Plant Physiology Preview. Published on April 20, 2015, as DOI:10.1104/pp.15.00321

1

RUNNING HEAD

2

CrDOF and the origin of photoperiodic flowering

3 4

CORRESPONDING AUTOR

5

Federico Valverde.

6

Institute for Plant Biochemistry and Photosynthesis. Plant Development Unit.

7

CSIC and Universidad de Sevilla. 49th Americo Vespucio Av., 41092 Seville,

8

SPAIN

9

Tel ++34 954 489525

10

E-mail: [email protected]

11

Web:

12

metabolism

http://www.ibvf.csic.es/en/molecular-basis-flowering-photoperiod-and-

13 14 15 16

RESEARCH AREA Genes, Development and Evolution

17

1 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

Copyright 2015 by the American Society of Plant Biologists

18

An evolutionarily conserved DOF-CONSTANS

19

module controls plant photoperiodic signalling

20 21

Eva Lucas-Reina1, Francisco J. Romero-Campero2, José M. Romero1 and

22

Federico Valverde1,*

23 24

1

Institute for Plant Biochemistry and Photosynthesis. Plant Development Unit.

25

CSIC and Universidad de Sevilla. 49th Americo Vespucio Av., 41092 Seville,

26

SPAIN.

27

2

Departamento de Ciencias de la Computación e Inteligencia Artificial, Grupo

28

de Investigación en Computación Natural, Universidad de Sevilla, Reina

29

Mercedes Av., 41012 Seville, SPAIN

30 31 32

SUMMARY

33

Algal DOF transcription factor is involved in photoperiod response and this

34

function is conserved in angiosperms despite its wide functional amplification

35

and diversification.

36 37 38 39 40 41 42 43 44 2 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

45 46

FOOTNOTES

47 48

Financial sources

49

Work supported by projects BIO2011-28847-C02-00 and BIO2014-52425-P

50

(MINECO) and P08-AGR-03582 (Junta de Andalucía) and FEDER funding.

51

E.L.-R. fellowship is a CSIC–Junta de Ampliación de Estudios fellow, partly

52

supported by structural funding from the European Union (European Social

53

Fund). RNAseq data was performed with a help from the “Juan de la Cierva”

54

Program (MINECO) to F.J.R.-C.

55 56 57

Corresponding author

58

Federico Valverde

59

E-mail: [email protected]

60 61

3 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

62

Abstract

63

The response to day length is a crucial process that evolved very early in plant

64

evolution, entitling the early green eukaryote to predict seasonal variability and

65

attune its physiological responses to the environment. The photoperiod

66

responses evolved into the complex signalling pathways that govern

67

angiosperm floral transition today. The Chlamydomonas CrDOF gene controls

68

transcription in a photoperiod-dependent manner and its missexpression

69

influences algal growth and viability. In short days, CrDOF enhances CrCO

70

expression, a homolog of plant CONSTANS (CO) by direct binding to its

71

promoter, while it reduces the expression of cell-division genes in long days

72

independently of CrCO. In Arabidopsis, transgenic plants overexpressing

73

CrDOF show floral delay and reduced expression of the photoperiodic genes

74

CO and FLOWERING LOCUS T (FT). The conservation of the DOF-CO module

75

during plant evolution could be an important clue to understand diversification

76

by the inheritance of conserved gene toolkits in key developmental programs.

77 78 79 80 81 82

4 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

83

Introduction

84

Due to its sessile lifestyle, vascular plants have evolved intricate genetic

85

regulatory pathways to control essential developmental processes in response

86

to changes in the environment. One of the most distinctive processes is the

87

floral transition (Amasino, 2010) that integrates information from endogenous

88

and external signals to guarantee the reproductive success (Valverde, 2011). In

89

Arabidopsis the CO gene plays a central role in the control of the photoperiodic

90

flowering by triggering the expression of FT in the leaf vascular tissue (An et al.,

91

2004). FT codes for a major component of the florigenic signal that moves

92

through the phloem to induce flower differentiation at the shoot apex (Corbesier

93

et al., 2007; Tamaki et al., 2007). CO and FT are controlled at the

94

transcriptional level by a circadian and photoperiodic signal through the

95

FLAVIN-BINDING, KELCH REPEAT, F-BOX 1 (FKF1), GIGANTEA (GI) and

96

CDF genes (Song et al., 2012). In long days (LD), blue light induces the

97

assembly of the FKF1-GI complex that promotes the proteasome-mediated

98

degradation of the CDFs, a four-member family of DOF (DNA-binding with One

99

Finger) transcription factors (TFs) that bind to CO and FT promoters, repressing

100

their expression (Song et al., 2012). The photoperiodic pathway is probably the

101

most conserved of the flowering signalling responses in Spermatophytes

102

(Amasino, 2010) and can be traced back to Chlorophytes where CrCO, a CO

103

homologous gene from Chlamydomonas reinhardtii was identified (Serrano et

104

al., 2009). CrCO is involved in the regulation of starch synthesis and cell growth

105

in the alga (Serrano et al., 2009), two processes that are also regulated by CO-

106

Like (COL) gene family members in vascular plants (Valverde, 2011; Romero-

107

Campero et al., 2013). In addition, transgenic plants overexpressing CrCO

108

under a constitutive or phloem-specific promoter, flowered earlier than WT

109

plants and complemented co mutants (Serrano et al., 2009).

110

CDFs constitute a subgroup of DOF TFs (Imaizumi et al., 2005; Fornara

111

et al., 2009), characterized by the presence of a 52 amino acid DOF domain, a

112

single zinc finger that binds to the DNA consensus sequence AAAG. DOFs can

113

work as transcriptional activators or inhibitors depending on the protein and the

114

target gene concerned (Noguero et al., 2013). The DOF domain is bi-functional

115

as it can participate in DNA-binding and protein–protein interactions 5 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

116

(Yanagisawa, 1997). DOFs have been exclusively described in plants and are

117

involved in numerous processes from tissue differentiation to metabolic

118

regulation and seed development. They are widely distributed, with numerous

119

genomic copies, in both Gymnosperms and Angiosperms and constitute a small

120

gene family in some Briophytes such as in the moss Physcomitrella patens. A

121

single copy CrDOF gene from Chlamydomonas (Moreno-Risueño et al., 2007;

122

Shigyo et al., 2007) represents the ancestor of the DOF TFs family. DOFs are

123

absent in heterotrophic eukaryotes, red algae and heterokonts (Hernando-

124

Amado et al., 2012), so modern DOFs, similarly to COLs, constitute a family of

125

TFs exclusive of the green lineage (Serrano et al., 2009).

126

Here we show that CrDOF induces CrCO expression in short days (SD)

127

and is a general repressor of algal gene expression in LD. CrDOF expression in

128

Arabidopsis reproduces CDFs function by reducing CO and FT transcript levels

129

and delaying flowering. DOF-CO evolutionarily conserved signalling module

130

constitutes a direct clue of how developmental responses may have evolved

131

from unicellular algae to multicellular plants by the parallel evolution of

132

developmental gene toolkits.

133 134

6 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

135

Results

136

Chlamydomonas CrDOF defines a subfamily of DOF Transcription Factors

137

closely related to CDFs in Arabidopsis

138

In

139

(http://phytozome.jgi.doe.gov/pz/portal.html),

140

(Cre12.g521150) in position 4424576 – 4431469 on chromosome 12 (CrDOF),

141

showed a significant identity with DOF TFs (Moreno-Risueño et al., 2007). In

142

order to characterize its gene structure, three cDNA clones from the Kazusa

143

ORF collection (http://www.kazusa.or.jp), were sequenced. The CrDOF coding

144

sequence was 1875 bp long and was identical to the draft genome sequence

145

(Merchant et al., 2007) at the 3’ end, but shorter at the 5’ (Fig. 1A). This was

146

further confirmed by the alignment of the RNAseq data to the predicted gene

147

structure (Fig. 1B).

v5.5

of

the

C.

genome

reinhardtii a

annotation

single-copy

gene

148

Using CrDOF amino acid sequence, a blast search in several

149

representative plant species was run and 84 different putative DOF proteins

150

identified (Supplemental Table S1). These sequences were used to construct a

151

phylogenetic tree (Fig. 1C) using the Mega5 program (Tamura et al., 2011).

