Page 1
1
COUNTERCURRENT CHROMATOGRAPHY SEPARATION OF SAPONINS BY 1
SKELETON TYPE FROM Ampelozizyphus amazonicus FOR OFF-LINE UHPLC-2
HRMS ANALYSIS AND CHARACTERISATION. 3
4
Fabiana de Souza Figueiredoa, Rita Celano
b, Danila de Sousa Silva
c, Fernanda das 5
Neves Costaa, Peter Hewitson
d, Svetlana Ignatova
d, Anna Lisa Piccinelli
b, Luca 6
Rastrellib, Suzana Guimarães Leitão
c, Gilda Guimarães Leitão
a* 7
8 aUniversidade Federal do Rio de Janeiro, Instituto de Pesquisas de Produtos Naturais, 9
CCS, bloco H, Ilha do Fundão, 21941-590, RJ, Brazil 10 bUniversità di Salerno, Dipartimento di Farmacia, Via Giovanni Paolo II 132, 84084 11
Fisciano, Italy 12 cUniversidade Federal do Rio de Janeiro, Departamento de Produtos Naturais e 13
Alimentos, Faculdade de Farmácia, CCS, bloco A2, Ilha do Fundão, 21941-590, RJ, 14
Brazil 15 dAdvanced Bioprocessing Centre, Institute of Environment, Health & Societies, 16
CEDPS, Brunel University London, Middlesex UB8 3PH, UK 17
18
Corresponding author: *[email protected] 19
20
Keywords: Ampelozizyphus amazonicus, Rhamnaceae, saponin skeleton type, triterpene 21
saponins, countercurrent chromatography, HPLC-MS 22
23
24
Abstract 25
26
Ampelozizyphus amazonicus Ducke (Rhamnaceae), a medicinal plant used to prevent 27
malaria, is a climbing shrub, native to the Amazonian region, with jujubogenin 28
glycoside saponins as main compounds. The crude extract of this plant is too complex 29
for any kind of structural identification, and HPLC separation was not sufficient to 30
resolve this issue. Therefore, the aim of this work was to obtain saponin enriched 31
fractions from the bark ethanol extract by countercurrent chromatography (CCC) for 32
further isolation and identification/characterisation of the major saponins by HPLC and 33
MS. The butanol extract was fractionated by CCC with hexane - ethyl acetate - butanol - 34
ethanol - water (1:6:1:1:6; v/v) solvent system yielding 4 group fractions. The collected 35
fractions were analyzed by HPLC-HRMS and MSn. Group 1 presented mainly oleane 36
type saponins, and group 3 showed mainly jujubogenin glycosides, keto-dammarane 37
type triterpene saponins and saponins with C31 skeleton. Thus, CCC separated saponins 38
from the butanol-rich extract by skeleton type. A further purification of group 3 by CCC 39
(ethyl acetate - ethanol - water (1:0.2:1; v/v)) and HPLC-IR was performed in order to 40
obtain these unusual aglycones in pure form. 41
42
Page 2
2
1. Introduction 43
44
Ampelozizyphus amazonicus Ducke (Rhamnaceae) is a climbing shrub native to 45
the Amazonian region, where its barks and roots are used in the folk medicine to 46
prepare a beverage to cure and prevent malaria, as well as a tonic and fortifier [1, 2]. 47
The literature cites triterpenes (ursolic acid, betulinic acid, lupenone, betulin, lupeol, 48
melaleucic acid, 3β-hydroxylup-20(29)-ene-27,28-dioic acid, 2α,3β-dihydroxylup-49
20(29)-ene-27,28-dioic acid and 3β,27α-dihydroxylup-20(29)-en-28β-oic acid), 50
jujubogenin glycoside saponins (3-O-β-D-glucopyranosyl-20-O-α-L-51
rhamnopyranosyljujubogenin and 3-O-[β-D-glucopyranosyl(l2)α-L-52
arabinopyranosyl]-20-O-α-L-rhamnopyranosyljujubogenin) as well as, C30 and C31 53
dammarane-type triterpene saponins (ampelozigenin-l5α-O-acetyl-3-O-α-L-54
rhamnopyranopyranosyl-(12)-D-glucopyranoside) as main compounds in these 55
preparations [3-6]. Saponins are usually produced by plants as a complex mixture with 56
very similar structures and polarities. Each saponin is biosynthesized at low 57
concentration, which makes difficult their direct identification and isolation [7]. 58
Therefore, traditionally, multidimensional chromatography has been used, for example, 59
with column chromatography as the first dimension and countercurrent chromatography 60
(CCC) as the second dimension [8] or CCC as the first dimension and liquid 61
chromatography (LC) as the second one [9]. CCC is a type of liquid-liquid partition 62
chromatography technique with no solid support [10]. The use of liquid stationary phase 63
is advantageous for preparative separations because there is no irreversible adsorption, 64
and it allows a high sample loading, and good reproducibility with the scale-up [11, 12]. 65
For the structural characterisation of saponins mass spectrometry is the most often used 66
technique as it provides important information about skeleton type and number of 67
sugars but not sugar identity or linkage [7]. 68
Our previous studies revealed the presence of dammarane saponins with 69
jujubogenin and keto-dammarane skeletons in Ampelozizyphus amazonicus bark extract 70
[1, 2, 6]. A high complexity of the crude extract, due to the variability and similar 71
structures of saponins, however, did not allow a complete chromatographic separation 72
and identification of individual saponins by UHPLC-HRMS and MSn
[6]. Therefore, a 73
series of pre-purification steps were undertaken starting from consequent liquid-liquid 74
extraction (LLE) of the crude ethanol extract with hexane, ethyl acetate and butanol. 75
Analyses by TLC and HPLC-HRMS revealed the presence of saponins in the butanol 76
Page 3
3
extract (Figure 1A). The next step was CCC separation to produce less complex 77
samples for HPLC-HRMS characterisation and further isolation of saponins from 78
Ampelozizyphus amazonicus bark by CCC and semi-preparative HPLC. 79
80
