Dwarfism of high‐monolignol Arabidopsis plants is rescued by ectopic LACCASE overexpression

Abstract Lignin is a key secondary cell wall chemical constituent, and is both a barrier to biomass utilization and a potential source of bioproducts. The Arabidopsis transcription factors MYB58 and MYB63 have been shown to upregulate gene expression of the general phenylpropanoid and monolignol biosynthetic pathways. The overexpression of these genes also results in dwarfism. The vascular integrity, soluble phenolic profiles, cell wall lignin, and transcriptomes associated with these MYB‐overexpressing lines were characterized. Plants with high expression of MYB58 and MYB63 had increased ectopic lignin and the xylem vessels were regular and open, suggesting that the stunted growth is not associated with loss of vascular conductivity. MYB58 and MYB63 overexpression lines had characteristic soluble phenolic profiles with large amounts of monolignol glucosides and sinapoyl esters, but decreased flavonoids. Because loss of function lac4 lac17 mutants also accumulate monolignol glucosides, we hypothesized that LACCASE overexpression might decrease monolignol glucoside levels in the MYB‐overexpressing plant lines. When laccases related to lignification (LAC4 or LAC17) were co‐overexpressed with MYB63 or MYB58, the dwarf phenotype was rescued. Moreover, the overexpression of either LAC4 or LAC17 led to wild‐type monolignol glucoside levels, as well as wild‐type lignin levels in the rescued plants. Transcriptomes of the rescued double MYB63‐OX/LAC17‐OX overexpression lines showed elevated, but attenuated, expression of the MYB63 gene itself and the direct transcriptional targets of MYB63. Contrasting the dwarfism from overabundant monolignol production with dwarfism from lignin mutants provides insight into some of the proposed mechanisms of lignin modification‐induced dwarfism.


