This thesis describes the isolation and characterization in <em>Arabidopsis thaliana</em> (L.) Heynh. of induced mutants, deficient for gibberellins (GA's), abscisic acid (ABA) and photoreceptors.These compounds are known to regulate various facets of plant growth and differentiation, so mutants lacking one of these substances are expected to be affected in several aspects of their physiology. It is shown in this thesis that the earliest expression of these mutants occurs during seed development and seed germination. Therefore these processes form an excellent phase to screen for these mutants.<em>Planthormone and photoreceptor mutants in relation to seed physiology.</em>In general three major periods may be distinguished in the history of a seed: 1) Seed development and maturation, 2) developmental arrest of the mature seed, characterized either by a dormant state in which seeds even do not germinate under favourable environmental conditions, or by a quiescent state in which seeds only require rehydration, and 3) germination, starting with water uptake and often requiring breaking of dormancy, which is triggered by specific environmental factors such as light and temperature. Planthormones may play a regulatory role in all three phases.Non-germinating GA-responsive mutants as described in Chapter 1 have a strongly reduced gibberellin biosynthesis (Barendse, pers.comm.) which may lead to an increased level of dormancy and/or to the inability of the seeds tobreak dormancy after imbibition of mature seeds. Clearly the presence of GA's, either by <em>de novo</em> GA synthesis, or by hydrolysis of bound forms, is not always a prerequisite for seed germination: genotypes that combine GA- and ABA deficiency like the revertants of non-germinating <em>ga-1</em> mutants described in Chapter 2 do readily germinate.Apart from the absence of endogenous factors such as gibberellins, also the lack of receptors for environmental factors that normally break dormancy might prevent germination. An example are the <em>hy-1</em> and <em>hy-2</em> mutants (Chapter 4), which are characterized by an increased hypocotyl length in white light and the absence of detectable phytochrome in dark grown hypocotyls. It was shown by Spruit et al. (1980), that these mutants hardly show any germination and correspondingly, have strongly reduced levels of phytochrome in their seeds. Their reduced germination capacity is restored by (relatively high) concentrations of exogeneously applied GA _{<font size="-1">4+7</font>} (Koornneef et al., 1981). Consequently one might expect such phytochrome deficient mutants to occur among the GA responsive non-germination mutants in <em>Arabidopsis,</em> like van der Veen and Bosma actually found for a tomato mutant (see Koornneef et al., 1981). Remarkably this was not the case in <em>Arabidopsis.</em> The reason for this seems to be the absence of a light requirement in the <em>hy</em> mutants from the M _{<font size="-1">2</font>} populations screened for non-germinating mutants of <em>Arabidopsis.</em> It happened that these M _{<font size="-1">2</font>} seeds in all cases were harvested from M _{<font size="-1">1</font>} plants grown in winter, in contrast to the seeds studied by Spruit et al. (1980) which were harvested in summer. We have observed during a number of years that seeds (including wild-type seeds), which developed in winter (natural daylight with additional continuous light by Philips TL 57) were less dormant than seeds from summer grown mother plants (long days, high light intensity, no additional light). Relevant environmental factors in this respect may be light intensity, light quality (McCullough and Shropshire, <em>1970)</em> and daylength (Karssen, <em>1970;</em> Luiten, <em>1982).</em> The effect of light quality (McCullough and Shropshire, <em>1970)</em> indicates that phytochrome may be involved in the determination of the level of dormancy.To select mutants with a reduced or absent seed dormancy, one may choose those conditions, where the wild-type is clearly dormant. However, the high and probably complex environmental variability of this character and the relatively rapid change in the level of dormancy during dry storage of the seeds makes this selection system less attractive.Selection for revertants in the progeny of mutagen treated non-germinating <em>ga-1</em> mutants proved to be an effective procedure to isolate mutants with a reduced dormancy (Chapter <em>2).