Carlsberg Res. Commun. Vol. 46, p. 43-52, 1981

STRUCTURE ELUCIDATION OF SOME IN BARLEY BY 1H 270 MHz NMR SPECTROSCOPY by

HENRIK OUTrRUP

Department of Brewing Chemistry, Carlsberg Research Laboratory, Gamle Carlsberg Vej 10, DK-2500 Copenhagen Valby

and

KJELD SCHAUMBURG

Department of Chemical Physics, The H. C. Orsted Institute, University of Copenhagen, Universitetsparken 5, DK-2100 Copenhagen O

Keywords: Anthocyanogens, barley, NMR spectroscopy, proanthocyanidins, structure

In the present paper the composition of proanthocyanidins isolated from barley is investigated. Separation by affinity chromatography and HPLC yields a large number of substances. Six of these have been isolated in this work. Based on the chromatographic preparation they are expected to be two dimers DI and D2 and four trimeric substances TI - T4. Studies by means of IH NMR spectroscopy at 270 MHz confirmed DI and D2 as being dimeric compounds, a prodelphinidin and B3 respectively. The spatial structure of the prodelphinidin corresponds to procyanidin B3. The peracetates of all six compounds have also been studied by means of IH NMR spectrocopy. By comparison with data for the analogous peracetates of (+)- and (-)-epicatechin and with literature data, T1 - T4 have been identified as trimeric proanthocyanidins and their structures have been elucidated.

1. INTRODUCTION has been shown (10) that they participate in the Proanthocyanidins, in the brewing literature formation of haze in beer through reactions with normally termed anthocyanogens, constitute a proteins. The origin of most of the proanthocya- group of flavonoid compounds of significant nidins in beer is the malt, which contains all the practical interest to the brewing industry since it proanthocyanidins of the barley (11). From

Abbreviations: HPLC = high-pressure liquid chromatography, NMR = nuclear magnetic resonance, TMS = tetramethylsilane.

0105-1938/81/0046/0043/$ 02.00 H. OUTTRUP• K. SCHAUMBURG:IH NMR of proanthocyanidins barley and malt three and one were fractionated mainly according to molecular prodelphinidin have previously been character- weight (12). Further separation of groups of ized (9, 14). The constitution of the three proanthocyanidins was obtained by preparative procyanidins was deduced by proving their HPLC. identity to authentic species (9, 14). The prodel- In this manner it was possible to separate phinidin could not be identified in this manner prodelphinidins and procyanidins. The two since no reference substance was available, trimeric proanthocyanidins which are at the leaving open the possibilities that the dimeric same time prodelphinidins and procyanidins prodelphinidin had either a procyanidin B 1- or a could not be separated by HPLC. The separation procyanidin B3-1ike structure (9, 14). was accomplished by renewed fractionation on In the present work we have reinvestigated the Sephadex LH-20. composition of the fraction of For convenience in the following discussion barley with the primary aim of studying the labels have been defmed for the six compounds. prodelphinidins. The labels D and T refer to the number of The separation of the proanthocyanidins is catechin units, whereas the numbering (e.g. D1 accomplished by a combination of low pressure and D2) refers to the sequence of elution by affinity chromatography (Sephadex LH-20) and HPLC. Thus D1 is more polar than D2. high-pressure liquid chromatography (HPLC) on reversed phase. The identity and structure of the compounds prepared were ascertained by chemi- 2.2. Acetylation of proanthocyanidins cal methods and IH NMR spectroscopy. The proanthocyanidin (about 10 mg) was The reinvestigation showed that the structure dissolved in pyridine (200 ~1). The solution was of the previously mentioned dimeric prodelphini- treated at room temperature with acetic acid din (9, 14) is analogous to that of procyanidin B3. anhydride (200 lal). When deuterated acetates Observations leading to the same conclusion for spectroscopic purposes were needed have been published by PORTER et al. (7). Three (CD3CO)20 (Merck Sharpe & Dohme Canada new trimeric proanthocyanidins were isolated Limited, Montreal) has been used. After 16 from barley (12). hours the reaction mixture was poured on ice, The spectroscopic results and their interpreta- whereby the acetate precipitated. The precipitate tion are discussed in this paper resulting in was washed with cold water until the smell of proposed structures for these compounds. pyridine disappeared. After the washing the precipitate was dried in vacuum at room temperature at 10 -2 torr. 2. MATERIALS AND METHODS 2.1. Isolation of proanthocyanidlns from barley The barley variety investigated was Nordal, 2.3. Hydrolysis of proanthocyanidlns harvest 1978 and 1979. The method of prepara- A few milligrams of proanthocyanidin were tion of pure proanthocyanidins has been descri- treated with 0.1 N HCI in ethanol for 15 min at bed elsewhere (i 2) and will only be stated briefly. 60 ~ according to HASLAM et al. (! 3). By this Barley kernels were decorticated in the Carls- procedure the lower terminal unit was split off as berg Research Mill, the flour fraction being a flavan-3-ol. After oxydative hydrolysis in isolated for further investigations. The fine BuIOH - HC1 (5:1) for half an hour at 100 ~ particles of this decorticated material were the upper units are transformed into anthocyani- separated from coarse proanthocyanidin-free dins. particles by means of bolting, hereby leaving a flour containing 8-9 mg. g-1 of proanthocyani- dins, which were extracted with acetone-water 2.4. Preparation of NMR samples (3:1) (9). An initial fractionation was carried out The proanthocyanidins were dissolved in by low pressure chromatography on a Sephadex deuterioacetone-d6 (Merck, Darmstadt, purity >1 LH-20 column with 96% ethanol-water (3:2). 99.5%) and tetramethylsilane (TMS) (Merck, Following this procedure the proanthocyanidins Darmstadt) was added as internal standard. The

