DEVELOPMENTAL
BIOLOGY
72,62-72
(1979)
Isolation and Characterization of Six Embryo-Lethal Arabidopsis thaliana
Mutants of
DAVID W. MEINKE AND IAN M. SUSSEX Department
of
Biology, Yale University, New Haven, Connecticut 06520
Received November 13, 1978; accepted in revised form April 3, 1979 Mature seeds of Arabidopsis thaliana strain “Columbia” were soaked for 7.5 hr in an aqueous solution of the chemical mutagen ethyl methanesulfonate (0.05,0.10, or 0.50%, v/v). Embryo-lethal mutants were identified in the resulting M-l chimeral plants by screening the first five siliques of each plant and noting the frequency of aborted seeds. Three hundred sixty seeds were treated at each mutagen dose; the frequency of embryo-lethal mutants ranged from l-3% of the M-l plants grown from seeds exposed to 0.05% EMS, to 20-30% of the M-l plants at the highest mutagen dose. Six embryo-lethal mutants identified through screening of M-l plants were chosen for detailed studies in subsequent generations. All six mutants segregate as nonallelic, Mendelian recessive lethals, and are maintained as heterozygotes since homozygotes die as embryos. Fruits of heterozygous plants contain 25% aborted seeds and 75% phenotypically normal seeds (z/3 heterozygotes and % wild type). Segregation ratios are not temperature sensitive; the same frequency of aborted seeds is found in plants grown at 18, 25, and 32°C. Embryo arrest and eventual lethality in each mutant occur at a characteristic stage of early embryo development: globular-heart, globular, early globular, or preglobular. Arrested embryos from five of the six mutants resemble normal embryos at early stages of development. Developmental arrest of the embryo proper in the remaining mutant is followed by abnormal growth of the suspensor, an embryonic structure that attaches the embryo proper to the maternal tissue. INTRODUCTION
opment: globular-heart, globular, or preglobular.
Arabidopsis thaliana was described in the preceding report as a model system for mutant analysis of plant embryo development (Meinke and Sussex, 1979). Normal development of strain “Columbia” plants was characterized, procedures for identifying and classifying embryo-lethal mutants of Arabidopsis were detailed, and the potential application of these mutants to the study of normal embryo development was discussed. The purpose of this report is to present genetic and developmental data on six embryo-lethal mutants isolated following EMS mutagenesis of wild-type seeds. Each mutant segregates as a Mendelian recessive lethal; fruits of heterozygous plants contain 25% aborted seeds and 75% phenotypically normal seeds. Developmental arrest and eventual lethality of homozygous mutant embryos occur at a characteristic stage of early embryo devel-
MATERIALS
early
METHODS
Mutagenesis Wild-type seeds of Arabidopsis thaliana (L.) Heynh strain “Columbia” were kindly provided by Dr. G. P. Redei of the University of Missouri, Columbia. Seeds used for mutagenesis were assumed to be essentially isogenic because they were collected from a single wild-type plant following self-pollination. The chemical mutagen ethyl methanesulfonate (EMS) was vigorously suspended in distilled water, and mutagenesis was performed at room temperature in a well-ventilated hood. Since EMS is a dangerous mutagen and carcinogen, extreme care was taken to minimize the chance of accidental exposure, and to decontaminate and discard exposed equipment (Lewis and 62
0012-1606/79/090062-11$02.00/O Copyright 0 1979 by Academic Press, Inc. All rights of reproduction in any form reserved.
AND
globular,
MEINKE
AND SUSSEX
Embryo-Lethal
Bather, 1968). Mature seeds were soaked for 7.5 hr in glass vials containing 2.0 ml of EMS solution (0.05, 0.10, or 0.50%, v/v). Treated seeds were then washed with water for 12 hr before planting. Three hundred sixty seeds treated at each concentration were planted in 3-in. pots by individually transferring each seed to the soil surface with a Pasteur pipet. Each pot contained five seeds placed at known positions to simplify cataloguing of M-l plants. All M-l plants and subsequent generations were grown in an environmental growth room at 25°C in 16 l-u-/8 hr light/dark cycles. Details of growth conditions and procedures for the identification of aborted seeds and isolation of embryo-lethal mutants in Arabidopsis are presented in the previous report (Meinke and Sussex, 1979).
