1

Pollen wall development in mango (Mangifera indica L.,

Anacardiaceae)

Jorge Lora and José I. Hormaza

Department of Subtropical Fruit Crops. Instituto de Hortofruticultura Subtropical y

Mediterránea La Mayora (IHSM la Mayora-UMA-CSIC). Avenida Dr. Wienberg, s/n. 29750

Algarrobo-Costa, Málaga (Spain)

Tel: (+34) 952 54 89 90

Fax:(+34) 952 55 26 77

✉ Jorge Lora jlora@eelm.csic.es

ORCID ID: 0000-0001-9713-0431

2

Abstract

The mango (Mangifera indica) is a woody perennial crop currently cultivated worldwide

in regions with tropical and subtropical climates. Despite its importance, an essential process such

as pollen development, and, specifically, cell wall composition that influences cross-talk between

somatic cells and the male germline, is still poorly understood in this species and in the

Anacardiaceae as a whole. A detailed understanding of this process is particularly important to

know the effect of low temperatures during flowering on pollen development that can be a limiting

factor for fertilization and fruit set. To fill this gap, we performed a thorough study on the cell

wall composition during pollen development in mango. The results obtained reveal a clear

differentiation of the cell wall composition of the male germline by pectins, AGPs and extensins

from the early developmental stages during microsporogenesis and microgametogenesis

reflecting a restricted communication between the male germline and the surrounding somatic

cells that is very sensitive to low temperatures. The combination of the results obtained provides

an integrated study on cell wall composition of the male germline in mango that reveals the crucial

role of the sporophyte and the gametophyte and the vulnerability of the process to low

temperatures.

Keywords Anacardiaceae, cell wall, low temperature, Mangifera indica, mango, pollen

development.

3

Introduction

The mango (Mangifera indica L.) is one of the five most important fruit crops worldwide

(together with bananas, oranges, grapes and apples) reaching a production of nearly 50 million

tons in 2016 (FAO 2017). Probably mango was domesticated at least 6000 years ago (Mukherjee

and Litz 2009) independently in different regions of Asia (Bompard 2009), resulting in two

different ecogeographic races: monoembryonic mangoes from subtropical India and

polyembryonic mangoes from southeastern tropical Asia. Currently, India is the largest producer

of mango, contributing 34% of the world´s total production and, together with other Asian

countries such as China, Thailand and Indonesia, account for more than 70% of global mango

production (FAO 2017). Mango cultivation has also been spread to other tropical and subtropical

regions and can be found in over 90 countries (Evans and Mendoza 2009). Systematic collection

and crosses of mango genetic resources in south Florida resulted in the release of several new

cultivars that currently dominate international trade and, in fact, Florida is now considered as a

secondary center of mango diversity (Viruel et al. 2005; Mukherjee and Litz 2009).

The genus Mangifera belongs to the family Anacardiaceae in the Sapindales with other

species of agronomic interest such as pistachio (Pistacia vera L.) or cashew (Anacardium

occidentale L.) among their 73 genera and 850 species (Bompard 2009). The genus Mangifera

contains 69 species and it is subdivided in two subgenera, Limus and Mangifera (Kostermans et

al. 1993). Due to its agronomic interest the mango has been subjected to extensive studies mainly

focused on optimizing tree management and improving yield. In the reproductive phase, several

studies have been performed on the effect of the environment on pollen germination (de Wet and

Robbertse 1986; de Wet et al. 1989; Dag et al. 2000) and flowering (Ramírez and Davenport

2010), important factors for production since mango trees show a low fruit set generally lower

than 0.1% under open pollination (Singh 1960). The mango is an andromonoecious tree with a

variable proportion of hermaphrodite and male flowers borne in terminal inflorescences which

are composed thirsoids (Coetzer et al. 1995). The number of flowers and the ratio of

hermaphrodite to male flowers varies depending on the cultivar, climatic conditions, location in

the tree, and the production in the previous season (Mukherjee and Litz 2009). Low temperatures

affect fertilization and fruit set and, as a consequence, a usual practice in countries such as Spain

and Israel, in which a first bloom occurs with too low night temperatures, is to delay the flowering

season by removing the first inflorescences produced (Dag et al. 2000; Torres et al. 2009).

However, despite the commercial importance of mango and other crops of the Anacardiaceae, an

essential process in sexual reproduction such as pollen development has been slightly studied in

this family (Maheshwari 1934; Srinivasachar 1940; Kelkar 1958; Copeland 1959; Issarakraisila

and Considine 1994; Oliveira and Mariath 2001; Huang et al. 2010). Those studies have been

mainly focused on pollen morphology from a taxonomic point of view (Ibe and Leis 1979;

Belhadj et al. 2007; Pell et al. 2010) and on pollen germination from a agronomical point of view

(de Wet and Robbertse 1986; de Wet et al. 1989; Wunnachit et al. 1992; Issarakraisila and

Considine 1994; Dag et al. 2000; Huang et al. 2010; Ramírez and Davenport 2016). Issarakraisila

and Considine (1994) were the first to report the best temperature range for mango pollen

development, showing that the phase from meiosis to early microspore development was very

sensitive to low temperatures. Similar conclusions were also obtained in a later study (Huang et

al. 2010).

