Aperture number influences pollen survival in Arabidopsis mutants

HAL Id: hal-01311848

https://hal.sorbonne-universite.fr/hal-01311848

Submitted on 4 May 2016

HAL

is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire

HAL, est

destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Aperture number influences pollen survival in Arabidopsis mutants

Charlotte Prieu, Alexis Matamoro-Vidal, Christian Raquin, Anna Dobritsa, Raphaël Mercier, Pierre-Henri Gouyon, Béatrice Albert

To cite this version:

Charlotte Prieu, Alexis Matamoro-Vidal, Christian Raquin, Anna Dobritsa, Raphaël Mercier, et al..

Aperture number influences pollen survival in Arabidopsis mutants. American Journal of Botany,

Botanical Society of America, 2016, 103 (3), pp.452-459. �10.3732/ajb.1500301�. �hal-01311848�

1 Prieu et al.

1

Aperture number influences pollen survival in Arabidopsis mutants ¹.

2

3

Charlotte Prieu

^2,3,4,7

, Alexis Matamoro-Vidal

^{2, 3, 4}

, Christian Raquin

^2,3

, Anna Dobritsa

⁵

, Raphaël 4

Mercier

⁶

, Pierre-Henri Gouyon

^4,8

and Béatrice Albert

^{2,3, 8}

5

Acknowledgements: The authors thank L. Saunois and A. Dubois for plants caring, A. Ressayre and C.

6

Dillman for helpful discussions on statistics. The authors also thank S. Nadot for useful comments on 7

the text.

8 9 10 11 12 13 14 15 16 17 18

1 Manuscript received _______; revision accepted _______.

2 Université Paris-Sud, Laboratoire Ecologie, Systématique et Evolution, UMR 8079, Orsay F-91405, France

3 CNRS, Orsay F-91405, France

4 Institut de Systématique, Évolution, Biodiversité, ISYEB - UMR 7205 - CNRS, MNHN, UPMC, EPHE, Muséum national d'Histoire naturelle, Sorbonne Universités, 57 rue Cuvier, CP39, F-75005, Paris, France

5 Department of Molecular Genetics and Center for Applied Plant Sciences, The Ohio State University, 015 Rightmire Hall, 1060 Carmack Road, Columbus, OH 43210, USA

6 INRA, UR254, IJPB, Versailles, France

7 Corresponding author: charlotte.prieu@u-psud.fr

8 These authors contributed equally to this work

2



Premise of the study: Pollen grains are subject to intense dehydration before dispersal. They 19

rehydrate after landing on a stigma or when placed in humid environment by absorbing 20

water from the stigma or surroundings. Resulting fluctuations in water content cause pollen 21

grains to undergo significant changes in volume. This suggests that morphological or 22

structural adaptations might exist to help pollen adjust to sudden volume changes, though 23

little is known about the correlation between pollen morphology and its ability to 24

accommodate volume changes. We studied the effect of one morphological feature of pollen 25

grains, the aperture number, on pollen wall resistance to water inflow in Arabidopsis 26

thaliana.

27



Methods: We used three Arabidopsis thaliana mutants producing pollen with different 28

aperture numbers (zero, four, and a mix of four to eight) and the wild-type pollen with three 29

apertures. We tested pollen survival in solutions with various mannitol concentrations.

30



Key results: Our experiments show that the number of intact pollen grains increases with 31

increased mannitol concentration for all pollen morphs tested. At a given mannitol 32

concentration, however, an increase in aperture number is associated with an increase in 33

pollen breakage.

34



Conclusions: Aperture patterns influence the capacity to accommodate volume variations in 35

pollen grains. When subjected to water inflow, pollen grains with few apertures survive 36

better than pollen with many apertures. Trade-offs between survival and germination are 37

likely to be involved in the evolution of pollen morphology.

38 39

Key words: pollen performance, aperture pattern, Arabidopsis thaliana, harmomegathy 40

41

42

43

44

45

46

47

48

49

50

51

52

53

3 Competition between individuals is a key feature of all ecosystems. Plants compete for light, space, 54

water and nutrients, as well as for opportunities to reproduce. For vascular plants, studies on 55

competitive abilities usually focus on sporophytes (the dominant phase in the life cycle of most land 56

plants), and rarely concern gametophytes, although they are also involved in ecological interactions.

57

Pollen, the male gametophyte of all seed plants, is very diverse morphologically, having frequent 58

inter-species variations in its shape, size, wall ornamentation, and aperture patterns (Erdtman, 1952;

59

PalDat, 2015). Apertures are the thinned areas of pollen wall where exine, the extremely resistant 60

outer layer made of sporopollenin, is absent or reduced (Edlund et al., 2004). Aperture patterns are 61

defined by the number, shape and positions of apertures (Walker and Doyle, 1975; Hesse et al., 62

2009). In many species, apertures are the primary sites through which exchanges between pollen 63

and environment take place – particularly during dehydration, when the grain is released from the 64

anther, and rehydration, when it lands on a stigma and absorbs water from stigmatic cells or surface 65

(Heslop-Harrison, 1979b; Edlund et al., 2004). Apertures also play an important role in pollen tube 66

germination, since in many cases pollen tube exits through these sites (Edlund et al., 2004).

67

Additionally, apertures are involved in harmomegathy, a pollen-specific process defined by 68

Wodehouse as “volume change accommodation” (Wodehouse, 1935) and manifested, for example, 69

by pollen’s ability to in-fold itself at the time of water loss to prevent complete desiccation (Katifori 70

et al., 2010). Wall structures involved in this process were called harmomegathus by Wodehouse, 71

namely “an organ or mechanism which accommodates a semirigid exine to changes in volume”

72

(Wodehouse, 1935). Harmomegathic properties of pollen grains are expected to mainly depend on 73

apertures (Payne, 1972; Blackmore and Barnes, 1986; Scotland et al., 1990), although as the exine 74

layer has some elastic properties, it may also be able to deform to some extent and accommodate 75

some volume changes (Bolick and Vogel, 1992; Rowley and Skvarla, 2000).

