Multivariate spectroscopy for targeting phenolic choreography in wine with A-TEEM TM and NMR crosscheck non-targeted metabolomics

. Present work comprises the use of different multivariate spectroscopic methods for tracking novel metabolomics signatures related to red wine chemistry. It is presented for the ﬁrst time the proton nuclear magnetic resonance metabolomics ﬁngerprint of a monovarietal Mexican Merlot, obtained with acquisition improvements recently proposed to the OIV Methods of Analysis sub-commission. Effective multi-presaturation solvent schemes have revealed a rich (poly)-phenolics aromatic region, so far not exploited for wine-ﬁngerprinting or – targeted proﬁling routines. It is presented as well for the ﬁrst time the use of simultaneous absorbance-transmission and ﬂuorescence excitation-emission matrix “push-one-bottom” method (A-TEEM TM ) at speciﬁc chemical conditions for a rapid, effective and high-sensitivity characterization of phenolic choreography in wines, as novel observables to quantify oenological practices and aging.


Introduction
Economic and cultural importance of wine is chemically supported mostly by its phenolics content, as said primary and special metabolites' family contribute to wines' organoleptic properties such as colour, taint, mouth-feel and aromas [1]. Standard quantification of phenolics comprises colorimetric and/or chromatographic approaches, whereas despite their robustness, they present certain complexity in terms of sample preparation, chemical manipulations, being in turn time consuming, laborious, costly and require some level of analytical expertise. OIV cross-commissions shall promptly evaluate the ratio between wineries and oenological research institutes in all member states that could provide the manoeuvre of analytical experts certified to carry out said standard methods.
This work presents a "push-one-bottom", rapid, userfriendly and non-invasive spectroscopic solution to track a robust phenolic profile in wines, with simultaneous absorbance-transmittance (A-T) and fluorescence excitation -emission matrix spectroscopy (EEM), branded as A-TEEM TM [2], wherein the simultaneous AT & EEM acquisition is carried out at each excitation increment. Immediate applications of the A-TEEM TM device comprise the characterization of human consumption water quality, in terms of quantification of Total dissolved Organic Carbon (TOC) metabolites like low-and highmolecular aromatics, as well as protein compounds from humic/fluvic sources [3]. More recently, A-TEEM TM technology has been used to determine phenolic and anthocyanin profiles in fresh and oxidized Italian red wines, in the excitation-emission range between 250-800 nm [3,4]. Construction of reliable meta-data bases in terms of reproducible A-TEEM TM phenolicanthocyanin libraries for quantitative analysis of mostly special metabolites in wine with a fast/high sensitivity/pushone-bottom solution needs an orthogonal crosscheck with robust OIV methods. For instance, high-resolution proton Nuclear Magnetic Resonance Spectroscopy ( 1 H-NMR) [5][6][7] is used to obtain the first reported metabolomics fingerprint and profiling of a monovarietal Mexican Merlot (2018, Sala Vivé, Freixenet, Querétaro MX), targeting the most abundant primary metabolites, in agreement to an OIV resolution project under course [7]. 1 H-NMR metadata analysis is presented, for fingerprinting some spectral regions associated to phenolics that orthogonally correlates with A-TEEM TM observables. Sample preparation for A-TEEM TM spectroscopy was carried out by dissolving 30 µL of wine samples (3), (6) and (11) with a 200× dilution factor, using a 12% ethanol v/v solution as solvent at three different pH levels: 1, 3, 7. A final volume of 3 mL per sampling was versed in each case within a (1 × 1) cm path length Cuvette Quartz. A-TEEM TM lectures were done with a temperature of 25 • C.

Wine samples
Sample preparation for NMR studies comprised the addition of 100 µL of a mixture of D 2 O and chemicalshift reference sodium 3-(trimethylsilyl)-propionate-2, 2, 3, 3-d 4 (TSP), phosphonate buffer KH 2 PO 4 0.1% and 2% NaN 3 to 900 uL of wine sample, whereas pH was finally adjusted to a value of 3.9 for all samples. Samples were finally versed in standard 5 mm NMR tubes.

