GRAS transcription factors regulate cell division planes in moss overriding the default rule

Significance Plant cells are surrounded by a cell wall and do not migrate, making the regulation of cell division orientation critical to the establishment of plant body shape. A gene encoding a GRAS transcription factor was transferred from soil bacteria to plants, and its descendent genes now regulate formative cell divisions in flowering plants. Here, we show that three GRAS proteins participate in leaf vein formation by regulating cell division orientation in the moss Physcomitrium patens. We propose that the acquisition of GRAS genes contributed to the genetic regulatory networks controlling cell division orientation in the ancestor of land plants, and this gene family then underwent expansion and adaptation in flowering plant and moss lineages to specify body plans.

The entire dataset was aligned with einsi in MAFFT version 7.475 (3). The alignment was imported into Mesquite version 3.61 (4). Redundant entries were not added as Selaginella sequences were contained in both the representative genome set and the non-angiosperm nr dataset. All characters in the alignment were first set as "excluded". The alignments were then reviewed on Mesquite, resulting in the identification of well-aligned regions later set for "included". Sequences harboring large deletions in the conserved region were removed, yielding a SHR data matrix with 203 entries, for each of which a maximum likelihood tree was constructed as follows: after examining the result for each of the 203 entries, the group derived from a protein encoded by a single gene in the last common ancestor of land plants was estimated, a smaller dataset comprising that group and two outgroup clades was chosen (74 entries for SHR), and the alignment was reinvestigated for inclusion/exclusion. Each data matrix was saved as a nexus file and subjected to a custom maximum likelihood analysis pipeline involving RAxML-HPC (5). The "included" sites were extracted from the matrix, and sequences identical in the included regions were treated as a single OTU during the analysis and reverted to the original names at the final stage. The extracted matrix was converted to PHYLIP format. A molecular evolutionary model was chosen using ProteinModelSelection.pl, and in this case, LGF (LG+F) was chosen for others. LG stands for the amino acid substitution matrix according to Le and Gascuel (2008) (1) and F stands for using empirical amino acid frequency in the dataset. SEQBOOT (6) was used to prepare 1,000 resampling sets. The original and bootstrap replicates were individually processed with the "-f a -# 100" option of RAxML. The model was specified with "-m PROTGAMMALGF" according to the chosen model. Random number seeds -x and -p option were taken from the operating system with /dev/urandom. After all jobs had finished, the resulting trees were recovered and the bootstrap frequency was calculated using CONSENSE (6). To force the topology to be the same as that of the tree obtained from the original dataset, the original tree was amplified 1,000 times and combined with the bootstrap results; after consensus calculation, 1,000 was subtracted from the result.

Fig. S2. Construction of the ∆ppshr1, ∆ppshr2, and ∆ppshr1∆ppshr2 deletion plants. (A and B)
Schematic diagrams of the constructs targeting the PpSHR1 (A) and PpSHR2 (B) loci. White boxes represent exons. Magenta, cyan, and green boxes denote the neomycin phosphotransferase II expression cassette (nptII) (7), the aminoglycoside phosphotransferase IV expression cassette (aphIV) (8), and loxP sequences (9), respectively. Probes used in (C-E) are shown as red horizontal lines. To generate the PpSHR1 and PpSHR2 gene deletion constructs, genomic fragments containing the 5' and 3' flanking regions of each PpSHR gene were amplified and inserted into the 5' and 3' ends of the nptII expression cassette (A) in plasmid pTN182 (AB267706) or the aphIV expression