Orchid NAC Transcription Factors: A Focused Analysis of CUPULIFORMIS Genes

Plant transcription factors are involved in different developmental pathways. NAC transcription factors (No Apical Meristem, Arabidopsis thaliana Activating Factor, Cup-shaped Cotyledon) act in various processes, e.g., plant organ formation, response to stress, and defense mechanisms. In Antirrhinum majus, the NAC transcription factor CUPULIFORMIS (CUP) plays a role in determining organ boundaries and lip formation, and the CUP homologs of Arabidopsis and Petunia are involved in flower organ formation. Orchidaceae is one of the most species-rich families of angiosperms, known for its extraordinary diversification of flower morphology. We conducted a transcriptome and genome-wide analysis of orchid NACs, focusing on the No Apical Meristem (NAM) subfamily and CUP genes. To check whether the CUP homologs could be involved in the perianth formation of orchids, we performed an expression analysis on the flower organs of the orchid Phalaenopsis aphrodite at different developmental stages. The expression patterns of the CUP genes of P. aphrodite suggest their possible role in flower development and symmetry establishment. In addition, as observed in other species, the orchid CUP1 and CUP2 genes seem to be regulated by the microRNA, miR164. Our results represent a preliminary study of NAC transcription factors in orchids to understand the role of these genes during orchid flower formation.


Introduction
Plant transcription factors control many developmental processes, regulating the temporal and spatial expression of genes involved in plant growth and development. Among more than 2500 transcription factors present in plant genomes, a large number is involved in the development of the flower [1].
Their name derives from the NAM gene (No Apical Meristem) in Petunia, ATAF (A. thaliana Activating Factor), and CUC (Cup-shaped Cotyledon) in Arabidopsis. The Petunia nam mutants lack a shoot apical meristem and show defects in organ boundaries [4]. This same phenotype characterizes the Arabidopsis cuc1 and cuc2 mutants and the cupuliformis (cup) mutant of the snapdragon, A. majus [2,8,15].
In the A. thaliana genome, more than 100 genes encode NAC transcription factors, divided into sixteen subfamilies based on amino acid sequence similarity [16,17]. The typical structure of a NAC protein has a conserved NAC domain at the N-terminus, with a twisted antiparallel β-sheet enclosed between two α-helices with five subdomains (A-E) [2,18]. The NAC domain, often found in two tandem repeats, contains positively charged amino acids possibly involved in DNA binding [19]. The NAC domain also enhances the protein size that the molecular network driving orchid flower formation is more complex than previously thought.
Based on these premises and the role of the CUP genes in the flower organ formation of Antirrhinum and other species, we analyzed the orchid CUP genes to verify their possible involvement in orchid perianth formation.
We identified the orchid genes encoding NAC transcription factors and investigated their genomic organization, starting with the available orchid genomes and transcriptomes [35,42,43,[48][49][50]. In addition, we analyzed the expression patterns of the CUP genes and miR164 in the orchid P. aphrodite at the early and late stages of flower development. P. aphrodite is an ornamental epiphytic orchid with many advantageous characteristics for cultivation and manipulation. Its inflorescence has many large flowers, and its genome was sequenced [48]. For these reasons, we used this species to conduct our expression analysis of the CUP genes and miR164.

Plant Material
The plants used in this study were grown in the Department of Biology at the University of Naples Federico II greenhouse (Napoli, Italy) under natural light and temperature. P. aphrodite (subfamily Epidendroideae) is commercially available, while Orchis italica (subfamily Orchidoideae) was collected in Sassano (Salerno, Italy) and is part of the Department of Biology orchid collection.
We sampled P. aphrodite flowers from three different plants at the B2 developmental stage (early, flower bud size 1-1.5 cm) and soon after anthesis (open flower, OF) ( Figure 1). We dissected outer tepals, lateral inner tepals, lip, column, and ovary (not pollinated) from each P. aphrodite sample. In addition, we also collected leaf samples of P. aphrodite and O. italica. We stored the collected material in RNA-later (Ambion -Berlin, DE) until RNA extraction.
We then virtually translated the identified orchid NAC nucleotide sequences to verify the presence of indels and/or stop codons within the coding sequence and evaluate their completeness. We discarded the sequences containing a partial NAM domain and/or stop codons.
In addition, we identified the NAC sequences of selected non-orchid species (A. thaliana, A. majus, Amborella trichopoda, Asparagus officinalis, Zea mays) using BLAST search and the Plaza 4.0 database [53]. The species name and accession number of the sequences used for the analyses are listed in Supplementary Data S1.
We aligned the amino acid sequences of the NAC proteins of Arabidopsis and each orchid species using ClustalW [54]. After manually adjusting the alignments, we constructed the Maximum Likelihood (ML) trees using MEGA7 [55] for each orchid species and Arabidopsis. We classified the orchid NAC sequences based on the ML trees and the subfamily classification of Arabidopsis. We then aligned the NAM amino acid sequences of orchid and non-orchid species and constructed the ML trees with 200 bootstrap replicates using MEGA7.
Using the MEME v5.5 online tool (https://meme-suite.org/meme/) [57], we searched for shared amino acid motifs among the sequences of the orchid NAC proteins, setting the search parameters to a maximum number of 10 motifs and an optimum width from 6 to 50.

