SHORT VEGETATIVE PHASE up-regulates TEMPRANILLO2 floral repressor at low ambient temperatures

1 2 3

Running Head: SVP up-regulates TEM2 at low ambient temperatures

4

Genomics, CSIC-IRTA-UAB-UB, Campus UAB, Bellaterra (Cerdanyola del Vallès)

5

08193 BARCELONA, Spain. Phone: (34) 93 563 6600 e-mail:

6

[email protected]

Corresponding Author: Soraya Pelaz. Centre for Research in Agricultural

7 8 9

Research Area: Genes, Development and Evolution

1

10

SHORT VEGETATIVE PHASE up-regulates TEMPRANILLO2 floral

11

repressor at low ambient temperatures

12 13

Esther Marín-González, Luis Matías-Hernández, Andrea E. Aguilar-Jaramillo, Jeong

14

Hwan Lee, Ji Hoon Ahn, Paula Suárez-López and Soraya Pelaz*

15

Centre for Research in Agricultural Genomics, CSIC-IRTA-UAB-UB, Campus UAB,

16

Bellaterra (Cerdanyola del Vallès), 08193 Barcelona, Spain (E.M.-G., L.M.-H., A.E.A.-

17

J., P.S.-L., S.P.). Creative Research Initiatives; Department of Life Sciences; Korea

18

University, Seoul, South Korea. (J.H.L., J.H.A.). ICREA (Institució Catalana de

19

Recerca i Estudis Avançats), Barcelona, Spain (S.P.)

20

*Address correspondence: [email protected]

21 22

SUMMARY

23

The functional characterization of two Arabidopsis floral repressors unravels their role

24

and regulation at low ambient temperatures.

25

2

26

Financial Source:

27

This work was supported by a MINECO/FEDER grant (BFU2012-33746). S.P.’s

28

research group has been recognized as a Consolidated Research Group by the Catalan

29

Government (2014 SGR 1406). E. M-G was a recipient of an FPI fellowship from the

30

Spanish Government. A.E.A-J is a predoctoral fellow of the investigator training

31

program (FI) from the Catalonia Government. J.H.L. and J.H.A were supported by a

32

National Research Foundation of Korea grant funded by the Korea government

33

(Ministry of Science, ICT, and Future Planning) (2008-0061988).

34 35

Corresponding Author: Soraya Pelaz. [email protected]

36

3

37

ABSTRACT

38

Plants integrate day length and ambient temperature to determine the optimal timing for

39

developmental transitions. In Arabidopsis thaliana, the floral integrator FLOWERING

40

LOCUS T (FT) and its closest homologue TWIN SISTER OF FT (TSF) promote

41

flowering in response to their activator CONSTANS (CO), under long-day (LD)

42

inductive conditions. Low ambient temperature (16 ºC) delays flowering, even under

43

inductive photoperiods, through repression of FT, revealing the importance of floral

44

repressors acting at low temperatures. Previously we have reported that the floral

45

repressors TEMPRANILLO (TEM1 and TEM2) control flowering time through direct

46

regulation of FT at 22 ºC. Here we show that tem mutants are less sensitive than the

47

wild type to changes in ambient growth temperature, indicating that TEM genes may

48

play a role in floral repression at 16 ºC. Moreover, we have found that TEM2 directly

49

represses the expression of FT and TSF at 16 ºC. In addition, the floral repressor

50

SHORT VEGETATIVE PHASE (SVP) directly regulates TEM2, but not TEM1,

51

expression at 16 °C. Flowering time analyses of svp tem mutants indicate that TEM may

52

act in the same genetic pathway as SVP to repress flowering at 22º C, but SVP and

53

TEM are partially independent at 16 ºC. Thus, TEM2 partially mediates the

54

temperature-dependent function of SVP at low temperatures. Taken together, our results

55

indicate that TEM genes are also able to repress flowering at low ambient temperatures

56

under inductive LD conditions.

57

4

58 59

INTRODUCTION Plants constantly monitor environmental and endogenous signals to control their growth

60

and adjust developmental responses to daily and seasonal cues (Penfield, 2008). During

61

the juvenile phase, plants are not competent to flower; they are insensitive to inductive

62

environmental factors such as favourable conditions of day length or temperature. The

63

transition to the adult phase permits to reach the competence to respond to those signals,

64

which is essential to trigger flowering during the reproductive phase (Bergonzi and

65

Albani, 2011; Huijser and Schmid, 2011). Consequently, the control of flowering time

66

is a key determinant of reproductive success and plays an essential role in plant

67

adaptation to seasons and geography.

68

Flowering time is controlled by an intricate network of inter-dependent genetic

69

pathways that monitor and respond to both endogenous and environmental signals.

70

These pathways include age, photoperiod and light quality, gibberellin, thermosensory

71

(ambient temperature), vernalization and autonomous pathways (Fornara et al., 2010;

72

Srikanth and Schmid, 2011). In Arabidopsis thaliana, it is well documented the

73

noteworthy regulation of the timing of flowering by day length, or photoperiod, and

74

temperature (recently reviewed in Andrés and Coupland, 2012; Song et al., 2013; Chew

75

et al., 2014; Romera-Branchat et al., 2014). However, in contrast to the finely described

76

photoperiod and light quality pathways, the nature of the primary perception of

77

temperature and the molecular characterization of its signalling remains limited

78

(McClung and Davis, 2010).

79

Lately several studies have reported how changes in ambient temperature, defined as the

80

physiological non-stressful temperature range of a given species, modulate many

81

processes in plant development and, in particular, how they affect flowering time

82

(reviewed in Wigge, 2013; Capovilla et al., 2014). Genetic analyses unravelled the

83

existence of the ambient temperature pathway that mediates temperature responses in A.

84

thaliana (Blázquez et al., 2003; Balasubramanian et al., 2006; Lee et al., 2007; Kumar

85

et al., 2012). It has been described that a slight decrease from 23 ºC to 16 ºC is

86

sufficient to cause a remarkable delay in flowering even under inductive long-day (LD)

87

photoperiod (Blázquez et al., 2003). Temperature-dependent differences in flowering

88

time are controlled by multiple factors that mainly affect the expression levels of one of

89

the key floral activators, FT gene (Kardailsky et al., 1999; Kobayashi et al., 1999). Low

90

ambient temperatures reduce the expression of FT even though this decrease is not due

91

to changes in its transcriptional activator CONSTANS (CO), which reveals the

5

92

importance of floral repressors controlling flowering time under low ambient

93

temperatures (16 ºC) (Blázquez et al., 2003; Lee et al., 2013). However, the levels of

94

TWIN SISTER OF FT (TSF), the closest homologue of FT, are similar at both

95

temperatures resulting in a higher expression of TSF than FT at 16 ºC (Blázquez et al.,

96

2003; Lee et al., 2012; Lee et al., 2013). Although TSF plays a secondary but redundant

97

role, FT and TSF act as floral pathway integrators (Kardailsky et al., 1999; Kobayashi et

98

al., 1999; Yamaguchi et al., 2005; Jang et al., 2009) and their main function is the

99

positive promotion of photoperiodic flowering (Kardailsky et al., 1999; Kobayashi et al.,

100

1999) with TSF playing a secondary but redundant role. Similar relation is also

101

observed at 16 ºC (Kim et al., 2013; Lee et al., 2013) and flowering of the double

102

mutant ft tsf is insensitive to ambient temperature changes (Kim et al., 2013), which

103

indicates that FT and TSF play an important role in the regulation of ambient

104

temperature-responsive flowering (Kim et al., 2013).