Pollen development is a complex and highly conserved process in angiosperms

(McCormick 2004; Blackmore et al. 2007) that has been widely studied in model plants such as

Arabidopsis thaliana or rice (Oryza sativa) in which the molecular and genetic pathways involved

4

in pollen development have been partially revealed (Wilson and Zhang 2009; Gómez et al. 2015;

Shi et al. 2015). During pollen development, cell wall composition influences cell-cell

communication and, therefore, is likely to influence cross-talk between the somatic and the

reproductive cells. It is specifically important during microsporogenesis (the transition from

somatic to the germ line cell, the arquesporial cell) and later, when the arquesporial cell increases

in size becoming the microspore mother cell, that undergoes two meiotic divisions to form four

haploid microspores (microspore tetrad). During this period an intense cell-cell communication

takes place between the cells of the male germ line and the surrounding somatic cells. Moreover,

the pollen cell wall is the most complex structure among plant cell walls with an outer and inner

walls, the exine and the intine, respectively (Blackmore et al. 2007). The exine is more complex

and consists of an outer layer, sexine, and an inner nexine (Heslop-Harrison 1968). The exine

generally includes material from the tapetum and is composed mostly by a combination of

polymers that forms the sporopollenin (Blackmore et al. 2007; Quilichini et al. 2015). Recent

studies have reported that the inner layer, nexine, is composed of arabinogalactan proteins (AGPs)

in Arabidopsis thaliana (Lou et al. 2014, Jia et al. 2015). Indeed, previous functional studies

identified the genes encoding AGPs and their essential role in pollen development in Arabidopsis

(Levitin et al. 2008; Coimbra et al. 2009) that were also observed around cells in the reproductive

lineage and in the intine (Coimbra et al. 2007). AGPs have also been observed in the intine of an

eudicot, Quercus suber (Costa et al. 2015), and of the early divergent Trithuria submersa (Costa

et al. 2013).

Other component of the pollen cell wall are cellulose and non-cellulosic polysaccharides

such as pectins that are the main components of the intine (Heslop-Harrison 1968; Blackmore et

al. 2007), and also of the plant cell wall as a whole (Palin and Geitmann 2012). Pectins together

with callose, a polysaccharide that is also found around the microspore mother cells and may act

as a molecular filter, have also been revealed to play an important role on the apertures and on

the pollen dispersal unit in the genus Annona (Lora et al. 2009, 2014).

While the pollen development process has been described in several plant species,

detailed studies on changes in cell wall components during pollen development are scarce

(Coimbra et al. 2007; Lora et al. 2009, 2014, Costa et al. 2013, 2015) and, to the best of our

knowledge, absent in the Anacardiaceae. Thus, in this work we perform an integrated study on

pollen development in M. indica, giving special attention to the cell wall composition, and

specifically to pectins, AGPs, extensins and cellulose. Moreover, we evaluate the effect of low

temperatures on the cell wall composition during pollen development of the first bloom under

Mediterranean conditions.

Material and methods

Plant material

Adult trees of two M. indica cultivars, M. indica `Kensington´ and M. indica `Kent´ located in a

field germplasm collection at the IHSM La Mayora-CSIC, Málaga, Spain, were used for this

work, which was performed during the first and second flowering periods in March and May,

respectively, of 2017.

Light microscopy

5

To follow pollen development, we collected anthers from flower buds at a range of developmental

stages, from differentiation up to anther dehiscence. The anthers were fixed in 2.5% v/v glutaraldehyde in 0.03 M phosphate buffer (Sabatini et al. 1963), dehydrated in an ethanol series,

embedded in Technovit 7100 (Kulzer & Co, Wehrheim, Germany), and sectioned at 2 μm.

For general histological observations, sections were stained with 0.5% (w/v) periodic acid for 2 h, washed three times with water and maintained in Schiff’s reagent in the dark for 1.5 h

(Feder and O’Brien 1968). After three washes with water, the sections were stained with aqueous

0.2% (w/v) toluidine blue. Sections were also stained for cellulose with 0.007% calcofluor in

water (Hughes and McCully 1975).

Immunocytochemistry

Anthers were fixed in 4% v/v paraformaldehyde in phosphate-buffered saline (PBS) at pH 7.3,

left overnight at 4 oC, dehydrated in an acetone series, embedded in Technovit 8100 (Kulzer),

polymerized at 4 oC, and sectioned at 2 μm. Sections were placed in a drop of water on a slide covered with 2% v/v 3-aminopropyltrietoxy-silane in water (Sigma, St. Louis, Missouri, USA)

and dried at room temperature (Satpute et al. 2005; Solís et al. 2008).

Different antibodies were used to localize specific cell components: JIM7 and JIM5 rat

monoclonal antibodies (Carbosource Service, University of Georgia, Athens, Georgia, USA), which recognize methyl-esterified and unesterified pectins, respectively (Knox 1997); JIM8

(Pennell et al. 1991) and JIM13 (Knox et al. 1991) for arabinogalactan proteins (AGPs)

(Carbosource Service), JIM11 (Carbosource Service) for extensins (Smallwood et al. 1994) and an anti-callose mouse monoclonal antibody (Biosupplies, Parkville, Australia) for callose.

Following the protocol of Lora et al (2009) and Solís et al (2008), sections were incubated

with PBS for 5 min and later with 5% w/v bovine serum albumin (BSA) in PBS for 5 min. Then,

different sections were incubated for 1 h with the primary antibodies: JIM5, JIM7, JIM8, JIM11 and JIM13 undiluted and anti-callose diluted 1/20 in PBS. After three washes in PBS, the sections

were incubated for 1 h in the dark with the corresponding secondary antibodies (anti-rat for JIM5,

JIM7, JIM8, JIM11 and JIM13, and anti-mouse for anti-callose) conjugated with Alexa 488 fluorochrome (Molecular Probes, Eugene, Oregon, USA) and diluted 1/25 in PBS. After three

washes in PBS, the sections were stained with 4 ,́6-diamidino-2-phenylindole (DAPI, 0.1 mg/mL)

and washed three times in PBS. The sections were mounted in ProLong Gold Antifade Reagent (Invitrogen), examined with a Leica TCS SP5 II confocal microscope and with a Leica DM2500

epifluorescence microscope equipped with a Leica DFC310 FX camera. Filters were 470/525 nm

for the Alexa488 fluorescein label of the antibodies. Overlapping photographs were obtained with

the Leica Acquisition Station AF6000 E software.

