Search

The interest of research groups in desiccation tolerance (DT) has increased substantially over the last decades. The emergence of germinated orthodox seeds and resurrection plants as main research models has pushed the limits of our knowledge beyond boundaries. At the same time, new questions and new challenges were posed. The work presented in this thesis aims at shedding light on some of these questions, deepening our understanding of DT and providing relevant information to improve stress resistance in crops. Chapter 2 is a survey of the literature and discusses the ecological and evolutionary significance for seeds to be able to re-acquire DT after germination. This chapter also discusses recent progress in DT studies using developing and germinated seeds of the model plants Arabidopsis thaliana and Medicago truncatula. In Chapter 3 I used microarray data from a time series of DT re-acquisition, together with network analysis of gene expression, to gain temporal resolution and identify relevant genes involved in the re-acquisition of DT in germinated A. thaliana seeds by incubation in abscisic acid (ABA). Overall, genes related to protection, response to stresses, seed development and seed dormancy were up-regulated, whereas genes related to cell growth and photosynthesis were down-regulated with time. Genes that respond early to exogenous ABA were related to wax biosynthetic processes, lipid storage, seed development and response to ABA stimulus. Genes that respond late to exogenous ABA were related to syncytium formation and response to abiotic stimulus (mainly light stimulus). The robustness of the network was confirmed by the projection of sets of genes – related to the acquisition of DT, seed dormancy, drought responses of adult plants and re-induction of DT by polyethylene glycol – on this network. In Chapter 4 the relation between DT in germinated seeds and drought resistance in adult plants is analysed, using rice (Oryza sativa) as experimental model. Considering the predictions of a future with lower availability of fresh water, efforts to increase rice drought tolerance without reducing yield are increasingly important. The results presented in this chapter suggest that the intrinsic mechanisms of drought tolerance in adult plants are part of the mechanisms used by seeds to tolerate desiccation, but the molecular nature of these mechanisms remains elusive. Chapter 5 explores the relation between DT and longevity in germinated seeds of the Brazilian tree species Sesbania virgata as experimental model. DT and longevity are acquired by orthodox seeds during the maturation phase of development and lost upon germination. DT can be re-induced in germinated seeds by an osmotic and/or ABA treatment, but there is no information on how these treatments affect seed longevity. S. virgata seeds lose DT slowly upon radicle growth. The radicle appeared to be the most sensitive organ and the cotyledons the most resistant. The ability to produce lateral roots was key for whole seedling survival. An osmotic treatment improved DT in germinated S. virgata seeds, but not longevity. This implies that DT and seed longevity can be uncoupled. Xerophyta viscosa is one of the best studied resurrection species. Despite the fact that adult plants and mature seeds display DT, young X. viscosa seedlings are sensitive to fast drying. A treatment with ABA can induce DT early in shoots of these seedlings, but not in roots. Chapter 6 addresses the changes in the transcriptome and proteome of X. viscosa seedlings during induction of DT. A draft genome sequence of X. viscosa was used to improve transcript and protein identification and annotation. Differences in ABA signalling and the cross talk between ABA and ethylene were presented as determinant for shoot and root responses. Moreover, differences in the accumulation of late embryogenesis abundant proteins were also shown as being key for DT in shoots and roots. In Chapter 7, DT-transcriptomes of distantly related organisms are compared and surveyed for a core set of genes representing the signatures of critical adaptive DT mechanisms. A shortlist of 260 genes emerged, with a significant number of genes under the control of ABI3 and related to dormancy. The results reinforced the idea that core mechanisms and key regulators involved in DT developed early in the history of life and were carried forward by diverse species and life forms in a conserved manner and in conjunction with dormancy. In Chapter 8, the findings of this thesis are integrated, showing how they can contribute to future improvement of stress tolerance in crops. The ability of germinated seeds to re-acquire DT is discussed in combination with dormancy and longevity and related to seed survival under unfavourable environmental conditions. The relationship between drought- and desiccation tolerance and the role of ABA are presented briefly. Possible approaches to mine for new genes for crop improvement, such as searching for conserved genes and analysing new genome sequences, are addressed. Finally, a new perspective of the way to consider the evolution of DT is proposed.

