Chapter 1 Isolation Techniques tor

KEN. R. MARKHAM

1.1 Introduction 1.2 Plant preparation, extraction and preliminary purification 1.2.1 Plant preparation and extraction 1.2.2 Preliminary purification 1.3 Column chromatography 1.3.1 Silica gel 1.3.2 Magnesium silicate 1.3.3 Alumina 1.3.4 Cellulose 1.3.5 Polyamide 1.3.6 Sephadex gel 1.3.7 Ion exchange resins 1.4 Thin-Iayer chromatography 1.4.1 Preparative sca1e TLC 1.4.2 Spray reagents and detecting methods 1.4.3 Separations on TLC, general 1.4.4 Separations on silica gel 1.4.5 Separations on cellulose 1.4.6 Separations on polyamide 1.5 Paper chromatography 1.6 Miscellaneous techniques 1.6.1 Gas-liquid chromatography 1.6.2 Paper electrophoresis 1.6.3 Sublimation 1.6.4 High pressure liquid chromatography

J. B. Harborne et al. (eds.), The Flavonoids © Springer Science+Business Media Dordrecht 1975 2 THE FLA VONOIDS

1.1 Introduction Over the past decade, the techniques used for the isolation of flavonoids from plant material have not changed drastically Review articles on this general topic by Seshadri (1962) and Seikel (1962) emphasized the importance of solvent extraction, crystalliz• ation, and column and paper chromatography, and these techniques are still very much in use today. Thin-layer chromatography, sephadex gel chromatography and gas-liquid chromatography are perhaps the three major innovations since that time. The increasing use of these essentially sma!l scale methods (together with paper chromatography) has been brought about largely by two factors: (i) the need to isolate only small amounts of material for structure analysis and (ii) the availability of only limited quantities of plant materials for surveys. In this chapter, although some general background information on plant preparation, extraction, paper and column chromatography is presented, no attempt is made to summarize the bulk of the data published in previous reviews. Rather, it is proposed to update these earlier reviews where necessary and to discuss the application of new techniques in greater detail. The diverse nature of the subject and the relative inaccessibility of such information from abstract literature, however, have precluded the possibility of obtaining complete literature coverage.

1.2 Plant preparation, extraction and preliminary purification

1.2.1 Plant preparation and extraction Flavonoids occur in virtually all parts of the plant, the root, heartwood, sapwood, bark, leaf, fruit and flower, and the method of isolation depends to some extent both on the source material and the type of being isolated. In cases when flavonoids occur in the surface oils or waxes, they may be obtained simply by scraping or washing the surface with an appropriate solvent (Wollenweber and Egger, 1971; Star and Mabry, 1971: Wollenweber, 1972). In general, however, the plant material is ground up or macerated before extraction. The possibility of enzyme action occurring during this early period of isolation, leading in particular to hydro lysis of (Beck, 1964; Trim, 1955), may be avoided by plunging ISOLATION TECHNIQUES FOR FLAVONOIDS 3 the plant material into boiling solvent or by rapid drying prior to extraction (Seshadri, 1962). Pre-drying of plant material gene rally appears to increase the yield of extractives, possibly due to rupture of the cell structure and to the better solvent access provided as a consequence. Solvents used for extraction are chosen according to the polarity of the flavonoids being studied. The less polar solvents are particularly useful for the extraction of flavonoid aglycones, whilst the more polar solvents are used if flavonoid glycosides or antho• cyanins are sought. The less polar aglycones, such as isoflavones, and dihydroflavonols, or and which are highly methylated, are usually extracted with solvents such as benzene, chloroform, ether or ethyl acetate (Mukerjee et al., 1969; Jackson et al .. 1967; Henrick and Jefferies, 1964; Kupchan et al., 1969; Shamma and Stiver, 1969; Morgan and Orsler, 1967; Harborne et al., 1963; Thomas and Mabry, 1968; Herz et al., 1972). A pre-extraction with light petroleum or hexane is lrequently carried out to rid plant material of sterols, carotenoids, chlorophylls, etc. (Bhutani et al.. 1969a; Bhutani et al., 1969b; Clark-Lewis and Porter, 1972). However, it should be noted that even this treatment may cause the loss (by extraction) of certain flavonoid ag1ycones. For example, the isoflavones, ichthynone and rotenone, have been isolated by extraction of Piscidia erythrina root wood with hexane (Schwarz et al., 1964) and methylated flavones such as are known (Kupchan et al. , 1969) to be significantly soluble in light petroleum. Flavonoid glycosides and the more polar aglycones such as hydroxylated flavones, flavonols, biflavonyls, and cha1cones are gene rally is01ated from plant material by extraction with acetone, a1cohol, water or a combination of these (Bottom1ey et al.. 1966; Markharn et al. , 1970; Ohta and Yagishita, 1970; Bhutani et al. , 1969a; Cambie and J ames, 1967; Dreyer and Bertelli, 1967). Perhaps the most useful solvent for the extraction of this group of compounds is a 1 : 1 mixture of water and methanol. Boiling water has been found suitable for the extraction of polyglycosides such as 7-0-diglucuronide (Okigawa et al., 1971a) and the flavone• polysaccharide compound from Monoclea forsteri (Markharn, 1972), and also for the isolation of compounds such as flavandiols, catechins and procyanidins. Traces of acid are occasionally incorporated in the 4 THE FLA VONOIDS solvent for the extraction of flavonoid glycosides (Wallace et al. , 1969) aIthough this practice is normally reserved for the extraction of anthocyanins and (Seshadri, 1962; Harborne, 1967b). The use of acid, however, can lead to hydro lysis of glycosidic materials.

1.2.2 Preliminary purification When flavonoids of varying types are to be extracted from a single batch of plant material, a worthwhile method for preliminary separation is sequential solvent extraction with a number of solvents of varying polarity. This can lead to separation of glycosides from aglycones and to the separation of polar from non-polar aglycones (Bhutani et al. , 1969b; Thomas and Mabry, 1968; Morgan and Orsler, 1967). AIternatively, sequential solvent extraction of a cmde extract may be used to produce the same type of separation (Ohta and Yagishita, 1970; Bhutani et al. , 1969a; Wong and Francis, 1968; Kupchan ef al. , 1969). Counter current separation techniques may also be of value, particularly for the separation of flavanoids (Roux and Paulus, 1962; Drewes et al. , 1967b; Drewes et al., 1967a; Drewes and Ilsley, 1969; Seikel et al., 1962). Distributions between water and an organic phase such as ethyl acetate (Clark-Lewis and Dainis, 1968; Coxon et al., 1972; Clark-Lewis and Dainis, 1967; Clark-Lewis and Porter, 1972) or butanol: light petroleum (peIter et al. , 1969; Roux and Paulus, 1962; Clark-Lewis and Porter, 1972) have been found effective for this purpose. Precipitation with lead acetate has been widely used in the past as a method of isolating phenolics (especially those with o-dihydroxyl groups) from other extractives in cmde extracts (Seshadri, 1962; Seikel, 1964). It has the disadvantage, however, that it does not precipitate some phenols and may co-precipitate other compounds. Decomposition of the lead saIts formed is best carried out with sulphate or phosphate rather than with sulphide, in view of the highly adsorptive nature of the precipitated lead sulphide. Recently, a method using polyvinylpyrrolidone (PVP) as precipitant has been suggested for use in cases where lead acetate is unsatisfactory (Andersen and Sowers, 1968). The optimum conditions for PVP• phenol bonding were established as pH 3.5 in 1-10% methanol in water. Treatment of cmde plant extracts with charcoal powder is also a ISOLATION TECHNIQUES FOR FLA VONOIDS 5 useful method for the preliminary purification of flavonoids, particularly glycosides. In one example of this procedure (Mabry et al., 1970), flavonoids from Baptisia lecontei were eluted from activated charcoal by washing successively with methanol, boiling water, 7% aqueous phenol and 15% methanolic phenol. The bulk of the flavonoid material appeared in the 7% phenol fraction, which was subsequently ether - extracted to give a phenol-free aqueous solution rich in flavonoid glycosides. In certain circumstances dialysis has been found helpful as a preliminary c1ean-up procedure. It is of particular use in the isolation of high molecular weight flavonoids (and tannins) which are water soluble and are mixed with sugars or soluble inorganic material in the crude extract. Markharn (1972) found dialysis useful in the separa• tion of a flavone-polysaccharide compound (MW about 3000) from an aqueous plant extract of Monoclea forsteri, and considerable use has been made of this technique in the isolation of high molecular weight blue flavonoid 'complexes' from cornflower (MW 6200) (Bayer et al., 1966), Commelina communis (Takeda et al. , 1966) and Professor Blaauw iris (Asen et al. , 1970), all of which were non-dialysable through a cellulose membrane. Flavonoids produced on hydrolysis of the non-dialysable commelinin (Takeda et al., 1966) were dialysable under the same conditions.

1.3. Column chromatography Column chromatography remains the single most useful technique for the isolation of large quantities of flavonoids from crude plant extracts. As a technique for separating flavonoids, it has been discussed in depth in a number of previous articles (Seikel, 1962; Mabry et al. , 1970; Harborne, 1959b; Harborne, 1967a). The advent of the relatively new chromatographie media, polyamide and Sephadex, has had a dramatic effect ön the type and efficiency of separations achieved, and 'dry c01umn' techniques (see, TLC, Section 4.1) also seem to offer prospects of markedly improved chromato• graphie separations. These changes appear to at least partly fulfil the requirement expressed by Harborne (1959b) for new adsorbents for column chromatography. Adsorbents commonly used for the separation of flavonoids inc1ude silica gel, kieselguhr, magnesol, cellulose, alumina, polyamide, 6 THE FLA VONOIDS Sephadex and ion exchange resins. The adsorbents of choice have generally been silica gel, cellulose and polyamide.

1.3.1 Silica gel Eluotropic series (Woelm): Light petroleum - carbon tetrachloride - benzene - chloroform (alcohol free) - diethyl ether - ethyl acetate - pyridine - acetone - n-propanol - ethanol - methanol - water. Silica gel has traditionally been used for the separation of isoflavones, flavanones, dihydroflavonols and highly methylated (acetylated) flavones and flavonols. Many examples of such separa• tions are recorded but the following are typical. The isoflavones formononetin, afrormosin and texasin were isolated from extracts of Baptisia australis by chromatography on silica, using chloroform with gradually increasing levels of ether or ethyl acetate as the eluting solvent (Lebreton et al. , 1967; Markharn et al. , 1968b), and isoflavones such as daidzein and scandenone/scandinone (ex Derris scandens) have been eluted from silica columns using benzene-ethyl acetate (6 : 1) (Bhutani et al. , 1969a) and benzene-chloroform (1 : 1) (PeIter and Stainton, 1966) respectively. Methylated flavones from the acetone extract of Andrographis paniculata roots were separated on si li ca, elution with benzene-hexane (3: 1) giving 7,4' -dirn ethyl ether, with benzene giving 5-hydroxy- 7,8,2',3' -tetramethoxyflavone and with chloroform-methanol (20 : 0.8) giving andrographolide (Govindachari et al., 1969). Using chloroform and chloroform-methanol (99.5: 0.5), Thomas and Mabry (1968) were able to separate the highly methylated flavones, hymenoxin, scaposin and demethoxysudachitin on a column of

Si02 • With the amount of methanol in the chloroform increased to 5%(v/v), 6-methoxyluteolin and its methylated derivatives were successfully separated (Kupchan et al., 1969) from extracts of a number of Eupatorium species. Less polar solvents, such as benzene• chloroform (5: 2) and benzene-ether (5: 1), have been used successfully in the separation of the flavones nevadensin (5, 7-OH, 6,8,4'-OMe) and pectolinaringenin (5,6-0H,7,4'-OMe) from the chloroform extract of Iva nevadensis (Farkas et al. , 1966). Occasionally, even flavonoid glycosides have been purified on silica. For example the C- and O-glycosides of daidzein (puerarin and daidzin) were separated on silica using ethyl acetate and ethyl ISOLATION TECHNIQUES FOR FLA VONOIDS 7 acetate-methanol (19 : 1) as solvents (Bhutani et al., 1969a) and glycosides of the naringenin were eluted from silica using benzene-methanol (9 : 1) (Mizelle et al., 1965). Remarkably, even the separation of and 3-p-coumaryltriglycosides has been achieved on silica (Bottomley et al., 1966). However in this case the silica was impregnated with boric acid and the pH adjusted to 7 with KOH. The kaempferol derivative was eluted with n-butanol-boric acid and the quercetin derivative with acetic acid• methanol. It is clear that silica gel is a useful adsorbent for the separation of flavonoids of quite a wide range of polarities. In general terms, this range may be extended to include many of the more polar flavonoids simply by deactivation through the addition of water. In fact many of the variable chromatographic properties observed with silica adsorbents from different sources are undoubtedly attributable to the water content of the gel. An additional factor is the presence of metal ions in the silica. We have observed that a number of the commercially available silica-gels contain a metallic impurity (prob• ably iron) which can cause the more polar flavonoids to adhere strongly to the column. This impurity is removed by treatment of the silica with warm, conc. HCl thus making the adsorbent much more useful for the separation of polar compounds. Other workers (Smith, 1960; Bottomley et al., 1966) who have recommended pre-treatment of silica with HCl have presumably also observed this phenomenon.

