See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/276162457

Model for perianth formation in orchids

Article in Nature · April 2015 DOI: 10.1038/nplants.2015.46

CITATIONS READS 82 1,993

7 authors, including:

Yung-I Lee Wan-Ting Mao National Museum of Natural Science National Chung Hsing University

81 PUBLICATIONS 1,297 CITATIONS 4 PUBLICATIONS 94 CITATIONS

SEE PROFILE SEE PROFILE

Jun-Yi Yang National Chung Hsing University

60 PUBLICATIONS 1,905 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

I am working on orchid embryology, cytogenetics, mycorrhizal association, usingCypripedium, Paphiopedilum and as the experimental materials. View project

All content following this page was uploaded by Yung-I Lee on 07 January 2016.

The user has requested enhancement of the downloaded file. LETTERS PUBLISHED: 27 APRIL 2015 | ARTICLE NUMBER: 15046 | DOI: 10.1038/NPLANTS.2015.46

Model for perianth formation in orchids

Hsing-Fun Hsu1, Wei-Han Hsu1,Yung-ILee2,3, Wan-Ting Mao1, Jun-Yi Yang4, Jen-Ying Li1 and Chang-Hsien Yang1,5*

Orchidaceae, the orchid under the order , protein complexes involved in perianth formation and patterning. contains more than 20,000 accepted in approximately We comprehensively examined the expression of all identified A 880 genera1–3. In contrast to most flowers of actinomorphic (AP1-like), B (AP3-like and PI-like) and E (AGL6-like and SEP1/ symmetry, orchid flowers typically have zygomorphic sym- 3-like) OMADS box genes in mature flowers of O. Gower Ramsey metry with a striking well-differentiated labellum (lip) that (Fig. 1a-1, left), its parent species O. flexuosum (Supplementary acts as the main pollinator attractant by employing visual, Fig. 2a-1) and the closely related horticultural cultivars O. Sweet fragrance and tactile cues4–7. Genetics models controlling pat- Sugar (Supplementary Fig. 2a-2) and O. Lemon Heart terning formation of actinomorphic flowers, such as (Supplementary Fig. 3a-1, left). As shown in Fig. 1b-1 and Arabidopsis, are well known. However, the mechanisms of Supplementary Figs 2b-1, b-2 and 3b-1, OMADS8 (OPI) was //lip determination remain obscure. Here, we found to be universally expressed in all perianth organs. demonstrate a conserved principle, called the Perianth (P) OMADS5 (OAP3-1) and OMADS7 (OAGL6-1) were highly code, which involves competition between two protein com- expressed in / but undetectable in lips. OMADS9 plexes containing different AP3/AGL6 homologues to deter- (OAP3-2) was detected at high levels in lips and petals, and mine the formation of the complex perianth patterns in OMADS1 (OAGL6-2) was only expressed in lips, but its expression orchids. In the P code, the higher-order heterotetrameric SP was relatively low in lateral sepals. Further analysis indicated that (sepal/petal) complex (OAP3-1/OAGL6-1/OAGL6-1/OPI) the expression patterns of OPI, OAP3-1, OAGL6-1, OAP3-2 and specifies sepal/petal formation, whereas the L (lip) complex OAGL6-2 were very similar and highly conserved in the perianths (OAP3-2/OAGL6-2/OAGL6-2/OPI) is exclusively required of both young flower buds (2, 3, 5 and 8 mm) and mature flowers for lip formation. This model is validated by the conversion of of O. Gower Ramsey (Supplementary Fig. 4a–c). Other B, A and E lips into sepal/petal structures in and Phalaenopsis genes, such as OMADS3 (OAP3-3), OMADS6/11 (E, OSEP3/1) and orchids through the suppression of the proposed L complex OMADS10 (A, OAP1)22, were expressed in all perianths, without a activity in lips using the virus-induced gene silencing (VIGS) notable pattern in the sepal/petal/lips (Supplementary Figs 2b-1, strategy. A comprehensive examination of four different subfa- b-2 and 3e-1). This result suggested that these A and E OMADS milies of further validates the P code and signifi- box genes should involve in the general formation of perianth cantly extends the current knowledge regarding the organs. However, they do not participate further in the specification mechanism and pathways of perianth formation in orchids. of the lips and sepal/petal in orchid. In contrast, OMADS12 (OAP3-4) According to the ABCDE model, which predicts the formation of was only weakly expressed in the but was expressed in any flower organ through the interaction of five classes of homeotic the carpel, although its expression was completely absent in the genes in plants, the B class genes APETALA3 (AP3) and sepal/petal/lips (Supplementary Fig. 4d). The highly conserved PISTILLATA (PI) are of particular importance in petal for- expression pattern for the A, B and E OMADS homologues was mation8–11. The AP3 orthologues identified in lower eudicots, mag- also observed in the perianth organs of four Oncidium botanical nolid dicots and monocots make up the paleoAP3 lineage, whereas species: O. gutfreundianum, O. tipuloides, O. cheirophorum and the euAP3 lineage is composed of the AP3 orthologues identified in O.cebolleta (Supplementary Fig. 2a-3 to a-6, b-3 to b-6). dicots9. More than two duplicated AP3-like genes have been ident- On the basis of the gene expression patterns among these ified in many orchid species12–17 (Supplementary Fig. 1b). However, Oncidium species, we have proposed a Perianth code (P code) hypoth- as most studies have mainly focused on the expression and function esis for the regulation of identity of perianth organs in Oncidium, of individual B or E class genes18–21, it is difficult to reach a solid which involves the interaction of OPI with two different AP3/AGL6 conclusion on perianth determination in orchids. We have pre- homologues to form two protein complexes, named SP (sepal/petal viously found that expression variation in clade 1 (OMADS5, complex) and L (lip complex) (Fig. 1c). In this P-code model, OPI renamed as OAP3-1 in this study) and clade 3 (OMADS9, is expressed ubiquitously in perianths. OAP3-1 and OAGL6-1 form renamed as OAP3-2) AP3-like genes potentially affects labellum the determinant unit of the SP complex, promoting sepal/petal devel- determination in Oncidium Gower Ramsey16 (Supplementary opment, whereas OAP3-2 and OAGL6-2 function as the determinant Fig. 1b). Additionally, an AGL6 homologue of Oncidium, unit of the L complex to promote lip development (Fig. 1c).Inthis OMADS1 (renamed as OAGL6-2) (Supplementary Fig. 1a), shows model, the SP and L complexes oppositely regulate perianth develop- a synchronized increase in gene expression accompanying perianth ment (Fig. 1c).LossofeitherOAP3-1 or OAGL6-1 in the SP complex transition towards lip-like structures16, suggesting that perianth andhighexpressionoftheintactLunit(OAP3-2/OAGL6-2)will organ identity may be determined by a complicated complex. contribute to lip development (Fig. 1d-1). Conversely, loss of either We first used Oncidium orchids (subtribe , tribe OAP3-2 or OAGL6-2 from the L complex and high expression in subfamily ) to further identify levels of the intact SP unit (OAP3-1/OAGL6-1)willcontributeto