152

This phylogenetic tree presents four clusters, including Chlorophyte (cluster 1),

153

Bryophyte (cluster 2) and Spermatophyte (clusters 3a and cluster 3b-3c)

154

sequences. Cluster 3b-3c enclosed the Arabidopsis photoperiodic CDFs

155

(Imaizumi et al., 2005; Fornara et al., 2009), closely aligned with Chlorophyte

156

sequences including CrDOF. This protein family also comprised other DOFs

157

from Populus, Zea and Oriza, which may represent putative photoperiodic DOF

158

TFs, such as OsDOF12 (RDD4) that has been involved in photoperiodic

159

flowering in rice (Li et al., 2009). No Physcomitrella protein was included in this

160

cluster; rather they formed a group on their own bridging Chlorophyte with

161

Spermatophyte sequences. The subfamily 3a included DOF proteins from

162

Arabidopsis, poplar, maize and rice that have been involved in diverse

163

processes other than photoperiodic flowering, such as AtDAG1,2, involved in

164

germination in Arabidopsis (Papi et al., 2000; Gualberti et al., 2002) or ZmDOF3

165

(PBF1), involved in storage protein synthesis in maize seeds (Vicente-

166

Carbajosa et al., 1997).

7 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

167

CDF proteins have a characteristic domain structure (Fig. 1C, group 3b)

168

with a conserved DOF domain (blue) preceded by a nuclear localization signal

169

(NLS, red). Carboxy-terminal domains for GI (purple) and FKF1 (orange)

170

binding sites are specific for CDFs as they participate in the interactions that

171

control their posttranslational regulation (Kloosterman et al., 2013). A conserved 8 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

172

amino-terminal domain of unknown function (yellow) is also present. The

173

ancestral algal DOF protein from group 1 (Fig. 1C) contains the conserved DOF

174

domain and the NLS signal, but lacks any other clear domain. In Physcomitrella,

175

although GI and FKF1 proteins have not been identified in its genome (Rensing

176

et al., 2008), different DOF proteins (cluster 2) include regions that resemble the

177

GI, FKF1 binding site, that together with the yellow domains have later evolved

178

into the modern plant DOF domains in CDFs (clusters 3b and 3c) and other

179

DOFs (cluster 3a) (Kloosterman et al., 2013). Therefore, the original algal DOF

180

TF acquired new domains that defined their regulation and function (Fig. 1C) as

181

suggested by the close association between domain structure and function

182

observed in the phylogenetic tree.

183 184

CrDOF is differentially expressed under diverse photoperiod conditions in

185

Chlamydomonas

186

Arabidopsis CDFs are differentially expressed under diverse day lengths in

187

order to correctly time the expression of CO and confer a photoperiodic

188

response to flowering (Imaizumi et al., 2005). CrDOF expression in

189

Chlamydomonas was monitored under LD and SD every 4h for 24h starting at

190

Zeitgeber Time 0 (ZT0), the moment the lights are switched on. CrDOF mRNA

191

levels were higher in LD than in SD and the expression profile resembled that of

192

CDFs in Arabidopsis: A peak of expression in the late night and early morning

193

(Fornara et al., 2009) (Fig. 2A). When the expression was monitored during a

194

24h LD followed by 48h in continuous light (LL) or continuous dark (DD), the

195

expression of CrDOF continued oscillating (Fig. 2B). This was particularly true

196

for LL, but even a slight oscillation in DD was observed after 48 h, indicating a

197

circadian influence.

198

To monitor the presence of CrDOF protein we produced polyclonal

199

antibodies by immunizing rabbits with a recombinant CrDOF purified from E.

200

coli (Methods). These antibodies were used to detect CrDOF in extracts from

201

the recombinant bacteria and from Chlamydomonas (Fig. 2C). While the

202

recombinant protein showed a molecular mass (MM) of 60 kD, which is close to

203

its predicted size (62 kD), the extract from Chlamydomonas showed a unique

204

band of approximately 100 kD. Because we used SDS-PAGE conditions, the

205

result

suggested

that

the

100

kD

band

reflected

an

SDS-resistant

9 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

206

posttranslational modification, such as has been shown for other SDS-resistant

207

complexes from plants and animals described previously in the literature

208

(Lüders et al., 2003; Varet et al., 2003; Kubista et al., 2004). Using these

209

antibodies, we followed the presence of CrDOF in 24h courses in LD and SD

210

(Fig. 2D; Supplemental Fig. S1A), in protein extracts from the same samples as

211

in Fig. 2A. CrDOF protein showed a higher accumulation in LD than in SD and a

212

clear circadian oscillation. However, in contrast to the mRNA profile, CrDOF

213

minimal presence was monitored at ZT4 and its maximum at dusk (ZT16-20),

214

coinciding with CDF2 stability profile in Arabidopsis (Fornara et al., 2009). Thus,

215

the expression of CrDOF mRNA and its protein stability are displaced in time.

216

To study CrDOF modifications, protein extracts from Chlamydomonas

217

were incubated with different chemicals and immunodetected with α-CrDOF. 10 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

218

When proteins were extracted with 8 M urea (a potent chaotropic agent), or in

219

the presence of 10 mM DTT (that breaks cysteine disulphide bonds) or a

220

mixture of both (Supplemental Fig. S1B upper panel), only the 100 kD band was

221

detected. Nevertheless, when proteins were extracted with a buffer containing 6

222

M guanidine chlorhydrate (Gd), a potent ionic chaotrop, the band of 100 kD was

223

replaced by a band of 60 kD (Supplemental Fig. S1B, lower panel). To discard

224

ubiquitin modifications, that are involved in Arabidopsis CDF stability (Imaizumi

225

et al., 2005), extracts were incubated with 1 mM of the proteasome inhibitor

226

MG-132. However, the intensity of the signal of the 100 kD band did not

227

increase or showed any mobility shift (Supplemental Fig. S1B bottom right),

228

probably indicating that the proteasome was not involved in CrDOF stability. So,

229

it is possible that CrDOF establishes tight connections with other(s) protein(s)

230

that control its function. As CDF stability is reduced under blue light in

231

Arabidopsis (Imaizumi et al., 2005), we monitored if CrDOF stability was

232

influenced by light quality. Chlamydomonas cultures were grown under LL, DD,

233

constant blue light (BB) or constant red light (RR), cell extracts were prepared

234

and immunoblots run using CrDOF antibodies. No significant difference in

235

CrDOF accumulation was detected in any light condition (Supplemental Fig.

236

S1C) indicating that light quality has no effect on its posttranslational stability.

237

The presence of CrDOF protein was also monitored under the confocal

238

microscope in an algal line (CrDOFox:YFP, see below) that expressed the

239

Yellow Fluorescent Protein (YFP) fused to CrDOF carboxyl end (Fig. 2E). To

240

identify the cell nucleus we developed a novel method for Chlamydomonas

241

using the nucleic acid-specific dye SYTO®Blue45 (Molecular Probes) in order to

242

visualize nuclear material in vivo without interfering with YFP and chloroplast

243

fluorescence (Methods). SYTO®Blue45 dye in blue a nucleic-acid-rich region of

244

CW15, sited in the flagella pole and encircled by the unique cup-shaped

245

structure of the chloroplast, which we identified as the nucleus (Fig. 2E, upper

246

panel). The merged figure in the lower panel (Fig. 2E) showed that the YFP

247

signal coincided with the blue signal from the dye, indicating that CrDOF:YFP

248

fusion was preferentially localized in the nucleus.

