Quercus suber L. samples were taken from a 6-year-old tree from discs taken between 1.0 and 1.3 m of stem height. The xylem, phloem and cork tissues were manually separated from each other using a chisel. Each material was milled in a knife mill (Retsch SM 2000), passing through a 6 × 6 mm sieve, and sieved in a Retsch AS 200. One sample was taken from the 40–60 mesh fraction (250–425 μm) for chemical analysis. A mixture of all the granulometric fractions was successively extracted with dichloromethane, ethanol and water for 24 h each. The extracted samples were oven-dried at 60°C, and milled in a knife mill (IKA MF10) passing through a 100-mesh (<180 μm) sieve to obtain a fine granulate for lignin isolation.

Two aliquot samples from the 40–60 mesh fraction (250–425 μm) from cork, phloem and xylem were chemically characterized, following procedures adapted from TAPPI standard methods (TAPPI, 2004): ash content (TAPPI T211 om-02), total extractives determined from successive Soxhlet extraction with dichloromethane, ethanol and water (TAPPI T204 cm-07), total lignin determined as the sum of Klason lignin (TAPPI T222 om-11) and acid-soluble lignin (UM 205 om-83). Neutral monosaccharide composition was determined in the hydrolysate from the lignin analysis. The monosaccharides were separated by High Pressure Ion Chromatography using a Dionex ICS-3000 system equipped with an electrochemical detector; the mobile phase was NaOH (2 mM solution) with a flux of 1.0 mL/min at 25°C; and the column used was Aminotrap plus Carbopac SA10. The results were reported as percent of initial material. In the case of cork, the suberin content was determined in the extractive-free material by methanolysis, as described in Pereira (2013), and the lignin was determined using the suberin-free material (according to TAPPI T222 om-11).

A sample of each tissue was impregnated with DP1500 polyethylene glycol, and transverse microscopic sections of approximately 17 μm thickness were cut with a microtome (Leica SM 2400). The cork and phloem sections were stained with triple staining astra blue/crysoidine/sudan IV, and the xylem sections with safranin. All the sections were observed in a light microscopic using Leica DM LA and photomicrographs were taken with a Nikon Microphot-FXA.

Milled lignins from xylem, phloem and cork were isolated according to a procedure adapted from Björkman (1956). The granulates were finely ball-milled using a 500 mL agate jar and agate ball bearings (20 × 20 mm) in a Retsch PM100 planetarium ball mill, at 400 rpm, during 5 h with 5 min breaks after every 5 min of milling. The ball-milled powder was extracted with dioxane-water (96:4, v/v) using 25 mL of solvent g⁻¹ of sample under agitation for 12 h. The solution was centrifuged, and the supernatant evaporated to dryness at 40°C at reduced pressure. The residue, called raw milled lignin (raw ML) was dissolved into a solution of acetic acid:water (9:1, v/v) using 20 mL of solvent g⁻¹ of raw ML. The lignin was precipitated into stirred cold water (225 mL g⁻¹ of raw ML), the precipitate was centrifuged, dried and milled in an agate mortar. This residue was dissolved in a 1,2-dichloroethane:ethanol solution (2:1, v/v) using 25 mL of solvent/g of lignin. After centrifugation to remove undissolved matter, the lignin in the supernatant was precipitated by adding the solution drop wise into diethyl ether, and the obtained residue was separated by centrifugation. The solid residue was suspended in diethyl ether overnight, centrifuged, and finally suspended in petroleum ether overnight. The final purified milled lignin was recovered by centrifugation and dried under N₂ flow. The final yields ranged from 15 to 20% based on the Klason lignin content.

The milled lignins (1.7 mg) were pyrolysed in a EGA/Py-3030D micro-furnace pyrolyzer (Frontier Laboratories Ldt.), connected to an Agilent 7820A GC system equipped with a DB-1701 fused-silica capillary column (60 m × 0.25 mm i.d. × 0.25 μm film thickness), and to a Agilent 5975 Mass detector (EI at 70 eV). The pyrolysis was performed at 500°C (1 min). The oven temperature was programmed from 45°C (4 min) to 280°C at a heating rate of 4°C min⁻¹, and held at 280°C during 10 min. The GC/MS interface was kept at 280°C and the injector at 250°C. The carrier gas was Helium with a flow of 2 mL min⁻¹. The compounds were identified using the literature (Faix et al., 1990; Ralph and Hatfield, 1991) and the Wiley and NIST libraries. Peak molar areas were calculated for each compound (by dividing the peak area by the respective molecular weights), the summed molar areas were normalized and the data expressed as percentage.

