The ancestral duplicated DL/CRC orthologs, PeDL1 and PeDL2, function in orchid reproductive organ innovation

Journal of Experimental Botany, Vol. 72, No. 15 pp. 5442–5461, 2021 doi:10.1093/jxb/erab195 Advance Access Publication 7 May 2021 RESEARCH PAPER The ancestral duplicated DL/CRC orthologs, PeDL1 and PeDL2, function in orchid reproductive organ innovation 1 Department of Life Sciences, National Cheng Kung University, Tainan, 701, Taiwan Institute of Tropical Plant Sciences and Microbiology, National Cheng Kung University, Tainan, 701, Taiwan 3 Orchid Research and Development Center, National Cheng Kung University, Tainan 701, Taiwan 4 Department of Statistics, National Cheng Kung University, Tainan, 701, Taiwan 5 Institute of Molecular Biology, National Chung Hsing University, Taichung, 40227, Taiwan 6 Bioproduction Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki 305–8566, Japan 7 Shanghai Key Laboratory of Plant Functional Genomics and Resources, Chenshan Plant Science Research Center, CAS, Shanghai Chenshan Botanical Garden, Shanghai, China 8 Key Laboratory of National Forestry and Grassland Administration for Orchid Conservation and Utilization at College of Landscape Architecture, Fujian Agriculture and Forestry University, Fuzhou 350002, China 2 * Correspondence: tsaiwc@mail.ncku.edu.tw Received 23 February 2021; Editorial decision 23 April 2021; Accepted 27 April 2021 Editor: Daphne Goring, University of Toronto, Canada Abstract Orchid gynostemium, the fused organ of the androecium and gynoecium, and ovule development are unique developmental processes. Two DROOPING LEAF/CRABS CLAW (DL/CRC) genes, PeDL1 and PeDL2, were identified from the Phalaenopsis orchid genome and functionally characterized. Phylogenetic analysis indicated that the most recent common ancestor of orchids contained the duplicated DL/CRC-like genes. Temporal and spatial expression analysis indicated that PeDL genes are specifically expressed in the gynostemium and at the early stages of ovule development. Both PeDLs could partially complement an Arabidopsis crc-1 mutant. Virus-induced gene silencing (VIGS) of PeDL1 and PeDL2 affected the number of protuberant ovule initials differentiated from the placenta. Transient overexpression of PeDL1 in Phalaenopsis orchids caused abnormal development of ovule and stigmatic cavity of gynostemium. PeDL1, but not PeDL2, could form a heterodimer with Phalaenopsis equestris CINCINNATA 8 (PeCIN8). Paralogous retention and subsequent divergence of the gene sequences of PeDL1 and PeDL2 in P. equestris might result in the differentiation of function and protein behaviors. These results reveal that the ancestral duplicated DL/ CRC-like genes play important roles in orchid reproductive organ innovation. Keywords: DROOPING LEAF, gynostemium, orchid, ovule, Phalaenopsis equestris, YABBY gene. © The Author(s) 2021. Published by Oxford University Press on behalf of the Society for Experimental Biology. All rights reserved. For permissions, please email: journals.permissions@oup.com Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 You-Yi Chen1,2,3, Yu-Yun Hsiao3, Chung-I Li4, Chuan-Ming Yeh5,6, Nobutaka Mitsuda6, Hong-Xing Yang7, Chi-Chou Chiu2, Song-Bin Chang1,3, Zhong-Jian Liu8 and Wen-Chieh Tsai1,2,3,*, Function of orchid DL/CRC genes | 5443 Introduction floral meristem determination, organ identity, fruit maturation, seed formation and plant architecture, and the outcomes of the interactions are essential to regulate varied orchid tepal morphogenesis (Pan et al., 2014). Recently, it has been reported that AGAMOUS-LIKE6 (AGL6)-like MADS-box genes have functions related to labellum development, and differential interactions among AGL6-like genes and AP3-like genes are required for sepal, petal, and labellum formation (Hsiao et al., 2013; Hsu et al., 2015). The plant-specific YABBY (YAB) gene family of transcription factor-encoding genes specifies the adaxial-abaxial leaf polarity (Bowman, 2000). In Arabidopsis, the YAB gene family contains six sub-families: the FILAMENTOUS FLOWER (FIL) sub-family, the YAB2 sub-family, the YAB3 sub-family, the YAB5 sub-family, the DROOPING LEAF/CRABS CLAW (DL/CRC) sub-family and the INNER NO OUTER (INO) sub-family, which encode proteins possessing a typical zinc finger domain and a YAB domain (Bowman and Smyth, 1999; Sawa et al., 1999; Siegfried et al., 1999; Villanueva et al., 1999). Previous studies indicated that the Arabidopsis CRC is specifically expressed at the carpel and plays an important role in carpel development and determinacy of the floral meristem (Alvarez and Smyth, 1999; Bowman and Smyth, 1999). The rice drooping leaf (dl) mutants showed that drooping leaf phenotypes resulted from the lack of a leaf midrib. In addition, the mutants also demonstrated that carpels homeotically transformed into stamens, indicating that DL regulates carpel identity (Nagasawa et al., 2003). It also indicated that DL interacts antagonistically with class B genes and controls floral meristem determinacy (Yamaguchi et al., 2004). Ancestral expression patterns and functional studies performed with Petunia hybrida, Nicotiana tabacum and Pisum sativum support an ancestral role of DL/CRC genes in the specification of carpel development and floral meristem termination (Lee et al., 2005;Yamada et al., 2011; Fourquin et al., 2014). Interestingly, in the basal eudicot California poppy, EcCRC not only harbors ancestral functions of DL/CRC but is also involved in ovule initiation (Orashakova et al., 2009). In this study, we identified eight YAB genes, including two in the CRC clade, one in the INO clade, three in the YAB2 clade and two members in the FIL clade, from the whole genome sequence of P. equestris (Cai et al., 2015). We determined the expression pattern of PeDLs and investigated their function by overexpression of PeDLs in wild-type Arabidopsis and its crc-1 mutant.Virus-induced gene silencing (VIGS) and transient overexpression in Phalaenopsis orchids were also adopted to functionally characterize PeDLs. We concluded that PeDLs not only have conserved functions in floral meristem determinacy and carpel development, but also have novel functions in stigmatic cavity formation and ovule initiation. Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 The fascinating and complex structures of the labellum and gynostemium in orchids have attracted great interest from evolutionary and developmental biologists. In addition to unique floral morphology, orchids are unusual among flowering plants in that in many species the ovule is not mature at the time of pollination, and the ovule initiation is precisely triggered by the deposition of pollinia into the stigmatic cavity (Tsai et al., 2008a; Chen et al., 2012; Dirks-Mulder et al., 2019). Carpels, the female reproductive organs of angiosperms, are one of the key morphological innovations for the success of flowering plants. Orchidaceae represents one of the largest families of angiosperms and is known for its complexity of flowers. Orchid flowers are composed of five whorls of three segments each, including the outermost whorl of sepals, the second whorl of two petals and a highly elaborated labellum with distinctive shapes and color patterns, the third and fourth whorls of six staminal organs, and the central whorl of fused male and female reproductive organs called the gynostemium. In addition, the abaxial side of the gynostemium forms the stigmatic cavity for deposition of pollinia. The innovated floral organs of the labellum and gynostemium make the orchid flower zygomorphic and are commonly invoked as a crucial pairing for attracting and interacting with pollinators (Rudall and Bateman, 2002). Over the last couple of decades, genes involved in angiosperm floral organ specification were identified and the ‘ABCDE model’ has been proposed, with the combined effect of the A-, B-, C-, D-, and E-class of MADS-box [MCM1 (MINI CHROMOSOME MAINTENANCE 1), AG (AGAMOUS), DEF (DEFICIENS), SERUM RESPONSE FACTOR (SRF)] genes determining the floral organ identity (Theissen et al., 2000). In orchids, the Phalaenopsis species is used as a model plant to delineate the B-, C-, D-, and E-class functions of MADS-box genes participating in specialized floral organ development (Pan et al., 2014; Tsai et al., 2014). The most notable feature is that P. equestris contains four B-class APETALA3 (AP3)-like genes with differential expression, largely responsible for differentiation of the two closely-spaced tripartite perianth whorls into three sepals of the outer whorl, versus the inner whorl of two lateral petals, and a median labellum (Tsai et al., 2004). Later, the refined ‘orchid code’ and ‘homeotic orchid tepal (HOT)’ models were proposed to illustrate the regulation of perianth morphogenesis in orchids (MondragonPalomino and Theissen, 2011; Pan et al., 2011). The studies of floral terata from Phalaenopsis and Cymbidium agreed that the function of C- and D-class MADS-box genes correlated with gynostemium and ovule development, respectively (Wang et al., 2011; Chen et al., 2012). In addition, a comprehensive study on the E-class MADS-box genes in P. equestris revealed that Phalaenopsis equestris SEPALLATA (PeSEP) proteins combined with B-, C- and D-class MADS-box proteins, are involved in 5444 | Chen et al. Materials and methods Genome-wide identification of YABBY genes from three sequenced orchid genomes The conserved YAB domain based on the hidden Markov model (HMM; PF04690) was obtained from the Pfam protein family database (http:// pfam.sanger.ac.uk). To identify the genes encoding the YABBY