Expression of a bacterial 3‐dehydroshikimate dehydratase reduces lignin content and improves biomass saccharification efficiency

Summary Lignin confers recalcitrance to plant biomass used as feedstocks in agro‐processing industries or as source of renewable sugars for the production of bioproducts. The metabolic steps for the synthesis of lignin building blocks belong to the shikimate and phenylpropanoid pathways. Genetic engineering efforts to reduce lignin content typically employ gene knockout or gene silencing techniques to constitutively repress one of these metabolic pathways. Recently, new strategies have emerged offering better spatiotemporal control of lignin deposition, including the expression of enzymes that interfere with the normal process for cell wall lignification. In this study, we report that expression of a 3‐dehydroshikimate dehydratase (QsuB from Corynebacterium glutamicum) reduces lignin deposition in Arabidopsis cell walls. QsuB was targeted to the plastids to convert 3‐dehydroshikimate – an intermediate of the shikimate pathway – into protocatechuate. Compared to wild‐type plants, lines expressing QsuB contain higher amounts of protocatechuate, p‐coumarate, p‐coumaraldehyde and p‐coumaryl alcohol, and lower amounts of coniferaldehyde, coniferyl alcohol, sinapaldehyde and sinapyl alcohol. 2D‐NMR spectroscopy and pyrolysis‐gas chromatography/mass spectrometry (pyro‐GC/MS) reveal an increase of p‐hydroxyphenyl units and a reduction of guaiacyl units in the lignin of QsuB lines. Size‐exclusion chromatography indicates a lower degree of lignin polymerization in the transgenic lines. Therefore, our data show that the expression of QsuB primarily affects the lignin biosynthetic pathway. Finally, biomass from these lines exhibits more than a twofold improvement in saccharification efficiency. We conclude that the expression of QsuB in plants, in combination with specific promoters, is a promising gain‐of‐function strategy for spatiotemporal reduction of lignin in plant biomass.


