Stigma/style cell cycle inhibitor 1 (SCI1), a tissue-specific cell cycle regulator that controls upper pistil development

Authors

  • Henrique C. DePaoli,

    1. Departamento de Biologia, Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto 14040-901, Brazil
    2. Departamento de Genética, Faculdade de Medicina de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto 14049-900, Brazil
    Search for more papers by this author
  • Michael S. Brito,

    1. Departamento de Biologia, Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto 14040-901, Brazil
    2. Departamento de Genética, Faculdade de Medicina de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto 14049-900, Brazil
    Search for more papers by this author
  • Andréa C. Quiapim,

    1. Departamento de Biologia, Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto 14040-901, Brazil
    Search for more papers by this author
  • Simone P. Teixeira,

    1. Departamento de Ciências Farmacêuticas, Faculdade de Ciências Farmacêuticas de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto 14040-903, Brazil
    Search for more papers by this author
  • Gustavo H. Goldman,

    1. Departamento de Ciências Farmacêuticas, Faculdade de Ciências Farmacêuticas de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto 14040-903, Brazil
    2. Laboratório Nacional de Ciência e Tecnologia do Bioetanol (CTBE), Campinas 13083-970, Brazil
    Search for more papers by this author
  • Marcelo C. Dornelas,

    1. Departamento de Biologia Vegetal, Instituto de Biologia, Universidade de Campinas, Campinas 13083-970, Brazil
    Search for more papers by this author
  • Maria Helena S. Goldman

    1. Departamento de Biologia, Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto 14040-901, Brazil
    Search for more papers by this author

Author for correspondence:
Maria Helena S. Goldman
Tel: +55 16 3602 3702
Email: mgoldman@ffclrp.usp.br

Summary

  • A cDNA encoding a small lysine-rich protein of unknown function was identified in a tobacco (Nicotiana tabacum) stigma/style suppression subtractive hybridization cDNA library. After its characterization, the corresponding gene was designated stigma/style cell cycle inhibitor 1 (SCI1).
  • Fluorescence microscopy with an SCI1-GFP protein fusion demonstrated its nuclear localization, which was confined to the interchromatic region. Real-time RT-PCR and in situ hybridization experiments showed that SCI1 is stigma/style-specific and developmentally regulated.
  • SCI1 RNAi knockdown and overexpression plants had stigmas/styles with remarkably enlarged and reduced areas, respectively, which was attributable to differences in cell numbers. These results indicate that SCI1 is a tissue-specific negative cell cycle regulator.
  • The differences in cell division had an effect on the timing of the differentiation of the stigmatic papillar cells, suggesting that their differentiation is coupled to stigma cell divisions. This is consistent with a role for SCI1 in triggering differentiation through cell proliferation control. Our results revealed that SCI1 is a novel tissue-specific gene that controls cell proliferation/differentiation, probably as a component of a developmental signal transduction pathway.

Introduction

The efficiency of sexual plant reproduction depends on appropriate development of the reproductive organs. It also depends on transfer of pollen grains to a receptive stigma and on compatible pollen–pistil interactions, which take place in specialized tissues of the pistil (mainly the stigmatic secretory zone (SSZ) and stylar transmitting tissue (STT)). To fulfill their reproductive functions, the pistil cells should proliferate and differentiate properly, that is, promote pistil development and determine final organ size and form.

In plants, changes in organ identity genes (Ingram et al., 1995), altered phytohormone signaling (Nemhauser et al., 2000; Krizek, 2009) and direct alterations in cell division control (Mizukami & Fischer, 2000; Mizukami, 2001) are good examples of how cell proliferation/differentiation processes can be disrupted. Among these examples, MEGAINTEGUMENTA/AUXIN RESPONSE FACTOR 2, BIG BROTHER, DA1 and DA1-related (DAR1) genes are involved in organ size control through the regulation of cell number (Disch et al., 2006; Schruff et al., 2006; Li et al., 2008; for a review see Krizek, 2009). Genes such as HECATE, SPATULA and STYLISH/SHORT INTERNODES have been shown to be important for adequate pistil formation (for a review, see Crawford & Yanofsky, 2008). Although several genes have been identified as important for pistil development, as well as for the development of other plant organs, a key missing link in our understanding is how the molecular machinery of cell proliferation senses the developmental signals and coordinates organ growth.

The cell cycle machinery is governed by cyclin-dependent kinases (CDKs). CDKs complex with cyclins to gain protein kinase activity and phosphorylate targets that regulate entry to and progression through the cell cycle (Inzé & De Veylder, 2006). In plants, the cell cycle machinery includes members involved in cell cycle regulation in an organogenesis-independent manner, which means that their functional alteration affects cell number and/or size but not morphogenesis, uncoupling cell proliferation and development (Hemerly et al., 1995; Cockcroft et al., 2000; Zhou et al., 2003). However, there are members involved in cell cycle regulation in an organogenesis-dependent manner, as targets for coupling cell proliferation and development (Yamaguchi et al., 2003; Boudolf et al., 2004; Ramirez-Parra et al., 2005; del Pozo et al., 2006). Among the latter, the group of CDK inhibitors, which bind to CDK/cyclin complexes, specifically inhibiting their activities (Pei & Xiong, 2005; Verkest et al., 2005), includes members that have been proposed to function as integrators of development through cell division control (Himanen et al., 2002).

Plant CDK inhibitors are represented by the CKI (CDKInhibitors)/KRP(KIP-related proteins) family and by the more recently described plant-specific SIM family, which includes the SIAMESE and SIAMESE-related (SMR) proteins (Churchman et al., 2006; Peres et al., 2007). In Arabidopsis, the CKI/KRP family comprises seven members (KRP1–7) (De Veylder et al., 2001; for a review, see Verkest et al., 2005). In tobacco (Nicotiana tabacum), two CKI/KRPs have been identified, NtKIS1a/b and NtKIS2 (Jasinski et al., 2002a, 2003). All plant CKI/KRPs share a restricted C-terminal region of similarity (for reviews, see Verkest et al., 2005; Wang et al., 2008). The SIM/SMR family has five short conserved domains, among which domain 4 shares a high similarity with a CKI/KRP C-terminal region (Churchman et al., 2006). All plant CDK inhibitors are expressed in several different tissues, and overexpression of CKI/KRP genes or SIM impairs the correct development of several plant organs, leading to overall reduced growth (Wang et al., 2000; De Veylder et al., 2001; Jasinski et al., 2002b; Zhou et al., 2002; Churchman et al., 2006; Ren et al., 2008).

