- Email: [email protected]
Analysis of flavonoids by high-performance liquid chromatography
SUMMARY
Numerous acetatesof Bavonoids (avoneS, flavonols and flavanone aglycones and glycosides)have been separatedand determinedon silicagel, using fom diSerent liquid systems(&za.mt~amtone; btznzen~~toniaile; benzenwzthanol; isoocta~ ethmobacettitie) at temperaturesbetween 49 and O”, This -me&ad has been appliedto ffie &k&ation of amdyticallypure substancesand fruit and vegetable extracts,
ENTRODUclloN
Flavoaoids exist in almost alI plants and can be presentin considerableconcentrations,e.g., in leaves. In the past, thin-layerchromatographywas mainly used for their analysis. R_mtly some workers have applied reversed-phasehigh-performanceliquidchromato&phy (HPLC) to pasticulartlavonoids’” and to theflavonoids in citrusfruit?,tobacco5,soya6and speciesof Mx’ and Cet&u.P_ The gzeatadvantage of the commonly used gradientelutionprocessis the saving of time and solventsby the directinjectionof only a slightlypurifiedcomplex extict, but expensivecolumns are require& The lack of a prelimimry clean-up, however, may lead to errors in identificatiotiif only one analysis system is usedAnother possibilityis to acetylatethe compounds,the productsbe&g separated isomaticallyon silica gel. The selectiveclean-up and acetyl&tionmethod that is used guaranteesdmabiity of the cohmms, detite ide&Gcation and quantitativedetennination of tiavonoidsin plant material,Such a method is describedin this paper.
A Siemens S 200 high-performanoeSquid chromatografih(isocratic system with an injection dosage of 29 4) was used with a Pye U&am EC 3 W detector.The cohunnsystenco&stedofasfzsiesoftbreestainlesssteeicolumns (eachZQQx3rnm UX) (dire&y conmcted). packed with 5-~ Idchrosorb Si 60 (Merck, IBam&a&
RVEciE
218
G-F-R_)_ Berrpme-acetone,benx.ne-xetonitrife, b~&e-&ii.~ol and, %oocta&~+. et.banoLacctonitrsLe w-ereused as mobile phases.The flow-rateswerebetween0-g and . ._ C‘-’ ;_: :.( :. ::._ .._~.-“..: _‘; r i;: ~ ‘: _,I._ lJ&/m& ‘_C.( .’ A Pye Uticam SP 800 instrumentwas used for the detkikatitinksf the eIe& tronic absorption spectra of the three groups of cumpoumis anaiysed (see Fig_ 1). Other equipmentused collsistEdof an Wtraturrax (MA, Staufkn,G.F.R.], a I&ufuge 3 centrZ@e (4300 g) (HeraeusCkis& &term&, G_F_R;),-811 c2‘ fikez4z’tier (R’KF, Btida&- G F-R-) and a rotary evaporator(Ktido!pk:Kelhe& G_F.R_)_-Al.l _ evaporationswereperformedat a temperaturenot higherthan 40”. ‘.