44 Carlsberg Res. Commun. Vol. 46, p. 43-52, 1981 H. OUTTRUP8r K. SCHAUMBURG:IH NMR of proanthocyanidins solutions were placed in 5 mm o.d. sample tubes. 2.6. Nuclear Overhauser experiment (NOE) The acetylated proanthocyanidins were dissolved The proton proton NOE experiment was in CDC13 (Merck, Darmstadt, purity >/ 99 %) performed by alternating selective irradiation at and the solutions were placed in 5 mm o.d. the frequencies of the C4 protons in (+)-catechin sample tubes. The concentrations were in the with a total saturation period of 5 sec. The range 2-5 % w/v. reference spectrum was obtained with the decoupling frequencies outside the spectral re- gion. The total accumulation time was 15 hours. 2.5, tH NMR spectra The two spectra are swopped from disc after 16 The IH NMR spectra have been obtained scans in order to minimize the influence of long using a BRUKER HX270S spectrometer opera- term effects. ting in Fourier transform mode. Spectra were accumulated in 32K data points using a spectral width of 3000 Hz. The repetition time was ~ 5 2.7. Spectral interpretation sec and a pulsewidth of 10 l~sec corresponding to The IH NMR spectra consist in general of a 50 o flip angle was used. The number of large number of lines which may be assigned to accumulated scans varied from 100 to 500 partly overlapping small spin systems, since the depending on concentration and experimental protons in the individual rings do not interfere. conditions. Decoupling experiments have been As a consequence it has been possible to analyse used extensively to confirm the assignments most of the spectra by standard methods. reported in the tables and figures in section 3. The spectrometer was operated at 30 ~ The chemical shift data are referenced to CHC13 (b 3. RESULTS = 7.37 ppm) for CDC13 solutions and to TMS 3.1. Chemical evaluations for solutions in deuterioacetone-d6. In order to According to the Sephadex LH-20 chromatog- improve resolution Gaussian lineshapes have ram (12), the proanthocyanidins were divided been introduced as described by FERmGE and into two groups, the proposed dimers D1 and D2 LINDON (5). and the proposed trimers named T1-T4. Some

O Y

ct

Figure 1. Numbering and labelling in proanthocyanidins straightly C4-Cs-linked.