Genetic Characterization
of Mutants
Segregation ratios were obtained by screening immature, green siliques of M-2 heterozygotes, grown from seed recovered from the mutant sector of M-l plants. Siliques were split longitudinally with a razor blade and examined under a dissecting microscope. Aborted seeds were recognized by their different size and color. Fifteen siliques from each M-2 plant were screened and the percentage of aborted seeds was calculated. Siliques from recessive mutant heterozygotes were expected to contain two types of phenotypically normal seeds: heterozygous and wild type. The relative frequency of these two types was determined by counting the number of heterozygous and wild-type plants that grew from mutant stocks of phenotypically normal seeds. Each mutant was also tested for temperature sensitivity; heterozygous plants raised at 25°C were transferred to 18°C shortly after flowering, when they could be distinguished from wild-type plants, and the percentage of aborted seeds was determined in the first 15 siliques that developed at the lower temperature. Further information on the nature of
Mutants
OfArabidopsis
63
each embryo-lethal mutation was obtained from M-l mutant sector sizes estimated from the relative position and phenotype of the five screened siliques (see Fig. 6 in Meinke and Sussex, 1979). The mutant sector of a recessive mutant chimera was expected to cover roughly ?4 of the main M-l inflorescence. The size of a mutant sector was determined in part by the number of apical meristem cells “hit” during seed mutagenesis; unusually large mutant sectors were often a sign of multiple mutations. Pairwise crosses between heterozygous plants were used to check for allelism among six embryo-lethal mutants. Allelism tests usually involve crossing two different homozygous mutants, but with recessive lethals it was necessary to cross heterozygotes. In this case, a cross between allelic mutants will produce a silique containing 25% aborted seeds, whereas crosses between nonallelic mutants will produce only phenotypicahy normal seeds. These two conditions can be clearly distinguished in siliques containing 30 or more seeds. Allelism tests require controlled crossing, which in Arabidopsis is complicated by small flowers and natural self-pollination. Two different techniques were used to overcome the problem of self-pollination: emasculation and “early pollination.” To emasculate the female parent of a given cross, the largest closed bud from an inflorescence was positioned under a dissecting microscope; sepals and petals were pulled back with fine forceps, and anthers were removed by cutting or crushing each stamen filament. With practice this could be done without noticeable injury to the ovary. The early pollination method did not require the tedious removal of anthers; sepals were pulled back from closed buds, and the exposed stigma was covered with pollen. Self-pollination still occurred, but too late to result in selffertilization. Pollen for either method was obtained by removing dehiscent anthers from open flowers, and cross-pollination was achieved by repeatedly touching at least one anther to the stigma surface.
64 Developmental tants
DEVELOPMENTALBIOLOGY
Characterization
of Mu-
The first step in determining the stage of embryo arrest for each mutant was to dissect aborted seeds under a dissecting microscope. Mutant embryos arrested at or beyond advanced globular stages (>40 pm) were easily removed and measured with a stage micrometer. When arrested embryos were not visible under a dissecting microscope, aborted seeds were squashed in a drop of stain and examined under a compound light microscope. Early globular embryos could be identified by this technique. Mutations that arrested embryogeny even earlier were assumed to be preglobular lethals, and were studied in sectioned material. The length of an aborted seed was also used as an indication of the stage of developmental arrest. For examination of embedded material, mutant ovules were removed from heterozygous siliques as soon as they could be distinguished from phenotypically normal seeds, fixed for 6-12 hr in 6% glutaraldehyde, postfixed in 2% OsOl, and embedded in Luft’s Epon 812. Most ovules had to be nicked with a razor blade to allow sufficient penetration of the embedding medium. Serial sections through the entire ovule were cut at 0.5 pm with an LKB-III ultramicrotome, and stained with a mixture of azure B and methylene blue. Details of these histological procedures are included in a separate paper on abnormal suspensor development in an embryo-lethal mutant of Arabidopsis (Meinke and Marsden, in preparation). RESULTS
Mutagenesis and M-l Plants The results of mutagenesis and screening of M-l plants are summarized in Table 1. The high frequency of germination obtained after all three EMS treatments indicates that seed viability was not affected by the mutagenesis procedure, although seeds exposed to 0.50% EMS germinated l-
VOLUME 72, 1979 TABLE 1 SUMMARYOFMUTAGENESISRESULTSANDM-I SCREENING Mutagen
Treated seeds Germination Plants screened Frequency of embryo-lethal mutants” Mutants studied in M-2 generation
dose
0.05% EMS
0.10% EMS
0.50% EMS
360 95% 211 l-3%
360 96% 294 3-5%
360 93% 50 20-30%
0
4
2
a This is an estimated frequency because no attempt was made to test all suspected mutants in the M-2 generation.
