1-methylimidazole (99%), bromoethane (98%), potassium acetate (99%), 1.8-diazabicyclo[5.4.0]undec-7-ene (99%), bromobutane (98%), allyl chloride (98%) were purchased from Alpha Aesar and allyl bromide (99%) stabilized and chlorobutane (99.5%) were purchased from Fischer Scientific. Ethyl acetate (99%) was purchased from Carlo Erba, dichloromethane (99.9%) from SDS, absolute ethanol from VWR and DMSO d₆ with 0.03% tetramethylsylane (Me₄Si, 99.8%D) from Eurisotop.

¹H and ¹³C NMR spectra were recorded at 400.13 MHz and 100.62 MHz with a Bruker DPX-400 spectrometer using DMSO(d₆) as solvent at room temperature. The chemical shifts (δ) are expressed in ppm with Me₄Si as the internal standard (δ₀).J values are given in Hz.

Wood samples (1 - 2 cm thick disks) were collected at 1.3 m from the trunk of a 20-year-old Douglas-fir tree (Pseudotsugamenziesii) in October at Gimel, France (45˚18'N, 1˚50'E). Once the bark was removed, the softwood part of the disks was grounded into a fine powder (<250 μm) with a Dangoumeau steel ball mill. Wood powder samples were then homogenized and stored in a dry atmosphere until analysis.

Douglas-fir softwood was dissolved at 10 mg per 5.23 mmol in each IL previously synthetized ([Emim]Ac, [Emim]Br, [Amim]Br, [Amim]Cl, [DBUC₄]Br and [DBUC₄]Cl). Wood dissolution has been performed at 80˚C or 100˚C until complete solubilization of wood particles. Dissolution advancement was followed by optical microscopy.

Before The polysaccharide identification is made thanks to an ELISA test with a set of 14 monoclonal antibodies (mAbs) (Table 1).

Commercial polysaccharides, used as controls for ELISA tests, were dissolved in ionic liquid in the 6 pure ILs (10 mg∙mL⁻¹) thanks to a heating treatment at 80˚C at 3 h. The following polysaccharides was used: guar from Cyamopsis

tetragonoloba (Sigma-Aldrich), ß-(1-4)-mannopentaoses (Megazyme), xylans from birchwood (Sigma-Aldrich), xyloglucans from tamarin seeds (Megazyme), pectin from citrus (Sigma-Aldrich), arabic gum (Sigma-Aldrich), Karaya gum (Sigma-Aldrich), arabinans from sugar beet (Megazyme).

All dilutions of commercial or Douglas-fir polysaccharides were done in deionized water to decrease the viscosity of the IL-polysaccharide solution and facilitate an accuracy sampling. Diluted polysaccharides-IL or wood-IL solutions were applied to 96-well plates (Nunc, ref 269620 ThermoFischer) over a broad concentration range (0.5 ng to 5 µg for polysaccharides or 0.05 to 100 µg for wood solutions). Control wells were coated with deionized water with IL at the same concentration as in other polysaccharides-IL solutions wells. The liquid was then evaporated overnight at 37˚C. After a washing step with 300 µL Tris-buffer saline (TBS) (Tris-HCl 50 mM, NaCl 150 mM, pH 7.6), the plates were blocked with 300 µL of 1% (w/v) nonfat dry milk in TBS for 1 h. After having removed the blocking solution by manual turning, 50 µL of primary mAb (1:100 in TBS) were added to each well and incubated for 4 h at room temperature. The supernatant was removed and wells were washed three times with 300 µL of 0.1% (w/v) nonfat milk in TBS (wash buffer). Then, 50 µL of diluted (1:5000 in wash buffer) peroxidase-conjugated goat anti-mouse IgG or goat anti-rat IgG antibodies (Sigma-Aldrich), depending on the primary antibody used, were added to each well and incubated for 1 h. Wells were washed five times with 300 µL of wash buffers. 50 µL of 1-Step Ultra TMB (3.3’,5,5’-Tetramethylbenzidine) ELISA Substrate Solution (Life Technologies) were added to each well. The reaction was stopped by adding 50 µL of 0.5 N sulfuric acid. The OD of each was read well as the difference in A₄₅₀ and A₆₅₅ using a microplate reader (BMG-Labtech, FLUOstar Omega).

