Plant Biotechnology Journal (2015), pp. 1–12 doi: 10.1111/pbi.12316 Introduction of chemically labile substructures into Arabidopsis lignin through the use of LigD, the Ca-dehydrogenase from Sphingobium sp. strain SYK-6 Yukiko Tsuji1,‡, Ruben Vanholme2,3,‡, Yuki Tobimatsu4,5,†, Yasuyuki Ishikawa1, Clifton E. Foster5,6, Naofumi Kamimura7, Shojiro Hishiyama8, Saki Hashimoto1, Amiu Shino9, Hirofumi Hara10, Kanna Sato-Izawa1, Paula Oyarce2,3, Geert Goeminne2,3, Kris Morreel2,3, Jun Kikuchi9, Toshiyuki Takano11, Masao Fukuda7, Yoshihiro Katayama12, Wout Boerjan2,3, John Ralph4,5, Eiji Masai7 and Shinya Kajita1,* 1 Graduate School of Bio-Applications and Systems Engineering, Tokyo University of Agriculture and Technology, Tokyo, Japan 2 Department of Plant Biotechnology and Bioinformatics, Ghent University, Ghent, Belgium 3 Department of Plant Systems Biology, VIB, Ghent, Belgium 4 Department of Biochemistry, University of Wisconsin, Madison, WI, USA 5 US Department of Energy, Great Lakes Bioenergy Research Center, Wisconsin Energy Institute, Madison, WI, USA 6 Michigan State University, East Lansing, MI, USA 7 Department of Bioengineering, Nagaoka University of Technology, Niigata, Japan 8 Forestry and Forest Products Research Institute, Ibaraki, Japan 9 Center for Sustainable Resource Science, RIKEN, Kanagawa, Japan 10 Malaysia-Japan International Institute of Technology, Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 11 Graduate School of Agriculture, Kyoto University, Kyoto, Japan 12 College of Bioresource Sciences, Nihon University, Fujisawa, Japan Received 9 September 2014; revised 7 November 2014; accepted 25 November 2014. *Correspondence (Tel / fax +81 42 388 7391; email kajita@cc.tuat.ac.jp) † Present address: Graduate School of Agriculture, Kyoto University, Kitashirakawa-oiwakecho, Sakyo-ku, Kyoto, 606-8502, Japan. ‡ These authors contribute equally to this work. Keywords: Arabidopsis thaliana. Ca-dehydrogenase, lignin biosynthesis, NMR, Sphingobium sp. SYK-6. Summary Bacteria-derived enzymes that can modify specific lignin substructures are potential targets to engineer plants for better biomass processability. The Gram-negative bacterium Sphingobium sp. SYK-6 possesses a Ca-dehydrogenase (LigD) enzyme that has been shown to oxidize the a-hydroxy functionalities in b–O–4-linked dimers into a-keto analogues that are more chemically labile. Here, we show that recombinant LigD can oxidize an even wider range of b–O–4-linked dimers and oligomers, including the genuine dilignols, guaiacylglycerol-b-coniferyl alcohol ether and syringylglycerol-b-sinapyl alcohol ether. We explored the possibility of using LigD for biosynthetically engineering lignin by expressing the codon-optimized ligD gene in Arabidopsis thaliana. The ligD cDNA, with or without a signal peptide for apoplast targeting, has been successfully expressed, and LigD activity could be detected in the extracts of the transgenic plants. UPLC-MS/MS-based metabolite profiling indicated that levels of oxidized guaiacyl (G) b–O–4-coupled dilignols and analogues were significantly elevated in the LigD transgenic plants regardless of the signal peptide attachment to LigD. In parallel, 2D NMR analysis revealed a 2.1- to 2.8-fold increased level of G-type a-keto-b–O–4 linkages in cellulolytic enzyme lignins isolated from the stem cell walls of the LigD transgenic plants, indicating that the transformation was capable of altering lignin structure in the desired manner. Introduction Lignin is a complex polymer with phenylpropanoid units linked together by various carbon–oxygen and carbon–carbon bonds. It is deposited in plant secondary walls and plays important roles in mechanical support, water transport and stress responses. Although lignin is essential for plant growth and development, it negatively affects the use of lignocellulosic biomass. In many cases, lignin has to be removed to isolate cellulosic and noncellulosic polysaccharides, and this process requires a great deal of energy and chemicals. To reduce the lignin recalcitrance through a modification of lignin content and structure, down- or up-regulation of individual genes in the monolignol biosynthetic pathway has been extensively studied over the last two decades (Demura and Ye, 2010; Li et al., 2008; Vanholme et al., 2012; Zhao and Dixon, 2011). The pulping yield from transgenic plants, such as those deficient in cinnamyl alcohol dehydrogenase, cinnamoyl-CoA reductase or 4-coumarate:CoA ligase, may increase due to the modified lignin content and/or composition (Baucher et al., 2003; Lapierre et al., 2000; Kajita et al., 2002; Please cite this article as: Tsuji, Y., Vanholme, R., Tobimatsu, Y., Ishikawa, Y., Foster, C. E., Kamimura, N., Hishiyama, S., Hashimoto, S., Shino, A., Hara, H., Sato-Izawa, K., Oyarce, P., Goeminne, G., Morreel, K., Kikuchi, J., Takano, T., Fukuda, M., Katayama, Y., Boerjan, W., Ralph, J., Masai, E. and Kajita, S. (2015) Introduction of chemically labile substructures into Arabidopsis lignin through the use of LigD, the Ca-dehydrogenase from Sphingobium sp. strain SYK-6. Plant Biotechnol. J., doi: 10.1111/pbi.12316 ª 2015 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd 1 2 Yukiko Tsuji et al.
et al., 2007). Pilate et al., 2002; O’Connell et al., 2002; Leple Furthermore, quantitative and qualitative modifications of lignins could positively affect the enzymatic hydrolysis of cell wall polysaccharides for more efficient biofuel production from biomass (Chen and Dixon, 2007; Fu et al., 2011; Li et al., 2010; Van Acker et al., 2014). Genetic engineering of lignin, however, sometimes brings negative effects to plant growth and development such as stunted growth and abnormal plant morphology (Kajita et al., 2002; Voelker et al., 2011; Zhou et al., 2010; Bonawitz and Chapple, 2010). These negative influences might be reduced and diminished via fine-tuning of the lignin biosynthetic pathway at spatial and temporal levels or via finetuning the responses invoked by shifts in the pathway (Bonawitz et al., 2014; Eudes et al., 2012; Yang et al., 2012). Lignin is derived from the oxidative polymerization of mainly pcoumaryl, coniferyl and sinapyl alcohols, collectively known as monolignols. Monolignols are synthesized in the cytosol from phenylalanine through the general phenylpropanoid and monolignol-specific pathways. These monomers are transported to the apoplastic space and then polymerized by the action of phenol oxidases such as laccases and peroxidases. The three monolignols give rise to p-hydroxyphenyl (H), guaiacyl (G) and syringyl (S) lignin subunits after their incorporation into the polymer (Boerjan et al., 2003; Ralph et al., 2004b). Subunit composition differs among plant species and is also controlled within a plant in a spatial and temporal manner. Based on numerous structural studies of lignins in both wild-type and transgenic plants, incorporation of substantial amounts of nonclassical monolignols into lignin is prominent (Ralph et al., 2004b; Vanholme et al., 2012). Various c-acylated lignin substructures derived from c-acylated monolignols, for example monolignol acetates, p-hydroxybenzoates and p-coumarates, are integral components of lignins in many plant species (Del Rio et al., 2007; Lu and Ralph, 2008; Ralph et al., 2004a; Smith, 1955). The natural diversity of lignification is further illustrated by recent discoveries of lignins derived solely from caffeyl alcohol or 5-hydroxyconiferyl alcohol present in the seed coats of several monocot and dicot plants (Chen et al., 2012; Tobimatsu et al., 2013). In transgenic plants, modified expression of monolignol biosynthetic pathway genes can also lead to unusual lignins incorporating pathway OH (a) OH OMe R O RO HO HO OMe GGE (R = H, erythro) VGE (R = Me, erythro) GGE-glc (R = β-D-glucose, erythro) PDA response (a.u.) (c) α-Keto-GGE OMe intermediates and/or other phenolic metabolites derived from them (Boerjan et al., 2003; Vanholme et al., 2010a,b; Weng et al., 2010). The plasticity of lignification even allows manipulation of lignin polymerization and structures through the introduction of exogenes targeting essentially new lignin precursors, as conceptualized a while ago (Grabber et al., 2008; Ralph, 2010; Vanholme et al., 2012) and recently demonstrated with a few successful studies (Zhang et al., 2012; Eudes et al., 2012; Wilkerson et al., 2014). The most abundant linkage unit in typical native dicot lignin is the b-aryl ether (b–O–4) unit, which can make up to 90% of the total units (Figure 1). The benzylic a-positions of b–O–4 units are usually hydroxy-substituted. The a-keto-b–O–4 units, with carbonyl groups at the benzylic positions, can also be found in € € natural lignins at quite low levels (Amm alahti et al., 1998; Lu and Ralph, 1998; Marita et al., 1999). These minor subunits most likely arise from nonenzymatic postpolymerization oxidations of the typical a-hydroxy b–O–4 units. Such a-keto-b–O–4 units can be cleaved under alkaline and/or oxidative conditions more easily and faster than the typical b–O–4 units with benzylic hydroxyl groups (Criss et al., 1998; Gierer and Ljunggren, 1979; Gierer et al., 1980; Gierer and Nor
