2.04 (broad s, 1H);H-3= 1.96 (dd;12.25, 4.98 Hz, 1H) and 1.69 (s, 1H); H-5 = 4.55(broad s); H-6 = 2.17 (m, 2H); H-8 and H-9 = 2.52 (broad s, 2H), 2.04(broad s, 1H), 1.82 (broad s, 1H); H-12 and H-13 = 1.73 (s) and1.69 (s); H-14 = 1.49 (s); H-15 = 1.53 (s). 13C-NMR (CDCl3; d, ppm):137.13 (C-10), 133.24 (C-11), 131.61 (C-4), 128.21 (C-5), 126.37 (C-1),126.37 (C-7), 38.89 (C-3 and C-6), 32.53 (C-8 and C-9), 25.79 (C-2),20.75 and 20.44 (C-12 and C-13), 16.24 (C-14 and C-15).

Analytical GC

Analyses were performed on a Clarus 500 PerkinElmer (Perkin Elmer,Courtaboeuf, France) system equipped with a FID and two fused-silicacapillary columns (50m � 0.22mm, film thickness 0.25mm), BP-1(polydimethyl siloxane) and BP-20 (polyethylene glycol). The oven tem-perature was programmed from 60�C to 220�C at 2�C/min and then heldisothermal at 220�C for 20min; injector temperature: 250�C; detectortemperature: 250�C; carrier gas: helium (0.8mL/min); split: 1/60; injectedvolume: 0.5mL. The relative proportions of the oil constituents wereexpressed as percentages obtained by peak-area normalisation, withoutusing correcting factors. Retention indices (RI) were determined relativeto the retention times of a series of n-alkanes with linear interpolation(Target Compounds software from PerkinElmer).

GC–MS analysis

The three oil samples A, B and C were analysed with a PerkinElmerTurboMass (Perkin Elmer, Courtaboeuf, France) detector (quadrupole),directly coupled to a PerkinElmer Autosystem XL (Perkin Elmer,Courtaboeuf, France), equipped with a fused-silica capillary column(60m� 0.22mm i.d., film thickness 0.25mm), Rtx-1 (polydimethylsiloxane).Carrier gas, helium at 1mL/min; split, 1/80; injection volume, 0.2 mL;injector temperature, 250�C; oven temperature programmed from 60�Cto 230�C at 2�C/min and then held isothermal (45min); ion sourcetemperature, 150�C; energy ionisation, 70 eV; electron ionisation massspectra were acquired over the mass range 35–350Da.

Nuclear magnetic resonance

All NMR spectra were recorded on a Bruker AVANCE (Bruker,Wissembourg, France) 400 (resonance frequencies, 400.132MHz for 1Hand 100.623MHz for 13C) equipped with a 5-mm probe, in deuteratedchloroform (CDCl3), with all shifts referred to internal tetramethylsilane

Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/pca

5

Z. A. Ouattara et al.

576

(TMS). The 1H-NMR spectra were recorded with the following parame-ters, pulse width (PW), 4.3ms; acquisition time, 2.6 s for 32 K data tablewith a spectral width (SW) of 6000Hz (15 ppm). 13C-NMR spectra wererecorded with the following parameters: pulse width (PW), 4ms (flipangle 45�); acquisition time, 2.7 s for 128 K data table with a spectralwidth of 24000Hz (240 ppm); total repetition time, 2.8 s; CPD (compositepulse decoupling) mode decoupling; digital resolution, 0.183Hz/point.Standard pulse sequences from the Bruker library were used for two-dimensional spectra. Gradient-enhanced sequences were used for theheteronuclear two-dimensional experiments.