249 250

Modification of CrDOF expression alters Chlamydomonas growth

11 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

251

To clarify the role of CrDOF in Chlamydomonas, algal lines with altered levels of

252

CrDOF mRNA were produced using Gateway technology vectors (Invitrogen)

253

containing constitutive or inducible promoters (Methods). Thus, CrDOF genomic

254

sequence was fused to the pnia2 promoter (CrDOFin) for inducible expression

255

depending on the nitrogen source (Camargo et al., 2007). Figure 3A upper

256

panel left, shows that CrDOF expression was significantly increased in CW15

257

transformed with the CrDOFin construct under inductive (+NO3-) conditions,

258

detecting also an increase in protein levels by Western blots (Fig. 3A lower

259

panel left). CrDOF expression levels increased when CrDOF was fused to YFP

260

in its carboxyl end and expressed under the chimeric promoter HSP70/RbcS2

261

(CrDOFox:YFP) for constitutive expression (Sizova et al., 2001) (Fig. 3A right

262

upper panel). Extracts from these algae presented an extra 120 kD band

263

corresponding to the MM of CrDOF plus the YFP protein (see arrow, Fig. 3A

264

right lower panel). CrDOF expression was significantly reduced employing an

265

artificial microRNA (amiCrDOF, Methods) strategy (Molnar et al., 2009) that was

266

reflected in a decrease in mRNA and protein levels in extracts from the two

267

independent silenced lines (amiCrDOF #1 and #4) (Fig. 3B). CrDOF

268

overexpression from the constitutive and the induced promoters had a strong

269

effect in algal growth (Fig. 3C left). WT and recombinant cultures growth in

270

liquid medium (Sueoka NO3-, Methods) were monitored by measuring

271

chlorophyll content during 4 weeks in LD. Whereas CW15 reached the

272

exponential growth phase in the second week, it was retarded to the third week

273

in the CrDOFox:YFP line and even further in the CrDOFin line. The amiCrDOF

274

#1 and #4 cultures grew also slower than CC-4351 wild type (Fig. 3C right

275

panel). The growing capacity of the lines was further measured in LD and SD,

276

both in rich (TAP NO3-) and normal (Sueoka NO3-) media agar plates, by droplet

277

growth monitoring (Supplemental Fig. S1D). In all conditions, decreasing or

278

increasing CrDOF transcript levels, slowed algal growth, indicating that changes

279

in CrDOF levels affected the growing conditions of the algae.

280

To

assess

the

systemic

effect

of

CrDOF

overexpression

in

281

Chlamydomonas, a transcriptomic approach was used. Chlamydomonas cells

282

(CW15) and two overexpression lines (CrDOFin and CrDOFox:YFP) were

283

grown in Sueoka NO3- media in LD and SD, RNA was extracted at ZT4 and

284

sequenced using the NGS platform Illumina Hiseq2000. After all quality 12 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

285

parameters were verified, a stepwise protocol to analyse the transcriptome,

286

using the Chlamydomonas genome as reference, was run (Methods). Several

287

software tools for read-mapping (Bowtie and TopHat), transcriptome assembly

288

(Cuffmerge and Cufflinks) and differential gene expression analysis (Cuffdiff

289

and CummeRbund) were used. Differentially Expressed Genes (DEGs)

290

between CW15 and the two overexpression lines were identified and Venn 13 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

291

diagrams representing these differences drawn (Supplemental Fig. S1E). The

292

RNAseq data analysis showed that CrDOF overexpression in both lines

293

affected the expression of thousands of genes: 2399 genes induced and 2431

294

repressed in CrDOFin and 3255 induced and 2632 repressed in CrDOFox:YFP

295

in both LD and SD. When we separately analysed the intersection between

296

induced and repressed genes of both overexpression lines compared to WT in

297

LD and SD and performed Venn diagrams (Fig. 3D) a tendency in gene

298

expression, which was day length-dependent, emerged. CrDOF overexpression

299

caused the specific induction of 245 genes in LD and 408 genes in SD while it

300

caused the specific repression on 295 genes in LD and 163 genes in SD.

301

Therefore, the RNAseq data from the two independent overexpression lines

302

indicated that CrDOF would function preferentially as a repressor of gene

303

expression in LD and as an inducer in SD. These data can be further analysed

304

in the webpage: (http://viridiplantae.ibvf.csic.es/crdof_cell_cycle.html).

305 306

CrDOF induces CrCO expression in Chlamydomonas

307

In Arabidopsis and potato, CDFs reduce CO expression and delay flowering

308

(Imaizumi et al., 2005; Fornara et al., 2009; Kloosterman et al., 2013). To show

309

that CrDOF could have a similar role in algae, CrCO expression was monitored

310

in 24h experiments in CW15, CrDOFin and CrDOFox:YFP lines (Fig. 4A). Both

311

in LD and SD CrDOFin and CrDOFox:YFP showed an increase in CrCO

312

expression during the morning and night. This was also observed in

313

experiments in which CrDOFin lines were transferred from not induced

314

(Supplemental Fig. S2A left) to induced conditions (Supplemental Fig. S2A

315

right) for 2 h, resulting in a 3-fold increase in CrCO expression. Constitutive

316

CrDOFox:YFP lines that presented a moderate activation of CrDOF expression,

317

also showed a moderate increase in CrCO expression (Supplemental Fig. S2B).

318

On the contrary, in amiCrDOF #1 and #4 (Fig. 4B), reduced levels of CrDOF

319

produced a reduction in CrCO mRNA levels. Thus, both incrementing and

320

reducing CrDOF expression levels had a parallel effect on CrCO, but rather

321

than inhibiting CO expression as in Arabidopsis, CrDOF behaved as an

322

activator of CrCO expression in Chlamydomonas.

323

As we had observed a 60 kD and a 100 kD CrDOF protein band in

324

Chlamydomonas immunoblots, we wondered if this had a biological 14 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

325

significance, so we performed gel filtration experiments in a Superose 12

326

10/300 GL (GE Healthcare) column attached to an FPLC (GE Healthcare). Total

327

extracts from CW15 or CrDOFin lines in LD and SD conditions, at ZT4 and

328

ZT24 were run in the column, the elution profile separated by size and 15 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

329

monitored for CrDOF presence (Fig. 4C; Supplemental Table S2). In CW15, a

330

protein complex containing CrDOF eluted only in two fractions corresponding to

331

a MM of 464 kD and 314 kD (Supplemental Fig. S2C-D). The 100 kD band was

332

observed in all conditions, but the 60 kD protein was only detected at ZT4 in the

333

fraction with the higher MM complex (Fig. 4C). At ZT24, the 60 kD band was

334

absent in CW15, but was present in the CrDOFin extract in both fractions.

335

Because CrCO expression in WT cells reach its highest peak at ZT4 and in

336

CrDOFin was induced at ZT24, we think that this correlates well with the 60 kD

337

band representing the free, active protein that is able to modify CrCO

338

expression. Moreover, the ratio between CrDOF and partners seems to have an

339

important role in the complex size, as the 60 kD band presence always

340

coincided with the higher MM complex.

341

The capacity to interact with itself was also investigated by bimolecular

342

fluorescence (BiFC) (Supplemental Fig. S3A). A translational fusion of CrDOF

343

to the YFP N-terminal part (YFN-CrDOF) was co-infected by Agrobacterium-

344

mediated infiltration in Nicotiana benthamiana leaves together with CrDOF

345

fused to the carboxyl part of the YFP (YFC-CrDOF). The presence of a strong

346

yellow signal in the nucleus of the co-infiltrated Nicotiana cells showed that

347

CrDOF was able to interact with itself to form dimers or higher MM complexes.

