The role of DORNRÖSCHEN-LIKE in early floral organogenesis

Abstract

Positional signals that specify founder cells and determine where lateral organs initiate and how these signals are perceived by cells that transition to the periphery of the meristem is a challenging problem. We recently showed that expression of the AP2 ERF transcription factor DORNRÖSCHEN-LIKE (DRNL) marks all floral organ founder cells and pre-patterns lateral stamen and petal, or medial stamen founder cells by two regions of expression that we propose represent morphogenetic fields, that subsequently resolve into discrete foci. The spatio-temporal expression pattern of DRNL allows speculation concerning evolutionary aspects of plant developmental biology and the control of the floral plant body. It further paves the way to use DRNL as a tool to address fundamental questions of cell type specification.

The Problems of Founder Cell Specification

Lateral organ founder cell specification involves the perception of positional information by cells or groups of cells in the peripheral zone of the meristem, so-called founder cells. The number of these cells is conceptually based on estimations derived from sector analysis, and remains unconfirmed for many organs such as floral organs, due to the lack of founder-cell-specific markers. These markers are hard to pinpoint, since founder cells are not histologically or morphologically identifiable. Although lateral root founder cell specification has been well-studied,¹ specification signals for floral organs remain largely obscure. Auxin can act as a trigger for organ specification, such as lateral roots² and auxin maxima correlate with sites of incipient leaf initiation at the flanks of the SAM.³ It is, however, becoming clear that competencies need to be in place for cellular responses to signals such as auxin, for example, conferred by opposing waves of oscillating transcription factor expression for lateral roots.⁴ Further evidence suggests that repression of cytokinin signalling might be as important a signal as auxin⁵ and opens debates concerning the correlative and interactive nature of signals involved in founder cell specification. Broad competencies that enable cells to respond to more specific signals might reside in morphogenetic fields from which founder cells are specified. The generation of auxin gradients⁶ suggests it might act as a morphogen, but functional evidence for this is lacking, due to difficulties in demonstrating that cellular responses are dependent on differential concentrations. Recently, we characterised the expression of DORNRÖSCHEN-LIKE (DRNL) which encodes an AP2/ERF protein and marks floral organ founder cells and we believe that it contributes to positional determination for floral organ initiation.⁷ This addendum focuses on the evolution and evolutionary developmental implications of DRNL.

DRNL Expression Marks Floral Morphogenetic Fields and Founder Cells

3-D video imaging revealed two important aspects to DRNL expression. Firstly, it identifies two distinct domains—one for lateral stamens and petals and a ring-shaped domain for medial stamens—and subsequently marks groups of cells for all floral organs that correspond to the number estimated for floral organ founder cells by sector analysis.⁸ Thus, DRNL marks what can be considered as a broad potential morphogenetic field that subsequently resolves into discrete foci that correlate with founder cell specification. Boundary formation is a concomitant and inseparable process to organ initiation and outgrowth, since surrounding cells must be repressed from coordinated divisions. Synergistic deficiencies resulting from loss of DRNL function together with that of individual CUC functions that help to specify floral organ boundaries, also demonstrates that DRNL function contributes to early organogenesis. The fact that DRNL expression precedes maxima in auxin transcription response as monitored by the DR5 reporter and that mutation of AuxREs in the DRNL promoter does not affect early DRNL expression, suggests that at least canonical auxin responses are not the earliest events in floral founder cell specification and that DRNL represents a novel signal involved in the earliest organogenesis events.

Evo-Devo Implications of DRNL Function

Using DRNL expression to monitor organ floral organ initiation leads to important conclusions concerning organ initiation in Arabidopsis, which occurs in a clear series, beginning with the abaxial sepal, then lateral sepals are simultaneously initiated in a decussate pattern, followed by the adaxial sepal. Separate developmental programmes exist for lateral and medial stamens, the former arising in tandem proximity with petals. The recruitment of lateral stamen and petal founder cells in close proximity from within adjacent or even overlapping domains and medial stamens from within a single domain is reflected in the stamen fusion phenotypes of the drnl mutant, which are preferentially restricted to amongst medial stamens. The sensitized organ fusions arising from drnl cuc double mutants supports a model whereby the resolution of organ founder cells from broader domains is compromised by loss of DRNL function. Further functional corroboration for a model of organ specification from broad competency domains comes from wild type flowers, which demonstrate stamen or petal fusions or mixed petal/stamen identities at low frequency (<1% flowers) (Fig. 1), suggesting that even the normal recruitment of founder cells involves imprecise realisation of signals.

Figure 1.

Wild type floral phenotypes, interpreted as problems in resolving morphogenetic fields. A single petalloid stamen (arrow) (A), two stamens completely fused along their filaments, with two anthers clearly visible (arrows) (B), two stamens partially fused along their filaments.