Identification of individual components

Identification of the individual components was based: (i) on comparisonof their GC RI on apolar and polar columns, determined relative tothe retention times of a series of n-alkanes with linear interpolation(Target Compounds software of PerkinElmer), with those of authenticcompounds, or literature data (Joulain and König, 1998; Ogunwandeet al., 2004); (ii) on computer search using digital libraries of massspectral data (Adams, 2001; König et al., 2001); and (iii) by 13C-NMRspectroscopy, following a computerised method developed in ourlaboratory, using homemade software, by comparison of the chemicalshift values of the signals in the essential oil spectrum with those ofreference compounds compiled in a laboratory-built library (Tomi andCasanova, 2006; Bighelli and Casanova, 2009). The signals of all thecarbons of identified compounds have been observed except those ofquaternary carbons of minor components. The number of overlappedsignals was always limited and the difference of chemical shift in theoil spectrum and in the reference spectra was lower or equal to0.05 ppm for almost all the signals. A compound was considered asidentified when at least 50% of its signals belonging solely to thatmolecule was observed. Two compounds have been identified bycomparison of their 13C-NMR chemical shift values with those reportedin the literature: (2E,6E)-methyl farnesoate (Ho and Millar, 2001) andjuvenile hormone III (Mori and Mori, 1987). For identification and quanti-fication of germacrene isomers and elemene isomers, see text. Thismethod allows the identification of individual components of theessential oil at contents as low as 0.3–0.4%.

Results and discussionThree essential oil samples have been isolated by water distilla-tion using a Clevenger-type apparatus from the leaves (sampleA) and the trunk bark (samples B and C) of C. patens trees,growing wild in forests of the Ivory Coast. Yields, calculated ona fresh-weight basis (leaves) or dry-weight basis (trunk bark)were 0.24%, 0.16% and 0.18%. Sample B was repeatedly fraction-ated on silica gel and a semi-purified fraction containing 80% ofgermacrene B was obtained.

Nowadays, the identification of individual components ofnatural extracts is performed by comparison of spectral data(including NMR) with those of reference compounds after isola-tion of the component. In contrast, according to the literature,the analysis of essential oils is based on the identification ofindividual components by MS (GC–MS), in combination withretention indices on a chromatographic column and preferablyon two columns of different polarities (apolar and polar). Despitethe improvement of analytical techniques, misidentification ofsesquiterpenes (or diterpenes) still occurs. Difficulties in theidentification of terpenes are encountered such as the identifica-tion of compounds (i) that co-elute on both chromatographiccolumns, (ii) that exhibit insufficiently differentiated massspectra (stereoisomers, diastereoisomers), (iii) that are not elutedon GC (heavy compounds, polar compounds). Heat-sensitivecompounds constitute another class of compounds for which

Copyright © 2013 Johnwileyonlinelibrary.com/journal/pca

identification by GC-RI and GC–MS is not easy. Indeed, thesecompounds may be damaged or they may undergo thermalrearrangement either in the injector port or in the chromato-graphic column. We experienced such problems for the quanti-fication of ascaridole that isomerise to isoascaridole on GCanalysis in the essential oil of Chenopodium ambrosioides (Cavalliet al., 2004). A similar behaviour was encountered withfuranodiene that rearranges to furanoelemene, in the essentialoil of Smyrnium olusatrum (Baldovini et al., 2001). In all cases,accurate results were obtained using a quantitative determina-tion by NMR using an internal standard or a combination ofGC-FID and 13C-NMR.

The composition of the three oil samples (A–C) has beeninvestigated by combination of GC -RI, GC–MS and 13C-NMRchromatographic and spectroscopic techniques. In total, 58 com-pounds have been identified. They accounted for 90.6–95.1% ofthe compositions.Most compoundswere identified by comparisonof their spectral data (MS and/or 13C-NMR) with those of authenticcompounds compiled for in-house MS and 13C-NMR spectral datalibraries. A few compounds were identified by comparison withliterature data. Various components accounting for 0.5–26.5% ofthe whole composition have been identified by the three tech-niques. Although almost all the same components were found inthe three oil samples, their composition varied dramatically fromsample to sample from a quantitative point of view.

Before providing an overview of the whole composition of thethree samples, we focused our attention on the identificationand quantification of germacrene isomers A, B and C, and conse-quently on the corresponding elemenes, b-elemene, g-elemeneand d-elemene, respectively. Indeed, the thermal rearrangementof 1,5-cyclodecadienes into 1,2-divinylcyclohexanes was reporteda long time ago and is well documented. The rearrangement goesthrough a [3.3]-sigmatropic reaction (Cope rearrangement) and itis particularly exemplifiedwith (E,E)-1,5-cyclodecadiene derivativesthat rearrange to trans-1,2-divinylcyclohexane derivatives. In thefield of natural products, the sesquiterpenes germacrene A, Band C have been reported to rearrange into b-, g- and d-elemenes (Fig. 1), respectively (Adio, 2009).