Around 100 mg of the whole cell wall (CW) material and 30 mg of the isolated milled lignins were dissolved in 1 mL and 0.75 mL of DMSO-d₆, respectively, for the NMR analysis. HSQC (heteronuclear single quantum correlation) spectra were recorded at 300 K on a Bruker AVANCE III 500 MHz spectrometer (Bruker Biospin, Fallanden, Switzerland), equipped with a cryogenically cooled 5 mm TCI gradient probe with inverse geometry (proton coils closest to the sample). The 2D ¹³C-¹H correlation spectra were obtained using an adiabatic HSQC pulse program (Bruker standard pulse sequence “hsqcetgpsisp2.2”). The spectral widths were from 10 to 0 ppm (5000 Hz) in F₂ for ¹H dimension, with an acquisition time of 100 ms (CW) or 145 ms (ML), and a recycle delay (d1) of 1 s. For the ¹³C dimension, the spectral width was from 200 to 0 ppm (25,168 Hz) in F₁, being collected 256 increments of 32 scans for a total acquisition time of 2 h 34 min (CW) and 2 h 40 min (ML). The ¹J_CH used was 145 Hz. Processing used typical matched Gaussian apodization in ¹H and a squared cosine bell in ¹³C. The central solvent peak was used as an internal reference (δ_C 39.5; δ_H 2.49 ppm). 2D NMR HSQC cross-signals were assigned after comparison with data from literature (Ralph et al., 1999, 2004b; Capanema et al., 2005; Rencoret et al., 2011; del Río et al., 2012a,b). A semiquantitative analysis of the volume integrals of the HSQC correlation peaks was performed using Bruker's Topspin 3.1 processing software. The relative abundances of side-chains involved in the different inter-unit linkages were estimated in the aliphatic oxygenated region from the C_α–H_α correlations, except for α-oxidized β−O−4 substructures (Aox) and cinnamyl alcohol end-groups (structure I, Figure 5), for which C_β–H_β and C_γ–H_γ correlations were used. In the aromatic/unsaturated region, C₂–H₂ correlations from H, G and S lignin units and from ferulates were used to estimate their relative abundances.

To evaluate the incorporation of acetylated monolignols into the lignin of the three materials, resulting in γ-acetylated lignin side-chains, a modification of the standard DFRC method was used (Ralph and Lu, 1998). Milled lignins (5 mg) were stirred for 2 h at 50°C with propionyl bromide in propionic acid (8:92, v/v). The solvents and excess bromide were removed by rotary evaporation. The products were dissolved in dioxane/propionic acid/water (5:4:1, v/v/v), and 50 mg powdered Zn were added. The mixture was maintained for 40 min at room temperature with stirring, and transferred into a separator funnel with dichloromethane and saturated ammonium chloride. The aqueous phase was adjusted to pH < 3 by adding 3% HCl, the mixture was vigorously mixed and the organic layer separated. The water phase was extracted twice with dichloromethane. The combined dichloromethane fractions were dried over anhydrous NaSO₄ and the filtrate was evaporated to dryness using a rotary evaporator. The residue was propionylated for 1 h in 1.1 mL of dichloromethane containing 0.2 mL of propionic anhydride and 0.2 mL pyridine. The propionylated (and naturally acetylated) lignin degradation compounds were collected after rotary evaporation of the solvents, and subsequently analyzed by GC/MS. The GC analyses were performed with a GCMS-QP2010 Ultra instrument (Shimadzu Co.) using a capillary column (DB-5HT, 30 m × 0.25 mm I.D., 0.10 μm film thickness). The oven was heated from 140°C (1 min) to 250°C at 3°C min⁻¹, then ramped at 10°C min⁻¹ to 300°C and held for 10 min at the final temperature. The injector was set at 250°C and the transfer line was kept at 300°C. Helium was used as the carrier gas at a rate of 1 mL min⁻¹.