UV-VIS absorbance-transmittance coupled with excitation emission matrix fluorescence
Simultaneous UV-VIS absorbance-transmission and fluorescence excitation-emission matrix spectra were carried out on an A-TEEM TM Aqualog system (Horiba Jobin Yvon, Inc.) with simultaneous absorption -excitation wavelength spans from 240-800 nm (5 nm interval) and emission wavelengths spanning from 248-826 nm with an average increment of 4.66 nm. Analysis of the fully corrected A-TEEM TM data was carried out with the Aqualog DataStream package based on the multivariate routine known as parallel factor analysis (PARAFAC, Solo + MIA package from Eigenvector Research Inc.) [10]. Best fit of the data was achieved with a five component model.

Nuclear Magnetic Resonance (NMR) spectroscopy
All spectra were recorded on a Bruker 600 AVANCE III HD equipped with a 5 mm 1 H/ D TXI probehead with z-gradient. 1D-1 H experiments with water-to-ethanol solvent presaturation were carried out as elsewhere reported [5]. Figure 1 presents the raw and 1 st derivative absorbance and % transmittance spectra of a set of Mexican Merlot 2018 red wines with no barrel aging (Tank, red) and three months aged, even within a 2018-Tonnellerie d'Aquitaine French barrel (blue) and with a 2018-Demptos American barrel (green), at three different acidic conditions (pH = 1, 3, 7). Peaks on raw spectra, attenuated in the derivative graphs, accentuate the following bands: a major extinction peak at 275 nm, a second minor peak at 520 nm and a third residual peak at around 715 nm, well observed at pH = 1, in a lesser extent at pH = 3, but only for aged samples in barrels. No residual 715 nm peak is observed, neither for Tank samples at any pH value and at a pH = 7 for the rest of the aged Merlot wines. Whereas the 275 nm absorbance -transmittance lines are commonly associated with simple phenolic compounds and the 520 nm peak region has been associated to stable anthocyanin compounds [12], the 715 nm absorbance peak could strongly be associated to flavylium cations of most common anthocyanidines in red wines that serve as dyes [13]. Deep inspection of first derivative absorbance and % transmittance spectra of Figs. 1 and 2, reveals a noticeable increase of the 520 nm band for not aged or poorly aged species, whilst absorption band at 715 nm is present in samples with presumably better aging processes, an effect that is better appreciated at acidic conditions. The last correlates with previous EEM studies [12] that claim the possibility to distinguish monomeric and polymeric anthocyanin species. Present results opens the venue to distinguish monomeric (at 520 nm) and polymeric (715 nm) anthocyanin species with first derivative absorption spectra, whereas at acidic conditions, said spectroscopic signature is more pronounced.