Six ILs, [Emim]Ac, [Emim]Br, [Amim]Br, [Amim]Cl, [DBUC₄]Br and [DBUC₄]Cl were synthesized to compare their efficiency for dissolving Douglas-fir sapwood as described in material and methods. These ILs have been chosen because of their good ability to solubilize either cellulose, lignin or even wood. Firstly, [Emim]Br was synthesized as a precursor for the synthesis of [Emim]Ac. [Emim]Ac was synthesized from [Emim]Br through the ion exchange reaction according to [23] adapted by [24] plus an increase of the reaction time (40 h instead of 5 h).

Two other imidazolium-based ILs were synthesized, [Allmim]Cl and [Allmim]Br. The first one was synthesized according to [25] and the second one, [Allmim]Br, result of the reaction of allyl bromide and 1-methylimidazole. These two ILs were obtained with a 100% purity, and in 97% and 96% yields for [Allmim]Cl and [Allmim]Br respectively.

DBU-based ILs, [DBUC₄]Cl and [DBUC₄]Br were synthesized from 1 equivalent of DBU and 1.2 equivalents of chlorobutane or bromobutane. They were obtained pure, in 96% yield.

Douglas-fir softwood dissolution has been performed at two heating temperatures: 80˚C and 100˚C (Figure 1).

The wood particles dissolution was followed by optical microscopy as shown in Figure 1 for to [Emim]Br. Identical results were obtained with each of the 6 ionic liquids (Table 2).

Our dissolution results show that the time to obtain a complete dissolution varies greatly depending on the IL and the treatment chosen. Usually for the 5 ILs, the time to dissolve the wood particles decreases with the increase of temperature from 80˚C to 100˚C. At a temperature dissolution of 80˚C, the ILs the most efficient are [Emim]Ac and [Allmim]Br for which 54 h is enough to completely dissolve wood. At this temperature, the least efficient are [Emim]Br and [DBUC₄]Br. At 100˚C, [Allmim]Br solubilizes Douglas-fir wood in only 14 h, whereas this time is 2-fold higher for [Emim]Ac and at least 6-fold higher for the other ILs.

The ELISA tests were carried out on a panel of pure commercial polymers previously dissolved in the 6 ionic liquids as described in material and methods. Labeling was observed for each of the commercial polymers, independently of

the ionic liquid used. The 6 ILs tested do not interfere with the polymer recognition by antibodies and the labeling is proportional to the quantity of polymer deposited on the ELISA plate. For commercial pure polymers, the regression lines have regression coefficients greater than 0.95. The labeling is thus proportional to the polymer amount deposited, whatever the IL. However, variations in the sensitivity of the method (minimum amount of polymer detected) have been identified. Moreover, the pectin solubilized in [Emim]Ac induced some chemical modifications, such as caramelization of the solution, suggesting strong pectin polysaccharide degradation or chemical reactions between the polymers and the IL in our experimental conditions.

This technique has been then applied to Douglas-fir wood samples. The results are presented for the same antibody in the form of relative percentage of labeling (DO) for each of the ILs (Figure 2). For the same antibody, labeling intensity differs depending on the IL used for dissolving the wood. In the case of [DBUC₄]Br with a dissolution temperature of 80˚C (Figure 2), low labeling is observed with all antibodies. Some pectin and XG labeling disappear with the increase of temperature to 100˚C. At 80˚C, the labeling is generally greater with [DBUC₄]Cl, IL for which the pectin labeling (HGs and RGI) are among the strongest of the 6 ILs studied. Moreover, the [DBUC₄]Cl and the [DBUC₄]Br solution show a high viscosity level after wood dissolution. Wood dissolved in the