1Institute of Biotechnology, National Chung Hsing University, Taichung, Taiwan 40227 ROC. 2Biology Department, National Museum of Natural Science, Taichung, Taiwan 40227 ROC. 3Department of Life Sciences, National Chung Hsing University, Taichung, Taiwan 40227 ROC. 4Institute of Biochemistry, National Chung Hsing University, Taichung, Taiwan 40227 ROC. 5Agricultural Biotechnology Center, National Chung Hsing University, Taichung, Taiwan 40227 ROC. *e-mail: [email protected]

NATURE PLANTS | VOL 1 | MAY 2015 | www.nature.com/natureplants 1 LETTERS NATURE PLANTS DOI: 10.1038/NPLANTS.2015.46

a 1 DS P 2 3 P 4 DS 5 DS 6 DS P P P P P P P P LS LS Lip Lip Lip Lip LS Lip Lip GR GRtrip GRpl Amp Macro Pp Oc Op

b 123456 4.0 PI OPI 1.4 OPI 1.6 OPI 5.0 OPI 1.6 OPI OPI 1.2 4.0 3.0 1.4 4.0 1.4 1.0 1.2 1.2 3.0 2.0 0.8 1.0 3.0 1.0 0.6 0.8 0.8 2.0 0.6 2.0 0.6 1.0 0.4 0.4 0.4 1.0 1.0 0.2 0.2 0.2 0.0 0.0 0.0 0.0 0.0 0.0

1.2 OAP3-1 OAP3-1 OAP3-1 OAP3-1 OAP3-1 OAP3-1 1.2 2.5 2.5 1.2 1.0 * 4.0 1.0 2.0 2.0 1.0 0.8 0.8 * 3.0 1.5 0.8 0.6 1.5 0.6 * 0.6 2.0 1.0 1.0 0.4 0.4 0.4 * * 1.0 0.2 0.2 0.5 0.5 0.2 0.0 0.0 0.0 0.0 0.0 0.0

SP 1.8 1.6 OL6-1 OL6-1 OL6-1 OL6-1 OL6-1 OL6-1 2.5 1.6 1.8 2.5 5.0 1.4 1.6 1.4 2.0 1.2 1.4 2.0 1.2 4.0 1.0 1.2 1.0 * 1.5 1.5 3.0 0.8 1.0 0.8 0.6 0.8 1.0 1.0 0.6 0.6 * 2.0 0.4 * * 0.4 0.4 * 0.5 0.5 * 1.0 0.2 0.2 0.2 0.0 0.0 0.0 0.0 0.0 0.0

2.0 OAP3-2 OAP3-2 OAP3-2 OAP3-2 OAP3-2 OAP3-2 1.8 2.0 * * 1.4 1.2 1.2 1.6 * 1.8 * 1.2 3.5 * 1.0 1.0 * 1.6 * 1.4 * 3.0 1.2 1.0 1.4 * 0.8 0.8 1.2 2.5 1.0 0.8 * 0.6 0.6 1.0 2.0 0.8 0.6 0.8 1.5 0.6 0.4 0.4 0.4 0.6 0.4 0.4 1.0 0.2 0.2 0.2 0.2 0.2 0.5 0.0 0.0 0.0 0.0 0.0 0.0

L 1.2 OL6-2 OL6-2 OL6-2 OL6-2 OL6-2 OL6-2 * 1.0 1.4 1.2 1.4 1.8 3.5 1.2 1.0 * 1.2 1.6 * * 0.8 * * * 1.4 3.0 1.0 1.0 0.8 1.2 2.5 0.6 0.8 0.8 * 0.6 1.0 2.0 0.4 0.6 0.6 0.8 1.5 0.4 0.6 0.4 0.4 1.0 0.2 * 0.2 * 0.4 0.2 0.2 0.2 0.5 0.0 0.0 0.0 0.0 0.0 0.0 Lip P DS LS Lip P DS LS Lip P DS LS Lip P DS LS Lip P DS LS Lip P DS LS Lip P DS LS Lip P DS LS GR GRtrip GRpl Amp Macro Pp Oc Op

c d

L AP3 SP Sepal/petal L6 Lip -2 -1 L6 AP3 -2 -1 1 OPI 2 OPI Lip Sepal/petal

AP3-2/L6-2 AP3-1/L6-1 L SP L AP3 L6 AP3 L6 SP -2 -2 -15 -17 OPI PI 3 OPI 4 OPI Intermediate

Figure 1 | OMADS box gene expression profiles in Oncidium reveal the P-code model. a,Theflowers of wild-type O. Gower Ramsey (GR) and the corresponding peloric mutant with two petals transformed into lips (GRtrip and GRpl) (a-1), O. ampliatum (Amp) (a-2), O. macropetalum (Macro) (a-3), papilio (Pp) (a-4), O. carthaginense (Oc) (a-5)andOdm. pulchellum (Op) (a-6). Scale bars, 10 mm. b, Total RNA samples isolated from the lips (Lip), petals (P), dorsal sepals (DS) and lateral sepals (LS) of mature orchid flowers (a-1 to a-6) were used as templates to detect the corresponding (b-1 to b-6) expression of OPI, OAP3-1/OAGL6-1 (SP) and OAP3-2/OAGL6-2 (L) by quantitative real-time PCR. To help visualize the different combination of SP and L units in the lips and lip-like petals/sepals (b-1 to b-4) and in the sepal/petal-like lips (b-5, b-6), the detected expression levels of only OAP3-1/OAGL6-1 (SP) and OAP3-2/OAGL6-2 (L) are marked with blue and red stars, respectively. c, In the P-code model, OPI serves as a base for the perianth identity. OAP3-1/OAGL6-1 (AP3-1/L6-1) is grouped in the SP complex, promoting sepal/petal identity. These genes antagonize the lip identity promoted by the L complex OAP3-2/OAGL6-2 (AP3-2/L6-2) in O.GowerRamsey.d,Inthe‘P-code’ balance, the presence of L complex (AP3-2/L6-2) in red only, resulting in lip formation (d-1), and the presence of SP complex (AP3-1/L6-1) in blue only, resulting in sepal/petal formation (d-2). The coexistence (d-3) or co-absence (d-4) of the SP and L complexes tipped the P-code balance and produced sepal/petal and lip intermediate structures. sepal/petal development (Fig. 1d-2). To explore whether the P-code in O. Gower Ramsey. Very similar patterns for OAP3-1, OAP3-2 hypothesis is also applicable to stamen and carpel formation, the and OPI and for OAGL6-1 and OAGL6-2 were observed in expression pattern of the SP (OAP3-1 and OAGL6-1)andL(OAP3-2 and carpels (Supplementary Fig. 4b,c), confirming that the P-code and OAGL6-2) components in stamens and carpels was analysed hypothesis is specifically applicable for perianth formation only.