It is clear that integrated cell proliferation and differentiation are necessary for proper organ development. The CDK inhibitors are the main group of genes involved in organ development through control of the basic cell cycle machinery (CDKs and cyclins). Identifying the molecular bridges that interconnect cell differentiation control and development through cell cycle regulation is of primary importance. Here, we report the characterization of a novel stigma/style-specific gene, which inhibits cell proliferation and coordinately controls upper pistil differentiation; for this reason it was named stigma/style cell cycle inhibitor 1 (SCI1). The SCI1 protein is distinct from the two families of CDK inhibitors that have already been described in plants: KRP and SIM. SCI1 is a new type of cell cycle regulator that acts in a tissue-specific manner and links cell proliferation/differentiation control to development, probably acting as a component of a signal transduction pathway. The finding of this tissue-specific cell proliferation regulator suggests that similar proteins occur in other plant tissues.

Materials and Methods

Plant material

Wild-type (Nicotiana tabacum L. cv Petit Havana SR-1) and transgenic tobacco plants were grown under standard greenhouse conditions. Samples for RNA extraction were collected, immediately frozen in liquid nitrogen and stored at −70°C. Samples were excised from tobacco flowers at the defined developmental stages (as indicated in each figure or part of the text), as described previously (Koltunow et al., 1990).

Nucleic acid manipulations and gene constructions

All the basic nucleic acid manipulations were as described in Sambrook & Russell (2001). Primer sequences are given in Table 1. The SCI1-ORF fragment was amplified from stigma/style cDNAs (as prepared for qRT-PCR) using the primers Ct11attB1FW/Ct11attB2RV. The product was re-amplified with the primers BP1/BP2 and then recombined into the pDONR201 vector (Gateway; Invitrogen, USA), generating the SCI1-ORF, which was completely sequenced. The SCI1-ORF, TOBSH1-002E06 and TOB068F10 cDNA sequences were deposited in GenBank under accession numbers GQ272329, GQ272330 and GQ272331, respectively.

Table 1.   Oligonucleotides used for PCR amplifications, gene cloning and expression analyses
Primer namePrimer sequence (5′–3′)
  1. aattB sites are in lowercase.

  2. bThe AvaII restriction site (bold underlined) created by the excision of the stop codon.

  3. cPrimers used to quantify stigma/style cell cycle inhibitor 1 (SCI1) gene expression in wild-type and transgenic plant tissues by qRT-PCR.

  4. dPrimers used to quantify β-actin gene expression in wild-type and transgenic plant tissues by qRT-PCR.

  5. *Primers labeled with 6-carboxyfluroscein (FAM), incorporated at the italic lowercase underlined nucleotide. Italic lowercase letters (at the 5′ end) represent the nucleotides added to the sequence according to the LUX system for real-time PCR. These primers were also used, after 6 months of storage, with the Power SYBR Green Master Mix (Applied Biosystems), providing reproducible results.

aBP1ggggacaagtttgtacaaaaaagcaggcttc
aBP2ggggaccactttgtacaagaaagctgggtc
aCt11attB1FWgcaggcttcaccATGGGGAGCGATAAGAAGAC
aCt11attB2RVaagctgggtcTTACTTTTTGATATTCCAAGCGTG
a,bCt11attB2GFPRWaagctgggtcCTTTTTGATATTCCAAGCGTG
c,*Ct11New_66FLcgatttGCCACAAACATAAGGATAAATcG
cCt11New_66FL/112RUCAGCCACGTAGCAAATTCATTGTT
dRT-ACTHM-FWCATGCCAACCATCACACCAGT
d,*RT-ACTHM-RVcacagcGATGATGCTCCAAGGGCTGtG

To obtain the 35Spro:SCI1-GFP construction, the SCI1-ORF was amplified using the primers Ct11attB1FW/Ct11attB2GFPRW and BP1/BP2, and recombined into the pDONR201 vector and then into the pK7FWG2 (Karimi et al., 2002), using BP and LR clonases (Invitrogen, USA), respectively. To obtain the SCI1 RNAi and overexpression constructions, 300 μg of DNA from the pDONR201-SCI1-ORF was recombined into pK7GWIWG2(I) and pK7WG2 (Karimi et al., 2002), generating the SCI1Ri and SCI1OE constructs. Each construction was sequenced to confirm the correct recombination and absence of mutations, and then individually introduced into the Agrobacterium tumefaciens C58C1RifR (pGV2260).

RNA extractions

Total RNA was extracted from frozen samples, as described by Dean et al. (1985). For the comparisons of transgenic and wild-type plants, four samples of stigmas/styles at stage 4 of flower development were collected from each plant and total RNA was extracted using the Plant RNA Purification Reagent (Invitrogen, USA). The RNA samples were quantified by measuring optical density at 260 and 280 nm and their integrity was evaluated using gel electrophoresis.

Quantitative RT-PCR and in situ hybridization

qRT-PCR was performed as described previously (Quiapim et al., 2009), except for transgenic plants, for which all of the RNA from each sample was subjected to DNAse treatment and cDNA synthesis, with the final volume of cDNA adjusted to 35 μl. The qPCR reactions were performed using either the LUX system (Lux System, Invitrogen, USA) as previously described (Quiapim et al., 2009) or the Power SYBR Green PCR Master Mix (Applied Biosystems, Warrington, UK), using the primers shown in Table 1. As the results obtained for the expression studies in at least two biological replicates of wild-type plants were very similar with the two systems, the comparisons between wild-type and transgenic plants were made using only the SYBR Green system. All analyses were performed in triplicate, in a final volume of 10 μl. The PCR conditions were an initial step at 50°C for 2 min, followed by 10 min at 95°C, 40 cycles of 15 s at 95°C and 1 min at 60°C. When using the SYBR Green system a final dissociation curve was added, with the temperature gradually rising from 60°C to 95°C. The expression level in arbitrary units (au) corresponds to the calculated log of the number of copies of the SCI1 gene divided by the calculated log of the number of copies of β-actin, which was used as a reference gene (Quiapim et al., 2009).

