SEPALLATA­-like genes of Isatis indigotica can affect the architecture of the inflorescences and the development of the floral organs

Plant materials and bacterial strains

Woad plants, wild-type Arabidopsis ecotype Col-0, wild-type Arabidopsis ecotype Ler, and ap1 cal double mutant in Ler genetic background (CS67157) were used as materials in the present work, and were grown under normal greenhouse conditions (23 ± 2 °C, 16 h light/8 h dark). Woad flowers and leaves were used as materials in extraction of RNA and DNA. pMD 18-T was used in cloning of cDNA and genomic DNA fragments. pCAMBIA1302 was used in construction of the binary expression vectors. Escherichia coli strain DH5α and Agrobacterium tumefaciens strain GV3101 were used as host in isolation of IiSEP2 and in transformation of Arabidopsis, respectively.

Cloning of the coding sequence and the promoter of IiSEP2

Total RNA was extracted from floral organs of I. indigotica with Trizol reagent (Invitrogen, Waltham, MA, USA). Genomic DNA was eliminated by RNase-free DNase I (Takara, Tokyo, Japan). First strand cDNA was synthesized with PrimeScript 1st Strand cDNA Synthesis Kit (Takara, Tokyo, Japan). The coding sequence of IiSEP2 was obtained by PCR and the primers were designed according to the open reading frames of the MADS-box genes in different plant species (Table S1).

Promoter sequences were isolated with Universal Genome Walker 2.0 Kit (Clontech, Mountain View, CA, USA). The primers were designed according to the sequence in the first exon of IiSEP2 (Table S1). Genomic DNA was extracted from leaves of I. indigotica by CTAB method (Stewart & Via, 1993). Uncloned libraries were created after restriction digestion of the genomic DNA with Dra I, EcoR V, Pvu II or Stu I, and ligation of the adaptors provided by Universal Genome Walker 2.0 Kit (Clontech, Mountain View, CA, USA). Upstream sequences were amplified with Adaptor Primer 1 and IiSEP2-Upstream1 with Advantage 2 PCR Kit (Clontech, Mountain View, CA, USA). Nested PCR was carried out with Adaptor Primer 2 and IiSEP2-Upstream2. The PCR products were ligated into pMD 18-T and were sequenced after identification by colony PCR.

Analyses of the expression pattern of IiSEP 2 in I. indigotica

Total RNA was extracted from different tissues and floral organs of I. indigotica with RNAprep Pure Plant Kit (TIANGEN, Beijing, China) and the first strand cDNA was synthesized with PrimeScript 1st Strand cDNA Synthesis Kit (Takara, Tokyo, Japan). Quantitative real-time PCR (qRT-PCR) was conducted in BIO-RAD real-time PCR system (BIO-RAD, Hercules, CA, USA) with the primers shown in Table S1 (Huang et al., 2008; Ma et al., 2019). Woad actin (GenBank accession No. AY870652.1) was used as a reference gene. Triplicate qRT-PCRs were carried out and relative quantification of the transcript levels was accomplished using the comparative threshold cycle (Ct) method. Relative quantification refers to that the PCR signals of the IiSEP2 transcript in leaves, stem apexes, early inflorescences (the inflorescence shrinks into a ball), later inflorescences (the inflorescence is completely unfolded), flowers in full bloom, floral organs (include sepals, petals, stamens, and pistils), and young silicles are normalized to the PCR signal in roots. The fold change was calculated using the following formula: fold change = 2^−∆∆Ct, where ∆∆Ct = (Ct_IiSEP2 − Ct_Iiactin) _{different sample} − (Ct_IiSEP2 − Ct_Iiactin) _root.