Insert Figure 1 here 81
82
2. Materials and methods 83
84
2.1. Chemical reagents and solvents 85
86
Organic solvents used to prepare extracts were analytical grade from Tedia 87
(Tedia Brazil, Rio de Janeiro, Brazil). Organic solvents used for TLC analyses and CCC 88
separations were analytical grade from Fisher Chemicals (Loughborough, UK). MS-89
grade acetonitrile and water were supplied by Romil (Cambridge, UK). Organic 90
solvents used in HPLC-IR separations were HPLC grade from Sigma Aldrich (Milan, 91
Italy) and ultrapure water (18 MΩ) was prepared by a Milli-Q purification system 92
(Millipore, Bedford, USA). 93
94
2.2. Plant material and preparation of the extracts 95
96
The stem barks of A. amazonicus were collected in the Brazilian Amazon, at 97
“quilombola” communities of Oriximiná (State of Pará) [1, 2]. Plants were identified by 98
Mr. José Ramos (parataxonomist) and a voucher specimen, INPA 224161, was 99
deposited at the herbarium of Instituto Nacional de Pesquisas da Amazônia (INPA) 100
(Manaus, AM, Brazil) [1,2]. We received authorization for this study from the Brazilian 101
Directing Council for Genetic Heritage (Conselho de Gestão do Patrimonio Genético) 102
through Resolution n.213 (6.12.2007) published in the Federal Official Gazette of 103
|Brazil on December 27, 2007. 104
The stem barks were dried in a ventilated oven (Marconi, model MA037) and 105
ground in a hammer mill (Marconi, model MA340, serial 9304176). The powder 106
material of bark was extracted by percolation with ethanol. The extract was filtrated and 107
the ethanol was removed by rotary evaporation at 40 ºC under reduced pressure. Then 108
the bark ethanol extract (346.5 g) was sequentially partitioned by hexane/ water, ethyl 109
acetate/ water and butanol/ water in a separatory funnel. The solvents were removed by 110
Page 4
4
rotary evaporation. The liquid-liquid extraction afforded 0.2 g of hexane, 44.7 g of ethyl 111
acetate and 72.5 g of butanol partitions. 112
113
2.3. Countercurrent chromatography apparatus 114
115
Two high performance countercurrent chromatography (HPCCC) centrifuges 116
were used for CCC separations, a Spectrum (semi-preparative) and a MIDI 117
(preparative), both from Dynamic Extractions Ltd. (DE, Tredegar, UK). The Spectrum 118
was equipped with a polytetrafluorethylene (PTFE) column of 143.5 ml and 1.6 mm 119
tubing I.D. The MIDI had a PTFE column of 912.5 ml and 4.0 mm tubing I.D. All 120
separations were performed at the maximum rotation speed of both instruments, 1600 121
rpm (Revolution radius (R) = 85 mm) and 1400 rpm (R = 110 mm) respectively. The 122
semi-preparative set up had a HPLC pump Agilent HP1200 (Santa Clara, California, 123
USA) and a fraction collector Agilent HP1200 (Santa Clara, California, USA). The 124
preparative chromatographic system had a HPLC pump Knauer K-1800 (Berlin, 125
Germany) and a fraction collector Gilson FC202 (Villiers-le-Bel, France). 126
127
2.4. Thin layer chromatography 128
129
Analyses of A. amazonicus bark extracts, solvent systems and CCC fractions 130
were done by thin layer chromatography (TLC) with silica gel TLC Plates 60F254 131
(Merck Art. 05554, Darmstadt, Germany). The mobile phase used for TLC analyses 132
was butanol – acetic acid – water (8:0.5:1.5; v/v). To visualize the compounds spots, the 133
universal spray-reagent, H2SO4 in methanol (5%, v/v) with vanillin in methanol (1%, 134
v/v), and Komarovisky specific spray-reagent for saponins [3,4] with subsequent 135
heating at 105 ºC on a hot plate were used. 136
137
2.5. Solvent system tests 138
139
The solvent systems tests were performed as follows: small amounts of a sample 140
extract were dissolved in a test tube containing a two-phase solvent system. After 141
shaking and allowing compounds to partition between the two phases, equal aliquots of 142
each phase were spotted beside each other separately on silica gel TLC plates. 143
Distribution coefficients (KD) were determined visually. 144
Page 5
5
145
2.6. CCC separations 146
147
Solvent systems used in all separations by CCC were prepared in a separatory 148
funnel at room temperature. After the equilibrating, the two phases were separated and 149
degassed by sonication for 5 min. In each separation run, a CCC column was first filled 150
with the stationary phase, after set the rotation, the mobile phase was pumped in. 151
Samples were dissolved in equal volumes of each phase and were injected after the 152
hydrodynamic equilibrium inside the column was reached. The column temperature was 153
maintained at 30° C. 154
155
2.6.1. CCC fractionation of the butanol extract of A. amazonicus 156
157
The solvent system chosen for fractionation of the butanol extract of A. 158
amazonicus was hexane – ethyl acetate – butanol – ethanol – water (1:6:1:1:6; v/v). 159
160
Semi-preparative fractionation: 161
The fractionation was performed on the Spectrum machine with the organic 162
upper phase as stationary phase and the aqueous lower phase as mobile phase (reversed 163
phase mode). Fractions of 4 ml were collected during elution (72 fractions, 2 ml/min, 2 164
Vc) and extrusion (36 fractions, 20 ml/min, 1 Vc). The sample was injected using an 165
Upchurch low pressure injection port (Model V-450, with 1/16 in. fittings) and a loop of 166
7.2 ml. The sample concentration was 100 mg/ml. The stationary phase retention (Sf) 167
