Polyphenols from Allanblackia floribunda seeds: Identification, quantification and antioxidant activity

Aberystwyth University Polyphenols from Allanblackia floribunda seeds Akpanika, Grace A.; Winters, Anne; Wilson, Thomas; Ayoola, Gloria A.; Adepoju-Bello, Aderonke A.; Hauck, Barbara Published in: Food Chemistry DOI: 10.1016/j.foodchem.2016.12.002 Publication date: 2017 Citation for published version (APA): Akpanika, G. A., Winters, A., Wilson, T., Ayoola, G. A., Adepoju-Bello, A. A., & Hauck, B. (2017). Polyphenols from Allanblackia floribunda seeds: Identification, quantification and antioxidant activity. Food Chemistry, 222, 35-42. https://doi.org/10.1016/j.foodchem.2016.12.002 Document License CC BY-NC-ND General rights Copyright and moral rights for the publications made accessible in the Aberystwyth Research Portal (the Institutional Repository) are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the Aberystwyth Research Portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the Aberystwyth Research Portal Take down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. tel: +44 1970 62 2400 email: is@aber.ac.uk Download date: 16. Jun. 2020 Accepted Manuscript Polyphenols from Allanblackia floribunda seeds: identification, quantification and antioxidant activity Grace A. Akpanika, Ana Winters, Thomas Wilson, Gloria A. Ayoola, Aderonke A. Adepoju-Bello, Barbara Hauck PII: DOI: Reference: S0308-8146(16)32010-6 http://dx.doi.org/10.1016/j.foodchem.2016.12.002 FOCH 20294 To appear in: Food Chemistry Received Date: Revised Date: Accepted Date: 3 August 2016 29 November 2016 2 December 2016 Please cite this article as: Akpanika, G.A., Winters, A., Wilson, T., Ayoola, G.A., Adepoju-Bello, A.A., Hauck, B., Polyphenols from Allanblackia floribunda seeds: identification, quantification and antioxidant activity, Food Chemistry (2016), doi: http://dx.doi.org/10.1016/j.foodchem.2016.12.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. 1 Polyphenols from Allanblackia floribunda seeds: identification, quantification and 2 antioxidant activity 3 4 Grace A. Akpanika a, Ana Winters b, Thomas Wilson b, Gloria A. Ayoola a, Aderonke A. 5 Adepoju-Bello b and Barbara Hauck b* 6 7 a 8 Nigeria 9 b Department of Pharmaceutical Chemistry, Faculty of Pharmacy, University of Lagos, Institute of Biological, Environmental and Rural Sciences, Aberystwyth University, 10 Gogerddan, Aberystwyth, SY23 3EB, UK 11 * Corresponding author. Phone +44 1970823000, email bdh@aber.ac.uk 12 13 email addresses: 14 gakpanika@yahoo.com (G. A. Akpanika) 15 alg@aber.ac.uk (A. Winters) 16 tpw2@aber.ac.uk (T. Wilson) 17 gayoola@unilag.edu.ng (G. A. Ayoola) 18 aadepoju-bello@unilag.edu.ng (A. A. Adepoju-Bello) 19 bdh@aber.ac.uk (B. Hauck) 20 21 Abstract 22 Oil rich seeds of Allanblackia floribunda, a tree from tropical Africa, have traditionally been 23 used in food preparation. Furthermore, the therapeutic properties of various parts of this 24 tree have long been exploited in traditional medicine. As both food and pharmaceutical 1 25 industries show growing interest in tropical tree crops, this study aimed to investigate 26 whether A. floribunda seeds could also be used as a source of potentially bioactive 27 compounds. The polyphenol profile revealed six predominant compounds which were 28 identified by HPLC-PDA-ESI/MSn as the biflavonoids morelloflavone, Gb-2a and 29 volkensiflavone and their respective glucosides. A range of less abundant flavones, flavonols 30 and flavan-3-ols was also detected. All six major compounds showed antioxidant activity, 31 with the activity of morelloflavone, its glucoside and Gb-2a-glucoside comparable with that 32 of ascorbic acid. The main compounds accounted for approximately 10% of dry weight, 33 making the seeds used for oil production a rich source of biflavonoids as a by-product. 34 35 Keywords: Allanblackia floribunda, phenolic compounds, biflavonoids, HPLC-PDA-ESI/MSn, 36 morelloflavone, volkensiflavone, Gb-2a, antioxidant capacity 37 38 Chemical compounds studied in this article 39 morelloflavone (PubChem CID: 5464454); volkensiflavone (PubChem CID: 23844069); Gb-2a 40 (PubChem CID: 176988) 41 42 1. Introduction 43 Allanblackia floribunda Oliv. (Clusiaceae or Guttiferae) is an evergreen tree which grows 44 in the tropical rainforests of Africa to a height of up to 30 m and is traditionally used in a 45 variety of ways (Orwa & Munjuga, 2007; Orwa, Mutua, Kindt, Jamnadass & Anthony, 2009). 46 Whilst the seeds consist to over 60% of oil and provide edible vegetable fat (Orwa et al., 47 2009; Wilfred, Adubofuor & Oldham, 2010), various parts of the tree, including roots and 48 bark, are used in traditional medicine for the treatment of a range of ailments such as 2 49 toothache, dysentery and coughs (Betti, 2004; Olowokudejo, Kadiri & Travih, 2008; Kayode, 50 2006). Furthermore, Allanblackia floribunda extracts have been reported to exhibit 51 antimicrobial, anti-inflammatory and antioxidant activity (Ajibesin, Rene, Bala & Essiett, 52 2008; Ayoola et al., 2009; Ayoola, Ipav, Sofidiya, Adepoju-Bello, Coker & Odugbemi, 2008; 53 Kuete et al., 2011) as well as potential antimalarial (Azebaze, Teinkela, Nguemfo, Valentin, 54 Dongmo & Vardamides, 2015) and anticancer effects (Fadeyi, Fadeyi, Adejumo, Okoro & 55 Myles, 2013). To date, a range of potentially bioactive compounds have been reported in a 56 variety of extracts from Allanblackia floribunda and related species, including 57 benzophenones, xanthones and biflavonoids (Locksley & Murray, 1971; Fuller, Blunt, 58 Boswell, Cardellina ΙΙ & Boyd, 1999; Nkengfack, Azebaze, Vardamides, Fomum & van 59 Heerden, 2002). In the seed, the focus has mainly been on oil, with some information 60 available on antioxidative properties of A. floribunda seed cake (Boudjeko, Ngomoyogoli, 61 Woguia & Yanou, 2013). With growing interest in Allanblackia species as tree crop in 62 particular for the food industry (IUCN, 2014), the aim of this study was to investigate the 63 polyphenol profile of A. floribunda seeds in order to establish whether seed remnants after 64 plant oil extraction could be a source of bioactive compounds with interest to the 65 pharmaceutical sector. 