Subcellular localization by confocal laser scanning microscopy

The coding sequence of IiSEP2 was inserted into pCAMBIA1302 containing CaMV 35S promoter after high-fidelity amplification with the primers shown in Table S1. The recombinant construct was introduced into competent cells of A. tumefaciens GV3101 by freeze-thaw method (Höfgen & Willmitzer, 1988). Transformed A. tumefaciens cells were selected on YEB medium containing rifampicin (50 μg/L), gentamicin (50 μg/L), and kanamycin sulfate (50 μg/L). Positive colony carrying recombinant plasmid was propagated in liquid medium for 24 h at 30 °C in a shaker incubator (~200 rpm). Precipitated bacterial cells were resuspended with fresh YEB medium containing the same antibiotics and were grown to an OD₆₀₀ value of 0.8. The bacterial cells were collected by centrifugation at 6,000 rpm for 10 min under room temperature. The pellet was resuspended with agroinfiltration buffer (10 mmol/L 2-(N-morpholino)ethanesulfonic acid, 200 μmol/L acetosyringone, 10 mmol/L MgCl₂) to an OD₆₀₀ value of 0.7. The bacterial suspension was infiltrated into the abaxial leaf epidermis of 6-week-old Nicotiana benthamiana plants using a syringe without needle. The quality of transient expression was estimated 60 h after agroinfiltration. Confocal laser scanning microscopy of the living plant tissues was performed using an OLYMPUS FLUOVIEW FV1000 microscope. Green fluorescence of GFP was observed with an exciting light of 488 nm.

Identification of IiSEP2 transgenic Arabidopsis plants

The recombinant binary expression vector pCAMBIA1302-IiSEP2, in which IiSEP2 is controlled by CaMV 35S promoter, was introduced into A. tumefaciens GV3101. Agrobacterium-mediated transformation of Arabidopsis was performed essentially according to the protocol of Clough & Bent (1998). Col-0 transgenic seedlings were selected on hygromycin-containing medium, and the expression levels of IiSEP2 in T2 generation were detected by RT-PCR using IiSEP2 specific primers (Table S1), and by Western blotting using rabbit monoclonal antibody of GFP (Sino Biological Inc., Beijing, China). Transgenic phenotypes were confirmed in T1 and T2 generations. Influence of IiSEP2 on Arabidopsis MADS-box genes, including AP3 (APETALA3), PI (PISTILLATA), SHP1 (SHATTERPROOF1) and SHP2, was determined by qRT-PCR.

Preparation of IiSEP4 transgenic plants in the genetic background of Ler

The binary expression vector pCAMBIA1302-IiSEP4 constructed in our previous work (Pu et al., 2020) was introduced into A. tumefaciens GV3101 and was used in genetic transformation of wild-type Arabidopsis plants (Ler) or ap1 cal double mutant (Ler). The expression of the transgene was analyzed by qRT-PCR with IiSEP4-qRT-F (5′-AGATAGCCGGGATGGGAGTG-3′) and IiSEP4-qRT-R (5′-AATCATCGACCGGGCCTTTG-3′) (Pu et al., 2020). To compare the phenotypic differences, 25 IiSEP4 transgenic Arabidopsis plants were grown in greenhouse together with 25 wild-type plants every time.

Silencing of IiSEP4 by VIGS

To investigate the function of IiSEP4 in woad, its expression was downregulated by Agrobacterium-mediated and TRV-based (tobacco rattle virus-based) VIGS (Liu, Schiff & Dinesh-Kumar, 2002). A. tumefaciens GV3101 was cultured in LB liquid medium containing 50 mg/L kanamycin, 25 mg/L rifampicin and 25 mg/L gentamicin to OD₆₀₀ = 2.0. After centrifugation, the bacterial cells were resuspended with infiltration solution (100 mL was composed of 100 µL 0.2 mol/L acetosyringone, 2 mL 0.5 mol/L 2-(N-morpholino)ethanesulfonic acid, 1 mL 1 mol/L MgCl₂). During four-leaf stage of growth, the leaves of woad plants were infiltrated with a mix of A. tumefaciens GV3101 carrying pTRV1 and A. tumefaciens GV3101 carrying recombinant pTRV2-IiSEP4 containing a 300 bp specific fragment coming from 5’-end of IiSEP4 coding sequence (Pu et al., 2020). The fragment of IiSEP4 was amplified by high-fidelity PCR with primers VIGS-IiSEP4-F (5′-CGCGAATTCCGATGATTGATCAACTATCG-3′) and VIGS-IiSEP4-R (5′-CGCGGATCCTCTGGGATGTTGTTGCAGAG-3′). In each VIGS experiment, the phenotypes of 30 woad plants in treatment group and 30 woad plants in control group were compared.

