Plant material, growing conditions and compounds treatments
The plant material was yellow lupine (Lupinus luteus L.) cv. Taper. Obtained from the Plant Breeding Station Wiatrowo (Poland) the seeds were prepared and sown as described by [34] in the experimental field in Pędzewo, Kuyavian-Pomeranian Voivodeship in north-central Poland (53°05′02″N 18°21′28″E). The plants were cultivated on the 5th soil class as recommended by the manufacturer. The anthers from particular stages of late development (1–4 LAD, Fig. 1A) were dissected from flowers with a sharp scalpel. For each phase of LAD not less than 80 plants were used, and additionally, part of the collected material was previously treatment with gibberellic acid (GA3, 100 µM) or GA biosynthesis inhibitor – paclobutrazol (PAC, 100 µM) in a 0.05% Tween 20 solution using a sprayer. The anthers in the corresponding developmental phases were used as control group, and they were treated with 0.05% Tween 20 solution only. According to the planned experiment, 3 h after applications the anthers were harvested and processed fresh or frozen in liquid nitrogen (then stored at -80 °C until use).
Identification of cDNAs
The cDNA sequences of individual genes have been identified differently. Molecular cloning of LlGA2ox1 cDNA was performed as described by [33]. The identification of LlGAMYB cDNA was as follows: the 1 g of yellowish-green flowers was pounded in liquid nitrogen using a mortar and pestle. Then, according to the manufacturer's instructions, total RNA was isolated by the column method (NucleoSpin® RNA kit, Macherey-Nagel, Düren, Germany), while the reverse transcription was made using 1 µg of RNA, oligo(dT)18 primers, and Transcriptor High Fidelity cDNA Synthesis Kit (Roche, Mannheim, Germany). Touchdown PCR was performed using 1.25 U Perpetual Taq DNA PolymeraseHOT START (EURx, Gdańsk, Poland), 2 µl of first-strand cDNA, 1 x buffer B, 0.2 mM dNTP mix, 3.0 mM Mg2+, 1 µM degenerated primers (Tab. S1) and deionized water (up to final volume of 50 µl) in the T3 Thermocycler (Biometra, Göttingen, Germany). An amplified cDNA fragment (704 bp, Fig. S11A) was isolated and purified from an agarose gel (GeneMATRIX Agarose Out DNA Purification Kit, EURx). In the insertion site, PCR product was ligated into the pCRII-TOPO vector (Fig. S11B, TOPO TA Cloning Kit, Invitrogen, Carlsbad, USA), and transferred into One Shot Mach1-T1 E. coli in the form of plasmid. The bacterial cells were plated on Petri dishes, which contained S-Gal/LB Agar Blend (Sigma-Aldrich, St. Louis, MO, USA) with addition of ampicillin (50 µg/ml) (Fig. S11C). Unlike the dark blue bacterial colonies, white ones were selected and cultured in liquid LB medium with ampicillin (50 µg/ml) for 12 h. Finally, in accordance with the manufacturer's guidelines, the plasmid DNA was isolated (GeneMATRIX Plasmid Miniprep DNA purification kit, EURx), digested with the restriction enzyme EcoRI (Fermentas) (Fig. S11D), and sequenced by Genomed (Warsaw, Poland). The 485 bp (Fig. S12A) and 768 bp (Fig. S12B) fragments were obtained from the 3’RACE-PCR (FirstChoice RLM-RACE Kit, SuperTaq-Plus Polymerase, Ambion, Austin, USA) with using two different pairs of primers (Tab. S1). The amplicons were isolated, purified, cloned, digested (Fig. S12C/D) and sequenced as described above. Due to difficulties arising from the experimental identification of the 5’ end of LlGAMYB cDNA, it was obtained based on sequences derived from a later RNA-Seq experiment deposited at NCBI in the Sequence Read Archive (SRA) database under accession number PRJNA285604 (BioProject) [77] and experiment accession number SRX1069734. All fragments had overlapping nucleotide sequences which allowed obtaining the complete LlGAMYB sequence. De novo assembled transcriptome of yellow lupine from RNA-seq experiment was also used to identify LlCAD, all LlIRXs, LlQRT2, LlPCS1, LlGA3ox1 and precursor of Ll-miR159 (Ll-MIR159).
Expression analysis
Quantitative real-time PCR (qPCR) was used to establish the expression pattern of all identified genes. 80 mg of frozen anthers (being at a specific stage of development with or without GA3/PAC treatment) were homogenized in a sterile mortar with a pestle. According to the manufacturer’s instructions, first total RNA using Isolate II RNA Plant Kit (Bioline, London, UK) was isolated, and next reverse transcription with matrix (2 µg), oligo(dT)18 (Roche) and the Transcriptor High Fidelity cDNA Synthesis Kit was performed. qPCRs were carried out using LightCycler 2.0 Carousel-Based System (Roche) in 20 µl capillaries which contained the mix of 0.1 µg of cDNA, 0.2 µM gene-specific primers (Tab. S1), 0.05 µM Universal Probe Library (UPL) hydrolysis probes (Roche) (Tab. S1), 1 × LightCycler TaqMan Master Mix (LightCycler TaqMan Master Kit, Roche) and H2O. The following program was used: 600 s at 96 °C; 45 cycles of 10 s at 96 °C, 20 s at specific annealing temperature (Tab. S1), 1 s at 72 °C; and 30 s at 40 °C. The negative no template controls (NTC) were included. As an endogenous control the actin gene (LlACT, GenBank accession number KP257588) was selected [32, 33, 34]. The absolute quantification was designed from the serial dilutions of cDNAs generating standard curves, whereas the relative gene expression (the data of the studied gene relative to internal control gene and calibrator) was determined using the 2 (-DeltaDeltaC(T)) method [78].