Carlsberg Res. Commun. Vol. 46, p. 43-52, 1981 45 H. OUTTRUP& K. SCHAUMBURG:IH NMR of proanthocyanidins

a c

7;0 6'.5 6'.0 plDm -6;2 6",1 6.0 ,S.9 5.8 ppm

b d

710 6.5 6".0 ~1 6~2 6.t 6.0 5.9 5.8 ppm Figure 2. The 270 MHz IH spectra of the aromatic regions of two dimeric proanthocyanidins. 2a shows the spectrum of synthetic procyanidin B3, 2b shows the spectrum of prodelphinidin B3 isolated from barley. 2c and 2d show the high field part of 2a and 2b, respectively, on an expanded scale. The impurity observed in the spectrum of B3 is assumed to arise from taxifolin.

of their chemical properties were as follows: ring signals and the aliphatic OH signals is also Common to all six components was the found. These effects prevent the use of integra- liberation of (+)-catechin on mild acid hydrolysis tion in the identification beyond the qualitative (2.3). level. The abundance of signals observed in DI On stronger acid hydrolysis (2.3)the identified and D2 arise from the presence of two conforma- anthocyanidins were the following: tions of the molecules under the experimental D1: Delphinidin conditions. D2 : Cyanidin The spectra of the peracetylated compounds in TI : Delphinidin CDCI 3 show only one dominant isomer and the T2: Delphinidin and cyanidin absence of OH signals makes the assignment T3: Delphinidin and cyanidin possible. In Figure 3 the chemical shifts are T4.. Cyanidin reported for the alicyclic parts of the ring systems in the peracetylated proanthocyanidins studied. The corresponding aromatic part of the 3.2. IH NMR analysis chemical shifts is reported in Figure 4. The In Figures 2a and 2b the aromatic region of accuracy of the spectral analysis was better than the IH NMR spectrum of DI and D2 is 10 -2 ppm. It was found, however, that concen- reproduced. Figure 2c shows the high field part tration and temperature dependent variations of of 2a including the H6 and H8 protons. The the chemical shifts exceeded this value by a factor broad lines interfering with the sharp signals two to five. from the aromatic protons are due to phenolic In order to keep track of the signals due to aU OH. A similar interference between the pyrane the proanthocyanidins investigated in a consis-

46 Carlsberg Res. Commun. Vol. 46, p. 43-52, 1981 H. OUTTRUP & K. SCHAUMBURG: t H NMR of proanthocyanidins

3~' 2ct' 2~'Z~ t. '~ l,ea ' I t.act' 3y 3a. 2~ OCCH3 T3 3 I 3al' ~,,;;.~ t.'~ t, ect, I t.aQt, 3'y 2o. 3~ OCCH3

T2 | , ! ! 3~ I 20.1 2'yI4"y /,'p /.eo. I l.aot. 3'y 3ct 21~ OCCH3

TI 3i3~' 30. 2~;,~ L:p /.ect' I /.oct' 3"y 20. 213 OCCH3

DI 3'p 2r 4p &e= I OCCH3 3a. P ~ar /92 3'p 3dtI 2'p Lp 4c=0. I O(~CH3 2cL /,aa

Procyonidin B~

Epicatechin I t.a| I OCCHI 3 w 2 ~e

Catechin 32I ~e 2a O(~CH3

I I I I 1 6.37 5.37 /.37 3.37 2.37 1.37 ppm Figure 3. Correlation of chemical shifts of (+)-catechin (-)-epicatechin, procyanidin Bh DI-D2 and TI-T4 in CDCI 3 solutions, aliphatic part. tent way we have numbered the atoms in the is possible due to the difference in distance from monomer units from 1-14 and labelled the units H4 to H6 and H8. The ring puckering motion in a, [~, y as indicated in Figure I. (+)-catechin is not well known, but a coarse The assignment of H6 and H8 in (+)-catechin estimate based on Dreiding models makes

Carlsberg Res. Commun. Vol. 46, p. 43-52, 1981 47 H. OUTTRUP~C K. SCHAUMBURG:IH NMR of proanthocyanidins