The interest of research groups in desiccation tolerance (DT) has increased substantially over the last decades. The emergence of germinated orthodox seeds and resurrection plants as main research models has pushed the limits of our knowledge beyond boundaries. At the same time, new questions and new challenges were posed. The work presented in this thesis aims at shedding light on some of these questions, deepening our understanding of DT and providing relevant information to improve stress resistance in crops. Chapter 2 is a survey of the literature and discusses the ecological and evolutionary significance for seeds to be able to re-acquire DT after germination. This chapter also discusses recent progress in DT studies using developing and germinated seeds of the model plants Arabidopsis thaliana and Medicago truncatula. In Chapter 3 I used microarray data from a time series of DT re-acquisition, together with network analysis of gene expression, to gain temporal resolution and identify relevant genes involved in the re-acquisition of DT in germinated A. thaliana seeds by incubation in abscisic acid (ABA). Overall, genes related to protection, response to stresses, seed development and seed dormancy were up-regulated, whereas genes related to cell growth and photosynthesis were down-regulated with time. Genes that respond early to exogenous ABA were related to wax biosynthetic processes, lipid storage, seed development and response to ABA stimulus. Genes that respond late to exogenous ABA were related to syncytium formation and response to abiotic stimulus (mainly light stimulus). The robustness of the network was confirmed by the projection of sets of genes – related to the acquisition of DT, seed dormancy, drought responses of adult plants and re-induction of DT by polyethylene glycol – on this network. In Chapter 4 the relation between DT in germinated seeds and drought resistance in adult plants is analysed, using rice (Oryza sativa) as experimental model. Considering the predictions of a future with lower availability of fresh water, efforts to increase rice drought tolerance without reducing yield are increasingly important. The results presented in this chapter suggest that the intrinsic mechanisms of drought tolerance in adult plants are part of the mechanisms used by seeds to tolerate desiccation, but the molecular nature of these mechanisms remains elusive. Chapter 5 explores the relation between DT and longevity in germinated seeds of the Brazilian tree species Sesbania virgata as experimental model. DT and longevity are acquired by orthodox seeds during the maturation phase of development and lost upon germination. DT can be re-induced in germinated seeds by an osmotic and/or ABA treatment, but there is no information on how these treatments affect seed longevity. S. virgata seeds lose DT slowly upon radicle growth. The radicle appeared to be the most sensitive organ and the cotyledons the most resistant. The ability to produce lateral roots was key for whole seedling survival. An osmotic treatment improved DT in germinated S. virgata seeds, but not longevity. This implies that DT and seed longevity can be uncoupled. Xerophyta viscosa is one of the best studied resurrection species. Despite the fact that adult plants and mature seeds display DT, young X. viscosa seedlings are sensitive to fast drying. A treatment with ABA can induce DT early in shoots of these seedlings, but not in roots. Chapter 6 addresses the changes in the transcriptome and proteome of X. viscosa seedlings during induction of DT. A draft genome sequence of X. viscosa was used to improve transcript and protein identification and annotation. Differences in ABA signalling and the cross talk between ABA and ethylene were presented as determinant for shoot and root responses. Moreover, differences in the accumulation of late embryogenesis abundant proteins were also shown as being key for DT in shoots and roots. In Chapter 7, DT-transcriptomes of distantly related organisms are compared and surveyed for a core set of genes representing the signatures of critical adaptive DT mechanisms. A shortlist of 260 genes emerged, with a significant number of genes under the control of ABI3 and related to dormancy. The results reinforced the idea that core mechanisms and key regulators involved in DT developed early in the history of life and were carried forward by diverse species and life forms in a conserved manner and in conjunction with dormancy. In Chapter 8, the findings of this thesis are integrated, showing how they can contribute to future improvement of stress tolerance in crops. The ability of germinated seeds to re-acquire DT is discussed in combination with dormancy and longevity and related to seed survival under unfavourable environmental conditions. The relationship between drought- and desiccation tolerance and the role of ABA are presented briefly. Possible approaches to mine for new genes for crop improvement, such as searching for conserved genes and analysing new genome sequences, are addressed. Finally, a new perspective of the way to consider the evolution of DT is proposed.

Keywords: Abscisic acid, Arabidopsis thaliana, calcium, CYP707As, desiccation, environmental factors, guard cells’ signalling pathway, hydrogen peroxide, natural variation, nitric oxide, photosystem II efficiency, RD29A, relative water content, secondary messengers, stomata, vapour pressure deficit, Vicia faba In short-term, guard cells close stomata in response to an increase in vapour pressure deficit (VPD) and they open the stomata after exposure to low VPDs. However, in long-term responses to low VPD, adaptation processes occur which make stomata less sensitive to stimuli which usually induce stomatal closure (stomatal malfunctioning). Cellular mechanism(s) leading to occurrence of stomatal malfunctioning is (are) still unknown. The aim of this project was to elucidate the processes that are involved in the malfunctioning of stomata after long-term exposure to low VPD. To elucidate whether the problem of stomatal malfunctioning is due to alterations in stomatal morphology and leaf anatomy or in the ABA signalling pathway, fava bean plants were grown at low or moderate VPDs and some plants that developed their leaves at moderate VPD were then transferred for four days to low VPD. Leaf anatomical and stomatal morphological alterations due to low VPD were not the main reason of stomatal malfunctioning in response to ABA and desiccation. Within one day exposure to low VPD, the level of foliar ABA decreased to the same level as in low VPD-grown plants, while the level of ABA-glucose ester was not affected. Spraying ABA during a 4-day exposure to low VPD maintained closing ability of the stomata after 4-day low VPD-exposure. Therefore, alteration in the signalling pathways due to low foliar ABA level was recognized as the main reason for stomatal malfunctioning after long-term low VPD-exposure. Coincidence in changes of Ca2+, ABA receptors, and positive and negative regulators of ABA signalling are proposed as early steps for stomatal malfunctioning induced by low VPD-exposure. Transcriptional activators, transcriptional repressors as well as E3 ligases are proposed for long-term adaptation of cellular processes which consequently cause decreased stomatal response to closing stimuli afterwards. In order to find the molecular mechanism(s) of stomatal malfunctioning, possible variation in stomatal response to closing stimuli was studied among Arabidopsis thaliana accessions after a 4-day low VPD-exposure. Accessions could be grouped to very sensitive, moderately sensitive and less sensitive to closing stimuli using principle component analysis. A positive correlation was found between foliar ABA level (before desiccation) and stomatal closure response to ABA (but not to desiccation) after exposure to different VPDs. Stomatal response to desiccation was positively correlated with the foliar ABA level after desiccation. In order to elucidate the molecular network underlying stomatal malfunctioning in response to ABA due to long-term low VPD-exposure, two groups of Arabidopsis accessions were used as accessions that maintained responsiveness to ABA after low VPD-exposure and accessions with low VPD induced non-ABA-responsive stomata. The foliar ABA content in all accessions correlated with the stomatal response to ABA: only when the ABA level was above a threshold value, stomata responded to ABA. After low VPD-exposure, mainly due to catabolism of ABA, the foliar ABA content decreased. This decrease in ABA level resulted in down regulation of RD29A, which caused decreased stomatal responsiveness to ABA.