before sample injection was 62%. Fractions were analysed by TLC and HPLC-HRMS 168
and MSn analyses (Figure 2). 169
170
Preparative fractionation: 171
The preparative fractionation of the butanol extract of A. amazonicus was 172
performed on the MIDI machine. Fractions of 24 ml were collected during elution (76 173
fractions, 12 ml/min, 2 column volume (Vc)) and extrusion (38 fractions, 24 ml/min, 1 174
Vc). The sample was injected using an Upchurch low pressure injection port (Model V-175
450, with 1/16 in. fittings) and loops of 45 ml and 90 ml. The sample concentration was 176
100 mg/ml. The stationary phase retention (Sf) before injection was 67%. After TLC and 177
HPLC-HRMS and MSn analyses, fractions were combined in groups (Figure 2). 178
Page 6
6
179
2.6.2. CCC fractionation of group 3 from the butanol extract of A. amazonicus 180
181
Purification of group 3 (Frs. 81 – 101 from the first CCC butanol extract 182
fractionation) was done with ethyl acetate – ethanol – water (1:0.2:1; v/v). The aqueous 183
lower phase was used as stationary phase and the organic upper phase as the mobile 184
phase (normal phase mode). The semi-preparative purification of group 3 was first 185
performed on the Spectrum. Fractions of 4 ml were collected during elution (36 186
fractions, 1 ml/min, 1 Vc) and extrusion (36 fractions, 2 ml/min, 1 Vc). The sample was 187
injected using a loop of 3.66 ml. The sample concentration was 27.5 mg/ml. The 188
stationary phase retention (Sf) before injection was 77%. The preparative purification of 189
group 3 was performed on the MIDI machine. Fractions of 24 ml were collected during 190
elution (38 fractions, 12 ml/min, 1 Vc) and extrusion (38 fractions, 24 ml/min, 1 Vc). 191
The sample was injected using a loop of 45 ml. The sample concentration was 27.5 192
mg/ml. The stationary phase retention (Sf) before injection was 90%. After TLC and 193
HPLC-HRMS and MSn analyses, fractions were combined in groups (Figure 2). 194
195
2.7. HPLC separations 196
197
Fractions from Group 3 CCC separation were combined in different groups 198
(Figure 2). Groups C and D2 were separated further by semi-preparative HPLC-IR. The 199
column used was a HyPurity Aquastar, 150 x 10 mm; particle size 5µ (Thermo Electron 200
Corporation). The semi-preparative HPLC system was composed of a pump Knauer 201
Smartline 1000 (Labservice Analytica, Bologna, Italy) and a refraction index (RI) 202
detector Knauer Smartline 2300 (Labservice Analytica). The mobile phase used was 203
aqueous methanol, 5.9:4.1; v/v, in isocratic mode. For group C, the flow rate was 3.5 204
ml/min, the sample was dissolved in methanol (0.1 mg/µl) and the sample solution 205
injected in each run was 35 µl. For group D2, the flow rate was 2.5 ml/min, the sample 206
was dissolved in methanol (0.1 mg/µl) and the sample solution injected in each run was 207
50 µl. All fractions were analysed by HPLC-HRMS and MSn. 208
209
Insert Figure 2 here 210
211
2.8. UHPLC-HRMS analyses 212
Page 7
7
213
The butanol extract, fractions from CCC and HPLC-IR separations were 214
analysed on an LTQ OrbiTrap XL mass spectrometer (LTQ OrbiTrap XL, 215
ThermoFisher Scientific) connected to a Platin Blue UHPLC system (KNAUER GmbH, 216
Berlin, Germany). This UHPLC system was composed by two ultra-high-pressure 217
pumps, an auto sampler, a diode array detector and a column temperature manager. The 218
LC parameters used were: a Kinetex C18 column (2.1 x 50 mm, 1.7 µm; Phenomenex, 219
Bologna, Italy), flow rate of 0.5 mL min–1
, column temperature of 30 °C and, water (A) 220
and ACN (B), both containing 0.1% formic acid, as mobile phase. The gradient elution 221
program used was: 10-20% B in 3 min, 20–25% B in 4 min, 25–30% B in 13 min and 222
30–50% B in 5 min. After each injection, the column was washed with 98% B for 4 min 223
and re-equilibrated for 4 min. The mass spectrometer, with ESI source, was operated in 224
negative mode. High purity nitrogen (N2) was used as sheath gas (40 arbitrary units) and 225
auxiliary gas (arbitrary units). High purity helium (He) was used as collision gas. Mass 226
spectrometer parameters used were: 3.5 KV of source voltage, -37 V of capillary 227
voltage, –225 V of tube lens voltage and 280 ºC of capillary temperature. Full scan data 228
acquisition (mass range: m/z 350 – 2000) and data dependant MS scan were performed. 229
The resolution was 60000 and 7500 for the full mass and the data dependant MS scan, 230
respectively. The normalised collision energy of the collision-induced dissociation 231
(CID) cell was set at 35 eV and the isolation width of precursor ions was set at 2.0. 232
Saponins were characterized according to the corresponding spectral characteristics: 233
mass spectra, accurate mass, characteristic fragmentation, and retention time. Xcalibur 234
software (version 2.2) was used for instrument control, data acquisition and data 235
analysis. 236
237
3. Results and discussion 238
239
3.1. Butanol extract separation by CCC 240
241
Previous studies on separation of dammarane saponins by CCC used ethyl 242
acetate – butanol – water (1:1:2; v/v) and hexane – ethyl acetate – ethanol – water 243
(1:1:1:1; v/v) solvent systems [13-14]. Therefore, they were selected for preliminary 244
tests. In search for the best solvent system showing a good distribution of compounds 245
