Printer Friendly

Semigamy gene is associated with chlorophyll reduction in cotton.

HAPLOID PRODUCTION has many attractive applications in plant breeding, as it not only enables selection for desirable traits at the haploid level but also provides a means to produce genetically uniform homozygous DH lines from a heterozygous cross in the shortest possible time, thereby shortening the time required for conventional breeding (Choo et al., 1985). Double haploids provide for accuracy in selection because of the lack of heterozygosity, and more rapid genetic gain when additive and additive x additive interactions predominate. Homozygous DH lines can be tested in multiple locations to improve accuracy in phenotypic evaluation of quantitative and qualitative traits. Since haploidy also offers the advantages of sampling gametes in a random fashion, DH lines represent an ideal genetic population for reliable gene and DNA marker mapping (Zhang et al., 2002). Even though in other species haploid plants have been produced via various methods, such as anther or microspore culture (in vitro androgenesis), ovule culture (in vitro gynogenesis), wide (interspecific or intergeneric) hybridization, and semigamy, currently the only feasible means for production of haploids in cotton is via semigamy.

Semigamy is a type of facultative apomixis in which the male sperm nucleus does not fuse with the egg nucleus after penetrating the egg in the embryo sac. Subsequent development can give rise to an embryo containing haploid chimeral tissues of paternal and maternal origins. In cotton, the semigametic phenomena was first reported by Turcotte and Feaster (1963), who developed the Pima line 57-4 that produced haploid seeds at a high frequency. The semigametic line was a DH from a natural haploid mutant found in a field of Pima S-1. In two presentations, Turcotte and Feaster (1974, 1975) indicated that semigamy in Pima appeared to be controlled by one dominant gene (Se). Despite wide acceptance by the cotton research community (Kohel and Lewis, 1984; Smith and Cothren, 1999), no comprehensive genetic data have been published. Semigametic lines can produce 30 to 60% haploids when self pollinated, and about 0.7 to 1.0% androgenic haploids when used as female parents in crosses with normal nonsemigametic cottons (Chaudhari, 1978, 1979; Turcotte and Feaster, 1967). A unique feature of semigamy is that the inheritance of the gene is conveyed by both male and female gametes, but expression of the trait in terms of haploid production occurs only in the female parent. As a consequence, for example, in reciprocal crosses between SeSe and sese parents, haploids will be produced only when SeSe or Sese is the female parent.

DH lines have been developed from cultivars and intra- and interspecific hybrids between Upland cotton (G. hirsutum L.) and American Pima cotton (G. barbadense L.) using semigamy (Feaster and Turcotte, 1973; Barrow and Chaudhari, 1976; Chaudhari, 1979; Stokes and Sappenfield, 1981; Mahill, 1982; Turcotte and Feaster, 1982; Mahill et al., 1983, 1984a, 1984b; Jenkins et al., 1984). The semigametic trait has also been transferred into different cotton cytoplasms (Mahill, 1982; Stewart, 1990), to facilitate rapid replacement of nuclei. Stelly et al. (1988) proposed a scheme called hybrid elimination and haploid production system using a strain with semigamy (Se), lethal gene (LefaV), virescent (117), and male sterility or glandless ([gl.sub.2][gl.sub.3]). This system proved to be successful (Stelly and Rooney, 1989); however, it has not been used by others. Zhang et al. (1998) reported that semigametic lines produced a substantial number of small seed and seed weight was significantly correlated with haploid percentage, indicating that seed weight could be used as a first indicator for haploid selection.

Pima 57-4 produced 32 to 43% haploids in the field across years and produced as high as 61% haploids in the greenhouse in Turcotte and Feaster's test. Researchers in China also found that semigamy was unstable, where Zhang (1993) reported that haploid percentage varied from 0 to 58.3% among progeny-rows of a virescent semigametic line, Vsg-7, and concluded that the semigametic trait was lost in some plants. The results could not be explained based on the one gene model as proposed by Turcotte and Feaster (1974, 1975). The reasons for variation in haploid production in semigamy strains are not presently understood, although environment appears to play a role. Both the genetic basis of semigamy and environmental effects on semigamy expression should be investigated in detail to solve the controversies.

The objectives of this study were to: (i) evaluate genetic and environmental effects on expression of semigamy; (ii) determine if differences beyond semigamy expression exist between 57-4 and its natural isogenic line, S-1; and (iii) detail a genetic study on semigamy and its association with leaf chlorophyll content.


Pedigree Selection in Semigametic Strains

Two semigametic lines, 57-4 and Sev7, were used in the study. Ser7 was originally selected from a single plant progeny row of Vsg-7 that had high ([approximately equal to] 50%) haploid production. Vsg-7 was a G. barbadense semigametic marker strain developed by combining the virescent-7 ([v.sub.7]) gene with semigamy (Turcotte and Feaster, 1969, 1973). Before the present study, Sev7 and 57-4 had been maintained through self-pollination for several years in Fayetteville, AR. The two genotypes were grown in the field and selfed in Fayetteville in 1996. To maintain purity of the two semigametic lines and minimize any genetic variation that might affect semigametic expression, pedigree selection was practiced for several generations from 1996 to 1999. Individual selfed plants were initially randomly selected in 1996 and grown in progeny rows (approximately 50 plants each) in 1997. Individual tetraploid plants in each of the progeny rows were further selfed and harvested to be grown as families in 1998. Then, individual flowers from each of the plants were selfed, harvested, and grown in 1999 for haploid plant identification. In this way, variation in haploid production could be partitioned into the sources due to boils, plants within a family, and families by ANOVA.