| INTRODUC TI ON
As society increasingly seeks to transition toward renewable resources, plant biomass is increasingly viewed as a potential source of commodity chemicals and novel bioproducts (Ralph, Lapierre, & Boerjan, 2019). Lignin is a key chemical component of plant biomass, conferring strength to cell walls, and resistance to degradation. For uses of plant biomass in forage and potential bioenergy applications, efforts have largely focused on reducing or altering lignin content, as it generally impedes accessibility to the carbohydrate cell wall fractions (DeMeester, de Vries, & Özparpucu, 2018). More recently, lignin is viewed positively as a valuable potential source of chemical feedstocks, and engineering lignin modification has been a major target (Mahon & Mansfield, 2019;Mottiar, Vanholme, Boerjan, Ralph, & Mansfield, 2016;Umezawa, 2018). In studies where lignin has been modified with the aim of biotechnological applications, and in mutant studies where monolignol biosynthesis has been perturbed, unexpected stunted growth phenotypes, termed ligninmodification-induced dwarfism (LMID), have been observed (reviewed by (Muro-Villanueva, Mao, & Chapple, 2019). Understanding why modifications in lignin content affect plant growth represents an important challenge.
Lignin is a polymer primarily made up of hydroxycinnamyl alcohols. Lignin also has a remarkable capacity to incorporate a wide range of other components including ferulic acid, tricin, and coniferyaldehyde (Mottiar et al., 2016). Hydroxycinnamyl alcohols are derived from the core phenylpropanoid pathway, commencing with the amino acid phenylalanine and resulting in the production of lignin monomers, or monolignols . p-coumaroyl-CoA represents a branch point between lignin biosynthesis and the biosynthesis of other phenolic compounds such as flavonoids (Dixon & Barros, 2019;Liu, Luo, & Zheng, 2018;Vanholme, De Meester, Ralph, & Boerjan, 2019). From p-coumaroyl-CoA, a series of subsequent hydroxylations, methylations, and reductions results in the formation of the three canonical monolignols (p-coumaryl-, coniferyl-, and sinapyl alcohols). The enzymes catalyzing this series of reactions have been biochemically characterized, and their functions inferred through mutant analysis in Arabidopsis thaliana (Arabidopsis; Van-Acker et al., 2013;Vanholme, Storme, & Vanholme, 2012) and gene knockdowns in Medicago sativa (alfalfa; Chen & Dixon, 2007) and poplar (Coleman, Park, Nair, Chapple, & Mansfield, 2008a. In addition to elucidating the biosynthetic pathways, the regulation of lignin production has also been studied in Arabidopsis. Both the general phenylpropanoid pathway and the monolignol-specific pathway are positively regulated by two MYB transcription factors, MYB58 and MYB63 (Zhou, Lee, Zhong, & Ye, 2009). These transcription factors are themselves controlled within the context of vessel and fiber development by upstream master transcription factors (Ohtani & Demura, 2019).
It remains uncertain exactly how monomers move from their site of synthesis to the wall where they are polymerized. Several possible mechanisms have been proposed, but there is no genetic evidence for transporter-mediated export of the major monolignols, coniferyl alcohol and sinapyl alcohol (Perkins, Smith, & Samuels, 2019). This may reflect a non-transporter-mediated mechanism, such as diffusion (Vermaas et al., 2019). Alternatively, active transport of monolignols by ABC transporters (Alejandro, Lee, & Tohge, 2012;Miao & Liu, 2010) or transport of monolignol glucosides have also been proposed (Tsuyama et al., 2013).
Once in the wall, the monomers are oxidized by secreted laccase and/or peroxidase to form monolignol radicals that ultimately polymerize into lignin. Simultaneous mutation of the two most highly expressed laccases in Arabidopsis inflorescence stems, LAC4 and LAC17, resulted in irregular xylem and reduced Klason lignin phenotypes (Berthet, Demont-Caulet, & Pollet, 2011). The mutation of a third laccase, lac11, in addition to lac4 and lac17 resulted in severe dwarfism and nearly absent lignin, suggesting these laccases play a major role in lignification (Zhao, Nakashima, & Chen, 2013). In the absence of polymerization machinery in the laccase triple mutant, excess monomers led to the formation of monolignol glucosides (Zhao et al., 2013), likely as a detoxification product (Väisänen et al., 2015).
These monolignol glucosides are thought to be sequestered in plant vacuoles. Both nonspecific monolignol export and activity of oxidative enzymes in the cell wall confer flexibility in a plants' ability to incorporate non-canonical monomers into lignin (Mottiar et al., 2016).
These factors suggest that manipulation of lignin content should be possible, but the deleterious effects on plant growth due to LMID remain an issue.
The mechanisms underlying LMID are currently the topic of intense interest (DeMeester et al., 2018;Panda, Li, Wager, Chen, & Li, 2020). Some mechanisms underlying LMID have been proposed, and then challenged. For example, HCT RNAi (HYDROXYCINNAMOYL