66 67 2. Materials and methods 68 2.1 Plant material and reagents 69 Fruits of Allanblackia floribunda were collected in three different years from uncultivated 70 farmland in Oke Igbo, Ondo State, South Western Nigeria. The fruits were authenticated at 71 the Forestry Research Institute of Nigeria (FRIN), Ibadan (voucher specimen number 72 FHI107929). Seeds were separated from the mesocarp, and air-dried seeds were crushed, 3 73 freeze-dried and milled into a fine powder (≤1.0mm) with an automated bespoke ball mill 74 robot (Labman Automation, Middlesborough, UK). Briefly, three stainless steel balls (3mm 75 diameter) were added to approximately 200mg seeds in a 2mL microtube and shaken in the 76 robot at the predetermined speed and duration to achieve the desired particle size. 77 78 HPLC grade solvents for extraction and analysis of phenolics were purchased from VWR 79 (Lutterworth, Leicestershire, UK). Response factors for quantification were obtained using 80 flavonoid standards with a minimum of 98% purity (Carbosynth Ltd, Compton, Berkshire, 81 UK). All other analytical standards and chemicals were purchased from Sigma Aldrich 82 (Gillingham, Dorset, UK). 83 84 2.2 Extraction of soluble phenolics 85 Soluble phenolics were extracted from approximately 20 mg seed powder by shaking 86 with 5 ml 70% methanol for 15 min at room temperature. The sample was then centrifuged 87 for 10 min at 1700g, the supernatant decanted and the pellet extracted twice more. 88 Methanol was removed from the combined supernatants under vacuum at 60 °C before 89 extracts were partially purified by solid phase extraction (SPE) using Sep-Pak C18 cartridges 90 (Waters Ltd, Elstree, UK) as described by Hauck, Gallagher, Morris, Leemans & Winters 91 (2014) and dried under vacuum at 60 °C. Prior to further analysis, samples were typically 92 dissolved in 0.5 ml 70% methanol and diluted as necessary. 93 94 2.3 Liquid chromatography-tandem mass spectrometry 95 Secondary metabolites were analysed by reverse-phase high performance liquid 96 chromatography with online photodiode array detection and electrospray ionisation–ion 4 97 trap tandem mass spectrometry (HPLC-PDA-ESI/MSn). Structural elucidation was performed 98 on a Thermo Finnigan LC-MS system (Thermo Electron Corp, Waltham, MA, USA) comprising 99 a Finnigan PDA Plus detector, a Finnigan LTQ linear ion trap with ESI source and a Waters C18 100 Nova-Pak column (3.9 x 100 mm, particle size 4 µm), with column oven temperature 101 maintained at 30 °C. The PDA scan range was set to 240–400 nm, and injection volume was 102 typically 10 µl. The mobile phase consisted of water with 0.1% formic acid (solvent A) and 103 methanol with 0.1% formic acid (solvent B). The column was equilibrated with 95% solvent 104 A at a flow rate of 1ml min-1, with 10% going to the mass spectrometer, and the percentage 105 of solvent B increased linearly to 65% over 60 min. MS parameters were as follows: sheath 106 gas 30, auxiliary gas 15 and sweep gas zero (arbitrary units), spray voltage -4.0 kV in 107 negative and 4.8 kV in positive ionisation mode, capillary temperature 320 °C, capillary 108 voltage -1.0 V and 45 V, respectively, tube lens voltage -68 and 110 V, respectively, and 109 normalised collision energy (CE) typically 35%. 110 Accurate mass measurements only were carried out on an Orbitrap Fusion Tribrid mass 111 spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) with heated ESI source coupled 112 to a Dionex UltiMate 3000 ultra high performance liquid chromatography (UHPLC) system 113 (Thermo Fisher Scientific). Chromatographic separation was performed on a reverse-phase 114 C18 Hypersil Gold column (20 x 2.1 mm, particle size 1.9 μm; Thermo Fisher Scientific) which 115 was maintained at a temperature of 60 °C and with solvents A and B as described above. 116 The column was equilibrated with 95% solvent A at a flow rate of 0.4 ml min-1, and the 117 percentage of solvent B increased linearly to 100% over 7 min, followed by isocratic elution 118 at 100% solvent B for 3.5 min. High resolution mass spectra were acquired in both negative 119 and positive ionisation mode at a resolution of 120,000 with an automatic gain control 120 (AGC) target of 2e5 and a maximum injection time of 100 ms, with other MS parameters as 5 121 follows: vaporiser temperature 358 °C, spray voltage -2.5 kV in negative and 3.5 kV in 122 positive ionisation mode, sheath gas 45, auxiliary gas 13 and sweep gas 1 (arbitrary units) 123 and capillary temperature 342 °C. 124 125 2.4 Quantification of predominant compounds 126 For quantification of the predominant compounds, seed samples harvested in three 127 different years were extracted in triplicate as described in section 2.2, followed by reverse- 128 phase HPLC. The system comprised a Waters 996 PDA detector and a Waters C18 Nova-Pak 129 radial compression column (8 x 100 mm, particle size 4 µm), with chromatographic 130 conditions as described by Hauck et al. (2014) and detection wavelength set to 280nm. 131 Response factors (RF) for the flavonoid monomers naringenin, eriodictyol, apigenin and 132 luteolin were obtained from linear standard curves over the range of 0.2 – 10 µg of standard 133 and calculated as 134 135 RF = amount [µg] / peak area280nm. All response curves had linear regression coefficients > 0.996. 