3 days after those exposed to 0.05 and 0.10% EMS. Vegetative growth, flowering, and silique development appeared normal in most M-l plants, and the general reproductive effort was similar to that of wild-type plants. Embryo-lethal mutants could be identified only by examining individual siliques under a dissecting microscope because siliques containing both aborted and normal seeds were the same length and had the same maturation time as wild-type siliques. Nearly 3000 siliques from approximately 600 M-l plants were screened, including most of the plants from the 0.05 and 0.10% EMS treatments and a representative sample from the high mutagen dose. One to five percent of the M-l plants from the 0.05 and 0.10% EMS treatments appeared to be segregating for recessive embryo-lethal mutations, based on examination of the first five siliques. These suspected mutant plants had three or four wild-type siliques with normal seeds and one or two mutant siliques that contained approximately 75% phenotypically normal seeds and 25% aborted seeds. A few plants were difficult to classify as either wild type or mutant in the M-l generation because they had one or two siliques in which the frequency of aborted seeds was significantly lower than 25%. The remaining 90-95% of
MEINKE
AND SUSSEX
Embryo-Lethal
the plants were phenotypically normal and showed no significant increase in the number of aborted seeds. The frequency of suspected embryo-lethal mutants was noticeably higher among plants grown from seeds exposed to 0.50% EMS. Many of these plants were still clearly wild type, but an increase in physiological abortion rates and the appearance of plants with more than one mutant phenotype made classification and screening more difficult. Some plants contained a single recessive mutation with a typical mutant sector and one phenotypic class of aborted seeds; a representative sample of these mutants was recovered and studied in the M-2 generation. Other plants appeared to be segregating for two different embryo-lethal mutations in the same mutant sector, and a few plants had unusually large mutant sectors with several types of aborted seeds. These complex mutants were found only at the high mutagen dose, and they were not studied in more detail.
Mutant Sectors of M-l Plants The shoot apical meristem of mature plant embryos is a multicellular structure composed of diploid cells. M-l plants grown from mutagenized seeds will therefore be mericlinal chimeras with heterozygous mutant cells and siliques occupying only a wedge-shaped sector of the primary reproductive inflorescence (see Fig. 6 in Meinke and Sussex, 1979). The size of a mutant sector is determined by five factors: (1) number of target cells in the mutagenized seed that contribute to the M-l reproductive meristems; (2) number of mutated target cells; (3) cell cycle stage of the mutated cell; (4) molecular nature of the mutation; and (5) competitive growth advantage or disadvantage of the mutant sector in M-l plants. When the shoot apical meristem of a mature embryo contains “x” number of target cells, induction of a recessive mutation in one of these cells should result in a mutant sector covering l/x to 1/4x of the
Mutants
Of Arabidopsis
65
circumference of the primary M-l reproductive meristem. Li and Redei (1969) have estimated from mutagenesis studies that mature seeds of Arabidopsis contain an average of two target cells. The initial size of a mutant sector containing cells heterozygous for an induced recessive mutation was therefore not expected to exceed ‘/z the circumference of the primary reproductive meristem. M-l plants with larger mutant sectors were likely to contain complex or multiple mutations. Lateral and additional branches of a mutant M-l plant could be chimeral, wild type, or mutant, depending on the genotype of contributing cells. Although the size of a mutant sector can change during development of an inflorescence, most evidence suggests that cells of the mutant sector do not have a competitive growth disadvantage (Muller, 1965; Grinikh et al., 1974). Information on the size of a mutant sector must be obtained in the M-2 generation because heterozygous and wild-type cells of the M-l plant are phenotypically identical. With embryo-lethal mutants, however, it was possible to screen M-2 embryos in the fruits of M-l plants, and estimate the size of a mutant sector from the number and position of mutant siliques segregating for aborted seeds. When two or more screened siliques were mutant, both the minimum and the maximum sizes of a mutant sector were estimated, assuming a normal phyllotaxy of 137” between consecutive siliques. The minimum size of a mutant sector could not be determined when only one mutant silique was identified. Mutant sector sizes were determined for 21 M-l plants that appeared to be segregating for an embryolethal mutation. Each suspected mutant plant contained a single mutant sector that included one or two of the first five siliques and covered approximately % to Yt of the primary M-l inflorescence. These results are consistent with the induction of single recessive mutations in seeds containing two target cells. Six mutants were chosen for detailed
66
DEVELOPMENTAL BIOLOGY
studies in subsequent generations. Characteristics of the M-l plants that yielded these mutants are summarized in Table 2. Segregation data for the M-l siliques have been included to show how the mutants were identified. Analysis of the mutant sector data and segregation ratios suggests that each of the six M-l plants carried a single recessive embryo-lethal mutation that was induced in one of the two target cells during mutagenesis.
