Volume 229, Issue 3 p. 1582-1597
Full paper
Free Access

WOX9 functions antagonistic to STF and LAM1 to regulate leaf blade expansion in Medicago truncatula and Nicotiana sylvestris

Tezera W. Wolabu,

Tezera W. Wolabu

Institute for Agricultural Biosciences, Oklahoma State University, Ardmore, OK, 73401 USA

Noble Research Institute, LLC, Ardmore, OK, 73401 USA

Search for more papers by this author
Hui Wang,

Hui Wang

Institute for Agricultural Biosciences, Oklahoma State University, Ardmore, OK, 73401 USA

College of Grassland Science and Technology, China Agricultural University, Beijing, 100193 China

Search for more papers by this author
Dimiru Tadesse,

Dimiru Tadesse

Institute for Agricultural Biosciences, Oklahoma State University, Ardmore, OK, 73401 USA

Search for more papers by this author
Fei Zhang,

Fei Zhang

Institute for Agricultural Biosciences, Oklahoma State University, Ardmore, OK, 73401 USA

Department of Molecular, Cellular and Developmental Biology, Yale University, New Haven, CT, 06520-8104 USA

Search for more papers by this author
Marjan Behzadirad,

Marjan Behzadirad

Institute for Agricultural Biosciences, Oklahoma State University, Ardmore, OK, 73401 USA

Search for more papers by this author
Varvara E. Tvorogova,

Varvara E. Tvorogova

Institute for Agricultural Biosciences, Oklahoma State University, Ardmore, OK, 73401 USA

Department of Genetics and Biotechnology, St Petersburg State University, St Petersburg, 199034 Russia

Search for more papers by this author
Haggag Abdelmageed,

Haggag Abdelmageed

Institute for Agricultural Biosciences, Oklahoma State University, Ardmore, OK, 73401 USA

Department of Agricultural Botany, Faculty of Agriculture, Cairo University, Giza,, 12613 Egypt

Search for more papers by this author
Ye Liu,

Ye Liu

School of Life Science, University of Science and Technology of China, Hefei, Anhui, 230027 China

Search for more papers by this author
Naichong Chen,

Naichong Chen

Institute for Agricultural Biosciences, Oklahoma State University, Ardmore, OK, 73401 USA

Department of Biochemistry and Molecular Biology, Oklahoma State University, Stillwater, OK, 74078 USA

Search for more papers by this author
Jianghua Chen,

Jianghua Chen

CAS Key Laboratory of Topical Plant Resources and Sustainable Use, CAS Center for Excellence in Molecular Plant Sciences, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Kunming, Yunnan, 650223 China

Search for more papers by this author
Randy D. Allen,

Randy D. Allen

Institute for Agricultural Biosciences, Oklahoma State University, Ardmore, OK, 73401 USA

Department of Biochemistry and Molecular Biology, Oklahoma State University, Stillwater, OK, 74078 USA

Search for more papers by this author
Million Tadege,

Corresponding Author

Million Tadege

Institute for Agricultural Biosciences, Oklahoma State University, Ardmore, OK, 73401 USA

Department of Plant and Soil Sciences, Oklahoma State University, Stillwater, OK, 74078 USA

Author for correspondence:

Million Tadege


Search for more papers by this author
First published: 22 September 2020
Citations: 7


  • WOX family transcription factors regulate multiple developmental programs. The intermediate clade transcriptional activator WOX9 functions together with the modern clade transcriptional repressor WOX genes in embryogenesis and meristems maintenance, but the mechanism of this interaction is unclear.
  • STF and LAM1 are WOX1 orthologs required for leaf blade outgrowth in Medicago truncatula and Nicotiana sylvestris, respectively. Using biochemical methods and genome editing technology, here we show that WOX9 is an abaxial factor and functions antagonistically to STF and LAM1 to regulate leaf blade development.
  • While NsWOX9 ectopic expression enhances the lam1 mutant phenotype, and antisense expression partially rescues the lam1 mutant, both overexpression and knockout of NsWOX9 in N. sylvestris resulted in a range of severe leaf blade distortions, indicating important role in blade development. Our results indicate that direct repression of WOX9 by WUS clade repressor STF/LAM1 is required for correct blade architecture and patterning in M. truncatula and N. sylvestris.
  • These findings suggest that controlling transcriptional activation and repression mechanisms by direct interaction of activator and repressor WOX genes may be required for cell proliferation and differentiation homeostasis, and could be an evolutionarily conserved mechanism for the development of complex and diverse morphology in flowering plants.


WUSCHEL-related homeobox (WOX) factors are plant-specific transcriptional regulator proteins that contain a DNA binding homeodomain similar to WUSCHEL (WUS), the founding member of the family from Arabidopsis. Several elegant studies demonstrated that the WOX family is involved in the regulation of a wide range of key developmental programs. While WOX2, WOX8, and WOX9 function in zygotic development and embryogenesis (Haecker et al., 2004; Breuninger et al., 2008; Ueda et al., 2011), WUS and WOX5 modulate stem cell maintenance in shoot and root apical meristems, respectively (Mayer et al., 1998; Sarkar et al., 2007). WOX1 and/or WOX3/PRS, however, orchestrate leaf blade and floral organ development as demonstrated in several species (Matsumoto & Okada, 2001; Nardmann et al., 2004; Park et al., 2005; Shimizu et al., 2009; Vandenbussche et al., 2009; Tadege et al., 2011a; Nakata et al., 2012; Zhuang et al., 2012; Cho et al., 2013; Ishiwata et al., 2013; Niu, 2018).

WUS and its homologs in other species, including TERMINATOR (TER) in petunia, ROSULATA (ROA) in Antirrhinum, and HEADLESS (HDL) in Medicago, are required for shoot apical meristem (SAM) maintenance (Laux et al., 1996; Mayer et al., 1998; Stuurman et al., 2002; Kieffer et al., 2006; Meng et al., 2019; Wang et al., 2019). Loss of WUS function in the wus-1 Arabidopsis mutant results in premature termination and arrest of the SAM and floral meristem (FM), but the SAM re-establishes itself to resume growth while the process repeats itself, leading to altered plant morphology (Laux et al., 1996; Mayer et al., 1998). In the hdl mutant of Medicago truncatula, the terminated SAM is never fully re-established, and the few meristematic cells that try to re-establish are consumed in making leaves. As a result, hdl plants make only leaves throughout development (Tadege et al., 2015; Meng et al., 2019; Wang et al., 2019). The hdl mutant also shows altered leaf shape (Meng et al., 2019; Wang et al., 2019), which was not detected in the wus mutant. However, the wus wox1 prs triple mutant showed a stronger leaf phenotype than the wox1 prs double mutant (Zhang & Tadege, 2015), suggesting that WUS may also have a redundant function in leaf development. Although WUS transcript is specifically expressed in the organizing center (OC) of the SAM (Mayer et al., 1998), it is likely that a noncell autonomous signal from WUS may contribute to blade outgrowth, since the WUS protein itself is shown to move from the OC to the stem cell region (Yadav et al., 2011; Daum et al., 2014).