136 Response factors of biflavonoid aglycones were calculated as the mean of the relevant 137 flavonoid monomer RFs, and for the quantification of biflavonoid glucosides these values 138 were multiplied by the following correction factor to allow for the difference in molecular 139 weight (Mr): 140 (Mr aglycone + 162) / Mr aglycone. 141 142 2.5 Acid hydrolysis and analysis of sugar moieties and flavonoid aglycones 143 To identify the aglycone part of the flavonoid glycosides, acid hydrolysis of total extract 144 was carried out by combining equal volumes of aqueous extract and 2 mol L-1 HCl. After 6 145 heating the solution to 90 °C for 1 h, the pH was adjusted to 3 – 4 with NaOH prior to partial 146 purification by SPE as described in section 2.2. Flavonoid aglycones were then analysed by 147 HPLC-PDA-ESI/MSn as described in section 2.3 and identified by direct comparison with 148 relevant flavonoid standards. 149 For identification of the sugar moieties of the main glycosylated metabolites, compounds 150 were purified by collecting the HPLC eluent of individual peaks from the Waters system 151 described in section 2.4. After reducing the volume under vacuum at 60 °C acid hydrolysis 152 was performed by combining equal volumes of aqueous extract and 2 mol L-1 HCl. The 153 solution was heated to 90 °C for 1 h and subsequently the pH was adjusted to 6 - 7 with 154 NaOH. The samples were then loaded onto Strata-X-A strong anion exchange columns 155 (Phenomenex, Macclesfield, UK) which had been conditioned with 100% methanol and 156 equilibrated with water. The sugar containing non-binding fraction of the samples was 157 collected for further analysis. 158 Monosaccharides were analysed by high-performance anion-exchange chromatography 159 with pulsed amperometric detection (HPAEC-PAD). Separation of monosaccharides was 160 achieved on a Dionex ICS-5000 system (Thermo Fisher Scientific, Waltham, MA, USA) fitted 161 with an Eluent Generator and a Dionex CarboPac SA10 column (250 x 4 mm; Thermo Fisher 162 Scientific) with an SA-10 guard column (50 x 4 mm). The operating temperature of the 163 column was 45 °C with an isocratic mobile phase of 1 mmol L-1 KOH pumped at a rate of 1.5 164 ml min-1. Monosaccharides were identified by comparison of retention times with glucose 165 and galactose standards. 166 167 2.6 Antioxidant activity of selected metabolites 7 168 The predominant metabolites were purified by collecting the HPLC eluent of individual 169 peaks from the Waters system described in section 2.4 and then dried down and 170 resuspended in 50% methanol at a concentration of 100 µg mL-1. The antioxidant activity of 171 individual compounds was measured in triplicate by 1,1-diphenyl-2-picrylhydrazyl (DPPH) 172 radical scavenging assay (Brand-Williams, Cuvelier & Berset, 1995) and by ferric ion reducing 173 antioxidant power (FRAP) assay (Benzie & Strain, 1996). 174 For the DPPH assay, 0.5 mL of DPPH solution (140 µmol L-1 in 50% methanol) was 175 combined with an equal volume of sample at a concentration of 5-100 μg mL-1 and left to 176 stand at room temperature for 30 min. After this time absorbance (A) was measured on a 177 Pharmacia Biotech Ultraspec 4000 UV/Visible spectrophotometer (Amersham Pharmacia 178 Biotech, Little Chalfont, UK) at 517 nm, including 50% methanol as control, and % inhibition 179 (I) was determined as 180 181 182 I = (A - Ablank)control – (A - Ablank)sample / (A - Ablank)control x 100 with IC50 defined as the concentration which resulted in 50% inhibition. The FRAP assay measures reduction of the Fe3+ / tripyridyl-s-triazine complex (TPTZ) to 183 the blue ferrous form and was performed as follows: acetate buffer (300 mmol L-1, pH 3.6), 184 TPTZ (10 mmol L-1 in 40 mmol L-1 HCl) and ferric chloride (20 mmol L-1) were mixed in the 185 ratio of 10 : 1 : 1 to obtain the FRAP reagent. Sample volumes of 125 µL at a concentration 186 of 100 μg mL-1 were added to 1 ml FRAP reagent and after 10 min absorbance was 187 measured at 593 nm. A ferric chloride calibration curve ranging from 100 to 1000 mol L-1 188 was prepared to estimate Fe2+ concentration. 189 190 2.7 Statistical analysis 8 191 Quantitative analyses were carried out in triplicate and results are presented as mean ± 192 standard error of the mean. Where appropriate results were statistically analysed by one- 193 way ANOVA and significant differences were determined by the Bonferroni post hoc test, 194 GenStat 14th edition. 195 196 3. Results and discussion 197 3.1 Phenolic profile 198 Figure 1 shows the profile of soluble polyphenols extracted from Allanblackia floribunda 199 seeds which showed little variation between years. Analysis by HPLC-PDA-ESI/MSn revealed 200 six abundant and a range of minor compounds which were tentatively identified based on 201 their UV/vis absorbance and mass spectral characteristics summarised in tables 1 and 2 and 202 figure 2. The compounds which were identified all belonged to the flavonoids, with 203 biflavonoids as the main class. The main compounds were collected individually from peaks 204 1 to 6 (figure 1) and purified as pale yellow to orange crystalline powders. Molecular 205 weights (Mr) of the main compounds are based on accurate mass measurements in both 206 negative and positive ionisation mode, whilst Mr of less abundant compounds are nominal 207 mass based on general MS scans in both ionisation modes. MSn analyses were carried out in 208 negative ionisation mode unless stated otherwise. 