Dendrobium Chao Praya Smile (a hybrid of Dendrobium Pinky and Dendrobium Kiyomi Beauty) plants were grown in pots in the greenhouse at 28 ± 4°C under natural lighting conditions. In our in vitro culture system for Dendrobium Chao Praya Smile, protocorms developing from seeds were used as the starting materials. They were cultured in modified liquid Knudson C (KC) medium supplemented with 2% (w/v) sucrose, 15% (v/v) coconut water and 4.4 μM benzyladenine (BA) at 24°C under a 16-h photoperiod on rotary shakers at 120 rpm (Yu and Goh, 2000a; Ding et al., 2013). Plantlets at the floral transitional stage were transferred to the two-layer modified KC medium as previously described (Hee et al., 2007; Sim et al., 2007). Wild-type Arabidopsis thaliana ecotype Columbia (Col-0) and 35S:DOAP1 transgenic plants in both wild-type and ap1-10 Col-0 backgrounds were grown under long-day conditions (16 h light/8 h dark) at 23 ± 2°C.

Total RNA was extracted from inflorescences of Dendrobium Chao Praya Smile using the RNeasy^® Plant Mini Kit (QIAGEN). To isolate the putative AP1 orthologs, two degenerate primers, 5′-CAGCTGARGCGRATMGAGAAC-3′ and 5′-GCKMAGCATCCAWGGYGG-3′, were manually designed based on the conserved regions among AP1 orthologs in various plant species. Based on the partial cDNA sequence obtained, the full-length cDNA of DOAP1 was further amplified using the SMARTER^TM RACE cDNA Amplification Kit (BD Biosciences Clontech).

Protein sequences of AP1-like genes were retrieved from the National Center for Biotechnology Information (NCBI) database. Alignment of amino acid sequences was performed using the Clustal Omega multiple sequence alignment program¹ and BOXSHADE 3.21². The phylogenetic tree was generated by the neighbor-joining algorithm using the MEGA6 software³.

Total RNA was extracted from orchids or Arabidopsis plants using the RNeasy^® Plant Mini Kit (QIAGEN), and reverse-transcribed using the SuperScript^TM II Reverse Transcriptase Kit (Invitrogen) according to the manufacturer’s instructions. Real-time PCR reaction was performed in triplicates on the CFX384 Real-Time PCR Detection System (Bio-Rad) with the SYBR Green Master Mix (Toyobo). UBIQUITIN (DOUbi) in Dendrobium Chao Praya Smile was used as a reference gene (Ding et al., 2013). Calculation of relative gene expression levels was performed as previously reported (Liu et al., 2007). Semi-quantitative reverse transcription PCR (RT-PCR) was carried out as previously reported (Xu et al., 2006) using either Arabidopsis β-TUBLIN2 (TUB2) or orchid ACTIN (DOActin) as a reference gene.

The full-length DOAP1 cDNA fragment was cloned into pGreen0229-35S under the control of two CaMV 35S promoters (Yu et al., 2004). The resulting construct was introduced into Agrobacterium tumefaciens GV3101 competent cells by electroporation. 35S:DOAP1 transgenic Arabidopsis lines were created by Agrobacterium tumefaciens-mediated transformation and selected by 0.3 g/L Basta on soil.

Genetic transformation of Dendrobium Chao Praya Smile was performed according to the reported L-methionine sulfoximine (MSO) selection system with minor modifications (Yu et al., 2001; Chai et al., 2007; Ding et al., 2013). Dendrobium Chao Praya Smile calli were cut into small pieces 3–5 mm in diameter and then cultured in modified KC liquid medium supplemented with 2% (w/v) sucrose, 15% (v/v) coconut water and 4.4 μM BA. Agrobacterium pellet was re-suspended in KC liquid medium to be co-cultivated with prepared orchid calli. Acetosyringone (100 μM) was added to induce transformation. After co-cultivation for 2 h, calli were placed on solid KC medium in the dark for three nights. Agrobacterium-infected calli were rinsed by water containing 200 mg/L cefotaxime to remove the bacteria, and grown on solid KC medium containing 0.5 μM MSO as a selection agent. Surviving calli were subcultured onto fresh solid medium every 20 days. After four rounds of MSO selection, green calli were transferred to modified KC solid medium containing 2 μM MSO for lethal selection. After three rounds of selection, putative surviving transgenic lines were cultured on modified KC solid medium for further investigation.