An intimate connection exists between WUS and the phytohormone cytokinin, and this appears to be true for STF/WOX1 as well (Tadege & Mysore, 2011; Tadege, 2016; H. Wang et al., 2017). While WUS activity is modulated by cytokinin in the SAM and axillary meristem (J. Wang et al., 2017; Snipes et al., 2018), WUS promotes cytokinin activity in the shoot stem cell niche by repressing type-A ARABIDOPSIS RESPONSE REGULATOR (ARR) genes ARR5, ARR6, ARR7, and ARR15 (Leibfried et al., 2005) to activate cell proliferation. The WUS SAM maintenance activity and interaction with cytokinin were shown to be linked to its WUS box (Ikeda et al., 2009; Dolzblasz et al., 2016; Snipes et al., 2018), suggesting that WUS primarily functions as a transcriptional repressor. Interestingly, WUS is reported to be activated by WOX9/STIP, which is also required for shoot meristem maintenance (Wu et al., 2005) and embryo development (Wu et al., 2007; Ueda et al., 2011). However, WOX9 is reported to be a strong transcriptional activator (Lin et al., 2013a), and it is unclear whether WUS and WOX9 employ the same mechanism in shoot meristem maintenance.

The WOX9 gain-of-function stip-D mutant displays wavy leaf margins (Wu et al., 2005), and the stip loss-of-function mutant is arrested at the seedling stage (Wu et al., 2005), but it is unclear if WOX9/STIP plays a significant role in leaf blade development. The WOX genes that function in leaf blade development belong to the WUS clade transcriptional repressors as exemplified by WOX1 and PRS/WOX3 in Arabidopsis (Vandenbussche et al., 2009; Nakata et al., 2012), STF in M. truncatula, and LAM1 in Nicotiana sylvestris (Tadege et al., 2011a). Unlike leaf polarity factors that are either adaxial or abaxial-specific (Waites et al., 1998; Sawa et al., 1999; Siegfried et al., 1999; Kerstetter et al., 2001; McConnell et al., 2001; Iwakawa et al., 2002), the WOX genes STF, WOX1, and PRS are expressed in the middle at the adaxial–abaxial juxtaposition to control mediolateral outgrowth of the leaf blade (Tadege et al., 2011a; Tadege et al., 2011b; Nakata et al., 2012; Nakata & Okada, 2012), suggesting a novel mechanism for blade expansion.

STF/LAM1 is a transcriptional repressor (Lin et al., 2013a,b) and its repression activity is conferred by its WUS box and STF box motifs (Zhang et al., 2014; Zhang & Tadege, 2015). The DNA binding STF homeodomain (HD) and the repression motifs (WUS box and STF box) are critically required for blade outgrowth function (Lin et al., 2013a; Zhang et al., 2014, 2019). Interestingly, all of the WUS clade Arabidopsis WOX transcription factors (WUS and WOX1–WOX7), which have transcriptional repression activity, can substitute for LAM1 function (Lin et al., 2013a), suggesting that modern/WUS clade WOX members have a conserved transcriptional repression mechanism in meristem maintenance and lateral organ development, with specificity conferred by cis elements that drive specific expression patterns.

Intermediate clade WOX members have intact HD but lack repression domains. Here we show that homologs of the intermediate clade transcriptional activator WOX9, namely, MtWOX9-1, MtWOX9-2 and NsWOX9, function antagonistically to STF or LAM1, and exacerbate the lam1 phenotype when overexpressed. Suppression of NsWOX9 transcript levels by antisense technology partially rescues the lam1 mutant leaf blade phenotype while the introduction of knockout mutations in the native NsWOX9 gene by multiplex gRNA genome editing severely affected leaf blade symmetry and expansion. Our results suggest that direct and antagonistic interactions between transcriptional repressor and activator WOX genes may be important to maintain homeostasis in cell proliferation and differentiation in acquiring complex morphology in flowering plants.

Materials and Methods

Plant materials and growth conditions

Plant materials used for this study including M. truncatula R108, stf mutant (stf-2), and N. sylvestris wild type (WT) and lam1 mutant were grown in one gallon pots in a glasshouse under long-day (LD) conditions with 16 h : 8 h, light : dark cycle at 23–27°C, and in growth room under LD conditions of 16 h : 8 h, light : dark cycle at 23–25°C, 70–80% relative humidity, and a light intensity of 150 µmol m2.

Samples collection, RNA extraction

Medicago truncatula tissue samples were collected from 4-wk-old WT and stf mutant plants. For GR induction, 4-wk-old plants of p35S::YFP-GR-STF, p35S::YFP-GR-LAM1 and p35S::YFP-GR-NsWOX9 transgenic lines were treated with mock, DEX (10 mM), and CHX (10 mM) for 3 h in M. truncatula and 6 h in N. sylvestris, before leaf samples were collected. Collected samples were immediately snap-frozen in liquid nitrogen and stored at −80°C until processing. Total RNA from shoot apex and young leaf of M. truncatula R108 (WT) and stf-2 mutant, N. sylvestris (WT) and lam1 mutant, Arabidopsis Ler (WT), wox1 prs double mutant and p35S::YFP-GR-STF, p35S::YFP-GR-LAM1 and p35S::YFP-GR-NsWOX9 transgenic lines were isolated using TRIzol Reagent (Invitrogen, Carlsbad, CA, USA) for cDNA synthesis.

Real-time PCR

Expression patterns analyses were performed using quantitative real time polymerase chain reaction (qPCR) and semi-quantitative PCR with specific forward and reverse primers (Supporting Information Table S1). Reverse transcription (RT) was performed using RNA treated with DNase I (Invitrogen), an oligo (dT) primer, and SuperScript III reverse transcriptase (Invitrogen) according to the manufacturer’s instruction. The RT-qPCR assays were performed with three biological and three technical replicates using SYBR Green real-time PCR Master Mix (Invitrogen). Medicago truncatula, Arabidopsis and N. sylvestris actin primers were used as standards. All specific forward and reverse primers used for gene cloning and expression and related molecular analyses in this study are listed in Table S1.

Gene isolation and transgene construction

MtWOX9-1 and MtWOX9-2 genes were isolated from the M. truncatula genome by BLAST search using the AtWOX9 sequence. Full-length MtWOX9-1 and MtWOX9-2 coding sequences were amplified by RT-PCR using total RNA extracted from leaf samples, and constructs were made via gateway cloning using pDONR207 and pMDC32 vectors. The resulting plasmids were transferred into Agrobacterium tumefaciens strains AGL1 and GV2260 for M. truncatula and N. sylvestris, respectively, transformation as previously described (Tadege et al., 2011a). For WOX9 antisense (NsWOX9-anti) construct, NsWOX9-anti-F and NsWOX9-anti-R primers were used to amplify the full length CDS from the NsWOX9-cDNA and clone into the Gateway vectors. All generated transgenic lines were confirmed by PCR using specific primers, and reduced expression lines were identified by RT-PCR.

