An SNW/SKI-INTERACTING PROTEIN influences endoreduplication and cell growth in Arabidopsis

Abstract

Endoreduplication plays an important role in cell growth and differentiation, but the mechanisms regulating endoreduplication are still elusive. We have previously reported that UBIQUITIN-SPECIFIC PROTEASE14 (UBP14) encoded by DA3 interacts with ULTRAVIOLETB INSENSITIVE4 (UVI4) to influence endoreduplication and cell growth in Arabidopsis (Arabidopsis thaliana). The da3-1 mutant possesses larger cotyledons and flowers with higher ploidy levels than the wild-type. Here, we identify the suppressor of da3-1 (SUPPRESSOR OF da3-1 3; SUD3), which encodes SNW/SKI-INTERACTING PROTEIN (SKIP). Biochemical studies demonstrate that SUD3 physically interacts with UBP14/DA3 and UVI4 in vivo and in vitro. Genetic analyses support that SUD3 acts in a common pathway with UBP14/DA3 and UVI4 to control endoreduplication. Our findings reveal an important genetic and molecular mechanism by which SKIP/SUD3 associates with UBP14/DA3 and UVI4 to modulate endoreduplication.

SNW/SKI-INTERACTING PROTEIN interacts with UBIQUITIN-SPECIFIC PROTEASE14 and ULTRAVIOLET-B INSENSITIVE4 to influence endoreduplication and cell growth in Arabidopsis.

Introduction

The mass of an organ is defined by the number and the size of its cells. The cell number depends on the mitotic activity of organ meristems, while plant cell growth can be achieved by cell expansion and endoreduplication (Barow, 2006; Sablowski and Carnier Dornelas, 2014). Endoreduplication (also known as endocycle) is an alternative version of the classical cell cycle during which cells replicate their nuclear DNA without subsequently dividing, thereby increasing their ploidy level. This process occurs in a variety of cell types and contributes to normal growth and development of plant cells as well as plant responses to biotic and abiotic stresses (Lang and Schnittger, 2020).

Many molecular players and mechanisms that take part in endoreduplication are conserved in all eukaryotes, particularly the role of cyclin-dependent kinase (CDK) activity in controlling transitions from one phase to another of the mitotic cell cycle (Harashima et al., 2013; Dante et al., 2014). A-type CDKs in Arabidopsis (Arabidopsis thaliana) comprise functional homologs of the yeast (Schizosaccharomyces pombe) p34^cdc2, and target Rb homolog RETINOBLASTOMA-RELATED1/RBR1 to regulate cell cycle (Nowack et al., 2012). B-type CDKs are unique to plants and contain a divergent cyclin binding motif (Boudolf et al., 2004). Plants overexpressing a dominant negative allele of CYCLIN-DEPENDENT KINASE B1;1 (CDKB1;1) exhibit the increased ploidy level (Boudolf et al., 2004). CDKB1;1 associates with A2-type cyclin CYCA2;3 to suppress endoreduplication onset (Boudolf et al., 2009). The anaphase-promoting complex/cyclosome (APC/C) is a multisubunit E3 ubiquitin ligase complex targeting cell cycle factors to regulate cell cycle progression from metaphase to S phase (Heyman and De Veylder, 2012). CELL CYCLE SWITCH PROTEIN 52 A1/FIZZY-RELATED 2 (CCS52A1/FZR2), an activator of APC/C, downregulates the protein abundance of CDKB1;1 and CYCA2;3 to control the transition of mitotic cycles to endoreduplication cycles (Boudolf et al., 2009). Loss-of-function alleles of CCS52A1/FZR2 show reduced endoreduplication and reduced cell expansion in trichomes (Larson-Rabin et al., 2008; Li et al., 2009). ULTRAVIOLET-B INSENSITIVE4 (UVI4), a functional homolog of Emi1, promotes the accumulation of CYCA2;3 by binding to the CCS52A1 activator subunits to inactivate APC/C (Heyman et al., 2011). The uvi4 mutant plants display an increased DNA ploidy level (Heyman et al., 2011).

UBIQUITIN-SPECIFIC PROTEASE 14 (UBP14) encoded by DA3 has been previously described to function with UVI4 to inhibit the APC/C and modulate endoreduplication and cell growth in Arabidopsis (Xu et al., 2016). The da3-1 mutant produces large cotyledons and flowers. However, the da3-1 plants exhibit delayed growth, reduced leaf numbers, and decreased branches, resulting in a small plant phenotype (Xu et al., 2016). Here we report the suppressor of da3-1 3 (sud3-1) mutant suppresses the endoreduplication and plant morphology of da3-1. SUD3 encodes SNW/SKI-INTERACTING PROTEIN (SKIP). The antisense transgenic lines of SKIP, atskip, exhibited reduced inflorescence stems, smaller rosette leaves, and reduced seed production (Lim et al., 2010). The skip-1 mutation lengthened the period of the circadian clock by impairing the alternative splicing of PSEUDO-RESPONSE REGULATOR7 and PSEUDO-RESPONSE REGULATOR9 in Arabidopsis (Wang et al., 2012). A T-DNA insertion allele of SKIP, skip-2, was severely dwarfed and infertile when homozygous (Wang et al., 2012). Overexpression of SKIP regulated the induction of salt tolerance and displayed an ABA-insensitive phenotype (Lim et al., 2010). SKIP acts as a component of spliceosome to be involved in regulation of alternative splice of circadian clock genes and control flowering time (Lim et al., 2010; Wang et al., 2012; Li et al., 2016; Cui et al., 2017). However, the role of SKIP in endoreduplication control has not been reported so far. Here, we demonstrate that SKIP/SUD3 participates in the regulation of endoreplication and cell growth in Arabidopsis by associating with the endocycle regulators UBP14/DA3 and UVI4.