PCR Amplification and Sequencing of the CUP Genes of O. italica
To obtain information about the genomic organization of the highest number of orchid CUP genes, we included O. italica in our analysis. The genome is unavailable for this species; however, we have access to samples of O. italica that are part of the orchid collection of the Department of Biology (University of Naples Federico II). We extracted genomic DNA from leaves of O. italica, PCR-amplified and sequenced the CUP genes, thus reconstructing their exon/intron structure.
Starting with 500 mg of O. italica leaf material, we extracted the genomic DNA following the Doyle and Doyle protocol [58]. We designed specific primer pairs to PCR amplify the CUP genes of O. italica based on the nucleotide sequence of the transcript identified (Table 1). We conducted the amplification reactions in a final volume of 50 µL with the following reagent concentrations: 1X reaction buffer, 0.2 mM of each primer, 0.2 mM dNTP mix Platinum™ GC Enhancer, and 2 U Taq Hot-Start DNA Polymerase Invitrogen™ Plat-inum™ II (Invitrogen-Waltham, MA). The thermal cycle was the following: 94 • C for 2 min, 35 cycles of 94 • C for 15 sec, 60 • C for 15 sec, and 68 • C for a time ranging from 1:30 sec to 2 min, depending on the amplicon size.
We cloned the amplicons into the pSC-A-amp/kan vector (Agilent, Santa Clara, CA, USA), sequenced them using the T3 and T7 primers (Eurofins Genomics), and deposited the genomic sequences of the CUP genes of O. italica in GenBank with the accession numbers OP752215, OP752217, and OP752219. After reconstructing the genomic organization of the O. italica CUP genes, we verified the presence of transposable/repetitive elements within introns, as described earlier.

PCR Amplification and Sequencing of P. aphrodite CUP Transcripts
We extracted total RNA from the floral buds of P. aphrodite at the B2 developmental stage using Trizol Reagent (Ambion -Berlin, DE) followed by DNase treatment. We reversetranscribed 500 ng of total RNA using an Advantage RT-PCR kit (Clontech, Mountain View, CA, USA) with oligo dT and random primers. Table 1. List of the primer sequences used for genomic (gDNA) and cDNA sequence isolation, qPCR analysis, miRNA and Yeast two-hybrid analysis.

Application
Gene We designed CUP-specific primer pairs starting with the sequences of P. aphrodite obtained from the in silico analysis (Table 1). We used 30 ng of first-strand cDNA to PCR amplify the transcripts of the CUP genes. We conducted each amplification reaction in a final volume of 50 µL, with the following reagent concentrations: 1X reaction buffer, 0.5 mM of each primer, and 2.5 U Wonder Taq DNA polymerase (Euroclone-Milan, IT). The thermal cycle was the following: 95 • C for 1 min, 35 cycles of 95 • C for 15 s, 57-59 • C for 15 s, and 72 • C for a time ranging from 30 s to 1:30 min, depending on the amplicon size.
We cloned and sequenced the amplicons as described before.

Expression Analysis
We extracted the total RNA from the outer and inner tepals, lips, columns, and ovaries collected from the P. aphrodite B2 stage and open flowers, and from the leaf tissue. We extracted, quantified and reverse-transcribed RNA as described in the previous paragraph.
We performed relative expression analysis of the P. aphrodite CUP genes by real-time qPCR on cDNA from flower organs and leaf tissue, using 18S as a reference gene [44,46]. We designed primer pairs specifically for the qPCR analysis (Table 1). We conducted the experiments in biological and technical triplicate using PowerUp™ SYBR™ Green Master Mix (Applied Biosystems™-Foster City, CA) following the manufacturer's instructions and applying the following thermal conditions: 50 • C for 2 min, 95 • for 2 min, 40 cycles of 95 • C for 15 s, 60 • C for 1 min.
We calculated the relative quantity (NRQ) ± SEM for each replicate normalized to the internal control gene [59]. After ANOVA analysis, we applied the Holm-Sidak post hoc test to assess the statistical significance of the differences in NRQ among the different tissues.