2 NATURE PLANTS |VOL1|MAY2015|www.nature.com/natureplants NATURE PLANTS DOI: 10.1038/NPLANTS.2015.46 LETTERS

a 1234567891011 CFP channel

YFP channel

Raw FRET

FRET efficiency

70

60

50

40

30

FRET efficiency (%) 20

10

0 CFP Cy OAP3-1 OAP3-1 OPI OAP3-1 OAGL6-1 OAP3-2 OAGL6-2 OAGL6-2 OAP3-2 OAGL6-2

YFP Y OAGL6-1 OAGL6-1 OAGL6-1 OPI OAGL6-1 OAGL6-2 OAP3-2 OPI OPI OAGL6-2

OPI OAP3-1 OPI OAP3-2

b 2 L SP OAP3-2 OPI OAP3-1 OPI (AP3) (PI) (AP3) (PI) OL6-2 OL6-2 OL6-1 OL6-1 (L6) (L6) (L6) (L6)

3

OL6 -2 OL6 -2 1 (L6) (L6) SP unit OAP3-1 (AP3) OL6 -1 OL6-1 OL6-1 N (L6) (L6) (L6) Lip program OAP3-1 OPI (AP3) (PI) L unit OAP3-2 (AP3) OAP3-2 OPI OL6-2 (AP3) (PI) (L6)

Sepal/petal Lip

Figure 2 | Detection of interaction among OMADS proteins. a, Analysis of interaction among OPI, SP (OAP3-1/OAGL6-1) and L (OAP3-2/OAGL6-2) MADS proteins using the FRET technique. In the top image, four channels of FRET imaging are shown. The CFP and YFP channels were excited with 440 nm and 515 nm lasers respectively, and these two channels were used to calculate the raw FRET signal. Finally, raw FRET values were divided by CFP signals to calculate FRET efficiency. The average FRET efficiency values (bottom) were quantified in multiple samples (n > 4). Empty CyPet and YPet protein pairs were used as FRET signal controls. Image frame, 20 × 20 µm2. The error bars represent the standard deviation of the values. b, Summary of protein interactions and efficiency by which the proteins enter the nucleus (N) as described in a. The OAP3-1/OAGL6-1 (SP unit) and OAP3-2/OAGL6-2 (L unit) heterodimers did not enter the nucleus (indicated as dashed lines) (b-1), whereas other combination of protein complexes did enter the nucleus (indicated as solid lines) (b-2, b-3). The L quartet (OAP3-2/OAGL6-2/OAGL6-2/OPI) switches on the ‘Lip program’ to promote lip development. In contrast, the SP quartet (OAP3-1/ OAGL6-1/OAGL6-1/OPI) switches off the ‘Lip program’ to promote petal/sepal development.CFP; cyan fluorescent protein, YFP; yellow fluorescent protein. To validate the P-code model, the expression of the proposed SP middle), OAP3-2/OAGL6-2 established an L complex, whereas the and L complex genes in the Oncidium peloric mutant was profiled. SP complex was dissociated due to a loss of OAP3-1 expression In the lip-like petals of the peloric mutant GRtrip (Fig. 1a-1, (Fig. 1b-1). The formation of the L complex alone in the petals of

NATURE PLANTS | VOL 1 | MAY 2015 | www.nature.com/natureplants 3 LETTERS NATURE PLANTS DOI: 10.1038/NPLANTS.2015.46

a Adaxial Abaxial bfAdaxial Abaxial 1 3 DS DS DS DS DS P P P P P

LS LS LS LS Mock LP LP LP LS Pe9-VIGS-1 Pe9-VIGS-2 LP Mock g LP-s Mock 2 4 LP-s DS DS LP-s P P

LS

LP-f LP LP-f LP-f LP Mock OAGL6-2 VIGS OAGL6-2 VIGS Pe9-VIGS-1 Pe9-VIGS-2

d Lip Petal Sepal c Mock OL6-2-VIGS i LP-f LP-s PS 1 3 25.0 1.2 OAGL6-2 PeM9 (OL6-2) 1.0 20.0 0.8 Lip 15.0 0.6 10.0 0.4 * 0.2 5.0 * 2 4 * * * 0.0 0.0 L 2.0 7.0 OAP3-2 Petal PeM3 (OAP3-2) 1.6 6.0 5.0 1.2 * 4.0 * 0.8 3.0 * e 1 2 2.0 0.4 Sepal Petal Lip Sepal Petal Lip * 1.0 * 0.0 SP SP SP SP SP 0.0 L L 6.0 OAGL6-1 1.4 * PeM10 (OL6-1) Mock Pe9-VIGS 5.0 h 1.2 * 1 3 1.0 4.0 0.8 3.0 0.6 * Lip-f * 2.0 0.4 1.0 * 0.2 0.0 0.0 SP 1.8 1.4 OAP3-1 2 4 PeM2 (OAP3-1) 1.6 * 1.2 1.4 1.0 1.2 * 0.8 Petal 1.0 0.6 0.8 * * 0.6 0.4 0.4 0.2 * 0.2 0.0 1 2 0.0 j Sepal Petal Lip Sepal Petal Lip Mock Mock Mock Mock

Mock Mock Mock SP SP SP SP SP SP Pe9-V-1 Pe9-V-1 L Pe9-V-1 Pe9-V-1 O6-2-V O6-2-V O6-2-V L Pe9-V-2 Pe9-V-2 Pe9-V-2 Pe9-V-2

