TCP transcription factor, BRANCH ANGLE DEFECTIVE 1 (BAD1), is required for normal tassel branch angle formation in maize

Characterization of the Upright Tassel Mutants.

Through screens of EMS-generated M2 populations, we identified five different recessive mutants in maize that affect tassel branch angle, namely SLO13, A619TR-501, SLO84, SLO271, and SLO365 (Fig. S3 A–F). Pair-wise crosses indicate that SLO13 and A619TR-501 are allelic to each other and to ra2 (Fig. S3G). Map-based localization and complementation tests showed that SLO84 is an allele of LIGULELESS1 (LG1) (Fig. S3H), which encodes a nuclear-localized protein required for induction of ligules and auricles during maize development (5). SLO271 and SLO365 were initially mapped to the same locus on chromosome 2 and subsequently were shown to be allelic. We named the SLO271 and SLO365 alleles branched angle defective 1-1 and 1-2 (bad1-1 and bad1-2), respectively.

The reference line used in our EMS mutagenesis experiment produces maize tassels with an average angle around 36° between the central spike and the lateral branch and is typical of many maize lines (Fig. 1A). bad1 mutants show the upright tassel phenotype, caused by a significant reduction in the first and second lateral branch angles compared with their wild-type siblings (Fig. 1 B and C). There are no apparent phenotypic differences in other parts of mutant plants. In mutant tassels, the pulvinus, which is clearly visible between the main spike and the lateral branch in the wild-type plant (Fig. 1D), is reduced in size or nearly absent (Fig. 1E).

To visualize the pulvinus at a cellular level, we cross-sectioned samples and stained with safranin O and fast green, resulting in nuclei, nucleoli, and xylem being stained red by safranin O, whereas other cellular components were stained green by fast green. The tissue corresponding to the pulvinus stained green and was obviously missing or reduced in mutants compared with wild-type siblings (Fig. 1 F and G).

To determine the distribution of lignin in the pulvinus, we hand-sectioned fresh tissue and stained with 18% (wt/vol) phloroglucinol in HCl, which stains lignified secondary cell walls red. The pulvinal tissue remained blue in the presence of phloroglucinol, indicating that it is composed of unlignified parenchyma cells (Fig. 1 H and I). To determine if enlargement of the pulvinus occurs by cell division or by cell expansion, or both, we sectioned the pulvinus at early and late stages of wild-type tassel development. Early in development [44 d after germination (DAG)] when the tassel is around 11 cm long, the pulvinus contains around 10–12 cell layers (Fig. S4 A and C), compared with 18–20 layers 18 d later when the tassel grows to around 30 cm long (Fig. S4 B and D). Because the size of the mature parenchyma cells in the pulvinus is similar at different stages, we conclude that the increase in size of the pulvinus between 44 DAG and 62 DAG is due to cell proliferation and not to cell expansion.

To determine if the pulvinus is composed of meristematic cells or, instead, a part of a differentiated lateral organ, we undertook in situ hybridization with the KNOTTED1 (KN1) homeobox gene on sections of developing tassels. KN1 is normally expressed in several types of indeterminate meristem tissues but not in determinate organs such as leaves (6, 7). As expected, we observed KN1 expression in meristems but not in the pulvinus. Thus, we infer that the pulvinus is an extension of a determinate lateral organ and is not meristematic.

Positional Cloning of BAD1 and Phylogenetic Tree.

BAD1 alleles were initially mapped to chromosome 2 on the basis of the linkage to the simple sequence repeat (SSR) markers umc1004 and umc1749. By screening a mapping population of 146 mutant individuals, BAD1-2 was mapped between the molecular marker Si603014H11 (two recombinants/292 chromosomes) and umc1749 (three recombinants/292 chromosomes). Syntenic regions in rice and sorghum were identified, and markers were developed for analyzing corresponding maize genes (Fig. 2A). The newly developed molecular markers PCO066968 and PCO100185 placed the BAD1 locus in a region ∼1.01 Mbp on contig 96. Several candidate genes fall in this region in rice and sorghum, including one encoding a predicted protein with unknown function, containing a conserved TCP domain common to a plant-specific TCP family of transcription factors named after TEOSINTE BRANCHED1 (TB1) from maize, CYCLOIDEA (CYC) from snapdragon, and PROLIFERATING CELL NUCLEAR ANTIGEN FACTOR (PCF) from rice (8, 9). Members in this gene family are involved in plant growth, cell proliferation, organ identity, and circadian clocks (10–15). Sequencing the candidate gene showed that the bad1-2 mutant contained a point mutation in the highly conserved TCP domain, causing a serine-to-phenylalanine change at amino acid 134 from the start codon (Fig. 2 B and C). The bad1-1 mutant contained another point mutation in the TCP domain resulting in a threonine-to-methionine substitution at amino acid 122 (Fig. 2 B and C). Both point mutations were caused by transitions from cytosine to thymine and are consistent with the expected mutations from EMS mutagenesis.

