MYB Transcription Factor161 Mediates Feedback Regulation of Secondary wall-associated NAC-Domain1 Family Genes for Wood Formation.

Wood formation is a complex process that involves cell differentiation, cell expansion, secondary wall deposition, and programmed cell death. We constructed a four-layer wood formation transcriptional regulatory network (TRN) in Populus trichocarpa (black cottonwood) that has four Secondary wall-associated NAC-Domain1 (PtrSND1) transcription factor (TF) family members as the top-layer regulators. We characterized the function of a MYB (PtrMYB161) TF in this PtrSND1-TRN, using transgenic P. trichocarpa cells and whole plants. PtrMYB161 is a third-layer regulator that directly transactivates five wood formation genes. Overexpression of PtrMYB161 in P. trichocarpa (OE-PtrMYB161) led to reduced wood, altered cell type proportions, and inhibited growth. Integrative analysis of wood cell-based chromatin-binding assays with OE-PtrMYB161 transcriptomics revealed a feedback regulation system in the PtrSND1-TRN, where PtrMYB161 represses all four top-layer regulators and one second-layer regulator, PtrMYB021, possibly affecting many downstream TFs in, and likely beyond, the TRN, to generate the observed phenotypic changes. Our data also suggested that the PtrMYB161's repressor function operates through interaction of the base PtrMYB161 target-binding system with gene-silencing cofactors. PtrMYB161 protein does not contain any known negative regulatory domains. CRISPR-based mutants of PtrMYB161 in P. trichocarpa exhibited phenotypes similar to the wild type, suggesting that PtrMYB161's activator functions are redundant among many TFs. Our work demonstrated that PtrMYB161 binds to multiple sets of target genes, a feature that allows it to function as an activator as well as a repressor. The balance of the two functions may be important to the establishment of regulatory homeostasis for normal growth and development.

undetermined in vivo. While these three members can activate PtrMYB021, we do not know 274 whether they can also activate PtrMYB074 or whether the activation is by direct  First, we determined whether PtrSND1-A1, PtrSND1-A2, and PtrSND1-B2 can activate 277 PtrMYB074. To do that, we transfected SDX protoplasts to overexpress each of these PtrSND1 278 members (pUC19-35S-PtrSND1-35S-sGFP) and analyzed the transcript abundance of 279 PtrMYB074 by RT-qPCR after 12 h of protoplast incubation. Protoplasts transfected with a 280 pUC19-35S-sGFP plasmid were used as an empty-vector control. The results showed that each 281 of the three PtrSND1 member genes could activate the expression of PtrMYB074 ( Figure 6A). 282 Therefore, all four PtrSND1 family members, PtrSND1-A1, -A2, -B1 and -B2, can activate the 283 expression of both PtrMYB021 and PtrMYB074. 284 Next, we analyzed whether PtrSND1-A1, PtrSND1-A2, and PtrSND1-B2 can directly 285 transactivate PtrMYB021 and PtrMYB074 by testing if each of the SND1 members can bind to 286 the promoter of PtrMYB021 and PtrMYB074, using anti-GFP antibody ChIP analysis in SDX 287 protoplasts (Chen et al., 2019;Li et al., 2019;Yeh et al., 2019). We isolated SDX protoplasts and 288 transfected a portion of the protoplasts with a plasmid DNA (pUC19-35S-PtrSND1 member-289 sGFP) for overexpressing PtrSND1 member-sGFP. Another portion of the protoplasts was 290 transfected with a pUC19-35S-sGFP plasmid as a mock control. Following crosslinking and 291 anti-GFP antibody purification, qPCR was performed for four fragments of the ~2,000 bp 292 chromatin DNA fragments (promoter sequences) upstream of the coding region of each tested 293 PtrMYB genes, with expected qPCR products ranging from 80 to 200 bp ( Figure 6B).  analysis of specific chromatin enrichments in protoplasts overexpressing the PtrSND1 member-295 sGFP fusion indicated that each of these three PtrSND1 members could bind to the ~2,000 bp 296 promoter in at least one location ( Figure 6C and Supplemental Figure 3). Previously, we had 297 shown that GFP fusion did not affect their native transactivation functions (Chen et al., 2019). 298 Therefore, all four PtrSND1 members can directly transactivate PtrMYB021 and PtrMYB074, 299 extending the previously established PtrSND1-B1 TRN into a PtrSND1 family-mediated TRN 300 ( Figure 6D). 301 302

