The use of extracted-ion chromatograms to quantify the composition of condensed tannin subunits

Mouls Laetitia (Orcid ID: 0000-0002-7861-9685)

The use of extract-ion chromatograms to quantify the composition of condensed

tannin subunits

Short title: A better analysis of tannin subunits by mass spectrometry

Julien Giribaldi ^a,b,c, Martin Besson ^a,b,c, Lucas Suc^a,b,c, Hélène Fulcrand ^a,b,c, and Laetitia Mouls ^a,b,c

^aINRA, UMR1083, F-34060 Montpellier, France



^bMontpellier SupAgro, UMR1083, F-34060 Montpellier, France

^cUniversité de Montpellier, UMR1083, F-34060 Montpellier, France

 Corresponding author. Tel.: +3-349-961-2454; fax: +3-349-961-2857; e-mail:

laetitia.mouls@supagro.fr

This article has been accepted for publication and undergone full peer review but has not been

through the copyediting, typesetting, pagination and proofreading process which may lead to

differences between this version and the Version of Record. Please cite this article as doi:

10.1002/rcm.8619

ABSTRACT

RATIONALE: The analysis of tannins is currently usually done by liquid chromatography

following their chemical depolymerization. One of the main limitations of this method,

however, is the difficulty with specifically detecting each constitutive tannin monomer in a

complex matrix, as numerous compounds co-elute with the monomers, thereby

compromising the analysis.

METHODS: Tannin depolymerization under acidic conditions and in the presence of

thioglycolic acid releases the various constitutive monomers, either as terminal or extension

units. The tannin subunits were then quantified by extracted-ion chromatogram (EIC) mass

spectrometry, which required determination of the ionization efficiency of the monomers.

Thus, we used AgBF4-assisted coupling for the hemisynthesis of the non-commercial ECG

dimer.

RESULTS: The EIC showed that the derivatives of the extension units were twice as ionized

as the terminal unit. Unlike the use of UV chromatograms, this new EIC-based method is

more specific and accurate because it is not impacted by the co-elution phenomenon. This

result, when added to the linearity obtained on a large range of the calibration curves, allowed

for quantification of tannin subunits from EIC in complex mixtures with less pretreatment of

the samples.

CONCLUSIONS: As a result of its specificity and sensitivity, this EIC-based method

represents a significant step toward improvement of the quantification of the tannin

composition of samples. The results of this study should allow the oxidation markers to be

quantified more accurately and taken into account, thereby providing a better indication of

the actual tannin composition.

Keywords: Tannin, polyphenol, average degree of polymerization aDP, extracted-ion

chromatogram EIC, ECG-ECG

INTRODUCTION

Condensed tannins (also called ‘proanthocyanidin’) are polyphenol biopolymers of flavan-3-

ol subunits and they are widely distributed in the plant kingdom. As a first approximation,

they can be considered to be a linear[1] or cyclic[2, 3] sequence of monomeric units such as

catechin (C), epicatechin (EC), epicatechin gallate (ECG), and epicatechin 3-O-gallate (EGC)

(Figure 1). The link between the monomeric subunits is called the interflavanic bond, and it is

usually a C4-C8 bond (rarely a C4-C6 bond)[4, 5]. Tannins are well known for their

antioxidant activity, which is thought to inhibit cellular aging[6]. They are, consequently,

highly sensitive to oxidation, which results in their chemical structure becoming more

complex.

Due to the high structural diversity of condensed tannins and the relatively low concentration

of some of their constituents, analyses of such polymers can be quite challenging, especially

for wines, as the chemical structure of tannins has a direct impact on their organoleptic

properties such as astringency[7], bitterness[8, 9], color[10, 11], and colloidal stability[12,

13].

The most common methodology to establish the average molecular composition of

condensed tannins, and to quantify them, starts with chemical depolymerization. This is

carried out in an acidic medium in the presence of a nucleophile. Chemical depolymerization

of condensed tannins leads to cleavage of the interflavan bonds, thereby releasing terminal

units and extension units that are substituted by the nucleophile at their C4 carbon. These

depolymerized products are then typically analyzed by liquid chromatography using UV

detection (Figure 1 and Table S1). The most commonly used nucleophiles are benzyl

mercaptan, thioglycolic acid, mercaptoethanol, and phloroglucinol[14-22]. The

depolymerization method does not provide access to the polydiversity polymeric distribution

in tannins, although it is a suitable way to generate ample information in regard to their

chemical characterization. The advantage of this method is that it readily allows for

determination of both the composition of constitutive flavan-3-ol (terminal and extension)

subunits of tannins, in addition to the average degree of polymerization (aDP). However, this

method fails to take into account the different transformations that tannins undergo

throughout the winemaking process and its storage. These transformations are mainly

induced by oxidation.

The oxidation of tannins results in several changes to their chemical structure. It generally

results in new covalent bonds that connect at least two flavan-3-ol constitutive subunits.

Since these new bonds are not broken under the depolymerization conditions, a set of biaryl-

C-C connected dimers, trimers, or higher oligomers of flavan-3-ol derivatives are generated

by this reaction[23]. These compounds are thought to be good markers of oxidation[23, 24].

They exhibit a very high level of structural diversity, and each of them is present at a very

low concentration. Hence, in chromatogram profiles, the corresponding peaks exhibit very

low intensities and poor resolution. They present as a broad “hump”, instead of separate

peaks. As none of these markers can be detected as a definite peak on UV chromatograms,

they cannot be quantified based on this method of detection[23, 25]. They co-elute with non-

oxidized depolymerized flavan-3-ol units and other phenolic compounds such as phenolic

acids and anthocyanins, and they are not taken into account in the analysis of the tannic

fraction.