Restriction of leaf water loss, by stomatal closure, is decisive for plant survival, especially under conditions of water deficit. This sensitivity of stomata to low water potential is attenuated by high relative air humidity (RH ≥ 85%) during growth, which impedes the plant’s ability to survive when subsequently exposed to lower humidities due to a negative water balance. This thesis focuses on the extent of the existing variation and the reasons underlying cultivar differences in their tolerance to high RH, as well as the rate and reversibility of stomatal adaptation to elevated RH in the course of leaf ontogeny. Cut rose was used as a model plant. An experiment on the postharvest water relations of three contrasting cultivars in their sensitivity to high RH showed that the sensitive cultivar (i.e. steepest decrease in the cut flower longevity) underwent a higher increase in the water loss compared to the tolerant cultivars. Preventing vascular occlusion considerably extended the time to wilting in the sensitive cultivar grown at high RH, showing that the high rate of water loss, as a result of plant growth at high RH, can only be detrimental for keeping quality under limiting water uptake conditions. Further investigation showed a large genotypic variation in the regulation of water loss, as a result of leaf development at high RH, and stomatal closing capacity was the key element in this process. The degree to which the stomatal anatomical features were affected and the extent that their functionality was impaired were not correlated. However, higher stomatal density, longer pore length and depth contributed to the higher water loss of high RH-grown leaves (16–30% of the effect depending on the cultivar). Reciprocal change in RH showed that stomatal functioning was no longer affected by the RH level after full leaf expansion. However, expanding leaves were always able to partly adapt to the new RH level. For leaves that started expanding at high RH but completed their expansion after transfer to moderate RH, the earlier this switch took place the better the regulation of leaf water loss. This behaviour of expanding leaves experiencing a shift from high to moderate RH was related with the increasing population of stomata exceeding a critical stomatal length. Contrary to this, leaves initially expanding at moderate RH and transferred to high RH exhibited poor stomatal functioning, even when this transfer occurred very late during leaf expansion. This suggests that stomata at various developmental stages were similarly prone to loss of closing ability, when these had been exposed to high RH prior to full leaf expansion. Key words: abscisic acid, cuticular permeability, heterogeneity, hydraulic conductivity, pore aperture, relative air humidity, Rosa hybrida, stomatal anatomy, stomatal conductance, stomatal growth, stomatal initiation, stomatal malfunctioning, stomatal population, stomatal proximity, vase life.

Restriction of leaf water loss, by stomatal closure, is decisive for plant survival, especially under conditions of water deficit. This sensitivity of stomata to low water potential is attenuated by high relative air humidity (RH ≥ 85%) during growth, which impedes the plant’s ability to survive when subsequently exposed to lower humidities due to a negative water balance. This thesis focuses on the extent of the existing variation and the reasons underlying cultivar differences in their tolerance to high RH, as well as the rate and reversibility of stomatal adaptation to elevated RH in the course of leaf ontogeny. Cut rose was used as a model plant. An experiment on the postharvest water relations of three contrasting cultivars in their sensitivity to high RH showed that the sensitive cultivar (i.e. steepest decrease in the cut flower longevity) underwent a higher increase in the water loss compared to the tolerant cultivars. Preventing vascular occlusion considerably extended the time to wilting in the sensitive cultivar grown at high RH, showing that the high rate of water loss, as a result of plant growth at high RH, can only be detrimental for keeping quality under limiting water uptake conditions. Further investigation showed a large genotypic variation in the regulation of water loss, as a result of leaf development at high RH, and stomatal closing capacity was the key element in this process. The degree to which the stomatal anatomical features were affected and the extent that their functionality was impaired were not correlated. However, higher stomatal density, longer pore length and depth contributed to the higher water loss of high RH-grown leaves (16–30% of the effect depending on the cultivar). Reciprocal change in RH showed that stomatal functioning was no longer affected by the RH level after full leaf expansion. However, expanding leaves were always able to partly adapt to the new RH level. For leaves that started expanding at high RH but completed their expansion after transfer to moderate RH, the earlier this switch took place the better the regulation of leaf water loss. This behaviour of expanding leaves experiencing a shift from high to moderate RH was related with the increasing population of stomata exceeding a critical stomatal length. Contrary to this, leaves initially expanding at moderate RH and transferred to high RH exhibited poor stomatal functioning, even when this transfer occurred very late during leaf expansion. This suggests that stomata at various developmental stages were similarly prone to loss of closing ability, when these had been exposed to high RH prior to full leaf expansion. Key words: abscisic acid, cuticular permeability, heterogeneity, hydraulic conductivity, pore aperture, relative air humidity, Rosa hybrida, stomatal anatomy, stomatal conductance, stomatal growth, stomatal initiation, stomatal malfunctioning, stomatal population, stomatal proximity, vase life.