en, 1982; Imai et al., 2007; Rahimi et al., 2014). Thus, increase of a-keto-b–O–4 units over the typical a-hydroxy-b–O–4 units in the lignin backbone should contribute to reducing the cost and energy penalty in chemical pulping and biomass pretreatment processes for cellulosic ethanol production. Although postharvest chemical processes for Caoxidation have long been pursued (Gierer and Nor
en, 1982; Rahimi et al., 2013; Shiraishi et al., 2013), a biotechnological approach could be an alternative strategy. Ca-dehydrogenase (LigD) from the Sphingobium sp. SYK-6 strain (also known as Pseudomonas paucimobilis SYK-6) is one of the best-characterized lignin degrading enzymes (Gall et al., 2014; Masai et al., 1993, 2007; Sato et al., 2009). This enzyme catalyses the benzylic oxidation of the (R)-isomer of the b–O–4type model compound guaiacylglycerol-b-guaiacyl ether (GGE, Figure 1a), during which NAD+ is reduced to NADH (Masai et al., 1993; Sato et al., 2009; Reiter et al., 2013). As lignin polymerization is initiated through oxidative coupling of monolignol radicals solely under chemical control, the resultant b–O–4 O (b) O HO OMe OH R RO VGE R HO HO OMe α-Keto-GGE-glc GGE-glc OMe O HO OMe α-Keto-GGE (R = H) α-Keto-VGE (R = Me) α-Keto-GGE-glc (R = β-D-glucose) G(β–O–4)G (R = H, erythro/threo) S(β–O–4)S (R = OMe, erythro/threo) GGE O OMe O Gox(β–O–4)G G(β–O–4)G OH R Gox(β–O–4)G (R = H) Sox(β–O–4)S (R = OMe) S(β–O–4)S Sox(β–O–4)S α-Keto-VGE With LigD With LigD With LigD With LigD With LigD Without LigD Without LigD Without LigD Without LigD Without LigD 1.0 2.0 3.0 Retention time (min) 2.0 3.0 4.0 Retention time (min) 1.0 1.5 2.0 2.5 Retention time (min) 1.0 1.5 2.0 2.5 Retention time (min) 1.5 2.0 2.5 Retention time (min) Figure 1 LigD activities against synthetic b–O–4-linked lignin dimers. Recombinant LigD, produced in E. coli, can convert various synthetic lignin dimers (a) to their corresponding a-keto dimers (b). The production of a-keto dimers by the recombinant LigD was determined (c) by UPLC equipped with PDA and ESI-MS detectors. The controls without LigD used crude extracts of E. coli transformed with empty vector pET-21a(+). Peak assignments were based on MS analysis and data of authentic standards (see the main text and supplemental information, Figure S1). ª 2015 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–12 Introduction of chemically labile units into lignin 3 dilignols and subsequent elongating oligolignols are racemic, that is, they are composed of compounds with (R)- and (S)-isomers. Therefore, LigD could in theory oxidize the a-hydroxyl groups in 50% of the b–O–4 units. Indeed, LigD, together with a b-etherase (LigF) and a glutathione lyase (e.g. LigG), was able to depolymerize synthetic lignin and hardwood alkali lignin, giving indirect evidence for the Ca-oxidation of b–O–4 units in lignin polymers by LigD (Sonoki et al., 2002; Reiter et al., 2013). Thus, assuming that sufficient NAD+ is supplied to the reaction site, LigD has the potential to oxidize a-hydroxyl groups in b–O–4 units in lignin dimers, oligomers and polymers in vivo. In this context, we hereby further investigated the substrate preference of LigD towards various types of b–O–4-linked lignin dimers and oligomers. In addition, to evaluate the use of LigD as a tool for introducing a-keto-b–O–4 units into lignins in planta, transgenic Arabidopsis plants expressing the gene for LigD (ligD), with or without an apoplast-targeting signal (ATS) peptide, were generated, and the changes in the phenolic metabolites and cell wall structures were delineated by UPLC-MS/MS and 2D NMR approaches. The potential for, and challenges to, implementing LigD strategies to engineer lignins for better biomass processability is discussed. Results LigD is active on b–O–4-linked lignin dimers and oligomers The a-oxidative activity of LigD on naturally occurring b–O–4linked units has been assumed based on the oxidation of the artificial substrate GGE (Masai et al., 1993; Sato et al., 2009; Reiter et al., 2013). To test whether LigD is also active on native plant metabolites and lignin, the enzymatic extracts prepared from recombinant Escherichia coli harbouring the ligD gene were reacted with different dimeric and oligomeric substrates containing b–O–4 linkages. The two authentic dilignols, guaiacylglycerol-b-coniferyl alcohol ether [G(b–O–4)G] and syringylglycerol-b-sinapyl alcohol ether [S(b–O–4)S] could be converted to their corresponding aketo derivatives by the recombinant LigD (Figures 1 and S1). The enzyme can also act on other artificial dimers derived from GGE, such as erythro-veratrylglycerol-b-guaiacyl ether (VGE) and erythro-GGE-b-D-glucoside (GGE-glc) (Figure 1a), in which the free phenolic hydroxyl group of GGE is etherified by methyl or glucosyl residues (Figures 1 and S1). The control samples prepared from a crude extract of E. coli transformed with empty vector pET-21a(+) did not show a-oxidative activity on these substrates. To test whether LigD could also act on substrates with longer chains of b–O–4-linked units, the recombinant LigD was incubated with synthetic b–O–4-linked lignin oligomers that were synthesized by an aldol-condensation oligomerization approach (Katahira et al., 2006) that included a pentamer, 4–O–2-hydroxyethyl-G(b–O–4)G(b–O–4)G(b–O–4)G (b–O–4)-vanillyl alcohol, as a main component. After incubation, UPLC-MS/MS detected the singly and doubly oxidized pentamers, that is 4–O–2-hydroxyethyl-G(b–O–-4)G(b–O–4)G(b–O–4)Gox(b– O–4)- and 4–O–2-hydroxyethyl-G(b–O–4)Gox(b–O–4)G(b–O–4)Gox(b–O–4)-vanillyl alcohols (Figure 2). These results clearly suggest that the recombinant LigD has the ability to introduce a-keto functionalities into the middle of b–O–4 unit chains of oligomeric lignin substrates. Collectively, these results demonstrate promiscuous catalytic activity of LigD on various types of b– O–4-linked dilignol and oligolignol substrates. Generation of Arabidopsis plants expressing LigD To test the consequences of in planta LigD activity, two different ligD overexpression constructs were designed for transformation into Arabidopsis, one construct for cytosolic localization and one for apoplastic localization of LigD protein (Figure S2). In both constructs, the codon of ligD cDNA was optimized according to the codon preference of A. thaliana using a codon usage database (www.kazusa.or.jp/codon/) and the resultant sequences were put under the control of the cauliflower mosaic virus 35S promoter. The 57-nucleotide-long sequence coding for ATS was derived from the INVERTASE INHIBITOR gene of tobacco (Greiner et al., 1998). To verify that the ATS can target the protein to cell wall, we also generated an additional construct for expression of YFP with ATS (sYFP). Although the ATS was shown to help in localizing a YFP protein to the cell walls of tobacco BY-2 cells, there was still a relatively high signal in the cytoplasm (Figure S3). Nine and six independent homozygous ligD overexpression lines without the ATS (designated as D lines) and with the ATS (designated as SD lines) were selected. Although no visible differences in plant growth and development were seen (Figure S4), semi-quantitative RT-PCR (sqRT-PCR) indicated that ligD was expressed in all the D and SD lines with various levels of expression (Figures S5a,b). Western blot analysis with an antibody raised against an oligopeptide derived from LigD showed diagnostic bands corresponding to LigD particularly in the inflorescence stem extracts prepared from five transgenic lines, D5, D8, D9, SD1 and SD9, whereas no diagnostic band was found in the wild-type line (Figure S5c). LigD activity was also detected in the crude protein extracts prepared from inflorescence stems of all five lines (Figure S6), further confirming the successful transformation and in planta expression of ligD in these transgenic lines. The extractable LigD activities detected in the SD lines were significantly lower than those detected in the D lines. The result may suggest that the ATS attachment inhibits the activity of LigD. Phenolic profiling of D and SD lines To determine the consequences of in vivo LigD activity on phenolic metabolism, methanol-soluble phenolics prepared from inflorescence stems of four selected transgenic lines, D5, D8, SD1 and SD9, were subjected to UPLC-MS phenolic profiling (Morreel et al., 2004, 2010) and compared with those of two control lines: wild-type and sYFP-transformed lines. The abundances of 27 compounds were higher in both D5 and D8 as compared to wild type and sYFP, and the abundances of 24 compounds were higher in both SD1 and SD9 lines as compared to both control lines (Table 1, Figure 3). The 24 compounds (1–24) had a higher abundance in both the D and SD lines. The three additional compounds (25–27) for which the abundance was found to be significantly higher in D lines also appeared significantly higher in SD9 lines but not in SD1 (Table 1). Remarkably, the order of magnitude of the increase was comparable across SD and D lines. These observations indicate that the influence of the ATS was minimal. Twenty-three of