209 210 3.1.1 Identification of biflavonoids 211 Based on accurate mass data (table 2) compound 5 was assigned the elemental 212 composition C30H20O11 with Mr 556.1005 Da. The corresponding UV/vis spectrum showed 213 absorbance maxima at 275(sh), 290 and 348 nm, a combination of the absorbance spectrum 214 typical of a flavone (with two major absorption bands in the region of 300-380 nm for band I 9 215 and 240-280 nm for band II; Mabry, Markham & Thomas, 1970) with that of a flavanone 216 (with predominant absorption band II and a much reduced band I) and in close agreement 217 with the absorbance characteristics reported for the biflavonoid morelloflavone, a 218 naringenin-luteolin conjugate (Herbin, Jackson, Locksley, Scheinmann & Wolstenholme, 219 1970). MS2 fragmentation of the parent ion at m/z 555 in negative mode resulted in loss of 220 126 Da, yielding one main product ion with m/z 429. This can be attributed to cleavage of 221 ring C (figure 3a), typical of the fragmentation of flavonoids (Cuyckens & Claeys, 2004; 222 Fabre, Rustan, de Hoffmann & Quetin-Leclercq, 2001), here breaking bonds 1 and 4 and 223 producing fragment ion [M-H-1,4A]- with the remainder of the upper moiety of the dimer still 224 attached to the lower moiety. A low intensity fragment with m/z 403 was also observed. 225 This corresponds to loss of 152 Da and stems from cleavage of bonds 1 and 3 of ring C. 226 Fragmentation of naringenin standard in negative mode showed cleavage of corresponding 227 bonds, yielding ion 1,3A- with m/z 151 as base peak and ion 1,3B - with m/z 119 among the less 228 intense fragment ions. In the dimer, ring B is still attached to the lower moiety, thus 229 resulting in fragment [M-H-1,3A]-. The minor fragments also included ions at m/z 449 (loss of 230 106 Da) and m/z 461 (loss of 94 Da, [M-H-B]-), and corresponding losses were also seen 231 when fragmenting naringenin standard. 232 With normalised collision energy (CE) set to default (35%), MS3 fragmentation of the 233 predominant MS2 ion at m/z 429 occurred only to a small extent, an observation also made 234 with luteolin standard. Raising CE to 70% resulted in the neutral loss of a number of small 235 molecules such as CO, CO2 and H2O, also typical of flavonoid fragmentation (Cuyckens & 236 Claeys, 2004; Fabre et al., 2001). A further MS3 product ion with m/z 295 (loss of 134 Da) 237 can be accounted for by cleavage of bonds 1 and 3 of the heterocycle of the luteolin moiety, 238 producing [M-H-1,4A-1,3E]-. Cleavage of corresponding bonds was also observed in luteolin 10 239 standard, resulting in a fragment with m/z 151. Overall, fragmentation of the flavonoid 240 dimer mirrored a combination of the fragmentation patterns observed for naringenin and 241 luteolin standards. The data presented here is in close agreement with mass spectral data 242 reported by Carrillo-Hormaza et al. (2016) and supports the tentative identification of 243 compound 5 as morelloflavone (naringenin-I,3-II,8-luteolin). 244 The UV/vis spectrum of peak 6 showed absorbance maxima at 274, 290 and 328 nm 245 (figure 2), similar to those of peak 5 but with absorbance band I at a lower wavelength, 246 consistent with the absorbance characteristics reported for volkensiflavone, a naringenin- 247 apigenin conjugate (Herbin et al., 1970). Accurate mass measurements (table 2) suggest the 248 elemental composition C30H20O10 with Mr 540.1056 Da for this compound. The 249 fragmentation pattern in negative mode was also similar to that of morelloflavone (table 1, 250 figure 3a) and yielded product ions consistent with a naringenin moiety, namely [M-H-1,4A]- 251 with m/z 413, [M-H-B]- with m/z 445, [M-H-106]- with m/z 433 and [M-H-1,3A]- with m/z 387, 252 whereas fragmentation of the MS2 base peak at m/z 413 with raised CE resulted in losses 253 similar to those observed with apigenin standard, yielding product ions at m/z 295 (loss of 254 118 Da, [M-H-1,4A-1,3E]-) and 293 (loss of 120 Da). The findings for compound 6 are in close 255 agreement with the fragmentation pattern reported for volkensiflavone (naringenin-I,3-II,8- 256 apigenin) by Carrillo-Hormaza et al. (2016). 257 The UV/vis spectrum of peak 3 (figure 2) showed a band II absorbance maximum at 293 258 nm with band I reduced to a small shoulder, a characteristic of flavanones (Mabry, Markham 259 & Thomas, 1970). Based on accurate mass measurements (table 2) the molecular formula 260 C30H22O11 with Mr 558.1162 Da was assigned to this compound. Fragmentation in negative 261 ionisation mode followed a pattern very similar to that described above (table 1, figure 3b). 262 MS2 analysis of the parent ion at m/z 557 yielded a predominant product with m/z 431 (loss 11 263 of 126 Da, [M-H-1,4A]-) as well as low abundance ions consistent with the loss of 152 Da (at 264 m/z 405), 106 Da (at m/z 451) and 94 Da (at m/z 463) from a naringenin moiety. MS3 265 analysis of the fragment at m/z 431 produced a base peak with m/z 295 (loss of 136 Da, [M- 266 H-1,4A-1,3E]-) and less intense ions at m/z 269 (loss of 162 Da, [M-H-1,4A-1,4E]-; Fabre et al., 267 2001) and 321 (loss of 110 Da, [M-H-1,4A-E]-), in agreement with the fragmentation pattern 268 observed for eriodictyol standard. Based on the data presented here compound 3 was 269 tentatively identified as Gb-2a (naringenin-I,3-II,8-eriodictyol), a conclusion supported by 270 the mass spectral data of Carrillo-Hormaza et al. (2016). 271 Peaks 1, 2 and 4 had UV/vis spectra which were almost identical to those of peaks 3, 5 272 and 6, respectively (figure 2) but with Mr 162 Da bigger than the corresponding biflavonoids 273 (tables 1 and 2), and they were also more polar as shown by the earlier retention times. The 274 MS2 spectra of the compound 1, 2 and 4 parent ions included products [M-H-162]-, 275 indicating the loss of an O-linked hexose (Vukics & Guttmann, 2010), while other product 276 ions resulted from fragmentation of the biflavonoid core with or without simultaneous loss 277 of the sugar moiety. In addition, MSn fragmentation of the presumed aglycone fragments at 278 m/z 557, 555 and 539, respectively, followed the common pattern seen for compounds 3, 5 279 and 6, thus confirming biflavonoids as core molecules of the glycosides, with neutral loss of 280 126 Da producing the diagnostic fragments [M-H-162-1,4A]-, while subsequent MSn analysis 281 included product ions [M-H-162-1,4A -1,3E]- with m/z 295. 