cassette (B) in plasmid pTN186 (AB542059), respectively. The generated constructs were digested with suitable restriction enzymes for gene targeting and introduced into the wild type to generate ∆ppshr1 and ∆ppshr2 single deletion mutant plants. To generate ∆ppshr1∆ppshr2 double deletion mutant plants, the PpSHR1-deletion construct was introduced into ∆ppshr2 plant #36. (C-E) DNA gel blot analyses of targeted plants with the probes indicated in (A and B). Genomic DNA of wild-type, ∆ppshr1 (#8, #19, and #34), ∆ppshr2 (#2, #23, #35, and #36), and ∆ppshr1∆ppshr2 (#4, #5, #7, and #8) plants was digested with EcoRV. Arrowheads and arrows indicate DNA fragments specific to wild-type and deletion mutant plants, respectively. (F) Representative gametophores and leaves of ∆ppshr1 and ∆ppshr2 single deletion mutants and wild type. The area delineated by a red box is shown at higher magnification in (G). (Scale bars, 1 mm in upper panels, 200 µm in lower panels.) (G) Quantification of the lengths of lamina and midrib in wild-type, ∆ppshr1, and ∆ppshr2 leaves. A magnified image of the wild-type leaf in (F) is shown above the plot. Yellow and red arrows indicate lamina and midrib, respectively. (Scale bar, 200 µm.) Lowercase letters indicate significant differences (one-way ANOVA and Tukey's test, p < 0.05).  (10) or Citrine (11) and the neomycin phosphotransferase II expression cassette (nptII) (7), respectively. Green and gray boxes denote the loxP sequence (9) and the terminator from nopaline synthase (nos-ter) (7). Probes used in (C and D) are shown as red horizontal lines. To insert mCitrine in frame with the PpSHR1 (A) or PpSHR2 (B) coding sequences, a genomic DNA fragment for each locus extending from the middle to the last codon (5' fragment) and containing the 3' flanking region of each gene (3' fragment) was PCR amplified from wild-type genomic DNA and inserted into the HindIII and SpeI sites of the pmCit-nptII plasmid (ON092626), using the In-Fusion HD Cloning Kit (Takara). To insert Citrine in frame with the PpSHR2 coding sequence, a genomic DNA fragment from the start codon to the last codon of the PpSHR2 coding sequence, a DNA fragment encoding Citrine, a genomic fragment of the 3' flanking region of PpSHR2, and pGEM-3Zf(+) (Promega) were individually PCR amplified and precisely assembled using the GENEART Seamless Cloning and Assembly Kit (Thermo Fisher Scientific). The generated constructs were digested with suitable restriction enzymes for gene targeting and introduced into wild-type plants.
(B) Induction of PpSHR2-Citrine in gametophores of GX8:PpSHR2-Citrine plants. Gametophores of wild-type, GX8:PpSHR2-Citrine#6, and GX8:PpSHR2-Citrine#7 plants were cultivated in liquid BCDAT medium with or without 1 µM β-estradiol for three days. Total RNA was purified from gametophores for RT-qPCR analysis. Relative PpSHR2, PpSCR1, PpSCR2, and PpSCR3 transcript levels were obtained by normalization to that of Pp3c9_17670 transcript encoding a thiosulfate sulfurtransferase (15); the value for each transcript in wild-type plants without β-estradiol was set to 1.0. The data points shown in red circles are the averages of three technical replicates in each plant, and the bars indicate the means of the relative transcript levels from three biological replicates. Error bars indicate standard deviations (SD). (C-G) Representative gametophores of the GX8:PpSHR2-Citrine#6 and GX8:NGG#4 plant transiently treated with DMSO or 1 µM -estradiol. Gametophores of the GX8:PpSHR2-Citrine#6 and GX8:NGG#4 plants grown on the solid BCDAT medium for 21 days after propagation were cultivated in the liquid BCDAT medium with DMSO (C) or 1 µM -estradiol (E and G) for 7 days. The chemical-treated gametophores were washed with sterilized milliQ water and further cultivated on solid BCDAT medium for 10 days. Leaves were detached from the upper half of each gametophore and arranged in order from apical to basal (D and F). An asterisk and arrows in F indicate a leaf without midrib and partially formed midribs, respectively. Side and top views of a representative gametophore of the GX8:NGG#4 plant transiently treated with -estradiol in the same conditions are also shown (G).   (12) and the PpSHR2pro:NLS-eGFP-GUS construct into the PIG1 putative neutral genomic locus (12). Green arrows denote the 1,182-bp PpSHR2 promoter sequence. Boxes represent 5' untranslated regions (PpSHR2 5'-UTR: black), PpSHR2 exons (white), a synthetic nucleotide sequence encoding a linker (KGGRADPAFLYKVVITG: purple), mCitrine (yellow) (10), the rbcS terminator (rbcS-ter: gray) (14), loxP sequences (green) (9), the zeocin resistance cassette (zeo: orange) (16), the uidA gene (GUS: light blue) (17), a synthetic nucleotide sequence encoding the SV40 nuclear localization signal (NLS: magenta) (18), eGFP (light green), and the blasticidin S deaminase expression cassette (BSD: brown) (19). To generate the PpSHR2pro:PpSHR2-mCitrine and PpSHR2pro:PpSHR2-mCitrine-GUS constructs, the DNA fragment containing 1,182 bp of the PpSHR2 promoter and the entire coding region until the terminator codon was PCR amplified from wild-type genomic DNA and cloned into pENTR/D-TOPO to generate pENTR-PpSHR2promoter-PpSHR2. The pENTR-PpSHR2promoter-PpSHR2 plasmid was subjected to LR reaction using the destination vectors pTPGmY (OP142292) and pTPGmYG (OP142293) with Gateway LR Clonase II enzyme mix (Thermo Fisher Scientific) to generate the pTPGmY-PpSHR2promoter-PpSHR2 and pTPGmYG-PpSHR2promoter-PpSHR2 plasmids, respectively. To generate the PpSHR2pro:NLS-eGFP-GUS lines, the fragment containing the 1,182-bp promoter fragment and 1,696 bp of 5'-UTR region from PpSHR2 was PCR amplified from wild-type genomic DNA and cloned into pPIG1-NGGII (20) digested with SmaI using In-Fusion HD Cloning Kit (Takara) to generate the pPIG1-PpSHR2pro-NGG plasmid. The generated constructs were digested with PmeI for gene targeting and introduced into wild-type plants.  , and PpSCR3 are indicated with arrows. The maximum likelihood tree under the JTTF model with empirical amino acid frequency was searched with RAxML-HPC, as explained below. The bootstrap probabilities (%) based on 1,000 resamplings are shown on the branches where bootstrap probability was over 50%. The horizontal branch lengths are proportional to the estimated number of substitutions per site, and are drawn to scale. INSD accession numbers or genome-specific identifiers are shown with the species name in which the sequence was present, sometimes followed by the protein name. The OTUs are colored according to the classification: blue, eudicots; blue violet, monocots; dark magenta, other seed plants; brown, other vascular plants (monilophytes and lycophytes); dark green, bryophytes. SCR homologs were searched from a representative genome collection and nr databases with restriction to the Viridiplantae but excluding angiosperms. The representative genome collection consisted of "Araport11" (Arabidopsis thaliana from TAIR), "Atrichopoda1" (Amborella trichopoda v 1.0 from Phytozome), "Mpo31" (Marchantia polymorpha v 3.1 from Phytozome), "Pita_2_01" (Pinus taeda v 2.01, Pita.2_01.peptides.fa.gz on Dec 28 2017 containing 36,732 sequences), "Ppa33" (Physcomitrella patens v 3.3 from Phytozome), "rice-IRGSP-1.0" (IRGSP-1.0_protein_2017-08-04.fasta), "Azolla" (Azolla fillicoides v1.1 from fernbase [https://www.fernbase.org]), "Salvinia" (Salvinia v 1.2 from fernbase), and "Selmo_ncbi" (GCF_000143415.4_v1.0_protein.faa). Sphagnum fallax v1.1 (DOE-JGI, http://phytozome.jgi.doe.gov/) was additionally searched at the Phytozome 13 site. The Anthoceros punctatus dataset was downloaded from https://www.hornworts.uzh.ch/en/download.html. BLASTP searches were performed with a minimum word size of 2 and allowing up to 1,000 hits of which the top 100 hit target amino acid sequences were retrieved. Thus, the SCR dataset of 100 representative plant + 100 non-angiosperm plants + 51 Anthoceros punctatus + 3 Sphagnum entries was obtained.