the twenty-seven compounds were structurally characterized via MS/MS analysis (Figures S7 and S8). In agreement with the in vitro activity of LigD, Gox(b–O–4)G (1) was clearly detected in the D and SD lines, whereas the compound was below detection limit in wild-type plants. In addition, also a-keto-guaiacylglycerol-b-coniferaldehyde ether (Gox(b-O-4)G0 , 2), and Gox 4-O-hexoside (b-O-4)G (3) accumulated in D and ª 2015 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–12 4 Yukiko Tsuji et al. (a) 4-O-2-hydroxyethyl G( -O-4)G( -O-4)G( -O-4)G( -O-4)vanillyl alcohol 50 m/z 391.140 OH OH m/z 195.066 CHOO- m/z 1027.382 0.060 Oligomer without LigD Oligomer with LigD O HO MeO O HO MeO O MeO O HO MeO O HO MeO HO m/z 981.376 m/z 741.276 OH OH OH m/z 545.202 m/z 349.128 m/z 153.055 m/z 631.240 m/z 827.313 981.381 100 165.052 195.065 545.202 -H2O -CH2O -H2O -CH2O 391.139 153.055 150.031 % Relative intensity m/z 587.213 MS/MS m/z 1027.382 100 741.280 -H2O -CH2O 301.111 349.130 343.121 497.187 587.223 631.235 827.325 693.246 723.242 0 798.301 1027.429 0 8 9 10 11 12 13 14 100 15 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 m/z 1050 Time (b) 4-O-2-hydroxyethyl G( -O-4)G( -O-4)G( -O-4)Gox( -O-4)vanillyl alcohol m/z 389.124 O HO MeO O HO MeO O MeO O HO MeO O HO MeO HO OH m/z 979.361 50 OH m/z 195.066 CHOO- m/z 739.260 -H2O 347.115 165.052 150.033 0 195.059 OH m/z 543.187 317.104 329.101 100 % Relative intensity m/z 585.198 MS/MS m/z 1025.370 m/z 1025.370 0.060 Oligomer without LigD Oligomer with LigD 100 OH m/z 347.113 -CH2O -CH2O -CH2O 301.104 961.351 949.364 -H2O 721.248 709.251 543.191 585.198-CH2O 739.257 545.200 389.124-CH2O 359.119 -H2O -H2O -CH2O 513.177 525.171 979.413 795.260 0 8 9 10 11 12 13 14 100 15 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 m/z 1050 Time (c) 4-O-2-hydroxyethyl G( -O-4)Gox( -O-4)G( -O-4)Gox( -O-4)vanillyl alcohol MS/MS m/z 1023.351 m/z 1023.351 0.060 Oligomer without LigD Oligomer with LigD 50 m/z 195.066 CHOOO HO MeO O HO MeO O MeO O HO MeO O HO MeO HO m/z 977.345 -CH2O 329.116 525.188 -H2O -H2O 153.064 m/z 347.113 OH OH m/z 153.055 -H2O -CH2O -CH2O -CH2O 543.196 544.193 347.106 0 m/z 543.187 513.181 317.110 100 m/z 737.245 OH % Relative intensity m/z 583.182 m/z 389.124 100 389.124-CH2O 359.118 -H2O 977.345-CH2O-H2O 929.319 707.242 689.203 719.249 917.337 737.212 195.065 977.345-CH2O 947.354 1023.388 0 8 9 10 11 12 13 14 15 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 m/z 1050 Time Figure 2 LigD activity on a b–O–4-type oligomer model. (a) MS/MS spectrum of 4–O–2-hydroxyethyl G(b–O–4)G(b–O–4)G(b–O–4)G(b–O–4)vanillyl alcohol and selected ion chromatogram of the compound reacted with (red) and without (black) LigD. (b) MS/MS spectrum of single oxidized 4–O–2hydroxyethyl G(b–O–4)G(b–O–4)G(b–O–4)G(b–O–4)vanillyl alcohol and selected-ion chromatogram of the compound reacted with (red) and without (black) LigD. (c) MS/MS spectrum of double oxidized 4–O–2-hydroxyethyl G(b–O–4)G(b–O–4)G(b–O–4)G(b–O–4)vanillyl alcohol and selected-ion chromatogram of the compound reacted with (red) and without (black) LigD. Single and double oxidized 4–O–2-hydroxyethyl G(b–O–4)G(b–O–4)G(b–O–4) G(b–O–4)vanillyl alcohol was detected in the samples (n = 4) treated with crude E. coli protein extract containing LigD, but not in those treated with crude protein extracts from transformed E. coli with empty vector pET-21a(+) (n = 3). SD lines. However, most of the compounds elevated in the transgenic lines were derivatives of a-keto-guaiacylglycerol-bferulic acid ether (Gox(b–O–4)ferulic acid, compounds 4–20). These compounds are likely derived from guaiacylglycerol-bferulic acid ether (G(b–O–4)ferulic acid) via a-oxidation catalysed by LigD. Also, two malated a-keto-guaiacylglycerol-b-sinapic acid ethers (Gox(b–O–4)sinapoyl malate, 21 and 22) were significantly higher in abundance. In contrast to the increased levels of the compounds noted above, the abundances of seven and six compounds were significantly lower in both D5 and D8, and both SD1 and SD8 compared to the wild-type, respectively (Figure 3 and Table 1). Again, there appeared a large overlap between the two lists of differentials: five compounds (28–32) were significantly lower in both D and SD lines, as compared to both control lines. Also the abundances of the three remaining compounds (33–35) appeared to be lower in both S and SD, but were not found to be significant in every comparison with the control lines, and were therefore missed by the selection criteria. These observations are again indicative of a minimal influence of the ATS. Five of the compounds with decreased levels in D and SD lines (28–31 and 34) were structurally characterized as malated derivatives of G(b–O–4)ferulic acid, and one compound (33) was characterized as a malated derivative of G(b–O–4)sinapic acid (33, Figure S8). These compounds are possible in vivo substrates of LigD, and their decrease in D and SD lines would nicely explain the increase of the oxidation products Gox(b–O–4)ferulic acid and Gox(b–O–4)sinapic acid derivatives. Also, the abundance of the hexosylated dilignol G 4-O-hexoside (b–O–4)S (35) was found to be significantly deceased in SD lines (Table 1). Unfortunately, a targeted search showed that the known plant metabolites G(b– O–4)G, G(b–O–4)S and S(b–O–4)S remained below the detection limit in all samples tested. In addition, potential products such as Gox(b–O–4)S, Sox(b–O–4)S and trimers with a-keto-b–O–4 units were also below the detection limit in the D and SD lines. Cell wall composition and lignin structure Cell wall composition of senesced inflorescence stems of D5, D8, SD1 and SD9 lines was characterized by wet-chemical and 2D NMR methods and compared with those from wild-type and sYFP-transformed control lines. Crystalline cellulose, the composition of matrix polysaccharides (hemicelluloses and pectins) and acetyl bromide-soluble lignins were overall similar among all the plants analysed (Table S1). For further in-depth analysis of cell ª 2015 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–12 Introduction of chemically labile units into lignin 5 Table 1 Identities of significantly differential metabolites and their peak areas in D5, D8, SD1 and SD9 transgenics and in wild-type and sYFPtransformed control lines Ret. nr m/z Time (min) Identity Wild type sYFP D5 D8 SD1 SD9 average average average average average average (s.d.) (s.d.) (s.d.) (s.d.) (s.d.) (s.d.) Significant up in both D and SD 1 373.1278 11.76 Gox(b-O-4)G 79 (22) 106 (46) 13479 (2940) 11124 (2209) 5981 (1256) 2 371.1123 14.12 Gox(b-O-4)G’ 89 (22) 129 (30) 6777 (1517) 5600 (1530) 3201 (907) 12365 (1975) 3594 (933) 3 581.1877 8.34 Gox 4-O-hexoside(b-O-4)G 11 (4) 82 (33) 10204 (2120) 6517 (1406) 9447 (1894) 12883 (2098) 4 549.1612 8.64 Gox(b-O-4)feruloyl hexose 236 (82) 628 (162) 23943 (2204) 17484 (1229) 16737 (1859) 22688 (2083) 5 549.1616 9.14 Gox(b-O-4)feruloyl hexose 12 (10) 2 (2) 6088 (2067) 3600 (342) 4111 (962) 5324 (944) 6 549.1609 9.67 Gox(b-O-4)feruloyl hexose 217 (50) 378 (85) 19633 (1098) 16470 (1145) 16094 (1182) 19768 (749) 7 549.1622 10.04 Gox(b-O-4)feruloyl hexose 4 (2) 5 (5) 4750 (580) 3534 (307) 2208 (479) 3363 (551) 8 773.2332 13.95 Gox(b-O-4)feruloyl hexose 59 (26) 329 (150) 29057 (4902) 15525 (3095) 11229 (2781) 26722 (4728) 9 773.2329 15.27 Gox(b-O-4)feruloyl hexose 0 (0) 3 (3) 11238 (1587) 5899 (1112) 4461 (1093) 8797 (1096) 13 (13) 93 (63) 23544 (4315) 14528 (3882) 9121 (2538) 14979 (3143) 494 (167) 41 (23) 7821 (1357) 4533 (1340) 3195 (921) 4138 (962) 0 (0) 8 (5) 8185 (1504) 4549 (1421) 2634 (766) 5599 (1184) (formic acid adduct) + 224 Da + 224 Da 10 935.2910 11.09 Gox(b-O-4)feruloyl hexose +hexose + 224 Da 11 935.2999 12.45 Gox(b-O-4)feruloyl hexose +hexose + 224 Da 12 935.2922 11.66 Gox c-O-hexoside(b-O-4)feruloyl hexose + 224 Da 13 503.1187 12.14 Gox(b-O-4)feruloyl malate 1308 (262) 547 (76) 36388 (3098) 36622 (1573) 29377 (2873) 33927 (3087) 14 503.1187 12.76 Gox(b-O-4)feruloyl malate 768 (179) 485 (85) 25238 (1838) 27122 (1715) 22373 (2034) 24003 (1460) 15 665.1724 9.03 Gox(b-O-4)feruloyl malate 12 (12) 10 (7) 10728 (1198) 8437 (555) 8041 (1695) 7864 (1067) 187 (83) 456 (195) 27180 (2151) 22506 (1684) 18999 (3703) 22297 (2658) 10 (5) 26 (15) 4979 (434) 4769 (550) 4317 (951) 4455 (589) 24 (15) 40 (16) 10378 (1352) 10573 (889) 7531 (1731) 10133 (1404) 771 (443) 282 (188) 5102 (400) 5089 (511) 3931 (904) 5034 (463) + hexose 16 665.1727 9.16 Gox(b-O-4)feruloyl malate + hexose 17 665.1732 9.47 Gox(b-O-4)feruloyl malate + hexose 18 665.1734 10.11 Gox c-O-hexoside(b-O-4)feruloyl malate 19 665.1721 11.01 Gox c-O-hexoside(b-O-4)feruloyl malate 20 516.1510 10.08 Gox(b-O-4)feruloyl glutamate 21 417.1210 12.22 Gox(b-O-4)sinapoyl malate 22 417.1182 13.14 Gox(b-O-4)sinapoyl malate 6 (6) 7 (3) 7751 (590) 8468 (953) 6163 (468) 8505 (911) 584 (157) 518 (100) 3480 (436) 3337 (476) 3410 (606) 4299 (658) 353 (135) 289 (86) 7018 (827) 6742 (764) 5193 (747) 7843 (1102) 2 (2) 2 (2) 4672 (1815) 4911 (1429) 705 (435) 4836 (1177) 12 (6) 5 (4) 3469 (537) 4282 (949) 2147 (346) 3550 (537) (-malate) (-malate) 23 545.1301 8.17 Unknown C26H25O13 24 665.1739 7.32 