WO1996018911A1 - Method for identifying plant species and hybrids thereof - Google Patents

Abstract

A method wherein a polyphenol compound extract from a plant material of a plant species, a hybrid thereof or lower-order taxons thereof, or from a product produced from said material, is analysed by multidimensional NMR, whereafter quantitative signal analysis and statistical integration data processing are performed to give data characterising the plant species, the hybrid and the lower-order taxons. The method is useful for identifying a plant species and hybrids or lower-order taxons thereof in order to determine their origin.

Description

 METHOD FOR IDENTIFYING PLANT SPECIES AND THEIR HYBRIDS

The subject of the invention is a method of identifying plant species and their hybrids, and lower rank taxa, based on the use of a reference system.

By taxa of lower rank, we mean in particular subspecies, varieties, sub-varieties, forms and sub-forms, as well as cultivars or grape varieties and clones, without being linked to strict definitions of these. Taking Vitis as an example _________

__ • labrusca, __. riparia or _. rupestris which are vines as a plant species, it is commonly considered that Vitis corresponds to the genus, vinifera, riparia, ________________________, or rupestris corresponds to the species, that these species include several cultivars or grape varieties, among which various clones.

The expressions "plant species" and

"hybrids", as used in the description and the claims include, unless otherwise indicated, the lower rank taxa which they contain.

In the vineyard, for example, a lot of research is being carried out to develop methods for recognizing grape varieties and, among these, clones, allowing the control services to ensure that a vine stock, even a bottle of wine, corresponds well to the indications given. Recognizing them by ampelographic observation alone is difficult - and necessitates the use of laboratory methods. Separation techniques by chromatography have thus been proposed. However, they do not allow a global and reproducible analysis of the products, part of the tannins being retained on the column. In addition, these techniques do not make it possible to differentiate the clones between them.

The use of NMR or genetic engineering techniques has also been reported.

However, the work described, with regard to NMR, mainly concerns wines and not the organs of the vine.

More specifically, these are studies on products resulting from the fermentation of yeasts in white wines, such as glycerol, or fermentable products, such as sugars, which are therefore not specific to the vine.

Regarding NMR, these are one-dimensional techniques (1H and 3c) for which the spectral lines overlap, which induces errors in the integration data.

As for genetic engineering techniques, they have the disadvantage of being difficult to implement. In addition, they require highly qualified personnel and laboratory equipment meeting strict standards, which does not allow applications on an industrial scale, given the resulting costs.

The search for reliable means of characterization led the inventors to take into account a type of compound present in the vine and many other plant species, and to submit it to a succession of analyzes carried out under determined conditions. The work carried out has thus shown that, unexpectedly, polyphenolic compounds have qualitative and quantitative constants characteristic of a plant species and its hybrids and, inside a plant species and its hybrids, characteristics of their lower rank taxa as their subspecies, varieties, sub-varieties, forms and sub-forms as well as cultivars or grape varieties and even among them, clones. The invention is based on the use of these phenolic compounds to obtain information constituting an identity card of a given plant species or of a hybrid and of their taxa of lower rank.

The invention therefore aims to provide a new method of recognizing plant species and their hybrids, their taxa of lower rank based on the use of a frame of reference allowing high reliability discrimination.

The invention thus aims to provide more specifically a reference system, or database, for a plant species, or its hybrids and their taxa of lower rank, developed on the basis of characteristics of the phenolic compounds which they contain. According to another aspect, the invention aims to provide means for verifying the nature of the rank of a taxon of a plant species or of a hybrid to be checked.

It also relates to means for controlling from a product derived from a plant material, the affiliation of the latter to a lower rank taxon of a plant species or of a hybrid.

The method according to the invention for identifying a plant species, or its hybrids, or a lower rank taxon of this species or its hybrids, is characterized in that an extract essentially consisting of polyphenolic compounds, originating from a given plant material of said species or hybrid, from their taxa of lower rank, or from a product produced from these materials, is subjected to minus a multidimensional NMR analysis, followed by a quantitative analysis of the signals and the statistical processing of the integration data, leading to obtaining data characteristic of the plant species, the hybrid, or their rank taxa inferior.