In light of the limitations of the current analytical procedure for quantification of the

composition of condensed tannin subunits of a sample, there is a clear need for a specific

analytical procedure that is not compromised by this co-elution phenomenon and the

overlapping of peaks.

New and improved approaches using mass spectrometry have been developed in recent years.

Mass spectrometry has proven to be a more efficient method than UV detection for the

quantification of flavonoids[26-28]. Indeed, mass spectrometry appears to be a more accurate

method with greater selectivity and sensitivity than UV detection[26]. However, due to the

complexity of natural samples, an LC-separation step is required prior to MS analysis in

order to limit the ionization suppression[29, 30] and to increase the quality of the MS

analysis[31]. Liquid chromatography coupled to mass spectrometry is a very attractive

method for flavonoid analysis of complex samples[32]. A UPLC system is also used

nowadays, in part to reduce the time required for the analysis[33-36].

More recently, flavonoids have been quantified by various mass spectrometry methods such

as MRM[27] or extract-ion chromatograms (EIC)[26].

Engström et al.[34] developed a method to analyze tannins by depolymerization directly in

the ion source after separation by liquid chromatography. This innovative method has many

advantages because no chemical depolymerization is required and because of the tannin

content, the procyanidin/prodelphinidin ratio, and the average degree of polymerization can

be estimated based on mass spectrometry analysis. However, this method still suffers from

co-elution of tannins in complex mixtures because a large hump containing a mixture of

tannins is observed in the UPLC chromatograms of some sample. The separation of tannins is

one of the limiting factors of liquid chromatography. Moreover, a single cone voltage cannot

be used for optimal detection of all tannin types and the results can be compromised based on

which tannins are co-eluted.

The high-throughput analysis of polyphenols using the MRM method quantifies constitutive

units of tannins as being equivalent to catechin for flavan-3-ols due to the lack of standard

products, as described by Lambert et al.[27]. Depending on the various monomers present in

tannins, the ionization efficiency can nevertheless differ and skew the result.

The EIC method offers several advantages over the previously used methods. The method

developed by Stanoeva et al. has shown that EIC had a higher selectivity and sensitivity

compared to UV spectrophotometry when it was used for HPLC detection in the analysis of

20 flavonoid monomers[26].

In light of the difficulty analyzing tannins in complex mixtures, the novel method developed

in this work was aimed at improving the characterization of tannins in such mixtures. We

decided to use UPLC coupled to ion trap MS. Compared to HPLC, the higher column

pressure with UPLC increases the working flow and thus improves the sensitivity of the

analysis. This method is based on the relative quantification of the different tannin subunits,

released after chemical depolymerization, from the EIC. The EIC method offers several

advantages. As a result of its high specificity and sensitivity, it provides better detection of

tannin subunits. By overcoming the issues related to co-elution, the tannin samples need to

undergo less pretreatment. However, development of this method requires determination of

the ionization efficiency of the extension subunit derivatives (EC-Nu, C-Nu, EGC-Nu, and

ECG-Nu) and of the terminal (EC, C, EGC, and ECG) units. For this reason,

proanthocyanidin dimers were depolymerized and investigated by means of their EIC. This

method was then applied to wine samples, for which the condensed tannins are composed of

catechin, gallocatechin, epicatechin, epigallocatechin, and epicatechin gallate[4, 37-40].

In the present work, this method was applied using non-oxidized units. Unlike oxidation

markers, standards tend to be more readily available for non-oxidized units. Hemisynthesis of

oxidation markers is required to obtain standards of oxidation markers, but this is an arduous

process. Once the method has been implemented with non-modified subunits, it can then be

applied to oxidized units after calibration with the standard obtained by hemisynthesis.

The aim of this work was to facilitate tannin characterization by improving the analytical

procedure.

EXPERIMENTAL

Reagents and samples

(+)-Catechin (≥ 99%), (-)-Epicatechin (≥ 99%), (-)-Epicatechin gallate (≥ 99%), and

procyanidin B2 (≥ 90%) were purchased from Extrasynthese S.A. (Genay, France).

Thioglycolic acid (minimum 98%) and MgSO₄(≥ 99.5%) were purchased from Sigma-

Aldrich. EtAcO (pure) was purchased from Carlo Erba Reagents S.A.S.(Peypin France).

Grape seed polyphenols were purchased from Sucren Groupe UDM (Vallon Pont d’Arc

France). A polyphenolic extract was obtained from red wine (a six-year-old wine made by

traditional winemaking using the Marselan grape variety, produced by the INRA Pech Rouge

experimental unit, Gruissan, France) according to the following process: 1 mL of red wine

was added to a reverse-phase C18 cartridge (20 cc, 5 g) after column conditioning (addition

of 2 x 5 mL of MeOH then 2 x 5 mL of H₂O/CH₃COOH, 98/2, v/v, this operation was carried

out twice). Two 5 mL volumes of H₂O/CH₃COOH, 98/2, v/v were introduced to remove the

sugar fraction. The polyphenol extract was eluted with methanol (2 x 5 mL), and the

methanol was then removed under vacuum. Fractionation of the polyphenolic extract from

the one-year-old red wine was performed using fractogel chromatography[41, 42] (methyl

acrylate copolymer solution in aqueous ethanol 20%, HW-50 F) with a flow rate of 5

mL.min^-1. The fractionation was achieved with EtOH/TFA (99.95/0.05, v/v), H₂O/TFA