1.3.2 Magnesium silicate Eluotropic series: as for silica (Woelm) Magnesol, a hydrated magnesium acid silicate, and Florisil, a 15.5 : 84 mix of magnesium oxide and silica, are somewhat more basic than silica. They have not been used extensively in recent years for the separation of flavonoids probably because they offer little or no advantage over silica. In early work by Ice and Wender (1952), separations of a number of simple mixtures of flavones, flavonols, flavanones and their glycosides were achieved using Magnesol. In particular, good separation of flavonoid aglycones from the more strongly adsorbed glycosides was obtained. More recently Watkin (1960) found that flavonoids applied in water could be displaced by 8 THE FLA VONOIDS water containing up to 5% of an organic solvent such as ethanol, acetone or ether. However, individual flavonoids were not separated by this procedure. Florisil has been used by Clark-Lewis and Dainis (1967) for the partial separation of flavonoids in fractions from counter-current distributions of Acacia obtusifolia extracts. The separations were obtained using acetone-ether (1 : 1) as solvent and led to the partial separation of 7,8,4' -trihydroxyflavanone, 7,8,4' -trihydroxyflavono1 and 'phlobaphene'. In addition these workers used Florisil, with acetone as solvent, for a general 'c1ean-up' procedure. Some separation of isoflavones from Derris scandens on F10risil has been reported (PeIter and Stainton, 1966) using benzene-chloroform (1 : 1) and chloroform. Flow rate problems may be overcome by using magnesium silicate-celite mixtures of about 3 : 1 (Watkin, 1960; Hörhammer et al., 1959).

1.3.3 Alumina Eluotropic series (Woe1m): Light petroleum - carbon tetrachloride• benzene - chloroform (a1cohol free) - diethyl ether - ethyl acetate• pyridine - acetone - n-propanol - ethanol - methanol - water. As with silica-gel, alumina may be deactivated by the addition of water. The so-called 'Brockman scale' of activities (Brockman and Schodder, 1941) ranges from I to V, and these categories correspond to water contents of 1,4, 7, 10 and 19% (w/w) respective1y. Alumina is marketed in three types, acid (pH about 4), neutral (pH about 7.5) and basic (pH about 10). Alumina has generally found little use in the separation of flavonoids, largely because of the problem of complex formation. Aluminium (111) ions are known to complex strongly with the 4-keto-5-hydroxyl and 4-keto-3-hydroxyl systems found in most flavones and flavonols (Porter and Markharn, 1970a; Porter and Markharn, 1970b) and also with o-dihydroxyl groups (Markharn and Mabry, 1968b) on the flavonoid nuc1eus. Thus flavonoids containing any of these functional groups tend to be adsorbed very strongly, so strongly in fact that they generally cannot be eluted at all. Neutral alumina (activity I) has been used successfully for the separation of completely methy1ated and fully substituted flavonols (MukeIjee et al., 1969). The solvents used for elution were benzene, benzene-ch10roform and chloroform. The fully methylated flavone ISOLATION TECHNIQUES FOR FLA VONOIDS 9 zapotin (5,6,2'6'-tetramethoxyflavone) was separated on acid washed a1umina (Dreyer and Bertelli, 1967) and the 5-deoxyisoflavones, 6,7,4'-trimethoxyisoflavone and 7-hydroxy-6,4'-dimethoxyflavone have been separated on an alumina (activity IV) column eluted with benzene (Harborne et ai., 1963). Cambie and J ames (1967) have reported the preliminary purification of biflavonoids, containing some of the functional groups mentioned above, on an alumina column. Chlorophylls and xanthophylls were first e1uted with benzene and ether, and then the co1umn was extruded in order to obtain the yellow biflavonoid band remaining at the top. Soxh1et extraction with ethyl acetate or methanol was found to remove some biflavonoids from this band but others were exceedingly difficult to extract. I t is c1ear that alumina is best suited to the purification of fully derivatized flavonoids. The successful separation of 5-deoxyiso• flavones (Harborne et al., 1963) suggests that as an adsorbent it might also be useful for the se1ective separation of certain 5-deoxy• flavonoids (not flavono1s) from their 5-hydroxylated equiva1ents. In princip1e flavonoids which form only acid-labile complexes with aluminium, for example 3' ,4'-dihydroxyflavones, 6,7- or 7,8-di• hydroxyflavanones, and dihydroflavonols lacking 5-hydroxylation (Markham and Mabry, 1968b; Porter and Markham, 1972), should be selectively removed from a1umina with acid containing solvents, although no report of this application seems to have appeared to date. 1.3.4 Cellulose Cellulose column chromatography can be considered as a scaled-up form of paper chromatography. As such, it is suited to the separation of all c1asses of flavonoids and their glycosides. I t is used for separations based on both absorption and partition (depending upon the solvent used), although a distinction between the two is often difficult to make (Cassidy, 1957). In principle the full range of solvents deve10ped for use in paper partition chromatography is availab1e for partition column chromatography and many have been used with success (Seike1, 1962). Rarely, however, have separations comparab1e to those obtained by paper chromatography been achieved by this method. Cellulose powder has a 10w capacity (Seike1, 1962; Harborne, 1967b) and limited reso1ving power and 10 THE FLA VONOIDS although this may be compensated for to some extent by tight column packing, many workers favour scaling up of paper chro• matography itself for large scale separations. The use of dextran gels in the adsorption mode (see section 3.6) has the potential to overcome many of these limitations. Notwithstanding this, in recent years there has been a noticeable trend away from cellulose to polyamides which offer equivalent versatility but much higher capacity and resolving power. As in paper chromatography, the solvents most favoured for use with cellulose columns are of the aqueous alcohol and acid types. Some typical examples include the separation of flavones from Spartium junceum flower extracts by elution with water-saturated butanol (Spada and Cameroni, 1958), the isolation of flavonols and their glycosides using the solvents ethyl acetate-methanol• water (50: 3 : 10) and ethyl acetate-methanol-acetone-water (50 : 3 : I : 10) (Olechnowicz-Stepien, 1967; Bodalski and Cisowski, 1967), and the partial separation of complex mixtures of flavone• C-glycosides from Vitex lucens extracts (Seikel et al.. 1966) using 5% acetic acid and BAW (n-butanol-acetic acid-water, 4 : I : 5). In this latter example, rolled paper columns were also used and it was claimed that in general, cellulose columns were preferable to polyamide for the separation of these compounds (Seikel et al" 1966; Seikel and Bushnell, 1959). C-Glycosides of the lucenin and vicenin types have recently been isolated from extracts of the green alga Nitella hookeri, using cellulose column chromatography with water as eluent (Markharn and Porter, 1969). Particularly effective use was made of cellulose chromatography by Clark-Lewis and Porter (1972) in the separation of heartwood flavonoids of Acacia species. In this application the ether extractives of A. kempeana (later corrected to A. cuthbertsoni) heartwood were applied to a cellulose column in 2% aqueous acetic acid. Elution with this solvent yielded constituent flavonoids in the following order: -3,4-diols, dihydroflavonols, flavanones, 3-methoxyflavones, and flavonols and chalcones. These flavonoids consisted of three basic substitution types: 7,8,4' -trihydroxy-, 7,8,3' ,4' -tetrahydroxy-, and 7 ,8-dimethoxy-3' ,4' -dihydroxy-, and some separation of these types, within each flavonoid group, was observed. For example 7,3',4'-trihydroxy-3,8-dimethoxyflavone (eluted with 2% acetic acid) preceded 3,7,3' ,4'-tetrahydroxy-8-methoxyflavone which in turn ISOLATION TECHNIQUES FOR FLA VONOIDS 11 preceded 3,7,8,3'4'-pentahydroxyflavone (both eluted with aqueous methanol). Cellulose column chromatography has been used for the separation of anthocyanins in the past, but variable results, limited column capacity and production of dilute eluates, led Harborne (l967b) to conclude that the method offered little if any advantage over paper chromatography.

1.3.5 Polyamide Eluotropic series (Endres, 1969): Water-methanol-acetone-aq. sodium hydroxide - formamide - dimethylformamide - aq. urea (Andersen and Sowers, 1968). Polyamides commercially available for chromatography are mainly of the Perlon-type (polycaprolactam), Nylon-type (polyhexamethyl• enediamine adipate) or Polyclar-type (polyvinylpyrrolidone, PVP). All have a high capacity for phenolic materials and all form strong hydrogen bonds with phenolic hydroxyl groups via their amide carbonyl functions (Andersen and Sowers, 1968; Endres and Hormann, 1963; Endres, 1969). The elution of phenolics from columns of polyamide depends either on the ability of the solvent to rep1ace the phenol on the hydrogen bonding site, in which ca se hydrophilie solvents are used, or on the ability of the solvent to form even stronger bonds with the phenol than does the polyamide (e.g. urea). The strength of attachment of phenolic compounds depends to a large extent on the number and type of hydrogen bonds formed (Endres, 1969; Achrem and Kuznetsova, 1963). Although polyamides are readily available they are very often not in a physical (or chemical) form ideal for column chromatography. With some (e.g. Polyclar AT* and Polypenco 66Dt) simple sieving to remove fines sm aller than 0.002 cm is sufficient to produce an acceptable adsorbent (Markharn and Mabry, 1968a). With others it is necessary to convert pellets (e.g. Durethan BK 40F:j:) into a powder by dissolution in hot conc. HCl and controlled precipitation with methanol-water (Mabry et al., 1970; Rösler, 1960; Hörhammer,

*General Aniline and Film Corp., Polyvinylpyrrolidone tPolymer Corp. Pennsylvania :j:Bayer Co., West Germany. Polycaprolactam 12 THE FLA VONOIDS

1964; Wyler et al., 1967). Strict1y controlled conditions are essential for the production of a powder which has no water/methanol soluble monomers or oligomers, uniform grain size, forms a column with a satisfactory flow rate and which has a high adsorption capacity . However, to some extent flow rate difficulties may be overcome by admixture with celite (Mabry et al., 1970; Okigawa et al., 1971 b) and low molecular weight polyamide may be partly removed by pre-washing with 50% aqueous methanol (Mabry et al. , 1970) or 10% HCl (McFarlane and Vader, 1962; Van Teeling et al., 1971). Like cellulose, polyamide is suitable for the separation of all types of flavonoids (Hörhammer, 1964; Mabry et al., 1970). However, it has the advantages over cellulose of higher capacity and higher resolution. In addition, solvents different from those commonly used in paper chromatography are employed and hence the two methods complement, rather than duplicate, each other. The type and extent of separation possible on a single application of a crude plant extract to a polyamide column is exemplified by the separation of the flavonoids in Baptisia lecontei (Markharn and Mabry, 1968a). This is presented graphically in Table 1.1, and the elution order highlights a number of factors which influence chromatography on this adsorbent. For example, within the same dass of flavonoid, the diglycosides precede the monoglycosides which in turn precede the aglycones. This is true for the isoflavones, flavones, dihydroflavonols () and flavonols (and cou• marins). Further, the inhibiting effect of increasing hydroxylation on the rate of elution is demonstrated by the isoflavone sequence: pseudobaptigenin glycosides (0) > calycosin (1) > daidzein (1) > genistein (2) > orobol (3). Some separation of flavonoid types is also apparent in that isoflavones tend to precede dihydroflavonols which precede flavones which in turn precede flavonols. The current literature abounds with examples in which polyamide column chromatography has been used for the separation of flavonoids, the following being typical. A number of excellent separations of anthocyanins (and anthocyanidins) have recently been achieved. Using an acid-washed PVP Van Teeling et al. (1971) separated anthocyanins from co-occurring flavonols, flavonol glyco• sides and anthocyanidins, all of which moved more slowly than the anthocyanins when 0.25N HCI was used as solvent. With the same system, good separations of the following were also achieved: (i) ISOLATION TECHNIQUES FOR FLA VONOIDS 13

Table 1.1 Composition of fractions from PVP column chromatography of a B. Iecontei extract (Markham and Mabry, 1968a).