The exploitation of the NMR map data obtained according to the invention, concerning the polyphenolic extracts of the materials and products studied, makes it possible to have a technique of highly reliable recognition of the taxa of lower rank to which these materials belong, or from from which these products are obtained. This reliability implies carrying out the extractions and analyzes under the same conditions. Advantageously, these data prove to be constant for a given material, which allows their use to develop a reference system or database constituting a signature of the genetic potential of the taxa of lower rank studied.

According to a preferred embodiment of the invention, the polyphenolic extracts are subjected to a heteronuclear 1 H-13c NMR analysis, providing information on the structure of the polyphenols present in the extracts.

More specifically, a bidi ensional analysis (2D) is used.

By means of proton detection, advantageously using a so-called reverse detection 1H-13C, therefore of high sensitivity 3-H, the correlations between 1H and! 3c (correlation spots) are established by means of their mutual coupling. Advantageously, use is made more particularly of HMBC (Heteronuclear Multiple Bond Connectivity) analysis, which establishes heteronuclear correlations! H-13C at long distance. We thus see the coupling between protons and carbons 13 distant from two or three bonds.

The pulse sequence used is more specifically that of an HMBC-GAS (Gradient Accelerated Spectroscopy). Pulses of magnetic field gradients are sent during the experiment, which allows a satisfactory suppression of the central peaks, ie horizontally on the middle of the spectrum containing no information, and this with a time d 'reduced experience in the conditions of the implementation of the process compared to HMBC without gradients. In addition, HMBC-GAS cards have the advantage of greater readability. Indeed, they do not contain vertical streaks ("tl noise) located at the frequency of each signal and coming from the imperfection of the system of elimination of such signals in the case of" detection in quadrature "in the experiments without gradients .

The interferograms (FID) are first multiplied by a square sinusoidal function to increase the signal / noise ratio (S / B) and undergo a Fourier transform in magnitude.

As a variant, or in addition to the HMBC-GAS analysis, the polyphenolic extract is subjected to an HMQC (Heteronuclear Multiple Quantum Correlation) analysis. This analysis makes it possible to study the hetero- nuclear between 3-H and 3c located at a single link. The interferograms (FID) are first multiplied by a square sinusoidal function to increase the S / N ratio, and undergo a Fourier Transform, but in this case in amplitude, then they are phases.

In another variant, used alone or in combination with the HMBC-GAS and / or HMQC analysis, the extract to be studied is subjected to an HMQC-HOHAHA analysis (HOmonuclear HArtman HAhn). This analysis makes it possible to see not only the coupling of a carbon 13 with the proton which it carries directly, but also with all the protons which are coupled to this latter. The main experimental parameters correspond to those of the HMQC to which a "mixing time" (spin-loc) is added.

To carry out these experiments, it is advantageous to use a spectrometer equipped with a very stable temperature regulation system, in order to avoid disturbances on the NMR maps. The correlation spots of the NMR analyzes are then integrated and the data obtained are the subject of a multidi ensional statistical analysis.

Integration corresponds to the calculation of the volume of each correlation peak by adding all the points present in the sector delimited around each correlation spot, each correlation spot being digitized in points of variable intensity.

All the integration data is subjected to a multidimensional analysis. More specifically, we use variance analysis (ANOVA), and / or principal component analysis (PCA), and / or discriminant factor analysis (AFD). ANOVA allows you to select the correlation spots (variables) that provide the most information.

The PCA represents individuals (clones) in a two-dimensional space, the individuals being initially present in an N-dimensional space (N = number of variables selected). This analysis makes it possible to study individuals in the space of the principal components, so as to be able to determine the possible existence of groups and anomalies.

AFD is used to optimally discriminate between predefined groups of clones.

