SEUSS integrates transcriptional and epigenetic control of root stem cell organizer specification

Abstract

Proper regulation of homeotic gene expression is critical for stem cell fate in both plants and animals. In Arabidopsis thaliana, the WUSCHEL (WUS)‐RELATED HOMEOBOX 5 (WOX5) gene is specifically expressed in a group of root stem cell organizer cells called the quiescent center (QC) and plays a central role in QC specification. Here, we report that the SEUSS (SEU) protein, homologous to the animal LIM‐domain binding (LDB) proteins, assembles a functional transcriptional complex that regulates WOX5 expression and QC specification. SEU is physically recruited to the WOX5 promoter by the master transcription factor SCARECROW. Subsequently, SEU physically recruits the SET domain methyltransferase SDG4 to the WOX5 promoter, thus activating WOX5 expression. Thus, analogous to its animal counterparts, SEU acts as a multi‐adaptor protein that integrates the actions of genetic and epigenetic regulators into a concerted transcriptional program to control root stem cell organizer specification.

The transcriptional adaptor protein SEUSS promotes expression of the quiescent center determinant WOX5 via recruitment of the histone methyltransferase SDG4 in Arabidopsis thaliana.

Introduction

Despite the huge evolutionary distance between plant and animal kingdoms, stem cell niches in both living forms contain organizer cells that maintain the adjacent stem cells (Dolan et al, 1993; Scheres, 2007; Dinneny & Benfey, 2008). In the model plant Arabidopsis thaliana, root stem cells are maintained by a small group of slowly dividing organizer cells called the quiescent center (QC) (Aichinger et al, 2012; Petricka et al, 2012). The QC generates signals that prevent differentiation of abutting stem cells, and it also acts as a reservoir to replace injured stem cells (van den Berg et al, 1995; Xu et al, 2006; Cruz‐Ramirez et al, 2013). In turn, the pluripotent stem cells undergo formative asymmetric division to generate specific tissue layers of the whole root system (Aichinger et al, 2012; Petricka et al, 2012). QC is first initiated in the embryo by asymmetric division of the hypophyseal cell during the early‐to‐mid‐globular embryo stage; the upper lens‐shaped daughter cell acquires QC identity, whereas the lower daughter cell becomes columella stem cells (CSCs) (Jürgens et al, 1994; Scheres & Benfey, 1999; Jürgens, 2001; Weigel & Jürgens, 2002). Remarkably, QC and the entire root stem cell niche can be readily re‐established in response to internal cues and external stresses during the lifelong post‐embryonic growth, which, as in the case of some trees, can extend beyond several thousand years (Chen et al, 2011; Marhava et al, 2019; Zhou et al, 2019).

Decades of molecular genetic studies have identified key transcription factors that regulate the acquisition of QC identity. Among these, the homeodomain transcription factor WUSCHEL (WUS)‐RELATED HOMEOBOX 5 (WOX5) is the best‐studied molecular marker of QC identity (Sarkar et al, 2007). WOX5 expression coincides with the embryonic formation of QC progenitors and persists specifically in the QC during post‐embryonic root growth (Haecker et al, 2004; Sarkar et al, 2007). WOX5 suppresses CYCLIN D activity to establish the quiescence of the QC and coordinates with several hormonal signals and transcriptional regulators to maintain the identity of CSCs (Stahl et al, 2009; Ding & Friml, 2010; Chen et al, 2011; Stahl et al, 2013; Forzani et al, 2014; Pi et al, 2015). Given that WOX5 is exclusively expressed in the QC, extensive research has been conducted to identify factors that confine WOX5 expression to such a narrow domain (Zhang et al, 2015; Long et al, 2017). However, the molecular mechanism underlying WOX5 expression activation remains largely unknown.

In contrast to WOX5, which is expressed exclusively in the QC, genes encoding the SCARECROW (SCR)/SHORT ROOT (SHR) transcription factors (belonging to the GRAS family) and PLETHORA (PLT) transcription factors [belonging to the APETALA2 (AP2) family] are expressed in larger domains including the QC (Di Laurenzio et al, 1996; Helariutta et al, 2000; Wysocka‐Diller et al, 2000; Sabatini et al, 2003; Aida et al, 2004; Heidstra et al, 2004; Galinha et al, 2007). The scr single mutant and plt1 plt2 double mutant displayed similar root stem cell defects (Sabatini et al, 2003; Aida et al, 2004), leading to the hypothesis that the SCR/SHR and PLT pathways converge to specify the QC identity as follows: The SCR/SHR pathway provides positional information along the radial axis, while the PLT pathway provides apical–basal information (Scheres, 2007; Dinneny & Benfey, 2008). Consistent with this hypothesis, a recent study proposed that PLT and SCR form a protein complex through their interaction with the teosinte‐branched cycloidea PCNA (TCP) transcription factor, and the PLT–TCP–SCR complex is essential for WOX5 expression and QC specification (Shimotohno et al, 2018). However, a previous study showed that the expression of WOX5 was reduced or undetectable in shr and scr mutants, but expanded to regions abutting the QC in the plt1 plt2 double mutant (Sarkar et al, 2007), suggesting that the SCR/SHR and PLT1/2 transcription factors might regulate WOX5 expression via distinct modes of action.

The glutamine (Q)‐rich SEUSS (SEU) protein contains a conserved domain, which shares high sequence similarity with the dimerization domain of the LIM‐domain‐binding (LDB) transcriptional co‐regulator proteins in animals (Franks et al, 2002). The animal LDB proteins, such as LDB1 in mouse and Chip in Drosophila, play fundamental roles in the transcriptional regulation of cell‐fate determination in versatile developmental processes (Agulnick et al, 1996; Morcillo et al, 1997; Matthews & Visvader, 2003; van Meyel et al, 2003; Bronstein et al, 2010; Bronstein & Segal, 2011; Love et al, 2014; Liu & Dean, 2019). Similar to its animal counterparts, SEU associates with cis‐regulatory elements through its interaction with specific transcription factors to regulate gene expression during multiple developmental processes (Franks et al, 2002; Pfluger & Zambryski, 2004; Sridhar et al, 2004, 2006; Grigorova et al, 2011; Gong et al, 2016; Huai et al, 2018).

Here, we report that SEU assembles a transcriptional complex to regulate root stem cell‐fate determination. SEU functions in the SCR signaling pathway to promote WOX5 expression for QC specification. SCR physically interacts with and recruits SEU to the WOX5 promoter. Then, SEU recruits the ASH1‐RELATED 3 (ASHR3) methyltransferase SET DOMAIN GROUP 4 (SDG4) (Cartagena et al, 2008; Kumpf et al, 2014) to the WOX5 promoter, which induces trimethylation of histone H3 lysine (K) 4 (H3K4me3), leading to WOX5 expression activation. Thus, SEU plays a fundamental role in the cell‐fate determination of root stem cell organizers by coordinating the formation of a functional SCR–SEU–SDG4 transcriptional complex.

Results

SEU positively regulates WOX5 expression and QC specification

To investigate the role of SEU in root stem cell determination, we generated transgenic plants expressing SEU fused to the green fluorescence protein (GFP) gene under the control of the SEU promoter (pSEU::SEU‐GFP). The SEU‐GFP fusion was localized to the nucleus and broadly expressed in the root meristem of post‐embryonic seedlings (Fig EV1A). During embryogenesis, the expression of SEU‐GFP was initiated early at the dermatogen stage and enriched broadly in different cells of the developing embryo (Fig 1A).

Figure EV1. SEU regulates QC specification.