SO LVENTS ..,. ..., .... :=> 7" ::r: N N N N N W W VI 0\ 0 w VI 0 0 0 0 0 0 0 0 0 0 VI 0 " 0 Cfl.~~~~~,*~~~C)Q.~ Z Z z 2l: 2l: 2l: 2l: 2l: 2l: 2l: 2l: 2l: 2l: 2l: 2l: G G

+ + + Pseudobaptigenin 7-O-rhamnoglucoside + Calycosin 7-0-rhamnoglucoside + Daidzein 7-0-rhamnoglucoside + + + Genistein 7-0-rhamnoglucoside + + + 7,4'-Dihydroxyflavone 7 -O-rhamnoglucoside + + + + Calycosin 7-0-glucosidc + + Daidzein 7-O-glucoside + + + + 7,3',4'-Trihydroxyflavone 7-O-rhamnoglucoside + + + + + Orobol7-O-rhamnoglucoside + + + + + + 7,4'-DihydroxyflavanonoI3-O-glucoside + + + + + Apigcnin 7-0-rhamnoglucoside + + + 7,4' -Dihydroxyflavone 7-O-g1ucoside + + + + + + + 7-O·rhamnoglucoside + Luteolin 7-0-glucoside + + 3-0-glucoside + + 7,4' -Dihydroxyflavanonol + 7-0-rhamnoglucoside + Fisetin 3-0-glucoside + + + Pseudobaptigenin + + + Daidzein + + Calycosin + + + 7,3',4'-Trihydroxyflavone 7 -O-glucoside + Fustin + + + + + 7,4'-Dihydroxyflavonol + + Fisetin

mono-, di- and triglycosides of the same (ü), the same glycosides of different anthocyanidins and (üi) anthocyanidins differing only in B-ring substitution. In the case of the free anthocyanidins, the rate of elution depended upon the availability of free phenolic hydroxyl groups for hydrogen-bonding, the order of elution being malvidin (3' ,5'-OMe, 4'-OH) > peonidin (3:-0Me, 4'-OH) > pelargonidin (4'-OH) > petunidin (3'-OMe, 4',5'-OH) > 14 THE FLA VONOIDS (3',4'-OH) > delphinidin (3',4',5'-OH). Similarly, Hrazdina (1970) used PVP for the separation of the 3,5-diglucosides of malvidin, peonidin, cyanidin and delphinidin from grape juice. Glycosides of isoflavones, flavanones, dihydroflavonols, flavones and flavonols have also been successfully chromatographed on polyamide (e.g. Markharn and Mabry, 1968a; Härhammer et al. , 1961; Härhammer et al. , 1962). In the isolation of flavanone glycosides from grapefruit (Mizelle et al., 1965), the non-aqueous solvent mixtures benzene-methanol, (3: I), (7: 3), (6: 4) and (l : 1) were used to good effect to separate naringenin ru tinoside from naringenin neohesperidoside on a PVP column. Flavonol glycosides are separated satisfactorily using water-methanol mixtures (Bhandari, 1964; Bandyukova and Shinkarenko, 1966) as also are isoflavone glycosides (Beck and Knox, 1971; Markharn and Mabry, 1968a) and flavones (Markharn et al., 1969; Markharn and Mabry, 1968a). For the separation of flavonoid aglycones, non-aqueous solvents seem to be the most satisfactory. For instance, a mixture of mono-, di- and tri-methylated kaempferol and quercetin derivatives from Larrea cuneifolia was resolved on polyamide with Egger's solvent (chloroform: methanol: butanone : 2,4-pentanedione, 20 : 10 : 5 : 1) as eluent (Valesi et al., 1972); methylated luteolin and quercetin derivatives were separated (Abdel-Baset, 1973) by the use of chloroform-ethyl acetate (3 : 1) followed by increasingly more polar solvents such as chloroform-methanol-butanone-acetone (20 : 10 : 5 : 1) and (10: 10 : 5 : 1) and chloroform-methanol (l : 3); and flavonols from the fern Pityrogramma chrysoconica were separated using benzene with increasing quantities of butanone (and methanol) as solvent (Wollenweber, 1972). Aurones and chalcones (Asen and Plimmer, 1972) and flavonols (Markharn and Mabry, 1968a) have been purified on a PVP column using methanol as eluting solvent. However, water-methanol combinations were re• quired for the elution of biflavones (ex Selaginella) from a polyamide-celite column (Okigawa et al., 1971b), and for the separation of flavanones from chalcones (Neu, 1959).

1.3.6 Sephadex gel Sephadex is a highly cross-linked dextran on which separations are ideally obtained on the basis of molecular size (Fischer, 1969; Gelotte, 1960). Thus substances are eluted in order of decreasing ISOLA TION TECHNIQUES FOR FLA VONOIDS 15 molecular size. The gel must be swollen in water prior to use (Fischer, 1969; Gelotte, 1960) and the extent to which individual gels swell (water regain) determines the molecular weight range of compounds which can be separated on that gel (see Table 1.2 for Sephadex G series). However, solvents other than water generally swell the gel to a lesser extent (Porter and Wilson, 1972; Nordstrom, 1967), and compounds other than dextran may have different affinities for the gel, thus introducing effects due to adsorption (Gelotte, 1960; Nordstrom, 1967; Repas et al. , 1969; Sachs and Painter, 1972). The hydroxypropylated dextran gel, Sephadex LH-20, is designed for use with organic solvents or water/solvent mixtures and the exclusion limit for this gel is normally considered to lie between MW 2000 and 10000 (Fischer, 1969). Both of these Sephadex types have been used with success in the separation of flavonoids. Adsorption on dextran gels is known to occur with aromatic compounds, and phenols in particular, and it is thus not surprising that this is commonly encountered with flavonoids, especially the aglycones. Adsorption of flavonoids appears to be associated particularly with the free phenolic hydroxyl groups. Hence Johnston et al. (1968), using Sephadex LH-20 with methanol as solvent, observed a general correlation between the elution volume of flavonoid aglycones and the number (but not the acidity) of the free phenolic hydroxyl groups present (see Table 1.3); and Woof and Pierce (1967) were unable to remove a wide range of flavonoid aglycones (including flavonols, dihydroflavonols, flavanones and

Table 1.2 Water regain and [ractionation range tor Sephadex G. *

Sephadex type Water regain Fractionation range (g H2 Ojg Sephadex) for Dextrans (MW)

G-IO 1.0 0-700 G-15 1.5 0-1 500 G-25 (medium) 2.5 100-5000 G-50 (medium) 5.0 500-10000 G-75 7.5 1000-50000 G-IOO 10.0 1000-100 000 G-150 15.0 1000-150 000 G-200 20.0 1000-200000

*Data taken from 'Sephadex-gelfiltration in theory and practice', Pharmacia. 16 THE FLA VONOIDS

Table 1.3 Ve/Vo o[ Flavonoids on Sephadex LH-20 in methanol*

Compounda Substituents

Flavones and Flavonols Apigenin 5,7,4'-OH 5.3 Luteolin 5,7,3',4'-OH 6.3 Resokaempferol 3,7,4'-OH 5.9 Fisetin 3,7,3',4'-OH 6.6 Robinetin 3,7,3',4',5'-OH 7.4 Kaempferol 3,5,7,4'-OH 7.5 Quercetin 3,5,7,3',4'-OH 8.3 3,5,7,2',4'-OH 4.4 3,5,7,3',4',5'-OH 9.2 3-0-Methylquercetin 5,7,3',4'-OH; 3-0Me 5.6 3,7,3',4'-OH; 5-0Me 5.9 3,5,7,4'-OH; 3'-OMe 7.0c 3,5,7,3'-OH; 4'-OMe 7.3 3,5,3',4'-OH; 7-0Me 7.7 Penta-O-methylquercetin 3,5,7,3',4'-OMe 3.4 Hexa-O-methylmyricetin 3,5,7,3',4',5'-OMe 3.1 Quercetin penta-acetate 3,5,7,3',4'-OAc 2.7 5,7,3',4'-OH; 3-rhamnoside 4.9 5,8,3' ,4' -OH; 3-rhamnoglucoside 4.0 3',4',5'-OH; 3-galactorhamnoside; 7-rhamnoside 3.3 Flavanones and Dihydroflavonols Naringenin 5,7,4'-OH 5.4 Eriodictyol 5,7,3',4'-OH 5.8 3,5,7,3',4'-OH 5.5 Hesperetin 5,7,3'-OH; 4'-OMe 5.2c Naringin 5,4'-OH; 7-rhamnoglucoside 3.6 5,7,3',4'-OH; 3-rhamnoside 4.2 Hesperidin 5,3'-OH; 4'OMe; 7-rhamnoglucoside 3.4c (+) -Catechin 3,5,7,3',4'-OH 5.2 aSamples: 2.5 mg/D.S ml; flow rate, 3-5 ml min,-l bUnder these conditions, Ve/Vo = 2.2 K D + l. cSample dissolved in dioxane-methanol (I : 1). *Reproduced by permission of the original authors (Johnston er al., 1968). catechin) from Sephadex G-25 with water or 0.5 M sodium chloride due to adsorption, and thus had to use 0.1 M ammonium hydroxide for elution. Where adsorption of flavonoids is an overriding factor, molecular size plays Httle or no part in the separation. Separations based on adsorption nevertheless are still very worthwhile and have been used successfully for the separation of ISOLATION TECHNIQUES FOR FLA VONOIDS 17 most classes of flavonoid. Specific examples include the separation of: anthocyanins on G-25 in 60% aqueous alcohol (+ acid) (Somers, 1966b) and in aqueous acetone (+ acid) (Somers, 1967), and on LH-20 in methanol (+ acid) (Asen et al. , 1970); flavanol digallates on LH-20 in methanol-chloroform (1 : I) and methanol-chloroform• light petroleum (2 : I : 1) (Coxon et al. , 1972); flavone and flavonol glycosides on G-I0 in methanol-water (1 : 1) (Nordstrom, 1967); molybdate complexes of flavonols, flavanones and dihydroflavonols on G-25 in 0.001 M sodium molybdate (Woof and Pierce, 1967); the isoflavones daidzein, genistein, formononetin and biochanin A on G-25 in 0.1 M ammonium hydroxide (Nilsson, 1962); flavan-3-01s, catechins, procyanidin dimers and procyanidin oligomers, in that order on LH-20 in ethanol or ethanol-propan-I-ol (1 : 1) (Thompson et al., 1972): and a wide range of flavonoids and their glycosides on LH-20 in methanol (Johnston et al., 1968). The separation of flavonoids achieved in the last example above is detaHed in Table 1.3, where relative elution rates are presented as Ve/Va (Vc = elution volume or total amount of solvent required to elute sampie, and Va = void volume of column). Other workers (Nordstrom, 1967; Woof and Pierce, 1967) have preferred to present elution rates in terms of K n (Gelotte, 1960), the partition coefficient between the two phases Vi and Va, where Vi = the inner volume within the gel grains, and Va = the void volume sUITounding the gel grains.

Hence Ve = Va + Kn Vi Ve - Va and K n = V- I Thus when a high molecular weight compound is completely excluded Ve = Va, Ve/Va = 1 and Kn = 0 and when a low MW compound is retained either by partition or by adsorption, Ve > Va, Ve/Va > land Kn > O. In Table 1.3, the high Ve/Va values for quercetin and myricetin (8.3 and 9.2), when compared with the low Ve/Vo values for their methyl ethers (3.4 and 3.1), clearly indicate that adsorption is a major factor in the retention of these aglycones on the column. The problem of adsorption may sometimes be overcome by a change of solvent. For example, Somers (1966a) was able to 18 THE FLA VONOIDS eliminate adsorptive effects by the use of aqueous acetone mixtures for the elution of wine tannins, and King and Pruden (1970) noted a decrease in the tendency of tannins to be adsorbed when water was replaced by an acetone-water mixture. Conversely, the use of methanol-water by Nordstrom (1967) failed to eliminate adsorption of flavone glycosides on G-l O. With flavonoid glycosides the effect of adsorption does not seem to dominate, as it does with the aglycones. For example, the Ve/Vo values (Table 1.3) for the flavonol glycosides, quercitrin, rutin and robinin (4.9, 4.0,3.3) and the flavanone glycosides, astilbin, naringin and hesperidin (4.2, 3.6 and 3.4) are in the relative order expected on a molecular weight basis and so suggest that molecular sieving is at least a contributing factor in their separation. Abdel-Baset (1973) considered molecular sieving to be an important factor in the separation of Vernonia baldwinii flavonoids on LH-20 in methanol. On elution, a quercetin 3,7-diglycoside preceded luteolin 7- and 7,4'-diglycosides, which in turn preceded quercetin 3-g1ucoside, and luteolin and apigenin 7-g1ucosides, in that order. The further fractionation of these compounds using t-butanol-methanol (3 : 1) was thought to involve less molecular sieving, since it brought about the separation of luteolin 7-g1ucoglucoside from luteolin 7-galacto• glucoside, and from luteolin 7-arabinoglucoside. However molecular sieving of flavonoid glycosides does appear to be a fairly general phenomenon in solvents such as water (Woof and Pierce, 1967) methanol (Johnston et al., 1968; Abdel-Baset, 1973), water-acetone (Porter and Wilson, 1972) or water-methanol (Nordstrom, 1967). It is probable that the higher the ratio of sugar to flavonoid in flavonoid polyglycosides, the greater is the contribution of molecular sieving to the separation. Thus, the recently reported separation of a flavone-polysaccharide compound (MW about 3000) from con• taminating polysaccharides and other flavone-polysaccharide material, on Sephadex G-SO with water as solvent (Markharn, 1972), probably involved little, if any, significant adsorption to the gel. Approximate MW values may be obtained from Sephadex gel filtration in cases where separation is based primarily on molecular size (Fischer, 1969). These are established by measuring the elution volume of the unknown and by applying this figure to the straight line graph of elution volume versus log MW, predetermined for a range of chemically similar compounds. Such a graph has been published ISOLATION TECHNIQUES FOR FLA VONOIDS 19 for a range of dihydroflavonoids by Porter and Wilson (1972), who established an exclusion limit of about MW 1000 for these compounds on Sephadex G-25 with 50% aqueous acetone. This exclusion limit differs markedly from that determined by King and Pruden (1970) using dyestuffs as MW standards, and this emphasizes the necessity to use as standards only compounds chemically similar to the unknown whose approximate MW is sought. Thus, the assumption that the exclusion limit is directly proportional to the percentage swelling of the gel, which has been used in the estimation of tannin molecular weights (Somers, 1966b; F orrest and Bendall, 1969b), is not valid unless compound type is also taken into consideration (Porter and Wilson, 1972). The method of determining an approximate MW by applying the sam pie to a range of gels with pre-established exclusion limits, has also been used with tannins (Somers, 1966b; Forrest and Bendall, 1969b). The use of Sephadex gels for large scale separation of flavonoids has not as yet been fully explored. Johnston et al. (1968) successfully separated 250 mg mixtures of quercetin and rutin on LH-20 and Porter (1973) has separated a 2 g mixture of catechin and dihydroquercetin on G-25 (column dimensions 30 cm x 2 cm) using water saturated sec-butanol. Even higher loadings have been used in the isolation of procyanidins on LH-20 (Thompson et al. , 1972), and the preliminary separation of luteolin and quercetin glycosides by Abdel-Baset (1973) mentioned earlier involved the application of 6 g of a crude methanol plant extract to a column of 350 g of LH-20. There is no reason therefore why upgrading of this technique should not be successful, especially when the gel is operating as an adsorbent. The G-series gels, due to their chemical nature, would be expected to be closely similar to cellulose in their general chromato• graphie properties. However, due to the regular spherical form of the particles these gels should offer markedly improved capacity and resolution (as compared with the essentially fibrous cellulose) when organic or slightly aqueous organic solvents are used. Thus they appear to have considerable potential as substitutes for cellulose in cases where high capacity or high resolution is required.