The location of the three tissues (cork, phloem and xylem) in the plant stem and their anatomical structure are shown in Figure 1. While cork is a homogeneous tissue of phellem cells, the phloem presents sieve elements, parenchyma, and sclerenchyma cells with sclereids, and the xylem has rays and axial parenchyma, vessels and fibers. The chemical summative composition of the xylem, phloem and cork tissues was determined in the 40–60 mesh fraction and the results are presented in Table 1. The different tissues presented great differences in composition, with cork having suberin as the major structural component. The lignin content also differed among the different tissues, accounting for 27.1% in cork, 38.4% in phloem and 23.6% in xylem.

The composition of the lignins from the different tissues was first addressed by Py-GC/MS. The pyrograms of the ML preparations isolated from cork, phloem and xylem are presented in Figure 2. The identities and relative molar abundances of the released lignin-derived phenolic compounds are listed in Table 2. The pyrograms revealed strong differences in lignin composition among the three tissues. Pyrolysis of cork lignin released predominantly phenolic compounds derived from G-lignin units, with guaiacol (compound 1, Table 2, Figure 2), 4-methylguaiacol (2) and 4-vinylguaiacol (4) as the major compounds released, with only few amounts of S-lignin units, and presenting a strikingly low S/G ratio of 0.10. The pyrograms of the lignins isolated from phloem and xylem released more S-lignin units, particularly syringol (7) and 4-methylsyringol (10), and which were more abundant in the xylem. These differences are reflected in their higher S/G molar ratio, accounting for 0.62 in phloem lignin and 1.66 in xylem lignin.

The whole cell-walls of cork, phloem and xylem were analyzed by 2D HSQC NMR (Figure 3) and the spectra compared with those of their isolated lignins (Figure 4). The main lignin cross-signals assigned in the spectra are listed in Table 3, and the main lignin substructures found are represented in Figure 5. The spectra of the whole cell-walls (Figure 3) presented signals from carbohydrates, including xylan correlations in the range δ_C/δ_H 60–85/2.5–5.5 (for X₂, X₃, X₄, and X₅) and signals from acetylated xylan moieties (X′₂ and X′₃) as well as signals from lignin, whereas the spectra of the isolated lignins (Figure 4) presented only signals from lignin. In general terms the signals of isolated lignins match those observed in the whole cell-walls, indicating that the MWL preparations used for the present work are representative of the native lignins in the cell-walls.

The aliphatic-oxygenated region of the spectra (around δ_C/δ_H 50−90/2.5−6.0, Figures 3A–C, 4A–C) gives information on the inter-unit linkages in lignin. In this region, cross-signals from methoxyl groups and from β−O−4′ alkyl-aryl ethers (structure A) are predominant in all samples, although differing in their intensities. Signals from other lignin substructures were also detected in the HSQC spectra, although with lower intensities, including signals from phenylcoumarans (B), resinols (C), dibenzodioxocins (D), open β-1 structures (E), spirodienones (F), and cinnamyl alcohol end-groups (I). This region of the spectrum can also provide information on the acylation degree in lignin. The HSQC spectrum of cork lignin (Figure 4A) clearly showed the occurrence of intense signals in the range from δ_C/δ_H 63.5/3.83-4.30 and at 64.3/4.63, that correspond to the C_γ-H_γ correlations of γ-acylated β−O−4′ alkyl-aryl ethers units (structures A′) and γ-acylated cinnamyl alcohol end-groups (I′). This indicates that cork lignin is partially acylated at the γ-position of the lignin side-chain. The estimation of the γ-acylation was accomplished by integration of the signals corresponding to the C_γ-H_γ correlations of the γ-hydroxylated (A) vs. γ-acylated (A′) structures, and indicated a 48% acylation degree of the lignin side-chains in cork. The HSQC spectra of the whole cell-walls of cork (Figures 3A,D) and their isolated ML (Figures 4A,D) showed the presence of a signal at δ_C/δ_H 80.7/4.51 characteristic for the C_β-H_β correlations of γ-acylated β−O−4′ substructures (A′) linked to G-units (del Río et al., 2012a,b, 2015), indicating a significant degree of γ-acylation of G-lignin units in the cork lignin. Signals from acylated lignin were not observed in the spectra of phloem (Figures 4B,E) and xylem (Figures 4C,F), indicating that these lignins are not acylated, or only to a very low extent. The aromatic region of the spectra (around δ_C/δ_H 100−155/6.0−8.0, Figures 3D–F, 4D–F) shows the signals from the aromatic rings and unsaturated side-chains of the different H-, G-, and S-lignin units, as well as from ferulates (structure FA). Signals from cinnamyl alcohol (I) and cinnamyl aldehyde end-groups (J) are also present in this region of the spectra. The content in cinnamaldehyde end-groups was estimated after comparing the intensities of C_β-H_β correlations in cinnamyl alcohols (I) and aldehydes (J).