Multiplex gRNA-CRISPR/Cas9 construction

To generate the multiplex gRNA-CRISPR/Cas9 NsWOX9 construct, we designed three multiplex gRNAs targeting multiple sites of exon 2 (gRNA1-CTTCAAGAATATGGCCA AGT; gRNA2-CTTCAAGAATATGGCCAAGT and gRNA3-TCTCCTGCTGTTATCA CACA) at the upstream of PAM (TGG, TGG, and AGG) sites, respectively, using the web-based tool CRISPR-P (http://cbi.hzau.edu.cn/cgi-bin/CRISPR; Lei et al., 2014). All designed gRNAs were inserted between tRNA and gRNA scaffolds and clustered in tandem using the Golden Gate assembly method (Engler et al., 2008). The pGTR plasmid, which contains a tRNA-gRNA fragment, was used as a template to synthesize polycistronic tRNA-gRNA (PTG) (Xie et al., 2015). The overlapping PCR products were separated and purified by the Spin Column PCR Product Purification Kit (Wizard SV Gel and PCR Clean-Up System) following manufacturer’s instruction (Promega, Madison, WI, USA). Then, the chain of multiplex tRNA-gRNA with three NsWOX9-spacers was inserted into an optimized vector with AtU6-tRNA-gRNAs-AtUBQ10-Cas9-pRGEB31-bar backbone by digestion and ligation using Fok I (NEB) and BsaI enzymes (Xie et al., 2015). Target gRNAs and primers used in this study are listed in Table S1.

CRISPR/Cas plant transformation and analysis

The subsequent multiplex NsWOX9spacer-gRNA-CRISPR/Cas9 binary vector construct was transformed into Agrobacterium strain GV2260. First, transgenic lines were screened by PCR using genomic DNA and specific primers (PPT-F+PPT-R) of Barsta selection marker. Then, putative Nswox9-CRISPR mutants were identified through amplification of the target region by PCR using extracted genomic DNA as a template with specific primers designed from the border of the target site. Amplified fragments of the mutated region were sub-cloned into pGEM-T easy plasmid by TA-Cloning and 10 colonies for each locus were subjected to Sanger sequencing. Reads were analyzed by aligning with the reference sequence using the SeqMan Pro 15.0.1 (Dnastar software for life scientists) (https://www.dnastar.com/quote-request/).

Electrophoretic mobility shift assay

The MtWOX9-1 promoter with 37, 30 and 32 bp oligonucleotides, corresponding to starting sites at −22, −226 and −491, respectively, upstream of the start codon were labeled using the Biotin 3ʹ End DNA Labeling Kit according to the manufacturer’s instructions (Thermo Scientific/Pierce, Waltham, MA, USA). Electrophoretic mobility shift assay (EMSA) was performed using Light Shift Chemiluminescent EMSA Kit (Thermo Scientific/Pierce). Unlabeled probes with a 50-fold higher concentration were used as competitors in each of the competing assays. Purified maltose binding protein (MBP) and MBP-STF were used in the EMSA on 10% native polyacrylamide gels as described (Zhang et al., 2014). The biotin-labeled DNA was detected by chemiluminescence and exposed to X-ray film (Kodak, Rochester, NY, USA).

Chromatin immunoprecipitation (ChIP) assays

Chromatin immunoprecipitation (ChIP) assays were performed as described previously (Xiong et al., 2013; Chen et al., 2018). Protoplast extracted from 14-d-old pMtWOX9-1:MtWOX9-1/wox1prs transgenic Arabidopsis leaves were transformed with 10 µg of 35S::STF-YFP using the polyethylene glycol-mediated transformation method. Protoplast crosslinking, chromatin isolation and preclearing were performed as described (Chen et al., 2018). Precleared chromatins were incubated with 5 μl of anti-GFP antibody (ab290) overnight at 4°C, after which Protein A agarose beads (40 μl) were added, and the samples were incubated at 4°C for 2 h. After reversing the crosslinks with Proteinase K at 65°C overnight, DNA was purified and analyzed by qPCR amplification using specific primers. The input DNA and HA antibody-precipitated DNA were used as PCR templates for the positive and negative controls, respectively. Experiments were repeated three times.

Dual luciferase assay

For effector plasmids, the coding sequence of STF or GUS was cloned into pDONR207 entry vector and then transferred into p2GW7 using the Gateway system (Invitrogen). Construction of the reporter pMtWOX9-1-mini-35S-LUC plasmid containing 1 kb of the MtWOX9-1 promoter was performed as previously described (Zhang et al., 2014). Transient expression assays were performed in Arabidopsis protoplasts as described (Asai et al., 2002). For normalization, 0.5 µg of plasmid pRLC was used as an internal control.

Histological analysis

Leaf samples were fixed in formaldehyde for 48 h and dehydrated in an ethanol series (60%, 70%, 85% and 95%). Then, leaves were embedded in Paraplast (Sigma-Aldrich, St Louis, MO, USA) and tissue sections (15 μm thick) were cut with a Reichert-Jung 2050 microtome. Specimens were mounted on slides and stained with Safranine O and Light Green as previously described (Tadege et al., 2011a). Images were captured with digital camera mounted on an Olympus BX-51 compound microscope.

RNA in situ hybridization

Templates of RNA probes were derived from MtWOX9-1 cDNA with gene specific primers containing T7 promoter sequence at the 5′ end, and the RNA probes were end-labeled with DIG oligonucleotide 3′ end-labeling kit. The vagetative shoot buds of 3-wk-old R108 plants grown in a glasshouse under LD conditions were used for RNA in situ hybridization. In situ hybridization was performed according to a standard protocol (Weigel & Glazebrook, 2002).

Sequence alignment and phylogenetic tree construction

Multiple protein sequence alignment was performed using BioEdit software and the ClustalW program. Species refer to At (Arabidopsis thaliana), Ns (Nicotiana sylvestris), Pc (Phaseolus coccineus), Gm (Glycine max), Vv (Vitis vinifera), Ph (Petunia × hybrid), Cs (Cucumis sativus), Sl (Solanum lycopersicum) and Ca (Capsicaum annuum). A neighbor-joining phylogenic tree was constructed using MEGA-X default settings with 1000 bootstrap replications.

Accession numbers

Sequence data from this article can be found in the National Center for Biotechnology Information (NCBI) Genbank database, or M. truncatula Genome Database (http://www.medicagogenome.org/).

Corresponding to: MtWOX9-1, Medtr2g015000; MtWOX9-2, Medtr7g026130; AtWOX9, AT2G33880; AtWOX8, AT5G45980; NsWOX9, XM_009794999; NsLAM1, AEL30893; MtSTF, JF276252; AtWOX1, AT3G18010; AtWOX3/PRS, AT2G28610; PcWOX9, ACL11801; GmWOX9, XP_006594207; GmWOX9, XP_003541514; VvWOX9, XP_002273188; CaWOX9, XP_016562050; CsWOX9, XP_004134676; PhEVG, ABO93066; PhSOE; ABO93067; SlWOX9/S, NP_001234072.