transcription factor in orchids, the HMM profile of the YABBY domain was used as a query for an HMMER search (http://hmmer.janelia.org/) of the P. equestris, Dendrobium catenatum, and Apostasia shenzenica genome sequences (E-value=0.00001; Cai et al., 2015; Zhang et al., 2016; 2017). Sequence alignment and phylogenetic analysis All the DL/CRC and other YAB protein sequences from diverse plant species were downloaded from the NCBI database using protein BLAST tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Multiple sequence alignment of the amino acid sequences of YAB proteins was performed by using ClustalW with default settings. The phylogenetic tree was generated using the neighbor-joining method in MEGA6 software (Tamura et al., 2013), and with the bootstrap of 1000 replications. The protein accession numbers for related proteins are listed in Supplementary Table S1. Sample collection and RNA isolation The five stages of developing flower buds were defined as B1 (0.5– 1.0 cm), B2 (1.0–1.5 cm), B3 (1.5–2.0 cm), B4 (2.0–2.5 cm), and B5 (2.5–3.0 cm), based on description by Pan et al. (2014). Definitions of various stages of the developing ovary and ovule from 0 to 64 days after pollination (DAP) were based on a previous study (Chen et al., 2012). The floral organs (sepal, petal, labellum, and gynostemium), pedicel, floral stalk, young leaf, root, and all the orchid P. subsp. formosana samples were frozen in liquid nitrogen and stored at −80°C. Total RNA was extracted following the guanidium thiocyanate method (O’Neill et al., 1993). Quantitative real-time PCR Total RNA was treated with DNase (NEB, Hertfordshire, UK) to remove remnant DNA. First-strand cDNA was synthesized using the Superscript III kit (Invitrogen, CA, USA). The quantitative real-time PCR was performed using SYBR GREEN PCR Master Mix (Applied Biosystems, Warrington, UK) on ABI 7500, Applied Biosystems System. The PCR was performed with the following reaction conditions: 95 °C for 10 min, 40 cycles of 95 °C for 15 s and 60 °C for 1 min. For realtime RT–PCR, each gene was analyzed in three biological and technical replicates. PeActin4 (PACT4, AY134752) and PeUbiquitin10 (Peq013061, XM_020735697) of Phalaenopsis were used as internal controls (Chen RNA in situ hybridization P. equestris flower buds, developing ovaries and ovules were fixed in 4% (v/v) paraformaldehyde and 0.5% (v/v) glutaraldehyde for 16–24 h at 4 °C. Samples were dehydrated with a graded ethanol series (20%, 30%, 50%, 70%, 95%, 100%). Tissues were sectioned between 6–8 μm with a rotary microtome (MICROM, HM 310, Walldorf, Germany), and stuck to poly-L-lysine-coated slides. Tissue sections were deparaffinized with xylene, rehydrated through an ethanol series (100%, 95%, 70%, 50%, 30%, 20%), and pre-treated with proteinase K (1 µg ml−1) in 1 × PBS at 37 °C for 30 min. Pre-hybridization and hybridization followed previous protocols with slight modification of wash conditions: twice of 1 × SSC at 45 °C for 20 min, and twice of 0.5 × SSC at 42 °C for 15 min (Tsai et al. 2005). The sense and antisense RNA probes used Digoxigenin-labeled UTP-DIG (Roche Applied Science, Branford, CT, USA). The probes containing partial specific regions (PeDL1, 150 bp long and PeDL2, 180 bp long) were produced following the manufacturer’s instructions (Roche Applied Science). PCR products were amplified with the primers listed in Supplementary Table S2. Arabidopsis transformation The gynostemium cDNA of P. equestris was used for PeDL1 and PeDL2 genes isolation and cloning. Full-length sequences of PeDL1 and PeDL2 were cloned into the pBI121 (Clontech, USA) vector digested with SmaI restriction enzyme to generate the construct 35S:PeDL1-GUS and 35S:PeDL2-GUS, respectively. The plasmids were transformed into Agrobacterium tumefaciens (strain GV3101), and was used to transform Arabidopsis thaliana ecotypes Columbia (Col) by floral dip mothod (Clough and Bent, 1998). After gene transformation in Arabidopsis, the seeds (T0) were sown on Murashige and Skoog (MS) media with 50 µg ml-1 kanamycin (Sigma-Aldrich, USA). Kanamycin segregation in the T1 generation was analyzed by χ 2 test. The resistant plants of the T2 generation were checked to confirm the integration fragment by PCR for each transgenic line.The T3 homozygous plants were used for subsequent phenotype observation and functional analysis.The transgenic plants were transferred to grow on soil as described previously (Chen et al., 2012).In this study, we analyzed six independently transformed PeDL1 transgenic lines (T3) and eight independently transformed PeDL2 transgenic lines (T3) for statistical analysis of phenotypic data. The total seed numbers per silique and silique length were measured using mature siliques. For statistical analysis, twenty samples of WT, 35S::PeDL1 and 35S::PeDL2 plants were collected and analyzed from three strong severity T3 lines. All the primers used in this study are listed in Supplementary Table S2. Complementation test To generate 35S:PeDL1 crc-1 and 35S:PeDL2 crc-1 plants, Arabidopsis thaliana ecotype Columbia (Col) pistils of transgenic 35S::PeDL1 and 35S::PeDL2 homozygous plants were pollinated with pollen grains from crc-1 homozygous mutants of Landsberg erecta (Ler). All plants in the F1 generation were kanamycin resistant. In the F2 generation (by selfing of F1), approximately 75% of the plants were kanamycin resistant. Six of the 35S:PeDL1 crc-1 and 35S:PeDL2 crc-1 F2 generation plants were used for subsequent phenotype observation and functional analysis. All the F2 generation independent lines showed a similar phenotype. For identification of crc-1 homozygous plants carrying 35S:PeDL1 or 35S:PeDL2, genomic DNA was extracted and verified by both PCR and sequencing. Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Plant materials Native species Phalaenopsis equestris, P. aphrodite subsp. formosana, and a Phalaenopsis cultivar P. Sogo Yukidian ‘V3’ were grown in the glasshouse at National Cheng Kung University (NCKU,Taiwan) under natural light (photosynthetic photon flux density, 90 μmol m−2 s−1) and controlled temperature from 23 °C to 27 °C. The floral organ cDNA of P. equestris was used for YABBY genes isolation and cloning. Spatial and temporal expression was examined by quantitative real-time PCR (qRT-PCR) using various organs of the P. aphrodite subsp. formosana plants. A Phalaenopsis cultivar of P. Sogo Yukidian V3’ was used for virus-induced gene silencing (VIGS) and transient overexpression. The Arabidopsis thaliana ecotypes Columbia (Col) was used for PeDL1 and PeDL2 gene overexpression. The Landsberg erecta (Ler) was used for complementation tests. et al., 2005; Cai et al., 2015), and the data were analyzed with the Sequencing Detection System v1.2.3 (Applied Biosystems). All the primers used in this study are listed in Supplementary Table S2. Function of orchid DL/CRC genes | 5445 Transient overexpression of PeDLs in Phalaenopsis orchid The gynostemium cDNA of P. equestris and the primers (Supplementary Table S2) including the attB1 or attB2 adapter sequence were used to amplify PeDLs with the use of Phusion® High-Fidelity DNA Polymerase (NEB, Beverly, MA, USA). The amplified sense full-length cDNA fragments and 35S-CymMV-Gateway vector were used for BP recombination with the BP clonase of the Gateway system (Lu et al., 2007). The vectors carrying the genes that were transferred to Phalaenopsis orchids were the same as those used in VIGS. We generated 24 independent PeDL1 and 24 PeDL2 transiently overexpressed plants and 24 mock plants (empty 35S-CymMV-Gateway vector) of P. ‘Sogo Yukidia V3’. Each treatment involved eight plants, and repeated three times individually. Real-time RT-PCR was used to examine the overexpression of PeDLs in gynostemium. Scanning electron microscopy Arabidopsis transgenic plants, VIGS, and transient overexpression Phalaenopsis plants were analyzed by using scanning electron microscopy (Hitachi TM3000, Tokyo, Japan). The plant samples were frozen in liquid nitrogen and low vacuum conditions with unfixed material. The rosette leaf epidermal cell and rosette leaf area image data of transgenic and wild-type (WT) Arabidopsis thaliana were analyzed using ImageJ software (http://rsb.info.nih.gov/ij/) for 35-day-old plants. Ten Arabidopsis samples of WT, 35S::PeDL1, and 35S::PeDL2 plants were collected and analyzed from three T3 lines that showed strong phenotypes . The Arabidopsis transgenic plants among T3 showed the most obvious phenotype compared to that of wild type. Yeast two-hybrid assay For systematically analyzing interaction behavior of PeDLs, PeDL1 and PeDL2 were cloned individually into a pDEST_GBKT7 bait vector, which was obtained from the Arabidopsis Biological Resource Center (ABRC, Columbus, OH, USA) from the Gateway entry clone. The Y2H Gold yeast strain (Takara Bio USA Inc.) was employed for the assay. After selecting the yeast harboring, PeDL1 or PeDL2 was constructed on bait plasmid (pDEST_GBKT7), the prey library composed of approximately 1920 transcription factor cDNAs of Arabidopsis thaliana that was constructed on a pDEST_GADT7 vector (Mitsuda et al., 2010). These yeast strains were additionally transformed and spotted onto positive control medium [6.7 g l-1 yeast nitrogen base without amino acids (Difco Laboratories, Detroit, Mich., USA), 20 g l-1 dextrose, 20 g l-1 agar)], which lacked leucine and tryptophan, and test medium, which lacked histidine, leucine, and tryptophan.Yeast growth was observed daily for several days. The results showed that PeDL1 and PeDL2 were able to interact with several Arabidopsis