The control of axillary meristem fate in the maize ramosa pathway

Abstract

Plant axillary meristems are composed of highly organized, self-renewing stem cells that produce indeterminate branches or terminate in differentiated structures, such as the flowers. These opposite fates, dictated by both genetic and environmental factors, determine interspecific differences in the architecture of plants. The Cys(2)-His(2) zinc-finger transcription factor RAMOSA1 (RA1) regulates the fate of most axillary meristems during the early development of maize inflorescences, the tassel and the ear, and has been implicated in the evolution of grass architecture. Mutations in RA1 or any other known members of the ramosa pathway, RAMOSA2 and RAMOSA3, generate highly branched inflorescences. Here, we report a genetic screen for the enhancement of maize inflorescence branching and the discovery of a new regulator of meristem fate: the RAMOSA1 ENHANCER LOCUS2 (REL2) gene. rel2 mutants dramatically increase the formation of long branches in ears of both ra1 and ra2 mutants. REL2 encodes a transcriptional co-repressor similar to the TOPLESS protein of Arabidopsis, which is known to maintain apical-basal polarity during embryogenesis. REL2 is capable of rescuing the embryonic defects of the Arabidopsis topless-1 mutant, suggesting that REL2 also functions as a transcriptional co-repressor throughout development. We show by genetic and molecular analyses that REL2 physically interacts with RA1, indicating that the REL2/RA1 transcriptional repressor complex antagonizes the formation of indeterminate branches during maize inflorescence development. Our results reveal a novel mechanism for the control of meristem fate and the architecture of plants.

Fig. 1.

Inflorescence development in normal and ra1-RS plants. (A) Wild-type tassel. (B) ra1-RS tassel. (C) Wild-type ear. (D) ra1-RS ears; arrowhead indicates the masculinized tip. (E-H) Scanning electron microscopy (SEM) of wild-type and ra1-RS inflorescences. (E) Wild-type tassel with branch meristems (arrowheads) and spikelet-pair meristems (arrows). (F) ra1-RS tassel with extra developing branch meristems (arrowheads). Wild-type (G) and ra1-RS (H) ears. Scale bars: 500 μm.

Fig. 2.

The rel2 mutant is a genetic enhancer of ramosa1 and ramosa2 mutants. (A,B) Double and single rel2 mutant inflorescences (tassel, left; ear, right). (C) SEM of an immature rel2;ra1-RS ear, showing a proliferation of branch meristems (arrowheads) in place of spikelet-pair meristems. Scale bar: 200 μm. (D) rel2;ra1-63.3359 ear phenotype. (E) ra1-63.3359 mutant ear. (F) Heterozygous effect of rel2 mutation in ra1 mutants. Example of a dose effect of rel2 in ear branching. (G) rel2 enhances (left) the branched ear phenotype of a weak allele of ra2 (right). (H) Triple rel2;ra2;ra1 mutant ear. Primary branches form indeterminate secondary (right) and occasionally tertiary branches (arrowheads).

Fig. 3.

The upright tassel branch phenotype of rel2 and ra2 mutants. (A-F) Cross-sections of the branch nodes in rel2 and rel2;ra1-RS mutants, stained with Saffranin O and Alcian Blue. (G) ra2 mutant tassel. (H,I) Cross-section of ra2 tassel branch nodes. White arrowheads indicate vascular strands (B,D) or lack of vascular strands (C). Black arrowheads indicate pulvinal cells. Scale bars: 200 μm.

Fig. 4.

REL2 encodes a transcriptional co-repressor. (A) Schematic representation of the REL2 gene (above) and of the encoded protein (below). The three isolated independent rel2 mutant alleles and the corresponding mutations are indicated. White boxes represent exons. Colored boxes represent different domains of the REL2 protein: lyssencephaly type1-like homology (LisH); C-terminal to LisH (CTLH). (B-F) In situ hybridization of REL2 during inflorescence development. (B) Young developing ear. (C) Spikelet-pair meristems. (D) Floral meristems developing floral organs; REL2 expression is also visible in the vasculature (arrowheads). (E,F) Branch meristem (arrowhead) of a young developing tassel; antisense (E) and sense control (F). Scale bars: 50 μm. (G) REL2 rescues the tpl-1 phenotype of Arabidopsis. One representative line is shown (tpl-1 R). (H-J) Confocal images of wild-type, tpl-1 and tpl-1 embryos rescued by REL2-YFP (tpl-1 R). (K,L) In planta repression assay. We transformed an Arabidopsis line carrying the reporter construct 2xUAStCUP::GUS (Szemenyei et al., 2008) (K) with another construct expressing REL2 fused to the DNA-binding domain (DB) of the yeast GAL4 transcriptional activator (pTPL::REL2:GAL4DB:HA). Staining in cotyledons shows that REL2:GAL4DB (L) is capable of repressing the expression of GUS.

Fig. 5.

The ra1RSenh allele is an intragenic enhancer of ra1-RS. (A) ClustalW alignment of the EAR repressor motifs of ra1-RSenh, ra1-RS and the SUPERMAN (SUP) proteins of Arabidopsis.(B) ra1-RSenh highly branched ear. (C) Scanning electron microscopy image of an immature ra1-RSenh ear showing a proliferation of branch meristems (arrowheads). Scale bar: 500 μm.

Fig. 6.

REL2 and RA1 interact in vivo and in vitro. (A) Schematic representations of the modified versions of REL2 and RA1 proteins used in protein-protein interaction assays. (B) Targeted yeast 2-hybrid assay. Yeast transformants are grown in the selection medium (–LT). The interaction is tested on histidine-lacking medium (–LTH) and by β-galactosidase (βgal) activity. AD, GAL4 activation domain; DB, GAL4 DNA-binding domain. (C) In planta bi-fluorescent complementation assay. Transient expression in tobacco leaves. (D) Pull-down assay of in vitro transcribed/translated RA1:HA3 (input). (E) Pull-down assay of in vivo transcribed/translated RA1:MYC:NYFP (input). In D,E, input lanes represent one-tenth of the volume used for the assay.

Fig. 7.

Proposed model for the repression of the indeterminate fate of spikelet-pair meristems. REL2 and RA1 form a complex, interacting via the two EAR motifs (grey) and the CTLH domain (yellow). This complex represses the expression of target genes (grey solid bar), resulting in the determinacy of spikelet-pair meristems. RA2 and RA3 have been previously shown to regulate RA1 transcript levels (Bortiri et al., 2006; Satoh-Nagasawa et al., 2006). REL2 and RA2 also differently affect, independently from RA1, the angle of tassel branches.