348

To test if CrDOF could bind to CrCO promoter, Agrobacterium-mediated

349

co-infiltration experiments were run. The sequence (846 bp) expanding the

350

intergenic region between CrCO and the upstream gene (Cre06.g278158,

351

Supplemental Fig. S3B) was cloned before the GFP coding sequence in a plant

352

transformation vector (pCrCO:GFP). This construct did not show any

353

expression in Nicotiana cells under the confocal microscope (Fig. 4D panel 1),

354

indicating that no plant factor could bind to the Chlamydomonas CrCO

355

promoter. When a construct overexpressing CrDOF under a plant constitutive

356

promoter (35S:CrDOF) was co-infiltrated with pCrCO:GFP, a marked increase

357

in GFP fluorescence was observed (Fig. 4D panel 2), demonstrating that

358

CrDOF could bind to CrCO promoter. In order to quantify the signal, more than

359

100 nuclei were monitored under the confocal microscope and their

360

fluorescence registered (representing 100% signal). To identify the minimal

361

promoter that still showed CrDOF binding, a 527 bp fragment pCrCO2

362

(underlined in Supplemental Fig. S3B) and a 285 bp fragment pCrCO3 (blue in 16 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

363

Supplemental Fig. S3B) were cloned before the GFP CDS and co-infiltrated

364

with 35S:CrDOF construct in Nicotiana leaves. As observed in Fig. 4D (panels

365

3-4) GFP fluorescence did not decrease, indicating that the minimal promoter

366

pCrCO3 contained all sequences needed for CrDOF binding. In the 285 bp

367

fragment, three canonical DOF AAAG binding sites were found (grey in

368

Supplemental Fig. S3B). The three motives were mutated individually to confer

369

pCrCO3:GFP*; pCrCO3:GFP** and pCrCO3:GFP*** constructs (Supplemental

370

Fig. S3B). Only pCrCO3:GFP* and pCrCO3:GFP** showed a significant

371

decrease in GFP fluorescence when co-infiltrated with 35S:CrDOF, indicating

372

that the motive closer to CrCO initial ATG codon was not involved in CrDOF

373

binding (Fig. 4D panels 5-7). The two elements with a positive binding activity

374

conserved the (A/T)AAAG or (A)CTTT(A) extra nucleotides characteristic of

375

DOF-binding sites (Yanagisawa and Schmidt, 1999) not present in the third site.

376

Therefore, we had identified the binding site sequence of the primitive DOF

377

gene and showed that it recognized the same core region as in higher plants.

378

This constitutes another evidence of the conservation of DOF TFs function

379

during plant diversification.

380 381

CrDOF alters algal cell cycle progression in a photoperiod-dependent

382

manner

383

CrCO

384

Chlamydomonas and thus the capacity to synchronize cell division and growth

385

in particular environmental and photoperiod conditions (Serrano et al., 2009).

386

As CrDOF overexpression and amiRNA constructs showed growth defects and

387

altered CrCO expression, we examined if the cell cycle was also affected in

388

these lines. Cell cycle defects are associated with a change in size due to cell

389

division alterations (Umen and Goodenough, 2001). When we analysed

390

CrDOFin and CrDOFox:YFP cells in LD by flow cytometry (Fig. 5A upper

391

panels) they showed an increase in size compared to CW15, whereas in SD

392

they were smaller. On the contrary, amiCrDOF #1 and #4 lines did not show

393

any significant alteration in cell size in LD but showed a marked decrease in

394

SD. Therefore, cell size in Chlamydomonas was severely affected by changes

395

in CrDOF expression and was influenced by photoperiod.

missexpression

was

shown

to

affect

synchronous

division

17 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

in

396

Cyclins and cyclin-dependent kinases are key proteins involved in cell

397

cycle progression in eukaryotes, and this function is conserved in green

398

microalgae (Bisova et al., 2005). In Chlamydomonas it was shown that cyclin

399

A1 (CrCYCA1) and cyclin-dependent kinase B1 (CrCDKB1) transcript levels

400

increased when CrCO levels increased in SD in order to promote the highly

401

synchronized mitotic divisions at the end of the day characteristic of

402

synchronous growth (Serrano et al., 2009). In order to assess the role of CrDOF

403

in this process we measured the capacity to activate the expression of 18 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

404

CrCYCA1 and CrCDKB1 in LD and SD in the inducible line. CrDOF induction in

405

two independent lines greatly enhanced the expression of CrCYCA1 in SD (Fig.

406

5B left lower panel), while in LD it had the opposite effect (Fig. 5B left upper

407

panel), inhibiting the small peak of expression present in CW15. CrCDKB1

408

expression showed a similar response, as it was significantly enhanced in SD,

409

but was strongly inhibited in LD by CrDOF overexpression (Fig. 5B right).

410

Therefore, CrDOF overexpression altered cell cycle progression by increasing

411

CrCYCA1 and CrCDKB1 transcript levels in SD and decreasing them in LD.

412

Because CrDOF activates CrCO expression, this effect may be associated with

413

the induction in CrCO mRNA levels observed in CrDOFin during the morning

414

(Fig. 4A), but this could not explain the inhibitory effect in LD. To confirm these

415

results we performed a clustering and functional enrichment using the RNAseq

416

data from CW15, CrDOFin and CrDOFox:YFP lines in SD and LD at ZT4. The

417

functional enrichment of genes inhibited by CrDOF overexpression in LD

418

showed that a significant part of the genes were involved in cell cycle

419

progression

420

(http://viridiplantae.ibvf.csic.es/crdof_cell_cycle.html). Therefore, CrDOF inhibits

421

cell cycle progression in LD independently of CrCO and activates its

422

progression in SD by inducing CrCO expression (Fig. 5C). In this scenario, cells

423

would be larger in LD as cell cycle is inhibited during the day (Fig. 5C left), while

424

in SD cells maintain a small size due to the activation of the cell cycle promoted

425

by CrCO (Fig. 5C right) as we had previously shown in the flow cytometer

426

experiments.

including

CrCYCA1

and

CrCDKB1

427 428

CrDOF in Arabidopsis delays flowering through the interaction with CDFs

429

In Arabidopsis, CDFs inhibit CO expression during the morning to avoid FT

430

missexpression outside the evening coincidence window that sets the exact

431

synchronization of flowering by day length. To test the capacity of CrDOF to

432

alter flowering time in Arabidopsis we expressed CrDOF under the control of the

433

constitutive promoter 35S (35S:CrDOF). CrDOF heterologous expression in

434

Arabidopsis delayed flowering in LD, 35S:CrDOF plants flowering with ten

435

leaves more than in Col-0, and had no effect in SD (Fig. 6A; Table I). We then

436

followed CO and FT expression during a 24h course in LD (Fig. 6B) and SD

437

(Supplemental Fig. S4A) in Col-0 and 35S:CrDOF plants. While CO expression 19 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

438

was reduced during the daytime in LD, particularly during the evening

439

(shadowed in Fig. 6B above left), the expression in SD was not affected

440

(Supplemental Fig. S4A above). In 35S:CrDOF plants in LD, FT expression was

441

strongly inhibited during the whole photoperiod (Fig. 6B above right), explaining

20 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

442

the late flowering phenotype in LD, but in SD no effect was observed

443

(Supplemental Fig. S4A below).

444

CDFs work additively to repress CO and FT expression, so that in the

445

quadruple

mutant

446

constitutively high (Fornara et al., 2009). We tested if CrDOF could complement

447

the early flowering phenotype observed in 4cdf (Fig. 6C; Table I). While 4cdf

448

mutant flowered earlier than Col-0 in LD and SD, ectopic expression of CrDOF

449

in 4cdf background had no effect on flowering time. Therefore, CrDOF have no

450

effect on the floral transition in the absence of CDFs. This suggests that CrDOF

451

may interact with the CDFs and enhance their capacity to inhibit CO and FT

452

expression. To test this hypothesis we performed BiFC experiments in

453

Nicotiana (Supplemental Fig. S4B). Both CDF1 and CDF2 were able to interact

454

with each other and CDF2 with itself (Panels 3-4) but they could also interact

455

with CrDOF (Panels 1-2). When we used CrDOF specifc antibodies in Col-0

456

nuclear extracts we could detect several bands around and below 60 kD that

457

could account for Arabidopsis DOF TFs, including CDFs, which present smaller

458

size than other DOFs (Supplemental Fig. S4C left panel). In fact, these signals

459

increased in extracts from ZT0 to ZT8, where the maximum amount of CDFs is

460

supposed to accumulate (Fornara et al., 2009). In 35S:CrDOF plants, the

461

immunodetection of the 60 kD band increased significantly both in col-0 and in

462

the 4cdf mutant background (Supplemental Fig. S4C right panel), due to

463

accumulation of CrDOF, but we could never detect the 100 kD band that we

464

had observed in Chlamydomonas. The exclusive presence of the 60 kD band in

465

the recombinant E. coli cultures (Fig. 2C), the detection of the 60 and 100 kD

466

bands in the gel filtration experiments in Chlamydomonas (Fig. 4C) and the lack

467

of a 100 kD band in Arabidopsis, strongly suggest that the higher MM must be a

468

specific characteristic of algae DOFs and that the active form must be the 60 kD

469

protein. Using a 35S:GFP:CrDOF construct, we could also show the nuclear

470

localization of CrDOF in Arabidopsis (Supplemental Fig. S4D).

cdf1/cdf2/cdf3/cdf5

(4cdf)

CO

transcript

levels

are

471

To further demonstrate the effect of CrDOF over CO and FT expression

472

we infiltrated Nicotiana cells with Agrobacterium carrying constructs expressing

473

GFP under the control of CO (pCO:GFP) and FT (pFT:GFP) promoters (Fig.