1.3.7 Ion exchange resins Ion exchange resins comprise a polymerie matrix to which are attached functional groups capable of ionization to provide fixed 20 THE FLA VONOIDS charge sites on the resin. These fixed charge sites are associated with bound ions which are replaceable with ions of the same charge in the eluting medium. Ion exchange resins are available in both cation and anion types and each of these types may act essentially as exchangers, sorbents or as catalysts (Cassidy, 1957; Webster et al., 1967). Ion exchange resins have been used very little in recent years for the isolation of flavonoids. Early work, which has been thoroughly reviewed by Seikel (1962), involved the use of cation-exchange resins en tirely. These resins were gene rally used in preliminary clean-up procedures in which flavonoids were held on the column while other water soluble impurities were washed off with water. The flavonoid was subsequently eluted with methanol. This method was found useful (Gage et al., 1951; Williams and Wender, 1952; Williams and Wender, 1953) for concentrating large volumes of dilute aqueous solutions of flavonoids. Wender carried out much of the early work (see Cassidy, 1957), and used the weakly acid Amberlite IRC50 cation exchanger for the purification of a wide variety of flavonoids as their salts or complexes. For example, rutin was applied to the column in the form of an aluminium chloride complex or as a salt (in potassium acetate) and morin was applied as a lead or potassium salt (Wender, 1954; Morris et al., 1951). The pure flavonoid was then eluted with ethanol after thorough washing of the column with water. Perhaps more suited to purification by ion exchange are the naturally occurring and readily ionizable flavonoid sulphates. The recent isolation and. purification of quercetin 3,7,3',4'-tetrasulphate (from Flaveria bidentis) (Pereyra and luliani, 1972), by the use of the strongly acid Amberlite IR 120(H) followed by Amberlite IRC50(Na) to yield the pure sodium salt, demonstrates this nicely. Mixtures of flavonoids have been separated on Amberlite CG50(H), a cation exchanger, by application in aqueous solution and elution with increasing amounts (5-40%) of isopropyl alcohol (Vancraenenbroeck et al. , 1963; Seikel, 1962). This type of separation however is probably due to adsorption or partition rather than ion exchange since under these conditions the flavonoids would be virtually unionized. In this category too, can probably be placed the impressive separation reported by Hori (1969). Hori used Amberlite XAD-2 for the separation of a wide variety of flavones, flavonols, flavanones, dihydroflavonols and their glycosides, applying ISOLATION TECHNIQUES FOR FLA VONOIDS 21 the sampies in water and using a gradient elution system with increasing amounts of ethanol. Amberlite XAD-2 is a resin matrix lacking ionic groups and as such presumably produces separations based on phenomena other than ion exchange. Nevertheless the separations achieved by Hori by controlling temperature and flow rate appear to be most useful. Recovery from the resin was quantitative and in general, the more polar compounds were eluted first, e.g. flavone uronide esters preceded flavonoid glycosides which in turn preceded aglycones. It is rather surprising that there appears to have been no attempt to use anion exchange resins for the separation of flavonoids. In principle a basic anion exchange resin will retain flavonoids applied as anions. These could then be eluted selectively by gradient elution with buffers of decreasing pH. Thus it should be possible to separate flavonoids on the basis of the pKa values of their free phenolic hydroxyl groups. Such separations have been most successful with simpler phenols (Skelly, 1961; Logie, 1957; Thomas and Thomas, 1969) using resins such as De-acidite FF (acetate form) and Dowex 2-X8 (acetate form). In these cases the phenols were ionized by amines in methanol solution for application to the resin. Selective elution according to pKa was then achieved by graded elution with acetic acid-methanol mixtures, or more effectively, by pH control using triethylamine-acetic acid or diethylamine-acetic acid buffer solutions in methanol.

1.4 Thin-Iayer chromatography Numerous excellent review artic1es have recently been published on the general techniques of thin-layer chromatography (TLC) (Kirchner, 1967; Stahl, 1969; Truter, 1963; Stahl and Mangold, 1967), and for this reason the comments in this artic1e will be confined largely to the types of flavonoid separation achieved by this method. TLC is a technique which has developed rapidly during the last decade and to a limited extent it has replaced paper chromatog• raphy in analytical and small scale separations of flavonoids. However, it is also complementary to paper chromatography in that it provides new media for the separation of flavonoids on a small scale, and permits the use of a wider variety of detecting reagents. As in column chromatography, the adsorbents of choice for the 22 THE FLA VONOIDS separation of flavonoids are silica, polyamide and cellulose. The mechanism of separation and the eluotropic series for each chro• matographie medium are as described earlier for column chromatog• raphy, but the solvent systems used often vary widely from those used for columns. In general, flavonoids are much more strongly held on thin-layers and as a result solvents of higher polarity are required for their elution.

1.4.1 Preparative scale TLC TLC is essentially a technique for the separation of milligram quantities of material. However, it can be upgraded to handle up to a gram when layers of from 1-5 mm thick are used in conjunction with plates of up to 20 x 100 cm in size (see Kirchner, 1967; Stahl, 1969; Stahl and Mangold, 1967). A more promising method is that of 'dry column' chromatography (Loev and Snader, 1965; Loev and Goodman, 1967; Loev and Goodman, 1970), which involves the use of columns packed with fine-powdered, dry adsorbent. Although this has not as yet been used for the separation of flavonoids, resolution with other compounds is c1aimed to be as good as TLC when low sampIe loadings (e.g. 1 : 500 (Truter, 1963)) are used.

1.4.2 Spray reagents and detecting methods Apart from the anthocyanins and some of the more intensely coloured cha1cones and aurones, flavonoids are not sufficiently coloured to be visible to the naked eye on a thin-layer plate; thus, some form of visualization is necessary for spot detection. In many cases this is achieved by viewing the plate in UV light (366 nm) either in the presence or absence of ammonia vapour (Mabry et al., 1970). Detection under UV is often assisted by the use of layers which contain a UV-fluorescent indicator (e.g. silica gel GF2 54). Flavonoids appear as dark spots against a fluorescent green back• ground. Another useful method of detection is brief exposure of the plate to iodine vapour which produces yellow-brown spots against a white background with most flavonoids. These techniques all have the advantage that they are non-destructive. Most flavonoids are detectable by one or other of the above techniques; however, spray reagents are available and some of the more frequently used are listed in Table 1.4. In principle this list of spray reagents could be extended to inc1ude most of those developed ISO LA TION TECHNIQUES FOR FLA VONOIDS 23 for use in paper chromatography (see e.g. Harborne, 1959b, 1967a; Seikel, 1962).

1.4.3 Separations on TLC, general In recent years at least two major reviews (Egger, 1969; Kirchner, 1967) and several minor compilations have been published (Har• borne, 1967a; Mabry et al., 1970; Harborne, 1967b; Randerath, 1963) on the subject of the separation of flavonoids by TLC. The data reviewed in these publications are summarized and updated in the present artic1e.

1.4.4 Separations on silica gel Flavones, Flavonols and Biflavonoids. Highly methylated or acetyl• ated flavones and flavonols require relatively non-polar solvents for

TLC on silica gel (Si02 ). Thus, flavones such as hymenoxin (5,7-diOH, 6,8,3',4'-tetraOMe), scaposin (5,7,3'-triOH, 6,8,4',5'• tetraOMe), and demethoxysudachitin (5,7,4'-triOH, 6,8,-diOMe) have been chromatographed using chlorofonn-methanol (15 : 1) (Thomas and Mabry, 1968), and digicitrin (5,3'-diOH, 3,6,7,4',5'• pentaOMe) and a number of related compounds were separated using benzene-ethyl acetate (3: 1) (Meier and Fuerst, 1962). Flavonol polyacetates and polymethyl ethers have been successfully chromato• graphed by Egger (1969) using benzene-acetone (9 : 1) and toluene• acetone (19: 1). Other highly methoxylated flavones and flavonols have been chromatographed with benzene-methanol-n-butyl acetate (20 : 4 : 1) (Govindachari et al. , 1965), hexane-acetone-n-butano1 (8 : 1 : 1 and 17 : 2 : 1) (Tatum and Berry, 1972), benzene-acetone (3 : 1), (9 : 1), (49 : 1), (92.8 : 7.2) (Mukerjee et al. , 1969; Egger, 1969; Tatum and Berry, 1972), and chloroform-ethyl acetate (1 : 1) (Dreyer and Bertelli, 1967), the degree of solvent polarity depending 1argely upon the extent of methylation in the flavonoids. More polar flavones and flavonols require more polar solvents. Thus apigenin, luteolin, , kaempferol, quercetin, myricetin, isorhamnetin (3,5,7,4',tetra-OH, 3'-OMe), datiscetin (3,5,7,2'• tetraOH) and morin (3,5,7,2' ,4' -pentaOH) separate weIl in toluene• chlorofonn-acetone (8 : 5 : 7) (Egger, 1969), Rf values being 0.43, 0.28,0.62,0.39,0.27,0.13,0.26,0.36 and 0.06 respectively. Similar flavonoid mixtures were separa ted by Hörhammer eta!. (1964) using benzene-pyridine-formic acid (36: 9 : 5). We have found IV """

Table 1.4 Spray reagents for the detection of fiavonoids on thin-Iayer plates

Reagent* Referencet Flavonoid type detected

1. Diphenyl-boric acid-ethanolamine Dass and Weaver, 1972; Quarmby, 1968; complex (1 % in ethanol) Gupta, 1968b; Neu, 1957. 2. Ferric chloride (methanolic) Spiegi, 1969; Kirchner, 1967; Beck and Knox, 1971 3. Ferric chloride-potassium ferricyanide Dass and Weaver, 1972; Quarmby, 1968; (1 % aq. solns. mixed 1 : 1) Kirchner, 1967; Hillis and Isoi, 1965 most fIavonoids 4. Ferric chloride, then cx.,o/-dipyridyl Barton, 1965 5. Antimony chloride (10% in CHC13 ), UV Hörhammer et al., 1964 6. Lead acetate-basic (25% aq.), UV Hörhammer et al., 1964 7 Aluminium chloride (2% in MeOH), UV Gupta, 1968b; Spiegi, 1969; Silvestri, 1970; Dass and Weaver, 1972 8. Zirconium oxychloride (2% in MeOH), Stahl, 1969 UV 9. Ceric sulphate (70% in conc. H2 S04) Cambie and James, 1967; Kawano et al., bifIavonoids 1964 10. Sulphuric acid (conc.) Camp bell et al., 1969 isofIavones 11. Hydrogen chloride (gas) Dreyer and Bertelli, 1967 fully methylated fIavone 12. Zinc-hydrochloric acid Barton, 1968 dihydrofIavonols 13. Oxalic acid (10% in acetone: Kirchner, 1967 anthocyanins and H2 0,1 : 1) anthocyanidins 14. Diazotized sulphanilic acid Smith, 1960; J angaard, 1970 15. Diazotized p-nitroaniline Hillis and Isoi, 1965; J angaard, 1970 16. Fast Blue Salt B (tetra-azotized Pastuska, 1961; Hörhammer and most flavonoids di-o-anisidine) 0.5%, then 0.1 N Wagner, 1962 NaOH 17. Fast Red Salt B (diazotized 5- Pastuska, 1961 nitro-2-aminoanisole) 0.5%, then O.IN NaOH 18. Bis-diazotized benzidine F orrest and Bendall, 1969a and flavonols 19. (1 % in Harborne, 1967a; Gupta, 1968b; flavanones, isopropanoI), HCI vapour or AICh Hillis and Isoi, 1965; isoflavanones 20. Sodium carbonate (5% aq.), UV Gupta, 1968b most flavonoids (not isoflavones, flavanones) 21. Ammonia (vapour), UV Dass and Weaver, 1972; Kirchner, 1967; most flavonoids Mabry et al., 1970 (except some isoflavones, flavanones) 22. Ammoniacal silver nitrate Harborne, 1967a o-dihydroxyflavonoids 23. Sodium hydroxide solution (1 % Dass and Weaver, 1972 most flavonoids in MeOH

*Most reagents, but not 10 or 11, can be used for detecting flavonoids on paper chromatograms. t References are to examples of the use of the reagent and not necessarily to the original work.