76

In angiosperms, pollen with a single distal aperture is ancestral for the clade (Furness and Rudall, 77

2004). This pollen type, called monosulcate, is the main type in early diverging angiosperms and 78

monocots, though derived types with other aperture patterns can be found at the generic or the 79

specific level. In eudicots, a very large clade with approximately 75% of the extant angiosperm 80

species, pollen grains are characterized by three furrow-like equatorial apertures (Doyle and Hotton, 81

1991; Furness and Rudall, 2004). The tricolpate pollen grain is an evolutionary innovation of the clade 82

and is the only morphological synapomorphy of eudicots known so far. Some authors have suggested 83

that the acquisition of tricolpate pollen could be the key innovation of eudicots that had been 84

essential for the evolutionary success of the clade (Furness and Rudall, 2004).

85

Previous studies have highlighted the connections between the aperture distribution on pollen 86

surface and characteristics of cell division during the male meiosis. Apertures are usually formed at 87

the last points of contact between the microspores (Wodehouse, 1935; Ressayre et al., 2002) and 88

variations in key aspects of pollen development, such as the type of cytokinesis or the geometrical 89

form of a tetrad, result in differences in aperture patterns (see, for example, Albert et al., 2010, 90

2011). Although the relationship between pollen development and aperture patterns has been 91

studied for some time, the link between aperture patterns and pollen fitness has received less 92

attention.

93

When focusing on apertures, some assumptions can be made regarding aperture pattern and pollen 94

performance. Since apertures are involved in pollen tube germination, an increase in aperture

95

4 number is expected to accelerate pollen germination on stigma. Moreover, high number of apertures 96

could enable faster pollen rehydration, especially in the case of dry stigmas (an assumption made, for 97

example, by Heslop-Harrison, 1979b). If the presence of many apertures indeed had a positive effect 98

on pollen fitness, then pollen grains with a high number of apertures should be widespread in wild 99

species. However this is not the case, since the vast majority of angiosperms produce pollen with one 100

or three apertures, whereas pollen grains with larger numbers of apertures occur rarely (Erdtman, 101

1952). Some studies carried out on Viola diversifolia, a species producing both three- and four- 102

colpate pollen, give some clues on this issue. These pollen types have different properties: pollen 103

with three apertures live longer than pollen with four apertures, whereas pollen with four apertures 104

germinate faster on stigma (Dajoz et al., 1991, 1993). It thus seems that increasing aperture number 105

is advantageous for the on-stigma competition between pollen grains, but could provide 106

disadvantage for pollen survival. Aquatic species dispersing pollen directly in water constitute 107

another example of aperture pattern adaptation. These species tend to produce pollen lacking 108

localized apertures; however, the exine layer of these pollen grains is uniformly thin and the entire 109

wall functions as an aperture (omniaperturate pollen; Pettitt and Jermy, 1974; Furness and Rudall, 110

1999). The reduced thickness of exine is interpreted as an adaptation to water dispersal: pollen 111

grains do not dehydrate before release, thus the protective outer layer becomes unnecessary and 112

tends to disappear (Pettitt and Jermy, 1974). Apart from these few examples, the connection 113

between aperture patterns and pollen fitness remains largely unexplored.

114

Apertures are involved in harmomegathy, but the precise effects of aperture number or shape on 115

volume accommodation are not completely understood. Theoretical approaches have shown that 116

aperture pattern and exine ornamentation can influence the way in which deformation of the pollen 117

wall takes place during dehydration (Katifori et al., 2010). In particular, that study has shown that 118

pollen grains with pore-shaped apertures are not able to deal with strong volume decrease very 119

efficiently, while in pollen grains with furrow-shaped apertures the wall can adapt to changes in 120

pollen size by folding inwards along the furrows as the volume decreases. Empirical studies have 121

shown that harmomegathy depends on aperture pattern (Payne, 1972; Blackmore and Barnes, 1986;

122

Scotland et al., 1990; Volkova et al., 2013), but it might also be influenced by other characteristics, 123

such as the size or shape of the grain (Halbritter and Hesse, 2004).

124

The aim of this study was to investigate the connection between aperture patterns and 125

harmomegathy. We are particularly interested in the link between aperture number and 126

accommodation of volume increase during water inflow, which happens during the on-stigma 127

rehydration phase (Heslop-Harrison, 1979a; Edlund et al., 2004) or is brought by variations in 128

humidity during pollen dispersal. In order to study the relationship between the aperture number 129

and the wall mechanical resistance, we used the wild-type Arabidopsis thaliana and three 130

Arabidopsis mutants with different numbers of apertures: this method enabled us to test the 131

influence of aperture number, since other pollen features remained largely the same among the 132

different morphs, except for the size of the grain. The different mutants came from two different 133

genetic backgrounds, this was taken into account when analyzing the results. The different plants 134

used here produce pollen grains with either 0, 3, 4 or a mix of 4 to 8 apertures, largely covering the 135

range of aperture numbers observed in wild species of angiosperms (Erdtman, 1952; PalDat, 2015).

136

We studied the ability of pollen grains to withstand the effects of water inflow when placed in 137

solutions of variable osmolarity. Pollen grains subjected to water inflow can either survive, if the wall

138

5 is resistant to swelling, or die, if the wall or plasma membrane breaks. In our study, we considered 139

pollen grains to be dead if wall and/or plasma membrane were broken, and we assumed that pollen 140

grains survived if they remained intact, regardless of their germination abilities (which can be 141

affected by several intrinsic and extrinsic factors). We evaluated the number of surviving pollen 142

grains for each genotype and for three different levels of osmolarity.

143

In theory, aperture number can influence pollen resistance to breakage during hydration in several 144

possible ways. Because apertures are more flexible than other parts of the pollen wall, their 145

increased number could potentially enable the wall to accommodate deformations more easily, 146

therefore leading to higher resistance to breakage. On the other hand, since water is absorbed 147

through apertures, more of these sites could result in faster swelling of the grain, compromising its 148

ability to cope with increased volume; also, increased aperture number could weaken the wall. These 149

possibilities are not exclusive. Our aim was to study the effect of aperture number on resistance to 150

osmotic stress and to test whether an increase in aperture number is associated with higher or lower 151

pollen mortality. We show here that the number of intact pollen grains decreases with hydration 152

intensity and that an increase in aperture number is associated with an increase in pollen breakage.