Segregation and Heterozygote Ratios All six mutants segregated as Mendelian recessive lethals with approximately 25% aborted seeds in the siliques of M-2 heterozygotes (Table 3). The slightly aberrant segregation ratio observed in mutant 124D was correlated with a reduction in the number of aborted seeds in the basal portion of heterozygous siliques. Available evidence suggests that in this particular mutant, pollen tubes carrying the mutant allele grow TABLE
2
M-l CHARACTERISTICS OF SIX EMBRYO-LETHAL MUTANTS STUDIED IN SUBSEQUENT GENERATIONS” Total Mutant Mutant Mutant P-;;Ff seeds siliques” silique number’ seeds 123B
1
5
53
24.5
79A
1
5
55
27.3
124D
1
3
58
15.5
87A
2
2 5
44 26
13.6 26.9
71E
2
3 5
27 10
22.2 20.0
50B
3
2 3 5
39 33 5
23.1 33.3 40.0
n M-l plants were grown from seeds treated with 0.10% EMS (mutants 123B, 79A, 124D, and 87A) or 0.50% EMS (mutants 7lE and 50B). b Number of sfiiques, out of five screened, that appeared to be segregating for an embryo-lethal mutation. ’ Silique 1 developed from the first flower in the main inflorescence, 2 from the second flower, etc.
VOLUME 72.1979 TABLE
3
SEGREGATION OF ABORTED SEEDS IN SILIQUES OF M-2 HETEROZYGOTES Percentage Mutant-plant Seeds Chi square screened aborted seeds 87A-61M 123B-93M 79A-54M 50B-1llM 71E-102M -102L Total 124D-85M -85L -288L Total * Significantly * * Significantly
711 500 701 773 476 636 1112 475 607 836 1918 different different TABLE
24.5 26.0 24.0 25.4 23.1 26.9 25.3 22.1 22.2 21.5 21.9
0.08 0.22 0.34 0.04 0.81 1.11 0.03 1.99 2.33 5.18: 9.68; *
from 25.0 at the 5% level. from 25.0 at the 1% level. 4
SEGREGATION OF ABORTED SEEDS AT 18°C Mutant
Seeds screened
87A 123B 79A 50B 71E 124D
762 658 853 728 690 814
Percentage aborted seeds 23.2 24.6 26.1 25.1 25.5 25.8
Chi square
1.18 0.03 0.53 0.002 0.07 0.24
more slowly than pollen tubes carrying the wild-type allele, and this difference in growth rates causes a slight distortion in the segregation ratio (Meinke, in preparation). The other five mutants, with the possible exception of mutant 79A, showed a random distribution of aborted seeds along the length of heterozygous siliques. Segregation ratios were not altered by lower temperatures (Table 4). Siliques of heterozygous plants grown at 18°C segregated 25% aborted seeds, and the stage of developmental arrest was similar to that found in plants grown at 25°C. Normal segregation was also detected in plants grown at 32°C. There is no evidence, therefore, that these mutants are temperature sensitive. Phenotypically normal seeds from each mutant heterozygote produced both heterozygous and wild-type plants in approxi-
MEINKE AND SUSSEX
Embryo-Lethal
mately the 2:l ratio expected for single recessive mutants (Table 5). Heterozygous plants were phenotypically normal except for the presence of aborted seeds, and wildtype plants grown from mutant stocks were indistinguishable from other wild-type plants. Allelism Tests Pair-wise crosses between all six mutant heterozygotes produced 95-100% phenotypically normal seeds (Table 6). Mutants 87A, 123B, 79A, 50B, 71E, and 124D are therefore nonahelic and represent six separate loci. The linkage relationships of these six loci are currently being studied using trisomics and standard morphological markers. Nineteen abortants were identified among 2432 wild-type seeds produced by a total of 43 separate crosses. Seventeen of these abortants were probably the result
Mutants
67
OfArabidopsis
of injury or stress to the developing silique because they were phenotypically unlike abortants from either parent plant. The remaining two abortants probably resulted from accidental self-pollination because development was arrested at a stage characteristic of the female plant. The low frequency of accidental self-pollination with either method of cross-pollination suggests that controlled crosses in Arabidopsis can be accomplished without emasculation if the stigma of mature buds is saturated with pollen from the male parent plant. Stages of Developmental Arrest Aborted seeds from each mutant contained homozygous mutant embryos arrested at a characteristic stage of early embryo development (Table 7). Aborted seeds from the globular-heart mutant (87A) always contained either a late globular or an TABLE
TABLE
5
FREQUENCYOFHETEROZYGOUSPLANTSGROWN FROMPHENOTYPICALLYNORMALSEEDS" Mutant
Plants scored
87A 123B 79A 50B 71E 124D Total
126 136 133 102 142 140 779
66.7 65.4 60.9 67.6 66.2 62.9 64.8
n Expected 66.7%.