Data analysis

SPSS 22.0 (SPSS Inc., Chicago, IL, USA) was used for data analysis. The data of rosette leaf number, the angle between cauline leaf and stem, flower bud number, stamen number in IiSEP2 overexpressing Arabidopsis plants, and the data about the expression pattern of IiSPE2 in woad plants, were analyzed by one-way ANOVA, followed by Tukey’s post-hoc test. The data obtained in qRT-PCR analysis of the MADS-box genes (including AP3, PI, SHP1 and SHP2) in IiSEP2 overexpressing Arabidopsis plants, IiSEP4 in 35S::IiSEP4-GFP transgenic plants in wild-type Ler genetic background and in 35S::IiSEP4-GFP transgenic plants of ap1 cal double mutant in Ler genetic background, and IiSEP4 in distal noninfiltrated leaves in woad plants infiltrated with pTRV1 + pTRV2-IiSEP4, were analyzed by two-sided Student’s t-test (Pu et al., 2020).

Identification of the coding sequence and the promoter of IiSEP2

The coding sequence of IiSEP2 was amplified with high-fidelity DNA polymerase by RT-PCR. Sequencing result showed that the open reading frame of IiSEP2 was 753 bp in length and encoded a 250-amino-acid protein (Fig. S1). IiSEP2 comprises the four regions of typical plant MIKC-type MADS-box proteins, namely MADS domain (M, 2-60aa), intervening region (I, 61-83aa), keratin-like domain (K, 84-171aa), and C-terminal region (C, 172-250aa). At the same time, the protein encoded by IiSEP2 shows high identity in amino acid sequence with the homologous proteins of other plant species. For instance, the identity between IiSEP2 and Arabidopsis SEP2 is 95% (Fig. S2A). Eight conserved motifs with a length of 6~50 amino acids could be found in IiSEP1, IiSEP2 and IiSEP2-like proteins by the online tool in MEME website (http://meme-suite.org/) (Bailey et al., 2009), and two motifs (Motif 1 and Motif 9) in MADS domain and three motifs (Motif 3, Motif 4 and Motif 5) in K-box domain existed in all the analyzed proteins (Fig. S2B). The logo of these motifs discovered by MEME is shown in Fig. S2C.

To clarify the relationship of IiSEP2 and other plant MADS-box proteins, we analyzed the phyletic evolution of IiSEP2 (Fig. S3). The results showed that IiSEP2 was closely related to Arabidopsis SEP2. These results demonstrated that IiSEP2 was an E-class floral homeotic gene and further confirmed that IiSEP2 was the orthologous gene of Arabidopsis SEP2.

In 1 kb regulatory sequence upstream of the initiation codon characterized by Genome Walking method (Sequence Data S1), there are three CArG-boxes, indicating the expression of IiSEP2 can be affected by other MADS-box transcriptional factors.

Spatial and temporal expression of IiSEP 2 in I. indigotica

The abundances of IiSEP2 mRNA in various tissues and floral organs of I. indigotica were analyzed by qRT-PCR. The transcriptional levels of IiSEP2 were low in roots and leaves. However, the quantity of IiSEP2 transcript was significantly increased in stem apex and was accumulated to a high level in flowers and young silicles. Further testing in floral organs showed that IiSEP2 mRNA was highly expressed in sepals, petals, stamens, and pistil (Fig. 1). Taken together, the qRT-PCR results illustrate that IiSEP2 expression occurs from the early stages of flower development to young silicles.

Subcellular localization of IiSEP2

To determine the subcellular localization of IiSEP2 in vivo, the coding sequence of IiSEP2 without the stop codon was fused in frame with GFP reporter gene in pCAMBIA1302 and was driven by CaMV 35S promoter. A. tumefaciens carrying the recombinant expression vector was used to perform a transient expression assay in N. benthamiana leaves. The GFP fluorescence of the control experiment using empty vector evenly distributed throughout the cell, whereas the GFP fluorescence of IiSEP2-GFP was observed only in the nucleus (Fig. S4). The result is consistent with the predictions using PSORT II online server (http://psort.hgc.jp/form2.html), namely KRIENKINRQVTFAKRR in N-terminus of IiSEP2 is a bipartite nuclear localization signal (NLS).

Overexpression of IiSEP2 leads to morphologic changes in transgenic Arabidopsis plants

To examine the function of IiSEP2, transgenic Arabidopsis plants under the control of the constitutive CaMV 35S promoter were prepared in this work. Seeds were selected on MS medium containing hygromycin, and the putative transgenic plants were confirmed by PCR and Western blotting (Fig. S5).