Unknown Significant up in D only 25 357.1133 4.47 Unknown C16H21O9 2531 (284) 2938 (278) 4777 (553) 5770 (646) 3155 (438) 3965 (317) 26 667.1892 6.86 G(b-O-4)feruloyl malate 2369 (381) 3206 (302) 5170 (322) 4568 (302) 3377 (539) 4856 (544) 27 535.1750 9.39 Unknown 1580 (172) 2020 (219) 4202 (463) 3973 (363) 2658 (285) 4661 (365) + hexose Significant down in D and SD 28 505.1339 9.39 G(b-O-4)feruloyl malate 37291 (3960) 30115 (1036) 21092 (1349) 22361 (1019) 20364 (2056) 22972 (1197) 29 505.1340 9.64 G(b-O-4)feruloyl malate 20196 (2483) 16977 (760) 12227 (1006) 12792 (267) 11766 (1549) 13122 (1010) 30 505.1347 10.24 G(b-O-4)feruloyl malate 16428 (2210) 12600 (693) 8936 (518) 9638 (728) 8808 (688) 10221 (344) 31 505.1344 10.44 G(b-O-4)feruloyl malate 17586 (2157) 14729 (907) 11019 (595) 12025 (774) 10526 (923) 12309 (542) 32 383.0967 6.83 9189 (850) 8630 (234) 5271 (453) 5232 (145) 4878 (636) 6167 (417) Ferulic acid + 74 Da + malate Significant down in D only 33 535.1466 9.13 G(b-O-4)sinapoyl malate 6384 (1032) 3931 (254) 2547 (305) 2212 (202) 2599 (520) 3238 (276) 34 701.2108 11.91 G(b-O-4)feruloyl malate 6173 (885) 7791 (1255) 2619 (412) 2191 (424) 2041 (464) 4749 (987) + 196 Da ª 2015 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–12 6 Yukiko Tsuji et al. Table 1 Continued Ret. nr m/z Time (min) Identity Wild type sYFP D5 D8 SD1 SD9 average average average average average average (s.d.) (s.d.) (s.d.) (s.d.) (s.d.) (s.d.) Significant down in SD only 35 613.2141 7.67 G 4-O-hexoside(b-O-4)S 3982 (788) 6482 (1061) 3092 (500) 1970 (472) 1525 (433) 1954 (367) (formic acid adduct) The corresponding structures and structural elucidation are given in Figure 3 and Figure S7, respectively. (a) Significantly increased compounds in both D and SD lines SD1 and SD9 D5 and D8 O C OMe MeO RO 0 24 3 O O MeO OH HO HO O HO O HO 1 R=H 3 R=Hex 2 O OMe MeO OMe O O O HO HO MeO OR2 R1O OMe MeO O Malate O O 21, 22 4-7 R1=H, R2=Hex 8, 9 R1=H, R2=Hex + 224 Da 10, 11 R1=H, R2=Hex + Hex + 224 Da 12 R1=Hex, R2=Hex + 224 Da 13, 14 R1=H, R2=malate 15-17 R1=H, R2=malate + Hex 18-19 R1=Hex, R2=malate 20 R1=H, R2=glutamate (b) Significantly decreased compounds in both D and SD lines D5 and D8 SD1 and SD9 2 OH HO 1 OMe OH O MeO 5 Malate O HO HO MeO O HO O 28-31 34 + 196 Da 33 OH MeO OMe MeO Malate O O OMe O Figure 3 Chemical structures of metabolites that are increased (a) or decreased (b) both in D and SD lines (Table 1). Figure 4 Aromatic subregions of short-range 1H–13C correlation (HSQC) NMR spectra of enzyme lignins (ELs) isolated from stem cell walls of D5, D8, SD1 and SD9 transgenics and wild-type and sYFP-transformed control lines. Volume integrals are given for the lignin aromatic units that are colour-coded to match their assignments in the spectrum. Box with 92 indicates regions that were vertically scaled twofold. wall lignins by NMR, enzyme lignins (ELs) were isolated via digestion of the stem cell walls by cellulases, leaving all of the lignin and minimum amounts of residual polysaccharides (Bonawitz et al., 2014; Wagner et al., 2011, 2013). The isolated ELs were then acetylated and subjected to solution-state 2D 1H–13C short-range correlation (HSQC) analysis. The aromatic regions of HSQC spectra reveal the lignin monomeric composition (Figure 4), whereas the aliphatic regions reveal the lignin interunit linkages (Figure 5). The relative contribution of the lignin substructures was deduced from the volume integrals of relevant HSQC contour peaks (Table S2). Overall, our NMR analyses revealed that the lignins in all the transgenic lines expressing ligD were very similar to those from the wild-type and sYFP control lines. The S/G ratios were between 0.27 and 0.31, showing that the lignins were relatively rich in G units (Figure 4, Table S2). In addition, about 75% of the linkage groups were Hex O HO MeO OH 35 ª 2015 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–12 Introduction of chemically labile units into lignin 7 Bβ Wild type A: B: C: D: E: A′/A: Bβ D5 Cβ 75.1% 15.1% 7.7% 0.8% 1.3% 0.15% X1 γ A: B: C: D: E: A′/A: Eβ Aγ (+ Dγ, Eγ...) Bγ Cγ Aα Cγ Eβ′ A’ β A’ β Bβ A’ β Eβ Cγ 6.0 Cγ A’ β Eβ 60 X1γ Cγ Aα 5.0 4.0 3.0 A’ β 5.0 80 Eα Dβ Dα Bα Cα (+ Eα′...) 6.0 70 Eβ′ Dβ Bα Cα (+ Eα′...) Cγ Aβ X2 Dα Aγ (+ Dγ, Eγ...) Bγ Eα Dβ 50 Cβ Eβ′ Eα Bα 75.3% 15.3% 7.4% 0.7% 1.3% 0.30% Aβ X2 Dα Bβ A: B: C: D: E: A′/A: Aγ (+ Dγ, Eγ...) Cγ Eβ′ A’ β Cα (+ Eα′...) SD9 Eβ Aα Aβ X2 Bα Cβ Bγ Cγ Dβ Dα 75.8% 14.8% 7.6% 0.8% 1.0% 0.31% X1 γ Aγ (+ Dγ, Eγ...) Bγ 80 Eα Cα (+ Eα′...) Bβ A: B: C: D: E: A′/A: 70 Eβ′ SD1 Cβ Cγ Aβ X2 Dα 74.5% 15.6% 7.6% 1.0% 1.3% 0.13% X1 γ Aα Cγ Dβ Bα Aγ (+ Dγ, Eγ...) Aα Eα Cα (+ Eα′...) sYFP A: B: C: D: E: A′/A: 60 Eβ′ Dβ Dα Eβ Bγ Aβ X2 Eα Bα Cγ 50 Cβ 74.7% 15.4% 7.7% 0 .9 % 1.3% 0.32% X1 γ Aγ (+ Dγ, Eγ...) Aα Aβ X2 A: B: C: D: E: A′/A: Eβ Bγ Cγ Bβ D8 Cβ 75.5% 15.1% 7.3% 0.8% 1.3% 0.37% X1 γ 4.0 3.0 6.0 Cα (+ Eα′...) 5.0 4.0 13 3.0 1 C ppm H 5 HO HO α 5 5 γ HO β O 4 γ β γ α O MeO β α O OMe A β-Aryl ether β–O–4 O β B Phenylcoumaran β–5 C Resinol β–β O O α β OMe OH D Dibenzodioxocin 5–5/β–O–4 α HO β O α OH γ OMe β O α β HO O γ β α 1 OH OMe O4 OMe Methoxyl Not assigned O carbohydrates, A’ E X1 Spirodienone Cinnamyl alcohol α-Keto-β-aryl ether solvents, etc. β–O–4 β–1 end-units Figure 5 Aliphatic subregions of short-range 1H–13C correlation (HSQC) NMR spectra of enzyme lignins (ELs) isolated from stem cell walls of D5, D8, SD1 and SD9 transgenics and wild-type and sYFP-transformed control lines. Volume integrals based on three biological replicates are given for the lignin aromatic units that are colour-coded to match their assignments in the spectrum. Box with 92 indicates regions that were vertically scaled twofold. b-aryl ether units (A), about 15% phenylcoumaran units (B) and about 7% resinol units (C) (Figure 5, Table S2). Dibenzodioxocin (D) and spirodienone (E) units made up around 1% each of the total units (Figure 5, Table S2). The signals from a-keto lignin substructures appeared in the regions separated from where the typical a-hydroxyl lignin signals appeared (Marita et al., 1999; Rahimi et al., 2013; Stewart et al., 2009). The relative levels of the oxidized versus nonoxidized lignin units appeared to be very low (a few %) in all the transgenic lines characterized. However, our NMR analysis revealed that the anticipated a-keto lignin units were indeed significantly elevated in all the transgenic lines compared to the wild-type and sYFP control lines. First, the aromatic HSQC region displayed a 1.6- to 2.6-fold increase in the intensities from a-keto G units (G0 ) in D and SD lines, whereas there were no significant increases in the signals from a-keto S units (S0 ) (Figure 4, Table S2). Notably, these aromatic signals may over-estimate the a-keto lignin units because of the possible overlapping presence of other naturally occurring oxidized lignin subunits such as benzaldehydes and benzoic acid end-units (Rahimi et al., 2013). Second, the aliphatic HSQC region revealed a 2.1- to 2.8-fold increase in a-keto b-aryl ether units (A0 ) (Stewart et al., 2009) in D and SD lines, compared with the wild-type and sYFP control lines (Figure 5, Table S2). Collectively, these results support our contention that expressing the ligD indeed augments the a-keto lignin units in the lignin polymer, albeit at fairly low levels. To examine effect of the increase in a-keto lignin units on enzyme saccharification efficiency, we performed enzymatic digestion of pulverized stem tissues from each plant line with or without alkaline pretreatment (Figure S9). However, no general significant differences could be detected between the transgenic (D5, D8, SD1 and SD9) and control lines (wild type and sYFP). Discussion LigD catalyses the oxidation of genuine b–O–4-linked dilignols and oligolignols In previous studies, LigD has been characterized as an enantioselective catalyst for the Ca-oxidation of GGE, a b–O–4-type model compound (Figure 1a) (Gall et al., 2014; Masai et al., 1993; Sato et al., 2009; Reiter et al., 2013). However, the use of LigD as an enzyme to engineer lignin in planta would only be plausible if it recognizes native oligolignols and/or lignin as substrates. In the present study, we demonstrated that LigD ª 2015 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–12 8 Yukiko Tsuji et al. oxidizes genuine dilignols, that is G(b–O–4)G and S(b–O–4)S, as well as the etherified GGE derivatives such as VGE and GGE-glc, into their respective a-keto products (Figure 1 and supplemental Figure S1). These results indicate that the enzyme tolerates variations in the side chains and aromatic ring phenolic substitution in the b–O–4-type substrates. In addition, we showed that LigD catalyses the oxidation of a-hydroxyl groups in synthetic b–O–4 lignin oligomers (Figure 2), which further shows that LigD has the potential to accept lignin polymers as substrates and can carry out both end- and internal-unit oxidations; this supports the indirect evidence for a-oxidation of synthetic lignins and hardwood alkali lignin by LigD described previously (Sonoki et al., 2002; Reiter et al., 2013). Increase of a-keto-b–O–4-linked dilignols and