Results

Microsporogenesis

1. Specific labelling of the microspore mother cell

We first evaluated the initial stages of microsporogenesis using semithin sections that were

stained with periodic acid-Shiff’s reagent (PAS) and toluidine blue. The early primordial stamens and ovules showed a cytoplasm with strong PAS staining (Fig. 1A). When the filament reached

about 131 ± 22 µm (n = 7) long, the cytoplasm of microsporangium cells showed strong PAS

staining (Fig. 1B). During the early stages of anther development, the hypodermal archesporial cell divided by a periclinal wall to produce an inner primary sporogenous cell and an outer primary

parietal cell. Those two cells underwent several mitotic division and the primary sporogenous cell

produced microspore mother cells whereas the primary parietal cell produced somatic cell layers

of the anther wall (binucleate tapetum, endotecium and epidermis). The morphology and the size of the somatic and reproductive cells were initially similar. The microspore mother cells were

only clearly observed when their size started to increase (Fig. 1C-D). Concomitantly, the tapetum

6

formed by the sporophytic binucleate cells surrounding the microspore mother cells underwent

vacuolization (Fig. 1D). To evaluate the main cell wall components during microsporogenesis, AGPs and

extensins were labelled using the monoclonal antibodies JIM5 and JIM7 for unesterified and

methyl-esterified pectins, respectively, JIM8 and JIM13 for AGPs and JIM11 for extensins. The microspore mother cell wall was clearly distinct from the surrounding somatic tapetum cells by

reacting positively to JIM5, JIM8, JIM13 and JIM11 (Fig. 2B, E, H and supplementary Fig. 1E).

The methyl-esterified pectins JIM7-specific labelling was equally observed in the tapetum and

the reproductive cells (Supplementary Fig. 1B). AGPs reacted positively to JIM13 and JIM8 in the cell wall of the microspore mother cell (Fig. 2F and supplementary Fig. 1F).

Callose, a polysaccharide that acts as a physical filter of molecules (Tucker et al. 2001),

was also observed by immunodetection and reacted positively to anti-callose antibodies around the microspore mother cell prior to meiosis (Fig. 2L).

Cellulose is also a main component of the plant cell wall (Cosgrove 2005). Indeed, the

primexine is mainly composed of cellulose and is essential for the cell wall patterning (Blackmore et al. 2007). Thus, cellulose staining with calcofluor was also evaluated during pollen

development. Cellulose was not observed in the early microspore mother cells but was observed

around them at later developmental stages and concomitant with the presence of callose

(Supplementary Fig.2).

2. Partial isolation of the reproductive cells during meiosis by a special cell wall

The microspore mother cell underwent meiosis followed by simultaneous cytokinesis (Fig. 3A,

C, E and H). The immunodetection and the cellulose staining revealed callose, AGPs (mainly

detected by JIM13) and cellulose around the meiotic products (Fig. 3A, C, and supplementary

Fig. 2D) but they were gradually reduced during the release of the microspores although some remnants of callose and cellulose could be observed around the microspores (Fig. 3B and

supplementary Fig. 2E). Pectins (JIM5 and JIM7), AGPs (JIM8) and extensins (JIM11) reacted

positively to their specific antibodies in the cell wall of the microspore mother cell during meiosis (Fig. 3E, H and supplementary Fig. 3A, D) but, while pectins and extensins did not react to these

antibodies in the cell wall of the connected microspores during simultaneous cytokinesis, they

were detected in the cell wall of the released microspores (Fig. 3E-J and supplementary Fig. 3A-C). Later, when the wall of the microspore mother cell was absent, the signal against AGPs,

pectins and extensins remained mainly at the aperture sites of the microspores (Fig. 3G, I, J and

supplementary Fig. 3C, E). Insoluble carbohydrates stained with PAS were also observed around

the microspore mother cells during the simultaneous citokinesis and some remained in the aperture site of the microspores (Fig. 4A-C). The microspore mother cell was, therefore,

surrounded by a special cell wall formed by callose, cellulose, AGPs and insoluble carbohydrates

and, at the cell periphery, by pectins.

Microgametogenesis

The microspores underwent vacuolization after their release and the nucleus was displaced to the

lateral side of the microspores, in which the first mitosis took place (Fig. 4D). In this

developmental stage, the intine showed a darker blue color, that allowed a clear differentiation

from the exine at later developmental stages (Fig. 4D-F). Following the first mitosis, the vacuoles decreased in size and, concomitantly, accumulation of starch grains took place (Fig. 4E). The

starch grains were observed close to anther dehiscence while most of the pollen grains were

starchless at anther dehiscence (Fig. 4F). The immunodetection of the cell wall components during microgametogenesis revealed

the composition, mainly, of the intine. Additionally, the nuclei were observed using 4 ,́6-

diamidino-2-phenylindole (DAPI) that showed autofluorescence of the exine but in a different

fluorescence wavelength (blue color) than the fluorescent marker of the AlexaFluor 488 antibodies, which emitted green fluorescence (Fig. 5). During microgametogenesis, unesterified

and methyl-esterified pectins reacted positively to JIM5 and JIM7 respectively, and they were

7

observed at the aperture sites. However, while the JIM5 signal was mainly observed at the mature

stage throughout the intine (Supplementary Fig. 4B-C), JIM7 signal was weakly observed in the intine at the vacuolate stage (Fig. 5A) and the signal was more prominent in starchy pollen (Fig.

5B) and in mature pollen at anther dehiscence (Fig. 5C). JIM7 signal was also observed around

the generative cell at the periphery of the pollen grain (Fig. 5B), and in the cytoplasm of the pollen grain at anther dehiscence (Fig. 5C). Similarly, the intine also reacted to JIM13 and JIM8 (AGPs)

at the vacuolate stage and at later stages (Fig. 5D-F and supplementary Fig. 4D-F). Based on the

new studies in Arabidopsis (Lou et al. 2014, Jia et al. 2015), this signal showing the AGPs JIM13

and JIM8-specific labelling may also derive in mango from the inner layer of the nexine. The pattern of extensins was similar to the JIM5 signal, with a stronger JIM11 signal in the cell wall

of the starchy pollen, in which it was also observed around the generative cell (Fig. 5G-I).

Cellulose was observed in the epidermis, endotecium and tapetum but was faintly observed in the vacuolate microspores or in pollen, once the first mitosis had taken place (Supplementary Fig.

2F). Cellulose was more apparent in the intine at later developmental stages (Supplementary Fig.