Using ANOVA makes it possible to select the variables richest in information. It is recalled that this technique is based on the calculation of a factor F proportional to the ratio of the variance calculated for all groups combined over the sum of the variances calculated within each pre-defined group. In the case of the vine, these groups are the grape varieties, or the clones. The variables for which F is the highest are used for the PCA and the AFD. Reducing this number of variables before proceeding with the multidimensional analysis ensures that reliable results are obtained. Such results are obtained with a number of samples / number of variables ratio greater than 3.

Once the number of variables has been reduced, a principal component analysis, PCA, is performed on the data. This technique calculates new variables (orthogonal main components) by linear combination of the original variables, from the set of individuals (grape varieties, clones) with which quantitative characters are associated. These main components make it possible to best distinguish individuals possible, without presupposing the existence of groupings among individuals.

The analysis is performed on raw data. We obtain a representation of individuals in a two-dimensional plane where the first calculated axis has the greatest variance leading to the greatest dispersion of individuals on this axis. This analysis makes it possible to detect the existence of atypical groups or individuals. To optimally discriminate individuals made up of defined subgroups (varieties, cultivars, clones, etc.), these distinctions being neither exhaustive nor limiting, by the results of the PCA, AFD is advantageously performed. To this end, new variables are calculated, linear combinations of the original variables, by seeking a center of gravity for each group as far as possible. AFD is then used to assign additional individuals to one of the preceding groups. The foregoing steps are advantageously carried out on the total polyphenolic extracts of a plant material of the plant species or of the hybrid to be studied, or of a product produced from this plant species or this hybrid, this material and this product being called raw materials below.

These raw materials are subjected to one and, advantageously, several extractions carried out under identical conditions. Organic solvents are used, if necessary added with water. Suitable solvents include acetone, ethyl acetate, ether, alone or as a mixture.

For handling convenience, it is advantageous to use these extracts in lyophilized form. The raw materials come from plant species or their hybrids having plant organs rich in polyphenolic compounds. Mention will be made in particular of the vine, the field of the invention however extending to the recognition of other tannins than those of the vine, such as the oenological tannins of pine, oak, or gall nut.

The differentiation technique of the invention is also of great interest for the recognition of different varieties of fruit trees (for example of apple trees), or of floral varieties (in particular of roses), or of medicinal plants. The plant organs used for obtaining extracts are advantageously chosen, taking into account their richness in polyphenolic compounds and their easy processing. These are, for example, leaves, seeds, seeds or, if necessary, bark or flower petals.

Products made from the plant species or organ include fruit juice, or products from their fermentation, such as wine or beer.

The invention thus provides the means, from the polyphenolic content of a plant extract, to establish a set of data characteristic of a plant species or of its hybrids and, according to an aspect of great interest, of taxa of rank of this plant species or its hybrids.

Given the constant nature obtained for these data, under the conditions set out above, the latter are used according to the invention to constitute a bank serving as a repository.

This benchmark is therefore established, by accumulation of results, operating under conditions identical, for a plant species and its hybrids, or for taxa of lower rank, depending on the plant organ or the product produced from the plant organ, from which the polyphenolic compounds are extracted. The invention relates to a method of identifying a plant species or its hybrids and their lower rank taxa, characterized in that an extract from a plant organ, or from a product made from of this organ or plant species, its hybrids and their lower rank taxa, to multidimensional NMR analysis and to the processing of NMR data as indicated above, and the results obtained are compared with data from a benchmark corresponding to the species. The invention thus provides the means of controlling with great precision an origin and, thereby, a set of qualities of a plant material or of a product produced from this material, such as fruit juices or products derived from it. of their fermentation like wine or beer.

Other characteristics and advantages of the invention are given in the examples which follow for the purposes of illustration.