Figure 3 | Conversion of lips into sepal/petal-like structures in Oncidium and Phalaenopsis orchids by suppressing OAGL6-2 orthologue expression through the VIGS strategy. a,TheOAGL6-2 VIGS flower (bottom row) containing green sepal/petal-like sectors in the lips (arrowed) is smaller than wild- type-like control (Mock) (top row) Oncidium Lemon Heart. Scale bar, 10 mm. LP, lips; P, petals; DS, dorsal sepals; LS, lateral sepals. b, Close-up of the green sepal/petal-like sectors on the adaxial side (b-2) and abaxial side (b-4)ofOAGL6-2 VIGS lips, which are distinct from wild-type like control (Mock) lips (b-1, b-3). c, The wild-type-like control (Mock) O. Lemon Heart lips exhibit conical cell morphology in the abaxial side (c-1); the petals exhibit elongated and flattened cell morphology (c-2). The OAGL6-2 VIGS green sepal/petal-like lip exhibits elongated and flattened cell morphology in the abaxial side (c-3), similar to petals (c-4). Scale bar, 50 μm. d, Detection of OAP3-2/OAGL6-2 (L) and OAP3-1/OAGL6-1 (SP) expression in wild-type-like control (Mock) and OAGL6-2 VIGS (O6-2-V) O.LemonHeartflowers. Expression levels of OAP3-2/OAGL6-2 and OAP3-1/OAGL6-1 in OAGL6-2 VIGS lips are marked with red and blue stars, respectively. e, Expression patterns of SP (brown) and L (yellow) complexes in the perianths of wild-type (e-1)andOAGL6-2 VIGS (e-2) O. Lemon Heart. In sepal, petal and lip, the higher expression for SP (brown) or L (yellow) complexes, the darker box colour is indicated. f,Theflowers of PeMADS9 (OAGL6-2 orthologues) VIGS hybrid (Pe9-VIGS-1, middle; Pe9-VIGS-2, right) contain larger sepal/petal-like lips, which are much more spread out than that in the wild-type-like control flower (Mock, left). Scale bar, 20 mm. LP, lips; P, petals; DS, dorsal sepals; LS, lateral sepals. g, Close-up of the front lobe (LP-f) and side lobe (LP-s) of the lips from f. Scale bar, 10 mm. h, The epidermal cells of the front lobe (LP-f) for wild-type like control (Mock) P. amabilis are flat and ridged with cuticles (h-1); the petals exhibit a conical cell morphology with extra holdfast papillae on top (h-2). The PeMADS9 VIGS (Pe9-VIGS-1) sepal/petal-like front lobe (LP-f) exhibits conical cell morphology with extra holdfast papillae on top (h-3), similar to petals (h-4). Scale bar = 50 μm. i, The detection of the expression of PeMADS3/PeMADS9 (L) and PeMADS2/PeMADS10 (SP) in the front lobe (LP-f) and side lobe (LP-s) of the lips, petals (P) and sepals (S) of wild-type like control (Mock) and PeMADS9 VIGS (Pe9-V-1, Pe9-V-2) P. amabilis hybrid flowers. The expression levels of PeMADS3/PeMADS9 and PeMADS2/PeMADS10 in PeMADS9 VIGS front (LP-f) and side (LP-s) lobes are marked with red and blue stars, respectively. j, Expression patterns of SP (brown) and L (yellow) complexes in perianths of wild-type (j-1)andPeMADS9 VIGS (j-2) P. amabilis hybrid. In sepal, petal and lip, the higher expression for SP (brown) or L (yellow) complexes, the darker box colour is indicated.

4 NATURE PLANTS |VOL1|MAY2015|www.nature.com/natureplants NATURE PLANTS DOI: 10.1038/NPLANTS.2015.46 LETTERS

Distinct lip formation

Oncidiinae 1 1611 12 15

Eulophiinae Cymbidieae 2

Cyrtopodiinae 3

Oncidiinae DendrobieaeOrchideae 4 Intermediate lip formation (I) 5 2 3 7 8

6 Laeliinae 7 Epidendroideae Eulophiinae Cyrtopodiinae Laeliinae 8 9 13 14 Arethuseae 9

10 Arethuseae Cranichideae Paphiopedilae 11

Orchideae 12 Intermediate lip formation (II) Orchidoideae Cranichideae 13 4 5 10

Paphiopedilae Cypripedioideae 14

Vanilloideae 15 Aerangidinae Aeridinae Dendrobieae Orchidaceae Apostasioideae

Tepal Sepal Petal Lip Sepal Petal Lip formation SP in sepal/petal SP SP SP L L L in lip

SP in sepal/petal SP SP SP Liliales L L strongly/SP weakly in lip Lily SP in sepal/petal SP SP SP L Petal L/SP equal in lip specification

SP-like in petal SP-like in all petals and sepals Eudicots Arabidopsis L duplicated from SP

Figure 4 | Possible evolutionary relationships between the SP and L complexes of the P-code model involved in regulating sepal/petal/lip formation in orchids. In eudicots such as Arabidopsis, SP-like complexes determine petal specification. In monocots, such as Liliales, the SP-like complex was expressed in all perianths and specified petal-like sepal formation. In Orchidaceae, the L complex genes were further duplicated from SP complex genes. The co- expression of strong levels of L with different weak levels of the SP complex in one of the petals converted it into a lip-like intermediate organ (blue and red dots) in most orchids. In some orchids, the SP complex was completely omitted from this lip-like intermediate organ, and the sole expression of L complex genes converted it into a more distinct lip structure (black dot). the GRtrip mutant resulted in the full transformation of the petals Fig. 3a-2, left, b-2) and its peloric mutant Bllra-Trip (Supplementary into lip-like structures. In GRpl (Fig. 1a-1, right) and a peloric Fig.3a-2,right,b-2). mutant O. May Fair for O. Lemon Heart (Supplementary Fig. 3a-1, We next examined Oncidium species with different types of peri- right), which exhibit partial conversion of petal to lip-like structures, anth conversion. The epidermal cells of the sepals and petals in the expression level of the SP unit (OAP3-1/OAGL6-1)was O. Gower Ramsey share identical morphology of elongated and flat- reduced; nonetheless, the L complex was formed due to an increase tened shapes, and the lips have conical cells with extra holdfast in OAGL6-2 expression along with normal expression of OAP3-2 in papillae on top (Supplementary Fig. 5, first row, GR). O. ampliatum the lip-like petals (Fig. 1b-1 and Supplementary Fig. 3b-1). The and O. macropetalum are different from O. Gower Ramsey in coexistence of the SP and L complexes in the peloric petals of their lip-like petal morphology (Fig. 1a-2, Amp, and a-3, Macro) GRpl and May Fair shifted the P-code balance to the middle, result- and have similar epidermal cell morphology of the lips ing in the production of a sepal/petal-lip intermediate structure (Supplementary Fig. 5). In the lip-like petals of O. ampliatum, the (Fig. 1d-3 and Supplementary Fig. 3c). This result also occurred coexistence of SP (OAP3-1/OAGL6-1) and L (OAP3-2/OAGL6-2) for the horticultural cultivar x Eurostar (Supplementary units (Fig. 1b-2) tended to transform petals into lip-like structures