As a result of contemporary horticultural practices relative air humidities (RH) in greenhouses are often very high. In particular in cut flowers, this results in quality problems after harvest when flowers are transferred to low RH conditions at the consumers. The quality problems are related to excessive water loss caused by a disturbance in normal functioning of stomata. The general aim of this study was to investigate the effects of high RH during growth on the stomatal response characteristics of Tradescantia virginiana L. To reach this aim, the quantitative effects of moderate (55%) and high (90%) RH during growth on the stomatal anatomy and responses of well-watered plants in response to desiccation, abscisic acid (ABA) application and light/dark transition were studied. Higher variability of stomatal closure and the presence of some non-closing stomata in high RH grown plants were striking. The dynamics of spatial heterogeneity of stomatal closure within a leaf altered by growth at high RH were studied using a chlorophyll fluorescence imaging system under non-photorespiratory conditions, and the estimates of stomatal closure obtained by means of Φ PSII measurements was correlated with direct measurements of stomatal closure and water relation parameters in plants subjected to water stress. It was shown that following desiccation, leaves grown at high RH had both a greater heterogeneity and a higher PSII efficiency compared to leaves grown at moderate RH. The stomata of high RH grown leaves were less sensitive to decreases in leaf relative water content and water potential. Moreover, the role ofABAin the disturbed stomatal responses of plants during growth at high RH was studied. The concentration ofABAin leaves grown at moderate and high RH was compared and stomatal responses to short-termABAapplication, and stomatal responses to long-termABAapplication was analysed.The results reinforce the proposal that a lowABAconcentration in well-watered plants during growth at high RH could be a reason for the structural or physiological changes of stomata. In addition, the consequences for stomatal responses of maintaining a high or low RH around only one leaf are described. The rate and the reversibility of adaptation of stomatal behaviour to moderate or high RH was also investigated and correlated with leaf ABA concentration and leaf hydraulic properties of the adapted plants. The main achievements and practical implications of this study are discussed, and suggestions for further research are presented.

As a result of contemporary horticultural practices relative air humidities (RH) in greenhouses are often very high. In particular in cut flowers, this results in quality problems after harvest when flowers are transferred to low RH conditions at the consumers. The quality problems are related to excessive water loss caused by a disturbance in normal functioning of stomata. The general aim of this study was to investigate the effects of high RH during growth on the stomatal response characteristics of Tradescantia virginiana L. To reach this aim, the quantitative effects of moderate (55%) and high (90%) RH during growth on the stomatal anatomy and responses of well-watered plants in response to desiccation, abscisic acid (ABA) application and light/dark transition were studied. Higher variability of stomatal closure and the presence of some non-closing stomata in high RH grown plants were striking. The dynamics of spatial heterogeneity of stomatal closure within a leaf altered by growth at high RH were studied using a chlorophyll fluorescence imaging system under non-photorespiratory conditions, and the estimates of stomatal closure obtained by means of Φ PSII measurements was correlated with direct measurements of stomatal closure and water relation parameters in plants subjected to water stress. It was shown that following desiccation, leaves grown at high RH had both a greater heterogeneity and a higher PSII efficiency compared to leaves grown at moderate RH. The stomata of high RH grown leaves were less sensitive to decreases in leaf relative water content and water potential. Moreover, the role ofABAin the disturbed stomatal responses of plants during growth at high RH was studied. The concentration ofABAin leaves grown at moderate and high RH was compared and stomatal responses to short-termABAapplication, and stomatal responses to long-termABAapplication was analysed.The results reinforce the proposal that a lowABAconcentration in well-watered plants during growth at high RH could be a reason for the structural or physiological changes of stomata. In addition, the consequences for stomatal responses of maintaining a high or low RH around only one leaf are described. The rate and the reversibility of adaptation of stomatal behaviour to moderate or high RH was also investigated and correlated with leaf ABA concentration and leaf hydraulic properties of the adapted plants. The main achievements and practical implications of this study are discussed, and suggestions for further research are presented.