Introduction
Plant cells walls are the primary source of terrestrial biomass and mainly consist of cellulosic and hemicellulosic polysaccharides impregnated with lignins. Lignins are polymers of p-hydroxycinnamyl alcohols (i.e. monolignols), which are synthesized inside the cells, exported to the cell wall and ultimately undergo oxidative polymerization via laccase and peroxidase activities. The main monolignolsp-coumaryl, coniferyl and sinapyl alcoholsgive rise to the p-hydroxyphenyl (H), guaiacyl (G) and syringyl (S) lignin units, respectively . Lignification generally confers mechanical strength and hydrophobicity in tissues that develop secondary cell walls, such as sclerenchyma (i.e. fibres) and xylem vessels. In addition to its essential role for upright growth, lignin also serves as a physical barrier against pathogens that degrade cell walls (Boudet, 2007).
Lignocellulosic biomass is used for pulp and paper manufacture, ruminant livestock feeding, and more recently has been considered an important source of simple sugars for fermentative production of intermediate or specialty chemicals and biofuels (Keasling, 2010). It is well-documented that lignin in plant biomass negatively affects pulp yield, forage digestibility and polysaccharide saccharification Chen and Dixon, 2007;Taboada et al., 2010). This has prompted major interest in developing a better understanding of lignin biosynthesis to reduce biomass recalcitrance by modifying lignin content and/or composition.
The shikimate pathway, which is located in plastids in plants, provides a carbon skeleton for the synthesis of phenylalanine, the precursor of the cytosolic phenylpropanoid pathway responsible for the biosynthesis of monolignols ( Figure 1). All the metabolic steps and corresponding enzymes for both pathways are known and well-conserved across land plants (Fraser and Chapple, 2011;Tohge et al., 2013;Umezawa, 2010). Classic approaches to lignin reduction have relied on genetic modifications, such as transcript reduction and allelic variation of specific genes from the phenylpropanoid pathway (Li et al., 2008;Vanholme et al., 2008). However, these strategies often result in undesired phenotypesincluding dwarfism, sterility and increased susceptibly to environmental stressesdue to loss of cell wall integrity, depletion of other phenylpropanoid-related metabolites, accumulation of pathway intermediates or the constitutive activation of defence responses (Bonawitz and Chapple, 2013;Voelker et al., 2011). Such negative effects are unfortunately difficult to avoid because of the nontissue specificity of the strategies employed: allelic variations are transmitted to every cell of the plant during cell divisions, and small interfering RNAs generated for gene silencing generally move from cell-to-cell and over long distance in vegetative tissues (Brosnan and Voinnet, 2011).
Alternatively, there are novel and promising gain-of-function strategies that involve expression of specific proteins to reduce the production of the three main monolignols or change their ratios. Using specific promoters with restricted expression patterns, these strategies would enable the alteration of lignin at later developmental stages or, for example, only in certain tissues such as fibreswithout compromising the functionality of conductive vessels for the transport of water (Voelker et al., 2011). Examples of such expressed proteins are transcription factors that act as negative regulators of lignin biosynthesis (Fornal e et al., 2010;Iwase et al., 2009;Shen et al., 2012;Yan et al., 2013); enzymes that produce alternative lignin monomers (Eudes et al., 2012;Wilkerson et al., 2014); engineered enzymes that modify monolignols into their nonoxidizable forms (Zhang et al., 2012); or proteins that mediate the post-transcriptional degradation of enzymes from the lignin biosynthetic pathway (Zhang et al., 2014).
In this study, we report for the first time on the expression of a bacterial 3-dehydroshikimate dehydratase in Arabidopsis (Teramoto et al., 2009). We selected QsuB from C. glutamicum and targeted it to the plastids to convert the shikimate precursor 3-dehydroshikimate into protocatechuate (Figure 1), with the aim of reducing lignin content and modifying its composition as shikimate is required for lignin biosynthesis. Metabolomic analysis of plants expressing QsuB revealed higher amounts of p-coumarate and of the two direct precursors of H-lignin units: p-coumaraldehyde and p-coumaryl alcohol. Conversely, the direct precursors of G and S unitsconiferaldehyde, coniferyl alcohol, sinapaldehyde and sinapyl alcoholwere reduced. Lignin content was severely reduced in these transgenic lines and exhibited an enrichment of H units at the expense of G units and a lower polymerization degree. Compared to those of wild-type plants, cell walls from lines expressing QsuB released significantly higher amounts of simple sugars after cellulase treatment and required less enzyme for saccharification. Collectively, these results support the hypothesis that expression of a plastidic QsuB affects the lignin biosynthetic pathway.

Targeted expression of QsuB in Arabidopsis
A sequence encoding QsuB was cloned downstream of the sequence encoding for a plastid-targeting signal peptide (SCHL) for expression in plastids. Using transient expression in tobacco, we first confirmed that QsuB was correctly targeted to the plastids by analysing its subcellular localization when fused at the C-terminus to a YFP marker ( Figure S1). The schl-qsuB sequence was cloned downstream of the Arabidopsis C4H promoter for expression in lignifying tissues of Arabidopsis. Western blot analysis confirmed that QsuB was expressed in stems of several T3 plants homozygous for the pC4H::schl::qsuB (thereafter C4H::qsuB) construct ( Figure 2). Based on the migration of molecular weight markers, QsuB was detected at around 70 kDa, which corresponds to the theoretical size of its native sequence after cleavage of the chloroplast transit peptide (Figure 2). Four homozygous lines with different QsuB expression levels (C4H::qsuB-1, -3, -6 and -7) were selected for biomass measurement. Although a height reduction was observed for these lines, only C4H::qsuB-1 showed a slight decrease (À18%) of biomass yield (Table 1).

