Most angiosperm flowers are organized in four concentric whorls: sepals (W1), petals (W2), stamens (W3) and carpel/s (W4), (Smyth et al., 1990). The number of floral organs and the placement of the organs within each whorl are genetically determined and MADS-box floral homeotic genes play a crucial role in the specification of floral organ identity (Bowman et al., 2012). Other classes of genes, the ones that determine the boundaries of different cell identities, are also crucial players during floral development (Yu and Huang, 2016). They were classified as “cadastral genes”, to which the Arabidopsis SUPERMAN (SUP) gene was assigned (Bowman et al., 1992; Sakai et al., 1995).

SUP is a transcriptional repressor, extensively studied in Arabidopsis thaliana, that encodes a plant-specific EPF-like protein with one Cys₂-His₂ zinc finger DNA binding domain and a C-terminus EAR-like (DLELRL) motif (Sakai et al., 1995; Hiratsu et al., 2002; Hiratsu et al., 2003; Hiratsu et al., 2004). The specific expression of SUP at the boundary between W3 and W4 (Sakai et al., 1995; Prunet et al., 2017) led to its classification as a cadastral gene specifying the stamens-carpel boundary. The supernumerary male organs at the expense of the female one of the sup-1 mutant ( Figure 1A ) (Bowman et al., 1992) was initially associated with the expansion of the MADS-box genes APETALA3 (AP3) and PISTILLATA (PI) expression closer to the centre of the floral meristem (Sakai et al., 1995; Prunet et al., 2017). Initially, models to explain the SUP function were based on a single allele: sup-1 (flo-10) (Bowman et al., 1992; Sakai et al., 1995; Prunet et al., 2017). However, the study of other sup alleles displaying phenotypes deviating from sup-1 ( Figure 1A ) has shed light on the SUP functions (Breuil-Broyer et al., 2016; Prunet et al., 2017; Xu et al., 2018). SUP is a gene controlling the stamens-carpel boundary setting and is linked to floral meristem termination (FMT) at the early stages of flower development.

It has been described that SUP is required for the correct timing to turn off WUSCHEL (WUS) from the floral meristem centre (FMC), thus controlling the floral meristem termination. WUS activity is required for the stem cell division at the floral meristem centre, a prolonged expression of WUS would lead to a delayed floral meristem termination, and more floral organs could be produced. However, SUP and WUS do not show an overlapping spatial expression and, in sup mutants, WUS expression is prolonged (Prunet et al., 2017; Xu et al., 2018). Moreover, SUP contributes to carpel medial region formation and the tissues derived from this region (Breuil-Broyer et al., 2016).

These studies provided new information to generate different models to explain SUP functions. One of the models proposes that SUP indirectly promotes floral meristem termination by repressing B-class genes. This model explains the different sup alleles phenotypes ( Figure 1A ), showing an indistinct male-female boundary and a sporadic carpel development (sup-1). By contrast, the increased number of stamens and carpels in the sup-5 mutant ( Figure 1A ) supports a second model that proposes that SUP controls the balance of cell proliferation and differentiation at W3 and W4 (Breuil-Broyer et al., 2016). To this regard, the effect of the overexpression of SUP-like genes supports the activity of SUP as a cell proliferation control gene (Nandi et al., 2000; Bereterbide et al., 2001; Hiratsu et al., 2002; Kazama et al., 2009; Nibau et al., 2011; Zhao et al., 2014). Recent studies demonstrated that SUP regulates both stem cell proliferation in the floral meristem and floral organogenesis through fine-tuning auxin biosynthesis (Xu et al., 2018). This mechanism might explain all sup mutant phenotypes ( Figure 1A ). Studies in SUP have shown its broad spectrum of action, highlighting how different are the floral phenotypes according to the type of mutation (Bowman et al, 2012; Breuil-Broyer et al, 2016).

In addition to the well-known capacity to fix nitrogen symbiotically, some distinctive features of legumes are the presence of compound leaves and inflorescences and a complex floral development (Ferrándiz et al., 1999; Singer et al., 1999; Benlloch et al., 2003). All these traits make them of interest for the study of unique developmental processes (Hofer and Ellis, 2014; Cañas and Beltrán, 2018).