neolignanlike compounds by expression of LigD in Arabidopsis The LigD gene without (D) and with (SD) sequence coding for an apoplast-targeting signal peptide was expressed in Arabidopsis plants. Both the D and SD constructs lead to significant accumulation of the a-oxidized products of two typical dilignols G(b–O–4)G and G(b–O–4)G0 , that is Gox(b–O–4)G and Gox(b–O– 4)G0 (Table 1, Figure 3). In addition, the levels of a series of aoxidized neolignan-like compounds such as Gox(b–O–4)ferulic acid and Gox(b–O–4)sinapic acid conjugated with hexose(s), glutamate, malate and/or a 224 Da moiety were also elevated to equivalent levels in both types of transgenic lines compared to the wild-type and sYFP transgenic lines (Table 1, Figure 3). G(b–O–4) ferulic acid and G(b–O–4)sinapic acid, and derivatives of these compounds, are known metabolites of wild-type Arabidopsis (Matsuda et al., 2010; Morreel et al., 2014). The accumulation of Gox(b–O–4)ferulic acid and Gox(b–O–4)sinapic acid derivatives coincides with a reduction in G(b–O–4)ferulic acid and G(b–O–4) sinapic acid derivatives (Table 1, Figure 3). It is currently unknown whether G(b–O–4)ferulic acid and G(b–O–4)sinapic acid are oxidized by LigD and then further decorated, or whether the derivatives themselves are accepted by LigD as substrates. However, based on the pronounced activity of the recombinant LigD on the synthetic lignin dimers and oligomers (Figures 1, 2 and S1), it seems likely that LigD would accept bulky substitutions (such as hexoses, malate and a 224 Da moiety) on the 4-O-linked unit and therefore that both the decorated and non-decorated compounds could act as substrates for LigD. Because almost all differentially accumulating Ca-oxidized products are present both in D and SD lines, it is most likely that SD lines still have considerable LigD activity in the cytosol. Indeed, the ATS-fused YFP showed cytoplasmic localization in tobacco BY-2 cells in addition to the anticipated apoplastic localization (Figure S3). Furthermore, the fact that the a-oxidized dilignols, Gox(b–O–4)G and Gox(b–O–4)G0 accumulated to the same extent in both D and SD lines suggests that these metabolites derived from cytosolic substrate pools in both lines. This idea is in line with the accumulation in both lines of Gox 4-Ohexoside(b–O–4)G, a metabolite that depends on the activity of nucleotide-diphosphate-sugar (hexosyl) transferases, which are known to be cytoplasmic enzymes (Lanot et al., 2006; Wang et al., 2013). The presence of dilignols in the cytoplasm further suggests that the monolignols couple in cytoplasm. These data are in line with conclusions taken from analysing transgenic poplar trees down-regulated for the cytoplasmic phenylcoumaran benzylic ether reductase (PCBER). These trees accumulate several thousandfold higher levels of cysteine adducts of dilignols. These adducts are formed by nucleophilic attack, by cysteine, on the quinone methide intermediate in the formation of (b–O–4)-linked dimers (Niculaes et al., 2014). The heterologous expression of LigD in the cytoplasm and the resulting accumulation of oxidized G(b–O–4)G dimers and neolignan-like compounds further underpin the hypothesis that monolignols not only couple in the apoplast, but also in the cytoplasm. The absence of (or low) LigD activity in apoplastic space could be attributed to the lack of the cofactor NAD+, unfavourable pH or the presence of inhibitors. Although NAD+ is localized mainly in the cytosol and in mitochondria, its apoplastic localization has been reported in the literature (Otter and Polle, 1997; Shinkle et al., 1992). LigD expression resulted in increased levels of a-keto substructures in Arabidopsis lignin As shown in Figures 4 and 5, the HSQC data from ELs prepared from each of the transgenic lines are typical of normal lignins from Arabidopsis (Marita et al., 1999; Vanholme et al., 2010b) except for the significant increases in oxidized a-ether unit (A0 ) levels. Elevated levels of the a-keto structure have also been observed in syringyl (S)-rich lignins deposited in transgenic Arabidopsis (Marita et al., 1999) and poplar (4.6-fold) (Stewart et al., 2009) in which the ferulate 5-hydroxylase gene was overexpressed. S-enriched lignins more easily succumb to a-oxidization because the oxidation preferentially occurs in syringyl units during the ball-milling procedure for pulverization of cell walls—S units have lower oxidation potentials than their guaiacyl coun€ € terparts (Amm alahti et al., 1998). Importantly therefore, as indicated in Figure 4 and Table S2, no significant differences could be observed between the transgenic and wild-type lines in the concentrations of a-keto S units, nor in the S/G ratio, implying that the frequency of a-keto-guaiacyl units is not artefactually inflated by the ball-milling procedure during the preparation of acetylated EL. In contrast to S units, levels of a-oxidization were increased significantly in guaiacyl (G) units of both transgenic lines compared to those in the wild-type and the sYFP transgenic plants (Figure 5 and Table S2). Given that S units protect G units from oxidation during the mechanical operations, (and that oxidized G units are always minor-to-undetectable in dicot lignins), the observed a-keto G units (G0 ) should then originate from LigD activity. The origin of a-keto substructures in the lignin of LigDexpressing Arabidopsis Based on the phenolic profiling data described above, it is most likely that LigD is mainly active in the cytoplasm. It is therefore striking that the a-keto structures were slightly but significantly increased in the cell wall lignins isolated from LigD lines (Figures 4 and 5, and Table S2), especially in the D lines where LigD is only in the cytoplasm. The observed a-keto-guaiacyl (G0 ) units must thus be due to incorporation of the a-oxidized metabolites listed in Table 1. Indeed, all of the structurally characterized a-oxidized dilignols and neolignan-like compounds have G residues. It is doubtful that the neolignan-like compounds conjugated with malate and hexose can be incorporated actively into the lignin polymer during an inherent process for lignin biosynthesis. More likely, the a-oxidized dilignols, Gox(b–O–4)G and Gox(b–O–4)G0 , detected in D and SD transgenic lines, are incorporated into the lignins and contribute to increased a-keto-b–O–4 units. Whether these dimers are translocated to the cell wall while the cell is alive as recently observed for the nontraditional monolignol coniferyl ferulate (Wilkerson et al., 2014), or are only incorporated during post-mortem lignification (Pesquet et al., 2013; Smith et al., ª 2015 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–12 Introduction of chemically labile units into lignin 9 2013), is currently unknown. The low level of the a-keto units in the lignins of the LigD transgenic lines might be due to the low frequency of the incorporation of the a-oxidized dilignols into the lignin polymer. In an endwise polymerization of lignin that is believed to predominate in the native lignification process, radicals from dilignols mainly couple with monolignol radicals, not with another dilignol or oligolignol radical. Unfortunately, no oligolignols such as trilignols or tetralignols with a-keto-b–O–4 substructure could be detected in our metabolomic analysis of the LigD transgenic lines. Although LigD accepted S(b–O–4)S as in vitro substrate, neither phenolic profiling nor the cell wall analyses revealed any significant changes in the a-oxidation levels for syringyl (S)-type oligolignols and cell wall lignin components. Given that LigD was mainly active in the cytoplasm, this observation is indicative of a much smaller cytoplasmic pool of S(b–O–4)S compared to that of G(b–O–4)G. In conclusion, although LigD is able to catalyse the a-oxidation of b–O–4-units in vitro and in the cytoplasm in planta, there is no evidence for in planta apoplast activity even when the enzyme was targeted to the apoplast with an ATS. As a consequence, aketo-b–O–4 units in the lignin were below target amounts, and their effect on enzymatic saccharification after alkaline pretreatment was not detected (Figure S9). However, given that a-ketob–O–4 units were increased in the lignin of LigD plants, LigD still represents a promising gene for lignin modification if its expression or activity could be optimized, for example, by defining the apoplastic LigD inhibitor(s) and by improving the catalytic properties of LigD. Materials and methods Synthetic lignin dimers and oligomers GGE was purchased from Tokyo Kasei Kogyo Co., Ltd. (Tokyo, Japan). Dilignols, G(b–O–4)G and S(b–O–4)S (Tanahashi et al., 1975), and the b–O–4-linked lignin oligomers (Katahira et al., 2006) were synthesized according to literature methods. Gox(b– O–4)G and Sox(b–O–4)S were synthesized from G(b–O–4)G and S (b–O–4)S via a three-step procedure described in the Supporting Information (Data S1). GGE-glc, a-keto-GGE, VGE were also newly synthesized and the method will be reported elsewhere. Substrate specificity test Substrate specificity of LigD was tested using cell extracts of a recombinant E. coli