(99.95/0.05, v/v) gradients. The highest molecular weight phenols were then eluted with a

CH₃COCH₃/H₂O/TFA solution (70/29.95/0.05 v/v/v). The gradient conditions were solvent A

(H₂O/TFA, 99.95/0.05, v/v); solvent B (Ethanol/TFA, 99.95/0.05, v/v); solvent C

(Acetone/H₂O/TFA, 69.97/29.97/0.05, v/v/v); 0-10.2 min isocratic 100% A; 10.2-11 linear

84% A and 16% B; 11-12.2 linear 60% A and 40% B; 12.2-22.2 isocratic 60% A and 40% B;

22.2-24.2 linear 45% A and 55% B; 24.2-34.2 isocratic 45% A and 55% B; 34.2-36.2 linear

100% C; 36.2-46.2 isocratic 100% C. The fraction studied in this work was the one that

eluted with H₂O/TFA (99.95/0.05, v/v) 45%; EtOH/TFA (99.95/0.05, v/v) 55%.

Analytical procedures

UPLC-ESI/MS analyses

The analyses were performed by liquid chromatography coupled to a mass spectrometer

(UPLC-ESI-MS) controlled by HyStar 3.2 software. The compounds were separated on a

UPLC HSS C18 column (1.8 μm; 1.0 x 150 mm) with a flow of 0.10 mL.min^-1using an

Acquity UPLC liquid chromatography system (Waters, Milford, MA) equipped with a diode

array detector. The gradient conditions were solvent A (H₂O/HCOOH, 99/1, v/v); solvent B

(CH₃CN/H₂O/HCOOH, 80/19/1, v/v/v); initial 0.1% B; 0-3 min linear 25% B; 3-5 min linear

35% B; 5-7 min isocratic 65% B; 7-8 min linear 99.9% B; 8-9 min isocratic 99.9% B; 9-10

min linear 0.1% B; and 10-11 min isocratic 0.1% B. The MS analyses were performed with

an amaZon X ESI Trap mass spectrometer (Bruker Daltonics, Bremen, Germany). In source,

the nebulizer pressure was 44 psi, the temperature of the dry gas was set at 200 °C with a

flow of 12 L.min^-1, and the capillary voltage was set at 5.5 kV. The mass spectra were

acquired over a mass range of 70-2000 Th in positive ionization mode. The speed for the

mass spectrum acquisition was set at 8.1 Th.min^-1.

Synthesis of ECG propylsulfanyl derivative (1)

Ninety-nine milliliters of MeOH, 9.61 mL (13.5 mmol, 15 eq) of propan-1-thiol, and 1 mL of

HCl 37% w/v were added to 2 g of grape seed polyphenols in a round-bottom flask. The

mixture was stirred for 1 hour at 40 °C and 100 mL of water was added. The methanol was

removed under vacuum and the aqueous phase was extracted three times with 100 mL of

ethyl acetate. The combined organic layers were dried over anhydrous MgSO₄and the EtAcO

was removed under vacuum, yielding 2.2 g of a brown solid. Five-hundred milligrams of this

crude mixture were purified on a C18 reverse-phase silica column (50 g of C18 silica, 40-60

μm particle size) at a flow rate of 35 mL.min^-1. The gradient conditions were solvent C

(H₂O); solvent D (CH₃CN); 0-2 min isocratic 0% D; 2-4 min linear 25% D; 4-8 min isocratic

25% D; 8-9 min linear 35% D; 9-15 min isocratic 35% D; 15-19 min linear 100% D; and 19-

24 min isocratic 100%. After CH₃CN evaporation and freeze-drying for 24 hours, 46.49 mg

(purity ≥ 90%) of a light brown fine powder of ECG propylsulfanyl derivative were

recovered. Analysis: UPLC-MS rt=7.4 min; m/z (calculated)=517.11 Th;

[M+H]⁺(found)=517.2 Th.

Synthesis of ECG-ECG dimer (2)

Thirteen milligrams of ECG (29.41 μmol, 1 eq) and 15.18 mg (29.41 μmol, 1 eq) of

epicatechin gallate propylsulfanyl derivative were added to a round-bottom flask (1). The

substrates were dissolved in 433 μL of THF at 0 °C. Then, 14.31 mg of AgBF₄were added

(73.56 μL, 2.5 eq). The reaction mixture was stirred for 10 min at 0 °C, and then 8 mL of

water was added. The aqueous phase was extracted three times with 10 mL of EtAcO. The

combined organic layers were dried over anhydrous MgSO₄and the EtAcO was removed

under vacuum, yielding a pale yellow-orange solid. The entire crude mixture was purified by

preparative HPLC. A 940 HPLC preparative LC Varian device controlled by Galaxy software

was used. A UV-visible dual channel detector was used and the column was a Microsorb

100-3 μm C18 (100 x 21.4). The pressure was 2,400 psi at the beginning of the run. The

gradient conditions were solvent A1 (H₂O/CH₃COOH 99/1 v/v) and B1

(CH₃CN/H₂O/CH₃COOH 80/19/1 v/v/v); initial 20% B1; 0-10 min linear 22% B1; 10-13 min

linear 25% B1; 13-15 min linear 80% B1; and 15-17 min linear 20% B1.