The relative abundances of the main lignin inter-unit linkages and end-groups, as well as the percentage of γ-acylation, the molar abundances of the different lignin units (H, G, and S) and ferulates, and the S/G ratios of the lignins in the cork, phloem and xylem of Q. suber, estimated from volume integration of contours in the HSQC spectra, are shown in Table 4. Important differences were observed in the composition and structure of the lignins from the three tissues. The lignin from cork is enriched in G-units, with a H:G:S molar composition of 2:85:13, whereas the lignin from phloem has less G-units (H:G:S of 1:58:41), and the lignin from xylem is enriched in S-units (H:G:S of 1:45:55). The S/G ratios estimated by 2D-NMR were 0.1 in cork, 0.7 in phloem, and 1.6 and 1.2 in xylem (respectively in cell walls and isolated lignin). These values match quite closely those determined by Py-GC/MS, as reported above. These compositional differences were also reflected in the relative abundances of the different inter-unit linkages. β−O−4′ alkyl-aryl ethers are the most predominant linkages in the three lignins, but their relative abundances increase from 68% in cork, to 71% in phloem and to 77% in xylem, consistent with the enrichment in S-lignin units. Cork lignin is enriched in condensed linkages such as phenylcoumarans (20%), dibenzodioxocins (5%) and resinols (4%). On the opposite, phloem and xylem present a lignin with less phenylcoumarans (13 and 9% respectively), dibenzodioxocins (2 and 1%) but more resinols (7 and 8%) and a small amount of open β−1 structures (2 and 1%). Signals from cinnamyl alcohol (I) and cinnamaldehyde end-groups (J) were also observed, particularly in cork lignin, which is enriched in end-groups (5% of acylated cinnamyl alcohol, 8% of cinnamyl alcohol and 11% of cinnamaldehyde). These end-groups are also present, although in lower amounts, in the lignins from phloem and xylem. Finally, Table 4 shows the high extent of γ-acylation of the cork lignin (48%) that contrasts with the absence of lignin acylation in phloem and xylem tissues.

As mentioned before, cork lignin is partially acylated at the γ-position of the side-chain (48% of the units), while the lignins in phloem and xylem were not acylated. However, the nature of the acylating group could not be assessed by HSQC. Information regarding the nature of the acylation of the γ-OH was obtained from DFRC, a degradation method that cleaves α- and β-ether linkages in the lignin polymer leaving γ-esters intact and therefore is appropriate for analysis of γ-acylated lignins (Lu and Ralph, 1997a,b, 1998). The method was slightly modified (so called DFRC′) by replacing acetylating reagents with propionylating ones in order to evaluate the presence of acetate groups originally acylating the lignin γ-OH (Ralph and Lu, 1998; del Río et al., 2007b).

The GC-MS chromatograms of the DFRC′ degradation products of the lignins isolated from cork, phloem and xylem are presented in Figure 6. The compounds released were the cis- and trans- isomers of guaiacyl (cG, tG) and syringyl (cS, tS) lignin monomers (as their propionylated derivatives) arising from normal γ-OH units in lignin. In addition, the presence of originally γ-acetylated guaiacyl (cG_ac and tG_ac) and syringyl lignin units (cS_ac and tS_ac) were also detected in the chromatograms, confirming that acetylation at the γ-OH of the side-chain occurred in these lignins, being particularly abundant in cork lignin, and, to a much lower extent, also in phloem and xylem lignins. In cork lignin, acetylation occurred predominantly over the guaiacyl units (28% of the total G units are acetylated), as already advanced by 2D-NMR, whereas only minor amounts of syringyl units were acetylated (4% of the S-units). In contrast, the lignins from phloem and xylem were acetylated only to a very minor extent (not observed by 2D-NMR) and predominantly over S-units.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.