[Amim]Cl has a low labeling intensity for pectin (HGs and RGI) and xylans. Labeling intensities are also weak or nonexistent in the case of [Emim]Ac whatever the temperature. Comparing the results for the [Emim]Br and [Amim]Br at 80˚C, a more important labeling is observed for pectin, AGP and galactan epitopes for the wood dissolved in [Amim]Br, while the labeling is higher in the case of [Emim]Br for arabinan, XG and mannan epitopes. At a dissolution temperature of 100˚C, some pectin labeling disappears in the case of [Emim]Br. Only wood dissolved at 100˚C in [Amim]Br has a label with CCRC-M14, LM20 and JIM7 antibodies. However, in this LI, the labeling is lower than in [Emim]Br for xylan, mannan, arabinan and XG epitopes. These two ILs have interesting characteristics for the solubilization and detection of wood polymers. A dissolution temperature of 80˚C is preferred because of a lack of labeling for some polymers after dissolution at 100˚C. This fact involves a possible degradation of these polymers or modifications of the epitopes recognized by the antibodies after a dissolution at 100˚C. This degradation/modification is also possible at 80˚C.

However, to the extent that all the wood samples will be treated in the same way (same temperature and dissolution time), this degradation/modification should be in the same proportion for all samples enabling the comparison of their biochemical composition using IL/ELISA method.

Six days at 80˚C in [Emim]Br allow a total dissolution of the biomass previously reduced to powder in the Dangoumeau steel ball mill. In contrast, only 3 days should be necessary after liquid nitrogen grinding. New tests were therefore conducted with the [Emim]Br from earlywood and latewood powder by combining the treatment at 80˚C and 100˚C in order to see if there was an improvement in the sensitivity of the method (only some of the 14 antibodies were chosen, one per polymer family). The following treatments were therefore compared: 3 days of dissolution at 80˚C, 4 days of dissolution at 80˚C, 3 days of dissolution at 80˚C followed by 2 h at 100˚C and 4 days of dissolution at 80˚C followed by 2 h at 100˚C (Figure 3(a) and Figure 3(b)).

Wood dissolution and ELISA analyses were done in triplicate. The standard deviation of repeated measurements is rather low (Figure 3(a)), even if some standard deviations are higher for some samples (for example LM11, 3 days at 80˚C) or for one antibody (particularly LM25). Whatever, the method stays reproducible for all the antibodies. Regarding the results for the sapwood earlywood (Figure 3), labeling is comparable regardless of the dissolution treatment for CCRC-M14, LM25, LM19, LM11 and LM21 antibodies. The LM5 labeling is more intense for the 3-day treatment at 80˚C in the case of the sapwood earlywood (Figure 3(a)) and the heartwood latewood (Figure 3(b)) than the other treatments. For the sapwood earlywood, a labeling difference is also observed for LM21 and LM11 for which labeling are more intense after 4 days at 80˚C followed

or not by 2 hours at 100˚C. However, the labeling differences between treatments are low, which shows that these differences in dissolution time and temperature only slightly impact the dissolution. process or polymer degradation in this IL.

The composition of Douglas-fir was analyzed in triplicate after separating earlywood from latewood of each of the rings (Figure 4). Concerning the HGs (CCRC-M38, LM19 and JIM7 antibodies, Figure 4), the labeling trends remain the same. This test also made it possible to see differences in terms of the proportion

of HGs according to the type of wood. Indeed, higher proportions are observed in earlywood compared to latewood. No labeling could be visualized for the LM20 antibody. The labeling trends observed for the RG I (Figure 4) are almost identical for the three repeats in the case of the CCRC-M14 antibody. Variations appear for one of the repeats in the case of CCRC-M35, but the trends remain similar. Higher proportions of RG I are noted in early wood. Regarding xyloglucans, the technique is reproducible for the LM25 antibody.