The compound was eluted with 22% of B1 from 8.6 min to 9.2 min. Following elution, the

solvent was removed under vacuum. The aqueous layer was extracted three times with 50 mL

of EtAcO. The combined organic layers containing ECG-ECG were dried under vacuum. The

ECG-ECG dimer was dissolved in H₂O and lyophilized, yielding 2.85 mg of a light-brown

solid resin of ECG-ECG dimer (purity ≥ 95%; yield=11%). Analysis: UPLC-MS rt=4.4 min;

m/z (calculated)=883.16 Th; [M+H]⁺(found)=883.2 Th.

Chemical depolymerization

Thioglycolysis: A 1 mg. L^-1solution of each sample (B2, ECG-ECG, the polyphenolic

extract, and the polyphenolic fraction) was prepared in methanol. One-hundred microliters of

the solution was introduced into a 250 μL glass insert and 100 μL of thioglycolic acid

solution (20 μL of thioglycolic acid, 4,880 μL of methanol, and 100 μL of HCl 37% w/v) was

added. After sealing the glass insert with an inert cap, the depolymerization reactions were

carried out at 90 °C for 6 min.

For each tannin fraction, the thioglycolysis reactions were performed in triplicate (indices 1,

2, or 3 on Table 1) and, after depolymerization, each sample was injected by UPLC-MS three

times (indices a, b, or c on Table 1).

RESULTS AND DISCUSSION

Development of the MS method

The way that laboratories currently characterize the tannin fraction of samples is by

determination of the aDP[43-47]. The aDP is defined, after the chemical depolymerization of

∑

푇표푡푎푙 푠푢푏푢푛푖푡푠

the tannins, as

, and is determined by integration of the extension and

∑

푇푒푟푚푖푛푎푙 푠푢푏푢푛푖푡푠

terminal unit areas from the UV chromatogram obtained after HPLC analysis (Figure 1).

The quantification of subunits of tannins has been based on integration of the various peak

areas from the UV chromatogram. This UV-based quantification is distorted to a certain

degree as a result of many co-elution phenomena due to the presence of other compounds in

the samples, as well by oxidation of a portion of the tannins.

This limitation is linked with the use of UV analysis as a quantification method for tannins,

and it is possible to overcome this by developing a new method based on EIC generated by

mass spectrometry (Figure 2). After chemical depolymerization, the various released subunits

are separated by UPLC and analyzed by ESI-MS. In order to be able to transpose the work

already done, this current work is based on the same UPLC-MS method used previously to

detect and identify oxidation markers [23, 24].

The quantification of the subunits requires evaluation of the ionization efficiency ratio

between the terminal units and the extension derivatives generated after the

depolymerization. Estimation of the ionization efficiency based on each unit obtained after

purification is not sufficiently accurate. Indeed, a large number of errors could arise in the

purification and in the preparation of the standards, which could distort the analysis. To be

sure of the ionization efficiency ratio, pure tannin dimers, in which the amounts of terminal

and extension units are equivalent, were depolymerized.

In this work, the ionization ratios were determined after depolymerization of the

commercially available B2 (EC-EC) dimer and the synthesized ECG-ECG dimer (which is

not commercially available).

Hemisynthesis of ECG-ECG dimer

The synthesis of proanthocyanidin dimers has already been described in the literature by

Steynberg et al.[48]. The optimized protocol is based on reacting 4β-benzenesulfanyl

epicatechin and catechin in the presence of AgBF₄(2.5 equiv.) for 1 h at 0 °C to obtain

proanthocyanidin B1 with a yield of 38%. We altered this reaction scheme by using ECG

propylsulfanyl derivative (1) and epicatechin gallate to generate ECG-ECG dimers (2) with a

yield of 11% (Figure 3). Propanethiol was used instead of benzenethiol due mainly to the

polarity of the former, thereby enabling good purification and isolation.

Analytical study of dimers by mass spectrometry

The calibration curves for each monomer showed that catechin and epicatechin had a similar

degree of ionization, but this differed for EGC and ECG. Indeed, the EGC unit was 1.14

times less ionized than catechin and epicatechin and the ECG was 1.35 times more ionized

than C and EC.

The UV chromatogram at 280 nm generated before and after the chemical depolymerization

(by means of thioglycolysis) of the ECG-ECG dimer showed that the area of the dimer was

equal to the sum of areas of the extension and the terminal units released after the

depolymerization. Furthermore, the area for the terminal units was equal to that of the

extension units. The ratio between the terminal and the extension units in the UV profile was,

therefore, equal to one. The sulfanyl 2-methyl acetate group was not expected to change the

absorbance coefficient at 280 nm of the extension unit derivative. Therefore, the same area

should be obtained for the terminal and extension units. However, the EIC indicated that the

area of the extension units was two-fold greater than that of the terminal units (Figure 4 and

Table S2). This was also observed in the case of B2. We can, therefore, assume that the

extension units EC and ECG were ionized twice as much as the terminal units. As C and EC

are two epimeric forms of one monomer, and as EGC is structurally closer to EC and C than

to ECG, it can be expected that their respective extension units should exhibit similar

ionization properties. Thus, we can make the assumption that the extension units of C and

ECG were twice as ionized as the terminal as EC and ECG. The UPLC-MS analysis allows

comparison of the ionization factor of the extension and terminal units by integration of the

EIC. Moreover, the ratio of ionization of each monomer (C, EC, EGC, and ECG) must be

taken into account in the non-oxidized subunits estimation.