During peak flowering stage, each plant was scored for ploidy based on morphological factors such as plant, leaf, and flower size; pollen production; boll-setting; and branch thickness. Turcotte and Feaster (1969) used plant appearance and lack of pollen shed as the basis for haploid identification. In our tests, haploid plants were unambiguously identified since they were profoundly shorter, slimmer, with much smaller leaf size and slimmer branches, and with no pollen shed or boll set. Even though some of the haploid plants did occasionally set one to two very small bolls, the bolls contained only one or two seeds per boll. Haploid percentage was calculated for each progeny line.

Comparison between Pima 57-4 and Pima S-1

The semigametic line 57-4 and its natural isogenic nonsemigametic line S-l, together with 12 additional semigametic alloplasmic lines, were grown in the field at the University of Arkansas Main Experiment Station at Fayetteville in 1996 and 1997. These alloplasmic lines, containing cytoplasms from [A.sub.2] (G. arboreum L.), [B.sub.1] (G. anomalum Wawra), [C.sub.1] (G. sturtia num J.H. Willis), [D.sub.2-2] (G. harknessii Brandegee), [D.sub.3-d] (G. davidsonii Kellogg), [D.sub.8] [G. trilobum (DC.) Skovst.], [E.sub.1] (G. stocksii Mast.), [F.sub.1] (G. longicalyx J.B. Hutch. & B.J.S. Lee), A[D.sub.1] (G. hirsutum), AD3 (G. tomentosum Nutt. ex Seem.), A[D.sub.4] (G. mustelinum Miers ex G. Watt), and AD5 (G. darwinii G. Watt), were developed through a backcrossing scheme using 57-4 as the recurrent parent (Stewart, 1990; Zhang and Stewart, 1999). Pima S-1 was kindly provided by Dr. R. G. Percy in 1996. The experimental design in both years was a randomized complete block with three replications. The plot size was 1 row x 15 m with plant spacing of 1 plant 0.3 [m.sup.-1]. In mid-August, six leaf disks, each with an area of 0.38 [cm.sup.2], were taken from the fourth main-stem leaf from the top of six normal tetraploid plants (1 disk [plant.sup.-1]) and put in a test tube containing 10 mL of 80% acetone at 4[degrees]C for 24 to 72 h. Chlorophyll was extracted and its content was determined spectrophotometrically (Zhao and Oosterhuis, 1998). Photosynthetic rate and related traits were determined on the fourth main-stem true leaf with a LI-COR-6200 (LI-COR Inc., Lincoln, NE) portable photosynthesis system following Zhao and Oosterhuis (1998). Measurements were taken four times during July and August in 1996 and two times during the same period in 1997. On each date, measurements were taken and completed between 1000 and 1200 h during sunny days. At plant maturity, 25 open bolls were hand harvested for estimation of boll size, seed index, and fiber quality. Fiber quality traits were measured using high volume instrument testing at Starlab (Starlab, Inc., Knoxville, TN). The data were subjected to ANOVA, and LSD was used for mean comparisons.

Genetic Analysis of Semigamy and Chlorophyll Content

To minimize any effects of genetic background on the inheritance of semigamy, the isogenic line S-1 was used as male or female to cross with 57-4. The resulting [F.sub.1] was then selfed and also backcrossed with both parents to generate [F.sub.2] and B[C.sub.1][F.sub.1] populations, respectively. The individual [F.sub.2] and B[C.sub.1][F.subb..1] plants were grown in the field and used as female to testeross with the virescent ([v.sub.7]) semigametic line Sev7. The individual [F.sub.2]- and B[C.sub.1][F.sub.1]-derived testcross progenies (30 to 50 seedlings per progeny) were grown in the greenhouse for haploid plant identification based on chimeric leaf tissues. Seedlings were classified as haploids when haploid virescent leaf tissues of paternal origin were observed (Turcotte and Feaster, 1969, 1973), which indicated that the [F.sub.2] or B[C.sub.1][F.sub.1] plants, from which the testcross progenies were derived by crossing as female with Sev7, carried the semigametic gene. Although completely green gynogenic haploids occasionally occur, for scoring purposes green seedlings were considered to be nonhaploid hybrids between the individual [F.sub.2] or B[C.sub.1][F.sub.1] plants and Sev7 carrying the recessive homozygous v7 gene. A chi-square test was used to test goodness-of-fit to expected ratios.

Chlorophyll content of the fourth main-stem leaves from the top of plants from the three segregating populations ([F.sub.2] and two B[C.sub.1][F.sub.1] generations), together with their parents and [F.sub.1] plants, were measured on an individual plant basis using the 80% acetone extraction method as described previously. Generation-mean analysis was employed to separate genic effects and estimate heritabilities and minimum number of genes for chlorophyll content (Mather and Jinks, 1982). To investigate the relationship between semigamy and chlorophyll content, the individual [F.sub.2] plants from the cross between 57-4 x S-1 were selfed to generate [F.sub.2]-derived [F.sub.3] families ([F.sub.2.3]). The individual [F.sub.2.3] families were grown in the field in a randomized complete design for haploid and chlorophyll content determination using the methods described previously. One leaf disc from each of 10 plants per 15 m progeny row (1 plant 0.3 [m.sub.-1]) were sampled in August and bulked for chlorophyll extraction. The homozygous semigametic [F.sub.2.3] lines characterized as having high haploid percentage were compared with the homozygous nonsemigametic [F.sub.2.3] lines that did not produce any haploid plants.