In 28 35S::IiSEP2 independent transgenic lines, nine transgenic lines showed obvious macroscopic differences in comparison with the wild-type Arabidopsis plants. 35S::IiSEP2 transgenic Arabidopsis plants were extremely early in bolting. The flowering time was evaluated by counting the number of the leaves produced when the first flower bloomed. The transgenic lines generated 4–5 leaves prior to the onset of flowering, whereas wild-type plants produced approximately nine leaves (Fig. 2A).

In wild-type Col-0 Arabidopsis plants, cauline leaves are generated at the lower part of the inflorescence stems, and the oval-shaped blades are slightly curved outward. Moreover, the angles between the cauline leaves and the stems in wild-type Col-0 Arabidopsis plants are larger (Figs. 2B, 3A and 3B). Different from wild-type Col-0 Arabidopsis plants, the cauline leaves of IiSEP2 transgenic Arabidopsis plants are clearly curling inward and the margin is involute. Simultaneously, the angles between the cauline leaves and the stems become smaller in transgenic Arabidopsis plants, and the whole blade looks like an inverted cone (Figs. 2B and 3C–3G).

The process that the plant switches from vegetative growth to reproductive growth is named as floral transition. At the early stage of floral transition, the vegetative shoot apical meristem is transformed into an inflorescence meristem, and floral meristems are formed in the peripheral regions of the inflorescence meristem. Then the primordia of the floral organs will be produced along with the development of floral meristems, which is accompanied by definition of the boundaries between different floral organs. The inflorescence meristem of wild-type Col-0 Arabidopsis plants is highly active during the growth process and can produce a lot of flower buds continuously (Figs. 4A and 4B). However, the activity of the inflorescence meristem in IiSEP2 transgenic Arabidopsis plants is reduced remarkably. In general, only a few flower buds could be generated at the top of the inflorescence stem in IiSEP2 transgenic Arabidopsis plants (Figs. 2C and 4C–4F), illustrating that constitutive expression of IiSEP2 could reduce the number of flowers and the number of lateral branches. Namely, IiSEP2 can affect the architecture of the inflorescences.

Under normal circumstances, the flowers of the wild-type Col-0 Arabidopsis plants consist of four whorls of floral organs, including four sepals in the outermost whorl, four petals in the second whorl, six stamens in the third whorl, and a gynoecium in the fourth whorl (Figs. 5A–5C). By observation under stereomicroscope, it was found that overexpression of IiSEP2 in Arabidopsis resulted in seriously anomalous development of the floral organs. A number of flowers were male sterile (Fig. 4D), owing to the abnormality of stamens in development. At the same time, the inflorescence meristems were transformed into defective terminal flowers without sepals and petals, and the number of stamens was reduced (Figs. 4E–4G). In Fig. 4E, only three stamens could be observed under the pistil in the middle. In many flowers, the sepals were converted homeotically into carpelloid structures with stigmatic papillae and ovules in transgenic lines 7 and 8 (Figs. 5E and 5J).

In transgenic lines 3 and 4, secondary flower could be generated in terminal flowers. The number of sepals, petals and stamens was decreased in these terminal flowers, whereas the secondary flowers were nearly normal in structure (Figs. 5D and 5F). Different from four petals in wild-type flowers, a few apical flowers of the IiSEP2 overexpressing plants produced five petals or lacked petal, together with anomalous development of the carpels in transgenic lines 5 and 6 (Figs. 5G and 5H). Quantitative expression analyses showed that AP3 and PI, two Arabidopsis MADS-box genes related to petal formation, were suppressed by ectopic expression of IiSEP2 (Fig. S6). On the contrary, the transcripts of SHP1 and SHP2 were increased in IiSEP2 transgenic Arabidopsis plants.

In order to fully understand the function of IiSEP2, transgenic line 8 was genetically crossed with the wild-type Col-0 plants, and the phenotypes of the crossing progenies were compared to that of the IiSEP2 transgenic line 8. The phenotypes of the IiSEP2 transgenic lines were maintained in filial generations of the crossing lines. These offsprings exhibited the phenotypic variations of the IiSEP2 overexpressing plants, including the secondary flowers, the loss of petals, and the reduction of stamens (Figs. 5I, 5K, and 5L). All these phenotypic changes indicated that constitutive expression of IiSEP2 could affect the morphogenesis of the flowers.