miR164 Analysis
We used the psRNATarget (2017 Update) online tool from The Zhao Lab © Ardmore, OK [60] with stringent search parameters (maximum expectation 0.0) to search for the presence of the miR164 target site in the orchid CUP sequences.
To identify the eventual specific fragments produced by the cleavage of miR164 on the P. aphrodite CUP transcripts, we performed a modified 5 -RACE experiment. We designed  (Table 1). We ligated the 5 -adaptor in the RLM-RACE GeneRace kit (Invitrogen-Waltham, MA) to the 5 -end of total RNA (500 ng) extracted from B2-stage buds of P. aphrodite without removing the 5 -cap. After reverse transcription, we amplified the cDNA using the reverse primers specific for PaCUP1 and PaCUP2 (Table 1) and the GeneRace 5 forward primer.
We conducted a stem-loop real-time qPCR experiment to evaluate the expression pattern of the miR164 in P. aphrodite. We reverse-transcribed total RNA (150 ng) from floral organs of B2 floral buds, open flower (OF), and leaf tissue using a specific miR164 or 18S stem-loop primer ( Table 1). We conducted the qPCR amplifications in biological and technical triplicate using the specific miR164 and 18S forward primers, and the universal stem-loop reverse primer. Amplification conditions, thermal cycle, NRQ calculation, and statistical analysis were performed as described in the previous paragraph.

Yeast Two-Hybrid Analysis
To check whether the proteins PaCUP1 and PaCUP2 of P. aphrodite interact with each other and form homodimers, we used the GAL4-based yeast two-hybrid (Y2H) system (Matchmaker two-hybrid system 3, Clontech-Mountain View, CA).
We amplified the full-length coding regions of the PaCUP1/2 transcripts using specific primer pairs (Table 1). We cloned both the amplified coding regions into the pGADT7 (prey) and PGBKT7 (bait) expression vectors containing the activation (AD) and DNA-binding (BD) domains of the yeast transcription factor GAL4, respectively. We transformed the Saccharomyces cerevisiae strain AH109 [61] with all the combinations of pray and bait vectors, conducting each transformation in triplicate. After the yeast transformation, we verified the presence of the recombinant vectors by growing the yeast cells in Synthetic Defined (SD) medium lacking tryptophan and leucine in the presence of different concentrations (from 0 to 70 mM) of 3-aminotriazole (3AT). We transformed the empty vectors pGBKT7 or pGADT7 in combination with the recombinant vectors as negative controls.
In addition, we checked the autoactivation ability of the PaCUP1/2 proteins. To evaluate this, we grew the yeast transformed only with the recombinant PaCUP1 or PaCUP2 vector pGBKT7 (BD) on a selective medium lacking tryptophan and histidine in the presence of different concentrations (from 0 to 70 mM) of 3AT.