frequency
for a recessive
Percentage heterozygous plants
Mutant
Stage of arrest
Chi square
87A 123B 79A 50B 71E 124D
0.00 0.05 1.75 0.01 0.001 0.74 1.10 mutation
is
TABLE
7
DEVELOPMENTALARRESTOFHOMOZYGOUSMUTANT EMBRYOS
Globular-heart Globular Early globular Preglobular Preglobular Preglobular
Aborted seed length (pm)”
Arrested embryo length (pmY
450-670 430-670 430-640 400-640 230-540 250-510
70-280 40-130 40 t30 t30
BIOLOGY
72,62-72
(1979)
Isolation and Characterization of Six Embryo-Lethal Arabidopsis thaliana
Mutants of
DAVID W. MEINKE AND IAN M. SUSSEX Department
of
Biology, Yale University, New Haven, Connecticut 06520
Received November 13, 1978; accepted in revised form April 3, 1979 Mature seeds of Arabidopsis thaliana strain “Columbia” were soaked for 7.5 hr in an aqueous solution of the chemical mutagen ethyl methanesulfonate (0.05,0.10, or 0.50%, v/v). Embryo-lethal mutants were identified in the resulting M-l chimeral plants by screening the first five siliques of each plant and noting the frequency of aborted seeds. Three hundred sixty seeds were treated at each mutagen dose; the frequency of embryo-lethal mutants ranged from l-3% of the M-l plants grown from seeds exposed to 0.05% EMS, to 20-30% of the M-l plants at the highest mutagen dose. Six embryo-lethal mutants identified through screening of M-l plants were chosen for detailed studies in subsequent generations. All six mutants segregate as nonallelic, Mendelian recessive lethals, and are maintained as heterozygotes since homozygotes die as embryos. Fruits of heterozygous plants contain 25% aborted seeds and 75% phenotypically normal seeds (z/3 heterozygotes and % wild type). Segregation ratios are not temperature sensitive; the same frequency of aborted seeds is found in plants grown at 18, 25, and 32°C. Embryo arrest and eventual lethality in each mutant occur at a characteristic stage of early embryo development: globular-heart, globular, early globular, or preglobular. Arrested embryos from five of the six mutants resemble normal embryos at early stages of development. Developmental arrest of the embryo proper in the remaining mutant is followed by abnormal growth of the suspensor, an embryonic structure that attaches the embryo proper to the maternal tissue. INTRODUCTION
opment: globular-heart, globular, or preglobular.
Arabidopsis thaliana was described in the preceding report as a model system for mutant analysis of plant embryo development (Meinke and Sussex, 1979). Normal development of strain “Columbia” plants was characterized, procedures for identifying and classifying embryo-lethal mutants of Arabidopsis were detailed, and the potential application of these mutants to the study of normal embryo development was discussed. The purpose of this report is to present genetic and developmental data on six embryo-lethal mutants isolated following EMS mutagenesis of wild-type seeds. Each mutant segregates as a Mendelian recessive lethal; fruits of heterozygous plants contain 25% aborted seeds and 75% phenotypically normal seeds. Developmental arrest and eventual lethality of homozygous mutant embryos occur at a characteristic stage of early embryo devel-
MATERIALS
early
METHODS
Mutagenesis Wild-type seeds of Arabidopsis thaliana (L.) Heynh strain “Columbia” were kindly provided by Dr. G. P. Redei of the University of Missouri, Columbia. Seeds used for mutagenesis were assumed to be essentially isogenic because they were collected from a single wild-type plant following self-pollination. The chemical mutagen ethyl methanesulfonate (EMS) was vigorously suspended in distilled water, and mutagenesis was performed at room temperature in a well-ventilated hood. Since EMS is a dangerous mutagen and carcinogen, extreme care was taken to minimize the chance of accidental exposure, and to decontaminate and discard exposed equipment (Lewis and 62
0012-1606/79/090062-11$02.00/O Copyright 0 1979 by Academic Press, Inc. All rights of reproduction in any form reserved.
AND
globular,
MEINKE
AND SUSSEX
Embryo-Lethal
Bather, 1968). Mature seeds were soaked for 7.5 hr in glass vials containing 2.0 ml of EMS solution (0.05, 0.10, or 0.50%, v/v). Treated seeds were then washed with water for 12 hr before planting. Three hundred sixty seeds treated at each concentration were planted in 3-in. pots by individually transferring each seed to the soil surface with a Pasteur pipet. Each pot contained five seeds placed at known positions to simplify cataloguing of M-l plants. All M-l plants and subsequent generations were grown in an environmental growth room at 25°C in 16 l-u-/8 hr light/dark cycles. Details of growth conditions and procedures for the identification of aborted seeds and isolation of embryo-lethal mutants in Arabidopsis are presented in the previous report (Meinke and Sussex, 1979).