Ectopic expression of IiSEP4 in Ler genetic background can affect the development of floral organs

In order to analyze the regulatory function of IiSEP4 in floral transition and flower development, transgenic Arabidopsis plants were prepared. The transcriptional levels of IiSEP4 in transgenic plants were detected by qRT-PCR. The results showed that IiSEP4 could express effectively in transgenic plants of wild-type Arabidopsis in Ler genetic background, and in transgenic plants of ap1 cal double mutant in Ler genetic background (Fig. S7).

All the 35S::IiSEP4-GFP transgenic plants displayed abnormal phenotypes, and the number of the floral organs was reduced. In particular, sepals were curly and were converted into carpelloid structures, and were obviously accompanied by the generation of ovules (Figs. 6B and 6C). In addition, dwarf and abnormal petals and stamens were produced (Fig. 6D), and the size of the silique was also decreased significantly in comparison with the wild-type Ler plants (Fig. 6F).

On account of the apetalous phenotype of ap1 cal double mutant, ectopic expression in this mutational material was conducted to clarify the roles of IiSEP4 in petal differentiation. In ap1 cal double mutant in Ler genetic background, floral meristems were transformed into inflorescence meristems, and new meristems were elaborated continuously. As a result, the inflorescence resembles a cauliflower in ap1 cal double mutant. In addition, ap1 cal plants produced apetalous flowers, namely only sepals, stamens and pistil could be formed (Figs. 7A and 7B). In 35S::IiSEP4-GFP transgenic plants of ap1 cal double mutant, the cauliflower phenotype was attenuated significantly, and the petals could be recovered (Figs. 7C–7H). Occasionally, chimeric organ composed of petaloid and sepaloid tissues (Figs. 7F and 7G), and chimeric organ composed of petaloid and stamineous tissues (Fig. 7E), could be observed.

Phenotypic variations of woad plants after downregulation of IiSEP4 by VIGS

To analyze the moderating effects on woad flowering and to verify the results obtained in Arabidopsis, the expression level of IiSEP4 was downregulated by VIGS mediated by Agrobacterium. The results of RT-PCR showed that RNA molecules of TRV1 and TRV2 could be detected in distal noninfiltrated leaves of the control woad plants treated with pTRV1 + pTRV2, and in distal noninfiltrated leaves of the woad plants treated with pTRV1 + pTRV2-IiSEP4 (Fig. S8). qRT-PCR data showed that the mRNA abundance of IiSEP4 was reduced dramatically in woad plants treated with pTRV1 + pTRV2-IiSEP4 (Fig. S9), indicating IiSEP4 could be silenced effectively by VIGS.

Silencing of IiSEP4 can delay the flowering time. Compared to woad plants infiltrated with a mixture of A. tumefaciens GV3101 carrying pTRV1 and A. tumefaciens GV3101 carrying the empty pTRV2, the woad plants infiltrated with a mixture of A. tumefaciens GV3101 carrying pTRV1 and A. tumefaciens GV3101 carrying pTRV2-IiSEP4 start bolting nearly half a month later, indicating downregulation of IiSEP4 can lead to late-flowering phenotype in woad. Specifically, wild-type woad plants begin to bloom in first ten-days of May, whereas IiSEP4-silenced lines bear flowers in late May.

More than 90% of the woad plants in treatment group showed phenotypic variations. Observation of the flowers showed that the floral organs were unchanged after infiltration with pTRV1 + pTRV2 (Fig. 8A), whereas the floral organs in woad plants infiltrated with pTRV1 + pTRV2-IiSEP4 presented pronounced anomalous phenotypes, and the number of sepals, petals and stamens was reduced (Figs. 8B–8E). In comparison with the control group, the size of the floral organs in woad plants infiltrated with pTRV1 + pTRV2-IiSEP4 was much smaller (Figs. 8F–8H). In some woad plants treated with pTRV1 + pTRV2-IiSEP4, the number of petals was increased. For instance, one flower in a woad plant infiltrated with pTRV1 + pTRV2-IiSEP4 produced five petals, but the number of stamens was reduced from six to five. In these plants, except the smaller sepals, no obvious change could be observed in the size of petals and stamens (Figs. 8I–8L).