The NAC Transcription Factors of Orchids
Through BLAST analysis, we identified 423 orchid NAC transcripts. All sequences are listed in Supplementary Data S1 and Supplementary Table S1. Almost all (370) have a complete coding sequence; partial sequences (53) containing the nucleotide region encoding for the NAC and TAR conserved motifs were included in the analysis. In addition, we downloaded the NAC protein sequences of the model species A. thaliana (111) and A. majus (77), the basal angiosperm A. trichopoda (42), and the monocots A. s officinalis (95) and Z. mays (for this latter only the CUP proteins).
The number of NAC transcripts we identified in the different orchid species was variable, from 37 to 74. The species with the smallest number of sequences identified are O. italica (37) and C. formosanum (47). There is no assembled genome available for both species, so we used their transcriptomes to search the NAC sequences. The small number of NACs detected could be due to the incompleteness of their transcriptomes (e.g., obtained from specific tissues where some NAC genes are not expressed).
Based on sequence similarity, the NAC sequences can be classified into the subfamilies previously described in Arabidopsis [16] (Supplementary Material, Table S1). As observed in other species, no orchid NAC belongs to the AtNAC3 and ANAC001 subfamilies, where only Arabidopsis sequences are included [16,62,63]. The analysis of the conserved domains of the NAC proteins reveals the presence of the five conserved subdomains (A-E) and the TAR regions specific to different subfamilies (Supplementary Data S1).
Scanning the available orchid genomes, we reconstructed the genomic organization of the orchid NAC genes. The intron number and size vary among the various orchid NAC subfamilies. Intron number ranges from one to ten, with the highest number found in the NAC2 subfamily; intron size ranges from 52 bp (intron 1 of the A. shenzenica As014187 gene, SENU5 subfamily) to 20,287 bp (intron 1 of the D. catenatum Dc021911 gene, ANAC063 subfamily).
Compared with A. thaliana, the NAC genes of orchids generally have a higher number of introns with a larger intron size in almost all cases (Supplementary Datas S1 and S2). CENSOR analysis revealed traces of transposable/repetitive elements (e.g., DNA transposons, LTR, and Non-LTR Retrotransposons) within the introns of the orchid NAC genes displaying an intron number higher than Arabidopsis (Supplementary Data S3). The presence of repetitive elements and transposons in the NAC introns of orchids reflects a common feature of the orchid genomes. Indeed, transposable and repetitive elements can cover up to 60% of the orchid genomes [64], and orchid genes are often characterized by large introns containing repetitive and transposable elements, as observed in KNOX [65], OitaAP2, OitaAG, OitaSTK [66,67], and MYB genes [44].
The NAM subfamily of the NAC genes includes the highest number of sequences in almost all the species examined. Among them, the CUP transcription factors [68] are involved in the flower development of A. majus and other Lamiales [15]. For this reason, we focused our study on the NAM subfamily, in particular on the CUP genes. Figure 2 shows the ML tree of the NAM transcription factors of orchids A. thaliana, A. majus, A. officinalis, and Z. mays. The tree topology shows a well-distinguished group that includes CUP-like proteins. In addition to the NAC domain, NAM proteins have two transcriptional activation regions, TAR1 and TAR2 [16]. The Cf000662 protein in C. formosanum is the only orchid NAM protein containing both the TARs, whereas 34 orchid NAM proteins lack both regions (Supplementary Data S1), as observed in the A. thaliana AT1G76420, AT2G24430, and AT3G18400 sequences. The other orchid NAM sequences show only the TAR1 or TAR2 motif.
As in the other orchid NAC subfamilies, the NAMs have an intron size generally higher than Arabidopsis, as illustrated in Figure 3. Intron 2 shows the highest intron size variability. For example, in D. catenatum, P. aphrodite, P. equestris, and V. planifolia, the size In addition to the NAC domain, NAM proteins have two transcriptional activation regions, TAR1 and TAR2 [16]. The Cf000662 protein in C. formosanum is the only orchid NAM protein containing both the TARs, whereas 34 orchid NAM proteins lack both regions (Supplementary Data S1), as observed in the A. thaliana AT1G76420, AT2G24430, and AT3G18400 sequences. The other orchid NAM sequences show only the TAR1 or TAR2 motif.
As in the other orchid NAC subfamilies, the NAMs have an intron size generally higher than Arabidopsis, as illustrated in Figure 3. Intron 2 shows the highest intron size variability. For example, in D. catenatum, P. aphrodite, P. equestris, and V. planifolia, the size of these introns ranges between 2001 and 9000 bp, values not observed in the Arabidopsis genes ( Figure 3). In contrast, the coding region size is generally conserved between Arabidopsis and orchid species (Figure 3).

