Modified expression of ZmMYB167 in Brachypodium distachyon and Zea mays leads to increased cell wall lignin and phenolic content

One of the challenges to enable targeted modification of lignocellulosic biomass from grasses for improved biofuel and biochemical production lies within our limited understanding of the transcriptional control of secondary cell wall biosynthesis. Here, we investigated the role of the maize MYB transcription factor ZmMYB167 in secondary cell wall biosynthesis and how modified ZmMYB167 expression in two distinct grass model species affects plant biomass and growth phenotypes. Heterologous expression of ZmMYB167 in the C3 model system Brachypodium led to mild dwarf phenotypes, increased lignin (~7% to 13%) and S-lignin monomer (~11% to 16%) content, elevated concentrations of cell wall-bound p-coumaric acid (~15% to 24%) and reduced biomass sugar release (~20%) compared to controls. Overexpression of ZmMYB167 in the C4 model system Zea mays increased lignin (~4% to 13%), p-coumaric acid (~8% to 52%) and ferulic acid (~13% to 38%) content but did not affect plant growth and development nor biomass recalcitrance. Taken together, modifying ZmMYB167 expression represents a target to alter lignin and phenolic content in grasses. The ZmMYB167 expression-induced discrepancies in plant phenotypic and biomass properties between the two grass model systems highlight the challenges and opportunities for MYB transcription factor-based genetic engineering approaches of grass biomass.

vary substantially depending on the biomass feedstock 9 , thus imparting a role to secondary cell wall properties. Grasses also incorporate considerable amounts of ferulic acid (FA) and p-coumaric acid (p-CA) into the secondary wall, both phenolics derived from the phenylpropanoid pathway. FA, in particular, is involved in cross-linking to GAX and lignin, forming a covalently linked carbohydrate-lignin complex 10 . The functional role of p-CA in grass cell walls is less clear, but these phenolics appear to be predominantly attached to lignin through its ester linkages to monolignols, primarily sinapyl alcohol, and are also found to be acylated to GAX 11 .
The relative abundances, cross-linkages, interactions and arrangements of secondary cell wall components within a dense and hydrophobic matrix collectively lead to inherent resistance to cell wall deconstruction, known as biomass recalcitrance 12 . This aspect represents a critical biological as well as a technical barrier for biorefining lignocellulose. Accordingly, studies uncovering the molecular and genetic mechanisms underpinning secondary cell wall biosynthesis and biomass recalcitrance have helped drive traditional plant breeding practices and biotechnological approaches aimed at developing crops for different end-use applications. The well-studied herbaceous feedstock maize (Zea mays) represents a powerful and versatile genetic model system and grass crop to address these matters with its completed genome sequence, past breeding success as well as the development of its genetic tools and resources 13 . It is also a member of the highly photosynthetic-efficient C 4 clade of grasses sharing a close evolutionary phylogeny and degree of gene synteny to the prime biorefining feedstocks Miscanthus, switchgrass (Panicum virgatum), sorghum (Sorghum bicolor), and sugarcane (Saccharum officinarum) 14 . Additionally, smaller-genome model grasses such as Brachypodium distachyon could facilitate the processes of gene discovery and translational genomics to the genetically more challenging grasses, adding value as a comparative model to increase our knowledge of secondary cell wall biosynthesis and biomass production in grasses 15 .