Sources of Variation in Haploid Production in 57-4 and Sev7

Haploid cotton plants produced by semigamy could be readily identified without ambiguity in the field in that haploid cotton seedlings grew slower than normal cotton plants and showed identifiable differences in plant, branch, and leaf size even before flowering. During the flowering and boll-setting stage, because of irregular meiosis without homologous chromosome pairing, haploid plants did not have pollen shed when anthers were squeezed between fingers and thereby did not set normal sized-bolls, if any. According to Turcotte and Feaster (1967), 57-4 produced 30 to 60% haploids across years. Zhang et al. (1998) reported that 57-4 and Sev7 populations produced 46.3 to 49.9 and 36.1 to 43.2% haploid plants in 1996 and 1997, respectively. However, the haploid percentage varied among plant progenies, ranging from 14.3 to 87.5%. We speculated that the low haploid percentage in some of these progenies could be due to impurity of the initial semigametic lines, lower germination of small haploid seeds, developmental factors, and effects of flowering date, boll position, temperature, and sampling error. Also, the small number of seeds sampled could give biased estimates of haploid production. Furthermore, since semigametic lines must be maintained through self-pollination of nonexpressing plants (normal, fertile), selection against semigametic expression is unavoidable.

An ANOVA on haploid percentages from different families selected from 57-4 and Sev7 revealed that the majority of the variation was among bolls within a plant, accounting for about two-thirds of the total variance (Table 1). The among-plant variance within a family contributed 28% of the total variance, but <6% of the total variation was due to variation from families. Therefore, our systematic approach verified that the two semigametic lines were relatively uniform and homozygous in semigamy genotype. The variation seen in haploid production was attributed to developmental and environmental factors. Therefore, our data do not support the previous observation that the semigametic traits could be lost in some of the semigametic plants (Zhang, 1993), unless seed impurity due to outcrossing or seed mixture occurred. Further study on flowering date, boll position, and environmental factors such as temperature, humidity, and sunlight with reference to semigamy could reveal optimum conditions for producing haploids.

Differences between Pima 57-4 and Pima S-1

Since the semigametic DH line 57-4 was a natural haploid mutant found in the normal nonsemigametic S-1 of Pima breeding program where selfing was used as a breeding strategy (Turcotte and Feaster, 1963), S-1 and 57-4 should be considered as isogenic lines. These two lines should not significantly differ in their performance except for semigamy-related traits, such as haploid production, unless natural selection also existed on other traits or natural outcrossing occurred during the origin of 57-4. However, this was unlikely. Surprisingly, when 57-4 and S-1 were grown side-by-side in the field, we noted a difference in leaf color. The visual difference was confirmed through measurement of chlorophyll a and b (chl a and chl b) content (Table 2). Compared with S-l, chl a and chl b contents in 57-4 were lower by 18.3 and 29.3%, respectively. The reduction in chlorophyll in 57-4 also resulted in a trend for lower net Pn than S-l, with an average reduction of 16.4% (Table 3). The differences in Pn were significant on three of the six dates. The differences between the two isogenic lines in the other photosynthesis-related traits (i.e., stomatal conductance and transpiration rate) were not significant in most cases, but the trend for lower values in 57-4 was still clear (data not shown). High environmental variation in determining these traits was encountered in the field conditions that mostly resulted in the nonsignificant differences between 57-4 and S-1. Even though all the measurements from one to two plants per plot per replication were taken in a short period of time, light intensity, temperature, and humidity still varied from plant to plant. Under a more controlled environment, such as greenhouse or growth chamber conditions, physiological differences between the two isogenic lines could be more reliably determined.

Differential growth responses related to genotype across a season can be reflected in agronomic yield and quality traits. Compared with S-l, 57-4 produced significantly smaller bolls (3.55 vs. 4.90 g for S-1 in 1996) and a higher lint percentage (37.42 vs. 34.17% for S-1) because it had a higher number of smaller seed (Zhang et al., 1998). Its significantly larger seed index (14.20 vs. 13.85 g for S-1 in 1996) was probably related to the reduced number of seeds per boll and reduced intraboll competition from the smaller haploid seeds. Interestingly, its fiber was significantly shorter (28.70 to 29.46 mm) and coarser (with micronaire readings of 4.4 to 4.6) than S-1 (33.02 to 34.54 mm and 3.2 to 4.1, respectively). However, 57-4 had significantly higher fiber elongation in 1997 (9.03 vs. 7.73 % for S-l), but similar fiber strength (average 379.61 vs. 362.65 kN m [kg.sup.-1] for S-1 in 1996 and 1997).

As shown above, the most profound and consistent differences between the two isogenic lines were chlorophyll content and fiber quality traits including fiber length and micronaire, which are associated with fiber cell elongation and secondary wall deposition. Semigamy might be involved in cell cycle function, affecting not only fertilization, but also other cell functions in other plant tissues such as leaf functions and single-celled fiber development. Zhang et al. (1998, 1999a, 1999b) used differential display of cDNA (from mRNA transcripts) to identify genes that were differentially expressed in ovules and pollen between 57-4 and S-1. Some differentially expressed sequences had significant homology to genes coding for known proteins related to cell division. However, the lower chlorophyll content, and the shorter and coarser fibers in 57-4 remain unexplained as to whether they were pleiotropic effects of the semigamy gene, or occurred by coincidence. The question was partially addressed through analysis of a segregating population.