w; I
g
f
‘k_
1
,__.J
w:
z,FLAVASCINES 01 es--FLAVONO!! u”:_ RAVONES “: ! !LlMmcx=TRAwsA~
-Commercialproducts(seeTribleIi)or cokpoids qxtlxskd by Koeppen9or isolated by our group were used. The commercXproducts’usuaIIy neededchemical pIlrifkatiol?_ : i3olvents of analytical-reagentgrade prere p&i&d with Extrelut (Merck) (aSways1 lpcrc&~~). PO&amid&SC G was obtainedfrom Macherey, Nagei &Co. @iken, G.F_R-j_ Clan up Of plant t?rah~P-~‘. The mat&id was komogenized with methanol in an UItraW, followed by cektrifi~gatioi~ and a triple methanolextractionof the remainder of tissue_The methan was cvapoxatcd, kaving an aqsous’ soktion, then fipoid substanceswere extract& by seps%ed shakingwith light petroleum(b-p_ 40-60”). The aqueous extractwas placed on a polyamide WIwhich wzs wztskd with water @IO-loo0 mQ_The fLzvonoidswere &ted with metband (XKI-10 ml), and glucuronidesby addition of ammonia? tke latter were not analysfxi in this study. The eIuatewas evaporatedto~dsyness;freeze drying is possiile after evaporaticmof methanol_Finally acetylation-wzsskied out .’ Acetylmiani2. Equal potions of pyridineand ace& a&y&& (2 ml-pez &I mg) wereadded to the dry pklt extractor t6xhti pure s&sti~; anB di@oived withy
stirring & boiling ten36 me soltia wasthen moled kld amwed m stand. Esterikatior~ was complete after several hours at ambient temperature (ovtigbt) OFSier P h in & drier a~ 70”. A 34%fold surpks of tild -k&et was added, the acetates being t5mpkf&y precipit&ed GM% Bre aIlowed to stand, preGeral& overnight in a (nOrmaUy 1-2 6 wi@ be adequate).The precipitatewetsf%mcd through a GA &ii and Washedwitb d&ii& water. hncomplete ac&ylation oft&e hydmxyl groups re&eratm
wan n~tcixemed~anyexperir~~t.The~ residuewaseiutedwith 15Ctml ofhenazntxazetone (85:15). The eluate was evaporated to dryness and the residue, dissolved in a solvent and suitablyconcentrated,was then subjectedto HFLC.
RE%JLTSAND DISCUSSION Table I givesthe retentiontimesof the examined substances in diEkent systems bat with tie same CQ~UEUL This comparisonshows occasionalreversalof retentionin TABLE I R=ENmON
TIhSES OF ACEXATES
B/N = knzae-acetmitriIe; B/A = -one; ethaookce%onitrik_
Peak NO.
: 3 4 5 6 ii 9 a0 11 :f 14 15
16 17 18 19 zi 22 23 24 25 z 28
corn
B/E = knzene~ol;
I/E/N = kooaane-
BIN (85:22).
WA (90:15),100b.iu
BE (8&0_8).
UEij++J (70:1655J.
I70 bar, C5’
20”
220 bar I6”
155bar, 45”
a0
450 483 309 343 362 392 392 446 447 4+8 485 489 520 520 541 583 6im 623 7oQ 707 70s 745
774 820 832 9rs mm
52s
639
639
774 785
774
945
774
!3m
837
1026
1052 954
1170
618
627 699 674
824
680
824
685
I040 1124 1025 950 102 942 950 1359 12Ll
760 803 870
1680
860 855 980 860 990
1682
1030
‘
Jrj 1
z 15 2
L 5
TABLEII DETECXI0~LIMll-S
i
x)
Min.
’
5
OF AcETAlEs
Injection: 2OpL Eiuents: see T&e
I.
15
Min
32 Min
Benzene-acetoni~e was the most suitable mobile phase, especially at 45”. Measurement up to 285 mu cm be made withoutdi&uky, and fiavanonescan therefore also be determined.The extremelydifferentpropertiesof flavono’raad fiavanone giyckdes on the one band and of Bavone glymsides owlthe other are uses from ffie ana@tical point of view. Bemene-acetonitriIe(85:40) separatedthe ftavoneglycusides very well, whereasBavonoi and ftavanonegIycosiaeshave very short retentiontimes. Interestingresu&swerkobtainedby cooling the column (Fig. 5). These res&s could not be achieved by changing the Ilow-rate or the proportions of the ratio coruponetnts in the mobile phase. They a~ based on differmes in the soIubllitiesof the saccharides(wbicb are aiso aceta6+, which dependon temperature.Cooling was employedequallysu~y with the other systemsand plant extracts.Cooling could also beuseful in 33SeSed-phaseHPLC. Bezmn~tiol often gave the best separation,but the sensitivi~ was lower than with B/N. Small difkem3 in the proportions of the components gave totally differentretention times, With benzemz-ethand it was possible to separateR- and S-isomers of fbvmone glycosidesif cooling was applied (Fig_ 6), but long retention timesresuxted.