2G-H). Thus, the intine was composed of AGPs (JIM8 and JIM13) and methyl-esterified pectins (JIM7) at the earlier stages and cellulose, pectins (JIM5 and JIM7), AGPs (JIM8 and JIM13) and

extensins (JIM11) at later stages.

Pollen development was anomalous under low temperatures

Under the climatic conditions of the Mediterranean coast of Southern Spain, the first bloom in

mango took place in March with an average daily temperature of 14.8 oC and reaching a minimum value of 9.4 oC at night. Conversely, the average daily temperature during the second bloom in

May was 19.6 oC with a minimum temperature of 14.7 oC. To study the effect of low temperatures

on pollen development during the first bloom, we first evaluated pollen development using

semithin sections that were stained with PAS and toluidine blue. Microsporogenesis followed a similar developmental pattern as that observed in the second bloom. However, differences were

reported during microgametogenesis and they were more prominent at later developmental stages.

The mature pollen showed numerous morphological anomalies such as a size 2-3 times larger than normal, irregular shape and/or unreleased microspores that resulted in grouped pollen (Fig.

6A-B). The cell wall components are key players in the morphology of plant cell and specifically

in the complex pollen grain wall. Indeed, a recent study using an experimental approach reported the key role of the pollen cell wall on survival and maintenance of pollen integrity (Matamoro-

Vidal et al. 2016). Therefore, we also performed a similar immunolocalization approach as that

explained above. Similarly, cell wall composition was unaffected during microsporogenesis and

the main anomalies were observed during microgametogenesis, particularly, in the intine, which was absent in several pollen grains (Fig. 6C-G).

Discussion

The results obtained on the transition from somatic to reproductive cells, microspore development, and pollen release and behavior at low temperatures in mango show a conserved

pollen development process, that has been slightly studied in the Anacardiaceae.

The reproductive cells are clearly differentiated from the surrounding somatic cells at the

early stages of development.

During early stages of anther development, the meristematic cells from the microsporogenous

tissue seem to be in a similar stage of differentiation showing dense PAS reagent-positive

cytoplasm. Soon the microspore mother cell increases in size and is clearly differentiated from

the neighbouring cells. Moreover, the cell wall of the microspore mother cell is also clearly identified by JIM5 (unesterified pectins), JIM8 and JIM9 (AGPs) and JIM11 (extensins). A

similar labelling pattern has also been observed during microsporogenesis in other tropical woody

perennial species, although phylogenetically distant from M. indica, such as the early divergent

8

Annona genus (Lora et al. 2009, 2014). AGPs were also observed in the cell wall of the microspore

mother cell in the model plant Arabidopsis thaliana (Coimbra et al. 2007) and Quercus suber (Costa et al. 2015). In the early stages of microsporophyte development, the tapetum cells were

generally binucleated in M. indica; this polinucleated feature, which is not very common in seed

plants, is generally found in the tapetum and also in the nucellus of some gymnosperms and in the endosperm of many angiosperms (D’Arcy 1996; Furness and Rudall 2001).

Male germline cells are not only differentiated but also partially isolated from the somatic

cells during meiosis

After this first clear differentiation of the microspore mother cell from the somatic cells of the

sporophyte, several components are deposited around the microspore mother cell, constituting a special cell wall. Among them, callose, a β1,3-glucan, is the most extensively reported in many

angiosperms families (Mascarenhas 1989; McCormick 1993; Hafidh et al. 2016), although some

exceptions have been shown (Periasamy and Amalathas 1991). In addition to callose, we also observed the co-localisation of AGPs revealed by JIM8 and JIM13. Although they have not been

reported as widely as callose, AGPs have also been observed around the microspore mother cell

during meiosis in some angiosperms such as Beta vulgaris (Nothnagel et al. 2001; Majewska-

Sawka and Rodriguez-Garcia 2006) and Arabidopsis thaliana (Otegui and Staehelin 2004). Moreover, cellulose has also been observed around the microspore mother cell in some species

such as several of the genus Annona (Lora et al. 2009, 2014)

The partial self-isolation of the microsporocyte from the sporophytic tapetum is crucial for its development and related to meiosis. M. indica undergoes simultaneous cytokinesis in which

the four microspores are not separated until the two meiosis are complete. Simultaneous division

was also reported in other species of the Anacardiaceae such as Schinus molle (Copeland 1959),

Semecarpus anacardium (Srinivasachar 1940) or Anacardium occidentale (Oliveira and Mariath 2001). After meiosis, this transient callose wall is digested by an enzyme cocktail from a

degenerating tapetum, a common feature in species with a secretory tapetum (Pacini et al. 1985;

Pacini 2010). The other components of the special cell wall are also digested at the same time by enzymes from the tapetum. The special cell wall is also deposited in particular areas of the

microspore cell wall that blocks the deposition of the primexine, a microfibrilar material with the

sporopollenin precursor (Blackmore et al. 2007). This special area will form the apertures that are essential for the absorption of nutrients from the sporophyte and for pollen germination on the

stigma (Pacini et al. 1985; Furness and Rudall 2004). After the microspores are released from the

tetrad stage, our results show the presence of remnants of the special cell wall with pectins, AGPs

and extensins at the aperture region and they can also be slightly observed in the young intine. These remnants of the special cell wall determine the aperture pattern of M. indica that becomes

tricolpate at anther dehiscence, a common aperture type found in the pollen of eudicots

(Blackmore and Crane 1998), including Anacardiaceae (Kelkar 1958; Copeland 1959; Ibe and Leis 1979; Oliveira and Mariath 2001; Pell et al. 2010). Similarly, JIM8 and JIM13-specific

labelling were observed at the aperture sites in Trithuria (Costa et al. 2013), a basal angiosperm.

Interestingly, a recent study performed in Arabidopsis revealed the presence of AGPs in the inner layer of the exine, the nexine, in young microspores (Jia et al. 2015).