Germacrene A and b-elemene (leaf oil, sample A)

Analysis of a leaf oil sample of C. patens (sample A, Table 1) hasbeen carried out by GC-RI, GC–MS and 13C-NMR. The majorcomponents of that oil were mainly sesquiterpene hydrocarbons:(E)-b-caryophyllene (26.5%), germacrene D (12.7%), a-selinene(4.1%) and a-humulene (3.8%) in addition to (E)-b-ocimene(8.3%) and linalool (8.1%), the major oxygenated compoundpresent in the oil sample. b-Elemene was also identified by thethree techniques and its content was evaluated by GC-FID at6.4%. In parallel, germacrene A has also been identified in thesample by GC–MS and GC-RI, and according to GC-FID itaccounted for only 0.1% of the whole composition. However,according to the relative intensities of its signals in the 13C-NMRspectrum, with respect to those of other components, the contentof b-elemene was obviously overevaluated by GC-FID. Moreover,the 13C-NMR spectrum of the sample also contained the 15 signalsof germacrene A (d, ppm: 153.62, C-11; 138.10, C-10; 131.73, C-5;128.89, C-4; 126.44, C-1; 107.35, C-13; 51.42, C-7; 41.74, C-9; 39.58,C-3; 34.87, C-6; 33.68, C-8; 26.63, C-2; 20.29, C-12; 16.70/16.17,C-14/C-15), in agreement with the chemical shift valuesreported by Faraldos et al. (2007). It should be pointed out thatgermacrene A and b-elemene exhibit close mass spectral

Phytochem. Anal. 2013, 24, 574–580Wiley & Sons, Ltd.