282 HPAEC analysis of the purified, hydrolysed compounds revealed glucose as the sugar 283 moiety of all three glycosides. Although it is not possible to determine the precise position 284 of the glucose part with the methods employed here, fragmentation data for compounds 1 285 and 2, which have different substitution patterns on rings B and E, suggests that the glucose 286 may be attached to one of the hydroxyl groups of the lower flavonoid unit. In the MS2 12 287 spectra of both compounds there were product ions characteristic for the fragmentation of 288 the naringenin moiety with the sugar unit still attached, namely [M-H-1,4A]- (loss of 126 Da), 289 [M-H-1,3A]- (loss of 152 Da) and low intensity ions indicating the loss of ring B ([M-H-B]-, loss 290 of 94 Da). 291 Analysis of the smaller peaks revealed further compounds of this type. In particular, a 292 compound with Mr 542 Da eluting at 41.6 min followed a similar fragmentation pattern as 293 described above (figure 3b) and was tentatively identified as dinaringenin, and a hexoside of 294 this with Mr 704 Da was found at tR 36.4 min. A further biflavonoid with Mr 574 Da eluted at 295 tR 30.9 min and had a fragmentation pattern consistent with naringenin as upper and an 296 unidentified flavonoid with Mr 304 Da as lower unit, and a hexoside of this with Mr 736 was 297 seen at 21.8 min. In addition to morelloflavone and morelloflavone-glucoside, a compound 298 with Mr 880 Da eluting at 27.5 min showed a fragmentation pattern consistent with 299 morelloflavone-dihexoside. Some larger related molecules were also detected amongst the 300 low abundance compounds. For example, fragmentation of a compound with Mr 1006 Da at 301 tR 28.4 min produced a fragment ion with m/z 717. Further MS3 analysis of this fragment 302 was consistent with morelloflavone-hexoside, and the compound was tentatively identified 303 as morelloflavone conjugate. Similarly, MS2 analysis of two compounds with Mr 1008, 304 eluting at 25.6 and 30.2 min, produced fragments with m/z 719, and MS3 analysis of this ion 305 product was consistent with the fragmentation pattern of Gb-2a-hexoside. 306 Both morelloflavone and volkensiflavone have previously been reported in A. floribunda 307 heartwood (Locksley & Murray, 1971) and stem bark (Brusotti et al., 2016) as well as in 308 various tissues of related genera such as fruits, leaves and heartwood of Garcinia (Herbin et 309 al., 1970; Yang et al., 2010; Stark, Lösch, Wakamatsu, Balemba, Frank & Hofmann, 2015; 310 Carrillo-Hormaza et al., 2016). Whilst Gb-2a is also well documented in Clusiaceae (Herbin et 13 311 al., 1970; Stark et al., 2015; Carrillo-Hormaza et al., 2016), this is to our knowledge the first 312 report of Gb-2a in Allanblackia. 313 314 3.1.2 Identification of other flavonoids 315 In addition to the biflavonoids discussed in section 3.1.1, a range of low abundance 316 compounds from different flavonoid classes were also detected (table 1). Epicatechin and 317 luteolin were identified by direct comparison with analytical standards. Due to the absence 318 of catechin in the extract, the flavan-3-ol oligomers found at several retention times were 319 assumed to consist of epicatechin units. There was also a range of flavone and flavonol 320 glycosides which were tentatively identified by their fragmentation patterns following the 321 principles outlined by Vuciks & Guttman (2010) and Ferreres, Gil-Izquierdo, Andrade, 322 Valentão & Tomás-Barberán (2007), whilst their flavonoid cores were confirmed by MSn 323 analyses and acid hydrolysis. Interestingly, A. floribunda seed extract did not contain 324 detectable amounts of xanthones or benzophenones which were previously reported in 325 other tissues of A. floribunda and other species of the Clusiaceae family (Locksley & Murray, 326 1971; Fuller et al., 1999; Nkengfack et al., 2002; Azebaze et al., 2009; Yang et al., 2010). 327 328 3.2 Characterisation of biflavonoids 329 3.2.1 Quantification 330 The six main biflavonoids were quantitated in µg mg-1 seed powder using response 331 factors based on HPLC standard curves obtained with analytical standards as described in 332 section 2.4. The total biflavonoid content was high, constituting approximately 10% of dried 333 seed powder in comparison to seeds of the related species Garcinia madruno where 334 Carrillo-Hormaza et al. (2016) reported a total biflavonoid content of less than 2%. 14 335 Morelloflavone and its glucoside were the main biflavonoids present in A. floribunda seeds 336 (table 2). Similarly, Locksley & Murray (1971) reported morelloflavone as the predominant 337 metabolite in A. floribunda heartwood. 338 339 3.2.2 Antioxidant activity 340 The antioxidant activities of the six main biflavonoids were analysed by the DPPH-radical 341 scavenging and the ferric reducing antioxidant power (FRAP) assays. All six compounds 342 demonstrated antioxidant activity which differed significantly (P≤0.05) and showed a similar 343 ranking of activity with both assays apart from Gb-2a aglycone and volkensiflavone- 344 glucoside whose order was reversed with the DPPH compared with the FRAP assay (table 3). 345 The highest activities were observed with Gb-2a and morelloflavone-glucosides and the 346 lowest activities with volkensiflavone aglycone, volkensiflavone-glucoside and Gb-2a 347 aglycone. Morelloflavone-glucoside and aglycone and GB-2a-glucoside showed radical 348 scavenging activities comparable with ascorbic acid (IC50 values 16.87, 21.26, 18.38 and 349 19.36 µg ml-1, respectively). Radical scavenging and FRAP activities observed with 350 morelloflavone aglycone and glucoside also reflect results reported by Kuete et al. (2011) 351 and Carillo-Hormaza et al. (2016) where the aglycone showed lower activity than the 352 glucoside. This pattern was observed with all biflavonoids tested here, indicating that 353 glycosylation enhances antioxidant activity. The higher IC50 values reported by Kuete and co- 354 workers may be due to the higher DPPH concentration used in their assays. 