</em> As the reverting effect (restored germination) was caused by a mutation at a different locus, the <em>ga-1</em> allele could be replaced by its wild-type allele by crossing the revertant with the wild-type parent followed by selection in F _{<font size="-1">2</font>} . <em></em> These newly selected monogenic recessive mutants had a reduced level of ABA in the leaves and in both the developing and ripe seeds. Correspondingly the mutant allele was called <em>aba</em> (ABA-types are <em>aba/aba</em> plants).The germination of seeds collected at different stages of their development on both ABA- and wild-type plants showed that dormancy developed during the last part of seed maturation in wild-type, but not in the <em>aba</em> -mutant. This shows that the function of ABA is dormancy induction. ABA determinations in unripe siliquae showed a peak level of ABA at about 10-12 days after anthesis, followed subsequently by a decrease, a short period at a constant level and a further decrease (Chapter 3). In addition to ABA-type mothers with ABA-type embryo's and wild-type mothers with wild-type embryo's, one can also obtain by means of the appropriate reciprocal crosses ABA-type mothers with wild-type embryo's and wild-type mothers with (50%) ABA type embryo's. So the effects of maternal and embryonic genotype can be separated. It was found (Chapter 3) that the genotype of the mother plant regulated the sharp rise in ABA content halfway seed development (maternal ABA). The genotype of the embryo and endosperm was responsible for a second ABA fraction (embryonic ABA), which reached lower levels; but persisted for some time after the maximum in maternal ABA. The onset of dormancy showed a good correlation with the presence of the embryonic ABA fraction and not with the maternal ABA.Another category of mutants which also may give some understanding of the role of ABA in seed germination are the ABA tolerant mutants recently isolated by us in <em>Arabidopsis.</em> Compared to wild-type these mutants require an upto 20 fold higher concentration of exogeneously applied ABA to inhibit seed germination. These mutants too are characterized by a reduced seed dormancy.Other genetically determined factors than those mentioned above are certainly also involved in seed development and seed germination. Thus in <em>Arabidopsis</em> mutations leading to the absence of seed coat pigments (transparent testa) and simultaneously to the absence of a mucilage layer around the seed have a reduced dormancy (Koornneef, 1981). The latter seedcoat characters are determined purely by the maternal genotype.<em>Planthormone and photoreceptor mutants in relation to other physiological effects.</em>Non-germinating mutants at the loci <em>ga-1, ga-2</em> and <em>ga-3,</em> when made to germinate by adding gibberellin, initially develop into normal looking seedlings. Later on they become dark green bushy dwarfs with reduced petals and stamens. Regular GA-spraying from the seedling stage onwards maintains the wild-type phenotype completely or nearly so (Chapter 1). The strong and quick response of the dwarfs to GA sprays (the elongation of the petals of older dwarfs becomes visible within two days) clearly demonstrates the essential role of gibberellin in elongation growth.Recently the non-germinating ga alleles were shown to have a strongly re duced kaurene synthetase activity in young siliquae compared to wild type. These analyses were performed by Dr. G.W.M. Barendse (pers.comm.). This indicates that these genes control some early step(s) in GA biosynthesis.Apart from mutants that do not germinate without GA, also more or less normally germinating GA responsive dwarfs were isolated. Half of these were found to be allelic to the non-germinating <em>ga-1</em> , <em>ga-2</em> and <em>ga-3</em> mutants. These mutant alleles behave like so called "leaky alleles", i.e. the alleles are only partly defective and produce sufficient GA for seed germination, but not enough to give normal elongation growth.GA sensitive dwarfs were also found at two other loci ( <em>ga-4</em> , <em>ga-5</em> )of which no non-germinating alleles have been isolated so far (Chapter 1). These mutants have normal or slightly reduced kaurene synthetase activity (Barendse, pers.comm.), which indicates that these genes regulate steps beyond kaurene, or affect GA metabolism in another way. It is also possible that in the mutants cell elonga tion factors are blocked for which the relatively high concentration of exogeneously applied GA may substitute. Locus <em>ga-4</em> seems to control interconversions between GA's, which is suggested by the insensitivity of <em>ga-4</em> dwarfs to GA _{<font size="-1">9</font>} . which gibberellin is effective with mutants at the other 4 loci.Abscisic acid (ABA) deficient mutants are characterized not only by reduced seed dormancy but also by disturbed water relations (wiltiness, withering), probably as a result of failure to close the stomata upon conditions of water stress (Chapter 2). This is characteristic for ABA deficient mutants in tomato (Tal and Nevo, 1973) and potato (Quarrie, 1982). ABA deficient mutants in maize are in addition to a reduced seed dormancy (viviparous mutants, gene symbol op) characterized by the failure to synthesize carotenoids and they accumulate