COENZYMEA:SHIKIMATE HYDROXYCINNAMOYL TRANSFERASE)
mutants (Hoffmann et al., 2005) were proposed to be dwarf due to flavonoid-induced inhibition of growth and auxin transport (Besseau et al., 2007). This model was not supported by the observation that loss of flavonoids in chalcone synthase mutants did not result in a dwarf phenotype (Li, Bonawitz, Weng, & Chapple, 2010).
In CINNAMOYL-COENZYME A REDUCTASE1 (ccr1-4) mutants, dwarf growth was reported to be due to ferulic acid accumulation creating disruption of the cell cycle during leaf development (Xue et al., 2015). This model is not supported by the observation that wild-type growth is possible in ccr1 mutants when a wild-type copy of the CCR1 gene is expressed exclusively in xylem cells (DeMeester et al., 2018). Other possible mechanisms leading to LMID include loss of vascular integrity, accumulation of pathway intermediates or derivatives, or triggering cell wall integrity sensing (Bonawitz & Chapple, 2013;Gallego-Giraldo, Liu, & Pose-Albacete, 2020;Muro-Villanueva et al., 2019). Forward genetic screens have identified that LMID requires subunits of the transcriptional coregulator Mediator (Bonawitz et al., 2014), as well as an importin-beta protein required to bring the MYB4 transcriptional repressor into the nucleus (Panda et al., 2020). These results highlight components of pathways that regulate the interaction between lignin defects and associated phenotypes, but leave large gaps in our understanding of the interaction between the pathways and the majority of the other factors involved.
Most studies examining LMID concern mutants with perturbations in the monolignol biosynthesis pathway, leading to phenotypes such as irregular xylem (DeMeester et al., 2018) or lower lignin levels ( Van-Acker et al., 2013). Paradoxically, overexpression of the regulatory MYB58 and MYB63 transcription factors that led to increased gene expression in the phenylpropanoid and monolignol biosynthetic pathways also led to impaired plant growth (Zhou et al., 2009). These plants with overabundant monolignol production and impaired growth provide opportunities to examine some of the proposed mechanisms of LMID. The first objective of this study was to examine the phenotypes of Pro35S::MYB58 (MYB58-OX) and Pro35S::MYB63 (MYB63-OX) plants that are relevant to LMID, such as vascular integrity, soluble phenolic profiles, and transcriptomes. One striking result was the high monolignol glucoside levels in these plants. As high monolignol glucosides are also found in loss of function lac4 lac11 lac17 triple mutants (Zhao et al., 2013), we hypothesized that additional copies of the lignin-related genes LAC4 or LAC17 might rescue the LMID in MYB58-OX or MYB63-OX lines. Co-expression of LAC4 or LAC17 was able to rescue the LMID growth phenotypes of MYB58-OX or MYB63-OX, in addition to reducing monolignol glucoside levels and transcriptome changes away from "stress-related" genes.
These data have interesting implications for understanding potential causes of LMID (Bonawitz & Chapple, 2013), arguing against dwarfism from the loss of a monolignol-derived growth-promoting molecules or the loss of lignin in vascular bundles, because

| Plant growth
Arabidopsis seeds were sown on ½ MS agar plates, were vernalized at 4°C for 2-3 days, and were transferred to soil after 7 days of growth together with WT seedlings. Growth conditions for all plants were set to 21°C, 16 hr light/8 hr dark, and 210 μmol m −2 s −1 light intensity.

| Microscopy
A Leica DMR epifluorescence microscope using 350/50 excitation and 455 nm longpass emission filter sets was used to document ultraviolet autofluorescence of Arabidopsis leaves and stems. A Perkin-Elmer UltraView VoX spinning disk confocal mounted on a Leica DMI6000 inverted microscope and a Hamamatsu 9100-02 CCD camera were used to image fluorescent proteins in living plant cells using the following excitation and emission filters: GFP (488 and 525), YFP (514 and 540), and RFP/m-Cherry (561 and 595). To study cell wall localization of LAC4, surface sterilized seeds were plated on GM media (MS media supplemented with 1% Sucrose and 1x Gamborg's Vitamin mix; Phytotechnology labs), vernalized at 4ºC for 2-3 days before being transferred to growth chambers.
Seeds were induced to germinate in 8 hr of light, then wrapped in foil and kept in darkness for 7 days. 7-day-old etiolated cotyledons were plasmolyzed in 0.4M D-mannitol (Sigma-Aldrich) for 1 hr.
Three seedlings were imaged as above for three independent lines of each construct for two independent experiments. Hand sections of Arabidopsis stem were Mäule stained (Mitra and Loqué, 2014).
To assess cell wall localisation of Lac4-mCherry, 7 day old etiolated hypocotyls were plasmolyzed in 0.4M mannitol for 1 hour. A line of prUBQ10-sec-mCherry (Chou et al 2018) was used as a control for cell wall localized signal.

| Molecular biology
Genomic sequences containing the MYB58 or MYB63 coding sequences were amplified using Phusion® High-Fidelity DNA Polymerase (New England Biolabs) using the primers listed in Table S4a. These PCR fragments were cloned using gateway cloning methodology (Invitrogen) using the pDONR221 vector as an entry clone and subsequently shuttled into the pK2GW7 binary vector (Karimi, Inzé, & Depicker, 2002).