I I 13p,, 1 %13y l(]p 100, 6p 60. 1/.1~1i/.0. 6Y I 8"/ I I 130. I T3 I ' ' ' ' ' ' 13y 1 lOy 100. 140. 8y 6y 130. 1013

T2 1 130. 10y 100. 140.I 8y 14y lOp 6r 14p

T1 1 0t 10~/' 100.' 6~' 1~0. 140.' I 6y' 14y lOp 6~, 1413 DI 13r' 16p 10o.' 14o.' 60.' ~p lt.p 8p

D2 130. 10 100. lt.a a 813I 13p lt.p 6p

Procyanidin B1 ~a ~p

Epicotechin 10 1~ 1'3 8

Cotechin 1!4 13 10 8

I I J I J 737 7.17 6.97 6.77 6.57 6.37 ppm Figure 4. Correlation of chemical shifts of (+)-catechin, (-)-epicatechin, procyanidin B 1, DI-D2 and TI-T4 in CDC 13 solutions, aromatic part.

expected NOE on H6 about five times larger than The measured coupling constants are reported on Hs. The NOE experiment (2.6) resulted in a in Table I. The main interest is focused on NOE effect 4 times larger on H6 than on Hs. coupling constants in the pyrane ring. The (+)-

48 Carlsberg Res. Commun. Vol. 46, p. 43-52, 1981 H. OUTTRUP ~s K. SCHAUMBURG:IH NMR of proanthocyanidins

Table I

Spin spin coupling constants in peracetylated proanthocyanidins dissolved in CDC 13, J q in Hz

i j BI a B3 a D1 D2 TI T2 T3 T4

2a 3a 10 ...... 3a 4ea - - 4 6.0 .... 3a 4aa - - 5.5 7.5 .... 4ea 4aa - - 16 16.5 17.5 17.5 17.1 16.5 10a 14a - - 2.2 2.0 ~ 1 2.0 2.0 2.0 13a 14ct - - 8.4 7.8 8.8 8.3 8.5 - 21~ 31~ <1 9.5 10.5 9.2 10.2 10.1 10.0 10.2 31~ 4~ 2.1 9.5 9.5 9.2 9.1 8.8 8.8 9.0 61~ 81~ - - 2 2.2 .... 10l~ 14~ ..... 1.8 - 1.8 13~ 14~ - - - 7.8 - 8.3 - - 2v 3v .... 10.2 10.3 10.3 lo.o 3`/ 4V .... 8.4 8.3 8.6 8.5 63, 8`/ .... 2.1 2.4 2.2 2.3 103, 14-/ ...... 2.0 - a Data from ref. 15 and 16. catechin units all give rise to axial-axial orienta- TI. The remaining trimeric compounds appear tion of the proton pairs H2/H3 and H3/H4. The to be procyanidin and prodelphinidin simultane- magnitude of these coupling constants have ously. Thus, the aim of this work was to previously been reported by WEINGES et al. (1 6) establish the configuration of the upper unit of to be approx. 10 Hz. In contrast (-)-epicatechin is the dimeric prodelphinidin and to elucidate the characterized by very small coupling constants structures of the previously unknown three (16). trimeric proanthocyanidins by IH NMR spectro- scopic analyses. From Table I and Figure 3 it is seen that in all 4. DISCUSSION the proanthocyanidins isolated from barley the The chromatographic separations resulted in configuration of the methylene protons in the a- the purification of six different proanthocyani- unit, although shifted towards high field, is as in dins being present in decorticated barley flour. (+)-catechin, i.e. opposite the configuration in The retention data indicate that two dimeric (DI (-)-epicatechin. This is in full agreement with the and D2) and four trimeric (T1-T4) compounds observation that all these proanthocyanidins split are present. The acid hydrolysis (2.3) permits the off (+)-catechin on acid hydrolysis. identification of the lower terminal unit (a in In Figure 2 the aromatic part of the spectra of Figure l) as a (+)-catechin in all six compounds. unacetylated DI and D2 are represented. The For the upper units (~, etc.) the chemical multitude of signals in both compounds correla- identification previously performed (12) is tes closely with the observation made by HASLAM unable to distinguish (+)- and (-)-configurations. et al. (13), that dimeric proanthocyanidins with a Thus it can only be established whether these (+)-oriented ~-unit in acetone exist as a mixture units are catechin or gallocatechin. The polarity of two slowly interchanging conformers. In of six substances differ, judged on the basis of Figure 2c it is seen that two secondary structures chromatography. The polarity is found to are present in almost equal quantities (6). Since increase with the proportion of gatlocatechin. D2 in every respect is identical with procyanidin The dimeric compounds are DI, a prodelphinidin B3 (9, 12, 13, 16, 17), it is assumed that the and D2, a procyanidin. The four trimeric procyanidin D2 is identical with this compound. compounds include a less polar procyanidin T4 The prodelphinidin DI possesses the same and a strongly polar compound prodelphinidin configuration as D2 in the upper unit judged by

