Introduction

After germination, the plant shoot apex passes through two different phases: a vegetative phase and a reproductive phase. The vegetative phase could be further divided into a juvenile vegetative phase and an adult vegetative phase. In addition to their difference in reproductive competence, the juvenile and adult vegetative phases can also be distinguished by a series of distinct phase-specific traits including morphological, physiological, and biochemical traits (Poethig 2003, 2013). The transition from the juvenile to the adult vegetative phase is referred to as vegetative phase change. In Arabidopsis, vegetative phase change is marked by an increase in the degree of serration of the leaf margin, an increase in the leaf length/width (L/W) ratio, a decrease in cell size, and the production of trichomes on the abaxial side of leaf blades (Poethig 1990; Telfer et al. 1997).

Vegetative phase change in plants is regulated by a decrease in the level of miR156/157 and a corresponding increase in their targets, the SPL genes (Wu and Poethig 2006; Wu et al. 2009a, b). SBPs (SQUAMOSA promoter binding proteins, SBPs) were originally identified in Antirrhinum majus as factors that bind to the promoter of the floral meristem identity gene SQUAMOSA (Klein et al. 1996). Later, they were identified in all green plants, and they constitute a plant-specific transcription factor family (Guo et al. 2008). Except for the conserved SBP domain with ~ 75 amino acids, plant SPLs are diverse in their primary structures. Nuclear magnetic resonance (NMR) results showed two, non-interleaved Zn2+ coordinating structures within SBP domains, and they are formed by eight conserved cysteine and histidine residues (Yamasaki et al. 2004). Amino acid exchange experiments indicate that each Zn2+ structure is essential for the binding to the GTAC core motif (Birkenbihl et al. 2005).

The Arabidopsis genome encodes 17 SPL genes, and 10 SPL genes including SPL2, SPL3, SPL4, SPL5, SPL6, SPL9, SPL10, SPL11, SPL13, and SPL15 are shown to be repressed by miR156 by transcript cleavage or translational inhibition (He et al. 2018). Based on their functional diversity, miR156-targeted genes can be categorized into 3 functionally different groups. SPL9, SPL13, SPL15, SPL2, SPL10 and SPL11 function to promote vegetative phase change with SPL9, SPL13 and SPL15 playing much more important roles than others; SPL6 functions in certain physiological processes, and SPL3, SPL4, SPL5 function to regulate floral meristem identity (Xu et al. 2016).

It has been shown that nuclear import of transcription factors is essential for their function, it is usually initiated by the binding of importin α to NLS (Görlich and Kutay 1999; Kaffman and O’Shea 1999; Jans et al. 2000). A putative bipartite NLS overlapping with the second Zn2+ motif was identified at the C terminal of the SBP domain, characterized by two basic amino acid clusters spaced by 11 amino acid residues (Stone et al. 2005; Birkenbihl et al. 2005). This putative bipartite NLS is required for the nuclear import of SPL3 and SPL8, and a truncation of the complete second part of the bipartite NLS of SPL3 decreased the nuclear import of SPL3 (Birkenbihl et al. 2005). Although this putative bipartite NLS is conserved in all SPLs, the nuclear import efficiency is different among different SPLs, implying that some additional basic regions may function as NLS (Birkenbihl et al. 2005). How different putative NLSs in SPLs affect nuclear import and therefore their functions in plant development and growth remains unclear.

In this study, we identified a loss-of-function mutant with delayed vegetative phase change phenotypes in a forward genetic screen, which we named delayed juvenile-to-adult phase transition mutant 6 (del6). Map-based cloning, the whole genome re-sequencing, and genetic analysis indicate that the Arginine-to-glutamine substitution at the conserved SBP domain of SPL9 is responsible for the phenotype. This mutation disrupted the conserved NLS of SPL9, and caused abnormal nuclear localization pattern of SPL9 in the plant cell, thus abolishing the normal function of SPL9 in vegetative phase change. Our work here therefore demonstrated that the NLS at the C-terminus of SBP is critical for SPL9 normal nuclear localization and function in vegetative phase change.

Materials and methods

Genetic stocks and growth conditions

Genetic stocks used in this study were all in the Columbia (Col) genetic background. spl9-4 (SAIL_150_B05, a null allele mutant) was obtained from the Arabidopsis Biological Resource Center (Columbus, OH, United States). Seeds were sown in a moistened 1:1 mixture of soil:vermiculite and left at 4 °C for 48 h before being transferred to the growth chamber. Plants were grown in short day (SD, 10-h light, 14-h darkness, 140 µm/m2/s, 22 °C). Plant age was measured from the time when seeds were transferred to the growth chamber. Abaxial trichomes were scored 2–3 weeks after planting with a stereomicroscope. For leaf shape analysis, fully expanded leaves were removed, attached to cardboard with double-sided tape and flattened with transparent tape, and then scanned in a digital scanner.