105

Previously, the TEMPRANILLO (TEM) genes were identified as main players in the

106

control of flowering time at 22 ºC and they were shown to directly repress FT

107

(Castillejo and Pelaz, 2008).

108

Recent genome-wide analysis has identified TEM1 and TEM2 as direct targets of the

109

MADS-box transcription factor SHORT VEGETATIVE PHASE (SVP) under LD at 22

110

ºC (Tao et al., 2012). TEM1 and TEM2 are expressed at low levels in svp-41 plants and

111

high levels in SVP-overexpressing plants compared to wild-type plants, indicating a

112

positive regulation of TEMs by SVP, more evident on TEM2 than on TEM1 (Tao et al.,

113

2012). Interestingly, previous genetic studies identified svp mutants as insensitive to a

114

wide range of ambient temperature changes (5 ºC to 27 ºC) (Lee et al., 2007, Lee et al.,

115

2013). Thus, svp mutants show an early flowering phenotype, producing almost the

116

same number of leaves at flowering at all temperatures (Lee et al., 2013). Similar to

117

TEM, SVP delays flowering by direct repression of FT (Lee et al., 2007; Li et al., 2008).

118

Moreover, SVP represses TSF in the vascular tissue of leaves and plays an antagonistic

119

role

120

OVEREXPRESSION OF CONSTANS 1 (SOC1), in the meristem (Li et al., 2008; Jang

121

et al., 2009).

122

Here, we characterised the role of TEM genes as repressors of flowering at moderate

123

low ambient temperature of 16 ºC under LD conditions. We show that TEM genes act as

124

floral repressors at 16 ºC under LD conditions by regulating both FT and TSF

125

expression. Furthermore, we show that SVP specifically regulates TEM2 at 16 ºC to

with

other

MADS-box

transcription

factor,

SUPPRESSOR

OF

6

126

repress flowering under LD conditions. Therefore, our results provide additional

127

information regarding SVP and TEM genetic relation at low temperatures.

128 129

7

130

RESULTS

131

tem mutants are early flowering at 16 ºC but still sensitive to low temperature

132

To study the function of TEM1 and TEM2 as floral repressors at low ambient

133

temperature, we first analysed the flowering phenotype of the loss-of-function double

134

mutant tem1 tem2 at 16 ºC in comparison with 22 ºC under LD conditions. tem1 tem2

135

showed early flowering at 22 ºC, in agreement with previous results (Castillejo and

136

Pelaz, 2008; Osnato et al., 2012), but also at 16 ºC showed a bigger difference with

137

wild-type plants at 16 ºC than at 22 ºC. However, tem1 tem2 plants grown at 16 ºC

138

produced more leaves than at 22 ºC (Fig. 1 and Supplemental Table S1) indicating that

139

this double mutant is still thermo-sensitive but less sensitive than wild-type plants,

140

which produced double number of leaves at 16 ºC than at 22 ºC. In contrast, svp mutants,

141

described as insensitive to temperature (Lee et al., 2007), flowered with a similar

142

number of leaves at both temperatures. We found that tem1 tem2 plants flowered with

143

slightly fewer leaves than svp plants at 22 ºC, although this difference was not

144

statistically significant; whereas svp plants were clearly earlier than tem1 tem2 at 16 ºC

145

(Fig. 1A and Supplemental Table S1). tem1 tem2 plants were also earlier than wild-type

146

plants in terms of the number of days to flowering at both temperatures (Fig. 1B).

147

Interestingly, tem1 tem2 plants grown at 16 ºC flowered with a similar number of leaves

148

and days to wild-type plants grown at 22 ºC. These flowering time data are directly

149

correlated with the FT expression levels observed in wild-type, tem1 tem2 and svp

150

mutant plants at different temperatures (Fig. 1C; Fig. 1D and Supplemental Fig. S1). In

151

tem1 tem2 plants, FT expression was slightly higher than in svp mutants and clearly up-

152

regulated when compared with wild-type plants at 22 ºC (Fig. 1C). At 16 ºC FT levels

153

of tem1 tem2 plants exhibited a clear increase compared with wild-type levels, but this

154

increase was lower than in svp plants (Fig. 1D). Both results clearly correlated with the

155

flowering time of those plants at both temperatures. Interestingly, this similar flowering

156

time phenotype observed in wild-type plants grown at 22 ºC and tem1 tem2 plants

157

grown at 16 ºC was associated with similar FT levels.

158

All these results indicate that TEM genes have a role in the control of flowering time at

159

low temperature.

160 161

Low temperature keeps high TEM levels before floral transition

162

To characterise the response of TEM1 and TEM2 to low temperatures, we first

163

performed diurnal analyses in 9-day-old wild-type plants grown under LD conditions at 8

164

16 ºC and 22 ºC. We harvested samples every 4h for 24h and added an extra point of

165

collection at ZT18, during the peak of TEM1 and TEM2 expression. Our results indicate

166

that TEM mRNA levels were not affected by low temperature at this stage, as TEM

167

daily oscillation showed a similar pattern at both temperatures (Supplemental Fig. S2). 9

168

In particular, TEM1 exhibited a clear peak of expression at ZT18 as previously

169

described (Osnato et al., 2012), and a small peak during the day between ZT8 and ZT12

170

at 22 ºC and also at 16 ºC; TEM2 showed high levels during the night until the

171

beginning of the day, when it was gradually reduced. Due to the importance of TEM as

172

repressors of flowering time genes along development (Castillejo and Pelaz, 2008;

173

Osnato et al., 2012), we then analysed TEM expression pattern later in development

174

under LD conditions. As we already knew, under LD at 22 ºC, TEM genes showed high

175

expression levels during early stages of development, which prevents a precocious

176

activation of FT and a consequent early flowering (Castillejo and Pelaz, 2008). After

177

that, there is a gradual decline of their levels until TEMs reach their minimum

178

expression around day 12th, when FT activation takes place (Supplemental Fig. S3).

179

To test our hypothesis of a possible thermal regulation, we compared TEM expression

180

in 12-day-old wild-type plants grown at 16 ºC and 22 ºC. Our results show that plants

181

grown at 16 ºC keep high TEM1 and TEM2 levels longer than at 22°C, maintaining high

182

levels at 16 ºC at a developmental phase in which they normally reached their minimum

183

at 22 ºC (Fig. 2 and Supplemental Fig. S3A). This indicates that a thermal regulation of

184

TEM exists under LD conditions. Moreover, the increased TEM levels at 16 ºC (Fig.