• A
  Expression pattern of pSEU::SEU‐GFP in WT at 5 DAG.
• B–D
  Root apical meristem phenotypes of the indicated genotypes.
• E, F
  Expression pattern of pWOX5::WOX5‐GFP in the indicated genotypes at 5 DAG. Insets: GFP channel (Insets scale bars: 10 μm).
• G
  Quantification of the QC number in the indicated genotypes at 5 DAG. Data represent mean ± SD of three independent replicates. Different lowercase letters indicate significant differences by one‐way ANOVA followed by Tukey's multiple comparison test (P < 0.01).

Data information: In (A–F), scale bars: 20 μm.

Figure 1. Ablation of SEU reduces WOX5 expression and impairs quiescent center (QC) specification.

• A
  Expression patterns of pSEU::SEU‐GFP and pWOX5::GFP in the embryos of the indicated genotypes at dermatogen, early globular, and heart stages. White arrows indicate the QC precursor cell, and white dashed lines indicate embryos. Scale bars: 10 μm.
• B
  Quantification of pWOX5::GFP GFP fluorescence in wild‐type (WT) and seu‐3 mutant embryos. Fluorescence intensity at the early globular stage of WT embryos was set to 1.
• C, D
  Expression pattern of pWOX5::GFP in WT (C) and seu‐3 (D) embryos at 5 days after germination (DAG). Scale bars: 20 μm.
• E
  Quantification of GFP fluorescence in the QC of pWOX5::GFP transgenic plants, as shown in (C) and (D). Fluorescence intensity was normalized to the WT.
• F
  RT–qPCR analysis of the relative expression levels of WOX5 in WT and seu‐3 roots. Total RNA was extracted from 5 mm root tip sections of seedlings at 5 DAG.
• G–J
  Modified pseudo‐Schiff propidium iodide (mPS‐PI) staining of stem cell niche areas in the indicated genotypes at 5 DAG. Blue arrows indicate the QC, and red arrows indicate the columella stem cells (CSCs). The numbers denote total number of scored samples, with similar phenotypes showing in (G–J). Scale bars: 20 μm.
• K
  Quantification of the CSC layer in the indicated genotypes at 5 DAG.
• L, M
  Double staining of the QC184 β‐glucuronidase (GUS) marker (light blue) and starch granules (dark brown) in WT (L) and seu‐3 (M) seedlings at 5 DAG.
• N, O
  Expression pattern of J2341 in WT (N) and seu‐3 (O) seedlings at 5 DAG.
• P, Q
  Expression pattern of CS9227 in WT (P) and seu‐3 (Q) seedlings at 5 DAG.

Data Information: In (B), (E), (F), and (K), data represent mean ± SD of three independent replicates. n denotes the total number of scored samples. Individual values (black dots) are shown. **P < 0.01, *P < 0.05 (Student's t‐test). In (L–Q), white arrows indicate the CSCs. Scale bars: 20 μm.

We then investigated the effect of SEU on the expression of WOX5, which is specifically expressed in and required for QC specification and function (Sarkar et al, 2007), by expressing the pWOX5::GFP construct (Blilou et al, 2005) in wild‐type (WT) and seu‐3 mutant (Pfluger & Zambryski, 2004) plants. In WT embryos, pWOX5::GFP expression was initiated in the QC progenitors at the early globular stage (Fig 1A). However, in seu‐3 mutant embryos, pWOX5::GFP expression initiation was delayed to the heart stage, and the level of pWOX5::GFP expression was significantly reduced compared with the WT (Fig 1A and B).

At 5 days after germination (DAG), seu‐3 seedlings displayed markedly reduced pWOX5::GFP expression compared with the WT (Fig 1C–E). Consistently, reverse transcription‐quantitative PCR (RT–qPCR) assays showed that the WOX5 transcript levels were significantly reduced in seu‐3 seedlings compared with WT seedlings (Fig 1F). Together, these results indicate that SEU plays an essential role in promoting WOX5 expression during both embryogenesis and post‐embryonic development.

Next, we investigated whether the observed delay and reduction in WOX5 expression in seu‐3 were accompanied by defects in QC specification and function. Similar to the wox5‐1 mutant (Sarkar et al, 2007), seu‐3 seedlings showed supernumerary cells with nonstereotyped shapes in the QC position (Fig 1C, D and G–I). Consistently, expression of the QC‐specific marker QC184 was markedly reduced in seu‐3 seedlings compared with the WT (Fig 1L and M), suggesting a loss of QC identity. The CSCs adjacent to the QC in seu‐3 seedlings were also abnormal in shape and size and showed ectopic accumulation of starch granules, indicating that they had undergone differentiation (Fig 1L and M). Consistently, expression of the CSC‐specific marker J2341 was largely abolished (Fig 1N and O), while that of the columella marker CS9227 expanded to the CSCs (Fig 1P and Q). Complementation of the seu‐3 mutant by the introduction of the pSEU::SEU‐GFP construct confirmed that the seu‐3 mutation caused the observed phenotype (Fig EV1A–D). Together, these observations revealed that SEU positively regulates WOX5 expression and QC specification.

To determine the genetic relationship between SEU and WOX5, we generated a seu‐3 wox5‐1 double mutant line. The seu‐3 wox5‐1 double mutant exhibited similar QC defects and CSC differentiation phenotypes as the wox5‐1 mutant (Fig 1G–K). In addition, expression of pWOX5::WOX5‐GFP (Pi et al, 2015) in seu‐3 partially rescued the QC defects of the mutant (Fig EV1E–G). These results collectively support that SEU acts in the same pathway with WOX5 to regulate QC specification.

SEU functions in the SHR/SCR pathway to promote WOX5 expression and QC specification

Next, we asked whether and how SEU interacts with the master transcription factors SHR/SCR and PLTs, which converge in the root stem cell niche and play essential roles in QC specification (Sabatini et al, 2003; Aida et al, 2004; Scheres, 2007; Dinneny & Benfey, 2008). In the scr‐3 mutant embryo, the expression of the pWOX5::GFP construct did not initiate until the heart stage, and the level of pWOX5::GFP expression was significantly lower than that in the WT embryo (Fig EV2A and B). By contrast, the plt1‐4 plt2‐2 double mutant showed higher pWOX5::GFP expression compared with the WT (Fig EV2A and B). Similarly, at 5 DAG, pWOX5::GFP expression was strongly reduced in scr‐3 seedlings compared with the WT (Fig 2A, B and D) but expanded to the CSCs surrounding the QC in plt1‐4 plt2‐2 double mutant seedlings (Fig 2A, C and D). Consistently, RT–qPCR assays showed that the WOX5 transcript levels were significantly reduced in the scr‐3 mutant but increased in the plt1‐4 plt2‐2 double mutant compared with the WT (Fig 2E). These observations uncovered that, similar to SEU, SCR positively regulates WOX5 expression, whereas PLT1/2 negatively regulate WOX5 expression.

Figure EV2. SCR and PLT1/2 differentially regulate WOX5 expression during embryogenesis, and the SHR transcript levels were reduced in the seu‐3 mutant.