On the other hand, some labeling variations are observed between the repetitions in the case of fucosylated xyloglucans (CCRC-M1) and the technique is not reproducible with the LM15 antibody (Figure 4). This can be explained by a sensitivity limit of the method due to low proportions of xyloglucan epitopes recognized by CCRC-M1 and LM15 in the wood (the intensities of DO do not exceed 0.3 for a deposited quantity of 20 μg of dissolved wood) or by a lack of homogeneity of the solubilization in the IL of the XGs with XLFG, XFFG, XXFG and XXXG pattern or even by a lack of affinity of these polymers for the polycarbonate of the ELISA plates. The results obtained with the LM25 antibody show that the proportion of xyloglucans is greater in early wood than in latewood. In the case of AGPs (LM2, Figure 4), variations are observed between the labeling of the three repetitions, the technique is therefore not reproducible for these polymers. The same reasons as those previously invoked for the XGs can explain this lack of reproducibility. Even if labeling variations are observed for some samples for galactans and arabinans, overall, the labeling follows similar trends (LM5 and LM6, Figure 4). Regarding galactans, higher proportions are observed in the latewood of the sapwood, whereas for the heartwood, this trend is reversed. In the case of arabinans, the marking is much more intense in the early wood of the sapwood than in its latewood, on the other hand the trends are less clear-cut in the heartwood. The labeling intensities for the xylans (LM10 and LM11, Figure 4) show variations between the repetitions for the same ring. These variations are all the more marked in the case of LM11. A lack of homogeneity in the dissolution of the wood by the LI, in particular in the solubilization of this polymer, can therefore occur. Be that as it may, labeling trends remain visible with the LM10 antibody. Moreover, these results would tend to show that the xylans are in greater proportion in the early wood, in particular at the level of the sapwood (the trends are less clear for the rings most in the center of the heartwood). In the case of mannans (LM21 and LM22 antibodies, Figure 4), the first repetition gives different results for many tree rings compared to the data collected for repetitions 2 and 3, in particular for the LM22 antibody. On the other hand, the trends in terms of labeling intensity observed for repetitions 2 and 3 are similar, despite some differences in values. The solubilization of mannans in [Emim]Br therefore does not always appear homogeneous. The labeling intensities with LM21 show that the mannans are in greater proportion in the latewood than in the early wood.

200 monoclonal antibodies were able to detect a large amount of polysaccharide [40]. In this study, we focused on 14 antibodies able to react against hemicelluloses, pectin, and parietal proteins (AGPs) epitopes.

Solubilized polysaccharides have better accessibility of epitopes to antibodies [13]. However, the solvents used to solubilize or purify the cell wall polymers may lead to damage or changes in the chemical structure of epitopes recognized by antibodies [40]. We observed, after the dissolution of Douglas fir wood in various ILs at different temperatures, different labeling for the same antibody according to the IL.

Before analyzing by ELISA the Douglas fir wood dissolved in the various ILs, we realized commercial polymer ranges in order to observe the proportionality of the labeling depending on the amount of polymer deposited. We found that the 6 ILs used ([Emim]Br, [Emim]Ac, [Amim]Br, [Amim]Cl, [DBUC₄]Br and [DBUC₄]Cl) do not interfere with the polymer recognition by the antibodies and the labeling is proportional to the amount of polymer deposited on ELISA plates, despite some differences in the detection threshold based on ILs.

