NUTCRACKER orchestrates cortical cell divisions and sustains stem cell niche integrity in the rice root meristem

Abstract

In Arabidopsis thaliana, BIRD nuclear factors, also known as the INDETERMINATE DOMAIN (IDD) protein family, regulate asymmetric cell division and tissue patterning in the root meristem. The BIRD protein JACKDAW (JKD) forms a regulatory complex with the GRAS transcription factors SHORT-ROOT (SHR) and SCARECROW (SCR) to maintain the stem cell niche and stabilize tissue boundaries. While BIRD protein functions are well characterized in Arabidopsis, their role in other plant species remains unclear. Here, we show that in rice, the JKD ortholog NUTCRACKER (NUC) restricts cell division in the root meristem. nuc knockout mutants display ectopic divisions in the ground tissue and vasculature, a loss of quiescence in the quiescent center, and premature differentiation of columella stem cells. NUC associates with SHR and SCR and regulates the expression of SCR and QUIESCENT CENTER-SPECIFIC HOMEOBOX (QHB/WOX5). NUC is expressed in the ground tissue and exodermis and complements the nuc mutant when driven by its native promoter. These findings reveal a role for the BIRD protein NUC in controlling cell division and maintaining the stem cell niche in the rice root meristem.

The Bird protein NUTCRACKER is a component of the core regulatory network that controls the stem cell niche and cell division the rice root meristem.

Introduction

During the development of a multicellular organism, cell fate determination and tissue specification are driven by regulatory networks involving morphogen gradients and transcription factors. In plants, as in animals, spatial and temporal dynamics of transcription factor networks control tissue patterning during development. In the root meristem, the precise organization of distinct cell files and the establishment of boundaries between tissue types are essential for proper root growth (Long et al., 2015). The root meristem is a reservoir of actively dividing cells that function to sustain longitudinal growth (Dolan et al., 1993; Bäurle and Laux, 2003; Jiang and Feldman, 2005), and mutants with dysfunctional meristems develop abnormal roots with impaired growth (Benfey et al., 1993). The meristem contains the stem cell niche from where all tissue types are produced through formative divisions (Clowes, 1954). The stem cells are located at the root tip and surround the organizing center, also termed the quiescent center (QC). Cells within the QC have a low mitotic activity and are required for stem cell maintenance (Clowes, 1978; Scheres et al., 1994; van den Berg et al., 1995; Ni et al., 2014). One of the best characterized examples of stem cells in the roots is the cortex–endodermis initial cell that gives rise to the cortex and the endodermis, which make up the ground tissue (GT) (Di Laurenzio et al., 1996). This initial divides anticlinally to generate a daughter cell, which then divides periclinally to form one layer of endodermis and one-to-multiple layers of cortex, depending on root type and plant species (Di Laurenzio et al., 1996; Ron et al., 2013; Ni et al., 2014). This process is controlled by the interplay of the transcription factors SCARECROW (SCR) and SHORTROOT (SHR) and the INDETERMINATE DOMAIN (IDD) proteins, also called BIRD proteins (Di Laurenzio et al., 1996; Helariutta et al., 2000; Welch et al., 2007; Long et al., 2015; Moreno-Risueno et al., 2015). In Arabidopsis, SHR is expressed in the vasculature and the SHR protein moves outward into the adjacent cell layer, the endodermis, where it promotes SCR expression. SCR together with the BIRD proteins JACKDAW (JKD) and BALDIBIS retain SHR in the nucleus and prevent further movement (Nakajima et al., 2001; Cui et al., 2007; Wu et al., 2014; Long et al., 2015).

The BIRD protein JKD is expressed in the QC and the GT where it associates with SCR and SHR (Welch et al., 2007; Long et al., 2015, 2017). SCR, SHR, and JKD activate the expression of CYCLIN D6 (CYCD6) to induce the periclinal division that leads to the separation of the cortex and the endodermis (Sozzani et al., 2010; Cruz-Ramírez et al., 2012; Long et al., 2017). In the endodermis, the SCR–SHR–JKD complex has a different conformation and induces endodermis-specific promoters (Long et al., 2015, 2017). Roots of jkd mutants display aberrant divisions originating from the cortex (Welch et al., 2007). This mutant fails to maintain a functional QC and the distal columella stem cells differentiate, while misexpression of JKD under the QC-specific promoter WUSCHEL-RELATED HOMEOBOX5 induces divisions in the QC (Welch et al., 2007; Long et al., 2017). Additionally, JKD has also been shown to control root epidermal patterning (Hassan et al., 2010).

While the function of the SHR/SCR pathway in specifying the QC, the endodermis, and the cortex seems to be conserved across plant species (Lim et al., 2000; Kamiya et al., 2003a; Cui et al., 2007; Henry et al., 2017; Ortiz-Ramírez et al., 2021), its regulatory mechanism to achieve the final tissue organization clearly differs between plant species. For example, in rice the cortex–endodermis initial gives rise to a single endodermal layer and multiple cortex layers (Ni et al., 2014), unlike the pattern seen in Arabidopsis. Whether BIRD proteins contribute similarly to root development in other species remains an open question.

Here, we describe the function of BIRD/IDD proteins in rice roots. We found that single-mutants of JKD, RAVEN (RVN), and MAGPIE (MGP) have a root meristem similar to the wild type. In nutcracker (nuc) mutants, however, the root meristem showed ectopic divisions in the GT and vasculature. In addition, the QC divided more frequently while the columella stem cells differentiated. Consistent with these findings, we observed a reduction of expression of QUIESCENT CENTER-SPECIFIC HOMEOBOX (OsQHB)/WOX5 in nuc mutants. We also found that the NUC protein is expressed in the QC, the GT initials, the inner cortex, and the endodermis, and it associates with SCR and SHR. Together, these results suggest that NUC functions in the root meristem in conjunction with OsSCR and OsSHR. Thus, our study highlights NUC as a major regulator for cell division in the rice root meristem and a key factor in controlling the stem cell niche.