Inheritance of Semigamy

The distribution of haploid production in the testcross progeny of the three segregating populations is shown in Fig. 1, 2, and 3. According to Turcotte and Feaster (1974, 1975), semigamy is controlled by one dominant gene (Se), thus, three genotypes, SeSe, Sese, and sese, should exist in the 57-4 x S-1 [F.sub.2] generation in a 1:2:1 ratio. However, the three [F.sub.2] genotypes could not be differentiated directly but required that individual F2 plants be progeny tested by selfing or by testcrossing as female to another semigamy line with the virescent (rT) phenotype, since semigametic lines do not produce haploids when used as male in crosses with normal lines. Green (maternal) or virescent (paternal) haploid plants or chimeric leaf tissues seen in an [F.sub.2]-derived [F.sub.3] ([F.sub.2.3]) or testcross progeny would be an indication that the [F.sub.2] or B[C.sub.1][F.sub.1] plant contained Se. Theoretically, 1/4 of the [F.sub.2] plants should not produce haploids. From 71 [F.sub.2] tetraploid plants, 23 did not produce haploids and 48 produced haploids ranging from 1.5 to 41.5% in their [F.sub.2.3] progeny. Recall that 3.7% of the plants produced in the [F.sub.2] were haploid (Table 4), derived as a result of incomplete fertilization between [F.sub.1] Se female gametes and [F.sub.1] Se or se male gametes. Under circumstances of normal fertilization, these haploid plants in [F.sub.2] would have been SeSe or Sese. Therefore, when using [F.sub.2.3] progeny to determine the ratio of genotypes (SeSe, Sese, and sese) in the [F.sub.2] generation, the number must be adjusted to include the haploids produced in the [F.sub.2] from which [F.sub.2.3] plants were not produced. Otherwise, the frequency of semigametic genotypes would be underestimated. Thus, the 3.7% haploids were added to the [F.sub.2] as semigametic genotypes (SeSe or Sese). The adjusted number of [F.sub.2] plants was 73.7 (50.7 semigametic plants and 23 nonsemigametic plants) since 71 tetraploid [F.sub.2] plants accounted for 96.3% of the [F.sub.2] population. The [chi square] test indicated that the semigamy (Se-) to nonsemigamy (sese) ratio in [F.sub.2] was 3:1 for an one-gene model ([chi square] = 1.51, P > 0,05). The unadjusted number also did not violate the expected ratio ([chi square] = 2.07, P > 0.05).


The results from the [F.sub.2] population were confirmed by testing progeny of the two B[C.sub.1][F.sub.1] populations. In [(57-4 x S-1)[F.sub.1] x S-1]B[C.sub.1][F.sub.1] (Fig. 2), the unadjusted number of plants producing haploid (Sese) and nonhaploid testcross progeny (sese) in the B[C.sub.1][F.sub.1] were 24 and 26, respectively (1:1 ratio, [chi square] = 0.08, P > 0.01). Since 3.8% haploids were produced when the [F.sub.1] was back crossed as female to S-1 (Table 4), they were undoubtedly from the incomplete fertilization between Se female gametes from [F.sub.1] and se male gametes from S-1. These plants would be the Sese genotype under normal fertilization. Therefore, the 3.8% haploids should be added to the B[C.sub.1][F.sub.1] population to give an adjusted number of 25.9 haploid-producing Sese vs. 26 sese (1:1 ratio, [chi square] = 0.0, P > 0.01). Similar to the [F.sub.2] population, the heterozygous individual B[C.sub.1][F.sub.1] plants, when testcrossed with Sev7, produced progeny with a wide range of haploid percentage (4 to 26%), except for one B[C.sub.1][F.sub.1] plant that produced >40% haploid plants. The average haploid percentage from the heterozygous Sese was 14.5% when testcrossed as female to Sev7. For another backcross population, [(57-4 x S-1)[F.sub.1] x 57-4]B[C.sub.1][F.sub.1] with two genotypes (SeSe and Sese in an equal ratio), all its progeny should produce haploids but at differing percentages. However, no haploids were detected in two of the B[C.sub.1][F.sub.1] progeny lines (Fig. 3), perhaps because of a low number of testcross progeny. These two should be the heterozygous genotype that normally gives a low haploid percentage.


As seen above from the [(57-4 x S-1)[F.sub.1] x S-1]BC1[F.sub.1] population, the haploid-producing heterozygous plants produced an average of 14.5% haploids when tested as female with Sev7. Furthermore, when the reciprocal [F.sub.1]'s of 57-4 and S-1 were crossed as female to 57-4, 11.9 to 14.0% (with an average of 13.0%) haploids were produced (Table 4). Thus, a dividing point around the average haploid percentage (13.0 to 14.5% for Sese) could be used to distinguish the two semigametic genotypes (between Sese and SeSe). The haploid percentages from individual [F.sub.2] plants (Fig. 1) with a bimodal distribution suggested that the semigamy genotypes could be divided into two groups. If the dividing point is set at 13.5%, the observed frequencies for SeSe:Sese:sese are 22:29:23 in the [F.sub.2]-derived testcross progeny, which fits a 1:2:1 ratio ([chi square] = 3.48, P > 0.05). Here, we ignored 3.7% haploids produced in [F.sub.2], which were either SeSe or Sese. However, in [(57-4 x S-1)[F.sub.1] x 57-4]B[C.sub.1][F.sub.1] derived testcross progeny, the observed frequencies for Sese: SeSe were 28:25 (1:1, [chi square] = 0.17, P > 0.05). Since 13.0% of the plants in the B[C.sub.1][F.sub.1] population were haploids, derived from the incomplete fertilization between Se female gametes from [F.sub.1] and Se male gametes from 57-4, this percentage should be added to the SeSe genotype in the population. This resulted in 28 Sese and 33 SeSe with a 1:1 ratio ([chi square] = 0.39, P > 0.05). The estimated average haploid percentages for Sese and SeSe produced when testcrossed as female to Sev7 were 9.9 and 30.1%, respectively.