• A
  Expression pattern of pWOX5::GFP in WT, scr‐3, and plt1‐4 plt2‐2 embryos at the globular and heart stages. The white dashed lines indicate embryos. Scale bars: 10 μm.
• B
  Quantification of GFP fluorescence in pWOX5::GFP transgenic and WT embryos shown in (A). Fluorescence intensity in the WT early globular stage embryos was set to 1. Data represent mean ± standard deviation (SD) of three independent replicates.
• C
  RT–qPCR analysis of the relative expression levels of SHR in WT and seu‐3 roots. Total RNA was extracted from 5 mm root tip sections of seedlings at 5 DAG. Data represent mean ± SD of three independent replicates. **P < 0.01 (Student's t‐test).
• D–G
  Representative confocal images of the indicated genotypes at 5 DAG. Scale bars: 20 μm.
• H
  The yeast transformants were plated on SD/-2 and SD/-4 media to assess protein–protein interactions.

Figure 2. SEU acts in the SHR/SCR pathway to promote WOX5 expression and QC specification.

• A–C
  Expression pattern of pWOX5::GFP in WT (A), scr‐3 (B), and plt1‐4 plt2‐2 (C) at 5 DAG.
• D
  Quantification of GFP fluorescence in the QC of pWOX5::GFP transgenic seedlings, as shown in (A–C). GFP signal intensity in each genotype was normalized relative to that in the WT.
• E
  RT–qPCR analysis of the relative expression levels of WOX5 in WT, scr‐3, and plt1‐4 plt2‐2 roots. Total RNA was extracted from 5 mm root tip sections of seedlings at 5 DAG.
• F
  Primary root length of the indicated genotypes at 5 DAG.
• G–J
  Representative confocal images of the indicated genotypes at 5 DAG. White arrows indicate the QC.
• K
  Quantification of GFP fluorescence in the QC of pSCR::GFP and pSCR::GFP‐SCR transgenic seedlings, as shown in (G–J). GFP signal intensity of each genotype was normalized relative to that of the WT.
• L–Q
  Root stem cell niche phenotypes of the indicated genotypes at 5 DAG. White dashed lines indicate the QC region. The numbers denote total number of scored samples, with similar phenotypes showing in (L–Q).

Data information: In (D), (F), and (K), n denotes the total number of scored samples. Individual values (black dots) are shown. Data represent mean ± SD of three independent replicates. In (D) and (F), different lowercase letters indicate significant differences by one‐way ANOVA followed by Tukey's multiple comparison test (P < 0.01). In (E) and (K), **P < 0.01 (Student's t‐test). Scale bars: 20 μm.

The finding that both SEU and SCR promote WOX5 expression suggests that these proteins function in the same pathway. To test this possibility, we investigated the genetic interaction between SEU and SCR by comparing the root growth defects of the seu‐3 scr‐3 double mutant with its parental lines. While the seu‐3 scr‐3 double mutant displayed similar root growth defects as the scr‐3 single mutant, seu‐3 significantly enhanced the root growth defects of the plt1‐4 plt2‐2 (Fig 2F), indicating that SEU and SCR act genetically in the same pathway, which is independent of the PLT pathway. Consistent with these results, the expression of pSCR::GFP‐SCR and pSCR::GFP was dramatically reduced in the QC region of seu‐3 roots (Fig 2G–K), revealing that mutation of the SEU gene leads to a significant reduction in SCR expression in the QC, which affects QC identity. We also observed reduced expression of pSHR::GFP and pSHR::SHR‐GFP in seu‐3 roots compared with WT (Fig EV2D–G). Consistently, RT–qPCR assays showed that the SHR transcript levels were reduced in the seu‐3 mutant (Fig EV2C). These results are in line with recent observations that SEU acts as an upstream transcriptional regulator of SHR (Gong et al, 2016; Clark et al, 2020). The seu‐3 shr‐2 double mutant displayed comparable root growth defects as shr‐2 (Fig 2F). Additionally, the QC defects observed in the seu‐3 scr‐3 and the seu‐3 shr‐2 double mutants were similar to those observed in their parental lines, scr‐3 and shr‐2, respectively (Fig 2L–Q). Together, our results support that SEU acts in the SHR/SCR pathway to regulate QC specification.

SCR physically recruits SEU to promote WOX5 expression and QC specification

The finding that SEU and SHR/SCR functionally and genetically interact to regulate WOX5 expression and QC specification prompted us to test their possible physical interaction. Yeast two‐hybrid (Y2H) assays showed that SEU interacts with SCR (Fig 3A) but not SHR (Fig EV2H). To confirm this observation, we performed in vitro pull‐down experiments using purified maltose‐binding protein (MBP)‐tagged SEU (SEU‐MBP) and FLAG epitope‐tagged SCR (SCR‐FLAG) or SHR (SHR‐FLAG). SEU‐MBP pulled down SCR‐FLAG but not SHR‐FLAG (Fig 3B), indicating that SEU interacts specifically with SCR in vitro. Furthermore, in co‐immunoprecipitation (Co‐IP) experiments performed in Nicotiana benthamiana leaves, GFP‐tagged SCR (SCR‐GFP) was immunoprecipitated by SEU‐myc (Fig 3C). In Co‐IP assays performed in transgenic Arabidopsis plants expressing SCR‐GFP using anti‐SEU antibody, SCR‐GFP pulled down endogenous SEU (Fig 3D), confirming that SEU interacts with SCR in planta.

Figure 3. SCR recruits SEU to the WOX5 promoter to promote its expression.

• A
  Yeast two‐hybrid (Y2H) assays showing the interaction between SEU and SCR. The yeast transformants were plated on synthetic defined (SD) media lacking Leu and Trp (SD/-2) or lacking Ade, His, Leu, and Trp (SD/-4) to assess protein–protein interactions. AD, GAL4 activation domain; BD, GAL4 DNA‐binding domain.
• B
  In vitro pull‐down assays showing that SEU directly interacts with SCR but not with SHR. SCR‐FLAG was pulled down by SEU‐MBP immobilized on amylose resin. Protein bound to the amylose resin was eluted and analyzed by immunoblotting using anti‐FLAG antibody. The asterisk indicates the position of SEU‐MBP.
• C
  Verification of in vivo interactions between SCR and SEU in Nicotiana benthamiana leaves via Co‐IP assays. SCR‐GFP and SEU‐myc were transiently coexpressed in N. benthamiana leaves. Protein samples were immunoprecipitated using anti‐myc antibody.
• D
  Co‐IP assays of SEU with SCR in Arabidopsis. Protein extracts from WT and SCR‐GFP roots were isolated at 5 DAG and immunoprecipitated with anti‐GFP antibody.
• E
  Schematic diagram of the WOX5 and PCR amplicons (indicated as letters A, B, and I) used for ChIP‐qPCR. TSS, transcription start site.
• F, G
  SCR and SEU physically bind to the WOX5 promoter, as shown by ChIP‐qPCR analysis. Chromatin was isolated from 5 mm root tip sections of seedlings at 5 DAG, sonicated, and immunoprecipitated using anti‐GFP antibody. The precipitated DNA was used as a template for qPCR analysis. A, B, and I indicated the PCR amplicons as shown in (E). ACT7, control.
• H
  Mutation of the SCR gene impairs the recruitment of SEU to the WOX5 promoter, as shown by ChIP‐qPCR analysis. Chromatin was extracted from SEU‐GFP and SEU‐GFP scr‐3 seedlings at 5 DAG and precipitated with anti‐GFP antibody. ChIP signals were quantified by qPCR as a percentage of total input DNA.
• I
  SEU stimulated SCR‐mediated WOX5 promoter activation in transient expression assays in Nicotiana benthamiana leaves. The pWOX5::LUC reporter was cotransformed with the indicated effector constructs. The pWOX5::LUC activity was normalized relative to the internal control [LUC/Renilla luciferase (REN)]. The schematic diagram shows the construct used in the transient expression assays. Arrows indicate promoter regions, and boxes indicate coding sequences. Different lowercase letters indicate significant differences by one‐way ANOVA followed by Tukey's multiple comparison test (P < 0.01).
• J–M
  Representative confocal images of the indicated genotypes at 5 DAG. Scale bars: 20 μm. Insets show the QC region in which the solid white lines indicate the QC of WT, and the dashed white lines indicate the QC of seu‐3. Insets scale bars: 5 μm.
• N
  Quantification of GFP signal intensity in the QC of WT and seu‐3 seedlings expressing pWOX5::SEU‐GFP and pWOX5::SCR‐GFP. GFP signal intensity in seu‐3 seedlings was normalized relative to the WT. Individual values (black dots) are shown. n denotes the total number of scored samples.