Table 1. Composition of leaf and trunk bark oils from C. patensa

Compounds RIab Ripb A B C Identification

1 a-Thujene 926 1029 – – 0.2 RI, MS2 a-Pinene 934 1027 0.1 1.2 0.2 RI, MS, 13C-NMR3 Camphene 947 1073 tr 0.1 0.2 RI, MS4 Sabinene 969 1126 tr 0.3 tr RI, MS5 b-Pinene 974 1115 0.2 1.8 0.1 RI, MS, 13C-NMR6 Myrcene 984 1164 1.2 1.4 0.5 RI, MS, 13C-NMR7 a-Phellandrene 1000 1169 – 19.5 6.9 RI, MS, 13C-NMR8 d -3-Carene 1008 1152 – 2.3 0.9 RI, MS, 13C-NMR9 a-Terpinene 1009 1184 – 0.8 0.4 RI, MS, 13C-NMR10 p-Cymene 1015 1274 0.2 7.4 6.1 RI, MS, 13C-NMR11 Limonene* 1024 1205 0.5 5.7 2.4 RI, MS, 13C-NMR12 1,8-Cineole* 1024 1212 – 2.7 – RI, MS, 13C-NMR13 b-phellandrene* 1024 1214 0.1 0.9 0.4 RI, MS, 13C-NMR14 (Z)-b-ocimene 1028 1235 0.6 0.6 0.1 RI, MS, 13C-NMR15 (E)-b-ocimene 1039 1253 8.3 0.3 0.1 RI, MS, 13C-NMR16 g-Terpinene 1051 1253 – 1.0 0.4 RI, MS, 13C-NMR17 cis-Linalool oxide THF 1057 1439 0.7 – – RI, 13C-NMR18 trans-Linalool oxide THF 1072 1467 0.7 – – RI, 13C-NMR19 Terpinolene 1081 1286 0.2 4.3 1.0 RI, MS, 13C-NMR20 Linalool 1085 1544 8.1 0.9 0.4 RI, MS, 13C-NMR21 Citronellal 1131 1478 – – 1.6 RI, MS, 13C-NMR22 Terpinen-4-ol 1164 1600 – tr 0.1 RI, MS23 a-Terpineol 1173 1691 1.8 – – RI, MS, 13C-NMR24 Nerol 1209 1795 0.3 – – RI, MS, 13C-NMR25 Geraniol 1233 1840 1.0 – 2.4 RI, MS, 13C-NMR26 Thymol 1268 2198 0.2 – 1.5 RI, MS, 13C-NMR27 d-Elemene 1336 1472 – 0.2e 5.6c RI, MS, 13C-NMR28 Germacrene C – e 2.6c 13C-NMR29 a-Cubebene 1349 1459 0.4 – – RI, MS, 13C NMR30 a-Copaene 1377 1494 1.3 1.5 1.1 RI, MS, 13C-NMR31 b-Elemene 1389 1592 4.2c 0.5d 1.5d RI, MS, 13C-NMR32 (E)-b-caryophyllene 1420 1600 26.5 2.3 2.1 RI, MS, 13C-NMR33 g-Elemene 1428 1635 – tr tr RI, MS34 b-Copaene 1429 1592 0.6 – – RI, MS, 13C-NMR35 (E)-b-farnesene 1447 1667 0.6 – – RI, 13C-NMR36 a-Humulene 1452 1669 3.8 0.3 0.4 RI, MS, 13C-NMR37 g-Muurolene 1474 1688 1.8 0.2 1.2 RI, MS, 13C-NMR38 Germacrene D 1477 1708 12.7 2.3 25.4 RI, MS, 13C-NMR39 b-Selinene 1482 1716 2.5 – – RI, MS, 13C-NMR40 a-Selinene 1491 1721 4.1 – – RI, MS, 13C-NMR41 Bicyclogermacrene 1494 1727 0.1 0.5 RI, MS, 13C NMR42 Germacrene A 1502 ND 2.3c d d RI, MS, 13C-NMR43 g-Cadinene 1507 1758 0.9 – 0.7 RI, MS, 13C-NMR44 d-Cadinene 1515 1756 1.9 0.3 1.2 RI, MS, 13C-NMR45 Selina-4(15),7(11)-diene 1528 1775 – 1.0 0.1 RI, 13C-NMR46 b-Elemol 1537 2075 0.6 5.3 10.5 RI, MS, 13C-NMR47 Germacrene B 1552 1826 – 19.6c 3.2c RI, MS, 2D-NMR48 Caryophyllene oxide 1571 1977 1.3 0.1 0.2 RI, MS, 13C-NMR49 Alismol 1611 2264 – – 1.4 RI, 13C-NMR50 g-Eudesmol 1619 2177 0.2 tr 0.6 RI, 13C-NMR51 t-Cadinol 1626 2180 – 0.1 1.1 RI, 13C-NMR52 b-Eudesmol 1636 2243 0.2 1.5 0.8 RI, 13C-NMR53 a-Cadinol 1638 2244 – – 1.0 RI, 13C-NMR54 a-Eudesmol 1641 2234 0.9 1.1 0.7 RI, MS, 13C-NMR55 (2Z,6E)-farnesal 1689 ND – 0.7 – RI, 13C-NMR56 (2E,6E)-farnesal 1715 ND – 1.3 – RI, 13C-NMR57 (2E,6E)- methyl farnesoate 1762 2227 – 3.0 0.2 RI, 13C-NMR

(Continues)

Composition of Cleistopholis patens Leaf and Trunk Bark Oils

Phytochem. Anal. 2013, 24, 574–580 Copyright © 2013 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/pca

577

Table 1. (Continued)

Compounds RIab Ripb A B C Identification

58 Juvenile hormone III 1864 2496 – 2.6 2.6 RI, 13C-NMRTotal 90.9 95.1 90.6

Bold typeface: compound identified by 13C-NMR; tr< 0,05%; ND, not detected.aOrder of elution and relative percentages of individual components are given on apolar column (BP-1), excepted those with anasterisk (*), percentages on polar column (BP-20).bRIa, RIp = retention indices measured on apolar and polar capillary column, respectively.cPercentage calculated by combination of GC and NMR (see text).dPercentage measured by GC-FID, probably a mixture of germacrene A and b-elemene.ePercentage measured by GC-FID, probably a mixture of germacrene C and d-elemene.