Map-based cloning and whole genome re-sequencing

To map the DEL6 locus, the del6 (Col background) mutant was crossed to Landsberg erecta (Ler), and the selfed F2 individuals exhibiting the del6 phenotype, including rounder leaves with fewer serrations and abaxial trichomes later than leaf fourteen were chosen for mapping. For the whole genome re-sequencing, the del6 mutant was backcrossed to Col plants, and sixty different F2 individuals with the del6 phenotype were chosen to take leaf samples, and DNA from the mixed leaf samples was sent for the whole genome re-sequencing. Based on the above mapping and the whole genome re-sequencing data, we searched for G-to-A and/or C-to-T mutations in the candidate region to pin down the mutation responsible for the del6 phenotype.

NLS and phosphorylation site prediction

The putative NLSs in different SPLs were predicted using the PSORTII program (http://psort.hgc.jp). The potential phosphorylation sites were predicted using the NetPhos 3.1 Server (http://www.cbs.dtu.dk/services/NetPhos/).

Vector construction and plant transformation

35S::GFP plasmid is a lab stock. The CDS sequence of SPL9 was amplified using cDNA samples from both WT and del6, and then were fused to GFP under the control of the 35S promoter to generate 35S::GFP-SPL9WT and 35S::GFP-SPL9del6 plasmids, respectively. Transient expression and subcellular localization analysis of GFP-fused proteins were done following a previous study (Wu et al. 2009). About a five-kb genomic fragment of SPL9 from both WT and del6 mutant were amplified and cloned into pCAMBIA1305.1 to generate gSPL9WT and gSPL9del6 expression vectors. These two expression vectors were then transformed into Agrobacterium tumefaciens strain GV3101 to transform Arabidopsis using the floral dipping method. Seeds of transformed plants were collected and sown on half strength MS medium containing hygromycin (50 mg/L) to select for transgenic plants. Eight independent homozygous transgenic lines (gSPL9WT 1-8 lines, gSPL9del6 1-8 lines) were selected randomly for further characterization in this study.

Results

The del6 mutant delays vegetative phase change

We identified a mutant with delayed vegetative phase change phenotypes in a forward genetic screen of an ethylmethanesulfonate (EMS)-mutagenized M2 population (Fig. 1a). In short day, wild type produced abaxial trichomes on leaf 7.23 ± 0.66, the del6 mutant had abaxial trichomes on leaf 13.73 ± 1.41, which was significantly later than wild type (Fig. 1b). Measurement of leaf length to width ratio (L/W ratio) showed that the del6 mutant had significantly smaller ratio than wild type, suggesting that leaves from the del6 mutant were much rounder than those from wild type, especially for those after the third leaf (Fig. 1c, d). The late-formed leaves from del6 also had fewer serrations than their counterparts from wild type (Fig. 1c). In addition, the del6 mutant had significantly faster leaf initiation rate than wild type (Fig. 1a, e). These results indicate the del6 mutant exhibits typical delayed vegetative phase change phenotypes.

Fig. 1
figure 1

del6 delays vegetative phase change in Arabidopsis.a 5-week-old plants grown in short day (SD). Scale: 1 cm. b The first leaf with abaxial trichomes in wild type (WT) and del6 in SD. The asterisks denote the significant difference from WT at P < 0.01 using Student’s t-test (Data are mean ± SD from n = 37 plants). c Leaf shape of WT and del6 in SD. Red triangles indicate leaf serrations on the 6th leaf margin. Scale: 1 cm. d Leaf length/width (L/W) ratio of different leaves from WT and del6 in SD. Asterisks denote the significant difference from WT at P < 0.01 using Student’s t-test (Data are mean ± SD from n = 37 plants). e Leaf initiation rate of WT and del6 in SD. Asterisks denote the significant difference from WT at P < 0.01 (Data are mean ± SD from n = 37 plants)