153 154

MATERIALS AND METHODS 155

Arabidopsis lines 156

We chose several mutants that differ in their pollen aperture number from the wild type, which has 157

tricolpate pollen grains (Fig. 1A). The inp1-1 mutant produces inaperturate pollen (Fig. 1B; Dobritsa 158

et al., 2011; Dobritsa and Coerper, 2012; length=18.07 µm, sd=1.62 µm, n=12). This mutant exhibits 159

well-developed siliques and a normal number of seeds per silique, indicating normal performance of 160

this pollen type during reproduction (Dobritsa et al., 2011). The lsq6 mutant (Dobritsa et al., 2011) 161

produces mostly tetracolpate pollen grains (Fig. 1C), and a few tricolpate grains. Pollen grains of lsq6 162

are larger (length=27.86 µm, sd=2.02 µm, n=9) than the wild-type Columbia (col) pollen grains 163

(length=23.49 µm, sd=1.12 µm, n=9). The osd1-1 (d’Erfurth et al. 2009) produces a mix of four- to 164

eight-aperturate pollen grains (Fig. 2). This genotype has a mutation in the OSD1 gene that leads to 165

the production of functional diploid pollen grains due to the absence of the second meiotic division 166

(D’Erfurth et al., 2009). These pollen grains are a little larger (length=24.25 µm, sd=2.00 µm, n=11) 167

than wt (see below) Arabidopsis pollen studied here (length=22.24 µm, sd=0.68 µm, n=9). Apertures 168

in the six- and eight-aperturate pollen grains also differ in their distribution on the surface of the 169

pollen grain compared to tricolpate pollen grains, where apertures are equatorial (Fig. 1A). In the six- 170

and eight-aperturate osd1-1 pollen, apertures form the edges of a tetrahedron (Fig. 2C) and a 171

square-based pyramid (Fig. 2E), respectively. The osd1-1 plants were obtained from the progeny of 172

osd1-1/+ plants. In this progeny, wild-type and osd1 phenotypes segregated. These wild-type plants, 173

called wild type (wt) here, were used as control plants for osd1-1 plants, as they have the same 174

genetic background (Nooseen ecotype). Another wild-type line, called col here, corresponds to the 175

Columbia ecotype (NASC N60000) and was used as a control for the inp1-1 and lsq6 mutants, which 176

both have the Columbia genetic background.

177

Other mutants presenting wall anomalies were available (for example, tam mutants from Magnard 178

et al., 2001, or the collection of mutants from Dobritsa et al., 2011), but we restricted our study to

179

6 mutants presenting normal exine and having aperture shape and size comparable to those of the 180

wild type, in order to limit the differences between mutants and controls to aperture number.

181

Growth conditions 182

Plants were grown in a climatic chamber under the following conditions: 8 hours of dark at 16°C, 16 183

hours of light at 20°C. Experiments were carried out as soon as plants produced flowers (usually 6 184

weeks after sowing). Seeds were sown during summer 2013 (in July and in August), and experiments 185

were done in September.

186

Microscopy 187

Each flower line was represented by four to eight different individuals. We used flowers picked from 188

different individuals for the experiments. These experiments lasted three weeks, since after this time 189

plants start to show signs of senescence, which could affect pollen. Open flowers were removed 190

from each plant the day before each experiment, in order to work with freshly open flowers, since a 191

majority of pollen grains are viable at this stage. On the day of the experiment, anthers from freshly 192

opened flowers were dissected and placed in a 20 µL drop of mannitol solution on a glass slide to 193

disperse pollen grains. We used one flower for each slide. Anthers were removed after a few minutes 194

and the drop with the pollen grains was gently sealed with a cover glass.

195

Mannitol is a non-metabolic sugar, which allowed us to specifically study the effects of osmolarity, 196

without affecting pollen metabolism. Three mannitol concentrations were used in our experiments:

197

0.2 mol.L

^-1

, 0.45 mol.L

^-1

and 0.7 mol.L

^-1

. These concentrations were chosen in an attempt to 198

reproduce osmolarity conditions faced by pollen grains in vivo. Because almost all pollen grains 199

explode in solutions devoid of mannitol, this condition was not used. As there are no quantitative 200

data available for stigmatic sugar concentrations, it is difficult to precisely match the osmolarity 201

conditions faced by pollen grains on the stigma. Therefore, we chose mannitol concentrations similar 202

to sugar concentrations of the previously used germination media. Since these concentrations are 203

suitable for germinations, they may resemble natural conditions occurring on stigmas. The following 204

mannitol concentrations were taken from the literature: low mannitol concentration (0.2 mol.L

^-1

) 205

corresponds to sucrose concentration used by Mouline (Mouline et al., 2002), and lies in the optimal 206

range defined by Boavida (5% to 15% or 0.15 mol.L

^-1

to 0.44 mol.L

^-1

in Boavida and McCormick, 2007).

207

The intermediate mannitol concentration (0.45 mol.L

^-1

) is close to the sucrose concentration used by 208

Li (18% or 0.53 mol.L

^-1

in Li et al., 1999), and the high mannitol concentration (0.7 mol.L

^-1

) is close to 209

the osmolarity used by Fan (about 0.68 mol.L

^-1

in Fan et al., 2001).

210

Solutions were prepared with 3 mL of buffer (see below), 6 g of Ficoll, a variable quantity of mannitol 211

(1.09 g for the 0.2 mol.L

^-1

solution, 2.46 g for the 0.45 mol.L

^-1

solution and 3.8 g for the 0.7 mol.L

^-1

212

solution), and deionized water (amount necessary to bring up solution to 30 mL). Buffer was adapted 213

from a medium routinely used to germinate pollen grains in vitro (Bergamini-Mulcahy and Mulcahy, 214

1983). This mineral salt solution contained 1.62 mmol.L

^-1

of H

3

BO

3

, 1.27 mmol.L

^-1

of Ca(NO

3

)

2

·4H

2

O, 215

0.81 mmol.L

^-1

of MgSO

4

·7H

2

O. The solution was buffered to pH 6 with 0.2 mmol.L

^-1

of KH

2

PO

4

and 216

0.05 mmol.L

^-1

of K

2

HPO

4

·3H

2

O. Ficoll is a polysaccharide that increases the viscosity of the solution 217

without changing the osmolarity. If the solution is not viscous enough, the cytoplasm leaking out of 218

broken pollen grains can disperse in the solution, and pollen with exploded cytoplasm can be 219

wrongly taken for intact pollen grains.

220

7 After 2 hours (most pollen grains rehydrate within 30 min according to our observations), 100 pollen 221

grains on each slide were scored as being either intact (having both intact wall and cytoplasm, Fig.