I'-l Vl 26 THE FLA VONOIDS chloroform-methanol (96 : 4) useful for distinguishing flavones such as apigenin, and luteolin, and Hillis and Isoi (1965) have had success with chloroform-acetie acid (9: 1) and toluene-ethyl formate-formic acid (5 : 4 : 1) for the chromatography of C-methyl• flavones such as sideroxylin and eucalyptin (the 7- and 7,4'-methyl ethers of 6,8-di-C-methylapigenin respectively). Other solvents, such as benzene-dioxan-acetic acid (90: 25 : 4) (Voirin, 1972), benzene• pyridine-ammonia (80 : 20 : 1) (Voirin, 1972) and acetone-benzene (1 : 3) (Radhakrishnan et al. , 1965) have also been used with flavones and flavonols. Biflavonoids are chromatographed in solvents similar to those above; for example, toluene-ethyl formate-formie acid (5 : 4 : 1) (Cambie and James, 1967; Kawano et al. , 1964) and benzene• pyridine-formic acid (36 : 9 : 5) (Chexal et al., 1970). Fully methyl• ated biflavonoids have been separated by Chexal et al. (1970) in benzene-pyridine-ethyl formate-dioxan (5 : 1 : 2 : 2). This solvent even permitted separation of the isomerie pairs of hinokiflavones (i.e. those containing the 4-0-8' and 4-0-6' interflavonoid linkages) and of amentoflavones (i.e. with the 3-8' and 3-6' linkages). Flavone and flavonol glycosides are not commonly chromato• graphed on Si02 • However, where this has been done, polar solvents such as ethyl acetate-butanone-formie acid-water (5: 3 : 1 : 1), n-butanol-2N HCI (1 : 1, upper phase) and isoamyl a1cohol-acetic acid-water (2: 1 : 1) have been used (Stahl and Schom, 1961; Kawano et al. , 1964; Egger, 1969; Gupta, 1968b). Mono-, di-and triglycosides of a variety of flavones and flavonols have been chromatographed in the first of these solvents (Egger, 1969) but distinction of flavonoids within each group was not good. However, monoglycosides (with Rf values of 0.46-0.65) were clearly separated from diglycosides (with Rf values of 0.21-0.36) and triglycosides (with Rf values of 0.06-0.16). Chopin (1971) found

Si02 more sensitive than paper for distinguishing flavone C-glyco• sides. In particular, the solvent system, ethyl acetate-pyridine-water• methanol (80 : 12 : 10 : 5) separated the paper chromatographically indistinguishable C-glucosides and C-galactosides. Rf values obtained for acacetin glycosides were as folIows: 6- and 8-C-glucosides 0.44, 0.63: 6- and 8-C-galactosides 0.32, 0.46: 6- and 8-C-xylosides 0.62, 0.70: 6,8-di-C-glucoside 0.11; 6,8-di-C-xyloside 0.30; 6-C-xyloside-8- C-glucoside 0.19. Magnesol has been considered superior to silica far ISOLATION TECHNIQUES FOR FLA VONOIDS 27 flavonol glycosides (Forrest and Bendall, 1969a) and quercitrin was separated from isoquercitrin using toluene-ethyl formate-formic acid (5 : 4 : 1). Silica gel G impregnated with complexing or buffering anions such as borate, molybdate, tungstate and acetate has been used successfully for the isolation of kaempferol and quercetin triglycosides and their p-coumaryl derivatives from Pisum sativum (Harper and Smith, 1969). Silica gel-cellulose (1 : 1) (V an Sumere et al., 1965) and silica gel-starch (Paris and Paris, 1963) mixtures have also been used with some success and it is c1aimed that they combine the advantages of Si02 and cellulose. Isoj7avones, j1avanones, and dihydroj7avonols. These flavonoids are generally chromatographed using less polar solvents than those required for the common flavones and flavonols. For example, the isoflavones daidzein, formononetin, genistein and biochanin A have been separated using chloroform-methanol (92 : 8, 3 : 1 and 1 : 1) (Guggolz et al., 1961; Beck and Knox, 1971) and ethyl acetate-light petroleum (3 : 1 and 1 : 1) (Guggolz et al., 1961). Isoflavones from the heartwood extracts of Cladrastis lutea have been chromato• graphed with chloroform-methanol (4 : 1) and ether (Shamma and Stiver, 1969). With the latter solvent, the order of elution was related to the extent of substitution in that the Rf value of formononetin (7-0H, 4'-OMe) > c1adrin and afrormosin (7-0H, 3'4'-diOMe and 7-0H, 6,4'-diOMe) > c1adrastin (7-0H, 6,3',4'-triOMe). The highly methylated isoflavonoids of Cordyla ajricana heartwood were iso• lated by preparative TLC with chloroform-benzene-acetone (10: 10: 1) (Campbell et al. , 1969). F1avanones such as naringenin, hesperetin and isosakuranetin have been distinguished and identified by Mizelle et al. (1965), using benzene-acetic acid-water (125: 72 : 3), benzene-nitromethane• water (3 : 2 : 5, upper layer), chloroform-acetic acid-H2 0 (2 : 1 : 1) and by Hörhammer et al. (1964), using benzene-pyridine-formic acid (36 : 9 : 5). Dihydroflavonols such as taxifolin and are conveniently separated (Barton, 1968) with chloroform-methanol• acetic acid (7 : 1 : 1). Glycosides of isoflavones and flavanones have also been chromato• graphed successfully on Si02 • The 8-C- and 7-0-glycosides of daidzein were separated using n-butanol-acetic acid-water (4 : I : 5, lower layer) (Bhutani et al. , 1969a) as also were the 7-0-rhamno• glucosides of the flavanones hesperetin, eriodictyol and naringenin 28 THE FLA VONOIDS

(Hörhammer and Wagner, 1962). However for the separation of these 1atter glycosides Mizelle et al. (1965) preferred to use cellulose and polyamide 1ayers. Mixtures of glycosides and acety1ated glycosides of formononetin, genistein and biochanin A (ex. Trifolium species) have been successfully separated (Beck and Knox, 1971) using ch10ro• form-acetone-methano1 (20 : 6 : 5) and ethyl acetate-methano1-water (100 : 16.5 : 13.5). Chalcones and aurones. A number of synthetic aurones have been chromatographed by Hanse1 et al. (1963), on Si02 G plates buffered with sodium acetate, in benzene-ethy1 acetate-formic acid (9 : 7 : 4), chloroform-ethyl acetate-formic acid (6: 3 : 1) and to1uene-ethy1 formate-formic acid (5 : 4 : 1). The same workers separated cha1- cones on Si02 -kieselguhr plates with cyc1ohexane-ethy1 acetate (7 : 1) saturated with formamide-water (2: 1). Harbome (1966) used benzene-ethyl acetate-formic acid (9 : 7 : 4) in the identification of isosalipurposide (chalcononaringenin 2' -glucoside), and Dhar (1972) separated isomeric chalcones and flavanones with ligroin-ethyl acetate (1 : 1). Anthocyanidins and anthocyanins. Although chromatography on paper or TLC-cellu10se is the method of choice for the separation of anthocyanins and anthocyanidins (Kirchner, 1967; Harbome, 1967b; Gupta, 1968a), a number of separations have been achieved on silica. Harbome (1967b) recommends ethyl acetate-formic acid-2N HCI (85 : 6 : 9) for the separation of anthocyanidins generally, and in particu1ar for the separation of ma1vidin (3,5,7,4'-tetraOH, 3'5'• diOMe) and peonidin (3,5,7,4'-tetraOH, 3'-OMe) which are difficult to distinguish by paper chromatography. Solvents such as n-butanol• acetic acid-water (4 : 1 : 2) and ethyl formate-methyl ethyl ketone• formic acid-water (3 : 4 : 1 : 2) are satisfactory for the separation of both anthocyanins and anthocyanidins (Kirchner, 1967; Tanner et al., 1963). Effective separation of anthocyanidin derivatives is thought to be dependent on the presence of trace metals in the Si02 which retard the movement of catecho1 derived pigments, such as cyanidin and de1phinidin, by complex formation (Harbome, 1967b). The anthocyanins of species of Medicago, e.g. the 3,5-diglucosides of de1phinidin, petunidin and ma1vidin separate in ethyl acetate• butanone-formic acid-water (5 : 3 : 3 : 1) (Gupta, 1968a), but some hydrolysis of the anthocyanins occurred with solvents containing higher levels of formic acid. The anthocyanins from Vitis vinifera skins (peonidin, malvidin, petunidin and de1phinidin 3-glucosides and ISOLATION TECHNIQUES FOR FLA VONOIDS 29 their acyl derivatives) have been effectively separated using n-buta• nol-ethyl acetate-benzene-formic acid (l : 1 : 1 : 1) saturated with paraformaldehyde on p1ates buffered with sodium acetate (Conradie and Neethling, 1968). Acid-washed Si02 was used for the preparative separation of the 3-0-glucoside and 3-0-rhamnoglucoside of cyanidin from blackcurrant juice with ethyl acetate-butanone-formic acid• water (6: 3 : I : 1) (Morton, 1967). Asen (1965) used 1 mm thick layers of silica gel-cellulose (2 : 1) and n-butanol-2N HCl (l : 1) to separate cyanidin (3,5,7,3' ,4'-pentaOH) glycosides from their pelar• gonidin (3,5,7,4'-tetraOH) equivalents, and c1aimed that the separa• tions achieved were better than with Si02 alone. Individual glycosides were further purified using a water-2N HC1-formic acid (8 : 4: 1).

1.4.5 Separations on cellulose Flavones and flavonols. TLC on cellulose layers has to some extent replaced paper chromatography in analytical work, since the high surface area, fine-grained cellulose thin-layers offer the advantages of greater speed (e.g. 1-2 h (Gupta, 1968b)) and better resolution (Egger, 1969; Jangaard, 1970; Gupta, 1968b; Kirchner, 1967). Flavone and flavonol aglycones were found by Egger (1969) to be conveniently separated on cellulose with ch10roform-acetic acid• water (10 : 9 : 1) whereas the glycosides were better reso1ved with butano1-acetic acid-water (4: 1 : 5). J angaard (1970) found that modest separations could be achieved with 2% formic acid, isopropyl alcohol-ammonium hydroxide-water (8 : 1 : 1), and 10% acetic acid. Two-dimensional cellulose TLC has been carried out with solvent pairs of the type commonly used in paper chromatography (e.g. water, or 5% methanol in water, and n-butanol-acetic acid-water, 4 : 1 : 5). Similar solvents were also used for the separation of flavone C-glycosides on microcrystalline cellulose (Asen et al. , 1970). However, twelve solvents developed for use in paper chromatography were tested by Gupta (1968b) for the separation of Medicago flower petal flavonoids, and only three were found useful for TLC, namely, n-butanol-2N HCl (l : 1), n-amyl alcohol-acetic acid-water (2 : 1 : 1) and ethyl acetate-formiC acid-water (8 : 2 : 3). Cellulose, when admixed with 3% by weight of polyamide, was found useful for the chromatography of a wide range of flavones and flavonols and their mono- and diglycosides (Spiegi, 1969), using 15, 40 or 60% acetic acid as solvents. Flavone and flavonol glycosides 30 THE FLA VONOIDS have also been separated with limited success on cellulose (or starch)•

Si02 (1 : 1) plates (Van Sumere et al., 1965; Paris and Paris, 1963). Flavonoids other than flavones and f7avonols. Micro-crystalline cellulose layers have been used in conjunction with polyamide by Mizelle et al. (1965) for the identification of the flavanone rhamnoglucosides from grapefruit. Solvents used included benzene• ethyl acetate-formic acid-water (9: 21 : 6 : 5), methyl isobutyl ketone-formic acid-water (14 : 3 : 1) and n-butanol-acetic acid-water (6 : 1 : 2). Aurones and cha1cones have recently been isolated from extracts of Limonium flowers on layers of micro-crystalline cellulose 2 mm thick, using a range of common solvent mixtures (Asen and Plimmer, 1972). Huke et al. (1969) used the solvents, n-propanol• acetic acid-water (1 : 1 : 1) and isopropanol-acetone-water (5 : 1 : 4) to determine the effect of hydroxylation pattern on Rf value for hydroxyaurones. Good separation of anthocyanidins and anthocyanins have been achieved using cellulose TLC. Nybom (1964), for example, used formic acid-HC1-water (10 : 1 : 3) and amyl a1cohol-acetic acid-water (2 : 1 : 1) to produce excellent two-dimensional separations of anthocyanidins, while Paris and Paris (1963) used Forestal (acetic acid-HC1-water, 30: 3 : 10) and 60% acetic acid for the same purpose. Deibner (1968) preferred TLC on cellulose (using 2% acetic acid and t-amyl a1cohol-acetic acid-water, 6: 0.1 : 94) to paper chromatography for the separation of malvidin and peonidin diglucosides.

Using a variety of Si02 -cellulose mixtures, Asen (1965) purified and isolated (in mg quantities) the anthocyanins of Euphorbia pulcherrima with acetone-0.5N HCl (1 : 3), n-butanol-2N HCl (1 : 1) and water-HCI-formic acid (8 : 4 : 1). Cellulose has also been mixed with polyamide (PVP) to advantage in the separation of antho• cyanins (Wro1stad, 1968). The addition of 10% PVP to cellulose layers led to more compact spots without markedly affecting Rf values, and layers of this mixture were preferred to PVP alone, Si02 or paper. The solvents used were n-butanol-acetic acid-water (4: 1 : 5) and acetic acid-water-HCI (15 : 82: 3).