Genomic DNA isolated from wild-type and 35S:DOAP1 transgenic orchids was digested by EcoRI for 16 h, resolved on 1.2% (w/v) agarose gels, and then transferred onto nylon membranes. Blots were hybridized overnight with the specific digoxigenin-labeled DNA, washed and detected in accordance with the manufacturer’s instructions (DIG Application Manual for Filter Hybridization, Roche) as previously reported (Yu and Goh, 2000b).

To isolate AP1-like genes from Dendrobium Chao Praya Smile, we designed a pair of degenerate primers based on the conserved amino acid sequences of AP1 orthologs in various plant species. A partial cDNA fragment was obtained by reverse transcription PCR using RNA extracted from inflorescence apices of Dendrobium Chao Praya Smile. Since this fragment showed high sequence similarity with other AP1-like genes, we further designed primers based on this fragment and obtained the corresponding full-length cDNA sequence, designated DOAP1 (GenBank accession No. KY471451), using the rapid amplification of cDNA ends (RACE) method.

DOAP1 cDNA is 897 bp in length with a 639 bp coding region. The deduced DOAP1 amino acid sequence contains a highly conserved MADS domain and a less conserved K domain, and a diverse C terminal region as found in other AP1-like genes from various plant species (Figure 1). Multiple sequence alignment showed that DOAP1 shared the highest sequence identity with other orchid AP1 orthologs, such as CeAP1 (Cymbidium ensifolium; 69.95% identity) and PjAP1 (Phalaenopsis japonica; 62.56% identity). DOAP1 also had higher sequence identity with monocot AP1 orthologs, such as TaAP1 (Triticum aestivum; 58% identity), OsAP1 (Oryza sativa; 56.72% identity) and ZmAP1 (Zea mays; 54.59% identity), than Arabidopsis AP1 (53.3% identity).

To determine the evolutionary relationship between DOAP1 and other AP1-like proteins from other plant species, we constructed a phylogenetic tree based on the analysis of amino acid sequences of MIK regions (Figure 2). The tree showed that DOAP1 was clustered together with other orchid AP1-like proteins, such as OMADS10 (Chang et al., 2009) and DOMADS2 (Yu and Goh, 2000b), in the monocotyledonous subgroup of SQUA that includes AP1-like proteins isolated from monocots (Purugganan et al., 1995; Theissen et al., 1996).

To study the expression pattern of DOAP1 in Dendrobium Chao Praya Smile, cDNA was synthesized from RNA extracted from various orchid tissues, including leaves, roots, inflorescence apices, floral buds at different stages (stages 1–3, at which buds were 0.5–1 cm, 1–2 cm, and over 2 cm in length, respectively), half-bloomed flowers and fully bloomed flowers (Figure 3A). Quantitative real-time PCR analysis showed that DOAP1 transcripts were detected at the highest level in inflorescence apices and at the lowest level in roots (Figure 3B). Notably, DOAP1 expression was significantly higher in inflorescence apices than in vegetative shoot apices. Moreover, DOAP1 expression gradually increased in developing floral buds at various stages and half-bloomed flowers, but decreased afterward in fully bloomed flowers (Figure 3B). These expression patterns indicate that DOAP1 function may be closely associated with flowering time control and flower development in Dendrobium Chao Praya Smile.

In order to investigate the biological function of DOAP1, we firstly generated transgenic Arabidopsis plants harboring the DOAP1 coding sequence driven by the CaMV 35S promoter. Among 50 independent 35S:DOAP1 transgenic lines created at the T1 generation, 46 lines showed an early flowering phenotype typically producing only 5–7 rosette leaves as compared to wild-type plants, which developed 9–12 rosette leaves under long day conditions (Figures 4A,B). Semi-quantitative RT-PCR using DOAP1-specific primers confirmed that DOAP1 was overexpressed in a representative 35S:DOAP1 transgenic Arabidopsis line with a single T-DNA insertion (Figure 4C), implying that the early flowering phenotype of 35S:DOAP1 is associated with overexpression of DOAP1. In addition to the defect in flowering time, the inflorescence meristems of 35S:DOAP1 transgenic lines usually terminated as flowers after producing only a few flowers, while these flowers produced floral organs indistinguishable from those of wild-type flowers (Figures 4D,E), suggesting that overexpression of DOAP1 promotes the formation of floral meristems, but does not affect floral organ identity in Arabidopsis.

