Dissection of the transcriptional program regulating secondary wall biosynthesis during wood formation in poplar.

Wood biomass is mainly made of secondary cell walls; hence, elucidation of the molecular mechanisms underlying the transcriptional regulation of secondary wall biosynthesis during wood formation will be instrumental to design strategies for genetic improvement of wood biomass. Here, we provide direct evidence demonstrating that the poplar (Populus trichocarpa) wood-associated NAC domain transcription factors (PtrWNDs) are master switches activating a suite of downstream transcription factors, and together, they are involved in the coordinated regulation of secondary wall biosynthesis during wood formation. We show that transgenic poplar plants with dominant repression of PtrWNDs functions exhibit a drastic reduction in secondary wall thickening in woody cells, and those with PtrWND overexpression result in ectopic deposition of secondary walls. Analysis of PtrWND2B overexpressors revealed up-regulation of the expression of a number of wood-associated transcription factors, the promoters of which were also activated by PtrWND6B and the Eucalyptus EgWND1. Transactivation analysis and electrophoretic mobility shift assay demonstrated that PtrWNDs and EgWND1 activated gene expression through direct binding to the secondary wall NAC-binding elements, which are present in the promoters of several wood-associated transcription factors and a number of genes involved in secondary wall biosynthesis and modification. The WND-regulated transcription factors PtrNAC150, PtrNAC156, PtrNAC157, PtrMYB18, PtrMYB74, PtrMYB75, PtrMYB121, PtrMYB128, PtrZF1, and PtrGATA8 were able to activate the promoter activities of the biosynthetic genes for all three major wood components. Our study has uncovered that the WND master switches together with a battery of their downstream transcription factors form a transcriptional network controlling secondary wall biosynthesis during wood formation.

Wood, composed of cellulose, hemicelluloses, and lignin, is the most abundant biomass produced by plants. It is an important raw forest product that has traditionally been used for myriad applications, including construction, pulping, paper making, direct burning for energy, and so on. Recently, wood from tree species, such as poplar (Populus spp.), has been proposed to be a renewable source for biofuel production (Carroll and Somerville, 2009). Therefore, there is a surge of interest in elucidating the molecular mechanisms underlying the process of wood formation in the hope of developing strategies for increasing wood biomass production and/or modifying wood composition tailored for biofuel production. Since wood biomass is mainly made of secondary walls, elucidation of the mechanisms underlying the coordinated activation of secondary wall biosynthetic genes will undoubtedly contribute to our understanding of the molecular control of wood formation.
Molecular and genomic studies in tree species have uncovered a number of wood-associated transcription factors that might be involved in the coordinated regulation of wood formation (Patzlaff et al., 2003a(Patzlaff et al., , 2003bKarpinska et al., 2004;Schrader et al., 2004;Goicoechea et al., 2005;Prassinos et al., 2005;Andersson-Gunnerå s et al., 2006;Bedon et al., 2007;Legay et al., 2007;Bomal et al., 2008;Pavy et al., 2008;Du et al., 2009;Wilkins et al., 2009;Grant et al., 2010;Zhong et al., 2010b;Robischon et al., 2011). Among them, the best-characterized ones are a set of transcription factors belonging to the NAC and MYB families. It has been shown that a group of poplar wood-associated NAC domain proteins (PtrWNDs; Zhong et al., 2010b;Zhong and Ye, 2010) and a Eucalyptus wood-associated NAC (EgWND1; Zhong et al., 2010a) are functional orthologs of the Arabidopsis (Arabidopsis thaliana) secondary wall NACs, including SND1, NST1/2, and VND6/7 . When overexpressed in Arabidopsis, these wood-associated NACs were able to induce the expression of secondary wall biosynthetic genes and a concomitant ectopic deposition of secondary walls, indicating that they are master transcriptional switches activating secondary wall biosynthesis.
PtrWNDs have been shown to be able to activate the promoter activities of a number of other woodassociated transcription factors (Zhong et al., 2010b), and two of them, PtrMYB3 and PtrMYB20, were demonstrated to be direct targets of PtrWNDs and functional orthologs of the Arabidopsis MYB46 McCarthy et al., 2010). When overexpressed in Arabidopsis, PtrMYB3/20 were also capable of activating the biosynthetic pathways of cellulose, xylan, and lignin, indicating that they function as second-level master switches in the transcriptional program regulating secondary wall biosynthesis. Several other wellcharacterized wood-associated transcription factors include the pine (Pinus taeda) PtMYB4/8 (Patzlaff et al., 2003a;Bomal et al., 2008) and the Eucalyptus EgMYB2 (Goicoechea et al., 2005). PtMYB4 and Eg-MYB2 have been shown to be functional orthologs of Arabidopsis MYB46, and overexpression of EgMYB2 in Arabidopsis resulted in ectopic deposition of secondary wall components in leaf epidermal cells (Zhong et al., 2010a). In addition, several other woodassociated transcription factors have been demonstrated to be transcriptional regulators of lignin biosynthesis (Patzlaff et al., 2003b;Legay et al., 2007;Bomal et al., 2008;Zhong and Ye, 2009). Based on these findings, it has been proposed that a transcriptional program encompassing WNDs and their downstream transcription factors is involved in the coordinated activation of secondary wall biosynthetic genes during wood formation (Zhong et al., 2010b). However, a direct proof demonstrating the existence of such a transcription program during wood formation in tree species is still lacking.
In this report, we provide direct evidence establishing that WNDs regulate the expression of a suite of their downstream transcription factors and that together they are involved in the regulation of secondary wall biosynthesis during wood formation in poplar. We show that dominant repression of PtrWND2B/6B causes a drastic reduction in secondary wall thickening in the wood of transgenic poplar and that their overexpression results in ectopic deposition of secondary walls in transgenic poplar plants. Using PtrWND overexpressors as a tool, we have uncovered a suite of wood-associated transcription factors that are up-regulated by WNDs, many of which were not reported previously and whose Arabidopsis counterparts were not known to be involved in secondary wall formation. We further demonstrate that WNDs directly bind to the secondary wall NAC-binding element (SNBE) sites present in the promoters of several of the WND-regulated transcription factors as well as a number of genes involved in secondary wall biosynthesis, cell wall modification, and programmed cell death. In addition, we reveal that, along with the previously characterized PtrMYB3/20 (McCarthy et al., 2010), several other WND-regulated transcrip-tion factors are able to activate the promoters of the biosynthetic genes for all three major wood components in a transient transactivation system, indicating that these transcription factors display a function different from their Arabidopsis counterparts. Together, our study provides a framework for further dissection of the transcriptional program regulating the biosynthesis of secondary walls during wood formation.