Results

sud3-1 partially suppresses the phenotypes of da3-1

To further understand the role of UBP14/DA3 in endoreduplication, we performed a genetic screen for the modifiers of da3-1. Several suppressors of da3-1 (sud) from the ethyl methane sulfonate (EMS)-treated M2 populations of da3-1 were isolated (Jiang et al., 2022). We previously identified one of them as sud6-1, which was epistatic to CDKB1;1.N161 to control endoreduplication and cell growth (Jiang et al., 2022). In this study, we identified another suppressor of da3-1 as sud3-1, which suppressed large cotyledon size and plant morphologies of da3-1 (Figure 1; Supplemental Figure S1). Phenotypic analysis demonstrated that the sud3-1 mutant suppressed the enlarged rosette leaves of da3-1 (Supplemental Figure S1B). The size of cotyledons in sud3-1 da3-1 double mutant was smaller than that in da3-1 (Figure 1C). As da3-1 cotyledons contained larger cells than wild-type cotyledons, we carried out cellular analysis of da3-1 and sud3-1 da3-1 cotyledons. Palisade cells in sud3-1 da3-1 cotyledons were smaller than those in da3-1 cotyledons (Figure 1C). We then tested whether small cells in sud3-1 da3-1 cotyledons are associated with changes in ploidy level. Flow cytometry analysis of nuclear DNA content showed that the number of 16C and 32C cells in sud3-1 da3-1 cotyledons was significantly reduced compared with da3-1 cotyledons (Figure 1D). Endoreduplication index (EI) represents the mean number of endocycles per cell (Eloy et al., 2011). We calculated the EI of sud3-1 da3-1 and da3-1. We found a lower EI in sud3-1 da3-1 compared with that in da3-1 (Figure 1E). Collectively, these results indicated that sud3-1 suppresses the ploidy level, cell size, and organ growth phenotypes of da3-1.

Figure 1.

The sud3-1 mutant suppresses the enhanced organ growth and high ploidy level of da3-1. A–B, The sud3-1 mutation suppresses organ growth phenotypes of da3-1. Eleven-day-old seedlings (A) and 30-day-old plants (B) of Col-0, sud3-1, da3-1, and sud3-1 da3-1 (from left to right). C, Cotyledon area (CA) and cotyledon cell area (CCA) of 10-day-old Col-0, sud3-1, da3-1, and sud3-1 da3-1 seedlings (n = 40 biological replicates for CA; n = 20 biological replicates for CCA). D, Estimation of nuclear DNA ploidy in cotyledons of Col-0, sud3-1, da3-1, and sud3-1 da3-1 seedlings at 12 DAG (n = 3 biological replicates). E, EI of Col-0, sud3-1, da3-1, and sud3-1 da3-1 cotyledons at 12 DAG (n = 3 biological replicates). F, Trichome branch distribution in the first pair of leaves of Col-0, sud3-1, da3-1, and sud3-1 da3-1 seedlings at 20 DAG (n = 20 biological replicates); br indicates number of branches. Values in (C) and (E) are given as mean ± se relative to the wild-type values, set at 100%. Values in (D) and (F) are given as mean ± se. se, standard error. Different lowercase letters above the columns indicate a significant difference among different groups in (C) and (E), as determined by ANOVA (analysis of variance) and Tukey’s post-hoc test (P < 0.05). Different letters above columns of the same color indicate significant differences among different groups in (D) and (F), as determined by ANOVA and Tukey’s post-hoc test (P < 0.05). Bars = 1 mm in (A) and 2 cm in (B).

SUD3 encodes a SNW/SKIP

We used MutMap approach to identify the sud3-1 mutation by using a F2 population of a cross between sud3-1 da3-1 and da3-1 (Abe et al., 2012). In the F2 population, the progeny segregation indicated that a single recessive mutation determines the phenotypes of sud3-1 (Supplemental Table S1). We detected 4,800 SNPs (single nucleotide polymorphisms) and 3,205 INDELs between the pooled F2 samples and da3-1. Considering that all mutant plants in the F2 population should possess the causative SNP/InDel, the SNP/InDel ratio for this causative mutation in bulked F2 plants should be 1. In all, three SNPs have an SNP/INDEL-index = 1. They were identified in exons (Supplemental Table S2). Next, we developed dCAPs (derived Cleaved Amplified Polymorphic Sequences) markers based on these mutations and found that only SNP1806 was cosegregated with the sud3-1 da3-1 phenotypes. SNP1806 occurs in the AT1G77180 gene. The sud3-1 had a single-nucleotide base substitution at nucleotide position 209 bp of AT1G77180 (CCG to CTG) (Figure 2A), resulting in an amino acid change from proline to leucine (P to L) (Figure 2B). These results suggest that the AT1G77180 gene is the candidate gene of SUD3.

Figure 2.

SUD3 encodes SKIP. A, Gene structure of SUD3. The start codon (ATG) and the stop codon (TAA) for this gene are indicated. The mutation of sud3-1 results in a single nucleotide C to T (CCG/CTG) transition. B, The protein structure of SUD3. SUD3 contains a SNW domain. The mutation results in an amino acid change from proline to leucine (P to L). C–D, 11-day-old seedlings (C) and 28-day-old plants (D) of Col-0, da3-1, sud3-1 da3-1, gSUD3-GFP;sud3-1 da3-1#1, and gSUD3-GFP;sud3-1 da3-1#2 (from left to right). gSUD3-GFP;sud3-1 da3-1 plants were transformed with vector harboring genomic sequence of SUD3 gene tagged with GFP in the sud3-1 da3-1 background. E, CA of 11-day-old Col-0, da3-1, sud3-1 da3-1, gSUD3-GFP;sud3-1 da3-1#1 and gSUD3-GFP;sud3-1 da3-1#2 seedlings (n = 50 biological replicates). F, Nuclear DNA ploidy distribution of cells in cotyledons of 11-day-old Col-0, da3-1, sud3-1 da3-1, gSUD3-GFP;sud3-1 da3-1#1, and gSUD3-GFP;sud3-1 da3-1#2 seedlings (n = 3 biological replicates). Values in (E) are given as mean ± se relative to the wild-type values, set at 100%. Values in (F) are given as mean ± se. Different lowercase letters above the columns indicate a significant difference among different groups in (E), as determined by ANOVA and Tukey’s post-hoc test (P < 0.05). Different letters above columns of the same color indicate significant differences among different groups in (F), as determined by ANOVA and Tukey’s post-hoc test (P < 0.05). Bars = 1 mm in (C) and 1 cm in (D).