between the two phases (K visually close to 1), different solvent proportions were tested 246
Page 8
8
in order to change system’s polarity and polarity difference between phases. Some 247
solvents were added or replaced, in order to change the selectivity of systems [15]. 248
Table 1 lists all solvent systems, i-iv, tested for the butanol extract purification. The 249
distribution of compounds between the two phases in each solvent system was analysed 250
by TLC [16]. 251
In the first solvent system, (i) ethyl acetate – butanol – water (1:1:2; v/v), 252
compounds were practically all concentrated in upper phase and in the second system, 253
(ii) hexane – ethyl acetate – ethanol – water (1:1:1:1; v/v), compounds were 254
concentrated mainly in the lower phase. To increase polarity of the second system, ii, 255
the proportions of hexane and ethanol were changed to (iii) hexane – ethyl acetate – 256
ethanol – water (5:6:5:6; v/v), causing a drop in the sample solubility. To resolve this 257
issue and to increase polarity, other solvents such as acetone were added to a solvent 258
system, (iv) hexane - ethyl acetate - acetone - ethanol - water (1:1:0.5:1:1; v/v), and also 259
systems i and ii were combined , (v) hexane - ethyl acetate - butanol - ethanol - water 260
(6:6:1:6:6; v/v). In (iv), the sample has limited solubility and in (v) it was soluble and 261
compounds were more concentrated in lower phase like system (ii). After testing 262
various solvent ratios aiming to achieve a K visually close to 1, the best solvent system 263
appeared to be (vi) hexane - ethyl acetate - butanol - ethanol - water (1:6:1:1:6; v/v). 264
Every other CCC fractions were analysed by TLC and HPLC-MS before being 265
combined in groups (Figure 2) according to TLC profile and mass distribution. 266
267
Insert Table 1 here 268
269
UHPLC-HRMS analyses of CCC groups 1-4 and their fractions revealed the 270
presence of saponins mainly in groups 1 and 3. Although UHPLC-HRMS analyses 271
helped to characterize the groups from CCC separation of butanol extract, the groups 1 272
and 3 still showed a high complexity (Figure 1B and 1C). Thus, CCC fractionation of 273
the butanol extract was scaled-up in order to obtain a higher amount of each group 274
fraction for subsequent purification steps. 275
The scale-up factor (6.36) was calculated as the ratio between the column 276
volumes of Spectrum (143.5 ml) and MIDI (912.5 ml), according to CCC volumetric 277
scale-up [17,18]. This scale-up factor was used to adjust the flow rate and the sample 278
volume. Stationary phase retention (Sf) before injection were 62% and 67% 279
respectively. Based on Sutherland and co-workers (2005) theory, which stated that 280
Page 9
9
larger tubing bore provides a better stationary phase retention and therefore, larger 281
scale-up factors can be reached, the sample loading was increased by doubling the 282
sample volume. The reproducibility of runs was analysed by TLC. 283
284
3.2. UHPLC-HRMS analyses of butanol extract of CCC groups 285
286
The combination of high resolution mass spectrometry and MSn experiments 287
was employed to identify the main constituents of groups 1 and 3. 288
Group 1 showed a complex profile with two main metabolite classes (Figure 289
1B). The first consisted of polar phenolic compounds (0-5 min), while triterpene 290
glycosides, possibly with the oleane skeleton as aglycone, were inferred as the second 291
metabolite class (7-16 min) (data not shown). A complete elucidation of the group 1 292
saponin structures is currently in progress. 293
Dammarane saponins are the major constituents of the group 3 (Figure 1C). 294
Table 2 report the HRMS data of the main saponins of this CCC group and their 295
proposed molecular formulas. HRMS and MSn data (Table 2) allowed to identify 296
saponins with C30 and C31 keto-dammarane and jujubogenin skeletons (Figure 3), 297
according to our previous studies [2, 6], and dammarane-types saponins. Three saponins 298
(2−4) were tentatively identified as 16-keto-tetrahydroxydammar-23-ene triglycosides 299
(C30H50O5, Figure 3A), based on the presence in MS/MS spectra of the diagnostic 300
product ion [M−C8H14O2]− due to the loss of the side chains by a McLafferty 301
rearrangement [6,20]. Comparison with literature data suggested for the compounds 3 302
and 4 structures superimposable to those of hoduloside VIII and VII, respectively, 303
isolated from Rhamnaceae [21]. Also 5, 9-10, 12−13 and 16 produced a similar 304
fragmentation pathway of 2−4 yielded by McLafferty rearrangement. The dominant 305
product ions [M−C9H16O2]− correspond to the loss of an alkyl side chain with an 306
additional methylene group than to 16-keto-tetrahydroxydammar-23-ene glycosides. 307
Based on this fragmentation pathway and the occurrence of a C31 dammarane-type 308
saponin in A. amazonicus [4], the 16-keto-tetrahydroxydammar-24-methylene structure 309
(C31H52O5, Figure 3B) was proposed as aglycon of saponins 5, 9-10, 12−13 and 16. 310
This aglycone is not reported in the literature and further studies are needed to confirm 311
unambiguously the proposed structure. 312
Compounds 1, 7, 11, 14−15 and 17 were tentatively characterized as glycosides 313
of jujubogenin (C30H48O4, Figure 3C), primarily due to the presence of the product ion 314
Page 10
10
at m/z 471.3469 in MSn spectra, corresponding to the deprotonated jujubogenin 315