harbouring pETDa that contains coding sequence of LigD (Sato et al., 2009). The extract of recombinant E. coli harbouring an empty vector pET-21a(+) was used as negative control. Induction of gene expression and preparation of the cell extract were described previously (Sato et al., 2009) with a slight modification in the buffer composition. The cell extracts (200 lg of protein) were incubated with 200 lM of each substrate in the presence of 1 mM NAD+ in 300 lL Tris-HCl buffer (50 mM, pH 8.0) at 30 °C for 30 s for oxidations of G(b–O– 4)G, S(b–O–4)S and GGE-glc or 10 min for GGE. An aliquot (90 lL) of each reaction mixture was mixed with the same volume of methanol, centrifuged, filtered, and subjected to UPLC-MS as described previously (Kamimura et al., 2010). The reaction conditions for the b–O–4-linked lignin oligomers were essentially the same as described above, except for the substrate concentration (195 lg/mL), and the reaction products were analysed by UPLC-MS/MS as described previously (Vanholme et al., 2013). Phenolic profiling D5, D8, SD1 and SD9 transgenic lines were grown together with wild type and sYFP as control genotypes for 8 weeks in short-day conditions (9 h light/15 h dark, 22 °C), which allowed for the development of a rosette but suppressed inflorescence stem development. After 8 weeks, they were moved to long-day conditions (16 h light/8 h dark, 22 °C). These conditions allowed the development of a single, thick inflorescence stem. Stems were harvested at a height of about 45 cm. Eight biological replicates were used for each of the genotypes. Methanol extraction, sample preparation and UPLC-MS settings were as described in Vanholme et al. (2013). From the resulting chromatograms, 7067 deisotoped peaks were integrated and aligned via ProgenesisQI (Nonlinear, a Waters company, Newcastle, UK), each peak having characterized by its m/z and retention time. We first filtered the peaks in Microsoft Excel based on their presence in the biological replicates: a peak was retained if it was present in at least 7 of the 8 replicates in at least one of the six genotypes and if the average peak area was at least 4000 counts in at least one of the six genotypes. Applying this filter resulted in 654 peaks. Next, ANOVA was applied to all peaks in Microsoft Excel with the genotype as a factor. The Benjamini and Hochberg multiple test correction was used on the resulting P-values and calculated with the function mt.rawp2adjp (proc = ‘BH’) in the R package multtest (Benjamini and Hochberg, 1995). 87 peaks had a P-value <0.01, and these were used for further selection. T-tests performed in Microsoft Excel were used as post hoc tests. The P-value had to be below 0.05 for both D5 and D8 as compared to both wild type and sYFP (making a total of four t-tests: D5 vs wild type, D5 vs sYFP, D8 vs wild type and D8 vs sYFP). In addition, the change had to be in the same direction (higher peak areas in both D5 and D8 as compared to the controls or lower peak areas in both D5 and D8) to retain the peak for further analysis. Thirty-two peaks had a significantly higher area in the transgenic lines and 8 peaks had a lower area (Table S1). In addition, peaks that had a P-value below 0.05 for both SD1 and SD9 as compared to both wild type and sYFP (making a total of four t-tests: SD1 vs wild type, SD1 vs sYFP, SD9 vs wild type and SD9 vs sYFP) were retained for further analysis. Again, the change had to be in the same direction (higher peak areas in both SD1 and SD9 as compared to the controls or lower peak areas in both SD1 and SD9) to retain the peak for further analysis. Twenty-eight peaks were significantly up in the transgenic lines and 6 down (Table S1). Some of the significantly different peaks appeared in-source derived fragments of parent ions that were also significantly different in abundance. Therefore, the number of differential compounds is lower than the number of differential peaks. Cell wall analysis Plant growth conditions were the same as those for metabolic profiling. Extractive-free plant cell walls of fully senesced inflorescence stems were prepared for wet-chemical analyses as previously described (Bonawitz et al., 2014; Wagner et al., 2011, 2013). In brief, senesced Arabidopsis stems (~300 mg) were cut into small pieces, preground using a Retch MM400 (frequency, 30 s1 for 1 min) mill and extracted with 80% aqueous ethanol (sonication 3 9 20 min). Lignin content was determined by the acetyl bromide method (Hatfield et al., 1999). The distribution of amorphous sugars (hemicelluloses and pectins) and crystalline glucan (cellulose) was determined by two-step acid hydrolysis of ª 2015 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–12 10 Yukiko Tsuji et al. the cell walls using trifluoroacetic acid and sulphuric acid as described previously (Tobimatsu et al., 2013, 2012). Lignin characterization by 2D NMR Cellulolytic enzyme lignin (EL) samples were prepared as described previously (Bonawitz et al., 2014; Wagner et al., 2011, 2013). In brief, pre-extracted cell walls of senesced stems (~150 mg) were ball-milled (3 9 5 min milling and 5 min cooling cycles) using a Fritsch Planetary micro mill Pulverisette 7 vibrating at 600 rpm with ZrO2 vessels containing ZrO2 ball bearings. The ball-milled walls were then transferred to centrifuge tubes and digested at 30 °C with crude cellulases (Calbiochem Cellulysin; lot no. D00074989; 30 mg/g of sample, in pH 5.0 acetate buffer; three times over 24 h; fresh buffer and enzyme added each time), leaving ELs comprised of all of the lignin and residual polysaccharides (yield, 22–28%). The isolated ELs (~30 mg) were subjected to solubilization and acetylation in DMSO/NMI/acetic anhydride (Lu and Ralph, 2003; Mansfield et al., 2012) to afford acetylated ELs (yield, 110–121%). The acetylated ELs were completely dissolved in 0.5 ml of chloroform-d and subjected to NMR on a Bruker Biospin AVANCE 700 MHz spectrometer fitted with a cryogenically cooled 5-mm TXI gradient probe with inverse geometry (proton coils closest to the sample). The central chloroform peak was used as internal reference (dC, 77.0; dH, 7.26 ppm). HSQC experiments using Bruker’s adiabatic pulse version of the experiment (hsqcetgpsisp2.2) were carried out using the parameters described previously (Mansfield et al., 2012). Processing used typical matched Gaussian apodization in F2 (LB = 0.5, GB = 0.001) and squared cosine-bell apodization and one level of linear prediction (32 coefficients) in F1. Volume integration of contours in HSQC plots used Bruker’s TopSpin 3.2 (Mac) software; no linear prediction was applied for integral determination. For quantification of lignin aromatic distributions (Figure 5), only the carbon/proton-2 correlations from G and G’ units and the carbon/proton-2/6 correlations from S and S’ units were used, and the G and G’ integrals were logically doubled. For lignin interunit linkage types (Figure 5), the well-resolved sidechain contours (Aa, Ba, Ca, Da, and A’b) were integrated. These quantifications used no correction factors, that is the data represent volume integrals only. NMR analyses were performed on three biological replicates for D and SD lines, and on two biological replicates for wild-type and sYFP lines. Significance of the quantification data was evaluated by unpaired t-tests as summarized in Table S2. Acknowledgements This work was supported in part by the Japan Science and Technology Agency (Advanced Low Carbon Technology Research and Development Program), the New Energy and Industrial Technology Development Organization of Japan (Development of Preparatory Basic Bioenergy Technology), the European Commission’s Directorate-General for Research within the 7th Framework Program (FP7/2007-2013) under the grant agreement N° 270089 (MULTIBIOPRO), the Hercules program of Ghent University for the Synapt Q-Tof (grant no. AUGE/014), the ‘Bijzondere Onderzoeksfonds-Zware Apparatuur’ of Ghent University for the FTICR-MS instrument (174PZA05) and the Multidisciplinary Research Partnership ‘Biotechnology for a Sustainable Economy’ (01MRB510W) of Ghent University. YT, CEF and JR acknowledge funding from the US Department of Energy (DOE) Great Lakes Bioenergy Research Center (DOE Office of Science BER DE-FC02- 07ER64494). PO is indebted to the National Commission for Scientific and Technological Research (of Chile) for a predoctoral fellowship, and RV is indebted to the Research FoundationFlanders (FWO) for a postdoctoral fellowship. The authors also thank Takahito Mizukami for his assistance with preparation of LigD substrates. References € €alahti, E., Brunow, G., Bardet, M., Robert, D. and Kilpel€ainen, I. (1998) Amm Identification of side-chain structures in a poplar lignin using three-dimensional HMQCHOHAHA NMR spectroscopy. J. Agric. Food Chem. 46, 5113–5117. Baucher, M., Halpin, C., Petit-Conil, M. and Boerjan, W. (2003) Lignin: genetic engineering and impact on pulping. Crit. Rev. Biochem. Mol. Biol. 38, 305– 350. Benjamini, Y. and Hochberg, Y. (1995) Controlling the false discovery rate - a practical and powerful approach to multiple testing. J. R. Stat. Soc. Series B Stat. Methodol. 