After the partial isolation during meiosis, the role of the sporophyte is more conspicuous

during microgametogenesis

The tetrad stage reveals the dominant role of the microspore on the restricted communication with

the somatic sporophyte that is reflected in a controlled meiosis and an established pattern of the cell wall. The contribution of the sporophytic tapetum to pollen development becomes more

relevant after this stage with clear consequences in pollination (Hesse 2000; Pacini and Hesse

2004) and pollen-pistil interaction in the stigma (Lora et al. 2016). In fact, besides its implication

in the nutrition of the developing microspores, the tapetum has a clear role in the deposition of material in the exine (Heslop-Harrison 1975a).

9

While the role of the sporophyte is essential for the formation of the outer layer of pollen

grains, it is reduced in the formation of the inner layer, the intine, a pectin-cellulosic layer produced by the released microspores from the tetrads (Heslop-Harrison 1975b; Blackmore et al.

2007). In this sense, and concomitant with the rest of the special cell wall in the aperture region,

the young intine of M. indica reacted with antibodies against pectins, AGPs and extensins as well as with calcofluor which revealed a blue cellulosic intine in the vacuolated pollen. However, the

pectin signal was reduced or even absent further in pollen development but becomes evident again

at later stages. A similar pattern of intine development was also observed in the Annonaceae (Lora

et al. 2009, 2014) and the thinning out of the pectins could be related to the substantial increase of the surface area of the pollen grain.

Low temperatures affect the normal pollen development process in mango

Pollen development in mango was critically compromised in the first bloom under Mediterranean

conditions with low night temperatures. While the vulnerability of the sexual reproductive phase by rising temperatures has been shown in a range of species and constitutes a clear threat under a

global warming scenario (Hedhly et al. 2009), the effect of low temperatures has also been

reported as critical (Hedhly 2011). In mango, low temperatures during meiosis causes meiotic

chromosomal irregularities (Huang et al. 2010) and low pollen viability (Issarakraisila and Considine 1994; Huang et al. 2010). Issarakraisila and Considine (1994) established an optimum

range of temperature in this species for pollen development during the phase from meiosis to

young microspores ranging from 15oC to 33oC. Temperatures during the first bloom under Mediterranean conditions are out of this range with an average of 14.8 oC and a minimum of 9.4 oC. Under these conditions, we observed numerous anomalies during pollen development and,

interestingly, cell wall composition, and mainly the intine, was also affected. These anomalous

pollen grains were mainly reported during microgametogenesis and more prominently at anther dehiscence, but these anomalies could be the result of earlier abnormal development. Indeed, the

pattern of pollen cell wall has been established mainly from meiosis to the young microspore

stage (Heslop-Harrison 1968; Hess 1993; Jia et al. 2015) and coincident with the phase most sensitive to temperature. Moreover, this pattern of pollen cell wall that determines cell shape

could also be behind the irregular shape reported in the pollen of mango.

Conclusion

Our results in M. indica show a conserved pollen development process that has been

slightly studied in the Anacardiaceae. First, the transition of somatic to reproductive cells revealed a clear differentiation by pectins, AGPs and extensins. Second, a self-partial isolation of the

microspore mother cell determined a crucial restricted communication during meiosis and cell

wall patterning, a process highly influenced by the tapetum. Third, the intine was mainly composed of cellulose, pectins, AGPs and extensins, a process highly influenced by the male

gametophyte. And finally, this pollen development was critically affected by low temperatures.

The combination of these results provide an integrated study on cell wall composition of the male germline in mango that reveal the crucial role of the sporophyte and the gametophyte and its

vulnerability to low temperatures.

Author contribution statement JL and JIH conceived and designed the study project. JL

performed the experiments. JL and JIH wrote, reviewed and edited this manuscript.

Acknowledgments This work was supported by Ministerio de Economía y Competitividad—

European Regional Development Fund, European Union (AGL2015-74071-JIN, AGL2016-77267-R). Distribution of the JIM antibodies used in this work was supported in part by NSF

grants DBI-C421683 and RCN009281.

10

Compliance with ethical standards

Conflict of interest The authors declare that they have no conflict of interest.

Figure legends

Fig. 1 Microsporogenesis in Mangifera indica revealed with periodic acid-Shiff’s reagent (PAS)

and toluidine blue. A Anther and ovule primordia showing dense PAS reagent-positive cytoplasm.

B Early stage of anther development showing a microsporogenous tissue with dense PAS reagent-

positive cytoplasm (asterisk). C Microsporogenous tissue with undifferentiated cells. D

Microspore mother cells surrounded by a binucleate tapetum showing cytoplasmic vacuolization

and the outer layers of the anther wall, the endotecium and epidermis. Scale bars correspond to

25 µm in A, C, D; 100 µm in B. a anther primordia; e epidermis, en endotecium, f filament, m

microspore mother cell, o ovule primordia, t tapetum.

Fig. 2 Microsporogenesis in Mangifera indica revealed by specific labelling of cell wall

components with monoclonal antibodies; JIM5 for unesterified pectins (A-C), JIM13 for

arabinogalactans (AGPs, D-F), JIM11 for extensins (G-I) and anti-callose for callose (J-L). A

JIM5-specific labelling (unesterified pectins) was equally observed in the anther primordium. B

The microspore mother cells reacted positively to JIM5 but no JIM5 labelling was observed in

the surrounding tapetum. C The microspore mother cells continued to increase in size and

maintain a visible JIM5-specific labelling. D JIM13-specific labelling (AGPs) was homogenously

observed in the cell wall of the anther primordium. E JIM13-specific labelling was observed in

the microspore mother cells wall and undetected in the surrounding tapetum. F Later, intense

JIM13-specific labelling was observed around the microspore mother cells. G JIM11-specific

labelling (extensins) was homogenously observed in the anther primordium. H JIM11-specific

labelling was more intensely observed in the microspore mother cell walls compared to the cell

walls of the tapetum. I Similarly, the differential fluorescence signal of JIM11 was conserved

when the microspore mother cells increase in size. J Anti-callose labelling was not observed in

the anther primordium. K Anti-callose labelling was also not observed during the early

development stages of the microspore mother cells. L Later, anti-callose was observed around the

microspore mother cell. All scale bars = 25 µm. m microspore mother cell, t tapetum.