Ectopic expression of WOX9 enhances the stf and lam1 mutant phenotypes

We have previously shown that, while the WUS clade repressor WOX genes of Arabidopsis including WUS and WOX1WOX7 complement the lam1 mutant phenotype when driven by the STF promoter, the intermediate clade WOX9 expression exacerbates the lam1 mutant phenotype (Lin et al., 2013a). Therefore, we decided to investigate whether the unique activity of WOX9 is conserved in M. truncatula and N. sylvestris and determine its biological significance in leaf development. We isolated orthologous coding sequences for AtWOX9 from M. truncatula (MtWOX9) and N. sylvestris (NsWOX9) and created expression constructs that placed these sequences under control of the STF promoter to ectopically express these genes in the stf-2 and lam1 mutant plants. The STF promoter drives expression specifically at the adaxial–abaxial juxtaposition of leaf primordia in M. truncatula (Tadege et al., 2011a) similar to the expression pattern of WOX1 in Arabidopsis (Nakata et al., 2012). Stf-2 and lam1 are severe leaf blade mutants in M. truncatula and N. sylvestris, respectively, caused by mutation of the WOX1 orthologs STF and LAM1 (Fig. 1b,e). The M. truncatula genome contains two WOX9-like sequences here designated as MtWOX9-1 and MtWOX9-2 (Supporting Information Figs S1, S2). We introduced MtWOX9-1 driven by the STF promoter (STF::MtWOX9-1) first into the stf-2 M. truncatula plants and eight independent transgenic lines were generated. Expression of MtWOX9-1 in the stf-2 mutant background was confirmed by RT-PCR assays. All of these transgenic lines displayed strongly enhanced mutant phenotype with much narrower leaves and thinner stems compared to the stf mutant (Fig. 1a–c). In addition, leaves and stems were significantly shorter in length, leading to a dwarf phenotype that was not characteristic of the stf mutant phenotype (Fig. 1b,c). Similarly, introduction of this construct into the lam1 mutant background severely affected both leaf length and width, exacerbating the lam1 mutant phenotype (Fig. 1d–f). These results indicate that both the stf and lam1 mutants respond similarly to the activity of MtWOX9-1, consistent with the effect of AtWOX9 expression driven by the STF promoter in the lam1 mutant (Lin et al., 2013a). We also transformed 35S::MtWOX9-1, 35S::MtWOX9-2 and 35S::NsWOX9 into the lam1 mutant and obtained severely enhanced mutant phenotypes similar to the STF::MtWOX9-1 expressing lam1 lines (Fig. 1g–i), indicating that these three genes have similar effects on leaf blade outgrowth. In most cases, these transgenic leaves displayed approximately five-fold reduction in leaf length, but this effect appeared to be dependent on the level of WOX9 transgene expression since plants with high level of transgene expression showed more severe phenotypes compared to lam1 plants with low level of transgene expression (Fig. S3).

Details are in the caption following the image
Ectopic expression of WOX9 enhances stf and lam1 mutant phenotypes. (a) Untransformed Medicago truncatula wild type (WT) (R108) plant. (b) Phenotype of untransformed stf mutant. (c) stf mutant transformed with STF::MtWOX9-1. (d) Untransformed Nicotiana sylvestris WT plant. (e) Phenotype of untransformed lam1 mutant. (f) lam1 mutant transformed with STF::MtWOX9-1. Insets in (b), (c), (e) and (f) show magnified areas of individual leaves. (g) lam1 mutant transformed with 35S::MtWOX9-1. (h) lam1 mutant transformed with 35S::MtWOX9-2. (i) lam1 mutant transformed with 35S::NsWOX9. Plants were 10-wk (e and f) or 5-wk old (all the rest). Bars, 10 cm.

The enhancement of stf and lam1 mutant phenotypes associated with WOX9 expression suggested to us that WOX9 acts in opposition to STF/LAM1 in leaf blade development. Therefore, we introduced an NsWOX9-antisense construct into the lam1 plants to see if reduced levels of NsWOX9 transcripts could alleviate the mutant phenotype. Indeed, expression of an NsWOX9-antisense construct had the opposite effect of NsWOX9 overexpression, partially rescuing the lam1 mutant leaf phenotype (Fig. 2). However, this partial complementation was limited, and the antisense plants still appeared bushy and failed to make stems. Nonetheless, unlike the untransformed lam1 mutant control, the NsWOX9-antisense leaves showed distinct petioles and blades especially at early stages of development (Fig. 2a,b), and the blades were variously branched and curled resulting in unusual leaf structure at maturity (Fig. 2c–e). These leaf phenotypes suggest that blade outgrowth initiation has significantly progressed in the NsWOX9-antisense plants but perhaps aborted before completion.

Details are in the caption following the image
NsWOX9-antisense partially rescues the lam1 mutant phenotype of Nicotiana sylvestris. (a) Phenotype of lam1 mutant transformed with 35S::GUS as control at 3 wk of age. Inset on the right is a detached leaf close up. (b) Partially complemented lam1 phenotype transformed with 35S::NsWOX9-antisense construct at 3 wk, the inset is a close up of a partially complemented leaf blade. Inset on the right is a detached leaf close up. (c) Partially complemented lam1 phenotype transformed with 35S::NsWOX9-antisense construct at 7 wk. (d) Representative individual leaves from 35S::NsWOX9-antisense/lam1 plants. (e) A magnified view of a leaf in (d). Note the branching and curling of leaves especially in older 35S::NsWOX9-antisense/lam1 plants. Bars: (a–d) 5 cm; (e), 1.5 cm.

To more completely evaluate the effects altered WOX9 expression on the lam1 mutant phenotype, we carried out structural examination of leaf tissues of WT, lam1 mutant, WOX9 overexpressing lam1, and WOX9 antisense lam1 plants (Fig. 3a–d). Transverse sections through the leaf blades showed that the lam1 leaves had vestigial blade strips at the position of WT blades (Fig. 3f), but these strips were completely absent and blades became fully radialized in NsWOX9 overexpressing lam1 lines (Fig. 3g). In NsWOX9-antisense lam1 lines, however, distinct blade outgrowth was apparent but the nascent blades were not fully expanded compared to the WT leaves (Fig. 3e,h), confirming that blade development in the NsWOX9-antisense plants initiated more effectively than in the lam1 plants but was not completed. Taken together, these results indicate that WOX9 functions oppose those of STF/LAM1 in two unrelated eudicot species M. truncatula and N. sylvestris.

Details are in the caption following the image
Transverse section of the leaf blade showing enhancement of the lam1 blade by 35S::NsWOX9 and partial complementation by 35S::NsWOX9-antisense. (a) Nicotiana sylvestris wild-type (WT) leaf blade. (b) Leaf blade of untransformed lam1 mutant control. (c) Leaf blade of lam1 transformed with 35S::NsWOX9. (d) Partially complemented leaf blade of lam1 transformed with 35S::NsWOX9-antisense. (e) Transverse section of N. sylvestris WT leaf blade. (f) Transverse section of untransformed lam1 mutant leaf blade. (g) Transverse sections of lam1 leaf blade transformed with 35S::NsWOX9 showing radialized blade. (h) Transverse sections of lam1 leaf blade transformed with 35S::NsWOX9-antisense showing blade outgrowth. Arrows indicate blade tissue in (e) and (h), vestigial blade stripes in (f) and position of blade in (g). Bars, 50 μm.