57, 289–300. Boerjan, W., Ralph, J. and Baucher, M. (2003) Lignin biosynthesis. Annu. Rev. Plant Biol. 54, 519–546. Bonawitz, N.D. and Chapple, C. (2010) The genetics of lignin biosynthesis: connecting genotype to phenotype. Annu. Rev. Genet. 44, 337–363. Bonawitz, N.D., Kim, J.I., Tobimatsu, Y., Ciesielski, P.N., Anderson, N.A., Ximenes, E., Maeda, J., Ralph, J., Donohoe, B.S., Ladisch, M. and Chapple, C. (2014) Disruption of mediator rescues the stunted growth of a lignin-deficient Arabidopsis mutant. Nature, 509, 376–380. Chen, F. and Dixon, R.A. (2007) Lignin modification improves fermentable sugar yields for biofuel production. Nat. Biotechnol. 25, 759–761. Chen, F., Tobimatsu, Y., Havkin-Frenkel, D., Dixon, R.A. and Ralph, J. (2012) A polymer of caffeyl alcohol in plant seeds. Proc. Natl Acad. Sci. USA, 109, 1772–1777. Criss, D.L., Fisher, T.H. and Schultz, T.P. (1998) Alkaline hydrolysis of nonphenolic a-carbonyl b-O-4 lignin dimers substituted on the leaving phenoxide ring: comparison with benzylic hydroxyl analogues. Holzforschung, 52, 57–60. Del Rio, J.C., Marquesm, G., Rencoret, J., Martinez, A.T. and Gutierrez, A. (2007) Occurrence of naturally acetylated lignin units. J. Agric. Food Chem. 55, 5461–5468. Demura, T. and Ye, Z.H. (2010) Regulation of plant biomass production. Curr. Opin. Plant Biol. 13, 299–304. Eudes, A., George, A., Mukerjee, P., Kim, J.S., Pollet, B., Benke, P.I., Yang, F., €l, O.P., Chabout, S., Mouille, G., Soubigou-Taconnat, Mitra, P., Sun, L., Cetinko L., Balzergue, S., Singh, S., Holmes, B.M., Mukhopadhyay, A., Keasling, J.D., Simmons, B.A., Lapierre, C., Ralph, J. and Loqu
e, D. (2012) Biosynthesis and incorporation of side-chain-truncated lignin monomers to reduce lignin polymerization and enhance saccharification. Plant Biotechnol. J. 10, 609–620. Fu, C., Mielenz, J.R., Xiao, X., Ge, Y., Hamilton, C.Y., Rodriguez, M., Chen, F., Foston, M., Ragauskas, A., Bouton, J., Dixon, R.A. and Wang, Z.-Y. (2011) Genetic manipulation of lignin reduces recalcitrance and improves ethanol production from switchgrass. Proc. Natl Acad. Sci. USA, 108, 3803–3808. Gall, D.L., Kim, H., Lu, F., Donohue, T.J., Noguera, D.R. and Ralph, J. (2014) Stereochemical features of glutathione-dependent enzymes in the Sphingobium sp. strain SYK-6 ß–aryl etherase pathway. J. Biol. Chem. 289, 8656–8667. Gierer, J. and Ljunggren, S. (1979) The reactions of lignins during sulfate pulping. Part 16. The kinetics of the cleavage b-aryl ether linkages in structures containing carbonyl group. Svensk Papperstidn. 82, 71–81. Gierer, J. and Nor
en, I. (1982) Oxidative pretreatment of pine wood to facilitate delignification during Kraft pulping. Holzforschung, 36, 123–130. Gierer, J., Ljunggren, S., Ljungquist, P. and Noren, I. (1980) Reactions of lignin during sulfate pulping. Part 18. The significance of a-carbonyl groups for the cleavage of b-aryl ether structures. Svensk Papperstidn. 83, 75–82. Grabber, J.H., Hatfield, R.D., Lu, F. and Ralph, R. (2008) Coniferyl ferulate incorporation into lignin enhances the alkaline delignification and enzymatic degradation of cell walls. Biomacromolecules, 9, 2510–2516. ª 2015 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–12 Introduction of chemically labile units into lignin 11 Greiner, S., Krausgrill, S. and Rausch, T. (1998) Cloning of a tobacco apoplasmic invertase inhibitor. Proof of function of the recombinant protein and expression analysis during plant development. Plant Physiol. 116, 733–742. Hatfield, R.D., Grabber, J., Ralph, J. and Brei, K. (1999) Using the acetyl bromide assay to determine lignin concentrations in herbaceous plants: some cautionary notes. J. Agric. Food Chem. 47, 628–632. Imai, A., Yokoyama, T., Matsumoto, Y. and Meshitsuka, G. (2007) Significant lability of guaiacylglycerol b-phenacyl ether under alkaline conditions. J. Agric. Food Chem. 55, 9043–9046. Kajita, S., Ishifuji, M., Ougiya, H., Hara, S., Kawabata, H., Morohoshi, N. and Katayama, Y. (2002) Improvement in pulping and bleaching properties of xylem from transgenic tobacco plants. J. Sci. Food Agric. 82, 1216–1223. Kamimura, N., Aoyama, T., Yoshida, R., Takahashi, K., Kasai, D., Abe, T., Mase, K., Katayama, Y., Fukuda, M. and Masai, E. (2010) Characterization of the protocatechuate 4,5-cleavage pathway operon in Comamonas sp. strain E6 and discovery of a novel pathway gene. Appl. Environ. Microbiol. 76, 8093– 8101. Katahira, R., Kamitakahara, H., Takano, T. and Nakatsubo, F. (2006) Synthesis of b-O-4 type oligomeric lignin model compound by the nucleophilic addition of carbanion to the aldehydes group. J. Wood Chem. 52, 255–260. Lanot, A., Hodge, D., Jackson, R.G., George, G.L., Elias, L., Lim, E.K., Vaistij, F.E. and Bowles, D.J. (2006) The glucosyltransferase UGT72E2 is responsible for monolignol 4-O-glucoside production in Arabidopsis thaliana. Plant J. 48, 286–295. Lapierre, C., Pollet, B., MacKay, J.J. and Sederoff, R.R. (2000) Lignin structure in a mutant pine deficient in cinnamyl alcohol dehydrogenase. J. Agric. Food Chem. 48, 2326–2331. Lepl
e, J.C., Dauwe, R., Morreel, K., Storme, V., Lapierre, C., Pollet, B., Naumann, A., Kang, K.Y., Kim, H., Ruel, K., Lefebvre, A., Joseleau, J.P., Grima-Pettenati, J., De Rycke, R., Andersson-Gunner as, S., Erban, A., Fehrle, I., Petit-Conil, M., Kopka, J., Polle, A., Messens, E., Sundberg, B., Mansfield, S.D., Ralph, J., Pilate, G. and Boerjan, W. (2007) Downregulation of cinnamoyl-coenzyme A reductase in poplar: multiple-level phenotyping reveals effects on cell wall polymer metabolism and structure. Plant Cell, 19, 3669–3691. Li, X., Weng, J.K. and Chapple, C. (2008) Improvement of biomass through lignin modification. Plant J. 54, 569–581. Li, X., Ximenes, E., Kim, Y., Slininger, M., Meilan, R., Ladisch, M. and Chapple, C. (2010) Lignin monomer composition affects Arabidopsis cell-wall degradability after liquid hot water pretreatment. Biotechnol. Biofuels 3, 27. Lu, F. and Ralph, J. (1998) The DFRC methods for lignin analysis. 2. Monomer from isolated lignins. J. Agric. Food Chem. 46, 547–552. Lu, F. and Ralph, J. (2003) Non-degradative dissolution and acetylation of ball-milled plant cell walls; high-resolution solution-state NMR. Plant J. 35, 535–544. Lu, F. and Ralph, J. (2008) Novel tetrahydrofuran structures derived from b-b-coupling reactions involving sinapyl acetate in kenaf lignins. Org. Biomol. Chem. 6, 3681–3694. Mansfield, S.D., Lu, F. and Ralph, J. (2012) Whole plant cell wall characterization using solution-state 2D-NMR. Nat. Protoc. 7, 1579–1589. Marita, J.M., Ralph, J., Hatfield, R.D. and Chapple, C. (1999) NMR characterization of lignins in Arabidopsis altered in the activity of ferulate 5-hydroxylase. Proc. Natl Acad. Sci. USA, 96, 12328–12332. Masai, E., Kubota, S., Katayama, Y., Kawai, S., Yamasaki, M. and Morohoshi, N. (1993) Characterization of the Ca-dehydrogenase gene involved in the cleavage of b-aryl ether by Pseudomonas paucimobilis. Biosci. Biotechnol. Biochem. 57, 1655–1659. Masai, E., Katayama, Y. and Fukuda, M. (2007) Genetic and biochemical investigation on bacterial catabolic pathways for lignin-derived aromatic compounds. Biosci. Biotechnol. Biochem. 71, 1–15. Matsuda, F., Hirai, M.Y., Sasaki, E., Akiyama, K., Yonekura-Sakakibara, K., Provart, N.J., Sakurai, T., Shimada, Y. and Saito, K. (2010) AtMetExpress development: a phytochemical atlas of Arabidopsis development. Plant Physiol. 152, 566–578. Morreel, K., Ralph, J., Kim, H., Lu, F., Goeminne, G., Ralph, S., Messens, E. and Boerjan, W. (2004) Profiling of oligolignols reveals monolignol coupling conditions in lignifying poplar xylem. Plant Physiol. 136, 3537–3549. Morreel, K., Kim, H., Lu, F., Dima, O., Aliyama, T., Vanholme, R., Niculaes, C., Goeminne, G., Inz
e, D., Messens, E., Ralph, J. and Boerjan, W. (2010) Mass spectrometry-based fragmentation as an identification tool in lignomics. Anal. Chem. 82, 8095–8105. Morreel, K., Saeys, Y., Dima, O., Lu, F., Van de Peer, Y., Vanholme, R., Ralph, J., Vanholme, B. and Boerjan, W. (2014) Systematic structural characterization of metabolites in Arabidopsis via candidate substrate-product pair networks. Plant Cell, 26, 929–945. Niculaes, C, Morreel, K., Kim, H., Lu, F., McKee, L., Ivens, B., Haustraete, J., Vanholme, B., De Rycke, R., Hertzberg, M., Fromm, J., Bulone, V., Polle, A., Ralph, J. and Boerjan, W. (2014) Phenylcoumaran benzylic ether reductase (PCBER) prevents accumulation of compounds formed under oxidative conditions in poplar xylem. Plant Cell, 26, 3775–3791. O’Connell, A., Holt, K., Piquemal, J., Grima-Pettenati, J., Boudet, A., Pollet, B., Lapierre, C., Petit-Conil, M., Schuch, W. and Halpin, C. (2002) Improved paper pulp from plants with suppressed cinnamoyl-CoA reductase or cinnamyl alcohol dehydrogenase. Transgenic Res. 11, 495–503. Otter, T. and Polle, A. (1997) Characterisation of acidic and basic apoplastic peroxidases from needles of Norway spruce (Picea abies L. Karsten) with respect to lignifying substrates. Plant Cell Physiol. 38, 595–602. Existence of NAD+ in apoplast of needles was discussed in this study. Pesquet, E., Zhang, B., Gorzs