Genetic Characterization
of Mutants
Segregation ratios were obtained by screening immature, green siliques of M-2 heterozygotes, grown from seed recovered from the mutant sector of M-l plants. Siliques were split longitudinally with a razor blade and examined under a dissecting microscope. Aborted seeds were recognized by their different size and color. Fifteen siliques from each M-2 plant were screened and the percentage of aborted seeds was calculated. Siliques from recessive mutant heterozygotes were expected to contain two types of phenotypically normal seeds: heterozygous and wild type. The relative frequency of these two types was determined by counting the number of heterozygous and wild-type plants that grew from mutant stocks of phenotypically normal seeds. Each mutant was also tested for temperature sensitivity; heterozygous plants raised at 25°C were transferred to 18°C shortly after flowering, when they could be distinguished from wild-type plants, and the percentage of aborted seeds was determined in the first 15 siliques that developed at the lower temperature. Further information on the nature of
Mutants
OfArabidopsis
63
each embryo-lethal mutation was obtained from M-l mutant sector sizes estimated from the relative position and phenotype of the five screened siliques (see Fig. 6 in Meinke and Sussex, 1979). The mutant sector of a recessive mutant chimera was expected to cover roughly ?4 of the main M-l inflorescence. The size of a mutant sector was determined in part by the number of apical meristem cells “hit” during seed mutagenesis; unusually large mutant sectors were often a sign of multiple mutations. Pairwise crosses between heterozygous plants were used to check for allelism among six embryo-lethal mutants. Allelism tests usually involve crossing two different homozygous mutants, but with recessive lethals it was necessary to cross heterozygotes. In this case, a cross between allelic mutants will produce a silique containing 25% aborted seeds, whereas crosses between nonallelic mutants will produce only phenotypicahy normal seeds. These two conditions can be clearly distinguished in siliques containing 30 or more seeds. Allelism tests require controlled crossing, which in Arabidopsis is complicated by small flowers and natural self-pollination. Two different techniques were used to overcome the problem of self-pollination: emasculation and “early pollination.” To emasculate the female parent of a given cross, the largest closed bud from an inflorescence was positioned under a dissecting microscope; sepals and petals were pulled back with fine forceps, and anthers were removed by cutting or crushing each stamen filament. With practice this could be done without noticeable injury to the ovary. The early pollination method did not require the tedious removal of anthers; sepals were pulled back from closed buds, and the exposed stigma was covered with pollen. Self-pollination still occurred, but too late to result in selffertilization. Pollen for either method was obtained by removing dehiscent anthers from open flowers, and cross-pollination was achieved by repeatedly touching at least one anther to the stigma surface.
64 Developmental tants
DEVELOPMENTALBIOLOGY
Characterization
of Mu-
The first step in determining the stage of embryo arrest for each mutant was to dissect aborted seeds under a dissecting microscope. Mutant embryos arrested at or beyond advanced globular stages (>40 pm) were easily removed and measured with a stage micrometer. When arrested embryos were not visible under a dissecting microscope, aborted seeds were squashed in a drop of stain and examined under a compound light microscope. Early globular embryos could be identified by this technique. Mutations that arrested embryogeny even earlier were assumed to be preglobular lethals, and were studied in sectioned material. The length of an aborted seed was also used as an indication of the stage of developmental arrest. For examination of embedded material, mutant ovules were removed from heterozygous siliques as soon as they could be distinguished from phenotypically normal seeds, fixed for 6-12 hr in 6% glutaraldehyde, postfixed in 2% OsOl, and embedded in Luft’s Epon 812. Most ovules had to be nicked with a razor blade to allow sufficient penetration of the embedding medium. Serial sections through the entire ovule were cut at 0.5 pm with an LKB-III ultramicrotome, and stained with a mixture of azure B and methylene blue. Details of these histological procedures are included in a separate paper on abnormal suspensor development in an embryo-lethal mutant of Arabidopsis (Meinke and Marsden, in preparation). RESULTS
Mutagenesis and M-l Plants The results of mutagenesis and screening of M-l plants are summarized in Table 1. The high frequency of germination obtained after all three EMS treatments indicates that seed viability was not affected by the mutagenesis procedure, although seeds exposed to 0.50% EMS germinated l-
VOLUME 72, 1979 TABLE 1 SUMMARYOFMUTAGENESISRESULTSANDM-I SCREENING Mutagen
Treated seeds Germination Plants screened Frequency of embryo-lethal mutants” Mutants studied in M-2 generation
dose
0.05% EMS
0.10% EMS
0.50% EMS
360 95% 211 l-3%
360 96% 294 3-5%
360 93% 50 20-30%
0
4
2
a This is an estimated frequency because no attempt was made to test all suspected mutants in the M-2 generation.