NATURE PLANTS | VOL 1 | MAY 2015 | www.nature.com/natureplants 5 LETTERS NATURE PLANTS DOI: 10.1038/NPLANTS.2015.46

(Fig. 1a-2, d-3). In the lip-like petals of O. macropetalum, the the significantly decreased expression of OAGL6-2 in the OAGL6-2 absence of both SP and L complexes due to undetectable OAGL6-1 VIGS lips (Fig. 3d), which caused a reduction in the expression and OAGL6-2 expression (Fig. 1b-3) appeared to also convert the level of the L unit (OAP3-2/OAGL6-2) (Fig. 3d, e-1, e-2). We also petals into intermediate lip-sepal/petal-like structures (Fig. 1a-3, found that the SP complex did form due to increased expression d-4). This hypothesis is also true for Psychopsis papilio (Fig. 1a-4, of OAGL6-1 and OAP3-1 in the OAGL6-2 VIGS sepal/petal-like Pp), a member of Oncidiinae, which produces lateral sepals with lip (Fig. 3d, e-2). Thus, coexistence of the SP and L complexes intermediate epidermal cells and a lip-sepal/petal phenotype shifted the P-code balance to the middle, resulting in production (Supplementary Fig. 5). In these lip-like lateral sepals, neither the of a sepal/petal-like lip structure. SP complex nor the L complex formed because of undetectable To extensively test the P-code hypothesis beyond Oncidiinae levels of OAP3-1 and OAP3-2 expression (Fig. 1b-4). Thus, the exist- orchids, additional orchid species from four different subfamilies ence (GRpl, May Fair and O. ampliatum)orabsence(O. macropetalum (Epidendroideae, Orchidoideae, Cypripedioideae and Vanilloideae) and P. papilio) of both the SP and the L complexes may induce the were analysed (Fig. 4). In Epidendroideae, and formation of similar intermediate lip-sepal/petal-like organs orchids (tribe Cymbidieae) (Supplementary Fig. 6), and (Fig. 1d-3, d-4). We next tested the P-code model in the Phalaenopsis orchids (tribe Vandeae) (Supplementary Fig. 7), Oncidiinae species O. carthaginense and pulchellum, and orchids (tribe Epidendreae) (Supplementary Fig. 8), both of which exhibit sepal/petal-like lips (Fig. 1a-5, a-6 and , orchids (tribe Arethuseae) and Supplementary Fig. 5). Unsurprisingly, the coexistence of the SP orchids (tribe Dendrobieae) (Supplementary Fig. 9) were included. (OAP3-1/OAGL6-1) and L (OAP3-2/OAGL6-2) units was found in Spiranthes (tribe Cranichideae) and Habenaria orchids (tribe the lips (Fig. 1b-5, b-6), tipping the P-code balance and converting Orchideae) in subfamily Orchidoideae (Supplementary Fig. 10) the lips into a more sepal/petal-like intermediate structure in these were used, and Vanilla orchids from subfamily Vanilloideae two species (Fig. 1d-3). (Supplementary Fig. 10) were utilized. For subfamily Cypripedioideae, We then performed fluorescence resonance energy transfer Paphiopedilum orchids (known as lady slipper orchids) (tribe (FRET) analyses to observe physical interactions of the proposed Paphiopedilae) (Supplementary Fig. 11) were included. Only the P-code protein complexes in vivo. In tobacco cells, the SP proteins SP complex (OAP3-1/OAGL6-1 homologues) was detected in the OAP3-1/OAGL6-1 formed heterodimers in the nucleus with the sepals/petals of all orchids tested (Supplementary Figs 6b–11b). In assistance of the OPI protein (unlabelled) (Fig. 2a, columns 2, 3, the lips, the L complex (OAP3-2/OAGL6-2 homologs) was either b-1, b-2). The same result was observed for the L proteins OAP3- highly and exclusively expressed (Supplementary Figs 6b-2, 8b-5, 2/OAGL6-2 (Fig. 2a, columns 7, 8, b-1, b-2). These results suggest 9b-4, 10b-2 to b-5, 11b-3) or expressed in combination with the existence of OAP3-1/OAGL6-1/OPI and OAP3-2/OAGL6-2/ various degrees of weak SP complex expression (Supplementary OPI complexes. Based on the floral quartet model described for Figs 6b-1, b-4, 7b-1 to b-4, 8b-1 to b-4, 9b-1 to b-3, 10b-1, 11b-1, Arabidopsis, the MADS complex controlling petal development is b-2). Similar to Oncidium orchids, a coexistence (Supplementary composed of one AP3, one PI and two A/E class proteins23–26.We Figs 6b-3, 7b-3-2, 8b-3, b-4, 9b-5, 11b-3) or co-absence found that the SP protein OAGL6-1 and the L protein OAGL6-2 (Supplementary Figs 10b-2, b-3, 11b-2) of the SP and L complexes form homodimers and localize to the nucleus without the OPI was found in the lip-like petals of peloric mutants. These results are protein (Fig. 2a, columns 6, 11, b-3). Furthermore, the heterodimers summarized in Fig. 4. OAP3-1/OPI and OAP3-2/OPI formed and also localized to the To further functionally test the P-code model in orchid species nucleus (Fig. 2a, columns 5, 10, b-3). Thus, the most likely other than Oncidium, a similar VIGS strategy was also applied for higher-order heterotetrameric SP complex is OAP3-1/OAGL6-1/ the Phalaenopsis amabilis hybrid by suppressing PeMADS9 OAGL6-1/OPI, which is composed of one AP3/PI (OAP3-1/OPI) (OAGL6-2 orthologue) expression and reducing the activity of the heterodimer and one A/E (OAGL6-1/OAGL6-1) homodimer proposed L complex (Supplementary Table 2). Interestingly, (Fig. 2b-2). Similarly, the most likely higher-order heterotetrameric PeMADS9 VIGS lips were significantly spread out and expanded L complex is OAP3-2/OAGL6-2/OAGL6-2/OPI, which is com- to a larger size than in wild-type lips, which consisted of front posed of one AP3/PI (OAP3-2/OPI) heterodimer and one A/E and side lobes (Fig. 3f,g), resulting in a more sepal/petal appearance (OAGL6-2/OAGL6-2) homodimer (Fig. 2b-2). distinct from normal lips (Fig. 3g). Further examination indicated We herein propose a model to explain how P-code complexes that this sepal/petal-like lip has epidermal cell morphology similar regulate Oncidium orchid perianth formation (Fig. 2b). The SP to sepals/petals (Fig. 3h). The production of this sepal/petal-like and L complexes may have evolved to have opposite functions in lip structure was correlated with the significantly decreased the regulation of a specific ‘Lip program” that is able to promote expression of PeMADS9 in PeMADS9 VIGS lips (Fig. 3i). These lip development and suppress petal/sepal development (Fig. 2b). decreases caused a reduction in the expression level of the L unit The L quartet (OAP3-2/OAGL6-2/OAGL6-2/OPI) activates this (PeMADS3/PeMADS9) (Fig. 3i) to a level lower than that of the ‘Lip program” and promotes lip formation, whereas the SP SP unit (PeMADS2/PeMADS10) in the PeMADS9 VIGS sepal/ quartet (OAP3-1/OAGL6-1/OAGL6-1/OPI) suppresses this ‘Lip petal-like lip (Fig. 3i, j-2) and shifted the P-code balance to sepal/ program” and promotes sepal/petal formation (Fig. 2b). To petal, resulting in the production of a sepal/petal-like lip structure. further test this model, virus-induced gene silencing (VIGS) was uti- In contrast, the expression level of the L unit was higher than that lized in O. Lemon Heart to downregulate the activity of the pro- of the SP unit in the wild-type lip (Fig. 3i, j-1). Thus, the P-code posed L complex by suppressing OAGL6-2 in the lips model is clearly supported by these direct functional data (Supplementary Table 1). Interestingly, the size of the lip was for orchids. reduced in OAGL6-2 VIGS flowers compared with the wild-type In conclusion, the scenario that the P-code model describes is lip (Fig. 3a). Furthermore, sepal/petal-like green sectors similar to conserved and can be precisely applied to explain diverse perianth those observed in sepals/petals were produced in the lip of organ formation in orchids. We propose a hypothesis to explain OAGL6-2 VIGS flowers (Fig. 3a,b). This resulted in a green sepal/ the evolutionary modification of the P code model in orchids petal within a lip structure (Fig. 3a, bottom, b-2, b-4) that was dis- (Fig. 4). An ancestral form of an SP (PI + euAP3 + A/E)-like tinct from wild-type yellow lips (Fig. 3a, top, b-1, b-3). Further complex regulating petal formation was maintained in most examination indicated that this sepal/petal-like green sector has epi- eudicot plants, such as Arabidopsis, during evolution. In monocots, dermal cell morphology similar to that of sepals/petals (Fig. 3c). The such as Liliaceae (in Liliales), the SP (PI + paleoAP3 + AGL6-like)- production of this sepal/petal-like lip structure was correlated with like complex extended to all the perianth and converted sepals