as, A., Puhakainen, T., Serk, H., Escamez, S., Barbier, O., Gerber, L., Courtois-Moreau, C., Alatalo, E., Paulin, L., Kangasj€arvi, J., Sundberg, B., Goffner, D. and Tuominen, H. (2013) Non-cell-autonomous postmortem lignification of tracheary elements in Zinnia elegans. Plant Cell, 25, 1314–1328. Pilate, G., Guiney, E., Holt, K., Petit-Conil, M., Lapierre, C., Lepl
e, J.-C., Pollet, B., Mila, I., Webster, E.A., Marstorp, H.G., Hopkins, D.W., Jouanin, L., Boerjan, W., Schuch, W., Cornu, D. and Halpin, C. (2002) Field and pulping performances of transgenic trees with altered lignifications. Nat. Biotechnol. 20, 607–612. Rahimi, A., Azarpira, A., Kim, H., Ralph, J. and Stahl, S.S. (2013) Chemoselective metal-free aerobic alcohol oxidation in lignin. J. Am. Chem. Soc. 135, 6415–6418. Rahimi, A., Ulbrich, A., Coon, J. J. and Stahl, S.S. (2014) Formic-acid-induced depolymerization of oxidized lignin to aromatics. Nature, 515, 249–252. in press (doi:10.1038) Ralph, J. (2010) Hydroxycinnamates in lignification. Phytochem. Rev. 9, 65–83. Ralph, J., Bunzel, M., Marita, J.M., Hatfield, R.D., Lu, F., Kim, H., Schatz, P.F., Grabber, J.H. and Steinhart, H. (2004a) Peroxidase-dependent cross-linking reactions of p-hydroxycinnamates in plant cell walls. Phytochem. Rev. 3, 79–96. Ralph, J., Lundquist, K., Brunow, G., Lu, F., Kim, H., Schatz, P.F., Marita, J.M., Hatfield, R.D., Ralph, S.A., Christensen, J.H. and Boerjan, W. (2004b) Lignins: natural polymers from oxidative coupling of 4-hydroxyphenylpropanoids. Phytochem. Rev. 3, 29–60. Reiter, J., Strittmatter, H., Wiemann, L.O., Schieder, D. and Siever, V. (2013) Enzymatic cleavage of lignin b-O-4 aryl ether bond via net internal hydrogen transfer. Green Chem. 15, 1373–1381. Sato, Y., Moriuchi, H., Hishiyama, S., Otsuka, Y., Oshima, K., Kasai, D., Nakamura, M., Ohara, S., Katayama, Y., Fukuda, M. and Masai, E. (2009) Identification of three alcohol dehydrogenase genes involved in the stereospecific catabolism of arylglycerol-b-aryl ether by Sphingobium sp strain SYK-6. Appl. Environ. Microbiol. 75, 5195–5201. Shinkle, J.R., Swoap, S.J., Simon, P. and Jones, R.L. (1992) Cell wall free space of Cucumis hypocotyls contains NAD and a blue light-regulated peroxidase activity. Plant Physiol. 98, 1336–1341. Existence of NAD+ in apoplast of Cucumis hypocotyls was discussed in this study. Shiraishi, T., Sannami, Y., Kamitakahara, H. and Takano, T. (2013) Comparison of a series of laccase mediators in the electro-oxidation reactions of non-phenolic lignin model compounds. Electrochim. Acta, 106, 440–446. Smith, D.C.C. (1955) p-Hydroxybenzoates groups in the lignin of Aspen (Populus tremula). J. Chem. Soc. 2347–2351. Smith, R.A., Schuetz, M., Roach, M., Mansfield, S.D., Ellis, B. and Samuels, L. (2013) Neighboring parenchyma cells contribute to Arabidopsis xylem lignification, while lignification of interfascicular fibers is cell autonomous. Plant Cell, 25, 3988–3999. ª 2015 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–12 12 Yukiko Tsuji et al. Sonoki, T., Iimura, Y., Masai, E., Kajita, S. and Katayama, Y. (2002) Specific degradation of b-aryl ether linkage in synthetic lignin (dehydrogenative polymerizate) by bacterial enzymes of Sphingomonas paucimobilis SYK-6 produced in recombinant Escherichia coli. J. Wood Chem. 48, 429–433. Stewart, J.J., Akiyama, T., Chapple, C., Ralph, J. and Mansfield, S.D. (2009) The effects on lignin structure of overexpression of ferulate 5-hydroxylase in hybrid poplar. Plant Physiol. 150, 621–635. Tanahashi, M., Takeuchi, H. and Higuchi, T. (1975) Dehydrogenative polymerization of 3,5-disubstituted p- coumaryl alcohols. Wood Res. 61, 44–53. Tobimatsu, Y., Elumalai, S., Grabber, J.H., Davidson, C.L., Pan, X. and Ralph, J. (2012) Hydroxycinnamate conjugates as potential monolignol replacements: in vitro lignification and cell wall studies with rosmarinic acid. ChemSusChem. 5, 676–686. ~o, L.L., Jackson, L., Tobimatsu, Y., Chen, F., Nakashima, J., Escamilla-Trevin Dizon, R.A. and Ralph, J. (2013) Coexistence but independent biosynthesis of catechyl and guaiacyl/syringyl lignin polymers in seed coats. Plant Cell, 25, 2587–2600. Van Acker, R., Lepl
e, J.-C., Aerts, D., Storme, V., Goeminne, G., Ivens, B., L
eg
ee, F., Lapierre, C., Piens, K., Van Montagu, M.C.E., Santoro, N., Foster, C.E., Ralph, J., Soetaert, W., Pilate, G. and Boerjan, W. (2014) Improved saccharification and ethanol yield from field-grown transgenic poplar deficient in cinnamoyl-CoA reductase. Proc. Natl Acad. Sci. USA, 111, 845– 850. Vanholme, R., Morreel, K., Ralph, J. and Boerjan, W. (2010a) Lignin biosynthesis and structure. Plant Physiol. 153, 895–905. Vanholme, R., Ralph, J., Akiyama, T., Lu, F., Jorge Rencoret Pazo, J.R., Kim, H., Jørgen Holst Christensen, J.H., Van Reusel, B., Storme, V., De Rycke, R., Rohde, A., Morreel, K. and Boerjan, W. (2010b) Engineering traditional monolignols out of lignin by concomitant up-regulation of F5H1 and down-regulation of COMT in Arabidopsis. Plant J. 64, 885–897. Vanholme, R., Storme, V., Vanholme, B., Sundin, L., Christensen, J.H., Goeminne, G., Halpin, C., Rohde, A., Morreel, K. and Boerjan, W. (2012) A systems biology view of responses to lignin biosynthesis perturbations in Arabidopsis. Plant Cell, 24, 3506–3529. Vanholme, R., Cesarino, I., Rataj, K., Xiao, Y., Sundin, L., Goeminne, G., Kim, H., Cross, J., Morreel, K., Araujo, P., Welsh, L., Haustraete, J., McClellan, C., Vanholme, B., Ralph, J., Simpson, G.G., Halpin, C. and Boerjan, W. (2013) Caffeoyl shikimate esterase (CSE) is an enzyme in the lignin biosynthetic pathway. Science, 341, 1103–1106. Voelker, S.L., Lachenbruch, B., Meinzer, F.C., Kitin, P. and Strauss, S.H. (2011) Transgenic poplars with reduced lignin show impaired xylem conductivity, growth efficiency and survival. Plant, Cell Environ. 34, 655–668. Wagner, A., Tobimatsu, Y., Phillips, L., Flint, H., Torr, K., Donaldson, L., Pears, L. and Ralph, J. (2011) CCoAOMT suppression modifies lignin composition in Pinus radiata. Plant J. 67, 119–129. Wagner, A., Tobimatsu, Y., Goeminne, G., Phillips, L., Flint, H., Steward, D., Torr, K., Donaldson, L., Boerjan, W. and Ralph, J. (2013) Suppression of CCR changes metabolite profile and cell wall composition in Pinus radiata tracheary elements. Plant Mol. Biol. 81, 105–117. Wang, Y., Chantreau, M., Sibout, R. and Hawkins, S. (2013) Plant cell wall lignification and monolignol metabolism. Front. Plant Sci. 9, 220. Weng, J.-K., Mo, H. and Chapple, C. (2010) Over-expression of F5H in COMT-deficient Arabidopsis leads to enrichment of an unusual lignin and disruption of pollen wall formation. Plant J. 64, 898–911. Wilkerson, C.G., Mansfield, S.D., Lu, F., Withers, S., Park, J.Y., Karlen, S.D., Gonzales-Vigil, E., Padmakshan, D., Unda, F., Rencoret, J. and Ralph, J. (2014) Monolignol ferulate transferase introduces chemically labile linkages into the lignin backbone. Science, 344, 90–93. Yang, F., Mitra, P., Zhang, L., Prak, L., Verhertbruggen, Y., Kim, J.S., Sun, L., Zheng, K., Tang, K., Auer, M., Scheller, H.V. and Loqu
e, D. (2012) Engineering secondary cell wall deposition in plants. Plant Biotechnol. J. 11, 325–335. Zhang, K., Bhuiya, M.-W., Pazo, J.R., Miao, Y., Kim, H., Ralph, J. and Liua, C.-J. (2012) An engineered monolignol 4-O-methyltransferase depresses lignin biosynthesis and confers novel metabolic capability in Arabidopsis. Plant Cell, 24, 3135–3152. Zhao, Q. and Dixon, R.A. (2011) Transcriptional networks for lignin biosynthesis: more complex than we thought? Trends Plant Sci. 16, 227–233. Zhou, R., Jackson, L., Shadle, G., Nakashima, J., Temple, S., Chen, F. and Dixon, R.A. (2010) Distinct cinnamoyl CoA reductases involved in parallel routes to lignin in Medicago truncatula. Proc. Natl Acad. Sci. USA, 107, 17803–17808. Supporting information Additional Supporting information may be found in the online version of this article: Figure S1 UPLC-MS data of synthetic b–O–4-linked lignin dimers treated with LigD. Figure S2 Schematic representation of T-DNA regions of the ligD and ATS-ligD expression cassettes. Figure S3 The tobacco invertase inhibitor derived apoplasttargeting signal aids to localize protein to the cell walls of BY-2 cells. Figure S4 Stem phenotypes of the transgenic plants. Figure S5 Expression and Western blot analysis. Figure S6 LigD activity in crude extracts prepared from transgenic Arabidopsis stems. Figure S7 Chromatograms and MS/MS spectra of Gox(b–O–4)G and Sox(b–O–4)S standards. Figure S8 MS-based structural elucidation of the differentially accumulating peaks in the D- and SD-transformed plants as compared to wild type. Figure S9 Enzymatic saccharification efficiency of pulverized stem tissues with or without alkaline pretreatment. Table S1 Cell wall composition determined by wet-chemical analyses. Table S2 NMR-derived lignin aromatic unit and inter-unit linkage distributions. Table S3 Nucleotide sequence of primers used for RT-PCR. Data S1 Materials and methods for generation and analysis of transgenic plants, and for chemical synthesis. ª 2015 Society for Experimental Biology, Association of Applied Biologists and John Wiley & Sons Ltd, Plant Biotechnology Journal, 1–12