| Real-time quantitative PCR
Gene-specific primers amplifying 190-240bp amplicons of were designed using primer 3 software (Koressaar & Remm, 2007) and are listed in Table S4b. The APT1 (AT1G27450) gene was used as a reference gene as described in Guénin et al. (2009). Real-time PCR was performed using iQ SYBR Green supermix (Biorad) and CFX connect real-time PCR detection system (Biorad). Efficiencies of PCR amplification and quantifications were performed according to the manufacturer's specification, as described in Schmittgen and Livak (2008). Total RNA was isolated from 5-week-old plants using TRIzol reagent (Invitrogen) and cDNA was synthesized using SuperScript III reverse transcriptase (Invitrogen) according to the manufacturer's instructions.

| Soluble phenolics extraction and analysis
Three-to four-week-old plants were harvested into liquid nitrogen and ground using a mortar and pestle. Fifty to 100 milligrams of ground tissue was combined with 1 ml of "methanol water" solution (49.5% methanol:1% acetic acid in water) and incubated 45°C for 4 hr to extract soluble phenolic compounds. The samples were centrifuged at 15 000 rpm for 15 min and the supernatant was transferred to glass vials. Phase partitioning with 1 ml ethyl ether was performed three times, with the upper layer transferred to new glass vials. The lower phase water-soluble phase was also transferred to separate glass vials. Samples were allowed to dry out overnight before resuspension in 50% methanol, sonication for 15-20 min and incubation at 35°C for 1 hr. Samples were filter-sterilized into HPLC vials and soluble phenolic compounds were separated by running approximately 10 μL of sample on an LC30 Chromatography Oven HPLC fitted with a Symmetry C14 column and PDA-100 Photodiode Array Detector (Dionex). The samples were examined at the wavelengths 280, 320, and 510 nm.
Methanol extracts were eluted from the column over a gradient from 95% A (100% water:0.1% trifluoroacetic acid (TFA)) to 45% B (75% acetonitrile:25% methanol: 0.1% TFA) over 50 min, followed by a 10 min wash with 75% B and re-acclimation of the column with 95% A for 10 min. The flow rate was 1 ml/minute; the column temperature was set to 40°C. Coniferin and syringin HPLC standards were prepared at a concentration of 0.05 mg/ml methanol and run on the HPLC using the conditions listed above.
To analyze the soluble phenolic phase by liquid chromatography/ mass spectroscopy, 10 μL of the same samples was run through an Agilent Zorbax Eclipse XDB C18 column (4.6 × 70 mm, particle size 1.8 μm) with a flow rate of 0.7 ml/min at 30°C. The samples were eluted with an increasing concentration of acetonitrile in 5% formic acid from 10% to 25% over 16 min, and from 25% to 100% over 9 min. The detection and analysis of metabolites was performed using a Bruker maXis Impact Ultra-High Resolution tandem TOF (UHR-Qq-TOF) mass spectrometer in positive electrospray ionization mode, temperature 220°C, drying gas flow rate 10 L/min, nebulizer pressure 4 bars, capillary voltage 3800 V, and using sodium formate as a calibrant.

| Structural chemistry analysis
Arabidopsis stems from 8-to 10-week-old plants were used to determine lignin and carbohydrate content following a modified Klason method (Cullis, Saddler, & Mansfield, 2004). Samples were ground in a Wiley mill to pass a 40 mesh screen, treated with acetone overnight using a Soxhlet, and then dried for 48 hr at 50°C.
Approximately 150 mg of dried extractive-free tissue was treated with 72% sulfuric acid for 2 hr, diluted to ~3% with 112 ml DI water and autoclaved at 121°C for 60 min. The mixture was filtered through a medium coarseness crucible and the retentate dried at 105°C. The acid-insoluble lignin was determined by weighing the retentate, while the acid-soluble lignin was measured from an aliquot of the filtrate using an UV spectrophotometer at 205 nm.
Carbohydrate contents were determined by HPLC analysis of the filtrate. Glucose, xylose, mannose, galactose, arabinose, and rhamnose were analyzed using a Dx-600 anion-exchange HPLC (Dionex) fitted with a CarboPac PA1 column (Dionex) at 1 ml/min and post column detection (100 mM NaOH min −1 ). Sugar concentrations were calculated from standard curves created from external standards.