GA

BE

GABE

BE BE

GA

125 115120 110 105

50.552.5 48.5 46.5 ppm

GA : germacrene A BE : beta-elemene

ppm

Z. A. Ouattara et al.

578

fragmentations, being differentiated only by the intensities ofthe fragments (Joulain and König, 1998). Comparing the resultsobtained by GC-RI and GC(MS) on the one hand and by 13C-NMRon the other hand, it is evident that, under our GC experimentalparameters (see Experimental section), germacrene A rearrangedpractically in total to b-elemene during GC-FID and GC–MSanalysis. A more accurate quantification was obtained bycombination of GC and NMR data (spectrum recorded at roomtemperature). Indeed, the total percentage of the two compoundshas been obtained by adding their relative percentages(6.4%+0.1%) and the content of each one was calculated withrespect to the ratio of the mean intensity values of the non-overlapped signals of protonated carbons in the 13C-NMRspectrum (germacrene A (d, ppm): 126.44 (CH); 107.35 (CH2);51.42 (CH); 33.68 (CH2); 16.70 (CH3); 16.17 (CH3); b-elemene:150.28, C-1(CH); 112.11, C-3(CH2); 109.86, C-2(CH2); 108.30,C-13(CH2); 52.74, C-5(CH); 45.73, C-7(CH); 21.02, C-12(CH3);16.61, C-14(CH3); IgermacreneA/Ib-elemene = 35/65; Fig. 2). Then, therespective contents of germacrene A and b-elemene werecalculated and evaluated at 2.3% and 4.2%, respectively.

In samples B and C, germacrene A was not detected whileb-elemene accounted for 0.5% and 1.5% by GC-FID, respectively.Only b-elemene was identified by NMR in sample C. It could beassumed, by comparison with sample A, that the measuredpercentages on GC-FID of b-elemene are overevaluated.

Figure 2. Parts of the 13C-NMR spectrum of the leaf essential oil fromCleistopholis patens (sample A).

Germacrene B and g-elemene (trunk bark oil, sample B)

Analysis of a trunk bark oil sample of C. patens (sample B, Table 1)has also been carried out by GC-RI, GC–MS and 13C-NMR. Thatoil sample contained an appreciable content of monoterpenes,including the major component, a-phellandrene (19.5%). b-Elemol(5.3%) was the major sesquiterpene identified.

However, a discrepancy appeared again between GC-RI andGC–MS analysis on the one hand and 13C-NMR analysis on theother hand. Indeed, two sesquiterpenes (RIa/RIp = 1428/1635,6.1% and 1552/1826, 13.5%) have been identified by GC-RI andGC–MS as g-elemene and germacrene B. Once again, it shouldbe pointed out that the mass spectra of both compounds arevery similar (Joulain and König, 1998). However, none of thesignals of g-elemene, reported by Brauchli and Thomas (1991),were observed in the 13C-NMR spectrum of the essential oil. Thisobservation led us to suspect that g-elemene is either notpresent in the sample investigated or is present at very lowcontent (less than 0.3–0.4%). A side-product coming from thethermal rearrangement of germacrene B is probably being

Copyright © 2013 Johnwileyonlinelibrary.com/journal/pca

observed in GC, as demonstrated by Adio (2009). In parallel, ninesignals (d, ppm: 131.62; 128.22; 126.39; 38.91; 32.53; 25.80; 20.75;20.43; 16.24) remained unassigned in the 13C-NMR spectrum.Among them, three signals (d, ppm: 38.91; 32.53; 16.24)exhibited higher intensity than the other six and they probablybelong to several carbons. Unfortunately, the NMR data ofgermacrene B (Adio et al., 2004) came from a spectrum recordedin C6D6 and it was not easy to compare the chemical shift values.As all the reference spectra compiled in our computerised 13C-NMR library are recorded in CDCl3, it was crucial to confirm the13C-NMR chemical shift values of germacrene B in that solventand to assign these values to the corresponding carbons.

The essential oil was repeatedly chromatographed on silicagel (see Experimental section). The hydrocarbons were firstseparated from the oxygenated compounds. Then, most ofthe monoterpenes contained in the hydrocarbon fraction (F1)were concentrated under vacuum, leading to the subfraction

Phytochem. Anal. 2013, 24, 574–580Wiley & Sons, Ltd.

Composition of Cleistopholis patens Leaf and Trunk Bark Oils

F1’, which was dominated by germacrene B (80% by GC, percent-age obtained by summing the contents of germacrene B andg-elemene) (Fig. 1). Subfraction F1’was submitted to the full set ofNMR experiments in CDCl3 (

1H, 13C, DEPT, HSQC, HMBC) and the13C-NMR spectrum was also recorded in C6D6. Correlation plotsin the HSQC spectrum demonstrated that signals of carbons 3and 6 (CH2) on the one hand (38.89 ppm), and those of carbons8 and 9 (CH2) on the other hand (32.53 ppm), as well as those ofcarbons 14 and 15 (CH3, 16.24 ppm) overlap. Moreover, correlationplots in the HMBC spectrum between the signals at 1.69and 1.73 ppm (H-12 and H-13) and signals at 133.24ppm (C-11)and 126.37 (C-7) demonstrated that the signals for C-1 and C-7overlap. Therefore, the 13C-NMR spectrum of the sesquiterpenegermacrene B, in CDCl3, exhibits only 11 signals. It should bementioned that the intensities of signals of quaternary carbonsC-10 (137.13 ppm) and C-11 (133.24ppm) are very low in spectrarecorded routinely with respect to that of the signal of C-4(131.61 ppm), probably due to a poor relaxation of these carbons.