Genes in the TCP family fall into two classes on the basis of their protein sequences. Class I proteins have a four-amino-acid deletion in their conserved TCP domains compared with class II members (16). Class II can be further divided into two subclades, the CYCLOIDEA/TB1 (CYC/TB1) clade and the CINCINNATA (CIN) clade, on the basis of conserved residues in each subclade (Fig. S5) (16). Signature residues place BAD1 in a subclade of CYC/TB1 (named as TCP CII) (Fig. 2C), a placement confirmed by phylogenetic analysis (Figs. S5 and S6). The phylogeny shows two successive duplication events within the grasses (Fig. S6). The first event occurred at the base of the grasses, giving rise to one clade with TB1-like sequences, and another clade that is again duplicated, resulting in two subclades, one containing OsTCP24-(REP1) from rice, GRMZM2G110242 (BAD1) and GRMZM2G064628 from maize, and Sb02g024450 from sorghum. The other subclade contains OsTCP22 from rice, GRMZM2G166687 and GRMZM2G059636 from maize, and Sb07g021140 from sorghum. Interestingly, BAD1 is homologous to OsTCP24, also called RETARDED PALEA1 (REP1) (17). REP1 in rice controls palea identity and development and is responsible for floral zygomorphy. Whether REP1 regulates lateral branch angle in rice has yet to be determined.

Expression of BAD1.

BAD1 expression was detected by RT-PCR in all tissues analyzed, with higher expression in the developing tissue of tassel and leaf (Fig. 3A). To investigate the site of BAD1 expression, RNA in situ hybridization was performed on immature tassels and ears. Early in development, BAD1 expression was observed in the tips of the leaf primordia (Fig. 3 B and C), spikelet pair meristems, spikelet meristems (Fig. 3E), and the tips of husk leaves (Fig. 3 F and G). To analyze the accumulation of BAD1 protein, an antibody was generated against the full-length BAD1 protein. Immunolocalization shows BAD1 protein in the same tissues as the RNA. BAD1 is undetectable in the tassel branch axis before the pulvinus is formed (Fig. 3I), but later appears in a few cells adaxial to the lateral branch where the pulvinus is developing (Fig. 3J).

BAD1 Is Down-Regulated in ra2 Mutants.

Because the mutations that we uncovered in BAD1, LG1, and RA2 all gave rise to upright tassel phenotypes, we wanted to know whether these genes work in the same or in parallel pathways. To this end, we performed both genetic and gene expression analysis. Quantitative RT-PCR (qPCR) showed that BAD1 expression in developing tassels is reduced by about half in a ra2 mutant compared with that in the isogenic A619 inbred background (Fig. 4A). Western blots using the BAD1 antibody confirmed that the amount of BAD1 protein is reduced in ra2-developing inflorescences compared with wild-type plants, suggesting that RA2 is an upstream positive regulator of BAD1 in controlling tassel lateral branch angle development (Fig. 4 B and J). To further confirm the epistasis between BAD1 and RA2, double mutants were made between bad1-1 and ra2. The tassels of bad1-1 single mutants, ra2 single mutants, and bad1 ra2 double mutants are all upright (Fig. 4 D–F). Lateral branch number, first lateral branch angle, and second lateral branch angle were quantified in wild-type plants, bad1-1 single mutants, ra2 single mutants, and bad1 ra2 double mutants. Lateral branch number was almost twofold higher in ra2 mutants compared with the isogenic A619 inbred line and was reduced nearly 60% in the bad1 mutant (Fig. 4G). However, the bad1 ra2 double mutant is intermediate in branch number compared with ra2 and bad1, suggesting an additive effect of these two genes on tassel branch number. Although lateral branches are more upright in ra2 and bad1 mutants than in wild type, the branch angle is significantly more acute in bad1 than in ra2 (Fig. 4 H and I). Significantly, lateral branch angle in the bad1 ra2 double mutants is nearly identical to that observed in bad1 mutants alone (Fig. 4 H and I), suggesting that bad1 is epistatic to ra2 with regard to tassel branch angle.

To determine the relationship between LG1 and BAD1, qPCR and Western blots were also done for lg1 mutant inflorescences. BAD1 expression level was similar in lg1 mutants and wild-type plants (Fig. 4 A and B). Also, no significant differences were found in tassel lateral branch angle or lateral branch number in bad1-1 mutants, lg1 mutants, and bad1 lg1 double mutants. The bad1 lg1 double mutants have not only an upright tassel phenotype but also upright leaves; the latter phenotype is similar to the lg1 mutant, but is not observed in bad1 mutants. We infer that ra2 and lg1 work in concert to control pulvinus development.

Plant Materials and Traits Analysis.

bad1, ra2, and lg1 mutants were identified as segregating upright tassel branch mutants in genetic screens for an ethyl methanesulfonate (EMS)-generated M2 family. SLO13 (ra2), SLO84 (lg1), and two bad1 alleles, SLO271 (bad1-1) and SLO365 (bad1-2), were isolated from EMS-mutagenized ra1-RS lines originally constructed to screen for modifiers of the weak ra1-RS allele (22). The ra2 (03IL-A619 TR-501) point mutation was identified by sequencing the ra2 gene from mutant plants segregating in an EMS-generated M2 family (http://www.maizegdb.org/mip). Both bad1-1 and bad1-2 were backcrossed five times into B73 and A619 for phenotypic analysis. Crosses between bad1 alleles resulted in the same phenotype, demonstrating that they are allelic.