TABLE KIWI OPERATENG
54,s
3 5 6 7 1%
p&ce
CONDl3XONS
z 155
330 300 335 312 3cmp70
180 15s
270 300
B/-Pi I/E/N(80:35) (7O:l6:5.5)
z37
180 180
BjN (85:20) B/N (8552)
45
B/A(!JO:lS)
izzdgme
B/E (8wO.5) I/-!qN (70:16:6)
15
None of these eluents pen&s useful detectionbelow 285 nm because of the of benzene. Therefore, isooctan~tbanol-acetonitrile was used for the
F&_ 6_ scpyation of R- and Sisoses
(fat peak see Tabk f ; for conditions see Tab&zm).
(b)
To
20
Min
1
?o
20
Mm.
20
Mm,
I IO’
20 Min.
6
IF-m-w
T-my
JL-L 6
E?
Fi8,8, Chromatogrnmof cxtrnctof skinsof ripotomatoes(“PrUhzaubcr”), Por pcakaaceTnbloI (28 w i\-#annlstnndnrd);l$r ppnditiongSC+Tn\&,fAI, Fig 9, Chromnio~ramof cxtrnctof cclcrylcnvcs(“Albn”).Pcnks: 1 = npigenin-7-Blucosidc; 2 = Iutcoliq-%@uqoside; 3 = puob@lyChrysoer~pI~‘lvllJiob~~~~r 4 = nplin; 5 = luteolin-7.npiosyl-glucoside; G= probablychrysoeriol-7-npiosylglucosidc, Par condltionpscqTgbleIII, r Fig, 10,ChromutoBrnm ofcxtrnctof ornqe juice,For pcnksseeTableI (1 - internalstnndnrd);for conc)itiqaa set Tnblc111,
10
4-J A
28
24
z4
R.GAU%S&ZC_H
sektive detectionof Bavanonesat 270 nm bezause there are speci.6~possibiities of detemGn&m at this wavekngth, where flavones and fiavoncls show a minimum (Fig_ 7a aud I$_ The increasein the sensitivity of measuringflavanonesat 270 mn, mentsat 300 or 312 nm, permitkd the presenceof Bavanones, compa!Xdwithmezisure ~&WOWS or Bavonolsin the examinedextractsto be established_ Applications
Chromatograms of an extsactof tomato, an extractof celery0 and an extractof orange juice are shown in Figs_ 8,9 and IO, respectively.Detailed studies of tbs applicationsto tomato and orangejuice will be publishedseparately. CONCLUSION
The expensive clean-up (imcomparisonwith reversed-phasemethod3)is offset by the long durabilityof the columns_The use of the four mobile phase systemsat d.ifknst temperaturesand wavelengthsmade possiblethe id~&kation of&;avcmoids eVenillUtlbOWT.tmixtures.
1 G_J_Nianarm and J. vza lhmkcde, J. tzhmmwm_, 152 (1978) 523. 2 L- W. Wdi and C. W_ Nzgef, L Chwogr_, 116 (1976) 271, 3 LtBe&a,G_WilkiqandK_Hosttttmann, L Ckromarogr_, 136 (1977) 174 4 J. F. Fii, L Agr. Fd CT.., 26 (197s) 1459. 5 w. k court. /. ciuomoiagr.., 130 (1977) 2m_ 6 L- G- West., P_ M, Bi and D. E. Pratt, L CArcunarogr.,150 (1978) 266. 7 G. 3. P&mam and J. W. Koersdmaa- Kc0y, Ph?Iua Med, 31 (1977) 297. 8 G-J-Nz_ Naturfonch. c., 32 (1977) 1015. 9 B. H. Koeppcq Unkrs&y of Stefknm South Africa persa& coinmunication, 10 W. WZdager and K Hexmxuq J_ ChTonaarogt., 76 (1973) 433. 11 ftTeubecaud XC.Herrmum, Z_ E&mum.-Untem_-For., 165 (1977) 147. 12 R Galwsa and K- He-, 2, ZrbawF -Uitters_-FimcF~. 166 (1978) 355. 13RaGalmsaandK.H~ i?. Lebcnsn;-Unters.-F*~==h., 169 (1979) 170.
