MADS-box, MYB, and GRAS genes are transcriptional regulators that are important for regulating the morphogenesis of flowering organs during the development of plant flowering organs through PPI or protein binding to target DNA sequences. The differences in the regulation of these transcription factors in different species may be related to differences in the final morphology of flowering organs.
Deletion of AP3-3 genes may affect labellar diversity
AP3 and PI genes in the MADS-box family can effectively regulate petal and stamen development (Enrico and Elliot 1991; Liu et al. 2013a), whereas class E genes are significant throughout the four rounds of floral organ formation (Kramer and Hall 2005). Class E genes are required from the initiation of floral meristematic tissue to the designation of floral organs and determination of ovule identity (Smaczniak et al. 2012). The evolutionary tree and heatmap results demonstrated that the expression of the AP3 and SEP4 genes was lacking in Apostasia odorata, which is consistent with previous findings (Zhang et al. 2017). To identify the specific missing AP3 genes, all gene sequences were compared using BLAST, and it was found that Apostasia odorata lacked the expression of AP3-3 genes, which might be related to the absence of specialized labellar structures.
Class E genes and their interactions maintain normal floral organ development
The relationships between genes within the MADS-box family have been extensively studied in model plants. Notably, the "ABCDE " model (Enrico and Elliot 1991; Immink et al. 2009; Nam et al. 2003) and the "four-molecule " model (Theissen and Saedler 2001) were applied to most dicotyledons, and the " fading bordersʺ model (Buzgo et al. 2004) can explain some monocotyledons. Nevertheless, the floral structure of orchids is exceptional, and their regulatory mechanisms are different. In Arabidopsis thaliana, class E genes can interact with other MADS-box subfamily genes to form multimers that exert regulatory effects. Similarly, in the current study, we found that class E genes were present in the PPI network of MADS-box proteins in Dendrobium officinale, Apostasia odorata, Erycina pusilla, and Phalaenopsis equestris, indicating that SEP-like proteins are still at the core of the network in orchids and may regulate multiple developmental stages. However, no relationship was found between SEP and class A, B, C, and D genes, indicating that the expression pattern of floral organ characteristic genes of the MADS-box gene family varies in different plants. For example, the lack of SEP2 expression in meristems and the expression of SEP1 in sepals in Armeniaca mume may be due to the different subfunctionalization and neofunctionalization of homologous genes of SEP-like in different species (Zhou et al. 2017).
In addition, SEP-like protein interactions are similar in different orchids, mainly interacting with histone acyltransferases, mitogen-activated protein kinases, two-component response regulators, and other transcription factors. Among these, histone acyltransferases can neutralize the positive charge of lysine in histone tails and drive the initiation of gene transcription by changing the specific conformation of local chromatin (Lee and Workman 2007; Morończyk et al. 2022). Therefore, the interactions between it and SEP-like proteins may alter the transcription initiation time of such transcription factors and affect protein expression, which may indirectly affect the normal development of floral organs and the formation of floral meristematic tissues. Members of the MAPK family are proline-targeted serine/threonine kinases that play an important role in signal transduction by regulating the level of gene transcription in the nucleus in response to changes in the cellular environment (Turjanski et al. 2007). It has been demonstrated that MAPK phosphorylation can regulate transcription factor expression by altering the activity, localization, and stability of transcription factors (Turjanski et al. 2007). E2FB and RR21 proteins are involved in cell division regulation and affect the size and number of floral organs. The expression of these genes and their corresponding reciprocal relationships exist in different species and may be crucial for normal flower organ development. However, additional experiments are required to verify whether or not they play significant roles.
Differences in MADS-box protein interactions may lead to orchid floral phenotypic diversity
The existence of large phenotypic differences between species relies on the differential expression of specific genes and gene interactions. It is evident from the analysis results that specific interactions are present in different species. In combination with the flower phenotypes, it is speculated that the key difference lies in the absence of interactions among SEP3, SYD, and BRM in Apostasia odorata. Since the BRM and SYD belong to SWI2/SNF2 chromatin remodeling ATPases, some experimental studies have found that these two genes are recruited to the regulatory regions of class B and C genes during Arabidopsis thaliana flower development and physically interact with the direct transcriptional activators SEP3 and LFY to form multimers, which in turn, regulate the expression of AP3 and AG class genes (Wu et al. 2012). This suggests that Apostasia odorata may lack such multimer formation, which affects the normal expression of class B genes and results in the absence of a special labellum.