The identity of the SUD3 gene was further confirmed by genetic complementation analysis. We transformed the genomic sequence of At1g77180 containing 2,724-bp promoter and 1,842-bp gene region fused with GFP (gSUD3–GFP) into sud3-1 da3-1 plants. The plant morphology, cotyledon size, and ploidy levels of gSUD3–GFP;sud3-1 da3-1 transgenic lines were similar to those of da3-1, demonstrating that AT1G77180 is the SUD3 gene (Figure 2, C–F). SUD3 encodes SKIP containing an SNW domain (Figure 2B). Phylogenetic analysis demonstrated that SUD3 is the putative paralog of AtSKIP19, ZmPRP45, and OsSKIPa (Supplemental Figure S2).

Expression and subcellular localization of SUD3

To investigate the expression pattern of SUD3 in Arabidopsis, we generated reporter lines (ProSUD3:GUS) using endogenous SUD3 promoter region to drive the expression of β-glucuronidase (GUS). The tissue-specific expression patterns of SUD3 were examined at cotyledons, leaves, inflorescences, and siliques (Supplemental Figure S3, A–F). By the real-time reverse transcriptase-polymerase chain reaction (RT-qPCR) analysis, SUD3 was predominantly expressed in older leaves and had an expression overlap with DA3 and UVI4 (Supplemental Figure S3G). In young leaves of seedlings, SUD3 was more expressed in expansion phases than the proliferation phases during leaf development (Supplemental Figure S3B). In inflorescence, the SUD3 expression was detected in flower buds and flowers (Supplemental Figure S3C). In developing stamens, sepals, and siliques, older tissues exhibited higher GUS activity than younger ones (Supplemental Figure S3, D–F).

To determine the subcellular localization of SUD3, we expressed a SUD3–GFP fusion protein under the control of the SUD3 promoter in sud3-1 da3-1 plants (gSUD3–GFP;sud3-1 da3-1). Transgenic lines gSUD3–GFP;sud3-1 da3-1 recovered the phenotypes of sud3-1 da3-1 to da3-1, indicating that the SUD3–GFP fusion protein is functional (Figure 2, C–F). As shown in Supplemental Figure S3H, GFP fluorescence in transgenic plants was observed exclusively in nuclei. Thus, these results show that SUD3 is a nuclear-localized protein as reported previously, consistent with the role of SUD3 in endoreduplication (Lim et al., 2010).

Overexpression of SUD3 increases endoreduplication

Since sud3-1 can partially suppress growth phenotypes of da3-1, we next investigated the effects of the sud3-1 single mutation on plant growth and development. To gain this aim, we obtained a single mutant sud3-1 from the F2 population of a cross between sud3-1 da3-1 and Col-0. The sud3-1 mutant exhibited similar plant morphology to the wild-type (Figure 3A). The cotyledon area and cotyledon cell area of sud3-1 were not significantly changed compared to wild-type (Figure 3B). We then measured ploidy levels in the wild-type and sud3-1. In 11-day after germination (DAG) seedlings, the ploidy levels of sud3-1 cotyledon cells were also similar to those of wild-type cotyledon cells (Figure 3C).

Figure 3.

Phenotype analysis of transgenic lines overexpressing SUD3. A, Eleven-day-old Col-0, sud3-1, 35S:SUD3#1, 35S:SUD3#2, and 35S:SUD3#3 seedlings (from left to right). B, The measurement of CA and CCA of Col-0 and sud3-1 at 11 DAG (n = 30 biological replicates for CA and n = 20 biological replicates for CCA). C, Estimation of nuclear DNA ploidy in cotyledons of Col-0 and sud3-1 seedlings at 11 DAG (n = 3 biological replicates). D, Relative expression levels of SUD3 in the cotyledons of Col-0, 35S:SUD3#1, 35S:SUD3#2, and 35S:SUD3#3 seedlings (n = 3 biological replicates). E, The measurement of CA and CCA of 11-day-old Col-0, 35S:SUD3#1, 35S:SUD3#2 and 35S:SUD3#3 seedlings (n = 30 biological replicates for CA and CCA). F, Nuclear DNA ploidy distribution of cells in cotyledons of Col-0, 35S:SUD3#1, 35S:SUD3#2 and 35S:SUD3#3 seedlings (n = 3 biological replicates). G, EI of Col-0, 35S:SUD3#1, 35S:SUD3#2, and 35S:SUD3#3 cotyledons at 12 DAG (n = 3 biological replicates). Values in (B), (D), (E), and (G) are given as mean ± se relative to the wild-type values, set at 100%. Values in (C) and (F) are given as mean ± se. Different lowercase letters above the columns indicate a significant difference among different groups in (B), (E), and (G), as determined by ANOVA and Tukeys post-hoc test (P < 0.05). Different letters above columns of the same color indicate significant differences among different groups in (C) and (F), as determined by ANOVA and Tukey’s post-hoc test (P < 0.05). Bars = 1 mm in (A).

For better understanding of the SUD3 function in the regulation of endocycle and cell growth, we transformed Col-0 plants with the SUD3 gene driven by CaMV 35S promoter (35S:SUD3). The expression levels of SUD3 in transgenic lines (35S:SUD3) were strongly increased compared with those in Col-0 (Figure 3D). 35S:SUD3 transgenic plants formed large leaves, in contrast to wild-type leaves (Figure 3A). We then investigated the phenotypes of 35S:SUD3 cotyledons and the first pair of leaves. The average cotyledon area and cotyledon cell area of 35S:SUD3 were increased compared with that of the wild-type (Figure 3E). Flow cytometry analysis was performed to measure ploidy levels of these overexpressing lines. Notably, the 16C and 32C fractions were higher in 35S:SUD3 cotyledons than those in wild-type cotyledons (Figure 3F). The EIs were significantly higher in 35S:SUD3 cotyledon cells than in Col-0 cotyledon cells (Figure 3G). We next measured the leaf area and leaf cell area in the first pair of leaves in 35S:SUD3 plants (Supplemental Figure S4). The average leaf area and leaf cell area of 35S:SUD3 transgenic plants were increased compared with those of Col-0 (Supplemental Figure S4). Taken together, these results indicated that overexpression of SUD3 increases endoreduplication and cell and organ growth.