Reduction in Secondary Wall Thickening in Poplar Wood by Dominant Repression of PtrWND2B and PtrWND6B Functions
To investigate the effect of inhibition of PtrWND functions on wood formation, we employed the EAR dominant repression approach (Hiratsu et al., 2004) to overcome the functional redundancy of PtrWND genes. The EAR-induced dominant repression approach has been successfully applied to study the functions of transcription factors regulating secondary wall biosynthesis in Arabidopsis (Kubo et al., 2005;Mitsuda et al., 2005Mitsuda et al., , 2007Zhong et al., 2006Zhong et al., , 2008McCarthy et al., 2009). The EAR repression domain was fused with PtWND2B or PtrWND6B at the C terminus and the chimeric genes were expressed under the control of the cauliflower mosaic virus (CaMV) 35S promoter in poplar (Populus alba 3 Populus tremula) plants. At least 50 independent transgenic lines for each transgene were generated and analyzed for any potential phenotypes. Five transgenic lines with high expression of the transgene (Fig.  1A) were chosen for detailed examination of plant morphology and wood anatomy. Dominant repression of either PtrWND2B or PtrWND6B was found to cause reduced plant growth with shorter stems compared with the control transgenic plant (Fig. 1B). Examination of wood cell morphology from the mature secondary xylem taken from the bottom part of stems of 6-month-old plants revealed alterations in secondary wall thickness and vessel morphology in both PtrWND2B-DR and PtrWND6B-DR transgenic lines ( Fig. 1, C-H; Supplemental Fig. S1). Similar wood cell phenotypes were observed in five lines that were chosen for this study, and representative images are shown here. Wood from the control stems transformed with the empty vector had vessels with regular shapes and xylary fibers with thick walls (Fig. 1, C and F). In contrast, the wall thickness of xylary fibers was reduced by 56% and 68%, respectively, in the wood of PtrWND2B-DR and PtrWND6B-DR compared with the control (Fig. 1, D , E, G, and H). In addition, a slight deformation of vessels was evident in the wood of PtrWND2B-DR (Fig. 1D). However, the length of xylary fibers was not apparently shortened in the wood of PtrWND2B-DR and PtrWND6B-DR compared with the control (Supplemental Fig. S2), indicating that the reduced stem height is caused by retardation of plant growth rather than reduced cell elongation. Although the actual underlying cause for the reduced plant growth is still unknown, this phenotype is reminiscent of many Arabidopsis mutants with secondary wall defects (Zhong et al., 2005;Peñ a et al., 2007). Further examination of putative downstream genes of PtrWNDs showed that dominant repression of PtrWND2B and PtrWND6B resulted in a significant reduction in the expression of PtrMYB3, a known direct target of PtrWNDs (Zhong et al., 2010c), and several other poplar homologs of the Arabidopsis SND1 direct targets, including PtrMYB21, PtrNAC157, PtrMYB128, PtrKNAT7, PtrLBD15, and PtrNAC118 (Supplemental Fig. S3). These results directly demonstrate the essen-tial roles of PtrWNDs in secondary wall deposition during wood formation in poplar.

Ectopic Deposition of Secondary Walls by Overexpression of PtrWND2B and PtrWND6B
As a first step toward utilizing PtrWNDs as a molecular tool to dissect the transcriptional program regulating secondary wall biosynthesis during wood formation, we generated transgenic poplar plants overexpressing PtrWND2B and PtrWND6B under the control of the CaMV 35S promoter. More than 70 independent transgenic poplar lines were generated for each gene, and six lines with a high-level expression of PtrWND2B or PtrWND6B ( Fig. 2A) were selected for further character- ization. These transgenic plants showed reduced growth with shorter stems (Fig. 2B), and the leaves of the PtrWND2B-OE lines were much smaller and slightly curled compared with the control (Fig. 2C). Five transgenic lines were chosen for detailed examination of ectopic deposition of lignin, cellulose, and xylan, and representative images are shown here. Lignin staining of the cross-sections of stems revealed that in the control stems, lignin staining was only seen in the secondary xylem and the phloem fibers that were present in patches and loosely aligned into two layers (Fig. 2D). In contrast, patches of lignified cells were present in the cortical regions in the stems of the PtrWND2B-OE and PtrWND6B-OE plants (Fig. 2, E and F). Furthermore, ectopic deposition of lignin was also evident in some secondary phloem cells that are normally parenchyma- tous (Fig. 2, E and F). Examination of the stem sections with immunodetection of xylan and staining of cellulose revealed that the walls of cells with ectopic deposition of lignin in PtrWND2B-OE and PtrWND6B-OE were also heavily impregnated with xylan and cellulose (Fig. 2, G-L). These results demonstrate that ectopic overexpres-sion of PtrWNDs is sufficient to activate the secondary wall biosynthetic program in poplar, which in turn leads to the ectopic deposition of secondary walls.
The curly-leaf phenotype observed in the PtrWND2B-OE plants was most prominent in some transgenic poplar seedlings (Fig. 3, A-C) that were  3N). In contrast, no staining of lignin, xylan, and cellulose was seen in the control mesophyll cells, except for the veins in which the helical secondary thickening was evident (Fig. 3, E, F, and I-K). Consistent with the ectopic deposition of secondary walls, the expression of genes involved in the biosynthesis of wood components, including cellulose (PtrCesA4/7/8/17/18), xylan (PtrGT43A and PtrGT47C), and lignin (PtrCCoAOMT1 and PtrCOMT2), was highly induced in these PtrWND2B-OE seedlings compared with the control (Fig. 3D, right panel). These genes were previously shown to be highly expressed in developing xylem and involved in the biosynthesis of wood components (Supplemental Fig. S4A; Zhong et al., 2000;Suzuki et al., 2006;Lee et al., 2009Lee et al., , 2011. In contrast, the expression of primary wall cellulose synthase genes, including PtrCesA6 and PtrCesA15 (Supplemental Fig. S4B; Samuga and Joshi, 2004), was not up-regulated in the PtrWND2B-OE seedlings (Fig. 3D, right panel).