The Orchid CUP Genes: Phylogeny, Structure, and Expression Pattern
The CUP genes are involved in plant and flower development. In Arabidopsis, are three CUP gene homologs [2,6]. CUC1 and CUC2 promote apical meristem g during pre-and post-embryo development [5,31]. All three Arabidopsis CUC gen involved in the maturation of flower reproductive organs; mutants of these genes r a phenotype strongly defective in the gynoecium and ovule structures [6,69]. In add the CUC2 and CUC3 genes regulate the formation of the border region of leaf prim and their subsequent development [6,31]. In A. majus, the CUP gene is expressed, early stage of development, at a high level in the bud lateral-ventral region, regu the boundary for lip development [15].
The ML tree obtained from the amino acid sequences of the orchid CUPs revea presence of two clades: the main includes the CUP1 and CUP2 proteins, and the contains the CUP3 proteins (Figure 4a). This clade separation is not unique to orchid is observed in other monocots, dicots, and early diverging angiosperm sp [27,68,70,71]. The CUP1-2 clade is the most ancient, while duplication events occurr  The CUP genes are involved in plant and flower development. In Arabidopsis, there are three CUP gene homologs [2,6]. CUC1 and CUC2 promote apical meristem growth during pre-and post-embryo development [5,31]. All three Arabidopsis CUC genes are involved in the maturation of flower reproductive organs; mutants of these genes reveal a phenotype strongly defective in the gynoecium and ovule structures [6,69]. In addition, the CUC2 and CUC3 genes regulate the formation of the border region of leaf primordia and their subsequent development [6,31]. In A. majus, the CUP gene is expressed, at an early stage of development, at a high level in the bud lateral-ventral region, regulating the boundary for lip development [15].
The ML tree obtained from the amino acid sequences of the orchid CUPs reveals the presence of two clades: the main includes the CUP1 and CUP2 proteins, and the other contains the CUP3 proteins (Figure 4a). This clade separation is not unique to orchids and is observed in other monocots, dicots, and early diverging angiosperm species [27,68,70,71]. The CUP1-2 clade is the most ancient, while duplication events occurring in angiosperms after divergence from gymnosperms generated the CUP3 clade [72].   Figure S1); (c) genomic organization of orchid CUP genes and other species with available genomes.
The analysis of the conserved motifs reveals different patterns between the two clades ( Figure 4b). Almost all the sequences have the characteristic NAM domain (motifs from 1 to 5) with few exceptions, possibly due to the incomplete CDS sequences isolated from the available transcriptomes. Motifs 6-7-8-10 are generally conserved within the CUP1-2 clade and absent in CUP3, whereas motif 9 is present only within the CUP3 clade.
Almost all the orchid CUP genes have three exons and two introns of variable size (Figure 4c). The As001107 gene of the basal orchid A. shenzenica is the only orchid CUP gene with four exons and three introns.
To evaluate whether the CUP genes could be involved in orchid flower development, we conducted a preliminary analysis of their expression patterns in the early and late developmental stages of P. aphrodite.
The three PaCUP genes show different expression profiles in the perianth during development. At the early stage, the PaCUP1-2 genes are mainly expressed in the lip, while PaCUP3 is also highly expressed in the inner tepals (Figure 5a). After anthesis, the expression of the three PaCUPs decreases in the perianth (Figure 5a). At this stage, PaCUP2 has higher expression in the inner tepals, while PaCUP3 maintains high expression in the lip (Figure 5a).
All three PaCUP genes are highly expressed in the reproductive organs (column and ovary) at the early stage of development (Figure 5b). After anthesis, PaCUP1 and PaCUP3 decrease their expression level in the column and ovary, whereas the expression of PaCUP2 increases (Figure 5b).
Only the PaCUP3 gene is expressed in leaf tissue, as previously reported in A. thaliana, suggesting its possible role in the leaf [73] (Figure 5c). The analysis of the conserved motifs reveals different patterns between the two clades ( Figure 4b). Almost all the sequences have the characteristic NAM domain (motifs from 1 to 5) with few exceptions, possibly due to the incomplete CDS sequences isolated from the available transcriptomes. Motifs 6-7-8-10 are generally conserved within the CUP1-2 clade and absent in CUP3, whereas motif 9 is present only within the CUP3 clade.
Almost all the orchid CUP genes have three exons and two introns of variable size (Figure 4c). The As001107 gene of the basal orchid A. shenzenica is the only orchid CUP gene with four exons and three introns.