In these examples, reference is made to FIGS. 1 to 17, which represent respectively: FIGS. 1 to 3, respectively represent the maps of the HMBC-GAS, the HMQC and the HMQC-HOHAHA of seed extracts of grape of clone Merlot Noir 343, - Figure 4, the results of 1 ANOVA on

1 'HMBC of seed extracts from different grape varieties,

- Figure 5, the hierarchical classification for these grape varieties (on HMBC-GAS), FIGS. 6 and 7, the results of the ACP and of the AFD respectively on the HMBC-GAS of these grape varieties,

- Figures 8, 9 and 10, the results of ANOVA on HMBC-GAS, respectively of Merlot Noir,

Cabernet Franc and Cabernet Sauvignon,

- Figures 11, 12 and 13, the results of the PCA on the HMBC-GAS, respectively of Merlot Noir, Cabernet Franc and Cabernet Sauvignon, - Figures 14, 15 and 16, the corresponding results of the AFD, FIG. 17, the map of the HMBC-GAS gradients of vine leaves of Cabernet Sauvignon 337, and - FIG. 18, the results of the ANOVA on the HMBC-GAS of vine leaves.

Example 1: Study of grape seed extracts.

We report an operating protocol applied to grape seeds from clones of three grape varieties.

These are Cabernet Franc (abbreviated to CF), Cabernet Sauvignon (abbreviated to CS), and Merlot Noir (abbreviated to MN) varietals and clones: N ^β 312, 332, 331, 330, 187 to Blanquefort and, N ^β 312, 332, 331 de Montagne, for CF; N ^β 335, 337, 339, 341, 191 for CS and N ° 181, 343, 346, 347, 348 for MN.

This operating protocol includes: I. the extraction of polyphenols; II. NMR analysis of the extracts obtained; III. integration of NMR analysis data; and IV. a statistical analysis of these results.

Each of these steps is detailed below. I. Extraction of polyphenols

Three extracts from each clone were prepared as follows:

50 g of raisin seeds are macerated for 24 hours, protected from light, in 60 ml of acetone / water mixture (2/3).

The mixture is subjected to a leaching step for 15 minutes. We recover a first fraction.

310 ml of the same solvent mixture is added to the macerated seeds and subjected to leaching for 10 to 12 hours. We recover a second fraction.

The two fractions are combined and the acetone is evaporated under reduced pressure at 30 ° C.

The remaining solution is extracted with ethyl acetate (AcOEt), according to the following scheme:

AcOEt volume (ml) Agitation

200 10 min

100 5 min

100 5 min

50 5 min

The two layers are then centrifuged to eliminate the insoluble interface.

After evaporation of the ethyl acetate, the extract is dissolved in water and then lyophilized. The extraction yield is approximately 2g / 100g of dry seeds.

The polyphenolic extracts of the seeds mainly contain flavanols, catechin and epicatechin, corresponding to the following formulas, as well as their polymers.

II. NMR analysis

An identical amount of extract for each clone (100 mg) is dissolved in deuterated dimethyl sulphoxide (DMS0-d6) and analyzed by 2D NMR. Three experiences are recorded for each extract: A so-called inverted detection probe 1H-13c is used, which has the advantage of very high sensitivity 1H.

We obtain

- the proton spectrum by projection on the horizontal axis,

- the carbon 13 spectrum by projection on the vertical axis,

the correlations between these two nuclei (correlation spots of the 2D map) through their mutual coupling.

Experiment 1: HMBC analysis (ref 1) giving heteronuclear correlations (1H-13C) at long distances (2JC-H and 3JC-H). The numerical references "ref. 1" and following are specified in the "Bibliography" section at the end of the description.

The pulse sequence used is that of an HMBC -GAS.

The acquisition parameters are as follows.

The excitation frequency in dimension H is 500.14021 MHz and 125.7728 MHz in dimension 13c. The spectral window is 5319.15 Hz in the 1H dimension and 27670.02 Hz in the 13c dimension.

Each free precession signal (FID) is digitized on 2048 points and 512 experiences are recorded in dimension 13c. The acquisition data are first multiplied by a square sine wave to increase the signal / noise ratio (S / B) and undergo a Fourier Transform in magnitude. For each experiment, 16 FIDs are accumulated. The experience time is 4 hours.