1.4.6 Separations on polyamide Commercially available polyamide powders for TLC vary widely in their physical and chromatographie properties. Thus some are water ISOLATION TECHNIQUES FOR FLA VONOIDS 31 repellent, some form fragile layers and others adhere poorly to the glass plates. An excellent polyamide powder, however, may be prepared from polyamide pellets by the method of Rösler (1960), which is described also by Hörhammer (1964), Wyler et al. (1967) and Mabry et al. (1970). Polyamide separates flavonoids either by partition or adsorption processes depending upon the solvent type used (Egger, 1969). The adsorption process is favoured with water-alcohol mixtures and as a result retention on the layer is determined largely by the number of isolated (not o-dihydroxy or vicinal) phenolic hydroxyl groups and the number and nature of sugars present. Such solvents are generally unsatisfactory for the separation of aglycones but distinguish different glycosides weIl (Egger, 1964); thus Rf values for flavonol glycosides increase in the order: glucuronide, rhamnoside, glucoside, rhamnoglucoside, diglucoside, 3-galactoside-7-rhamnoside, 3-g1uco• side-7-g1ucoside, 3-rhamnoglucoside-7-g1ucoside, 3-xyloglucoside-7- glucoside (Egger, 1969). However, when lipophilic solvents such as chloroform-methanol-butanone (60: 26 : 14, water saturated) are used, partition processes are also operative. Hence under these conditions, mixtures of aglycones and also of the same glycosides of different aglycones are resolved. Two-dimensional separations on polyamide should thus preferably involve the use of an aqueous solvent in one direction and a lipophilic solvent in the other. Flavones and flavonols. As mentioned above, lipophilic solvents are preferable to aqueous solvents for the separation of aglycones according to their degree of oxidation or methylation. A striking demonstration of this is the separation by Egger (1969) of a range of flavonol aglycones in water-ethanol-butanone-acetylacetone and chloroform-ethanol-butanone-acetylacetone. With the aqueous solvent (16: 10 : 5 : I) no separation of flavonol aglycones was achieved, whereas with the chloroform containing solvent (16 : 10 : 5 : I) the following approximate Rf values were obtained: rhamnocitrin (3,5,4'-triOH, 7-0Me) 0.80, galangin (3,5,7-triOH) 0.60, rhamnetin (3,5,3',4' -tetraOH, 7-0Me) 0.50, isorhamnetin (3,5,7,4'-tetraOH, 3'-OMe) 0.45, kaempferol (3,5,7,4'-tetraOH) 0.30, quercetin (3,5,7,3',4' -pentaOH) 0.18, and myricetin (3,5,7,3' ,4' ,5'• hexaOH) 0.13. The flavonols with the highest Rf values are those with the least number of free hydroxyl groups, although an additional methoxyl appears to increase the Rf value. Egger also 32 THE FLA VONOIDS separated flavone and flavonol aglycones using benzene-butanone• methanol (3: I : .1), and kaempferol 7-methyl ether has been isolated (Star and Mabry, 1971) by polyamide TLC using a similar solvent chloroform-butanone-methanol (100: 0.3 : 0.6). PVP has been used for the separation of kaempferol, quercetin and myricetin; however, in this case the solvents 90% formic acid and acetic acid-water-HCI (30: 10: 3) were used (Quarmby, 1968). Flavones and flavonols which lack phenolic hydroxyl groups, e.g. polyacetyl• ated or polymethylated derivatives are better chromatographed (Egger, 1969) with solvents of thetype light petroleum-benzene• butanone-methanol (50 : 40 : 5 : 5) and (60 : 30 : 5 : 5). A solvent recommended by Mabry et al. (1970) for the separation of flavone and flavonol glycosides (as weIl as aglycones) is methanol• acetic acid-water (18 : I : I), and this has been used extensively in the isolation of flavones from liverworts and algae (Markharn, 1972; Markharn et al., 1969; Markharn and Porter, 1969; Markharn et al., 1972). Kaempferol, quercetin and myricetin mono-, di- and triglyco• sides have been separated on perlon using ethanol-water (3 : 2), water-ethanol-acetylacetone (4 : 2 : I) and water-ethanol-butanone• acetylacetone (13: 3 : 3 : 1) (Egger, 1961b; Egger, 1969). Other solvents found suitable for the separation of flavone and flavonol glycosides include: methanol-water (3 : 7) (Davidek, 1960), water saturated n-butanol-acetic acid (100 : land 50 : I), acetone-water (1 : I), ethanol-acetic acid (50: 1), isopropanol-water (3: 2), and butanone-toluene-acetic acid-methanol-water (80 : 10 : 2 : 5 : 6) (Bhandari, 1964, Birkofer et al., 1962). The removal of compounds from polyamide may on occasions be difficult. Markharn (1972) overcame this problem by using a plate, one half of which was spread with polyamide, and the other with cellulose. The compound to be isolated was chromatographically separated on the polyamide, run through to the cellulose with the same solvent, and then eluted from the cellulose powder with methanol. Flavonoids other than fiavones and fiavonols. The solvents mentioned above for flavones, flavonols and their glycosides are applicable, in the main, to other flavonoids also. Hence, hesperidin methyl chalcone has been satisfactorily chromatographed by Silvestri (1970) on polyamide with water-ethanol-acetylacetone (4 : 2 : I), and catechin, naringenin -7-rhamnoglucoside, and delphinidin and ISOLATION TECHNIQUES FOR FLA VONOIDS 33 cyanidin chlorides have been successfully separated in solvents used for the separation of flavonol aglycones (Quarmby, 1968). The rhamnoglucosides of the flavanones, naringenin, hesperetin and isosakuranetin from Texas Ruby Red grapefruit were weIl separated by Mizelle et al. (1965) on polyamide with nitromethane-methanol (5 : 2). Notably, rutinosides were distinguished from their isomerie neohesperidosides in this solvent system. Huke et al. (1969) have chromatographed a wide range of hydroxyaurones in the solvents, chloroform-methanol-acetone-dimethylformamide (4: 4: 2 : 1) and benzene-methanol-acetone-formic acid (4: 4 : 2 : 1), and have correlated the effect of changing hydroxylation pattern on Rf values. Polyamides are generally considered unsatisfactory for the chro• matographie separation of anthocyanins and anthocyanidins (Egger, 1969; Wrolstad, 1968), although Paris and Paris (1963) achieved some separation of the common anthocyanins with acetic acid-HCl• water (10 : 1 : 3) and 60% acetic acid. Limited success has also been obtained using mixtures of polyamide with either polyacrylonitrile or cellulose. For example, Birkofer et al. (1962) used a layer of polyamide-polyacrylonitrile (2: 7) buffered with 0.05 M primary potassium phosphate solution to separate both anthocyanins and anthocyanidins. U sing n-butanol-n-pen tanol-n-propanol-acetic acid• water (2: 3 : 2 : 2 : 1), they obtained the following Rf values: pelargonidin 0.41, peonidin 0.3 9, malvidin 0.37, cyanidin 0.31, petunidin 0.27, and delphinidin 0.21.

1.5 Paper chromatography The technique of paper chromatography still occupies a dominant position in the field of flavonoid analysis and separation. However, it is a technique which has been reviewed extensive1y and exhaustive1y in the past (Egger, 1961a; Harborne, 1958, 1959a,b, 1967a,b; Jiracek and Prochazka, 1963; Mabry et al., 1970; Seikel, 1964), and in view of this, it is. proposed to deal only briefly with this topic in the present chapter. Paper chromatography is suitable for the separation of complex mixtures of all types of flavonoids and their glycosides. Its continuing appeal is probab1y due mainly to this feature, but it must also be due to the ease of obtaining acceptable chromatographie separations by this method, its convenience for isolating both small and relatively 34 THE FLA VONOIDS large amounts of flavonoids, and the low cost of the necessary equipment and materials. Paper chromatographic analyses are com• monly carried out on Whatman No. l, No. 3 or 3MM paper and for optimum resolution, two-dimensional chromatography is recom• mended. For the separation of flavonoid glycosides generally , the first dimension is normally run using an 'alcoholic' solvent (such as: n-butanol-acetic acid-water, 4 : I : 5 (upper phase, BAW); t-butanol• acetic acid-water, 3 : 1 : I (TBA) or water saturated butanol) which produces separations based largely on partitioning (Roux et al., 1961). The second dimension is commonly run using an aqueous solvent (such as: water, 2-60% aqueous acetic acid, 3% NaCl or acetic acid-conc.HC1-water (30 : 3 : 10, Forestal)) in order to achieve complementary separations based on adsorption (Roux et al., 1961). Flavonoid aglycones are generally separated satisfactorily from one another by use of the 'alcoholic' solvents (above) or with benzene-acetic acid-water (125 : 72 : 3) (Wong and Taylor, 1962), chloroform-acetic acid-water (13 : 6 : 1), phenol-water (4: 1), or Forestal. Many of these solvents are also useful for the chromatog• raphy of highly methylated flavonoids although hydrocarbon solvents of the type benzene-ligroin-methanol-water (50 : 50 : 1 : 50) and benzene-nitromethane-water (3 : 2 : 5) have been recommended for this purpose. Solvents containing HCI or acetic acids are required for the chromatography of anthocyanins and anthocyanidins. Compilations of Rf values for a wide range of flavonoids can be found in the references quoted at the beginning of this section. Most flavonoids appear as coloured spots on paper chromatograms viewed in UV light, and fuming with ammonia often produces significant changes in these colours (Mabry et al., 1970). A variety of general and specific spray reagents is available for the detection of flavonoids on paper (Table 1.4).

1.6 Miscellaneous techniques 1.6.1 Gas-liquid chromatography Although GLC has not been used extensively for the analysis or isolation of flavonoids, it is an acceptable method provided the flavonoid is derivatized to increase its volatility. Trimethylsilyl ether derivatives have been found most effective for this purpose, although ISOLA nON TECHNIQUES FOR FLA VONOIDS 35 methyl ether and acetate derivatives have also been used (Nara• simhichari and Von Rudloff, 1962; Nordstrom and Krone1d, 1972). The stationary phases, SE-30 and OV -1, are the most commonly used for the separation of flavonoids. Furuya (1965) separated a mixture of flavones, flavono1s, flavanones, dihydroflavono1s, cha1- cones, and isoflavones (as their TMS ethers) on 1.5% SE-30 (240°) and found that the retention time increased with increasing oxidation and/or hydroxylation. Thus the retention time (for equivalent compounds) of flavanones < cha1cones; leucoantho• cyanidins < dihydroflavonols < flavonols, and flavanones < flavones < flavono1s. However, AI-Shakir (1968) is reported (Pierce, 1968) to have observed a decrease in retention time with increasing B-ring oxygenation in anthocyanidins. Anthocyanidins and anthocyanins, as their TMS ethers, have been chromatographed on SE-30 (0.05% and 3%, 210-250°) and SE-52 (3%, 210-250°) with some success, as also have a number of flavonols and their glycosides (Keith and Powers, 1966). Other examples of the use of SE-30 and SE-52 include the separations of flavonoid TMS ether derivatives by Hemingway and Hillis (1969) and Pier ce (1968), and of flavonoid methyl ethers and acetates by Narasimhichari and Von Rudloff (1962). In this latter case it was shown that milligram quantities of flavonoid could be isolated by this method. The stationary phase OV-1 has been found suitable for GLC of the TMS ethers of flavanols (Collier and Mallows, 1971; Pierce et al., 1969), flavones (Nordstrom and Kroneid, 1972), flavonols (Andersen and Vaughn, 1970: Pellizzari et al., 1969; Nordstrom and Krone1d, 1972), flavanones and cha1cones (Pellizzari et al. , 1969), and OV-101 was used by Andersen and Vaughn (1970) to carry out quantitative estimations of quercetin at the 300 ng level. Andersen and Vaughn used an electron capture detector for their work and found it to be 2-3 times more sensitive than the flame ionization detector. Both OV-1 and OV-17 were used successfully by Nordstrom and Krone1d (1972) for the separation of a wide variety of flavone and flavono1 methyl ethers, TMS ethers and mixed methyl and TMS ethers. It was found that on OV-1, absolute and relative retention times of both ether types were similar. On the OV-17 stationary phase however, TMS ether derivatives had shorter retention tim es than the equivalent methyl ethers. In addition, the TMS ethers of flavone-flavonol pairs were more effectively separated. 36 THE FLA VONOIDS Flavonoids appear to undergo very Httle decomposition (apart from flavanone-chalcone isomerization (Furuya, 1965; Nara• simhichari and Von Rudloff, 1962» during GLC at normal temper• atures (190-250°) as evidenced by GLC-mass spectrometry (Pelliz• zari et al., 1969). 1.6.2 Paper electrophoresis Electrophoresis is another technique which has found only limited application in the field of flavonoid isolation and separation (Seikel, 1964; Harborne, 1967a; Paris, 1961). This is largely because, except for a few specific cases, it offers little or no advantage over paper chromatography. Flavonoids must be ionized or complexed with a metal ion to be mobile in an electric field and for this reason the technique is admirably suited to the isolation of flavonoids which occur naturally in these states, e.g. 7-potassium bisulphate (which was readily separated from the 7-glucoside) (Saleh et al., 1971), the pectin-metal-anthocyanin cornflower pigment (Bayer et al, 1966) and the metallo-anthocyanin-flavone Commelina pigment (Take da et al, 1966). In each of these cases acidic acetate buffers were used in conjunction with high voltages (3000 V, 110 V and 400 V respectively) to produce migration of the flavonoids to the anode. The positively charged anthocyanins (in acids) are separated only slowly by electrophoresis (Markakis, 1960). Ionization of flavonoids with alkali tends to cause oxidative decomposition in air and for this reason, complex formation with a borate buffer has been preferred as a method of producing charged species. Such complexes have been useful in the separation of flavanone rutinosides from the equivalent neohesperidosides (pH 9.5) (Mizelle et al.. 1965; Gentili and Horowitz, 1964), for the isolation and purification of flavan-3,4-diols (pH 8.8, 150-170 V, 33 mA cm- 1 ) (Drewes and Roux, 1966) and of bileucofisetinidins (pH 8.8, 170 V, 31 mA cm- 1 ) (Drewes et al., 1967b; Drewes et al., 1967a), and far the separation of flavonoids containing vicinal hydroxyl groups (Drewes and Roux, 1964; Hashimoto et al., 1952; Cooper and Roux, 1965) or ionized 7-hydroxyl groups (Cooper and Roux, 1965; Paris, 1961; Coulson and Evans, 1958). The effect of the number of vicinal hydroxyl groups on mobility was nicely demonstrated by Hashimoto et al. (1952), who recarded relative mobilities of 25, 16 and 2 respectively for rutin (2), quercetin (1) ISOLATION TECHNIQUES FOR FLA VONOIDS 37 and morin (0), (500 V, 1.5 mA cm -1). The separation of cis from trans flavan-3,4-diols (Drewes and Roux, 1964) in O.lM sodium borate (170 V, 5 mA cm -1) was also achieved by this technique, the cis isomers having positive mobility and the trans having zero or negative (due to electroendosmotic flow) mobility. Cellulose thin-layer e1ectrophoresis of some flavonols and their glycosides has been investigated at an analyticallevel by Walker and Thompson (1969). Using aborate buffer they found that the glycosides were 2-3 times more mobile than the aglycones, but that separation of different ag1ycones was unimpressive. This technique, however, has the advantage that it may be combined with conven• tiona1 TLC (at right angles to the direction of electrophoresis) for improved analytical separations.