Coffee seeds display slow and variable germination which severely hampers the production of seedlings for planting in the following growth season. Little work has been done with the aim to understand the behavior of coffee seeds during germination and there is a lack of information concerning the regulation of the germination process. This thesis addresses questions concerning the mechanism and regulation of coffee seed germination.Initial experiments showed that radicle protrusion in the dark at 30 °C was initiated at around day 5 of imbibition. At day 10, 50% of the seed population displayed radicle protrusion and at day 15 most of the seeds had completed germination. The water uptake by the coffee seeds followed a common triphasic pattern as described for many other species (Chapter 2 and 3). During imbibition the coffee embryo grew inside the endosperm. The cotyledons increased in length by 35% and the axis by 40%, resulting in the appearance of a protuberance in the endosperm cap region. There was an increase in the embryo pressure potential up to day 5 of imbibition followed by a release of turgor thereafter, indicating relaxation of embryonic cell walls (Chapter 3). Light microscopy demonstrated that the cells of the embryonic axis displayed isodiametric growth (swelling) at the beginning of the imbibition process followed by both isodiametric and longitudinal growth later during imbibition. The isodiametric growth coincided with a random orientation of the microtubules whereas longitudinal growth was accompanied by a transversal orientation. Accumulation ofb-tubulin, an increase in the number of 4C nuclei and DNA replication were evident during imbibition. These cell cycle events coincided with the growth of the embryo and the appearance of cell division prior to radicle protrusion. However, cell division was not a pre-requisite for radicle protrusion in coffee seeds (Chapter 5).The endosperm of the coffee seeds possesses polygonal and rectangular cell types located in different parts of the endosperm. The endosperm cap cells have smaller and thinner cell walls than the rest of the endosperm, suggesting that the region where the radicle will protrude is predestined in coffee seed. Low temperature scanning microscopy revealed that during imbibition cells in the endosperm cap became compressed which was followed by a loss of cell integrity, appearance of a protuberance and occurrence of cell wall porosity (Chapter 3). As in many other species, the hemi-cellulose fraction of endosperm cell walls of coffee seeds consists mainly of mannans and galacto-mannans. These polysaccharides are commonly deposited in the cell walls as food reserve. Upon germination, these galacto-mannans are degraded through the action of hydrolytic enzymes, including endo-b-mannanase,b-mannosidase anda-galactosidase resulting in a weakening of the cell walls. The coffee endosperm cap weakens in two steps: cellulase activity correlated with the first step and endo-b-mannanase activity with the second step. Endo-b-mannanase activity appeared first in the endosperm cap and only later in the rest of the endosperm, and coincided with a decrease in the required puncture force and appearance of cell wall porosity. Different isoforms of endo-b-mannanase were found in the endosperm cap and in the rest of the endosperm. The activity ofb-mannosidase increased predominantly in the endosperm cap. However, low levels of endo-b-mannanase andb-mannosidase activities were also observed in the rest of the endosperm and in the embryo prior to radicle protrusion (Chapter 3 and 6). Two partial length cDNA clones encoding for endo-b-mannanase andb-mannosidase, respectively, were isolated from coffee endosperm caps. The deduced amino acid sequences exhibited high homology with those of other endo-b-mannanases andb-mannosidases from plants (Chapter 6).Abscisic acid (ABA) inhibited germination of coffee seeds but not their water uptake, isodiametric growth, increase in 4C nuclei and DNA synthesis in the embryo cells. In the endosperm cap ABA inhibited the second step of endosperm cap weakening, presumably by inhibiting the activities of at least two endo-b-mannanase isoforms. However, ABA had no effect on endo-b-mannanase activity in the rest of the endosperm or on cellulase activity. Two peaks of endogenous ABA occurred in the embryo cells during germination. The first peak was observed at day 2 of imbibition and the second (smaller) peak at day 5 of imbibition. The occurrence of these ABA peaks coincided with the increase in the embryo growth potential and the second step of endosperm cap weakening, which makes these processes possible targets of ABA action (Chapter 3).Exogenous gibberellin (GA 4+7 ) inhibited coffee seed germination. The response to GA 4+7 showed two sensitivity thresholds: a lower one between 0 and 1mM and a higher one between 10 and 100mM. However, it was shown that radicle protrusion of coffee seeds depended on de novo synthesis of GAs. Endogenous GAs were required for embryo cell elongation and the second step of endosperm cap weakening. Incubation of seeds in exogenous GA 4+7 resulted in a loss of embryo viability and the occurrence of dead cells, as observed by low temperature scanning microscopy. We suggest that the inhibition of germination by exogenous GAs is caused by factors that are released from the endosperm cap during or after its weakening. Exogenous GAs greatly accelerated the degradation of the endosperm cap. Factors that are involved in (normal) programmed cell death of the endosperm may reach the embryo during germination, causing cell death in the embryonic axis and, hence, inhibition of radicle protrusion. The results presented in this thesis show that coffee seed germination is controlled both by embryo growth and the second step of endosperm cap weakening (Chapter 4).Finally, the sequence of events during coffee seed germination and their interrelationships are presented and discussed (Chapter 7). The events occurring in embryo and endosperm all followed a two-phase pattern. The first phase occurred during the first 5 days of imbibition and the second phase thereafter, until radicle protrusion. The results make clear that the germination processes are temporally and spatially coordinated and that disturbance of this coordination, as in the presence of GAs, may severely affect seed behaviour.