Of course, the dividing point (13.5%) for distinguishing the two semigametic genotypes (between Sese and SeSe) arguably may be subjective since the Sese genotype in [(57-4 x S-1)[F.sub.1] x S-1]B[C.sub.1][F.sub.1], when testcrossed with Sev7, produced haploids with a range extending up to 26% (Fig. 2). Because of varying numbers of plants in the testcross progeny, experimental errors may prevent the existing data from clearly distinguishing between the SeSe and Sese genotypes. Furthermore, testcross hybrids were made during different fruiting periods, on different fruiting branches and boll positions, all of which could confound the genotype differentiation. As seen previously, the number of haploids obtained from individual self-pollinated plants of the two semigametic lines, 57-4 and Sev7, ranged from 14.3 to 87.5%. To generate data in distinguishing SeSe from Sese, sufficient testcross hybrid seeds should be made during the same fruiting period and at similar fruiting positions.

Nevertheless, the three segregating populations clearly showed that the heterozygous genotype (Sese) only produced an average of 10 to 15% haploids, approximately half that (25 to 30%) produced by the homozygous genotype (SeSe), when testcrossed as female with Sev7. Therefore, the semigamy gene (Se) is incompletely dominant. Our analysis also pointed out that haploids produced in [F.sub.2] or B[C.sub.1][F.sub.1] should be counted toward SeSe or Sese genotypes when segregating ratios are calculated, a factor not considered by Turcotte and Feaster (1974, 1975). In their two conference presentations, one was an abstract and another in a proceeding, which were difficult to access and contained limited information on the inheritance of semigamy in cotton. As a result, semigamy has been mistakenly known as a dominant trait (Kohel and Lewis, 1984; Smith and Cothren, 1999). High variation of haploid production in semigametic lines observed by various researchers also led to speculations that more than one gene and environmental effects could be involved in semigamy expression, warranting a necessary revisit to clarify the genetic basis of semigamy in cotton. Our data from the three segregating populations provided compelling evidence that verified that the semigamy trait is indeed controlled by one gene, Se, as Turcotte and Feaster (1974, 1975) originally proposed. However, the detailed genetic analyses revealed that the Se gene allele is incompletely dominant, not dominant as commonly believed. The haploid data from various crosses in the next section will convincingly explain that its incomplete dominance nature is due to its actions at the sporophytical and gametophytical levels.

Genetic Model of Semigamy

In our test, 57-4 produced 44% haploid progeny when selfed, but produced only 11% haploids when crossed as female to S-1 (Table 4). However, no haploids were detected when 57-4 was crossed as male to S-1, indicating that a special microenvironment in the embryo sac provided by the semigametic genotype is essential for haploid production. Also, a similar condition in male gametes with the Se genotype can substantially facilitate semigamy expression. Unexpectedly, the (57-4 x S-1)[F.sub.1] produced only 3.7% haploids in the [F.sub.2]. The haploid production in other crosses is listed in Table 4. There are two possible genetic models that could explain semigamy expression in cotton based on the empirical observation that self-pollination of a homozygous genotype (SeSe) will yield 44% haploid progeny, whereas crosspollinating SeSe as female with sese genotype will yield 11% haploid progeny in [F.sub.2].

Model 1--gametophytic control where haploid production is controlled only by the genotype of the gametes. In this model, a Se female gamete in the [F.sub.1] (57-4 x S-1) has 44% probability to produce haploids in the [F.sub.2] when fertilized by a Se male gamete, as exemplified by selfing 57-4. The chance of producing haploids is reduced to 11% when fused with a se male gamete, as occurs when 57-4 is crossed as female to a nonsemigametic line (e.g., S-1). Of course, se female gametes will not produce any haploids regardless of male gamete genotype (Se or se). Therefore, the expected haploid percentage in [F.sub.2] would be 1/2 x 1/2 x 44% + 1/2 x 1/2 x 11% = 13.8%. The expected haploid-producing percentages for other crosses can be similarly calculated (Table 4).

Model 2--gametophytic and sporophytic control. This model takes quantitative differences into consideration and states that effectiveness in haploid production of Se female gametes from heterozygous (Sese) [F.sub.1] plants when fused with Se or se male gametes from homozygous SeSe or sese plants are only half that of female gametes from homozygous (SeSe) plants (22 vs. 44% with Se male gametes and 5.5 vs. 11% with se male gametes). However, during selfing when Se female gametes are fused with Se male gametes in heterozygous (Sese) [F.sub.1] plants, haploid production is reduced further to 1/4 of 44%. In other words, the heterozygous condition decreases the ability of the Se gametes to aid in haploid production. Even though the se male gametes might receive a certain amount of Se product produced before meiosis from the heterozygous [F.sub.1] sporophytes, the haploid production efficiency for the se male gametes in heterozygous (Sese) [F.sub.1] plants is unchanged, same as the homozygous (sese) condition when fusing with Se female gametes, so the percentage haploid production is dependent upon the genotype of the plant donating the Se female gamete (11% for SeSe female). As shown before, the se female gametes in the heterozygous [F.sub.1] do not induce any haploids. Therefore, in the heterozygous Sese condition, the possibility for a Se female gamete (with a frequency of 1/2) to produce a haploid when fused with a Se male gamete (with a frequency of 1/2) is (1/2 x 1/2) x 1/4 x 44% = 2.75%; and its chance of producing a haploid seed when fused with a se male gamete is (1/2 x 1/2) x 1/2 x 11% = 1.38%. The haploid percentage in the [F.sub.2] would be the sum of these two, which is 4.1%. For [(57-4 x S-1)[F.sub.1] x 57-4] B[C.sub.1][F.sub.1], the haploid percentage is 1/2 x 1/2 x 44% = 11.0%, since only 1/2 of the Se female gametes will induce haploids with 1/2 of the haploid-producing efficiency observed in the homozygous status (SeSe). For its reciprocal B[C.sub.1][F.sub.1], 57-4 x (57-4 x S-1)[F.sub.1], the Se male gametes will produce 1/2 x 1/2 x 44% = 11.0% haploids, and the se male gametes will induce 1/2 x 11% = 5.5% haploids for a total of 16.5% haploids in the progeny. For [(57-4 x S-1)[F.sub.1] x S-1] B[C.sub.1][F.sub.1], the expected percentage of haploids will be 1/2 x 1/2 x 11% = 2.8%. Table 4 lists the expected values based on Model 2 for a comparison. Interestingly, when (57-4 x S-1)[F.sub.1] x 57-4 and (S-1 x 57-4)[F.sub.1] x 57-4 were compared, the former with 57-4 as female in the [F.sub.1] produced a higher percentage of haploids (14.0 vs. 11.9%). The same difference occurred between (57-4 x S-1)[F.sub.1] x S-1 and (S-1 x 57-4)[F.sub.1] x S-1 (4.8 vs. 3.8%, respectively). The significance of this observation and possible implications of maternal effects need further investigation.