Materials and methods

Plant material and growth conditions

Oryza sativa ssp. japonica cv. Nipponbare was used as the wild type (WT) for all experiments. For phenotypic analysis of CRISPR/Cas9 mutants, homozygous seedlings isolated by genotyping of T2 or T3 progenies were used. For expression analysis of the OsNUC reporter line in the WT background, seedlings of T2 were used, for the OsNUC reporter line in the nuc mutant background and OsQHB reporter lines in the WT, seedlings of T1 were used.

Rice seedlings were grown on plates in Percival E-36L2 plant growth chambers or to maturity in pots in a greenhouse under long-day conditions of 16/8 h light/dark, at 27/24 °C (300 μmol m^–2 s^–1). For growth on agar plates, rice seeds were sterilized by dehusking the seeds manually, washing them briefly in 100% ethanol (EtOH), and incubating them in 80% liquid bleach (Clorox) for 60 min. After washing them three times with sterile dH₂O, five seeds per plate were placed on square Petri dishes (120×120 mm) that contained 50 ml of half-strength Murashige and Skoog (MS) medium [1% (w/v) sucrose, 1% (v/v) MES (stock 50 g l^–1, pH 5.7), 0.22% (w/v) MS, 0.8% agar, pH 5.7). For staining with 5-ethynyl-2′-deoxyuridine (EdU), seedlings were grown on agar plates and transferred to 15 ml falcon tubes containing 2 μM EdU in dH₂O, so that the roots were completely submerged in the solution. For all experiments, the plants were grown for 5 d after germination.

Vibratome sections and staining

For examination of cross-sections, roots were fixed in SR2200 staining solution [0.1% (v/v) SR2200, 1% (v/v) DMSO, 0.05% (w/v) triton-X100, 5% (w/v) glycerol and 4% (w/v) para-formaldehyde in PBS buffer, pH 8.0) for 3–4 h under vacuum at room temperature, embedded in 5% (w/v) low-melting point agar, and cut into 80 μm sections using a VT1000S vibratome (Leica Microsystems). Sections were taken in the root meristem and shootwards at ∼1 cm distance from the root tip. The sections were stained with basic fuchsin as described previously (Ursache et al., 2018). For staining with calcofluor white, the roots were fixed in 3.7% formaldehyde in PBS overnight at 4 °C and subsequently cleared with ClearSee for 2 weeks at 4 °C in darkness (Kurihara et al., 2015). After being subjected to a vacuum for 1 h, they were then incubated overnight in 10% (v/v) calcofluor in ClearSee at room temperature in darkness. Before microscopy, the sections were again incubated in ClearSee solution for a few days at 4 °C in darkness. mPS-PI staining was performed as described previously (Kirschner et al., 2017). For the EdU staining, roots were fixed in SR2200 staining solution as described above and the staining was performed as described previously (Kirschner et al., 2017). For quantifying the percentage of the cells at the QC region, QC cells were defined as those that were shootwards of the root cap border and within the cortex–endodermis initial.

Microscopy

Vibratome sections and whole roots were imaged using an inverted confocal microscope (LSM 710, Zeiss). For the SR2200 and calcofluor white stains, a 405 nm laser was used with a detection range of 410–502 nm; for basic fuchsin, a 561 nm laser was used with a detection range of 600–650 nm; for EdU, a 561 nm laser was used with a detection range of 600–650 nm; and for mPS-PI, a 561 nm laser was used with a detection range of 566–718 nm. The samples for RNA fluorescence in situ hybridization (FISH) were imaged using a SP8 TCS STED confocal microscope (Leica); for Alexa647, a white-light laser set to 650 nm was used, with a detection range of 654–707 nm and gating between 1 ns and 6 ns. For bimolecular fluorescence complementation (BiFC) analysis, and for promoter and protein fusion YFP imaging, a Leica Stellaris Falcon microscope was used, with a 514 nm laser a detection range of 520–550 nm.

Cloning

For CRISPR/Cas9 editing, the expression of Cas9 was driven by OsUBIQUITIN. The pRGEB32 vector was used for callus transformations. The single-guide (sg)RNAs were designed to target the first or second exon of each locus (Table 1), and they were synthesized as oligonucleotides with BsaI overhangs, GGCA in the forward oligonucleotides and AAAC in the reverse (for primers see Supplementary Table S1). The pRGEB32 plasmid was digested with BsaI and the oligonucleotides were annealed and ligated in the BsaI-digested vector.

Table 1.

List of the rice BIRD/IDD genes targeted by CRISPR/Cas9

Gene	Identifier	Allele	GuideRNA	Mutation in coding sequence	Resulting amino acid sequence
MGP	LOC_Os01g09850	mgp	GAACATGTCGAACCTGACGT (exon 1)	1 bp (A) insertion behind G75	Frameshift behind AA25 and premature stop codon behind AA56 (before ZF1)
NUC	LOC_Os01g14010	nuc-1	AGGTTCGTGTGCGAGATATG (exon 2)	1 bp (T) insertion behind A193	Frameshift behind AA65 (in ZF1) and premature stop codon behind AA475
NUC	LOC_Os01g14010	nuc-2	AGGTTCGTGTGCGAGATATG (exon 2)	5 bp deletion behind C189	Frameshift behind AA63 AA65 (in ZF1) and premature stop codon behind AA474
JKD	LOC_Os02g31890	jkd	TTTGTCACCTAAGACACTCC (exon 2)	2 bp deletion behind C176	Frameshift behind AA59 (before ZF1) and premature stop codon behind AA83
STR	LOC_Os03g10140	str	CGTTTGCGAGATCTGCAACA (exon 2)	2 bp deletion behind T199	Premature stop codon at position of mutation, behind AA66 (in ZF1)
RVN	LOC_Os08g44050	rvn	AAGAAGAAGAGGAATCAGCC (exon 1)	1 bp (A) insertion behind T135	Frameshift behind AA45 (before ZF1) and premature stop codon behind AA52

ZF, zinc finger.