222

3A), having broken plasma membrane (visible as leaking cytoplasm, Fig. 3B), or having broken wall 223

(Fig. 3C). In Columbia lines (inp1-1, col and lsq6), 14 slides were counted for each concentration and 224

each genotype, which yields a total of 126 slides. In Nooseen lines (wt and osd1-1), 16 to 19 slides 225

were counted for each concentration, for a total of 104 slides.

226 227

Statistical analysis 228

R software was used to carry out statistical analysis (R Core Team, 2014). The number of intact 229

pollen grains over intact and non-intact pollen grains being a binomial variable, we used a 230

generalized linear model (GLM, which is an extension of the linear model to non-normal 231

variable) with Binomial distribution and the classical logit link function. We tested the effect of 232

genotype, osmotic concentration, and genotype-osmotic concentration interaction. To limit the 233

number of tests, we only conducted pairwise comparisons between genotypes for each osmotic 234

concentration, and between osmotic concentrations for each genotype.

235 236

RESULTS 237

Pollen grains with a broken wall were very rarely observed in the conditions tested. Therefore, for 238

each experiment we only present the proportion of intact pollen grains; the rest of the grains being 239

mostly grains with a broken plasma membrane.

240

The pollen grains of inp1-1 and lsq6 were compared to col, as they all share the Columbia genetic 241

background, while the pollen from osd1-1 was compared to wt, as they share the Nooseen genetic 242

background.

243

A GLM test compared the number of intact and damaged pollen grains between genotypes with zero 244

(inp1-1), three (col) or four (lsq6) apertures; it showed a significant effect of genotype, mannitol 245

concentration, and interaction between the two (Tab. 1). Similarly, a GLM test comparing the 246

number of intact and damaged pollen grains between the lines with three (wt) and four to eight 247

(osd1-1) apertures showed a significant effect of genotype, mannitol concentration, and interaction 248

of both (Tab. 2).

249



Influence of mannitol concentration 250

For pollen grains with zero (inp1-1), three (col, wt), four (lsq6) or four to eight (osd1-1) 251

apertures, an effect of mannitol concentration was detected (Fig. 4 and Fig. 5): the 252

proportion of intact pollen grains increases with increased mannitol concentration, showing 253

that pollen grains tend to survive better in higher mannitol concentrations (p < 0.005, Tab. 3 254

and Tab.4). This result implies that varying mannitol concentration is a useful way to study 255

pollen harmomegathy.

256

257

258

259

260

8



Influence of aperture number 261

When we compared the results for the Columbia background lines with zero (inp1-1), three 262

(col) or four (lsq6) apertures, we found that pollen mortality increases with increased 263

aperture number (p < 0.005, Tab. 3, except between inaperturate and triaperturate pollen at 264

high mannitol concentration), regardless of mannitol concentration (Fig. 4): the percentage 265

of intact pollen grains was the highest for pollen with no apertures, and the lowest for pollen 266

with four apertures. Triaperturate pollen survival rate was intermediate between those two.

267

Differences among the three genotypes are stronger at low mannitol concentration than at 268

high concentration. In general, pollen grains survive better in high mannitol concentration.

269

Therefore, it is not surprising that differences among pollen types are more obvious at low 270

concentration than at high concentration.

271 272

Similarly, pollen from Nooseen with three (wt) apertures had a higher survival rate than 273

pollen with four to eight (osd1-1) apertures (p < 0.005, Tab. 4) and this was true for all three 274

mannitol concentrations tested (Fig. 5). Here again, the contrast in mortality between the 275

two pollen types is stronger at low mannitol concentration than at high concentration: as the 276

overall mortality rate is reduced at high mannitol concentration, the differences between 277

morphs become less pronounced. Taken together, our results suggest that aperture number 278

has a negative impact on pollen survival rate under the osmotic stress conditions.

279 280

DISCUSSION 281

In nature, pollen grains are exposed to various environmental conditions, including changing 282

atmospheric humidity and rainfall events, which can affect pollen viability (Lisci et al., 1994). Also, 283

during the rehydration on stigma, pollen grains can either be immediately submerged into stigmatic 284

liquid covering the stigma surface in the case of plants with wet stigmas, or can gradually rehydrate 285

by transferring water from stigma cells in the case of plants with dry stigmas (Heslop-Harrison and 286

Shivanna, 1977; Edlund et al., 2004). Therefore, during both pollen dispersal and pollen-stigma 287

interactions, volume accommodation is a critical issue.

288

Our experiments reveal that both aperture number and mannitol concentration affect pollen 289

survival. The more apertures the pollen has, the more likely its plasma membrane to break. The 290

intensity of this effect depends on mannitol concentration, and is more pronounced at low 291

concentration. Survival rate of pollen grains is higher at high mannitol concentration than at low 292

concentration for all aperture numbers.

293

In Arabidopsis thaliana, as in most angiosperm species, pollen grains are dispersed in a dehydrated 294

state (Edlund et al., 2004), i.e. their cytoplasmic osmolarity is very high. Therefore, the difference in 295

osmolarity between pollen cytoplasm and the solutions used in our study is likely to be higher with 296

low mannitol concentration than with high concentration. At low mannitol concentration, water 297

inflow is expected to be more pronounced and may occur faster. This is consistent with our 298

observations that at low mannitol concentration plasma membrane of pollen grains tends to break 299

more easily. Moreover, some studies have shown the presence of potassium in the pollen grain 300

cytoplasm (Rehman et al., 2002, 2004), indicating that a potassium gradient might be involved in 301

pollen grain osmolarity. Potassium gradient are present in other plant cells, for example in stomata

302

9 guard cells (Talbott and Zeiger, 1996), and this cation is often associated with osmotic water 303

movements.

304

By comparing lines with zero, three, four, and four to eight apertures, we demonstrated that 305

aperture number has a clear effect on pollen survival at all mannitol concentrations and in both 306

genetic backgrounds: an increase in aperture number is associated with higher vulnerability to 307

osmotic stress. We wanted to find out if an increase in aperture number would weaken the wall, or if 308

more apertures would allow the wall to accommodate volume variations more easily, since apertural 309

areas are more flexible than exine. In most cases, dead pollen grains exhibited a broken plasma 310

membrane without visible exine breakage. Since exine rarely breaks, we can conclude that aperture 311

pattern does not affect the resistance of the exine layer, which might have some elastic properties of 312

its own (as reported in Bolick and Vogel, 1992 and Rowley and Skvarla, 2000). An increase in aperture 313

number, however, increases the number of sites where the plasma membrane can break. Therefore, 314

aperture pattern in Arabidopsis thaliana has an effect on the harmomegathic properties of pollen 315

grains, but not directly on the resistance of the exine layer.