For in situ hybridization, stigmas/styles at defined stages of flower development (Koltunow et al., 1990) were analyzed, as described by Jackson (1991). The sense and antisense riboprobes were prepared by in vitro transcription from the TOBSH1-002E06 clone, using the DIG RNA Labeling Kit SP6/T7 (Roche, Germany). Digoxigenin (DIG)-labeled probes were detected using anti-DIG antibody conjugated with alkaline phosphatase (AP) (Roche, Germany) and the one-step nitro blue tetrazolium chloride and 5-bromo-4-chloro-3 indolyl phosphate (NBT/BCIP) detection mixture (Thermo Fisher Scientific, Rockford, IL, USA).

Plant transformation

For transient expression assays, plants c. 6 wk old were infiltrated with the A. tumefaciens culture containing the appropriate construct, as described by Sawers et al. (2006), and were analyzed 2–5 d after infiltration.

To prepare stable transgenic plants, tobacco leaf discs were co-cultivated with the A. tumefaciens culture containing the appropriate construction and regenerated in vitro, according to a previously described protocol (Brasileiro & Carneiro, 1998). The primary transgenic plants (T0 generation) were analyzed in the qRT-PCR and light microscopy experiments. The next generation (T1) was produced by controlled self-pollination; the seeds were sown in MS medium (Sigma-Aldrich) containing kanamycin; resistant seedlings were transferred to soil and the first flowering used for the scanning electron microscopy (SEM) analyzes.

Microscopy experiments

For SCI1-GFP protein subcellular localization, leaves were infiltrated as described by Sawers et al. (2006), cut into small pieces, fixed in 3.7% formaldehyde, 50 mM NaH2PO4 and 0.2% Triton X-100 for 30 min, stained with 4,6′-diamidino-2-phenylindole (DAPI) for 15 min (1.25 μg ml−1 in phosphate-buffered saline (PBS)) in darkness, washed in PBS and mounted on slides with water, just before analyses. GFP and DAPI fluorescences were imaged under a Leica TCS SP2 confocal laser scanning microscope (Leica Microsystems, Mannhein, Germany). The individual stacks were assembled with Adobe Photoshop CS (Adobe Systems, San Jose, CA, USA). At least three additional analyses without fixed material were performed to confirm that formaldehyde self-fluorescence was not interfering with the results.

For light microscopy analysis, stigmas/styles and anthers were dissected from floral buds at the relevant developmental stages, fixed and embedded in paraffin as described for in situ hybridization. The embedded material was sliced into 8-μm sections, mounted on slides and stained with 0.05% toluidine blue in phosphate buffer at pH 4.4 (O’Brien et al., 1964). Photographs were taken using a Leica DM4500B microscope equipped with a Leica DFC320 camera. The tissues analyses were performed using Leica Q-Win Plus v3.4.0 software (Leica Microsystems) and the statistical analyses were performed using SigmaStat v3.5 software (Aspire Software International, Ashburn, VA, USA).

For SEM, the floral buds from wild-type and T1 transgenic plants were collected from stages −4 to 1. One anther of each floral bud was collected for light microscopy analysis. The rest of the floral bud (pistil and four anthers) was fixed for 48 h in formaldehyde, acetic acid and 50% ethanol (1 : 1 : 18), dehydrated in an ethanol series, subjected to critical-point drying, and coated with 30 nm of gold. Preparations were examined using a Jeol JSM-5200 scanning electron microscope (Jeol Ltd.,Tokyo, Japan). At least two biological replicates of each stage were analyzed from the transgenic and wild-type plants.

Results

Identification of the SCI1 gene and analysis of the subcellular localization of its protein

Two suppression subtractive hybridization cDNA libraries allowed the identification of several genes that are specifically or preferentially expressed in the pistil of Nicotiana tabacum (H. C. DePaoli et al., unpublished). Clone TOBSH1-002E06 was chosen for further characterization. As this clone was a partial cDNA, we performed a BLAST search in the TOBEST database (Quiapim et al., 2009) and found two overlapping cDNA clones (TOB048C05 and TOB068F10). Clone TOB068F10 was completely sequenced and found to contain part of the coding region, the 3′-UTR and the polyA tail (Supporting Information Fig. S1). The corresponding gene was designated stigma/style cell cycle inhibitor 1 (SCI1), based on the results presented in this paper.

A homology search using TBLASTX (Altschul et al., 1997) showed that the encoded protein is similar to proteins of unknown function from potato (Solanum tuberosum; AC233353), tomato (Solanum lycopersicum; AC212309), barley (Hordeum vulgare; AK252156), corn (Zea mays; EU958997), rice (Oryza sativa; NM_001052562) and grapevine (Vitis vinifera; AM424297). The putative homologous sequence in the Arabidopsis genome is At1g79200 (NM_106571), which shares 65% identity and 86% similarity with SCI1. Alignment of these proteins allowed the recognition of three domains: (1) a short conserved sequence at the N-terminus; (2) an intermediary lysine-rich domain with low sequence conservation, and (3) a highly conserved C-terminal domain (Fig. S2).

A detailed in silico analysis using the complete protein sequence of 154 amino acids identified an N-terminal lysine-rich domain (LRD) (MotifScan, e-value = 6.5e-07), with lysine residues predicted to be Nε-acetylated, a nuclear localization signal (NLS) (PSORT, 0.992 affirmative; PredictNLS), two cyclin interaction domains (ELM prediction) and 15 predicted phosphorylation sites (NetPhos, ≥ 96%) (Fig. S1b). The identified cyclin interaction domains, consensus [RK].L.{0,1}[FYLIVMP], are present in a wide range of cyclin/CDK interacting proteins (Takeda et al., 2001) that commonly share different phosphorylation sites targeted by different signaling molecules such as casein kinase 1, casein kinase 2 and Mitogen-activated protein kinase (MAPKs), all three present in SCI1 (NetPhosK; ELM prediction; Table S1).