Coffee seeds display slow and variable germination which severely hampers the production of seedlings for planting in the following growth season. Little work has been done with the aim to understand the behavior of coffee seeds during germination and there is a lack of information concerning the regulation of the germination process. This thesis addresses questions concerning the mechanism and regulation of coffee seed germination.Initial experiments showed that radicle protrusion in the dark at 30 °C was initiated at around day 5 of imbibition. At day 10, 50% of the seed population displayed radicle protrusion and at day 15 most of the seeds had completed germination. The water uptake by the coffee seeds followed a common triphasic pattern as described for many other species (Chapter 2 and 3). During imbibition the coffee embryo grew inside the endosperm. The cotyledons increased in length by 35% and the axis by 40%, resulting in the appearance of a protuberance in the endosperm cap region. There was an increase in the embryo pressure potential up to day 5 of imbibition followed by a release of turgor thereafter, indicating relaxation of embryonic cell walls (Chapter 3). Light microscopy demonstrated that the cells of the embryonic axis displayed isodiametric growth (swelling) at the beginning of the imbibition process followed by both isodiametric and longitudinal growth later during imbibition. The isodiametric growth coincided with a random orientation of the microtubules whereas longitudinal growth was accompanied by a transversal orientation. Accumulation ofb-tubulin, an increase in the number of 4C nuclei and DNA replication were evident during imbibition. These cell cycle events coincided with the growth of the embryo and the appearance of cell division prior to radicle protrusion. However, cell division was not a pre-requisite for radicle protrusion in coffee seeds (Chapter 5).The endosperm of the coffee seeds possesses polygonal and rectangular cell types located in different parts of the endosperm. The endosperm cap cells have smaller and thinner cell walls than the rest of the endosperm, suggesting that the region where the radicle will protrude is predestined in coffee seed. Low temperature scanning microscopy revealed that during imbibition cells in the endosperm cap became compressed which was followed by a loss of cell integrity, appearance of a protuberance and occurrence of cell wall porosity (Chapter 3). As in many other species, the hemi-cellulose fraction of endosperm cell walls of coffee seeds consists mainly of mannans and galacto-mannans. These polysaccharides are commonly deposited in the cell walls as food reserve. Upon germination, these galacto-mannans are degraded through the action of hydrolytic enzymes, including endo-b-mannanase,b-mannosidase anda-galactosidase resulting in a weakening of the cell walls. The coffee endosperm cap weakens in two steps: cellulase activity correlated with the first step and endo-b-mannanase activity with the second step. Endo-b-mannanase activity appeared first in the endosperm cap and only later in the rest of the endosperm, and coincided with a decrease in the required puncture force and appearance of cell wall porosity. Different isoforms of endo-b-mannanase were found in the endosperm cap and in the rest of the endosperm. The activity ofb-mannosidase increased predominantly in the endosperm cap. However, low levels of endo-b-mannanase andb-mannosidase activities were also observed in the rest of the endosperm and in the embryo prior to radicle protrusion (Chapter 3 and 6). Two partial length cDNA clones encoding for endo-b-mannanase andb-mannosidase, respectively, were isolated from coffee endosperm caps. The deduced amino acid sequences exhibited high homology with those of other endo-b-mannanases andb-mannosidases from plants (Chapter 6).Abscisic acid (ABA) inhibited germination of coffee seeds but not their water uptake, isodiametric growth, increase in 4C nuclei and DNA synthesis in the embryo cells. In the endosperm cap ABA inhibited the second step of endosperm cap weakening, presumably by inhibiting the activities of at least two endo-b-mannanase isoforms. However, ABA had no effect on endo-b-mannanase activity in the rest of the endosperm or on cellulase activity. Two peaks of endogenous ABA occurred in the embryo cells during germination. The first peak was observed at day 2 of imbibition and the second (smaller) peak at day 5 of imbibition. The occurrence of these ABA peaks coincided with the increase in the embryo growth potential and the second step of endosperm cap weakening, which makes these processes possible targets of ABA action (Chapter 3).Exogenous gibberellin (GA 4+7 ) inhibited coffee seed germination. The response to GA 4+7 showed two sensitivity thresholds: a lower one between 0 and 1mM and a higher one between 10 and 100mM. However, it was shown that radicle protrusion of coffee seeds depended on de novo synthesis of GAs. Endogenous GAs were required for embryo cell elongation and the second step of endosperm cap weakening. Incubation of seeds in exogenous GA 4+7 resulted in a loss of embryo viability and the occurrence of dead cells, as observed by low temperature scanning microscopy. We suggest that the inhibition of germination by exogenous GAs is caused by factors that are released from the endosperm cap during or after its weakening. Exogenous GAs greatly accelerated the degradation of the endosperm cap. Factors that are involved in (normal) programmed cell death of the endosperm may reach the embryo during germination, causing cell death in the embryonic axis and, hence, inhibition of radicle protrusion. The results presented in this thesis show that coffee seed germination is controlled both by embryo growth and the second step of endosperm cap weakening (Chapter 4).Finally, the sequence of events during coffee seed germination and their interrelationships are presented and discussed (Chapter 7). The events occurring in embryo and endosperm all followed a two-phase pattern. The first phase occurred during the first 5 days of imbibition and the second phase thereafter, until radicle protrusion. The results make clear that the germination processes are temporally and spatially coordinated and that disturbance of this coordination, as in the presence of GAs, may severely affect seed behaviour.