This dataset was aligned with einsi in MAFFT version 7.475 (3). The alignment was imported to Mesquite version 3.61 (4). Redundant entries were not added, as Selaginella sequences were contained in both the representative genome set and the non-angiosperm nr dataset. All characters in the alignment were first set as "excluded". The alignments were reviewed on mesquite, and well-aligned regions were identified and set for "included". Sequences harboring large deletion(s) in the conserved region were removed. Thus, a 211-entry SCR data matrix was produced and a maximum likelihood tree was constructed. After examining the result, the group derived from a single protein encoded by the gene in the last common ancestor of land plants was estimated, a smaller dataset comprised of the group and two clades of outgroup were chosen (79 entries for SCR), and the alignment was reinvestigated for inclusion/exclusion. The data matrix was saved as a nexus file and subjected to a custom maximum likelihood analysis pipeline involving RAxML-HPC (5). The "included" sites were extracted from the matrix and sequences identical in the included regions were treated as a single OTU during the analysis and reverted to the original names at the final stage. The extracted matrix was converted to PHYLIP format. At first, molecular evolutionary model was chosen using ProteinModelSelection.pl, and in this case, the JTTF (JTT + F) model was selected. JTT stands for amino acid substitution matrix (22) and F stands for using empirical amino acid frequency in the dataset. SEQBOOT (6) was used to prepare 1,000 resampling sets. The original and bootstrap replicates were individually processed with the "-f a -# 100" option in RAxML. The model was specified with "-m PROTGAMMALGF" or "-m PROTGAMMAJTTF" according to the chosen model. Random number seeds -x and -p option were taken from the operating system with /dev/urandom. After all jobs had finished, the resulting trees were recovered and the bootstrap frequency was calculated using CONSENSE (6). To force the topology to be the same as that of the tree obtained from the original dataset, the original tree was amplified 1,000 times and combined with the bootstrap results; after consensus calculation, 1,000 was subtracted from the result. (B) Quantification of relative PpSCR1, PpSCR2, and PpSCR3 transcript levels in gametophores. The value of PpSCR2 transcripts was set to 1.0. Three biological replicates (blue, orange, and green plots) with three technical replicates were performed. The horizontal bars indicate the means.  (11), respectively. Probes used in (B) are indicated as red horizontal lines. To insert the Citrine gene in-frame with the PpSCR1 coding sequence, a genomic DNA fragment spanning from the start codon to the stop codon of PpSCR1 was inserted into the BsrGI site of the pCit-aphIV plasmid (LC703380), resulting in the generation of the in-frame Citrine-PpSCR1 fusion. The Citrine-PpSCR1 fragment was then PCR amplified and inserted into the pENTR/D-TOPO vector (Thermo Fisher Scientific) to generate the pENTR:Citrine-PpSCR1 plasmid. A genomic DNA fragment including a partial sequence of the PpSCR1 promoter and the 5' untranslated region was amplified from wild-type genomic DNA and inserted at the 5' end of Citrine in the plasmid pENTR:Citrine-PpSCR1. The generated construct was digested with suitable restriction enzymes for gene targeting and introduced into wild-type plants.  exons. Magenta, orange, and green boxes denote the neomycin phosphotransferase II expression cassette (nptII) (7), the zeocin resistance cassette (zeo) (16), and loxP sequences, respectively. Probes used in (C and D) are indicated as red horizontal lines. To delete PpSCR1, two types of deletion constructs, ∆ppscr1-nptII and ∆ppscr1zeo, were generated. In the plasmid ∆ppscr1-nptII (A), the coding region encoding the GRAS domain of the PpSCR1 gene was deleted. Genomic fragments containing the partial coding sequence and 3' flanking regions of the PpSCR1 gene were amplified and inserted into the 5' and 3' ends, respectively, of the nptII expression cassette of plasmid pTN182. In the plasmid ∆ppscr1-zeo (B), the full-length PpSCR1 gene was deleted. Genomic fragments containing the 5' and 3' flanking regions of the gene were amplified and inserted into the 5' and 3' ends, respectively, of the zeo expression cassette of the plasmid p35S-loxP-Zeo (AB540628). The two constructs were digested with suitable restriction enzymes for gene targeting and introduced into wild-type plants to generate ∆ppscr1 (C) and ∆ppscr1-zeo (D) single deletion mutants. (C) DNA gel blot analysis of ∆ppscr1 plants. Genomic DNA of wild-type and ∆ppscr1 (#2, #3, #5, #8, and #11) plants was digested with XbaI. (D) DNA gel blot analysis of ∆ppscr1∆ppshr1∆ppshr2 plants. Genomic DNA of wild-type, ∆ppscr1-zeo#8, ∆ppshr1∆ppshr2 (#7 and #8) and ∆ppscr1∆ppshr1∆ppshr2 (#8, #10, and #11) plants was digested with HindIII. To generate ∆ppscr1∆ppshr1∆ppshr2 triple deletion mutants, the ∆ppscr1-zeo plasmid (B) was introduced into the ∆ppshr1∆ppshr2#7 double mutant plant.  (11), the aminoglycoside phosphotransferase IV expression cassette (aphIV) (8), the nopaline synthase terminator (nos-ter) (7), and loxP sequences (9), respectively. Probes used in (C and D) are indicated as red horizontal lines. To insert Citrine inframe with the PpSHR1 coding sequence, a PpSHR1 genomic DNA fragment extending from the middle to the last codon and the 3' flanking region were separately PCR amplified from wild-type genomic DNA. The two amplified fragments and the Citrine expression construct including Citrine and the aphIV expression cassette (8) from the pCit-aphIV plasmid were precisely inserted into the pBluescriptII plasmid (Agilent). The generated construct was digested with suitable restriction enzymes for gene targeting and introduced into wild-type plants.  (23,24). To introduce mutations in the PpSHR1 and PpSHR2 genes by the CRISPR/CAS9 system, primers were designed by CRISPRdirect (http://cridpr.dbcls.jp) to produce single guide RNAs (sgRNAs) targeting PpSHR1 or PpSHR2. The primer sequences are listed in Table S1. The primers for PpSHR1 and PpSHR2 were annealed and cloned at the BsaI site of the sgRNA expression plasmid pPpU6-sgRNA (LC494193) to generate pPpU6-PpSHR1-sgRNA and pPpU6-PpSHR2-sgRNA, respectively. Schematic diagrams illustrating the design of the sgRNA targeting PpSHR1 (E) or PpSHR2 (F) are shown. White boxes denote exons. The two plasmids were simultaneously introduced into Citrine-PpSCR1#38 plants. The resulting mutant plants were screened for deletions or insertions at the PpSHR1 and PpSHR2 loci using Sanger sequencing and genomic PCR analyses. Sequences of the PpSHR1 locus in the wild type and mutants are shown at the bottom in (E). Light green, light blue, and red letters denote the start codon of PpSHR1, the protospacer adjacent motif (PAM), and the sgRNA target sequences, respectively. Genomic DNA of wild type and Citrine-PpSCR1ppshr1ppshr2 (#13, #20, #24) plants was purified and subjected to Sanger sequencing (E) and genomic PCR analyses with six primer pairs (1 to 6) shown as arrows (F). The primer sequences for genomic PCR to evaluate the PpSHR2 locus are described in Table S2.  (11), PpSCR1 (green), the pea rbcS3A terminator (black: pea3A-ter) (14), the putative ProGX8 promoter (orange) (12), a DNA fragment encoding an XVE fusion protein derived from pER8 (purple) (13), the rbcS terminator (gray, rbcS-ter) (14), and the aminoglycoside phosphotransferase IV expression cassette (aphIV: cyan) (8) are shown in different colors. To generate the Citrine-PpSCR1 expression construct, the pENTR:Citrine-PpSCR1 plasmid (Fig. S7A) was subjected to LR reaction using the destination vector pPGX8 (AB537482) (12) to generate the GX8:Citrine-PpSCR1 plasmid. The generated construct was digested with the restriction enzyme PmeI for gene targeting and introduced into wildtype plants.
(B) Accumulation pattern of Citrine-PpSCR1 protein in gametophore apices from GX8:Citrine-PpSCR1 plants. Young gametophores of each line were cultivated in liquid BCD medium with or without 1 µM β-estradiol for three days and stained with calcofluor white. Optical longitudinal sections of apical parts of gametophores with leaf primordia are indicated. Magenta and cyan indicate the signals of Citrine-PpSCR1 and calcofluor, respectively. (Scale bar, 50 µm.) (C) Gametophores (upper) and gametophore apices (lower) of the GX8:Citrine-PpSCR1 and wild-type plants cultivated with or without 1 µM β-estradiol for 31 days. To observe gametophore apices, leaves surrounding gametophore apices were removed. Magnified images of these gametophore apices are shown in the lower panels. Red arrows indicate young leaves. Note that young leaves did not grown in the GX8:Citrine-PpSCR1 plants cultivated with 1 µM β-estradiol. (Scale bars, 1 mm in upper panels, 200 µm in lower panels.)  6) were streaked on the synthetic complete (SC) medium without leucine and tryptophane (SC-Leu-Trp; left) or the SC medium without leucine, tryptophane, and histidine (SC-Leu-Trp-His; middle). The latter contains 20 mM 3-amino-1,2,4-triazole (3AT) to suppress the auto-transcriptional activity of PpSHR1, PpSHR2, and PpSCR1 used as baits. Beta-galactosidase assay was also preformed (right).

