American Journal of Botany 84(6): 781–791. 1997.

GENETIC AND CYTOLOGICAL ANALYSES OF A PARTIAL-FEMALE-STERILE MUTANT (PS-1) IN SOYBEAN (GLYCINE MAX; LEGUMINOSAE)1,2 TELMA N. S. PEREIRA,3 NELS R. LERSTEN,4

AND

REID G. PALMER5,6

3

Department of Agronomy, Iowa State University, Ames, Iowa 50011; Department of Botany, Iowa State University, Ames, Iowa 50011; and 5 USDA ARS FCR and Departments of Agronomy and of Zoology/Genetics, Iowa State University, Ames, Iowa 50011 4

Soybean partial-female-sterile mutant 1 (PS-1) was recovered from a gene-tagging study. The objectives were to study the inheritance, linkage, allelism, and certain aspects of the reproductive biology of the PS-1 mutant. For inheritance and linkage tests, PS-1 was crossed to flower color mutant Harosoy-w4 and to chlorophyll-deficient mutant CD-1, also recovered from the gene-tagging study. For allelism tests, reciprocal crosses were made with PS-1 and three other partial-sterile mutants (PS-2, PS-3, and PS-4) recovered from the same gene-tagging study. The PS-1 mutant is inherited in a 3:1 ratio and is a single recessive gene. Linkage results indicated that the gene for partial sterility in PS-1 is not linked either to the w4 locus or to the CD-1 locus. Allelism tests showed that the gene in PS-1 is nonallelic to the gene in PS-2, PS-3, and PS-4. Investigations of developing and mature pollen indicated no differences in morphology, stainability, or fluorescence between normal and partial-sterile genotypes. The PS-1 mutant is completely male fertile. Confocal scanning laser microscopy was used to determine that early embryo abortion in PS-1 is due indirectly to abnormal migration of the fused polar nucleus, which prevented it from being fertilized. Subsequent absence of endosperm development leads directly to abortion of the proembryo. Key words: double fertilization; embryo abortion; endosperm; female sterility; genetics; Glycine max; Leguminosae; pollen viability.

Transposable elements are segments of DNA that possess the ability to move to new locations in the genome. At the molecular level, the insertion of a transposable element results in a duplication of the target site. The excision of the transposable element removes its DNA sequence from the insertion site. During excision, the target site duplication sequence can undergo a number of changes. As a result, the movement of a transposable element can generate mutations or chromosomal rearrangements that can affect the expression of other genes. Deletions, duplications, translocations, and inversions are some of many chromosomal rearrangements that can be produced by chromosome breakage caused by transposition (Saedler and Nevers, 1985; Nevers, Shepherd, and Saedler, 1986). An unstable mutation for anthocyanin pigmentation in soybean was identified (Groose, Weigelt, and Palmer, 1988) whose mutability is conditioned by a mutable allele

at the w4 locus that is recessive to wild type (Palmer et al., 1989). The population containing the mutable allele is called the w4-mutable line. Mutable plants produce both near-white and purple flowers, as well as flowers of mutable phenotype (w4-m) with purple sectors on nearwhite petals (Groose, Schulte, and Palmer, 1990). The w4-m allele exhibits many features typical of an allele controlled by a transposable element, such as (1) variation in reversion frequency, (2) loss of mutability, (3) partial reversion to ‘‘pale’’ alleles, and (4) recovery of new mutants (Groose, Schulte, and Palmer, 1990). To provide evidence for transposition, Palmer et al. (1989) conducted a gene-tagging study to recover new mutants. Several were isolated, such as mutants for chlorophyll deficiency (CD-1 to CD-8), mutants for root necrosis (NR-1 to NR-3), mutants for partial sterility (PS-1 to PS-4), and a mutant for near sterility. Partial-sterile 1 (PS-1) soybean mutant was recovered from an F11 family. The F11 family, descended from a single F9 plant of the Asgrow Mutable line, was constituted of 12 plants, and all were partial-sterile (Groose and Palmer, 1987). The partial-sterile plants had a reduced number of seeds per pod, which could be the result of either ovule abortion or very early embryo abortion. The objectives of this study were (1) to determine whether the gene for partial sterility in PS-1 mutant is linked to the flower color mutant, w4 locus or to the chlorophyll-deficient mutant, CD-1 locus; (2) to determine whether the gene in PS-1 is an alternative form (allele) of the gene in partial-sterile mutants PS-2, PS-3, and PS-4 or whether it is a new gene; and (3) to compare pollen and embryo sac (megagametophyte) development of the PS-1 mutant with fertile plants.

1 Manuscript received for publication 9 July 1996; revision accepted 26 November 1996. The authors acknowledge Dr. Harry Horner and Bruce Wagner from the Bessey Microscopy Facility, Iowa State University, and Drs. Philip Haydon and Raj Lartius from the Confocal Microscopy Facility, Iowa State University, for expertise in microscopy, and Dr. Hilal Ilarslan for help on the light and confocal photographs. This research was supported by CEPLAC-CNPq-Brazil. 2 Joint contribution of the Iowa Agriculture and Home Economics Experiment Station, Ames, Iowa, Journal paper No. J-15904, project no. 3352, and the USDA ARS FCR. The mention of a trademark or proprietary product does not constitute a guarantee or warranty of the product by Iowa State University or the United States Department of Agriculture and does not imply its approval to the exclusion of other products that may be suitable. 6 Author for correspondence; telephone, 515-294-7378; FAX, 515294-2299.