Most legumes show complex raceme inflorescences with more than one branching. In the model legume Medicago truncatula the primary inflorescence meristem (I1) differentiates a secondary inflorescence meristem (I2). The existence of the I2 is linked to the compound inflorescence development and is a distinctive feature compared to Arabidopsis, which produces a unique inflorescence meristem (IM) before differentiating the floral meristem (FM) (Tucker, 2003; Benlloch et al., 2007). The I2 is a transient meristem, and its identity is given by a genetic function of VEGETATIVE1 (VEG1) in Pisum sativum and MtFRUITFULLc (MtFULc) in M. truncatula (Cheng et al., 2018). It has been proposed that this function was derived from the sub-functionalization of the AGL79 MADS-box gene clade within the AP1/SQUA/FUL family (Berbel et al., 2012). The perpetual activity of the I2 meristem will define the number of flowers per inflorescence and its termination as a residual vegetative organ (stub or spike) in the legume compound inflorescences (Benlloch et al., 2003; Benlloch et al., 2015). In the model legume M. truncatula, the identity of the I1 and FM, also involved in this developmental process, are specified by MtTERMINAL FLOWER1 (MtTFL1) and MtAPETALA1 (MtAP1) or MtPROLIFERATING INFLORESCENCE MERISTEM (MtPIM), respectively (Benlloch et al., 2006; Cheng et al., 2018). Their spatial and temporal expression and mutual repression control the compound inflorescence development in M. truncatula (Cheng et al., 2018).

The wild type flower of M. truncatula ( Figure 1B ) displays pentamerous floral organs per whorl: five sepals in W1, five petals in W2 (a keel petal formed by two fused petals, two wing petals and one standard or vexillum), 10 stamens in W3 (nine fused in a staminal tube and one free) and a single carpel in W4 (Benlloch et al., 2003; Cañas and Beltrán, 2018). In contrast to Arabidopsis, organ differentiation shows a high degree of spatial and temporal overlapping. Even more characteristic is the presence of common primordia (CP), ephemeral meristems from which petals and stamens will differentiate, and the early carpel differentiation (Ferrándiz et al., 1999; Tucker, 2003; Benlloch et al., 2003; Roque et al., 2018). The model species A. thaliana ( Figure 1C left) shows a centripetal and sequential organ differentiation. First, the sepal primordia are differentiated, then the petal primordia, followed by the stamen primordia, and finally, the carpel primordium. In contrast, M. truncatula ( Figure 1C right) shows unidirectional differentiation of the organ primordia with a high degree of overlapping. Unique differences are the presence of four common primordia and the early carpel primordium differentiation ( Figures 1D, E ). Despite the functional divergence of the duplicated floral homeotic MADS-box genes in M. truncatula, the specification of the floral organs is conserved in this model legume (Roque et al., 2018).

The SUP gene has been widely studied in A. thaliana. However, there is scant information on the role of SUP orthologs in other plant species. The petunia PhSUP gene (Nakagawa et al., 2004) had been the only SUP ortholog functionally characterized on its own species until it was studied in the model legume M. truncatula (Rodas et al., 2021). Recently, the SMALL REPRODUCTIVE ORGANS (SRO) gene was described as the SUP ortholog in rice (Xu et al., 2022).

The results obtained from the functional characterization of MtSUP in Medicago (Rodas et al., 2021) uncovered new context-specific functions in a different plant species. This information may have changed not only the previously proposed idea of SUPERMAN as a boundary gene but also made MtSUP a key player of the complex regulatory network behind the compound inflorescence development, being an undescribed function for a SUP ortholog in eudicots. Nevertheless, there are also similarities between MtSUP and other SUP orthologs regarding flower development.

The floral phenotypes of mtsup mutants ( Figure 1B ) are different to Arabidopsis, Petunia and rice mutants in several respects. However, sup, phsup and mtsup mutants have in common the increase in the numbers of both stamens and carpels in their respective flowers. Thus, the early floral meristem function of SUP is conserved in these three species (Nakagawa et al., 2004; Rodas et al., 2021). However, the rice SUP ortholog controls the size of male and female organs but not their number (Xu et al., 2022).