Identification of Downstream Transcription Factors Induced by PtrWND2B
The finding that overexpression of PtrWNDs induces ectopic deposition of secondary walls demonstrates that PtrWNDs are transcriptional master switches regulating the secondary wall biosynthetic program during wood formation and thus could be used for the identification of their downstream transcription factors involved in secondary wall biosynthesis. Since PtrWNDs are specifically expressed in wood-forming cells (Zhong et al., 2010b), we reasoned that their downstream transcription factors should also be expressed in the same cell types, and ectopic overexpression of PtrWNDs in transgenic poplar leaves would lead to an induction of the expression of these downstream transcription factors in leaves. Previous transcriptome analysis of wood-forming cells in poplar revealed a number of transcription factor genes that are preferentially expressed in developing  Transcriptional Regulation of Wood Formation wood (Wilkins et al., 2009). Gene expression analysis in the leaves of PtrWND2B-OE seedlings that exhibited ectopic deposition of secondary walls revealed that 41 transcription factors were highly induced in the PtrWND2B-OE leaves compared with the control (Figs. 4 and 5). Among them, 29 are close homologs of the Arabidopsis secondary wall-associated transcription factors induced by the master transcriptional switch, SND1. These include the MYB46 homologs (PtrMYB2/3/20/21), SND2 homologs (PtrNAC154/156), SND3 homologs (PtrNAC105/157), KNAT7 homolog (PtrKNAT7), MYB20/43 homolog (PtrMYB18), MYB42/ 85 homologs (PtrMYB75/92/125/199), MYB52/54 homologs (PtrMYB90/161/167/175), MYB58/63 homologs (PtrMYB28/192), MYB69 homologs (PtrMYB26/31/158/ 189), MYB103 homologs (PtrMYB10/128), LBD15 homolog (PtrLBD15), and XND1 homolog (PtrNAC118). Thirteen additional poplar transcription factors belonging to different families were also found to be induced by PtrWND2B. These include two NACs (PtrNAC150/151), two MYBs (PtrMYB74/121), two WRKYs (PtrWRKY12/13), one IAA (PtrIAA11), two WUSCHEL-related homeobox genes (PtrWUS1 and PtrWOX13), one BEL1-like homeobox gene (PtrBLH3), one ULTRAPETALA-like gene (PtrULT1), and two zincfinger transcription factors (PtrZF1 and PtrGATA8). None of these 13 poplar transcription factors or their Arabidopsis homologs was previously known to be involved in secondary wall biosynthesis. The finding that the expression of these transcription factors is induced by the overexpression of PtrWND2B in poplar suggests that they may be part of the transcriptional program regulating secondary wall biosynthesis during wood formation. It was interesting that overexpression of PtrWND2B also induced the expression of its own endogenous gene and several other PtrWND genes (Fig. 4), indicating that PtrWNDs can directly or indirectly up-regulate their own gene expression.
We next investigated whether the PtrWND2Binduced transcription factors are also activated by other NAC master switches, including PtrWND6B and the Eucalyptus EgWND1 (Zhong et al., 2010a). To do this, we ligated the 2-kb promoter sequences of 23 representative PtrWND2B-induced transcription factors upstream of the GUS reporter gene in an expression vector and then tested the activation of GUS reporter gene expression by PtrWND6B and EgWND1 in Arabidopsis protoplasts (Fig. 6). Although EgWND1 was able to activate the promoters of all the transcription factors tested to a level similar to PtrWND2B (Fig. 6), several gene promoters were not activated by PtrWND6B, possibly due to a lower activation strength. Nevertheless, these results indicate that the woodassociated NAC master switches regulate common downstream transcription factors, a scenario shared with the Arabidopsis secondary wall NAC master switches.
To substantiate the hypothesis that these WNDinduced transcription factors are part of the transcriptional network regulating secondary wall biosynthesis during wood formation, we examined the expression patterns of WND-regulated downstream transcription factors in the developing wood of poplar (Populus trichocarpa). A total of 21 representative genes were selected, including one gene from each group of transcription factors that are orthologs of Arabidopsis secondary wall-associated transcription factors (Fig. 4) and those that are newly identified in poplar (Fig. 5). All of them were found to be expressed in woodforming cells, including vessels, xylary fibers, and ray parenchyma cells, albeit at different levels (Fig. 7). Positive hybridization signals were also seen in the phloem fiber cells. A previous microarray study of poplar genes (http://www.bar.utoronto.ca/efppop/ cgi-bin/efpWeb.cgi; Wilkins et al., 2009) showed that all the WND-induced transcription factors are expressed in developing xylem of wood and that, among them, many display a preferential expression in the xylem (Supplemental Figs. S5 and S6). The finding that these WND-regulated downstream transcription factors exhibit an expression pattern similar to that of WNDs is consistent with the fact that their expression is up-regulated in the PtrWND overexpressors. Together, these results indicate that these PtrWNDinduced transcription factors are likely parts of the PtrWND-mediated transcriptional network regulating wood formation.