Fig. 3 The tetrad stage and the release of the microspores in Mangifera indica revealed by specific

labelling of cell wall components with monoclonal antibodies; anti-callose for callose (A-B),

JIM13 for arabinogalactans (AGPs, C-D), JIM5 for unesterified pectins (E-G) and JIM11 for

extensins (H-J). A Anti-callose labelling was prominently observed around the four microspores

during the simultaneous cytokinesis (asterisk). B After the microspore tetrad release, weak anti-

callose fluorescence signals (asterisks) was visible between the microspores. C The JIM13-

specific labelling (AGPs) was more conspicuous around the tetrad. D Remnants of the JIM13

fluorescence signal were also observed at the aperture sites (arrows). E The JIM5-specific

labelling (unesterified pectins) was observed in the microspore mother cell wall during

cytokinesis (asterisk). F Later in development, JIM5 fluorescence signal was also observed in the

microspore wall (arrows). G After the release of the microspores from the tetrads, JIM5

fluorescence signal was observed at the aperture site and in the early intine (arrows). H The

JIM11-specific labelling (extensins) was observed in the microspore mother cell wall during

cytokinesis (asterisk). I-J JIM11 fluorescence signal was later observed at the aperture sites and

11

in the intine (arrows), just after the disappearance of the cell wall of the microspore mother cell

(I) and after the release of the microspores (J). All scale bars = 25 µm. t tapetum.

Fig. 4 Citokinesis (A-B) and microgametogenesis (C-F) in Mangifera indica. Semithin sections

of anthers were stained with periodic acid-Shiff’s reagent (PAS) and toluidine blue. A-B Insoluble

carbohydrates were observed around the microspore mother cells before (A) and during the

simultaneous citokinesis (B). C The aperture site showed a dense staining of insoluble

carbohydrates (arrows) that was also faintly observed (asterisks) around the microspores after

their release. D Later, the microspore underwent cytoplasmic vacuolization and the nucleus was

displaced laterally (asterisk). The first mitosis (arrow) was also observed in the vacuolated pollen

grain. E At later developmental stages, the pollen grain showed a central vegetative nucleus

(white asterisk) and a lateral generative nucleus (black asterisk); the pollen grain also showed

starch grains stained in dark blue. F Number of starch grains (dark blue) decreased at anther

dehiscence. All scale bars = 25 µm. c cytokinesis, m microspore mother cell, t tapetum, v vacuole.

Fig. 5 Microgametogenesis in Mangifera indica revealed by specific labelling of cell wall

components with monoclonal antibodies; JIM7 for methyl-esterified pectins (A-C), JIM13 for

arabinogalactans (AGPs, D-F) and JIM11 for extensins (G-I). A JIM7-specific labelling (methyl-

esterified pectins) was faintly observed in the intine (white arrow) of the young pollen grains. B

The intine reacted more intensely against JIM7 at later developmental stages and specifically at

the aperture sites (white arrow). The generative nucleus was also surrounded by JIM7-specific

labelling (asterisk). C JIM7 signal was observed in the intine (white arrow) and in the cytoplasm

at anther dehiscence. D-F The intine showed JIM13 fluorescence signal (AGPs) at early pollen

grain developmental stages (D) and continued at later stages (E) up to anther dehiscence (F) and,

more prominently, at the aperture sites. E-F The wall of the generative nucleus (asterisks) reacted

to JIM13. G JIM11-specific labelling (extensins) was not observed in the young pollen grains. H

The intine of the pollen grains reacted to JIM11 (extensins) and more prominently in the aperture

site. JIM11 signal was also detected around the generative nucleus (asterisks). I The intine showed

a faint signal of JIM11 at anther dehiscence. A-I The exine was observed in blue

(autofluorescence, blue arrows). All scale bars = 25 µm. v vacuole.

Fig. 6 Cytology of anomalous pollen developed at low temperatures in Mangifera indica. A-B

Anther showing anomalous pollen grains such as unusual sizes and shapes (asterisk) (A) or

without exine (arrow) (B) at later developmental stages. C-D Anomalies were also observed in

the cell wall composition: JIM7-specific labelling for methyl-esterified pectins (arrow) (C) and

JIM5-specific labelling for unesterified pectins (arrow) (D) in the intine were only observed in

some pollen grains at later developmental stages. E-F Similar anomalies were also observed with

the JIM8 (E) and JIM13 (F) specific labelling (antibodies for arabinogalactans, arrows). G The

staining of cellulose by calcofluor also revealed anomalous intine development. Scale bars

correspond to 25 µm in A-F; 50 µm in G. v vacuole.

Supplementary figures

Supplementary Figure 1 JIM7-specific labelling for methyl-esterified pectins (A-C) and JIM8-

specific labelling for arabinogalactans (AGPs, D-F) during Mangifera indica microsporogenesis.

Supplementary Figure 2 Visualization of cellulose using calcofluor during microsporogenesis

and microgametogenesis in Mangifera indica.

12

Supplementary Figure 3 Visualization of JIM7-specific labelling for methyl-esterified pectins

(A-C) and JIM8-specific labelling for AGPs (D-E) during microspore development in Mangifera

indica.

Supplementary Figure 4 JIM5-specific labelling for unesterified pectins (A-C) and JIM8-

specific labelling for arabinogalactans (AGPs, D-F) during microgametogenesis in Mangifera

indica.