Wood samples completely dissolved in each of the ILs were then analyzed with the set of the 14 antibodies. In the presence of [DBUC₄]Br, the labelling are weak whereas [DBUC₄]Cl are the best, the [DBUC₄]Cl and the [DBUC₄]Br were not selected because their too high viscosity making the pipetting complicated. Low labeling was observed for pectin after wood dissolution in [Amim]Cl and [Emim]Ac at 80˚C and 100˚C. Wood solution dissolved in [Emim]Br and [Amim]Br at 80˚C present interesting labeling levels for all polymers. For higher temperature pectin labeling disappears in [Emim]Br. For xylans and mannans higher temperature is higher labelling is observed. This shows that, in the solutions obtained after wood dissolution at 80˚C, some mannan and xylan epitopes remain inaccessible to antibodies. Thus, a higher heating temperature allows a better dissociation of these hemicelluloses from other cell wall components. In softwood, (galacto)glucomannans exhibited similar orientation than cellulose microfibrils increasing linkeage [41] [42]. Moreover, (Arabino)glucuronoxylan and (glucurono)arabinoxylans are linked to lignin via ester bonds at the uronic acids of their side chains [43] [44]. Hence, higher wood dissolution temperatures in ILs would better separate mannans and xylans from cellulose and lignin.

In some case, high temperature treatment can lead to damage of certain polymers such as pectin or chemical changes of the epitopes recognized by antibodies. Microwave treatment can also lead to the same result. Walnut wood dissolves in [Emim]Ac, exhibited a decreased labeling for homogalacturonan (HGs) after a microwave treatment before conventional heating [2]. Sosha et al. [45] demonstrate that high temperatures (160˚C) degrease immunolabelling of HG (CCRC-M38, JIM5, JIM7 among others), rhamnogalacturonan I (RGI) and some arabinogalactan epitopes. However, after a pretreatment with the [Emim]Ac, galactomannan epitopes could be detected as xylan and xyloglucan.

Pine and oak cellulose solubilized [Emim]Ac, found minor degradation after 16 h at 110˚C [12]. The acetate of [Emim]Ac can also can modify polymers and hide epitopes. This has been shown for cellulose: after 20 min at 150˚C, the cellulose has an acetyl degree of substitution of 0.017 [46]. Yang et al., [47] shown partial degradation of hemicelluloses and lignin during bamboo solubilization in [Amim]Cl. The Douglas-fir wood solution dissolved in [Amim]Cl at 80˚C has one of the lowest xylan labeling. Pre-treatments with AFEX (ammonia fiber expansion) before sequential extraction induced increase of substituted or unsubstituted xylan epitopes are important, whereas the pectin epitopes (HGs and RGI) disappear [48].

Six days at 80˚C in [Emim]Br were necessary to completely dissolve the wood powder obtained after grinding with the Dangoumeau. In contrast, only 3 days in this IL were enough to completely dissolve sapwood or heartwood ground in liquid nitrogen in the ball-mill Retsch, due to a finer powder. By comparing immunostaining obtained after a dissolution of earlywood and latewood powder for 3 or 4 days at 80˚C, followed or not by 2 h at 100˚C, we found few differences with a limited number of antibodies (6). Increasing the processing time causes a decrease in the galactan epitope labeling and an increase for xylan and mannan epitope labeling. Labeling differences between treatments are, however, low, indicating that these differences in dissolution time and temperature impact very little solubilization and/or polymer degradation. Increasing dissolution time allows a better solubilization for mannans and xylans. Adding a step of 2 hours at 100˚C after a treatment at 80˚C has no impact on the immunolabeling result.

Different imidazolium or DBU based IL was synthesized and their efficiency to dissolve Douglas fir wood powder by heating at 80˚C and 100˚C was compared. [Amim]Cl and [Emim]Ac are the most efficient. However, these ILs lead to degradation or chemical modifications of wood polymers. The dissolution process at 80˚C in [Amim]Br, [Emim]Br and [DBUC₄]Cl needs more time but polymers seem to be less modified. We showed, moreover, with [Emim]Br that the IL/ELISA method is reproductive to analyze plant cell wall composition after IL dissolution. Hence, this method was used to analyze Douglas fir wood composition from the outer sapwood to the inner heartwood, ring by ring, from wood cores, after a treatment of 4 days at 80˚C in [Emim]Br.