PtrWNDs Directly Bind to the SNBE Sites in the Promoters of Their Direct Targets
PtrWNDs were previously found to be capable of activating the same transcriptional program leading to secondary wall biosynthesis as is SND1 when overexpressed in Arabidopsis (Zhong et al., 2010b), indicating that PtrWNDs may bind to the same DNA elements on the target promoters as SND1 and thereby activate the expression of target genes. SND1 has been shown to specifically bind to a 19-bp semipalindromic DNA element named SNBE, and this binding is essential for SND1 activation of its target genes (Zhong et al., 2010c). To test whether PtrWNDs also bind to the SNBE sites on their direct targets, we first investigated the abilities of WNDs to activate the Arabidopsis SNBE-driven expression of the GUS reporter gene (Figs. 8 and 9). Cotransfection of the WND overexpression construct together with the MYB46 SNBE1driven GUS reporter construct in Arabidopsis protoplasts revealed that WNDs effectively activated the expression of the GUS reporter gene (Fig. 8). . Expression patterns of PtrWND-regulated transcription factors in the developing wood of wildtype poplar stems. Cross-sections of stems were hybridized with digoxigenin-labeled antisense (A-U) or sense (V) RNA probes, and the hybridization signals were detected with alkaline phosphatase-conjugated antibodies and are shown as purple color. Stem sections from three different plants were hybridized with each probe, and representative data are shown. The hybridization signals for each gene (labeled on each panel) were evident in vessels, xylary fibers, and ray parenchyma cells in the developing wood but not in the mature wood. Phloem fibers also showed positive hybridization signals. Stem sections hybridized with the control sense probes of each gene showed an absence of hybridization signals. One representative image for the sense probe of PtrNAC156 is shown (V). pf, Phloem fiber; sx, secondary xylem. Bar in A = 74 mm for A to V.
Although mutations of the noncritical nucleotides (M1) did not reduce the activation of the GUS reporter gene, mutations of the nucleotides (M2-M5) that were previ-ously shown to be critical for SND1 binding (Zhong et al., 2010c) reduced or abolished the ability of WNDs to activate GUS reporter expression (Fig. 8). Furthermore, WNDs were also able to activate GUS reporter gene expression driven by various SNBE sites from the promoters of other SND1 direct targets, including MYB83, SND3, MYB103, KNAT7, XCP1, and RNS3 (Fig. 9). These results indicate that PtrWNDs most likely bind to the SNBE sites in the promoters of their direct targets, thereby activating their expression and the secondary wall biosynthesis program during wood formation.
PtrMYB3 and PtrMYB20 have previously been shown to be direct targets of PtrWNDs, as they are directly activated by PtrWNDs in a transient system using Arabidopsis protoplasts (McCarthy et al., 2010). Analysis of the promoters of PtrMYB3 and PtrMYB20 revealed multiple SNBE sequences (Supplemental Table S1). To test the direct binding of PtrWNDs to the SNBE sites, we first performed electrophoretic mobility shift assay (EMSA) using recombinant PtrWND2B and PtrWND6B proteins and a PtrMYB3 promoter fragment (2470 to 2365) containing an SNBE site. It was found that both PtrWND2B and PtrWND6B caused specific retardation in the mobility of the PtrMYB3 promoter fragment (Fig. 10A), indicating their effective binding to the DNA fragment. In addition, the Eucalyptus EgWND1 was also able to bind to the PtrMYB3 promoter fragment, suggesting the conservation of the DNA-binding sequences of WNDs from different tree species.
To prove that WNDs bind to the PtrMYB3 promoter fragment via the SNBE site, we employed the EMSA competition assay (Fig. 10, B and C). Addition of the 19-bp SNBE sequence from the PtrMYB3 promoter fragment completely abolished the retardation of the labeled DNA fragment by WNDs, demonstrating its efficient competition with the binding of WNDs to the promoter fragment. We next examined for the presence of SNBE sites in the promoters of several other PtrWND-regulated poplar transcription factors that are close homologs of SND1 direct targets. We reasoned that poplar close homologs of SND1 direct targets are strong candidates of PtrWND direct targets due to the conservation of the DNA-binding elements for both SND1 and PtrWNDs. These PtrWND-regulated downstream transcription factors include PtrNAC105/ 157 (SND3 homologs), PtrMYB10/128 (MYB103 homologs), PtrKNAT7 (KNAT7 homolog), PtrLBD15 (LBD15 homolog), and PtrNAC118 (XND1 homolog). The promoters of all these genes and their Eucalyptus close homologous genes also contain various numbers of SNBE sites (Supplemental Tables S1 and S2). EMSA analysis with one representative SNBE arbitrarily chosen from each of these poplar genes demonstrated that they effectively competed with the binding of WNDs to the labeled PtrMYB3 promoter fragment (Fig. 10, B and C; Supplemental Fig. S7) as well as exhibited efficient binding by PtrWND2 (Supplemental Fig. S8), indicating that these poplar genes are direct targets of PtrWNDs. Direct activation of these transcription factors The GUS reporter constructs consist of the GUS reporter gene driven by three copies of the wild-type or mutated Arabidopsis MYB46 SNBE1 sequence. The bottom panel depicts the WND-activated expression of the GUS reporter gene driven by the wild-type or the mutated MYB46 SNBE1 sequence as indicated at left. The wild-type SNBE sequence was mutated by altering all the noncritical nucleotides (M1) or changing one or more critical nucleotides (M2-M5; Zhong et al., 2010c). The GUS activity in the protoplasts transfected with the GUS reporter construct only is set to 1. Error bars represent the SE of three biological replicates. by PtrWNDs was verified by the steroid receptor-based inducible system, showing that upon estradiol treatment in the presence of the protein synthesis inhibitor cycloheximide, PtrWND2B was still able to induce the reporter gene expression driven by the promoters of these PtrWND-regulated transcription factors (Supplemental Fig. S9).