For the reporter lines, we used the Greengate cloning system as described previously (Lampropoulos et al., 2013) Using the primers listed in Supplementary Table S1. For complementation of rice nuc mutants, we amplified a 2932 bp fragment upstream of the start codon of LOC_Os01g14010 and cloned it into the pGGA entry vector. The coding sequence (CDS) from the start codon to the stop codon (excluding the stop codon) was amplified from genomic DNA and cloned into the pGGC entry vector. The CDS of SYFP2 (GenBank identifier DQ092361.1) was cloned into the pGGD entry vector. For the final construct, we used the pGGA entry vector with the OsNUC promoter, pGGB003 as the dummy N-tag sequence, pGGC with the OsNUC CDS, pGGD with YFP, pGGE009 as the terminator, and pGGF005 for hygromycin resistance, and combined them into the destination vector pGGZ003. For the transcriptional reporter line of OsQHB (LOC_Os01g63510), we synthesized the promoter sequence of OsQHB from 1950 bp upstream of the start codon in pUC57 (GenScript) and added the BsaI sites and appropriate flanking sequences. For the final construct, we used pUC57 with the OsQHB promoter, pGGB003 as the dummy N-tag sequence, pGGC012 for GFP-NLS, pGGD002 as the dummy C-tag sequence, pGGE009 as terminator, and pGGF005 for hygromycin resistance, and combined them into the destination vector pGGZ003.

For yeast two-hybrid (Y2H) and BiFC analyses, the coding sequences of OsNUC, OsSCR1, and OsSHR1 were synthesized in the Gateway entry vector pTwist (Twist Bioscience), and cloned via single Gateway LR reactions using Gateway LR Clonase II (Invitrogen) into the binary vectors p35S_GW_VYCE and p35S_GW_VYNE as described by Long et al. (2017) to generate C- and N-terminal fusions to the YFP fragments, respectively. AtMPK6 was amplified from Arabidopsis cDNA and cloned into the BiFC expression vector p35S_GW_VYNE.

Rice transformation

Rice was transformed as described previously for Japonica Type 1, using the strain EHA105 (Hiei and Komari, 2008; Butt et al., 2019). WT or nuc mutants without the CRISPR/Cas9 cassette were used for transformation. Briefly, callus induction was performed in 2N6 medium. Agrobacterium cells adjusted to OD₆₀₀=0.3 were incubated with the calli for 5 min. After co-cultivation in darkness at 25 °C for 3 d, rice calli were selected on media supplemented with hygromycin (50 mg l^–1) and Timentin (200 mg l^–1). After the screening stage, actively growing calli were sub-cultured onto a regenerative medium as described by Lee et al. (2002) for regeneration under continuous light. After 2–3 weeks, the transgenic seedlings were transferred to sterile plastic containers containing fresh rooting medium (Lee et al., 2002) and grown for 2–3 weeks before being transferred into soil.

Genotyping of mutant plants

DNA was extracted from leaf samples that were immediately frozen in liquid nitrogen after collection. The frozen material was homogenized into a fine powder for DNA extraction as described by Richards et al. (1994). PCR was conducted using gene-specific primers (Supplementary Table S1). Purified PCR products were cloned using a CloneJET PCR Cloning Kit (K1231, ThermoFisher Scientific). Recombinant clones were selected, and DNA was purified and subjected to Sanger sequencing.

Bimolecular fluorescence complementation assays

The DNA constructs were transformed into A. tumefaciens strain GV3101 using a Bio-Rad Gene Pulser Xcell Electroporation System. Positive clones were inoculated into LB medium supplemented with appropriate selective antibiotics and grown overnight at 28 °C. The growing culture was then pelleted at 1600 g for 10 min, and resuspended with 10 mM MgCl2, 200 μM acetosyringone, and 10 mM MES (pH 5.6), and adjusted to a final OD₆₀₀=1.5. The p19 helper plasmid was adjusted to OD₆₀₀=1.00. Then, the constructs containing the respective CDS with pBIFC2 and pBIFC3 were mixed with p19 in a 1:1:1 ratio and used for infiltration into leaves of Nicotiana benthamiana. Expression was detected 3 d after infiltration.

Yeast two hybrid assays

The Y2H assays were performed using a ProQuest Two-Hybrid System (Invitrogen). The entry clones containing the CDSs of OsNUC, OsSCR, and OsSHR were each cloned into both the pDEST32 BD and pDEST22 AD vectors. Autoactivation of yeast containing the bait was tested using selective medium lacking His and Leu and supplied with 2, 5, 10, 25, 50, 75, or 100 mM 3-amino-12,4-triazol (3AT, Sigma). The pDEST32 BD and pDEST22 AD vectors containing the CDSs were transformed into the yeast strains Pj694α and Pj694a, respectively. Mating and selection for interactions were performed as described in by James et al. (1996).