References

Belhadj S, Derridj A, Civeyrel L, Gers C, Aigouy T, Otto T (2007) Pollen morphology and

fertility of wild Atlas pistachio (Pistacia atlantica Desf., Anacardiaceae). Grana 46:148–156

Blackmore S, Crane PR (1998) The evolution of apertures in the spores and pollen grains of

Embryophytes. In: Owens JS, Rudall PJ (eds) Reproductive biology. Royal Botanic Gardens, Kew, pp 159–182

Blackmore S, Wortley AH, Skvarla JJ, Rowley JR (2007) Pollen wall development in flowering plants. New Phytol 174:483–498

Bompard JM (2009) Taxonomy and systematics. In: Litz RE (ed) The mango: botany, production and uses. CABI, Wallingford, pp 19–41

Coetzer LA, Oosthuyse SA, Wishart DL, Robbertse PJ (1995) Influence of pruning on the

flower sex ratio in some mango cultivars. South African Mango Grow Assoc Yearb 15:26-30

Coimbra S, Almeida J, Junqueira V, et al (2007) Arabinogalactan proteins as molecular markers in Arabidopsis thaliana sexual reproduction. J Exp Bot 58:4027–4035

Coimbra S, Costa M, Jones B, et al (2009) Pollen grain development is compromised in Arabidopsis agp6 agp11 null mutants. J Exp Bot 60:3133–3142

Copeland H (1959) The reproductive structure of Schinus molle (Anacardiaceae). Madroño 15:14–25

Cosgrove DJ (2005) Growth of the plant cell wall. Nat Rev Mol Cell Biol 6: 850–861

Costa M, Pereira AM, Rudall PJ, Coimbra S (2013) Immunolocalization of arabinogalactan proteins (AGPs) in reproductive structures of an early-divergent angiosperm, Trithuria (Hydatellaceae). Ann Bot 111:183–190

Costa ML, Sobral R, Ribeiro Costa MM, et al (2015) Evaluation of the presence of

arabinogalactan proteins and pectins during Quercus suber male gametogenesis. Ann Bot 115:81–92

D’Arcy WG (1996) Anthers and stamens and what they do. In: D´Arcy WG, Keating RC, (eds)

The anther: form, function, and phylogeny. Cambridge University Press, Cambridge, pp 1–25

Dag A, Eisenstein D, Gazit S (2000) Effect of temperature regime on pollen and the effective pollination of “Kent” mango in Israel. Sci Hortic 86:1–11

de Wet E, Robbertse PJ (1986) A preliminary study of the pollen of Mangifera indica L. cv.

13

Haden in South Africa. South African J Plant Soil 3:87–89

de Wet E, Robbertse PJ, Groeneveld HT (1989) The influence of temperature and boron on pollen germination in Mangifera indica L. South African J Plant Soil 6:228–234

Evans EA, Mendoza OJ (2009) World mango trade and the economics of mango production. In: Litz RE (ed) The mango: botany, production and uses. CABI, Wallingford, pp 606–627

FAO (2017) Agricultural data: FAOSTAT. http://www.fao.org/faostat/en/?#data/QC. Accessed 20 Feb 2017

Feder N, O’Brien TP (1968) Plant microtechnique: some principles and new methods. Am J Bot 55:123–142

Furness CA, Rudall PJ (2001) The tapetum in basal angiosperms: early diversity. Int J Plant Sci 162:375–392

Furness CA, Rudall PJ (2004) Pollen aperture evolution - a crucial factor for eudicot success? Trends Plant Sci 9:154–158

Gómez JF, Talle B, Wilson ZA (2015) Anther and pollen development: a conserved developmental pathway. J Integr Plant Biol 57:876–891

Hafidh S, Fíla J, Honys D (2016) Male gametophyte development and function in angiosperms: a general concept. Plant Reprod 29:31–51

Hedhly A (2011) Sensitivity of flowering plant gametophytes to temperature fluctuations. Environ Exp Bot 74:9–16

Hedhly A, Hormaza JI, Herrero M (2009) Global warming and sexual plant reproduction. Trends Plant Sci 14:30–36

Heslop-Harrison J (1968) Pollen Wall Development. Science (80- ) 161:230–237

Heslop-Harrison J (1975a) The adaptive significance of the exine. In: Ferguson JK, Muller J

(eds) The evolutionary significance of’ the exine, Linnean Society of london, London, pp 27–37

Heslop-Harrison J (1975b) Incompatibility and the pollen-stigma interaction. Ann Rev Plant Physiol 26:403–425

Hess M (1993) Cell-wall development in freeze-fixed pollen: intine formation of Ledebouria socialis (Hyacinthaceae). Planta 189:139–149

Hesse M (2000) Pollen wall stratification and pollination. Plant Syst Evol 222:1–17

Huang J-H, Ma W-H, Liang G-L, Zhang L-Y, Wang W-X, Cai Z-J, Wen S-X (2010) Effects of low temperatures on sexual reproduction of “Tainong 1” mango (Mangifera indica). Sci Hortic 126:109–119

Hughes J, McCully ME (1975) Use of an optical brightener in study of plant structure. Stain Technol 50:319–329

Ibe RA, Leis RA (1979) Pollen morphology of the Anacardiaceae of northeastern north America. Bull Torrey Bot Club 106:140-144

Issarakraisila M, Considine J (1994) Effects of temperture on pollen viability in Mango cv. `Kensington .́ Ann Bot 73:231–240

14

Jia Q-S, Zhu J, Xu X-F, Lou Y, Zhang Z-L, Yang Z-N (2015) Arabidopsis AT-hook protein

TEK positively regulates the expression of arabinogalactan proteins for nexine formation. Mol Plant 8:251–260

Kelkar S (1958) Embryology of rhus mysurensis heyne. J Indian Bot Soc 37:114–122

Knox JP (1997) The use of antibodies to study the architecture and developmental regulation of plant cell walls. Int Rev Cytol 171:79–120

Knox JP, Linstead PJ, Cooper JPC, Roberts K (1991) Developmentally regulated epitopes of

cell surface arabinogalactan proteins and their relation to root tissue pattern formation. Plant J 1:317–326

Levitin B, Richter D, Markovich I, Zik M (2008) Arabinogalactan proteins 6 and 11 are required for stamen and pollen function in Arabidopsis. Plant J 56:351–363

Lora J, Herrero M, Hormaza JI (2014) Microspore development in Annona (Annonaceae): Differences between monad and tetrad pollen. Am J Bot 101:1508–1518

Lora J, Hormaza JI, Herrero M (2016) The Diversity of the pollen tube pathway in plants: toward an increasing control by the sporophyte. Front Plant Sci 7:107

Lora J, Testillano PS, Risueno MC, et al (2009) Pollen development in Annona cherimola Mill. (Annonaceae). Implications for the evolution of aggregated pollen. BMC Plant Biol 9:129

Lou Y, Xu X, Zhu J, et al (2014) The tapetal AHL family protein TEK determines nexine formation in the pollen wall. Nat Commun 5: 3855

Maheshwari P (1934) The Indian Mango. Curent Sci 3:97–98

Majewska-Sawka A, Rodriguez-Garcia MI (2006) Immunodetection of pectin and arabinogalactan protein epitopes during pollen exine formation of Beta vulgaris L. Protoplasma 228:41–47

Mascarenhas JP (1989) The male gametophyte of flowering plants. Plant Cell 1:657–664.