In plants, SKIP is a bifunctional regulator of gene expression that works as a component of a splicing complex and as a transcriptional regulator (Li et al., 2019). AtSKIP regulates the splicing of serrated leaves and early flowering (SEF) pre-messenger RNA (mRNA) to control flowering time (Cui et al., 2017). Consistent with previous study, we found that the sud3-1 mutation regulates the alternative splicing of SEF in Arabidopsis (Supplemental Figure S5A). We then asked whether SUD3 affects the splicing of cell cycle genes. We tested several cell cycle genes and found that the sud3-1 did not influence the splicing of these cell cycle genes (Supplemental Figure S5, B and C). However, we found that the expression levels of CDKD;1, CENP-C, CYCA3;4, CYCB2;1, CYCB2;2, CYCB2;4, CYCD1;1, CYCH;1, DPA, E2FB, and EDA25 were higher in sud3-1 da3-1 than those in da3-1 (Supplemental Figure S6). These results suggest that the altered expression of these cell cycle genes might partially contribute to the effect of sud3-1 on endoreduplication in da3-1 background.

UBP14/DA3 acts genetically with SUD3 to regulate endoreduplication

Because UBP14/DA3 is a negative regulator of endoreduplication, and the mutation in the SUD3 gene partially suppresses the ploidy and cell and organ growth phenotypes of da3-1, we examined the genetic relationship between DA3 and SUD3 with respect to endoreduplication and cell and organ growth. We first measured the size of Col-0, da3-1, sud3-1, and sud3-1 da3-1 cotyledons. The size and palisade cell size of sud3-1 da3-1 cotyledons were smaller than that of da3-1 cotyledons (Figure 1C). The flow cytometry analyses revealed that ploidy level in cotyledons of the double mutant sud3-1 da3-1 was significantly lower than that in the da3-1 (Figure 1, D and E). The changes in endoreduplication are often associated with trichome branching pattern in Arabidopsis. To find whether endoreduplication in sud3-1 da3-1 is functionally linked with variations in trichome branch number, we counted trichome branch numbers from the first pair of leaves from Col-0, sud3-1, da3-1, and sud3-1 da3-1. The trichomes of sud3-1 had possessed similar number of branches to those of the wild-type (Figure 1F). The sud3-1 da3-1 trichomes with 4 and 5 branches were significantly lower than da3-1, indicating that the sud3-1 mutation suppresses the trichome branch phenotype of da3-1. Collectively, these results suggest that sud3-1 is partially epistatic to da3-1 with respect to endoreduplication, trichome branch and cell size, and organ growth.

SUD3 physically interacts with UBP14/DA3 in vitro and in vivo

The genetic analyses indicate that SUD3 acts in a common genetic pathway with UBP14/DA3 to regulate endoreduplication and organ growth in Arabidopsis. We therefore asked whether SUD3 could physically interact with UBP14/DA3 to control endoreduplication. We performed bimolecular fluorescence complementation (BiFC) assay to test the interaction of SUD3 with DA3 in leaves of Nicotiana benthamiana. We transiently co-expressed nYFP-SUD3 with cYFP-DA3 in leaves of N. benthamiana. The leaves co-expressing nYFP-SUD3 and cYFP-DA3 displayed strong YFP signals in epidermal cells of N. benthamiana leaves, whereas no signal was detected in negative controls (Figure 4A). These results indicate that SUD3 physically associates with UBP14/DA3 in vivo.

Figure 4.

SUD3 interacts with DA3 in vivo and in vitro. A, The BiFC assays indicate that SUD3 interacts with DA3 in N. benthamiana. nYFP-SUD3 and cYFP-DA3 were coexpressed in leaves of N. benthamiana.  Bars =20 μm. B, SUD3 binds DA3 in vitro. DA3 fused to MBP tag was pulled down by His-SUD3 immobilized on His beads and then probed with His and MBP antibodies. C, SUD3 associates with DA3 in Arabidopsis. Total proteins extracted from 35S:GFP and gSUD3-GFP;Col-0 Arabidopsis plants were immunoprecipitated (IP) with GFP-Trap-A beads and then probed with anti-GFP and anti-DA3 antibodies, respectively. IB, immunoblot.

To test whether SUD3 could directly interact with DA3 in vitro, we performed pull-down assay in vitro. The SUD3 was expressed as a His fusion protein, while DA3 was expressed as maltose-binding protein (MBP) fusion protein in Escherichia coli. His-SUD3 directly bound MBP, but His-SUD3 did not bind MBP. This result demonstrates that SUD3 physically interacts with UBP14/DA3 in vitro (Figure 4B).

We then performed co-immunoprecipitation analyses to validate the physical association of SUD3 with DA3 in Arabidopsis. For this purpose, we generated transgenic lines expressing gSUD3–GFP in Col-0 background. The transgenic plants carrying 35S:GFP were used as a negative control. Total proteins were isolated from 35S:GFP and gSUD3–GFP;Col-0 seedlings and incubated with GFP-Trap-A agarose beads to immunoprecipitate GFP-binding complexes. DA3 was detected in the immunoprecipitated SUD3–GFP complexes but not in the negative control, indicating that SUD3 associates with UBP14/DA3 in Arabidopsis (Figure 4C).

SUD3 physically interacts with UVI4 in vivo and in vitro

We previously demonstrated that DA3/UBP14 associates with UVI4 to regulate the activity of APC/Complex and repress endoreduplication in A. thaliana (Xu et al., 2016). CDKB1;1 and CYCA2;3 act downstream of APC/C to repress endoreduplication. Therefore, we asked whether SUD3 could interact with these cell cycle regulators. To test this, we generated cYFP-UVI4 and cYFP-CYCA2;3 constructs and co-expressed them with nYFP-SUD3 in N. benthamiana leaves, respectively. We observed strong YFP signals in epidermal cells in N. benthamiana leaves when we co-expressed nYFP-SUD3 and cYFP-UVI4 (Figure 5A). In contrast, no YFP signals were detected when we co-expressed nYFP-SUD3 with cYFP and cYFP-CYCA2;3 (Supplemental Figure S7A).

Figure 5.