781

782

AMERICAN JOURNAL MATERIALS AND METHODS

Genetic study—Linkage tests—The genetic materials used in this study were partial-sterile soybean mutant, PS-1, a true breeding partialfemale-sterile mutant, chlorophyll-deficient mutant, CD-1, and flower color isoline mutant w4 in the cultivar Harosoy, Harosoy-w4. PS-1 and CD-1 mutants were found in a gene-tagging study (Groose and Palmer, 1987). CD-1 is a chlorophyll-deficient mutant, and upon self-pollination of heterozygous plants, it gives 3 green:1 yellow-green plants. The yellow-green plants have reduced vigor. Seeds from PS-1, CD-1, and Harosoy-w4 were sown in the field in summer 1990 and 1991. At flowering, PS-1 was crossed with Harosoyw4, and CD-1. Harosoy-w4 was always used as the female parent because it has white flowers, a recessive trait that can be used as a morphological marker. The F1 seeds were used to produce the F2 generation. In the F2 generation, the number of purple/white fertile and purple/ white partial-sterile plants were recorded at harvest to estimate the linkage between PS-1 and Harosoy-w4. The linkage estimation was calculated by using the Linkage-1 computer program, which uses the maximum likelihood method (Suiter, Wendel, and Case, 1983). Fertile and partial-sterile F2 plants were threshed individually in 1991 and 1992. These F2:3 progenies or F3 generation were evaluated each succeeding year, and data were recorded for number of segregating and nonsegregating progenies. The x2 test was calculated to see whether the observed data fit the expected ratio. The F2 yellow-green foliage plants (CD-1) were weak and set too few pods to be classified as either fertile or partial-sterile. Individual F2 green-foliage plants were threshed and evaluated as F2:3 progenies. The F2:3 progenies were classified as either segregating for foliage color or not segregating. For the PS-1 locus, the F2:3 families were classified as segregating for either fertility (normal and partial-sterile) or nonsegregating (normal or partial-sterile). Allelism tests—PS-1 is a true-breeding mutant for reduced number of seeds per pod, whereas PS-2, PS-3, and PS-4 mutants segregate for normal and reduced number of seeds per pod in a 1:1 ratio. The phenotype of PS-1 plants is indistinguishable from the phenotype of PS-2, PS-3, and PS-4 plants, but the breeding behavior is different. Seeds from PS-1, PS-2, PS-3, and PS-4 were sown in the field in summer 1990 and 1991. At flowering, five plants in each row of PS mutants were chosen at random and tagged. Reciprocal cross-pollinations for allelism tests were made among PS mutants. The PS-1 to PS-4 tagged plants were classified phenotypically at maturity for fertility or partial sterility on the basis of the number of seeds per pod. The F1 seeds were used to generate the F2 generation; the F2 seeds to generate the F2:3 generation. The x2 test was calculated to see whether the observed data fit the expected ratio. Embryo abortion—Partial-sterile plants from PS-1 were field-harvested and brought into the laboratory in 1991, 1992, and 1993. A podby-pod record of seed, aborted seed, and aborted ovules, recorded as basal, middle, or apical in position, was made for all tagged plants. Cytological study—Plants were grown in both greenhouse and growth chamber. To obtain vegetative and reproductive growth, the temperature was 298C during day and 268C during night; the photoperiod was 18 h during 4 wk, 16 h during 1 wk, and 14 h until maturity. Pollen development—Flower buds of various sizes were fixed in formalin-acetic acid-alcohol (FAA) at room temperature (RT 5 228C), then dehydrated through a graded ethanol/xylene series; the final xylene stage was gradually infiltrated with Paraplast over a minimum of 3 d. Longitudinal sections of the embedded buds were cut on a rotary microtome at 10 mm and stained with safranin O, counterstained with fast green.

OF

BOTANY

[Vol. 84

Pollen viability—Three methods were used to estimate pollen viability/fertility. Samples of mature pollen were collected at anthesis from each PS-1 tagged plant and from fertile plants and stored in 70% ethanol at cold temperature (CT 5 48C). Pollen diameter was measured by using a calibrated 103 millimetric eyepiece reticule. Pollen grains were classified as normal/viable (stained) or abnormal/nonviable (remained unstained) on the basis of their reaction to an iodine potassium iodide (I2KI) solution (Johansen, 1940). A second differential pollen staining solution also was used (Alexander, 1969). When dusted into a drop of differential staining solution, normal/viable pollen grains stained red and abnormal/nonviable pollen grains stained green. The fluorochrome diacetate reaction (FCR) method (Gwyn and Stelly, 1989) was used to distinguish pollen of normal plants from pollen of partial-sterile plants. Fresh pollen was dusted into a drop of freshly prepared solution, and a coverslip was applied. After 2 min, the slides were observed under a Zeiss fluorescence microscope equipped with epifluorescence by using barrier filter LP 510–550 nm and excitation filter BG 12 (blue). Normal/viable pollen grains had a natural blue-green fluorescence and abnormal/nonviable pollen grains would not fluoresce or fluoresced unevenly and dimly.

Pollen germination—To test for in vitro pollen tube germination, a factorial boric acid 3 sucrose combination was used (Pereira and Palmer, 1996). Optimum results were obtained with 30 mg/kg of boric acid and 10% sucrose. Newly opened flowers were collected from growthchamber-grown plants early in the morning. Pollen grains from ten individual flowers (two flowers each from five plants) were sprinkled onto a drop of solution on each slide. After 1 h at room temperature (RT), germination was recorded if pollen tube length was . 4 times grain diameter. Percentage of germinated and nongerminated pollen was recorded.

Ovule and embryo sac development—For paraffin serial sections, flower buds of different sizes were fixed in FAA at RT for at least 24 h. A gentle vacuum at 15 psi (6.89 kPa) enhanced the penetration of fixative. After infiltration, gynoecia were removed, and both ends were cut off with a razor blade. Gynoecia were dehydrated, infiltrated with Paraplast, sectioned, and stained as described previously. For resin sections, flower buds of different sizes were fixed in 3% glutaraldehyde and 2% paraformaldehyde in sodium cacodylate buffer (0.1 mol/L, pH 7.2) at RT. Gynoecia with both ends cut off were placed in the fixative under vacuum at 15 psi for 1 h, placed in fresh cold fixative overnight, rinsed three times in buffer, postfixed in 1% osmium tetroxide (OsO4) in same buffer for 4 h at RT, and dehydrated in a graded ethanol/acetone series. Gynoecia were embedded in Spurr’s resin (Spurr, 1969), sectioned 1–2 mm thick on a Reichert Ultracut E ultramicrotome, and stained with toluidine blue.

Whole-ovule clearing—Ovules dissected out of a different sample of living floral buds and segregated by ovary, were fixed in FAA at RT, and stored for 24 h at CT. After rinses in water, ovules were stained in aqueous Mayer’s Hemalum (Johansen, 1940) for 20–30 min and destained in 2% acetic acid for 10 min, then dehydrated in a graded ethanol series to 100% ethanol, which was gradually replaced with methyl salicylate (Stelly et al., 1984). Ovules from the same ovary were mounted in methyl salicylate on slides and sealed with nail polish. Slides were stored at CT to inhibit evaporation. The mounted ovules were observed in a Noram Odyssey Confocal Scanning Laser Microscope (CSLM) with attached NIKON-DIAPHOT microscope with a 603/1.4 NA objective and an argon laser. The emission wavelength was 515 nm; the excitation wavelength 488 nm. The thickness of each recorded optical section was ;1 mm, and the microscope was refocused at 2 mm between successive images.

June 1997]

PEREIRA

ET AL.—GENETICS AND CYTOLOGY OF A SOYBEAN MUTANT

TABLE 1. Linkage test of crosses between Harosoy-w4 as female parent and soybean partial-sterile 1 mutant (PS-1) as male parent; 1991 and 1992 data.