WOX9 overexpression severely affects leaf architecture

To further examine the effect of WOX9 in leaf blade outgrowth, we introduced 35S::MtWOX9-1, 35S::MtWOX9-2 and 35S::NsWOX9 into WT N. sylvestris. Analysis of over 20 independent transgenic lines for each construct revealed that all transgenic lines displayed an array of leaf phenotypes that can generally be grouped into severe and mild based on the phenotype strength. The WT N. sylvestris leaf blade is a well-expanded flat lamina with smooth margin and distinctive pinnate venation pattern (Fig. 4a). WOX9 overexpression disrupted this pattern resulting in highly disorganized and distorted leaf forms. These include narrow and downward curling blades, deep margin serrations, disorganized venation patterns, uneven blade surfaces, as well as retarded plant growth with two to five tillers and additional leaves leading to a bushy appearance until stem elongation at a later developmental stage (Figs 4b,d,e,g, S4). While most overexpressing plants showed this strong phenotype, some exhibited a mild phenotype where the blade margin, shape and venation patterns were largely intact but with puckered and uneven blade surfaces (Figs 4c,f, S4). MtWOX9-1 overexpression in M. truncatula also produced downward curling and narrower leaves similar to that seen in N. sylvestris (Fig. S5), albeit more mild, with an insignificant effect on tillering. Microscopic examination of the leaf epidermal cells revealed that NsWOX9 overexpressing cells were larger in size, expanded especially in the width direction and the number of cells per unit area was 2–3 times fewer than WT cells (Fig. S6), suggesting that WOX9 may be involved in cell differentiation/expansion. These WOX9 overexpression phenotypes together suggest that WOX9 antagonizes the function of STF/LAM1 in leaf blade outgrowth probably by regulating cell differentiation unlike STF’s primary role in cell proliferation (Tadege et al., 2011).

Details are in the caption following the image
WOX9 ectopic expression in wild-type (WT) Nicotiana sylvestris alters leaf architecture. (a) WT N. sylvestris control plant at 7 wk. (b) Phenotype of 35S::MtWOX9-1/WT (severe phenotype) at 7 wk after regeneration. (c) Phenotype of 35S::MtWOX9-1/WT (mild-phenotype) at 7 wk after regeneration. (d) Phenotypes of different individual representative leaves from different 35S::MtWOX9-1/WT plants showing severe phenotypes at variable stages. (e) Phenotype of 35S::MtWOX9-2/WT (severe phenotype) at early growth stage. (f) Phenotype of 35S::MtWOX9-2/WT (mild) at early growth stage. (g) Phenotype of 35S::NsWOX9/WT (severe) at early growth stage. Bars: (a–c) , (e–g) 10 cm; (d) 3 cm.

Deleting NsWOX9 using multiplex gRNA CRISPR/Cas9 genome editing in N. sylvestris alters blade symmetry and expansion

To gain insight into the function of the endogenous WOX9 gene in WT plants, we disrupted NsWOX9 in N. sylvestris using CRISPR/Cas9 genome editing technology. We constructed an NsWOX9-multiplex gRNA-CRISPR/Cas9 vector containing three guide RNAs; gRNA1, gRNA2 and gRNA3 (Fig. 5a), and introduced this construct into N. sylvestris. Sixteen putative mutant lines were identified by PCR amplification of the target regions using specific primers and Sanger sequencing. Out of these 16 putative mutants, five representative lines were selected for further characterization of their mutant phenotypes (Fig. 5b). Target site sequence analysis of the five NsWOX9-CRISPR-mutants labeled here as NsWOX9-1, NsWOX9-2, NsWOX9-13, NsWOX9-18 and NsWOX9-22 using SeqMan Pro 15.0.1 (Dnastar software) revealed four different patterns of deletions ranging from 14 to 183 bp (Fig. 5b). NsWOX9-1 and NsWOX9-13 had identical 14 bp deletions in gRNA3 and considered as one line. NsWOX9-2 showed two deletion events at gRNA2 and gRNA3 with 7 and 21 bp deletions, respectively. NsWOX9-22 showed a 73 bp deletion spanning the upstream and downstream region of gRNA3, while the largest deletion was detected in line NsWOX9-18 where a 183 bp region between gRNA1 and gRNA2 was removed, which included the PAM region of gRNA1 and extended to three nucleotides upstream of the PAM of gRNA2 (Fig. 5b). All of these CRISPR-derived mutant lines displayed malformed leaves including narrow blades, blade asymmetry, partial deletion in the blade half, rough blade surface, leaf shape distortions, multiple tillers, early flowering and sterility (Fig. 5c–i), indicating that NsWOX9 is required for bilateral symmetry and proper leaf blade development in N. sylvestris.

Details are in the caption following the image
Knock out of NsWOX9 with multiplex gRNA-CRISPR/Cas9 in Nicotiana sylvestris alters leaf architecture and symmetry. (a) Schematic representation of the three guide RNAs NsWOX9-gRNA1, -gRNA2 and -gRNA3 inserted in pRGEB31-bar-AtUbi10-AtU6-tRNA-gRNA vector. (b) Mutation events detected at the corresponding target sites of gRNA1, gRNA2 and gRNA3 in five independent CRISPR/Cas lines (NsWOX9-1, NsWOX9-2, NsWOX9-13, NsWOX9-18 and NsWOX9-22). (c–i) Phenotype of CRISPR/Cas9 edited plants and leaves; wild-type (WT) control (c), edited NsWOX9-13 mutant (d), edited NsWOX9-22 mutant (e), edited NsWOX9-18 mutant (f), edited NsWOX9-2 mutant (g), control WT leaf blade (h), representative individual leaf blades from edited plants (i); left, half blade deleted (NsWOX9-22), and right, narrow and asymmetric blade (NsWOX9-2). Red arrows point to blade defects, white arrows show multiple shoots. Bars: (c–g) 5 cm; (h), (i) 2.5 cm.

The MtWOX9-1 transcript is detected at the leaflet initiation site and abaxially expressed in leaf primordia but not detected in the SAM

RT-qPCR analysis showed that expression of MtWOX9-1 was relatively weak in most plant tissues, including leaves, while higher levels of expression were detected in the shoot apex, flowers, pods and immature seeds, with highest levels detected in developing seeds 10 d after anthesis, followed by expression in flowers (Fig. 6a). Highest expression of MtWOX9-2, however, was detected in the leaves followed by expression in shoot apices (Fig. S7). However, MtWOX9-2 is distantly related to MtWOX9-1 (Fig. S2) and it is not clear at this stage if it performs a redundant function other than the overexpression phenotypes (Figs 1, 4). To further examine expression in different tissues with a different method, we fused a 3 kb promoter region of MtWOX9-1 upstream of the translational start codon to the β-glucuronidase (GUS) coding region, and transformed it into Medicago R108 leaf explants. GUS staining analysis revealed that expression in the mature leaf was relatively weak with particularly strong expression in the pulvinus at the base of the leaflets (Fig. S8a). Expression was detected in the flowers and stems (Fig. S8b,c) but, not in the anthers (Fig. 8c). Intense staining was detected in immature pods and seeds at early stages of development (Fig. S8d,e). Expression in the seed became progressively restricted as the seed develops, and confined only to the hilum in the mature seed (Fig. S8f).

Details are in the caption following the image
MtWOX9-1 is abaxially expressed in leaf primordia and strongly upregulated in the stf mutant leaf of Medicago truncatula and corresponding mutants in Nicotiana sylvestris and Arabidopsis. (a) RT-qPCR analysis showing relative expression of MtWOX9-1 in different tissues of M. truncatula. Leaf, stem, and shoot apex were from 4-wk-old plants, pods and seeds were 10 d after pollination, flower at anthesis, all the rest were at seedling stage. (b–f) RNA in situ hybridization in vegetative shoot apex of 3-wk-old M. truncatula seedlings using MtWOX9-1 antisense probe. (b, c) Continuous slice of the SAM in a longitudinal view. (d) Longitudinal section of the SAM viewed from a different angle to show the AM. (e) Expanded young leaf. (f) Unexpended immature leaf. (g) Control with MtWOX9-1 sense probe. Bars, 50 µm. P1–P5: different developmental stages of the leaf primordia. Pet, petiole; SAM, shoot apical meristem; AM, axillary meristem; St, stipule; TL, terminal leaflet; LL, lateral leaflet; AD,adaxial; AB, abaxial. (h) Relative expression of MtWOX9-1 in the leaf of 4-wk-old M. truncatula stf-2 mutant. (i) Relative expression of NsWOX9 in the leaf of 4-wk-old Nicotiana sylvestris lam1 mutant. (j) Relative expression of AtWOX9 in the leaf of 4-wk-old Arabidopsis wox1 prs double mutant. Error bars indicate ± SE (n = 3). Asterisks indicate significant difference from the control (*, P < 0.05; **, P < 0.01; Student's t-test).

To examine the expression pattern of MtWOX9-1 in more detail at early stages of leaf development, we performed RNA in situ hybridization in the shoot apiece of 3-wk-old M. truncatula seedlings using MtWOX9-1 specific antisense probe. We found that MtWOX9-1 transcript was detected in leaf primordia and at the boundary region between the SAM and leaf primordia, but absent from the central SAM and axillary meristem (Figs 6b–d, S9). A serial section through the SAM and primordia revealed that the transcript was first detected in the peripheral region of the SAM that marks the position of the emerging primordium (P0) and continued to be expressed in leaf primordia to P5 (Figs 6b,c, S9). At P2 and early P3 stages, expression was detected at the leaflet (LL) initiation site, between leaf primordia and stipule (Fig. S9b), and at the boundary region between the terminal leaflet (TL) and emerging LLs (Fig. S9c). In late P3 and P4, the signal was detected at the basal region of the TL and LLs but not in the petiole (Pet) Fig. S9d,e). Interestingly, in expanded and unexpanded young leaves (late P3 to late P5), expression was found restricted to the abaxial side of the leaf with no signal whatsoever on the adaxial side (Fig. 6e,f). No signal was detected in SAM or leaf primordia in the negative control using sense probe (Fig. 6g). These results indicate that MtWOX9-1 is an abaxial factor expressed starting very early in initiating leaf primordia but conspicuously absent from the central SAM, suggesting that it may be involved in cell differentiation during leaf morphogenesis and development.

MtWOX9-1 is directly repressed by STF in M. truncatula

To investigate the mechanistic relationship between WOX9 and key regulators of blade outgrowth, we examined the leaf blade expression levels of WOX9 in stf-2, lam1, and wox1 prs, mutants of M. truncatula, N. sylvestris, and Arabidopsis thaliana respectively. RT-qPCR analyses showed that expression of MtWOX9-1, NsWOX9, and AtWOX9 was upregulated by 2–4 fold in the leaves of stf-2, lam1 and wox1 prs mutants compared to their respective WT levels (Fig. 6h–j), indicating that WOX9 may be directly or indirectly repressed by the action of STF/LAM1/WOX1 in WT leaves.

To examine whether MtWOX9-1 is a direct target of STF, we performed a dexamethazone (DEX) induction experiment using the glucocorticoid (GR) system in the presence of the protein synthesis inhibitor, cycloheximide (CHX). Analysis was performed in 4-wk-old stf mutant plants transformed with the 35S::YFP-GR-STF construct. The shoot apex and young leaves of the transgenic plants were treated with both DEX and CHX for 3 h, and MtWOX9-1 transcript accumulation was monitored by RT-qPCR in the leaves with and without the induction treatment. Our results showed that the expression of MtWOX9-1 was reduced by c. 60% in the DEX and CHX treated lines compared to the control CHX alone (Fig. 7a). Since new protein synthesis is inhibited by CHX in the treated lines, this result suggests that MtWOX9-1 may be directly repressed by STF. We repeated this experiment in N. sylvestris using 35S::YFP-GR-LAM1 c. 60% repression of NsWOX9 upon LAM1 induction by DEX plus CHX treatment (Fig. 7b). However, in the reciprocal experiment, induction of NsWOX9 expression had no significant effect on LAM1 expression (Fig. 7c), indicating that NsWOX9 does not regulate LAM1 transcription. While NsWOX9 induction gave a typical WOX9 overexpression phenotype in the sprayed leaf after 1 wk (Fig. S10a,b), LAM1 induced leaves looked normal (Fig. S10c,d). We also looked at the transcript level of LAM1 and NsWOX9 by semi-quantitative RT-PCR using UBI as a loading control in two NsWOX9 stable overexpression lines, two NsWOX9 CRISPR lines and lam1 mutant. Consistent with the GR induction results, we found that LAM1 expression was not affected in the NsWOX9 overexpression and knockout lines (Fig. S11).

Details are in the caption following the image
WOX9 expression is directly repressed by GR induction of STF or LAM1 in the presence of cyclohexamide, and by STF in Dual Luciferase assay. (a) Relative expression of MtWOX9-1 in 35S::YFP-GR-STF transformed Medicago truncatula lines with CHX or DEX + CHX treatment for 3 h in 4-wk-old leaves. (b) Relative expression of NsWOX9 in two independently transformed 35S::YFP-GR-LAM1 Nicotiana sylvestris lines with CHX alone, CHX + DEX or DEX alone treatments for 6 h in 4-wk-old leaves. (c) Relative expression of LAM1 in 35S-YFP-GR-NsWOX9 transformed two independent N. sylvestris lines with CHX alone, CHX + DEX or DEX alone treatments for 6 h in 4-wk-old young leaves. Relative gene expression was determined by RT-qPCR analyses. (d) Schematic representation of reporter and effector constructs used in the dual luciferase assay. (e) Relative expression of luciferase activity (luminescence) in the presence of 35S::STF effector compared with the 35S::GUS control in Arabidopsis protoplasts. Error bars indicate ± SE (n = 3). Asterisks indicate significant difference from the control (*, P < 0.05; **, P < 0.01; ***,P < 0.001; Student's t-test).

To determine if STF reduces MtWOX9-1 transcript accumulation by targeting its promoter, we performed dual luciferase assay in Arabidopsis protoplasts using the Firefly-Renilla Dual-Luciferase assay system (Promega). In the reporter construct, a 1 kb promoter region of MtWOX9-1 upstream of the translation start codon was fused to a mini 35S promoter driving the luciferase reporter gene (Fig. 7d), while the effector constructs were made using either STF, or GUS as control, both driven by the 35S promoter (Fig. 7d). Consistent with the earlier results, co-expression of the STF effector in the protoplast almost fully abolished luciferase luminescence compared to the GUS effector control (Fig. 7e), indicating that this 1 kb region of the MtWOX9-1 promoter is sufficient for STF-dependent repression of transcription.

To determine if STF indeed binds to the MtWOX9-1 promoter in vitro and in vivo, we performed EMSA using biotin labeled probes, and ChIP assay, using anti GFP antibody. We previously reported that STF binds preferentially to ‘AT-rich’ DNA elements without a strong consensus sequence (Zhang et al., 2014). We screened six such selected regions in the 3 kb upstream region of the MtWOX9-1 promoter, and found that the MBP-STF fusion protein was able to bind to three of them (Fig. 8a). These STF-binding elements are located at −22, −226 and −491 bp upstream of the MtWOX9-1 CDS while binding was not detected with the control MBP alone (Fig. 8b). Each of these sites were significantly competed by addition of 50-fold excess of the respective unlabeled probes, indicating binding specificity. This shows that, at least in vitro, STF can directly bind to multiple sites within the 1 kb fragment of the MtWOX9-1 promoter, consistent with the dual luciferase assay and GR induction experiments.

Details are in the caption following the image
STF directly binds to the MtWOX9-1 promoter in EMSA and ChIP assays. (a) Schematic representation of the MtWOX9-1 promoter and CDS regions tested for EMSA and ChIP assays. The three promoter regions tested are indicated as P1, P2 and P3 and the MtWOX9-1 coding region as C1. (b) EMSA showing MBP-STF bound to the biotin-labeled probe at P1, P2 and P3 promoter fragments but not the MBP control alone. Fifty-fold excess of unlabeled P1, P2 or P3 DNA was used to compete with the respective labeled probe (right lanes). (c) MtWOX9-1 promoter enrichment at P1, P2, P3 regions and CDS C1. Chromatin precipitated with anti-GFP antibody from 35S::STF-GFP and 35S::GFP control Arabidopsis lines were compared. Purified DNA from the chromatin were used as templates for qPCR. Note that P2 is highly enriched in 35S::STF-GFP samples. Error bars indicate ± SE (n = 3). Asterisks indicate statistical significance (*, P < 0.05; ***, P < 0.001; Student's t-test).

To confirm that STF binds to the MtWOX9-1 promoter in vivo, we performed ChIP assays in leaves of Arabidopsis wox1 prs double mutant plants transformed with pMtWOX9-1::MtWOX9-1 construct. Protoplasts were isolated from these transgenic plants and transformed with 35S::STF-GFP fusion from which chromatin was isolated for analysis. ChIP-qPCR analysis revealed that among the three tested promoter regions (P1, P2 and P3), STF was highly enriched at P2 with significant enrichment also at P3 (Fig. 8c). By contrast, no significant enrichment was detected at P1 or within the MtWOX9-1 coding region (C1), indicating that STF binds strongly at the P2 position and to a limited extent at the P3 region in planta. Taken together, these results indicate that STF directly binds to the proximal region of the MtWOX9-1 promoter in M. truncatula, and represses its activity, and that STF and WOX9 function antagonistically to regulate leaf blade outgrowth.


Plant specific WOX transcription factors regulate a variety of plant developmental programs from embryogenesis to shoot apical meristem maintenance and lateral organ development (Mayer et al., 1998; Schoof et al., 2000; Lohmann et al., 2001; Matsumoto & Okada, 2001; Nardmann et al., 2004; Sarkar et al., 2007; Breuninger et al., 2008; Shimizu et al., 2009; Vandenbussche et al., 2009; Ji et al., 2010; Tadege et al., 2011a; Nakata et al., 2012). WOX family members are also known for their promiscuous ability to substitute for each other’s functions. For example, in Arabidopsis, WUS complements the prs/wox3 and wox5 mutants, which are defective in floral organ development and root apical meristem maintenance, respectively (Sarkar et al., 2007; Shimizu et al., 2009). Conversely, members of the WUS clade WOX genes (WOX1–WOX7), with the exception of WOX4, can substitute for WUS function in stem cell maintenance (Dolzblasz et al., 2016). Arabidopsis WUS and WOX1–WOX7 can also complement the lam1 leaf blade mutant of N. sylvestris (Tadege et al., 2011a; Lin et al., 2013a). Here we show that the M. truncatula and N. sylvestris WOX9 homologs, MtWOX9-1, MtWOX9-2 and NsWOX9 function antagonistically to STF/LAM1 in regulating leaf blade outgrowth. Ectopic expression of these genes enhanced the stf and lam1 leaf mutant phenotypes, and severely affected blade expansion and morphology in WT N. sylvestris with a range of phenotypes (Figs 1, 4). Conversely, reducing NsWOX9 transcript levels in the lam1 mutant with antisense technology partially complemented the mutant phenotype (Fig. 2), indicating that WOX9 antagonizes the STF/LAM1 function. Complete knockout of NsWOX9 by CRISPR/Cas9 genome editing in the WT background resulted in a range of leaf blade deformations including lack of bilateral symmetry, altered venation patterns, narrow blades and bushy shoots (Fig. 5), indicating that WOX9 function is required for proper leaf blade development.

In petunia and tomato, WOX9 homologs are involved in inflorescence development and architecture (Lippman et al., 2008; Rebocho et al., 2008; Costanzo et al., 2014). Both the evergreen (evg) mutant in petunia (Rebocho et al., 2008) and compound inflorescence (s) in tomato (Lippman et al., 2008), which dramatically alter the WT inflorescence architecture are caused by lesions in WOX9 homologs. The s allele in tomato results in a highly branched structure with hundreds of flowers, which increases fruit production and may have been selected by breeders (Lippman et al., 2008). In the evg mutant of petunia, however, the inflorescence stem often fails to bifurcate after the formation of bracts and continues to grow as a single thickened stem without physical separation of the FM and inflorescence meristem (IM), leading to a fasciated appearance (Rebocho et al., 2008). Unlike s, the evg FM also fails to produce floral organs suggesting that EVG is required for inflorescence bifurcation and floral organ identity, though evg mutants are indistinguishable from WT during early vegetative growth (Rebocho et al., 2008). Thus, in tomato and petunia, WOX9 homologs appear to have opposite effects specific to inflorescence development and architecture. However, both tomato and petunia have a cymose inflorescence pattern (determinate growth) and it is unclear whether these inflorescence-associated defects are specific to cymose or are also exhibited by racemose (indeterminate growth) and panicle (mixed inflorescence) inflorescences. At least in Arabidopsis (racemose inflorescence), the role of WOX9 appears not to be restricted to inflorescence development, and in rice (mixed inflorescence), WOX9 is involved in uniform tiller growth and development (Wang et al., 2014; Fang et al., 2020).

In Arabidopsis, WOX9, also called STIMPY (STIP), is required for meristem growth and maintenance and positively regulates WUS (Wu et al., 2005). Futhermore, stip mutants display arrested growth at an early stage of development but can be fully rescued by sucrose (Wu et al., 2005). STIP/WOX9 is shown to mediate cytokinin signaling during shoot meristem establishment and, together with WOX2 and WOX8, regulates zygote and embryo polarity patterning (Wu et al., 2007; Breuninger et al., 2008; Skylar et al., 2010; Ueda et al., 2011). WOX9 homologs in other species are also reported to be involved in promoting somatic embryogenesis (Gambino et al., 2011; Tvorogova et al., 2019). The Arabidopsis gain-of-function mutant stip-D displays wavy leaf margins and increased number of axillary shoots leading to a bushy phenotype (Wu et al., 2005). This phenotype is similar to the wavy margins and bushy shoots seen in all MtWOX9-1, MtWOX9-2 and NsWOX9 overexpressing N. sylvestris transgenic lines (Fig. 4). We uncovered that MtWOX9-1 is an abaxial factor expressed on the abaxial side of leaf primordia excluded from the central SAM and axillary meristem (Figs 6, S9). At early stages of development, expression was detected at the boundary region between SAM and leaf primordia, leaf primordia and stipules, and between terminal and lateral leaflet. In older leaf primordia (late P3 and afterwards), expression was detected in the basal region of the terminal and lateral leaflets, notwithstanding abaxial-specific expression. In addition, overexpression of NsWOX9 in N. sylvestris resulted in fewer but laterally enlarged cells compared to WT (Fig. S6), suggesting that WOX9 may be involved in organ separation and cell differentiation in the abaxial side of leaf primordia. This is consistent with the observed effect of WOX9 in leaf development and its antagonistic activity to STF. STF is expressed at the adaxial–abaxial junction and promotes cell proliferation (Tadege et al., 2011a). Thus, one aspect of the STF and WOX9 anatagonistic function is that STF promotes cell proliferation in the middle domain while WOX9 promotes cell differentiation in the abaxial domain. Such antagonistic function has been demonstrated between STF and the adaxial factor AS2 (Zhang et al., 2014). It is conceivable that a similar scenario may exist between STF and the abaxial factor WOX9 since STF/WOX1 is likely to negatively regulate both adaxial and abaxial factors to protect the cell proliferation zone in the middle domain from differentiation (Nakata et al., 2012).

The Arabidopsis WOX family has been divided into three clades based on phylogenetic analysis: the modern/WUS clade (WUS, WOX1–7), intermediate clade (WOX8, WOX9, WOX11, WOX12), and ancient clade (WOX10, WOX13, WOX14) (van der Graaff et al., 2009). This classification is largely consistent in other species as well (Zhang et al., 2010; Hao et al., 2019; Wu et al., 2020). The WUS clade members are characterized by an intact WUS box motif (Haecker et al., 2004; Lin et al., 2013a), which is a transcriptional repression motif (Ikeda et al., 2009; Lin et al., 2013a). Members of this group function primarily as transcriptional repressors, are able to complement the lam1 mutant phenotype (Lin et al., 2013a), and are capable of substituting for WUS function in maintaining vegetative and FMs (Dolzblasz et al., 2016). The WOX1 homologs M. truncatula STF and N. sylvestris LAM1 belong to this clade and function as master regulators of leaf blade outgrowth through a transcriptional repression mechanism in association with the co-repressor TOPLESS (Tadege et al., 2011a; Lin et al., 2013a,b; Zhang et al., 2014, 2019). The intermediate and ancient clade members have partial or no WUS box, and do not have transcriptional repression activity in Dual Luciferase assays. As a result, they are unable to rescue the lam1 mutant (Lin et al., 2013a) nor substitute for WUS function (Dolzblasz et al., 2016). Among the intermediate and ancient clades, AtWOX9 is unique in that it displays the strongest activation activity, and strongly enhances the lam1 mutant phenotype (Lin et al., 2013a), indicating that transcriptional activation activity modulated by AtWOX9 is antagonistic to LAM1 function. This is consistent with the observation that activation activity at the STF expression domain antagonizes STF function in blade outgrowth (Zhang et al., 2014, 2019).

The results presented here demonstrated that the WOX9 transcript is upregulated in three leaf blade mutants; stf in M. truncatula, lam1 in N. sylvestris and wox1 prs in Arabidopsis (Fig. 6j–l), indicating that WOX9 transcription may be suppressed by STF/LAM1/WOX1 in these species to allow cell proliferation. Several lines of evidence including a GR inducible system in the presence of DEX and CHX, Dual Luciferase assay, EMSA, and ChIP confirmed that STF/LAM1 directly binds to the MtWOX9-1 promoter to repress WOX9 transcription (Figs 7, 8), demonstrating that modern clade WOX1/STF/LAM1-mediated repression of intermediate clade WOX9 is required for proper leaf blade outgrowth in eudicots. In Arabidopsis, WOX9 functions upstream of WUS and is supposed to activate WUS to promote vegetative meristem growth (Wu et al., 2005), although it is unclear whether this activation is direct or indirect. WOX9 is activated by cytokinin signaling (Skylar et al., 2010), and type-B ARRs directly activate WUS (Meng et al., 2017; J. Wang et al., 2017; Zhang et al., 2017; Zubo et al., 2017; Xie et al., 2018), while WUS promotes cytokinin activity by repressing type-A ARRs (Leibfried et al., 2005). Thus, cytokinin signaling provides a potential connection between WOX9 and WUS in Arabidopsis, but whether WUS can directly affect WOX9 activity via negative or positive feedback loop is yet to be determined. Our work clearly demonstrates that in Medicago and woodland tobacco, the WUS clade member STF/LAM1 directly represses MtWOX9-1 or NsWOX9, but significant STF/LAM1 activation by WOX9 was not detected (Fig. 7c). Although MtWOX9-1 may not activate STF, it is likely to activate other targets in leaf development. WOX9 amino acid sequences from different species show a highly conserved acidic domain at the C-terminus (Fig. S1), which could mediate transcriptional activation. Thus, our current model is that STF represses the transcription of key targets at the adaxial–abaxial junction to promote cell proliferation, and these targets include cell differentiation factors abaxial MtWOX9-1 and adaxial MtAS2 (Zhang et al., 2014). WOX9, however, may positively regulate cell differentiation by directly activating targets independent of STF and/or by activating targets repressed by STF (Fig. 9) but in doing so may negatively regulate cell proliferation. The hypothesis that STF and WOX9 may oppositely regulate common targets providing homeostasis between cell proliferation and differentiation during leaf morphogenesis, explains why WOX9 ectopic expression enhances the stf/lam1 mutant phenotype. Since both STF and WUS promote cell proliferation via a transcriptional repression mechanism in the leaf primordium and SAM, respectively, our findings suggest that direct control of WOX9 activity by WUS may be required for SAM maintenance as well, uncovering a mechanistic framework for WOX modulated control of robust plant growth and developmental programs. It would be interesting to investigate if repression of the intermediate and ancient clade WOX transcriptional activators by modern clade WOX transcriptional repressors is a universal strategy exploited during the evolution of land plants and resulting in the complex morphological architecture of flowering plants.

Details are in the caption following the image
Schematic representation of hypothetical model for the regulation of cell proliferation-mediated leaf blade outgrowth by the interaction of STF/LAM1 and WOX9. STF/LAM1 directly represses WOX9, AS2, and other unidentified cell differentiation factors to promote leaf blade outgrowth through cell proliferation. WOX9, however, antagonizes cell proliferation by activating cell proliferation repressors and/or by promoting cell differentiation. The model proposes that STF/LAM1 and WOX9 may have a common target(s) repressed by STF/LAM1 and activated by WOX9 to achieve homeostasis in cell proliferation and differentiation required for proper blade development during leaf morphogenesis.


This work was supported by the National Science Foundation (NSF) grant IOS-1354422, Agriculture and Food Research Initiative Grant no. 2015-67014-22888 from the USDA National Institute of Food and Agriculture, and the Agricultural Experiment Station. Work in VET’s laboratory at St Petersburg State University, Russia, was supported by the Russian Science Foundation project no. 16-16-10011 and by grant from the Russian Foundation for Basic Research no. 20-016-00124.

    Author contributions

    TWW, HW, and MT designed the research. TWW, HW, DT, FZ, MB, VET, HA, YL, and NC performed the experiments. TWW, HW, FZ, YL, JC, RDA, and MT analyzed the data. TWW, RDA, and MT wrote the manuscript.