Histone deacetylases negatively affect gene expression by restoring the positive charge of lysine in histones, which triggers chromatin condensation (Martin et al. 2007). It affects the transcriptional activity of proteins and leads to altered functions by altering local chromatin conformation. It is present only in Erycina pusilla and interacts with SEP1 and SEP2, suggesting that this protein may play an important role in floral organ morphogenesis. REF6 was the first reported H3K27me3 demethylase in plants that can directly bind and activate the expression of CUC1, which is a key gene for regulating cotyledons and other organ boundaries (Lu et al. 2011; Cui et al. 2016). Thus, REF6 plays an important role in floral organ development and boundary formation. The SEP class genes in Erycina pusilla and Phalaenopsis equestris interact with REF6. It is possible that REF6 is a downstream target gene of MADS-box proteins and is regulated by their expression. It then achieves flower organ boundary formation through multiple regulations, resulting in more complex petal, sepal, and labellum morphology in Erycina pusilla and Phalaenopsis equestris.
DIV different interactions may regulate flower symmetry formation in orchids
Regarding floral symmetry, CYC and DICH genes were first identified in the model plant Antirrhinum majus (Luo et al. 1999), and subsequent studies have also identified DIV and RAD in the MYB family that regulate floral symmetry development (Galego and Almeida 2002; Corley et al. 2005). These four genes synergistically regulate the overall asymmetry or asymmetry of individual petals. Radially symmetric ancestors in different angiosperm lineages have often evolved floral bilateral symmetries (Endress 1999). Apostasia odorata lacks a specialized labellum and has sepal-like petals of similar size, typical of radially symmetric flowers. Dendrobium officinale, Erycina pusilla, and Phalaenopsis equestris became bilaterally symmetric because of the production of a specialized labellum, suggesting that Apostasia odorata is the ancestor of the evolution of floral symmetry in orchids. In addition to the lowest number of interactions, analysis of the PPI network of DIV revealed significant differences in the interaction patterns in Apostasia odorata. However, the types of interacting proteins are distinct. Various functional proteins, such as ZIP3, DJA6, LOS1, GRIK1, and the like, have recently been identified in the other three species. The emergence of these identical new interactions may be an important reason for the differences in DIV expression, which enables orchids to evolve and maintain bilaterally symmetrical flowers by regulating the expression pattern of DIV.
MADS-box and MYB Co-Regulate Pollen Development
The MIKC* genes of the MADS-box family have maintained a relatively conserved role in gametophytes during land plant evolution (Kwantes et al. 2012). Five MIKC* genes (AGL30、AGL65、AGL66、AGL94, and AGL104) are expressed in pollen in Arabidopsis thaliana and regulate pollen development by repressing immature pollen genes and activating mature pollen genes (Verelst et al. 2007a, 2007b). Further studies have depicted that the P- and S-subclasses of MIKC*-type genes are primary regulators of male gametophytic development (Kwantes et al. 2012; Liu et al. 2013b). P-subclass genes were found only in Apostasia odorata. Although most orchids have pollen grains, only Apostasia odorata has dispersed pollen, similar to that of rice and Arabidopsis thaliana. This suggests that the loss of the P-subclass members is related to pollen evolution (Zhang et al. 2017). Apostasia odorata has powdered pollen in this study, whereas the remaining three had massive composite pollen. Based on previous studies, further construction of the interaction network of these key proteins regulating pollen development revealed no interaction between Apostasia odorata and Dendrobium officinale. These genes may play a relatively singular role in the regulation of pollen development. Therefore, the pollen structure of Apostasia odorata is probably at the initial stage of evolution because of the presence of P-subclass genes and the absence of interactions with MIKC*-type proteins. Erycina pusilla and Phalaenopsis equestris have different degrees of interaction owing to their higher levels of evolution. It is speculated that the evolutionary relationships of pollen are, in that order, Apostasia odorata, Dendrobium officinale, Phalaenopsis equestris, and Erycina pusilla.