MtSUP transcript is firstly detected in the I2 meristem and later in the FM ( Figures 2A, B ). The expression pattern of MtSUP during floral organogenesis (Rodas et al., 2021) showed that even before the carpel primordium is initiated MtSUP transcript is already detected in the floral meristem centre and later in the common primordia ( Figure 2C ). The proliferation of extra petals was a distinctive feature discovered for the mtsup mutants during floral development ( Figure 1B ). At the common primordia, the meristematic cells that will produce petals and stamens coexist, and a given number of meristematic cells will give place to the organ primordia (Bossinger and Smyth, 1996).

As a role already described for SUP, MtSUP might control cell proliferation in the common primordia. In other words, MtSUP is involved in the determinacy of the common primordia as prolonged maintenance of these meristematic cells can give rise to extra organs (Bowman et al., 1989; Bossinger and Smyth, 1996). The supernumerary petals, stamens and carpels in mtsup mutants ( Figure 1B ) might also be explained by a delayed floral meristem termination linked to MtWUS persistence, as also occurs for WUS in sup mutants of A. thaliana (Prunet et al., 2017; Xu et al., 2018). MtWUS expression is not detected in the wild type after the floral apex flattens (Rodas et al., 2021). This is consistent with the early carpel initiation in legumes (Ferrándiz et al., 1999), as floral meristem termination happens when the pool of stem cells of the floral meristem centre is set to a female fate (Prunet et al., 2009). In mtsup-1, the expression of MtWUS is prolonged, thus the pool of stem cells remains undifferentiated during more time at the floral meristem centre. Contrary to SUP, which is not expressed in the floral meristem centre and plays a non-cell-autonomous function there (Prunet et al., 2017). MtSUP and MtWUS expression overlaps, both spatially and temporally at the I2 and the floral meristem centre, which would allow them to interact physically (Rodas et al., 2021). MtSUP has a novel function in the common primordia determinacy and seems to conserve its cell antiproliferative role in this unique feature of legumes.

A proper carpel primordium formation requires a correct floral meristem termination (Sakai et al., 2000; Prunet et al., 2017). MtSUP is expressed in the carpel marginal tissue that will develop the parietal placenta. It agrees with the defects in the marginal derived tissues of the gynoecium in mtsup mutants. Defects in placenta morphogenesis were also observed in the Petunia phsup mutants (Nakagawa et al., 2004), and the strong sup-5 mutant ( Figure 1A ) of A. thaliana (Gaiser et al., 1995). Therefore, the SUP orthologs PhSUP and MtSUP are required for proper floral meristem termination and the correct development of the carpel marginal tissues (Nakagawa et al., 2004; Rodas et al., 2021). Common aberrancies in the development of the placenta impacted ovule development in mtsup mutants, reducing fertility. Similar phenotypes were reported for Arabidopsis sup-5 (Gaiser et al., 1995) and Petunia phsup mutants (Nakagawa et al., 2004). Thus, the late floral function of SUP controlling ovule development is conserved in these three species (Gaiser et al., 1995; Nakagawa et al., 2004; Breuil-Broyer et al., 2016; Rodas et al., 2021).

Unlike Arabidopsis, MtSUP was first detected in the whole I2 ( Figures 2A, B ), similar to the expression of the I2 identity gene MtFULc ( Figures 2D, E ). This expression matched with the multiflowered phenotype in mtsup mutants, assigning MtSUP a determinant role in controlling the maturation rate of the I2. This novel function has not been described for any SUP-like gene. In M. truncatula cv.R108, the I2 derived from the I1 divides to produce one or two floral meristems (Benlloch et al., 2003). The I2 is a transient state between the vegetative and the reproductive tissue that remains immature until the floral identity acquisition (Prusinkiewicz et al., 2007). After producing a floral meristem, the remaining cells of the I2 enter senescence and produce the spike (Tucker, 1989; Benlloch et al., 2007). In mtsup mutants, the I2 gives rise to more floral meristems than the wild type because the residual cells of the I2 terminate as a floral meristem instead of a spike ( Figures 2F–J ). The I2 determinacy could also be linked to the gradual turn-off of MtWUS in the I2. As MtWUS prolongs its expression, the I2 could extend its activity in mtsup mutants, or MtSUP could influence the I2 activity by controlling cell proliferation. MtFULc transcript occupies a broader area in mtsup mutants compared to the wild type ( Figures 2D, E ), and there might be more cells expressing MtFULc. According to the expression analysis, MtSUP also seems to restrict MtPIM expression to the floral meristem and this restriction could be considered another way to control the determinacy of the I2. In mtsup mutants, MtPIM invades the expression domain of MtFULc in the I2 and the remnant cells that lose their vegetative nature and acquire floral identity (Rodas et al., 2021).