precursors of these pigments (Robichaud et al., 1980). As ABA deficient mutants in <em>Arabidopsis,</em> tomato and potato have normal pigments, it is suggested that in the latter species the ABA biosynthesis may be blocked in the last part of the pathway, whilst in the maize mutants it is blocked at an earlier stage, i.e. where ABA and carotenoids still have a common pathway.Some of the photoreceptor mutants are affected in their germination behaviour as discussed above. However, the most conspicuous effect observed is the partial lack of light induced inhibition of hypocotyl elongation (Chapter 4). Mutants in <em>Arabidopsis,</em> and also in tomato and cucumber (Koornneef et al., 1981; Koornneef et al., unpublished), that have an elongated hypocotyl when grown in white light, were shown to have locus-specific alterations in the spectra of light inhibition when grown in light of restricted spectra] regions. In these "colour blind" mutants at two loci <em>(hy-1</em> and <em>hy-2)</em> little or no spectrophotometrically detectable phytochrome was present in dark grown hypocotyls, nor was it in the seeds. In these mutants the inhibitory effect of red and farred was almost absent. Mutants of other genes were characterized by the absence only of red inhibition <em>(hy-3)</em> or by a decreased sensitivity to the shorter wavelengths of the spectrum <em>(hy-4, hy-5). Hy-5</em> also showed a reduced inhibitory effect of far-red light. The differential sensitivity of the genotypes to specific spectral regions strongly suggests the involvement of more than one pigment in the inhibition by light of hypocotyl elongation and probably also in other photomorphogenetic processes. Some authors ascribed this role solely to phytochrome (Schäfer, 1976).Since under specific conditions phytochrome could nevertheless be detected in so called phytochrome deficient mutants (Koornneef and Spruit, unpublished) the genes <em>hy-1</em> and <em>hy-2</em> probably do not represent the structural genes of the phytochrome protein or the phytochrome chromophore, but instead may play a role in the regulation of phytochrome metabolism.<em>Further genetic aspects of plant hormone and light receptor mutants.</em>Mutation frequencies for the different groups of loci were estimated for ethylmethanesulphonate (EMS), fast neutrons and X-rays (Chapter 5). Average mutation frequencies calculated per diploid cell, per locus and per MM EMS during 1 hr at 24 °C, were for <em>ga-1, ga-2, ga-3</em> 8.0 ± 1.8 x 10 <sup><font size="-1">-6</font></SUP>, for <em>hy-1, hy-2, hy-3</em> 4.2 ± 1.4 x 10 <sup><font size="-1">-6</font></SUP>and for the <em>aba</em> locus about 27 x 10 <sup><font size="-1">-6</font></SUP>. These mutation frequencies are relatively high compared to other loci studied by us and others. It is not excluded that in these categories loci escaped detection simply because of a low mutation frequency.It is a good custum to locate newly induced mutations on the organisms gene map, especially when they are the basis of extensive research like our <em>ga, aba</em> and <em>hy</em> mutants. Unfortunately, the gene map of <em>Arabidopsis</em> was rather fragmentary, and contradictory or wrong conclusions about linkage relations could be found in literature. Since we had gradually built up the complete set of 5 primary trisomics supplemented with a number of telotrisomics (one chromosome arm extra) and also made a collection of mutations at many loci, induced in the course of various experiments at our department and supplemented with mutants described in literature, we had a good starting point to construct a more representative gene map for <em>Arabidopsis.</em> The required scale of operations was only feasable thanks to the accurate assistance of many students who performed trisomic analysis and gene mapping as part of their university training program. Important further additional data were obtained from the department of Genetics of Groningen University and from literature.The trisomic analysis aimed at assigning linkage groups (via representative markers) to the different chromosomes is described in Chapter 6. The gene maps in centimorgans for the five <em>Arabidopsis</em> chromosomes is presented in Chapter 7. On the basis of 76 loci mapped the genetic length of the <em>Arabidopsis</em> chromosomes now compares well with that of individual chromosomes in e.g. tomato and maize. This notwithstanding the small size of the <em>Arabidopsis</em> chromosomes.Genes with a similar mutant phenotype (and probably comparable functions) seem to be distributed at random over the <em>Arabidopsis</em> genome.Our set of mutants at the <em>ga-1</em> locus of <em>Arabidopsis</em> provides an excellent opportunity for fine structure analysis of the gene. The system has a very high resolving power, for the intragenic recombinants are found as the rare wild-type seedlings among thousands of non-germinating seeds per petri dish. The results show (Chapter 8) that 8 different alleles could be arranged into an internally consistent map on the basis of the frequencies of intragenic recombinants. One fast neutron induced allele behaved as an intragenic deletion. The order of the sites with respect to other genes on chromosome 4 could be established.