A sophisticated, extensive and multi-level network of transcription factors (TFs) has emerged over the last decade, controlling the secondary cell wall biosynthesis programme in a highly coordinated and orchestrated fashion [16][17][18] . Of these, lower network tier MYB (MYELOBLASTOSIS) TFs act as crucial transcriptional regulators of secondary cell wall biosynthesis in several plant species, most frequently studied in the pioneer dicot model plant Arabidopsis thaliana 19 . Substantially less published literature is available on bioengineering the next generation of biotechnology grasses using MYB TF-based strategies 20 . For instance, OsMYB103L overexpression (OX) and RNA interference (RNAi) in transgenic rice resulted in a ~13% increase and ~15% to 30% decrease in cellulose content respectively 21 . Other work identified PvMYB4 as a transcriptional repressor of phenylpropanoid biosynthesis and its overexpression in switchgrass resulted in ~50% reduction in lignin and phenolic content, which in turn improved ethanol yields by ~2.5-fold 22,23 . The overexpression of SbMYB60, a transcriptional activator regulating lignin and possibly cellulose/hemicellulose biosynthesis, in sorghum resulted in a ~10% increase in lignin content, leading to a higher energy content of the biomass 24 . In another study, overexpression of ZmMYB42 was accompanied by a ~8% to 21% reduction in lignin content and ~30% more glucose release in transgenic sugarcane 25 . Overexpression of ZmMYB42 or ZmMYB31 in Arabidopsis reduced lignin content resulting in dwarfed plants and more enzymatically degradable cell walls 26,27 . Syntelogs of MYB31 and MYB42 across three different grass species (maize, sorghum and rice) were shown to bind to phenylpropanoid gene promoters in vivo, although with some subfunctionalisation between the species 28 . ZmMYB46 or OsMYB46, when overexpressed in Arabidopsis, were able to activate the entire secondary cell wall biosynthetic programme 29 . ZmMYB5 and ZmMYB152 may also play a regulatory role in phenylpropanoid biosynthesis but their function was not evaluated in planta 30 . Overall, these transgenic approaches have proven informative to understand the regulatory roles of MYB TFs in secondary cell wall biosynthesis of grasses and underline the potential of MYB TFs for developing advantageous lignocellulosic biomass qualities.
A previously published maize transcriptome analysis revealed several uncharacterised MYBs with potential involvement in the transcriptional regulation of secondary cell wall biosynthesis in grasses 31 . However, their potential to alter lignocellulose properties for improved biomass processing and biochemical production have remained unexplored to date. Here, we utilised two distinct grass model species to investigate the role of a maize MYB TF, designated as ZmMYB167 (GRMZM2G037650) by the GRASSIUS TF database 32 , in secondary cell wall biosynthesis. Modified expression of ZmMYB167 in both transgenic Brachypodium distachyon and maize plants led to a higher abundance of cell wall lignin and phenolics but with distinct effects on plant growth phenotype and biomass processing properties. These results are informative for TF-based bioengineering strategies aimed at improving the economic value of bioenergy grasses via carbon-neutral production of biofuels and value-added phenolic crude matter.

Results
ZmMYB167 is a potential orthologue of OsMYB42/85, PvMYB42/85A and AtMYB85. A phylogenetic relationship of known secondary cell wall regulatory MYB TFs was modelled with a maximum likelihood approach to forecast the functional role of the maize MYB TF ZmMYB167 in secondary cell wall biosynthesis. The phylogenetic tree topology revealed that MYB TFs with similar roles in secondary cell wall biosynthesis clustered together (Fig. 1a). Thus, ZmMYB167 may function similarly to OsMYB42/85, PvMYB42/85A and AtMYB85, which have been demonstrated as transcriptional activators of phenylpropanoid biosynthesis [33][34][35] . Amino acid sequence alignment showed that the ZmMYB167 protein is 64% and 61% identical to its potential orthologue in rice (OsMYB42/85) and switchgrass (PvMYB42/85A), respectively, and 50% identical with AtMYB85 (Figs 1b and S1). In addition, the phylogenetic analysis revealed ZmMYB17 as a syntelog of ZmMYB167 (Fig. 1a). Motif analysis of ZmMYB167 with the characterised OsMYB42/85, PvMYB42/85A and AtMYB85 proteins further indicated conserved DNA-binding R2 and R3 MYB sites within the N-terminal region and potential functional motifs within the C-terminal region (Fig. S2), the latter usually containing transcriptional activator or repressor activity for the regulation of gene expression 36 . Based on these results coupled with the up-regulation of ZmMYB167 in a maize internode in which many cells were undergoing secondary cell wall deposition 31 , we hypothesised that ZmMYB167 is an important transcriptional regulator of secondary cell wall biosynthesis in grasses.
