*3.2. Prediction of MADS-Box Genes Involved in the Regulation of Flower Development and Floral Organ Identity*

An investigation of the genetic and molecular basis of flower development and floral organ identity in Arabidopsis and petunia suggested that MADS-box genes play fundamental roles in floral organ identity and flower development [82]. It has been confirmed that five classes of MADS-box genes (A–E) were involved in specifying floral organ identity [83–85]. In Arabidopsis, the class A genes (*AP1* and *AP2*), the class B genes (*AP3* and *PI*), the class C gene (*AG*), the class D gene (*AGL11*), and the class E gene (*SEP1*, *SEP2*, *SEP3,* and *SEP4*) were MADS-box genes, which have been reported to be involved in the regulation of floral organ development [83,86]. In petunia, lots of MADS-box genes, including the class A genes PETUNIA FLOWERING GENE (*PFG*), FLORAL BINDING PROTEIN 26 (*FBP26*), and *FBP29*, the class B genes *TM6*, *PMADS1*/*GP*, *PMADS2,* and *FBP1*, the class C genes *PMADS3*, *FBP6,* and *FBP24*, the class D genes *FBP11* and *FBP7,* as well as *FBP2*, *FBP4*, *FBP5*, *FBP9*, and the class E genes *FBP23*, *PMADS4,* and *PMADS12* played important roles in flower development [17,87].

In this paper, we investigated the tomato MADS-box genes' phylogenetic relationships with the petunia hybrid to select 15 tomato MADS-box genes that may play specific roles in flower development (Table S2). According to the expression profile predictions shown in Figure 4, the highest expression values for most of the genes (*MADS-MC TAP3*, *TM6*, *SlMBP1*, *SlMBP2*, *TAG1*, *SlMBP22*, *SlMBP21,* and *SlMBP6*) were observed in flower development stages. Furthermore, qPCR was used to study the expression patterns of four whorls of floral organs (sepal, petal, stamen, and carpel) in these 15 tomato MADS-box genes.

*AP1* is an Arabidopsis A class gene, which conferred sepal identity in the first floral [88]. In petunia, the three genes *PFG*, *FBP26,* and *FBP29* have been identified, which were orthologs of *AP1*/*SQUA* in Arabidopsis [17]. Our phylogenetic analysis showed that *MADS-MC* and *SlMBP20* belonged to this clade (Figure S3) and their expression were particularly high in sepal, suggesting that they might play a similar role to *AP1* (Figure 5A,B). The class B genes were involved in the identification of petal and stamen in angiosperms [89]. Regarding the class B genes, five close homologs of petunia—TOMATO MADS-BOX GENE6 (*TM6)*, GREEN PETAL (*GP*)/ PETUNIA MADS BOX GENE 1 (*PMADS1*), *PMADS2,* and *FBP1*—were found in tomato [77]. The qPCR analysis showed that *TAP3*, *SlMBP2,* and *SlMBP1* have petal and stamen specific expression, while the *TM6*/*TDR6* transcripts were mainly detected in the petal and carpel (Figure 5C–F). These results were similar to the homologous genes of that in petunia [77]. Two tomato MADS-box genes *TAG1* and *TAGL1*, which were involved in C functions, were from the monophyletic *AGAMOUS* (*AG*) subfamily. These two genes were mainly expressed in stamens and carpels (Figure 5G,H), which is consistent with their function in specifying stamen and carpel development [52]. *SlMBP3* and *SlMBP22*, which are highly homologous to two petunia class D MADS-box genes, *FBP11* and *FBP7*, were shown to be separately and exclusively expressed in carpel (Figure 5L,I). The result suggested that *SlMBP3* and *SlMBP22* may have similar functions to the petunia *FBP7* and *FBP11* genes, which are related to the establishment of real ovules or carpel-like structures [90]. Arabidopsis *SEP* genes, the typical class E genes, were necessary for the specification of sepal, petal, stamen, and carpel identity with interaction with the class A, B, C, and D genes [16]. In petunia, seven class E genes (*FBP2*, *FBP4*, *FBP23*, *FBP5*, *FBP9*, *pMADS12,* and *pMADS4*) have been determined that belong to the *SEPALLATA* (*SEP*) clade [32]. The tomato *TAGL2, TM5*, *SlMADS1*, *SlMBP21,* and *SlMBP6* genes were homologous to these petunia class E genes (Figure S3), and some differences in

expression patterns have been observed (Figure 5J–N), indicating that these five tomato class E genes may be involved in multiple floral organ identity. Thus, we believe that the expression patterns of the tomato MADS-box genes identified in our study will be an important tool for understanding the flower development mechanisms in tomato. Previous reports have found that the MADS-box proteins are able to form multiple homologous or heterodimeric complexes in plants, and the combinatorial MADS-box proteins are often deriving their regulatory specificity from other DNA binding or accessory factors. To understand how the tomato MADS-box genes can act in flower development and floral organ identity, it is necessary to identify the network of protein–protein interactions amongst them. Therefore, the predicted interaction networks of the 15 tomato MADS-box proteins, which are involved in floral organ development, were analyzed in our report (Figure S4). In many domesticated crops, it's an important way to select inflorescence architecture with improved flower production and yield. In tomato, SlMBP21 forms protein complexes with JOINTLESS and MACROCALYX as a transcription activator for tomato flower abscission zone development [50], because SlMBP21, J, and MC have a common function in the development of the tomato flower abscission zone. In breeding, altering any of these genes will have the function on plant growth. In this study, the predicted interaction networks may help us to understand how the tomato MADS-box genes can act in flower development and floral organ identity.