783

TABLE 3. Linkage test of crosses between soybean partial-sterile 1 mutant (PS-1) as female parent and CD-1 as male parent. Number of F2:3 progenies

Number of F2 plants Phenotypes

Phenotypes

1991

1992

Total

Purple normal Purple partial-sterile White normal White partial-sterile Total

694 213 211 94 1 212

576 192 193 68 1 029

1270 405 404 162 2 241

x2 (9:3:3:1) Pa % Rb SE

6.49 0.09 55.0 62.00

0.23 0.75 51.00 62.00

4.41 0.22 53.00 62.00

Purple : white x2 (3:1) P

907:305 0.02 0.89

768:261 0.07 0.79

1 675:566 0.08 0.78

Normal : partial-sterile x2 (3:1) P

905:307 0.07 0.79

769:260 0.04 0.84

1 674:567 0.11 0.784

a b

Normal Normal and partial-sterile Partial-sterile

P: probability. %R: percentage recombination 6 standard error (SE).

RESULTS Genetic study—Linkage tests—Segregation data and x2 values for the F2 generation from crosses between PS-1 and Harosoy-w4 were calculated for two consecutive years (Table 1). In both years, the results fit the ratio of 9:3:3:1, the expected ratio for two independent loci. The percentage recombination was 55% 6 2 and 51% 6 2, respectively, for both years. The results of F2:3 segregation data (Table 2), confirmed two independent loci, each one segregating an 1:2:1 ratio. The results of linkage tests of PS-1 with CD-1 are shown in Table 3. Due to the weakness of F2 yellowgreen plants, tests were conducted on F2:3 progenies. From a total of 219 F2:3 families, the observed foliagecolor data fit the expected 1:2 ratio of nonsegregating to segregating. The seed-set data fit the expected 1:2:1 ratio. Allelism tests—The allelism tests data showed that the gene that causes partial sterility in the PS-1 mutant is

a

Purple : segregating : white x2 (1:2:1) Pa

Segregating

White

17 34 15

46 70 36

22 40 20

66:152:82 1.76 0.41

Normal : normal and partial-sterile : partial-sterile x2 (1:2:1) P a

P: probability.

85:144:81 1.80 0.41

43 67 35 74:145 0.02 0.89 57:101:61 1.31 0.60

P: probability.

nonallelic to the gene(s) that causes partial sterility in PS-2, PS-3, and PS-4 mutants. In the F2 population from crosses between PS-1 and PS-2, two types of ratios were observed; one segregated 3:1 (normal:partial sterile), and the other segregated 3:5 (normal:partial sterile). From 572 F2 individual plants, the pooled x2 value fit the expected 3:1 ratio (Table 4). The homogeneity x2 value indicated that all progenies came from the same population (Table 4). Also within this same F2 population (Table 5), 786 individual plants were classified, and the results gave a 3:5 ratio. For the F2 individual plants, results from crosses of PS-1 with PS-3 and PS-4 were similar to F2 results of crosses of PS-1 with PS-2. With PS-3 (Table 4), 294 plants fit the 3:1 ratio and gave a x2 value of 0.22, whereas 483 plants fit the expected 3:5 ratio and gave a x2 value of 0.72 (Table 5). The homogeneity x2 value was 0.46, which suggested that the families were homogeneous. The crosses with PS-4 also had F2 families that segregated 3:1 (Table 4) and 3:5 (Table 5). The segregation data fit the expected ratios. The F2:3 progeny tests confirmed the F2 results. The gene in PS-1 is nonallelic to the gene in PS-2, PS-3, and PS-4. The F2 families that segregated 3:1 were expected TABLE 4. Allelism test of crosses between partial-sterile soybean mutants in the F2 generation. The F2 plants are expected to segregate 3:1 (normal : partial-sterile). Fertility Normal

Partialsterile

PS-1 3 PS-2 Total Pooled x2 (1 df) Homogeneity x2 (2 df)

433

139

PS-1 3 PS-3 Total Pooled x2 (1 df) Homogeneity x2 (2 df)

224

PS-1 3 PS-4 Total Pooled x2 (1 df) Homogeneity x2 (2 df)

38

Cross

Purple

14 34 26

Normal : normal and partial-sterile : partial-sterile x2 (1:2:1) P

Number of F2:3 progenies Phenotypes

Green and yellow-green

Green : green and yellow-green x2 (1:2) Pa

TABLE 2. Linkage test of crosses between Harosoy-w4 as female parent and soybean partial-sterile 1 mutant (PS-1) as male parent.

Normal Normal and partial-sterile Partial-sterile

Green

a

P: probability.

70

15

x2(3:1)

Pa

0.29 0.15 0.14

0.70 0.93

1.16 0.22 0.94

0.64 0.62

0.33 0.31 0.02

0.60 0.89

784

AMERICAN JOURNAL

TABLE 5. Allelism test of crosses between partial-sterile soybean mutants in the F2 generation. The F2 plants are expected to segregate 3:5 (normal : partial-sterile).

OF

BOTANY

[Vol. 84

TABLE 7. Allelism test between soybean partial-sterile mutants segregating 3:5 in the F2, based upon F2:3 families. Families are expected to segregate 3:5 (nonsegregating : segregating) or 1:2:5 (nonsegregating normal : nonsegregating partial-sterile : segregating).

Fertility Cross

PS-1 3 PS-2 Total Pooled x2 (1 df) Homogeneity x2 (2 df)

Normal

Partialsterile

297

489

PS-1 3 PS-3 Total Pooled x2 (1 df) Homogeneity x2 (2 df)

172

PS-1 3 PS-4 Total Pooled x2 (1 df) Homogeneity x2 (2 df)

41

a

311

79

Number of entries x2(3:5)

Pa

Nonsegregating Cross

0.00 0.02 0.17 1.18 0.72 0.46 0.96 0.57 0.39

Na

PSb

Segregating N and PS x2(1:1)

Pc

x2(1:1:2)

P

0.89 0.92

PS-1 3 PS-2 Total

6

10

25

0.11

0.76

0.14

0.93

0.40 0.80

PS-1 3 PS-3 Total Pooled x2 (1 df) Homogeneity x2 (3 df)

14

25

81

1.86 1.28 0.58

0.26 0.90

1.38

0.50

0.45 0.82

PS-1 3 PS-4 Total Pooled x2 (1 df) Homogeneity x2 (1 df)

13

1.22 0.29 0.93

0.59 0.63

0.35

0.85

P: probability.