3 days after those exposed to 0.05 and 0.10% EMS. Vegetative growth, flowering, and silique development appeared normal in most M-l plants, and the general reproductive effort was similar to that of wild-type plants. Embryo-lethal mutants could be identified only by examining individual siliques under a dissecting microscope because siliques containing both aborted and normal seeds were the same length and had the same maturation time as wild-type siliques. Nearly 3000 siliques from approximately 600 M-l plants were screened, including most of the plants from the 0.05 and 0.10% EMS treatments and a representative sample from the high mutagen dose. One to five percent of the M-l plants from the 0.05 and 0.10% EMS treatments appeared to be segregating for recessive embryo-lethal mutations, based on examination of the first five siliques. These suspected mutant plants had three or four wild-type siliques with normal seeds and one or two mutant siliques that contained approximately 75% phenotypically normal seeds and 25% aborted seeds. A few plants were difficult to classify as either wild type or mutant in the M-l generation because they had one or two siliques in which the frequency of aborted seeds was significantly lower than 25%. The remaining 90-95% of
MEINKE
AND SUSSEX
Embryo-Lethal
the plants were phenotypically normal and showed no significant increase in the number of aborted seeds. The frequency of suspected embryo-lethal mutants was noticeably higher among plants grown from seeds exposed to 0.50% EMS. Many of these plants were still clearly wild type, but an increase in physiological abortion rates and the appearance of plants with more than one mutant phenotype made classification and screening more difficult. Some plants contained a single recessive mutation with a typical mutant sector and one phenotypic class of aborted seeds; a representative sample of these mutants was recovered and studied in the M-2 generation. Other plants appeared to be segregating for two different embryo-lethal mutations in the same mutant sector, and a few plants had unusually large mutant sectors with several types of aborted seeds. These complex mutants were found only at the high mutagen dose, and they were not studied in more detail.
Mutant Sectors of M-l Plants The shoot apical meristem of mature plant embryos is a multicellular structure composed of diploid cells. M-l plants grown from mutagenized seeds will therefore be mericlinal chimeras with heterozygous mutant cells and siliques occupying only a wedge-shaped sector of the primary reproductive inflorescence (see Fig. 6 in Meinke and Sussex, 1979). The size of a mutant sector is determined by five factors: (1) number of target cells in the mutagenized seed that contribute to the M-l reproductive meristems; (2) number of mutated target cells; (3) cell cycle stage of the mutated cell; (4) molecular nature of the mutation; and (5) competitive growth advantage or disadvantage of the mutant sector in M-l plants. When the shoot apical meristem of a mature embryo contains “x” number of target cells, induction of a recessive mutation in one of these cells should result in a mutant sector covering l/x to 1/4x of the
Mutants
Of Arabidopsis
65
circumference of the primary M-l reproductive meristem. Li and Redei (1969) have estimated from mutagenesis studies that mature seeds of Arabidopsis contain an average of two target cells. The initial size of a mutant sector containing cells heterozygous for an induced recessive mutation was therefore not expected to exceed ‘/z the circumference of the primary reproductive meristem. M-l plants with larger mutant sectors were likely to contain complex or multiple mutations. Lateral and additional branches of a mutant M-l plant could be chimeral, wild type, or mutant, depending on the genotype of contributing cells. Although the size of a mutant sector can change during development of an inflorescence, most evidence suggests that cells of the mutant sector do not have a competitive growth disadvantage (Muller, 1965; Grinikh et al., 1974). Information on the size of a mutant sector must be obtained in the M-2 generation because heterozygous and wild-type cells of the M-l plant are phenotypically identical. With embryo-lethal mutants, however, it was possible to screen M-2 embryos in the fruits of M-l plants, and estimate the size of a mutant sector from the number and position of mutant siliques segregating for aborted seeds. When two or more screened siliques were mutant, both the minimum and the maximum sizes of a mutant sector were estimated, assuming a normal phyllotaxy of 137” between consecutive siliques. The minimum size of a mutant sector could not be determined when only one mutant silique was identified. Mutant sector sizes were determined for 21 M-l plants that appeared to be segregating for an embryolethal mutation. Each suspected mutant plant contained a single mutant sector that included one or two of the first five siliques and covered approximately % to Yt of the primary M-l inflorescence. These results are consistent with the induction of single recessive mutations in seeds containing two target cells. Six mutants were chosen for detailed
66
DEVELOPMENTAL BIOLOGY
studies in subsequent generations. Characteristics of the M-l plants that yielded these mutants are summarized in Table 2. Segregation data for the M-l siliques have been included to show how the mutants were identified. Analysis of the mutant sector data and segregation ratios suggests that each of the six M-l plants carried a single recessive embryo-lethal mutation that was induced in one of the two target cells during mutagenesis.