From Table 4, it can be seen that, except for the [(57-4 x S-1)[F.sub.1] x S-l] B[C.sub.1][F.sub.1] population, the expected haploid percentages for various crosses based on Model 1 are violated, indicating that semigamy expression in cotton is not adequately explained by the gametophytic control model. However, Model 2 does provide a reasonable prediction of the empirical results from all the crosses tested and is a valid model for explaining the genetic mode of semigamy expression in cotton.

Inheritance of Chlorophyll Content and Its Association with Semigamy

Additive effects were highly significant for chl a, chl b, and total chlorophyll content, while dominant effects were also detected for chl b and total chlorophyll content (Table 5). Also, heritabilities for chlorophyll content ranged from moderate to high, indicating the effectiveness of selection for this trait. Interestingly, the minimum number of genes controlling chlorophyll content (a, b or total) was estimated to be only one, indicating that the difference in chl a, chl b, and total chlorophyll content between 57-4 and S-1 is controlled by one gene. The association between this gene and the semigametic gene (Se) was further examined by comparing chlorophyll contents of homozygous semigametic SeSe [F.sub.3] lines and homozygous nonsemigametic sese [F.sub.3] lines (Fig. 4) and by correlation analysis. On average, chl a, chl b, and total chlorophyll content in the semigametic [F.sub.3] lines were lower by 11.8, 14.4, and 12.6%, respectively, compared with the nonsemigametic [F.sub.3] lines. The differences were significant and similar to the differences between the near isogenic parental lines, semigametic 57-4 and nonsemigametic S-1 (see Table 2). Haploid production was significantly negatively correlated with chl a (correlation coefficient r = -0.51, P < 0.05), chl b (r = -0.48, P < 0.05), and total chlorophyll content (r = -0.51, P < 0.05). Therefore, the genetic data clearly revealed that the semigamy gene and the gene responsible for reduced chlorophyll content are associated. Whether they are the same gene or tightly linked remains for further investigation.


Unfortunately, when this genetic study was conducted, the focus was on chlorophyll content and we were not aware that the differences in the fiber quality traits were sufficiently high to warrant a genetic investigation. If the association between the differences in fiber quality and semigamy is proven to be true, the Se gene will be the first gene in cotton known to reduce fiber quality from that typical of Pima cotton to values typical of Upland cotton. Knowledge of the metabolic activities associated with the semigamy gene could provide insight into the biological factors involved in the determination of fiber quality. With reference to the semigamy trait, fiber length possibly could be used as a criterion to differentiate the SeSe genotype from the sese genotype in an [F.sub.2] population of 57-4 x S-1 since fiber length values of 57-4 and S-1 did not overlap. Otherwise, either cytological evaluation on fertilization and embryogenesis or progeny test will be required to distinguish the genotypes.


The original semigametic line Pima 57-4 is a DH line from a haploid plant found in the commercial cultivar Pima S-1. The two genotypes were considered to be isogenic for these comparative studies. Although the selfed 57-4 and its derived semigametic line (Sev7) with virescent gene (v7) were consistent in producing haploids in 1996 and 1997, ranging from 36.1 to 49.9%, the variation seen in haploid percentage among individual plant progenies was largely because of developmental and environmental effects during pollination and boll development.

Compared with the original cultivar S-1, the semigametic mutant 57-4 had lower chlorophyll content, lower Pn, and lower boll weight, while it had higher lint percentage, higher seed index, and shorter, stronger, and coarser fiber. Genetic data demonstrated that semigamy expression is sporophytically and gametophytically controlled by one incompletely dominant gene, Se. The difference in chlorophyll content between 57-4 and S-1 is also conditioned by one gene. Their semigametic [F.sub.3] sibling lines had significantly lower chlorophyll content than the nonsemigametic [F.sub.3] sibling lines. The correlation between haploid production and chlorophyll content was significant, indicating that semigamy and chlorophyll reduction are associated. The possibility that the semigamy gene conditions reduced chlorophyll is strongly indicated, but such a pleiotropic effect must be confirmed.

Abbreviations: chl, chlorophyll; DH, doubled haploid; Pn, photosynthetic rate.
Table 1. Sources of variation in haploid production of
Pima 57-4 and Sev7 based on ANOVA.

                            57-4               Sev-7
Source of                         %
variance              Var      ([double   Var      %
                   ([dagger])  dagger])

Among families        37.6       4.0     28.9     7.1
Among plants
  within families    261.2       27.7    115.3   28.3
Among bolls
  within plants      644.7       68.3    262.8   64.6
Total variance       943.5      100.0    407.0  100.0

([dagger]) Variance of haploid percentage.

([double dagger]) Percentage of variance
for haploid percentage.

Table 2. Comparison between Pima 57-4 and
Pima S-1 for chlorophyll content in the
fourth main-stem leaf from the topmost leaf.