24

55

a

N: normal. PS: partial-sterile. c P: probability. b

to segregate 1:1 or 1:2:1 in F2:3 (Table 6), and those that segregated 3:5 were expected to segregate 3:5 or 1:2:5 in F2:3 (Table 7). The observed data fit the expected ratios. Embryo abortion—Table 8 shows the percentage of embryo abortion for the PS-1 mutant in 1990, 1991, and 1993. The percentage of embryo abortion in two-ovule pods and three-ovule pods was similar and random with respect to position within the pod. Cytological study—Pollen development—Pollen development in PS-l mutant was normal. During the sporogenous stage (Fig. 1), the mass of densely cytoplasmic sporogenous cells was surrounded directly by the tapetum. Microspore mother cells (MMC) became isolated by callose shortly before meiosis (Fig. 2). After meiosis and cytokinesis, the four microspores remained for a short time within the original callose layer as a tetrad. Callose dissolved later, releasing the individual microspores, TABLE 6. Allelism test between soybean partial-sterile mutants segregating 3:1 in the F2, based upon F2:3 families. Families are expected to segregate 1:1 (nonsegregating : segregating) or 1:1:2 (nonsegregating normal : nonsegregating partial-sterile : segregating).

which gradually developed numerous vacuoles, large plastids lacking starch, and a large nucleus oppressed to the microspore wall. Mitosis and cytokinesis within the microspores produced binucleate pollen grains with a large vegetative cell and a small generative cell. No differences were noted between mature pollen grains from partial-sterile plants and pollen grains from fertile plants. All pollen grains were round and stained very well with safranin O–fast green (Fig. 3). Pollen viability—The two pollen viability techniques did not show any distinguishable differences between the pollen grains of fertile plants and of partial-sterile plants. All pollen grains from PS-1 plants were well stained, round, engorged with starch, and appeared identical to pollen from fertile plants. No differences in grain diameter were observed; the average diameter was 30 mm. Pollen germination—Pollen germination was optimal at 10% sucrose combined with 30 mg/kg boric acid (Pereira and Palmer, 1996). The in vitro pollen germination from PS-1 mutant (Fig. 4) was 87.8% and for normal plants was 89.0%.

Number of entries Nonsegregating

Ovule and embryo sac development—Megagametogenesis was observed by using standard light microscopy

Segregating N and PS x2(1:1)

Na

PSb

7

7

17

0.00

PS-1 3 PS-3 Total Pooled x2 (1 df) Homogeneity x2 (3 df)

15

12

33

0.66 0.60 0.06

PS-1 3 PS-4 Total Pooled x2 (1 df) Homogeneity x2 (1 df)

11

Cross

PS-1 3 PS-2 Total

a

N: normal. PS: partial-sterile. c P: probability. b

15

27

0.03 0.02 0.01

Pc

x2(1:1:2)

P

1.00

0.29

0.85

TABLE 8. Embryo abortion in soybean partial-sterile mutant 1 (PS-1) for 3 yr (Ames, Iowa). Percentage of embryo abortion in pods with

0.44 0.99

0.89 0.92

0.90

0.55 Phenotypes

0.62

0.75

PS-1 Normal PS-1 Normal PS-1 Normal a

n.d.: no data.

Year

1990 1991 1993

Number of plants

Two seeds

Three seeds

30 14 30 32 5 n.d.a

34.1 17.8 34.9 21.3 31.0 n.d.

44.2 25.1 41.5 17.2 45.0 n.d.

June 1997]

PEREIRA

ET AL.—GENETICS AND CYTOLOGY OF A SOYBEAN MUTANT

785

Figs. 1–4. Pollen development in PS-1 soybean mutant. 1. Sporogenous mass stage. MMC nucleus with nucleolus (arrow). Bar 5 10 mm. 2. MMC nuclei are enlarged preceding meiosis, and the tapetum is differentiated. Callose is being deposited in white area around each MMC. 3. Oblique anther section with mature pollen grains. Tapetal cells already disintegrated, 4. Pollen-tube germination after 1 h at room temperature in 10% sucrose 3 30 mg/kg boric acid solution. Bars 5 25 mm.

and confocal scanning laser microscopy (CSLM). The soybean pistil is unicarpellate and contains from one to four ovules. The ovules, which alternate along the placental suture, are bitegmic, campylotropous, and crassinucellate. Ovules from the PS-1 mutant exhibited the Polygonum type of gametophyte development as in normal soybean. The archesporium first was distinguished as a group of two to six cells arising two or three layers beneath the epidermis of the developing ovule. One of these cells enlarged and became the archesporial cell. At this stage, the integuments were initiated as small outgrowths at the base adjacent to the funiculus. The archesporial cell continued to enlarge into an ovoid megaspore mother cell (MMC) (Fig. 5). Meiosis resulted in a linear tetrad of haploid megaspores, of which the three micropylar ones disintegrated. The chalazal megaspore enlarged and underwent three successive mitoses, resulting in an eightnucleate megagametophyte. At time of fertilization, the mature embryo sac had an egg cell and two synergids at the micropylar end, and the already fused polar nuclei in the middle portion of the central cell. The egg cell presented a large vacuole at its micropylar end and dense cytoplasm with nucleus at its chalazal end (Fig. 6). The fused polar nucleus exhibited two unfused nucleoli (Fig. 7). The synergids, each with a filiform apparatus (FA), (Fig. 8) were in close association with the egg cell. Before fertilization the antipodals (Fig. 9) at the

chalazal end degenerated, and the central cell became engorged with starch. The fused polar nuclei in PS-1 ovules were not always close to the egg cell (Figs. 10, 11). The distance between the egg cell and fused polar nucleus was measured under CSLM by using a microcalibrator. Ninety-three ovules from PS-1 mutant and 93 from normal plants were observed. Three distinct distance classes between fused polar nucleus and egg cell were recorded from ovules of PS-1 plants (Table 9). The closest was ;86.5 mm, the intermediate was ;111.0 mm, and the farthest was ;171.1 mm. Ovules from normal plants had only the closest two distance classes. In ovaries from PS-1 at 5 d postanthesis, ovules of different sizes could be detected. The large ovules had a well-developed proembryo (Fig. 12), while small ovules with fused polar nuclei all of the far distance class (Table 10) were already aborted (Figs. 13–16). These observations indicate that in the small ovules there was no fusion of second male sperm and fused polar nucleus. Calcium oxalate crystals were observed under polarization in tissues of normal and PS-1 ovules, but the PS-1 ovules had fewer crystals than did normal ovules. The normal ovules from PS-1 mutant showed typical proembryos through the globular stage, and normal endosperm development. DISCUSSION Genetic study—Linkage tests—PS-1 was obtained from a soybean population containing a putative trans-

786

AMERICAN JOURNAL

OF

BOTANY

[Vol. 84

Figs. 5–10. Megasporogenesis in ovules from PS-1 soybean mutant (CSLM preparations). 5. Megaspore mother cell showing enlarged nucleus (arrow) with nucleolus and newly initiated integuments. 6. Egg cell with a large micropylar vacuole (white arrow) and chalazal cytoplasm with nucleus (black arrow). 7. Middle position of central cell showing the fused polar nucleus (arrow) with two unfused nucleoli. 8. Synergids (arrow) at micropylar end, dark area within each is the FA. 9. Three antipodals at chalazal end (arrow). 10. Fused polar nucleus (arrow) near the egg cell at micropylar end in ovules from normal soybean plant. Bars 5 10 mm.