To study the SCI1 subcellular localization, we fused the green fluorescent protein to the SCI1 C-terminus (35Spro:SCI1-GFP). Fluorescence microscopy revealed that the SCI1-GFP protein was exclusively localized in the nucleus (Fig. 1a,b), in contrast to the 35Spro:GFP construct, which had a uniform distribution of GFP in both the nucleus and the cytoplasm (data not shown). Interestingly, the SCI1-GFP fluorescence was present as spots within the nucleoplasm, where DAPI was not detected, corresponding to the interchromatin and nucleolus (Fig. 1e–h). We found that the reported patterns of the fusion protein occurred at both low and high expression levels, indicating that its level of expression does not influence its localization. In conclusion, SCI1 is a nuclear protein localized in the interchromatic regions.

Figure 1.

Stigma/style cell cycle inhibitor 1 (SCI1) subcellular localization and gene expression analysis in Nicotiana tabacum. (a, b) 35Spro:SCI1-GFP expression; (c, d) 4,6′-diamidino-2-phenylindole (DAPI) staining; (e, f) superposition of upper and middle upper panels plus chloroplast (red) fluorescence (merge); (g, h) superposition of 10 (g) and 12 (h) stacks of the GFP, DAPI and chloroplast signals through a cell slice (‘3D’ merge). Bars, 4 μm. cl, chloroplast; arrow, interchromatin space; n, nucleus; nu, nucleolus. (i) SCI1 expression level in the different tobacco organs as determined by qRT-PCR, using the β-actin reference gene. (j) Expression level in stigmas/styles during the 12 stages of tobacco flower development (Koltunow et al., 1990). (k, l) In situ hybridization on longitudinal sections of stage 4 stigmas/styles, using sense (k) and antisense (l) probes prepared from clone TOBSH1-002E06. The magnification factor is × 40; p, parenchyma; ssz, stigmatic secretory zone; stt, stylar transmitting tissue; vb, vascular bundle.

The SCI1 gene is exclusively expressed in stigma/style specialized tissues related to plant reproduction

To determine the expression pattern of the SCI1 gene, we performed qRT-PCR experiments, using RNA from roots, stems, leaves, sepals, petals, stamens, stigmas/styles and ovaries. The expression of SCI1 was at least 34-fold higher in the stigma/style than in the other organs (Fig. 1i). We also evaluated the temporal expression pattern of the SCI1 gene in the stigma/style during the 12 tobacco flower developmental stages (Koltunow et al., 1990). At stage 1, the pistil is completely differentiated; thereafter, it will grow and prepare itself for receptivity until stage 12, when anthesis occurs. Fig. 1(j) shows that SCI1 expression levels were high at the early stages (stages 1 to 6) and decreased towards anthesis (stages 7 to 12). Its high expression during the earliest developmental stages suggests that the SCI1 gene product is involved in events taking place early rather than late in pistil development.

In situ hybridization experiments were performed to determine which cell types transcribe the SCI1 gene (Fig. 1k,l). At stage 4, we observed a hybridization signal in the stigmatic secretory zone, including the papillar cells, and in the stylar transmitting tissue (Fig. 1l). The SCI1 transcript was expressed in stigma/style specialized tissues at initial stages of flower development; no signal was detected in stigmas/styles at stages 7 and 10 (data not shown). Although these tissues are intrinsically related to pollen–pistil interactions, most of the genes related to these interactions have a higher expression level at later stages of flower development (Goldman et al., 1992; Wang et al., 1993; McClure et al., 1999). Thus, SCI1 is a developmentally regulated stigma/style-specific gene; its temporal and spatial expression patterns suggest involvement of SCI1 in the morphological and/or physiological development of the pistil, rather than contributing directly to pollen–pistil interactions.

The SCI1 gene product controls stigma/style size

To unravel the SCI1 protein function, we generated tobacco transgenic plants by either silencing or overexpressing the SCI1 gene. We obtained 13 independent transgenic plants possessing the RNAi construct (SCI1Ri) and nine independent transgenic plants harboring the overexpression cassette (SCI1OE). RNAi-mediated silencing efficiently reduced transcript levels, as monitored by qRT-PCR (Fig. 2a). The plants with the lowest transcript levels had enlarged stigma (Fig. 3B) and considerable style elongation (Fig. 3A). This phenotype was observed in nine out of 13 RNAi transgenic plants and was visible when the SCI1 transcript levels were below 27.8% of that of the wild-type plants (Fig. 2a). There was an overall correspondence between the degree of alteration in the phenotypes and the SCI1 mRNA levels, with the plant SCI1Ri 14.2 (8.4% of SCI1 mRNA) exhibiting the most noticeable phenotype (Fig. 3A). The differences between the wild-type and RNAi phenotypes were more evident at stage 10 of flower development, when the stigma/style remained enclosed by the corolla limb in wild-type plants and protruded in the RNAi plants (Fig. 3A). In accordance with the phenotype generated by silencing SCI1, the SCI1OE plants revealed an opposite modification in the stigma, that is, a smaller stigma (Fig. 3B). Among nine independent overexpression plants, five showed the same phenotype, which was associated with the highest SCI1 transcript levels (at least 48.8-fold higher than that of wild-type plants; Fig. 2b). However, we did not observe alterations in the style length of the SCI1OE plants, in which the stigmas were positioned at the same height as the tops of the stamens and petals at anthesis. The plant SCI1OE 1.1 showed an RNAi-like phenotype and, accordingly, the SCI1 mRNA level in this transgenic plant was only 17.7% of the wild-type level (Fig. 2b), a co-suppression effect. No other organ showed a macroscopically altered phenotype in comparisons of wild-type and transgenic plants. Together, these results strongly imply involvement of the SCI1 protein in controlling stigma/style size.

Figure 2.

Expression level of the stigma/style cell cycle inhibitor 1 (SCI1) gene in stage 4 stigmas/styles of SCI1Ri and SCI1OE transgenic Nicotiana tabacum plants, determined by qRT-PCR, using the β-actin gene as reference. The hatched bars represent the transgenic plants that showed altered phenotypes with elongated styles and a larger stigma surface; the cross-hatched bars correspond to plants with a smaller stigma surface; open bars represent wild-type-like phenotypes; and solid bars represent the wild type. The expression levels of independent (a) RNAi and (b) overexpression (OE) transgenic plants are shown.