6 NATURE PLANTS |VOL1|MAY2015|www.nature.com/natureplants NATURE PLANTS DOI: 10.1038/NPLANTS.2015.46 LETTERS into petal-like structures. In Orchidaceae, except for the Transcript levels for these MADS box genes were determined using three Apostasioideae subfamily in which the flowers have weakly to replicates and were normalized using reference genes ACTIN for Phalaenopsis18,21 and α-tubulin for other orchids16. By referring to several α-tubulin genes to non-differentiated lips, L genes consisting of new paleoAP3 locate a conserved sequence region, the universal primers for α-tubulin of orchids (OAP3-2-like) and AGL6 (OAGL6-2-like) genes emerged after were designed. To exclude false amplification from genomic DNA contamination, duplication of SP genes (OAP3-1-like and OAGL6-1-like genes). primers were designed in two exons separated by one intron. To ensure primer Additional expression of L complex genes in one petal resulted in specificity, we only adopted orchid samples with Ct (threshold cycle) values of α fi the coexistence of minor SP and major L complexes and in for- -tubulin genes between 15 and 30. Also, the PCR products were detected with ne melting curves and subjected to autosequencing to confirm the sequence identity. mation of various forms of intermediate lip organs in most orchid Each experiment was repeated twice or more if enough samples were available. species. With further loss of the SP complex in this intermediate lip organ, exclusive expression of the L complex converted the Virus-induced gene silencing experiment. VIGS experiments in orchids were organ into the distinct lip organ found in Oncidium (O. Gower performed as described previously28,29. DNA fragments for OAGL6-2 and PeMADS9 (OAGL6-2 homologue) for insertion into the VIGS vector pCymMV–Gateway30 Ramsey), Dendrobium (D. Sonia), Habenaria (H. rhodocheila), were obtained by PCR amplification using the following primers: OAGL6-2 (in the Vanilla (V. planifolia) and Epidendrum orchids (Fig. 4). C domain, 160 bp), O1-gateway-F, 5′-GGGGACAAGTTTGTACAAAAA AGCAGGC TAGCAAATGGTGGGTCATC-3′; O1-gateway-R, 5′- Methods GGGGACCACTTTGTACAAGAAAGCTGGGTAAATGGTTGCTTCAGAAG-3′. Plant materials. Species, cultivars and peloric mutants used in this study, including PeMADS9 (in the C domain, 150 bp), PeM9-VIGS-F, 5′-GGGGACAAGTTT the Oncidium Gower Ramsey, the parental species (O. flexuosum), cultivars GTACAAAAAAGCAGGCTCAGGTGATATTAACAAGCAGCTTAAA CA-3′; (O. Sweet Sugar, O. Lemon Heart, O. May Fair and x Beallara Eurostar), botanical PeM9-VIGS-R, 5′-GGGGACCACTTTGTACAAGAAAGCTGGGTCTTTGA species (O. gutfreundianum, O. tipuloides, O. cheirophorum and O. cebolleta), related AGAGTGGGTTCTGTATCCAT G-3′. The underlined sequences are attB sites for species (O. ampliatum, O. macropetalum, Psy. papilio, O. carthaginense and Odm. in vitro recombination with attP sites in the VIGS vector pCymMV–Gateway to pulchellum), moth orchids (P. equestris, P. amabilis hybrid and the associated peloric generate recombinant clones using Gateway BP Clonase II Enzyme Mix (Invitrogen, mutants), wind orchids (A. fastuosa and A. biloba), Cymbidium spp., Geodorum Life Technologies). pCymMV–Gateway–OAGL6-2, pCymMV–Gateway–PeMADS9 densiflorum, Cattleya spp., Dendrobium spp., Epidendrum spp., , and the empty pCymMV–Gateway as a control were transformed into Arundina graminifolia, Spiranthes sinensis, Habenaria spp., Vanilla spp. and Agrobacterium tumefaciens EHA105 for further inoculation. For Oncidium Lemon Paphiopedilum spp. (lady slipper orchids) were maintained in the greenhouse of the Heart infiltration, suspensions were injected into the third leaf (L3) just below National Chung-Hsing University and the National Museum of Natural Science, the site where the inflorescence emerged. For P. amabilis hybrid leaf infiltration, Taichung, Taiwan. The voucher information (herbarium and spirit collections of all suspensions were injected into the leaf just above the site where the inflorescence plants analysed) has been made (Supplementary Table 3) and stored in the National emerged. For every infiltration, at least three plants were inoculated with each Museum of Natural Science, Taichung, Taiwan. pCymMV–Gateway construct. Flower samples were collected and analysed at 45 DPI (days post inoculation), when the buds at the end of the Oncidium Cryoscanning electron microscopy. At the time of anthesis, perianths were inflorescences or the last bud of the Phalaenopsis inflorescences bloomed. dissected and loaded onto stubs. Samples were frozen using liquid nitrogen and transferred to the sample preparation chamber at −160 °C. Samples were etched for Received 19 January 2015; accepted 18 March 2015; 10 min at −85 °C. After gold coating, samples were observed under a cryoscanning published 27 April 2015 electron microscope (FEI Quanta 200 SEM, Quorum Cryo System PP2000TR FEI; FEI Company). References FRET analysis. The procedure used to prepare FRET-associated fusion constructs 1. Cameron, K. M. et al. A phylogenetic analysis of the Orchidaceae: evidence from was described in a previous study27. To fuse OAGL6-2/OAP3-2/OAGL6-1/OAP3-1/ rbcL nucleotide. Am. J. Bot. 86, 208–224 (1999). OPI with CFP or YFP, the complementary DNAs for OAGL6-2/OAP3-2/OAGL6-1/ 2. Chase, M. W., Cameron, K. M., Barrett, R. L. & Freudenstein, J. V. in Orchid OAP3-1/OPI were obtained by polymerase chain reaction (PCR) amplification using Conservation (eds Dixon, K. M., Kell, S. P., Barrett, R. L. & Cribb, P. J.) gene-specific primers and cloned into the pEpyon-36 K and pEpyon-37 K vectors 69–89 (Natural History Publications, 2003). upstream of the CFP or YFP sequence under the control of the CaMV 35S promoter. 