SUD3 physically interacts with UVI4. A, BiFC assays showing that SUD3 interacts with UVI4 in N. benthamiana. cYFP-UVI4 was coexpressed with nYFP-SUD3 in leaves of N. benthamiana. Bars = 20 μm. B, Interaction between SUD3 and UVI4 in the co-immunoprecipitation assays. Nicotiana benthamiana leaves were transformed by injection of Agrobacterium GV3101 cells harboring Pro35S:Myc-UVI4 and Pro35S:GFP, Pro35S:Myc-UVI4 and Pro35S:GFP-DA3 or Pro35S:Myc-UVI4 and Pro35S:GFP-SUD3 plasmids. Total proteins were immunoprecipitated with GFP-Trap-A, and the immunoblot was probed with anti-GFP and anti-MYC antibodies, respectively. C, SUD3 physically interacts with UVI4 in vitro. GST-UVI4 was incubated with MBP-SUD3 and pulled down by MBP-SUD3 and detected by immunoblot with anti-GST antibody.

We then performed co-immunoprecipitation analysis to validate the association of SUD3 with UVI4 in leaves of N. benthamiana. Leaves of N. benthamiana were transformed by injection of Agrobacterium GV3101 cells harboring 35S:GFP-SUD3 and 35S:Myc-UVI4 plasmids. Total proteins were immunoprecipitated with GFP-Trap-A and the immunoblot was probed with anti-GFP and anti-Myc antibodies, respectively. Myc-UVI4 was detected in the GFP-SUD3 complexes but not in the negative control (GFP), further supporting that SUD3 associates with UVI4 in vivo (Figure 5B).

To test whether SUD3 could directly interact with UVI4 in vitro, pull-down assay was performed. SUD3 was expressed as a MBP fusion protein, while UVI4 was expressed as GST fusion protein in E. coli. The GST-UVI4 was pulled down by MBP-SUD3 immobilized on MBP beads. As shown in Figure 5C, GST-UVI4 physically interacted with MBP-SUD3, but not the negative control (MBP) in vitro. Meanwhile, pull-down assay was performed to test whether SUD3 could directly interact with CDKB1;1 in vitro. SUD3 was expressed as a His fusion protein, while CDKB1;1 was expressed as MBP fusion protein in E. coli. The MBP-CDKB1;1 cannot be pulled down by His-SUD3 immobilized on His beads (Supplemental Figure S7B). Thus, this result indicates that SUD3 directly binds to UVI4 in vitro.

sud3-1 partially suppresses the endoreplication phenotype of uvi4

Previous studies demonstrated that uvi4 loss-of-function plants display increased DNA ploidy level (Heyman et al., 2011). Considering that SUD3 interacts with UVI4 in vivo and in vitro, we further investigated the genetic relationship between SUD3 and UVI4 with respect to endoreduplication. We generated the sud3-1 uvi4 double mutant and investigated several morphological and cellular parameters in Col-0, uvi4, sud3-1, and sud3-1 uvi4. Under our growth conditions, the cotyledon cell area in uvi4 was repressed by sud3-1 (Figure 6A). The flow cytometry analysis of 10-day-old cotyledons revealed that sud3-1 partially suppresses the high DNA ploidy phenotype of uvi4 (Figure 6B). The EI in sud3-1 uvi4 was obviously lower than those of uvi4 (Figure 6C). Considering that DA3 acts genetically with UVI4 to regulate endoreduplication, and SUD3 interacts with both DA3 and UVI4, it would be of interest to see the effect of the higher order mutant sud3-1 uvi4 da3-1. We obtained the sud3-1 uvi4 da3-1 triple mutant and measured the ploidy level. Flow cytometry analysis of nuclear DNA content showed that the number of 32C and 64C cells in sud3-1 uvi4 da3-1 cotyledons was significantly reduced compared with uvi4 da3-1 cotyledons (Figure 6D). Furthermore, the EI in sud3-1 uvi4 da3-1 was reduced compared with uvi4 da3-1 (Figure 6E). Collectively, these results indicated that SUD3 acts downstream of DA3 and UVI4 to control endoreduplication.

Figure 6.

SUD3 acts in a common pathway with UVI4 and DA3 to control endoreduplication. A, The measurement of CCA of Col-0, sud3-1, uvi4 and sud3-1 uvi4 seedlings at 11 DAG (n = 20 biological replicates). B, Nuclear DNA ploidy distribution of cells in cotyledons of Col-0, sud3-1, uvi4, and sud3-1 uvi4 seedlings at 10 DAG (n = 3 biological replicates). C, EI of Col-0, sud3-1, uvi4, and sud3-1 uvi4 cotyledons at 10 DAG (n = 3 biological replicates). D, Nuclear DNA ploidy distribution of cells in cotyledons of Col-0, sud3-1, uvi4 da3-1, and sud3-1 uvi4 da3-1 seedlings at 12 DAG (n = 3 biological replicates). E, EI of Col-0, sud3-1, uvi4 da3-1, and sud3-1 uvi4 da3-1 cotyledons at 11 DAG (n = 3 biological replicates). Values in (A), (C), and (E) are given as mean ± se relative to the wild-type values, set at 100%. Values in (B) and (D) are given as mean ± se. Different lowercase letters above the columns indicate a significant difference among different groups in (A), (C), and (E), as determined by ANOVA and Tukey’s post-hoc test (P < 0.05). Different letters above columns of the same color indicate significant differences among different groups in (B) and (D), as determined by ANOVA and Tukey’s post-hoc test (P < 0.05).

Discussion

It is now a well-established concept that the coordinated regulation of cell proliferation and cell expansion determines organ growth and development in plants. Many studies suggest that endoreduplication is coupled with cell expansion to meet the developmental and physiological requirements of the cell. Although several genes with the capacity to influence endoreduplication have been identified in plants, the mechanisms of plant endoreduplication have not been fully understood (Dewitte and Murray, 2003; Sugimoto-Shirasu and Roberts, 2003; De Veylder et al., 2011; Edgar et al., 2014). In previous studies, UBP14/DA3 interacts with UVI4 to modulate endoreduplication and cell and organ growth (Xu et al., 2016). In this study, we prove that SKIP/SUD3 acts genetically and physically with cell cycle regulators DA3 and UVI4 to regulate endoreduplication that influences cell and organ growth.