Numerous acetatesof Bavonoids (avoneS, flavonols and flavanone aglycones and glycosides)have been separatedand determinedon silicagel, using fom diSerent liquid systems(&za.mt~amtone; btznzen~~toniaile; benzenwzthanol; isoocta~ ethmobacettitie) at temperaturesbetween 49 and O”, This -me&ad has been appliedto ffie &k&ation of amdyticallypure substancesand fruit and vegetable extracts,
ENTRODUclloN
Flavoaoids exist in almost alI plants and can be presentin considerableconcentrations,e.g., in leaves. In the past, thin-layerchromatographywas mainly used for their analysis. R_mtly some workers have applied reversed-phasehigh-performanceliquidchromato&phy (HPLC) to pasticulartlavonoids’” and to theflavonoids in citrusfruit?,tobacco5,soya6and speciesof Mx’ and Cet&u.P_ The gzeatadvantage of the commonly used gradientelutionprocessis the saving of time and solventsby the directinjectionof only a slightlypurifiedcomplex extict, but expensivecolumns are require& The lack of a prelimimry clean-up, however, may lead to errors in identificatiotiif only one analysis system is usedAnother possibilityis to acetylatethe compounds,the productsbe&g separated isomaticallyon silica gel. The selectiveclean-up and acetyl&tionmethod that is used guaranteesdmabiity of the cohmms, detite ide&Gcation and quantitativedetennination of tiavonoidsin plant material,Such a method is describedin this paper.
A Siemens S 200 high-performanoeSquid chromatografih(isocratic system with an injection dosage of 29 4) was used with a Pye U&am EC 3 W detector.The cohunnsystenco&stedofasfzsiesoftbreestainlesssteeicolumns (eachZQQx3rnm UX) (dire&y conmcted). packed with 5-~ Idchrosorb Si 60 (Merck, IBam&a&
RVEciE
218
G-F-R_)_ Berrpme-acetone,benx.ne-xetonitrife, b~&e-&ii.~ol and, %oocta&~+. et.banoLacctonitrsLe w-ereused as mobile phases.The flow-rateswerebetween0-g and . ._ C‘-’ ;_: :.( :. ::._ .._~.-“..: _‘; r i;: ~ ‘: _,I._ lJ&/m& ‘_C.( .’ A Pye Uticam SP 800 instrumentwas used for the detkikatitinksf the eIe& tronic absorption spectra of the three groups of cumpoumis anaiysed (see Fig_ 1). Other equipmentused collsistEdof an Wtraturrax (MA, Staufkn,G.F.R.], a I&ufuge 3 centrZ@e (4300 g) (HeraeusCkis& &term&, G_F_R;),-811 c2‘ fikez4z’tier (R’KF, Btida&- G F-R-) and a rotary evaporator(Ktido!pk:Kelhe& G_F.R_)_-Al.l _ evaporationswereperformedat a temperaturenot higherthan 40”. ‘.