To evaluate whether the CUP genes could be involved in orchid flower development, we conducted a preliminary analysis of their expression patterns in the early and late developmental stages of P. aphrodite.
The three PaCUP genes show different expression profiles in the perianth during development. At the early stage, the PaCUP1-2 genes are mainly expressed in the lip, while PaCUP3 is also highly expressed in the inner tepals (Figure 5a). After anthesis, the expression of the three PaCUPs decreases in the perianth (Figure 5a). At this stage, PaCUP2 has higher expression in the inner tepals, while PaCUP3 maintains high expression in the lip (Figure 5a). At the early stage of flower development, the P. aphrodite CUP genes have an expression pattern similar to that of A. majus [15], being expressed in the lip and inner tepals. However, while in A. majus the CUP genes have a lateral-ventral profile, in the orchid flower the CUP expression is lateral-dorsal due to resupination. In most orchids, including P. aphrodite, resupination occurs before anthesis and is the 180° rotation of the pedicel and ovary, resulting in the reversal position of the dorsal and ventral organs in the mature flower. Consequently, although the lip is a dorsal structure, it has a ventral position [34].
As observed in A. majus, the expression pattern of the CUP genes in the orchid flower suggests their possible role in flower formation. The CUP genes might interact with other transcription factor encoding genes, e.g., MYB and TCP, involved in flower development and symmetry establishment [26,[44][45][46][74][75][76]. Similar interactions were demonstrated in A. majus, where the CUPs can interact with the TCP proteins, suggesting their role in the boundary flower formation [8]. In addition, the phenotype of the cup mutant resembles that of the double mutant cyc dich, suggesting a functional correlation between CUP and TCP genes [8]. Furthermore, the expression of the MYB gene, DIV, could positively regulate the CUP expression in the labellum of Antirrhinum [15].
Recently, the orchid YABBY transcription factor, DL2, was proposed as a possible regulator of orchid lip formation [47]. In P. equestris there is a correlation between the expression pattern of DL2 and the class-B DEF-like genes, PeMADS2-PeMADS5, responsible for orchid perianth formation [47]. Curiously, the expression pattern of All three PaCUP genes are highly expressed in the reproductive organs (column and ovary) at the early stage of development (Figure 5b). After anthesis, PaCUP1 and PaCUP3 decrease their expression level in the column and ovary, whereas the expression of PaCUP2 increases (Figure 5b).
Only the PaCUP3 gene is expressed in leaf tissue, as previously reported in A. thaliana, suggesting its possible role in the leaf [73] (Figure 5c).
At the early stage of flower development, the P. aphrodite CUP genes have an expression pattern similar to that of A. majus [15], being expressed in the lip and inner tepals. However, while in A. majus the CUP genes have a lateral-ventral profile, in the orchid flower the CUP expression is lateral-dorsal due to resupination. In most orchids, including P. aphrodite, resupination occurs before anthesis and is the 180 • rotation of the pedicel and ovary, resulting in the reversal position of the dorsal and ventral organs in the mature flower. Consequently, although the lip is a dorsal structure, it has a ventral position [34].
As observed in A. majus, the expression pattern of the CUP genes in the orchid flower suggests their possible role in flower formation. The CUP genes might interact with other transcription factor encoding genes, e.g., MYB and TCP, involved in flower development and symmetry establishment [26,[44][45][46][74][75][76]. Similar interactions were demonstrated in A. majus, where the CUPs can interact with the TCP proteins, suggesting their role in the boundary flower formation [8]. In addition, the phenotype of the cup mutant resembles that of the double mutant cyc dich, suggesting a functional correlation between CUP and TCP genes [8]. Furthermore, the expression of the MYB gene, DIV, could positively regulate the CUP expression in the labellum of Antirrhinum [15].
Recently, the orchid YABBY transcription factor, DL2, was proposed as a possible regulator of orchid lip formation [47]. In P. equestris there is a correlation between the expression pattern of DL2 and the class-B DEF-like genes, PeMADS2-PeMADS5, responsible for orchid perianth formation [47]. Curiously, the expression pattern of PaCUP1 in the lip is similar to that of DL2 in Phalaenopsis. Both genes have high expression in the lip during the early developmental stages, followed by a decrease until the mature flower [47].
Our results suggest that, together with MADS-box and other transcription factors, the orchid CUP genes might be involved in the complex molecular pathway of orchid perianth formation.