474

6D). Both constructs showed a clear nuclear GFP signal when infiltrated alone.

475

When a negative control (AKIN10) was co-transformed with pCO:GFP, no 21 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

476

reduction in GFP fluorescence was observed. As expected, co-transformation of

477

pCO:GFP with a construct ectopically expressing CDF1 produced a dramatic

478

decrease in GFP fluorescence. Ectopic co-expression of CrDOF had a lower,

479

but still significant effect in GFP fluorescence. In a similar way to what was

480

shown in the expression experiments, CrDOF also reduced GFP fluorescence

481

when GFP was driven by the FT promoter (Fig. 6D right), suggesting a role for

482

CrDOF (and CDFs) in FT expression.

483 484

22 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

485

Discussion

486

DOF proteins are present in the genomes of higher plants and unicellular green

487

algae (Chlorophytes) and are absent from all other algal divisions and, in fact,

488

from any other prokaryote or non-green-lineage eukaryote (Moreno-Risueño et

489

al., 2007). This way, we could identify CrDOF sequences in several

490

Chlorophytes orders such as Chamydomonadales (Chlamydomonas and

491

Volvox genres); Trebouxiophyceae (Chlorella and Coccomyxa genres),

492

Mamiellales (Ostreococcus, Bathyococcus and Micromonas genres) and

493

Chlorodendrales (Tetraselmis). CrDOF sequences were absent from other

494

major algal divisions such as the primary endosymbionts Rhodophytes

495

(Cyanidioschyzon) and Glaucophytes (Cyanophora), but also from other

496

secondary endosymbionts such as Euglenophytes (Euglena), Haptophytes

497

(Emiliania), Heteroconts (Phaeodactylum, Thalassiosira and Navicula genres),

498

Cryptophytes (Guillardia), Chloraracniophytes (Bigelowiella) and Dynophytes

499

(Symbiodinium). Therefore, they constitute a good model to study the evolution

500

of specific regulatory pathways in plants. DOFs are found as single copy genes

501

in the genome of Chlorophytes but in Physcomitrella patens we have identified

502

16 genes homologous to CrDOF, while others have identified a family of 19

503

members (Shigyo et al., 2007). In rice and maize around 31 and 10 DOF genes

504

have been identified, respectively, while in Arabidopsis and Populus numbers

505

are even higher: 36 and 61, respectively (Noguero et al., 2013). Therefore, this

506

family has increased from a single copy gene in algae, to a multi-copy family

507

with numerous functions, including the control of photoperiodic flowering, in

508

Angiosperms. Therefore, the study of the ancestral function of the algal DOF

509

was important to understand how different regulatory pathways are controlled,

510

being the photoperiod pathway of special interest.

511

In Arabidopsis, potato and rice, several DOF genes have been classified

512

as CDFs because they are regulated by the clock and at the same time regulate

513

photoperiodic flowering (Imaizumi et al., 2005; Fornara et al., 2009; Li et al.,

514

2009; Kloosterman et al., 2013). In its domain structure, ancient DOF proteins

515

contained a single DOF domain with a NLS in its amino terminal part. This

516

structure is repeatedly found in green microalgae. In the course of evolution, as

517

observed in Physcomitrella DOFs, different domains were incorporated to the 23 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

518

basic pattern that added new regulatory elements to the protein and recruited

519

novel regulators to the pathway (Fig. 1C). As a culmination to this evolutionary

520

process, higher plant CDFs possess an established structure that is the

521

combination of several domains present in Bryophytes, such as the GIGANTEA

522

and FKF1 binding elements, responsible for their complex posttranslational

523

regulation. Therefore, the evolution of DOF factors is a good example of the

524

Innovation-Amplification-Divergence model of evolution by gene duplication

525

described also for the family of COL genes in photosynthetic eukaryotes

526

(Romero-Campero et al., 2013).

527

DOF transcription factors have multiple roles in plants, from the control of

528

nitrogen and carbon metabolic balance, as in maize DOF1 (Yanagisawa, 2000),

529

to the control of the photoperiodic flowering by Arabidopsis CDFs (Imaizumi et

530

al., 2005; Fornara et al., 2009). At the molecular level they function as DNA

531

binding proteins that regulate transcription, either inducing or repressing gene

532

expression (Noguero et al., 2013). In this paper, we have shown that the

533

Chlamydomonas single-copy DOF gene already had this double function. We

534

have provided evidence for gene induction at individual level, such as CrCO or

535

the repression of CrCYCA1 and CrCDKB1 in CrDOF overexpression lines, but

536

we have also provided systemic-level evidence using RNAseq data from WT

537

and CrDOF overexpression lines under different photoperiod conditions. We

538

have further shown that CrDOF preferentially induces or represses gene

539

expression depending on photoperiod, such that it is mainly an activator of gene

540

expression in SD and an inhibitor in LD. Both functions can indeed be recruited

541

to achieve the same goal, such as the regulation of synchronous growth. In

542

Chlamydomonas, and other microalgae, cell division is synchronized to the cell

543

cycle (Bisova and Zachleder, 2014) and this is affected by day length. This way,

544

in LD or LL, with an excess of resources, algae can divide several times and no

545

peak in cell cycle genes is detected. On the contrary, in SD, in stress conditions

546

or nutrients limitation, algae divide once a day and the cell cycle genes peak at

547

dusk (Serrano et al., 2009). Therefore, we can now compare CrDOF

548

overexpression lines data with CrCO transgenic lines data (Serrano et al.,

549

2009), where CrCO positively regulated CrCYCA1 and CrCDKB1 transcript

550

levels in SD conditions. So, we can propose that CrDOF induces CrCO

551

expression in SD at the end of the day promoting the up regulation of cell cycle 24 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

552

factors such as CrCYCA1 and CrCDKB1 (Fig. 7). This peak of expression

553

activates the cell cycle and assures cell division in that particular time, allowing

554

all algae in the culture to divide synchronously. In LD, on the contrary, CrDOF

555

inhibits the expression of cell cycle genes in specific conditions when no

556

synchronous division is necessary, thus allowing indeterminate cell division

557

throughout the day. In fact, in Arabidopsis, the DOF factor OBP1 is involved in

558

the control of cell cycle by inducing the expression of cyclin CYCD3.3 and its

559

overexpression causes the reduction of G1 phase, accelerating growth (Skirycz

560

et al., 2008), suggesting that this algal DOF function is conserved in plants.

561

Our data suggest that a posttranslational regulatory step is necessary to

562

explain CrDOF function, as CrDOF can bind to DNA as a single 60 kD form, but

563

it is mostly detected as a 100 kD band in Chlamydomonas extracts. We

564

propose that the most plausible explanation is that a partner, or more than one,

565

may bind CrDOF in a protein complex. Bound CrDOF would be unable to

566

promote or inhibit gene expression. In the CrDOF induced line this regulation is

567

lost, probably due to an excess of CrDOF in the complex, so that free CrDOF

568

could then activate CrCO expression. The experiments in Nicotiana cells that

569

show that CrDOF can bind directly to CrCO promoter, the absence of the 100 25 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

570

kD band in plant immunodetections and the data from the chromatographic

571

experiments in Chlamydomonas, where the 60 kD appear only when CrDOF

572

induces CrCO expression, suggest that CrDOF is active whenever the 60 kD

573

band is detected. Therefore, we propose that CrCO mRNA levels induction by

574

CrDOF in algae is a direct effect caused by DOF binding to its promoter.