The da3-1 mutant produces large cotyledons and flowers but small and abnormal plant morphology due to increased cell size and DNA ploidy in an organ-dependent way (Xu et al., 2016). The uvi4 mutant plants display an increased DNA ploidy level (Heyman et al., 2011). Morphology and cellular analyses showed that the sud3-1 mutation partially suppressed the phenotypes of da3-1 and uvi4 with respect to endoreduplication (Figures 1 and 6; Supplemental Figure S1). Overexpression of SKIP/SUD3 significantly increases ploidy level, indicating that SKIP/SUD3 is a positive regulator of endoreduplication (Figure 3, F and G). Interestingly, the sud3-1 single mutant did not obviously affect endoreduplication and plant growth, although this mutation can partially suppress the phenotypes of da3-1 and uvi4. Similar genetic phenomena have been reported in plants. The T-DNA insertion mutants of UBIQUITIN-CONJUGATING ENZYME 32 did not show obvious growth phenotypes, but ubc32 mutants partially suppressed the growth phenotype of brassinosteroid insensitive 1-9 (Cui et al., 2012). The hypocotyl length of the T-DNA insertion mutant for THESEUS1 (the1-3) seedlings was comparable to that of the wild-type, but the1-3 partially suppressed the hypocotyl-elongation defect of procuste1-8 and several other cellulose-deficient mutants, indicating that PROCUSTE 1 and THESEUS1 play crucial function in the hypocotyl elongation (Hematy et al., 2007). The null allele skip-2 is severely dwarfed and sterile, indicating SKIP is required for plant normal growth and development (Wang et al., 2012). This also suggests that sud3-1 is a weak allele.

In plants, SKIP is a component of spliceosome and mediates alternative splicing for clock and stress tolerance genes (Wang et al., 2012; Feng et al., 2015; Cui et al., 2017). SKIP also works as a transcriptional regulator by interacting with EARLY FLOWERING 7 to regulate FLOWERING LOCUS C transcription (Cao et al., 2015). However, the genetic and molecular mechanism of SKIP/SUD3 in endoreduplication control has not been reported in plants. In this study, we find the sud3-1 did not influence the splicing of several cell cycle genes (Supplemental Figure S5, B and C). Interestingly, the expression levels of several cell cycle genes elevated in sud3-1 da3-1 in comparison to those in the da3-1 (Supplemental Figure S6). These data suggest that the decreased ploidy level in sud3-1 da3-1 might, in part, result from the altered expression of these cell cycle genes. Considering SKIP/SUD3 has been reported to function as a transcriptional regulator, we propose that SKIP/SUD3 may associate with an unknown transcription factor to regulate expression of these cell cycle genes. Considering that SUD3 and SUD6 have been known to influence mRNA splicing (Wang et al., 2012; Li et al., 2016; Cui et al., 2017; Cavallari et al., 2018; Nibau et al., 2019), it will be interesting to identify commonly alternatively spliced genes of SUD3 and SUD6 and investigate their functions in endoreduplication in the future.

In this study, the genetic and biochemical data indicate that SUD3 may act downstream of DA3 and UVI4 to control endoreduplication. The expressions of DA3 and UVI4 were present in both the proliferation and expansion phases during leaf development (Hase et al., 2006; Xu et al., 2016). The expression of SUD3 was more present in expansion phases than in proliferation phases (Supplemental Figure S3B). We assume that SUD3 may plays a key role in expansion phases during leaf development. It is possible that SUD3 associates with DA3 and UVI4 to regulate endoreduplication mainly in expansion phases. DA3 interacts with UVI4 to influence the protein abundance of two downstream components, CYCA2;3 and CDKB1;1, to regulate endoreduplication (Xu et al., 2016). We therefore measured the protein levels of CDKB1;1 in da3-1 and sud3-1 da3-1 seedlings by immunoblot analysis (Supplemental Figure S8). The protein levels of CDKB1;1 in da3-1 and sud3-1 da3-1 have no significant difference (Supplemental Figure S8D). We proposed that SUD3 is not involved in the regulation of CDKB1;1 abundance. How SUD3 positively regulate endoreduplication remains to be investigated in the future. Thus, our findings discover a mechanism by which SUD3 interacts with two endoreduplication regulators UBP14/DA3 and UVI4 to regulate endoreduplication (Figure 7).

Figure 7.

A model for UBP14/DA3-UVI4-SKIP/SUD3 complex-mediated control of endoreduplication in Arabidopsis. UBP14 interacts with UVI4 to repress endoreduplication and cell growth in Arabidopsis. SKIP/SUD3 associates with UBP14 and UVI4 in vitro and in vivo. SUD3 acts antagonistically in a common pathway with UBP14 and UVI4 to control endoreduplication.

Materials and methods

Plant material and growth conditions

In this study, all Arabidopsis (A. thaliana) mutants were in the Col-0 background. The da3-1 and uvi4 were described previously (Xu et al., 2016). The suppressor of da3-1 (sud3-1 da3-1) was screened from an EMS mutagenized M2 population of da3-1. The single mutant sud3-1 was obtained from F2 population of a cross between sud3-1 da3-1 and Col-0. Seeds were surface sterilized as previously described and plated on 1/2 Murashige and Skoog medium (Jiang et al., 2022). Plants were grown at 22°C under a 16-h light (28 W/6500 K)/8-h dark cycle and 55% relative humidity.

Morphological and cellular analysis

Cotyledons were photographed under a Leica microscope (Leica S8APO) with a Leica CCD camera (DFC420). To measure cell area, cotyledons and the first pair of leaves were mounted in clearing buffer, as described previously (Jiang et al., 2022). Differential interference contrast optics on a Leica DM2500 microscope was used to observe samples, and a Spot Flex Cooled CCD digital image system (SPOT Imaging) was used to photograph the samples. ImageJ software (http://rsb.info.nih.gov/ij/) was used to measure cotyledons area, cotyledon cell area, the leaf area, and leaf cell area in the first pair of leaves. The GFP fluorescence in roots was observed by the Zeiss LSM 710 NLO Confocal Microscope system (Chameleon Ultra II) using an excitation 488-nm laser with an emission wavelength of 505–550 nm for GFP.