Expression Pattern of miR164
The transcripts of the CUP1-2 genes have a target site for the microRNA, miR164. This site, located within the region upstream of the 3 UTR, is absent in the CUP3 lineage [71]. In Arabidopsis, microRNAs regulate the CUP1-2 genes in many plant life processes. For example, the miR164c mutant has an altered number of petals, suggesting a role for miR164 in flower structure formation [77]. For these reasons, we investigated the presence of the miR164 target site on the orchid CUP1-2 transcripts and compared the expression profile of this microRNA to that of the PaCUP1-2 genes.
Using the psRNATarget online tool [60], we verified the presence of the putative miR164 target site on the CUP transcripts. Our results show that 12 orchid CUP1-2 transcripts have the miR164 target site ( Figure 6). As expected, no CUP3 transcript had the miR164 target site. PaCUP1 in the lip is similar to that of DL2 in Phalaenopsis. Both genes have high expression in the lip during the early developmental stages, followed by a decrease until the mature flower [47].
Our results suggest that, together with MADS-box and other transcription factors, the orchid CUP genes might be involved in the complex molecular pathway of orchid perianth formation.

Expression Pattern of miR164
The transcripts of the CUP1-2 genes have a target site for the microRNA, miR164. This site, located within the region upstream of the 3′UTR, is absent in the CUP3 lineage [71]. In Arabidopsis, microRNAs regulate the CUP1-2 genes in many plant life processes. For example, the miR164c mutant has an altered number of petals, suggesting a role for miR164 in flower structure formation [77]. For these reasons, we investigated the presence of the miR164 target site on the orchid CUP1-2 transcripts and compared the expression profile of this microRNA to that of the PaCUP1-2 genes.
Using the psRNATarget online tool [60], we verified the presence of the putative miR164 target site on the CUP transcripts. Our results show that 12 orchid CUP1-2 transcripts have the miR164 target site ( Figure 6). As expected, no CUP3 transcript had the miR164 target site.
We conducted the modified 5′-RACE experiment on the PaCUP transcripts of P. aphrodite in the tissues examined and did not detect any miR164 cleavage fragment of the expected size. The apparent absence of the cleavage product in the selected tissues may be due to the high instability of the mRNA fragment produced after cleavage by miR164.  We conducted the modified 5 -RACE experiment on the PaCUP transcripts of P. aphrodite in the tissues examined and did not detect any miR164 cleavage fragment of the expected size. The apparent absence of the cleavage product in the selected tis-sues may be due to the high instability of the mRNA fragment produced after cleavage by miR164.
To compare the expression patterns of miR164 and PaCUP genes, we conducted stemloop qPCR experiments of miR164 on floral tissues from the B2 and OF stages of P. aphrodite (Figure 7a,b). To compare the expression patterns of miR164 and PaCUP genes, we conducted stem-loop qPCR experiments of miR164 on floral tissues from the B2 and OF stages of P. aphrodite (Figure 7a,b). In the perianth, miR164 shows an expression trend opposite to that of the PaCUP1-2 genes. At the B2 stage, the miR164 transcripts are almost undetectable, whereas they are expressed in the floral tissues after anthesis (Figure 7a), suggesting that the expression of miR164 reduces the number of the PaCUP1-2 transcripts. These results are in agreement with those obtained in A. majus, where at the first stage of development, the CUP1-2 genes regulate the boundary formation and are expressed in all flower primordia. Their expression then decreases before anthesis, limiting the presence of the CUP transcripts to specific parts of the flower [15]. In the perianth, miR164 shows an expression trend opposite to that of the PaCUP1-2 genes. At the B2 stage, the miR164 transcripts are almost undetectable, whereas they are expressed in the floral tissues after anthesis (Figure 7a), suggesting that the expression of miR164 reduces the number of the PaCUP1-2 transcripts. These results are in agreement with those obtained in A. majus, where at the first stage of development, the CUP1-2 genes regulate the boundary formation and are expressed in all flower primordia. Their expression then decreases before anthesis, limiting the presence of the CUP transcripts to specific parts of the flower [15].
In the reproductive structures of P. aphrodite, miR164 is not expressed at the B2 stage, allowing a high expression of the PaCUP genes. Later, at the OF stage, mir164 is expressed; however, only the transcript levels of PaCUP1 decrease, in particular in the ovary, whereas PaCUP2 is still highly expressed, suggesting that after anthesis, the levels of PaCUP2 transcripts are not regulated by miR164. The activity of miR164 on the transcript levels of the homologs of the CUP genes was also proposed for A. thaliana, where miR164 is involved in reproductive structure formation and possibly modulates the transcript levels of the CUC1-2 genes (homologs of the CUPs) involved in ovary and column formation [31,69].
Finally, miR164 is expressed in the leaf tissue of P. aphrodite, where the PaCUP1-2 transcript levels are low. The miR164 of A. thaliana regulates the CUC1-2 genes during leaf shape formation and senescence. Mutants of miR164 have an unusual leaf shape, with leaves showing a more pronounced dentition than the wild type [78]. Our results suggest that, also in orchids, miR164 seems to regulate the PaCUP1-2 expression in the leaf.