| RNA extraction, RNAseq analysis, and defining MYB63 target genes
Total RNA was isolated from three replicated samples of sev- Surviving reads were aligned against the A. thaliana reference genome (TAIR10 genome release; Swarbreck, Wilks, & Lamesch, 2008) with bowtie2 v2.2.7 using the very sensitive-local parameter (Langmead & Salzberg, 2012). Count matrices were obtained with htseq-count v0.6.1 with default parameters (Anders, Pyl, & Huber, 2015). Differential analysis and normalization of count data were performed with DESeq2 (Love, Huber, & Anders, 2014). False discovery rate (FDR) < 0.05 and an absolute log2 fold change >1. Differentially expressed genes whose promoter region (1.5 kb upstream of transcription start site) contains MYB63 motif peaks are defined as high-confidence MYB63 target genes in this study. The raw sequence reads were deposited in NCBI Sequence Read Archive (http://www.ncbi.nlm.nih.gov/sra). lines produced inflorescence stems that were shorter than the WT ( Figure 1f). These data are consistent with the dramatic changes in growth and development, as well as ectopic deposition of lignin reported by Zhou et al. (2009), who discovered these monolignolassociated transcription factors.

| Overexpression of MYB58 and MYB63 produces dwarf plants with intact xylem
In studies examining dwarfism associated with downregulation of lignin, one of the proposed mechanisms leading to decreased growth was insufficient reinforcement of the xylem vessels (Bonawitz & Chapple, 2013;DeMeester et al., 2018). In this study,  (Bonawitz et al., 2014;Li et al., 2010). This suggests that the stunted growth of MYB58-OX and MYB63-OX is due to factors other than irregular xylem.

| Increased monolignol glucosides and sinapoyl esters, but decreased flavonoids, in MYB58-OX and MYB63-OX lines
The consequences of MYB58 or MYB63 overexpression on the soluble phenolic metabolites were not previously examined. Based on their ectopic lignification, we hypothesized that these lines also have elevated levels of monolignols, which could manifest in high monolignol glucoside levels (LeRoy, Huss, Creach, Hawkins, & Neutelings, 2016). The appearance of monolignol glucosides is proposed to be a homeostatic mechanism used by plant cells to balance monolignol metabolism (Lin et al., 2016), analogous to hexosylation of xenobiotics (Vanholme et al., 2012). HPLC-MS of extracted soluble phenolics from MYB63-OX leaves was used to examine the major

| Co-overexpression of lignin-related LACCASES in MYB58-OX and MYB63-OX lines makes soluble phenolic pools similar to wild type
Accumulations of monolignol glucosides, seen here in the MYB58-OX and MYB63-OX lines, were also observed in lac11 lac4 lac17 triple mutants (Zhao et al., 2013), which were severely impaired in lignification and dwarfed. We hypothesized that overexpression of lignin-related laccases such as LAC4 or LAC17 could lead to reduc- ( Figure S3). Clearly, the co-overexpression of either lignin-related LAC4 or LAC17 in the MYB58-OX and MYB63-OX lines had a strong impact on the soluble phenolics, shifting them toward a WT profile.

| Cell wall composition is restored by cooverexpression of LAC17 in MYB58-OX and MYB63-OX backgrounds
In addition to reverting the phenolic profiles toward WT levels, ectopic lignin (detected by Mäule staining) found in dwarf