Flow cytometry analysis

Cotyledons were chopped with a razor blade in 500-μL Galbraith GS buffer as described before (Jiang et al., 2022). The nuclei suspension was then filtered through a 40-μm mesh and stained with 5-μL (1 μg mL⁻¹) 4,6-diamidino-2-phenylindole staining solution. The nuclear DNA content distribution was estimated with BD FACSAria II flow cytometer, and values were analyzed against relative fluorescence intensities of the wild-type. The experiment was performed in three biological replicates with three measurements for each biological sample. A total of 10,000 nuclei were analyzed per experiment. EI represents the mean number of endocycles per cell and was calculated using the following equation: EI =  [4C] + 2[8C] + 3[16C] + 4[32C] + 5[64C] where [4C] is the percentage of 4C nuclei, [8C] is the percentage of 8C nuclei, and so on. Statistical analysis was performed using a one-way ANOVA followed by Tukey’s post-hoc test.

Identification of the SUD3 gene

The MutMap approach was used to identify the sud3-1 mutation by using an F2 population of a cross between sud3-1 da3-1 and da3-1. We extracted DNA from 42 plants of an F2 population that showed the sud3-1 da3-1 phenotypes, and the same amount of DNA was mixed for whole-genome sequencing. DNA from da3-1 was also sequenced as a control. The NovaSeq 6000 platform system (Illumina, San Diego, CA, USA) was used for whole-genome resequencing.

Phylogenetic analysis

The protein sequence of SKIP/SUD3 was used to identify its homologs via BLAST search. The phylogenetic tree shown in Supplemental Figure S2 was constructed with MEGA6 software using the Neighbor-Joining method with the following parameters: Poisson model, Uniform rates, Complete deletion, and 1,000 replicates Bootstrap. NCBI (National Center for Biotechnology Information) reference sequence numbers are indicated before the specie names. The scale bar at bottom represents the genetic distance. Branch lengths are shown proportional to the amino acid variation rates.

RNA extraction, RT–PCR, and RT–qPCR analysis

Total RNA was extracted from various plant organs using an RNAprep Pure Plant kit (Tiangen, Beijing, China). The mRNA was reversely transcribed into cDNA using SuperScript III Reverse Transcriptase (Genstar). For RT–qPCR in Supplemental Figure S3, total RNA was extracted from the first pair of leaves in 8 DAG, 9 DAG, and 10 DAG wild-type seedlings. Quantitative assays were performed on 3 μL of each cDNA dilution using the 2× RealStar Green Fast Mixture. The internal control is ACTIN2 (Supplemental Table S4). Relative quantitative analysis was determined using the cycle threshold method. The primers used for RT–qPCR are listed in Supplemental Table S4.

For RT–PCR in Supplemental Figure S5 and RT–qPCR in Supplemental Figure S6, total RNA was extracted from 11-DAG cotyledons. The primers used to analyze the alternative splicing of the different genes are listed in Supplemental Table S3.

Constructs and plant transformation

The 4,566-bp SUD3 genomic sequence was amplified from Col-0 using the primers gSUD3–GFP-F/R and cloned into the pMDC107 vector with PacI and KpnI sites using In-Fusion enzyme (Biomed, China) to generate the gSUD3–GFP plasmid (Supplemental Table S4). The gSUD3–GFP construct was transferred into Col-0 and sud3-1 da3-1 plants by Agrobacterium tumefaciens-mediated transformation using the floral dip method, respectively (Clough and Bent, 1998). The primers of SUD3–CDS-F/R (Supplemental Table S4) were used to amplify the 1,842-bp cDNA sequence of the SUD3 gene. The full-length SUD3 cDNA was then inserted into the pMDC32 vectors with AscI and SacI sites by using infusion enzyme (Genebank Biosciences, Jiangsu, China) to generate the 35S:SUD3 construct. The plasmid 35S:SUD3 was inducted into Col-0 plant by using A. tumefaciens GV3101 to generate 35S:SUD3 transgenic lines. The primers ProSUD3–GUS-F/R were used to amplify the 2,073-bp sequence of SUD3 promoter (Supplemental Table S4). The SUD3 promoter was inserted into the binary vector pMDC164 containing GUS reporter gene with PacI and AscI sites by using infusion enzyme to generate the ProSUD3:GUS construct. The plasmid ProSUD3:GUS was transformed into Col-0 A. thaliana seedlings using Agrobacterium GV3101.

BiFC assay

The coding sequence of SUD3 was amplified by primers nYFP-SUD3-F/R (Supplemental Table S4). The SUD3 CDS was linked to the N-terminal fragment of YFP (nYFP) and cloned into the pGWB414 vector (Invitrogen, Waltham, MA, USA) with XbaI and SalI sites by using infusion enzyme (Genebank Biosciences). The primers cYFP-735-F/R were used to amplify C-terminal fragment of YFP (cYFP) from pSY735, and fused with the DA3 gene. Then the infusion enzyme (Genebank Biosciences) was used to subclone fragment into the pGWB414 vector (Invitrogen). nYFP-SUD3 and cYFP-DA3 plasmids were transformed into Agrobacterium GV3101. Similarly, we constructed the plasmids cYFP-UVI4 and cYFP-CYCA2;3. Combinations of nYFP-SUD3/cYFP, nYFP-SUD3/cYFP-DA3, nYFP-SUD3/cYFP-UVI4, and nYFP-SUD3/cYFP-CYCA2;3 were coinfiltrated into N. benthamiana leaves by Agrobacterium strain GV3101-mediated transformation respectively. Plants were grown for 2 days after infiltration before observation. Fluorescent signals were detected under a Zeiss LSM 710 confocal microscope (Zeiss LSM710 META, Germany).