TEOSINTE-BRANCHED/CYCLOIDEA/PCF (TCP) transcription factors, including members in Proliferating Cell Factor (PCF), CINCINNATA (CIN) and CYCLOIDEA (CYC) sub-families (Supplementary Tables S3; S4; Fig. S11). Because of low expression of several TCP genes in Phalaenopsis (Lin et al., 2016), we selected three P. equestris CYCLOIDEA (PeCYC) genes, two P. equestris CINCINNATA (PeCIN) genes (PeCIN7 and PeCIN8), , and four P. equestris Proliferating Cell Factor (PePCF) genes with relative high expression for further confirmation.Y2H assays were conducted with the MATCHMAKER II system (Clontech, Palo Alto, CA, USA). The target gene full-length sequences (PeDL1, PeDL2 and PePCF4 with restriction enzyme site of SmaI, PeCIN7, PeCIN8, PePCF5, PePCF7, PeCYC1, PeCYC2, and PeCYC3 with restriction enzyme site of BamHI, and PePCF10 with restriction enzyme site of NdeI) were cloned into pGBKT7 bait vectors (GAL4 DNA-binding domain; BD) and pGADT7 prey vectors (GAL4 activation domain; AD). All the primers used in this study are listed in Supplementary Table S2. The assay was performed as previously described (Pan et al. 2014) with slight modification. Two plasmids were co-transformed into AH109 yeast cells and selected on medium lacking leucine and tryptophan (SD-Leu/-Trp; Sigma-Aldrich, St. Louis, MO, USA). The yeast cells were evaluated by serial dilutions and spotting assays.The yeast cells were spotted on SD-Trp-Leu and SD-TrpLeu-His (in the absence or presence of 3-amino-1,2,4-aminotriazole, 3-AT) plates and incubated until visible colonies were formed. All spotting assay sets were performed as at least three independent experiments. Preparation of Phalaenopsis petal protoplasts Petal protoplasts were isolated from the fully blooming flowers of P. aphrodite subsp. formosana. Approximately 5 g of flower petals were sterilized in 70% ethanol for 1 min, and washed in sterilized distilled water three times. Petals were cut into small pieces and mixed with 8 ml enzyme solution (4% Onozuka R-10 cellulase, 2% Macerozyme R-10, 0.6 M sucrose and cell protoplast washing salt solution (1 mM KH2PO4, 5 mM NH4NO3, 5 mM sodium citrate, 0.6 M mannitol, pH 5.8). The digestion was carried out in the dark with gentle shaking at 50 rpm at 25 °C for 2 h. After incubation, the solution was filtered through steel sieves with 100 μm and 50 μm pore sizes to remove the undigested materials. The filtrate was transferred to 15 ml centrifuge tubes and centrifuged at 50 × g for 5 min. The viable protoplasts floating on the surface were collected. The purified protoplasts were suspended in 2 ml W5 solution (154 mM NaCl, 125 mM CaCl2, 5 mM KCl, 2 mM MES pH 5.7, 5 mM glucose). The protoplasts were assessed for viability using a haemocytometer and a fluorescence microscope (Leica, DMI 3000B, Germany). Bimolecular fluorescence complementation (BiFC) assay The gynostemium cDNA of P. equestris and the primers (Supplementary Table S2) including the attB1 or attB2 adapter sequence were used to amplify PeDL1 and PeCIN8 (without stop codon) with the use of Phusion® High-Fidelity DNA Polymerase (NEB, Beverly, MA, USA). The amplified full-length cDNA fragments of PeDL1 and PeCIN8 and the pDONR 221 vector (Invitrogen, ,Carlsbad, CA, USA) was used for BP recombination with the BP clonase of the Gateway system. BP reaction products were transformed into E. coli DH5α cells, and bacteria were plated on LB medium containing kanamycin. Gateway-compatible vectors [pSAT5(A)-DEST-cEYFP-N1 and pSAT5-DEST-cEYFP-C1; Lin et al., 2016) were used to generate expression vectors by a LR recombination reaction. The N-terminus (YN) of yellow fluorescent protein Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Virus-induced gene silencing (VIGS) of PeDLs in Phalaenopsis orchids The gynostemium cDNA of P. equestris and the primers (Supplementary Table S2) containing the attB1 or attB2 adapter sequence were used to amplify antisense PeDLs with the use of Phusion® High-Fidelity DNA Polymerase (NEB, Beverly, MA, USA).The amplified antisense full-length cDNA fragments and 35S-CymMV-Gateway vector were used for BP recombination with the BP clonase of the Gateway system (Lu et al., 2007). The recombinant vectors were transformed into Agrobacterium tumefaciens (strain EHA105). The bacterial cells with 35S-CymMV-Gateway expression vectors containing antisense PeDL1 or PeDL2 were individually injected into the P. ‘Sogo Yukidian V3’ inflorescence spike and leaf right above the inflorescence. An empty 35S-CymMV-Gateway vector was also transformed into Agrobacterium tumefaciens (strain EHA105), and the bacteria were injected into the P. ‘Sogo Yukidian V3’as a negative control (mock) to confirm that any flower phenotype changes were not due to the viral infection. This experiment was performed as described by Hsu et al. (2015). We generated seven independent PeDL1, 7 PeDL2 and nine mock genesilenced plants of P. ‘Sogo Yukidian V3’. Detection of Cymbidium mosaic virus (CymMV) in VIGS-treated plants was performed by using RT– PCR. Real-time RT–PCR was used to examine the expression of silenced genes in the gynostemium and ovary. All the primers used in this study are listed in Supplementary Table S2. 5446 | Chen et al. (YFP) was cloned upstream of PeDL1 and PeCIN8 in the pE-SPYNE vector, and the C-terminus of yellow fluorescent protein (YC) was fused upstream of PeDL1 and PeCIN8 in the pE-SPYCE vector. Empty vectors YN and YC were used as negative controls.The PeMADS6 (YN) and PeMADS4 (YC) were used as positive controls (Tsai et al., 2008b). LR reaction products were transformed into E. coli DH5α cells, and bacteria were plated on LB medium containing ampicillin. The plasmids were transfected into Phalaenopsis petal protoplasts by PEG transformation, as previously described (Lin et al. 2018).The signals were visualized by confocal laser microscopy (Carl Zeiss LSM780, Jena, Germany). Primers used for this study are listed in Supplementary Table S2. Identification and phylogenetic analysis of orchid YABBY genes The P. equestris, D. catenatum, and A. shenzhenica genomes encode eight, seven, and five YAB genes, respectively.The number Fig. 1. Sequence analysis of plant YABBY genes. (A) Protein sequence alignment of the C2C2 zinc finger domain and YABBY domain from rice, Arabidopsis and P. equestris. Consensus is denoted by the color black. Conserved cysteine residues in the zinc-finger domain are indicated with red asterisks. (B) Phylogenetic analysis of YABBY-related proteins from the angiosperms and gymnosperms. The PeYABBY-related proteins are marked by red asterisks. The tree was generated using MEGA6.0 neighbor-joining software with 1000 bootstrap trials. Accession numbers are listed in Supplementary Table S1. Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Results of YAB genes of P. equestris and D. catenatum is comparable to that of Arabidopsis (six members) and rice (eight members). The multiple sequence alignment analysis showed that protein sequences translated by eight YAB genes from P. equestris harbor a standard zinc finger domain and YAB domain (Fig. 1A; Bowman and Smyth, 1999).To determine the phylogenetic relationships among the orchid and other plant YAB genes, the phylogeny of known YAB genes was reconstructed using the conceptual amino acid sequences of respective genes as input data. The topology of the phylogenetic tree obtained indicated that the orchid and rice genes reported here resided well in the four clades including YAB2, DL/CRC, INNER NO OUTER (INO) and FILAMENTOUS FLOWER (FIL), using gymnosperm YAB genes as an outgroup (Fig. 1B; Finet et al., 2016). The results suggest that the last common ancestor of monocots might have lost the YAB5 sub-family (Fig. 1B). Furthermore, PeYAB1 and PeFIL belong to the FIL clade; PeYAB2, PeYAB3 Function of orchid DL/CRC genes | 5447 Expression profiles of PeDL genes in P. aphrodite subsp. formosana The presence of two DL/CRC-like paralogs in Phalaenopsis raises the question of whether the two genes are functionally redundant or are regulated differentially, either temporally or spatially during reproductive organ development. To distinguish between these two possibilities, P. aphrodite subsp. formosana plant tissues of the various developing flower buds (Fig. 2A, B), flower organ (Fig. 2C), pedicel (Fig. 2D), floral stalk (Fig. 2B), young leaf (Fig. 2E), root (Fig. 2E) and ovary (Fig. 2F-O) were harvested at normal greenhouse growth conditions. The two expression patterns of PeDL paralogs were investigated by quantitative real-time RT-PCR. (Fig. 2P-U; Supplementary Fig. S2A-F). The results showed that both PeDL1 and PeDL2 were predominantly expressed at early flower bud developmental stages (Fig. 2P, Q; Supplementary Fig. S2A, B). In addition, transcripts of both PeDL1 and PeDL2 could be specifically detected in gynostemium (Fig. 2R, S; Supplementary Fig. S2C, D).The expression of these two genes was not detectable in vegetative tissues, such as pedicels, floral stalks, leaves and roots (Fig. 2R, S; Supplementary Fig. S2C, D). According to the highly consistent spatial and temporal expression patterns of these two PeDL genes in flowers, it is possible that both PeDL genes have redundant functions for gynostemium development. YAB genes are also involved in ovule development, and pollination is a key regulatory event in orchid ovule initiation. We determined the temporal mRNA expression patterns of both PeDLs in developing ovules triggered by pollination (Fig. 2F-O). Expression of both PeDL1 and PeDL2 were detected from zero (before pollination) to 32 days after pollination (DAP). After