                          Chlorophyll content

                             Pima     57-4/
                    57-4      S-1     Pima     LSD
                                       S-1    (0.05)

                              Mg [cm.sub.-2]

Chlorophyll a        8.69   10.63 *    81.8    0.81
Chlorophyll b        1.59    2.26 *    70.8    0.20
Total chlorophyll   10.28   12.89 *    79.8     0.99 *

* Significant at the 0.05 probability level.

Table 3. Comparison between Pima 57-4 and Pima
S-1 for photo-synthetic rate measured on the
fourth main-stem leaf from the topmost leaf.

Measuring               Pima          LSD
date year      57-4      S-1        (0.05)

                 [micro]mol C[O.sub.2]
                 [m.sup.-2] [s.sup.-1]

25 July 1996   17.79   26.20 *       6.19
2 Aug. 1996    15.80   19.73          NS ([dagger])
6 Aug. 1996    14.51   15.81          NS
19 Aug. 1996   20.44   17.83          NS
30 July 1997   25.30   32.41 *       1.91
4 Aug. 1997    26.33   31.03 *       1.41

* Significant at the 0.05 probability level.

([dagger]) NS, no significance at the
0.05 probability level.

Table 4. Haploid plant percentage produced in different crosses.

Cross                          No.          No.
                              total     tetraploids

57-4 selfed                    419          235
(57-4 [female] ([dagger])
 x PS-1) [F.sub.1]             45           40
(PS-1 [female] x 57-4)
 [F.sub.1]                     69           69
(57-4 x PS-1) [F.sub.2]        107          103
(57-4 x PS-1) [F.sub.1]
 [female] x 57-4               86           74
(PS-1 x 57-4) [F.sub.1]
 [female] x 57-4               67           59
(57-4 x PS-1) [F.sub.1]
 [female] x PS-1               81           77
(PS-1 x 57-4) [F.sub.1]
 [female] x PS-1               83           81
57-4 [female] x
 (57-4 x PS-1)[F.sub.1]        99           84

Cross                          No.       Observed
                            haploids     haploids

57-4 selfed                    184         43.9
(57-4 [female] ([dagger])
 x PS-1) [F.sub.1]              5          11.1
(PS-1 [female] x 57-4)
 [F.sub.1]                      0           0.0
(57-4 x PS-1) [F.sub.2]         4           3.7
(57-4 x PS-1) [F.sub.1]
 [female] x 57-4               12          14.0
(PS-1 x 57-4) [F.sub.1]
 [female] x 57-4                8          11.9
(57-4 x PS-1) [F.sub.1]
 [female] x PS-1                4           4.8
(PS-1 x 57-4) [F.sub.1]
 [female] x PS-1                3           3.8
57-4 [female] x
 (57-4 x PS-1)[F.sub.1]        15          15.2

                            Expected     Expected
Cross                       haploids:    haploids:
                             Model 1      Model 2

57-4 selfed                   44.0         44.0
(57-4 [female] ([dagger])
 x PS-1) [F.sub.1]            11.0         11.0
(PS-1 [female] x 57-4)
 [F.sub.1]                     0.0          0.0
(57-4 x PS-1) [F.sub.2]       13.8          4.1
(57-4 x PS-1) [F.sub.1]
 [female] x 57-4              22.0         11.0
(PS-1 x 57-4) [F.sub.1]
 [female] x 57-4              22.0         11.0
(57-4 x PS-1) [F.sub.1]
 [female] x PS-1               5.5          2.8
(PS-1 x 57-4) [F.sub.1]
 [female] x PS-1               5.5          2.8
57-4 [female] x
 (57-4 x PS-1)[F.sub.1]       27.5         16.5

([dagger]) Used as female in crossing.

Table 5. Estimators of genetic parameters based on generation mean
analysis on chlorophyll a and b (chl a and chl b) content of the
fourth leaf from the topmost leaf.

Genetic parameter           (a + b)    chi a     chi b

Additive effect             0.22 **   0.76 **   0.22 **
Dominant effect             0.98 **   0.76      0.98 **
Broad-sense heritability    0.70      0.64      0.81
Narrow-sense heritability   0.39      0.31      0.79
Minimum no. genes           0.50      0.54      0.44
Minimum no. genes           0.91      1.13      0.45 **

** Significant at the 0.01 probability level.


Barrow, J.R., and H.K. Chaudhari. 1976. A homozygous interspecific F2 hybrid of Gossypium barbadense x G. hirsutum via the semigametic haploid method. Crop Sci. 16:441-442.

Chaudhari, H.K. 1978. Use of semigamy in the production of cotton haploids. Bull. Torrey Bot. Club 105:98-103.

Chaudhari, H.K. 1979. The production and performance of doubled haploids of cotton. Bull. Torrey Bot. Club 106:123-130.

Choo, T.M., E. Reinbergs, and K.J. Kasha. 1985. Use of haploid in breeding barley. Plant Breed. Rev. 3:219-252.

Feaster, C.V., and E.L. Turcotte. 1973. Yield stability in doubled haploids of American Pima cotton. Crop Sci. 13:232-233.

Jenkins, J.N., J.C. McCarty, Jr., J.F. Mahill, and J. Fallieri. 1984. Registration of fifteen doubled haploid lines of cotton, Gossypium spp. germplasm. Crop Sci. 24:624-625.

Kohel, R.J., and C.F. Lewis. (ed.) 1984. Cotton. Agron. Monogr. 24. ASA, CSSA, and SSSA, Madison, WI.

Mahill, J.F. 1982. Semigamy in cotton breeding. Diss. Abs. Int. 43:586.