This thesis describes the isolation and characterization in Arabidopsis thaliana (L.) Heynh. of induced mutants, deficient for gibberellins (GA's), abscisic acid (ABA) and photoreceptors.These compounds are known to regulate various facets of plant growth and differentiation, so mutants lacking one of these substances are expected to be affected in several aspects of their physiology. It is shown in this thesis that the earliest expression of these mutants occurs during seed development and seed germination. Therefore these processes form an excellent phase to screen for these mutants.Planthormone and photoreceptor mutants in relation to seed physiology.In general three major periods may be distinguished in the history of a seed: 1) Seed development and maturation, 2) developmental arrest of the mature seed, characterized either by a dormant state in which seeds even do not germinate under favourable environmental conditions, or by a quiescent state in which seeds only require rehydration, and 3) germination, starting with water uptake and often requiring breaking of dormancy, which is triggered by specific environmental factors such as light and temperature. Planthormones may play a regulatory role in all three phases.Non-germinating GA-responsive mutants as described in Chapter 1 have a strongly reduced gibberellin biosynthesis (Barendse, pers.comm.) which may lead to an increased level of dormancy and/or to the inability of the seeds tobreak dormancy after imbibition of mature seeds. Clearly the presence of GA's, either by de novo GA synthesis, or by hydrolysis of bound forms, is not always a prerequisite for seed germination: genotypes that combine GA- and ABA deficiency like the revertants of non-germinating ga-1 mutants described in Chapter 2 do readily germinate.Apart from the absence of endogenous factors such as gibberellins, also the lack of receptors for environmental factors that normally break dormancy might prevent germination. An example are the hy-1 and hy-2 mutants (Chapter 4), which are characterized by an increased hypocotyl length in white light and the absence of detectable phytochrome in dark grown hypocotyls. It was shown by Spruit et al. (1980), that these mutants hardly show any germination and correspondingly, have strongly reduced levels of phytochrome in their seeds. Their reduced germination capacity is restored by (relatively high) concentrations of exogeneously applied GA 4+7 (Koornneef et al., 1981). Consequently one might expect such phytochrome deficient mutants to occur among the GA responsive non-germination mutants in Arabidopsis, like van der Veen and Bosma actually found for a tomato mutant (see Koornneef et al., 1981). Remarkably this was not the case in Arabidopsis. The reason for this seems to be the absence of a light requirement in the hy mutants from the M 2 populations screened for non-germinating mutants of Arabidopsis. It happened that these M 2 seeds in all cases were harvested from M 1 plants grown in winter, in contrast to the seeds studied by Spruit et al. (1980) which were harvested in summer. We have observed during a number of years that seeds (including wild-type seeds), which developed in winter (natural daylight with additional continuous light by Philips TL 57) were less dormant than seeds from summer grown mother plants (long days, high light intensity, no additional light). Relevant environmental factors in this respect may be light intensity, light quality (McCullough and Shropshire, 1970) and daylength (Karssen, 1970; Luiten, 1982). The effect of light quality (McCullough and Shropshire, 1970) indicates that phytochrome may be involved in the determination of the level of dormancy.To select mutants with a reduced or absent seed dormancy, one may choose those conditions, where the wild-type is clearly dormant. However, the high and probably complex environmental variability of this character and the relatively rapid change in the level of dormancy during dry storage of the seeds makes this selection system less attractive.Selection for revertants in the progeny of mutagen treated non-germinating ga-1 mutants proved to be an effective procedure to isolate mutants with a reduced dormancy (Chapter 2). As the reverting effect (restored germination) was caused by a mutation at a different locus, the ga-1 allele could be replaced by its wild-type allele by crossing the revertant with the wild-type parent followed by selection in F 2 . These newly selected monogenic recessive mutants had a reduced