pollination, the expression of PeDL1 gradually decreased up to 32 DAP (Fig. 2T; Supplementary Fig. S2E). However, the expression of PeDL2 gradually increased to 4 DAP, and then decreased, and then was not detectable after 32 DAP (Fig. 2U; Supplementary Fig. S2F). Therefore, the functions of both PeDLs may be associated with ovule development. Based on the substantially differential expression patterns of the two PeDL paralogs, it is possible that they have functional differences in ovule initiation. The detailed spatial and temporal expression patterns of the PeDL genes in gynostemium and ovule development was further confirmed by using in situ hybridization antisense RNA as a probe. Both PeDL transcripts were detected in the inflorescence meristem and floral primordia (Fig. 3A; Supplementary Fig. S3A [antisense probes]; Fig. 3B; Supplementary Fig. S3B [sense probes]). At early flower bud stage, transcripts of PeDL1 and PeDL2 were strongly detected in the gynostemium (column) and column foot, the basal protruding part of the column (Fig. 3C; Supplementary Fig. S3C [antisense probes]; Fig. 3D; Supplementary Fig. S3D [sense probes]). At the later developmental stages, the transcripts of PeDL1 were more strongly concentrated at the rostellum, the projecting part of the column that separates the male androecium from the female gynoecium, and the column foot (Fig. 3E; Supplementary Fig. S3E [antisense probes]; Fig. 3F; Supplementary Fig. S3F [sense probes]). In the process of ovule development, PeDL transcripts were detected in the pericarp and placenta before pollination (Fig. 3G; Supplementary Fig. S3G [antisense probes]; Fig. 3H; Supplementary Fig. S3H [sense probes]). At 4 and 8 DAP, both PeDLs were expressed in the pericarp and developing ovules (Fig. 3I, K; Supplementary Fig. S3I, K [antisense probes]; Fig. 3J, L; Supplementary Fig. S3J, L [sense probes]). At 16 DAP, the signals of both PeDLs could be detected in developing ovules and the first cell layers of the placenta (Fig. 3M; Supplementary Fig. S3M [antisense probes]; Fig. 3N; Supplementary Fig. S3N [sense probes]). These results suggest that both PeDL genes might be involved in gynostemium development as well as ovule initiation. Functional characterization of the two PeDL genes using transgenic Arabidopsis To further investigate the role of the putative function of PeDL genes, we overexpressed PeDL1 and PeDL2 in Arabidopsis. Thirteen independently transformed PeDL1 transgenic Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 and PeYAB4 are classified in the YAB2 clade; PeINO is in the INO clade; and PeDL1 and PeDL2 are members of the DL/ CRC clade (Fig. 1B). It has been indicated that the DL/CRC sub-family represents a single orthologous lineage, without ancient duplications (Lee et al., 2005). Interestingly, we found that both P. equestris and D. catenatum have two DL/CRC-like genes and only one was identified in Apostasia (Fig. 1B).We further compared the gene structures of DL/CRC genes among P. equestris, A. shenzhenica, D. catenatum, Cabomba caroliniana, Oryza sativa and A. thaliana. All the DL/CRC genes have one intron located in the zinc finger domain, three introns located in the YAB domain and one intron situated in the non-conserved region between the zinc finger domain and YAB domain (Supplementary Fig. S1A, B). Interestingly, the sixth intron could be found in the region downstream of the YAB domain in most of the plants, except in Cabomba caroliniana CRABS CLAW (CcCRC) and Dendrobium catenatum DROOPING LEAF2 (DcaDL2). The multiple alignments of protein sequences showed that both PeDL1 and PeDL2 contain the typical zinc finger domain and YAB domain which are consistent with other plant species with DL/CRC orthologous proteins (Supplementary Fig. S1C). The phylogenetic tree showed that orchid DL/CRC form a monophyletic clade with two sub-clades, including AshDL, PeDL1 and DcaDL1 in sub-clade I, and PeDL2 and DcaDL2 in sub-clade II (Supplementary Fig. S1D). Paralogue retention and subsequent divergence of gene sequences of the DL/CRC-like genes in P. equestris and D. catenatum might have resulted in the functional differentiation of the DL/CRC-like genes. 5448 | Chen et al. Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Fig. 2. Analyses of spatial and temporal expression patterns of PeDL1 and PeDL2 in various tissues of P. aphrodite subsp. formosana by quantitative real-time RT-PCR. (A) Flower bud development stages from B1 to B5. Scale bar =1 cm. (B) Flower buds, fully blooming flowers and floral stalk in mature Phalaenopsis plants. Scale bar =10 cm. (C, D) Structure of P. aphrodite subsp. formosana flower. Scale bar =1cm. (E) Leaf and root at the vegetative Function of orchid DL/CRC genes | 5449 F, G; Bowman and Smyth, 1999). Overexpression of PeDL1 and PeDL2 in crc-1 plants dramatically influenced the development of the gynoecium. The carpels of 35S:PeDL1 crc-1 plants showed restored phenotype with smaller gaps at the apical region of the gynoecium compared with that of WT plants (Fig. 5A, D, H). The 35S:PeDL2 crc-1 plants presented perfect fused carpels, although the styles and replum shriveled severely (Fig. 5A, E, I). These results suggest that the PeDLs play a conserved role in controlling carpel development as eudicot CRCs, and can partially restore the defect of crc-1 mutant to different extent. Virus-induced gene silencing (VIGS) of PeDLs in Phalaenopsis To characterize the function of PeDLs in Phalaenopsis orchids, we performed VIGS of PeDL genes in P. ‘Sogo Yukidian V3’. We generated seven independent PeDL1-silenced plants, seven PeDL2-silenced plants, and nine mock-treated (empty 35S-CymMV-Gateway vector) plants of P. ‘Sogo Yukidian V3’ by using the VIGS strategy. The RT-PCR results showed that the infection rate was 89% (eight out of nine plants) in mock-treated plants, 78% (seven out of nine plants) in VIGSPeDL1 plants and 78% (seven out of nine plants) in VIGSPeDL2 plants (Supplementary Fig. S7). PeDL1 and PeDL2 expression in gynostemium was reduced in PeDL1 and PeDL2-silenced plants (Supplementary Figs S8G, H; S9A, B), but PeDL1 expression in gynostemium was not obviously affected in VIGS-PeDL2 plants (Supplementary Figs S8H; S9B). However, no phenotypic differences were observed in flower and gynostemium of both PeDL1- and PeDL2-silenced plants (Supplementary Fig. S8A-F). Furthermore, we detected reduced expression of both PeDL1 and PeDL2 in ovaries of silenced plants at 16 DAP (Fig. 6C, E; Supplementary Fig. S9C, D). We further analyzed the phenotype of developing ovaries of PeDL-silenced plants by scanning electron microscopy (SEM), although the ovary length and width 16 DAP showed no difference compared with control plants (Fig. 6A, B, D). The results indicated that the size of the placenta in PeDL1- and PeDL2-silenced plants were increased, compared with those of mock plants (Fig. 7A-I). The number of protuberant ovule initials differentiated from the placenta was obviously increased in PeDL1- and PeDL2-silenced plants (mock: 19.1;VIGS-PeDL1: 27.6 and VIGS-PeDL2: 32; Fig. 7J). These results suggest that both PeDL1 and PeDL2 were involved in ovule initiation in Phalaenopsis orchids. seedling stage. Scale bar =1 cm; (F) Morphological changes in the ovary at the indicated days after pollination (DAP). Scale bar = 1 cm. (G, H) Expression patterns of PeDL1 and PeDL2 in developmental stages of flower buds. (I, J) Expression patterns of PeDL1 and PeDL2 in various plant organs; (K, L) Expression patterns of PeDL1 and PeDL2 in various ovule developmental stages. B1, stage 1 flower buds (0.5–1 cm); B2, stage 2 flower buds (1–1.5 cm); B3, stage 3 flower buds (1.5–2 cm); B4, stage 4 flower buds (2–2.5 cm); B5, stage 5 flower buds (2.5–3 cm); Co, column (gynostemium); DAP, days after pollination; fs, floral stalk; Lf, leaf; Li, lip; Ova, ovary; Pe, petals; pedi, Pedicel; Rt, root; Se, sepals. The expression patterns of PeDL1 or PeDL2 were determined using three replicates and were normalized using PeActin4. Error bars indicate standard error of means (n=3) and different letters indicate significant differences at P=0.05 (Tukey’s honestly significant difference test). Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 lines showed kanamycin resistance and a similar phenotype. Among 13 transgenic lines, six showed a 3:1 segregating kanamycin resistance phenotype. A total of 21 independently transformed PeDL2 transgenic lines showed kanamycin resistance with similar phenotype. Among 21 transgenic lines, eight showed a 3:1 segregating kanamycin resistance phenotype. Both overexpressed plants showed reduced plant size, and 35S::PeDL1 transgenic plants were smaller than 35S::PeDL2 (Supplementary Fig. S4A). In addition, both transgenic plants produced rosette leaves curled towards the abaxial side with wrinkled surfaces (Supplementary Fig. S4A). The observations of Arabidopsis epidermal cells in rosette leaf revealed that in both 35S::PeDL1 and 35S::PeDL2 transgenic lines, the epidermal cell size was decreased (Supplementary Fig. S4B-D, H), along with increased cell density in the adaxial side of rosette leaves (Supplementary Fig. S4C, D, I). However, no significant changes in abaxial epidermal cells were observed among wild type (WT) and transgenic plants (Supplementary Fig. S4E-G, I). Both overexpressing transgenic plants showed short siliques containing reduced numbers of seeds, compared with those of WT plants (Fig. 4A, E-J). The seed size of 35S::PeDL1 plants was larger than that of WT plants and the seed weight was significantly increased (P<0.001) (Supplementary Fig. S5AE). Interestingly, both 35S::PeDL1 and 35S::PeDL2 plants presented shriveled styles and replum, and the style cell size was smaller than that of WT (Fig. 4B-D). These observed gynoecium abnormalities of overexpressing plants suggest that PeDL1 and PeDL2 are important for gynoecium development. We also noticed that the primary inflorescence apices were abnormally terminated and showed a cluster of flower buds in 35S::PeDL1 and 35S::PeDL2 plants (Supplementary Fig. S6A-D).This phenotype implies that PeDL1 and PeDL2 might be involved in termination of the primary inflorescence meristem. Previous studies also indicated that DL/CRC genes are involved in floral meristem determinacy (Alvarez and Smyth, 1999; Li et al., 2011). To investigate the functions of the PeDL genes, the Arabidopsis transgenic 35S::PeDL1 and 35S::PeDL2 homozygous plants were pollinated with pollen grains of crc-1 Arabidopsis homozygous mutants. The six 35S-PeDL1 crc-1 and six 35S-PeDL2 crc-1 F2 generation plants were used for subsequent phenotype observation and functional analysis. The results revealed that both PeDL genes could partially rescue the crc-1 mutant. The crc-1 mutant flower was similar to that of WT plants, except the carpels were not fused at the apical region (Fig. 5A-C, 5450 | Chen et al. Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Fig. 3. RNA in situ hybridization of PeDL1 in longitudinal sections of developing floral buds, and cross sections of developing ovules of P. equestris. In panels A, C, E, G, I, K, M, antisense probes were used to detect PeDL1 transcripts. In panels B, D, F, H, J, L, N, hybridization was done with PeDL1 sense probes (negative controls). Sections were hybridized with the antisense 3’-specific PeDL1 RNA probes or sense RNA probes. (A, B) Inflorescence meristem and floral primordia; (C, D) longitudinal section of the flower buds at an early stage; (E, F) longitudinal section of the flower buds at a late stage; (G, H) ovary tissue before pollination; (I, J) placenta with ovule primordia at 8 DAP; (K, L) placenta with ovule primordia at 4 DAP; (M, N) placenta with ovule primordia at 16 DAP. Ac, anther cap; b, bract; cf, column foot; co, column; en, endocarp; fp, floral primordia; im, inflorescence meristem; li, lip; ovp, ovule primordia; pe, petal; p, placenta; po, pollinium; ro, rostellum; se, sepal; vb, vascular bundles. (A-N) Scale bar =100 μm. Transient overexpression of PeDL genes in Phalaenopsis orchid Although VIGS of PeDL1 and PeDL2 showed phenotypic changes in Phalaenopsis ovary, further confirmation was obtained with a gain-of-function approach. To characterize the function of PeDLs in Phalaenopsis orchids, we performed transient overexpression of PeDL genes in P. ‘Sogo Yukidian V3’. The flower size and the morphology of sterile floral organs were similar between mock and OE-PeDL1 plants Function of orchid DL/CRC genes | 5451 Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Fig. 4. Morphology of siliques in WT (Col), 35S::PeDL1 (Col) and 35S::PeDL2 (Col) Arabidopsis transgenic plants. (A) Comparison of immature siliques in WT (left), 35S::PeDL1 (middle) and 35S::PeDL2 (right) transgenic plants. Scale bar=0.25 cm. SEM of (B) WT, (C) 35S::PeDL1 and (D) 35S::PeDL2 transgenic Arabidopsis. Scale bar=500 μm. (E) Comparison of mature siliques in WT (left), 35S::PeDL1 (middle) and 35S::PeDL2 (right) transgenic plants. Scale bar=0.25 cm. (F–H) Siliques of (F) WT plants, (G) 35S::PeDL1 and (H) 35S::PeDL2 transgenic plants. (F, G) Images of siliques with valves removed. Scale bar=0.25 cm. (I) Mean number of silique length in WT, 35S::PeDL1 and 35S::PeDL2 plants. Asterisks indicate statistically significant differences (*** P<0.001, compared with WT by Student’s t-test). Error bars represent the SD of three biological repeats (n=20 each). (J) Mean number of seeds per silique in WT, 35S::PeDL1 and 35S::PeDL2 plants. Asterisks indicate statistically significant differences (***P<0.001, compared with WT by Student’s t-test). Error bars represent the SD of three biological repeats (n=20 each). 5452 | Chen et al. Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Fig. 5. Ectopic expression of PeDL1 and PeDL2 in the crc-1 Arabidopsis mutant. (A) Comparison of immature siliques in WT (Ler), crc-1, overexpression 35S-PeDL1 crc-1 and overexpression 35S-PeDL2 crc-1 transgenic plants. Scale bar=0.5 cm. (B) Immature siliques of a WT plant. (C) Immature silique in crc-1 mutant. (D) Immature silique of a transgenic crc-1 mutant plant ectopically expressing PeDL1. (E) Immature silique of a transgenic crc-1 mutant plant ectopically expressing PeDL2. (F) SEM of WT immature silique. (G) SEM of crc-1 mutant immature silique. (H) SEM of immature silique of a transgenic crc-1 mutant plant ectopically expressing PeDL1. (I) SEM of immature silique of a transgenic crc-1 mutant plant ectopically expressing PeDL2. (B–E) Scale bar =100 μm; (F–I) Scale bar =500 μm. Function of orchid DL/CRC genes | 5453 (Fig. 8A–D). In addition, 24 transient overexpression lines of PeDL2 (OE-PeDL2) did not cause obvious phenotype changes. Twenty-four plants with transient overexpression of PeDL1 (OE-PeDL1) including three strong (Fig. 8F), six moderate (Fig. 8G), and eight weak (Fig. 8H) phenotypic plants with abnormal stigmatic cavity morphology, were obtained. The stigmatic cavities with strong severity showed that a protruding tissue grew from the bottom and almost covered the cavity (Fig. 8F). The stigmatic cavities with moderate severity had a sharp (Fig. 8J) or flat tissue (Fig. 8G; Supplementary Fig. S10B) extending from the bottom of the cavity. The ones with weak severity presented unapparent v-shapes (Fig. 8H, K; Supplementary Fig. S10C) at the bottom of the stigmatic cavity, compared with the v-shapes of mock plants (Fig. 8E, I; Supplementary Fig. S10A). Comparing the expression pattern between the mock (control) and OE-PeDL1 plants, PeDL1 had higher expression in OE-PeDL1 plants compared with that of the mock (control) plants (Fig. 8L). These results showed that the expression of PeDL1 in OE-PeDL1 was positively correlated with the phenotype severity of the stigmatic cavity (OE-PeDL1-2, strong severity; OE-PeDL1-4, moderate severity; OE-PeDL1-5, weak severity), suggesting that the restoration of phenotypes was dosage-dependent. These results indicated that the function of PeDL1 might be specifically associated with the morphogenesis of the stigmatic cavity. PeDL1 influences endocarp development and ovule initiation One of the remarkable reproductive characteristics of most orchid species is that ovary and ovule development are precisely initiated following pollination (Tsai et al., 2008a). Interestingly, Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Fig. 6. Phenotypic and expression analysis of the ovary in mock, VIGS-PeDL1 and VIGS-PeDL2 plants. (A) Comparison of mock (upper), VIGS-PeDL1 (middle) and VIGS-PeDL2 (lower) capsules at 16 DAP. Scale bar=1 cm. (B) Mean number of ovary lengths in mock-treated, VIGS-PeDL1 and VIGSPeDL2 plants. (C) Relative expression of PeDL1 and PeDL2 in 16 DAP capsules of the mock and VIGS-PeDL1 plants. (D) Mean number of capsule widths in mock, VIGS-PeDL1 and VIGS-PeDL2 plants. (E) Relative expression of PeDL1 and PeDL2 in 16 DAP capsules of the mock and VIGSPeDL2 plants. (B, D) Error bars indicate SE of means (n=5) and different letters indicate significant differences at P=0.05 (Tukey’s honestly significant difference test). (C, E) The expression of PeDL1 or PeDL2 were determined using three replicates and was normalized using PeActin4. Mock: an empty 35S-CymMV-Gateway vector was transformed into Agrobacterium tumefaciens (strain EHA105), and injected into the P. Sogo Yukidian ‘V3’. Expression of each gene in various plants was relative to that in the mock plant that was set to 100%. Error bars indicate SE of means (n=3) and different letters indicate significant differences at P=0.05 (Tukey’s honestly significant difference test). 5454 | Chen et al. OE-PeDL1 plants showed that ovary growth was affected after pollination. Before pollination, the ovary morphology was not different between OE-PeDL1 and mock plants (Supplementary Fig. S12A-D). SEM showed that the mock plants presented almost completely filled endocarps (Fig. 9A, C). However, SEM observations revealed that several empty spaces existed among three lobes of the endocarp in OE-PeDL1 plants (Fig. 9B, D). At 16 DAP, the ovaries of OE-PeDL1 plants did not obviously enlarge compared with those of mock plants (Supplementary Fig. S12E, F). The number of protuberant ovule initials differentiated from the placenta was obviously decreased in OE-PeDL1 (Fig. 9E-H). These results suggest that PeDL1 is involved in endocarp development and ovule initiation in Phalaenopsis orchids. Interaction behaviors of PeDL1 It has been previously reported that CRC could form homodimers and heterodimers with INNER NO OUTER (INO; Gross et al., 2018). However, our yeast-two hybrid assays indicated that PeDL1 and PeDL2 did not show the ability to form homodimers or heterodimers (Supplementary Fig. S13A, C). To identify transcription factors that interact with PeDL proteins, systematic screening of an Arabidopsis yeast-two hybrid transcription factor library composed of approximately 1920 transcription factors (Mitsuda et al., 2010), was performed with PeDL proteins as preys. The results showed that PeDL1 and PeDL2 were able to interact with several Arabidopsis TCP transcription factors (Supplementary Fig. S14; Tables 3, 4). In plants, TCP genes participate in regulating various features of plant development, for example, in flower development (Koyama et al., 2011; Uberti-Manassero et al., 2012), seed germination (Tatematsu et al., 2008; Rueda-Romero et al., 2012), and leaf development (Kieffer et al., 2011). We inferred that PeDL1 and PeDL2 might have the ability to form heterodimers with TCP proteins in Phalaenopsis orchids. We further examined the protein interactions among PeDLs and Phalaenopsis TCP proteins. Genes encoding three PeCYCs (PeCYC1, PeCYC2, PeCYC3), two PeCINs (PeCIN7, PeCIN8) and four PePCFs (PePCF4, PePCF5, PePCF7, PePCF10) were cloned and examined using the Y2H system followed by a spotting assay of serial dilutions with cultured Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Fig. 7. Phenotypic effects of PeDL1-silenced plants on ovule and placenta development at 16 days after pollination. (A, D, G) SEM of longitudinal sections of the ovary in mock plants. (B, E, H) SEM of longitudinal sections of the ovary in VIGS-PeDL1. (C, F, I) SEM of the longitudinal sections of the ovary in VIGS-PeDL2. Mock: an empty 35S-CymMV-Gateway vector was transformed into Agrobacterium tumefaciens (strain EHA105), and injected into the P. Sogo Yukidian ‘V3’. (A–C) Scale bar =500 um. (D–I) Scale bar = 250 µm. (J) Mean number of ovule primordia per placenta in mock, VIGSPeDL1 and VIGS-PeDL2 plants. Asterisks indicate statistically significant differences (**P<0.01 compared with mock plants by Student’s t-test). Error bars represent the SD of three biological repeats (n=10 each). ovp, ovule primordia; p, placenta; en, endocarp. Function of orchid DL/CRC genes | 5455 Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Fig. 8. Transient overexpression of PeDL1 in P. Sogo Yukidian ‘V3’. (A) Mock flower, with empty 35S-CymMV-Gateway vector. (B) OE-PeDL1-2 flower. (C) The OE-PeDL1-4 flower. (D) The OE-PeDL1-5 flower. (E) Gynostemium from mock plants (P. Sogo Yukidian ‘V3’). (F) The OE-PeDL1-2 gynostemium (strong phenotype). (G) The OE-PeDL1-4 gynostemium (moderate phenotype). (H) The OE-PeDL1-5 gynostemium (weak phenotype). SEM of the moderate and weak phenotypes on the gynostemium of (I) mock-treated and (J, K) OE-PeDL1 plants. (L) Relative expression of PeDL1 in mock and OE-PeDL1 plants. The expression of PeDL1 was determined using three replicates and was normalized using PeActin4. Asterisks indicate statistically significant differences (***P<0.001, compared with mock plants by Student’s t-test). (A–D) Scale bar =2 cm; (E–H) Scale bar =0.1 cm; (I–K) Scale bar =1 mm. Mock flower: an empty 35S-CymMV-Gateway vector was transformed into Agrobacterium tumefaciens (strain EHA105), and injected into the P. ‘Sogo Yukidian V3’. 5456 | Chen et al. Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Fig. 9. Phenotypic effects of transient overexpression PeDL1 on ovule and placenta development in the ovary before pollination and 16 days after pollination. SEM of cross sections of the ovary of mock plants (A-C) and OE-PeDL1 plants (B, D) before pollination. (B) and (D) show the shriveled placenta. SEM of (E) mock and (F) OE-PeDL1 plants 16 days after pollination. (G) Enlarged region of the white arrow in panel (E). (H) Enlarged region of the white arrow in panel (F). Compare with mock (E and G), the OE-PeDL1 (F and H) reduced development of ovule primordia. en, endocarp; ovp, ovule primordia; p, placenta. Function of orchid DL/CRC genes | 5457 Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Fig. 10. Interaction behaviors of PeDL1. (A) PeDL1 interacts with PeCIN8 in a yeast two-hybrid assay. PeDL1 and PeCIN8 genes were cloned into the binding domain vector pGBKT7(BD) and the activation domain vector pGADT7(AD). Yeast AH109 strains expressing the combination BD alone or fusion, and AD alone or fusion, were spotted on SD-Trp-Leu and SD-Trp-Leu-His (in the absence or presence of 3-amino-1,2,4-aminotriazole, 3-AT) plates and 5458 | Chen et al. Discussion An ancient orchid-specific whole-genome duplication (WGD) event is shared by all extant orchids, which might be correlated with orchid diversification (Zhang et al., 2017). Interestingly, it has been reported that DL/CRC represents a single orthologous lineage, without ancient duplications (Lee et al., 2005). Phylogenetic analysis indicates that orchid DL/CRC could be divided into two sub-clades. Sub-clade I includes primitive Apostasia DL, Phalaenopsis DL1, and Dendrobium DL1, whereas sub-clade II contains only Phalaenopsis DL2 and Dendrobium DL2 (Supplementary Fig. S1D). These results support that an ancient orchid-specific WGD event generated two DL/CRC paralogs in the last common ancestor of orchids, and one of the copies has been lost in Apostasia, one of two genera that form a sister lineage to the rest of the Orchidaceae. The Apostasia might have lost one DL gene in sub-clade II, and duplicated DL genes were retained in Epidendroideae orchids. Although PeDL1 and PeDL2 have similar expression patterns in reproductive tissues, PeDL1 per se presents much higher expression than that of PeDL2 in the gynostemium (Fig. 2; Supplementary Fig. S2). Both sequence divergence and differential expression were associated with the functional differentiation of the two DL/CRC paralogous genes in Phalaenopsis. The expression patterns of eudicot DL/CRC genes are consistent with their functions, including termination of the floral meristem, promotion of gynoecium growth, and elaboration of the abaxial carpel wall structures (Yamada et al., 2011). PsCRC in legumes is not only significantly expressed in carpels through all of the floral buds, but also in the ovary chamber, style–stigma junction, stigmatic tissues, and ovary wall (Fourquin et al., 2014). Some of the core eudicots have recruited DL/CRClike genes for nectary development (Lee et al., 2005). In rice, DL is expressed in the floral meristem, whole carpel tissue, and leaf mid-rib (Yamaguchi et al., 2004). Both PeDL transcripts detected in the Phalaenopsis floral meristem and carpel tissue are consistent with that of ancestral DL/CRC genes involved in the floral meristem determinacy and carpel specification (Yamada et al., 2011). In addition, expression of both PeDLs could be significantly measured in the placenta and the ovule primordia at early stages of ovary development. It is possible that expression of PeDLs is acquired in the ovule. In California poppy (Eschscholzia californica) the CRC ortholog EcCRC was also independently recruited for additional functions in ovule initiation (Orashakova et al., 2009). In this study, we provide evidence that overexpression of PeDL1 and PeDL2 in Arabidopsis show loss of inflorescence indeterminacy (Supplementary Fig. S6). The observed phenotypes agree with DL/CRC ancestral function in floral meristem determinacy of angiosperms. Interestingly, overexpression of AGAMOUS (AG) orthologous CeMADS genes from Cymbidium orchid in Arabidopsis also presented similar inflorescence structures (Wang et al., 2011). These results are consistent with the fact that AG can directly bind to CRC promoter sequence (Gomez-Mena et al., 2005). In Arabidopsis, expression of CRC driven by its UBQ10 promoter in crc1 mutants could restore the apical fusion of the gynoecium (Gross et al., 2018).To the best of our knowledge, complementation of an Arabidopsis CRC gene null mutant by monocot DL/CRC-like genes has not been achieved so far. Transgenic Arabidopsis crc-1 plants overexpressing individual PeDL1 and PeDL2 demonstrated that both PeDLs have conserved functions shared with CRC in regulating gynoecium development. Previous studies have shown that the California poppy (Eschscholzia californica) EcCRC gene is involved in floral meristem determinacy, carpel polarity and ovule initiation (Orashakova et al., 2009). The EcCRC-VIGS plants showed obvious defects in the carpel polarity and ovule initiation. incubated until visible colonies were formed. The yeast strain AH109 transformed with vectors pGADT7 + pGBKT7 was used as negative controls. (B) Bimolecular fluorescence complementation (BiFC) assay of protein-protein interactions of PeDL1 and PeCIN8 in transiently transformed P. aphrodite protoplasts. The yellow fluorescent protein (YFP) was split into two non-overlapping N-terminal (Yn) and C-terminal (Yc) fragments. Combinations of nonfused Empty (EYFPn) with PeDL1(EYFPc) or PeCIN8(EYFPc) were unable to reconstitute the fluorescent YFP chromophore. Combination of non-fused Empty (EYFPc) with PeDL1(EYFPn) or PeCIN8(EYFPn) were also unable to reconstitute the fluorescent YFP chromophore. The YFP fluorescence formed through the interaction between PeDL1(EYFPn) + PeCIN8(EYFPc) or PeDL1(EYFPc) + PeCIN8(EYFPn) were observed by confocal microscopy. Empty (EYFPn) + Empty (EYFPc), and PeMADS6(EYFPn) + PeMADS4 (EYFPc) were negative and positive controls, respectively. The images were obtained from the YFP fluorescent protein channel, bright field, and merged image of YFP fluorescence, and protoplasts stained with DAPI represented in blue. Scale bars=20 μm. Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 yeast cells (Supplementary Fig. S13). The results showed that PeDL1 could only interact with PeCIN8, whereas PeDL2 could not interact with the Phalaenopsis TCP proteins tested (Fig. 10A,;Supplementary Fig. S13A-F). To reconfirm the interaction relationship between PeDL1 and PeCIN8 proteins in vivo, the bimolecular fluorescence complementation (BiFC) assay was adopted. The N-terminus (Yn) of yellow fluorescent protein (YFP) was cloned upstream of PeDL1 or PeCIN8 in the pE-SPYNE vector, and the C-terminus of yellow fluorescent protein (Yc) was fused upstream of PeDL1 or PeCIN8 in the pE-SPYCE vector. The emitted EYFP (enhanced yellow fluorescent protein) signals revealed that PeDL1 and PeCIN8 could form heterodimers in the nucleus s (Fig. 10B). In our previous studies, we demonstrated that PeCIN8 plays important roles in orchid ovule development by modulating cell proliferation (Lin et al., 2016). These results suggest that the PeDL1-PeCIN8 heterodimer could function in the regulation of orchid gynostemium and ovule development. Function of orchid DL/CRC genes | 5459 LEUNIG_HOMOLOG (LUH) to specify abaxial-adaxial patterning of lateral organs (Stahle et al., 2009). In addition, it has been reported that CRC could form complexes with transcription factors, INDETERMINATE DOMAIN15, ANGUSTIFOLIA3, GROWTH-REGULATING FACT OR2, NGATHA4 or TEOSINTE BRANCHED1CYCLOIDEA-PCF15 (TCP15), involved in carpel development and auxin biosynthesis and homeostasis (Trigg et al., 2017; Lee et al., 2018). Although we observed that both PeDL1 and PeDL2 could not form homodimers or heterodimers with each other (Supplementary Fig. S13), PeDL1 could interact with type II TCP protein PeCIN8 (Fig. 10; Supplementary Fig. S13).The reason the two proteins do not interact the same way as the other DL/CRC proteins might be because sequence diversification followed by duplication of the two genes may release some constraints of the interaction behavior of the corresponding proteins. However, we cannot exclude the possibility that the transcription factor library of Arabidopsis adopted for high throughput screening of PeDL interaction targets, limited the feasibility of identification of possible candidates. Actually, the expression pattern of PeDL1 at the early stage of ovule development is parallel with that of PeCIN8 (Lin et al., 2016). It has been indicated that PeCIN8 plays important roles in orchid ovule development by modulating cell proliferation (Lin et al., 2016). It is possible that the PeDL1PeCIN8 heterodimer could have functions for the regulation of orchid gynostemium and ovule development. Further study of the downstream targets of PeDL1-PeCIN8 heterodimers may illuminate the molecular basis for unique orchid reproductive development. Supplementary data The following supplementary data are available at JXB online. Fig. S1. Sequence analysis of plant CRC/DL genes. Fig. S2. Expression patterns of PeDL1 and PeDL2 in P. aphrodite subsp. formosana. Fig. S3. RNA in situ hybridization of PeDL2 in longitudinal sections of developing floral buds and cross sections of developing ovules of P. equestris. Fig. S4. Comparison of rosette leaves in WT, 35S::PeDL1and 35S::PeDL2 transgenic Arabidopsis plants. Fig. S5. Seed phenotype of WT, 35S::PeDL1 and 35S::PeDL2 transgenic Arabidopsis plants. Fig. S6. The inflorescence phenotypes of WT, 35S::PeDL1 and 35S::PeDL2 transgenic Arabidopsis plants. Fig. S7. Detection of Cymbidium mosaic virus (CymMV) in P. Sogo Yukidian ‘V3’ VIGS plants by using RT-PCR. Fig. S8. Phenotypes of the flower and gynostemium of mock-treated, VIGS-PeDL1 and VIGS-PeDL2 plants in P. Sogo Yukidian ‘V3’. Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 In this study, PeDL1-silenced and PeDL2-silenced plants displayed a larger placenta (Fig. 7A–I). In particular, the number of differentiated protuberant ovule initials was obviously increased, in which VIGS-PeDL2 plants apparently had more severe phenotype than that of the VIGS-PeDL1 plants (Fig. 7J). Our previous research demonstrated that placental protuberances differentiated between 16 DAP to 32 DAP, which corresponds to the stage of developing ovule primordia (Chen et al., 2012). Therefore, we concluded that the two PeDL paralogs have functional redundancy in ovule initiation. The orchid gynostemium is a unique floral organ, which is fused by three carpels of female and male reproductive organs. One median carpel is initiated on the abaxial side and two lateral carpel apices protrude on the adaxial side, with the two lateral carpels partly uniting with the median carpel, forming the stigmatic cavity (Tsai et al., 2004). The stigmatic cavity provides an elaborate sticky space for accepting pollinium. In our results, transient overexpression of PeDL1 but not PeDL2 in Phalaenopsis flowers showed altered morphology with protuberances from the bottom of the stigmatic cavity, suggesting functional differentiation of the two paralogous genes. Notably, only one DL/CRC-like gene was found in the genome of primitive A. shenzhenica, which has a gynostemium without stigmatic cavity formations (Kocyan and Endress, 2001). As the ancestral function of DL/CRC plays a role in the regulation of floral meristem termination and carpel development, one of the duplicated DL/CRC members in the last common ancestor of orchids might have been co-selected as a regulator of the stigmatic cavity within the Orchidaceae, except for the species in primitive Apostasioideae. DL/CRC-like genes in monocots are also recruited for additional functions; for example, DL has an important function in the differentiation of the Poaceae leaf midrib, different than the ones observed in dicots (Yamaguchi et al., 2004). Thus, retention of duplicated DL/CRC genes and functional diversification in the sister lineage to the Apostasioideae could indicate an important role for DL/CRC genes in orchid evolution. Knowledge of the behavior of YABBY proteins originates from the study of model plants such as Arabidopsis, Antirrhinum, and rice. In Arabidopsis, YAB2, YAB3, YAB5, INO, and CRC can form homodimers (Stahle et al., 2009; Simon et al., 2017; Gross et al., 2018). Heterodimerization was also observed between different YABBY proteins (Stahle et al., 2009; Gross et al., 2018). Interestingly, YABBY proteins could also interact with co-activators as well as co-repressors to implement their functions (Liu and Meyerowitz, 1995; Stahle et al., 2009). In Antirrhinum, a heterodimer formed by GRAMINIFOLIA (GRAM) and STYLOSA (STY) was described (Navarro et al., 2004). In rice, OsYABBY4 physically interacts with SLENDER RICE 1 (SLR1) to inhibit Gibberellin-dependent degradation of SLR1 (Yang et al., 2016). Arabidopsis CRC could interact with the co-repressors LEUNIG (LUG) and 5460 | Chen et al. Acknowledgements We thank Professor Chiou-Rong Sheue (Department of Life Sciences, National Chung Hsing University) for assisting with the SEM experiment. This work was supported by the Ministry of Science and Technology, Taiwan [grants MOST 107-2313-B-006-002-MY3, MOST 108-2622-B-006-006-CC1 and MOST 110-2811-B-006 -512] and Key Laboratory of National Forestry and Grassland Administration for Orchid Conservation and Utilization Construction Funds (nos. 115/118990050; 115/KJG18016A). Author contributions W-CT and Y-YC designed the research;Y-YC and Y-YH performed the research; C-IL and H-XY conducted transcriptome sequencing and analysis; NM conducted Y2H screening and analysis; W-CT, Z-JL, Y-YC, C-MY, C-CC, C-IL, H-XY and S-BC analyzed the data; W-CT and Y-YC prepared and revised the manuscript. Conflict of interest The authors declare no conflict of interests. Data availability The sequences of PeDL2, PeFIL, and AshINO have been deposited in NCBI under GenBank accession numbers: BankIt2305555 PeDL2 MN969999, BankIt2305555 PeFIL MN970000, and BankIt2305574 AshINO MN974477. References Alvarez J, Smyth DR. 1999. CRABS CLAW and SPATULA, two Arabidopsis genes that control carpel development in parallel with AGAMOUS. Development 126, 2377–2386. Bowman JL. 2000. The YABBY gene family and abaxial cell fate. Current Opinion in Plant Biology 3, 17–22. Bowman JL, Smyth DRJD. 1999. CRABS CLAW, a gene that regulates carpel and nectary development in Arabidopsis, encodes a novel protein with zinc finger and helix-loop-helix domains. Development 126, 2387–2396. Cai J, Liu X, Vanneste K, et al. 2015. The genome sequence of the orchid Phalaenopsis equestris. Nature Genetics 47, 65–72. Chen YH, Tsai YJ, Huang JZ, Chen FC. 2005. Transcription analysis of peloric mutants of Phalaenopsis orchids derived from tissue culture. Cell Research 15, 639–657. Chen YY, Lee PF, Hsiao YY, Wu WL, Pan ZJ, Lee YI, Liu KW, Chen LJ, Liu ZJ, Tsai WC. 2012. C- and D-class MADS-box genes from Phalaenopsis equestris (Orchidaceae) display functions in gynostemium and ovule development. Plant & Cell Physiology 53, 1053–1067. Clough SJ, Bent AF. 1998. Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana. The Plant Journal 16, 735–743. Dirks-Mulder A, Ahmed I, Krol L, Menger N, Snier J, van Winzum A, de Wolf A, van’t Wout M, Zeegers J, Butôt R. 2019. Morphological and molecular characterization of orchid fruit development. Frontiers in Plant Science 10, 137. Finet C, Floyd SK, Conway SJ, Zhong B, Scutt CP, Bowman JL. 2016. Evolution of the YABBY gene family in seed plants. Evolution & Development 18, 116–126. Fourquin C, Primo A, Martínez-Fernández I, Huet-Trujillo E, Ferrándiz C. 2014. The CRC orthologue from Pisum sativum shows conserved functions in carpel morphogenesis and vascular development. Annals of Botany 114, 1535–1544. Gomez-Mena C, de Folter S, Costa MM, Angenent GC, Sablowski R. 2005. Transcriptional program controlled by the floral homeotic gene AGAMOUS during early organogenesis. Development 132, 429–438. Gross T, Broholm S, Becker A. 2018. CRABS CLAW acts as a bifunctional transcription factor in flower development. Frontiers in Plant Science 9, 835. Hsiao YY, Huang TH, Fu CH, Huang SC, Chen YJ, Huang YM, Chen WH, Tsai WC, Chen HH. 2013. Transcriptomic analysis of floral organs from Phalaenopsis orchid by using oligonucleotide microarray. Gene 518, 91–100. Hsu CC, Chen YY, Tsai WC, Chen WH, Chen HH. 2015. Three R2R3MYB transcription factors regulate distinct floral pigmentation patterning in Phalaenopsis spp. Plant Physiology 168, 175–191. Hsu H-F, Hsu W-H, Lee Y-I, Mao W-T, Yang J-Y, Li J-Y, Yang C-H. 2015. Model for perianth formation in orchids. Nature Plants 1, 15046. Kieffer M, Master V, Waites R, Davies BJTPJ. 