316

Another possible reason for the reduced survival of pollen with higher number of apertures is that 317

more apertures may enable faster water inflow into pollen grains, since water enters pollen through 318

these structures (Heslop-Harrison, 1979a). This may lead to faster cytoplasmic swelling, which in 319

some cases might be too quick to allow the cell membrane to accommodate volume increase. This 320

hypothesis could be tested through observations of water inflow dynamics. Some studies have 321

shown that pollen hydration might be quick, lasting less than a second in some cases (Matsui et al., 322

1999; Rehman et al., 2002, 2004), though some of these observations were made in vitro. They still 323

point out that quick hydration is not necessarily lethal for pollen grains. Other factors can influence 324

rehydration intensity of the pollen grain, such as the type of stigma (wet or dry: Heslop-Harrison and 325

Shivanna, 1977), or the osmolarity of the stigma solution. We can also expect that rehydration 326

intensity depends on the volume of the pollen grain, since small pollen grains are likely to swell faster 327

than large pollen grains (a small volume being more affected than a large volume with the same 328

water quantity added).

329

In addition to aperture number, other pollen characteristics, such as shape of pollen grains (Muller, 330

1979; Halbritter and Hesse, 2004), morphology of apertures (Payne, 1972; Katifori et al., 2010), and 331

presence of pseudocolpori (Scotland et al., 1990; Volkova et al., 2013), have been suggested to 332

influence harmomegathy. However, these studies compared pollen from different species that 333

differed in multiple pollen characteristics, including exine thickness or patterning. In our study, 334

comparisons of Arabidopsis mutants with constant wall characteristics enabled us to primarily test 335

the effect of aperture number. We note, however, that we cannot completely rule out the effects of 336

pollen size on harmomegathy. The lsq6 and osd1-1 mutants are somewhat larger than the wild-type 337

pollen grains, and such a difference in size could also have an impact on their lower resistance to 338

osmotic stress.

339

Connection between pollen performance and aperture numbers has been studied previously by 340

looking at germination and longevity of pollen grains in Viola diversifolia, a species in which 341

individuals produce both three-aperturate and four-aperturate pollen. That study showed that the 342

four-aperturate pollen grains germinate faster, but possess reduced longevity than the three- 343

aperturate pollen (Dajoz et al., 1991, 1993). In this Viola species, there is a trade-off between

344

10 longevity (a component of survival) and competitive ability in germination. A study on another 345

heteromorphic Viola species, Viola calcarata (which produces pollen with four and five apertures), 346

has shown that the proportions of the two morphs vary according to altitude (Till-Bottraud et al., 347

1999). This result has been interpreted as an adaptation to pollinator abundance. Pollinators become 348

scarcer at high altitude, and pollination is thus uncertain. Flowers at high altitude produce a higher 349

proportion of four-aperturate pollen grains, which survive longer than the five-aperturate ones 350

(which, in turn, germinate faster). This strategy is likely to be safer in an environment with few 351

pollinators, where visits are rare and competition between pollen grains to fertilize female gametes 352

is weak. Empirical results on Viola are backed by theoretical approaches: a game theory model, 353

taking into account germination and longevity in a competitive context, has been developed, and the 354

results suggest that heteromorphism is an Evolutionarily Stable Strategy (Till-Bottraud et al., 2001).

355

Pollen resistance to osmotic stress studied here constitutes another component of survival. We thus 356

point out a new trade-off between germination, longevity, and pollen wall resistance. Triaperturate 357

pollen, the dominant type in eudicots, could result from the trade-off between the germination 358

ability, longevity, and harmomegathic properties. Since the tricolpate pollen is the only 359

morphological synapomorphy of the clade, it has been suggested that this pollen type could be the 360

key innovation of the eudicot clade (Furness and Rudall, 2004). This hypothesis is supported by the 361

studies carried out on the species-rich family Euphorbiaceae (>6700 species) and on a representative 362

set of eudicot species in which the developmental sequence leading to the formation of triaperturate 363

pollen is under strong stabilizing selection (Matamoro-Vidal et al., 2012, 2015).

364

Some species seem to have evolved particular reproductive strategies to avoid strong volume 365

changes due to intense dehydration and rehydration (Nepi and Pacini, 1993; Nepi et al., 2001). In 366

these species, pollen grains are dispersed in a partially hydrated state: their water content is above 367

30% (Nepi et al., 2001), whereas in most angiosperm species it is generally between 15% and 30%

368

(Heslop-Harrison, 1979a). These partially hydrated pollen grains do not undergo strong volume 369

decrease before anthesis, and they are able to germinate quickly on the stigma. They are, however, 370

usually short-lived, which means pollination has to occur quickly after anther dehiscence (Nepi et al., 371

2001; Franchi et al., 2002). This particular reproductive strategy appeared several times 372

independently in angiosperms (Nepi et al., 2001; Franchi et al., 2002) and could be another way to 373

achieve a compromise between the germination and survival. These species almost always have 374

porate apertures (Nepi et al., 2001; Franchi et al., 2002), suggesting a possible link between aperture 375

pattern and reproduction syndrome.

376

To conclude, we have shown that aperture number in Arabidopsis has an effect on pollen survival 377

under osmotic stress, and that pollen grains with few apertures survive better than pollen with many 378

apertures. Pollen with many apertures is probably maintained in the long run, thanks to trade-offs 379

between the survival rate and the speed of germination. Apertures are involved in several aspects of 380

pollen life, and what is favorable for reproduction is not necessarily good for survival. Many selective 381

pressures might contribute to generation of various morphologies produced during pollen 382

development. As the ancestral state of pollen grains in flowering plants is monoaperturate, the 383

number of apertures has generally increased over time. The great success of the tricolpate pollen of 384

eudicots seems to indicate that it constitutes a good compromise between fast germination and 385

survival. Thanks to the existence of mutants with defective pollen traits, the selective pressures that

386

11 have been proposed in broad-scale studies to influence the evolution of pollen morphology can now 387

be studied experimentally.