Full size image

Map-based cloning of DEL6

To map the del6 mutation, the del6 mutant was first crossed to Landsberg erecta, and the selfed F2 individuals with the del6 phenotype were selected for map-based cloning. The mutation was narrowed down to a ca.3200 Kb region on chromosome 2 between position 16291 K and 19542 K) (Fig. S1). Next, we used the whole genome re-sequencing method to identify mutations in this region. The sequencing result indicated that there were six different mutations within this candidate region, and one of them is located within the SPL9 coding region (Fig. S2). SPL9 has been shown to promote vegetative phase change, and loss-of-function mutation of SPL9 delays vegetative phase change (Wu et al. 2009a, b; Xu et al. 2016). Therefore, we proposed that the mutation in the SPL9 coding region might be responsible for the del6 mutant phenotype. Based on the availability of different spl9 alleles (four T-DNA insertional mutant alleles, spl9-1 to spl9-4) (Schwarz et al. 2008; Xu et al. 2016), we renamed the del6 mutant as spl9-5. In spl9-5 (del6), there is a G-to-A substitution at the +434 bp from the start codon in the SPL9 sequence (Fig. S3), this causes a replacement of 145-Arg by a Gln within the SBP conserved domain (Fig. 2a).

Fig. 2
figure 2

A mutation in the SPL9 gene is responsible for the delayed vegetative phase change phenotype in del6.a Available loss-of-function mutant alleles for SPL9. del6 contains a G-to-A substitution in the SPL9 gene. spl9-1 ~ spl9-4 are T-DNA insertional mutants. The white box indicates non-coding region, the light black box indicates coding region, and the dark black indicates SBP box. b The genetic complementation test result for del6 and spl9-4. Leaf shape and the first leaf with abaxial trichomes in 3-week-old WT, del6, spl9-4, and F1 plants from a cross between del6 and spl9-4 in SD. Different letters indicate significant difference between genotypes using one-way ANOVA at P < 0.01 (Data are mean ± SD from n = 37 plants). Leaf serration on the 6th leaf blade is marked with red triangles. Scale bar: 1 cm

Full size image

del6 acts as a new allelic mutant of the SPL9 gene

To further conform that the mutation in the SPL9 gene is responsible for the del6 mutant phenotype, we did a genetic complementation test by crossing del6 to spl9-4. In short day, del6 and spl9-4 produced abaxial trichomes on leaf 13.73 ± 1.41 and 12.22 ± 1.16, respectively; whereas wild type produced abaxial trichomes on leaf 7.23 ± 0.66, which was significantly earlier than del6 and spl9-4, and del6 and spl9-4 also lacked serrations on leaf margin compared with wild type (Fig. 2b). The F1 progeny plants from the cross between del6 and spl9-4 had almost similar abaxial trichome production to spl9-4 (Fig. 2b), this indicates that del6 was unable to complement the spl9-4 phenotype, suggesting that del6 and spl9-4 are allelic. This result demonstrates that the del6 mutant phenotype is truly attributable to the mutation in the SPL9 gene.

del6 has a mutation in the conserved putative NLSs

Previous studies have shown that there is a putative bipartite NLS within the SBP domain at the C terminals of SPL proteins (Fig. 3a), and it overlaps with the second Zn-coordinating structure (Zn-2 motif) (Stone et al. 2005; Birkenbihl et al. 2005). In this study, we identified three putative NLSs in SPL9 using the POSRTII program. These include a monopartite NLS 1 between amino acid aa-71 and aa-74 in a context of RRRK, a monopartite NLS 2 between aa-72 and aa-75 in a context of RRKP, and a bipartite NLS 1 between aa-58 and aa-74 in a context of KRSCRRRLAGHNERRRK (Fig. 3b). These NLS are located at the C-terminal of the SBP domain, and the bipartite NLS 1 was also described by previous studies (Stone et al. 2005; Birkenbihl et al. 2005) and characterized by clusters with two basic amino acids separated by 11 amino acids. In del6, the mutation is located in the conserved putative bipartite NLS (Fig. 3a, b), resulting in the replacement of the conserved basic amino acid arginine at aa-72 by a glutamine (Fig. 3b). This putative bipartite NLS was shown to be required for the nuclear import of SPL3 and SPL8 (Birkenbihl et al. 2005), but how this mutation in del6 affects SPL9 nuclear import therefore its function remains unknown.