level of ABA in the leaves and in both the developing and ripe seeds. Correspondingly the mutant allele was called aba (ABA-types are aba/aba plants).The germination of seeds collected at different stages of their development on both ABA- and wild-type plants showed that dormancy developed during the last part of seed maturation in wild-type, but not in the aba -mutant. This shows that the function of ABA is dormancy induction. ABA determinations in unripe siliquae showed a peak level of ABA at about 10-12 days after anthesis, followed subsequently by a decrease, a short period at a constant level and a further decrease (Chapter 3). In addition to ABA-type mothers with ABA-type embryo's and wild-type mothers with wild-type embryo's, one can also obtain by means of the appropriate reciprocal crosses ABA-type mothers with wild-type embryo's and wild-type mothers with (50%) ABA type embryo's. So the effects of maternal and embryonic genotype can be separated. It was found (Chapter 3) that the genotype of the mother plant regulated the sharp rise in ABA content halfway seed development (maternal ABA). The genotype of the embryo and endosperm was responsible for a second ABA fraction (embryonic ABA), which reached lower levels; but persisted for some time after the maximum in maternal ABA. The onset of dormancy showed a good correlation with the presence of the embryonic ABA fraction and not with the maternal ABA.Another category of mutants which also may give some understanding of the role of ABA in seed germination are the ABA tolerant mutants recently isolated by us in Arabidopsis. Compared to wild-type these mutants require an upto 20 fold higher concentration of exogeneously applied ABA to inhibit seed germination. These mutants too are characterized by a reduced seed dormancy.Other genetically determined factors than those mentioned above are certainly also involved in seed development and seed germination. Thus in Arabidopsis mutations leading to the absence of seed coat pigments (transparent testa) and simultaneously to the absence of a mucilage layer around the seed have a reduced dormancy (Koornneef, 1981). The latter seedcoat characters are determined purely by the maternal genotype.Planthormone and photoreceptor mutants in relation to other physiological effects.Non-germinating mutants at the loci ga-1, ga-2 and ga-3, when made to germinate by adding gibberellin, initially develop into normal looking seedlings. Later on they become dark green bushy dwarfs with reduced petals and stamens. Regular GA-spraying from the seedling stage onwards maintains the wild-type phenotype completely or nearly so (Chapter 1). The strong and quick response of the dwarfs to GA sprays (the elongation of the petals of older dwarfs becomes visible within two days) clearly demonstrates the essential role of gibberellin in elongation growth.Recently the non-germinating ga alleles were shown to have a strongly re duced kaurene synthetase activity in young siliquae compared to wild type. These analyses were performed by Dr. G.W.M. Barendse (pers.comm.). This indicates that these genes control some early step(s) in GA biosynthesis.Apart from mutants that do not germinate without GA, also more or less normally germinating GA responsive dwarfs were isolated. Half of these were found to be allelic to the non-germinating ga-1 , ga-2 and ga-3 mutants. These mutant alleles behave like so called "leaky alleles", i.e. the alleles are only partly defective and produce sufficient GA for seed germination, but not enough to give normal elongation growth.GA sensitive dwarfs were also found at two other loci ( ga-4 , ga-5 )of which no non-germinating alleles have been isolated so far (Chapter 1). These mutants have normal or slightly reduced kaurene synthetase activity (Barendse, pers.comm.), which indicates that these genes regulate steps beyond kaurene, or affect GA metabolism in another way. It is also possible that in the mutants cell elonga tion factors are blocked for which the relatively high concentration of exogeneously applied GA may substitute. Locus ga-4 seems to control interconversions between GA's, which is suggested by the insensitivity of ga-4 dwarfs to GA 9 . which gibberellin is effective with mutants at the other 4 loci.Abscisic acid (ABA) deficient mutants are characterized not only by reduced seed dormancy but also by disturbed water relations (wiltiness, withering), probably as a result of failure to close the stomata upon conditions of water stress (Chapter 2). This is characteristic for ABA deficient mutants in tomato (Tal and Nevo, 1973) and potato (Quarrie, 1982). ABA deficient mutants in maize are in addition to a reduced seed dormancy (viviparous mutants, gene symbol op) characterized by the failure to synthesize carotenoids and they accumulate precursors of these pigments (Robichaud et al., 1980). As ABA deficient mutants