PtrWNDs Activate the Expression of Genes Involved in Secondary Wall Biosynthesis, Cell Wall Modification, and Programmed Cell Death
A number of genes involved in secondary wall biosynthesis, cell wall modification, and programmed cell death have been previously shown to be direct targets of SND1 (Zhong et al., 2010c). Therefore, we next investigated whether the poplar close homologs of some of these genes are activated by overexpression of PtrWNDs. Real-time quantitative PCR analysis revealed a high level of induction in the expression of all the genes examined in the leaves of PtrWND2B overexpressor seedlings (Fig. 11A). These genes include those involved in secondary wall biosynthesis, such as three xylan-specific glycosyltransferases (PtrGT43C, PtrGT47A, and PtrGT8F), a laccase (PtrLAC3), a UDP-Xyl synthase (PtrUXS5), and a peroxidase (PtrPO3), those involved in cell wall modification, such as a pectin methylesterase (PtrPME1) and a pectinase (PtrPEC1), and those involved in programmed cell death, such as two Cys proteases (PtrXCP1 and PtrXCP2), an aspartyl protease (PtrASP1), an RNase (PtrRNS3), and a nuclease (PtrBFN1). Examination of the 1.5-kb promoter sequences of these genes and their Eucalyptus close homologous genes revealed the presence of multiple putative SNBE sites (Supplemental Tables S1 and S2). Representative SNBE sites (Fig. 11B) from the promoters of six respective poplar genes were tested for their ability to compete with the binding of WNDs to the PtrMYB3 promoter fragment using EMSA competition analysis. It was found that these SNBE sequences effectively competed the binding of WNDs to the labeled PtrMYB3 promoter fragment ( Fig. 11C; Supplemental Fig. S7) as well as exhibited efficient binding by PtrWND2 (Supplemental Fig. S8), indicating that these secondary wall/ programmed cell death-related genes are also direct targets of PtrWNDs.

Activation of the Secondary Wall Biosynthetic Genes by WND-Regulated Downstream Transcription Factors
To further elucidate the functional roles of WNDregulated downstream transcription factors, we investigated whether any of them might be involved in the activation of secondary wall biosynthetic genes specific to cellulose, xylan, or lignin. This was done by cotransfecting the effector construct containing the WND-regulated transcription factor under the control of the CaMV 35S promoter together with the reporter construct containing the GUS reporter gene driven by the secondary wall biosynthetic gene promoter in Arabidopsis protoplasts (Fig. 12). The activation of the promoters of representative poplar secondary wall biosynthetic genes, including those for cellulose (PtrCesA4/8/17), xylan (PtrGT43A and PtrGT47C), and lignin (PtrCCoAOMT1 and PtrCOMT2), was first tested with the known secondary wall master switches from poplar and Eucalyptus. All of these master switches, including PtrWND2B/6B, EgWND1, PtrMYB3, and EgMYB2, were able to activate the expression of the GUS reporter gene driven by the promoters of genes involved in the biosynthesis of cellulose, xylan, and lignin (Fig. 12), demonstrating the usefulness of these promoters for the transactivation analysis. Transactivation analysis with WNDregulated transcription factors revealed that a number Figure 9. Activation of the SNBE sites from a number of Arabidopsis SND1 direct targets by WNDs. The ability of WNDs to activate the SNBE sites was tested by cotransfecting Arabidopsis leaf protoplasts with the GUS reporter and the effector constructs as described in Figure  8. The GUS reporter constructs consist of the GUS reporter gene driven by three copies of the SNBE sequences from the promoters of various SND1 direct targets. The GUS activity in the protoplasts transfected with the GUS reporter construct only (control) is set to 1. Error bars represent the SE of three biological replicates.
of them were capable of activating the promoters of secondary wall biosynthetic genes. It was found that several of them, including PtrMYB18, PtrMYB75, PtrMYB128, PtrNAC150, PtrMYB74, PtrMYB121, PtrZF1, and PtrGATA1, were able to activate the promoters of genes for all three secondary wall biosynthetic pathways (Fig. 12). PtrNAC156 was shown to activate the promoters of PtrCesA8 and PtrCCoAOMT1, whereas PtrNAC157 only activated that of PtrCesA8. In addition, while PtrMYB26 activated both lignin biosynthetic genes tested, PtrMYB90 only activated PtrCCoAOMT1 (Fig. 12). Furthermore, PtrLBD15 was shown to slightly activate the promoters of several genes. Since the promoters of secondary wall biosynthetic genes used for the analysis were from poplar, their activation by poplar WNDregulated transcription factors should reflect the regulatory activities of these transcription factors on secondary wall biosynthetic genes. These findings suggest the involvement of an array of transcription factors in the regulation of secondary wall biosynthesis, some of which act as key activators of all three biosynthetic pathways of secondary wall components. They further indicate the complexity of the transcriptional program regulating wood formation in tree species.