Hybridization chain-reaction RNA fluorescence in situ hybridization

For hybridization chain-reaction RNA fluorescence in situ hybridization (HCR RNA FISH), roots were fixed overnight in 4% paraformaldehyde, 0.1% Tween-20, and 0.1% Triton-X-100 in PBS at 4 °C, then they were dehydrated in a 30%, 40%, 50%, 60%, 70%, 85% EtOH series in RNase-free water for 1 h each under vacuum. After incubation overnight in 95% ethanol and 0.1% eosin in RNase-free water at 4 °C, a Leica, TP1020 Tissue Processor was used for paraffin embedding with the following protocol: two times 1 h 100% histoclear; 1 h 75% EtOH/25% histoclear; 1 h 50% EtOH/50% histoclear; 1 h 25% EtOH/75% histoclear; two times 1 h 100% histoclear; 1 h in 100% histoclear supplemented with 25% volume of paraffin pellets; two times 1 h melted paraffin at 60 °C. Sections of 5 µm were then cut with a microtome. The HCR™ RNA-FISH protocol for FFPE tissue sections of Molecular Instruments was then used for hybridization and detection of the probes, with digestion using 0.125 mg ml^–1 pronase in PBST for 10 min at 37 °C. Probes were purchased from Molecular Instruments (https://www.molecularinstruments.com/), with the following lot numbers and adaptors: OsNUC, PRR102, B1; OsSCR1, PRR103, B2; OsSCR1 and OsSCR2, RTC802, B2; OsSHR1, PRR105, B4; OsQHB/WOX5, PRR107, B1; and cortex marker, RTD177, B1. B1-647, B2-647, or B4-647 were used as amplifiers and fluorophores. Negative controls were performed without probes, but with amplifiers and fluorophores.

Phylogenetic analysis

To create the phylogenetic tree, we used At5g03150 (AtIDD10/AtJKD) as the query in a BLASTp search of the Arabidopsis proteome in the Phytozome v12 database (https://phytozome-next.jgi.doe.gov/), with an e-value <1.8×10^–33. We used the Arabidopsis IDD whole-protein sequences together with the rice IDD whole-protein sequences (Colasanti et al., 2006). MRBAYES (Huelsenbeck and Ronquist, 2001) was used for Bayesian phylogenetic analysis at NGPhylogeny.fr (Lemoine et al., 2019). Four independent chains were run for 100 000 generations with a burn-in fraction of 0.25 and a sampling frequency of 500 generations.

Accession numbers and gene identifiers

Rice: MGP (LOC_Os01g09850); NUC (LOC_Os01g14010); JKD (LOC_Os02g31890); STR (LOC_Os03g10140); RVN (LOC_Os08g44050); IME (LOC_Os09g38340); SCR1 (LOC_Os11g03110); SCR2 (LOC_Os12g02870); SHR1 (LOC_Os07g39820); SHR2 (LOC_Os03g31880); QHB (LOC_Os01g63510); and Cortex marker (LOC_Os04g46810). Arabidopsis: MPK6 (AT2G43790).

Statistical analysis

Each experiment was repeated at least three times. Statistical analysis was performed using one-way ANOVA followed by Tukey's HSD test.

Results

Generation of CRISPR/Cas9 bird/idd mutants in rice

While the role of the BIRD/IDD genes for QC maintenance and GT development in the Arabidopsis roots is well-established (Welch et al., 2007; Long et al., 2015; Moreno-Risueno et al., 2015), their function in monocot root development is currently unknown. Based on their protein phylogeny in rice and Arabidopsis (Supplementary Fig. S1), we targeted different rice BIRD/IDD-encoding genes by CRISPR/Cas9 and named them according to their orthologues in Arabidopsis.

We isolated homozygous rice lines for non-synonymous mutations in the CDS for LOC_Os01g09850 [MAGPIE (MGP)/IDD2)], LOC_Os01g14010 [NUTCRACKER (NUC)/IDD8)], LOC_Os02g31890 (JKD/IDD7), LOC_Os03g10140 [STARLING (STR)/IDD1)], and LOC_Os08g44050 [RAVEN (RVN)/IDD6)] (Supplementary Fig. S1; Table 1).

BIRD/IDD proteins have a conserved ID domain with a putative nuclear-localization sequence at the N-terminal border and four distinct zinc-finger motifs (Colasanti et al., 2006). All mutations in the mutant alleles led either to frameshifts or premature stop codons before or at the beginning of the first zinc finger (Table 1), suggesting that the resulting protein is non-functional because it lacks all zinc-finger domains.

To assess the effects of the mutations on root growth and patterning, we first examined root lengths in 5-day-old seedlings (Fig. 1A). We found that the length of primary and crown roots of the rice mgp, nuc, jkd, and rvn single-mutants were comparable to the WT, and only str deviated by having shorter primary roots (Fig. 1A–C). In Arabidopsis, the root phenotypes of the bird mutant are mostly noticeable at the cellular level, and therefore we examined the anatomical structures of the root meristems. While mgp, jkd, str, and rvn had a well-organized meristem similar to the WT (Fig. 1D), the meristems of the two nuc mutant alleles (nuc-1 and nuc-2) showed misoriented cell divisions at the stem cell region and within the meristem when compared with the WT roots (Fig. 1E).

Fig. 1.

Rice nuc single-mutants show aberrant divisions in the root meristem while other bird mutants are similar to the wild type. (A) Representative images of wild type (WT) and bird/idd single-mutants grown for 5 d after germination; scale bar is 1 cm. (B) Primary root length (n>20) and (C) crown root length (n=20) of the WT and the mutants normalized to the WT root length, the mean values of which were set as 1. Significant differences between the WT and individual mutants are indicated and were determined using one-way ANOVA followed by Tukey’s HSD tests: *P<0.05, **P<0.01. (D) Representative images of crown root meristems of the WT and the mutants showing the stem cell niche. Cell walls are stained with calcofluor white. The arrowheads indicate aberrant periclinal cells whilst arrowheads indicate disorganized quiescent centers (QCs). The scale bar is 100 μm. (E) Higher magnification images of the WT and the nuc-1 and nuc-2 mutants, showing aberrant divisions around the QC and the ground tissue.