(C30H47O4). Compounds corresponding to the structure proposed for 1 and 17 were 316
previously reported in A. amazonicus [3], whereas the isomers 7 and 11 corresponded 317
presumably to hoduloside IV [22] and bacoside A3 [23], respectively, and the isomers 318
14 and 15 to bacopasaponin C [24]. In addition, the structure of hydroxymethylglutaryl 319
(HMG) jujubogenin glycosides [25] was established for 23 and 24 by the diagnostic 320
neutral loss of −144 Da and molecular formula of product ions [M−HMG]−. Finally, one 321
dammarane-types saponin, 6, and four acetylated derivatives, 20, 22, 25 and 28, were 322
detected in group 3. Their MS/MS spectra (Table 2) suggested the structure of acylated 323
tetrahydroxydammar-24-ene triglycosides, structurally related to ginseng saponins with 324
protopanaxatriol (C30H52O4) as aglycone [26, 27]. 325
The sugar residues of all identified saponins were established by characteristic 326
neutral losses (hexose −162 Da, deoxyhexose −146 Da, pentose −132 Da) and accurate 327
mass of corresponding product ions. Particularly, in the case of C30 (2−4) and C31 keto-328
dammarane saponins (5, 9-10, 12−13 and 16) the product ions at 479.3003 or 509.3109 329
in MSn spectra allowed to establish the nature of the sugar residue (pentose or hexose, 330
respectively) directly attached to the aglycone skeleton. 331
Other minor compounds (7, 8, 18, 19, 21, 26 and 27) detected in group 3 were 332
tentatively identified as saponins, but further studies are required to their detailed 333
characterization and identification of aglycones. 334
335
Insert Figure 3 here 336
Insert Table 2 here 337
338
3.3. Separation of Group 3 by CCC and HPLC-IR 339
340
UHPLC-HRMS analysis of butanol extract CCC groups showed the presence of 341
dammarane saponins only in the group 3. This saponin class is characteristic of A. 342
amazonicus [2-4] and it includes unusual aglycones as C30 and C31 keto-dammarane [4, 343
6, 20]. Thus, a further purification of group 3 by CCC and HPLC-IR was performed in 344
order to obtain as pure as possible these unusual compounds. 345
The same approach, as used for butanol extract, was applied to choose a suitable 346
solvent system for group 3 (Table 1; 1-9). Based on a previous work, where 347
dammarane saponins were isolated from Panax ginseng, the solvent system (1) 348
Page 11
11
dichloromethane - isopropanol – methanol – 5mM aqueous ammonium acetate (6:3:2:4; 349
v/v) was tested as a start [28]. Compounds were more concentrated in lower phase. 350
Further tests showed that in systems (2) ethyl acetate – butanol – methanol – water 351
(1:0.5:0.2:1; v/v) and (3) ethyl acetate – butanol – ethanol – water (1:0.5:0.2:1; v/v) 352
practically all compounds were in upper phase and any difference between system 353
selectivity was observed [29]. In (4) ethyl acetate – methanol – water (1:0.2:1; v/v) [30] 354
and (5) ethyl acetate – ethanol – water (1:0.2:1; v/v), was achieved a good distribution 355
of compounds between the two phases and a slight difference between compounds 356
selectivity. The system 5 showed visually slightly better selectivity than the system 4. 357
The replacement of ethanol with methanol in (6) ethyl acetate – isopropanol – methanol 358
– water (1:0.5:0.2:1; v/v) and (8) ethyl acetate – propanol – methanol – water 359
(1:0.5:0.2:1; v/v), provided a better distribution between two phases as in (7) ethyl 360
acetate – isopropanol – ethanol – water (1:0.5:0.2:1; v/v) and (9) ethyl acetate – 361
propanol – ethanol – water (1:0.5:0.2:1; v/v), because in ethanol containing systems 362
compounds were slightly more concentrated in upper phase due to ethanol polarity in 363
comparison with methanol. For the same reason, addition of propanol or isopropanol to 364
a solvent system reduced the its selectivity (Ks visually similar). Thus, the solvent 365
system selected for the purification of this group was (5) ethyl acetate – ethanol – water 366
(1:0.2:1; v/v). 367
Every two CCC fractions were analysed by TLC and HPLC-MS before being 368
combined in groups (Figure 2) according to TLC profile and mass distribution. 369
Betulinic acid was identified as main compounds of group A, while dammarane 370
saponins were detected in the other group 3 of CCC fractions. As shown in the 371
chromatograms reported in Figure 4, C, D1 and D2 groups were the most saponin-372
enriched groups. The main components of group C were the C31 dammarane-type 373
saponins 10, 13 and 16, whereas D groups were rich in C30 dammarane saponins 3-4 374
and 10, jujubogenin glycosides 11, 14, 15 and 17 and compound 9. 375
The fractionation of group 3 was also scaled-up to MIDI to obtain larger 376
amounts of enriched fractions for the successive purification by semi-preparative 377
HPLC-IR. Testing four different flow rates, 10, 12, 20 and 40 ml/min resulted in 378
stationary phase retention (Sf) of 86%, 90%, 81% and none, respectively. Therefore, 379
flow rate of 12 ml/min was selected for MIDI runs. The scale-up factor (6.36), applied 380
in CCC separation of butanol extract, was used to adjust the sample volume. 381
Page 12
12
In order to isolate the main dammarane saponins of A. amazonicus bark, 382
particularly saponins with C31 keto-dammarane-type skeleton, groups C and D2 from 383
the CCC purification of group 3 were selected for a subsequent purification by semi-384