Carlsberg Res. Commun. Vol. 46, p. 43-52, 1981 49 H. OUTTRUP& K. SCHAUMBURG:IH NMR of proanthocyanidins the magnitude of the coupling constants (Table accuracy with the data reported by WEINGES et I). Thus, the [~,unit must have a (+)-configura- al. (16). As all the trimeric proanthocyanidins tion, i.e. it must be a (+)-gallocatechin unit. In investigated in this study show a similar pattern, Figure 2d it is seen that the secondary structures it is suggested that CDCI3 interferes with their present in D1 correlate closely with those of molecular configurations yielding only one procyanidin B3. observable rotamer. Choosing CDC 13 as solvent The same prodelphinidin has previously been makes it relatively easy to interpret the spectra isolated from Ribes Sanguinea (7). The reported and to correlate the signals to distinct protons. NMR data coincide with our observations. In the In Figure 3 and Figure 4 the data for assignment of the signals the locations of H2~ procyanidin B3 (D2)and the dimeric prodelphini- and H4~ are, however, reversed. din (191) are represented on the same scale as The mentioned multitude of signals in the those of (+)-catechin, (-)-epicatechin and procya- aromatic region (Figure 2) is accompanied by a nldin B1 (15, 16). The 8-values for the four similar doubling of the line pattern in the trimeric compounds (T1-T4) are represented in aliphatic region. Furthermore the total spectrum the upper halves of the same figures. is hampered by a multitude of OH-signals (com- Significant differences between gallocatechin- pare with Figure 2). In a communication FEENEY containing and catechin-containing substances (4) reports the observation of slow exchange of are found in the aromatic region. Comparing phenolic OH protons with water present in the Figure 2a and Figure 2b it is seen that the acetone used as solvent. It would therefore gallocatechin unit is characterized by a strong appear promising to eliminate the OH-signals by singlet arising from H 10/H 14 protons (cf. Figure exchange with D20. When this experiment was 1). In prodelphinidin B 3 (DI) the Hi0/H 14-signal carried out it was, however, observed that the is situated at 8 = 7.03 ppm. Furthermore, when aromatic signals were diminished as well. It is regarding the locations of such signals, it is assumed that this phenomenon is due to the possible to deduce the number and sequence of existence of quinone tautomeric forms in proan- monomer units. thocyanidinic polyphenols allowing deuterium to Comparing with T1, yielding delphinidin on be incorporated in the aromatic rings. In addition acid hydrolysis, it is seen that this compound the free polyphenols are very sensitive to displays a singlet at 7.03 ppm. This is taken as oxidation, making it very difficult to maintain evidence of a gallocatechin unit as an upper the integrity of such a sample over a longer terminal unit (v-unit). An additional singlet is period necessary for experiments. All these found at 8 = 6.73 ppm. This signal is identified disadvantages made it advantageous to use as a gallocatechin in the [~-unit. These signals and protected compounds for spectroscopic studies. the integral of the spectrum define TI as a Several communications (7, 15, 16) have trimer. advocated the peracetylation adopted in this The singlet indicating a gallocatechin unit in study. the y-position is observed in T2. This informa- Referring to Figures 2, 3 and 4 it is interesting tion, combined with the integral defines T2 as a to notice that the peracetylated compounds only trimer containing only one gallocatechin unit. In display one set of IH NMR signals. HASLAMet al. T3 a similar argument leads to the conclusion (6) reported for solutions of the decaacetate of that this compound contains gallocatechin as the procyanidin B3 at 30 ~ in dimethylsulfoxide-d6 [~-unit. Such an argumentation also excludes the and in nitrobenzene-d5 two sets of signals presence of gallocatechin units in T4. indicating the presence of two conformers in The information obtained by hydrolysis of unequal quantities. Temperature studies yielded TI-T4 do not exclude the presence of (-)- an estimate for the interconversion barrier AG of epicatechin in these compounds. However, the 20 Kcal/mole. The observation of one set of location of the 6[~-signal in procyanidin BI (15, signals for solutions in CDCI 3 therefore would 16) excludes the presence of (-)-epicatechin as a be consistent with one dominating conformer unit in the four trimers investigated. The being present. In this respect our data for aliphatic signals arising from the pyrane rings procyanidin B3 concur within experimental are very similar in all four compounds (cf.