DISCUSSION
Wood formation involves a series of complex developmental processes, including cambial cell division and differentiation into xylem cells, cell elongation, secondary wall deposition, programmed cell death, and, finally, heartwood maturation (Plomion et al., 2001). Although past studies have identified a number of transcriptional regulators involved in the control of cambial cell activity, xylem cell differentiation, and secondary wall biosynthesis in tree species, our understanding of the underlying regulatory mechanisms of wood formation is still in its infancy. Our direct proof in transgenic poplar that PtrWNDs are master transcriptional switches controlling secondary wall biosynthesis during wood formation, together with our finding that they regulate a suite of downstream transcription factors implicated in wood formation, provide foundation knowledge for further elucidation of the transcriptional program controlling secondary wall biosynthesis during wood formation in tree species.
We have demonstrated that PtrWNDs activate the expression of a suite of downstream transcription Figure 11. Overexpression of PtrWND2B induces the expression of genes involved in secondary wall biosynthesis, cell wall modification, and programmed cell death in poplar. A, Real-time quantitative PCR analysis showing the induction in the expression of genes involved in secondary wall biosynthesis, cell wall modification, and programmed cell death in the leaves of PtrWND2B overexpressors. The expression of each gene in the control is set to 1. Error bars represent the SE of three biological replicates. B, Representative SNBE sequences from the promoters of several PtrWND-induced genes that are involved in secondary wall biosynthesis, cell wall modification, and programmed cell death. These SNBE sequences were used for the EMSA competition analysis in C. The consensus nucleotides in the SNBE sequences are shaded. The number shown at the left of each sequence is the position of the first nucleotide relative to the start codon. C, SNBE sequences from the promoters of several PtrWND-regulated genes efficiently competed with the binding of PtrWND2B (top panel), PtrWND6B (middle panel), and EgWND1 (bottom panel) to the labeled PtrMYB3 promoter fragment. D, Transactivation analysis of the SNBE sequences using the GUS reporter gene. The GUS reporter and the effector constructs are shown in the left panel. The GUS reporter constructs are made of the GUS reporter gene driven by two copies of the SNBE sequences from the promoters of various PtrWND targets (B). Activation of the SNBE sequences by WNDs was tested in Arabidopsis protoplasts by cotransfecting the GUS reporter and the effector constructs (right panel). The GUS activity in the protoplasts transfected with the GUS reporter construct only (control) is set to 1. Error bars represent the SE of three biological replicates.
factors, including those that are close homologs of the Arabidopsis SND1 downstream transcription factors. In addition, the Eucalyptus EgWND1 is also able to activate the promoters of these PtrWND-regulated transcription factors. These findings indicate that the major transcriptional programs regulating secondary wall biosynthesis are conserved in vascular plants. However, since wood anatomy and composition vary among different plant species, it is expected that some variations in the transcriptional regulation of secondary wall biosynthesis might have evolved. Indeed, we have found that some of the poplar wood-associated transcription factors have diversified their transcriptional regulatory activities from those of their Arabidopsis counterparts. For example, Arabidopsis MYB103 was shown to preferentially induce the expression of genes for the biosynthesis of cellulose but not xylan and lignin (Zhong et al., 2008), whereas its poplar homolog, PtrMYB128, was able to activate the promoters of the biosynthetic genes for all three secondary wall components in a transient transactivation system using Arabidopsis protoplasts.
In Arabidopsis, only two groups of transcription factors, namely secondary wall NACs (including SND1, NST1/2, and VND6/7) and their direct targets MYB46/83, have been shown to be master switches activating all three secondary wall biosynthetic pathways (Zhong et al., 2010a). In poplar, we revealed, to our knowledge for the first time, that a set of additional transcriptional factors, such as PtrNAC150, PtrNAC156, PtrNAC157, PtrMYB18, PtrMYB74, PtrMYB75, PtrMYB121, PtrMYB128, PtrZF1, and PtrGATA1, are able to activate the promoters of genes for all three secondary wall biosynthetic pathways in a transient transactivation system using Arabidopsis protoplasts (Fig. 13). Although further functional characterization of these transcription factors is needed to establish their roles in wood formation in poplar, it is intriguing to find that so many poplar wood-associated transcription factors are capable of activating all three secondary wall biosynthetic pathways. It is conceivable that poplar evolved to have a much more complex transcriptional network consisting of additional multiple levels of master controls to ensure secondary wall biosynthesis during wood formation, which requires the deposition of massive amounts of secondary wall components.
Among the PtrWND-regulated downstream transcription factors, PtrMYB26 and PtrMYB90 were found to specifically activate the promoters of lignin biosynthetic genes in a transient transactivation system using Arabidopsis protoplasts (Fig. 13), which is similar to PtrMYB28, which was previously shown to regulate lignin biosynthesis . Several other transcription factors, including PtrNAC156, PtrNAC157, and PtrLBD15, were shown to only activate the promoters of one or more specific genes tested in a transient transactivation system using Arabidopsis protoplasts. These findings indicate that some transcription factors may only regulate the expression of one or a few secondary wall biosynthetic genes instead of all genes in a specific biosynthetic pathway. Although many PtrWND-regulated downstream transcription factors did not activate the promoters of secondary wall biosynthetic genes, they might be involved in finetuning the transcriptional program through cooperative interaction with other transcription factors that activate secondary wall biosynthetic genes. These findings reflect the complexity of the transcriptional regulation of wood formation in tree species.
We have further revealed that PtrWNDs bind to the SNBE sites in the promoters of their direct targets and thereby directly activate their expression. The fact that WNDs from both poplar and Eucalyptus directly bind to and activate the consensus SNBE sites like Arabidopsis SND1 is consistent with the previous finding that PtrWNDs and EgWND1 were able to activate the secondary wall biosynthetic program when overexpressed in Arabidopsis (Zhong et al., 2010a(Zhong et al., , 2010b. The findings from these studies indicate that the secondary wall NAC transcriptional switches positioned at the top of the transcriptional network are highly conserved in their recognition of the binding elements. It is likely that these secondary wall NAC master switches were coopted for the regulation of secondary wall biosynthesis when vascular plants first evolved during the Silurian period, since their close homologs exist in seedless vascular plants such as Selaginella moellendorffii (Zhong et al., 2010a). After these secondary wall NACs were recruited to function in the regulation of secondary wall biosynthesis, there might have been little room left for revamping their DNA-binding elements, because of the essential roles of secondary walls in plant structure and function.
In summary, we have provided direct evidence demonstrating that PtrWNDs are master transcriptional switches controlling secondary wall biosynthesis during wood formation in poplar, and we have revealed that they regulate a suite of wood-associated transcription factors, a number of which are able to activate the expression of the biosynthesis genes for cellulose, xylan, and lignin. Our study indicates that although the basic secondary wall NAC-regulated transcriptional programs in the herbaceous Arabidopsis and the tree species exhibit conservation, the tree species evolved to have a more complex regulatory network controlling secondary wall biosynthesis than Arabidopsis ( Fig. 13; Zhong et al., 2010a). The identification of these WND-regulated transcription factors involved in secondary wall biosynthesis will also facilitate the functional characterization of many wood-associated transcription factors previously revealed in different tree species through transcriptome profiling (Bedon et al., 2007;Pavy et al., 2008). Further deciphering the interrelationships of the PtrWNDregulated downstream transcription factors and their functional roles in regulating the secondary wall biosynthetic program will provide new insight into the molecular mechanisms controlling wood formation in tree species. Because wood biomass is naturally recalcitrant for biofuel production, uncovering the transcriptional program regulating secondary wall biosynthesis during wood formation will potentially present novel strategies to engineer wood composition suited for a more efficient production of biofuels. For example, once the key transcriptional factors specifically regulating individual secondary wall biosynthetic pathways are identified, it would be possible to use one or a few transcription factors to custom design wood composition with an increased amount of cellulose, the major polysaccharide potentially used for the production of bioethanol.