Experiment 2: HMQC analysis (ref 2) giving direct heteronuclear correlations (1H-13c) (1JC-H).

The main parameters for data acquisition are as follows.

The excitation frequency in dimension 1H is 500.139658 MHz and 125.7672 MHz in dimension 13c. The spectral width is 3787.88 Hz in the 1H dimension and 16520.27 Hz in the 13c dimension.

Each FID is digitized on 2048 points and 256 experiences are recorded in dimension 13c against 512 experiences for the HMBC-GAS. This reduction in the number of experiments deteriorates the resolution of the correlation spots in the dimension 13c, but makes it possible to keep an experiment time identical to that of the HMBC. For each experiment, 32 FIDs are accumulated to achieve a suitable S / N ratio. The experience time is 5 hours.

The acquisition data are first multiplied by a square sinusoidal function to increase the S / N ratio and undergo a Fourier Transform in amplitude, then are phased.

Experiment 3: HMQC-H0HAHA (ref. 3) giving the heteronuclear correlations (1H-13c) relayed to the coupled protons.

The main experimental parameters correspond to those of HMQC to which is added "a mixing time" (spin-lock). The three experiences are recorded at 303

K. The spectrometer is equipped with a very stable temperature regulation system, which avoids disturbances on the NMR maps, such as streaks at the level of the correlation spots.

The NMR maps obtained at the end of each of these three experiments are given respectively in FIGS. 1 to 3.

III. Integration

The information contained in the 2D NMR maps of each clone is translated into digital data by calculating the volume of each correlation peak using a computerized system,

For this purpose, we delimit sectors around each correlation spot and the volume is calculated by adding all the points present in the sector

(each correlation spot is digitized in points of variable intensity).

The data obtained are then subjected to a multidimensional statistical analysis.

IV. Multidimensional statistical analysis of the data obtained with the HMBC-GAS

IV.A. analysis of grape varieties:

a) analysis of variance (ANOVA)

By analysis of variance, we select the correlation spots (variables) that provide the most information (ref 4 and 5). 17

For the data obtained with the HMBC-GAS, which are represented in FIG. 4, the most discriminating variables (F greater than 150) are the following: N ^β 22, 25, 27, 28, 31, 32, 34, 35, 37, 38, 42 and 44.

These variables correspond mainly to 1H correlation spots between 2 and 3 ppm on the one hand and between 6 and 7 ppm on the other.

b) hierarchical analysis of individuals

We then proceed to a hierarchical classification of individuals so that the distances between them are a function of their similarity and this without presupposing the existence of groups.

As shown in Figure 5, which represents the analysis of the clusters, the grape varieties are fairly well differentiated since only the samples marked with an asterisk are misclassified.

These samples correspond to clones 191 from CS and 343 from MN. The integration data show that they have different characteristics from the clones from the same grape varieties.

c) principal component analysis (PCA)

This analysis (ref 6, 7 and 8) allows an analysis of individuals (clones) in a two-dimensional space, the individuals being initially present in an N-dimensional space, N being the number of variables selected. We thus study the structure of individuals so as to be able to determine the possible existence of groups and anomalies. The results obtained represented in FIG. 6, for F greater than 30, show that the grape varieties are on the whole well differentiated.

The clones inside the CS grape are also well differentiated, with more or less great similarity of the clones.

It is not the same for the other grape varieties, which probably results from the choice of variables according to the grape variety criterion, which are not discriminating for the clones of the CF and MN grape varieties.

d) discriminating factor analysis (AFD) This analysis arranges the samples in a two-dimensional space so as to have the barycenters of the pre-defined groups (ref 6, 7 and 8) as far apart as possible, this by constructing discriminating axes which are linear combinations of the starting variables. The results of the AFD obtained from the HMBC-GAS data (F greater than 150) are reported in FIG. 7. We note that on the whole the grape varieties are well separated.

Study of statistical models

To validate the technique of discriminating grape varieties and clones, the robustness of the models was tested by statistical tests (Jacknife test and Cross validation).