preparative HPLC-IR. This isolation procedure allowed to obtain the jujubogenin 385
glycosides 1, 11 and 14−15, C31 dammarane saponins 9-10 and 13 and C30 dammarane 386
saponins 3 and 4, with a suitable purity grade (checked by NMR) for a detailed 387
characterization of their structures. 388
389
Insert Figure 4 here 390
391
4. Conclusions 392
393
The preparative purification procedure, based on CCC and HPLC-IR 394
separations, was successfully developed to isolate the main constituents of A. 395
amazonicus bark. The CCC reduced the complexity of butanol extract allowing a 396
characterization by HPLC-HRMS of saponins and allowed to isolate unusual C31 397
saponins by HPLC. CCC was able to separate saponins by skeleton type, mainly oleane 398
in group 1 and dammarane in group 3. The demonstrated scale-up methodology enables 399
more detailed chemical studies of compounds via future structure elucidation by NMR. 400
401
Acknowledgement 402
403
F.S. Figueiredo is indebted to Coordenação de Aperfeiçoamento de Pessoal de Nível 404
Superior (CAPES, Brazil) for the Ph.D scholarship. 405
F.N. Costa and S. Ignatova would like to thank Newton Advanced Fellowship project 406
funded by the Royal Society of the United Kingdom. 407
S.G. Leitão and G.G. Leitão are indebted to FAPERJ and CNPq for financial support. 408
The authors are deeply indebted to ARQMO (Associação de Comunidades 409
Remanescentes de Quilombolas do Município de Oriximiná), Oriximiná-PA, Brazil, for 410
supervising plant collection and for providing housing during the field trips. 411
412
413
414
Page 13
13
References 415
[1] D.R. de Oliveira, A.L.M.A. Costa, G.G. Leitão, N.G. Castro, J.P. dos Santos, 416
S.G. Leitão, Estudo etnofarmacognóstico da saracuramirá (Ampelozizyphus 417
amazonicus Ducke), uma planta medicinal usada por comunidades quilombolas 418
do Município de Oriximiná-PA, Brasil, Acta Amaz. 41 (2011) 383–392. 419
doi:10.1590/S0044-59672011000300008. 420
[2] L.M.T. Peçanha, P.D. Fernandes, T.J.M. Simen, D.R. De Oliveira, P.V. Finotelli, 421
M.V.A. Pereira, F.F. Barboza, T.D.S. Almeida, S. Carvalhal, A.P.T.R. Pierucci, 422
G.G. Leitão, L. Rastrelli, A.L. Piccinelli, S.G. Leitão, Immunobiologic and 423
antiinflammatory properties of a bark extract from Ampelozizyphus amazonicus 424
Ducke, Biomed Res. Int. 2013 (2013). doi:10.1155/2013/451679. 425
[3] M.G.L. Brandao, M.A. Lacaille-Dubois, M.A. Teixera, H. Wagner, Triterpene 426
saponins from the roots of Ampelozizyphus amazonicus, Phytochemistry. 31 427
(1992) 352–354. doi:10.1016/0031-9422(91)83076-W. 428
[4] M.G.L. Brandao, M.A. Lacaille-Dubois, M.A. Teixera, H. Wagner, A 429
dammarane-type saponin from the roots of Ampelozizyphus amazonicus, 430
Phytochemistry. 34 (1993) 1123–1127. 431
[5] L. V. Rosas, M.S.C. Cordeiro, F.R. Campos, S.K.R. Nascimento, A.H. Januário, 432
S.C. França, A. Nomizo, M.P.A. Toldo, S. Albuquerque, P.S. Pereira, In vitro 433
evaluation of the cytotoxic and trypanocidal activities of Ampelozizyphus 434
amazonicus (Rhamnaceae), Brazilian J. Med. Biol. Res. 40 (2007) 663–670. 435
doi:10.1590/S0100-879X2007000500009. 436
[6] T.J.M. Simen, P.V. Finotelli, F.F. Barboza, M.V.A. Pereira, A.P.T.R. Pierucci, 437
M.R.L. Moura, D.R. Oliveira, L.G. Abraçado, R. Celano, F.S. Figueiredo, A.L. 438
Piccinelli, L. Rastrelli, G.G. Leitão, L.M.T. Peçanha, S.G. Leitão, Spray-dried 439
extract from the Amazonian adaptogenic plant Ampelozizyphus amazonicus 440
Ducke (Saracura-mirá): chemical composition and immunomodulatory 441
properties, Food Res. Int. 90 (2016)100-110 doi: 442
10.1016/j.foodres.2016.10.040. 443
[7] W. Oleszek, Z. Bialy, Chromatographic determination of plant saponins-An 444
update (2002-2005), J. Chromatogr. A. 1112 (2006) 78–91. 445
doi:10.1016/j.chroma.2006.01.037. 446
[8] M. Kang, I.J. Ha, J. Chun, S.S. Kang, Y.S. Kim, Separation of two cytotoxic 447
saponins from the roots of Adenophora triphylla var. japonica by high-speed 448
counter-current chromatography, Phytochem. Anal. 24 (2013) 148–154. 449
doi:10.1002/pca.2394. 450
[9] D. Zhao, M. Yan, Y. Huang, X. Sun, Efficient protocol for isolation and 451
purification of different soyasaponins from soy hypocotyls, J. Sep. Sci. 35 (2012) 452
3281–3292. doi:10.1002/jssc.201200531. 453
Page 14
14
[10] W.D. Conway, Counter-Current Chromatography: Apparatus, Theory and 454
Applications, VCH Publishers Inc., NY, 1990. 455
[11] A. Berthod, Countercurrent Chromatography: From the Milligram to the 456
Kilogram, Adv. Chromatogr. (2008) 324–352. 457
[12] A. Berthod, T. Maryutina, B. Spivakov, O. Shpigun, I. A. Sutherland, 458
Countercurrent chromatography in analytical chemistry (IUPAC Technical 459
Report), Pure Appl. Chem. 81 (2009) 355–387. doi:10.1351/PAC-REP-08-06-05. 460
[13] X.L. Cao, Y. Tian, T.Y. Zhang, Q.H. Liu, L.J. Jia, Y. Ito, Separation of 461
dammarane-saponins from notoginseng, root of Panax notoginseng (Burk.) F. H. 462
Chen, by HSCCC coupled with evaporative light scattering detector, J. Liq. 463
Chromatogr. Relat. Technol. 26 (2003) 1579–1591. doi:10.1081/JLC-120021268. 464
[14] J. Peng, F. Dong, Y. Qi, X. Han, Y. Xu, L. Xu, Q. Xu, K. Liu, Z. Zhu, 465
Preparative separation of four triterpene saponins from Radix Astragali by high-466
speed counter-current chromatographycoupled with evaporative light scattering 467
detection, Phytochem. Anal. 19 (2008) 212–217. doi:10.1002/pca.1011. 468