Dominant Repression and Overexpression of PtrWNDs
The PtrWND dominant repression constructs (PtrWND2B-DR and PtrWND6B-DR) were generated by fusing the full-length PtrWND2B or Figure 13. Diagram of the transcriptional regulatory network controlling wood formation. The WNDs in poplar and Eucalyptus have been shown to function as the first-level master switches that directly activate a number of downstream transcription factors as well as many genes involved in secondary wall biosynthesis, cell wall modification, and programmed cell death. PtrMYB3 and PtrMYB20, direct targets of PtrWNDs, act as second-level master switches activating a secondary wall biosynthetic program (McCarthy et al., 2010). EgMYB2 from Eucalyptus and PtMYB4 from pine are functional orthologs of PtrMYB3/ 20. Among the PtrWND-regulated transcription factors, a number of them have been demonstrated to be able to activate the promoters of genes for all three secondary wall biosynthetic pathways, and a few others only act on the promoters of lignin biosynthetic genes in a transient transactivation system. In Arabidopsis, only the NAC and MYB master switches are known to activate all three secondary wall biosynthetic pathways. Note that for simplicity, many additional PtrWNDregulated transcription factors as listed in Figures  PtrWND6B cDNA in frame with the dominant EAR repression sequence (Hiratsu et al., 2004), which was ligated downstream of the CaMV 35S promoter in pBI121 (Clontech). The PtrWND overexpression constructs (PtrWND2B-OE and PtrWND6B-OE) were created by ligating the full-length PtrWND2B or PtrWND6B cDNA downstream of the CaMV 35S promoter in pBI121. These constructs were introduced into poplar (Populus alba 3 Populus tremula) by Agrobacterium tumefaciens-mediated transformation as described by Leple et al. (1992). The transgenic poplar seedlings were selected on Murashige and Skoog medium containing 50 mg L 21 kanamycin and 500 mg L 21 carbenicillin. After rooting, transgenic seedlings were transferred to soil and grown in a greenhouse. The control plants were transgenic poplar plants transformed with an empty vector. For each construct, at least 50 independent transgenic lines were used for morphological and histological analyses.

Gene Expression Analysis
Total RNA was isolated from plant tissues with a Qiagen RNA isolation kit (Qiagen). First-strand cDNAs were synthesized from total RNA treated with DNase I and then used as a template for PCR analysis. The expression level of a poplar actin gene served as an internal control to determine the reverse transcription (RT)-PCR amplification efficiency among different samples. Various PCR cycles were performed to determine the logarithmic phase of amplifications for the samples. RT-PCR was repeated three times, and similar results were obtained. Real-time quantitative PCR was performed with the QuantiTect SYBR Green PCR kit (Clontech) using first-strand cDNAs as templates. The relative expression level of each gene was calculated by normalizing the PCR threshold cycle number of each gene with that of a poplar actin reference gene. The data shown are averages of three biological replicates.

Histology
Tissues were fixed in 2% formaldehyde and embedded in low-viscosity (Spurr's) resin (Electron Microscopy Sciences) as described (Burk et al., 2006). Sections (1 mm thick) were cut with a microtome and stained with toluidine blue for light microscopy. For transmission electron microscopy, 85-nm-thick sections were cut, poststained with uranyl acetate and lead citrate, and observed using a Zeiss EM 902A transmission electron microscope (Carl Zeiss). Lignin was examined by staining the sections with phloroglucinol-HCl or visualized using a UV fluorescence microscope . Secondary wall cellulose staining was done by incubating 1-mm-thick sections with 0.01% Calcofluor White (Hughes and McCully, 1975). Under the conditions used, only secondary walls exhibited brilliant fluorescence. Xylan was detected by using the monoclonal LM10 antibody against xylan and fluorescein isothiocyanate-conjugated goat anti-rat secondary antibodies according to McCartney et al. (2005).