June 1997]

PEREIRA

ET AL.—GENETICS AND CYTOLOGY OF A SOYBEAN MUTANT

787

Figs. 11–16. Embryo sac comparison between normal soybean plant and PS-1 soybean mutant (CSLM preparations). 11. Fused polar nucleus (arrow) positioned far from the egg cell at micropylar end in abnormal ovules from PS-1 mutant. This type of ovule probably will abort. 12. Proembryo from normal soybean plant 5 d after anthesis. Note the free endosperm nucleus (arrows). 13–16. Aborted ovule at four optical levels with embryo sac showing traces of degenerated young proembryo in ovules from PS-1 mutant 5 d after anthesis. Bars 5 10 mm.

788

AMERICAN JOURNAL

TABLE 9. Ovules in three classes reflecting distance between egg cell and fused polar nucleus in partial-sterile 1 mutant (PS-1) and normal soybean plants. Fused polar nucleus/egg cell distance (mm) classes ,100

PS-1 Number of ovules Mean Range Normal Number of ovules Mean Range

28 86.5 64.9–99.2

100–140

20 113.6 100.5–135.1

.140

OF

Fused polar nucleus distance (mm) from egg cell or proembryo

1 DBAa

Anthesis

5 DAAb

1 DBA

Anthesis

,100 100–140 .140

14 12 7

14 8 8

0 0 30

22 26 0

25 20 0

45 172.3 151.2–189.2

b

46 120.3 101.8–140.6

posable element at the w4 locus, therefore we tested for linkage at the w4 locus. F2 segregation data, x2 values, and P values showed that the gene in the partial-sterile 1 mutant was not linked to the w4 locus. The F2 data fit the expected 9:3:3:1 ratio, and each locus segregated in a 3: 1 ratio. The F2:3 data supported the F2 data in that the two loci are inherited independently. Three chlorophyll-deficient (CD) mutants, as well as three necrotic root mutants, were obtained from the same population as PS-1, and all are inherited independently of the w4 locus (Hedges and Palmer, 1992; Kosslak et al., 1996). Because transposable elements are segments of DNA that move in the genome from one chromosome to another or to different sites within the same chromosome, they can transpose to linked or unlinked sites. On the basis of their results in maize, Van Schaik and Brink (1958) suggested that transposable elements often transpose to linked sites. Modulator (Mp) and Activator (Ac), both from maize, have preferential transposition to closely linked sites (McClintock, 1952; Greenblatt, 1984). Mu elements in maize preferentially insert into nonmethylated sequences (Bennetzen, Brown, and Springer, 1988). The PS-1 mutant and nine more mutants were recovered from a population containing a putative transposable element, and none was linked to the w4 locus. This suggests that, if a transposable element had inserted at the w4 locus, it moved 50 centiMorgans or more from the w4 locus. No molecular evidence is available to support the hypothesis that the w4 locus has a transposable element insertion, so any conclusions about transposition are tenuous. The linkage test between PS-1 and CD-1 also suggested that the two loci are inherited independently. If the gene for plant color in CD-1 was linked to the gene for partial sterility in PS-1, we would expect an excess of green and partial-sterile families as well as an excess of families segregating for plant color and fertility. This was not observed, indicating that the CD-1 locus is not linked to the PS-1 locus. The linkage tests gave information that PS-1 is a single recessive gene, confirming the true breeding status of PS-1. Because sterility is not complete in the PS-1 mutant, the gene in PS-1 could be a leaky mutant. In leaky mutants, generated by point mutation, the base substitution may reduce the activity of the protein, but it does not necessarily abolish it (Lewis, 1993). In the bt2 and sh2 maize leaky mutants, ADP-glucose pyrophosphorylase activity is 90–95% diminished in the endosperm but

[Vol. 84

TABLE 10. Number of ovules in partial-sterile 1 mutant (PS-1) and normal soybean plants observed in regard to fused polar nucleus position and ovary stage.

a

47 87.9 81.9–98.3

BOTANY

PS-1

Normal

DBA: day before anthesis. DAA: day after anthesis, ovules with aborted embryos.

is not significantly reduced in the embryo, pollen, and chloroplast (Hannah, Tuschall, and Mans, 1980). In the PS-1 mutant, the number of seeds per pod and number of pods per plant were reduced, but they were not abolished; otherwise, PS-1 would be completely sterile. Allelism tests—The gene in PS-1 is nonallelic to the gene(s) in PS-2, PS-3, and PS-4. Therefore, the expectation in the F2 generations is that families will segregate in a ratio of 3 fertile plants:1 partial-sterile plant, and plants within partial-sterile families will segregate in a ratio of 3 fertile:5 partial-sterile plants. These results were expected due to the different genetic behavior of the partial-sterile mutants. The gene in PS-1 is a single recessive gene, so the partial-sterile plants do not segregate upon self-pollination but are true breeding. In the next generation, the gene in heterozygous PS-2, PS-3, and PS-4 plants segregates fertile and partial-sterile plants in an approximate 1:1 ratio (Pereira, Ilarslan, and Palmer, 1997). The gene in PS-2, PS-3, and PS-4 plants was not transmitted through the female parent (Pereira, Ilarslan, and Palmer, 1997). The gene behaves as a lethal ovule, similar to the lethal ovule factors reported in maize by Singleton and Mangelsdorf (1940), Clark (1942), and Nelson and Clary (1952). All ovules in PS-2, PS-3, and PS-4 plants carrying the gene abort because of some abnormality that prevents fertilization. Therefore, when PS-2, PS-3, and PS-4 plants were used as female parents and PS-1 was the male parent, the partial-sterility trait was not transmitted to the next generation. The F1 was fertile and the F2 generation segregated 3:1 (fertile:partial-sterile) for PS-1. Thus, the data presented for the allelism tests with PS-1 have PS-2, PS-3, and PS-4 only as the male parent. When PS-2, PS-3, and PS-4 were used as male parent and PS-1 as female parent, two classes of F1 plants were obtained: one fertile and the other partial-sterile. The fertile F1 plants would segregate in a 3:1 ratio in the F2, and the F2:3 progenies would be segregating and nonsegregating. The F2:3 nonsegregating families would be truebreeding fertile or true-breeding partial-sterile. The F2:3 segregating families would segregate in a 3:1 ratio. The partial-sterile F1 plants would segregate in a 3:5 ratio (fertile:partial-sterile) in the F2, and the F2:3 progenies would be nonsegregating and segregating in a 3:5 ratio. In the 5 partial-sterile class, one class was double partial-sterile because it had both genes for partial sterility, one from PS-1 and the other from PS-2, PS-3, or PS-4. This double partial-sterile was easily identified in the field because it had green leaves and stem, delayed