Using the EIC method to assay the non-oxidized subunits of wine tannins

Two polyphenol samples that had been extracted from red wine were studied: the entire

polyphenol extract and a polyphenolic fraction, eluted in the first part of the fractionation of

the polyphenol extract. These samples were chosen because it is difficult to determine their

tannin composition by the routine UV method. Indeed, the polyphenolic extract is composed

of various classes of phenolic compounds (anthocyanins, phenolics acids, flavonols, etc.)[11]

that co-elute under UPLC separation conditions. To limit this co-elution phenomenon, it is

necessary to separate the tannins before the UV analysis takes place. However, the first

fractions are still difficult to analyze after fractionation because they often contain low

monomers (Figure 5a). The current UV analysis procedure is mainly only suitable for

fractions with high molecular weight entities (i.e., tannin fractions eluted at the end of a

CH₃COCH₃/H₂O/TFA solution gradient, Figure 5b) and not for the entire tannin fraction. In

addition, as a result of their reactivity, tannins undergo structural modifications, while a

portion of them can also be lost by adsorption on the stationary phases[49]. Thus, limiting

pretreatments appears to be essential in order to be able to determine tannin compositions

with a greater degree of accuracy.

The quantities of total non-oxidized subunits estimated by UV chromatogram integration

exceeded those obtained from EIC 18-fold and more than 200-fold for the polyphenolic

extract and the polyphenolic fraction, respectively (Table 1). This highlights the need to

replace UV detection with EIC when analyzing complex samples. Indeed, this overestimation

by UV chromatogram analysis is mainly due to the substantial co-elution phenomenon that

stems from the complexity of the samples. We have observed numerous co-elutions with

anthocyanins on UV chromatograms: malvidin-3-O-glucoside co-eluting with epicatechin,

malvidin-3-O-(6-O-acetyl-glucoside) co-eluting with the extension units of EGC (EGC-Nu),

and malvidin-3-O-(6-O-coumaroyl-glucoside) co-eluting with extension units of EC (EC-Nu)

(Supplementary data, Figure S2). For the samples analyzed in this work, the co-elution of

malvidin-3-O-(6-O-acetyl-glucoside) with the extension units of EGC (EGC-Nu) profoundly

skewed the estimation. Depending on the extent of these co-elutions, they tend to result in

varying degrees of over or underestimation of the amounts of non-oxidized tannin subunits

(co-elutions with terminal or extension units, respectively). In addition to the co-elutions with

anthocyanins, we also observed co-elutions with oxidation markers, which are sometimes not

insignificant depending on the oxidation state of the wine product. For example, the mass

spectrum obtained at the retention time of EGC-Nu revealed co-elution with malvidin-3-O-

(6-O-acetyl-glucoside) as well as with numerous compounds (Supplementary data Figure

S3). The analyses of these samples by the conventional HPLC system also revealed the

limitations of UV chromatograms for non-oxidized subunits of tannins. Indeed, with HPLC,

although anthocyanins did not co-elute with the EC-Nu subunits, other co-elutions could also

distort the results. The non-oxidized subunit content was also overestimated with HPLC

analysis of total polyphenolic extract. For the polyphenolic fraction, no subunits were

detected by the current UPLC method, while the EIC method allowed their detection.

These results show that, depending on the sample, the levels of non-oxidized subunits of

tannins from UV chromatograms can be significant. The EIC method allows the problems

associated with co-elution to be circumvented, thereby allowing more accurate determination

of the quantity of non-oxidized subunits of tannins.

Although ion trap mass spectrometry is not the most suitable system for quantification,

calibration curves were generated from the UV chromatograms and from the EIC for each

monomer (EC, EGC, ECG) (Supplementary data, Figure S4 and Table S3). The calibration

curves from the EIC were linear in the concentration ranges expected in wine samples. Thus,

one can envisage using EIC for quantification within a certain concentration range. However,

it should be pointed out that the calibration curves were generated for pure monomers. The

presence of several co-eluted molecules could impact their ionization in the source of the

mass spectrometer, thereby distorting the EIC and their quantification. In the present

analytical conditions, EC (which is the terminal unit of B2) was co-eluted with malvidin-3-O-

glucoside. To test the robustness of this EIC method, various ratios (4/1, 2/1, 1/1, 1/2, and

1/4) of a mixture of B2 and malvidin-3-O-glucoside within a concentration range from 7.5

μmol. L^-1to 120 μmol. L^-1were chemically depolymerized prior to mass spectrometry

analysis. This study showed that the co-elution of malvidin-3-O-glucoside with EC did not

alter the EIC response when the concentration of the malvidin-3-O-glucoside did not exceed

60 μmol. L^-1, irrespective of the B2/malvidin-3-O-glucoside ratio. The concentration of

malvidin-3-O-glucoside of the samples analyzed in this study was significantly lower than

this value (< 7.5 μmol. L^-1). Dilution will be required for samples with high concentrations (>

60 μmol. L^-1) of anthocyanins, malvidin-3-O-glucoside, malvidin-3-O-(6-O-acetyl-

glucoside), and/or malvidin-3-O-(6-O-coumaroyl-glucoside), in order to accurately determine

the amount/concentration of tannin subunits.