Figure 3.

Analyses of the stigma/style cell cycle inhibitor 1 (SCI1) transgenic Nicotiana tabacum plant phenotypes. (A) Flowers of independent RNAi transformants showing increased style length at stages 12, 10 and 9 (from left to right) of flower development. The inset marked ‘*’ shows a top view of a stage 10 wild-type flower (WT), showing the stigma/style enclosed by the corolla. Bars, 10 mm. The percentages correspond to the transcript level of the SCI1 gene in comparison with the wild-type, which was considered to be 100%. (B) Top view of the stage 12 stigma showing (a) the decreased size in overexpression plants (SCI1OE 11.1) and (c) the increased size in RNAi plants (SCI1Ri14.2), compared with (b) the wild-type (WT). Bar, 1 mm. (C) Longitudinal sections of stage 11 stigmas/styles. (a) Longitudinal section scheme representing the region where the stigmatic secretory zone (SSZ) cell layers were counted (≡, as indicated in the light grey area, left-most section) (D; Supporting Information Table S2). The SSZ/stylar transmitting tissue (STT) intersection was defined by a straight line at the abaxial curves of the stigmatic umbrella-like structure (green dotted line). The SSZ area measured in the longitudinal sections is shown in light gray (Table S2; longitudinal plane Ec1). The square indicates the region in which the SSZ cell numbers were counted in an area of 20 000 μm2. The STT is represented in dark gray and its thickness was measured at 400 μm below the SSZ/STT intersection. The rectangle represents the region shown in (Ea,b) for each plant and in which the STT thickness and parenchyma cell number were evaluated. (b, c, d) Longitudinal sections of stage 11 stigmas/styles stained with 0.05% toluidine blue for (b) the overexpression plant SCI1OE 3.1.1, (c) the WT and (d) the RNAi plant SCI1Ri 14.2. Bar, 500 μm. The blue arrow indicates the deeper carpel junction in the SCI1OE 3.1.1 plant. (D) Longitudinal sections of stage 11 stigmas/styles at a higher magnification (× 200), demonstrating the altered number of cells in the SSZ. (E) Histological sections of stage 11 styles. Longitudinal sections of the SCI1OE3.1.1, WT and SCI1Ri14.2 plants are shown, illustrating the thickness differences in (a) the STT and (b) the parenchyma tissue, through the longitudinal plane indicated by ‘2’ in (c). (c) Transversal section of the WT style showing the two different longitudinal planes where the SSZ area, cell layers and cell number, as well as the STT thickness, were determined (‘1’) and where the style thickness and parenchyma cell number were determined (‘2’). See also Fig. 4 and Table S2. The longitudinal plane ‘2’ corresponds to the first cell layer behind the vascular bundle. Bars: (Ea,b) 50 μm; (Ec) 500 μm. p, parenchyma; stt, stylar transmitting tissue; ssz, stigmatic secretory zone; vb, vascular bundle (black arrow).

The SCI1 protein is a negative regulator of cell proliferation

To elucidate SCI1 function and to determine if SCI1 affects cell proliferation or cell enlargement or both, we prepared longitudinal sections of stage 11 stigmas/styles from SCI1Ri 14.2, SCI1OE 3.1.1 and wild-type plants. Fig. 3(C,D) shows that the cell number was significantly increased (Student’s t-test; < 0.05) as a consequence of reduced SCI1 mRNA levels, whereas SCI1 overexpression resulted in a decreased cell number (< 0.05) in the SSZ. The SCI1Ri 14.2 transgenic plant showed an SSZ mean area of 367 355.3 μm2, 48% larger than the mean area (248 272 μm2) of the wild-type plant (Figs 3, 4; Table S2). By contrast, the SCI1OE 3.1.1 transgenic plant had an SSZ mean area of 191 381.3 μm2, 23% smaller than that of the wild-type plant (Figs 3, 4; Table S2). Accordingly, the number of cell layers at the top of the SSZ was significantly reduced to approximately eight in SCI1OE 3.1.1, while the wild-type showed 13 cell layers and the SCI1Ri 14.2 RNAi plant an increase to approximately 18 (Fig. 3Ca,D). In addition to the SSZ area and cell layer data, we also counted the number of cells in a defined area of 20 000 μm2 (Fig. 3Ca). The cell number of the SCI1OE 3.1.1 overexpression plant was lower than, but not significantly different from, that of the wild-type and SCI1Ri 14.2 plant, which had very similar values (Fig. 4; Table S2). In SCI1OE plants, we did not observe the phenomenon called ‘compensation’, in which cell size increases to balance the decrease in cell number. However, the SCI1OE 3.1.1 SSZ cells were more circular in shape and the intercellular spaces were slightly increased, consistent with a lower cell density. In contrast, the cells of the SSZ seemed smaller in RNAi plants, probably because of space-limited growth and/or a rapid re-entry into the cell cycle. In conclusion, the cells of the transgenic and wild-type plants were approximately the same size.

Figure 4.

Stigma/style tissue analyses in transgenic (SCI1OE, cross-hatched bars; SCI1Ri, hatched bars) and wild-type plants (closed bars) of Nicotiana tabacum. Longitudinal sections of stigmas/styles from each transgenic and wild-type plant were analyzed as described in Fig. 3(Ca) for stigmatic secretory zone (SSZ) cell layers, SSZ area, cell size (represented as SSZ cell number in a defined area of 20 000 μm2), style thickness (μm), stylar transmitting tissue (STT) thickness (μm) and parenchyma thickness (= 3 for each plant shown). The values represent percentages of the wild-type, which was considered to be 100%. The raw data can be found in Supporting Information Table S2. *Significantly different from the wild-type by Student’s t-test (< 0.05).