This thesis describes the influence of abscisic acid (ABA) on the transport of assimilates to seeds and the deposition of reserves in seeds. It is well-known from literature that ABA accumulates in seeds during development, and that ABA concentrations in seeds correlate rather well with seed size and seed growth rates. However, since ABA is at least partly synthesized in the leaves and transported to the seeds via the phloem, a correlation between ABA levels and growth rate can easily be explained as the result of the combined transport of ABA and assimilates. Reports about the effect of applied ABA on transport of assimilates to seeds are contradictory (Table 1.I). Moreover, application of ABA has several disadvantages: the application technique itself may cause artefacts, and the results are difficult to interpret since the endogenous ABA level after application depends on penetration, transport and metabolism in the tissue. For these reasons, we have chosen for a different approach, viz . the use of hormone mutants. Two species were used: Pisum sativum and Arabidopsis thaliana .Growth and development of the ABA-deficient ' wilty ' mutant of pea is described in detail (Chapter 2). A non-wilty isogenic line was obtained after six successive backcrosses of the mutant with a closely approximating line. The plants were grown at conditions of high relative humidity and cultured on hydroponics, since leaves of ABA-deficient plants fail to accumulate ABA at drought stress and consequently do not close their stomata- For the same reason, mutant leaves have a higher dry matter content than wild-type leaves. The mutant grew slower and especially root growth was reduced; this resulted in a considerably larger shoot/root ratio. Similar effects have been found in ABA-deficient mutants of several other species. This root-growth promotive effect of ABA can be explained as a measure to prevent an undesirable water status of the leaves by increasing the volume of soil explored under dry conditions.ABA-deficient plants had fewer and smaller seeds than wild-type plants, but since the mutants plants themselves were also smaller, the weight ratio of reproductive to vegetative parts was similar in both lines. The seeds of mutant plants contained about five times less ABA than wild-type seeds. It was concluded that the lower growth rate of both vegetative and reproductive parts was not directly caused by the lower ABA content of these organs, but by disturbed water relations.One of the reasons to choose the pea mutant was that transport of assimilates to legume seeds can be studied by the empty-seed-coat technique. After removal of a part of the pod wall and the seed coat, the embryo is replaced by a buffer, while leaving most of the maternal tissue intact. This buffer receives assimilates from the seed-coat and is regularly analysed for the presence of sucrose. The rate of sucrose efflux calculated from the seed-coat into the medium is assumed to be a measure for phloem import, especially during the period of near-constant sucrose release (4-10 hours after the start of the experiment). The effect of ABA on sucrose release was studied by applying various ABA concentrations to the buffer (Figure 3. 1) and expressing the amount of sucrose released into these buffers relative to the amount present in a control seed-coat (a surgically modified seed-coat containing buffer without ABA). It was shown that hardly any ABA leaked from one seed-coat to another. The experiments were performed with both wildtype and ABA-deficient plants, either or not at source-limited conditions, since it was assumed that a possible effect of ABA might be more pronounced in ABA-deficient plants and at source-limited conditions. Source-limiting indeed caused a reduction of the sucrose release- rate. However, no effect of ABA on sucrose release could be discerned, irrespective of the experimental conditions.Another advantage of the use of mutants is the possibility to study competition between genetically different seeds, for the same source of assimilates (Figure 1.3). In pea, this was achieved by crossing an ABA-deficient mother plant with pollen from plants that were heterozygous for this trait. Chapter 4 describes experiments on ABA-deficient pea plants bearing pods with both ABA-deficient and ABA-containing seeds in the same pod. Seeds in the same pod usually have the same growth rate. In these pods, the growth rate of the seeds was determined by measuring the diameter of the seeds with a pair of callipers. In a control experiment it was shown that these manipulations (opening of the pod and measuring the seeds) did not disturb the normal growth pattern of the seeds. No effect of the genotype on the growth rate of the seeds was detected.Similar studies were performed with Arabidopsis mutants (Chapter 5). In one series of experiments, successive flowers of a recombinant of an ABA-deficient and an ABA-insensitive mutant (aba,abi3) were alternatingly pollinated with pollen from either wildtype or double-mutant plants. In another series of experiments, a double-mutant that was both ABA-deficient and starchless was used as a mother plant; the amount of available assimilates in these plants was reduced by decreasing the light intensity. The growth rate of the seeds was determined by exposing the mother plants to radiolabelled CO 2 and detecting the amount of radioactivity in the seeds. The weight of the seeds of these crosses was determined on a high-precision balance. In these experiments, again no significant influence of the genotype on either the import of radioactivity or the weight of the seeds could be detected.The possible effect of ABA on the deposition of reserve material in seeds was studied with some Arabidopsis mutants. Arabidopsis is a crucifer and its seeds initially accumulate starch which is degraded and converted to lipids during seed maturation. Seeds of the ABA-deficient (aba) and the ABA-insensitive (abi3) mutant and their recombinant (aba,abi3) were collected during development and their lipid and carbohydrate composition was analysed and compared with wild-type seeds. The maximum dry and fresh weight of the seeds was not influenced by the genotype. All mutants had considerably reduced levels of eicosenoic acid (20: 1) in the triacylglycerol fraction as compared to wild- type seeds; it is concluded that ABA is involved in the regulation of elongation of fatty acids. The total amount of neutral lipids in seeds of the single mutants was similar to that in wild-type seeds (about 30-35 % on a dry weight basis), but doublemutant seeds contained only half this amount. On the other hand, double-mutant seeds had elevated levels of starch and soluble sugars. Apparently, the blockade in lipid synthesis in these mutants is so strong that it results in starch accumulation and finally in accumulation of soluble sugars. It is concluded that both the presence of ABA and the sensitivity to ABA are required for normal acyl-chain elongation and lipid accumulation; the absence of both factors results in a higher proportion of the imported assimilates being stored as carbohydrates.From the above-mentioned experiments, it was concluded that ABA has no major influence on the long-distance transport of assimilates, at least not in the species Pisum sativum and Arabidopsis thaliana. However, ABA appears to be involved in the distribution of assimilates over the various types of storage material during seed development.