1.6.3 Sublimation Most flavonoids are not sufficiently vo1atile be10w their decomposi• tion temperatures to be amenable to purification by this technique. The few examples of the use of this technique for the isolation or purification of naturally occurring flavonoids include: the isolation of 3'-chloro-2' ,5-dihydroxy-3, 7 ,8-trimethoxyflavone from a fungal

0 culture extract (at 200 , 0.1 mm) (Bird and Marshall, 1969), the purification of 7-hydroxy-6,4'-dimethoxyisoflavone from Cabreuva wood (Harborne et al. , 1963) and calycosin (7,3'-dihydroxy-4'•

0 methoxyisoflavone) from Baptisia lecontei (180 , 0.05 mm) (Mark• harn et al., 1968a). Sim (1967), in a general investigation of the sublimation of flavonoids, found that fractional sublimation could be used to separate certain mixtures, e.g. 3,5-dihydroxy-6, 7 ,4'-trimeth• oxyflavone and 5,4' -dihydroxy-3 ,6,7 -trimethoxyflavone from 3,5,7- trihydroxy-3' ,4' ,5'-trimethoxyflavone. In addition it was observed that glycosides such as pendulin may hydrolyse quantitatively under sublimation conditions to produce the aglycone, a phenomenon previously noted by Fischer (1937). Typical sublimation conditions used by Sim were 180-2400 at 0.2 mm.

1.6.4 High pressure liquid chromatography (HPLC) HPLC is a relatively new technique, somewhat akin to GLC except that the carrier gas is replaced by a solvent or solvent mixture (see for example: Perry, 1971; Done et al., 1972). In princip1e it is ideally 38 THE FLA VONOIDS suited to the chromatographie analysis (both qualitative and quanti• tative) of non-volatile compounds. The full potential of this method for the chromatography of flavonoids has not as yet been realized, but one good example of its usefulness has recently appeared. Charalambous et al. (1973) examined the in beers and wines by this method, using a VIDAC (silica) column (1 m x 2 mm) at 400 psi monitored by a UV detector. The solvent systems: hexane, followed by methanol-chloroform-acetic acid (30: 70 : 1); and hexane-chloroform (l : 1), followed by methanol-chloroform-acetic acid (50: 50 : 1), were found suitable for the separation of quercetin, kaempferol, catechin, catechin gallates, caffeic acid and coumaric acid, with retention tim es of 10-20 min. Advantages c1aimed for HPLC analysis inc1ude (i) short analysis time, (ii) high resolution, (iii) no derivatization required, (iv) no risk of thermal decomposition, and (v) easy quantification.

References Abdel-Baset, Z. M. E. H. (1973), Ph.D. dissertation, University of Texas at Austin, U.S.A. Achrern, A. A. and Kuznetsova, A. I. (1963), Russ. Chem. Rev. 32,366. AI-Shakir, S. H. (1968), Diss. Abstr. B28, 2892. Andersen, R. A. and Sowers, J. A. (1968), Phytoehemistry 7,293. Andersen R. A. and Vaughn, T. H. (1970),J. Chromat. 52,385. Asen, S. (1965),J. Chromat. 18,602. Asen, S. and Plirnrner, J. R. (1972), Phytoehemistry 11,2601. Asen, S., Stewart, R. N., Norris, K. H. and Massie, D. R. (1970),Phytoehemistry 9,619. Bandyukova, V. A. and Shinkarenko, A. L. (1966), Zh. Analit. Khim. 21,232. Barton, G. M. (1965),J. Chromat. 20, 189. Barton, G. M. (1968), J. Chromat. 34, 562. Bayer, E., Egeter, H., Nether, A. F. K. and Wegrnann, K. (1966), Angew. Chern. (internatl. Ed.) 5, 791. Beck, A. B. (I 964), Aust. J. Agrie. Res. 15,223. Beck, A. B. and Knox, J. R. (1971), Aust. J. Chem. 24, 1509. Bhandari, P. R. (1964),J. Chromat. 16,130. Bhutani, S. P., Chibber, S. S. and Seshadri, T. R. (1969a), Indian J. Chem. 7, 210. Bhutani, S. P., Chibber, S. S. and Seshadri, T. R. (I969b), Phytoehemistry 8, 299. Bird, A. E. and Marshall, A. C. (1969),J. ehem. Soe. (C) 2418. Birkofer, L., Kaiser, C., Meyer-Stoll, H. A. and Suppan, F. (1962), z. Naturf. 17b,352. ISOLATION TECHNIQUES FOR FLA VONOIDS 39

Bodalski, T. and Cisowski, W. (1967), Dissnes. Pharrn. Warsz. 19,99. Bottomley, W., Smith, H. and Galston, A. W. (1966),Phytoehernistry 5, 117. Brockman, H. and Schodder, H. (1941), Chern. Ber. 74B, 73. Cambie, R. C. and James, M. A. (1967), New Zeal. 1. Sei. 10,918. Campbell, R. V., Harper, S. H. and Kemp, A. D. (1969),J. ehern. Soe. (C) 1787. Cassidy, H. G. (1957). In Teehniques olOrganie Chernistry (ed. A. Weissberger), Vol. X, pp. 107-209,285-317. Interscience Publisher Inc., New York. Charalambous, G., Bruckner, K. J., Hardwick, W. A. and Linnebach, A. (1973), Master Brewers Association 01 Arneriea, Teehnieal Quarterly 10 (2), 74. Chexal, K. K., Handa, B. K. and Rahman, W. (1970),1. Chrornat. 48,484. Chopin, J. (1971). In Pharrnaeognosy and Phytoehernistry (eds. H. Wagner and L. Hörhammer), p. 111. Springer-Verlag, Berlin. Clark-Lewis, J. W. and Dainis, I. (1967), Aust. J. Chern. 20,2191. Clark-Lewis, J. W. and Dainis, I. (1968), Aust. J. Chern. 21,425. Clark-Lewis, J. W. and Porter, L. J. (1972), Aust. J. Chern. 25, 1943. Collier, P. D. and Mallows, R. (1971),1. Chrornat. 57, 19,29. Conradie, J. D. and Neethling, L. P. (1968),1. Chrornat. 34,419. Cooper, D. R. and Roux, D. G. (1965),1. Chrornat. 17,396. Coulson, C. B. and Evans, W. C. (1958),1. Chrornat. 1,374. Coxon, D. T., Holmes, A., Ollis, W. D. and Vora, V. C. (1972), Tetrahedron 28, 2819. Dass, H. C. and Weaver, G. M. (1972),1. Chrornat. 67,105. Davidek, J. (1960), Nahrung. 4, 661. Deibner, L. (1968),1. Chrornat. 34,425. Dhar, D. N. (1972),1. Indian ehern. Soc. 49,309. Done, J. N., Kennedy, G. J. and Knox, J. H. (1972), Nature 237, 77. Drewes, S. E. and Ilsley, A. H. (1969),Phytoehernistry 8,1039. Drewes, S. E. and Roux, D. G. (1964), Bioehern. J. 92,555. Drewes, S. E. and Roux, D. G. (1966), Bioehern. J. 98,493. Drewes, S. E., Roux, D. G., Eggers, S. H. and Feeney, J. (1967a),J. ehern. Soe. (C) 1217. Drewes, S. E., Roux, D. G., Saayman, H. M., Eggers, S. H. and Feeney, J. (1967b),1. ehern. Soe. (C) 1302. Dreyer, D. L. and Bertelli, D. J. (1967), Tetrahedron 23,4607. Egger, K. (1961a),1. Chrornat. 5,74. Egger, K. (1961b),Z. analyt. Chern. 182,161. Egger, K. (1964), Plan ta rned. 12,265. Egger, K. (1969). In Thin-Layer Chrornatography - A Laboratory Handbook (ed. E. Stahl), p. 687. Springer-Verlag, Berlin. Endres, H. (1969). In Thin-Layer Chrornatography (ed. E. Stahl), p. 41. George Allen and Unwin Ud, London; Springer-Verlag, Berlin. Endres, H. and Hormann, H. (1963),Angew. Chern. (internatl. Ed.) 2, 254. Farkas, L., Nogradi, M., Sudarsanam, V. and Herz, W. (1966),1. org. Chern. 31, 3228. Fischer, L. (1969). In Laboratory Teehniques in Bioehernistry and Moleeular Biology (eds. T. S. Work and E. Work), Vol. 1, p. 157. North-Holland Publ. Co. 40 THE FLA VONOIDS

Fischer, R. (1937), Areh. Pharm. 275,516. Forrest, G. I. and Bendall, D. S. (1969a), Bioehem. J. 113, 74l. Forrest, G. I. and Bendall, D. S. (1969b), Bioehem. J. 113,757. Furuya, T. (1965),1. Chromat. 19,607. Gage, T. B., Morris, Q. L., Detty, W. E. and Wender, S. H. (1951), Scienee 113, 522. Gelotte, B. (1960),1. Chromat. 3,330. Gentili, B. and Horowitz, R. M. (1964). Symposium on 'Recent Advances in Plant Polyphenolics', Delhi, India. Govindachari, T. R., Pai, B. R., Srinivasan, M. and Kalyanaraman, P. S. (1969), Indian J. Chem. 7, 306. Govindachari, T. R., Parthasarathy, P. C., Pai, B. R. and Subramaniam, P. S. (1965), Tetrahedron 21,3237. Guggolz, J., Livingston, A. L. and Bickoff, E. M. (1961),J. Agrie. Fd. Chem. 9, 330. Gupta, S. B. (1968a),1. Chromat. 36,115. Gupta, S. B. (1968b),1. Chromat. 36,258. HanseI, R., Langhammer, L., Frenzel, J. and Ranft, G. (1963),1. Chromat. 11, 369. Harborne, J. B. (1958), Bioehem J. 70,22. Harborne, J. B. (1959a), Chromat. Rev. 1, 209. Harborne, J. B. (1959b),1. Chromat. 2, 58l. Harborne, J. B. (1966), Phytoehemistry 5, 111. Harborne, J. B. (1967a), In Chromatography (ed. E. Heftmann), pp. 677-699. Reinhold Publ. Co., New York. Harborne, J. B. (1967b), Comparative Bioehemistry 01 the Flavonoids, Academic Press, London. Harborne, J. B., Gottlieb, O. R. and Magalhaes, M. T. (1963),1. org. Chem. 28, 88l. Harborne, J. B. and Williams, C. A. (1971),Phytoehemistry 10,367. Harper, D. B. and Smith, H. (1969),1. Chromat. 41, 138. Hashimoto, Y., Mori, I. and Kimura, M. (1952), Nature 170,975. Hemingway, R. W. and Hillis, W. E. (1969),1. Chromat. 43,250. Henrick, C. A. andJefferies, P. R. (1964), Aust. J. Chem. 17,934. Herz, W., Gibaja, S., Bhat, S. V. and Srinivisan, A. (1972), Phytoehemistry 11, 2859. Hillis, W. E. and Isoi, K. (1965), Phytoehemistry 4, 54l. Hörhammer, L. (1964). In Methods 01 Chemistry (ed. J. B. Pridham), p. 89. Pergamon Press, London. Hörhammer, L., Stich, L. and Wagner, H. (1961), Areh. Pharm. 294,685. Hörhammer, L. and Wagner, H. (1962),Dt. Apoth. Ztg. 102,759. Hörhammer, L., Wagner, h. and Beyersdorff, P. t1962), Naturwissensehaften 49, 392. Hörhammer, L., Wagner, H. and Dhingra, H. S. (1959), Areh. Pharm. 292,84. Hörhammer, L., Wagner, H. and Hein, K. (1964),1. Chromat. 13,235. Hori, M. (1969),Bull. ehem. Soe. Japan. 42,2333. ISOLATION TECHNIQUES FOR FLA VONOIDS 41