Currently, the use of aDP is not suitable for determination of tannin compositions because the

oxidized subunits are not taken in account and it is based only on the non-oxidized subunits

present in the tannin chains (Supplementary data Figure S1). The tannin fraction can be

altered to a certain extent by oxidation. The value of aDP from such fractions does not reflect

the actual composition of the entire tannin extract and should not be used for analysis of

products containing modified tannins.

To be as exhaustive as possible, the percentage of oxidized tannin should be taken in to

account for tannin characterization. For example, more than two-thirds of tannin polymeric

chains are modified in the polyphenolic fraction from red wine studied in this work. This can

be estimated from the yield of depolymerization. The tools and methods used to date do not

allow this modified part to be taken into account in the aDP estimation due to the high level

of complexity of the mixture of products. However, establishing the proportion of modified

tannin and accurate quantitation of the non-oxidized subunits of tannins should provide a

more accurate assessment of tannin compositions.

CONCLUSIONS

Quantification of tannin subunits is difficult due to the limitations of tannin analytical

procedures based on UV chromatography. In this work, we developed a new analytical assay

based on EIC. As a result of the specificity and the sensitivity of the EIC, the quantity of

tannin subunits could be accurately determined in chemically complex samples.

The depolymerization of flavan-3-ol dimers (B2 and the synthesized ECG-ECG) revealed

that sulfanyl-2-methyl acetate derivatives (extension units) were twice as ionized compared

to the terminal units. This result allowed for better determination of the tannin levels in wine

extract, which was greatly overrated using UV detection. Moreover, the use of EIC allows

characterization of complex samples, which have been difficult to study to date, and a

reduction in pretreatments. In a given concentration range, tannin quantification may also be

feasible.

This method, by using EIC, is a first step toward improvement of the quantification of the

tannin composition of samples. These outcomes show that EIC-MS analysis is more

amenable to quantification than UV chromatograms. As a result, this study will allow, in the

future, oxidation markers to be accurately quantified and taken into account so as to obtain a

more representative indication of the actual tannin composition.

ACKNOWLEDGMENTS

We thank Frédéric Véran for helping us with the tannin analyses by HPLC after chemical

depolymerization, Jean-Paul Mazauric for his generous assistance with the preparative

HPLC, and Chahinez Aouf for proofreading this manuscript.

FUNDING SOURCES

Masters funding was supported by the CASDAR project.

REFERENCES

1.

2.

3.

Cheynier, V., et al., Structure and properties of wine pigments and tannins. American Journal

of Enology and Viticulture, 2006. 57(3): p. 298-305.

Longo, E., et al., Distribution of crown hexameric procyanidin and its tetrameric and

pentameric congeners in red and white wines. Food Chemistry, 2019. 299: p. 1-9.

Zeng, L.M., et al., Crown Procyanidin Tetramer: A Procyanidin with an Unusual Cyclic

Skeleton with a Potent Protective Effect against Amyloid--Induced Toxicity. Molecules, 2019.

24(10): p. 1915.

4.

Santosbuelga, C., E.M. Franciaaricha, and M.T. Escribanobailon, COMPARATIVE FLAVAN-3-OL

COMPOSITION OF SEEDS FROM DIFFERENT GRAPE VARIETIES. Food Chemistry, 1995. 53(2):

p. 197-201.

5.

De Bruyne, T., et al., Condensed vegetable tannins: Biodiversity in structure and biological

activities. Biochemical Systematics and Ecology, 1999. 27(4): p. 445-459.

Lotito, S.B., et al., Metabolic conversion of dietary flavonoids alters their anti-inflammatory

and antioxidant properties. Free Radical Biology & Medicine, 2011. 51(2): p. 454-463.

Zanchi, D., et al., Colloidal stability of tannins: astringency, wine tasting and beyond. Journal

of Physics-Condensed Matter, 2008. 20(49).

6.

7.

8.

Robichaud, J.L. and A.C. Noble, ASTRINGENCY AND BITTERNESS OF SELECTED PHENOLICS IN

WINE. Journal of the Science of Food and Agriculture, 1990. 53(3): p. 343-353.

Llaudy, M.C., et al., New method for evaluating astringency in red wine. Journal of

Agricultural and Food Chemistry, 2004. 52(4): p. 742-746.

9.

10.

Sims, C.A. and J.R. Morris, A COMPARISON OF THE COLOR COMPONENTS AND COLOR

STABILITY OF RED WINE FROM NOBLE AND CABERNET SAUVIGNON AT VARIOUS PH LEVELS.

American Journal of Enology and Viticulture, 1985. 36(3): p. 181-184.

Alcalde-Eon, C., et al., Color Stabilization of Red Wines. A Chemical and Colloidal Approach.

Journal of Agricultural and Food Chemistry, 2014. 62(29): p. 6984-6994.

Poncet-Legrand, C., et al., Flavan-3-ol aggregation in model ethanolic solutions: Incidence of

polyphenol structure, concentration, ethanol content, and ionic strength. Langmuir, 2003.

19(25): p. 10563-10572.

11.

12.

13.

14.

Fabre, S., et al., Colloidal behavior of wine galloylated tannins. Comptes Rendus Chimie,

2010. 13(5): p. 561-565.

Souquet, J.M., et al., Phenolic composition of grape stems. Journal of Agricultural and Food

Chemistry, 2000. 48(4): p. 1076-1080.

15.

Rigaud, J., et al., Micro method for the identification of proanthocyanidin using thiolysis

monitored by high-performance liquid chromatography. Journal of Chromatography, 1991.