Each individual was removed successively, with delivery to the starting population, then an AFD was carried out and the coordinates of the removed individual were calculated from the discriminating equations. We then check whether the individual is assigned to the right group. We consider that the technique is validated when stable discriminant equations are obtained regardless of the individual removed and the percentage of well classified individuals is high.

With each operation, new discriminating axes were obtained:

Xl-anVari + ai2 ^Var 2 + ... + aι _p Var _p

χ ₂ --a2i ^Var l + a22 ^Var 2 + ... + a2p ^Var ρ

Xn = a _n i ^Var l + a _n 2 ^Var 2 + ... + a _n p ^Var P

X _n : discriminating axes

n: number of individuals

anp: coefficients of the discriminating axes

Var _p : variable part selected by ANOVA

From this data, several calculations were made:

To know the importance of the model coefficients and their stability, the mean (m) and, the standard deviation (EC) of each coefficient were calculated.

The barycenter of each group was also calculated with each operation and each individual assigned to the group whose center of gravity was closest to him. We can thus calculate the percentage of individuals classified well each time and study the rate of good classification and the stability of classification for each group and for all the groups combined. The means ₍ m ₎ and standard deviations ⁽ EC ⁾ of the 2 are given below ⁽ number of coefficients of axes 1 and of individuals: 45 raw data ⁾

Mo y. Axis 1 Coef. . m Axis 2 Coef. EC Var. EC Var.

Coef. 1 24.2 3.51 14.5 28.8 3.56 12.4

Ccef. 2 20.5 2.28 11.1 2.75 1.64 59.6

Coef. 3 13.7 4.17 30.4 42 3.89 9.25

Coef. 4 20.9 2.51 12 21 2.35 11.2

Coef. 5 2.94 2.36 80.3 2.11 2.22 105

Coef. 6 32.7 2.56 7.82 30.4 2.07 6.83

* _ *. 0.9

Coef. 7 11.4 3.75 23.3 2.54 1

Coef. 8 11.4 3.7 32.5 18.2 3.95 21.6

Coef. 9 13.9 2.15 15.5 5.79 2.19 37.8

Coef. 10 23.7 2.17 9.17 5.56 2.68 48.2

Coef. 11 4.23 1.34 31.7 12.8 1.84 14.3

Coef. 12 4.36 1.44 32.9 5.84 2.08 35.7

The calculations performed for the HMBC-GAS show that for axis 1, the coefficients 1, 2, 4, 6 and 10 are the most important in the model and have good stability ₍ low standard deviations). Given the values of the coefficients, we deduce that these are the variables 22, 25, 28, 32, 37 and 38 which are the most influential for the construction of axis 1.

For axis 2, the coefficients 1, 3, 4, 6 and 7 are the most important in the model and have good stability (low standard deviations). Given the values of the coefficients, we deduce that it is the variables 22, 27, 28, 32 and 34 which are the most influential for the construction of axis 2.

The percentages of well classified individuals are 99.85 for CFs and MNs and 100 for CSs, the clones CF312mg and CF331Mg are responsible for the percentage of CFs due to their proximity to MNs.

Discrimination of grape varieties using NMR data from HMBC-GAS is therefore a reliable technique in view of the statistical results.

IV.2 analysis of the clones

For the recognition of the clones, the grape varieties were analyzed separately.

- ANOVA analysis

Figures 8, 9 and 10 show the values of F in the analysis of variance, obtained from the NMR spectra HMBC-GAS, respectively for the clones of the grape varieties MN, CF and CS.

The value of F selected and the variables chosen are indicated below for each grape variety:

Merlot Noir: F>10; selected variables: 3, 4, 6, 9, 12, 34, 37; Cabernet Franc: F>30.5; selected variables: 7, 22, 23, 25, 34, 36, 37, 38, 39, 43;

Cabernet Sauvignon: F> 55; chosen variables: 3, 6, 7, 9, 39, 45, 47.