w; I
g
f
‘k_
1
,__.J
w:
z,FLAVASCINES 01 es--FLAVONO!! u”:_ RAVONES “: ! !LlMmcx=TRAwsA~
-Commercialproducts(seeTribleIi)or cokpoids qxtlxskd by Koeppen9or isolated by our group were used. The commercXproducts’usuaIIy neededchemical pIlrifkatiol?_ : i3olvents of analytical-reagentgrade prere p&i&d with Extrelut (Merck) (aSways1 lpcrc&~~). PO&amid&SC G was obtainedfrom Macherey, Nagei &Co. @iken, G.F_R-j_ Clan up Of plant t?rah~P-~‘. The mat&id was komogenized with methanol in an UItraW, followed by cektrifi~gatioi~ and a triple methanolextractionof the remainder of tissue_The methan was cvapoxatcd, kaving an aqsous’ soktion, then fipoid substanceswere extract& by seps%ed shakingwith light petroleum(b-p_ 40-60”). The aqueous extractwas placed on a polyamide WIwhich wzs wztskd with water @IO-loo0 mQ_The fLzvonoidswere &ted with metband (XKI-10 ml), and glucuronidesby addition of ammonia? tke latter were not analysfxi in this study. The eIuatewas evaporatedto~dsyness;freeze drying is possiile after evaporaticmof methanol_Finally acetylation-wzsskied out .’ Acetylmiani2. Equal potions of pyridineand ace& a&y&& (2 ml-pez &I mg) wereadded to the dry pklt extractor t6xhti pure s&sti~; anB di@oived withy
stirring & boiling ten36 me soltia wasthen moled kld amwed m stand. Esterikatior~ was complete after several hours at ambient temperature (ovtigbt) OFSier P h in & drier a~ 70”. A 34%fold surpks of tild -k&et was added, the acetates being t5mpkf&y precipit&ed GM% Bre aIlowed to stand, preGeral& overnight in a (nOrmaUy 1-2 6 wi@ be adequate).The precipitatewetsf%mcd through a GA &ii and Washedwitb d&ii& water. hncomplete ac&ylation oft&e hydmxyl groups re&eratm
wan n~tcixemed~anyexperir~~t.The~ residuewaseiutedwith 15Ctml ofhenazntxazetone (85:15). The eluate was evaporated to dryness and the residue, dissolved in a solvent and suitablyconcentrated,was then subjectedto HFLC.
RE%JLTSAND DISCUSSION Table I givesthe retentiontimesof the examined substances in diEkent systems bat with tie same CQ~UEUL This comparisonshows occasionalreversalof retentionin TABLE I R=ENmON
TIhSES OF ACEXATES
B/N = knzae-acetmitriIe; B/A = -one; ethaookce%onitrik_
Peak NO.
: 3 4 5 6 ii 9 a0 11 :f 14 15
16 17 18 19 zi 22 23 24 25 z 28
corn
B/E = knzene~ol;
I/E/N = kooaane-
BIN (85:22).
WA (90:15),100b.iu
BE (8&0_8).
UEij++J (70:1655J.
I70 bar, C5’
20”
220 bar I6”
155bar, 45”
a0
450 483 309 343 362 392 392 446 447 4+8 485 489 520 520 541 583 6im 623 7oQ 707 70s 745
774 820 832 9rs mm
52s
639
639
774 785
774
945
774
!3m
837
1026
1052 954
1170
618
627 699 674
824
680
824
685
I040 1124 1025 950 102 942 950 1359 12Ll
760 803 870
1680
860 855 980 860 990
1682
1030
‘
Jrj 1
z 15 2
L 5
TABLEII DETECXI0~LIMll-S
i
x)
Min.
’
5
OF AcETAlEs
Injection: 2OpL Eiuents: see T&e
I.
15
Min
32 Min
Benzene-acetoni~e was the most suitable mobile phase, especially at 45”. Measurement up to 285 mu cm be made withoutdi&uky, and fiavanonescan therefore also be determined.The extremelydifferentpropertiesof flavono’raad fiavanone giyckdes on the one band and of Bavone glymsides owlthe other are uses from ffie ana@tical point of view. Bemene-acetonitriIe(85:40) separatedthe ftavoneglycusides very well, whereasBavonoi and ftavanonegIycosiaeshave very short retentiontimes. Interestingresu&swerkobtainedby cooling the column (Fig. 5). These res&s could not be achieved by changing the Ilow-rate or the proportions of the ratio coruponetnts in the mobile phase. They a~ based on differmes in the soIubllitiesof the saccharides(wbicb are aiso aceta6+, which dependon temperature.Cooling was employedequallysu~y with the other systemsand plant extracts.Cooling could also beuseful in 33SeSed-phaseHPLC. Bezmn~tiol often gave the best separation,but the sensitivi~ was lower than with B/N. Small difkem3 in the proportions of the components gave totally differentretention times, With benzemz-ethand it was possible to separateR- and S-isomers of fbvmone glycosidesif cooling was applied (Fig_ 6), but long retention timesresuxted.