Results and discussion
Proton nuclear magnetic resonance ( 1 H-NMR) metabolomics fingerprint of Mexican monovarietal Merlot red wines, with the use of key methodological improvements [5,7] that noticeably increases spectral signal-to-noise ratio, is presented in Fig. 3.
Improvements of water-to-ethanol multi-presaturation schemes for having a full set of proton resonances of both primary and specialized wine metabolites towards fast acquisition NMR fingerprinting & targeting, has been recently presented at OIV SCMA experts' group [7]. Advantages of the use of said methodological 1 H-NMR aspects, is exemplified in Fig. 3 and should be read as follows: 1 H-NMR-OIV resolution project comprises the quantitation of solely six primary metabolites for wine analysis [7], most probably due to poor signal-tonoise ratio with the use of standard 1 H-NMR solventsupression schemes, that severely penalizes the limits of detection -quantification of lower concentration metabolites. Accurate water-to-ethanol multipresaturation schemes allowed the fingerprinting and profiling of at least 15 novel metabolites, having excellent agreements with respect prestigious plant metabolomics meta-data Repositories [14,15]. It is worth noting to highlight that with selected acquisition conditions [5], an important number of non-assigned resonances at the phenolics region (5.58-8.0 ppm) present an accurate signal-to-noise ratio for increasing known NMR fingerprinting & targeting of wines.
With the use of accurate multipresatuation schemes for water and ethanol intense signals, a rich (poly)phenolics aromatic region is exposed within the full NMR-fingerprint of studied Querétaro Merlot wines, with reasonable acquisition times per sample (i.e. 4 minutes per experiment, with 64 transients). Despite a full NMR pre-processing treatment (signal bucketing, integration and quantification) of present data will be elsewhere discussed in detail, mostly from rich   Fig. 4, was done by means of the UV-VIS Absorbance 715 nm peak intensity that each set of samples present (full UV-VIS Absorbance depicted in Fig. 2). As observed and expected, aged 2017-Merlot in a 2017-Tonnellerie d'Aquitaine barrel, present a maximum 715 nm Absorbance peak intensity. In counterpart, said sample present a minimum signal intensity of Absorbance at 520 nm wavelength (Fig. 2), strongly suggesting the presence of a major amount of poly-anthocyanin with a minor amounts of mono-anthocyanin moieties. In extreme contrast, Merlot 2018/tank samples present no Absorbance signal at 715 nm and a maximum at 520 nm, confirming the UV-VIS Absorbance antagonist mono/poly-anthocyanin profile. Last observations could be cross-checked with the use of the NMR-Merlot profiling. Despite the lack of spectral resolution at the novel revealed (poly)-phenolics region around 5.58-8.0 ppm, signal integration could shed light on phenolics content in agreement with UV-VIS observations. Finally, coupled A-T & EEMs could be regarded as genuine fingerprints of optical active (macro)-molecular composition of wines. As above mentioned, Absorbance -%Transmission profiles present maximum extinction peaks around 275, 520 and 715 nm (Figs. 1 and 2). In turn, EEMs produces broad intense florescence emission spectra spanning from 250 to 320 nm for a 240-325 nm excitation wavelength (Fig. 5, Top).
PARAFAC model decomposes 3D-signals into a fixed number of statistical components (scores) that in turn describe the variability of acquired AT-EEM. In our study, a five component PARAFAC model (C1-C5) best explained the variability of fluorescence signatures of red wines as a function of pH. The dominant component C4 present a maximum emission signal of 375 nm (emission) and 320 nm (excitation), whereas said emission/excitation profiles present accurate agreements with fluorescence of flavonoid like moieties [11]. C2 and C3 scores present emission/excitation peaks at respectively (375/260), (355/270) nm that most likely are due to Figure 4. Proton one-dimensional Nuclear Magnetic Resonance spectra ( 1 H-NMR) with improved water-to-ethanol multipresat. Scheme [5] of Mexican monovarietal Merlot wines, with different aging schemes (samples 1 to 11, refer to Materials and Methods). Novel rich polyphenolics exposed region (5.58-8.0 ppm) with multipresat scheme has been integrated in all cases, referenced with respect sample 1 (I = 1.0). First derivative ( ) UV-VIS absorbance spectra of each sample, at the 715 nm region is as well exposed, per case.

Conclusions
This work presents for the first time a NMR/ A-TEEM traceable molecular (polyphenolics) fingerprints, linked to the chemistry involved in aging processes, using a set of mono-varietal Querétaro Merlot samples as model system. First the use of raw and first derivative -pH dependent -UV-VIS ( )Absorbance spectroscopy is proposed to elucidate a simple-to-complex phenolics' profile within wine samples, in terms of ( ) Absorbance lines at 275, 520 and 715 nm. In parallel, methodological improvements allowed to obtain a proton NMR fingerprint of studied samples that in turn revealed a novel exploitable aromatics region, whereas signal integration of key regions, present excellent correlation with UV-VIS data, for cross-checking the novel method of analysis. Twodimensional Absorbance (Transmittance) -Excitation Emission Fluorescence Matrix and processing with a five component PARAFAC cluster accurately describe the variability of polyphenols in wines as a function of pH and different aging processes.