in Arabidopsis, tomato and potato have normal pigments, it is suggested that in the latter species the ABA biosynthesis may be blocked in the last part of the pathway, whilst in the maize mutants it is blocked at an earlier stage, i.e. where ABA and carotenoids still have a common pathway.Some of the photoreceptor mutants are affected in their germination behaviour as discussed above. However, the most conspicuous effect observed is the partial lack of light induced inhibition of hypocotyl elongation (Chapter 4). Mutants in Arabidopsis, and also in tomato and cucumber (Koornneef et al., 1981; Koornneef et al., unpublished), that have an elongated hypocotyl when grown in white light, were shown to have locus-specific alterations in the spectra of light inhibition when grown in light of restricted spectra] regions. In these "colour blind" mutants at two loci (hy-1 and hy-2) little or no spectrophotometrically detectable phytochrome was present in dark grown hypocotyls, nor was it in the seeds. In these mutants the inhibitory effect of red and farred was almost absent. Mutants of other genes were characterized by the absence only of red inhibition (hy-3) or by a decreased sensitivity to the shorter wavelengths of the spectrum (hy-4, hy-5). Hy-5 also showed a reduced inhibitory effect of far-red light. The differential sensitivity of the genotypes to specific spectral regions strongly suggests the involvement of more than one pigment in the inhibition by light of hypocotyl elongation and probably also in other photomorphogenetic processes. Some authors ascribed this role solely to phytochrome (Schäfer, 1976).Since under specific conditions phytochrome could nevertheless be detected in so called phytochrome deficient mutants (Koornneef and Spruit, unpublished) the genes hy-1 and hy-2 probably do not represent the structural genes of the phytochrome protein or the phytochrome chromophore, but instead may play a role in the regulation of phytochrome metabolism.Further genetic aspects of plant hormone and light receptor mutants.Mutation frequencies for the different groups of loci were estimated for ethylmethanesulphonate (EMS), fast neutrons and X-rays (Chapter 5). Average mutation frequencies calculated per diploid cell, per locus and per MM EMS during 1 hr at 24 °C, were for ga-1, ga-2, ga-3 8.0 ± 1.8 x 10 -6, for hy-1, hy-2, hy-3 4.2 ± 1.4 x 10 -6and for the aba locus about 27 x 10 -6. These mutation frequencies are relatively high compared to other loci studied by us and others. It is not excluded that in these categories loci escaped detection simply because of a low mutation frequency.It is a good custum to locate newly induced mutations on the organisms gene map, especially when they are the basis of extensive research like our ga, aba and hy mutants. Unfortunately, the gene map of Arabidopsis was rather fragmentary, and contradictory or wrong conclusions about linkage relations could be found in literature. Since we had gradually built up the complete set of 5 primary trisomics supplemented with a number of telotrisomics (one chromosome arm extra) and also made a collection of mutations at many loci, induced in the course of various experiments at our department and supplemented with mutants described in literature, we had a good starting point to construct a more representative gene map for Arabidopsis. The required scale of operations was only feasable thanks to the accurate assistance of many students who performed trisomic analysis and gene mapping as part of their university training program. Important further additional data were obtained from the department of Genetics of Groningen University and from literature.The trisomic analysis aimed at assigning linkage groups (via representative markers) to the different chromosomes is described in Chapter 6. The gene maps in centimorgans for the five Arabidopsis chromosomes is presented in Chapter 7. On the basis of 76 loci mapped the genetic length of the Arabidopsis chromosomes now compares well with that of individual chromosomes in e.g. tomato and maize. This notwithstanding the small size of the Arabidopsis chromosomes.Genes with a similar mutant phenotype (and probably comparable functions) seem to be distributed at random over the Arabidopsis genome.Our set of mutants at the ga-1 locus of Arabidopsis provides an excellent opportunity for fine structure analysis of the gene. The system has a very high resolving power, for the intragenic recombinants are found as the rare wild-type seedlings among thousands of non-germinating seeds per petri dish. The results show (Chapter 8) that 8 different alleles could be arranged into an internally consistent map on the basis of the frequencies of intragenic recombinants. One fast neutron induced allele behaved as an intragenic deletion. The order of the sites with respect to other genes on chromosome 4 could be established.