185

2A) were correlated with a reduction of FT expression at low temperature (Fig. 2B). In

186

addition to FT, we decided to include the analysis of TSF in our experiments as it is

187

known to play a role in the regulation of ambient temperature-responsive flowering

188

(Kim et al., 2013). We found first that TSF and FT display a similar expression pattern

189

throughout development in wild-type plants grown at 22 ºC under LD conditions

190

(Supplemental Fig. S3B), both opposite to TEM abundance (Supplemental Fig. S3A).

10

191

Next, to better characterise the TEM thermal regulation along development and to

192

determine when TEM abundance reaches the minimum at low temperatures, we

193

performed time course analyses of wild-type plants grown at 16 ºC during 5 weeks. The

194

relative mRNA levels of TEM1 and TEM2 showed the expected gradual decrease along

195

development. In contrast to what happens at 22 ºC (Supplemental Fig. S3A), at low

196

temperature their levels dropped later, around the 3rd-4th week (Fig. 2C and

11

197

Supplemental Fig. S4). In accordance to that, plants grown at 16 ºC showed a later rise

198

of FT and TSF levels than at 22 ºC and this rise occurred almost simultaneously with

199

the descent of TEM expression, around the 3rd-4th week (Fig. 2D and Supplemental Fig.

200

S4B). These results indicate that there is a correlation between the decrease of TEM

201

abundance and the increase of FT and TSF levels at 16 ºC and that this happens later

202

than at 22 ºC, in agreement with the delayed flowering at low temperatures.

12

203 204

TEMs directly repress FT and TSF at 16 ºC

205

To further confirm whether TEM genes regulate FT and/or TSF, we carried out

206

expression analyses in wild-type and tem1 tem2 mutant plants grown at 16 ºC. As we

207

expected, FT and TSF were clearly up-regulated in tem1 tem2 (Fig. 3 and Supplemental

208

Fig. S5). To understand whether the repression of FT and TSF was direct or indirect, we

209

performed chromatin immunoprecipitation (ChIP) experiments at low temperature

210

conditions. TEM proteins, as other RAV members, recognise and bind a canonical

211

sequence known as RAV binding site (Kagaya et al., 1999). Direct binding of TEM1 to

212

the RAV binding site of the 5’ UTR region of FT was previously reported under LD at

213

22 ºC (Castillejo and Pelaz, 2008). Here we show that TEM2 binds in vivo specifically

214

to both FT and TSF chromatin at 16 ºC. We found a significant enrichment of the 5’

215

UTR region of FT containing the canonical RAV-binding site (5’CAACAN9CACCTG

216

3’) (Fig. 4 and Supplemental Fig. S6), 43 nucleotides (nt) upstream of the ATG start

217

codon, in 35S::TEM2 plants, while only a slight enrichment was found in 35S::TEM1

218

plants. In addition, a clear significant enrichment of the TSF promoter was detected in

219

35S::TEM2, but not in 35S::TEM1 plants, in a region 321 nt upstream of the ATG,

220

which contains a non-canonical RAV-binding site (5’CAAGAN2CAAGTG 3’; Fig. 4B

221

and Supplemental Fig. S6B). Taken together, these data demonstrate that TEM2

222

specifically binds to both FT and TSF, and directly regulates their expression at 16 ºC.

223 224

SVP positively regulates TEM2 at 16 ºC

13

225

As a result of genome-wide analyses, TEM genes were identified as targets of SVP

226

under LD conditions at 23 ºC (Tao et al., 2012). Given that SVP is involved in the

227

thermosensory pathway, our following question was whether SVP regulates TEM genes

228

at low temperatures. To test the possibility that SVP protein regulates TEM1 and TEM2 14

229

expression via direct binding to the CArG motifs, where the MADS domain proteins are

230

known to bind, present in the TEM1 and TEM2 genomic loci, we performed ChIP

231

assays using pSVP::SVP:HA svp-32 plants (Lee et al., 2013, Supplemental Fig. S7)

232

under the two temperature conditions (23 °C and 16 °C). We chose two regions

15

233

containing CArG motifs (Tao et al., 2012) in the TEM1 and TEM2 promoter sequences

234

(Fig. 5 and Supplemental Fig. S8). A region lacking a CArG motif (NC) was used as a

235

negative control. Strong binding of SVP protein was observed in regions I and II of

16

236

TEM1 and TEM2; while interestingly, at 16 ºC we only observed a clear binding to

237

TEM2 but not to TEM1 regulatory regions (Fig. 5B).

238

To determine if the binding of SVP to TEM1 and TEM2 genomic loci affects TEM

239

expression, we carried out expression analysis in wild-type and svp mutant plants at 22 17

240

ºC and 16 ºC. A strong down-regulation of TEM2 was observed in svp plants at both 22

241

ºC and 16 ºC, whereas a slight or no TEM1 reduction was found in svp at 22 ºC and 16

242

ºC, respectively (Fig. 6 and Supplemental Fig. S9). Taken together, these data show that

243

SVP regulates the expression levels of TEM genes via direct binding to the CArG

18

244

motifs in TEM1 and TEM2 genomic loci at 22 ºC, but only to TEM2 at 16 ºC for the

245

regulation of ambient temperature-responsive flowering.

246

To test whether there could be reciprocal regulation between TEM and SVP, we

247

examined the expression levels of SVP in tem1 tem2 plants. However, we did not find

248

changes in the expression levels of SVP in tem1 tem2 compared to wild-type plants,

249

neither at 22 ºC nor at 16 ºC (Supplemental Fig. S10), suggesting that TEMs do not

250

regulate SVP levels.

251 252

Genetic interactions between svp and tem mutations at different temperatures.

253

Finally, to examine the genetic relationship between SVP and TEMs in the

254

thermosensory pathway, we measured the flowering time of tem single mutants, tem1

255

tem2 double mutant, svp mutant and the double and triple combinations of svp and tem

256

mutations, under LD conditions at 22 ºC and 16 ºC (Fig. 7 and Supplemental Table S2).

257

As shown above, at 16 ºC tem1 tem2 flowered earlier than wild-type plants; and tem

258

single mutants also showed an early flowering phenotype. Despite their early flowering,

259

we observed that tem double mutants flowered later than any combination with svp

260

mutation. svp, svp tem1, svp tem2 and the triple svp tem1 tem2 all flowered earlier at 16

261

ºC than tem1 tem2 plants (Fig. 7). By contrast, at 22 ºC tem1 tem2 flowered with a

262

similar number of leaves than mutants including tem and svp mutations, and the small

263

differences observed were not statistically significant (Fig. 7). Interestingly, tem2 single

264

mutants flowered with a similar number of leaves to svp at 22 ºC, but flowered later at

265

16 ºC, whereas tem2 showed a slight delay compared to svp tem2 at 22 ºC, more evident

266

at 16 ºC (Fig. 7). Taken together, these results indicate that TEMs regulate flowering at

267

22 ºC in a SVP-dependent manner and at 16 ºC in a partially SVP-independent manner.