| Co-overexpression of lignin-related LACCASES rescues growth in MYB58-OX and MYB63-OX lines
In addition to restoring soluble phenolic and cell wall composition toward the WT levels, growth was dramatically altered when LAC4 or LAC17  Other hypotheses related to LMID plants from previous studies were tested using the more severe dwarf growth phenotype of MYB63-OX lines. To test the potential contribution of flavonoids to the reduced growth of MYB63-OX lines, flavonoid biosynthetic mutant transparent testa4, tt4-2, deficient in chalcone synthase (Burbulis, Lacobucci, & Shirley, 1996) Conversely, loss of function of LAC4 and LAC17 did not modify the MYB63-OX phenotype. When the laccase double lac4 lac17 mutant (Berthet et al., 2011) was transformed with the 35S::MYB63 overexpression construct, the transformants also showed dwarf growth similar to MYB63-OX lines ( Figure S6c).  Table S2). Overall, the transcriptomic analysis highlights a significant upregulation of gene ontology terms relating to phenylpropanoid metabolic processes, including aromatic amino acid biosynthesis (p < 5E-12; Figure 10, Table S3). There were ab a also increases in expression of genes in GO categories "response to stress" and "response to abiotic stimulus" in the dwarf MYB63-OX line compared to the rescued double MYB63-OX/LAC17-OX lines ( Figure 10, Table S2). The ability of purified MYB63 to bind to sites in the Arabidopsis genome was previously mapped by O'Malley et al. (2016) using in vitro DNA affinity purification sequencing. There is strong overlap between that MYB63 target dataset and the genes modulated in the MYB63-OX line in this study (see starred genes in Figure 11; column A in Table S1). Transcriptomes of the rescued double MYB63-OX/LAC17-OX overexpression lines against WT showed that all of the direct target genes for MYB63 had significantly higher expression than WT (Figure 11, Table S3). This analysis confirmed MYB63 target genes continued to be upregulated, within the context of dramatic shifts in the transcriptome away from stress-responsive genes. The ability of additional laccase outside the cell membrane to trigger dramatic transcriptional shifts suggests monolignol oxidation outside the membrane changes the phenolic pools inside the cells.  Figure 4) but also the impaired plant growth was rescued (Figures 6   and 7). Using RNA-seq, we captured a more complete picture of the complex transcriptional changes elicited by MYB63 overexpression and the co-expression of MYB63 and LAC17. While there were many groups of differentially expressed genes, the stress-associated genes were strongly differentially expressed in dwarfed lines compared to rescued lines (Figure 10). The addition of LAC17 seems to broadly alleviate stress and dwarfism, and also alleviates the overaccumulation F I G U R E 4 Pro35S::MYB68 (MYB58-OX) and Pro35S::MYB63 (MYB63-OX) plants disrupted soluble phenolic phenotype reversed with LAC17 co-overexpression. Quantification of the soluble phenolics from Arabidopsis leaves of MYB58-OX and MYB63-OX overexpression lines, as well as MYB58-OX or MYB63-OX that cooverexpress Pro35S::LAC17 (LAC17-OX). (a) monolignol glucosides, coniferin and syringin; (b) hydroxycinnamate esters (HCE), sinapoyl glucose and sinapoyl malate; and (c) kaempferol glycosides. Bars indicate standard deviation. Three replicate experiments. n = 3-6 pooled samples of five leaves from five plants. Means with different letters represent statistically significant differences of total monolignol glucosides, HCEs, or flavonols (Tukey's pairwise comparison, p < .01). It has been suggested that one factor contributing to LIMD could be the disruption of the normal geometry and physiological functioning of xylem vessels (Bonawitz & Chapple, 2013;DeMeester et al., 2018). However, MYB58 and MYB63 overexpression lines do not have collapsed or irregularly shaped xylem vessels ( Figure 2).