[15] Y. Ito, Golden rules and pitfalls in selecting optimum conditions for high-speed 469
counter-current chromatography, J. Chromatogr. A. 1065 (2005) 145–168. 470
doi:10.1016/j.chroma.2004.12.044. 471
[16] A. Marston, K. Hostettmann, Developments in the application of counter-current 472
chromatography to plant analysis, J. Chromatogr. A. 1112 (2006) 181–194. 473
doi:10.1016/j.chroma.2005.10.018. 474
[17] P. Wood, S. Ignatova, L. Janaway, D. Keay, D. Hawes, I. Garrard, I.A. 475
Sutherland, Counter-current chromatography separation scaled up from an 476
analytical column to a production column, J. Chromatogr. A. 1151 (2007) 25–30. 477
doi:10.1016/j.chroma.2007.02.014. 478
[18] F.N. Costa, M.N. Vieira, I. Garrard, P. Hewitson, G. Jerz, G.G. Leitão, S. 479
Ignatova, Schinus terebinthifolius countercurrent chromatography (Part II): Intra-480
apparatus scale-up and inter-apparatus method transfer, J. Chromatogr. A. 1466 481
(2016) 76–83. doi:10.1016/j.chroma.2016.08.054. 482
[19] I. Sutherland, D. Hawes, S. Ignatova, L. Janaway, P. Wood, Review of progress 483
toward the industrial scale-up of CCC, J. Liq. Chromatogr. Relat. Technol. 28 484
(2005) 1877–1891. doi:10.1081/JLC-200063521. 485
[20] J.J. Ma, L.P. Kang, W.B. Zhou, H.S. Yu, P. Liu, B.P. Ma, Identification and 486
characterization of saponins in extract of Ziziphi spinosae semen (ZSS) by ultra-487
performance liquid chromatography-electrospray ionization-quadrupole time-of-488
flight tandem mass spectrometry (UPLC-ESI-QTOF-MS E), J. Med. Plant Res. 5 489
(2011) 6152–6159. doi:10.5897/JMPR11.339. 490
Page 15
15
[21] K. Yoshikawa, Y. Nagai, M. Yoshida, S. Arihara, Antisweet natural products. 491
VIII. Structures of hodulosides VI-X from Hovenia dulcis Thunb. var. tomentella 492
Makino, Chem. Pharm. Bull. 41 (1993) 1722–1725. 493
[22] K. Yoshikawa, S. Tumura, K. Yamada, S. Arihara, Antisweet natural products. 494
VII. Hodulosides I, II, III, IV, and V from the leaves of Hovenia dulcis Thunb., 495
Chem. Pharm. Bull. 40 (1992) 2287–2291. 496
[23] W. Phrompittayarat, W. Putalun, H. Tanaka, K. Jetiyanon, S. Wlttaya-Areekul, 497
K. Ingkaninan, Determination of pseudojujubogenin glycosides from Brahmi 498
based on immunoassay using a monoclonal antibody against bacopaside I, 499
Phytochem. Anal. 18 (2007) 411–418. doi:10.1002/pca.996. 500
[24] R. Higuchi, S. Kubota, T. Komori, T. Kawasaki, V.B. Pandey, J.P. Singh, A.H. 501
Shah, Triterpenoid saponins from the bark of Zizyphus joazeiro, Phytochemistry. 502
23 (1984) 2597–2600. doi:10.1016/S0031-9422(00)84106-0. 503
[25] A. Plaza, M. Cinco, A. Tubaro, C. Pizza, S. Piacente, New triterpene glycosides 504
from the stems of Anomospermum grandifolium, J. Nat. Prod. 66 (2003) 1606–505
1610. doi:10.1021/np030283j. 506
[26] D.G. Lee, Y. Lee, K. Kim, E.J. Cho, S. Lee, Novel dammarane-type triterpene 507
saponins from Panax ginseng root, Chem. Pharm. Bull. 63 (2015) 927–934. 508
[27] W. Yang, M. Ye, X. Qiao, C. Liu, W. Miao, T. Bo, H. Tao, D. Guo, A strategy 509
for efficient discovery of new natural compounds by integrating orthogonal 510
column chromatography and liquid chromatography/mass spectrometry analysis: 511
Its application in Panax ginseng, Panax quinquefolium and Panax notoginseng to 512
characterize 437 potencial new ginsenosides, Anal. Chim. Acta. 739 (2012) 56–513
66. doi:10.1016/j.aca.2012.06.017. 514
[28] X. Qi, S. Ignatova, G. Luo, Q. Liang, F.W. Jun, Y. Wang, I. Sutherland, 515
Preparative isolation and purification of ginsenosides Rf, Re, Rd and Rb1 from 516
the roots of Panax ginseng with a salt/containing solvent system and flow step-517
gradient by high performance counter-current chromatography coupled with an 518
evaporative light scattering detector, J. Chromatogr. A. 1217 (2010) 1995–2001. 519
doi:10.1016/j.chroma.2010.01.057. 520
[29] X. Zhang, Y. Ito, J. Liang, Q. Su, Y. Zhang, J. Liu, W. Sun, Preparative isolation 521
and purification of five steroid saponins from Dioscorea zingiberensis 522
C.H.Wright by counter-current chromatography coupled with evaporative light 523
scattering detector, J. Pharm. Biomed. Anal. 84 (2013) 117–123. 524
doi:10.1016/j.jpba.2013.02.005. 525
[30] R. Liu, L. Kong, A. Li, A. Sun, Preparative isolation and purification of saponin 526
and flavone glycoside compounds from Clinopodium chinensis (Benth) O. 527
Kuntze by high-speed countercurrent chromatography, J. Liq. Chromatogr. Relat. 528
Technol. 30 (2007) 521–532. doi:10.1080/10826070601093846. 529
530
Page 16
16
531 Figure 1. UHPLC-HRMS profiles of butanol extract (A) and its HPCCC groups 1 (B) 532
and 3 (C). 533
534
RT: 0.00 - 23.00 SM: 7G
0 5 10 15 20
Time (min)
0
10
20
30
40
50
60
70
80
90
100
Re
lative
Ab
un
da
nce
NL: 4.69E6
Base Peak m/z= 350.0000-1500.0000 F: FTMS - p ESI Full ms [350.00-1500.00] MS BUOH-2MGML
RT: 0.00 - 23.00 SM: 7G
0 5 10 15 20
Time (min)
0
10
20
30
40
50
60
70
80
90
100
Re
lative
Ab
un
da
nce
NL: 5.51E6
Base Peak m/z= 350.0000-1500.0000 F: FTMS - p ESI Full ms [350.00-1500.00] MS i-2mgml
RT: 0.00 - 23.00 SM: 7G
0 2 4 6 8 10 12 14 16 18 20 22
Time (min)
0
10
20
30
40
50
60
70
80
90
100
Rela
tive A
bundance
NL: 4.60E6
Base Peak m/z= 350.0000-1500.0000 F: FTMS - p ESI Full ms [350.00-1500.00] MS iva-2mgml
1
2
34 5
9
10
11
13
14,
15
16
18
20
67
12
17
19
21
24
22,
23
25
27
28
A B
C
Page 17
17
535 Figure 2. Separation by HPCCC of butanol extract and group 3. This scheme was based on information from a MIDI run. In Spectrum runs only 536