For the double-mutant analysis, we used the following alleles: bad1-1; ra2-ref introgressed in A619, and bad1-1;lg1-SLO84 in a ra1-RS background (1, 22). The number of lateral branches was determined by counting all primary branches. The first and second lateral branch angles were measured from the basal side of the tassel. The statistical significance of the differences in branch number and lateral branch angle were analyzed using Student’s t test.

Positional Cloning of BAD1.

For positional cloning, an F2 mapping population was generated by crossing the bad1-2;ra1-RS double mutant with the maize inbred line A619. The BAD1 region was delimited by screening 146 F2 homozygous mutants using SSR markers. Syntenic regions in rice and sorghum were identified using rice (http://www.gramene.org) and sorghum (http://www.phytozome.net) sequences. New maize markers were designed using dCAPS (http://helix.wustl.edu/dcaps/dcaps.html) and genome sequence information available at MAGI (http://magi.plantgenomics.iastate.edu).

RT-PCR and in Situ Hybridization.

For RT-PCR, five individual samples were pooled for each tissue tested (0.2-cm ears, 1.5-cm ears, 1.0-cm tassels, 2.0-cm tassels, 21-DAG root tips, 0.8-cm spikelets, 21-DAG shoot tips, immature leaves, 21-DAG cotyledons, internodes, and ligules). Total RNA was extracted using the RNeasy Plant Mini Kit (Qiagen). One microgram of RNA was treated with DNase I (Promega). Template cDNA samples were prepared using Moloney Murine Leukemia Virus (M-MLV) reverse transcriptase (Invitrogen) for reverse transcription with 1 μg of total RNA in a reaction volume of 20 μL. The cDNA synthesis reaction mixture was diluted by a factor of 10 before being used for PCR. Tissue preparation and in situ hybridization were performed as previously described (7).

Histochemical Staining of the Pulvinus in Wild-Type and Mutant Tassels.

For histological sections of the pulvinus, the first tassel branch node from the basal side of the tassel at different stages before anthesis was fixed in formalin–acetic acid–alcohol (FAA), dehydrated in an ethanol series, treated with HistoClear, and embedded in paraffin. Sections (8 μm) were placed on slides. After deparaffinization of the paraffin-embedded sections and serial rehydration steps, sections were stained with 1% safanin O for 25 min, then dehydrated to 95% ethanol, counterstained with 0.5% fast green in 95% ethanol for 15 s, washed with 95% ethanol, dehydrated, cleared, mounted, and examined by light microscopy. The nuclei, nucleoi, and xylem are red and all other components of the cells are green.

Fresh hand sections of the pulvinus at the first tassel branch were stained with 18% phloroglucinol in HCl. The lignified tissues appear in red (Fig. 1 H and I), and parenchyma tissues are not stained. Images were acquired with Adobe Photoshop 7.0 software.

Real-Time qPCR.

qPCR was performed using a LightCycler system from Roche Diagnostics. Five 1.0-cm tassels from the A619 inbred line, ra2 mutants in A619, and lg1 mutants in an unknown background (lg1 205B from Maize Genetics Cooperation Stock Center) were pooled for RNA isolation. The expression level of each gene represents the average of three replicates of three distinct biological pools relative to actin. The LightCycler FastStart DNA Master SYBR Green I Q-PCR kit (Roche Diagnostics) was used for PCR reactions. Quantification was performed with LightCycler Relative Quantification software 1.0 (Roche Applied Science).

Generation and Purification of BAD1-Specific Antibody.

Sequence encoding the full-length BAD1 was coupled to a HIS tag. Recombinant protein was isolated under denaturing conditions, gel purified, and injected into rabbits (Pacific Immunology). The peptide was coupled to Aminolink Plus Coupling Resin (Thermo Scientific) and affinity-purified using full-length purified BAD1 protein and elution buffers according to the manufacturer’s protocols (Pierce/ThermoFisher Scientific). Western blots and immunolocalizations were obtained using purified antibodies.

Immunolocalization and Confocal Microscopy.

The first tassel lateral branch from the basal side of the tassel was fixed and sectioned for immunolabeling. After blocking for 30 min in 1× PBS with 0.05% TWEEN 20 and 1% BSA, the tissue was incubated in rabbit anti-BAD1 diluted 1:500 in blocking solution at 4 °C overnight; the signal was detected with Alexa Fluor 488-conjugated anti-rabbit IgG (Invitrogen) diluted 1:200 in blocking solution for 4 h. Confocal microscopy was performed with a Leica SP/2 inverted microscope. Image analysis was done with the Leica SP/2 software package and the ImageJ bundle provided by the Wright Cell Imaging facility.