388 389

LITERATURE CITED 390

A

LBERT

, B., S. N

ADOT

, L. D

REYER

,

AND

A. R

ESSAYRE

. 2010. The influence of tetrad shape and intersporal 391

callose wall formation on pollen aperture pattern ontogeny in two eudicot species. Annals of 392

Botany 106: 557–564.

393

A

LBERT

, B., C. R

AQUIN

, M. P

RIGENT

, S. N

ADOT

, F. B

RISSET

, M. Y

ANG

,

AND A

. R

ESSAYRE

. 2011. Successive 394

microsporogenesis affects pollen aperture pattern in the tam mutant of Arabidopsis thaliana.

395

Annals of Botany 107: 1421–1426.

396

B

ERGAMINI

-M

ULCAHY

, G.,

AND

D. M

ULCAHY

. 1983. A comparison of pollen tube growth in bi- and 397

trinucleate pollen. In E. Ottaviano [ed.], Pollen: Biology and Implication for Plant Breeding., 29–

398

33. New York.

399

B

LACKMORE

, S.,

AND

S.H. B

ARNES

. 1986. Harmomegathic mechanisms in pollen grains. In S. Blackmore, 400

and I. Ferguson [eds.], Pollen and spores: forms and functions, 137–149. London.

401

B

OAVIDA

, L.C.,

AND

S. M

C

C

ORMICK

. 2007. Temperature as a determinant factor for increased and 402

reproducible in vitro pollen germination in Arabidopsis thaliana. The Plant Journal 52: 570–582.

403

B

OLICK

, M.R.,

AND

S. V

OGEL

. 1992. Breaking strengths of pollen grain walls. Plant Systematics and 404

Evolution 181: 171–178.

405

D’E

RFURTH

, I., S. J

OLIVET

, N. F

ROGER

, O. C

ATRICE

, M. N

OVATCHKOVA

,

AND

R. M

ERCIER

. 2009. Turning meiosis 406

into mitosis. PLoS Biology 7: e1000124.

407

D

AJOZ

, I., I. T

ILLBOTTRAUD

,

AND

P.H. G

OUYON

. 1993. Pollen aperture polymorphism and gametophyte 408

performance in Viola diversifolia. Evolution 47: 1080–1093.

409

D

AJOZ

, I., I. T

ILL

-B

OTTRAUD

,

AND

P.H. G

OUYON

. 1991. Evolution of pollen morphology. Science 253: 66–

410

68.

411

D

OBRITSA

,

A

.

A

.,

A

. G

EANCONTERI

, J. S

HRESTHA

,

A

. C

ARLSON

, N. K

OOYERS

, D. C

OERPER

, E. U

RBANCZYK

-W

OCHNIAK

, 412

ET AL

. 2011. A Large-Scale Genetic Screen in Arabidopsis to Identify Genes Involved in Pollen 413

Exine Production. Plant Physiology 157: 947–970.

414

D

OBRITSA

, A.

A

,

AND

D. C

OERPER

. 2012. The novel plant protein INAPERTURATE POLLEN1 marks distinct 415

cellular domains and controls formation of apertures in the Arabidopsis pollen exine. The Plant 416

cell 24: 4452–64.

417

D

OYLE

, J.A.,

AND

C. H

OTTON

. 1991. Diversification of early Angiosperms pollen in a cladistic context. In 418

S. Blackmore, and S. H. Barnes [eds.], Pollen and Spores: Patterns of Diversification, 169–195.

419

Oxford.

420

E

DLUND

, A.F., R. S

WANSON

,

AND

D. P

REUSS

. 2004. Pollen and Stigma Structure and Function: The Role of 421

Diversity in Pollination. The Plant cell 16: S84–S97.

422

12 E

RDTMAN

, G. 1952. Pollen morphology and plant taxonomy: Angiosperms (an introduction to

423

palynology). Almqvist and Wiksell, Stockholm.

424

F

AN

, L.M., Y.F. W

ANG

, H. W

ANG

,

AND

W.H. W

U

. 2001. In vitro Arabidopsis pollen germination and 425

characterization of the inward potassium currents in Arabidopsis pollen grain protoplasts.

426

Journal of Experimental Botany 52: 1603–1614.

427

F

RANCHI

, G.G., M. N

EPI

,

A

. D

AFNI

,

AND

E. P

ACINI

. 2002. Partially hydrated pollen: Taxonomic distribution, 428

ecological and evolutionary significance. Plant Systematics and Evolution 234: 211–227.

429

F

URNESS

, C.

A

,

AND

P.J. R

UDALL

. 1999. Inaperturate Pollen in Monocotyledons. International Journal of 430

Plant Sciences 160: 395–414.

431

F

URNESS

, C.

A

.,

AND

P.J. R

UDALL

. 2004. Pollen aperture evolution - A crucial factor for eudicot success?

432

Trends in Plant Science 9: 154–158.

433

H

ALBRITTER

, H.,

AND

M. H

ESSE

. 2004. Principal modes of infoldings in tricolp(or)ate Angiosperm pollen.

434

Grana 43: 1–14.

435

H

ESLOP

-H

ARRISON

, J. 1979a. An Interpretation of the Hydrodynamics of Pollen. American Journal of 436

Botany 66: 737.

437

H

ESLOP

-H

ARRISON

, J. 1979b. Pollen walls as adaptive systems. Annals of the Missouri Botanical Garden 438

66: 813–829.

439

H

ESLOP

-H

ARRISON

, Y.,

AND

K.R. S

HIVANNA

. 1977. The receptive surface of the angiosperm stigma. Annals 440

of Botany 41: 1233–1258.

441

H

ESSE

, M., M. W

EBER

, R. B

UCHNER

, A. F

ROSCH

-R

ADIVO

,

AND

S. U

LRICH

. 2009. Pollen Terminology - An 442

illustrated handbook. Springer.

443

K

ATIFORI

, E., S. A

LBEN

, E. C

ERDA

, D.R. N

ELSON

,

AND

J. D

UMAIS

. 2010. Foldable structures and the natural 444

design of pollen grains. Proceedings of the National Academy of Sciences of the United States of 445

America 107: 7635–7639.