We also searched for alterations in STT tissue. As the cells of the STT are more closely packed together, counting them individually is very difficult. Thus, we measured the STT thickness in transgenic and wild-type plants (Figs 3Ea, 4; Table S2) in longitudinal sections of stigma/style at 400 μm below the SSZ/STT intersection (for a definition, see Fig. 3Ca,E ). Consistently, the STT thickness was decreased in SCI1OE 3.1.1 plants (Student’s t-test; < 0.05) and increased in SCI1Ri 14.2 plants (< 0.05) compared with the wild-type (Fig. 4), indicating a lower and a higher number of cells, respectively. It is noteworthy that the RNAi plant SCI1Ri 14.2 also had a higher cell number in the parenchyma tissue, compared with wild-type and SCI1OE 3.1.1 plants (Figs 3Eb, 4). As SCI1 is not expressed in the parenchyma cells (Fig. 1l), our results suggest that the cells of the SSZ and STT produce signals capable of influencing cell division in adjacent tissues, maintaining a certain proportionality; the more cells from the specialized tissues (SSZ/STT), the more supporting (parenchyma) tissue. Taken together, our results reveal the involvement of the SCI1 protein in inhibiting the cell cycle in a tissue-specific manner.

SCI1 is expressed at the very early stages of stigma/style development

In order to better correlate SCI1 expression/function with the early events occurring during stigma/style differentiation (stages −4 to −1; Goldman et al., 1994; Wolters-Arts et al., 1996), we performed an SEM analysis on wild-type flower buds from stages −4 to 1, as well as collected stigmas/styles from the same developmental stages for qRT-PCR experiments. In addition to measuring the floral bud length, we also considered floral and anther morphological markers that have been previously described (Koltunow et al., 1990) to reliably distinguish each developmental stage (Fig. S3).

The tobacco pistil is formed by the fusion of two carpels that give rise to the stigma around stage −4 of flower development (Fig. 5Af). Stigmatic tissue differentiation starts from the region of carpel fusion outwards (Fig. 5Ak), where at stage −3 the stigma surface shows a rough aspect, contrasting with the smooth carpel epidermis (Fig. 5Ag). The style begins to elongate (Fig. 5Ab) and the upper pistil increases considerably in size, with a visible differentiation of papillar cells at the stigma surface (Fig. 5Ag,l). At stage −2, the style is definitely elongated (Fig. 5Ac), and the papillar cells turn into finger-like structures (Fig. 5Am) and become more evident at the central stigma area (Fig. 5Ah). In an upper view of the stigma, we clearly see the papillar cells developing from the central area of the carpel edges and bordering the carpel junction towards the outside (Fig. 5Bb,c, yellow). The epidermis is still visible at the central borders of the pistil at this stage (Fig. 5Ah,Bc, blue) and the carpel tip is almost straight, as a consequence of the growth pressure produced by SSZ cell proliferation. Stage −1 is characterized by an extended style (Fig. 5Ad) and a clear increase in stigma size (Fig. 5Ai), with the upper carpel epidermis and parenchyma starting to bend backward, creating the abaxial part of the stigma (Fig. 5Bd). Papilla differentiation occurs throughout the stigma surface. At stage 1, the SSZ is fully differentiated; the upper carpel epidermis and the parenchyma have bent completely, forming an umbrella-like structure in a side view (Fig. 5Ae).

Figure 5.

Very early upper pistil development and differentiation in the wild-type Nicotiana tabacum plant. (A) Scanning electron microscopy (SEM) of pistils at very early stages of flower development. (a–e) Flower buds at the indicated stages (stage −4 to stage 1, respectively, a to e) had the sepals and petals removed to allow visualization of the pistils. (f–j) Stigmatic secretory zone formation and development, and (k–o) epidermis differentiation in papillar cells. Bars: (a–e) 500 μm; (f–j) 200 μm; (k–o) 50 μm. (B) A scheme to illustrate the upper pistil development and stigmatic secretory zone formation, from stages −4 to 1 (a to e, respectively) of flower development. yellow, stigmatic tissue; blue, carpel epidermis. Bars, 500 μm. (C) Expression level of the stigma/style cell cycle inhibitor 1 (SCI1) gene determined by qRT-PCR in wild-type stigmas/styles from stages −4 to 1 of flower development. The graphic is representative of the results obtained in two biological replicates.

As shown in Fig. 5(C), the SCI1 gene is highly expressed at very early stages of stigma/style development, with the highest expression level observed at stage −3, when SSZ starts its differentiation. The SCI1 transcript level decreases at later stages (−2, −1 and 1) during which the stigmatic tissue is proliferating and continuously differentiating. This expression pattern is consistent with a role in controlling cell proliferation/differentiation in the upper pistil.

SCI1 influences stigma/style developmental timing through cell proliferation/differentiation control

To investigate how SCI1 affects stigma size and style elongation, we used SEM to compare wild-type and transgenic stigmas/styles at very early stages of development (stage −4 to stage −1). We analyzed the transgenic plants with the most striking phenotypes: RNAi SCI1Ri 14.2 and overexpression SCI1OE 3.1.1 (Fig. 6). Stage −4 pistils of the RNAi SCI1Ri 14.2 plant showed features similar to those of the wild-type pistil at stages −4/−3 of flower development (Fig. 6Ba,d). During the first steps of stigma/style formation, the SCI1Ri 14.2 plant clearly exhibited faster growth of the stigma through stages −3/−2, with the carpel epidermis bending backward, an event observed at stage −1 of the wild-type plant (Figs 6Aa and 5Ad, respectively). Stigma differentiation was also advanced in the SCI1Ri 14.2 plant, occurring as soon as the cell number increased (Fig. 6Cb), showing that cell differentiation is coupled to proliferation in SCI1 silencing plants. Additionally, pistils of the SCI1Ri 14.2 plant attained the umbrella-like structure at stage −2 (Fig. 6Ab,Bb), a phenotype observed only at stage 1 of the wild-type plant (Figs 5Ae, 6Af). Similarly, style elongation followed the accelerated growth of the stigma, with the style becoming longer and carrying the stigma surface to a higher position in relation to the anthers as early as stage 1 (Fig. 6Ac,f), showing that the stimuli for style elongation start at very early stages of its development. The developmental acceleration observed in the SCI1Ri 14.2 plant suggests that SCI1 is a key regulator in stigma/style developmental control. The size of each cell in the RNAi stigma was similar to that in the wild-type stigma, confirming that the larger area observed must be a result of a higher number of cells in the RNAi stigma. These results show that SCI1 is involved in final stigma size and style length determination through the control of cell divisions, beginning at the initial stages of organ formation.

Figure 6.