355 Burda & Oleszek (2001) observed a relationship between certain structural features of 356 flavonoids and antioxidant behaviour. Their studies demonstrated that a hydroxyl group in 357 the para position on the B-ring is essential for activity and that this activity is enhanced by a 358 second hydroxyl on the B-ring in the ortho position and a double bond between C2 and C3 15 359 on the C ring. These findings are consistent with aglycone activities observed in the current 360 study. Overall, the data presented here demonstrates that Allanblackia seeds are an 361 abundant source of highly active antioxidant phenolic components in common with seeds of 362 the fruits of other tropical species including jabuticaba (Myrciaria cauliflora; Hacke et al., 363 2016) and guaraná (Paullinia cupana; Majhenič, Škerget & Knez, 2007; Marques et al. 2016). 364 365 4. Conclusion 366 The comprehensive profile of soluble phenolics presented here confirms biflavonoids as 367 the main phenolic compound class present and as major constituents of A. floribunda seeds. 368 Due to their antioxidant activity and reported therapeutic properties, these compounds are 369 of increasing interest to the pharmaceutical industry. With oil from A. floribunda seeds 370 attracting attention from the food sector as a potential ingredient of margarine and other 371 products (Cernansky, 2015), they could provide an excellent source of biflavonoids as a by- 372 product. 373 374 375 376 Acknowledgements The authors would like to thank Manfred Beckman for assistance with accurate mass measurements and Sarah Spicer for support with carbohydrate analysis. 377 This work was supported by funding from the European Regional Development Fund 378 through funding provided for the BEACON project by the Welsh European Funding Office. 379 380 The authors declare that they have no conflict of interest. 381 382 References 16 383 384 385 Ajibesin, K. K., Rene, N., Bala, D. N., & Essiett, U. A. (2008). Antimicrobial activities of the extracts and fractions of Allanblackia floribunda. Biotechnology, 7, 129-133. Ayoola, G. A., Akpanika, G. A., Awobajo, F. O., Sofidiya, M. O., Osunkalu, V. O., Coker, H. A. B. 386 & Odugbemi, T. O. (2009). Anti-inflammatory properties of the fruits of Allanblanckia 387 floribunda Oliv. (Guttiferae). Botany Research International, 2, 21-26. 388 Ayoola, G. A., Ipav, S. S., Sofidiya, M. O., Adepoju-Bello, A. A., Coker, H. A. B. & Odugbemi, T. 389 O. (2008). Phytochemical screening and free radical scavenging activities of the fruits and 390 leaves of Allanblackia floribunda Oliv (Guttiferae). International Journal of Health 391 Research, 1, 87-93. 392 Azebaze, A. G. B., Menasria, F., Noumi, L. G., Nguemfo, E. L., Tchamfo, M. F., Nkengfack, A. 393 E. Kolb, J. P. & Meyer, M. (2009). Xanthones from the seeds of Allanblackia monticola and 394 their apoptotic and antiproliferative activities. Planta medica, 75, 243-248. 395 Azebaze, A. G. B., Teinkela, J. E. M., Nguemfo, E. L., Valentin, A., Dongmo, A. B. & 396 Vardamides, J. C. (2015). Antiplasmodial activity of some phenolic compounds from 397 Cameroonians Allanblackia. African Health Sciences, 15, 835-840. 398 399 400 Benzie, I. F. F. & Strain, J. J. (1996). The ferric reducing ability of plasma (FRAP) as a measure of “antioxidant power”: the FRAP assay. Analytical biochemistry, 239, 70-76. Betti, J. L. (2004). An ethnobotanical study of medicinal plants among the Baka pygmies in 401 the Dja biosphere reserve, Cameroon. African Study Monographs, 25, 1-27. 402 Boudjeko, T., Ngomoyogoli, J. E.K., Woguia, A. L. & Yanou, N. N. (2013). Partial 403 characterization, antioxidative properties and hypolipidemic effects of oilseed cake of 404 Allanblackia floribunda and Jatropha curcas. BMC complementary and alternative 405 medicine, 13, 352-360. 17 406 407 408 Brand-Williams, W., Cuvelier, M. E., & Berset, C. (1995). Use of a free radical method to evaluate antioxidant activity. LWT - Food Science and Technology, 28, 25-30. Brusotti, G., Papetti, A., Serra, M., Temporini, C., Marini, E., Orlandini, S., Kada Sanda, A., 409 Watcho, P. & Kamtchouing, P. (2016). Allanblackia floribunda Oliv.: An aphrodisiac plant 410 with vasorelaxant properties. Journal of Ethnopharmacology, 192, 480-485. 411 412 Burda, S. & Oleszek, W. (2001). Antioxidant and antiradical activities of flavonoids. Journal of Agricultural and Food Chemistry, 49, 2774-2779. 413 Carrillo-Hormaza, L., Ramírez, A. M., Quintero-Ortiz, C., Cossio, M., Medina, S., Ferreres, F., 414 Gil-Izquierdo, A. & Osorio, E. (2016). Comprehensive characterization and antioxidant 415 activities of the main biflavonoids of Garcinia madruno: A novel tropical species for 416 developing functional products. Journal of Functional Foods, 27, 503-516. 417 Cernansky, R. (2015). The African tree that could help replace palm oil. URL 418 http://www.takepart.com/article/2015/12/18/african-tree-could-help-replace-palm-oil. 419 Accessed 19.07.2016. 420 421 Cuyckens, F. & Claeys, M. (2004). Mass spectrometry in the structural analysis of flavonoids. Journal of Mass Spectrometry, 39, 1-15. 422 Fabre, N., Rustan, I., de Hoffmann, E. & Quetin-Leclercq, J. (2001). Determination of flavone, 423 flavonol, and flavanone aglycones by negative ion liquid chromatography electrospray 424 ion trap mass spectrometry. Journal of the American Society for Mass Spectrometry, 12, 425 707-715. 