Co-immunoprecipitation

Using primers Myc-UVI4-F/R to amplify the sequence of UVI4 and cloned into the pCAMBIA1300-221-Myc vector with KpnI and BamHI sites, we generated the transformation plasmid Pro35S:Myc-UVI4. The CDS of SUD3 was linked to the vector pMDC43 with AscI and SacI sites to generate the Pro35S:GFP-SUD3 construct. The Pro35S:GFP-DA3 construct was described previously (Xu et al., 2016). The specific primers for Pro35S:GFP-SUD3 construct were SUD3–CDS-GFP-F/R (Supplemental Table S4). Then N. benthamiana leaves were transformed by injection of Agrobacterium GV3101 cells harboring Pro35S:Myc-UVI4 and Pro35S:GFP-DA3 plasmids. Total proteins were extracted with extraction buffer (50-mM Tris/HCl, pH 7.5, 150-mM NaCl, 20% (v/v) glycerol, 2% (v/v) Triton X-100, 1-mM EDTA, Complete protease inhibitor cocktail [Roche, Basel, Switzerland], and 20 mg mL^–1 MG132) and incubated with GFP-Trap-A (Chromotek, Bayern, Germany) for 1 h at 4°C. Beads were washed 3 times with wash buffer (50-mM Tris/HCl, pH 7.5, 150-mM NaCl, 0.1% (v/v) Triton X-100, and Complete protease inhibitor cocktail [Roche]). The immunoprecipitates were separated in 10% (v/v) sodium dodecyl sulfate (SDS)–polyacrylamide gel and detected by immunoblot analysis with anti-GFP (Abmart, Shanghai, China) and anti-Myc (Abmart) antibodies, respectively. Total proteins from Pro35S:GFP-SUD3 and Pro35S:GFP transgenic plants were extracted with extraction buffer and incubated with Anti-GFP-Tag mouse mAb (Abmart) for 1 h at 4°C. Beads were washed 4 times with wash buffer. The immunoprecipitates were separated in 10% (v/v) SDS–polyacrylamide gel and detected by immunoblot analysis with anti-GFP (Abmart) and anti-DA3 antibodies, respectively (Xu et al., 2016).

Pull-down assays

The coding sequence of SUD3 was cloned into SacI and KpnI sites of the pHUE vector to generate His-SUD3 construct. The coding sequence of SUD3 was cloned into BamHI and HindIII sites of the pMALC2 vector to generate MBP-SUD3 construct using infusion enzyme (Genebank Biosciences). The specific primers for His-SUD3 and MBP-SUD3 were His-SUD3-F/R and MBP-SUD3-F/R, respectively (Supplemental Table S4). The constructs MBP-DA3, GST-UVI4, and MBP-CDKB1;1 were described previously (Xu et al., 2016; Jiang et al., 2022).

To verify protein–protein interaction in vitro in Figure 4 and Supplemental Figure S7B, bacterial lysates containing His-SUD3 fusion proteins were mixed with lysates containing MBP-DA3 fusion proteins and MBP-CDKB1;1 fusion proteins. Lysates were incubated with Dynabeads His-Tag (10104D, Invitrogen) at 4°C for 1 h. Beads were washed 4 times with 1-mL TGH buffer (50-mM HEPES, pH 7.5, 150-mM NaCl, 1.5-mM MgCl₂, 1-mM EGTA, pH 8.0, 1% (v/v) Triton X-100, 10% (v/v) glycerol, 1-mM PMSF, and Protease inhibitor cocktail tablet [Roche]). The isolated proteins were separated on a 10% (v/v) SDS–PAGE and detected by immunoblot analysis with anti-His anti-MBP antibodies, respectively. To verify protein–protein interaction in vitro in Figure 5C, bacterial lysates containing GST–UVI4 fusion proteins were mixed with lysates containing MBP–SUD3 fusion proteins and MBP proteins. Lysates were incubated with MBP beads at 4°C for 1 h. Beads were washed 4 times with 1-mL TGH buffer. The isolated proteins were separated on a 10% (v/v) SDS–PAGE and detected by immunoblot analysis with anti-GST and anti-MBP antibodies, respectively.

Total plant protein extraction and immunoblot assays

To examine the CDKB1;1 protein level, total proteins were isolated from da3-1 and sud3-1 da3-1 seedlings with extraction buffer. Immunoblot analysis was performed using anti-Actin (Abmart; M20009m) and anti-CDKB1;1 (PhytoAB, PHY0912S) antibodies. The intensities of CDKB1;1 bands and corresponding Actin bands on blots were measured using ImageJ software. CDKB1;1 protein levels were expressed relative to Actin. Values in sud3-1 da3-1 are given as mean ± SE (n = 3 biological replicates) relative to the value for da3-1, set at 1.

Accession numbers

Sequence data from this article can be found in the GenBank/EMBL library under the following accession numbers: UBP14/DA3 (AT3G20630), SKIP/SUD3 (AT AT1G77180), UVI4 (AT2G42260), CDKB1;1 (AT3G54180), SEF (AT5G37055), CDKD1;1 (AT1G73690), CENP-C (AT1G15660), CYCA3;4 (AT1G47230), CYCB2;1 (AT2G17620), CYCB2;2 (AT4G35620), CYCB2;4 (AT1G76310), CYCD1;1 (AT1G70210), CYCH;1 (AT5G27620), DPA (AT5G02470), E2FB (AT5G22220), EDA25 (AT1G72440), CDKB2;1 (AT1G76540), and APC6 (AT1G78770).

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. The sud3-1 mutant suppresses the plant phenotype of da3-1.

Supplemental Figure S2. The phylogenetic analysis of SUD3 protein from different species.

Supplemental Figure S3. Spatial expression pattern and subcellular localization of SUD3.

Supplemental Figure S4. Overexpression of SUD3 increases the leaf area and leaf cell area.

Supplemental Figure S5. Intron-retention analysis of pre-mRNA transcripts of several cell cycle genes.

Supplemental Figure S6. The expression levels of cell cycle genes listed in Supplemental Figure S5 by RT-qPCR.

Supplemental Figure S7. SUD3 could not form a complex with CYCA2;3/CDKB1;1.

Supplemental Figure S8. The protein levels of CDKB1;1 in da3-1 and sud3-1 da3-1 seedlings.

Supplemental Table S1. The statistics of progeny segregation in the F2 population of a cross between sud3-1 da3-1 and da3-1.

Supplemental Table S2. Identification of the sud3-1 mutation using the MutMap approach.

Supplemental Table S3. List of primers used in Supplemental Figure S5.

Supplemental Table S4. List of primers used in this study.

 

These authors contributed equally (S.J., B.M., and Y.Z.)

Y.L. conceived this project. Y.Z., S.J., and B.M. performed most experiments. L.Z. and R.X. mapped the SUD3 gene. S.J., B.M., Y.Z., X.Z., N.L., S.G., C.P.S., and Y.L. analyzed the data. S.J., B.M., S.G., C.P.S., and Y.L. wrote the manuscript.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/general-instructions) is Yunhai Li (yhli@genetics.ac.cn).