Fig. S13. Phylogeny of LAS in land plants.
Maximum likelihood tree of LAS-related proteins. The two Physcomitrium patens proteins PpLAS1 and PpLAS2 are indicated with arrows. Phylogenetic analysis was performed using amino acid sequences found by BLAST searches using tomato (Solanum lycopersicum) LS (Lateral suppressor) as a query. The maximum likelihood tree under the LG model (1) with empirical amino acid frequency was searched with RAxML-HPC. The horizontal branch lengths are proportional to the estimated number of substitutions / site, and are drawn to scale. Bootstrap values over 50% are shown on each branch. INSD accession numbers or genome-specific identifiers are shown with the species name in which the sequence was present, sometimes followed by the protein name. The OTUs are colored according to the classification: blue, eudicots; blue violet, monocots; dark magenta, other seed plants; brown, other vascular plants (monilophytes and lycophytes); dark green, bryophytes.
LAS homologs were searched at NCBI using the tomato LS amino acid sequence NP_001234179.1 as a query as of April 7, 2021. The search was divided in three targets: 1) Streptophyta (taxid:35493) but excluding: Spermatophyta (taxid:58024), 2) Spermatophyta but not angiosperms, and 3) refseq protein of Arabidopsis, rice (Oryza sativa), and Amborella trichopoda. Minimum word size was set to 2, with a maximum hit list size of 1,000. These searches recovered 20, 10, and 10 sequences, respectively. The PpLAS coding sequences were subjected to BLASTX search against Physcomitrium patens proteins and revealed that PpLAS1 encodes a protein identical to XP_024377672.1, whereas PpLAS2 encodes a large portion identical of XP_024375390.1, but with a different N-terminal structure. PHYPA_007838 was also effectively identical to PpLAS2. Therefore, these three entries were removed from the dataset. Hornwort (25,26) and Zygnematales (27) homologs were searched on their respective recently published sets. Each of the four hornwort protein sets contained one high similarity gproteinene with 288-294 bits score of similarity, while the next one dropped to 249 bits. Thus, the top hit from Anthoceros agrestis Oxford isolate was retained. The hit to the two Zygnematales algae were limited to 230 bits and no sequence was taken.
The dataset was aligned with einsi in MAFFT version 7.475 (3). The alignment was imported to Mesquite version 3.61. (4). Redundant entries were not added, as Selaginella sequences were contained in both the representative genome set and the non-angiosperm nr dataset. All characters in the alignment were first set as "excluded". The alignments were reviewed on Mesquite and well-aligned regions were identified and set to "included". Sequences having large deletion(s) in the conserved region were removed. For the LAS dataset, Adiantum sequence MBC9838938.1 was removed due to a lack of otherwise conserved "RF" sequence. After "excluding" all characters, only well-conserved blocks were included to result in 247 amino acid sites. Each data matrix was saved as a nexus file and subjected to a custom maximum likelihood analysis pipeline involving RAxML-HPC (5). The "included" sites were extracted from the matrix and sequences identical in the included regions were treated as a single OTU during the analysis and reverted to the original names at the final stage. The extracted matrix was converted to PHYLIP format. At first, a molecular evolutionary model was chosen using ProteinModelSelection.pl, and in this case, LGF (LG+F) was chosen, where LG stands for the amino acid substitution matrix by Le and Gascuel (1) and F stands for using empirical amino acid frequency in the dataset. SEQBOOT (6) was used to prepare 1,000 resampling sets. The original and bootstrap replicates were individually processed with the "-f a -# 100" option in RAxML. The model was specified with "-m PROTGAMMALGF" according to the chosen model. Random number seeds -x and -p options were taken from the operating system with /dev/urandom. After all jobs finished, the resulting trees were recovered and the bootstrap frequency was calculated using CONSENSE (6). To force the topology to be the same as the tree obtained from the original dataset, the original tree was amplified 1,000 times and combined with the bootstrap results; after consensus calculation, 1,000 was subtracted from the result. Fig. S14. Construction of ∆pplas1, ∆pplas2, and ∆pplas1∆pplas2 plants.  (A and B) Schematics of construct targeting PpLAS1 and PpLAS2 loci. White boxes represent exons. Pink and cyan boxes denote the neomycin phosphotransferase II expression cassette (nptII) (7) and the aminoglycoside phosphotransferase IV expression cassette (aphIV) (8), respectively. Probes used in (C-E) are indicated as red horizontal lines. To assemble the PpLAS1 and PpLAS2 gene deletion constructs, genomic fragments containing the 5' flanking region and a partial coding sequence of each PpLAS gene were inserted into the 5' end of the nptII expression cassette of plasmid pTN182 for ∆pplas1 and of the aphIV expression cassette of plasmid pTN186 for ∆pplas2. The 3' flanking region of each PpLAS gene was inserted into the 3' region of the nptII and the aphIV expression cassettes of the resulting plasmids. The generated constructs were digested with suitable restriction enzymes for gene targeting and introduced into wild-type plants to generate ∆pplas1 (C) and ∆pplas2 (D) single deletion mutants.
(C) DNA gel blot analysis of ∆pplas1 plants. Genomic DNA of wild-type and ∆pplas1 (#23, #26, #27, and #28) plants was digested with HindIII. The arrowhead and arrow indicate DNA fragments specific to wild-type and ∆pplas1 plants, respectively. (D) DNA gel blot analysis of ∆pplas2 plants. Genomic DNA of wild-type and ∆pplas2 (#1, #16, #27, and #31) plants was digested with HindIII. The arrowhead and an arrow indicate DNA fragments specific to wild-type and ∆pplas2 plants, respectively. (E) DNA gel blot analysis of ∆pplas1∆pplas2 plants. Genomic DNA of wild-type and ∆pplas1∆pplas2 (#1, #2, #3, #4, and #5) plants was digested with HindIII. The arrowhead and arrow indicate DNA fragments specific to wildtype and ∆pplas1∆pplas2 plants, respectively. The PpLAS2 deletion construct was introduced into ∆pplas1#23 and ∆pplas1#26 plants to generate ∆pplas1∆pplas2#1 and ∆pplas1∆pplas2#2 to #5 double deletion plants, respectively. The circular plasmid pTN75 (AB542060), which contains a Cre recombinase expression cassette, was transiently expressed in the ∆pplas1∆pplas2#1 double deletion mutant plant by polyethylene glycol (PEG)mediated transformation. The nptII and the aphIV expression cassettes located between the two loxP sites were removed by the Cre recombinase. Loss of the antibiotic expression cassettes was confirmed on the basis of failure to grow on medium containing G418 or hygromycin. The newly generated plants were renamed as ∆pplas1∆pplas2-mf.  (28), the nopaline synthase terminator (nos-ter: gray) (7), the aminoglycoside phosphotransferase IV expression cassette (aphIV: cyan) (8), and the neomycin phosphotransferase II expression cassette (nptII: pink) (7) are shown in different colors. Probes used in (C and D) are indicated as red horizontal lines. To insert mClover3 in frame with the PpLAS1 coding sequence, a genomic DNA fragment from the middle to the last codon was PCR amplified from wild-type genomic DNA. The amplified fragment was inserted into the 5' end of the coding region of Citrine in the pCit-aphIV plasmid (LC703380) in-frame. A genomic fragment containing the 3' flanking region of PpLAS1 was inserted into the 3' region of the aphIV expression cassette. Citrine in the resulting plasmid was replaced with mClover3 in the pmClo3-nptII plasmid (ON092627) to place mClover3 in frame with the PpAS1 coding sequence. To insert mClover3 in frame with the PpLAS2 coding sequence, a PpLAS2 genomic DNA fragment extending from the middle to the last codon was PCR amplified from wild-type genomic DNA. The amplified fragment