426 Fadeyi, S. A., Fadeyi, O. O., Adejumo, A. A., Okoro, C. & Myles, E. L. (2013). In vitro 427 anticancer screening of 24 locally used Nigerian medicinal plants. BMC Complementary 428 and Alternative Medicine, 13, 79-87 18 429 Ferreres, F., Gil-Izquierdo, A., Andrade, P. B., Valentão, P. & Tomás-Barberán, F. A. (2007). 430 Characterization of C-glycoslyl flavones O-glycosylated by liquid chromatography-tandem 431 mass spectrometry. Journal of Chromatography A, 1161, 214-223. 432 Fuller, R. W., Blunt, J. W., Boswell, J. L., Cardellina ΙΙ, J. H. & Boyd, M. R. (1999). Guttiferone 433 F, the first prenylated benzophenone from Allanblackia stuhlmannii. Journal of Natural 434 Products, 62, 130-132. 435 Hacke, A. C. M., Granato, D., Maciel, L. G., Weinert, P. L., do Prado-Silva, L., Alvarenga, V. O., 436 de Souza Sant’Ana, A., Bataglion, G. A., Eberlin, M. N. & Rosso, N. D. (2016). Jabuticaba 437 (Myrciaria cauliflora) Seeds: Chemical Characterization and Extraction of Antioxidant and 438 Antimicrobial Compounds. Journal of Food Science, 81, C2206-C2217. 439 Hauck, B., Gallagher, J. A., Morris, S. M., Leemans, D. & Winters, A. L. (2014). Soluble 440 phenolic compounds in fresh and ensiled Orchard grass (Dactylis glomerata L.), a 441 common species in permanent pastures with potential as a biomass feedstock. Journal of 442 Agricultural and Food Chemistry, 62, 468-475. 443 444 445 Herbin, G. A., Jackson, B., Locksley, H. D., Scheinmann, F. & Wolstenholme, W. A. (1970). The biflavonoids of Garcinia volkensii (Guttiferae). Phytochemistry, 9, 221-226. IUCN (International Union for Conservation of Nature, 2014). Allanblackia oil enters market 446 in new margarine. URL 447 https://www.iucn.org/news_homepage/news_by_date/?18536/Allanblackia-Oil-Enters- 448 Market-in-New-Margarine. Accessed 19.07.2016. 449 450 451 452 Kayode, J. (2006). Conservation of indigenous medicinal botanicals in Ekiti State, Nigeria. Journal of Zhejiang Universitv Science B, 7, 713-718. Kuete, V., Azebaze, A. G. B., Mbaveng, A. T., Nguemfo, E. L., Tshikalange, E. T., Chalard, P. & Nkengfack, A. E. (2011). Antioxidant, antitumor and antimicrobial activities of the crude 19 453 extract and compounds of the root bark of Allanblackia floribunda. Pharmaceutical 454 Biology, 49, 57-65. 455 Locksley, H. D. & Murray, I. G. (1971). Extractives from Guttiferae. Part XIX. The isolation and 456 structure of two benzophenones, six xanthones and two biflavonoids from the 457 heartwood of Allanblackia floribunda Oliver. Journal of the Chemical Society C: Organic, 458 1332-1340. 459 460 461 462 463 Mabry, T. J., Markham, K. R. & Thomas, M. B. (1970). The systematic identification of flavonoids. Springer Verlag, New York. Majhenič, L., Škerget, M., Knez, Ž. (2007). Antioxidant and antimicrobial activity of guarana seed extracts. Food Chemistry, 104, 1258-1268. Marques, L. L. M., Panizzon, G. P., Aguiar, B. A. A., Simionato, A. S., Cardozo-Filho, 464 L., Andrade, G., de Oliveira, A. G., Guedes, T. A. & de Mello, J. C. P. (2016). Guaraná 465 (Paullinia cupana) seeds: Selective supercritical extraction of phenolic compounds. Food 466 Chemistry, 212, 703-711. 467 468 469 Nkengfack, A. E., Azebaze, G. A., Vardamides, J. C., Fomum, Z. T. & van Heerden, F. R. (2002). A prenylated xanthone from Allanblackia floribunda. Phytochemistry, 60, 381-384. Olowokudejo, J. D., Kadiri, A. B. & Travih, V. A. (2008). An ethnobotanical survey of herbal 470 markets and medicinal plants in Lagos state of Nigeria. Ethnobotanical Leaflets, 12, 851- 471 865. 472 Orwa, C. & Munjuga, M. (2007). Allanblackia floribunda Oliv. Internet Record from 473 PROTA4U. van der Vossen, H.A.M. & Mkamilo, G.S. (editors). PROTA (Plant Resources of 474 Tropical Africa / Ressources végétales de l’Afrique tropicale), Wageningen, Netherlands. 475 URL http://www.prota4u.org/search.asp or http://uses.plantnet- 476 project.org/en/Allanblackia_floribunda_(PROTA). Accessed 19.07.2016. 20 477 Orwa C., Mutua A., Kindt R., Jamnadass R. & Anthony S. (2009). Agroforestree Database: a 478 tree reference and selection guide version 4.0. URL 479 http://www.worldagroforestry.org/treedb/AFTPDFS/Allanblackia_floribunda.PDF. 480 Accessed 19.07.2016. 481 Stark, T. D., Lösch, S., Wakamatsu, J., Balemba, O. B., Frank, O. & Hofmann, T. (2015). UPLC- 482 ESI-TOF MS-based metabolite profiling of the antioxidative food supplement Garcinia 483 buchananii. Journal of Agricultural and Food Chemistry, 63, 7169-7179. 484 485 486 Vukics, V. & Guttman, A. (2010). Structural characterization of flavonoid glycosides by multistage mass spectrometry. Mass Spectrometry Reviews, 29, 1-16. Wilfred, S., Adubofuor, J. & Oldham, J. H. (2010). Optimum conditions for expression of oil 487 from Allanblackia floribunda seeds and assessing the quality and stability of pressed and 488 solvent extracted oil. African Journal of Food Science, 4, 563-570. 489 Yang, H., Figueroa, M., To, S., Baggett, S., Jiang, B., Basile, M. J., Weinstein, I. B. & Kennelly, 490 E. J. (2010). Benzophenones and biflavonoids from Garcinia livingstonei fruits. Journal of 491 Agricultural and Food Chemistry, 58, 4749-4755. 492 21 493 494 495 496 497 498 499 500 501 502 503 504 505 Figure 1. HPLC chromatogram of soluble secondary metabolites from Allanblackia floribunda seeds detected by photodiode array at 280nm. Peak numbering refers to the six predominant compounds. Figure 2. UV/vis spectra of biflavonoids from Allanblackia floribunda seeds. Figure 3. Structures and fragmentation patterns of (a) flavanone-flavone dimers (volkensiflavone, morelloflavone) and (b) flavanone-flavanone dimers (Gb-2a, dinaringenin) extracted from Allanblackia floribunda seeds. 