3. Gorniak, M., Paun, O. & Chase, M. W. Phylogenetic relationships within The following gene-specific primers were used: OAGL6-2, OM1-ATG-F-XbaI Orchidaceae based on a low-copy nuclear coding gene, Xdh: congruence (5′-CCTCTAGAATGGGAAGGGGCAGAGTAG-3′) and OM1-R-KpnI (5′- with organellar and nuclear ribosomal DNA results. Mol. Phylogenet. Evol. CCGGTACCAACAGCCCATCCTGACATG-3′); OAP3-2, OM9-ATG-F-BamHI 56, 784–795 (2010). (5′-CCGGATCCATGGGAAGAGGAAAGATAGAA-3′) and OM9-R-SalI (5′- 4. Dressler, R. L. Phylogeny and classification of the Orchid family [electronic CCGTCGACGGCGAGACGAAGATCATG-3′); OAGL6-1, OM7-ATG-F-XbaI resource] (Dioscorides Press, 1993). (5′-CCTCTAGAATGGGGCGAGGAAGAGTTGAG-3′) and OM7-R-SmaI 5. Rudall, P. J. & Bateman, R. M. Roles of synorganisation, zygomorphy and (5′-CACCCGGGAGAGCATCCATCCAAGCAT-3′); OAP3-1, OM5-ATG-F- heterotopy in floral evolution: the gynostemium and labellum of orchids and BamHI (5′-CCGGATCCATGGGGAGGGGGAAGATA-3′) and OM5-R-SalI other lilioid monocots. Biol. Rev. Camb. Philos. Soc. 77, 403–441 (2002). (5′-CCGTCGACAGAAAGGCTAAGATCATG-3′); OPI, OM8-ATG-F-BamHI 6. Kocyan, A., Conti, E., Qiu, Y-L. & Endress, P. K. A phylogenetic analysis of (5′-CCGGATCCATGGGGCGCGGAAAGA-3′) and OM8-R-SalI (5′- Apostasioideae (Orchidaceae) based on ITS, trn L-F and mat K sequences. CCGTCGACCTTGTTTCCCTGCAAGTTGG-3′). These constructs were Plant Syst. Evol. 247, 203–213 (2004). transformed into the Agrobacterium strain C58C1. Different ectopic proteins were 7. Cozzolino, S. & Widmer, A. Orchid diversity: an evolutionary consequence expressed in tobacco cells, and fluorescence signals in the nucleus were detected with of deception? Trends Ecol. Evol. 20, 487–494 (2005). a confocal microscope. For the subcellular localization assay, Agrobacterium- 8. Krizek, B. A. & Meyerowitz, E. M. The Arabidopsis homeotic genes APETALA3 fi fi fi in ltrated Nicotiana benthamiana were vacuum in ltrated in 10 mM MgCl2 and PISTILLATA are suf cient to provide the B class organ identity function. at room temperature until immersed. Visualization of fluorophores and Development 122, 11–22 (1996). calculation of raw FRET and FRET efficiency values were performed using an 9. Kramer, E. M., Dorit, R. L. & Irish, V. F. Molecular evolution of genes controlling Olympus FV1000 confocal microscope (Olympus FV1000, Tokyo, Japan) and soft petal and stamen development: duplication and divergence within the APETALA3 FV-ASW 3.0 software as previously described28. The mean value of FRET efficiency and PISTILLATA MADS-box gene lineages. Genetics 149, 765–783 (1998). in the nucleus was calculated to evaluate the variation in protein interaction 10. Hernandez-Hernandez, T., Martinez-Castilla, L. P. & Alvarez-Buylla, E. R. distances among different protein complexes (n > 4). Functional diversification of B MADS-box homeotic regulators of flower development: adaptive evolution in protein-protein interaction domains after RNA isolation and real-time PCR analysis. Total RNA was isolated from various major gene duplication events. Mol. Biol. Evol. 24, 465–481 (2007). organs of orchid flowers as described previously12. Total RNA was used to synthesize 11. Irish, V. F. Evolution of petal identity. J. Exp. Bot. 60, 2517–2527 (2009). cDNA using the ImProm-II RT System (Promega). Quantitative real-time PCR was 12. Hsu, H. F. & Yang, C. H. An orchid (Oncidium Gower Ramsey) AP3-like conducted using a Mini OpticonReal-Time PCR Detection System and Optical MADS gene regulates floral formation and initiation. Plant Cell Physiol. 43, System Software version 3.0a (Bio-Rad Laboratories). KAPA SYBR FAST 1198–1209 (2002). quantitative (q)PCR Master Mix Universal (KAPA Biosystems) was used for 13. Tsai, W. C., Kuoh, C. S., Chuang, M. H., Chen, W. H. & Chen, H. H. Four DEF- transcript measurements. The cycling programme was as follows: one cycle at 95 °C like MADS box genes displayed distinct floral morphogenetic roles in for 1 min, then 40 cycles of 95 °C (15 s), 58 °C (15 s) and 72 °C (30 s), and plate Phalaenopsis orchid. Plant Cell. Physiol. 45, 831–844 (2004). reading after each cycle. The name and the sequence of the gene-specific primers for 14. Xu, Y., Teo, L. L., Zhou, J., Kumar, P. P. & Yu, H. Floral organ identity genes in MADS box genes of different orchids are listed in Supplementary Tables 4–6. the orchid . Plant J. 46, 54–68 (2006). The data were analysed using CFX Manager Software (Version 3.0; Bio-Rad 15. Mondragon-Palomino, M. & Theissen, G. MADS about the evolution of Laboratories, Inc.). orchid flowers. Trends Plant Sci. 13, 51–59 (2008).