was inserted into the 5' end of the coding region of mClover3 in the pmClo3-nptII plasmid in-frame. Genomic fragments containing the 3' flanking region of PpLAS2 were inserted into the 3' region of the nptII expression cassettes of the resulting plasmids. The generated constructs were digested by suitable restriction enzymes for gene targeting. (C) DNA gel blot analysis of PpLAS1-mClover3 lines. Genomic DNA of wild-type and PpLAS1-mClover3 (#1, #2, #3, #4, and #5) plants was digested with BamHI. The arrowhead and arrow indicate DNA fragments specific to wild-type and PpLAS1-mClover3 plants, respectively. (D) DNA gel blot analysis of PpLAS2-mClover3 lines. Genomic DNA of wild-type and PpLAS2-mClover3 (#1, #7, #15, #32, #51, and #53) plants was digested with HindIII. The arrowhead and arrow indicate DNA fragments specific to wild-type and PpLAS2-mClover3 plants, respectively.  (11), loxP (green) (9), the nopaline synthase terminator (nos-ter: gray) (7), and the aminoglycoside phosphotransferase IV expression cassette (aphIV: cyan) (8) are shown in different colors. Probes used in (D and E) are indicated as red horizontal lines. To insert Citrine at the 3' ends of PpSHR1 (A) and PpSHR2 (B), the PpSHR1-Citrine (Fig. S9A) and PpSHR2-Citrine (Fig. S3C) expression constructs were introduced into ∆pplas1∆pplas2-mf#12 plants (Fig. S14), respectively. (C) Schematic diagram of the strategy used to insert mClover3 at the 3' end of the endogenous PpLAS1 gene. White boxes represent exons. Green, gray, and orange boxes denote mClover3 (28), the nopaline synthase terminator (nos-ter) (7), and the zeocin resistance cassette (zeo) (16), respectively. The probe used in (F) is indicated as a red horizontal line. The aphIV expression cassette in the PpLAS1-mClover3 expression construct (Fig. S15A) was replaced with the zeocin resistance cassette. The generated construct was digested with suitable restriction enzymes for gene targeting and introduced into ∆ppshr1∆ppshr2#7 plants.   (12). The connected DNA fragment of the LexA operator and minimal 35S promoter (blue: LexAop:m35S) (13), PpLAS1 (green), Citrine (yellow) (11), the pea rbcS3A terminator (black: pea3A-ter) (14), the putative ProGX8 promoter (orange) (12), a DNA fragment encoding an XVE fusion protein derived from pER8 (purple) (13), the rbcS terminator (rbcS-ter: gray) (14), and the aminoglycoside phosphotransferase IV expression cassette (aphIV: cyan) (8) are shown in different colors. To produce the PpLAS1-Citrine induction construct, the PpLAS1-coding sequence was inserted into the KpnI and ClaI sites of the pCit-aphIV plasmid (LC703380), resulting in the generation of the in-frame PpLAS1-Citrine fusion. The PpLAS1-Citrine fragment was then PCR amplified and inserted into the pENTR/D-TOPO vector (ThermoFisher Scientific) to generate the pENTR:PpLAS1-Citrine plasmid. The pENTR:PpLAS1-Citrine plasmid was subjected to LR reaction using the destination vector pPGX8 (AB537482) (12) to generate the GX8:PpLAS1-Citrine plasmid. The generated construct was digested with the restriction enzyme PmeI for gene targeting and introduced into wild-type plants.   (7) and aminoglycoside phosphotransferase IV expression cassette (aphIV) (8), respectively. Green boxes denote loxP sequences (9). Probes used in (C and D) are indicated as red horizontal lines. To visualize microtubules, the GFP--Tubulin expression construct (8) was introduced into wild-type plants to generate the GFP--Tubulin#84 (GTU84) line. The PpSHR1 gene deletion construct (Fig. S2A) was introduced into the GTU84 plant to generate GTU84∆ppshr1∆ppshr2 plants (C). The GTU84∆ppshr1#16 plant was used to generate the GTU84∆ppshr1∆ppshr2 double deletion mutants. (C and D) DNA gel blot analyses of targeted plants with the probes indicated in (A and B). Genomic DNA of GTU84, GTU84∆ppshr1 (#16 and #26), and GTU84∆ppshr1∆ppshr2 (#1, #7, #10, #20, #33 and #34) was digested with EcoRV. Arrowheads and arrows indicate DNA fragments specific to GTU84 and GTU84∆ppshr1 or