50 Carlsberg Res. Commun. Vol. 46, p. 43-52, 1981 H. OUTTRUP & K. SCHAUMBURG:IH NMR of proanthocyanidins

Figure 3). This indicates that all these proantho- rise to two signals with a separation of 0.13 cyanidins possess the same configuration in all ppm. It may therefore be concluded that the the pyrane rings. The measured coupling con- influence of the secondary structure on the stants (Table I) are consistent with 2,3-trans- chemical shift of H6a is larger than the chemical configurations of the protons, i.e. (+)-catechin shift difference used by Roux (2) to differentiate configuration of the pyrane ring in all cases. between C4/C6 and C4/C8 isomers. Unless the When the sequences of units have been fixed, assumption of equal secondary structure is the question of linkage has to be discussed, i.e. invoked, it seems that the question of this whether the interflavan linkages connect C4 and isomerism in proanthocyanidins is still open for C6 carbon atoms or C4 and C8 atoms. discussion. HASLAM et al. (6, 13) have observed that among the two isomer procyanidins B3 and B6 the former, i.e. the C4--C8 interflavan bonded ACKNOWLEDGEMENTS compound, is the most widespread in nature, the The authors acknowledge the support of The latter representing only 15-20% of the total Danish Science Research council (SNF) making amount. The proof of B3 as being C4-C8 linked the HX270 spectrometer available for the work. was presented by WEINGES et al, (17). Accord- Mrs. JOAN WINTERBERG'S skillful technical as- ingly most authors have claimed investigated sistance is gratefully acknowledged. proanthocyanidins to exhibit C4--C8 linked struc- ture. When investigating more complex systems, REFERENCES i.e. dimeric and trimeric proanthocyanidins, several conformers are possible for the free 1. BATTERHAM, T. J. & R. J. HIGHET: Nuclear magnetic resonance spectra of flavonoids. Aust. forms, but in the case of peracetylated com- J. Chem. 17, 428-439 (1964) pounds, under the conditions employed in this 2. BOTHA, J. J., D. FERREIRA & D. G. Roux: study, only one rotamer was seen for each. The Condensed tannins: Direct synthesis, structure spectral patterns observed for T1-T4 ensure that and absolute configuration of four biflavanoids the isolated chromatographic fractions are ho- from Black Wattle bark b)Mimosa~0 extract. J. mogeneous with respect to interflavan linkages. Chem. Soc., Chem. Comm. 700-702 (1978) Since these compounds are the dominant trime- 3. ENGEL,S. W., M. HATTINGH,H. K. L. HUNOT& ric proanthocyanidins in barley (12) it is D. G. Roux: X-ray structure, conformation, and assumed that in any case the linkages are C4-C8 absolute configuration of 8-bromotetra-O-me- as anticipated in Figure 1. thyl-(+)-catecbin. J. Chem. Soc., Chem. Comm. In the literature several papers have been 695-696 (1978) 4. FEENEY, J. & A. HEINRICH: The use of nuclear concerned with the assignment of H6 and H8 in magnetic double resonance to detect and charac- (+)-catechin. In an early paper BATTERHAM et al. terize phenolic OH groups. J. Chem. Soc., Chem. (1) studied a large number of flavanols by IH Comm. 295-297 (1966) NMR. Based on consideration of substituent 5. FERRIGE, A. G. & J. C. LINDON: Resolution effects they concluded that i5 H8 > ~ H6. enhancement in FT NMR through the use of a Lately Roux et al. (3, 8) have synthesized pure double exponential function. J. Magn. Reson. 31, 6-bromo-(+)-catechin and pure 8-bromo-(+)- 337-340 (1978) catechin and used them to prepare a number of 6. FLETCHER,A. C., L. J. PORTER, E. HASLAM& R. known 6- or 8-substituted . The obser- K. GUPTA: Plant Proanthocyanidins. Part 3. vation of the IH NMR spectra of such deriva- Conformational and configurational studies of natural plant proanthocyanidins. J. Chem. Soc., tives, lead to the suggestion of an empirical rule Perkin I, 1628-1636 (1977) whereby i5 H8 - i5 H6 ~ 0. l ppm (2). In no case 7. Foo, L. Y. & J. PORTER; Prodelphinidin overlapping of these signal ranges was observed. polymers: Definition of structural units. J. In the previous discussion it has already been Chem. Soc., Perkin I, 1186-1190 (1978) mentioned that proanthocyanidins may exist in 8. HUNDT, H. K. L. & D. C. Roux: Condensed two long-lived secondary structures (6). This tannins: Determination of the point of linkage in was seen in Figures 2c and 2d, where H6a gives >>Terminal~ (+)-catechin units and degradative