June 1997]

PEREIRA

ET AL.—GENETICS AND CYTOLOGY OF A SOYBEAN MUTANT

maturity, and few two- and three-seeded pods. RobinsonBeers, Pruitt, and Gasser (1992), in complementation tests with two sterile mutants in Arabidopsis, reported the occurrence of double mutants that exhibited a largely additive phenotype for sterility. Robertson (1978) reported that only 20% of the Muinduced mutations in maize that appeared to be allelic were, in fact, allelic. Clark and Sheridan (1991) isolated 51 embryo-specific (emb) mutations from an active Robertson’s Mutator maize stock, of which at least 88% represented independent mutation events. In Arabidopsis, bel1 and sin are chemically induced mutants that have altered ovule development. Genetic analyses indicated that bel1 segregates as a single recessive gene, and complementation tests showed that the two mutants are nonallelic (Robinson-Beers, Pruitt, and Gasser, 1992). Embryo abortion—The embryo abortion data for the double partial-sterile plants were 41% for two-seeded pods and 51% for three-seeded pods. This compares with the embryo abortion in PS-1, which averaged 33% for two-seeded pods and 43.5% for three-seeded pods over 3 yr. In the present study, the embryo abortion for normal plants was between 19.5 and 19.9% for two- and threeseeded pods, respectively, over 3 yr. Palmer and Heer (1984) recorded 40.6% ovule abortion in heterozygous chromosome translocation plants and 15% ovule abortion and seed abortion in homozygous normal chromosome and homozygous translocated chromosome genotypes. They also observed that ovule abortion in heterozygous chromosome translocation plants was equally frequent among all ovule positions, both in two- and three-ovule pods. In PS-1 mutant the distribution of embryo abortions within the pod also was at random. In homozygous normal chromosome and homozygous translocated chromosome genotypes, seed abortions were more frequent in the basal position of the pod than in either the middle or apical positions. Our results for embryo abortion and distribution of ovule/embryo abortion are in approximate agreement with those of Palmer and Heer (1984). Cytological study—Pollen development, viability, and germination—Partial-sterile 1 mutant was recovered from a soybean population containing a putative transposable element. Transposable elements can cause chromosome breakage. These chromosome rearrangements can generate partial sterility detectable by pollen analysis. Any disruption during pollen formation leads to some degree of sterility (Albertsen and Palmer, 1979). All mature pollen grains from PS-1 mutant appeared normal and stained intensely red with safranin O–fast green. The two techniques we used to distinguish between pollen from fertile plants and pollen from PS-1 plants demonstrate that partial sterility in the PS-1 mutant is not due to defective pollen. Ovule and embryo sac development—According to Gottschalk and Kaul (1974), many genes affect the pathways for microsporogenesis and microgametogenesis, but nuclear male-fertile, female-sterile mutations are rare. Few female gametophyte mutations have been described to date, which most likely reflects the technical complexity of identifying and characterizing the mutations rather

789

than the actual number of genes involved (Reiser and Fisher, 1993). Some examples of crops in which female sterility has been reported are: cotton (Stroman, 1941), sorghum (Casady, Heyne, and Weibel, 1960), alfalfa (Bingham and Hawkins-Pfeiffer, 1984), common bean (Myers and Bassett, 1993), and birdsfoot trefoil (Nyrakabbi and Beuselinck, 1995). The sorghum mutant phenotype is due to the complementary effect of two dominant genes; the other reported cases seem to be caused by single recessive gene mutations. In the present study, we used CSLM to circumvent several procedures such as embedding, orientation, and sectioning. Unstained whole ovules in clearing fluid could be studied individually, and a large number of ovules were screened. Megagametogenesis in PS-1 ovules exhibited the normal Polygonum type of development, which results in an eight-nucleate megagametophyte (Kennell and Horner, 1985; Carlson and Lersten, 1987). The antipodal cells were ephemeral, degenerating before pollen tube entry, also typical of soybean megagametophytes. At fertilization, the central cell of PS-1 mutant ovules was filled with starch. Synergid degeneration at fertilization was variable; some ovules had both synergids, and others had only the persistent one. Kennell and Horner (1985) suggested that the viability of synergids at fertilization is variable among ovules and may be correlated with the time of pollen tube entry. The position of the fused polar nucleus in ovules of partial-sterile plants was not consistent. In a normal soybean embryo sac at fertilization, the fused polar nucleus migrates to a position against the top of the egg cell or directly above it (Kennell and Horner, 1985; Carlson and Lersten, 1987). We have shown that the fused polar nucleus in the PS-1 mutant may be far from the egg cell or close to it. Ovules with the fused polar nucleus far from the egg cell would abort, whereas ovules with a nearby fused polar nucleus were fertilized and became seed. The 15 ovules from PS-1 with fused polar nucleus far from the egg cell (.140 mm) were hypothesized to abort. The reason for nonmigration of the fused polar nucleus is not known. The young soybean central cell has a large vacuole at the late four-nucleate embryo sac stage (Kennell and Horner, 1985; Folsom and Cass, 1992). Cass, Peteya, and Robertson, (1985) suggested that, in barley, the large central vacuole influences nuclear position for subsequent mitotic divisions. Folsom and Cass (1992) reported that this vacuole diminished until only a few small scattered vacuoles remained. In the ovules that we studied, the chalazal polar nucleus was interpreted as migrating in a micropylar direction, the micropylar polar nucleus chalazally, until they became paired. If the central cell vacuole in PS-1 ovules did not shrink sufficiently, the migration of the fused polar nucleus could not occur. Our light-microscopic observations support this hypothesis, but ultrastructural studies of PS-1 ovules will be necessary to explain the nonmigration of fused polar nuclei. Vollbrecht and Hake (1995) used deficiency analysis of small cytologically defined chromosomal regions in maize and concluded that megasporogenesis is composed of distinct events. They believe that ‘‘polar nucleus migration is not likely an internuclear phenomenon; rather,