446

L

I

, H., Y. L

IN

, R.M. H

EATH

, M.X. Z

HU

,

AND

Z. Y

ANG

. 1999. Control of pollen tube tip growth by a Rop 447

GTPase-dependent pathway that leads to tip-localized calcium influx. The Plant cell 11: 1731–

448

1742.

449

L

ISCI

, M., C. T

ANDA

,

AND

E. P

ACINI

. 1994. Pollination Ecophysiology of Mercurialis annua L.

450

(Euphorbiaceae), an anemophilous species flowering all year round. Annals of botany 74: 125–

451

135.

452

M

AGNARD

, J.-L., M. Y

ANG

, Y.-C.S. C

HEN

, M. L

EARY

,

AND

S. M

C

C

ORMICK

. 2001. The Arabidopsis Gene Tardy 453

Asynchronous Meiosis Is Required for the Normal Pace and Synchrony of Cell Division during 454

Male Meiosis . Plant Physiology 127 : 1157–1166.

455

M

ATAMORO

-V

IDAL

,

A

., C.

A

. F

URNESS

, P.H. G

OUYON

, K.J. W

URDACK

,

AND

B. A

LBERT

. 2012. Evolutionary stasis 456

in Euphorbiaceae pollen: Selection and constraints. Journal of Evolutionary Biology 25: 1077–

457

1096.

458

13 M

ATAMORO

-V

IDAL

, A., C. P

RIEU

, C.

A

. F

URNESS

, B. A

LBERT

,

AND

P.H. G

OUYON

. 2015. Evolutionary stasis in 459

pollen morphogenesis due to natural selection. New Phytologist in Press.

460

M

ATSUI

, T., K. O

MASA

,

AND

T. H

ORIE

. 1999. Rapid Swelling of Pollen Grains in Response to Floret 461

Opening Unfolds Anther Locules in Rice (Oryza sativa L.). Plant Production Science 2: 196–199.

462

M

OULINE

, K., A.A. V

ERY

, F. G

AYMARD

, J. B

OUCHEREZ

, G. P

ILOT

, M. D

EVIC

, D. B

OUCHEZ

,

ET AL

. 2002. Pollen tube 463

development and competitive ability are impaired by disruption of a Shaker K(+) channel in 464

Arabidopsis. Genes & Development 16: 339–350.

465

M

ULLER

, J. 1979. Form and Function in Angiosperm Pollen. Missouri Botanical Garden Press 66: 593–

466

632.

467

N

EPI

, M., G.G. F

RANCHI

,

AND

E. P

ACINI

. 2001. Pollen hydration status at dispersal: cytophysiological 468

features and strategies. Protoplasma 216: 171–180.

469

N

EPI

, M.,

AND

E. P

ACINI

. 1993. Pollination, Pollen viability and pistil receptivity in Cucurbita pepo.

470

Annals of Botany 72: 527–536.

471

P

AL

D

AT

. 2015. PalDat. Society for the Promotion of Palynological Research in Austria. Available at:

472

http://www.paldat.org/ [Accessed January 1, 2015].

473

P

AYNE

, W.W. 1972. Observations of Harmomegathy in Pollen of Anthophyta. Grana 12: 93–98.

474

P

ETTITT

, J.M.,

AND A

. C. J

ERMY

. 1974. Pollen in hydrophilous angiosperms. Micron (1969) 5: 377–405.

475

R C

ORE

T

EAM

. 2014. R: A language and environment for statistical computing.

476

R

EHMAN

, S., K.J. L

EE

, E.S. R

HA

, S.J. Y

UN

,

AND

J.K. K

IM

. 2002. Mechanisms involved in rapid swelling of 477

sesame ( Sesamus indicum ) pollen. New Zealand Journal of Crop and Horticultural Science 30:

478

209–214.

479

R

EHMAN

, S., E.. R

HA

, M. A

SHRAF

, K.. L

EE

, S.. Y

UN

, Y.. K

WAK

, N.. Y

OO

,

AND

J.-K. K

IM

. 2004. Does barley 480

(Hordeum vulgare L.) pollen swell in fractions of a second? Plant Science 167: 137–142.

481

R

ESSAYRE

, A., B. G

ODELLE

, C. R

AQUIN

,

AND

P.H. G

OUYON

. 2002. Aperture pattern ontogeny in 482

angiosperms. Journal of Experimental Zoology 294: 122–135.

483

R

OWLEY

, J.R.,

AND

J. S

KVARLA

. 2000. The elasticity of the exine. Grana 39: 1–7.

484

S

COTLAND

, R.W., S.H. B

ARNES

,

AND

S. B

LACKMORE

. 1990. Harmomegathy in the Acanthaceae. Grana 29:

485

37–45.

486

T

ALBOTT

, L.D.,

AND

E. Z

EIGER

. 1996. Central Roles for Potassium and Sucrose in Guard-Cell 487

Osmoregulation. Plant Physiology 111 : 1051–1057.

488

T

ILL

-B

OTTRAUD

, I., P.H. G

OUYON

, D.L. V

ENABLE

,

AND

B. G

ODELLE

. 2001. The number of competitors 489

providing pollen on a stigma strongly influences intraspecific variation in number of pollen 490

apertures. Evolutionary Ecology Research 3: 231–253.

491

14 T

ILL

-B

OTTRAUD

, I., M. V

INCENT

, I. D

AJOZ

,

AND

A. M

IGNOT

. 1999. Pollen aperture heteromorphism.

492

Variation in pollen-type proportions along altitudinal transects in Viola calcarata. Comptes 493

Rendus de l’Académie des Sciences de Paris, Sciences de la Vie 322: 579–589.

494

V

OLKOVA

, O.

A

., E.E. S

EVEROVA

,

AND

S. V. P

OLEVOVA

. 2013. Structural basis of harmomegathy: Evidence 495

from Boraginaceae pollen. Plant Systematics and Evolution 299: 1769–1779.

496

W

ALKER

, J.W.,

AND

J.A. D

OYLE

. 1975. The bases of angiosperm phylogeny, palynology. Annals of the 497

Missouri Botanical Garden 62: 664–723.

498

W

ODEHOUSE

, R.P. 1935. Pollen grains: their structure, identification and significance, in science and 499

medicine. Hafner Publishing CO, New York.