Moreover, to ensure the occurrence of germacrene B in theessential oil, an aliquot of oil was kept at 200�C in an oven, for2 h. As expected, the same nine signals mentioned abovebelonging to germacrene B disappeared and the signals of g-elemene (d, ppm: 149.94, C-1; 109.91, C-2; 111.84, C-3; 147.94,C-4; 53.03, C-5; 31.30, C-6; 130.93, C-7; 25.45, C-8; 40.08, C-9;39.91, C-10; 120.99, C-11; 20.04/19.92, C-12/C-13; 16.76, C-14;24.76, C-15) were observed (Brauchli and Thomas, 1991). Itcan be concluded that, under our experimental conditions forGC-FID and GC–MS analysis, germacrene B partially rearrangedto g-elemene. Hence, it can be assumed that germacrene Brepresented 19.6% (sum of 6.1% and 13.5%) of the wholecomposition of C. patens trunk bark oil (sample B).

Germacrene B and g-elemene were not detected in sampleA. In sample C, they accounted for 1.9% and 1.3%, respectively(GC-FID). All the signals of germacrene B were observed in theNMR spectrum while those of g-elemene were not observed.Therefore, in parallel with the analysis of sample B, it can beassumed that the sum of the percentages should be attributedto germacrene B, the content of which is evaluated at 3.2%.

57

Germacrene C and d-elemene (trunk bark oil, sample C)

A second sample of C. patens trunk bark oil (sample C, Table 1)also has been analysed by GC-RI, GC–MS and 13C-NMR. Thecomposition of that oil sample was largely dominated bysesquiterpenes, germacrene D (25.4%) and b-elemol (10.5%),with a-phellandrene representing only 6.9%.

In the sample C, d-elemene was identified by GC–MS and by13C-NMR, and according to GC-FID it accounted for 8.2% of thewhole composition. In contrast, germacrene C was identifiedonly by 13C-NMR. Once again, germacrene C is not detected byGC–MS due to its thermal rearrangement to d-elemene, thecontent of which is overestimated. Taking into account themean values of the intensities of the signals belonging toevery molecule in the 13C-NMR spectrum, the ratio germacreneC/d-elemene was evaluated as 32/68 and the two compoundsaccounted for 2.6% and 5.6%, respectively, in this oil sample. Itcan be assumed that, under our experimental conditions,germacrene C is totally converted to d-elemene. Moreover, tothe best of our knowledge, the RI of germacrene C are notreported in the literature.

Germacrene C and d-elemene were not detected in sample A,while d-elemene accounted for 0.2% in sample B, identified by

Phytochem. Anal. 2013, 24, 574–580 Copyright © 2013 John

GC-RI and GC–MS. According to the results obtained on sampleC, it is likely that the essential oil contained a small amountof germacrene C, which was isomerised to d-elemene duringGC-FID and GC–MS analysis.

Chemical composition of C. patens leaf and trunk bark oils(samples A, B and C)

In total, 36 components have been identified in the leaf oilsample of C. patens (sample A), they accounted for 90.9%of the whole composition (Table 1). The major components ofthat oil were mainly sesquiterpene hydrocarbons: (E)-b-caryophyllene (26.5%), germacrene D (12.7%), a-selinene (4,1%)and a-humulene (3,8%). Germacrene A and b-elemene wereevaluated at 2.3% and 4.2%, respectively, by combination ofGC-RI, GC–MS and 13C-NMR. (E)-b-ocimene (8.3%) was themajor monoterpene hydrocarbon while linalool (8.1%) was themajor oxygenated compound present in the oil sample. IvorianC. patens leaf oil differed quantitatively from Nigerian andCameroonian oils. Indeed, the former was dominated by (E)-b-ocimene (31.0%, with (E)-b-caryophyllene of 12.8%) (Ekundayoet al., 1988), and the later contained mostly (E)-b-caryophyllene(17.5%), germacrene D (16.1%) and germacrene B 16.0%,whereas (E)-b-ocimene was not detected (Boyom et al., 2011).It should be pointed out that germacrene B was not detectedin our sample whereas an appreciable content of germacreneA was present.The two samples of Ivorian C. patens trunk bark oil differed