- PCA analysis

The results of the ACP analysis on the HMBC-GAS of the polyphenolic extracts of seeds for MN, CF and CS, are reported in Figures 11, 12 and 13 respectively.

There is a tendency for Merlot Noir to group together by clone except for individuals MN 181 bt and MN 346 bt 194.

For Cabernet Franc, clones tend to form groups with more or less similarities between them.

The clones differing only by terroir (312, 331, 332) differ very clearly.

In the case of Cabernet Sauvignon, we notice a good separation of the clones.

- AFD analysis

The results of the discriminating factor analysis on the HMBC-GAS of grape seed extracts are given in Figures 14, 15, and 16 respectively for MN (F> 10), CF (F> 30.5) and CS (F> 55).

We . obtains in the 3 cases a good differentiation of the clones. The differentiation of grape varieties and clones by integrating HMBC-GAS correlations therefore constitutes a reliable technique.

V - Statistical analysis of the data obtained with the HMOC on the one hand, with the HMOC - HOHAHA. other E__X-t.

The techniques reported above are applied to the data of the HMQC and HMQC-HOHAHA spectra for grape varieties and clones.

In both cases, good discrimination is obtained.

Example 2: Study of vine leaf extracts.

Polyphenolic extracts of vine leaves are prepared from the following grape varieties and clones:

Grapes Merlot Noir MN clones: 343 and 347

Cabernet Sauvignon CS clones: 335 and 337

Cabernet Franc CF clones: 312 and 187

i. Extraction des polyphénols

i. Polyphenol extraction

50 g of dry leaves are macerated for 24 hours, protected from light, in 500 ml of acetone / water mixture (2/3).

The mixture is subjected to a slow leaching step for 1 to 2 hours.

The acetone is evaporated under reduced pressure at 30 ° C. A green precipitate composed mainly of chlorophyll and waxes is formed. To remove the waxes and chlorophyll, a centrifugation step and an ether extraction are used.

The mixture is subjected to centrifugation to separate the two phases.

The remaining solution is extracted with ethyl ether (Et2θ) according to the following diagram:

V Et2θ (ml) Agitation

100 10 min

100 5 min

100 5 min

The remaining phase is subjected to an evaporation step to remove the residual ether, then the remaining solution is extracted with ethyl acetate, operating as follows: V AcOEt (ml) Ag: itation

200 10 min

100 5 min

100 5 min

100 5 min

100 5 min

100 5 min

100 5 min

100 5 min

The ethyl acetate is then evaporated off under reduced pressure and the solute is taken up in water to lyophilize.

The polyphenolic extracts obtained are essentially formed of flavonoids, σuercetin and isoquercitrin, these compounds corresponding to the following formulas:

quercétine isoquercitrine

quercetin isoquercitrin

The extraction yield is approximately 3g / 100g of dry leaves. II NMR analysis

50 mg of extract are dissolved in 0.5 ml of

DMS0-d6. For the first extract of each clone, three

HMBC-GAS are recorded, and for the second extract, only one experiment is recorded to verify the reproducibility of the extraction technique.

The experimental parameters are as follows:. 1H excitation frequency: 500.1412769 MHz and 13c: 125.7728 MHz; . 1H spectral window: 14.489 ppm and

13c: 220 ppm.

The number of accumulations per experiment is 12. Each interferogram (FID) is digitized on 2048 points and 512 experiments are recorded in dimension 3c. The duration of the experience is 2 hours. The data are multiplied by a square sinusoidal function and undergo a Fourier Transform in magnitude mode.

FIG. 17 shows the NMR map obtained.

III Statistical analysis

a) ANOVA

The results of the ANOVA analysis of the HMBC-GAS spectral data are reported in FIG. 18.

The most discriminating variables are the following (F> 60): Nos. 113, 115, 117 and 118.

These variables correspond to the correlation spots from 1H to ζj _. = 12.5 ppm with carbons at _ _> = 99 ppm.