NATURE PLANTS | VOL 1 | MAY 2015 | www.nature.com/natureplants 7 LETTERS NATURE PLANTS DOI: 10.1038/NPLANTS.2015.46

16. Chang, Y. Y. et al. Characterization of the possible roles for B class MADS 28. Hsieh, M. H. et al. Virus-induced gene silencing unravels multiple transcription box genes in regulation of perianth formation in orchid. Plant Physiol. 152, factors involved in floral growth and development in Phalaenopsis orchids. 837–853 (2010). J. Exp. Bot. 64, 3869–3884 (2013). 17. Mondragon-Palomino, M. & Theissen, G. Conserved differential expression of 29. Hsieh, M. H. et al. Optimizing virus-induced gene silencing efficiency with paralogous DEFICIENS- and GLOBOSA-like MADS-box genes in the flowers of Cymbidium mosaic virus in Phalaenopsis flower. Plant Sci. 201–202, Orchidaceae: refining the ‘orchid code’. Plant J. 66, 1008–1019 (2011). 25–41 (2013). 18. Pan, Z. J. et al. The duplicated B-class MADS-box genes display dualistic 30. Lu, H. C. et al. A high-throughput virus-induced gene-silencing vector for characters in orchid floral organ identity and growth. Plant Cell Physiol. screening transcription factors in virus-induced plant defense response in 52, 1515–1531 (2011). orchid. Mol. Plant Microbe Interact. 25, 738–746 (2012). 19. Su, C. L. et al. A modified ABCDE model of flowering in orchids based on gene expression profiling studies of the moth orchid . PLoS Acknowledgements ONE 8, http://dx.doi.org/10.1371/journal.pone.0080462 (2013). This work was supported by grants to C-H.Y. from the National Science Council, Taiwan, 20. Acri-Nunes-Miranda, R. & Mondragon-Palomino, M. Expression of paralogous ROC, grant number: NSC96-2752-B-005-007-PAE and NSC 100-2313-B-005-004-MY3. SEP-, FUL-, AG- and STK-like MADS-box genes in wild-type and peloric This work was also supported in part by the Ministry of Education, Taiwan, ROC under the Phalaenopsis flowers. Front. Plant Sci. 5, http://dx.doi.org/doi:10.3389/fpls.2014. ATU plan. We thank Drs Elena M. Kramer (Department of Organismic and Evolutionary 00076 (2014). Biology, Harvard University) and Kerstin Kaufmann (Institute of Biochemistry and 21. Pan, Z. J. et al. Flower development of Phalaenopsis orchid involves functionally Biology, University of Potsdam) for their helpful discussion of the results. divergent SEPALLATA-like genes. New Phytol. 202, 1024–1042 (2014). 22. Chang, Y. Y., Chiu, Y. F., Wu, J. W. & Yang, C. H. Four orchid (Oncidium Gower Ramsey) AP1/AGL9-like MADS box genes show novel expression patterns Author contributions and cause different effects on floral transition and formation in Arabidopsis C-H.Y. and H-F.H. developed the overall strategy, designed experiments and coordinated thaliana. Plant Cell. Physiol. 50, 1425–1438 (2009). the project. H-F.H. and W-T.M. performed gene expression analyses. W-H.H., H-F.H. and 23. Theissen, G. Development of floral organ identity: stories from the MADS J-Y.L. performed FRET analyses. Y-I.L., H-F.H. and C-H.Y. collected the orchid samples. house. Curr. Opin. Plant Biol. 4, 75–85 (2001). Y-I.L. and H-F.H. performed the cryoscanning electron microscopy. H-F.H., W-T.M. and 24. Theissen, G. & Saedler, H. Plant biology. Floral quartets. Nature 409, J-Y.Y. performed VIGS experiments. C-H.Y. and H-F.H. prepared and revised 469–471 (2001). the manuscript. 25. Melzer, R. & Theissen, G. Reconstitution of ‘floral quartets’ in vitro involving class B and class E floral homeotic proteins. Nucleic Acids Res. 37, Additional information 2723–2736 (2009). Supplementary information is available online. Reprints and permissions information is 26. Smaczniak, C. et al. Characterization of MADS-domain transcription factor available online at www.nature.com/reprints. Correspondence and requests for materials should complexes in Arabidopsis flower development. Proc. Natl Acad. Sci. USA be addressed to C.H.Y. 109, 1560–1565 (2012). 27. Hsu, W. H. et al. AGAMOUS-LIKE13, a putative ancestor for the E functional genes, specifies male and female gametophyte morphogenesis. Competing interests Plant J. 77, 1–15 (2014). The authors declare no competing financial interests.

8 NATURE PLANTS |VOL1|MAY2015|www.nature.com/natureplants

View publication stats