This thesis describes the influence of abscisic acid (ABA) on the transport of assimilates to seeds and the deposition of reserves in seeds. It is well-known from literature that ABA accumulates in seeds during development, and that ABA concentrations in seeds correlate rather well with seed size and seed growth rates. However, since ABA is at least partly synthesized in the leaves and transported to the seeds via the phloem, a correlation between ABA levels and growth rate can easily be explained as the result of the combined transport of ABA and assimilates. Reports about the effect of applied ABA on transport of assimilates to seeds are contradictory (Table 1.I). Moreover, application of ABA has several disadvantages: the application technique itself may cause artefacts, and the results are difficult to interpret since the endogenous ABA level after application depends on penetration, transport and metabolism in the tissue. For these reasons, we have chosen for a different approach, viz . the use of hormone mutants. Two species were used: Pisum sativum and Arabidopsis thaliana .Growth and development of the ABA-deficient ' wilty ' mutant of pea is described in detail (Chapter 2). A non-wilty isogenic line was obtained after six successive backcrosses of the mutant with a closely approximating line. The plants were grown at conditions of high relative humidity and cultured on hydroponics, since leaves of ABA-deficient plants fail to accumulate ABA at drought stress and consequently do not close their stomata- For the same reason, mutant leaves have a higher dry matter content than wild-type leaves. The mutant grew slower and especially root growth was reduced; this resulted in a considerably larger shoot/root ratio. Similar effects have been found in ABA-deficient mutants of several other species. This root-growth promotive effect of ABA can be explained as a measure to prevent an undesirable water status of the leaves by increasing the volume of soil explored under dry conditions.ABA-deficient plants had fewer and smaller seeds than wild-type plants, but since the mutants plants themselves were also smaller, the weight ratio of reproductive to vegetative parts was similar in both lines. The seeds of mutant plants contained about five times less ABA than wild-type seeds. It was concluded that the lower growth rate of both vegetative and reproductive parts was not directly caused by the lower ABA content of these organs, but by disturbed water relations.One of the reasons to choose the pea mutant was that transport of assimilates to legume seeds can be studied by the empty-seed-coat technique. After removal of a part of the pod wall and the seed coat, the embryo is replaced by a buffer, while leaving most of the maternal tissue intact. This buffer receives assimilates from the seed-coat and is regularly analysed for the presence of sucrose. The rate of sucrose efflux calculated from the seed-coat into the medium is assumed to be a measure for phloem import, especially during the period of near-constant sucrose release (4-10 hours after the start of the experiment). The effect of ABA on sucrose release was studied by applying various ABA concentrations to the buffer (Figure 3. 1) and expressing the amount of sucrose released into these buffers relative to the amount present in a control seed-coat (a surgically modified seed-coat containing buffer without ABA). It was shown that hardly any ABA leaked from one seed-coat to another. The experiments were performed with both wildtype and ABA-deficient plants, either or not at source-limited conditions, since it was assumed that a possible effect of ABA might be more pronounced in ABA-deficient plants and at source-limited conditions. Source-limiting indeed caused a reduction of the sucrose release- rate. However, no effect of ABA on sucrose release could be discerned, irrespective of the experimental conditions.Another advantage of the use of mutants is the possibility to study competition between genetically different seeds, for the same source of assimilates (Figure 1.3). In pea, this was achieved by crossing an ABA-deficient mother plant with pollen from plants that were heterozygous for this trait. Chapter 4 describes experiments on ABA-deficient pea plants bearing pods with both ABA-deficient and ABA-containing seeds in the same pod. Seeds in the same pod usually have the same growth rate. In these pods, the growth rate of the seeds was determined by measuring the diameter of the seeds with a pair of callipers. In a control experiment it was shown that these manipulations (opening of the pod and measuring the seeds) did not disturb the normal growth pattern of the seeds. No effect of the genotype on the growth rate of the seeds was detected.Similar studies were performed with Arabidopsis mutants (Chapter 5). In one series of experiments, successive flowers of a recombinant of an ABA-deficient and an ABA-insensitive mutant (aba,abi3) were alternatingly pollinated with pollen from either wildtype or double-mutant plants. In another series of experiments, a double-mutant that was both ABA-deficient and starchless was used as a mother plant; the amount of available assimilates in these plants was reduced by decreasing the light intensity. The growth rate of the seeds was determined by exposing the mother plants to radiolabelled CO 2 and detecting the amount of radioactivity in the seeds. The weight of the seeds of these crosses was determined on a high-precision balance. In these experiments, again no significant influence of the genotype on either the import of radioactivity or the weight of the seeds could be detected.The possible effect of ABA on the deposition of reserve material in seeds was studied with some Arabidopsis mutants. Arabidopsis is a crucifer and its seeds initially accumulate starch which is degraded and converted to lipids during seed maturation. Seeds of the ABA-deficient (aba) and the ABA-insensitive (abi3) mutant and their recombinant (aba,abi3) were collected during development and their lipid and carbohydrate composition was analysed and compared with wild-type seeds. The maximum dry and fresh weight of the seeds was not influenced by the genotype. All mutants had considerably reduced levels of eicosenoic acid (20: 1) in the triacylglycerol fraction as compared to wild- type seeds; it is concluded that ABA is involved in the regulation of elongation of fatty acids. The total amount of neutral lipids in seeds of the single mutants was similar to that in wild-type seeds (about 30-35 % on a dry weight basis), but doublemutant seeds contained only half this amount. On the other hand, double-mutant seeds had elevated levels of starch and soluble sugars. Apparently, the blockade in lipid synthesis in these mutants is so strong that it results in starch accumulation and finally in accumulation of soluble sugars. It is concluded that both the presence of ABA and the sensitivity to ABA are required for normal acyl-chain elongation and lipid accumulation; the absence of both factors results in a higher proportion of the imported assimilates being stored as carbohydrates.From the above-mentioned experiments, it was concluded that ABA has no major influence on the long-distance transport of assimilates, at least not in the species Pisum sativum and Arabidopsis thaliana. However, ABA appears to be involved in the distribution of assimilates over the various types of storage material during seed development.