Scanning electron microscopy (SEM) of the transgenic and wild-type Nicotiana tabacum plants. (A) The early style elongation of (a–c) the SCI1Ri14.2 plant compared with (d–f) the wild-type plant and (g–i) the SCI1OE3.1.1 plant. (a, d, g) Stage –3; (b, e, h) stage –2; (c, f, i) stage +1. Note that, at stage 1, the SCI1Ri14.2 stigma is already positioned above the anthers. Style elongation is delayed in the SCI1OE3.1.1 plant. (B) The premature enlargement of the upper pistil part of (a–c) the SCI1Ri14.2 transgenic plant, compared with (d–f) the wild-type plant, in contrast to the developmental delay observed for (g–i) the SCI1OE3.1.1 plant, at stages (a, d, g) −4, (b, e, h) −3 and (c, f , i) −2 of flower development. (C) The accumulation of cells at the stigmatic secretory zone (SSZ) and the differentiation of the papillar cells are initiated at stages −4/−3 in (a–c) the SCI1Ri14.2 transgenic plant, at stage −3 in (d–f) the wild-type plant and at stage −2 in (g–i) the SCI1OE3.1.1 plant. (a, d, g) Stage –4; (b, e, h) stage –3; (c, f, i) stage –2. Bars: (Aa–i) 500 μm; (Ba–i) 100 μm; (Ca–i) 50 μm.

The upper pistils of the overexpression SCI1OE 3.1.1 plant showed a phenotype comparable to that of the wild-type pistils at stage −4 (Fig. 6Bg,d). However, the upper pistil of this transgenic plant showed delayed development from stage −4 to stage −3, with a minimal differentiation of papillar cells (Fig. 6Ch). The typical stigma surface characteristics, observed in stage −3 pistils of wild-type plants, were visible only at stage −2 of the SCI1OE 3.1.1 pistils (Fig. 6Ci). It is interesting to observe that SSZ differentiation is strictly linked to cell proliferation. Pistils at stages −1 and 1 of the overexpression plant were morphologically and developmentally similar to those at stages −2 and −1 of the wild-type plant, respectively. Also, the SCI1OE 3.1.1 style was visible only at stage −2 (Fig. 6Ah), remaining shorter than the wild-type style at stage 1 (Fig. 6Ai). Consistently, stigmas of the SCI1OE 3.1.1 plant were smaller than those of the wild-type plant at stages −3, −2, −1 and 1. In summary, early pistils of the overexpression SCI1OE 3.1.1 plant showed a growth and differentiation delay of nearly one stage compared with wild-type pistils.

Anther development was affected in the SCI1OE 3.1.1 plant, as a result of the ectopic expression of the SCI1 gene (stage −3; Fig. 6Ag), although at stage 1 anthers had already recovered their normal size (Fig. 6Ai). At stage −3, the SCI1Ri 14.2 and the wild-type plants exhibited similar anther size (Fig. 6Aa,d). No other vegetative or floral organs were macroscopically altered. We propose that SCI1 inhibits cell division at very early stages of flower development in a tissue-specific manner and is sufficient to modulate the developmental timing of the stigma/style. These observations suggest that SCI1 may be part of a developmental pathway that influences the final organ size and its number of cells.

Discussion

SCI1 is a tissue-specific regulator of cell proliferation/differentiation that influences pistil development

We have identified a gene encoding a lysine-rich protein specifically expressed in stigma/style specialized tissues during early stages of development. The SCI1-GFP fusion protein was localized in both the nucleolus and nuclear bodies, always in the interchromatin region. It has already been shown that signaling molecules, such as MAPKs, enter the nucleus at specific interchromatin domains in plant differentiation and proliferation processes (Coronado et al., 2002). Despite the well-known involvement of the nucleolus in ribosomal RNA (rRNA) and small nucleolar RNA (snoRNA) biogenesis, it has recently been suggested to be a multifunctional nuclear structure with various other functions, for example in cell cycle control (Zimber et al., 2004). Many cell cycle regulators, such as cyclins and CDKs, are located at the nucleolus and nuclear bodies (Boruc et al., 2010). Compartmentalization to nuclear bodies was also described for AtKRP1 (Zhou et al., 2006), NtKIS1a (Le Foll et al., 2008) and transducin/WD-40 repeat proteins which interact with cell cycle regulators, which are sequestered in the nucleolus (Visintin & Amon, 2000; Brown et al., 2005). Although the SCI1 localization pattern is unique, a similar pattern has been observed for a cyclin A in Arabidopsis (Boruc et al., 2010). Additionally, changes in the architecture and functional organization of the nuclear bodies accompany the switch of the developmental program and the activation of proliferative activity in plant cells (Testillano et al., 2005; Echeverría et al., 2007). Therefore, SCI1-GFP distribution in the nucleolus and its relocation to nuclear bodies may be linked to different cell cycle phases, which is consistent with SCI1 involvement in cell proliferation/differentiation control.

Based on the phenotype of the transgenic plants, our results strongly indicate that SCI1 has cell proliferation inhibitory activity. The smaller SSZ observed in SCI1 overexpression plants was formed of cells of approximately the same size as cells of wild-type plants, showing that fewer cells accumulated during SCI1OE stigma development. Accordingly, the larger SSZ in SCI1Ri plants was directly associated with an increase in cell number and an earlier initiation of cell proliferation/differentiation activity. Our results demonstrate that SCI1 participates in the timing control of stigma/style cell proliferation/differentiation and seems to be an integrator of cell cycle regulation in the stigma/style developmental program. It is worth mentioning that, apart from the pistils, no other organ was macroscopically altered in SCI1OE plants. This suggests that SCI1 depends on tissue-specific post-translational modifications, such as phosphorylations, and/or on the interaction with tissue-specific partners to exert its effect on the more ubiquitous cell cycle machinery.