790

AMERICAN JOURNAL

nuclear migration within the central cell may respond primarily to extranuclear, positional cues, with each polar nucleus migrating towards the same micropylar location.’’ PS-1 ovules 5 d postanthesis had an aborted proembryo and a distantly positioned fused polar nucleus, which suggests that distance was critical in preventing fertilization by the second sperm. Ovaries from the PS-1 mutant showed ovules of different sizes at 5 d postanthesis. Small ovules were degenerating, whereas large ovules were fertilized. Serial and optical sections showed that the degenerated ovules ranged from zygote to the late proembryo stage. Depending on the stage, the degenerated ovules were found with the fused polar nucleus intact, suggesting that it was not fertilized. No endosperm formed. In the later stages of degeneration, such as late proembryo, no traces of endosperm were found. Thus, the embryo evidently aborted due to failure of endosperm formation, which resulted from failure of double fertilization. Similar results were reported by Cichan and Palser (1982) in Cichorium intybus. They found that aborting ovules at the four-celled to early globular proembryo stages lacked endosperm and had a fused, possibly unfertilized, polar nucleus in the central cell. The proembryos were normal in appearance up to about the 30-celled stage, probably because they were supplied with nutrients by the periendothelial zone. At the globular stage, when the Cichorium ovule is normally supplied with nutrients by the endosperm, these ovules aborted (Cichan and Palser, 1982). In soybean, Chamberlin, Horner, and Palmer (1994) observed that the early zygote lacked a cell wall and was connected to the central cell by plasmodesmata, suggesting that the zygote may absorb nutrients from the central cell. Folsom and Cass (1992) also reported that the plasmodesmata were responsible for the movement of nutrients (breakdown products of starch) from the central cell to the soybean proembryo. A supply of metabolites by the starch-filled central cell could explain why PS-1 soybean ovules appear normal from the zygote to the late proembryo stage. The proembryo might not require nutrients from the endosperm (Chamberlin, Horner, and Palmer, 1994), but endosperm is required for later embryo development (Brink and Cooper, 1947). In interspecific crosses in tobacco, young seeds aborted from an impaired capacity for growth of the endosperm, and excessive development of the adjacent maternal tissue (Cooper and Brink, 1940; Brink and Cooper, 1941). The endosperm might also be necessary to provide space in the central cell for the growing embryo. Smaller ovules and fewer calcium oxalate crystals in PS-1 suggest less metabolic activity. Davis (1961) speculated that calcium oxalate crystals store calcium for the developing embryo in Podolepis jaceoides. In chicory ovules (Cichan and Palser, 1982), these crystals also may be related to tissue dissolution because the integumentary crystals disappear coincidentally with the breakdown of cells, which leads to expansion of the periendothelial zone in normally developed seeds, whereas crystals persist in those ovules in which there is no expansion of the periendothelial zone. Several lines of evidence suggest that in embryo abortion the primary defect often is in the endosperm tissue,

OF

BOTANY

[Vol. 84

and histological studies have shown that developmental arrest of the embryos is often preceded by degeneration of the endosperm (Meinke, 1986). Embryo abortion is associated with an unfertilized fused polar nucleus that fails to migrate to or near the egg cell (Mogensen, 1982; Vallania, Botta, and Me, 1987; Arthur, Ozias-Akins, and Hanna, 1993). Ovule abortion in angiosperms has been associated with a variety of factors such as lack of fertilization, failure of the embryo sac to complete its development, high temperatures, reduction of cell divisions in the nucellus, competition due to heavy crop load, suppression of other ovules in a given ovary by the first one to be fertilized, and blockage of the vascular trace leading to the ovule (Mogensen, 1982). CONCLUSIONS Genetic evidence from F2 data indicated that the PS-1 mutant is a single nuclear recessive gene. The PS-1 gene is not linked to the w4 locus or to the CD-1 locus. Allelism tests of PS-1 with the three other partial-sterile mutants, recovered from the same gene-tagging study, showed nonallelism. Pollen from PS-1 plants can effect fertilization. The evidence presented here supports the view that embryo abortion in the PS-1 mutant is due to abnormal fused polar nucleus behavior. The fused polar nucleus does not migrate to its correct position, therefore double fertilization does not occur. This results in the lack of endosperm formation and, consequently, early embryo abortion and cessation of ovule growth. LITERATURE CITED ALEXANDER, M. P. 1969. Differential staining of aborted and nonaborted pollen. Stain Technology 44: 117–122. ALBERTSEN, M. C., AND R. G. PALMER. 1979. A comparative light- and electron-microscope study of microsporogenesis in male sterile (MS1) and male fertile soybeans (Glycine max (L.) Merr.). American Journal of Botany 66: 253–265. ARTHUR, L., P. OZIAS-AKINS, AND W. W. HANNA. 1993. Female sterile mutant in pearl millet: evidence for initiation of apospory. Journal of Heredity 84: 112–115. BENNETZEN, J. E., W. E. BROWN, AND P. S. SPRINGER. 1988. The state of DNA modification within and flanking maize transposable elements. In O. Nelson [ed.], Plant transposable elements, 237–250. Plenum Press, New York, NY. BINGHAM, E. T., AND J. HAWKINS-PFEIFFER. 1984. Female sterility in alfalfa due to a recessive trait retarding the integument development. Journal of Heredity 75: 231–233. BRINK, R. A., AND D. C. COOPER. 1941. Incomplete seed failure as a result of somatoplastic sterility. Genetics 26: 487–505. , AND . 1947. The endosperm in seed development. Botanical Review 13: 423–541. CARLSON, J. B., AND N. R. LERSTEN. 1987. Reproductive morphology. In J. R. Wilcox [ed.], Soybeans: improvement, production, and uses, 2d ed., 95–134. American Society of Agronomy, Madison, WI. CASADY, A. J., E. G. HEYNE, AND D. E. WEIBEL. 1960. Inheritance of female sterility in sorghum. Journal of Heredity 51: 35–38. CASS, D. D., D. J. PETEYA, AND B. L. ROBERTSON. 1985. Megagametogenesis development in Hordeum vulgare. 1. Early megagametogenesis and the nature of cell wall formation. Canadian Journal of Botany 63: 2164–2171. CHAMBERLIN, M. A., H. T. HORNER, AND R. G. PALMER. 1994. Early endosperm, embryo, and ovule development in Glycine max (L.) Merr. International Journal of Plant Science 155: 421–436.