| D ISCUSS I ON
As with some severe monolignol biosynthetic mutants (Panda et al., 2020), very young seedlings of MYB63 and MYB58 overexpressing plants are dwarfed even when grown inside highly moist sterile petri dishes where water should be readily available to all cells of the plant (Figure 1b), which also supports the conclusion that dwarfism in this case is due to other factors. In poplar, downregulation of C3'H resulted in irregular xylem, but measurement of physiological parameters such as xylem pressure potential and water-use efficiency was not consistent with a simple explanation of water stress (Coleman, Samuels, et al., 2008b). As the obvious explanation of water stress is not sufficient, we considered a constellation of other correlated factors among dwarfed and rescued lines.
In our dwarfed and rescued lines, plant size and monolignol glucosides are inversely proportional, with soluble phenolic analyses showing the highest coniferin and syringin in the most dwarf MYB63 overexpression lines. The more modestly dwarfed MYB58 overexpression lines also have modestly elevated coniferin and syringin levels ( Figure 4). In the double MYB and LAC17 overexpression lines, where growth was near WT (Figure 6), so too were the monolignol glucoside levels (Figure 4). Monolignol glucosides can accumulate as a method of detoxifying cytoplasmic accumulations of monolignols (Väisänen et al., 2015). When exogenous coniferyl alcohol was applied to BY-2 cell cultures, levels of coniferin and other related phenolics were elevated (Väisänen et al., 2015). Considering that, and the generally very low levels of free monolignols in cells ( stances, such as dehydro-diconiferyl glucosides (Binns, Chen, Wood, & Lynn, 1987;Lynn, Chen, Manning, & Wood, 1987), are playing a role in dwarfism in this case. Alternatively, it could be the case that the increase in glucosylation of monolignols limits the monolignols that could be oxidized and undergo combinatorial coupling to form dimers or oligolignols. Feeding experiments in Arabidopsis leaves have demonstrated the cytoplasmic oxidation and coupling of monolignols into oligomers in the cytoplasm (Dima et al., 2015).  Figure 11). We see a significant upregulation of many of the core components of the monolignol biosynthetic pathway, as well as a glucosyltransferase (UGT72E2) implicated in the formation of coniferin (Lanot et al., 2006). Although they did not accumulate monolignol glucosides, there was a similar upregulation of lines. It appears that the hydroxycinnamate esters, sinapoyl malate and sinapoyl glucose, move in concert with the monolignol glucosides. Interestingly, this was also observed in lac4 lac11 lac17 triple mutants, where accumulation of monolignol glucosides was associated with elevated hydroxycinnamoyl esters (Zhao et al., 2013).
These consistent patterns support the view that complex crosstalk exists in the phenylpropanoid and monolignol biosynthetic pathways as well as lignin polymerization mechanisms (Zhou et al., 2009).
F I G U R E 1 0 Gene Ontology (GO) enrichment analysis in the transcriptome of MYB63-OX compared to wild-type Arabidopsis rosette leaves or MYB63-OX/LAC17-OX double overexpression lines compared to wild type. Gene Ontology (GO) enrichment analysis, with GO SLIM terms having FDR < 0.05 considered significantly enriched. Dwarf MYB63-OX (left) have more up-and downregulated GO terms than rescued MYB63-OX/LAC17-OX (right) lines, details in Table S2. Response to stress indicated by a star. way. This is similar to the case of the c3'h mutants with decreased lignin levels, which could be rescued by the disruption of Mediator components MED5a and MED5b (Bonawitz et al., 2014). The quantity of lignin was restored to wild-type levels, even though the composition of the lignin remained altered, consisting almost exclusively of H subunits (Bonawitz et al., 2014). This demonstrated that the plant was chemically and physiologically capable of tolerating the C3'H mutation, but that the severe phenotype was due to other signals dependent on MED5a/b. It may be that, in the case of MYB58-OX and MYB63-OX, the cause and alleviation of severe dwarf phenotypes may be similarly dependent on intracellular or cell wall homeostatic mechanisms, and independent from the bulk lignification effects of laccase.
Our study of LMID in the context of MYB63, MYB58, and MYB/ LAC co-overexpressing lines has directly answered some questions about dwarfism related to mis-regulation of the lignin pathway, and raised the possibility of connections to other pathways and mechanisms worthy of further inquiry. We have ruled out the possibility that collapsed xylem is the cause of dwarfism in these lines. MYB/ LAC4 or LAC17 co-overexpression was able to normalize the hyperaccumulation of monolignol glucosides and restore growth, which is not consistent with dwarfism being due to loss of a monolignol-derived growth factor. Transcriptional analysis suggests that broad transcriptional changes are elicited by both the overexpression of MYB63 and the co-overexpression of MYB63 and LAC17, with the most striking difference between the two being stress-associated transcripts associated with dwarfism. We have not fully determined the mechanisms and signaling pathways responsible for the dramatic changes in phenotype that we observe. It is likely that other factors are involved, which may overlap with unknown factors implicated in other related studies of LMID-associated phenotypic rescues (Bonawitz et al., 2014;Gallego-Giraldo et al., 2020;Panda et al., 2020).

ACK N OWLED G M ENTS
Technical assistance from UBC Bioimaging Facility is gratefully acknowledged.