www.nature.com/scientificreports www.nature.com/scientificreports/ Expression of ZmMYB167 in Brachypodium distachyon. As a first step to examine the involvement of ZmMYB167 in the transcriptional regulation of secondary cell wall biosynthesis in grasses in vivo, we expressed the ZmMYB167 gene under control of the constitutive maize ubiquitin promoter (ZmUbi1) in Brachypodium distachyon. Four independent transgenic lines in the T 1 generation (named Bd6, Bd8, Bd9 and Bd10) exhibiting heterologous expression of ZmMYB167 transcripts (Fig. S3a) were evaluated for plant growth and cell wall characteristics. The ZmMYB167 Brachypodium expression lines showed reduced plant growth phenotypes 30 days after germination (Fig. 2a). Stem cross-sections from two out of the three randomly selected transgenic ZmMYB167 Brachypodium lines stained with phloroglucinol-HCl for lignin demonstrated noticeable increases in lignin deposition particularly in the epidermal cells and cortex (Fig. 2b), possibly induced via the constitutively active ZmUbi1 promoter. A double staining test with Calcofluor White and phloroglucinol-HCl of the same stem cross-sections showed noticeable staining intensity differences in all three transgenic lines when compared with null-segregant controls (non-transgenic progeny of transgenic parent lines), although the exact cell wall features underpinning these differences remain to be determined (Fig. 2c). Even though transgenic ZmMYB167 Brachypodium plants were fertile and produced seeds, plant height was significantly reduced by ~22% and biomass yield by ~43% on average, while tillering was unaffected compared to controls (Fig. 3). Similar results were observed for rice and switchgrass plants overexpressing OsMYB42/85 and PvMYB42/85A respectively which showed a mild dwarf phenotype 34,35 , though growth characteristics were not reported for AtMYB85-OX Arabidopsis plants 33 . Altered plant growth phenotypes have been reported in several other transgenic grasses because of changes in lignification 22,24 . Indeed, all four ZmMYB167 transgenic Brachypodium lines exhibited increased levels of acetyl bromide soluble lignin (ABSL) content by ~7% to 13% compared to controls ( Fig. 4a and Table 1). There were also higher levels of p-CA (~15% to 24%) and syringyl (S) lignin monomers (~11% to 16%), lower levels of guaiacyl (G) lignin monomers (~17% to 25%), and a concomitant increase in the S/G ratio (~32% to 53%) in at least three ZmMYB167 transgenic Brachypodium lines (Table 1). While the overall abundance of cell wall polysaccharides remained mostly unchanged (Table 1). Moreover, the glucose yields after 72 hrs of enzymatic hydrolysis of untreated biomass were reduced in all four ZmMYB167 transgenic Brachypodium lines; on average by ~20% compared to controls (Fig. 4b).
Overexpression of ZmMYB167 in Zea mays. We next generated independent F 1 maize progeny from five transformation events (Zm1 to Zm5) harbouring ZmMYB167 under the control of ZmUbi1 (Fig. S3b) to study the ZmMYB167 overexpression (OX) effects in the endogenous model system. Quantitative Real-time PCR analysis verified that ZmMYB167 expression levels were higher in four transgenic lines (Zm1, Zm2, Zm3 and Zm5), ranging from ~1.5 to 270-fold higher, relative to the respective null-segregant controls for the transgene (Fig. S4).    www.nature.com/scientificreports www.nature.com/scientificreports/ The expression levels of endogenous ZmMYB167 and the ZmMYB17 syntelog appeared not to be affected by the overexpression of ZmMYB167 (Fig. S5). Unlike the phenotypes observed in the C 3 transgenic ZmMYB167 Brachypodium lines, overexpression of ZmMYB167 in the C 4 maize lines did not affect plant growth and development (Fig. 5a). Although there was variation in plant height (~124 to 198 cm), flowering time (79 to 87 days), and stem biomass (~24 to 54 g) for the transgenic plants at the vegetative (VT) growth stage, the values for the individual ZmMYB167 maize lines were not substantially different from their corresponding null-segregant controls originating from the same transformation event (Table S1).
In contrast to the ZmMYB167 Brachypodium lines, ectopic lignin deposition was not apparent in phloroglucinol-HCl stained stem cross-sections of the ZmMYB167-OX lines when compared to control plants (Fig. 5b). There were also no apparent differences in the double staining with Calcofluor White/ phloroglucinol-HCl (Fig. 5c) or the Maüle staining (Fig. S6), as well as in the overall organisation of vascular bundles and fibre cells (Fig. 5b and c). However, ABSL content determined for stem biomass of F 1 generation plants at the V13 stage was significantly elevated in four of the ZmMYB167-OX lines (Zm1, ~8%; Zm2, ~4%; Zm3, ~13% and Zm5, ~7%) when compared to the corresponding control (Table 2). In contrast to the results in Brachypodium, the relative percentage of thioacidolysis released S and G lignin monomers, and hence the S/G ratio, were not substantially different between the ZmMYB167-OX lines and controls, averaging ~56% and ~42% respectively in both the ZmMYB167-OX lines and control plants (Table 2). However, concomitant with an increase in ABSL, the cell walls of these four ZmMYB167-OX lines contained significantly higher levels of p-CA (~8% to 52%) relative to control samples ( Table 2). There were also significantly higher levels of FA (~13% to 38%) in three ZmMYB167-OX lines ( Table 2). Since both ABSL and total p-CA and FA content was higher in at least three ZmMYB167-OX lines, we predicted these OX lines to exhibit alterations in Klason lignin. As shown in Table 2, total Klason lignin content was indeed elevated by ~7% to 13% in all the four ZmMYB167-OX lines.  www.nature.com/scientificreports www.nature.com/scientificreports/ The effects of ZmMYB167 overexpression on increased lignin and phenolic acids content led us to propose that the stem biomass of these transgenic maize lines may also be less susceptible to enzymatic hydrolysis. Interestingly, none of the ZmMYB167 transgenic maize lines showed considerable changes in glucose yields (~27% to 41%) after 72 hrs of enzymatic hydrolysis when compared to control samples (Fig. 6). Following a mild pre-treatment with alkaline (0.2 M NaOH), higher glucose yields (~72% to 95%) were obtained, illustrating that the removal of lignin, p-CA and FA by this alkaline pre-treatment exposes glucan to enzymatic hydrolysis. However, there were still no significant differences in glucose yields between the ZmMYB167-OX lines and controls (Fig. 6).

Discussion
A thorough understanding of the regulatory mechanisms underlying secondary cell wall biosynthesis is essential to tailor grasses for sustainable bio-based applications. An array of MYB TFs predominantly regulating phenylpropanoid biosynthesis and impacting secondary wall formation have been identified in dicot plants, including Arabidopsis and poplar 19,37 . In contrast, relatively few MYB TFs have been directly evaluated in monocots for a regulatory role in secondary cell wall biosynthesis and as a potential target for bioenergy crop improvement 20 .
Here, we highlight that modified ZmMYB167 expression in Brachypodium and Zea mays can have different implications on plant growth and development as well as cell wall composition and biomass processing efficiency, emphasising both the opportunities and challenges of using MYB TFs as genetic engineering tools.
Phylogenetic analysis revealed that ZmMYB167 is orthologous to the phenylpropanoid biosynthesis activators OsMYB42/85, PvMYB42/85A and AtMYB85 (Fig. 1a), and amino acid sequence analysis of ZmMYB167 indicated that it is a typical R2R3-MYB with conserved R2 and R3 motifs (Figs 1b and S2). A similar candidate gene identification procedure proved effective for identifying TFs involved in secondary cell wall biosynthesis from rice and Miscanthus 38,39 . To confirm our hypothesis that ZmMYB167 regulates lignin biosynthesis, we performed a functional analysis in two different grass model systems, Brachypodium and maize. Expression of ZmMYB167 (heterologous in Brachypodium and overexpression in maize) led to increased levels of ABSL in both Brachypodium (~7% to 13%) and maize (~4% to 13%) plants compared to control plants (Tables 1 and 2). Such   www.nature.com/scientificreports www.nature.com/scientificreports/ increases in lignin paralleled those of other overexpression approaches using MYB transcriptional activators of lignin biosynthesis. For instance, transgenic OsMYB42/85-OX rice plants accumulated a ~4% increase in thioglycolic acid lignin content of leaf blades 34 , whereas transgenic PvMYB42/85A-OX switchgrass plants accumulated on average a ~20% increase in ABSL lignin content in whole tillers 35 . Li et al. 40 demonstrated that both ABSL and Klason lignin content significantly increased by ~14% to 28% in stems of transgenic PtoMYB92-OX Populus plants, while Zhong et al. 33 did not report lignin content for the transgenic AtMYB85-OX Arabidopsis plants. Hussey et al. 41 postulated that TF overexpression might induce a phenotype within a limited range of overexpression, partly because of the limited amount of protein co-regulators. In our study, the alterations in lignin, p-CA and FA content did not correlate well with ZmMYB167 expression levels in maize plants (Table 2 and Fig. S4), which could imply a TF threshold beyond which there is no further induction of the phenylpropanoid pathway.
The overall abundance of cell wall polysaccharides in ZmMYB167 transgenic Brachypodium lines and ZmMYB167-OX maize plants remained largely unchanged (Table 1 and S2), suggesting ZmMYB167 specifically induces the biosynthesis of lignin and cell wall phenolics without modifying cellulose and hemicellulose biosynthesis. Indeed, Zhong et al. 33 demonstrated AtMYB85 to specifically induce GUS expression driven by the 4CL1 promoter (lignin), and not for the CesA8 (cellulose) and IRX9 (hemicellulose) promoters while Rao et al. 35 showed PvMYB42/85A to activate the COMT and F5H promoters, two major genes in lignin biosynthesis. Future studies will need to determine if ZmMYB167 directly binds to the AC elements ubiquitous in most promoters of lignin biosynthesis genes in monocots. In terms of TFs binding to their own gene promoters 28 , the expression levels of the endogenous ZmMYB167 and the ZmMYB17 syntelog did not appear to be affected by the overexpression of ZmMYB167 in transgenic maize plants (Fig. S5), suggesting that ZmMYB167 auto-regulation (i.e. ZmMYB167 binding to its own gene promoter) and ZmMYB17 cross-regulation is unlikely to have occurred in planta.
ZmMYB167 overexpression in transgenic maize plants had no impact on lignin composition and hence the S/G lignin monomer ratio ( Table 2) which can affect biomass recalcitrance 42 , indicative that the carbon flux of the phenylpropanoid pathway towards the biosynthesis of lignin monomers may have remained stable. Among the most distinctive traits of grasses are the phenolics p-CA and FA, participating in the composition, cross-linking and structural organisation of secondary cell walls. Concomitant with an increase in lignin content, the biomass of ZmMYB167-OX maize plants contained higher levels of cell wall-bound p-CA (~8% to 52%) and FA (~13% to 38%) relative to controls (Table 2), whereas the ZmMYB167 transgenic Brachypodium lines exhibited only higher levels of cell wall-bound p-CA (~15% to 24%) ( Table 1). These elevations in cell wall-bound phenolics are possibly related to the redirection of the metabolic flux along the phenylpropanoid pathway towards these phenolic intermediates. It may be that p-CA accumulation occurs in tandem with lignin deposition and could thus be a biochemical indicator to predict lignification 43 . Indeed, our data showed a positive correlation between ABSL and p-CA content (r = 0.78), and between ABSL and the p-CA: FA ratio (r = 0.71). Together, these results suggest that ZmMYB167 can fine-tune lignin biosynthesis pathway intermediates and thereby biomass composition.
Interestingly, the ZmMYB167-OX maize plants maintained similar growth phenotypes to controls (Table S1), suggesting that increases in the content of cell wall lignin and phenolics did not restrict the expansion of the cell wall during plant growth. In accordance, we also observed no adverse effect on biomass yield in the ZmMYB167-OX maize plants (Table S1), which could be relevant and beneficial from an agronomic and biorefining standpoint. Attempts to genetically alter lignin biosynthesis using MYB TFs are frequently accompanied by plant dwarfing or other developmental abnormalities. However, it remains unclear whether such undesirable phenotypic effects directly relate to MYB TF-induced changes in lignin deposition, hyperaccumulation of phenylpropanoid by-products or an indirect consequence of metabolic spill-over into developmental processes as well as signalling pathways involved in biotic and abiotic stress responses 27,44,45 . Although these TFs represent tools to modify lignin biosynthesis, conducting studies with tissue-specific promoters rather than constitutive promoters could address unintended pleiotropic effects in transgenic plants 46 . In this regard and in contrast to ZmMYB167-OX maize plants, heterologous expression of ZmMYB167 in Brachypodium led to a mild dwarf phenotype (Figs 2a and 3a). This phenotypic variation could be a result of different metabolic plasticity, intrinsically variable transcriptional regulatory circuits, changes in spatio-temporal expression of TFs, differences in cis-regulatory element composition of genes or protein-protein interactions controlling their distinct tissue organisation and patterning, cell wall formation and growth architecture [47][48][49][50] . The adverse phenotypic effects could also be due to the expression of ZmMYB167 in a heterologous system, with the Brachypodium orthologue BdMYB58 (Fig. 1a) showing 65% identity with ZmMYB167 (Fig. S1).
Transgenic approaches targeting MYB TFs have enriched our understanding of the regulatory mechanisms involved in secondary cell wall biosynthesis. Although the ultimate aim of such studies often is to tailor lignocellulose for improved processing and biorefinery application, the impact of transgenic interventions targeting the MYB clade of phenylpropanoid biosynthesis activators on biomass recalcitrance properties has received little attention (Table 3). Even if lignin is commonly referred to as one of the leading factors impeding enzymatic saccharification 51 , the glucose yields of untreated as well as alkaline pre-treated cell wall material from ZmMYB167-OX maize plants did not show significant alterations in biomass recalcitrance compared with that of control plants (Fig. 6). This indicates that the higher abundance of cell wall lignin and phenolics did not limit the accessibility of hydrolysing enzymes to matrix polysaccharides nor results in non-productive binding with hydrolysing enzymes and that these secondary wall components alone may not directly reveal the extent of biomass recalcitrance to enzymatic hydrolysis [52][53][54] . In fact, inhibition of enzymatic hydrolysis by phenolic compounds is more likely related to the presence of different phenolic functional groups 55 . In contrast, the glucose yields of untreated biomass from ZmMYB167 transgenic Brachypodium lines were decreased by ~20% on average (Fig. 4b) and may be explained by the increased S/G ratio (~32% to 53%), a key factor in determining biomass recalcitrance 42,56 that was not affected in transgenic maize plants (Table 2). Nonetheless, the variable saccharification efficiency observed for the ZmMYB167 transgenic C 3 and C 4 grasses continues to emphasise the complexity of www.nature.com/scientificreports www.nature.com/scientificreports/ secondary cell wall structures and limited fundamental understanding of how their properties collectively contribute towards biomass recalcitrance.
In summary, our findings highlight that ZmMYB167 expression levels can be modified to increase concentrations of lignin and cell wall-bound phenolics in grasses. We also demonstrate some of the potential challenges associated with MYB TF-based biomass engineering. Considering the economic and ecological importance of several perennial bioenergy grasses, more sophisticated strategies and functional analysis of additional TFs across grasses is needed to improve our understanding of which transcriptional regulatory genes are essential for controlling secondary cell wall biosynthesis and how alterations may impact lignocellulose quality, plant growth and fitness. Such knowledge is vital to help drive traditional plant breeding practices and biotechnological approaches for tailored and improved lignocellulosic biomass production.

Methods
Phylogenetic and protein motif analysis. Amino acid sequences were obtained by BLAST search of the NCBI database and analysed for a phylogenetic relationship via alignment using Clustal Omega 57 . Construction of a phylogenetic tree (Maximum likelihood method, Poisson correction model, bootstrap values of 1000) was done using the Molecular Evolutionary Genetics Analysis version 7.0 (MEGA7) program 58 . Protein sequence motifs were identified using the MEME (Multiple Expectation Maximisation for Motif Elicitation) program version 4.12.0 59 .

Construction of ZmMYB167 expression cassette and genetic transformation. ZmMYB167
(GRMZM2G037650; Zm00001d032032) was amplified from maize inbred line B73 cDNA and cloned into the Bb7m24GW destination vector 60,61 , containing the BAR selection marker, for overexpression of ZmMYB167 in maize. For heterologous expression in Brachypodium distachyon, ZmMYB167 was cloned into the pIPKb002 overexpression vector 62 , containing the hygromycin selection marker. For both vectors, expression of ZmMYB167 was under the control of the ZmUbi1 promoter. Overexpression constructs were introduced into Brachypodium inbred line Bd21-3 or maize Hi-II hybrid genotype (A188 X B73) by Agrobacterium-mediated transformation, using A. tumefaciens strain EHA108, or particle bombardment (transgenic maize line Zm3 only) as previously described [63][64][65] . Brachypodium and maize plants were selected for characterisation in the T 1 or F 1 generation respectively.
Plant material and growth conditions. Hi-II maize plants were grown in a glasshouse under standard conditions (16 h day; temperature range 22-26 °C; light intensity of 600 μmol m −2 s −1 ), and T 0 plants regeneration and F 1 seed germination were carried out per the Iowa State University "greenhouse care for transgenic maize plants" protocol. Backcrosses and development of F 1 segregating population were carried out as described by Scott (2013) 66 . All Brachypodium distachyon plants were grown in a transgenic glasshouse (16 h day; temperature range 21-23 °C and relative humidity range 40-43%) with a light intensity of 350 μmol m −2 s −1 .
Genomic DNA isolation and PCR analysis. Leaf tissue was harvested from maize (sixth leaf at V8 vegetative stage) and Brachypodium plants, frozen in liquid nitrogen and stored at −80 °C until use for genomic DNA (gDNA) isolation using the Qiagen DNAeasy 96 Plant Kit (Qiagen). Following extraction, gDNA concentration and quality were assessed using an Epoch Microplate Spectrophotometer (BioTEK). Gene-specific primers were designed using the NCBI primer designing tool to amplify (i) a DNA fragment covering the ZmUbi1 promoter region and the transgene-specific region and (ii) the entire DNA fragment of the transgene (Table S3). www.nature.com/scientificreports www.nature.com/scientificreports/ Mix (NEB) with the following thermal cycling steps: 95 °C for 2 mins, 35 cycles of 30 sec at 95 °C, 1 min at 60 °C, and 30 sec at 68 °C, final extension 68 °C for 5 mins. RT-PCR products were analysed on 1% agarose gels with S-adenosylmethionine decarboxylase or Peptidase C14 as reference genes for Brachypodium and maize respectively.

RNA isolation, RT-PCR and
Real-time PCR was performed on a LightCycler ® 480 II system (Roche). Relative quantitative analysis of gene expression was conducted using 1 μl of 2-fold diluted cDNA, 10 μl of SYBR Green I Master Mix, primer pairs and concentrations listed in Tables S4 and S5 and thermal cycling conditions as described above. Peptidase C14 was used as a reference gene 31 . The Roche Light Cycler ® 480 software 1.5 performed relative quantification and primer efficiency-corrected calculations. Data are expressed as means ± SE of at least three independent assays.
Preparation of cell wall residue. Stem biomass 1 cm above the seventh internode (IN7) from maize plants at vegetative stage 13 (V13) was harvested, prepared using the NREL LAP "Preparation of samples for compositional analysis" 67 and fractionated to an alcohol insoluble residue (AIR) 68 . Total aboveground leaf and stem biomass of fully senesced Brachypodium plants were sampled for preparation of AIR 69 .
Hydroxycinnamic acids and lignin content. The amounts of total hydroxycinnamic acid derivatives p-coumaric acid (p-CA) and ferulic acid (FA) was determined as described by Li et al. 70 . Acetyl bromide soluble lignin (ABSL) of AIR samples was quantified as described by da Costa et al. 68 . Extinction coefficients (g −1 L cm −1 ) of Brachypodium and maize for the ABSL method of lignin quantification were taken from Barnes and Anderson 71 .
Thioacidolysis of lignin. Thioacidolysis of AIR was performed as described by Foster et al. 69  Structural carbohydrates and Klason lignin. Analysis of cell wall carbohydrate content of Brachypodium AIR was performed as described by da Costa et al. 54 . Compositional analysis of maize AIR was based on the NREL LAP "Determination of structural carbohydrates and lignin in biomass" 72 . Structural carbohydrates were determined by HPAEC on a Dionex ICS-5000 system (Thermo Fischer Scientific) equipped with a pulsed amperometric detector (PAD) using the Dionex CarboPac SA10 column set at 45 °C and 1 mM KOH as eluent, with an eluent flow rate of 1.5 ml/min and 10 μl injection volume. Monosaccharide chromatograms were analysed and processed using the Chromeleon ™ 7.2 Chromatography Data System (CDS) software. The percentage of structural carbohydrates and Klason lignin was reported and calculated on an oven-dry weight and extractives-free basis.
Enzymatic hydrolysis and alkaline pre-treatment. Low solids enzymatic hydrolysis was carried out based on the NREL LAP "Low solids Enzymatic Saccharification of Lignocellulosic Biomass" 73 using Accellerase1500 enzymes (DuPont) at a dosage of 60 FPU/g cellulose. Soluble and enzyme-derived sugars were determined by HPAEC using conditions described above. Sugar yields were reported on an oven-dry weight basis, and the correction for hydration/water incorporated upon hydrolysis of cellulose to glucose monomers was applied. AIR samples were also subjected to a mild alkaline pre-treatment at 80 °C for 1 hr using an alkali loading of 0.08 g NaOH per gram of AIR 74 . Following pre-treatment, insoluble solids in pre-treated AIR were determined using the NREL LAP "Determination of Insoluble Solids in Pretreated Biomass Material" 75 . Enzymatic hydrolysis after alkaline pre-treatment was carried out using the NREL LAP "Low Solids Enzymatic Saccharification of Lignocellulosic Biomass". The data was plotted for glucose yield rather than glucose release to ensure that the results are not biased towards cell walls with higher glucan content, assuming 95% glucan recovery of pre-treated AIR post alkaline pre-treatment 76 . Histochemical staining of cellulose and lignin. Maize internode development was assessed and determined using the vegetative and reproductive stage identification system 77 . Maize and Brachypodium internode samples were collected in the glasshouse and stored in 70% EtOH at 4 °C until use. The middle portion of internode nine (IN9) from maize and IN1 of the 1st flowering tiller from Brachypodium plants were used as sectioning material. Transverse stem cross-sections were freehand-cut with a clean razor blade and stained with 0.01% (w/v) aqueous Calcofluor