Hrazdina, G. (1970),1. Agric. Fd. Chern. 18,243. Huke, M., Gorlitzer, K. and Schenck, G. (1969), Arch. Pharm. 302,401. lee, C. H. and Wender, H. (1952), Analyt. Chern. 24, 1616. Jackson, B., Locksley, H. D. and Scheinrnann, F. (1967), Tetrahedron Letters 787. Jangaard, N. O. (1970).J. Chrornat. 50,148. Jiracek, V. and Prochazka, Z. (1963). In Paper Chrornatography (eds. I. M. Hais and K. Macek), pp. 254-271. Acadernic Press, New York. Johnston, K. M., Stern, D. J. and Waiss, A. C. (1968). J. Chrornat. 33,539. Kawano, N., Miura, H. and Kikuchi, H. (1964). Yakugaku Zasshi 84, 469. Keith, E. S. and Powers, J. J. (1966),J. Fd. Sei. 31,971. King, H. G. C. and Pruden, G. (1970),J. Chrornat. 52,285. Kirchner, J. G. (1967), In Thin-Layer Chrornatography. Vol. XII of Techniques o/Organic Chernistry (eds. E. S. Perry and A. Weissberger), pp. 556-568. Interscience Publishers, New York. Kupchan, S. M., Siegel, C. W., Herningway, R. J., Knox, J. R. and Udayamurthy, M. S. (1969). Tetrahedron 25, 1603. Lebreton, P., Markharn, K. R., Swift, W. T., Oung-Boran and Mabry, T. J. (1967), Phytochernistry 6, 1675. Loev. B. and Goodrnan, M. M. (1967). Chern. Ind. 2026. Loev, B. and Goodman, M. M. (1970). In Progress in Purification and Separation (eds. E. S. Perry and C. J. Van Oss), Vol. 3, p. 73. Wiley-Interscience, New York. Loev, B. and Snader, K. M. (1965), Chern. Ind. 15. Logie, D. (1957),Analyst. 82,563. Mabry, T. J., Markham, K. R. and Thornas, M. B. (1970), The Systernatic Identi/ication 0/ Flavonoids, Springer-Verlag, Berlin. Markakis, P. (1960), Nature 187, 1092. Markharn, K. R. (1972),Phytochernistry 11,2047. Markharn, K. R., Hirshrnan, J. L., Kupferrnann, H. and Ma, T. S. (1970), Microchirnica Acta 590. Markharn, K. R. and Mabry, T. J. (1968a),Phytochernistry 7, 791. Markham, K. R. and Mabry, T. J. (1968b),Phytochernlstry 7,1197. Markharn, K. R., Mabry, T J. and Averett, J. E. (1972), Phytochernistry 11, 2875. Markharn, K. R., Mabry, T. J. and Swift, T. W. (1968a),Phytochernistry 7, 803. Markharn, K. R. and Porter, L. J. (l969), Phytochernistry 8, 1777. Markharn, K. R., Porter, L. J. and Brehm, B. G. (1969),Phytochernistry 8,2193. Markharn, K. R., Swift, W. T. and Mabry, T. J. (1968b),J. org. Chern. 33,462. McFarlane, W. D. and Vader, M. J. (1962),J. Inst. Brew. 68,254. Meier, W. and Fuerst, A. (1962), Helv. chirn. Acta 45, 232. Mizelle, J. W., Dunlap, W. J., Hagan, R. E., Wender, S. H., Lirne, B. J., Albaeh, R. F. and Griffiths, F. P. (1965), Analyt. Biochern. 12,316. Morgan, J. W. W. and Orsler, R. J. (1967), Chern. Ind. 1173. Morris, Q. L., Gage, T. B. and Wender, S. H. (1951), J. Arn. ehern. Soc. 73, 3340. 42 THE FLA VONOIDS

Morton, A. D. (1967),1. Chrornat. 28,480. Mukerjee, S. K., Sarkar, S. C. and Seshadri, T. R. (1969), Tetrahedron 25, 1063. Narasimhichari, N. and Von Rudloff, E. (1962), Canad. J. Chern. 40, 1123. Neu, R. (1957), rnicrochirn. Acta 196. Neu, R. (I959),Arch. Pharrn. 293,169. Nilsson, A. (1962), Acta chern. scand. 16,31. Nordstrom, C. G. (1967), Acta chern. scand. 21,2885. Nordstrom, C. G. and Kroneid, T. (1972), Acta chern. scand. 26,2237. Nybom, N. (1964),Physiol. Plantarurn 17,157. Ohta, N. and Yagishita, K. (I970),Agric. biol. Chern. 34,900. Okigawa, M., Hatanaka, H., Kawano, N., Matsunaga, I. and Tamura, Z. (1971a), Chern. pharm. Bull. 19, 148 Okigawa, M., Wu Hwa, C., Kawano, N. and Rahman, W. (1971b), Phyto- chernistry 10,3286. Olechnowicz-Stepien, W. (1967), Dissnes. Pharrn. Warsz. 19,91. Paris, R. (1961), Pharrn. Acta Helv. 36, 176. Paris, R. and Paris, M. (1963),Bull. Soc. chirn. France 1597. Pastuska, G. (1961),Z. analyt. Chern. 179,355. Pellizzari, E. D., Chuang, C-M., Kuc, J. and Williams, E. B. (1969), J. Chrornat. 40,285. PeIter, A., Amenechi, P. I. and Wanen, R. (1969),1. chern. Soc. (C) 2572. PeIter, A. and Stainton, P. (1966),1. chern. Soc. (C) 70l. Pereyra, de Santiago, O. J. and Juliani, H. R. (1972), Experientia 28, 380. Perry, S. G. (1971), Chern. Britain 7(9), 366. Pierce, A. E. (1968). In Silylation o/Organic Cornpounds, pp. 154-159. Pierce Chemical Co., Illinois. Pierce, A. R., Graham, H. N., Glassner, S., Madlin, H. and Gonzalez, J. G. (1969),Analyt. Chern. 41,298. Porter, L. J. (1973), Personal Communication. Porter, L. J. and Markham, K. R. (1970a),1. chern. Soc. (C) 344. Porter, L. J. and Markham, K. R. (1970b),1. chern. Soc. (C) 1309. Porter, L. J. and Markham, K. R. (1972),Phytochernistry 11, 1477. Porter, L. J. and Wilson, R. D. (1972),1. Chrornat. 71,570. Quarmby, C. (1968),1. Chrornat. 34,52. Radhakrishnan, P. V., Rao, A. V. R. and Venkataraman, K. (1965), Tetrahedron Letters 663. Randerath, K. (1963), Thin-Layer Chrornatography, pp. 180-182. Academic Press, London and New York. Repas, A., Nikolin, B. and Dursun, K. (1969),J. Chrornat. 44, 184. Rösler, H. (1960), Ph.D. dissertation, Univ. ofMunich. Roux, D. G., Maihs, E. A. and Paulus, E. (1961),J. Chrornat, 5,9. Roux, D. G. and Paulus, E. (1962), Biochern, J. 82,320,324. Sachs, D. H. and Painter, E. (1972), Science 175, 78l. Saleh, N. A. M., Bohm, B. A. and Ornduff, R. (1971), Phytochernistry 10, 61l. Schwarz, J. S. P., Cohen, A. 1., Ollis, W. D., Kaczka, E. A. and Jackman, L. M. (1964), Tetrahedron 20,1317. ISOLA TION TECHNIQUES FOR FLA VONOIDS 43

Seikel, M. K. (1962). In The Chemistry o[ Flavonoid Compounds (ed. T. A. Geissman), p. 34. Pergamon Press, New York. Seikel, M. K. (1964). In Bioehemistry o[ Phenolie Compounds (ed. J. B. Harborne), p. 37. Academic Press, London and New York. Seikel, M. K. and Bushnell, A. J. (1959),1. org. Chem. 24, 1995. Seikel, M. K., Bushnell, A. J. and Birzgalis, R. (1962),Areh. Bioehem. Biophys. 99,451. Seikel, M. K., Chow, J. H. S. and Feldman, L. (1966), Phytoehemistry 5,439. Seshadri, T. R. (1962). In The Chemistry o[ Flavonoid Compounds (ed. T. A. Geissman), p. 6. Pergamon Press, New York. Shamma, M. and Stiver, L. D. (1969), Tetrahedron 25,3887. Silvestri, S. (1970), Pharm. Acta. Helv. 45, 390. Sim, K. Y. (1967),1. ehem. Soe. (C) 976. Skelly, N. E. (1961), Analyt. Chem. 33,271. Smith, I. (ed.) (1960), Chromatographie and Eleetrophoretie Teehniques, Vol. 1, p. 71. W. Heinemann Medical Books Ltd., London, and Interscience Publ., New York. Somers, T. C. (1966a),1. Sei. Fd. Agrie. 17,215. Somers, T. C. (1966b), Nature 209, 368. Somers, T. C. (1967),1. Sei. Fd. Agrie. 18,193. Spada, A. and Cameroni, R. (1958), Gazz. Chim. ital. 88,204. Spiegi, P. (1969),J. Chromat. 39,93. Stahl, E. (1969), Thin-Layer Chromatography - A Laboratory Handbook, 2nd Ed., Springer-Verlag, Berlin. Stahl, E. and Mangold, H. K. (1967). In Chromatography (ed. E. Heftmann), p. 165. Reinhold Publishing Co., New York. Stahl, E. and Schom, P. 1. (1961), Z. physiol. Chem. 325,263. Star, A. E. and Mabry, T. J. (1971),Phytoehemistry 10,2817. Takeda, K., Mitsui, S. and Hayashi, K. (1966), Bot. Mag. Tokyo, 79,578. Tanner, H., Rentschler, H. and Senn, G. (1963), Chem. Abstr. 59,13094. Tatum, J. H. and Berry, R. E. (1972),Phytoehemistry 11,2283. Thomas, D. E. and Thomas, J. D. R. (1969),Analyst. 94, 1099. Thomas, M. B. and Mabry, T. J. (1968), Tetrahedron 24,3675. Thompson, R. S., Jacques, D., Haslam, D. J. E. and Tanner, R. J. N. (1972), J. ehem. Soe. (C) 1387. Trim, A. R. (1955). In Modern Methods o[ Plant Analysis (eds. K. Paech and M. V. Tracey), Vol. 2, p. 295. Springer-Verlag, Berlin. Truter, E. V. (1963), Thin-Film Chromatography, Cleaver-Hume Press Ltd., London. Valesi, A. G., Rodriguez, E., Vander Velde, G. and Mabry, T. J. (1972), Phytoehemistry 11, 2821. Vancraenenbroeck, R., Rogirst, A., Lemaitre, H. and Lontie, R. (1963), Bull. Soe. eh im. Belg. 72, 619. Van Sumere, C. F., Wolf, G., Teuchy, H. and Kint, J. (1965), J. Chromat. 20, 48. 44 THE FLAVONOIDS

Van Teeling, C. G., Cansfield, P. E. and Gallop, R. A. (1971), J. chromat. Sei. 505. Voirin, B. (1972),Phytochemistry 11,257. Walker, J. R. L. and Thompson, J. E. (1969), Laborat. Practice, 18,629. Wallace, J. W., Mabry, T. J. and Alston, R. E. (1969),Phytochemistry 8, 93. Watkin, J. E. (1960), Chem. Ind. 378. Webster, P. V., Wilson, J. N. and Franks, M. C. (1967), Analyt. chim. Acta 38, 193. Wender, S. H. (1954), V.S. Patent No. 2,681,907. Williams, B. L. and Wender, S. H. (1952), J. Am. chem. Soc. 74,4372. Williams, B. L. and Wender, S. H. (1953),J. Am. chem. Soc. 75,4363. Woelm, M. (Publisher). In Adsorbents Woelm, Adsorbents Woelm tor Column and Thin-Layer Chromatography, M. Woelm, West Germany. Wollenweber, E. (1972),Phytochemistry 11,425. Wollenweber, E. and Egger, K. (1971), Z. Pflanzenphysiol. 65,427. Wong, E. and Francis, C. M. (1968),Phytochemistry 7, 2131. Wong, E. and Taylor, A. O. (1962),1. Chromat. 9,449. Woof, J. B. and Pierce, J. S. (1967),1. Chromat. 28,94. Wrolstad, R. E. (1968),J. Chromat. 37,542. Wyler, H., Rösler, H., Mercier, M. and Dreiding, A. S. (1967), Helv. chim. Acta 50,545. 本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP 图书馆。 图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具