SUPERMAN was classified as a “cadastral gene” after studying the sup-1 (flo10) allele (Bowman et al., 1992). However, the results obtained in Medicago might support that the conserved ancestral function of SUP-like genes is the control of cell proliferation rather than a cadastral function. MtSUP does not have a typical boundary expression pattern since it is expressed in the whole I2 and floral meristem. This transcript localization correlates with the floral and inflorescence phenotypes of mtsup-1 mutant. Thus, the model that proposes the balance of cell proliferation explains mtsup mutants better than the model that considers that SUP is related to the repression of B-class MADS-box genes to the floral meristem centre. Indeed, MtPI expression (Roque et al., 2018) in mtsup-1 expands towards the W1 instead of expanding to the floral meristem centre (Rodas et al., 2021). In both flower and inflorescence development, there is no need to invoke a boundary function to explain mtsup mutants. Certainly, the phenotypic consequences of sup mutations in Arabidopsis are correlated to an over-proliferation of cells at W3 and the floral stem cells at the floral meristem centre (Prunet et al., 2009; Prunet et al., 2017; Xu et al., 2018). In addition, in P. hybrida the specific expression of PhSUP in the stamen primordia and the excessive proliferation of cells at the connective tissue in phsup1 anthers suggest that the control of cell division and growth is the function of this SUP ortholog (Nakagawa et al., 2004). Similar conclusions were reached with the SUP ortholog in rice. The authors stated that SRO is not a cadastral gene based on its expression pattern and the mechanisms through which SRO regulates reproductive organ development (Xu et al., 2018).

From an evolutionary point of view, SUP and MtSUP are orthologs that have functionally diverged through changes in their gene transcription patterns while keeping some common functions ( Figure 2K ). Such changes can occur through transposition, rearrangement, duplication or point mutations in the regulatory regions (Carroll, 2005), which are frequent after whole-genome duplications (WGD) events, a common phenomenon in the evolution of angiosperms (Cui et al., 2006). The WGD event that pre-dated the speciation of legumes ~50–60 million years ago had an essential role in structuring the M. truncatula genome and in the success of papilionoid legumes (Cannon et al., 2006). However, these rounds of polyploidization have contributed mainly to the gradual decline in the conserved synteny between species, as is the case for Arabidopsis and M. truncatula (Young et al., 2011). An example is the absence of collinearity in the flanking regions of MtSUP and SUP in their respective genomes (Rodas et al., 2021).

The functional study of SUP orthologous genes in other legume species (alfalfa, common bean) or plants with complex inflorescence (i.e. tomato, mustard) could help to understand SUP-like genes implications in the development of higher order meristems (i.e. I2). Alternatively, they could show the emergence of new functions for transcription factors when they are expressed in species with different architectures. Recently, the SINGLE FLOWER (SFL) gene, a MYB transcription factor expressed in the I2, was shown to perform a similar role to MtSUP in chickpea (Caballo et al., 2022). Multiflowered pea and chickpea mutants have been reported for three decades (Murfet, 1985; Singer et al., 1999; Gaur and Gour, 2002; Srinivasan et al., 2006; Devi et al., 2018). However, the correlation between the multiflowered mutants and the genes responsible requires further studies. The genes involved in the specification and determinacy of inflorescence meristems could be used as bioengineering tools to optimize inflorescence traits (Wang et al., 2021). In line with this, SUP-like genes in other crops should be studied to determine their possible roles in inflorescence development.

ER, AR and LC wrote the manuscript and RH, CG-M and JB contributed with valuable comments during the manuscript writing. All authors contributed to the article and approved the submitted version.