268 269

19

270

DISCUSSION

271

Plants have developed mechanisms to perceive and respond to environmental

272

fluctuations by adjusting their growth as well as to predict upcoming daily and seasonal

273

cues, which result in massive developmental plasticity (Franklin, 2009). Photoperiod 20

274

and ambient temperature provide relevant information for the adaptation to seasonal

275

changes, which would allow plants to respond to a cold snap or a sudden warm up, and

276

to optimise flowering time. Hence, light and temperature cues have a key role in

277

flowering time regulation (Andrés and Coupland, 2012).

278 279

TEM genes repress flowering at low ambient temperatures

280

It has been shown that a slight decrease from 23 ºC to 16 ºC down-regulates FT

281

expression, even under a favourable photoperiod (Blázquez et al., 2003, Fig. 1). This

282

suggests that under low ambient temperature and LD conditions, floral repressors gain a

283

relevant role in maintaining low FT expression levels despite the presence of its

284

activator CO. One of the most studied repressors acting in these conditions is SVP,

285

which interacts with other floral repressors, such as FLM (Lee et al., 2013), FLC (Lee et

286

al., 2007; Li et al., 2008) and possibly other members of the FLC-clade (MAF genes)

287

(Gu et al., 2013) to repress FT, TSF and SOC1.

288

The role of TEM genes as flowering repressors under LD at 22 ºC has been previously

289

reported (Castillejo and Pelaz, 2008), however their function at low ambient

290

temperatures was unknown. The early flowering observed in the tem1 tem2 mutant at

291

both 22 ºC and 16 ºC compared to wild-type plants indicates that TEMs act as floral

292

repressors also at low ambient temperature. However this double mutant flowered later

293

at 16 ºC than at 22 ºC (Fig. 1A and Fig. 7). Therefore, in contrast to other genes, whose

294

loss of function causes insensitivity to ambient temperature, such as SVP or FLM

295

(Balasubramanian et al., 2006; Lee et al., 2007; Lee et al., 2013; Posé et al., 2013), tem

296

mutants are still sensitive to low ambient temperature, although their response is

297

reduced compared to wild-type plants. Thus, TEM genes do not seem to control the low

298

temperature responsive pathway, but somehow act downstream of SVP in transmitting

299

the response signal. Indeed, there is a clear correlation between the flowering time of

300

tem1 tem2 and svp mutants and the FT relative expression, at both temperatures (Fig.

301

1C; Fig. 1D and Supplemental Fig. S1). At 22 ºC the FT upregulation displayed in tem1

302

tem2 and in svp plants gave rise to the same earlier flowering, probably because in both

303

mutant plants FT exceeded the expression threshold required to induce the floral

304

transition (Fig. 1A and Fig. 1C). By contrast, at 16 ºC the higher level of FT in svp

305

mutants leads to an earlier flowering than in tem1 tem2 (Fig. 1A and Fig. 1D).

306

Therefore, the later flowering of tem mutants relative to svp at 16 ºC seems to be due to

307

the presence of active SVP that keeps some repression on FT expression in tem mutants. 21

308

In that sense, SVP directly represses FT expression through direct binding to the FT

309

promoter (Lee et al., 2007; Lee et al., 2013).

310 311

Late decay in TEM gene expressions at 16 ºC delays flowering

312

Like SVP, which shows practically the same mRNA levels at 16 ºC and 22-23 ºC (Lee

313

et al., 2007; Supplemental Fig. S10), TEM expression is not increased at 16 ºC at early

314

stages of development as we did not detect changes in TEM1 or TEM2 mRNA levels in

315

9-day-old wild-type seedlings at 16 ºC relative to 22 °C (Supplemental Fig. S2).

316

However, our results showed that in 12-day-old seedlings, TEM genes are expressed at

317

higher levels at 16 ºC than at 22 °C, which correlated with a decrease in FT and TSF

318

expression at 16 °C (Fig. 2A and Fig. 2B). These data indicate that TEM gene

319

expression decays later at low ambient temperatures, a conclusion supported by the

320

analysis of TEM expression throughout development at 16 °C and 22 °C (Fig. 2C; Fig.

321

2D and Supplemental Fig. S3).

322

FT and TSF act as floral promoters at 22-23 ºC and at 16 ºC (Michaels et al., 2005;

323

Yamaguchi et al., 2005; Jang et al., 2009; Kim et al., 2013; Lee et al., 2013), with FT

324

being the main player. We previously reported that TEM genes are floral repressors

325

under LD and short day conditions at 22 ºC, by controlling FT (Castillejo and Pelaz,

326

2008) and GA biosynthetic genes GA3ox1/2 (Osnato et al., 2012). Here we report that

327

TEM genes also delay flowering and repress FT, as well as TSF, at 16 ºC under LD

328

conditions. The FT and TSF daily patterns of expression were mostly unchanged in

329

tem1 tem2 mutants, but their abundance was increased, both at 22 ºC and 16 ºC (Fig. 1;

330

Fig. 3; Supplemental Fig. S1 and Supplemental Fig. S5). This up-regulation of FT and

331

TSF in tem1 tem2 plants (Fig. 3), together with the binding of TEM to FT and TSF

332

regulatory regions (Fig. 4), indicate that TEM, and more specifically TEM2, directly

333

represses FT and TSF at low ambient temperatures. Thus, the later drop of TEM gene

334

expression at 16 ºC results in a longer FT and TSF repression and therefore in a later

335

flowering when compared with plants growing at 22 ºC.

336 337

SVP up-regulates TEM2 through direct binding in response to low temperatures

338

Previous high-throughput experiments indicated that SVP positively regulates TEM

339

genes at 22 ºC under LD conditions (Tao et al., 2012). Here we show that the effect of

340

low temperature on TEM2 can be explained by the positive and direct regulation that

341

SVP exerts over it also at 16 ºC (Fig. 5 and Fig. 6). Bioinformatic analyses detected 22

342

several MADS binding sites in the promoters of TEM1 and TEM2 where SVP could

343

putatively bind and that were experimentally tested by ChIP assays at 23 ºC and 16 ºC.

344

Indeed, our results confirmed the binding of SVP to TEM1 and TEM2 at 23 ºC

345

described by Tao et al (2012) through ChIP-chip analysis, but also provide new data on

346

the specific regions regulation of TEM2, but not TEM1, by SVP at 16 ºC (Fig. 5 and 6).

347

At both temperatures, SVP binding on TEM2 appeared to be stronger than on TEM1,

348

which correlated with the strong downregulation of TEM2 observed in svp mutants at

349

22 ºC and 16 ºC (Fig. 6). Furthermore, in svp mutants TEM2 presented practically the

350

same relative mRNA levels at 22 ºC and 16 ºC, reduced in both cases when compared to

351

wild-type plants. This indicates that, although TEM genes are redundant in its function

352

as repressors of FT and GA3ox1/2 (Castillejo and Pelaz, 2008; Osnato et al., 2012), they

353

are differentially regulated by SVP in the ambient temperature pathway.

354 355

TEM acts at least partially independently of SVP at low ambient temperatures

356

The analyses of tem1, tem2 and svp mutants in multiple combinations at 22 ºC indicate

357

that TEM1 and TEM2 act basically on the same genetic pathway as SVP. However, at 16

358

ºC the global analysis of our flowering time data indicates that a slight additive effect of

359

TEM and SVP exists, which suggests that they may act in a partially independent

360

manner at low ambient temperature (Fig. 7). Analysed in detail, at 22 ºC the flowering

361

time of tem2 and svp was almost the same, as the difference in the total leaves produced

362

was not statistically significant, and we did not find a significant difference either

363

between the double tem1 tem2 mutant and svp plants. However, when we compared

364

tem2 and svp tem2, we found that svp tem2 is slightly earlier, which indicates that not

365

all the effect of SVP is through TEM2. These flowering time data correlate with the

366

strong binding of SVP to TEM2 obtained by ChIP-qPCR (Fig. 5) and the

367

downregulation of TEM2 observed in svp mutant plants (Fig. 6). At 16 ºC svp tem1

368

tem2 is not much earlier than the single svp mutant and the difference obtained was not

369

statistically significant. This suggests that loss of SVP activity masks the effect of TEM

370

on flowering time at lower temperatures. Therefore, at 16 ºC SVP represses flowering

371

partly through TEM, specifically TEM2, and partly through direct binding to FT (Lee et

372

al., 2007; Lee et al., 2013).

373 374

CONCLUSION

23

375

In conclusion, we have identified new players, TEM genes, in the ambient temperature

376

pathway, as well as their regulation by SVP. SVP and TEM can reinforce the

377

temperature responses by signalling partially through distinct pathways to control

378

common outputs, such as FT and TSF. Moreover, SVP protein accumulation is higher

379

during the day than during the night under LD conditions (Yoshida et al., 2009), in

380

agreement with its mRNA expression (Supplemental Fig. S10), while, TEM1 protein,

381

and most probably TEM2, have the opposite pattern (Osnato et al., 2012). This could

382

suggest that SVP and TEM could regulate FT and TSF in different moments of the day.

383

SVP represses FT during the morning (reviewed in Song et al., 2013) and TEM would

384

repress FT, TSF and GA3ox1/2 during the night. This work, together with previous

385

reports, indicates that TEM genes are involved in several genetic pathways that regulate

386

flowering.

24

387

MATERIAL & METHODS

388

Plant material and growth conditions

389

Arabidopsis thaliana Col-0 ecotype was used as wild-type control in all the experiments.

390

All mutants and transgenic lines are in the Col-0 background. tem1-1, 35S::TEM1,

391

35S::TEM2 (Castillejo and Pelaz, 2008); tem2-2, tem1-1 tem2-2 (Osnato et al., 2012)

392

and svp-32 (Lee et al., 2007) have been described previously. The svp-41 mutant

393

(Hartmann et al., 2000) was kindly donated by Martin Kater (Università degli studi di

394

Milano, Milan, Italy). The tem1-1 svp-41, tem2-2 svp-41 and tem1-1 tem2-2 svp-41

395

combinations were generated by crosses. Genotypes were confirmed by PCR using

396

published oligonucleotides (Supplemental Table S3). Seeds were stratified in darkness

397

at 4 ºC for 3 days and sown on soil. All plants were grown in chambers under controlled

398

LD photoperiod (16 h light/ 8 h dark) at 22-23 ºC or 16 ºC, under a mixture of cool

399

white (TL5 54W, 965) and with warm white (TL5 54W, 840) fluorescent lights, with a

400

fluence rate of 80 to 90 μmol m-2 s-1.

401 402

Generation of transgenic plants

403

To generate the pSVP::SVP:HA construct, the open reading frame of SVP was amplified

404

via reverse transcription (RT)-PCR using complementary DNA (cDNA) produced from

405

8-day-old seedlings. The resulting amplicons were cloned into the pCHF3 vector

406

harboring the approximately 2.5 kb promoter fragment of SVP. This construct was

407

introduced into svp-32 plants (Lee et al., 2007) using the floral dip method with minor

408

modifications (Weigel and Glazebrook, 2002). Subsequently, transformants were

409

selected for kanamycin resistance and about 30-40 T1 seedlings were analysed (Lee et

410

al., 2013). Oligonucleotide primers used for cloning are listed in Supplemental Table S3.

411

In pSVP::SVP:HA svp-32 plants, the production of the SVP-HA protein was confirmed

412

(Supplemental Fig. S7A) and the early flowering and ambient temperature-insensitive

413

flowering

414

(Supplemental Fig. S7B), indicating that HA-tagged SVP protein is functional.

phenotypes

of

svp-32

mutants

were

rescued

by

pSVP::SVP:HA

415 416

Phenotypic and statistical analyses

417

For flowering time measurements, plants were randomized with the respective controls

418

and grown on soil in controlled environment growth chambers. Flowering time was

419

determined by counting the number of cauline and rosette leaves of at least 12

420

individual plants. The number of days to flowering was determined when the floral bud 25

421

was visible to the naked eye. Data are reported as a mean value of the total leaf number

422

± standard deviation (SD), for each genotype and experimental condition compared; we

423

use the mean value of the total leaf number ± standard error of the mean (SEM) when

424

comparing the mean of independent experiments. All flowering time assays were

425

performed at least twice. Flowering time data were subjected to analyses of variance

426

(ANOVA). Post-hoc tests were performed using Tukey’s multiple comparisons test

427

after two-way ANOVA. Statistical analyses were performed with Prism 6 software

428

(GraphPad Software, Inc).

429 430

RNA isolation and expression analysis

431

Samples consisted of pools of seedlings (from 12 to 59 individuals for each time point,

432

depending on the time of collection) sown on soil, which were quickly frozen in liquid

433

nitrogen and powdered prior to RNA extraction. RNA was extracted using PureLinkTM

434

Micro-to-Midi Total RNA Purification kit (Invitrogen-Ambion) and DNAse treated

435

using DNA-free kit (Ambion). RNA integrity was checked on agarose gels and

436

concentration was measured using ND-1000 Spectrophotometer (Thermo Scientific).

437

Between 1-1.5 μg of DNAse-treated RNA was used for cDNA synthesis by employing

438

SuperScriptTM III Reverse Transcriptase (Invitrogen), according to the manufacturer’s

439

instructions. The resulting cDNA was diluted prior to subsequent expression analyses.

440

qPCR was performed on a LightCycler® 480 (Roche) using SYBR® Premix ExTaqTM

441

(Takara). Three technical replicates were made per sample. The relative expression was

442

calculated using the 2-ΔΔCt method (Livak and Schmittgen, 2001). Ubiquitin 10 (UBQ10)

443

was used as a reference gene. Results from biological duplicates are shown.

444

Oligonucleotide primers used for qPCR are listed in Supplemental Table S3.

445 446

Chromatin immunoprecipitation (ChIP) assays

447

Two grams of pSVP::SVP:HA seedlings grown on soil under LD conditions at 23 °C or

448

16 °C were cross-linked in 1% formaldehyde on ice using vacuum infiltration. Nuclear

449

extracts were isolated and the immunoprecipitation assays were conducted as described

450

previously (Kim et al., 2012). After shearing chromatin via sonication, rabbit anti-HA

451

polyclonal antibody (about 5 µg) (Santa Cruz Biotechnology) was used to

452

immunoprecipitate genomic DNA fragments. qPCR was performed using DNA

453

recovered from immunoprecipitation or 10% input DNA with a number of primer sets

26

454

spanning the regulatory regions of TEM1 and TEM2 (Supplemental Table S3). The

455

relative enrichment of each fragment was calculated by comparing samples

456

immunoprecipitated with HA and cMyc (negative control) antibodies (Livak and

457

Schmittgen, 2001). ChIP experiments were performed in two biological replicates

458

(samples independently harvested on different days) with three technical replicates each,

459

with similar results.

460

One to 1.5 grams of 35S::TEM1:HA and 35S::TEM2:HA seedlings were grown on soil

461

under LD conditions at 22 ºC and 16 ºC to test direct binding of TEM to FT and TSF

462

loci. Crosslinked DNA was immunoprecipitated with an anti-HA antibody (Sigma),

463

purifed using Protein A-Agarose resin (Millipore) and tested by qPCR using specific

464

primer sets (see Supplemental Table S3) on regulatory regions of FT and TSF. ChIP

465

experiments were performed in at least two biological replicates (samples independently

466

harvested on different days) with three technical replicates each, with similar results.

467 468

ACKNOWLEDGEMENTS

469

We thank Martin Kater for svp-41 seeds.

470 471 472

AUTHOR CONTRIBUTIONS

473

E.M.-G. and S.P. conceived and designed the experiments. E.M.-G is the main

474

contributor to the experimental part of this manuscript. L.M.-H., A.A-J., J.H.L. and

475

J.H.A. performed some of the experiments. E.M.-G., P.S-L. and S.P. wrote the

476

manuscript with corrections from the rest of the authors. SVP ChIP assays were

477

performed at J.H.A. laboratory. Rest of the work was performed in the group of S.P.

478

who provided funding and supervised the research and the writing of the manuscript.

479 480 481 482

COMPETING INTEREST

483

None of the authors have any financial, personal or professional interests that have

484

influenced this present paper.

485 486 487 27

488

FIGURE LEGENDS

489

Figure 1.

490

tem1 tem2 mutant plants are early flowering at 16ºC but still sensitive to changes in

491

ambient growth temperature. Flowering time was measured as (A) the number of

492

total leaves produced at flowering, and (B) the number of days to flowering for wild-

493

type (Col-0) plants, tem1 tem2 and svp mutants grown under LD conditions at 22ºC

494

(black) or 16ºC (white). Data are reported as mean ± SEM of three independent

495

experiments (each dot plot represents an independent experiment; red dots indicate

496

experiments performed at 22ºC; blue squares, at 16ºC). A minimum of 12 plants per

497

genotype and experimental condition were analysed in each independent experiment.

498

The numbers below the bars denote the leaf number ratio (16°C/22°C). See

499

Supplemental Table S1 for more detail. (C) and (D), RT-qPCR analysis of FT

500

expression in wild-type (black triangles), tem1 tem2 (grey squares) and svp plants (grey

501

dots) in 9-day-old seedlings grown under LD conditions at (C) 22 ºC or (D) 16 ºC.

502

Samples were collected over a 24-h period. The dark period is denoted by the black bar.

503

Two independent experiments gave similar results (Supplemental Fig. S1), and one was

504

chosen as representative. RNA levels were normalized to UBQ10. Error bars show

505

standard deviation (SD) of three technical replicates.

506 507

Figure 2.

508

Opposite expression pattern of TEM and FT/TSF genes at 16ºC. Expression analysis

509

of TEM1, TEM2, FT and TSF in (A-B) 12 day-old wild-type (Col-0) plants grown at

510

22ºC or 16ºC, and (C-D) wild-type plants grown at 16ºC for 5 weeks. Fold change in

511

transcript levels at 16 ºC is depicted in comparison to 22ºC. All samples were collected

512

at ZT18. Three independent experiments gave similar results (Supplemental Fig. S4),

513

and one was chosen as representative. Error bars show SD of three technical replicates.

514

RNA levels were determined by RT-qPCR and normalized to UBQ10.

515 516

Figure 3.

517

TEMs regulate FT and TSF levels at 16ºC. Relative FT and TSF mRNA levels in

518

tem1 tem2 mutant compared to wild-type (Col-0) plants. Nine-day-old seedlings were

519

sampled at 4-h intervals, except from ZT16 to ZT20 when samples were collected every

520

2h. Two independent experiments gave similar results (Supplemental Fig. S5), and one

28

521

was chosen as representative. Error bars show SD of three technical replicates. RNA

522

levels were determined by RT-qPCR and normalized to UBQ10.

523 524

Figure 4.

525

Binding of TEM2 protein to FT and TSF regulatory regions at 16ºC. ChIP assay of

526

binding of TEM1-HA and TEM2-HA proteins to the RAV motifs in the FT and TSF

527

regulatory regions. (A,B) DNA fragments containing the RAV binding site were

528

analyzed by chromatin immunoprecipitation (ChIP) using an anti-HA antibody and 9-

529

day-old 35S::TEM1 and 35S::TEM2 plants carrying an HA tag. Precipitated chromatin

530

was used as a template in qPCR, using FT (A) or TSF-specific (B) primers.

531

Immunoprecipitated DNA enrichment is presented as percentage of input DNA. Two

532

(FT) or three (TSF) independent experiments gave similar results (Supplemental Fig.

533

S6), and one was chosen as representative. Error bars show SD of three technical

534

replicates. (B) Schematic diagrams of the FT and TSF regulatory regions are shown

535

below the graphs. Arrows indicate fragments amplified by qPCR after ChIP. These

536

fragments contain the canonical RAV binding site for FT and a putative RAV binding

537

site for TSF.

538 539

Figure 5.

540

Binding of SVP protein to the TEM1 and TEM2 genomic loci. (A) Schematic

541

diagram of the TEM1 and TEM2 genomic regions. Closed boxes and thin lines represent

542

exons and introns, respectively. Asterisks indicate the predicted CArG and variant

543

CArG motifs. Short horizontal lines indicate amplicons in ChIP-qPCR assays. Regions I

544

and II, carrying CArG motifs, were selected to amplify; NC was the amplicon used as a

545

negative control. (B) Chromatin immunoprecipitation (ChIP) analysis of binding of

546

SVP protein to the TEM1 and TEM2 genomic regions at 23°C and 16°C in 9-day-old

547

pSVP::SVP:HA svp-32 plants. An anti-HA antibody was used for immunoprecipitation.

548

Black bars denote the amplified fragments in qPCR: [region I (-1,005 to – 920, relative

549

to ATG); region II (-350 to -271); NC (+1,011 to +1,085) for TEM1] and [region I (-

550

1,429 to – 1,385, relative to ATG); region II (-434 to -345); NC (+1,005 to +1,065) for

551

TEM2]. Two independent experiments gave similar results (Supplemental Fig. S8), and

552

one was chosen as representative. Error bars show SD of three technical replicates.

553 554

Figure 6.

29

555

SVP positively regulates TEM2 expression at 22ºC and 16ºC. Relative mRNA levels

556

of TEM1 and TEM2 at 22°C (upper panels) and 16°C (lower panels) in svp mutant

557

compared to wild-type (Col-0) plants. Nine-day-old seedlings were sampled at 4-h

558

intervals, except from ZT16 to ZT20 when samples were collected every 2h. Two

559

independent experiments gave similar results (Supplemental Fig. S9), and one was

560

chosen as representative. Error bars show SD of three technical replicates. RNA levels

561

were determined by RT-qPCR and normalized to UBQ10.

562 563

Figure 7.

564

Genetic interaction between tem and svp mutants. (A) Flowering time measured as

565

the number of total leaves produced at flowering for wild-type (Col-0), tem mutants and

566

svp tem double and triple mutants grown under LD conditions at 22ºC or 16ºC. Data are

567

reported as mean ± SEM of two independent experiments (each dot plot represents an

568

independent experiment; red dots indicate experiments performed at 22ºC; blue squares,

569

at 16ºC). A minimum of 10 plants per genotype and experimental condition were

570

analysed in each independent experiment. The numbers below the bars denote the leaf

571

number ratio (16°C/22°C). See Supplemental Table S2 for more detail. (B, C)

572

Photographs of plants used in flowering time analysis, grown for 24 days at 22°C (B)

573

and for 32 days at 16°C (C).

574 575 576

30

Supplemental Figure S1. Upregulation of FT in tem1 tem2 and svp mutant plants at 22ºC and 16ºC. FT expression in wild-type (Col-0), tem1 tem2 and svp plants grown under LD conditions at 22ºC or 16ºC. Nine-day-old plants were collected at 4-hours intervals over a 24hperiod. Error bars show SD of three technical replicates. RNA levels were determined by RT-qPCR and normalized to UBQ10. Biological replicate of Fig. 1.

Supplemental Figure S2. Low temperature does not increase TEM levels early in development. Relative mRNA levels of TEM1 (left) and TEM2 (right) in wild-type plants at 22ºC and 16ºC. Nine-day-old seedlings were sampled at 4-h intervals, except from ZT16 to ZT20 when samples were collected every 2h (to add the ZT18 time point). Data are reported as mean ± SEM of two independent experiments represented by blue squares (16ºC) and red dots (22ºC). The dark period is denoted by the black bar. RNA levels were determined by RT-qPCR and normalized to UBQ10. ZT, zeitgeber time.

Supplemental Figure S3. Opposite expression pattern of TEM and FT/TSF genes along development at 22ºC. (A) TEM1 and TEM2 expression, and (B) FT and TSF expression in wild-type plants grown under LD conditions at 22ºC during 2 weeks. Samples were collected at ZT18. Error bars show SD of three technical replicates. RNA levels were determined by RTqPCR and normalized to UBQ10.

Supplemental Figure S4. Opposite expression pattern of TEM and FT/TSF genes along development at 16ºC. (A) TEM1 and TEM2 expression, and (B) FT and TSF expression in wild-type plants grown under LD conditions at 16ºC during 5 weeks. Samples were collected at ZT18. Error bars show SD of three technical replicates. RNA levels were determined by RTqPCR and normalized to UBQ10. Biological replicates of Fig. 2.

Supplemental Figure S5. TEMs regulate FT and TSF levels at 16ºC. Relative FT and TSF mRNA levels in tem1 tem2 mutant compared to wild-type (Col-0) plants. Nine-day-old seedlings were sampled at 4-h intervals, except from ZT16 to ZT20 when samples were collected every 2h. Error bars show SD of three technical replicates. RNA levels were determined by RT-qPCR and normalized to UBQ10. Biological replicates of Fig. 3.

Supplemental Figure S6. TEM2 protein binds to the FT and TSF promotors at 16ºC. ChIP assay of binding of TEM1-HA and TEM2-HA proteins to the RAV motifs in the (A) FT and (B) TSF regulatory regions. DNA fragments containing the RAV binding site were analyzed by chromatin immunoprecipitation (ChIP) using 9-day-old 35S::TEM1 and 35S::TEM2 plants carrying an HA tag. Precipitated chromatin was used as a template in qPCR. Immunoprecipitated DNA enrichment is presented as percentage of input DNA. Error bars show SD of three technical replicates. Biological replicates of Fig. 4.

Supplemental Figure S7. Characterization of pSVP::SVP:HA svp-32 transgenic plants. (A) SVP-HA protein expression in two independent pSVP::SVP:HA svp-32 lines (#9 and #16) grown at 23C under LD conditions. An anti-HA antibody was used to detect SVP-HA protein. RbcL was used as a loading control. (B) Flowering time of two independent homozygous pSVP::SVP:HA svp-32 plants at 23C and 16C under LD conditions. Error bars show SD. The numbers of plants used in this analysis are as follows: wild-type, n=30; svp-32, n=34; pSVP::SVP-HA svp-32 #9, n=20; pSVP::SVP-HA svp-32 #9, n=26.

Supplemental Figure S8. Binding of SVP protein to the TEM1 and TEM2 genomic loci. Chromatin immunoprecipitation (ChIP) analysis of binding of SVP protein to the TEM1 and TEM2 genomic regions at 23C and 16C in 9-day old plants pSVP::SVP:HA svp-32. Error bars show SD of three technical replicates. Biological replicates of Fig. 5.

Supplemental Figure S9. SVP positively regulates TEM2 expression at 22ºC and 16ºC. Relative mRNA levels of TEM1 and TEM2 in svp mutant compared to wild-type (Col-0) plants. Nine-day-old seedlings were sampled at 4-h intervals, except from ZT16 to ZT20 when samples were collected every 2h. Bars show SD of three technical replicates. RNA levels were determined by RT-qPCR and normalized to UBQ10. Biological replicate of Fig. 6.

Supplemental Figure S10. TEMs do not regulate SVP levels. Relative SVP mRNA levels at 22ºC (left) and 16ºC (right) in tem1 tem2 mutant compared to wild-type plants. 9-day-old seedlings were sampled at 4-h intervals, except from ZT16 to ZT20 when samples were collected every 2h. Two independent experiments gave similar results and one was chosen as representative. Error bars show SD of three technical replicates. RNA levels were determined by RT-qPCR and normalized to UBQ10.

Supplemental Table S1 (A)

Flowering time (corresponding to data in Figure 1) and (B) statistical analysis

of total leaf numbers. RL: rossete leaves; CL: cauline leaves; TL: total leaves; n: number of plants per genotype and experiment; NQ: not quantified; NS: not significant. For total leaves and days, mean ± SD are shown for each genotype and experimental condition. A 22 ºC Genotype RL Experiment 1 Col-0 12.2 tem1 tem2 5.9 6.1 svp Experiment 2 Col-0 14.5 tem1 tem2 6.1 6.2 svp Experiment 3 Col-0 11.9 tem1 tem2 5.6 7.9 svp

CL

TL±SD Days±SD

n

2.7 14.9±0.7 21.9±0.9 2.1 8.1±0.7 13.1±0.8 2.4 8.5±0.5 15.5±0.7

13 17 12

3.8 18.3±0.7 22.5±0.5 2.9 9.0±1.0 14.1±0.8 2.9 9.2±0.4 14.5±0.5

15 18 18

3.0 14.9±0.7 2.6 8.2±0.8 2.0 9.9±0.7

15 15 15

NQ NQ NQ

16 ºC RL CL Experiment 1 26.2 5.8 10.7 3.7 7.2 2.8 Experiment 2 28.0 5.8 10.9 3.8 7.4 2.9 Experiment 3 23.6 7.4 11.4 2.3 8.4 2.1

TL±SD

Days±SD

n

32.0 ± 2.1 14.4 ± 0.9 9.9 ± 0.5

34.5±1.4 20.3±0.5 18.4±0.5

12 13 12

33.8 ± 2.3 14.7 ± 0.8 10.3 ± 1.4

37.2±0.9 22.2±0.4 20.7±1.5

13 12 13

31.0 ± 1.2 13.7 ± 1.4 10.6 ± 1.1

NQ NQ NQ

15 15 15

B. Supplemental Table S1 (continued). Multiple Comparisons of Total Leaf Number 22 ºC: Col-0 Exp1 vs. 22 ºC: tem1 tem2 Exp1 22 ºC: Col-0 Exp1 vs. 22 ºC: tem1 tem2 Exp2 22 ºC: Col-0 Exp1 vs. 22 ºC: tem1 tem2 Exp3 22 ºC: Col-0 Exp2 vs. 22 ºC: tem1 tem2 Exp1 22 ºC: Col-0 Exp2 vs. 22 ºC: tem1 tem2 Exp2 22 ºC: Col-0 Exp2 vs. 22 ºC: tem1 tem2 Exp3 22 ºC: Col-0 Exp3 vs. 22 ºC: tem1 tem2 Exp1 22 ºC: Col-0 Exp3 vs. 22 ºC: tem1 tem2 Exp2 22 ºC: Col-0 Exp3 vs. 22 ºC: tem1 tem2 Exp3 16 ºC: Col-0 Exp1 vs. 16 ºC: tem1 tem2 Exp1 16 ºC: Col-0 Exp1 vs. 16 ºC: tem1 tem2 Exp2 16 ºC: Col-0 Exp1 vs. 16 ºC: tem1 tem2 Exp3 16 ºC: Col-0 Exp2 vs. 16 ºC: tem1 tem2 Exp1 16 ºC: Col-0 Exp2 vs. 16 ºC: tem1 tem2 Exp2 16 ºC: Col-0 Exp2 vs. 16 ºC: tem1 tem2 Exp3 16 ºC: Col-0 Exp3 vs. 16 ºC: tem1 tem2 Exp1 16 ºC: Col-0 Exp3 vs. 16 ºC: tem1 tem2 Exp2 16 ºC: Col-0 Exp3 vs. 16 ºC: tem1 tem2 Exp3 22 ºC: tem1 tem2 Exp1 vs. 16 ºC: tem1 tem2 Exp1 22 ºC: tem1 tem2 Exp1 vs. 16 ºC: tem1 tem2 Exp2 22 ºC: tem1 tem2 Exp1 vs. 16 ºC: tem1 tem2 Exp3 22 ºC: tem1 tem2 Exp2 vs. 16 ºC: tem1 tem2 Exp1 22 ºC: tem1 tem2 Exp2 vs. 16 ºC: tem1 tem2 Exp2 22 ºC: tem1 tem2 Exp2 vs. 16 ºC: tem1 tem2 Exp3 22 ºC: tem1 tem2 Exp3 vs. 16 ºC: tem1 tem2 Exp1 22 ºC: tem1 tem2 Exp3 vs. 16 ºC: tem1 tem2 Exp2 22 ºC: tem1 tem2 Exp3 vs. 16 ºC: tem1 tem2 Exp3 22 ºC: tem1 tem2 Exp1 vs. 22 ºC: svp Exp1 22 ºC: tem1 tem2 Exp1 vs. 22 ºC: svp Exp2 22 ºC: tem1 tem2 Exp1 vs. 22 ºC: svp Exp3 22 ºC: tem1 tem2 Exp2 vs. 22 ºC: svp Exp1 22 ºC: tem1 tem2 Exp2 vs. 22 ºC: svp Exp2 22 ºC: tem1 tem2 Exp2 vs. 22 ºC: svp Exp3 22 ºC: tem1 tem2 Exp3 vs. 22 ºC: svp Exp1 22 ºC: tem1 tem2 Exp3 vs. 22 ºC: svp Exp2 22 ºC: tem1 tem2 Exp3 vs. 22 ºC: svp Exp3 16 ºC: tem1 tem2 Exp1 vs. 16 ºC: svp Exp1 16 ºC: tem1 tem2 Exp1 vs. 16 ºC: svp Exp2 16 ºC: tem1 tem2 Exp1 vs. 16 ºC: svp Exp3 16 ºC: tem1 tem2 Exp2 vs. 16 ºC: svp Exp1 16 ºC: tem1 tem2 Exp2 vs. 16 ºC: svp Exp2 16 ºC: tem1 tem2 Exp2 vs. 16 ºC: svp Exp3 16 ºC: tem1 tem2 Exp3 vs. 16 ºC: svp Exp1 16 ºC: tem1 tem2 Exp3 vs. 16 ºC: svp Exp2 16 ºC: tem1 tem2 Exp3 vs. 16 ºC: svp Exp3 22 ºC: Col-0 Exp1 vs. 16 ºC: tem1 tem2 Exp1

Significant? **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** **** NS NS *** NS NS NS NS NS ** **** **** **** **** **** **** **** **** **** NS

p-value