506 507 508 22 2 RT: 0.00 - 60.00 100 100 NL: 4.05E5 nm=279.5280.5 PDA seed1_041 6 95 90 85 Relative absorbance (%) 80 80 75 70 65 60 60 1 55 5 50 45 40 40 35 30 25 20 20 3 15 4 6 10 5 00 0 0 5 10 10 15 20 20 25 30 30 35 40 40 45 Time (min) 509 510 23 50 50 55 60 60 peak 5 peak 2 peak 6 peak 4 peak 3 peak 1 240.0 280.0 320.0 360.0 240.0 280.0 320.0 360.0 400.0 nm 511 512 24 513 -1,3A a. -1,4A B 0 I A 1 C 2 -B 4 3 E 0 D F 4 2 3 -1,3E upper moiety: naringenin lower moiety: apigenin (R=H) or luteolin (R=OH) b. II 1 -1,3A -1,4A B A C -B E D -E F upper moiety: naringenin lower moiety: naringenin (R=H) or eriodictyol (R=OH) -1,3E -1,4E 514 515 25 516 517 518 519 520 521 522 523 524 525 526 527 528 Figure 1. HPLC chromatogram of soluble secondary metabolites from Allanblackia floribunda seeds detected by photodiode array at 280nm. Peak numbering refers to the six predominant compounds. Figure 2. UV/vis spectra of biflavonoids from Allanblackia floribunda seeds. Figure 3. Structures and fragmentation patterns of (a) flavanone-flavone dimers (volkensiflavone, morelloflavone) and (b) flavanone-flavanone dimers (Gb-2a, dinaringenin) extracted from Allanblackia floribunda seeds. 529 530 531 26 Table 1 Polyphenolic compounds detected in Allanblackia floribunda seed extract by HPLC-PDA-ESI/MSn 2 tR (min) 10.9 λmax (nm) 279 578 MS fragments in -ve mode unless stated otherwise (base peak in bold) 287, 289, 407, 425, 451, 559 13.7 279 290 179, 205, 245 14.4 279 866 407, 425, 451, 575, 577, 695, 713, 739 Mr 3 2 MS fragments of MS base peak (base peak in bold) n additional diagnostic MS fragments (parent ion in bracket) 179, 205, 245 (289) tentative ID (figure 1 peak no. in brackets) epicatechin dimer epicatechin 287, 289 (577) 1 epicatechin trimer 179, 205, 245 (289) 575, 577, 739, 863, 865, 983, 1001, 1027, 1135 14.9 279 1154 21.8 nd 736 429, 447, 573 22.8 nd 442 139, 143, 161, 179, 205, 217, 235, 330/331, 397 578 287, 289, 407, 425, 451, 559 23.6 nd 624 285, 327, 489 269, 385, 403, 419, 429 245, 287, 289 (577) epicatechin tetramer 419, 421, 429, 447, 467, 479 (573) hexoside of biflavonoid at tR 30.9 min unidentified compound 179, 205, 245 (289) epicatechin dimer luteolin-hexosyl-glucuronate 287, 463 (+ve mode) 24.9 nd 578 293, 311, 341, 413, 457 apigenin-2"-O-rhamnosyl-C-hexoside 25.6 nd 462 285 luteolin-glucuronate 26.8 nd 1008 608 431, 557, 719, 837, 845, 855, 881 431, 649, 675, 811, 837 431, 557, 593 (719) 269 Gb-2a conjugate apigenin-hexosyl-glucuronate 271, 447 (+ve mode) 27.5 nd 880 717 610 301 403, 429, 493, 537, 555, 565, 591 295 (429) morelloflavone-dihexoside quercetin-rhamnosyl-hexoside 303, 449, 465 (+ve mode) 28.4 nd 1006 29.6 nd 756 717, 809, 835, 853, 879 809, 835 429, 493, 537, 555, 565, 591, 623 (717) 285, 575, 593 morelloflavone conjugate luteolin-rhamnosyl-dihexoside 287, 449, 595, 611 (+ve mode) 30.2 nd 1008 431, 557, 719, 837, 845, 855, 881 431, 649, 675, 811, 837 431, 557, 593 (719) Gb-2a conjugate 30.9 nd 574 447 269, 295, 325, 403, 419, 429 421, 467, 479 (573) unidentified biflavonoid 31.1 288 720 431, 557, 593 269, 295, 413, 321 405, 431, 451, 463 (557) Gb-2a-glucoside (1) 31.3 nd 578 269 31.9 nd 594 625 (719) apigenin-rhamnosyl-hexoside 271, 433 (+ve mode) 285 kaempferol-rhamnosyl-hexoside 287, 449 (+ve mode) 594 (leading edge) 285 luteolin-rhamnosyl-hexoside 287, 449 (+ve mode) 34.1 274, 290, 351 718 429, 493, 537, 555, 565, 591, 623 (trailing edge) 403, 445 403, 449, 461 (555) morelloflavone-glucoside (2) 295 (429) 623 (717) 36.4 nd 704 389, 415, 447, 523, 541, 551, 577, 609 295, 321, 269 389, 415, 435, 447 (541) dinaringenin-hexoside 37.0 293 558 431 269, 295, 321, 413 405, 451, 463 (557) Gb-2a (3) 37.2 nd 610 463, 301 quercetin-rhamnosyl-hexoside 38.0 nd 286 151, 175, 197, 199, 201, 213, 217, 241, 243, 257, 267 luteolin 40.3 274, 290, 330 702 413, 433, 445, 521, 539, 575 41.2 nd 594 285, 447 41.6 nd 542 389, 415, 435, 447 269, 295, 309, 321 41.7 275(sh), 290, 348 556 429 295, 357, 385, 401 46.3 274(sh), 290, 328 540 387, 413, 433, 445 293, 295, 369, 385 nd: not detected, either due to low absorbance or coeluting compounds 1 : identified by direct comparison with reference compounds 387, 413, 433, 445 295 (413) 1 volkensiflavone-glucoside (4) kaempferol-rhamnosyl-hexoside dinaringenin 403, 449, 461 (555) morelloflavone (5) volkensiflavone (6) Table 2 Accurate mass and content of predominant biflavonoids in seeds of Allanblackia floribunda Figure 1 peak no. (compound ID) 1 (Gb-2a-glucoside) tR (min) 30.1 measured mass [M-H] 719.1615 calculated mass [M-H] 719.1618 mass difference (ppm) 0.42 measured mass + [M-H] 721.1774 calculated mass + [M-H] 721.1763 2 (morelloflavone-glucoside) 33.7 717.1459 3 (Gb-2a) 35.4 557.1087 717.1461 0.28 719.1616 557.1089 0.36 559.1238 4 (volkensiflavone-glucoside) 39.2 701.1514 701.1512 0.29 703.1666 703.1657 1.28 4.03 ± 0.297 5 (morelloflavone) 40.3 555.0929 555.0933 0.72 557.1079 557.1078 0.18 16.54 ± 0.935 541.1134 541.1129 0.92 2.02 ± 0.136 6 (volkensiflavone) 44.5 539.0984 539.0984 0.00 The content of individual compounds is expressed as mean ± standard error of the mean (n=9). mass difference (ppm) 1.53 content -1 (µg mg seed powder) 16.73 ± 0.906 719.1607 1.25 58.26 ± 3.872 559.1235 0.54 2.75 ± 0.146 Table 3 Antioxidant capacity of biflavonoids from A. floribunda seeds Figure 1 peak no. FRAP DPPH 2+ -1 -1 (compound ID) (µmol Fe mg ) IC50 (µg mL ) 1 (Gb-2a-glucoside) 13.67 ± 0.281 18.38 ± 0.112 2 (morelloflavone-glucoside) 15.35 ± 0.082 16.87 ± 0.337 3 (Gb-2a) 3.82 ± 0.350 27.20 ± 0.292 4 (volkensiflavone-glucoside) 2.61 ± 0.012 26.42 ± 0.006 5 (morelloflavone) 6.33 ± 0.016 21.26 ± 0.059 6 (volkensiflavone) 2.12 ± 0.272 33.92 ± 0.382 ascorbic acid 17.91 ± 0.243 19.36 ± 0.036 Results are expressed as mean ± standard error of the mean (n=3). 532 Highlights 533 • The phenolic profile of Allanblackia floribunda seeds was studied by LC-PDA-MSn. 534 • A. floribunda seeds contain approximately 10% biflavonoids on a dry weight basis. 535 • Morelloflavone, volkensiflavone and Gb-2a were the predominant biflavonoids. 536 • Antioxidant activity of the main biflavonoids was comparable with ascorbic acid. 537 27