Data information: In (F), (G), (H), (I), and (N), data represent mean ± SD of three independent replicates. **P < 0.01 (Student's t‐test).Source data are available online for this figure.

The above results suggest that SEU is physically recruited by SCR to the WOX5 promoter to promote its expression. Consistent with this presumption, chromatin immunoprecipitation (ChIP)‐qPCR assays using pSCR::GFP‐SCR transgenic Arabidopsis plants and anti‐GFP antibody showed the enrichment of WOX5 promoter at approximately −1,100 bp (Fig 3E, fragment B and 3F). Parallel ChIP‐qPCR assays using pSEU::SEU‐GFP transgenic Arabidopsis roots and anti‐GFP antibody revealed a similar enrichment pattern of the WOX5 promoter as that obtained using pSCR::GFP‐SCR plants (Fig 3E, fragment B and 3G). Together, these results indicate that SEU and SCR are recruited to the same region of the WOX5 promoter. However, in scr‐3 mutant plants expressing the pSEU::SEU‐GFP construct, the enrichment of the WOX5 promoter was markedly reduced (Fig 3E, fragment B and 3H), indicating that the recruitment of SEU on the WOX5 promoter is dependent on SCR.

Next, we used a dual‐luciferase (LUC) reporter system (Hellens et al, 2005) to examine the effect of SCR on WOX5 expression. To perform this experiment, a 3,100 bp fragment of the WOX5 promoter was cloned into the dual‐LUC reporter system to generate a pWOX5::LUC reporter construct (Fig 3I). Co‐expression of SCR with pWOX5::LUC in N. benthamiana leaves increased LUC activity, confirming that SCR positively regulates WOX5 expression (Fig 3I). When SEU and SCR were coexpressed with the pWOX5::LUC reporter, LUC activity was significantly enhanced further (Fig 3I), indicating that SEU acts as a co‐activator of SCR in regulating WOX5 expression.

We then asked whether QC‐enriched expression of SEU or SCR could rescue the WOX5 expression and QC defects of the seu mutant. For this purpose, we introduced the pWOX5::SEU‐GFP construct or the pWOX5::SCR‐GFP construct into the seu‐3 mutant. As expected, the pWOX5::SEU‐GFP construct restored WOX5 promoter activity and rescued the QC defects of the seu‐3 mutant (Fig 3J, K and N). By contrast, the pWOX5::SCR‐GFP construct failed to rescue the WOX5 expression and QC defects of the mutant (Fig 3L–N). Taken together, these results demonstrate that the transcriptional co‐regulator SEU is physically recruited by SCR to promote WOX5 expression and QC specification.

SEU recruits SDG4 to induce methylation of the WOX5 promoter

To understand how SEU co‐activates WOX5 expression along with SCR, we performed Y2H assays to identify SEU‐interacting proteins. In a Y2H screen, we found that the histone methyltransferases SDG4 and SDG25 interact with SEU (Figs 4A and EV3A). To determine whether SDG4 and SDG25 interact with SEU in planta, we conducted firefly LUC complementation imaging (LCI) assays (Chen et al, 2008) using N. benthamiana leaves. SDG4 was fused to the C‐terminal half of LUC (cLUC‐SDG4), while SEU was fused to the N‐terminal half of LUC (SEU‐nLUC). Co‐expression of cLUC‐SDG4 and SEU‐nLUC constructs in N. benthamiana cells resulted in a strong fluorescence signal, indicating that SDG4 interacts with SEU in plant cells (Fig EV3B). Parallel experiments indicated that SDG25 also interacts with SEU in plant cells (Fig EV3C).

Figure 4. SDG4 interacts with SEU and deposits H3K4me3 modification in the WOX5 promoter.

• A
  Y2H assays showing that SEU interacts with SDG4 and SDG25. The yeast transformants were plated on SD/‐2 and SD/‐4 media to assess protein–protein interactions.
• B
  Schematic diagram of the WOX5 and PCR amplicons (indicated as letters A–I) used for ChIP‐qPCR.
• C
  SDG4 physically bind to the WOX5 promoter, as shown by ChIP‐qPCR analysis. ACT7, control.
• D, E
  ChIP‐qPCR analysis indicating that mutations in SDG4 and SEU genes impair H3K4me3 deposition in the WOX5 promoter. ChIP signal was normalized to the WT. Individual values (black dots) are shown.
• F, G
  Expression pattern of pWOX5::GFP in the indicated genotypes at 5 DAG. Scale bars: 20 μm.
• H
  Quantification of GFP fluorescence in the indicated genotypes expressing pWOX5::GFP. GFP signal intensity in the WT was set to 1. n denotes the total number of scored samples.
• I
  RT–qPCR analysis of the relative expression levels of WOX5 in WT and ashr3‐1 roots. Total RNA was extracted from 5 mm root tip sections of seedlings at 5 DAG. Individual values (black dots) are shown.
• J–M
  mPS‐PI staining of stem cell niche in the indicated genotypes at 5 DAG. Blue arrows indicate the QC, and red arrows indicate the CSCs. The numbers denote total number of scored samples, with same phenotypes showing in (J–M). Scale bars: 10 μm.
• N
  Quantification of the CSC layer in the indicated genotypes at 5 DAG.

Data information: For (C–E), chromatin was isolated from 5 mm root tip sections of seedlings at 5 DAG, sonicated, and immunoprecipitated using anti‐GFP or anti‐H3K4me3 antibodies. The precipitated DNA was used as a template for qPCR analysis. PCR amplicons are shown in (B). In (C), (D), (E), (H), (I), and (N), data represent mean ± SD of three independent replicates. **P < 0.01, *P < 0.05 (Student's t‐test).

Figure EV3. SDG4 and SDG25 interact with SEU and regulate QC identity.

• A
  Yeast two‐hybrid (Y2H) assays showing the interaction of methyltransferases with SEU. The yeast transformants were plated on synthetic defined (SD) media lacking Leu and Trp (SD/‐2) or lacking Ade, His, Leu, and Trp (SD/‐4) to assess protein–protein interactions. SCR served as a positive control.
• B, C
  Luciferase (LUC) complementation assays showing that SEU interacts with SDG4 and SDG25. Nicotiana benthamiana leaves coinfiltrated with different combinations of constructs are shown. The pseudocolor bar shows the range of luminescence intensity.
• D, E
  mPS‐PI staining of stem cell niche in the indicated genotypes at 5 DAG. Scale bars: 10 μm.
• F
  Quantification of the CSC layer in the indicated genotypes at 5 DAG. Data represent mean ± SD of three independent replicates.
• G
  Expression of pSDG4::SDG4‐GFP in WT embryos at the dermatogen, globular, and heart stages. White arrows indicate the hypophyseal cell. Scale bars: 10 μm.
• H
  Expression pattern of pSDG4::SDG4‐GFP in WT seedlings at 5 DAG. Scale bars: 20 μm.