PtrMYB161 Mediated Direct Feedback Repression of Four PtrSND1 Member Genes and 303
PtrMYB021 but Not PtrMYB074 304 As suggested above, PtrMYB161 in OE-PtrMYB161 transgenic lines may negatively regulate 305 higher level TFs, such as PtrSND1s, initiating a cascade of feedback regulation for a broad 306 suppression of cell-wall component genes ( Figure 5 and Supplemental Dataset 1). We then 307 examined RNA-seq results and conducted RT-qPCR analysis of transcripts in SDX tissues of 308 OE-PtrMYB161-L8 and the wild-type. These analyses revealed that overexpression of 309 PtrMYB161 repressed the expression of the four PtrSND1 members, as well as PtrMYB021 and 310 PtrMYB074 ( Figure 7A and Supplemental Figure 4A for RNA-seq data). The results support our 311 hypothesis that PtrMYB161 acts as a feedback repressor in the PtrSND1 family-mediated TRN 312 for wood formation. We next tested whether these repressions are direct or indirect regulatory 313 effects. 314 We performed anti-GFP antibody ChIP, as we did above, using SDX protoplasts constitutively 315 expressing a PtrMYB161-sGFP fusion (pUC19-35S-PtrMYB161-sGFP). Protoplasts transfected 316 with a pUC19-35S-sGFP plasmid were used as the control. Four chromatin fragments in the ~2-317 kb promoter of each PtrSND1 member (P1 to P4 in Figure 7B) and PtrMYB021 and PtrMYB074 318 (P1 to P4 in Figure 6B) were amplified (by RT-qPCR) from the transfected protoplasts following 319 ChIP. We detected 2-to 12-fold enrichment of at least one sequence fragment within the ~2-kb 320 promoter of each PtrSND1 member genes and PtrMYB021 ( Figure 7C), indicating that 321 PtrMYB161 can bind to these upstream cell-wall TF genes for a direct feedback repression in the 322 wood formation TRN ( Figure 7D). The ChIP analysis suggests that PtrMYB161 can not directly 323 repress PtrMYB074 (Supplemental Figure 4B). The repressed transcript abundance of 324 PtrMYB074 observed in OE-PtrMYB161 lines ( Figure 7A) was most likely an indirect effect 325 through the repression of PtrSND1 members by PtrMYB161. 326 There is growing evidence that transcriptional repression events may result from histone 327 deacetylation (Shahbazian and Grunstein, 2007;Zentner and Henikoff, 2013;Liu et al., 2014;328 Gao et al., 2015;Han et al., 2016;Li et al., 2017;Zhang et al., 2018;Park et al., 2019;Ueda and 329 Seki, 2020;Zeng et al., 2020). We hypothesized that the repression mechanism of PtrMYB161 in 330 P. trichocarpa might involve corepressors such as histone deacetylase to effect transcriptional 331 repression via histone deacetylation. To test this hypothesis, we analyzed the expression of all 332 known P. trichocarpa histone deacetylase genes in OE-PtrMYB161 transgenics using the RNA-333 seq of SDX tissues. In all the OE-PtrMYB161 transgenic lines, the three epigenetic regulators in 334 the PtrHDT3 family (also known as the HD2C family) (Dangl et al., 2001;Pandey et al., 2002) 335 were drastically up-regulated ( Figure 8 and Supplemental Dataset 2), suggesting that these three 336

Wood Structure, and Wood Cell-Wall Composition 341
We demonstrated, through transgenic plant overexpression, that PtrMYB161 is a strong negative 342 regulator that affects plant growth, cell-wall biosynthesis, and wood formation (Figures 3,4,5 343 and Tables 1, 2). To further test PtrMYB161's regulatory roles, we generated loss-of-function 344 mutants of PtrMYB161 in P. trichocarpa using CRISPR-based genome editing with 345 Streptococcus pyogenes Cas9 (Deltcheva et al., 2011;Jiang et al., 2013;Heler et al., 2015). An 346 sgRNA was designed to target SNP-free regions in PtrMYB161's exon II (see Methods; Figure  347 9A). The sgRNA was cloned into the pEgP237- 2A-GFP vector (Osakabe et al., 2016;Ueta et al., 348 2017) for CRISPR-editing using our improved P. trichocarpa transformation protocol (Song et 349 al., 2006;Li et al., 2017;2019). We generated two independent biallelic mutants, ptrmyb161-2 350 and ptrmyb161-3, with each containing heterozygous edits for PtrMYB161 ( Figure 9B). These 351 mutants were clonally propagated for further analysis. 352 The ptrmyb161 lines had slightly improved growth in height, internode number, and stem 353 diameter but had similar internode lengths compared to wildtype ( Figure 9C). These growth 354 effects were sustained for the period (~4 months) of the study ( Figure 9D, Supplemental Figure  355 5). The morphology, size, and wall thickness of the three main stem xylem cells (fibers, vessels, 356 and rays) in the ptrmyb161 lines were similar to those in wildtype plants (Supplemental Figure 6). 357 Loss-of-function of PtrMYB161 in P. trichocarpa also did not significantly alter the lignin 358 quantity, lignin S/G ratio, and neutral sugar composition and contents in wood (Supplemental  359   Tables 1 and 2