Metabolite analysis of C4H::qsuB lines
Methanol-soluble metabolites from stems of the four homozygous C4H::qsuB lines were extracted for analysis (Table 2, Figure  S2). Compared to wild-type plants, protocatechuate content was increased 67-to 113-fold in the transgenic lines. However, no significant reduction was observed for the content of several metabolites derived from the shikimate pathway such as salicylate and aromatic amino acids (i.e. phenylalanine, tyrosine and tryptophan). Interestingly, several metabolites from the phenylpropanoid pathway were increased in the transgenic lines; p-coumaraldehyde and p-coumaryl alcohol, the two direct precursors of H-lignin units, were increased 5.7-16.4-fold and 12.2-13.7-fold, respectively. Similarly, p-coumarate content was increased 6.4-9.5-fold compared to wild type. In contrast, the direct precursors of G-and S-lignin units were negatively altered in transgenic lines. Coniferaldehyde and coniferyl alcohol were reduced by 33-50% and 36-68%, respectively. Sinapaldehyde and sinapyl alcohol were decreased by 45-77% and 73-87%, with the exception of line C4H::qsuB-1 which showed no significant difference for sinapaldehyde compared to wild type (Table 2).
Cell wall-bound p-coumarate and ferulate released from cell wall residues by mild alkaline hydrolysis were also analysed ( Table 3). The content of p-coumarate was significantly increased in the C4H::qsuB lines (1.75-3-fold), whereas ferulate was reduced (1.8-2.9-fold). In addition, bound p-coumaraldehyde could be detected in cell wall samples from the transgenic lines but not in those from wild type (Table 3).
Lignin content and monomeric composition in C4H:: qsuB lines The Klason method was used to measure the lignin content: a reduction ranging from 45% (C4H::qsuB-7) to 52% (C4H:: qsuB-1) was observed in stems of the C4H::qsuB lines compared to wild type (Table 4). Cell wall material from stems of wild-type and C4H::qsuB lines was analysed by pyro-GC/MS for the determination of the lignin monomer composition. For each line, identification and relative quantification of the pyrolysis products derived from H, G or S units allowed determination of H/G/S ratios (Table 4, Table S1). Compared to wild type, the relative amount of H units is increased between 3.3-fold (C4H::qsuB-3) and 6-fold (C4H::qsuB-6) in transgenics. The relative amount of S units is moderately increased 1.3-1.5-fold, whereas that of G units is reduced 1.5-1.8-fold in the C4H::qsuB lines. NMR (2D 13 C-1 H-correlated, HSQC) spectra of cell wall material from wild-type and C4H::qsuB-1 plants were also obtained for determination of lignin composition and structure. Analysis of the aromatic region of the spectra confirmed the higher relative amount of H units in C4H::qsuB-1 (27.2%) compared to wild type (3.8%), as well as a reduction of G units ( Figure 3). Moreover, analysis of the aliphatic region of the spectra indicated a diminution of phenylcoumaran (b-5) and resinol (b-b) linkages in the lignin of the C4H::qsuB-1 line ( Figure S3).

Lignins from C4H::qsuB plants have a lower polymerization degree
Lignin fractions were isolated from wild-type and C4H::qsuB-1 plants for analysis of their polydispersity using size-exclusion chromatography (SEC). Elution profiles acquired by monitoring UV-F fluorescence of the dissolved lignin revealed differences between wild-type and the transgenic line ( Figure 4). The total area of the three mass peaks, corresponding to the largest lignin fragments detected between 7.8 and 12.5 min, was significantly reduced in C4H::qsuB-1 compared to wild type. Similarly, intermediate molecular mass material, which elutes in a fourth peak between 12.5 and 18 min, was also less abundant in the C4H::qsuB line. Conversely, the area corresponding to the smallest lignin fragments, detected between 18 and 23.5 min, was increased in the transgenic line. These results demonstrate a reduction in the degree of polymerization of lignins purified from plants expressing QsuB compared to that of wild type.
Biomass from C4H::qsuB lines shows improved saccharification Saccharification assays on stem material were conducted to evaluate the cell wall recalcitrance of the C4H::qsuB lines. As shown in Figure 5a, higher amounts of sugars were released after 72 h enzymatic hydrolysis of biomass from the C4H::qsuB lines Table 1 Height and dry weight of the main inflorescence stem of senesced mature wild-type (WT) and pC4H::schl::qsuB (C4H::qsuB) plants. Number, n, of plants analysed Plant  Asterisks indicate significant differences from the wild type using the unpaired Student's t-test (*P < 0.005; **P < 0.001).  Figure 2 QsuB expression in Arabidopsis stems. Detection by Western blot of QsuB tagged with the AttB2 peptide (approximate size 70 kDa) using the 'universal antibody' and stem proteins from eight independent 6-week-old homozygous pC4H::schl::qsuB T3 transformants. A stem protein extract from wild type was used as a negative control (WT), and a Ponceau staining of Rubisco large subunit (rbcL) is shown as a loading control. compared to those of wild type in all pretreatments tested. Saccharification improvements ranged between 79-116% after hot water, 63-93% after dilute alkali and 26-37% after dilute acid pretreatments ( Figure 5a). Moreover, similar saccharification experiments using hot water-pretreated biomass, at 59 lower cellulase loadings, revealed that biomass from all C4H::qsuB lines releases more sugar than that of wild type hydrolysed with a typical enzyme loading (Figure 5b). Taken together, these data demonstrate that cellulose from the C4H::qsuB lines is less recalcitrant to cellulase digestion and requires a lower amount of enzyme to be converted into high yields of fermentable sugars.

Discussion
Gain-of-function strategies have several advantages for the manipulation of metabolic pathways. For example, they can be used to bioengineer lignin deposition in plants via better spatiotemporal control of monolignol production in lignifying cells and to adjust lignin composition and its biophysical properties (Eudes et al., 2014). Therefore, identification of proteins in which in planta-expression results in modifications of lignin content or composition is of particular interest and presents novel opportunities. In this work, we demonstrate that expression of the 3-dehydroshikimate dehydratase QsuB in plastids leads to drastic reduction and compositional changes of lignin in Arabidopsis (Table 4). As a result, biomass from these transgenic plants exhibits much higher saccharification efficiency after pretreatment (Figure 5a), which is a highly desired trait for several agroindustries and the bioenergy sector. Moreover, the efficiency of this approach to decrease lignin content in plant biomass allows a reduction of hydrolytic enzyme loadings by at least fivefold, while retaining greater saccharification potential than control plants hydrolysed at standard enzyme loading (Figure 5b). Consequently, the transfer of this technology to energy crops should have a great impact on the cost-effectiveness of cellulosic biofuels production, because enzyme cost is the major barrier in this process (Klein-Marcuschamer et al., 2012). In this study, as a proof of concept, we used the promoter of the AtC4H gene to ensure strong QsuB expression in all lignifying tissues of the plant. This resulted in a slight decrease of plant height for all the lines, but no significant reductions in biomass yield except for that of one transgenic line, which expressed QsuB strongly (Table 1; Figure 2) and exhibitedin some stem transverse sections ( Figure S4)evidence of vessel collapse that could impair xylem conductivity (Voelker et al., 2011). Nevertheless, our strategy offers the potential to overcome these defects by selecting more stringent promoters (e.g. fibre-specific) that would exclude QsuB expression from xylem-conductive elements (Eudes et al., 2014;Yang et al., 2013). Other particular phenotypes such as increased branching or thicker stems, which would explain the unaffected biomass yield despite a reduced height in some transgenic lines, were not observed ( Figure S5). Hypothetically, the biomass from the transgenic lines could be denser than that of wild type as previously shown in plants with lower lignin content (Nuopponen et al., 2006). Moreover, translation of our technology from model plant to crops is expected to be straightforward: it is based solely on the expression of QsuB and does not require any particular genetic backgrounds, and the lignin and shikimate pathways are well-conserved among vascular plants.
A direct consequence of QsuB expression is the accumulation of protocatechuate in the biomass of transgenic plants. Considering the beneficial properties of protocatechuate in the bio-based chemical industry, such de novo production adds extra commercial value to the biomass of plants expressing QsuB (Linger et al., 2014;Otsuka et al., 2006). Much higher amounts of protocatechuate were recovered after acid treatment of the methanolsoluble extracts from transgenic plants (data not shown), which suggests its conjugation in the cytosol after export from the plastids. Interestingly, in stems of 5-week-old plants, QsuB expression did not affect the overall level of metabolites derived from the shikimate pathway such as aromatic amino acids and salicylate, suggesting that plastidic 3-dehydroshikimate is not limiting ( Table 2). Although an alternative cytosolic pathway for salicylate has been described , the de novo biosynthesis of aromatic amino acids outside plastids remains undefined (Maeda and Dudareva, 2012), suggesting that 3-dehydroshikimate is not limiting at least for the biosynthesis of aromatic amino acids in plants expressing QsuB. On the other hand, a build-up of the pool of p-coumarate, p-coumaraldehyde Asterisks indicate significant differences from the wild type using the unpaired Student's t-test (*P < 0.05; **P < 0.01).  Figure S2). Analysis of the lignin monomeric compositionusing 2D NMR spectroscopy and pyro-GC/MSunequivocally demonstrated an increase in H units in plants expressing QsuB (Figure 3; Tables 4 and S1). These data could explain the reduced degree of polymerization of these lignins, which has been previously observed in various lignin mutants that exhibit high content of H units, incorporation of which typically slows or stops lignin-chain elongation (Sangha et al., 2014;Ziebell et al., 2010; Figure 4). Therefore, reduced lignin-polysaccharide cross-linking within the biomass of the transgenic lines is expected, and this could contribute to its superior enzymatic digestibility (Ralph et al., 2004).
A low lignin content rich in H units and higher S/G corresponds to a phenotype previously characterized in plants down-regulated for hydroxycinnamoyl-CoA shikimate/quinate hydroxycinnamoyl transferase (HCT), p-coumarate 3-hydroxylase (C3H) or caffeoyl shikimate esterase (CSE) (Ralph et al., 2006;Vanholme et al., 2013;Ziebell et al., 2010). Moreover, reduction of HCT activity results in the accumulation of free and bound p-coumaraldehyde in cucumber and of p-coumarate in alfalfa, presumably due to the build-up of coumaroyl-CoA (Gallego-Giraldo et al., 2011;Varbanova et al., 2011). This suggests that an alteration of these biosynthetic steps has occurred in the C4H::qsuB lines. However, no particular reduction of transcript abundance for HCT, C3H and CSE was observed in the transgenic lines compared to wild type ( Figure S6). A possible explanation is that QsuB activity, which consumes 3-dehydroshikimate in lignifying tissues, affects indirectly the amount of shikimate available for HCT in the cytosol. Although some enzymes of the shikimate pathway exist in the cytosol, there is so far no evidence for a complete alternative extra-plastidial shikimate biosynthetic pathway. Instead, a yetunidentified transporter probably mediates the export of shikimate from the plastid to the cytosol (Maeda and Dudareva, 2012). If such transport system is only active at a narrow range of concentrations, a reduction of shikimate content in plastids (as anticipated in plants expressing QsuB) would compromise its export to the cytosol. Moreover, it is possible that the large amount of protocatechuate generated by QsuB activity in plastid competes with shikimate export. The distribution of shikimate between plastids and the cytosol is still poorly understood, and shikimate levels were below the detection limit in our stem extracts from wild type and transgenic plants. Alternatively (Moglia et al., 2010;Sander and Petersen, 2011), the large accumulation of protocatechuate could act as competitive inhibitor of HCT, thus limiting the synthesis of coumaroyl shikimate required for the production of G-and S-lignin units.

Plant material and growth conditions
Arabidopsis thaliana (ecotype Columbia, Col-0) seeds were germinated directly on soil. Growing conditions were 150 lmol/ m 2 /s, 22°C, 60% humidity and 10 h of light per day. Selection of T2 and identification of T3 homozygous transgenic plants were made on Murashige and Skoog vitamin medium (PhytoTechnology Laboratories, Shawnee Mission, KS, USA), supplemented with 1% sucrose, 1.5% agar and 50 lg/mL kanamycin.

Western blot analysis
Proteins from Arabidopsis stems were extracted using a buffer containing 250 mM Tris-HCl pH 8.5, 25 mM EDTA, 2 mM DTT, 5 mM b-mercaptoethanol and 10% sucrose, and were quantified using the Bradford method (Bradford, 1976). Proteins (15 lg) were separated by SDS-PAGE, blotted and immunodetected using a universal antibody, as previously described (Eudes et al., 2011).

Methanol-soluble metabolites extraction
Arabidopsis stems of 5-week-old wild-type and T3 homozygous C4H::qsuB lines were collected in liquid nitrogen and stored at À80°C until further utilization. Prior the metabolite extraction, collected stems were pulverized in liquid nitrogen. For extraction of methanol-soluble metabolites, 700-1000 mg of frozen stem  4). Asterisks indicate significant differences from the wild type using the unpaired Student's t-test (*P < 0.05; **P < 0.005). (b) Amounts of sugars released from biomass after hot water pretreatment and 72-h enzymatic digestion using two different cellulase loadings (1% or 0.2% w/w). Values are means AE SE of four biological replicates (n = 4). Asterisks indicate significant differences from the wild type at 1% cellulase loading using the unpaired Student's t-test (*P < 0.05; **P < 0.005). powder was mixed with 2 mL of 80% (v/v) methanol-water and mixed (1400 rpm) for 15 min at 70°C. This step was repeated four times. Pooled extracts were cleared by centrifugation (5 min, 20 000 g, at room temperature), mixed with 4 mL of analytical grade water and filtered using Amicon Ultra centrifugal filters (10 000 Da MW cut-off regenerated cellulose membrane; EMD Millipore, Billerica, MA, USA). Filtered extracts were lyophilized and the resulting pellets dissolved in 200 lL 50% (v/v) methanolwater prior to LC-MS analysis. An acid hydrolysis of the samples was performed for the quantification of protocatechuate and salicylate; an aliquot of the filtered extracts was dried under vacuum, resuspended with 1 N HCl and incubated at 95°C for 3 h. The mixture was subjected to three ethyl acetate partitioning steps. Ethyl acetate fractions were pooled, dried in vacuo and resuspended in 50% (v/v) methanol-water prior to LC-MS analysis.

LC-MS analysis
Phenolic acids, phenolic aldehydes, and aromatic amino acids were analyzed using high-performance liquid chromatography (HPLC), electrospray ionization (ESI), and time-of-flight (TOF) mass spectrometry (MS) as previously described in Eudes et al. (2013) and Bokinsky et al. (2013), respectively. Aromatic alcohols were analysed by HPLCatmospheric pressure chemical ionization (APCI) -TOF MS. Their separation was conducted on an Agilent 1200 Series Rapid Resolution HPLC system (Agilent Technologies Inc., Santa Clara, CA, USA) using a Phenomenex Kinetex XB-C18 (100 mm length, 2.1 mm internal diameter and 2.6 lm particle size; Phenomenex, Torrance, CA, USA). The mobile phase was composed of 0.1% formic acid in water (solvent A) and methanol (solvent B). The elution gradient was as follows: from 5% B to 25% B for 6 min, 25% B to 5% B for 1 min and held at 5% B for a further 3 min. A flow rate of 0.5 mL/min was used throughout.
The column compartment and sample tray were set to 50 and 4°C, respectively. The HPLC system was coupled to an Agilent Technologies 6210 LC/TOF mass spectrometer with a 1:4 postcolumn split. Mass spectrometric detection was conducted using APCI in the positive ion mode. MS experiments were carried out in the full-scan mode, at 0.86 spectra/second, for the detection of [M-H 2 O+H] + ions. Drying and nebulizing gases were set to 10 L/min and 25 psi, respectively, and a drying gas temperature of 330°C was used throughout. The vaporizer and corona were set to 350°C and 4 lA, respectively, and a capillary voltage of 3500 V was also used. Fragmentor and OCT 1 RF voltages were each set to 135 V, while the skimmer voltage was set to 50 V. Data acquisition and processing were performed by the Mas-sHunter software package (Agilent Technologies Inc.). Metabolites were quantified via 10-point calibration curves of authentic standard compounds for which the R 2 coefficients were ≥0.99. Representative LC-MS chromatograms obtained from solutions of standard compounds and from plant metabolite extracts are illustrated in Figure S7.

Lignin content and composition
The biomass from senesced wild-type plants and T3 homozygous C4H::qsuB lines was used to determine lignin content and composition. Biomass was extracted sequentially by sonication (20 min) with 80% ethanol (three times), acetone (one time), chloroform-methanol (1:1, v/v, one time) and acetone (one time). The standard NREL biomass protocol was used to measure lignin content (Sluiter et al., 2008). The chemical composition of lignin was analysed by pyrolysis-gas chromatography (GC)/mass spec-trometry (MS) using a previously described method with some modifications (Del R ıo et al., 2012). Pyrolysis of biomass was performed with a Pyroprobe 5200 (CDS Analytical Inc., Oxford, PA, USA) connected with GC/MS (Thermo Electron Corporation with Trace GC Ultra and Polaris-Q MS) equipped with an Agilent HP-5MS column (30 m 9 0.25 mm i.d., 0.25 lm film thickness).
The pyrolysis was carried out at 550°C. The chromatograph was programmed from 50°C (1 min) to 300°C at a rate of 30°C/ min; the final temperature was held for 10 min. Helium was used as the carrier gas at a constant flow rate of 1 mL/min. The mass spectrometer was operated in scan mode and the ion source was maintained at 300°C. The compounds were identified by comparing their mass spectra with those of the NIST library and those previously reported (Del R ıo and Guti errez, 2006;Ralph and Hatfield, 1991). Peak molar areas were calculated for the lignin degradation products, and the summed areas were normalized.

Cell wall-bound aromatics extraction
The biomass from senesced wild-type plants and T3 homozygous C4H::qsuB lines was used to measure cell wall-bound aromatics. Extracted biomass (10 mg) was mixed with 500 lL of 2 M NaOH and shaken at 1400 rpm for 24 h at 30°C. The mixture was acidified with 100 lL of concentrated HCl and subjected to three ethyl acetate partitioning steps. Ethyl acetate fractions were pooled, dried in vacuo and suspended in 50% (v/v) methanolwater prior to LC-MS analysis.

Size-exclusion chromatography
Lignin solutions, 1% (w/v), were prepared in analytical grade 1-methyl-2-pyrrolidinone (NMP). The polydispersity of dissolved lignin was determined using analytical techniques involving SEC UV-F 250/400 as previously described (George et al., 2011). An Agilent 1200 series binary LC system (G1312B) equipped with diode-array (G1315D) and fluorescence (G1321A) detectors was used. Separation was achieved with a Mixed-D column (5 lm particle size, 300 mm 9 7.5 mm i.d., linear molecular mass range of 200 to 400 000 u, Agilent Technologies Inc.) at 80°C using a mobile phase of NMP at a flow rate of 0.5 mL/min. Absorbance of materials eluting from the column was detected using UV-F fluorescence (Ex 250 /Em 450 ). Spectral intensities were area-normalized, and molecular mass estimates were determined after calibration of the system with polystyrene standards.

Cell wall pretreatments and saccharification
Ball-milled senesced stems (10 mg) were mixed with 340 lL of water, 340 lL of H 2 SO 4 (1.2%, w/v) or 340 lL of NaOH (0.25%, w/v) for hot water, dilute acid or dilute alkali pretreatments, respectively, shaken at 1400 rpm (30°C, 30 min) and autoclaved at 120°C for 1 h. Samples pretreated with dilute acid were neutralized with 5 N NaOH (25 lL). Saccharification was initiated by adding 650 lL of 100 mM sodium citrate buffer pH 5 (for hot water-and dilute alkali-pretreated samples) or 625 lL of 80 mM sodium citrate buffer pH 6.2 (for dilute acid-pretreated samples) containing 80 lg/mL tetracycline and 1% w/w or 0.2% w/w Cellic CTec2 cellulase (Novozymes). After 72 h of incubation at 50°C with shaking (800 rpm), samples were centrifuged (20 000 g, 3 min) and 10 lL of the supernatant was collected for measurement of reducing sugars using the 3,5-dinitrosalicylic acid assay and glucose solutions as standards (Miller, 1959).

Conflict of interests
JDK has financial conflict of interests in Amyris, LS9 and Lygos. DL has financial conflict of interests in Afingen.