We then asked the functional relevance of SEU interaction with SDG4 and SDG25 in regulating QC specification. In line with previous observations (Kumpf et al, 2014), the ashr3‐1 mutant, which contains a T‐DNA insertion in the SDG4 gene, exhibited aberrant cellular organization in the QC and differentiated CSCs, similar to the seu‐3 mutant (Fig 4J and L). In comparison with ashr3‐1, the sdg25‐1 mutant (Berr et al, 2009; Tamada et al, 2009) displayed relatively weaker defects in the QC and CSC (Fig EV3D–F), we therefore focused on SDG4 to investigate its role in regulating WOX5 expression and QC specification by using transgenic plants expressing the pSDG4::SDG4‐GFP construct. During embryogenesis, the SDG4‐GFP fusion protein was expressed in the hypophyseal cell at the dermatogen stage (Fig EV3G), and during post‐embryonic root development, SDG4‐GFP was strongly expressed in the root meristem (Fig EV3H). ChIP‐qPCR experiment revealed that SDG4 indeed associated with the same region of the WOX5 promoter as SCR and SEU (Fig 4B, fragment B, C, compared to Fig 3E–G, fragment B; and Fig EV4A). Previous studies indicated that SDG4‐mediated histone H3 methylation is involved in diverse physiological processes including pollen tube growth and pathogen‐responsive gene expression (Cartagena et al, 2008; De‐La‐Pena et al, 2012). Our ChIP‐qPCR analyses revealed that the level of H3K4me3 modification in the WOX5 promoter was significantly decreased, whereas the level of H3K4me2 modification in the WOX5 promoter was slightly increased in the ashr3‐1 mutant as compared to the WT (Figs 4B and D, and EV4B). In parallel ChIP‐qPCR experiments, we found that the levels of H3K36me and the H3K36me3 in the WOX5 promoter of ashr3‐1 were comparable to those of WT (Fig EV4C and D). Consistently, the level of H3K4me3 modification in the WOX5 promoter, but not the H3K4me modification and the H3K4me2 modification in the WOX5 promoter, was significantly reduced in the seu‐3 mutant as compared to the WT (Figs 4E, and EV4E and F). Together, these results suggest that SEU physically recruits SDG4 to the WOX5 promoter, which deposits the H3K4me3 mark.

Figure EV4. ChIP‐qPCR analysis in the WOX5 promoter.

• A
  SDG4 physically bind to the WOX5 promoter, as shown by ChIP‐qPCR analysis.
• B–D
  ChIP‐qPCR analysis showing histone H3 methylation in the WOX5 promoter of WT and ashr3‐1. Chromatin of each sample was immunoprecipitated using anti‐H3K4me2, anti‐H3K36me, or anti‐H3K36me3 antibodies, respectively.
• E, F
  ChIP‐qPCR analysis showing histone H3 methylation in the WOX5 promoter of WT and seu‐3. Chromatin of each sample was immunoprecipitated using anti‐H3K4me or anti‐H3K4me2 antibodies, respectively.
• G
  ChIP‐qPCR analysis showing that the SDG4 mutation does not impair H3K4me3 levels of indicated genes.

Data information: The precipitated DNA was used as a template for qPCR analysis. PCR amplicons are shown in Fig 4B. Individual values (black dots) are shown. Data represent mean ± SD of three independent replicates. **P < 0.01, *P < 0.05 (Student's t‐test).

Considering that H3K4me3 is associated with gene activation (Cheng et al, 2020), we predicted that the reduced level of H3K4me3 modification in the WOX5 promoter in ashr3‐1 is correlated with decreased WOX5 expression and defective QC specification. Indeed, the expression levels of pWOX5::GFP in ashr3‐1 were decreased to levels comparable to those in seu‐3 (Fig 4F–H, compared with Fig 1C–E). Consistently, RT–qPCR results showed that the WOX5 transcript levels were significantly reduced in ashr3‐1 root tips compared with WT root tips (Fig 4I). Additionally, the seu‐3 ashr3‐1 double mutant exhibited aberrant cellular organization in the QC and differentiated CSCs, similar to the seu‐3 mutant (Fig 4J–N). Collectively, these biochemical and genetic data support our finding that SEU physically recruits SDG4 to the WOX5 promoter to regulate WOX5 expression and QC specification.

SEU coordinates the formation of the SCR‐SEU‐SDG4 transcriptional complex in planta

To determine the genetic relationship among SDG4, SEU, and SCR in regulating WOX5 expression and QC specification, we generated ashr3‐1 seu‐3 wox5‐1 and ashr3‐1 seu‐3 scr‐3 triple mutants (Fig 5A–G). The QC defects in ashr3‐1 seu‐3 wox5‐1 and ashr3‐1 seu‐3 scr‐3 triple mutants were similar to those observed in the wox5‐1 and scr‐3 single mutants, respectively (Fig 5D–G), corroborating that both SEU and SDG4 regulate WOX5 expression and QC specification through the SCR signaling pathway. Next, we transformed 35S::WOX5‐GFP into ashr3‐1 seu‐3 scr‐3 plants and observed that the 35S::WOX5‐GFP construct was able to complement the QC defects of the ashr3‐1 seu‐3 scr‐3 triple mutant in a dosage‐dependent manner (determined by 35S::WOX5‐GFP fluorescence signals) (Fig EV5), further confirming that SEU together with SCR and SDG4 function through WOX5 to control QC activity.

Figure 5. SEU cooperates with SCR and SDG4 to promote WOX5 expression and QC specification.

• A–G
  Representative confocal images of the indicated genotypes at 5 DAG. The white dashed lines indicate the QC region. Scale bars: 20 μm.
• H
  Domain mapping of SEU involved in interactions with SCR and SDG4. The yeast transformants were plated on SD/‐2 and SD/‐4 media to assess protein–protein interactions.
• I
  SEU associates with SCR and SDG4 in Nicotiana benthamiana leaves, as shown by Co‐IP assays. SDG4‐myc, SCR‐HA, and SEU‐GFP were coinfiltrated into Nicotiana benthamiana leaves. Protein samples were immunoprecipitated with anti‐GFP antibody and immunoblotted with anti‐myc and anti‐HA antibody. Nicotiana benthamiana leaves coinfiltrated with SDG4‐myc and SCR‐HA were used as a negative control.
• J
  Co‐IP assays of SCR with SEU and SDG4 in Arabidopsis. Proteins extracted from SDG4‐myc and SDG4‐myc SCR‐GFP plants were immunoprecipitated using anti‐GFP antibody and immunoblotted using anti‐myc and anti‐SEU antibodies.
• K
  Schematic representation of the role of SEU in SCR‐mediated activation of WOX5 expression. In WT roots, SEU is recruited to the promoter of WOX5 through the SEU–SCR interaction. SEU interacts with SDG4, which promotes H3K4me3 deposition in the WOX5 promoter. The SCR–SEU–SDG4 ternary complex plays an essential role in activating WOX5 expression and QC specification. Mutation of SEU impairs the formation of the SCR–SEU–SDG4 complex, which leads to dramatically reduced WOX5 expression and defective QC identity.

Source data are available online for this figure.

Figure EV5. QC defects in ashr3‐1 scr‐3 seu‐3 are rescued by WOX5 overexpression.

• A–C
  Representative confocal images of the indicated genotypes at 5 DAG. The numbers denote independent transgenic lines. The white dashed circles indicate the QC region. Insets: GFP channel (Insets scale bars: 20 μm). Scale bars: 20 μm.