2011. TCP14 and TCP15 affect internode length and leaf shape in Arabidopsis. The Plant Journal 68, 147–158. Kocyan A, Endress PK. 2001. Floral structure and development of Apostasia and Neuwiedia (Apostasioideae) and their relationships to other Orchidaceae. International Journal of Plant Sciences 162, 847–867. Koyama T, Ohme-Takagi M, Sato F. 2011. Generation of serrated and wavy petals by inhibition of the activity of TCP transcription factors in Arabidopsis thaliana. Plant Signaling & Behavior 6, 697–699. Lee JY, Baum SF, Oh SH, Jiang CZ, Chen JC, Bowman JL. 2005. Recruitment of CRABS CLAW to promote nectary development within the eudicot clade. Development 132, 5021–5032. Lee SJ, Lee BH, Jung JH, Park SK, Song JT, Kim JH. 2018. GROWTH-REGULATING FACTOR and GRF-INTERACTING FACTOR specify meristematic cells of gynoecia and anthers. Plant Physiology 176, 717–729. Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Fig. S9. Expression analysis of gynostemium and ovary in mock-treated, VIGS-PeDL1 and VIGS-PeDL2 plants in P. Sogo Yukidian ‘V3’. Fig. S10. Phenotypes of the gynostemium of mock-treated and OE-PeDL1 plants in P. Sogo Yukidian ‘V3’. Fig. S11. Expression analysis of gynostemium in mocktreated and transient overexpression PeDL1 plants in P. Sogo Yukidian ‘V3’. Fig. S12. Phenotypes of the ovary of mock-treated and OE-PeDL1 plants in P. Sogo Yukidian ‘V3’. Fig. S13. Interaction of PeDLs (PeDL1, PeDL2) with Phalaenopsis TCP related-proteins in yeast. Fig. S14. Interactome of PeDLs (PeDL1, PeDL2) analyzed by efficient Y2H screening using a library composed only of transcription factors in Arabidopsis. Table S1. YABBY-related genes analyzed by phylogenetic tree and protein alignment. Table S2. List of primers used in this work. Table S3. Results of Y2H assay between PeDL1 and Arabidopsis TFs. Table S4. Results of Y2H assay between PeDL2 and Arabidopsis TFs. Function of orchid DL/CRC genes | 5461 Mondragon-Palomino M, Theissen G. 2011. Conserved differential expression of paralogous DEFICIENS- and GLOBOSA-like MADS-box genes in the flowers of Orchidaceae: refining the ‘orchid code’. The Plant Journal 66, 1008–1019. Nagasawa N, Miyoshi M, Sano Y, Satoh H, Hirano H, Sakai H, Nagato Y. 2003. SUPERWOMAN1 and DROOPING LEAF genes control floral organ identity in rice. Development 130, 705–718. Navarro C, Efremova N, Golz JF, Rubiera R, Kuckenberg M, Castillo R, Tietz O, Saedler H, Schwarz-Sommer Z. 2004. Molecular and genetic interactions between STYLOSA and GRAMINIFOLIA in the control of Antirrhinum vegetative and reproductive development. Development 131, 3649–3659. O’Neill SD, Nadeau JA, Zhang XS, Bui AQ, Halevy AH. 1993. Interorgan regulation of ethylene biosynthetic genes by pollination. The Plant Cell 5, 419–432. Orashakova S, Lange M, Lange S, Wege S, Becker A. 2009. The CRABS CLAW ortholog from California poppy (Eschscholzia californica, Papaveraceae), EcCRC, is involved in floral meristem termination, gynoecium differentiation and ovule initiation. The Plant Journal 58, 682–693. Pan ZJ, Cheng CC, Tsai WC, Chung MC, Chen WH, Hu JM, Chen HH. 2011. The duplicated B-class MADS-box genes display dualistic characters in orchid floral organ identity and growth. Plant & Cell Physiology 52, 1515–1531. Pan ZJ, Chen YY, Du JS, Chen YY, Chung MC, Tsai WC, Wang CN, Chen HH. 2014. Flower development of Phalaenopsis orchid involves functionally divergent SEPALLATA-like genes. New Phytologist 202, 1024–1042. Rudall PJ, Bateman RM. 2002. Roles of synorganisation, zygomorphy and heterotopy in floral evolution: the gynostemium and labellum of orchids and other lilioid monocots. Biological Reviews 77, 403–441. Rueda-Romero P, Barrero-Sicilia C, Gómez-Cadenas A, Carbonero P, Oñate-Sánchez L. 2012. Arabidopsis thaliana DOF6 negatively affects germination in non-after-ripened seeds and interacts with TCP14. Journal of Experimental Botany 63, 1937–1949. Sawa S, Watanabe K, Goto K, Liu YG, Shibata D, Kanaya E, Morita EH, Okada K. 1999. FILAMENTOUS FLOWER, a meristem and organ identity gene of Arabidopsis, encodes a protein with a zinc finger and HMG-related domains. Genes & Development 13, 1079–1088. Siegfried KR, Eshed Y, Baum SF, Otsuga D, Drews GN, Bowman JL. 1999. Members of the YABBY gene family specify abaxial cell fate in Arabidopsis. Development 126, 4117–4128. Simon MK, Skinner DJ, Gallagher TL, Gasser CS. 2017. Integument development in Arabidopsis depends on interaction of YABBY protein INNER NO OUTER with coactivators and corepressors. Genetics 207, 1489–1500. Stahle MI, Kuehlich J, Staron L, von Arnim AG, Golz JF. 2009. YABBYs and the transcriptional corepressors LEUNIG and LEUNIG_HOMOLOG maintain leaf polarity and meristem activity in Arabidopsis. The Plant Cell 21, 3105–3118. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. 2013. MEGA6: molecular evolutionary genetics analysis version 6.0. Molecular Biology and Evolution 30, 2725–2729. Tatematsu K, Nakabayashi K, Kamiya Y, Nambara EJTPJ. 2008. Transcription factor AtTCP14 regulates embryonic growth potential during seed germination in Arabidopsis thaliana. The Plant Journal 53, 42–52. Theissen G, Becker A, Di Rosa A, Kanno A, Kim JT, Munster T, Winter KU, Saedler H. 2000. A short history of MADS-box genes in plants. Plant Molecular Biology 42, 115-149. Trigg SA, Garza RM, MacWilliams A, et al. 2017. CrY2H-seq: a massively multiplexed assay for deep-coverage interactome mapping. Nature Methods 14, 819–825. Tsai WC, Hsiao YY, Pan ZJ, Kuoh CS, Chen WH, Chen HH. 2008a. The role of ethylene in orchid ovule development. Plant Science 175, 98–105. Tsai WC, Kuoh CS, Chuang MH, Chen WH, Chen HH. 2004. Four DEF-like MADS box genes displayed distinct floral morphogenetic roles in Phalaenopsis orchid. Plant & Cell Physiology 45, 831–844. Tsai WC, Lee PF, Chen HI, Hsiao YY, Wei WJ, Pan ZJ, Chuang MH, Kuoh CS, Chen WH, Chen HH. 2005. PeMADS6, a GLOBOSA/ PISTILLATA-like gene in Phalaenopsis equestris involved in petaloid formation, and correlated with flower longevity and ovary development. Plant & Cell Physiology 46, 1125–1139. Tsai WC, Pan ZJ, Hsiao YY, Jeng MF, Wu TF, Chen WH, Chen HH. 2008b. Interactions of B-class complex proteins involved in tepal development in Phalaenopsis orchid. Plant & Cell Physiology 49, 814–824. Tsai WC, Pan ZJ, Su YY, Liu ZJ. 2014. New insight into the regulation of floral morphogenesis. International Review of Cell and Molecular Biology 311, 157–182. Uberti-Manassero NG, Lucero LE, Viola IL, Vegetti AC, Gonzalez DH. 2012. The class I protein AtTCP15 modulates plant development through a pathway that overlaps with the one affected by CIN-like TCP proteins. Journal of Experimental Botany 63, 809–823. Villanueva JM, Broadhvest J, Hauser BA, Meister RJ, Schneitz K, Gasser CS. 1999. INNER NO OUTER regulates abaxial- adaxial patterning in Arabidopsis ovules. Genes & Development 13, 3160–3169. Wang SY, Lee PF, Lee YI, Hsiao YY, Chen YY, Pan ZJ, Liu ZJ, Tsai WC. 2011. Duplicated C-class MADS-box genes reveal distinct roles in gynostemium development in Cymbidium ensifolium (Orchidaceae). Plant & Cell Physiology 52, 563–577. Yamada T, Yokota S, Hirayama Y, Imaichi R, Kato M, Gasser CS. 2011. Ancestral expression patterns and evolutionary diversification of YABBY genes in angiosperms. The Plant Journal 67, 26–36. Yamaguchi T, Nagasawa N, Kawasaki S, Matsuoka M, Nagato Y, Hirano HY. 2004. The YABBY gene DROOPING LEAF regulates carpel specification and midrib development in Oryza sativa. The Plant Cell 16, 500–509. Yang C, Ma Y, Li J. 2016. The rice YABBY4 gene regulates plant growth and development through modulating the gibberellin pathway. Journal of Experimental Botany 67, 5545–5556. Zhang GQ, Liu KW, Li Z, et al. 2017. The Apostasia genome and the evolution of orchids. Nature 549, 379–383. Zhang GQ, Xu Q, Bian C, et al. 2016. The Dendrobium catenatum Lindl. genome sequence provides insights into polysaccharide synthase, floral development and adaptive evolution. Scientific Reports 6, 19029. Downloaded from https://academic.oup.com/jxb/article/72/15/5442/6272287 by guest on 03 November 2022 Li H, Liang W, Yin C, Zhu L, Zhang D. 2011. Genetic interaction of OsMADS3, DROOPING LEAF, and OsMADS13 in specifying rice floral organ identities and meristem determinacy. Plant Physiology 156, 263–274. Lin HY, Chen JC, Fang SC. 2018. A protoplast transient expression system to enable molecular, cellular, and functional studies in Phalaenopsis orchids. Frontiers in Plant Science 9, 843. Lin YF, Chen YY, Hsiao YY, Shen CY, Hsu JL, Yeh CM, Mitsuda N, Ohme-Takagi M, Liu ZJ, Tsai WC. 2016. Genome-wide identification and characterization of TCP genes involved in ovule development of Phalaenopsis equestris. Journal of Experimental Botany 67, 5051–5066. Liu Z, Meyerowitz EM. 1995. LEUNIG regulates AGAMOUS expression in Arabidopsis flowers. Development 121, 975–991. Lu HC, Chen HH, Tsai WC, Chen WH, Su HJ, Chang DC, Yeh HH. 2007. Strategies for functional validation of genes involved in reproductive stages of orchids. Plant Physiology 143, 558–569. Mitsuda N, Ikeda M, Takada S, Takiguchi Y, Kondou Y, Yoshizumi T, Fujita M, Shinozaki K, Matsui M, Ohme-Takagi M. 2010. Efficient yeast one-/two-hybrid screening using a library composed only of transcription factors in Arabidopsis thaliana. Plant & Cell Physiology 51, 2145–2151.