500 501 502

TABLES 503

Table 1: Summary of GLM results testing the effect of genotype (col vs inp1-1 vs lsq6), mannitol 504

concentration, and the genotype-mannitol concentration interaction, on pollen resistance to osmotic 505

stress. Threshold value: α = 0.005, sample size: N = 126.

506 507

Df Deviance Residual Df Residual deviance p-value NULL 125 2452.22

genotype 2 685.67 123 1766.55 < 1

^e

-5***

genotype+concentration 2 936.18 121 830.37 < 1

^e

-5 ***

genotype+concentration+

genotype*concentration

4 17.18 117 813.19 < 0.005 **

508 509

Table 2: Summary of GLM results testing the effect of genotype (wt vs osd1-1), mannitol 510

concentration, and the genotype-mannitol concentration interaction, on pollen resistance to osmotic 511

stress. Threshold value: α = 0.005, sample size: N = 104.

512

model Df Deviance Residual Df Residual deviance p-value NULL 103 2108.64

genotype 1 243.68 102 1864.96 < 1

^e

-5***

genotype+concentration 2 1049.83 100 815.13 < 1

^e

-5***

15 genotype+concentration+

genotype*concentration

2 66.82 98 748.31 < 1

^e

-5***

513

Table 3: Pairwise comparisons between the different plant lines (col, inp1-1, and lsq6) and different 514

mannitol concentrations (0.2 mol.L

^-1

, 0.45 mol.L

^-1

, and 0.7 mol.L

^-1

). Global threshold value: α = 0.005.

515

Threshold value for each comparison was adjusted using Bonferoni correction. Sample size: N = 126.

516

DF X

²

p-value

col-inp1-1: 0.2

1 90.25 < 1

^e

-5***

col-lsq6: 0.2

1 60.41 < 1

^e

-5 ***

inp1-1-lsq6: 0.2

1 282.37 < 1

^e

-5 ***

col-inp1-1: 0.45

1 62 < 1

^e

-5***

col-lsq6: 0.45

1 100.45 < 1

^e

-5***

inp1-1-lsq6: 0.45

1 294.16 < 1

^e

-5 ***

col-inp1-1: 0.7

1 6.75 0.0094

col-lsq6: 0.7

1 64.61 < 1

^e

-5***

inp1-1-lsq6: 0.7

1 107.47 < 1

^e

-5***

col: 0.2-0.45

1 97.56 < 1

^e

-5 ***

col: 0.2-0.7

1 346.22 < 1

^e

-5***

col: 0.45-0.7

1 96.23 < 1

^e

-5***

inp1-1: 0.2-0.45

1 67.56 < 1

^e

-5***

inp1-1: 0.2-0.7

1 152.71 < 1

^e

-5***

inp1-1: 0.45-0.7

1 21.65 < 1

^e

-5 ***

lsq6: 0.2-0.45

1 58.13 < 1

^e

-5***

lsq6: 0.2-0.7

1 353.68 < 1

^e

-5***

lsq6: 0.45-0.7

1 139.5 < 1

^e

-5***

517

518

519

16 Table 4: Pairwise comparisons between the different plant lines (wt vs osd1-1) and different mannitol 520

concentrations (0.2 mol.L

^-1

, 0.45 mol.L

^-1

, and 0.7 mol.L

^-1

). Global threshold value: α = 0.005.

521

Threshold value for each comparison was adjusted using Bonferoni correction. Sample size: N = 104 522

523

DF X

²

p-value

osd1-1-wt: 0.2

1 261.04 < 1

^e

-5***

osd1-1-wt: 0.45

1 41.25 < 1

^e

-5***

osd1-1-wt: 0.7

1 12.47 0.00041 **

osd1-1: 0.2-0.45

1 334.83 < 1

^e

-5***

osd1-1: 0.2-0.7

1 693 < 1

^e

-5***

osd1-1: 0.45-0.7

1 121.11 < 1

^e

-5***

wt: 0.2-0.45

1 65.22 < 1

^e

-5***

wt: 0.2-0.7

1 246.96 < 1

^e

-5***

wt: 0.45-0.7

1 73.44 < 1

^e

-5***

524 525

FIGURE LEGENDS 526

Fig. 1: Pollen grains of Arabidopsis thaliana with different aperture patterns, in Columbia genetic 527

background. 1-A: Tricolpate pollen with three furrow-like equatorial apertures (col). 1-B:

528

Inaperturate pollen with no apertures (inp1-1). 1-C: Tetracolpate pollen with four equatorial 529

apertures (lsq6). Confocal images of pollen grains stained with auramine O. Scale bar: 10 µm.

530

Fig. 2: Pollen of osd1-1 mutants, in Nooseen genetic background. 2-A: Lower and upper focus of a 531

tetracolpate pollen. 2-B: Lower and upper focus of a pentacolpate pollen, with five furrow-like 532

apertures. The four edging furrows in the lower focus are in continuity with the furrows of the upper 533

focus. 2-C: Lower and upper focus of a six aperturate pollen, the apertures forming the edges of a 534

tetrahedron. 2-D: Lower and upper focus of a seven aperturate pollen. 2-E: Lower and upper focus of 535

an eight aperturate pollen, the apertures forming the edges of a square based pyramid.

536

Epifluorescence microscopy with FITC filter. Scale bar: 10 µm.

537

Fig. 3: Examples of pollen responses to osmotic stress in our experiments. 3-A: Intact pollen grain. 3- 538

B: Pollen with plasma membrane broken. 3-C: Pollen with exine breakage. Scale bar: 10 µm.

539

Fig. 4: Percentage of intact pollen for the genotypes inp1-1, col and lsq6, in solutions of different 540

mannitol concentrations (0.2 mol.L

^-1

, 0.45 mol.L

^-1

and 0.7 mol.L

^-1

).

541

17 Fig. 5: Percentage of intact pollen for the genotypes wt and osd1-1, in solutions of different mannitol 542

concentrations (0.2 mol.L

^-1

, 0.45 mol.L

^-1

and 0.7 mol.L

^-1

).

543

A B C

Figure 1

Figure 2

A B C

Figure 3

20 30 40 50 60 70 80 90 100

inaperturate tri-colpate tetra-colpate

Percentageofintactpollen

0.2

col

inp1-1 lsq6

0.45 0.7

Figure 4

20 30 40 50 60 70 80 90 100

tri-colpate four to eight-colpate

Percentageofintactpollen

0.2

0.7

wt osd1-1

0.45

Figure 5