Family-Mediated Transregulation in P. trichocarpa 363
We performed RNA-seq and RT-qPCR to examine SDX gene expression in ptrmyb161-2 and 364 ptrmyb161-3 lines. We found that mutation of PtrMYB161 in ptrmyb161-2 did not cause a 365 substantial change in the transcript abundances of the PtrSND1 members, PtrMYB021 and 366 PtrMYB074-the higher-level transregulators in the PtrSND1 family-mediated TRN 367 (Supplemental Figure 7A). The overall effect of PtrMYB161 mutation on the PtrSND1-mediated 368 wood formation TRN appeared to be minor in both ptrmyb161-2 and ptrmyb161-3 because the 369 RNA-seq results showed that transcript levels of monolignol, cellulose, and hemicellulose 370 biosynthetic pathway genes in these mutants were similar to those in the wildtype (Supplemental 371 In this study, we continued to explore the PtrSND1-B1 mediated TRN for wood formation in P. 380 trichocarpa. We revealed that the PtrSND1-B1 TRN in P. trichocarpa is mediated not only by 381 PtrSND1-B1 but also by all PtrSND1 family members ( Figure 6). The PtrSND1 members-382 mediated transregulation (of PtrMYB021 and PtrMYB074) was established in a P. trichocarpa 383 SDX protoplast system and the regulatory effect (direct transactivation) was validated by 384 PtrSND1 TF-sGFP fusion proteins in the system using anti-GFP antibody ChIP ( Figure 6). In our 385 previous studies, we used the same transregulation system to discover TF-DNA interactions for 386 wood formation and histone acetylation-mediated drought tolerance responses in P. trichocarpa 387 (Lin et al., 2013;Chen et al., 2019;Li et al., 2019;Yeh et al., 2019). In these studies, we 388 demonstrated that GFP-tagged TFs retain their native transactivation function, and tested 82 389 protoplast-inferred TF-DNA interactions in 68 genotypes of transgenic and CRISPR-edited P. and PtrMYB074-the six regulators of the top two layers of the TRN ( Figure 7D). The 400 transgenesis combined with ChIP assays using the protoplast system suggested that five of these 401 six regulators were directly repressed by PtrMYB161 (Figure 7) In Arabidopsis, AtSND1 directly regulates AtMYB46 (the homolog of PtrMYB021, (Zhong et 406 al., 2007;Chen et al., 2019)), which directly regulate AtMYB4, AtMYB7, and AtMYB32, three 407 MYB repressors (Ko et al., 2009;Chen et al., 2019). In this regulatory network, the three MYBs 408 were also demonstrated for functions as negative feedback regulators of AtSND1 ( Overexpression of PtrMYB161 decreased the cambial activities (wood differentiation), leading to 475 a reduced secondary xylem (wood) area ( Figure 4A and Supplemental Figure 2C). Therefore, 476 PtrMYB161 overexpression-induced high PtrHDT3 levels may repress wood differentiation 477 genes through histone deacetylation in the chromatin of such genes, causing reduced stem 478 secondary growth. High expression levels of epigenetic regulators HD2s (putative orthologs of 479 HDT3s) are known to provoke developmental abnormalities in Arabidopsis (Zhou et al., 2004). Our study suggested that high levels of PtrMYB161 may act as signals triggering a cascade of 494 regulations affecting adversely wood cell-wall formation and plant growth (Supplemental 495 Dataset 4). However, deletion of PtrMYB161 using CRISPR-Cas9 could not enhance these 496 developmental traits. In fact, the edited ptrmyb161-2 and ptrmyb161-3 mutants were similar to 497 the wildtype in all aspects examined, such as cell-wall structure, morphology and chemical 498 composition, and other growth developmental features (Figure 9,Supplemental Figures 5,6,and 499 Supplemental Tables 1, 2). The phenotypic similarity between the mutants and the wildtype 500 comports with the full transcriptome analysis showing that deleting PtrMYB161 had little effect 501 on the transcriptome. Only 1.7 to 4.2% of the expressed genes were altered for their transcript 502 levels in ptrmyb161-2 and ptrmyb161-3, respectively (Supplemental Table 3   The eighth stem internodes of P. trichocarpa were harvested for fixation with FAA solution 580 (50% ethanol, 5% acetic acid, 3.7% formaldehyde) (v/v). After dehydration, the fixed tissues 581 were embedded in paraffin (Sigma) and sectioned to a thickness of 10 µm using a rotary 582 microtome (Leica RM2245). The 175 bp region of PtrMYB161 was used as specific probes for in 583 situ hybridization. The antisense and sense probes were generated using T7 and SP6 RNA 584 polymerases, respectively. Digoxigenin RNA labeling kit (Roche) was used for probe labeling. 585 After pretreatment, the sections on slides were hybridized with the digoxigenin-labeled 586 Stem internodes of P. trichocarpa were cut into 2-mm fragments and fixed with FAA solution 610 (50% ethanol, 5% acetic acid, 3.7% formaldehyde) (v/v). The fixed stem segments were 611 transferred into a graded ethanol series (50%, 60%, 70%, 85%, and 100%) (v/v)   Fresh stem segments of 4-month-old P. trichocarpa plants were cut and immersed in 90% (v/v) 644 acetone at room temperature for 2 days. The stem segments were transferred into 100% acetone 645 at room temperature for 14 days, with fresh acetone replaced every 2 days. The acetone was then 646 discarded, and the stem segments were air-dried. The air-dried stem segments were used to 647 quantify the wood composition (acid-insoluble lignin, acid-soluble lignin, sugars) and lignin 648 composition (S-lignin, G-lignin, H-lignin) following established procedures (Abraham et al., 649 2013;Wang et al., 2018). 650