Since SEU physically interacts with both SCR and SDG4, we mapped the protein domains of SEU involved in these interactions. Y2H assays revealed that the LDB domain of SEU interacts with SCR, whereas the C‐terminal Q‐rich domain of SEU interacts with SDG4 (Fig 5H), indicating that SEU interacts with SCR and SDG4 via distinct domains. To determine whether SEU coordinates the formation of the SCR–SEU–SDG4 transcriptional complex in planta, we performed Co‐IP assays using N. benthamiana leaves co‐expressing SEU‐GFP, SDG4‐myc, and SCR‐HA. Both SDG4‐myc and SCR‐HA could be pulled down by SEU‐GFP (Fig 5I). Furthermore, in Co‐IP experiments using transgenic plants expressing SDG4‐myc and SCR‐GFP, both SDG4‐myc and endogenous SEU could be pulled down by SCR‐GFP (Fig 5J), confirming the existence of the SCR–SEU–SDG4 complex in planta. In summary, our results suggest that SEU functions as a scaffold protein to orchestrate the formation of the SCR–SEU–SDG4 transcriptional complex, which regulates WOX5 expression during QC specification (Fig 5K).

Discussion

SEU acts as an integrative hub to mediate QC specification and root patterning

The transcriptional program that determines stem cell fate in plants, as in animals, is controlled by a small number of master transcription factors. Previous studies suggest that master transcription factor‐mediated protein–protein interaction networks play an important role in maintaining the QC‐specific expression of WOX5 (Long et al, 2017; Shimotohno et al, 2018). However, the exact mode of action of specific master transcription factors and the mechanism underlying the regulation of WOX5 expression by these transcription factors remain elusive. Here, we show that whereas SCR promotes WOX5 expression, PLT1/2 might negatively regulates WOX5 expression. We demonstrate a mechanism whereby SEU assembles a functional SCR–SEU–SDG4 transcription complex to activate WOX5 expression. At the WOX5 promoter, SCR physically recruits SEU, which then physically recruits the epigenetic co‐activator SDG4 to activate the expression of WOX5 (Fig 5K). Interestingly, a 35S::WOX5‐GFP construct was able to complement the QC defects of the ashr3‐1 seu‐3 scr‐3 triple mutant (Fig EV5), suggesting that SEU together with SCR and SDG4 function through WOX5 to control QC activity. Our model highlights the mechanistic function of SEU as an interface that physically and functionally integrates master transcription factors and epigenetic regulators into a functional complex, which accurately regulates the transcription of genes controlling plant development.

Deposition of appropriate epigenetic marks is important for transcriptional regulation. Previous studies report that SDG4/ASHR3 is required for QC quiescence maintenance (Kumpf et al, 2014). They also reveal that SDG4 associates with H3K36me (Kumpf et al, 2014). Interestingly, we found that the levels of H3K36me and the H3K36me3 modification in the WOX5 promoter of WT and the ashr3‐1 mutant were comparable (Fig EV4C and D), but the level of H3K4me3 modification in the WOX5 promoter was significantly reduced in the ashr3‐1 mutant compared with the WT (Fig 4B and D), suggesting that SDG4 associates to H3K4me3 in the WOX5 promoter. The reduction in H3K4me3 on the WOX5 promoter could be explained as an effect of reduced expression, or genome‐wide down‐regulation. To exclude these possibilities, we examined the effect of the ashr3‐1 mutation on the deposition of H3K4me3 in the promoter of a group of genes richly expressed in the root meristem including PLT1, PLT2, PIN1, PIN3, and PIN4. Results showed that the levels of H3K4me3 modification in the transcriptional start site of these genes were comparable between ashr3‐1 and WT (Fig EV4G), indicating that the reduction in H3K4me3 on the WOX5 promoter may not be due to genome‐wide down‐regulation. In addition, it was reported that SDG4 is a direct target of E2Fa/E2Fb transcription factors that control G1‐to‐S‐phase transition (Kumpf et al, 2014). As WOX5 binds to promoters of D‐type cyclins CYCD3;3 and CYCD1;1 and represses their expression in the QC (Forzani et al, 2014), it is of significance in future studies to explore the functional relevance of SDG4 and WOX5 in regulating QC activity with respect to cell‐cycle regulation.

Notably, we found that the mutation of the SEU gene reduces SCR expression in the QC (Fig 2G–K), indicating that the SEU‐dependent transcriptional complex, SCR–SEU–SCR, positively regulates the expression of SCR. This suggests that SEU operates a positive feedback loop to activate and maintain WOX5 expression. In addition, during embryogenesis, SCR, SEU, and SDG4 were expressed earlier than WOX5, suggesting that the SCR–SEU–SDG4 complex is important for the initiation of the QC progenitor hypophyseal cell.

Besides regulating QC specification, SCR, and its interacting partners, SHR and MED31, also regulate root ground tissue patterning (Di Laurenzio et al, 1996; Helariutta et al, 2000; Zhang et al, 2018). It was shown that SEU regulates the expression of SCR and SHR, thus playing an important role in the post‐embryonic formation of the middle cortex in the ground tissue (Gong et al, 2016). Together, these observations suggest that, in addition to regulating QC specification, the SCR–SEU interaction module likely operates in other SHR/SCR‐directed developmental programs. Indeed, it was proposed very recently that SEU is involved in a protein complex that acts as an upstream regulator of SHR and SCR, thereby differentially regulating the division timing of distinct cell types of the root stem cell niche (Clark et al, 2020).

Considering that SEU interacts with SCR and SDG4 through distinct protein domains, it is plausible that SEU provides a flexible interface to physically recruit and integrate versatile transcription factors and their co‐factors to form linage‐specific transcriptional programs that drive cell‐fate determination and differentiation. Future studies are needed to identify these SEU‐dependent transcriptional programs controlling root patterning.

SEU is an evolutionarily conserved transcriptional adaptor of cell‐fate specification

Although we showed that SEU positively regulates WOX5 expression, SEU was first identified for its repressive effect on the expression of the floral homeotic gene AGAMOUS (AG) (Franks et al, 2002). SEU represses AG expression by physically interacting with LUNIG (LUG), a transcriptional co‐repressor containing a LUFS (LUG/LUH, Flo8, and SSBP/SSDP) domain, which is highly conserved in the Groucho (Gro)/Tup1 family of co‐repressors, including the mammalian single‐stranded DNA‐binding proteins (SSBPs, also known as SSDP in Drosophila) (Franks et al, 2002; van Meyel et al, 2003; Sridhar et al, 2004; Liu & Karmarkar, 2008). Notably, the LUFS domain of LUG is necessary and sufficient for its interaction with SEU (Sridhar et al, 2004; Liu & Karmarkar, 2008). Although the entire SEU protein is required for SEU–LUG interaction, SEU does contain a LDB1/Chip conserved domain (LCCD), which is essential for the interaction of LDB proteins with the LUFS domain of SSBPs (van Meyel et al, 2003; Sridhar et al, 2004; Bao et al, 2010). Analogous to the SEU–LUG interaction in plants, the LCCD–LUFS‐mediated interactions between LDB proteins and SSBPs play critical roles in the transcriptional regulation of cell‐fate decisions in diverse developmental systems including cardiogenesis, neurogenesis, and hematopoiesis (Matthews & Visvader, 2003; van Meyel et al, 2003; Love et al, 2014; Liu & Dean, 2019). The intriguing conservation of the LCCD–LUFS‐mediated adaptor–co‐repressor interactions across plants and vertebrates suggests that a fundamentally important protein–protein interaction mechanism enables regulated control of gene transcription. It is likely that plant‐ or animal‐specific transcription regulators recruited this ancient transcription regulatory mechanism to mediate cell‐fate determination and organ patterning.