Flowers vary enormously in their ground plan between species. Signals such as those provided by DRNL are relevant for understanding such variations; for example, the Brassicaceae contains much floral diversity,⁹ and evolutionary adaptations in boundary regulation from morphogenetic fields might be one explanation for the absence of lateral stamens in Lepidium species and fusion of medial stamens in Vella. Organ fusion has been a driving force in evolution, to provide fused corolla, androecia or carpels, and stamen fusions occur within 70 families of seed plants.¹⁰ DRNL expression sheds light on current whorl models of the floral ground plan in Arabidopsis, which could be refined to incorporate asynchronous initiation of single organs and temporally overlapping primordia initiation. Many plant families have flowers with more complicated ground plans than Arabidopsis: petals are unidirectionally initiated from abaxial to adaxial in the Caesalpinioideae or helically in Gleditsia triacanthos¹¹ and petal and stamen primordia arise from single common primordia in Pisum¹² and legumes.¹³ Models have to explain overlapping floral organ initiation, where different signals or competencies must exist simultaneously and how same-whorl organs differentiate different structures. Potentially, orthologous DRNL promoter activities could illuminate mechanisms underlying these different modes of initiation.

Using DRNL Address Fundamental Developmental Questions

The subtleties of spatio-temporal DRNL expression combined with that of its paralogue DRN, allow further speculation to be made concerning important developmental questions. One question concerns bract formation. Initial DRNL expression at the abaxial sepal position might mark a cryptic bract, which is repressed from outgrowth in Arabidopsis.¹⁴ This is supported by the subsequent bifurcation of this expression domain into two sub-domains,⁷ the upper of which might represent expression in the true abaxial sepal, which initiates in the rudimentary bract axils^15,16 and the lower, transcript persistence in the rudimentary bract. PUCHI, an AP2/ERF protein closely-related to DRNL within clade VIIIb¹⁷ has a role in bract formation and is transiently expressed on the adaxial side of the IM at stages 1 and 2, when the FM buttresses away from the IM.¹⁸ PUCHI, DRNL and DRN uniquely share a conserved C-terminal CMVIII-3 motif within this clade.¹⁷ It will be interesting to establish any functional link between DNRL and PUCHI in bract formation and characterize the extent of domain overlap for both genes. The PUCHI orthologues FRIZZY PANICLE (FZP) in rice¹⁹ and BRANCHED SILKLESS1 (BD1) in maize²⁰ have extreme floral phenotypes, demonstrating a conserved floral function for PUCHI. The weaker Arabidopsis puchi phenotype suggests possible redundancy with DRNL or DRN.

A second related question addresses the debate concerning de novo or detached meristems. Notably, the discontinuity in DRN expression from the IM centre to the FM centre with an absence in stage 1 FMs might correspond with bract initiation, and re-initiation of DRN transcription with de novo meristem formation from the bract initial as suggested by SHOOT MERISTEMLESS and AINTEGUMENTA expression¹⁴ as opposed to theories of the FM being a detached meristem from the IM.

Both the isolation of direct gene targets of DRNL and the use of DRNL as a tool to unravel early aspects of floral organ initiation, particularly via the isolation of upstream transcriptional DRNL regulators to discover what establishes and then restricts DRNL expression domains aims at the heart of understanding floral organ positional signals.

Evolution of DRNL

DRNL and DRN are paralogs that presumably have arisen by recent gene duplication in Arabidopsis and divergence. As DRN can functionally complement a strong drnl mutant under control of the DRNL promoter, divergence has mainly occurred at the level of transcription and both genes have acquired almost mutually exclusive expression domains in flowers.⁷ Phylogenetic analysis of the AP2 domains of most closely related sequences from sequence databases show two clades of putative DRN or DRNL orthologues (Fig. 2). Interestingly, monocot sequences align most closely only with the DRN AP2 domain, but no monocot mutants for these genes have been characterised. In contrast, mutation of FZP and BD1 in rice and maize already mentioned, have extreme floral phenotypes relative to that of the Arabidopsis PUCHI orthologue. It can be speculated that the regulation of floral organ development in monocots has evolved to be mainly via a PUCHI-type function and predominantly by DRNL in eudicots.

Figure 2.

Phylogenetic relationship between DRN and DRN L putative orthologues from different species. Unrooted phylograms of the AP2 domain of ER F proteins most closely related to DRNL from available species, obtained by pblastp and pblastn from NCBI. Sequences were aligned by Clustal W and the tree constructed using the Neighbour Joining method.

Abbreviations

DRNL

DORNRÖSCHEN-LIKE

DRN

DORNRÖSCHEN

IM

inflorescence meristem

FM

floral meristem

AuxRE

auxin response element

CUC

cup-shaped cotyledon

STM

SHOOT MERISTEMLESS

ANT

AINTEGUMENTA

FZP

FRIZZY PANICLE

BD1

BRANCHED SILKLESS1

Addendum to: Chandler J, Jacobs B, Cole M, Comelli P, Werr W. DORNRÖSCHEN-LIKE expression marks Arabidopsis floral organ founder cells and precedes auxin response maxima. Plant Mol Biol. 2011;76:171–185. doi: 10.1007/s11103-011-9.