SDX Protoplasts Isolation and DNA Transfection 652
The P. trichocarpa SDX protoplast isolation and transfection were carried out as described (Lin 653 et al., 2013;2014) with minor modifications. The debarked stem segments of 6-month-old P. 654 trichocarpa were immersed in cell wall digestion enzyme solution in a 50-mL centrifuge tube for 655 3 h at room temperature. The digested debarked stem segments were transferred into the 30-mL 656 MMG solution in another 50-mL centrifuge tube. The protoplasts were released by gently 657 shaking the 50-mL centrifuge tube for 30 s. The protoplasts were filtered by the 75-μm nylon 658 membrane and centrifuged at 500 g for 3 min at room temperature. The pelleted protoplasts were 659 resuspended in the MMG solution and the cell density was adjusted to 5 × 10 5 cells/mL. For 660 gene expression analysis, 2 mL of protoplasts were used for transfection with 0.2 mL plasmid 661 DNA (2 mg/mL) and 2.2 mL of PEG solution. For ChIP assay, 20 mL protoplasts were used for 662 transfection with 2 mL plasmid DNA (2 mg/mL) and 22 mL of PEG solution. The transfection 663 detail procedure was described as our previous studies (Lin et al., 2013;2014). 664 665

Table 1. Wood composition of OE-PtrMYB161 transgenic and wild-type P. trichocarpa
Four-month-old plants were tested. Three biological replicates from independent pools of P. trichocarpa stems were carried out. Data are mean of three independent assays. Asterisks indicate significant differences between each line of the transgenics and wild-type plants (*P < 0.05, **P < 0.01, Student's t test). Units: g/100g of dry extractive-free wood. C:L=Carbohydrate to Lignin Ratio.

Table 2. Lignin composition of OE-PtrMYB161 transgenic and wild-type P. trichocarpa
Four-month-old plants were tested. Three biological replicates from independent pools of P. trichocarpa stems were carried out. Data are mean of three independent assays. Asterisks indicate significant differences between each line of the transgenics and wild-type plants (**P < 0.01, Student's t test). H: H-subunits; G: G-subunits; S: S-subunits; %: percentage weight of total lignin.