Materials and Methods

Plant material and growth conditions

Arabidopsis thaliana ecotypes Columbia (Col‐0), C24, and Landsberg erecta (Ler) were used as the WT. Among the plant materials used in this study, the following were described previously: QC25, QC46, QC184, and pSCR::GFP (Sabatini et al, 1999); pSCR::GFP‐SCR (Gallagher et al, 2004); p35S::SCR‐GFP (Cruz‐Ramirez et al, 2012); pWOX5::GFP (Blilou et al, 2005); pWOX5::WOX5‐GFP (Pi et al, 2015); pSHR::GFP and pSHR::SHR‐GFP (Nakajima et al, 2001); seu‐3 (Pfluger & Zambryski, 2004); scr‐3 (Fukaki et al, 1998); shr‐2 (Helariutta et al, 2000); plt1‐4 plt2‐2 (Aida et al, 2004); and wox5‐1 (Sarkar et al, 2007). Seeds of CS9227 and J2341 were obtained from the Haseloff enhancer trap GFP line collection (http://www.plantsci.cam.ac.uk/Haseloff). Seeds of the ashr3‐1 (SAIL_804_D06) and sdg25‐1 (SALK_149692) mutants were obtained from the Arabidopsis Biological Resource Center (ABRC), Ohio, USA (http://www.arabidopsis.org/abrc/).

Seeds were surface‐sterilized for 10 min in 10% commercial kitchen bleach, washed three times with sterile water, and plated on half‐strength Murashige and Skoog (1/2MS) medium (Murashige & Skoog, 1962) supplemented with 1% sucrose and 0.8% agar. Plants were stratified in vertically or horizontally oriented Petri dishes at 4°C for 2 days in the dark and then transferred to a phytotron set at 22°C, 16 h light/8 h dark photoperiod, and 120 μmol photons/m²/s light intensity. Roots of seedlings were examined at 3–5 DAG, depending on the experimental requirements.

Plasmid construction and plant transformation

To construct pSEU::SEU‐GFP and pSDG4::SDG4‐GFP plasmids, the GFP coding sequence (CDS) and nopaline synthase terminator (NOS‐T) sequence were amplified from the pGFP‐2 vector, and cloned in‐frame at the 3′ end of the promoter and CDSs of SEU and SDG4, respectively, in the pCAMBIA1300 binary vector using restriction endonucleases. To generate pWOX5::SEU‐GFP and pWOX5::SCR‐GFP constructs, the WOX5 promoter and SEU and SCR CDSs were amplified by PCR and cloned into pCAMBIA1300‐GFP (Zhang et al, 2018). To generate the p35S::SEU‐GFP construct, the SEU CDS was PCR amplified and cloned into the pENTR vector using a pENTR Directional TOPO Cloning Kit (Invitrogen), and then recombined with the PGWB5 binary vector containing the GFP CDS under the control of the cauliflower mosaic virus (CaMV) 35S promoter. Primers used for plasmid construction are listed in Appendix Table S1.

Constructs were transformed into Agrobacterium tumefaciens strain GV3101 to generate transgenic Arabidopsis plants using the floral dip transformation method. Transformants were selected based on their resistance to hygromycin. Homozygous T3 or T4 lines were used to perform various experiments. The p35S::SEU‐GFP construct was introduced into the scr‐3 mutant background by crossing, and homozygous plants were selected by genotyping.

Histology and microscopy

β‐Glucuronidase (GUS) staining was performed as described previously (Zhou et al, 2010). Whole seedlings were immersed in the GUS staining solution [1 mM X‐glucuronide in 100 mM sodium phosphate (pH 7.2), 0.5 mM ferricyanide, 0.5 mM ferrocyanide, and 0.1% Triton X‐100], briefly vacuum‐infiltrated, and incubated at 37°C in the dark for 1 h. Differential interference contrast (DIC) images were captured using the Leica DM5000B microscope. Images were processed using the Spot Flex software. Modified pseudo‐Schiff propidium iodide (mPS‐PI) staining was performed as described previously (Zhang et al, 2018). To perform confocal laser scanning microscopy, root tips of 3–5 DAG seedlings were stained with 10 μg/ml PI (Sigma P‐4170) for 5 min and observed under a Zeiss LSM 710 confocal microscope system. PI staining and GFP signals were visualized at wavelengths ranging from 600–640 and 500–540 nm, respectively. Images were taken with the ZEN 2012 software (Zeiss). GFP signal intensity was quantified as described previously (Zhang et al, 2018). The same offset and gain settings were used for WT, seu‐3, and ashr3‐1 seedlings, and the GFP signal was measured using the ZEN software on unmodified root images. At least 20 seedlings or embryos per genotype were observed.

Yeast two‐hybrid (Y2H) assays

Y2H assays were based on the MATCHMAKER GAL4 Two‐Hybrid System (Clontech). To verify the interaction between SCR and SEU, the CDS of SCR was fused to the GAL4 activation domain (AD) in pGADT7, and CDSs of SEU and its derivatives were fused with the GAL4 DNA‐binding domain (BD) in pGBKT7. To investigate the interactions of SEU with methyltransferases, CDSs of ATX1, SDG4, CaM KMT, CMT2, DDM1, ERF6, DRM2, MRG1, MRG2, NRP2, ORC1A, ORC1B, ROS3, SDG25, MET1, and REF6 (Baumbusch et al, 2001) were fused to the GAL4 AD in pGADT7. To verify the domains of SEU involved in interactions with SCR and SDG4, CDSs of SCR and SDG4 were fused to the GAL4 BD in pGBKT7, and those of SEU and its derivatives were fused to the GAL4 AD in pGADT7. Primers used for generating Y2H constructs are listed in Appendix Table S1. The resulting constructs were cotransformed into yeast (Saccharomyces cerevisiae) strain AH109. The presence of transgenes in yeast cells was confirmed by growth on plates containing solid synthetic defined (SD) media lacking leucine (Leu) and tryptophan (Trp) (SD/‐2). To assess protein–protein interactions, the transformed yeast cells were spread on plates containing SD media lacking adenine (Ade), histidine (His), Leu, and Trp (SD/‐4). Plates were incubated at 30°C, and protein–protein interactions were observed after 3 days.

Antibodies

The SEU CDS was amplified from the WT cDNA using gene‐specific primers (Appendix Table S1). The PCR product was cloned into the pMAL‐c2X vector to express the SEU‐MBP fusion in Escherichia coli BL21 (DE3) cells. The recombinant proteins were used to raise polyclonal antibodies in mice.

Antibody for ChIP: anti‐GFP (Abcam, ab290), anti‐H3K4me (Abcam, ab8895), anti‐H3K4me2 (Abcam, ab7766), anti‐H3K4me3 (Abcam, ab8580), anti‐H3K36me (Abcam, ab9048), and anti‐H3K36me3 (Abcam, ab9050).

Antibody for Western blot: Anti‐SCR (Santa Cruz, sc‐12643), Anti‐myc (Abmart, M20002L), Anti‐HA (Abmart, M20003L), Anti‐FLAG (Abmart, M20008L), Anti‐GFP (YTHX, ZA009), Anti‐MBP (BioLabs, E8032L).