Fig. 3
figure 3

del6 disrupts a putative conserved nuclear localization signal in SPL9. a Cluster analysis of the SBP domain of different SPL proteins using MEME. The two Zn-coordinating structures (Zn-1 and Zn-2) are highlighted with grey according to previous studies (Yamasaki et al. 2004; Birkenbihl et al. 2005). The conserved basic amino acid clusters are highlighted with black, the conserved proline within Group I SPLs is highlighted with blue. Numbers indicate the relative amino acid position in the conserved SBP domain. b Three basic amino acid clusters (BR1 ~ 3) and the fourth putative NLSs at the C terminal of SBP domain. Monopartite NLS 1, monopartite NLS 2 and bipartite NLS 1 are predicted using the POSRTII program, bipartite NLS 2 is predicted according to a previous study (Zhu et al. 2015). R72-to-Q mutation in del6 (spl9-5) is marked with an arrow. Numbers indicate the relative amino acid position in the conserved SBP domain

Full size image

Mutation in the NLS in SBP domain of SPL9 changes the nuclear localization of SPL9

Since the mutation in del6 is located in the conserved putative NLS, we hypothesized that the phenotype of del6 might be attributable to its defects in nuclear localization. Therefore, we studied the nuclear localization of SPL9 and DEL6 using Arabidopsis mesophyll protoplast transformed with 35S::GFP, 35S::GFP-SPL9WT (SPL9 CDS from WT), and 35S::GFP-SPL9del6 (SPL9 CDS from the del6 mutant) plasmids, respectively. In the transient expression assay with 35S::GFP, the GFP signal was mainly concentrated in the nucleus, as well as some distribution of GFP signals in the cytoplasm (Fig. 4). In the assay with 35S::GFP-SPL9WT, GFP was specifically localized in the nucleus, implying that the GFP-SPL9WT protein was imported efficiently into the nucleus. In contrast, the GFP-SPL9del6 fusion protein was diffusible in the plant cell, and a large proportion of the GFP signal was retained in the cytoplasm in addition to its nuclear distribution (Fig. 4). This result suggests that the del6 mutation changes the specific nuclear localization of SPL9, and the conserved NLS in the SBP domain is indispensable for the correct nuclear localization of SPL9 in the plant cell.

Fig. 4
figure 4

Subcellular localization of eGFP, eGFP-SPL9WT and eGFP-SPL9del6 fusion proteins. Plasmids were delivered to protoplast using the PEG-mediated method. After 16 h of transformation, GFP distribution in cells were examined with a confocal fluorescent microscopy. Scale: 10 μm

Full size image

The NLS is required for the function of SPL9 in plants

To future conform if the conserved NLS is required for the function of SPL9, we amplified the genomic sequence of SPL9 from wild type and del6, and cloned them into the pCAMBIA1305.1 expression vector, and transformed wild-type Arabidopsis plants to generate gSPL9WT and gSPL9del6 transgenic plants, respectively. We randomly chose 8 independent homozygous transgenic lines with single T-DNA insertions from both gSPL9WT and gSPL9del6 transgenic plants, and characterized the vegetative phase change phenotype of these plants. In short day, gSPL9WT transgenic plants produced abaxial trichomes significantly earlier than wild type (6.23 ± 0.71 versus 7.54 ± 0.67), whereas gSPL9del6 transgenic plants had abaxial trichomes similar to wild type (7.48 ± 0.87 versus 7.54 ± 0.67) (Figs. 5, S4); moreover, leaves of gSPL9WT were significantly elongated and had more serrations on leaf margins than wild type and del6, whereas leaves of gSPL9del6 were similar to those of wild type (Fig. 5b, d). Leaves of gSPL9WT were also dark green (Fig. 5e). These results demonstrate that the mutation in the NLS of the SPL9 gene disrupts the normal function of SPL9 in vegetative phase change.

Fig. 5
figure 5

The G-to-A substitution in SPL9 impairs its function to promote vegetative phase change. a 4-week-old WT, del6, gSPL9WT, and gSPL9del6 transgenic plants grown in SD. gSPL9WT: transgenic plants transformed with the wild-type SPL9 genomic sequence under the control of its native promoter; gSPL9del6: transgenic plants transformed with the del-6 mutant SPL9 genomic sequence under the control of its native promoter. Scale: 1 cm. b Leaf shape and the first leaf with abaxial trichomes. Leaf serrations on the 6th leaf margin are marked with red triangles. Different letters indicate significant difference between genotypes using one-way ANOVA at P < 0.01 (Data are mean ± SD from n = 17 plants). Scale: 1 cm. c The first leaf with abaxial trichomes from WT and different gSPL9WT and gSPL9del6 transgenic lines. Asterisks denote the significant difference from WT at P < 0.01 (Data are mean ± SD from n = 17 plants). d Leaf L/W ratio of the 6th leaf from different genotypes. Different letters indicate significant difference between genotypes using one-way ANOVA at P < 0.01 (Data are mean ± SD from n = 10 individual plant). e Leaves from gSPL9WT transgenic lines are dark green. The 6th leaf from different genotypes are used. Scale: 1 cm