We further performed Agrobacterium-mediated transformation to introduce 35S:DOAP1 into ap1-10 mutants, which were generated from ethyl methanesulfonate-mutagenized populations of Arabidopsis ecotype Columbia (Schultz and Haughn, 1993), to examine whether DOAP1 could complement the loss function of AP1. While ap1-10 exhibited comparable flowering time to wild-type plants, most of the ap1-10 35S:DOAP1 lines flowered much earlier than wild-type and ap1-10 plants, a flowering pattern resembling 35S:DOAP1 in the wild-type background (Figure 4A). We further identified several ap1-10 35S:DOAP1 transgenic lines with single T-DNA insertion, which consistently displayed the similar early flowering phenotype at various generations, and selected one representative line for further investigations. Like other ap1 mutants, ap1-10 showed typical defects in the identity of perianth organs, such as homeotic transformation of petals into stamens or stamen-petal chimeric structures (Figure 4F). In contrast, petal formation was restored in ap1-10 35S:DOAP1 (Figure 4G), demonstrating that DOAP1 plays the same role as AP1 in regulating petal formation in Arabidopsis. Similarly, semi-quantitative RT-PCR detected high expression of DOAP1 in the representative ap1-10 35S:DOAP1 line (Figure 4C), substantiating a causal link between the observable phenotypes and overexpression of DOAP1 in ap1-10 35S:DOAP1.

To understand the endogenous function of DOAP1 in Dendrobium Chao Praya Smile, we created transgenic orchids bearing 35S:DOAP1 in a pGreen vector (Figure 5A) using an integrated orchid gene transformation and in vitro orchid culture system (Yu et al., 2001; Chai et al., 2007; Ding et al., 2013).

After Agrobacterium-mediated transformation of Dendrobium Chao Praya Smile calli, the transformed materials were selected on the medium containing 0.5 and 2 μM MSO as a selection agent for initial and lethal selection (Figure 5B), respectively. After a total of seven rounds of selection, most of the calli turned necrotic and eventually died, while some putative transformants survived and proliferated into protocorm-like bodies, which further developed into young plantlets. We screened out nine independent 35S:DOAP1 transgenic Dendrobium Chao Praya Smile lines, and confirmed the presence of the 35S:DOAP1 transgene in these lines by PCR genotyping using the specific primers from 35S and DOAP1 (Figure 5C).

We then compared the growth status of wild-type and 35S:DOAP1 Dendrobium Chao Praya Smile plants using our established in vitro culture system (Ding et al., 2013), which allows rapid in vitro development of orchid plants from the vegetative to reproductive phase. Under our growth conditions, it took wild-type orchid plants 25–30 weeks to grow from protocorm-like bodies to plantlets with the first visible inflorescence stalks (Figure 6A), whereas 8 out of 9 35S:DOAP1 transgenic lines displayed the first visible inflorescences at 10–17 weeks of culture (Figures 6A,B). These phenotypes indicate a role of DOAP1 in promoting the transition of vegetative shoot apical meristems into inflorescence meristems. As compared to wild-type inflorescences (Ding et al., 2013), the inflorescence apices of these 35S:DOAP1 transgenic plants were usually terminated as floral buds immediately after producing only 1–2 floral buds or even without generating any other floral structure (Figure 6C), suggesting a quick transformation of inflorescence meristems into floral meristems in 35S:DOAP1. In addition, Southern blot hybridization revealed the presence of the T-DNA region containing the bialaphos resistance (bar) gene in the genome of 35S:DOAP1 #1 and #2 lines (Figure 6D), which substantiated the results showing the presence of the 35S:DOAP1 transgene in the 35S:DOAP1 genome (Figure 5C). As expected, semi-quantitative RT PCR showed that DOAP1 was overexpressed in inflorescence apices of 35S:DOAP1 transgenic orchids compared to wild-type orchids (Figure 6E). These results imply that the phenotypes of 35S:DOAP1 transgenic orchids are associated with overexpression of DOAP1 in these plants.