Carlsberg Res. Commun. Vol. 46, p. 43-52, 1981 51 H. OUTTRUP & K. SCHAUMBURG:IH NMR of proanthocyanidins

bromination of 4-flavanylflavan-3,4-diols. J. N. TANNER: Plant proanthocyanidins. Part I. Chem. Soc., Chem. Comm. 696-698 (1978) Introduction; the isolation structure and distribu- 9. JERUMANIS, J.: Separation et identification de tion in nature of plant proanthocyanidins. J. flavanoides par Chromatographie Liquide a Chem. Soc., Perkin I, 1387-1399 (1972) Haute Performance (HPLC). Proceedings of the 14. VANCRAENENBROECK, R., H. GOmSSEN & R. European Brewery Convention Congress, Berlin LONTm: Separation of beer flavanoids by column 1979, 309-319. chromatography. Proceedings of the European 10. LAWS, D. R. J., J. D. McGuINNESS & N. A. Brewery Convention Congress, Amsterdam BAXH: The use of 14C-labelled dimeric eatechin 1977, 429--439 to study the fate of dimeric potyphenols during 15. WEINGES, K., H. GORISSEN & R. LONTm: Les malting and brewing. J. Am. Soc. Brew. Chem. Proanthocyanidines naturelles. Ann. Physiol. 34 (4), 170-173 (1976) v~g. 11 (1), 67-92 (1969) 11. McMuRROUGH, I.: Chromatographic procedures 16. WEINGES, K., K. GORITS& F. NADER: Konfigura- for measuring polyphenol haze precursors. tionsbestimmung von C30 H26 O12 - Proantho- Proceedings of the European Brewery Conven- cyanidinen und Structuraufklarung eines neuen tion Congress, Berlin 1979 (321-334) Procyanidins. Liebigs Ann., 715, 164-171 12. OOTXRUP, H.: Structure of prodelphinidins in (1968) barley. Proceedings of the European Brewery 17. WEINGES, K., J. PERNER & H.-D. MARX: Convention Congress, Copenhagen 1981 (in Synthese des Octamethyl-diacetyl-proanthocya- press) nidins B3. Chem. Ber. 103, 2344-2349 (1970) 13. Tr~OMPSON,R. S., D. JACOUES,E. HASLAM& R. J.

52 Carlsberg Res. Commun. Vol. 46, p. 43-52, 1981