Multiple MYB members are now found in Arabidopsis thaliana and regulate anther development and functional formation (Cheng et al. 2009; Mandaokar and Browse 2009). For example, GAMYB, a key activator of the GA signaling pathway, is involved in sporopollenin synthesis by regulating the expression of its downstream genes CYP703A3 and KAR (Kwon et al. 2015). MYB20 regulates the biosynthesis of phenylalanine and lignin to influence the formation of secondary cell walls (Geng et al. 2020). We did not find interactions of MYB20 or the apparent simplification of the GAMYB interaction network in Apostasia odorata. The complex regulatory mechanisms of the remaining three may affect the expression of GAMYB and MYB20 during pollen development, leading to earlier sporopollenin synthesis and secondary thickening of the pollen wall. Sporopollenin starts to attach to the surface of the microspore before the callose wall of the microspore tetrad is entirely disintegrated. Thus, the microspores are not free out and form tetrahymena pollen grains.
Hormone signaling pathways are directly or indirectly involved in the regulation of floral organ development
Studies have revealed that many plant hormone signaling pathways influence flower transformation, organ number, and floral meristem (FM) determination. Several studies have identified various transcription factors whose downstream target genes are related to plant hormone responses (Sessions et al. 1997; Leibfried et al. 2005; Shani et al. 2006). Auxin, cytokinin, and jasmonic acid are potential targets of MADS-box (Kaufmann et al. 2009; Li et al. 2021; Cai et al. 2014). They are directly and indirectly involved in plant growth and development, flower organ morphogenesis, and other processes. Nevertheless, the link between flower organ patterns and hormone signaling is not fully understood. In this study, we found that the SEP class in all four species interacted with B-type ARR proteins in the cytokinin signaling pathway and that some SEP proteins were indirectly regulated by the brassinosteroid pathway-related protein BRI1 via MAPKs. SEP1 was found to interact with ARR4 in Apostasia odorata, which is consistent with the finding that OsMADS1 directly acts on type A response factors during rice flower development (Khanday et al. 2013).
In addition to the MADS-box, responses to different plant hormone signaling factors were found in the MYB and GRAS transcription factors. In Arabidopsis thaliana, gibberellin was found to regulate the expression of three MYBs through jasmonic acid, which regulates stamen development (Cheng et al. 2009). MYB108 and MYB24 jointly regulate jasmonic acid-mediated stamen maturation (Mandaokar and Browse 2009). In our study, Multiple MYB members interacted directly with BRI1 and CUEL3, whereas ARR was associated with MYB through linking multiple intermediate proteins. In addition, MYB transcription factors are associated with CYCD3, which regulates flower organ development by influencing cell numbers through cell cycle regulation.
It has been demonstrated that DELLA proteins have an inhibitory effect on signaling processes such as ETH and JA (Lawit et al. 2010; Hou et al. 2010), whereas they facilitate ABA accumulation (Zentella et al. 2007). In this study, these hormone-related proteins were not found to be associated with GRAS members. However, RHT1 and GAI, members of the DELLA subfamily, were found to directly interact with GID1 and GID2 in the GA signaling pathway in Apostasia odorata. This finding was consistent with previous studies (Dill et al. 2001; Hirsch and Oldroyd 2009). In Phalaenopsis equestris, RHT1 and RGL1 were also found to interact with GID2. This indicates that GAI, RHT1, and RGL1 might be involved in flower formation induction by regulating GA signaling. GAI was also found to be associated with BZR1, a signaling factor of the brassinosteroid pathway that regulates cell elongation and division in Apostasia odorata.
DoSYD gene function validation
A mutation in flower phenotype was observed in positive Arabidopsis plants, which exhibited increased petal aperture, overlapping petals, and altered stamen distribution, indicating the regulatory function of the gene in the formation of floral diversity. However, it remains uncertain whether the DoSYD gene affects the development of the labellum in orchids, and further genetic transformation experiments in orchids are required to observe phenotypic differences. It should be noted that genetic transformation of orchids has difficulty and flowering requires approximately 3–4 years, therefore this experiment is limited to Arabidopsis transformation study only.