quantitatively. The 40 and 46 identified compounds accountedfor 95.1 and 90.6%, respectively of the whole composition(Table 1). Sample B exhibited a high content of monoterpenehydrocarbons, a-phellandrene (19.5%) being the major one, withp-cymene (7.4%) and limonene (5.7%). Germacrene B (19.6%)was the major sesquiterpene hydrocarbon and b-elemol (5.3%)the major oxygenated sesquiterpene. The composition ofsample C was dominated by germacrene D (25.4%, versus2.3% in sample B) and b-elemol (10.5%). Germacrene C (2.6%)was identified as well as d-elemene (5.6%). Both samplescontained methyl (2E, 6E)-farnesoate and juvenile hormone III(10,11-epoxy-(E)-b-farnesene = 10,11-epoxy-3-methylenedodeca-1,6(E)-diene), two insect hormones scarcely found in plants.The trunk bark oil from Nigeria differed from our samples by itshigh content of myrcene (12.0% versus 0.5-1.4% in our samples)as well as p-cymene (13.4%) and germacrene D (12.5%) (Ekundayoet al., 1988). In contrast, stem bark oil of C. patens from Cameroundiffered dramatically, being dominated by d-cadinene (28.7%) anda-copaene (16.9%) (Boyom et al., 2011).

Summary

The composition of leaf and trunk bark oils from IvorianC. patens is reported for the first time. Combined analysis ofthree oil samples by combination of chromatographic andspectroscopic techniques, GC-RI, GC–MS, 13C-NMR, demon-strated the originality of the compositions of these oils withrespect to those of other oil samples isolated from plantsgrowing in other African countries. Analysis of the samples by13C-NMR without isolation of the individual components allowedthe identification of various compounds scarcely found in essen-tial oils, such as juvenile hormone III. Finally, combined use ofGC-FID and 13C-NMR allowed identification as well as the correctquantification of the heat-sensitive compounds germacrenes A,

Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/pca

9

Z. A. Ouattara et al.

580

B and C and in parallel that of the corresponding b-, g- and d-elemenes. The thermal rearrangement of a given germacrene tothe corresponding elemene may occur in the injector port and/or in the chromatographic column, the degree of rearrangementbeing dependent on the experimental GC-FID and GC–MS param-eters used in every laboratory (temperature of injector, highesttemperature raised in the column, length of the capillary column,temperature program, etc.).

Acknowledgements

The authors wish to thank Professor L. Aké Assi for his valuablehelp in the identification of the plant, the CPN Laboratory ofthe University of Corsica where MS spectra have been recordedand the Ministère de l’Enseignement Supérieur de Côte d’Ivoirefor providing a research grant to Z. A. Ouattara.

ReferencesAdams RP. 2001. Identification of Essential Oil Components by Gas

Chromatography/Mass Spectrometry. Allured: Carol Stream, IL; 1–804.Adio AM. 2009. Germacrenes A–E and related compounds: thermal,

photochemical and acid induced transannular cyclizations. Tetrahedron65: 1533–1552.

Adio AM, Paul C, Tesso H, Kloth P, König WA. 2004. Absolute configura-tion of helminthogermacrene. Tetrahedron Asymmetr 15: 1631–1635.

Akendengue B, Ngou-Milama E, Koudogbo B, Roblot F, Laurens A,Hocquemiller R. 1999. 8-Hydroxysampangine a new alkaloid isolatedfrom Cleistopholis patens. Nat Prod Lett 13: 147–150.

Bighelli A, Casanova J. 2009. Analytical tools for analysing Cymbopogonoils. In Essential Oil Bearing Grasses – Cymbopogons, Akhila A (ed.).CRC Press: London; 195–221.

Baldovini N, Tomi F, Casanova J. 2001. Identification and quantitativedetermination of furanodiene, a heat-sensitive compound, inessential oils by 13C-NMR. Phytochem Anal 12: 58–63.

Boyom FF, Ngouana V, Kemgne EAM, Avam Zollo PH, Menut C, BessiereJM, Gut J, Rosenthal PJ. 2011. Antiplasmodial volatile extracts fromCleistopholis patens Engler & Diels and Uvariastrum pierreanum Engl.(Engl. & Diels) (Annonaceae) growing in Cameroon. Parasitol Res108: 1211–1217.

Brauchli R, Thomas AF. 1991. Comment on sesquiterpene composition ofbasil oil. Assignment of the 1H and 13C NMR spectra of b-elemenewith two-dimensional NMR. J Agric Food Chem 39: 431–431.

Copyright © 2013 Johnwileyonlinelibrary.com/journal/pca

Cavalli JF, Tomi F, Bernardini AF, Casanova J. 2004. Combined analysis ofChenopodium ambrosioides essential oil by GC, GC-MS and 13C-NMRspectroscopy. Quantitative determination of ascaridole, a heat-sensitive compound. Phytochem Anal 15: 275–279.

Ekundayo O, Oguntimein B. 1987. Major constituents of the essential oilof Cleistopholis patens root. Planta Med 53: 228–229.

Ekundayo O, Laakso I, Oguntimein B, Hiltunen R. 1988. Volatile compo-nents of Cleistopholis patens. Planta Med 54: 338–340.

Faraldos JA, Wu S, Chappell J, Coates RM. 2007. Conformational analysisof (+)-germacrene A by variable temperature NMR and NOE spectros-copy. Tetrahedron 63: 7733–7742.

Ho H-Y, Millar JG. 2001. Identification and synthesis of male-producedsex pheromone components of the stink bugs Chlorochroa ligataand Chlorochroa uhleri. J Chem Ecol 27: 2067–2095.

Hufford CD, Liu S, Clark AM. 1987. Anticandidal activity of eupolauridineand onychine alkaloids from Cleistopholis patens. J Nat Prod 50:961–964.

Joulain D, König WA. 1998. The Atlas of Spectral Data of SesquiterpeneHydrocarbons. E.B.-Verlag: Hamburg; 1–658.

König WA, Hochmuth DH, Joulain D. 2001. Terpenoids and RelatedConstituents of Essential Oils, Library of MassFinder 2.1. University ofHamburg, Institute of Organic Chemistry: Hamburg.

Le Thomas A. 1969. Flore du Gabon, Volume 16: Annonaceae. MuséumNational d’Histoire Naturelle: Paris; 91–94.

Liu SC, Oguntimein B, Hufford CD, Clark AM. 1990. 3-Methoxysampangine, a novel antifungal copyrine alkaloid fromCleistopholis patens. Antimicrob Agents Chemother 34: 529–533.

Mori K, Mori H. 1987. Synthesis of both enantiomers of juvenile hormoneIII. Tetrahedron 43: 4097–4106.

Ngnokam D, Tsopmo A, Ayafor JF, Nuzillard JM, Sterner O. 2003. 4,5-epoxide-1,6-dimethyl p-coumarate: a novel monoterpene derivativefrom Cleistopholis patens. Bull Chem Soc Ethiopia 17: 177–180.

Ogunwande IA, Olawore NO, Adeleke KA, Ekundayo O, König WA. 2004.Rare terpenoid esters from Hyparrhenia rufa (Nees) Stapf. growingwild in Nigeria. Flavour Fragrance J 19: 239–243.

Ouattara ZA, Boti JB, Ahibo CA, Tomi F, Casanova J, Bighelli A. 2012.Chemical composition of the leaf oil of Cleistopholis glauca Pierre exEngler & Diels from Côte d’Ivoire. J Essent Oil Res 24: 474–878.

Seidel V, Bailleul F, Waterman PG. 1999. Partially acetylated tri- andtetrarhamnoside dodecanyl ether derivatives from Cleistopholispatens. Phytochemistry 52: 465–472.

Tomi F, Casanova J. 2006. 13C-NMR as a tool for identification of individ-ual components of essential oils from Labiatae. A review. Acta Hortic723: 185–192.

Waterman PG, Muhammad I. 1985. Sesquiterpenes and alkaloids fromCleistopholis patens. Phytochemistry 24: 523–527.

Phytochem. Anal. 2013, 24, 574–580Wiley & Sons, Ltd.