NUC is required for stem cell niche integrity in the rice root meristem

In the rice root meristem, the stem cell niche contains QC cells that divide infrequently and are surrounded by the stem cells. The columella stem cells are located below the QC and differentiate to generate the differentiated columella layers that accumulate starch granules (Coudert et al., 2010; Ni et al., 2014). In nuc-1, we observed aberrant divisions in cells at the position of the QC and accumulation of starch granules in the QC and columella stem cells (Figs 1E, 2A–G). However, we did not observe changes in the meristem length when compared with the WT (Fig. 2E). These observed divisions prompted us to quantify the cell division rate at the QC position in primary and crown roots over 24 h using EdU staining. We found that the QC cells rarely showed an EdU signal in the WT, indicating that they divided infrequently, while in the nuc-1 mutant a higher percentage of cells at the QC position showed an EdU signal, suggesting that division was more frequent (Fig. 2H–J).

Fig. 2.

NUC is required for rice stem cell maintenance and for restricting cell divisions in the quiescent center. (A–D) Representative images of root meristem phenotypes of the wild type (WT) and the nuc-1 single-mutant. MR, main (primary) root; CR, crown root. Cell walls and starch granules are visualized by modified pseudo-Schiff-propidium iodide staining. Scale bars are 100 μm. (E) Root meristem lengths of the WT and the nuc-1 and nuc-2 mutants, normalized to the WT, the mean value of which was set as 1. Black corresponds to main (primary) roots and gray corresponds to crown roots. The numbers of replicates were: WT PR, n=20; WT CR, n=51; nuc-1 PR, n=18; nuc-1 CR, n=34; nuc-2 PR, n=10; nuc-2 CR, n=16. The significant difference between the WT and nuc-1 crown roots was determined using one-way ANOVA followed by Tukey’s HSD test: **P<0.01.(F, G) Magnifications of the crown roots in (B, D) showing the quiescent center (QC). region. (H, I) Cell divisions in the stem cell niche of crown roots, visualized by 5-ethynyl-2'-deoxyuridine staining (purple); cell walls are stained with calcofluor white. (H', I') show calcofluor staining only. Arrowheads indicate the position of the QC. (J) Percentage of cells with EdU-stained nuclei at the position of the QC in the WT and the nuc-1 and nuc-2 mutants. (K–M) RNA Fluorescent in situ hybridization in crown roots of (K) the WT and (L) the nuc-1 mutant using a probe targeting QHB; (K, L) show overlays of transmitted light and the fluorescent signal whilst (K', L') show the fluorescent signal only. (M) Control without the probe.

To understand the function of NUC in the rice QC, we used HCR RNA FISH with probes marking the QC and the endodermis. The rice orthologue QHB/WOX5 has been shown to be expressed in the QC and the metaxylem region (Kamiya et al., 2003b). In WT roots we detected its expression in the metaxylem and in a region including the QC and the surrounding stem cells (Fig. 2K, K'); this expression was similar to the QHB promoter pOsQHB::NLS-GFP activity in the roots (Supplementary Fig. S2) and was consistent with previous reports (Kamiya et al., 2003b). In the nuc-1 mutant, QHB/WOX5 RNA accumulated in the metaxylem, but we could not detect fluorescence in the region around the QC (Fig. 2L, L'), suggesting that NUC controls the activity of QHB. We did not detect fluorescence in the tissues of the negative controls with no probes (Fig. 2M).

NUC regulates cell division and tissue boundaries in the ground tissue of rice

In the WT, the structure of the root follows a circular arrangement In which the barrel-shaped endodermis encloses the vasculature and is surrounded by multiple cortex cell files (Fig. 3A). The GT in the WT develops from consecutive periclinal cell divisions of the cortex–endodermis initial and forms well-structured layers consisting of one layer of endodermis and multiple layers of cortex (Rebouillat et al., 2009; Ni et al., 2014). In both the nuc alleles, however, we observed ectopic and misoriented divisions in both tissue layers (Fig. 1E; Supplementary Fig. S3). These additional divisions resembled the jkd phenotype in Arabidopsis, where ectopic periclinal divisions in the cortex generate additional layers in the GT and vasculature (Welch et al., 2007). Hence, we evaluated the radial pattern in root cross-sections of the nuc mutants. In the WT, cell divisions within each layer were patterned regularly (Fig. 3A, A', C, C'), whereas cells within the layers in nuc-1 showed more divisions (Fig. 3B, B', D, D'). Quantification showed that the nuc mutants had more cells in the vasculature in both the meristem and the differentiation zone (Supplementary Fig. S3).

Fig. 3.

The rice nuc mutant shows irregular cell divisions in the inner ground tissue and more cells in the vasculature. (A, B) Cross-section of crown root meristems of (A) the wild type (WT) and (B) the nuc-1 mutant. (A', B') showed magnified versions of (A, B). Cell walls are stained with SR2200. Arrowheads indicate irregular periclinal cell divisions in the cortex. The images are representative of n=9 replicates for the WT and n=12 for nuc-1. (C, D) Cross-sections of crown roots ∼1 cm above the root meristem in (C) the WT and (D) nuc-1. (C', D') Magnifications of the area of inner ground tissue indicated in (C, D). Cell walls are stained with SR2200 (purple) and lignin is stained with basic fuchsin (green). Arrowheads indicate irregular periclinal cell divisions at the endodermis position. The images are representative of n=10 replicates for the WT and n=21 for nuc-1. (E, F) RNA fluorescent in situ hybridization (FISH) in crown roots of (E) the WT and (F) nuc-1 (F) crown roots using a probe targeting SHR1. (E', F') Magnifications of the inner ground tissue indicated in (E, F). The fluorescent signal and is overlayed with the transmitted light signal. The images are representative of at least n=10 replicates. (G, H) RNA FISH in crown roots of (G) the WT and (H) nuc-1 using a probe targeting SCR1 and SCR2. (G', H') Magnifications of the inner ground tissue indicated in (G, H). The fluorescent signal and is overlayed with the transmitted light signal. Arrowheads indicate SCR expression in the inner cortex. The images are representative of at least n=10 replicates. All scale bars are 100 μm.

To assess the identity of cells resulting from the additional divisions in the nuc mutants, we examined the expression of SHR and SCR homologues, and of a cortex marker in the meristem by HCR RNA FISH using probes targeting the rice orthologues. SHR and SCR expression is conserved among plant species (Di Laurenzio et al., 1996; Helariutta et al., 2000; Lim et al., 2000; Kamiya et al., 2003a; Cui et al., 2007). Consistent with these reports, we found that SHR1 (LOC_Os07g39820) was expressed in the vasculature, including the pericycle (Fig. 3E, E'); however, we did not detect a difference in its expression in the vasculature of the nuc-1 mutant (Fig. 3F, F').

In rice, there are two SCR homologues, SCR1 (LOC_Os11g03110) and SCR2 (LOC_Os12g02870) (Cui et al., 2007). Because of their high sequence similarity at the cDNA level, we designed RNA probes targeting both of them. In the WT, we detected a high signal in the endodermis as well as in patches in the innermost cortex cell file adjacent to the endodermis, and a lower signal in the outer cortex (Fig. 3G, G'). In nuc-1, a few cells adjacent to the endodermis had a higher level of SCR expression when compared with the WT (Fig. 3H, H'), suggesting that these cells might have undergone asymmetric cell division.

Next, we examined LOC_Os04g46810, which was identified as a marker of the cortex in a single-cell RNA-seq dataset (Liu et al., 2021). In the WT, we observed expression in the two or three middle cortex cell files, but not in the inner cortex, nor in the outer two cortex cell files (Supplementary Fig. S4A). We could not detect any changes in expression in the nuc-1 mutant (Supplementary Fig. S4B).

NUC is expressed in the ground tissue and associates with SHR1 and SCR1

To evaluate its expression pattern in the rice root meristem, we first used HCR RNA FISH with probes targeting NUC, and observed signals in the cortex–endodermis initials, the endodermis, the inner cortex, and the exodermis (Fig. 4A–F). We then created a translational reporter line using the native OsNUC promoter to generate a OsNUCpro::OsNUC-YFP construct and transformed it into the WT and the nuc-1 mutant. We observed a high YFP signal in the GT initials, in the first endodermal cells close to the QC, and in the inner cortex cell files, and the signal decreased towards the shootward side of the meristem (Fig. 4G–L). We also detected the OsNUC-YFP signal in the QC (Fig. 5). Transformation with OsNUCpro::OsNUC-YFP restored the phenotypes of the vasculature, GT, QC, and stem cells of the nuc-1 mutant (Fig. 5A–I; Supplementary Figs S5, S6), suggesting that the protein was functional and that the nuc mutant phenotypes were caused by the mutation in the NUC gene in the CRISPR/Cas9 lines and not by off-target mutations.

Fig. 4.

The rice NUC protein is expressed in the ground tissue initials, the endodermis, the inner cortex, and the exodermis. (A, B) RNA fluorescent in situ hybridization (FISH) in longitudinal sections of crown roots of the wild type (WT) using a probe targeting NUC; (A) shows the fluorescent signal (red) whilst (B) shows an overlay of transmitted light and the fluorescent signal. (C–F) RNA FISH of cross-sections of crown roots in the WT using a probe targeting NUC; again, (D, F) show overlays of transmitted light and the fluorescent signal. (C, D) Cross-sections were taken within the meristem. (E, F) Cross-sections were taken at a position above the quiescent center (QC). (G, H) OsNUCpro::OsNUC-YFP Expression of the OsNUCpro::OsNUC-YFP construct in (G) primary and (H) crown roots in transformed plants. YFP expression is in yellow; cell walls were stained with calcofluor white. (I–K) The cross-sections in (J, K) were taken at the positions indicated in (I). (L) Quantification of the mean gray value (MGV) of the YFP channel in crown roots in the endodermis (black) and inner cortex (gray) cell files measured in cells distant to the QC.; standard deviation is shown in Dark gray and light gray denote the SD for the endodermis and inner cortex, respectively. All images and data are representative of at least n=10 replicates. All scale bars are 100 μm.

Fig. 5.

Transformation with the OsNUCpro::OsNUC-YFP complements the rice nuc-1 phenotype, and OsNUC interacts with OsSHR1 and OsSCR1 in vivo. (A–D) Longitudinal sections of the root meristem of (A) the wild type (WT), (B) the nuc-1 mutant, (C) the WT expressing OsNUCpro::OsNUC-YFP, and nuc-1 expressing OsNUCpro::OsNUC-YFP. E, E' and F, F') are magnified images from (C) and (D), respectively. G–I) Cross-sections of roots of (G) the WT, (H) the nuc-1 mutant, and (I) nuc-1 complemented with OsNUCpro::OsNUC-YFP. The squares indicate regions of aberrant divisions in the nuc-1 mutant compared with the WT and the rescued lines. All scale bars are 50 μm. Images are representative of n=10 replicates for the WT, n=12 for nuc-1, and n=15 for the complemented lines. (J) Bifluorescence complementation in Nicothiana benthamiana leaves. The coding sequences were fused to the N-terminal or to C-terminal part of YFP (nYFP and cYFP, respectively) as indicated AtMPK6-nYFP and OsNUC-cYFP were used as negative controls. The pairs of images show the YFP fluorescence only and the YFP fluorescence and transmitted light. Scale bars are 50 μm. At least 20 nuclei were imaged per combination.

In Arabidopsis, JKD associates with SCR and SHR to constrain SHR protein movement through nuclear retention, and transcriptionally regulates downstream targets such as SCR and the cell-cycle regulator CYCLIN D6 (Long et al., 2015, 2017). Therefore, we examined the protein interactions among OsNUC, OsSHR1, and OsSCR1 using BiFC assays in leaves of N. benthamiana and by Y2H assays, and both indicated OsNUC interacts with OsSCR1 and OsSHR1 (Fig. 5J; Supplementary Fig. S7). These findings suggested that the interaction of BIRD proteins with SCR and SHR is conserved between Arabidopsis and rice.

Discussion

In Arabidopsis, the BIRD proteins play an important role in root tissue patterning. Through regulating SHR movement together with SCR, they control asymmetric cell division in the GT and tissue boundary delimitation (Welch et al., 2007; Long et al., 2015; Moreno-Risueno et al., 2015). The BIRD protein JKD associates with SCR and SHR to specify the QC and endodermis through regulating target genes in a cell-type-specific manner (Long et al., 2017) while IMPERIAL EAGLE (IME), MGP, and NUC are mainly involved in regulating formative divisions in the GT in conjunction with SCR (Long et al., 2015; Moreno-Risueno et al., 2015).

While the functions of BIRDs in controlling radial patterning in Arabidopsis are well described, their role in root development in other plant species remains to be characterized. Here, we studied the function of the BIRD proteins NUC, MGP, JKD, STR, and RVN in rice, designated based on their protein sequence similarity to the Arabidopsis BIRDs. Analysis of CRISPR/Cas9-generated mutants of the BIRD genes identified NUC as a key regulator of GT patterning in rice roots.

Similar to NUC expression in Arabidopsis, OsNUC mRNA and protein were detected in the endodermis and the inner cortex of rice (Figs 4, 6). Additionally, the rice NUC protein, but not its mRNA, was detected in the QC. This absence in the QC is most likely due to its low expression levels, making it undetectable by RNA in situ hybridization. An alternative explanation is that the OsNUC protein might move into the QC; however, this is unlikely, as BIRD proteins have not been shown to move between tissue layers. Testing this hypothesis would be challenging and would require misexpression of OsNUC under a GT-specific promoter or the induction of callose deposition in specific tissues to constrain protein movement. Furthermore, while NUC mRNA was detected in the rice exodermis, the corresponding protein was not. This discrepancy could be due to protein degradation in the exodermis or the possibility that the FISH probe was cross-reacting with another BIRD family member expressed in that tissue.

Fig. 6.

Schematic representations of the rice wild type (WT) and nuc mutant phenotypes and a model summarizing the function of NUC in the regulation of quiescent center (QC) division, stem cell maintenance, and cell division in the root meristem. (A) Schematic diagram of the WT rice root meristem structure with the QC surrounded by stem cells. The two layers of columella stem cells and lateral root cap stem cells can be recognized as described by Coudert et al. (2010) and Wang et al. (2014). Cells in gray are epidermis, cortex, and vascular stem cells and their daughters based on tracing their tissue lineages. (B) The nuc root meristem showing aberrant cell divisions and differentiation of stem cells based on the presence of starch granules. The diagrams in (A, B) are based on root confocal microscopy images. (C) Model illustrating the role of NUC in controlling cell division in the quiescent center (QC) and in promoting stem cell maintenance. Mobile SHR associates with NUC to possibly activate QHB, thereby maintaining stem cells and restricting cell division in the QC. (D) Model illustrating the potential mode of action of NUC in controlling cell division in the ground tissue. SHR moves from the vasculature to the ground tissue, where it binds to NUC to promote endodermal fate and restrict cell division in the cortex. Question marks indicate a suggested modes of action of NUC together with SCR and SHR in the QC and the ground tissue based on their protein interactions and transcriptional regulation of QHB.

Unlike Arabidopsis, where mutation of JKD induces loss of the QC function and ectopic division in the cortex that leads to additional layers in the meristem (Welch et al., 2007), we found that OsJKD was not expressed in roots in rice (Supplementary Fig. S8) and the jkd mutant did not display any changes in root phenotype (Fig. 1). However, a mutation in OsNUC led to a loss of the QC quiescence and differentiation of the columella stem cells. These phenotypes were similar to those observed in the Arabidopsis jkd mutant (Welch et al., 2007) and suggest that in rice it is the BIRD protein NUC that functions to regulate QC and GT patterning. Consistent with the observed phenotypes, we could not detect expression of the WOX5 orthologue OsQHB/WOX in the QC of the nuc mutant (Fig. 2), suggesting that NUC function is necessary for QC maintenance (Fig. 6). However, OsNUC does not seem to be required for root growth and development as we did not observe differences in root length and meristem size when compared with the WT. Our findings indicate that, similarly to WOX5 in Arabidopsis, OsQHB expression in the QC is required for restricting divisions and for columella stem cell maintenance but not for root growth. Reduction of OsQHB expression in the QC and vasculature has been shown to lead to reduction of the meristem size (Chu et al., 2013). Given that expression of QHB could still be observed in the vasculature of the nuc mutants, it is plausible that the function of QHB in regulating the meristem size is independent from NUC, and therefore the meristem length in nuc is not reduced.