Full size image

Discussion

SPLs belong to a plant specific transcription factors family with diverse roles in plant growth and development. In Arabidopsis, ten different SPLs are targets of miR156, and six of them (Group I) function to promote vegetative phase change (Fig. S5a, b) (Xu et al. 2016). SPL proteins are characterized by a conserved SBP domain at their C-terminals with three putative NLSs. The function of those NLSs has not been elucidated so far. Here, we showed that the NLSs of SPL9 are indispensable for its normal function. These NLSs affect SPL9 function by changing its nuclear localization in the plant cell. Typical NLSs contain one (monopartite) or two (bipartite) clusters of basic residues. Monopartite NLSs, exemplified by SV40 T-antigen, have a single cluster of four or five basic residues (Conti et al. 1998; Conti and Kuriyan 2000), whereas bipartite NLSs, exemplified by Xenopus laevis protein nucleoplasmin, have two clusters of basic residues separated by a flexible linker of 9–12 amino acids (Robbins et al. 1991; Dingwall and Laskey 1991; Jans et al. 2000). SPL9 contains a region rich in basic amino acids at the C terminal of the SBP domain (aa58-75 KRSCRRRLAGHNERRRKP), it contains three clusters with basic amino acids designated as BR1(aa58-59 KR), BR2 (aa62-64 RRR) and BR3 (aa71-74 RRRK) in the SBP domain (Fig. 3b). We identified three potential NLSs within this region using PSORT including a monopartite NLS 1 (aa71-74 RRRK) and a monopartite NLS 2 (aa72-75 RRKP) in BR3, and a bipartite NLS 1 (aa58-74 KRSCRRRLAGHNERRRK) consisting of BR1 and BR3 (Fig. 3b). In monopartite NLS 2, there is a conserved proline residue at aa75. Interestingly, this proline is only present in SPL proteins including SPL2, SPL9, SPL10, SPL11, SPL13, and SPL15 shown to play important roles in vegetative phase change (except for SPL6). Recent studies showed that the flexible linker for the bipartite NLSs can be up to 29 aa in Ty1 integrase (Lange et al. 2010), or to 7 aa in IRF3 (Zhu et al. 2015). This suggests that the traditional way of defining bipartite NLSs is too restrictive, and the linker size can vary depending on amino acid composition or conformation. According to this rule, the sequence of 62-74 RRRLAGHNERRRK in the SBP domain of SPL9 contains BR2 and BR3, and it can also possibly function as a bipartite NLS 2 with a 6 aa linker (Fig. 3b). Obviously, all above NLSs requires the presence of the BR3 region (72-74 RRK), as shown by the del6 mutation that the R72-to-Q substitution in BR3 completely disrupted the normal function of SPL9. Whether BR3 functions as a monopartite NLS, or together with BR1 or BR2 as a bipartite NLSs, needs further investigation.

Covalent modification of nuclear proteins, especially within and/or adjacent to the NLS, can affect the binding affinity of NLS for importin α, and change the nuclear import efficiency (Harreman et al. 2004; Birkenbihl et al. 2005). We identified two potential phosphorylation sites, Serine at aa60 (S60) and the Serine/Threonine at aa101 (S/T101), adjacent to the NLS in SPL9 and other SPL proteins using NetPhos (Fig. S5b). S60 is located in the conserved SBP domain, and the S/T101 is located ~ 20 aa downstream of NLS. The biological function of S60 phosphorylation has been investigated in a previous study (Birkenbihl et al. 2005). S60 phosphorylation in SPL3 doesn’t affect the nuclear import efficiency, but S60 phosphorylation negatively influences SPL8 nuclear import efficiency. In contrast, S60 phosphorylation in OsSPL14 (the SPL9 homolog in rice) changed its DNA binding activity in that the non-phosphorylated OsSPL14 preferentially binds to the GTAC motif, while the S60-phosphorylated OsSPL14 preferentially binds to the TGGGCC motif without altering its nuclear localization. S60 phosphorylation in OsSPL14 could be induced by fungus infection to balance between growth and immunity (Wang et al. 2018). Among different clades in SPL family, the Group II SPLs including SPL3, SPL4, and SPL5 contains a conserved putative phosphorylation site at S76, and Group I SPLs including SPL2, SPL10, and SPL11 contain a conserved potential phosphorylation site at S66 (Fig. S5a, b). Questions of whether S66, S76 and S/T101 site are subjected to phosphorylation and how this potential modification affects the function of different SPL genes will be an important task to deepen our understanding of how vegetative phase change is regulated in plants.