RNA extraction, reverse transcription (RT), and RT‐quantitative PCR (RT–qPCR) assays

To perform RT–qPCR analysis of WOX5, total RNA was extracted from approximately 5 mm long root tip sections of WT and mutant seedlings harvested at 5 DAG. Subsequently, cDNA was prepared from 2 μg total RNA using Superscript III reverse transcriptase (Invitrogen) and quantified on a Roche 480 cycler using the SYBR Green Kit (Takara). Expression levels of WOX5 were normalized relative to ACT7 expression. Statistical significance was evaluated with Student's t‐test. Primers used for RT–qPCR analysis are listed in Appendix Table S1.

Western blot analysis

Seedlings were ground to a fine powder in liquid nitrogen and transferred to the extraction buffer [50 mM Tris–HCl (pH 7.5), 150 mM NaCl, 10 mM EDTA, 50 mM DTT, 2% (v/v) Nonidet P‐40, and protease inhibitor cocktail (Roche)]. To perform western blot analysis, protein samples were boiled in sodium dodecyl sulfate (SDS) loading buffer for 5 min, separated by SDS–polyacrylamide gel electrophoresis (SDS–PAGE), and transferred to polyvinylidene fluoride (PVDF) membranes. Proteins of interest were detected using specific antibodies.

Co‐immunoprecipitation (Co‐IP) assays

To perform Co‐IP assays, p35S::SCR‐GFP and WT Col‐0 seedlings were harvested at 5 DAG and homogenized in protein lysis buffer [50 mM Tris–HCl (pH 7.5), 150 mM NaCl, 5 mM EDTA, 0.1% Triton X‐100, 0.2% Nonidet P‐40, 0.6 mM phenylmethylsulfonyl fluoride (PMSF), 20 μM MG132, and protease inhibitor cocktail (Roche)]. After protein extraction, 20 μl protein A/G plus agarose beads (Santa Cruz Biotechnology) was added to 2 mg protein extracts to reduce nonspecific immunoglobulin binding. After 1 h of incubation, the supernatant was transferred to a new tube. Then, anti‐GFP antibody‐bound agarose beads (Chromtek) were added to each reaction and incubated at 4°C for 4 h. Col‐0 seedlings were used as a negative control. The precipitated samples were washed at least four times with the lysis buffer and then eluted by boiling the beads in SDS protein loading buffer for 5 min. SEU and GFP proteins were detected with anti‐SEU antibody (1:2,000) and anti‐GFP antibody (1:2,000), respectively.

Co‐IP assays using N. benthamiana leaves were performed as described previously (Liu et al, 2010). Agrobacterium tumefaciens strain GV3101, carrying p35S::SCR‐GFP and p35S::SDG4‐myc constructs, or p35S::SEU‐GFP, p35S::SDG4‐myc and p35S::SCR‐HA constructs, was infiltrated into tobacco leaves. The transformed N. benthamiana leaves were ground into a fine powder and transferred to the lysis buffer [50 mM Tris‐MES (pH 8.0), 0.5 M sucrose, 1 mM MgCl₂, 10 mM EDTA, 5 mM DTT, 0.5 mM PMSF, 50 μM MG132, and protease inhibitor cocktail (Roche)]. The Co‐IP procedure was the same as that described for Arabidopsis.

In vitro pull‐down assays

To detect the SEU–SCR interaction using pull‐down assays, SHR‐FLAG and SCR‐FLAG fusion proteins were synthesized by in vitro transcription/translation reactions (Promega). The SEU‐MBP protein was affinity purified. Per reaction, 15 μl agarose beads bound by 1 μg SEU‐MBP was incubated with 10 μl SHR‐FLAG or SCR‐FLAG protein in 1 ml reaction buffer [25 mM Tris–HCl (pH 7.5), 100 mM NaCl, 1 mM DTT, and protease inhibitor cocktail (Roche)] at 4°C for 1 h. The beads were then collected and washed three times with washing buffer [25 mM Tris–HCl (pH 7.5), 150 mM NaCl, and 1 mM DTT]. The bound proteins were eluted off the agarose beads using the elution buffer [25 mM Tris–HCl (pH 7.5), 100 mM NaCl, 1 mM DTT, and 10 mM maltose]. The SHR‐FLAG and SCR‐FLAG proteins were detected by western blotting using anti‐FLAG antibody (1:2,000). Purified MBP was used as a negative control.

Transient expression assays in Nicotiana benthamiana leaves

The promoter of WOX5 was amplified from the genomic DNA of Col‐0 and cloned into the pGreenII 0800‐LUC (Hellens et al, 2005) vector to generate a reporter construct. The Renilla luciferase (REN) gene under the control of the CaMV 35S promoter was used as an internal control. To generate the effector constructs, CDSs of SEU and SCR were cloned into the pUC19‐35S-HA‐RBS vector (Li et al, 2005) under the control of the 35S promoter. Primers used for vector construction are listed in Appendix Table S1. Firefly LUC and REN activities were measured using the Dual‐LUC Reporter Assay System (Promega), according to the manufacturer's instructions.

Chromatin immunoprecipitation (ChIP)‐qPCR assays

Root tips were harvested from seedlings at 5 DAG and crosslinked using 1% formaldehyde at room temperature for 10 min. The crosslinking reaction was stopped by the addition of 0.125 M glycine. The chromatin complex was isolated and resuspended in lysis buffer [50 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1% SDS, 1% Triton X‐100, 0.1% sodium deoxycholate, 1 mM PMSF, and 1 × protease inhibitor cocktail (Roche)]. Then, the chromatin was sheared by sonication to an average size of approximately 200 bp. The sheared chromatin was pre‐cleared with Protein A salmon sperm‐coupled agarose (Millipore), and 10 μl of the pre‐cleared chromatin was set aside for use as an input control. The remaining chromatin complex was immunoprecipitated overnight at 4°C with the following antibodies: anti‐GFP, anti‐H3K4me, anti‐H3K4me2, anti‐H3K4me3, anti‐H3K36me, and anti‐H3K36me3. The immunoprecipitated chromatin complex was washed once with each of the following buffers in this order: low‐salt buffer [20 mM Tris–HCl (pH 8.0), 2 mM EDTA, 150 mM NaCl, 0.5% Triton X‐100, and 0.2% SDS]; high‐salt buffer [20 mM Tris–HCl (pH 8.0), 2 mM EDTA, 500 mM NaCl, 0.5% Triton X‐100, and 0.2% SDS]; LiCl buffer [10 mM Tris–HCl (pH 8.0), 1 mM EDTA, 0.25 M LiCl, 0.5% NP‐40, and 0.5% sodium deoxycholate]; and TE buffer [10 mM Tris–HCl (pH 8.0) and 1 mM EDTA]. After washing, the immunoprecipitated chromatin was eluted with elution buffer (1% SDS and 0.1 M NaHCO₃). The protein–DNA crosslinks were reversed by incubating the immunoprecipitated complexes at 65°C overnight. DNA was recovered using the QIAquick PCR Purification Kit (Qiagen). ChIP signals were quantified as the percentage of total input DNA and normalized relative to the control (ACT7). Primers used for qPCR are listed in Appendix Table S1.

Author contributions

XZ, HZ, and CL designed the research; HZ, XZ, YY, and LL performed the research; XZ, HZ, WZ, and CL analyzed the data; XZ, WZ, and CL wrote the paper; and CL conceived and supervised the study.

Conflict of interest

The authors declare that they have no conflict of interest.

Supporting information

Appendix

Expanded View Figures PDF

Review Process File

Source Data for Figure 3

Source Data for Figure 5

The EMBO Journal (2020) 39: e105047

Data availability

This study includes no data deposited in external repositories.