540: p. 401-405.

16.

17.

Kennedy, J.A. and A.W. Taylor, Analysis of proanthocyanidins by high-performance gel

permeation chromatography. Journal of Chromatography A, 2003. 995(1-2): p. 99-107.

Betts, M.J., et al., DEGRADATION OF CONDENSED TANNINS - STRUCTURE OF TANNIN FROM

COMMON HEATHER. Chemical Communications (Cambridge, United Kingdom), 1967(21): p.

1110-&.

18.

19.

20.

21.

Foo, L.Y., G.W. McGraw, and R.W. Hemingway, CONDENSED TANNINS - PREFERENTIAL

SUBSTITUTION AT THE INTERFLAVANOID BOND BY SULFITE ION. Journal of the Chemical

Society-Chemical Communications, 1983(12): p. 672-673.

Tanaka, T., et al., CHEMICAL EVIDENCE FOR THE DE-ASTRINGENCY (INSOLUBILIZATION OF

TANNINS) OF PERSIMMON FRUIT. Journal of the Chemical Society-Perkin Transactions 1,

1994(20): p. 3013-3022.

Kennedy, J.A. and G.P. Jones, Analysis of proanthocyanidin cleavage products following acid-

catalysis in the presence of excess phloroglucinol. Journal of Agricultural and Food Chemistry,

2001. 49(4): p. 1740-1746.

Gea, A., et al., In Situ Analysis and Structural Elucidation of Sainfoin (Onobrychis viciifolia)

Tannins for High-Throughput Germplasm Screening. Journal of Agricultural and Food

Chemistry, 2011. 59(2): p. 495-503.

22.

23.

Orejola, J., et al., Characterization of Proanthocyanidin Oligomers of Ephedra sinica.

Molecules, 2017. 22(8).

Mouls, L. and H. Fulcrand, UPLC-ESI-MS study of the oxidation markers released from tannin

depolymerization: toward a better characterization of the tannin evolution over food and

beverage processing. Journal of Mass Spectrometry, 2012. 47(11): p. 1450-1457.

Mouls, L. and H. Fulcrand, Identification of new oxidation markers of grape-condensed

tannins by UPLC–MS analysis after chemical depolymerization. Tetrahedron, 2015. 71(20): p.

3012-3019.

24.

25.

26.

Mouls, L. and H. Fulcrand, Identification of new oxidation markers of grape-condensed

tannins by UPLC-MS analysis after chemical depolymerization. Tetrahedron, 2015. 71(20): p.

3012-3019.

Stanoeva, J.P. and M. Stefova, Evaluation of the ion trap MS performance for quantification

of flavonoids and comparison to UV detection. Journal of Mass Spectrometry, 2012. 47(11):

p. 1395-1406.

27.

28.

29.

30.

Lambert, M., et al., A High-Throughput UHPLC-QqQ-MS Method for Polyphenol Profiling in

Rose Wines. Molecules, 2015. 20(5): p. 7890-7914.

March, R. and J. Brodbelt, Analysis of flavonoids: Tandem mass spectrometry, computational

methods, and NMR. Journal of Mass Spectrometry, 2008. 43(12): p. 1581-1617.

Fulcrand, H., et al., Direct mass spectrometry approaches to characterize polyphenol

composition of complex samples. Phytochemistry, 2008. 69(18): p. 3131-3138.

Fayeulle, N., et al., Fast Discrimination of Chocolate Quality Based on Average-Mass-Spectra

Fingerprints of Cocoa Polyphenols. Journal of Agricultural and Food Chemistry, 2019. 67(9):

p. 2723-2731.

31.

32.

Vallverdu-Queralt, A., et al., Targeted filtering reduces the complexity of UHPLC-Orbitrap-

HRMS data to decipher polyphenol polymerization. Food Chemistry, 2017. 227: p. 255-263.

de Villiers, A., P. Venter, and H. Pasch, Recent advances and trends in the liquid-

chromatography-mass spectrometry analysis of flavonoids. Journal of Chromatography A,

2016. 1430: p. 16-78.

33.

34.

Wojdylo, A., J. Oszmianski, and P. Bielicki, Polyphenolic Composition, Antioxidant Activity,

and Polyphenol Oxidase (PPO) Activity of Quince (Cydonia oblonga Miller) Varieties. Journal

of Agricultural and Food Chemistry, 2013. 61(11): p. 2762-2772.

Engstrom, M.T., et al., Rapid Qualitative and Quantitative Analyses of Proanthocyanidin

Oligomers and Polymers by UPLC-MS/MS. Journal of Agricultural and Food Chemistry, 2014.

62(15): p. 3390-3399.

35.

36.

37.

38.

39.

Laaksonen, O.A., et al., Proanthocyanidins and Their Contribution to Sensory Attributes of

Black Currant Juices. Journal of Agricultural and Food Chemistry, 2015. 63(22): p. 5373-5380.

Pinasseau, L., et al., A Fast and Robust UHPLC-MRM-MS Method to Characterize and

Quantify Grape Skin Tannins after Chemical Depolymerization. Molecules, 2016. 21(10).

Prieur, C., et al., OLIGOMERIC AND POLYMERIC PROCYANIDINS FROM GRAPE SEEDS.

Phytochemistry, 1994. 36(3): p. 781-784.

Souquet, J.M., et al., Polymeric proanthocyanidins from grape skins. Phytochemistry, 1996.