Matamoro-Vidal A, Raquin C, Brisset F, Colas H, Izac B, Béatrice A, Gouyon P-H (2016)

Links between morphology and function of the pollen wall: an experimental approach. Bot J Linn Soc 180:478–490

McCormick S (2004) Control of male gametophyte development. Plant Cell 16:S142–S153

McCormick S (1993) Male gametophyte development. Plant Cell 5:1265–1275

Mukherjee SK, Litz RE (2009) Introduction: botany and importance. In: Litz RE (ed) The mango: botany, production and uses. CABI, Wallingford, pp 1–18

Nothnagel E, Münster A, Majewska-Sawka A (2001) Arabinogalactan proteins in

differentiation of gamtes and somatic cells. In: Bohanec B (ed) Biotechnological

approaches for utilization of gametic cells. Office for Official Publications of the European Communities, Luxembourg, pp 237–246

Oliveira JMS de, Mariath JE de A (2001) Anther and pollen development in Anacardium occidentale (Anacardiaceae) clone CP76. Phytomorphology 51:91–100

Otegui MS, Staehelin LA (2004) Electron tomographic analysis of post-meiotic cytokinesis during pollen development in Arabidopsis thaliana. Planta 218:501–515

Pacini E (2010) Relationships between tapetum, loculus, and pollen during development. Int J

15

Plant Sci 171:1–11

Pacini E, Franchi GG, Hesse M (1985) The tapetum - its form, function, and possible phylogeny in Embryophyta. Plant Syst Evol 149:155–185

Pacini E, Hesse M (2004) Cytophysiology of pollen presentation and dispersal. Flora 199:273–285

Palin R, Geitmann A (2012) The role of pectin in plant morphogenesis. Biosystems 109:397–402

Pell SK, Mitchell JD, Miller AJ, Lobova TA (2010) Anacardiaceae. In: Kubitzki (ed) Flowering plants. Eudicots. Springer Berlin Heidelberg, Berlin, Heidelberg, pp 7–50

Pennell RI, Janniche L, Kjellbom P, et al (1991) Developmental regulation of a plasma Membrane arabinogalactan protein epitope in oilseed rape flowers. Plant Cell 3:1317–1326

Periasamy K, Amalathas J (1991) Absence of callose and tetrad in the microsporogenesis of Pandanus odoratissimus with well-formed pollen exine. Ann Bot 67:29–33

Quilichini TD, Grienenberger E, Douglas CJ (2015) The biosynthesis, composition and assembly of the outer pollen wall: a tough case to crack. Phytochemistry 113:170–182

Ramírez F, Davenport TL (2010) Mango (Mangifera indica L.) flowering physiology. Sci Hortic 126:65–72

Ramírez F, Davenport TL (2016) Mango (Mangifera indica L.) pollination: a review. Sci Hortic 203:158–168

Sabatini DD, Bensch K, Barrnett RJ (1963) Cytochemistry and electron microscopy - preservation of cellular ultrastructure and enzymatic activity by aldehyde fixation. J Cell Biol 17:19–58

Satpute GK, Long H, Seguí-Simarro JM, et al (2005) Cell architecture during gametophytic and

embryogenic microspore development in Brassica napus L. Acta Physiol Plant 27:665–674

Scott RJ, Spielman M, Dickinson HG (2004) Stamen structure and function. Plant Cell 16:S46–S60

Shi J, Cui M, Yang L, Kim YJ, Zhang D (2015) Genetic and biochemical mechanisms of pollen wall development. Trends Plant Sci 20:741–753

Singh U (1960) Studies in the fruit drop of mango (Mangifera indica L.) II. Nature and extent of fruit drop. Hortic Adv 4:142–154

Smallwood M, Beven A, Donovan N, Neill SJ, Peart J, Roberts K (1994) Localization of cell

wall proteins in relation to the developmental anatomy of the carrot root apex. Plant J 5:237–246

Solís M-T, Pintos B, Prado M-J, et al (2008) Early markers of in vitro microspore reprogramming to embryogenesis in olive (Olea europaea L.). Plant Sci 174:597–605

Srinivasachar D (1940) Morphological studies in the family Anacardiaceae. Half-Yearly J Mysore Univ Sect B–Science 1:83–91

Torres MD, Guirado E, Farré JM, Hermoso JM (2009) Pre-bloom shoot tipping of mango. Influence on yield, fruit size and quality. Acta Hortic 391–394

16

Tucker MR, Paech NA, Willemse MTM, Koltunow AMG (2001) Dynamics of callose

deposition and beta-1,3-glucanase expression during reproductive events in sexual and apomictic Hieracium. Planta 212:487–498

Viruel MA, Escribano P, Barbieri M, Ferri M, Hormaza JI (2005) Fingerprinting, embryo type and geographic differentiation in mango (Mangifera indica L., Anacardiaceae) with microsatellites. Mol Breed 15:383–393

Wilson ZA, Zhang D-B (2009) From Arabidopsis to rice: pathways in pollen development. J Exp Bot 60:1479–92

Wunnachit W, Jenner C, Sedgley M (1992) Pollen vigour and composition in relation to andromonoecy in cashew (Anacardium occidentale L.: Anacardiaceae). Sex Plant Reprod 5:264–269