total number of fractions change. 537
538
Page 18
18
HO
C30H48O4
C
O
HOOH
O
OH
OH
C30H50O5
A
HOOH
O
OH
OH
C31H52O5
B
O
OH
539 Figure 3. Proposed aglycone structures of saponins in group 3: (A) 16-keto-540
tetrahydroxydammar-23-ene, (B) 16-keto-tetrahydroxydammar-24-methylene and (C) 541
jujubogenin. 542
543
544
545
546 Figure 4. UHPLC-HRMS profiles of group 3 HPCCC groups (B, C, D1-2). 547
548
2 18 21 2423
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
0
20
40
60
80
100
Re
lative
Ab
un
da
nce
NL: 6.61E6
Base Peak m/z= 560.0000-1500.0000 F: FTMS - p ESI Full ms [350.00-1500.00] MS ivac-1mgml
Group 3 CCC-C
1 3 4
10
17 28
14
27
16
13
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
0
20
40
60
80
100
Rela
tive
Ab
un
da
nce
NL: 5.24E6
Base Peak m/z= 560.0000-1500.0000 F: FTMS - p ESI Full ms [350.00-1500.00] MS ivad1-1mgml
Group 3 CCC-D11 3
49
10
11
13 14, 15
17
2 518 21 25
28
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
0
20
40
60
80
100
Rela
tive
Ab
un
da
nce
NL: 6.73E6
Base Peak m/z= 560.0000-1500.0000 F: FTMS - p ESI Full ms [350.00-1500.00] MS ivad2-1mgml
Group 3 CCC-D2
1
3
4
9
11
14, 15
17
18
255
28
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
0
20
40
60
80
100
Rela
tive
Ab
un
da
nce
NL: 3.81E6
Base Peak m/z= 560.0000-1500.0000 F: FTMS - p ESI Full ms [350.00-1500.00] MS ivab-1mgml
Group 3 CCC-B13
16
Page 19
19
Table 1. Solvent systems tested with butanol extract and group 3. 549
Hex DCM EtOAc Acetone BuOH PrOH iPrOH EtOH MeOH H20 CH3COONH4
5 mM
i - - 1 - 1 - - - - 2 -
ii 1 - 1 - - - - 1 - 1 -
iii 5 - 6 - - - - 5 - 6 -
iv 1 - 1 0.5 - - - 1 - 1 -
v 6 - 6 - 1 - - 6 - 6 -
vi 1 - 6 - 1 - - 1 - 6 -
1 - 6 - - - - 3 - 2 - 4
2 - - 1 - 0.5 - - - 0.2 1 -
3 - - 1 - 0.5 - - 0.2 - 1 -
4 - - 1 - - - - - 0.2 1 -
5 - - 1 - - - - 0.2 - 1 -
6 - - 1 - - - 0.5 - 0.2 1 -
7 - - 1 - - - 0.5 0.2 - 1 -
8 - - 1 - - 0.5 - - 0.2 1 -
9 - - 1 - - 0.5 - 0.2 - 1 -
Hex: hexane. DCM: dichloromethane. EtOAc: ethyl acetate. BuOH: butanol. PrOH: 550
propanol. iPrOH: isopropanol. EtOH: ethanol. MeOH: methanol. 551
552
Page 20
20
553
Table 2. UHPLC-HRMS data of saponins detected in butanol extract HPCCC group 3. 554
Peak tR
(min)
[M-H]-
(m/z)
Molecular
Formula
ppm Diagnostic product ion a
(m/z)
Aglycone b
Sugar residue c
1 8.2 779.4577 C42H68O13 0.2 633 (-dHex), 617 (-Hex), 471 (C30H48O4) C30H48O4 1 Hex, 1 dHex
2 11.4 959.5211 C48H80O19 0.1 817 (-C8H14O2), 655 (-C8H14O2-Hex), 509d (-C8H14O2-Hex-dHex) C30H50O5 2 Hex, 1 dHex
3 12.2 915.4956 C46H76O18 0.9 773 (-C8H14O2), 611 (-C8H14O2-Hex), 641 (-C8H14O2-Pen), 479d (-
C8H14O2-Hex-Pen)
C30H50O5 2 Pen, 1 Hex
4 12.7 929.5113 C47H78O18 0.9 787 (-C8H14O2), 625(-C8H14O2-Hex), 479d (-C8H14O2-Hex-dHex) C30H50O5 1 Hex, 1 dHex,
1 Pen
5 13.0 959.5217 C48H80O19 0.8 803 (-C9H16O2) 641 (-C9H16O2-Hex), 479d (-C9H16O2-2Hex) C31H52O5 2 Hex, 1 Pen
6 13.6 931.5266 C47H80O18 0.6 799 (-Pen), 769 (-Hex), 637(-Hex-Pen) C30H52O4 2 Hex, 1 Pen
7 13.7 927.4951 C47H76O18 0.3 765 (-Hex-), 603(-2Hex) C30H48O4 2 Hex, 1 Pen
8 13.8 955.4901 C48H76O19 0.5 823 (-Pen), 793(-Hex), 661 (-Pen-Hex) 1 Pen, 1 Hex
9 13.9 959.5217 C48H80O19 0.7 803 (-C9H16O2), 641 (-C9H16O2-Hex), 479d (-C9H16O2-2Hex) C31H52O5 2 Hex, 1 Pen
10 14.4 929.5108 C47H78O18 0.4 773 (-C9H16O2), 611 (- C9H16O2-Hex), 479d (-C9H16O2-Hex-Pen) C31H52O5 1 Hex, 2 Pen
11 14.9 927.4957 C47H76O18 1 795 (-Pen), 765 (-Hex), 633(-Hex-Pen) C30H48O4 2 Hex, 1 Pen
12 15.3 929.5106 C47H78O18 0.2 773 (-C9H16O2), 611 (- C9H16O2-Hex), 479d (-C9H16O2-Hex-Pen) C31H52O5 1 Hex, 2 Pen
13 15.6 929.511 C47H78O18 0.6 773 (-C9H16O2), 611 (- C9H16O2-Hex), 479d (-C9H16O2-Hex-Pen) C31H52O5 1 Hex, 2 Pen
14 15.7 897.4835 C46H74O17 -0.9 765 (-Pen), 735 (Hex), 603(-Pen-Hex), 471d (C30H48O4) C30H48O4 1 Hex, 2 Pen
15 16.0 897.4854 C46H74O17 1.3 765 (-Pen), 735 (Hex), 603(-Pen-Hex), 471d (C30H48O4) C30H48O4 2 Hex, 2 Pen
16 16.2 943.5258 C48H80O18 -0.3 787 (-C9H16O2), 625 (-C9H16O2-Hex), 479d (-C9H16O2-Hex-dHex) C31H52O5 1 Hex, 1 dHex,
1 Pen
17 16.5 911.5001 C47H76O17 0.2 749 (-Hex), 603 (-Hex-dHex) C30H48O4 1 Hex, 1 dHex,
1 Pen
18 16.7 955.5257 C49H80O18 -0.3 793 (-Hex), 647 (-Hex-dHex) 1 Hex, 1 dHex
19 17.4 1013.532 C51H82O20 0.1 851 (-Hex), 705 (-Hex-dHex) 1 Hex, 1 dHex
20 18.0 1003.547 C50H84O20 -0.1 943(-C2H4O2), 841 (-Hex), 781 (-C2H4O2-Hex) C30H52O4 3 Hex
Page 21
21
21 18.2 1013.532 C51H82O20 0.3 851 (-Hex), 705 (-Hex-dHex) 1 Hex, 1 dHex
22 19.0 1003.548 C50H84O20 0.7 943(-C2H4O2), 841 (-Hex), 781 (-C2H4O2-Hex) C30H52O4 3 Hex
23 19.0 1041.527 C52H82O21 0.1 897 (-HMG), 879 (-Hex), 765 (-HMG-Pen), 735 (-HMG-Hex) C30H48O4 1 Hex, 2 Pen
24 19.3 1041.527 C52H82O21 0 897 (-HMG), 879 (-Hex), 765 (-HMG-Pen), 735 (-HMG-Hex) C30H48O4 2 Hex, 2 Pen
25 20.1 973.5371 C49H82O19 0.5 913 (-C2H4O2), 811 (-Hex), 781 (-C2H4O2-Pen), 751 (-C2H4O2-Hex),
619 (-C2H4O2-Pen-Hex)
C30H52O4 2 Hex 1 Pen
26 20.1 969.5062 C49H78O19 1.0 837 (-Pen), 807 (-Hex), 675 (-Pen-Hex) 1 Hex, 1 Pen
27 20.9 969.5058 C49H78O19 0.4 837 (-Pen), 807 (-Hex), 675 (-Pen-Hex) 1 Hex, 1 Pen
28 21.2 973.5372 C49H82O19 0.6 913 (-C2H4O2), 811 (-Hex), 781 (-C2H4O2-Pen), 751 (-C2H4O2-Hex),
619 (-C2H4O2-Pen-Hex)
C30H52O4 2 Hex 1 Pen
a Hex: hexose; dHex: deoxyhexose; Pen: pentose; HMG: hydroxymthylglutaryl;
b strucutures reported in Figure 3;
c in bold the sugar residue 555
attached to aglycone skeleton; d product ions detected in MS
3 spectra. 556
557