43(2): p. 509-512.

de Freitas, V.A.P., et al., Characterisation of oligomeric and polymeric procyanidins from

grape seeds by liquid secondary ion mass spectrometry. Phytochemistry, 1998. 49(5): p.

1435-1441.

40.

Santos-Buelga, C. and A. Scalbert, Proanthocyanidins and tannin-like compounds - nature,

occurrence, dietary intake and effects on nutrition and health. Journal of the Science of Food

and Agriculture, 2000. 80(7): p. 1094-1117.

41.

42.

43.

Derdelinckx, G. and J. Jerumanis, SEPARATION OF MALT AND HOP PROANTHOCYANIDINS ON

FRACTOGEL TSK HW-40 (S). Journal of Chromatography, 1984. 285(1): p. 231-234.

Ricardo-Da-Silva, J.M., et al., PROCYANIDIN DIMERS AND TRIMERS FROM GRAPE SEEDS.

Phytochemistry, 1991. 30(4): p. 1259-1264.

Cheynier, V., et al., The structures of tannins in grapes and wines and their interactions with

proteins, in Wine: Nutritional and Therapeutic Benefits, T.R. Watkins, Editor 1997, Amer

Chemical Soc: Washington. p. 81-93.

44.

Guyot, S., N. Marnet, and J.F. Drilleau, Thiolysis-HPLC characterization of apple procyanidins

covering a large range of polymerization states. Journal of Agricultural and Food Chemistry,

2001. 49(1): p. 14-20.

45.

46.

Vidal, S., et al., Changes in proanthocyanidin chain length in winelike model solutions. Journal

of Agricultural and Food Chemistry, 2002. 50(8): p. 2261-2266.

Vivas, N., et al., Differentiation of proanthocyanidin tannins from seeds, skins and stems of

grapes (Vitis vinifera) and heartwood of quebracho (Schinopsis balansae) by matrix-assisted

laser desorption/ionization time-of-flight mass spectrometry and thioacidolysis/liquid

chromatography/electrospray ionization mass spectrometry. Analytica Chimica Acta, 2004.

513(1): p. 247-256.

47.

Kassara, S. and J.A. Kennedy, Relationship between Red Wine Grade and Phenolics. 2. Tannin

Composition and Size. Journal of Agricultural and Food Chemistry, 2011. 59(15): p. 8409-

8412.

48.

49.

Steynberg, P.J., et al., Oligomeric flavanoids. Part 27. Interflavanyl bond formation in

procyanidins under neutral conditions. Tetrahedron, 1998. 54(28): p. 8153-8158.

Tarascou, I., et al., The hidden face of food phenolic composition. Archives of Biochemistry

and Biophysics, 2010. 501(1): p. 16-22.

Figure 1: Chemical depolymerization of condensed tannins.

Figure 2: a) UV chromatogram at 280 nm of the polyphenol extract from a red wine sample

generated before and after chemical depolymerization; b) EIC from mass spectrometry

analysis of each constitutive tannin monomer.

Figure 3: Hemisynthesis of the ECG-ECG dimer.

Figure 4: UV chromatogram at 280 nm and extracted-ion chromatogram of ECG-ECG after

chemical depolymerization.

Figure 5: Comparison of UV chromatograms at 280 nm and EIC of two polyphenolic

fractions of red wine: (a) the fraction eluted in the first part of the fractionation (b) the

fraction eluted in the last part of the fractionation.

Quantity of non-oxidized tannin subunits (mg/mL)

1a

1b

1c

2a

2b

2c

3a

3b

3c

1a

1b

1c

2a

2b

2c

3a

3b

3c

1a

1b

1c

2a

2b

2c

37

39

40

41

6

11

12

11

12

15

16

13

15

12

15

16

0

1

2

3

4

5

6

7

8

0

61

60

61

62

52

50

52

56

45

56

55

59

31

30

31

30

3

1

2

1

2

1

0

1598

1611

1603

1607

1616

1600

1618

1629

1639

23

1

3

2

48

51

1663 1711

1676 1726

1668 1718

1671 1722

1680 1731

1664 1714

1684 1735

1695 1746

1705 1758

50

51

U

V

51

50

51

53

22

79

77

75

76

82

69

81

85

101

102

98

8

23

25

7

19

22

6

19

25

101

108

92

EI

C

8

21

26

6

20

23

7

20

28

109

111

117

7

22

30

7

22

32

100

101

100

101

98

6702

6685

6653

6655

6652

6654

100

101

100

101

98

6735 6835

6717 6816

6685 6786

6688 6788

6684 6785

6686 6785

0

U

V

0

3a

3b

3c

1a

1b

1c

2a

2b

2c

3a

3b

3c

106

103

0

1

31

6

0

1

6718

6696

6661

10

2

106

103

30

6750 6855

6729 6832

6694 6797

0

2

0

2

13

20

30

0

6

10

30

0

6

10

30

0

6

10

30

EI

C

0

6

10

30

0

6

10

30

0

6

10

30

0

6

10

30

0

6

10

30

Table 1: quantity of non-oxidized subunits estimated from UV chromatograms or EIC.

Article Doi

The use of extracted-ion chromatograms to quantify the composition of condensed tannin subunits

DOI: 10.1002/rcm.8619

Source and publish data:

Authors:

Article abstract of DOI:10.1002/rcm.8619

Full text of DOI:10.1002/rcm.8619

Products guided by the article

R&D Labs maybe for 134054-57-2

Relevant to this article

Hot Product