PaCUP1-PaCUP2 Protein Interaction
The NAC transcription factors of Arabidopsis can form homo-and heterodimers due to their ability to interact through the NAC domains [25]. In addition, CUC1 and CUC2 in Arabidopsis can activate transcription due to the presence of acidic residues within the TAR domain [25]. In a Y2H assay, they activated the transcription of the yeast reporter genes [79]. For these reasons, we checked the autoactivation ability of the P. aphrodite PaCUP1-2 proteins and their ability to form homo-and heterodimers in the Y2H assay.
The yeast clones double-transformed with the recombinant vector pGBKT7 containing the full-length transcripts PaCUP1 or PaCUP2, and the empty vector PGADT7. The clones were grown on a selective medium without tryptophan, lysine, or histidine in the presence of different concentrations of 3AT (from 0 to 70 mM). Only the PaCUP1 protein alone activated the transcription of the yeast GAL4 gene in the absence of histidine and in the presence of up to 20 mM 3AT (Figure 8a). For this reason, the protein-protein interaction experiments were conducted in the presence of 3AT from 50 to 70 mM.
In the reproductive structures of P. aphrodite, miR164 is not expressed at the B2 stage, allowing a high expression of the PaCUP genes. Later, at the OF stage, mir164 is expressed; however, only the transcript levels of PaCUP1 decrease, in particular in the ovary, whereas PaCUP2 is still highly expressed, suggesting that after anthesis, the levels of PaCUP2 transcripts are not regulated by miR164. The activity of miR164 on the transcript levels of the homologs of the CUP genes was also proposed for A. thaliana, where miR164 is involved in reproductive structure formation and possibly modulates the transcript levels of the CUC1-2 genes (homologs of the CUPs) involved in ovary and column formation [31,69].
Finally, miR164 is expressed in the leaf tissue of P. aphrodite, where the PaCUP1-2 transcript levels are low. The miR164 of A. thaliana regulates the CUC1-2 genes during leaf shape formation and senescence. Mutants of miR164 have an unusual leaf shape, with leaves showing a more pronounced dentition than the wild type [78]. Our results suggest that, also in orchids, miR164 seems to regulate the PaCUP1-2 expression in the leaf.

PaCUP1-PaCUP2 Protein Interaction
The NAC transcription factors of Arabidopsis can form homo-and heterodimers due to their ability to interact through the NAC domains [25]. In addition, CUC1 and CUC2 in Arabidopsis can activate transcription due to the presence of acidic residues within the TAR domain [25]. In a Y2H assay, they activated the transcription of the yeast reporter genes [79]. For these reasons, we checked the autoactivation ability of the P. aphrodite PaCUP1-2 proteins and their ability to form homo-and heterodimers in the Y2H assay.
The yeast clones double-transformed with the recombinant vector pGBKT7 containing the full-length transcripts PaCUP1 or PaCUP2, and the empty vector PGADT7. The clones were grown on a selective medium without tryptophan, lysine, or histidine in the presence of different concentrations of 3AT (from 0 to 70 mM). Only the PaCUP1 protein alone activated the transcription of the yeast GAL4 gene in the absence of histidine and in the presence of up to 20 mM 3AT (Figure 8a). For this reason, the protein-protein interaction experiments were conducted in the presence of 3AT from 50 to 70 mM.
Our results reveal that PaCUP1 forms homodimers (Figure 8a), as observed in Arabidopsis [79], and, as expected, PaCUP1 and PaCUP2 proteins directly interact with each other (Figure 8b), showing that the CUP1-2 proteins have a conserved ability to crossinteract.  Our results reveal that PaCUP1 forms homodimers (Figure 8a), as observed in Arabidopsis [79], and, as expected, PaCUP1 and PaCUP2 proteins directly interact with each other (Figure 8b), showing that the CUP1-2 proteins have a conserved ability to cross-interact.

Conclusions
Orchidaceae represent approximately 10% of flowering plants, with many unique morphological, ecological, and physiological characteristics. Their cultivation and commercialization are a significant part of the worldwide floriculture trade. In this scenario, studies on the molecular basis of orchid flower development are particularly relevant, as they enrich the knowledge of how to develop specific breeding programs to obtain even more new cultivars and varieties.
For the first time in orchids, we analyzed the NAC transcription factor family, focusing on the CUP genes.
Our study reveals that orchid NAC proteins have a conserved domain organization and can be easily classified into different subfamilies based on their sequence.
The expression patterns of the P. aphrodite CUP genes suggest their possible involvement in flower formation and symmetry establishment. In addition, as previously described for different species, the transcript levels of orchid CUP1 and CUP2 genes seem to be regulated by miR164.
The data obtained in this work are preliminary and represent a starting point for understanding the role of these genes. Further investigations, in particular functional analyses, are needed to clarify the role of CUP genes during orchid flower formation.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/genes13122293/s1, Figure S1: Conserved motif for CUP proteins identified by MEME analysis, Table S1: Number of NAC genes isolated in all species analyzed in this work, Data S1: List of sequences, accession number, conserved motif and genome structures of the genes considered in this work; Data S2: Difference between the size of the intron-exon of the orchid gene species and A. thaliana; Data S3: CENSOR analysis results; Data S4: Statistical analysis results.

Conflicts of Interest:
The authors declare no conflict of interest.