0 = 145 ppm, = 161 ppm and J = 163 ppm. These characteristics are those of a proton of a hydroxyl which is linked by the establishment of a hydrogen bond with a carbonyl, this type of proton is found in flavonoids which are the majority compounds of extracts of vine leaves.

b) Hierarchical classification.

By operating as indicated for the extracts of grape seeds, one obtains a differentiation of the grape varieties, in particular of the grape variety CS.

c) ACP and AFD analyzes

In these two experiments, carried out as indicated above, the clones are generally well separated.

The results show that the leaf extracts also make it possible to detect a specificity of the polyphenolic content of each grape variety, or of each clone, since differences appear between the clones and the grape varieties.

Claims

1. A method of identifying a plant species or its hybrids and lower rank taxa of this plant species or this hybrid, characterized in that an extract is essentially submitted consisting of polyphenolic compounds, originating from a plant material of the said plant species or of the said hybrid and their lower rank taxa, or of a product produced from this material, at least one multidimensional NMR analysis, followed by a quantitative analysis of the signals and of the statistical processing of integration data, leading to obtaining data characteristic of the plant species, the hybrid and lower rank taxa. 2. Method according to claim 1, characterized in that the polyphenolic extracts are subjected to a two-dimensional heteronuclear analysis lH-13c. 3. Method according to claim 2, characterized in that an HMBC analysis is carried out, more especially an HMBC-GAS. 4. Method according to claim 2 or 3, characterized in that an HMQC analysis is carried out. 5. Method according to one of claims 2 to 4, characterized in that an HMQC-HOHAHA analysis is carried out. 6. Method according to one of the preceding claims, characterized in that the acquisition parameters, excitation frequency in the dimension 1H, excitation frequency in the dimension 13c and the spectral windows in the two dimensions are determined specifically for each type of analysis and that each interferogram (FID) is multiplied by a square sinusoidal function to increase the Signal / Noise ratio and undergoes a Fourier Transform in magnitude mode in the case of HMBC-GAS and in mode amplitude, then phase, in the case of an HMQC or an HMQC-HOHAHA. 7. Method according to one of the preceding claims, characterized in that the correlation spots of the NMR analyzes are then integrated and the data obtained are the subject of a multidimensional statistical analysis. 8. Method according to one of the preceding claims, characterized in that an analysis of variance (ANOVA) is carried out, and / or a principal component analysis (PCA), and / or a discriminating factor analysis (AFD). 9. Method according to one of the preceding claims, characterized in that it is carried out on the total polyphenolic extracts of a plant material of the species to be studied, or of a product produced from this species or a plant organ. 10. Method according to claim 9, characterized in that the plant material comes from species having plant organs rich in polyphenolic compounds or in tannins, such as vine, pine, oak, gall nut, fruit trees , floral varieties, or medicinal plants. 11. Method according to claim 10, characterized in that the plant organs are leaves, seeds, seeds and, where appropriate, barks, or petals. 12. Method according to claim 9, characterized in that the product produced from plant material is fruit juice, or products from their fermentation, such as wine or beer. 13. Method according to one of the preceding claims, characterized in that it is used to establish a frame of reference by accumulation of results of analyzes carried out, operating according to the provisions defined in the preceding claims, on a material or a product of a given plant species or hybrid, or lower rank taxa of that plant species or hybrid, the materials or products being treated under identical conditions. 14. Method for controlling the origin and / or quality of a plant species, of a hybrid, or of lower rank taxa of this plant species or of this hybrid, characterized in that an extract is submitted from a plant organ, or from a product produced from this organ, or from the plant species or hybrid, or taxa of lower rank, to an NMR analysis, followed by the processing of data from NMR as defined in one of claims 1 to 12, and of the comparison of the results with the data of a frame of reference corresponding to the plant species, to the hybrid, or to the taxa of lower rank of this plant species or this hybrid, to which the element to be analyzed is supposed to belong. 15. Method according to any one of the preceding claims, characterized in that it is applied to one identification of grape varieties and vine clones.