June 1997]

PEREIRA

ET AL.—GENETICS AND CYTOLOGY OF A SOYBEAN MUTANT

CICHAN, M. A., AND B. F. PALSER. 1982. Development of normal and seedless achenes in Cichorium intybus (Compositae). American Journal of Botany 69: 885–895. CLARK, F. J. 1942. Cytological and genetic studies of sterility in inbred and hybrid maize. Connecticut Agricultural Experiment Station Bulletin 465: 705–726. CLARK, J. K., AND W. F. SHERIDAN. 1991. Isolation and characterization of 51 embryo-specific mutations of maize. Plant Cell 3: 935–951. COOPER, D. C., AND R. A. BRINK. 1940. Somatoplastic sterility as a cause of seed failure after interspecific hybridization. Genetics 25: 593–617. DAVIS, G. L. 1961. The life history of Podolepis jaceoides. II. Megasporogenesis, female gametophyte and embryogeny. Phytomorphology 11: 206–219. FOLSOM, M. W., AND D. D. CASS. 1992. Embryo sac development in soybean: the central cell and aspects of fertilization. American Journal of Botany 79: 1407–1417. GOTTSCHALK, W., AND M. L. H. KAUL. 1974. The genetic control of microsporogenesis in higher plants. Nucleus 17: 133–166. GREENBLATT, I. 1984. A chromosome replication pattern deduced from pericarp phenotypes resulting from movement of the transposable element Modulator in maize. Genetics 108: 471–485. GROOSE, R. W., AND R. G. PALMER. 1987. Partially sterile mutants in a genetically unstable line of soybean. Soybean Genetics Newsletter 14: 169–171. , S. M. SCHULTE, AND R. G. PALMER. 1990. Germinal reversion of an unstable mutation for anthocyanin pigmentation in soybean. Theoretical and Applied Genetics 79: 161–167. , H. D. WEIGELT, AND R. G. PALMER. 1988. Somatic analysis of an unstable mutation for anthocyanin pigmentation in soybean. Journal of Heredity 79: 263–267. GWYN, J. J., AND D. M. STELLY. 1989. Method to evaluate pollen viability of upland cotton: tests with chromosome translocations. Crop Science 29: 1165–1169. HANNAH, L. C., D. M. TUSCHALL, AND R. J. MANS. 1980. Multiple forms of maize endosperm ADP-glucose pyrophosphorylase and their control by shrunken-2 and brittle-2. Genetics 95: 961–970. HEDGES, B. R., AND R. G. PALMER. 1992. Inheritance of malate dehydrogenase nulls in soybean. Biochemical Genetics 30: 491–502. JOHANSEN, D. A. 1940. Plant microtechnique. McGraw-Hill, New York, NY. KENNELL, J. C., AND H. T. HORNER. 1985. Megasporogenesis and megagametogenesis in soybean, Glycine max. American Journal of Botany 72: 1553–1564. KOSSLAK, R. M., J. R. DIETER, R. L. RUFF, M. A. CHAMBERLIN, B. A. BOWEN, AND R. G. PALMER. 1996. Partial resistance to root-borne infection by Phytophthora sojae in three allelic necrotic-root mutants in soybean. Journal of Heredity 87: 415–422. LEWIS, B. 1993. Genes IV. Oxford University Press, New York, NY. MCCLINTOCK, B. 1952. Mutable loci in maize. Carnegie Institute of Washington Year Book 51: 212–219. MEINKE, D. 1986. Embryo-lethal mutants and the study of plant embryo development. Oxford Survey of Plant Molecular & Cell Biology 3: 122–165.

791

MOGENSEN, H. L. 1982. Double fertilization in barley and the cytological explanation for haploid embryo formation, embryoless caryopses, and ovule abortion. Carlsberg Research Communications 47: 313–354. MYERS, J. R., AND M. J. BASSETT. 1993. Inheritance, allelism, and morphological characterization of unifoliate mutations in common bean. Journal of Heredity 84: 17–20. NELSON, O. E., JR., AND G. B. CLARY. 1952. Genic control of semisterility in maize. Journal of Heredity 43: 205–210. NEVERS, P., N. S. SHEPHERD, AND H. SAEDLER. 1986. Plant transposable elements. Advances in Botanical Research 12: 103–203. NYRAKABBI, I., AND P. R. BEUSELINCK. 1995. Female sterility of the vestigial corolla floral mutant of birdsfoot trefoil. American Society of Agronomy Annual Meeting, 98 (Abstract). American Society of Agronomy, Madison, WI. PALMER, R. G., B. R. HEDGES, R. S. BENAVENTE, AND R. W. GROOSE. 1989. w4-mutable line in soybean. Developmental Genetics 10: 542–551. , AND H. E. HEER. 1984. Agronomic characteristics and genetics of a chromosome interchange in soybean. Euphytica 33: 651–663. PEREIRA, T. N. S., H. ILARSLAN, AND R. G. PALMER. 1997. Genetic and cytological analyses of three lethal ovule mutants in soybean (Glycine max; Leguminosae). Genome, in press. , AND R. G. PALMER. 1996. Pollen germination studies. Soybean Genetics Newsletter 23: 54–55. REISER, L., AND R. L. FISHER. 1993. The ovule and embryo sac. Plant Cell 5: 1291–1301. ROBERTSON, D. S. 1978. Characterization of a mutator system in maize. Mutation Research 51: 21–28. ROBINSON-BEERS K., R. E. PRUITT, AND C. S. GASSER. 1992. Ovule development in wild-type Arabidopsis and two female-sterile mutants. Plant Cell 4: 1237–1249. SAEDLER, H., AND P. NEVERS. 1985. Transposition in plants: a molecular model. EMBO Journal 3: 585–590. SINGLETON, W. R., AND P. C. MANGELSDORF. 1940. Gametic lethals on the fourth chromosome of maize. Genetics 25: 366–390. SPURR, A. R. 1969. A low-viscosity epoxy resin embedding medium for electron microscopy. Journal of Ultrastructure Research 26: 368–370. STELLY, D. W., S. J. PELOQUIN, R. G. PALMER, AND C. F. CRANE. 1984. Mayer’s hemalum-methyl salicylate: a stain-clearing technique for observations within whole ovules. Stain Technology 59: 155–161. STROMAN, G. N. 1941. A heritable female-sterile type in cotton. Journal of Heredity 84: 97–104. SUITER, K. A., J. F. WENDEL, AND J. S. CASE. 1983. Linkage-1. Journal of Heredity 74: 203–204. VALLANIA, R., R. BOTTA, AND G. ME. 1987. Investigations on anomalies of ovule development and on pollination in mutated grapevines, cv. Barbera. Vitis 26: 1–8. VAN SCHAIK, N. W., AND R. A. BRINK. 1958. Transposition of modulator, a component of the variegated pericarp allele in maize. Genetics 44: 725–738. VOLLBRECHT, E., AND S. HAKE. 1995. Deficiency analysis of female gametogenesis in maize. Developmental Genetics 16: 44–63.