575

In Arabidopsis, CrDOF overexpression delays flowering by reducing CO

576

and FT levels exclusively in LD and not in SD, reflecting a strict photoperiodic

577

response. In our model, as CrDOF cannot alter flowering time in the 4cdf

578

mutant, it is possible that CrDOF could act in a complex with the CDFs,

579

protecting them from the proteasome degradation triggered by GI-FKF, as

580

CrDOF does not present the binding domains for these proteins (Fig. 7 right). In

581

Nicotiana experiments, CrDOF also inhibit CO and FT expression probably

582

through the interaction with the Nicotiana CDF homologs. This association

583

between CDFs and CrDOF could have interesting applications in plant

584

biotechnology to modify flowering time.

585

Conclusions

586

It is interesting to notice that both COL and DOF genes seem to have suffered a

587

similar evolutionary process and often cluster together in gene co-expression

588

networks constructed from different organisms representing different steps of

589

plant evolution (Romero-Campero et al., 2013). Both CO and CDF proteins are

590

part of the photoperiod pathway. It is likely that other partners of this route (i.e.

591

FLOWERING BHLHs [FBHs] [Ito et al., 2012], FKF, COP1…) had followed a

592

similar co-evolution process. If this was the case, and further, could be

593

demonstrated for other routes, it may constitute the basis of a general

594

evolutionary conserved process in which all members of a regulatory pathway

595

(toolkit) evolve together, restrained and connected by the functional links

596

established early in evolution in the original regulatory network (Romero-

597

Campero et al., 2013; Ichihashi et al., 2014; Della Pina et al., 2014). Our

598

demonstration that the DOF-CO module is conserved from algae to higher

599

plants provides direct evidence that this effect may be indeed be taking place.

600

The study of these conserved toolkits throughout the plant evolutionary lineage

601

may provide important hints to understand the processes that allowed the

602

differentiation of higher plants from simple unicellular algae. 26 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

603 604

Materials and methods

605

Biological Material and Growth Conditions

606

Chlamydomonas reinhardtii wild types CW15 (Davies and Plaskitt, 1971), CC-

607

4351 [From René Matagne, University of Liège, Belgium via Michael Schroda

608

(Max Planck Institute of Molecular Plant Physiology)] and transgenic lines, were

609

grown in flasks with minimum Sueoka media (Sueoka et al., 1967) or rich TAP

610

(Tris Acetate Phosphate) media (Harris, 1989) in LD or SD conditions at 50 µE

611

light intensity with temperatures ranging from 22°C (day) to 18°C (night). For

612

induction of nia2 promoter, algal cells grown to exponential phase in Sueoka

613

media supplemented with ammonium, were harvested by centrifugation (4 mins

614

at 3,200 g) and suspended in Sueoka media supplemented with nitrate.

615

Arabidopsis thaliana plants were grown in soil for phenotypic analysis or

616

in MS agar media supplemented with 1% (w/v) sucrose plates for QPCR

617

assays. In both cases, seeds were previously incubated during 4 days at 4°C

618

before sowing. Plants were grown in humidity controlled (80%) chambers with

619

100 µE light intensity and temperatures ranging from 22°C (day) to 18°C (night).

620

Different photoperiods were used: 16 hr light / 8 hr dark (long day, LD) and 8 hr

621

light / 16 hr dark (short day, SD).

622

Cloning and analysis of CrDOF

623

The complete CrDOF ORF was obtained from the cDNA Collection of Kazusa

624

DNA Research Centre (http://www.kazusa.or.jp). Sequencing of the cDNA

625

showed 100% identity with predicted locus Cre12.g521150 (position: 4425725

626

to 4430320) from Phytozome v.9.1. CrDOF deduced protein was 625 amino-

627

acid long and contained a DOF-type zinc finger domain (Moreno-Risueño et al.,

628

2007; Yanagisawa, 1997).

629

Phylogenetic analysis and identification of conserved motifs

630

Evolutionary relationships among DOF proteins from the green algae

631

Chlamydomonas reinhardtii (Cre), Volvox carteri (Vc), Ostreococcus tauri (Ot)

632

and Micromonas pusilla (Mp); the Bryophyte Physcomitrella patens (Pp); the

633

monocots Oriza sativa (Os) and Zea mays (Zm) and the dicots Arabidopsis

634

thaliana (At) and Populus trichocarpa (Pt) were analyzed using CrDOF full

635

length amino-acid sequence as a query to detect protein homologs in

27 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

636

Phytozome v9.1 database. All sequences were aligned using the program

637

MUSCLE (Edgar, 2004) and a phylogenetic tree was generated applying the

638

Neighbor-Joining algorithm (Saitou and Nei, 1987) with the JTT + G 0.54 as a

639

substitution model (Jones et al., 1992). The number of bootstrap replicates was

640

500. Phylogenetic analyses were conducted with MEGA5 (Tamura et al., 2011).

641

Accession numbers of sequences used in the alignment are shown in

642

Supplemental Table S1. A domain analysis of the same 84 DOF proteins in the

643

phylogenetic

644

meme.nbcr.net/meme/) (Bailey et al., 2009). The motifs domain were

645

characterized

646

(http://www.ncbi.nlm.nih. gov/Structure/cdd/wrpsb.cgi). Further identification of

647

the GI and FKF1 binding domains was done as in Kloosterman et al., 2013.

648

Phenotypic analysis

649

Growth curve and drop test. Algal cultures were grown to stationary phase in LD

650

conditions and then diluted to 0.5 µg mL-1 of chlorophyll with new media.

651

Culture growth was monitored by chlorophyll measures once a week for one

652

month. For drop tests, the culture in stationary phase was diluted to 4 to 5x106

653

cells mL-1 and 3 consecutive 1/10 dilutions were made. 3 µl drops were grown

654

in agar plates containing Sueoka NO3- or TAP NO3- media and incubated in LD

655

or SD.

656

Flow cytometry. WT and transgenic lines, samples were taken at ZT4 in LD and

657

SD conditions in Sueoka media supplemented with NO3- and analyzed on a BD

658

influx™ Cell Sorter flow cytometer (Becton Dickinson, Canada) equipped with a

659

laser with a 488 nm line at 200 mW and analysed with “BD FACS sortware”.

660

Analysis of flowering time. Flowering time was analyzed in LD and SD

661

conditions in controlled environment cabinets by scoring the number of rosette

662

(excluding cotyledons) and cauline leaves. Data are media of at least 30

663

individual ± standard error of the mean (SEM).

664

Algal and plant transformation

665

Chlamydomonas. For Chlamydomonas nuclear transformation, a modification of

666

the electroporation protocol from (Shimogawara et al., 1998) was used. In brief,

667

about 1 to 5x106 cells mL-1 were harvested by centrifugation at 1,500 g and

668

suspended in TAP medium supplemented with 60 mM sucrose in 1/100 the

669

original volume. 250 µl of cell suspension and 1 µg of the plasmid were placed

tree

was

using

done the

using

the

Conserved

MEME Domain

software Search

28 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

(http:// Service

670

into 0.4 cm gapped cuvette and incubated for 30 mins on ice. After a 800 v

671

pulse (15 µF and no shunt resistance) in an electroporator, cells were

672

transferred to 10 mL TAP media supplemented with 60 mM of sucrose and

673

incubated overnight with moderate shaking and light. Cells were harvested by

674

centrifugation, resuspended in 700 µl TAP media and spread into selective TAP

675

agar plates containing hygromicine 25 µg mL-1. Colonies started to appear after

676

1 week. For the nitrate inducible CrDOF expression vector construction, the

677

Gateway modified version of the pnia2 vector (Serrano et al., 2009) was used.

678

For the constitutive CrDOF vector expression construction, direct gene

679

synthesis from Invitrogen was used. In-silico sequence of Hsp70A/RbcS2

680

constitutive promoter, Gatteway casette B and 3`UTR Rubisco terminator were

681

performed and restriction enzymes sites present in gateway cassette and

682

terminator were changed without modifying the codon use. YFP and

683

hygromicine resistence (HYG) were cloned into EcoRV/BamHI and KpnI

684

restriction sites, respectively. YFP was adapted to the Chlamydomonas codon

685

preference (Matsuo and Ishiura, 2010).

686

Aradidopsis. For Arabidopsis Col-0 or 4cdfs transformation, the Agrobacterium-

687

mediated floral dip protocol was used. CrDOF cDNA was amplified by PCR (5'-

688

GGGGACAAGTTTGTACAAAAAAGCAGGCTTCGAAGGAGATAGAACCATGG

689

TAGACGGTGGTTCG-3'

690

GGGGACCACTTTGTACAAGAAAGCTGGGTCTCACCTAGCACCCGAGTAAG

691

C-3'

692

GGGGACCACTTTGTACAAGAAAGCTGGGTCCCTAGCACCCGAGTAAGCGG

693

C-3' without STOP codon were used for carboxyl-terminal fusions) and cloned

694

into pDONOR 207 plasmid. Finally, CrDOF was cloned in pEG100 vector to

695

obtain the 35S:CrDOF construct and in pMDC43 vector to obtain the

696

35S:GFP:CrDOF construct. Arabidopsis Col-0 plants were transformed with

697

these constructs and plants selected for antibiotic resistance. 4cdfs mutant

698

plants were transformed with 35S:CrDOF construct and positive plants selected

699

by genome PCR tests using 5'- ACTGCAAGACGTGTCAGCG-3' and 5'-

700

GCCAACTCACTGTTGAACTGC-3' primers.

701

RNA techniques

702

Expression analysis by QPCR. Total RNA from Chlamydomonas cells (20 mL of

703

an exponential phase curve) or Arabidopsis seedlings (0.1 g leave tissue) were

with

and STOP

codon

5'and

29 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

5'-

704

isolated by the TRIZOL (Invitrogen) procedure following the recommendations

705

of the manufacturer. The final RNA sample was suspended in 30 µl DEPC

706

water, quantified in a ND-1000 Spectrophotometer (Nanodrop) and stored at -

707

80°C.

708

Quantitec®Reverse kit (Qiagen), diluted to a final concentration of 10 ng µl-1

709

and stored at -20°C until QPCR was performed. Specific primers to amplify the

710

3´region of each gene analysed, plus CrTUB and UBQ-10 (Supplemental Table

711

S3) as housekeeping genes, were designed with Oligo analyzer program

712

(Integrated DNA technologies, www.idtdna.com). QPCR was performed with a

713

Multicolor Real-Time PCR Detection System iQTM5 (Bio-Rad) in 10 µl

714

reactions: primer concentration 0.2 µM, 10 ng cDNA and 5 µl SensiFAST TM

715

SYBR & Fluorescein Kit (Bioline). Each sample was measured by triplicate. The

716

QPCR program consisted in i) 1 cycle (95°C, 2 min); ii) 40 cycle (95°C , 5 s;

717

60°C, 10 s and 72°C, 6 s) iii) 1 cycle (72°C, 6 s). Fluorescence was measured

718

at the end of each extension step and the melting curve was performed

719

between 55°C and 95°C. The initial concentration of candidate and reference

720

gene was calculated by means of LingRegPCR software version 11.0 (Ruijter et

721

al., 2009). Normalized data was calculated dividing the average of three

722

replicates of each sample of the candidate and reference genes.

723

Transcriptomic analysis by RNAseq. The experimental design consisted of two

724

replicates for each genotype (CrDOFin, CrDOFox:GFP and wild type CW15)

725

and growth condition (LD and SD). RNA was extracted from each sample using

726

Trizol. Library preparation was carried out following the manufacturer’s

727

recommendations. Sequencing of RNA libraries was performed with the Illumina

728

HiSeq 2000 sequencer, yielding approximately 40 million 50 bp long reads for

729

each sample. The software package FastQC (Andrews) was used for quality

730

control. All sequencing samples were of high quality and no pre-processing of

731

the reads was required to remove low quality reads or read fragments. The

732

Chlamydomonas reference genome version 5.3 (Merchant et al., 2007) and

733

annotation were downloaded from the Phytozome database (Goodstein et al.,

734

2012). Mapping of reads to the reference genome, transcript assembly and

735

differential expression was performed with the software tools Bowtie, Tophat

736

and Cufflinks (Trapnell et al., 2012) using default parameters. The R package

737

from Bioconductor CummeRbund (Goff et al., 2012) was used for subsequent

1

µg

RNA

was

used

to

synthesize

cDNA

employing

30 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

the

738

analysis and graphical representation of the results. DEGs were selected as

739

those exhibiting an expression fold change greater than 2 when compared to

740

the wild type CW15 and a pValue < 0.05. Venn diagrams comparing the

741

different sets

742

(http://omics.pnl.gov/software/venn-diagram-plotter). The RNAseq raw data

743

generated in this study are publicly available from the European Nucleotide

744

Archive

745

(http://www.ebi.ac.uk/ena/data/view/PRJEB6682).

746

Protein and immunological techniques

747

Purification of recombinant CrDOF. CrDOF cDNA fragment was amplified by

748

PCR using specific primers designed with NdeI/BamHI restriction enzymes

749

adaptors

750

cccggatccTCACCTAGCACCCGAGTAAGC-3'). CrDOF was cloned into pET19b

751

vector, which incorporates an amino terminal Histidine tail (His-CrDOF) and

752

transformed into E. coli BL21 strain. 10 mL of an overnight culture of BL21

753

transformed with His-CrDOF was added to 1 L of LB medium containing 100 µg

754

mL-1 ampicillin and grown at 30ºC until the absorbance at 600 nm reached 0.7.

755

Recombinant gene expression was induced by addition of 1 mM IPTG and

756

incubation during 4 h at 30°C. After centrifugation, the pellet was resuspended

757

in lysis buffer (Tris-HCl 50 mM pH 8; KCl 50 mM; EDTA 1 mM pH 8; DTT 1 mM;

758

MgCl2 10 mM; Glicerol 10% (v/v)) to 3 mL g-1 of cells and broken by sonic

759

disruption. After centrifugation, the pellet was solubilized with buffer IB (Tris-HCl

760

50 mM pH 8; EDTA 10 mM pH 8; Triton X-100 0.5% (v/v); PMSF 0.1 mM),

761

washed twice in the same buffer and once with TALON K buffer (phosphate

762

buffer 0.05 M pH8; KCl 0.3 M). Between each wash, the pellet was sonicated

763

and centrifuged. Finally, the pellet was suspended in TALON buffer including 6

764

M Guanidine chlorhydrate and incubated for two hours. After this period, the

765

protein extract was disrupted by sonication, centrifuged and the supernatant

766

was used to purify CrDOF employing BD TALONTM resin following the

767

recommendations of the manufacturer in denaturing conditions (presence of 6

768

M guanidine chlorhydrate). After elution, proteins were precipitated with ethanol

769

100% (v/v) for 30 minutes at 4°C. The pellet was washed twice (0.3 M

770

guanidine chlorhydrate in 95% (v/v) ethanol and ethanol 90% (v/v) respectively).

771

Finally, the pellet was suspended in SDS-PAGE buffer and separated by

of

database

(5'-

DEGs

were generated

identified

with

with

accession

Venn

Diagram

number

gggcatATGGTAGACGGTGGTTCGCGTG

Plotter

PRJEB6682

-3,

31 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

5'-

772

electrophoresis in 10% (w/v) polyacrylamide gels. 1 mg of purified protein was

773

used to produce antibodies.

774

Protein extraction and immunodetections. Cells from 20 mL of exponentially

775

growing Chlamydomonas cultures were centrifuged and suspended in 3 mL per

776

g of extraction buffer: Tris-HCl 50 mM pH 7.5; PMSF 1mM; DTT 1 mM;

777

Quimostatine, Antipaine and Leupectine at 5 µg mL-1; β-Mercaptoethanol 10

778

mM and 1/1000 SIGMA plant protease inhibitor cocktail. Algae were broken by

779

two cycles of slow freezing to −80°C followed by thawing to room temperature.

780

Cell extracts were centrifuged at 15,000 g for 30 minutes at 4°C and

781

supernatant collected. Protein content was estimated by the Bradford method

782

(Bradford, 1976) using ovalbumin as a standard. Samples were run on SDS-

783

PAGE 10% (w/v) polyacrylamide gels and transferred to nitrocellulose

784

membranes using a Trans-Blot®TurboTMTransfer System (BioRad) according to

785

the manufacturer´s instructions. Western blot analysis was performed using

786

specific antibodies raised against CrDOF recombinant protein. To disaggregate

787

CrDOF complex, total protein was processed by different treatments: Urea 8 M,

788

DTT 0.1 M, Urea 8 M with DTT 0.1 M and Guanidine clorhydrate 6 M.

789

Cromatography. Total protein from CW15 and CrDOFin cultures grown in

790

Sueoka suplemented with NO3- at LD and SD was extracted as above. 1 mg of

791

protein was separated in a Superose 12 10/300 GL column (GE Healthcare). 1

792

mL fractions were collected, precipitated with TCA 10% (w/v) and run in SDS-

793

PAGE gels. Western blot was performed using anti-CrDOF antibodies. Protein

794

complex MM was calculated using a Gel Filtration Markers Kit (SIGMA-

795

MWGF1000) (Supplemental Fig. S2D).

796

Nuclear and protein extraction. Nuclear isolation of Col-0 and 35S:CrDOF (ZT0,

797

ZT8) were done as described in (Lazaro et al., 2012). Protein extraction was

798

carried out in native conditions using a high salt buffer (KCl 1.6 M; HEPES 50

799

mM pH 8; MgCl2 3 mM; DTT 5 mM; glycerol 1% (v/v); PMSF 1 mM) and run in

800

SDS-PAGE gels. Western blot was performed using anti-CrDOF antibodies.

801

Transitory expression in Nicotiana benthamiana

802

BiFC experiments. To verify CrDOF protein interactions in vivo, CrDOF, CDF1

803

and CDF2 complete ORFs were cloned in pYFN43 and pYFC43 to produce

804

fusions to the YTP N-terminal part (YFN-CrDOF, YFN-CDF1 and YFN-CDF2)

805

as well as to the YFP C-terminal part (YFC-CrDOF, YFC-CDF1, YFC-CDF2). 32 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

806

Specific

primers

were

used

for

each

gene

(CrDOF:

807

GGGGACAAGTTTGTACAAAAAAGCAGGCTTCGAAGGAGATAGAACCATGGTAGACGGTGG

808

TTCG-3', 5'-GGGGACCACTTTGTACAAGAAAGCTGGGTCCCTAGCACCCGAGTAAGCGGC-

809

3';

810

GGGGACAAGTTTGTACAAAAAAGCAGGCTTCGAAGGAGATAGAACCATGCTGGAAACTAAA

811

GATCCTGCG-3',

812

GGGGACCACTTTGTACAAGAAAGCTGGGTCTCACATCTGCTCATGGAAATTGAT-3';

813

CDF2:5'GGGGACAAGTTTGTACAAAAAAGCAGGCTTCGAAGGAGATAGAACCATGGCTGAT

814

CCGGCGATTAAGCTC-3',

815

GGGGACCACTTTGTACAAGAAAGCTGGGTCCTATGAGCTCTCATGGAAGTTTGC-3').

816

These constructs were introduced into Agrobacterium tumefaciens strain

817

GVG3101 pmp90. 4-week old Nicotiana plants were inoculated with the

818

following combinations: YFN-CrDOF and YFC-CrDOF, YFN-CrDOF and YFC-

819

CDF1, YFN-CrDOF and YFC-CDF2, YFN-CDF2 and YFC-CDF2. As negative

820

controls, pairs of YFC-CrDOF, YFC-CDF1 and YFC-CDF2 with YFN-AKIN10

821

were used. As positive controls, amino and carboxy parts of AKIN10 were used,

822

following protocols previously described (Voinnet et al., 2003). Fluorescent

823

interactions were visualized under a confocal microscope Leica TCS

824

SP2/DMRE using an excitation wavelength of 514 nm.

825

CrDOF binding assays. In order to assess if CrDOF was able to bind to CrCO,

826

CO and FT promoters, a similar transitory expression experiment in N.

827

benthamiana leaves was used. In this experiments, 864 nt from CrCO promoter

828

(5'-

829

GGGGACAAGTTTGTACAAAAAAGCAGGCTTCACCATGGTGCAGGCATGCCGTGGCTCGGC

830

AAG-3',

831

3');

832

GGGGACAAGTTTGTACAAAAAAGCAGGCTTCACCATGGTAGTATAGAGTATCATCATAAAC

833

CC-3' 5'-GGGGACCACTTTGTACAAGAAAGCTGGGTCAATAACTCAGATGTAGTAAGTTTG-

834

3')

835

GGGGACAAGTTTGTACAAAAAAGCAGGCTTCACCATGGATTATATGTATAGATAGATTACC

836

G-3',

837

promoters were cloned first in pDONOR207 and then in pMDC110, which

838

generates carboxyl-terminal fusions to GFP (pCrCO:GFP, pCO:GFP and

839

pFT:GFP). N. benthamiana leaves were inoculated with this constructs, while

840

other set of N.benthamiana leaves were co-inoculated with this constructs and

5'-

5'-

CDF1:

5'-

5'-

5'-GGGGACCACTTTGTACAAGAAAGCTGGGTCTCTCGCGTATAGAGGTGTGGT1

and

kb

from

1

kb

AtCO

from

(5'-

AtFT(5'-

5'-GGGGACCACTTTGTACAAGAAAGCTGGGTCCTTTGATCTTGAACAAACAGG-3')

33 Downloaded from www.plantphysiol.org on August 31, 2015 - Published by www.plant.org Copyright © 2015 American Society of Plant Biologists. All rights reserved.

841

35S:CrDOF. GFP was visualized at 488 nm wavelength. Three independent

842

inoculations were made for each studied condition. For each inoculation, three

843

leaves fragments of the same size were analysed (total of 9 leaves fragments in

844

three biological replicas) in which, eleven or twelve randomly picked nuclei with

845

positive signal were chosen and GFP fluorescence measured using MBF

846

ImageJ. Therefore, signals from more than 100 nuclei were quantified per

847

condition. Two versions of pCrCO including 527 nt (pCrCO2) and 285 nt

848

(pCrCO3) were constructed. Mutations at the binding sites were generated with

849

“Muta-directTM Site directed Mutagenesis” kit from iNtRON Biotechnology

850

following the instructions of the manufacturer.

851

Chlamydomonas Microscopy

852

CW15 and transgenic lines (CrDOFox:YFP) were observed under the confocal

853

microscope (Leica TCS SP2/DMRE) for nuclear localization experiments. To

854

visualize Chlamydomonas nuclei, a SYTO®Blue45 Fluorescent Nucleic Acid

855

Stains (Molecular Probes) was used. Algae were grown in SD conditions in

856

Sueoka media supplemented with NO3- 10 mM until lag phase (3-4 µg mL-1 of

857

chlorophyll). 1 mL was collected by centrifugation (4 mins, 5,500 g) and

858

suspended in 1 mL TBS. 1 µl SYTO®Blue45 and 1 and 5 µl Triton 10% (v/v) for

859

CW15 cells and transgenic lines respectively, were added. After incubation for

860

10 minutes, cells were centrifuged and suspended in 100 µl of the same buffer.

861

Finally, 3µL of cells were mixed with 10 µl of 1.2% (w/v) low point fusion

862

agarose at 30°C. The wavelengths used were 514 nm for YFP and 458 nm for

863

SYTO®Blue 45.

864

Statistical Analysis

865

The statistical data presented in the figures and table are marked with asterisk

866

and are the mean (±SEM) of three biological experiments or two biological

867

experiments (±SE). The statisticals significance between the means of the

868

different samples were calculated using a two-tailed Student´s t-test.

869

Differences observed were considered statistically significant with a pValue