Segregation and Heterozygote Ratios All six mutants segregated as Mendelian recessive lethals with approximately 25% aborted seeds in the siliques of M-2 heterozygotes (Table 3). The slightly aberrant segregation ratio observed in mutant 124D was correlated with a reduction in the number of aborted seeds in the basal portion of heterozygous siliques. Available evidence suggests that in this particular mutant, pollen tubes carrying the mutant allele grow TABLE
2
M-l CHARACTERISTICS OF SIX EMBRYO-LETHAL MUTANTS STUDIED IN SUBSEQUENT GENERATIONS” Total Mutant Mutant Mutant P-;;Ff seeds siliques” silique number’ seeds 123B
1
5
53
24.5
79A
1
5
55
27.3
124D
1
3
58
15.5
87A
2
2 5
44 26
13.6 26.9
71E
2
3 5
27 10
22.2 20.0
50B
3
2 3 5
39 33 5
23.1 33.3 40.0
n M-l plants were grown from seeds treated with 0.10% EMS (mutants 123B, 79A, 124D, and 87A) or 0.50% EMS (mutants 7lE and 50B). b Number of sfiiques, out of five screened, that appeared to be segregating for an embryo-lethal mutation. ’ Silique 1 developed from the first flower in the main inflorescence, 2 from the second flower, etc.
VOLUME 72.1979 TABLE
3
SEGREGATION OF ABORTED SEEDS IN SILIQUES OF M-2 HETEROZYGOTES Percentage Mutant-plant Seeds Chi square screened aborted seeds 87A-61M 123B-93M 79A-54M 50B-1llM 71E-102M -102L Total 124D-85M -85L -288L Total * Significantly * * Significantly
711 500 701 773 476 636 1112 475 607 836 1918 different different TABLE
24.5 26.0 24.0 25.4 23.1 26.9 25.3 22.1 22.2 21.5 21.9
0.08 0.22 0.34 0.04 0.81 1.11 0.03 1.99 2.33 5.18: 9.68; *
from 25.0 at the 5% level. from 25.0 at the 1% level. 4
SEGREGATION OF ABORTED SEEDS AT 18°C Mutant
Seeds screened
87A 123B 79A 50B 71E 124D
762 658 853 728 690 814
Percentage aborted seeds 23.2 24.6 26.1 25.1 25.5 25.8
Chi square
1.18 0.03 0.53 0.002 0.07 0.24
more slowly than pollen tubes carrying the wild-type allele, and this difference in growth rates causes a slight distortion in the segregation ratio (Meinke, in preparation). The other five mutants, with the possible exception of mutant 79A, showed a random distribution of aborted seeds along the length of heterozygous siliques. Segregation ratios were not altered by lower temperatures (Table 4). Siliques of heterozygous plants grown at 18°C segregated 25% aborted seeds, and the stage of developmental arrest was similar to that found in plants grown at 25°C. Normal segregation was also detected in plants grown at 32°C. There is no evidence, therefore, that these mutants are temperature sensitive. Phenotypically normal seeds from each mutant heterozygote produced both heterozygous and wild-type plants in approxi-
MEINKE AND SUSSEX
Embryo-Lethal
mately the 2:l ratio expected for single recessive mutants (Table 5). Heterozygous plants were phenotypically normal except for the presence of aborted seeds, and wildtype plants grown from mutant stocks were indistinguishable from other wild-type plants. Allelism Tests Pair-wise crosses between all six mutant heterozygotes produced 95-100% phenotypically normal seeds (Table 6). Mutants 87A, 123B, 79A, 50B, 71E, and 124D are therefore nonahelic and represent six separate loci. The linkage relationships of these six loci are currently being studied using trisomics and standard morphological markers. Nineteen abortants were identified among 2432 wild-type seeds produced by a total of 43 separate crosses. Seventeen of these abortants were probably the result
Mutants
67
OfArabidopsis
of injury or stress to the developing silique because they were phenotypically unlike abortants from either parent plant. The remaining two abortants probably resulted from accidental self-pollination because development was arrested at a stage characteristic of the female plant. The low frequency of accidental self-pollination with either method of cross-pollination suggests that controlled crosses in Arabidopsis can be accomplished without emasculation if the stigma of mature buds is saturated with pollen from the male parent plant. Stages of Developmental Arrest Aborted seeds from each mutant contained homozygous mutant embryos arrested at a characteristic stage of early embryo development (Table 7). Aborted seeds from the globular-heart mutant (87A) always contained either a late globular or an TABLE
TABLE
5
FREQUENCYOFHETEROZYGOUSPLANTSGROWN FROMPHENOTYPICALLYNORMALSEEDS" Mutant
Plants scored
87A 123B 79A 50B 71E 124D Total
126 136 133 102 142 140 779
66.7 65.4 60.9 67.6 66.2 62.9 64.8
n Expected 66.7%.
frequency
for a recessive
Percentage heterozygous plants
Mutant
Stage of arrest
Chi square
87A 123B 79A 50B 71E 124D
0.00 0.05 1.75 0.01 0.001 0.74 1.10 mutation
is
TABLE
7
DEVELOPMENTALARRESTOFHOMOZYGOUSMUTANT EMBRYOS
Globular-heart Globular Early globular Preglobular Preglobular Preglobular
Aborted seed length (pm)”
Arrested embryo length (pmY
450-670 430-670 430-640 400-640 230-540 250-510
70-280 40-130 40 t30 t30