Histological studies
A fixer containing 4% paraformaldehyde (w/v), 0.2% glutaraldehyde (v/v) and 3% N-ethyl-N′-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDAC) (w/v) (Sigma-Aldrich) in 1 × phosphate-buffered saline buffer (1 × PBS, pH 7.2) was prepared and applied into appropriate anther tissue small fragments for 12 h at 4 °C. Samples were washed out of fixative in triplicate with 1 × PBS (pH 7.2) for 10 min. Dehydrated was achieved by placing the material in increasing ethanol concentrations (30%, 50%, 70%, 90%, 100%) (v/v). Then, supersaturation and embedding in BMM resin (butyl methacrylate, methyl methacrylate, 0.5% (w/v) benzoin ethyl ether, 10 mM dithiothreitol) (Fluka, Buchs, Switzerland) were performed at − 20 °C under UV light. Using the ultramicrotome (Reichert-Jung, Germany) semi thin sections have been received, which were later placed on slides, stained with 0.05% toluidine blue (Sigma-Aldrich), and observed under LM Zeiss Axioplan (Carl Zeiss, Oberkochen, Germany) microscope with a ProGres C3 digital camera.
GA 3 immunolocalization
The anther tissue fragments were fixed, washed, dehydrated, supersaturated, embedded, polymerized and cut in the same way as described for histological studies. Next, semi thin sections were placed on slides with Biobond (BBInternational, Cardiff, UK), and after washing off the resin blocked in BlockAid TM Blocking Solution (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s instruction. Then, sections were incubated with polyclonal primary antibody anti-GA3 (Abbexa, Cambridge, UK) diluted 1:50 in 1% bovine serum albumin (BSA) in 1 × PBS (pH 7.2), and placed in a wet container at 4 °C for 12 h. Next, the primary antibodies by washing 3 times in 1 x PBS (pH 7.2) were removed and secondary antibody (MFP488 goat anti-rabbit IgG, MoBiTec, Goettingen, Germany) diluted 1:250 in PBS buffer were applied for 2 h at 37 °C. A negative control was done by omitting the incubation with primary antibody (Fig. S13). The next step was the incubation with DAPI (1:2500) for 10 min., and washing it out with distilled water. Finally, the samples were observed in a Leica DMI4000B inverted microscope using the BP365, FT395, and LP397 filters.
Is silico analyses
The integrated FastPCR v.6.5.99 [79] tool was used for degenerate and RACE-PCR primers design. The Universal Probe Library Assay Design Center [80] was used for design of qPCR specific primers and short hydrolysis UPL probes substituted with Locked Nucleic Acids. The identified cDNA sequences of all genes and predicted proteins were analyzed using Basic Local Alignment Search Tool (BLAST) [81] and the bioinformatics resource portal of the Swiss Institute of Bioinformatics [Expasy [82], including Translate tool [83] which allows the translation of a nucleotide sequences to a protein sequences, and ProtParam tool [84] for calculation of molecular weights and isoelectric points]. Alignments and phylogenetic reconstructions were performed using Python Environment for Tree Exploration3 (ETE3) v3.1.1 program as implemented on the GenomeNet [85]. Maximum likelihood phylogenetic trees were inferred using PhyML v20160115 ran with model and parameters: --pinv e -o tlr -f m --bootstrap − 2 --nclasses 4 --alpha e. Branch supports are the Chi2-based parametric values return by the approximate likelihood ratio test. Multiple alignments of different amino acid sequences found in BlastP and showing close association with yellow lupine sequences were carried out using ClustalW [86] program with default settings. The presence of functional domains were checked via NCBI's Conserved Domain Database (CDD) [87], which is a protein annotation resource, consisting of a collection of well-annotated multiple sequence alignment models for ancient domains and full-length proteins. These are available as position-specific score matrices (PSSM) for the rapid identification of conserved domains in protein sequences via RPS-BLAST. CDD content includes NCBI-curated domains, as well as domain models imported from a number of external source databases (Pfam, SMART, COG, PRK, TIGRFAMs). The cytoplasmic and transmembrane domains were also predicted using TMHMM v. 2.0 [88] and Protein Homology/analogY Recognition Engine V 2.0 (Phyre2) [89] web portal for protein modeling, prediction and analysis. Comparisons of proteins derived from different plant species were made using the DiAlign program (Genomatix) [90] with default parameters. The tertiary structures of the proteins derived from yellow lupine were constructed using protein structure prediction Robetta service [91], which uses the PDB100 template database and a co-evolution based model database (MDB). The obtained results were visualized using the UCSF ChimeraX [92], which is the next-generation molecular visualization program from the Resource for Biocomputing, Visualization, and Informatics (RBVI).
Statistical analysis
All presented data are the results of three separate samples (biological replications) with two repetitions of each (technical replications) and presented as mean ± standard error (SE). Statistical analysis were performed using one-way Anova followed by post-hoc Tukey’s HSD test, with differences accepted at p < 0.05. All analyses were performed using R version 3.5.3.