The nuc mutant also formed patches of cortex with more layers (Fig. 3G, G'), suggesting that OsNUC is required to restrict periclinal cell divisions similar to the function of JKD in Arabidopsis. In addition to the ectopic divisions in the meristematic cortex, we found ectopic periclinal cell divisions at the position of the endodermis and the innermost cortex layer in the meristem in the more differentiated part of the root (Fig. 1). These findings suggest that OsNUC is required to restrict cell division in the GT, possibly by restricting OsSHR movement.

SHR has been suggested to be more mobile in grasses than in Arabidopsis, and its movement is not restricted to the endodermis (Wu et al., 2014; Fig. 6). Outward movement of SHR could facilitate the formation of multiple cortex cell files, as opposed to one cortex file in Arabidopsis. Indeed, expression of rice and Brachypodium SHR in Arabidopsis induces the formation of more cortex cell files, facilitated by SHR movement to the outer cell files (Wu et al., 2014).

OsSHR1, OsSCR1, and OsNUC form heterodimers and their expression domains are similar to their orthologues in Arabidopsis (Cui et al., 2007). It is thus possible that OsNUC might contribute to SHR sequestration in the nucleus through physical association with OsSCR to limit its movement to the outer layers, as in Arabidopsis (Nakajima et al., 2001; Cui et al., 2007; Welch et al., 2007; Long et al., 2015, 2017; Fig. 6).

Our interaction and protein localization results together with the observed mutant phenotypes suggest a model in which OsNUC interacts with OsSCR1 and OsSHR1 in the inner GT (endodermis and inner cortex) to prevent the movement of SHR1 and to convey the identity of the ‘inner’ cortex. Given the role of the BIRD proteins JKD and BALDIBIS that has been described in regulating SHR movement, it remains to be established whether the ectopic divisions observed in the rice nuc mutant are mediated by excessive SHR movement beyond the endodermis. Testing this hypothesis would require a functional SHR translational reporter line, which is currently not available as the promoter sequences of OsSHR1 and OsSHR2 are high in GC- and A-rich regions and repeats, which prevented the creation of such a reporter line in this study.

Although NUC clearly plays a role in QC maintenance and GT development in rice, the functions of other BIRD genes in root growth and tissue patterning remain less defined. Single-cell RNA-seq analysis of rice roots has revealed that MGP, STR, and RVN are also expressed in the meristem, similar to NUC (Supplementary Fig. S8; Wang et al., 2021). However, since single-mutants of these genes did not display noticeable root meristem phenotypes (Fig. 1), it is possible that they function redundantly in meristem development. Alternatively, the role of BIRD proteins in stem cell maintenance and radial patterning might not be fully conserved in rice, with a different clade of BIRD/IDD proteins fulfilling this function.

Another possibility is that BIRD proteins in rice are more specialized for environmental responses. In maize, for example, the first identified IDD gene, STARLING/ID1, regulates flowering time (Colasanti et al., 2006) sucrose and starch metabolism (Coneva et al., 2012), and it influences plant height and tillering under short-day conditions in rice (Hu et al., 2013). BIRDs/IDDs in rice have also been implicated in responses to abiotic stress, including drought tolerance (Zhang et al., 2020), ammonium uptake, and nitrogen metabolism (Xuan et al., 2013) as well as cell wall formation (Huang et al., 2018). Furthermore, BIRD/IDD genes contribute to pathogen resistance in both Arabidopsis and rice (Völz et al., 2019; Sun et al., 2020).

Taken together, our findings identify NUC as a central regulator of GT patterning and QC maintenance in rice roots, performing functions analogous to Arabidopsis JKD but through a different regulatory network. While core components of the SHR–SCR–BIRD module are conserved, our results suggest that functional divergence has occurred among BIRD proteins in rice, with OsNUC assuming a primary developmental role. The lack of phenotypes that we observed in other BIRD single-mutants hints at redundancy or specialization for environmental responses, underscoring the evolutionary plasticity of the BIRD gene family across monocots and dicots.

Abbreviations

BiFC

bimolecular fluorescence complementation

FISH

fluorescence in situ hybridization

GT

ground tissue

IDD

INDETERMINATE DOMAIN

IME

IMPERIAL EAGLE

JKD

JACKDAW

MGP

MAGPIE

NUC

NUTCRACKER

QC

quiescent center

QHB

QUIESCENT CENTER-SPECIFIC HOMEOBOX

RVN

RAVEN

SCR

SCARECROW

SHR

SHORT-ROOT

WT

wild type

Y2H

yeast two hybrid

Supplementary data

The following supplementary data are available at JXB online.

Fig. S1. Phylogenetic relationships of the BIRD/IDD proteins in rice and Arabidopsis.

Fig. S2. Visualization of expression of rice QHB/WOX5 in the WT root meristem.

Fig. S3. Effects of nuc mutations on periclinal cell division in the rice root meristem.

Fig. S4. Expression of a cortex marker in the WT and nuc mutant.

Fig. S5. Expression of OsNUC can complement the nuc-1 root meristem phenotype.

Fig. S6. Expression of OsNUC rescues cell divisions in the stele and the cortex in the nuc-1 mutant.

Fig. S7. Yeast two-hybrid assays showing the interaction of OsNUC with OsSCR1 and OsSHR1.

Fig. S8. Transcript levels of OsBIRD genes in the root meristem, derived from Wang et al. (2021).

Table S1. Primers used for cloning and genotyping.

Funding

This research was supported by the KAUST research baseline fund (BAS/1/1081-01-01) to IB, and by the KAUST Competitive Research Grants (CRG 9) URF/1/4381-01-01 and CRG8 URF/1/4081-01-01 to IB.

Data availability

The data that support the findings of this study are available from the corresponding author, Ikram Blilou, upon request.