An interesting issue in organ development is whether cell division follows a developmental program (organismal theory) or whether cell division instructs organogenesis and development (cellular theory). As there are published results that are consistent with both theories, a ‘neo cell theory’ was proposed (Tsukaya, 2002; Beemster et al., 2003). In this theory, an integrating model between growth and division is supported. In SCI1 transgenic plants, alterations in cell division in stigma/style tissues affect both the final organ size and differentiation timing, which supports the cellular theory. However, while STT thickness is modified in both types of SCI1 transgenic plants as a result of altered cell division activity, style length is modified only in the RNAi plants. This phenotype represents an unaltered developmental program in SCI1OE styles, which is not influenced by reduced cell division activity in the SSZ/STT, supporting the organismal theory. SSZ cell ablation led to shorter styles in STIG1pro:barnase transgenic plants (Goldman et al., 1994), with no alteration in STT cells. Hence, these observations suggest that the style elongation program is a result of different signals, including those from stigma tissue (SSZ). Thus, one possible explanation is that the lower number of cells in SCI1OE stigmas produced sufficient ‘stigma-dependent elongation signal’ for SCI1OE styles to elongate to the wild-type length. However, the protruded style in SCI1Ri plants would be a consequence of the increased number of stigma cells, which in turn produced more of the ‘elongation signal(s)’ or, alternatively, had started to produce the ‘elongation signal(s)’ earlier. The effects of SCI1 modulation on stigma size and style length suggest that pistil development is governed by a variety of signals, in agreement with the proposal that growth processes are implemented at the cellular level but generate cross-talk between organizational levels, a scenario that does not wholly support either the organismal or the cellular theory. Considering this point of view, how consistent are the SCI1 results with the ‘neo cell theory’? In our SCI1Ri plants, an increase in cell division in the stigma and premature differentiation were observed. Thus, we propose that cell division inhibits cell differentiation until the stigmatic cell number reaches a threshold above which the differentiation signals accumulate and, thereafter, the cell division rate slows down and is coupled with cell differentiation (Fig. 7). Consistent with this proposal, the SCI1OE plants, in which an over-inhibition of stigma cell division occurred, began their stigma/style differentiation process later than the wild-type plants, but the subsequent steps of differentiation followed the wild-type developmental rate (Fig. 7). Our results show that the SCI1 gene encodes an important tissue-specific cell cycle regulator that influences upper pistil growth and development, without changing cell size.

Figure 7.

A proposed model for stigma/style cell cycle inhibitor 1 (SCI1) action in upper pistil development. After the two carpels are formed and fused, cells positioned at the upper pistil start dividing at the carpel junction site. When the appropriate number of cells is reached, they start differentiation in the stigmatic secretory zone (SSZ) and stylar transmitting tissue (STT). In the wild type (WT), the SCI1 level would regulate cell proliferation and determine the correct developmental stage in which differentiation occurs. In RNAi plants (SCI1Ri), the reduced SCI1 level would allow cells to start proliferating earlier or more quickly, reaching the threshold for differentiation at an earlier flower developmental stage. As these cells would stop dividing at the same developmental moment as in the WT, probably under the control of a yet unknown factor, a greater number of cells would accumulate. In overexpression plants (SCI1OE), the very high SCI1 level would reduce or delay cell divisions, and consequently cells would differentiate later. The unknown factor, which is independent of SCI1, would regulate the moment at which these cells would stop proliferation, resulting in a reduced number of stigma/style specialized cells (SSZ and STT).

The regulation of SCI1 expression levels is sufficient to influence stigma size and style length, which may have important consequences for plant reproduction. A decreased SCI1 level resulted in stigmas positioned above the anthers and in heterostyly in our population of tobacco transgenic plants. In nature, the regulation of expression levels of the SCI1 homologous sequences may contribute to the heterostyly observed in some species with perfect (bisexual) flowers and affect the rates of self- and cross-pollination.

Is SCI1 a tissue-specific CDK inhibitor?

Various aspects of plant development have been studied to uncover the molecular network behind organ growth and development, culminating with the discovery of different classes of genes, such as those involved in cell proliferation control. To achieve this control, the cell recognizes different signals from a variety of pathways that are transduced by cell cycle regulators and result in the activation or inhibition of the cell cycle. Plant CKI/KRPs constitute the major negative regulators of CDK activity during development (De Veylder et al., 2007). In addition to its effect in inhibiting cell proliferation, SCI1 has similarity with SIM (Fig. S4), in the same region of similarity between SIM and CKI/KRPs (Churchman et al., 2006), and it contains several domains recognized as important for the function of CDK inhibitors (Table S1). To date, all overexpressed plant CKI/KRPs have been found to cause overall reduced growth, serrated leaves and abnormal flowers (Wang et al., 2000; De Veylder et al., 2001; Jasinski et al., 2002b; Zhou et al., 2002; Ren et al., 2008), phenotypes not present in our SCI1OE plants. However, it is clear that SCI1 encodes a protein that negatively controls cell division and differentiation. Is SCI1 a new class of CDK inhibitor? Plant CDK inhibitors usually have a small molecular weight, contain conserved motifs, such as the cyclin binding domains, and are exclusively present in the nucleus of the cell (Wang et al., 2008), characteristics shared with SCI1. Despite the fact that a direct link between SCI1 and the cell cycle machinery is still missing, based on all the results presented here, it is reasonable to propose that SCI1 is a new CDK inhibitor that works in a tissue-specific manner.

Acknowledgements

The authors are grateful to Prof. Dr Maria-Cristina Pranchevicius (FM/UFT, Brazil) and Prof. Dr Luis Lamberti daSilva (FMRP/University of Sao Paulo – USP, Brazil) for stimulating discussions; to Prof. Dr Wagner Ferreira-Santos (FFCLRP/USP) and Marina F. B. Costa for their contributions to the optical microscopy; to José-Augusto Maulin and Maria-Dolores Ferreira (FMRP/USP) and Rodrigo Silva (FFCLRP/USP) for their technical assistance with the electron microscopy; to Dr Lenaldo Rocha (FMRP/USP) for his contribution to the confocal microscopy; to CEMEQ for the statistical analysis; and to the anonymous reviewers for their suggestions for improvements to this article. We thank the Botanical Garden of the University of Nijmegen (the Netherlands) for providing us with Nicotiana seeds and the Flanders Institute for Biotechnology (VIB, Belgium) for the Gateway vectors. This work was supported by grants from FAPESP and CNPq (Brazil). H.C.DeP, A.C.Q., M.S.B., G.H.G. and M.H.S.G. are grateful to CNPq and CAPES (Brazil) for their fellowships.

Ancillary