TABLE KIWI OPERATENG
54,s
3 5 6 7 1%
p&ce
CONDl3XONS
z 155
330 300 335 312 3cmp70
180 15s
270 300
B/-Pi I/E/N(80:35) (7O:l6:5.5)
z37
180 180
BjN (85:20) B/N (8552)
45
B/A(!JO:lS)
izzdgme
B/E (8wO.5) I/-!qN (70:16:6)
15
None of these eluents pen&s useful detectionbelow 285 nm because of the of benzene. Therefore, isooctan~tbanol-acetonitrile was used for the
F&_ 6_ scpyation of R- and Sisoses
(fat peak see Tabk f ; for conditions see Tab&zm).
(b)
To
20
Min
1
?o
20
Mm.
20
Mm,
I IO’
20 Min.
6
IF-m-w
T-my
JL-L 6
E?
Fi8,8, Chromatogrnmof cxtrnctof skinsof ripotomatoes(“PrUhzaubcr”), Por pcakaaceTnbloI (28 w i\-#annlstnndnrd);l$r ppnditiongSC+Tn\&,fAI, Fig 9, Chromnio~ramof cxtrnctof cclcrylcnvcs(“Albn”).Pcnks: 1 = npigenin-7-Blucosidc; 2 = Iutcoliq-%@uqoside; 3 = puob@lyChrysoer~pI~‘lvllJiob~~~~r 4 = nplin; 5 = luteolin-7.npiosyl-glucoside; G= probablychrysoeriol-7-npiosylglucosidc, Par condltionpscqTgbleIII, r Fig, 10,ChromutoBrnm ofcxtrnctof ornqe juice,For pcnksseeTableI (1 - internalstnndnrd);for conc)itiqaa set Tnblc111,
10
4-J A
28
24
z4
R.GAU%S&ZC_H
sektive detectionof Bavanonesat 270 nm bezause there are speci.6~possibiities of detemGn&m at this wavekngth, where flavones and fiavoncls show a minimum (Fig_ 7a aud I$_ The increasein the sensitivity of measuringflavanonesat 270 mn, mentsat 300 or 312 nm, permitkd the presenceof Bavanones, compa!Xdwithmezisure ~&WOWS or Bavonolsin the examinedextractsto be established_ Applications
Chromatograms of an extsactof tomato, an extractof celery0 and an extractof orange juice are shown in Figs_ 8,9 and IO, respectively.Detailed studies of tbs applicationsto tomato and orangejuice will be publishedseparately. CONCLUSION
The expensive clean-up (imcomparisonwith reversed-phasemethod3)is offset by the long durabilityof the columns_The use of the four mobile phase systemsat d.ifknst temperaturesand wavelengthsmade possiblethe id~&kation of&;avcmoids eVenillUtlbOWT.tmixtures.
1 G_J_Nianarm and J. vza lhmkcde, J. tzhmmwm_, 152 (1978) 523. 2 L- W. Wdi and C. W_ Nzgef, L Chwogr_, 116 (1976) 271, 3 LtBe&a,G_WilkiqandK_Hosttttmann, L Ckromarogr_, 136 (1977) 174 4 J. F. Fii, L Agr. Fd CT.., 26 (197s) 1459. 5 w. k court. /. ciuomoiagr.., 130 (1977) 2m_ 6 L- G- West., P_ M, Bi and D. E. Pratt, L CArcunarogr.,150 (1978) 266. 7 G. 3. P&mam and J. W. Koersdmaa- Kc0y, Ph?Iua Med, 31 (1977) 297. 8 G-J-Nz_ Naturfonch. c., 32 (1977) 1015. 9 B. H. Koeppcq Unkrs&y of Stefknm South Africa persa& coinmunication, 10 W. WZdager and K Hexmxuq J_ ChTonaarogt., 76 (1973) 433. 11 ftTeubecaud XC.Herrmum, Z_ E&mum.-Untem_-For., 165 (1977) 147. 12 R Galwsa and K- He-, 2, ZrbawF -Uitters_-FimcF~. 166 (1978) 355. 13RaGalmsaandK.H~ i?. Lebcnsn;-Unters.-F*~==h., 169 (1979) 170.