In Situ Hybridization
Young wild-type poplar (Populus trichocarpa) stems were fixed in 2.5% formaldehyde and 0.5% glutaraldehyde, embedded in paraffin, and sectioned (12 mm thick) for in situ mRNA localization according to McAbee et al. (2005) and Zhou et al. (2007). The 200-bp 3# untranslated sequences of poplar transcription factor cDNAs were used for the synthesis of digoxigenin-labeled antisense and sense RNA probes with the DIG RNA Labeling Mix (Roche). Stem sections were hybridized with the antisense and sense probes, and the hybridization signals were detected by incubation with alkaline phosphataseconjugated antibodies against digoxigenin and subsequent color development with alkaline phosphatase substrates.

Transactivation Analysis
For testing the ability of WNDs to activate the downstream transcription factor gene promoters, the reporter construct containing the GUS reporter gene driven by a 2-kb promoter of the poplar gene of interest and the effector construct containing WNDs driven by the CaMV 35S promoter were cotransfected into Arabidopsis (Arabidopsis thaliana) leaf protoplasts (Sheen, 2001). For examining the ability of WNDs to activate the SNBE sites, the reporter construct containing the GUS reporter gene driven by three copies of various SNBE sequences and the effector construct containing WNDs driven by the CaMV 35S promoter were cotransfected into Arabidopsis leaf protoplasts. For testing the ability of WND-regulated transcription factors to activate the promoters of secondary wall biosynthetic genes, the reporter construct containing the GUS reporter gene driven by a 2-kb promoter of the poplar gene of interest and the effector construct containing various WND-regulated transcription factors driven by the CaMV 35S promoter were cotransfected into Arabidopsis leaf protoplasts. Another construct containing the firefly luciferase gene driven by the CaMV 35S promoter was included in each transfection for the determination of transfection efficiency. After 20 h of incubation, protoplasts were lysed and the supernatants were subjected to assay of the GUS and luciferase activities (Gampala et al., 2001). The GUS activity was normalized against the luciferase activity in each transfection, and the data shown are averages of three biological replicates.
For testing the direct activation of the promoters of transcription factors by PtrWND2, the PtrWND2-HER expression construct (McCarthy et al., 2010) was cotransfected with the 2-kb promoter-driven GUS reporter construct (Fig.  6) into Arabidopsis leaf protoplasts. The transfected protoplasts were treated with estradiol and cycloheximide and analyzed for gene expression with quantitative PCR as described previously (Zhong et al., 2008).

EMSA
PtrWND2B, PtrWND6B, and EgWND1 were fused in frame with the maltose-binding protein (MBP) and expressed in Escherichia coli. The recombinant WND-MBP protein was purified using amylose resin and then used for EMSA with the PtrMYB3 promoter fragments. The PtrMYB3 promoter fragments were PCR amplified and biotin labeled at the 3# end. The biotin-labeled DNA fragments were incubated with 100 ng of WND-MBP in the binding buffer [10 mM Tris, pH 7.5, 50 mM KCl, 1 mM dithiothreitol, 2.5% glycerol, 5 mM MgCl 2 , 0.05% Nonidet P-40, and 100 ng mL 21 poly(dI-dC)]. For competition analysis, unlabeled promoter fragments or oligonucleotides were included in the binding reactions as competitors in 120-fold molar excess relative to the labeled probes. The WND-bound DNA probes were separated from the unbound ones by PAGE. The DNA was electroblotted onto nitrocellulose membranes and detected by the chemiluminescence method.

Statistical Analysis
The experimental data from the quantitative PCR analysis and GUS activity assay were subjected to statistical analysis using the Student's t test program (http://www.graphpad.com/quickcalcs/ttest1.cfm), and the quantitative difference between the two groups of data for comparison in each experiment was found to be statistically significant (P , 0.001).
The GenBank accession numbers for the genes used in this study are

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Cross-sections of mature secondary xylem from the stems of 6-month-old plants showing reduced wall thickness in xylary fibers and slightly deformed vessel morphology in PtrWND2B-DR (B) and PtrWND6B-DR (C) compared with the control (A).
Supplemental Figure S2. Longitudinal sections of the mature secondary xylem from the stems of 6-month-old plants showing the length of xylary fibers in PtrWND2B-DR (B) and PtrWND6B-DR (C) compared with the control (A).
Supplemental Figure S3. Quantitative PCR analysis of the expression level of poplar homologs corresponding to the Arabidopsis SND1 direct targets in the PtrWND2B-DR and PtrWND6B-DR lines.
Supplemental Figure S4. Transcript abundance of poplar cellulose synthase genes.
Supplemental Figure S5. Transcript abundance of PtrWND-regulated transcription factors in seedlings, young leaves, roots, and xylem.
Supplemental Figure S6. Transcript abundance of additional PtrWNDregulated transcription factors in seedlings, young leaves, roots, and xylem.
Supplemental Figure S7. EMSA competition analysis of PtrWND2B binding to the labeled PtrMYB3 promoter fragment by different concentrations of SNBE sequences from the promoters of PtrWND-regulated genes.
Supplemental Figure S8. EMSA analysis of PtrWND2B binding to the labeled SNBE sequences from the promoters of PtrWND-regulated genes.
Supplemental Figure S9. Direct activation of the promoters of PtrWNDregulated transcription factors (TFs) by PtrWND2B.
Supplemental Table S1. List of putative SNBE sequences in the 1.5-kb promoters of poplar PtrWND-regulated genes.
Supplemental Table S2. List of putative SNBE sequences in the 1.5-kb promoters of Eucalyptus (Eucalyptus grandis) close homologs of PtrWND direct targets.