Mahill, J.F., J.N. Jenkins, and J.C. McCarty, Jr. 1983. Registration of eight germplasm lines of cotton (Reg. Nos. GP 210 to GP 217). Crop Sci. 23:403-404.

Mahill, J.F., J.N. Jenkins, J.C. McCarty, Jr., and W.L. Parrott. 1984a. Performance and stability of doubled haploid lines of upland cotton derived via semigamy. Crop Sci. 24:271-277.

Mahill, J.F., J.N. Jenkins, W.L. Parrott, and J.C. McCarty, Jr. 1984b. Registration of four doubled haploid cotton germplasms. Crop Sci. 24:625.

Mather, K., and J.L. Jinks. 1982. Biometrical genetics. Chapman and Hall, London.

Smith, C.W., and J.T. Cothren. 1999. Cotton. John Wiley and Sons, New York.

Stelly, D.M., and W.L. Rooney. 1989. Delimitation of the [Le.sup.dav.sub.2] complementary lethality system of Gossypium to intracellular interaction. J. Hered. 80:100-103.

Stelly, D.M., J.A. Lee, and W.L. Rooney. 1988. Proposed schemes for mass-extraction of doubled haploids of cotton. Crop Sci. 28:885-890.

Stewart, J.McD. 1990. New cytoplasms for cotton, p. 55-58. Proc. 1990 Cotton Res. Meeting. Spec. Rep. No. 144. Arkansas Agric. Exp. Stn., Fayetteville, AR.

Stokes, L.G., and W.P. Sappenfield. 1981. Registration of BW76-31 cotton germplasm (Reg. No. GP 162). Crop Sci. 21:991.

Turcotte, E.L, and C.V. Feaster. 1963. Haploids: High-frequency production from single-embryo seeds in a line of Pima cotton. Science (Washington, DC) 140:1407-1408.

Turcotte, E.L., and C.V. Feaster. 1967. Semigamy in Pima cotton. J. Hered. 58:54-57.

Turcotte, E.L., and C.V. Feaster. 1969. Semigametic production of haploids in Pima cotton. Crop Sci. 9:653-655.

Turcotte, E.L., and C.V. Feaster. 1973. The origin of 2n and n sectors of chimeral Pima cotton plants. Crop Sci. 13:111-112.

Turcotte, E.L., and C.V. Feaster. 1974. Methods of producing haploids: Semigametic production of cotton haploids, p. 53-64. In K.J. Kasha (ed.) Haploids in higher plants: Advances and potential. Ontario Agric. College, Univ. of Guelph, ON, Canada.

Turcotte, E.L., and C.V. Feaster. 1975. Inheritance of semigamy in American Pima cotton (Gossypium barbadense L.). p. 65. In Agronomy abstracts. ASA, Madison, WI.

Turcotte, E.L., and C.V. Feaster. 1982. Doubled haploids of American Pima cotton. Agric. Reviews and Manuals. Western Series no. 32. USDA-ARS, Maricopa, AZ.

Zhang, H.Y. 1993. Studies on the semigamy of cotton and its application in breeding. Ph.D. diss. Nanking Agric. Univ. Library, Nanking, China.

Zhang, J., W.Z. Guo, and T.Z. Zhang. 2002. Molecular linkage map of allotetraploid cotton (Gossypium hirsutum L. x Gossypium barbadense L.) with a haploid population. Theor. Appl. Genet. 105:1166-1174.

Zhang, J.F., A. Nopomuceno, J.McD. Stewart, and R.B. Turley. 1998. Gene expression related to semigamy genotype in cotton (Gossypium barbadense), p. 1457-1462. In Proc. Beltwide Cotton Conf., San Diego, CA. 5-9 Jan. 1998. National Cotton Council of America, Memphis, TN.

Zhang, J.F., and J.McD. Stewart. 1999. Cytoplasmic effects on photosynthesis and chlorophyll content, p. 585. In Proc. Beltwide Cotton Conf., Orlando, FL. 3-7 Jan. 1999. National Cotton Council of America, Memphis, TN.

Zhang, J.F., J.McD. Stewart, and R.B. Turley. 1999a. Analysis of semigamy expression in cotton (Gossypium barbadense), p. 446-447. In Proc. Beltwide Cotton Conf., Orlando, FL. 3-7 Jan. 1999. National Cotton Council of America, Memphis, TN.

Zhang, J.F., J.McD. Stewart, and R.B. Turley. 1999b. Genetic and molecular characterization of semigamy expression in cotton (Gossypium barbadense), p. 275-277. In D.M. Oosterhuis (ed.) Proc. of the 1999 Cotton Res. Meeting and Summaries of Cotton Research in Progress. Spec. Rep. 193. Arkansas Agric. Exp. Stn., Fayetteville AR.

Zhao, D., and D.M. Oosterhnis. 1998. Cotton responses to shade at different growth stages: Nonstructural carbohydrate composition. Crop Sci. 38:1196-1203.

J. F. Zhang * and J. McD. Stewart

J.F. Zhang, Dep. of Agronomy and Horticulture, New Mexico State Univ., Las Cruces, NM 88003; J.McD. Stewart, Dep. of Agronomy, Univ. of Arkansas, Fayetteville, AR 72701. Received 27 Feb. 2004. J.F. Zhang, Corresponding author (
COPYRIGHT 2004 Crop Science Society of America
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2004 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Crop Breeding, Genetics & Cytology
Author:Zhang, J.F.; Stewart, J. McD.
Publication:Crop Science
Date:Nov 1, 2004
Previous Article:Screening for salinity tolerance in alfalfa: a repeatable method.
Next Article:Origin of the blue-aleurone gene in Sebesta Blue wheat genetic stocks and a protocol for its use in apomixis screening.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters