Characterization of Mutations That Feminize Gametophytes of the Fern Ceratopteris

Corresponding author: Jo Ann Banks, Department of Botany and Plant Pathology, Whistler Bldg., Purdue University, West Lafayette, IN 47907., banks{at}btny.purdue.edu (E-mail)

Communicating editor: T. W. CLINE

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Gametophytes of the fern Ceratopteris are either male or hermaphroditic. Their sex is epigenetically determined by the pheromone antheridiogen, which is secreted by the hermaphrodite and induces male and represses female development in other young, sexually undetermined gametophytes. To understand how antheridiogen represses the development of female traits at the genetic level, 16 new mutations that feminize the gametophyte in the presence of antheridiogen were identified and characterized. Seven are very tightly linked to the FEM1 locus previously described. Nine others define another locus named NOTCHLESS1 (NOT1), as several of the not1 mutants lack a meristem notch. Some not1 mutations also affect sporophyte development only when homozygous, indicating that the not1 mutations are recessive and that NOT1 is also required for normal sporophyte development. The epistatic interactions among FEM1, NOT1, and other sex-determining genes are described. This information was used to expand the genetic model of the sex-determining pathway in Ceratopteris. On the basis of this model, we can say that the presence of antheridiogen leads to the activation of the FEM1 gene, which not only promotes the differentiation of male traits, but also represses female development by activating the NOT1 gene. NOT1 represses the TRA genes necessary for the development of female traits in the gametophyte.

THE haploid gametophyte generation of the plant life cycle begins with the production of haploid spores through meiotic divisions of diploid spore mother cells. Like many nonseed plants, the fern Ceratopteris richardii produces only one type of spore (i.e., is homosporous), yet each Ceratopteris spore has the potential to develop as a free-living haploid gametophyte that produces only sperm or produces both egg and sperm. The sex of the gametophyte is epigenetically determined in many species of homosporous ferns, including Ceratopteris, by the pheromone antheridiogen (DOPP 1950 ; NAF 1959 , NAF 1979 ; NAF et al. 1975 ; WARNE and HICKOK 1991 ; BANKS 1997A ; YAMANE 1998 ). When the antheridiogen of Ceratopteris (or A_CE) is absent, a Ceratopteris spore develops as a hermaphroditic gametophyte, which produces and secretes the pheromone once it becomes insensitive to its male-inducing effects (BANKS et al. 1993 ). In the presence of A_CE, a spore develops as a male gametophyte. This mechanism of sex determination allows the ratio of males and hermaphrodites in a population to vary depending on population density rather than on genetic predisposition, such that the proportion of males increases as the population size increases.

The male and hermaphroditic gametophytes of Ceratopteris are morphologically distinct, which aids in genetic screens for sex-determining mutants. Hermaphroditic gametophytes have a distinct lateral multicellular meristem from which most cells of the hermaphrodite prothallus are derived (Fig 1A). This meristem forms a meristem notch that gives the hermaphrodite its heart-shaped appearance. A cell within the meristem has four possible cell fates. It can remain within the meristem as a stem cell to maintain the meristem, differentiate as an egg-forming archegonia or as a sperm-forming antheridia, or enlarge to form a photosynthetic, vegetative cell of the prothallus. Male gametophytes never develop a multicellular meristem or archegonia. Growth of the male gametophyte occurs by divisions of a single apical initial. Derivatives of the initial quickly differentiate as antheridia such that at maturity, almost all cells of the male gametophytic prothallus have differentiated as antheridia (Fig 1B). Antheridiogen therefore acts to repress meristem and archegonia formation as well as to promote the rapid differentiation of antheridia during gametophyte development.

View larger version (147K): In this window In a new window Download PPT slide   Figure 1. Wild-type and mutant phenotypes. (A) An 11-day-old wild-type hermaphrodite that is similar in appearance to her hermaphrodites. (B) An 11-day-old wild-type male that is similar in appearance to the tra2 male. (C) The meristematic region of an 11-day-old man1 gametophyte, which produces many more antheridia than wild-type or her hermaphrodites. (D) A 10-day-old fem1-1 female, which produces a meristem and archegonia but no antheridia. (E) A 16-day-old not1-7 (allele 15) gametophyte showing the centralized meristem and notchless phenotype. (F) An 11-day-old male gametophyte mutant for not1-4 (allele 8) and tra2, which is smaller and produces fewer antheridia than wild-type or tra2 males grown in the presence of ACE. (G) An 8-week-old sporophyte homozygous for not1-7 (allele 15) produced by self-fertilizing a not1-7 female; this sporophyte has a normal root but the shoot develops as a ball from which no leaves form. (H) A 4-week-old fem1-1 tra2 intersexual gametophyte. Although ruffled in appearance due to multiple but shallow meristems, the prothallus consists of a single layer of cells. (I) A 4-week-old asexual gametophyte mutant for fem1-1, not1-4 (allele 8), and tra2. Arrows point to the meristematic regions of the gametophyte that are growing in three dimensions. mn, meristem notch; ar, egg-forming archegonia; an, sperm forming antheridia. Bars: A–F and H–I, 0.1 mm; G, 2 mm.

To understand how A_CE regulates the sex of the fern gametophyte, a genetics approach has been used to identify genes that are involved in the sex-determining process in Ceratopteris (reviewed in BANKS 1997A ). Four phenotypic classes of mutations that alter the sex of the gametophyte have been characterized and include the hermaphroditic (her) mutants, which are hermaphrodites in the presence or absence of A_CE; the transformer (tra) mutants, which are male in the presence or absence of A_CE; the feminization1 (fem1) mutant, which is female in the presence or absence of A_CE; and the many-antheridia1 (man1) mutant, which is male in the presence of A_CE and a hermaphrodite that produces 10 times more antheridia than wild-type hermaphrodites when A_CE is absent. A model of the sex-determining pathway in Ceratopteris, based upon the epistatic interactions observed among these mutants (BANKS 1997B ), has been developed. In the present study, we have identified additional sex-determining mutants that specifically address the function of the FEM1 gene. On the basis of the phenotype of the previously described fem1 mutant (always female), FEM1 has two conceivable wild-type functions, both of which are consistent with the model of the sex-determining pathway. One function would be to promote the differentiation of male traits (antheridia) and the other would be to repress the differentiation of female traits (meristem and archegonia). If the FEM1 gene fulfills the latter function by activating a repressor of the TRA genes that are required for the development of female traits, we predict that a mutation that is defective only in its ability to repress the TRA genes could exist. If the current model of the sex-determining pathway is correct, such a mutant would have a fem phenotype but would be genetically unlinked to the FEM1 locus and have different but predictable epistatic interactions with other sex-determining genes. By isolating and characterizing 16 new feminizing mutant gametophytes, we have identified additional mutations that are tightly linked to fem1 as well as mutations that define at least one additional fem1-like locus called NOTCHLESS1 (NOT1), so named because many not1 gametophytes do not form a characteristic meristem notch. Tests of epistasis between the not1 mutants and other sex-determining mutants confirm that the function of the NOT1 gene is to repress the TRA genes. The function of the FEM1 gene in the sex-determining pathway is therefore not only to promote the differentiation of antheridia, but also to repress meristem and archegonia development indirectly by activating the NOT1 gene.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The origins of Hnn, the wild-type strain of Ceratopteris, and the tra, her, fem1, and man1 mutants used in this study are described in BANKS 1997A , BANKS 1997B . The conditions for gametophyte and sporophyte culture were the same as those described by BANKS 1994 . To isolate new fem mutants, 2 x 10⁶ Hnn spores were mutagenized with EMS according to BANKS 1994 and plated on medium containing A_CE, conditions that promote the development of male gametophytes in wild-type spores. Mutagenized gametophytes that developed a multicellular meristem but no or very few (fewer than five) antheridia (putative fem mutants) were transferred to fresh A_CE-containing medium and grown for 20 days. Those that formed no or few antheridia were backcrossed by Hnn twice and then crossed by other sex-determining mutants to test linkage and epistasis. In crosses using a wild-type hermaphrodite as the female parent or a her hermaphrodite as a male parent, the hermaphrodite carried the clumped chloroplast2 (cp2) mutation. Since the recessive cp2 mutation is visible in cells of the gametophyte and sporophyte, it is a useful marker for confirming the hybridity of a cross.

The phenotypes of progeny gametophytes were scored after 2 weeks of growth on A_CE-containing medium prepared by conditioning the medium with the growth of wild-type gametophytes. The phenotypes of gametophytes grown for the same period of time in the absence of A_CE were tested by pipetting individual spores suspended in nonconditioned fern medium into individual microtiter wells. When scoring gametophytes grown in the presence of A_CE, hermaphrodites were ignored if they occurred at a frequency of <1%. These hermaphrodites are likely to be wild-type gametophytes that developed as hermaphrodites even though A_CE was present, as the proportion of wild-type spores that develop as males in the presence of A_CE is always <100% but >95%.

The number of antheridia present on mutant and wild-type hermaphroditic gametophytes was determined by counting the number of antheridia on each of 40 gametophytes after 20 days of growth. The ability of a fem gametophyte to self-fertilize was determined by adding water to 40 individual gametophytes that had been placed individually into microtiter wells before reaching sexual maturity. Gametophytes were kept submerged until embryos could be observed or, if no embryos formed, until 25 days of growth.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Identification of new feminizing mutants:
By virtue of their large size and presence of a meristem compared to the smaller, ameristic male members of the population, putative fem mutants are easy to identify in a population of mutagenized gametophytes grown in the presence of A_CE. After selecting and backcrossing such gametophytes, each putative fem mutant was crossed by wild-type sperm to test the heritability of the fem phenotype. As shown in Table 1, diploid sporophytes heterozygous for each of 16 new mutations produced haploid progeny that segregated male and female or hermaphroditic gametophytes with fewer than nine antheridia in a 1:1 ratio, confirming that the partial or complete feminizing phenotypes segregate as single Mendelian traits. For convenience, these feminizing mutants were temporarily referred to as mutants 2–17.

Eleven of the mutants (2, 4, 5, 7, 8, 10, 11, 12, 13, 15, and 17) did not form visible antheridia nor did they self-fertilize, while five mutants (3, 6, 9, 14, and 16) formed between zero and eight antheridia (Table 1). Other aberrant gametophyte phenotypes were observed among the group of mutants that could form antheridia. Three mutants (2, 9, and 15) lost their meristem notch after 2 weeks of growth and eventually formed tubes through a centralized rather than marginal meristem. The notchless phenotype is illustrated in Fig 1E. Mutant 16 formed archegonia along the marginal cell layer of the meristem notch rather than a few cell layers below the margin of the notch (data not shown). Collectively, these phenotypes indicate that the wild-type functions of some of these genes are not only to promote antheridia development, but also possibly to organize patterns of cell division and differentiation within the meristem.

Among the five mutants that formed antheridia, four (3, 6, 14, and 16) could be self-fertilized to form sporophytes. Of these, only the sporophytes homozygous for mutation 3 developed normally. Sporophytes homozygous for mutations 6, 14, or 16 were abnormally small and either failed to form leaves (Fig 1G, for example) or formed up to 10 small leaves before dying. None of these sporophyte plants produced fertile sporophylls bearing sporangia. The abnormal sporophyte phenotypes associated with these gametophyte mutations indicate that these genes are also necessary for normal sporophyte development. Since sporophytes heterozygous for mutations 6, 14, or 16 are indistinguishable from wild-type sporophytes (data not shown), these mutations are likely to be recessive, complete, or partial loss-of-function mutations.

Seven mutations are linked to FEM1:
Although tests of allelism cannot be determined in the haploid gametophyte, it is possible to establish whether the new feminizing mutations are linked to the FEM1 locus previously characterized by crossing each mutant by the sperm of a fem1 tra2 intersexual gametophyte (illustrated in Fig 1H). As shown in Table 2, sporophytes heterozygous for tra2, fem1, and seven feminizing mutations (4, 5, 7, 10, 11, 12, and 13) segregated females and intersexes in a 1:1 ratio. The absence of males in >2000 progeny per cross (data not shown) indicates that no recombination between fem1 and this group of seven mutant alleles occurred. Although the absence of males does not prove that these new mutants are allelic to fem1, the data do indicate that they are very tightly linked and may be allelic. For this reason, these seven feminizing mutations are collectively referred to as the fem1 group of mutations. Sporophytes homozygous for fem1, generated by crossing a fem1 female by the sperm of a fem1 tra2 intersex, are indistinguishable from wild-type sporophytes, as are sporophytes mutant for both fem1 and each of eight new mutations in the fem1 group (data not shown).

To confirm that gametophytes mutant for both tra2 and each of the fem1 group mutations are intersex, each of the eight fem1 group mutant females was crossed by the sperm of a tra2 male. As shown in Table 2, the resulting heterozygous sporophytes produced progeny that segregated female, male, and intersexual progeny gametophytes in a 1:2:1 ratio as expected. This indicates that all gametophytes mutant for tra2 and each of the fem1 group mutations are phenotypically the same (intersex) and that all of the fem1 group mutations segregate independently of tra2. Similar results were obtained using other tra mutations (data not shown). The intersex, double-mutant gametophyte phenotype is novel, indicating a lack of epistasis between the FEM1 group genes and the TRA genes.

Epistasis and linkage among the remaining feminizing mutants:
Eight of the nine remaining mutants isolated (2, 3, 6, 8, 9, 15, 16, and 17) could be genetically distinguished from the fem1 mutants on the basis of the phenotypes of gametophytes also mutant for TRA2. Mutant females, when crossed by the sperm of tra2 males, formed heterozygous sporophytes whose progeny segregated females and males in a 1:3 ratio when grown in the presence of A_CE (Table 3). Among the male class of progeny, two phenotypically distinct types were observed: one indistinguishable from wild-type and tra2 males and the other having fewer antheridia than wild-type males (compare in Fig 1B and Fig F). The two types of males segregated in 1:2 ratios of small to wild-type males (data not shown), indicating that the odd male phenotype occurs in gametophytes mutant for these feminizing mutations and tra2. In the absence of A_CE, sibling gametophytes derived from the same crosses segregated females, hermaphrodites, and males in a 1:1:2 ratio. These ratios indicate that gametophytes mutant for tra2 and each of the 2, 3, 6, 7, 8, 15, 16, and 17 mutations are male in the absence or presence of A_CE. The TRA2 gene is therefore epistatic to the locus or loci defined by these feminizing mutations.

To examine the linkage between the 2, 3, 6, 7, 8, 15, 16, and 17 mutations and fem1, mutant females were crossed by the sperm of fem1 tra2 intersexual gametophytes. With one exception (mutant 3), the resulting heterozygous sporophytes produced progeny that segregated female, male, intersex, and asexual gametophytes in a 3:3:1:1 ratio when grown in the presence of A_CE (Table 3). This ratio indicates that fem1 is not linked to any of these mutations. Crosses using the females harboring mutation 3 as the female parent and fem1 tra2 intersexes as the male parent also resulted in the same four phenotypic classes of progeny (Table 3). However, they did not segregate in a 3:3:1:1 ratio for reasons that are unknown at this time.

The asexual gametophytes resulting from all crosses produced no antheridia or archegonia and did not form a marginal meristem typical of the female or hermaphroditic gametophytes, which is a novel phenotype. In the hermaphrodite and fem1 females, cell divisions occur in two planes (anticlinal and periclinal) such that the gametophyte prothallus consists of a single layer of cells in these individuals (see Fig 1A and Fig D). While the asexual gametophyte forms a distinct meristematic region, divisions within the meristem often occur in three planes and parts of the gametophyte prothallus are consequently three-dimensional, as illustrated in Fig 1I.

Eight possible progeny genotypes result from each of the crosses described above (i.e., mutant 2, 3, 6, 7, 8, 15, 16, or 17 females x fem1 tra2 intersex). The phenotypes associated with only two genotypes are unknown, including double mutants harboring fem1 and each of the feminizing mutations and triple mutants harboring fem1, tra2, and each feminizing mutation. On the basis of the uniform segregation ratios of phenotypes resulting from these crosses and the known phenotypes of the remaining six genotypes, we deduced that the double-mutant gametophytes are likely to be female and the triple-mutant gametophytes are likely to be asexual when grown in the presence of A_CE. To confirm the genotypes of the phenotypic classes resulting from the mutant 2, 3, 6, 7, 8, 15, 16, and 17 x fem1 tra2 intersex crosses, backcrosses or testcrosses were performed on the progeny gametophytes derived from the mutant 8 x fem1 tra2 cross (Table 4). All progeny intersexes, when crossed to fem1 females, formed sporophytes that segregated females and intersexes in a 1:1 ratio, indicating that the intersexes are genotypically fem1 tra2. Females crossed by wild-type sperm formed sporophytes that segregated females and males in a 3:1 or a 1:1 ratio. The 3:1 ratio indicates that the female parent harbored both fem1 and mutation 8. The 1:1 ratio indicates that the female parent carried either fem1 or mutation 8. Male progeny, when crossed to mutant 8 females, formed sporophytes whose progeny segregated females and males in either a 1:1 or a 1:3 ratio. One male crossed to a fem1 female resulted in a segregation ratio of three females:three males:one intersex:one asexual, which is the same as that observed from the original mutant 8 x fem1-1 tra2 cross. The results of the crosses using the males as sperm donors indicate that the male parent was wild type, a tra2 mutant, or doubly mutant for tra2 and mutation 8. The one of eight genotypes not accounted for (the mutation 8 fem1-1 tra2 triple mutant) is likely to result in an asexual gametophyte phenotype. Since the asexual gametophyte does not form antheridia or archegonia, crosses to confirm its genotype could not be performed.

To establish the linkage relationships among mutations 2, 3, 6, 7, 8, 15, 16, and 17, >100 females of each mutant were crossed by the sperm of small males derived from several of the mutant 2, 3, 6, 7, 8, 15, 16, and 17 x tra2 crosses, i.e., males carrying both tra2 and a feminizing mutation. Either no embryos or abnormal embryos giving rise to abnormal and sterile plants developed from these mutant females (data not shown). Since these plants produced no progeny, we could not establish the linkage relationships among these mutations.

Each mutant 2, 3, 6, 7, 8, 15, 16, and 17 female, when crossed by wild-type or tra2 males, formed sporophytes that were indisinguishable from wild-type sporophytes (data not shown). However, only abnormal sporophytes developed when sibling females were crossed by other males harboring one of the feminizing mutants of this group. Because no combination of any two mutations of this group could restore a wild-type sporophytic phenotype to the sporophyte, each of these feminizing mutations is likely to be recessive in the sporophyte. The inability of these mutations to complement one another in trans in the diploid sporophyte provides strong evidence that mutants 2, 3, 6, 7, 8, 15, 16, and 17 are allelic. Since they are allelic and behave similarly in their epistatic interactions with fem1, tra2, and other sex-determining mutations, as will be shown, these mutations are collectively referred to as notchless1 (or not1-1 to not1-9) alleles, so named because some of them (not1-2, -7, and -9) lack a meristem notch at maturity.

The not1-6 (mutant 14) mutant is unique. Sporophytes heterozygous for not1-6 and tra2 produced progeny that segregated females, males, and intersexes in a 1:2:1 ratio in the presence of A_CE, indicating that the not1-6 tra2 double mutant is intersex (Table 3). It resembles the fem1 group mutants in this regard. When crossed by the sperm of a fem1 tra2 male, however, its progeny gametophytes segregated females, males, intersexes, and asexuals, but not in a 3:2:2:1 ratio, as would be expected if the not1-6 tra2 double-mutant gametophyte were intersex (Table 3). Although the reasons for the observed segregation distortion are unknown, the presence of males and asexual gametophytes in this population is typical of the not1 mutants. We group this mutant with the not1 alleles until evidence showing that it represents another fem-like locus is obtained.

Epistasis among the FEM1 group, NOT1, and HER genes:
All mutant females were crossed by the sperm of her19 hermaphrodites to determine the phenotypes of the double mutants. As shown in Table 5, sporophytes heterozygous for fem1 or each of the not1 alleles and her19 produced progeny that segregated meristic gametophytes (which included hermaphrodites and females) and males in a 3:1 ratio in the presence of A_CE. After scoring, the meristic gametophytes were segregated and permitted to grow further to allow more time for antheridia to develop before determining their sex. Of the meristic gametophytes, hermaphrodites and females segregated in a 1:2 ratio. The segregation ratios of meristics to males and hermaphrodites to females indicate that the fem1 her19 and not1 her19 double mutants are female. Similar results were obtained using other fem1 group mutants (data not shown). On the basis of the double-mutant phenotypes, we conclude that the FEM1 group and NOT1 genes are epistatic to HER19.

Epistasis among the FEM1 group, NOT1, and MAN1 genes:
Several mutant females were crossed by the sperm of a man1 hermaphrodite to determine the phenotypes of the fem1 group man1 and not1 man1 double mutants (Table 6). The man1 phenotype is illustrated in Fig 1C. The fem1 x man1 cross resulted in progeny that segregated females and males in a 1:1 ratio in the presence of A_CE (Table 6). Similar results were obtained using other fem1 group mutants (mutants 4, 5, 10, and 11) as females (data not shown). In the absence of A_CE, sibling gametophytes segregated females, hermaphrodites, and man1 gametophytes in a 2:1:1 ratio (Table 6). These results indicate that the fem1 group man1 double-mutant gametophytes are female in the presence or absence of A_CE. The not1-2, 1-3, 1-4, and 1-8 x man1 crosses resulted in progeny that segregated females and males in a 1:3 ratio in the presence of A_CE; in the absence of A_CE, sibling gametophytes segregate females, hermaphrodites, and man1 gametophytes in a 1:1:2 ratio (Table 6). These results indicate that the phenotypes of the not1 man1 double mutants are the same as the man1 single mutant, i.e., a hermaphrodite with many antheridia in the absence of A_CE and male in its presence. On the basis of the double-mutant phenotypes, we can say that the FEM1 group genes are epistatic to MAN1, while MAN1 is epistatic to NOT1.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The feminizing mutants of Ceratopteris:
Sixteen independently derived mutations of Ceratopteris that completely or partially feminize the gametophyte have been isolated and characterized. Seven of them are very tightly linked and morphologically indistinguishable from the fem1 mutant previously characterized (BANKS 1994 ). None of these seven fem mutants form antheridia, even in the presence of A_CE, and all display the same double-mutant phenotype when combined with other sex-determining mutations (i.e., tra, her, and man). For these reasons, they are likely to be alleles of the FEM1 locus, although the possibility that they are alleles of other very closely linked loci cannot be ruled out. Sporophytes homozygous or heterozygous for these feminizing mutations are indistinguishable from wild-type sporophytes, indicating that they are not necessary for normal sporophyte development. On the basis of the mutant gametophyte phenotypes of these mutants, we conclude that the FEM1 gene(s) is an important regulator of sexual development that functions to promote the development of antheridia and/or to repress the development of the meristem and archegonium in the gametophyte.

Eight other feminizing mutants characterized define at least one other fem-like locus that is unlinked to FEM1. Unlike the fem1 mutants, all but one mutation affected both the gametophyte and sporophyte when homozygous. The lack of complementation among these alleles in the sporophyte indicates that all are alleles of another FEM-like locus. To distinguish them from fem1, they are referred to as notchless1 (not1) alleles because several of them lack the meristem notch typical of wild-type and her hermaphrodites and fem1 females. All of the not1 alleles had similar gametophytic phenotypes when combined with other sex-determining mutants, which is expected if they are alleles of the same locus.

The recessive not1 mutants vary in phenotype, depending on the allele, indicating that they form an allelic series. The strongest not1 alleles (not1-1, 1-4, 1-7, and 1-9) include those that form no antheridia, cannot be self-fertilized, and have aberrant notchless meristems. Gametophytes carrying the weakest allele, not1-2, produce a normal meristem and several antheridia (but much less than wild-type hermaphrodites) and, when self-fertilized, form normal sporophytes. Alleles with intermediate phenotypes, including not1-3, 1-6, and 1-8, can produce at least one antheridium, yet after self-fertilization, form abnormal, sterile sporophytes. The pleiotropic effects of most not1 mutations reveal that the NOT1 gene is an important regulator of sex expression in the gametophyte and of sporophyte development.

Interactions among the fem1, not1, and other sex-determining mutants:
The epistatic relationships among the new feminizing mutants and previously characterized sex-determining mutants were determined by comparing the phenotypes of double- or triple-mutant gametophytes to single-mutant gametophytes. These results are summarized in Table 7 and have been used as a basis for expanding the model of the sex determination pathway in Ceratopteris, illustrated in Fig 2. Since this model is based solely on genetic analyses, no assumptions regarding the molecular mechanisms involved are implied. According to this model, in Ceratopteris two major groups of sex-determining genes are responsible for the development of male traits (antheridia) or female traits (multicellular meristem and archegonia) during the sex-determining process. One group includes the TRA genes, of which there are at least two loci (BANKS 1997B ), and the other is the FEM1 locus (BANKS 1994 ). The TRA genes have two functions. One is to promote the development of female traits and the other is to repress the development of male traits. To demonstrate that these two functions could be genetically separated, a screen for mutants defective in their ability to repress maleness, but wild type in their ability to promote female traits, was carried out and resulted in the isolation of the man1 mutant (BANKS 1997B ). Since the man1 gametophyte is incapable of repressing antheridia development when A_CE is absent, it produces an order of magnitude of more antheridia than normal hermaphrodites. The identification and further characterization of this mutant led to the conclusion that in addition to its female-promoting function, TRA represses male development, but does so indirectly by activating the MAN1 gene. Once activated, MAN1 represses the gene responsible for male development (i.e., FEM1).

View larger version (19K): In this window In a new window Download PPT slide   Figure 2. A model of the sex-determining pathway in Ceratopteris. The potential functions and interactions among the sex-determining genes are shown in A. The activities of the sex-determining genes in wild-type gametophytes grown in the presence of ACE are shown in B, and, in the absence of ACE, in C. (D) The activities of the sex-determining genes in mutant not1 gametophytes grown in the presence of ACE. The line ending in a bar between the HER and TRA genes in D is dotted to indicate that although the HER genes are active and can repress the TRA genes, the TRA genes are not repressed. The TRA genes are not repressed because repression of TRA also requires FEM1 (discussed in EBERLE and BANKS 1996 ) and the NOT1 genes (this study). Female sex traits include the multicellular meristem and archegonia; the one male sex trait in this species is antheridia. Lines ending in arrows indicate positive regulation (activation) and lines ending in bars represent negative regulation (repression).

The other major sex-determining gene in Ceratopteris is the FEM1 gene. If FEM1 functions both to promote maleness and to repress female development, it should be possible to genetically uncouple these two functions. To test this, we sought a fem-like mutant that was defective in its ability to repress female development (i.e., able to repress the TRA genes), but wild type in its ability to promote the development of male traits. Although able to promote the development of antheridia, this mutant was expected to have no or few antheridia (a fem1 phenotype) because the TRA genes would not be repressed in this mutant, which would lead to the repression of FEM1, the gene required for the development of male traits. The results of this study show that, on the basis of their phenotypes and epistatic interactions with other sex-determining mutations, the not1 mutants fulfill these predictions. We conclude from these results that in addition to its male-promoting function, FEM1 also functions to repress female development, but does so indirectly by activating the NOT1 gene, which, in turn, represses the TRA genes.

Given that the FEM1 and TRA genes mutually exclude each other's expression via the NOT1 and MAN1 genes, respectively, the sex of the Ceratopteris gametophyte ultimately depends on which of two genes (FEM1 or TRA) is activated early in the sex-determining process. This is determined by A_CE, the male-inducing pheromone. According to the model, the presence of A_CE activates the HER genes, which repress the TRA genes (Fig 2B). Since the TRA genes are not active, MAN1 is also not active. The absence of active MAN1 leads to the activation (lack of repression) of FEM1. FEM1 in turn activates NOT1, which, together with the active HER genes (EBERLE and BANKS 1996 ), continues to repress the TRA genes. Since the FEM1 gene is active and the TRA genes are not when A_CE is present, the young gametophyte develops as a male in the presence of A_CE. When A_CE is absent (Fig 2C), the activities of the sex-determining genes are reversed; the TRA genes are not repressed, which ultimately leads to the repression of FEM1 by MAN1. The gametophyte therefore develops female traits when A_CE is absent.

This simple model of the sex-determining pathway in Ceratopteris is inherently flexible, allowing the sex of the gametophyte to be determined by its environment (specifically the presence or absence of pheromone in its surrounding medium) as the spore begins to germinate (BANKS 1994 ) rather than by genetic predetermination. Although consistent with all of the available genetic data thus far, this model does not explain the hermaphrodite, i.e., how antheridia form on an otherwise female gametophyte. One hypothesis is that the TRA and/or MAN1 gene products form a gradient within the meristem, the highest levels being in the center of the meristem where archegonia initiate, and the lowest levels in cells as they exit the meristem, where antheridia are know to initiate. As some cells exit the meristem, the levels of MAN1 or TRA would fall below a level necessary to repress FEM1. In this case, FEM1 would become active, TRA would be repressed, and the cell(s) would consequently differentiate as an antheridium. An alternative hypothesis is that another signal emanating from the hermaphrodite prothallus directly or indirectly activates the NOT1 gene in some cells as they exit the meristem, causing a change in gene activity that ultimately leads to the activation of the FEM1 gene. Testing of these and other hypotheses to explain how antheridia form on an otherwise female gametophyte will require cloning and molecular analyses of the sex-determining genes in Ceratopteris.

	FOOTNOTES

This article is dedicated to the memory of DeLill Nasser.
¹ Present address: Bioinformatics Research Center, North Carolina State University, Raleigh, NC 27695.

	ACKNOWLEDGMENTS

We thank Susan Lolle and Burt Bluhm for helpful discussions and Brody DeYoung and Drew Schultz for their help in the greenhouse. This research was supported by a grant from the National Science Foundation.

Manuscript received April 17, 2001; Accepted for publication August 20, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BANKS, J., 1994  Sex-determining genes in the homosporous fern Ceratopteris. Development 120:1949-1958[Abstract].

BANKS, J., 1997a  Sex determination in the fern Ceratopteris. Trends Plant Sci. 2:175-180.

BANKS, J., 1997b  The TRANSFORMER genes of the fern Ceratopteris simultaneously promote meristem and archegonia development and repress antheridia development in the developing gametophyte. Genetics 147:1885-1897[Abstract].

BANKS, J., L. HICKOK, and M. WEBB, 1993  The programming of sexual phenotype in the homosporous fern Ceratopteris richardii. Int. J. Plant Sci. 154:522-534.

DÖPP, W., 1950  Eine die antheridienbildung bei farnen fördernde substanz in den prothallien von Pteridium aquilinum L. Kuh. Ber. Dtch. Got. Ges. 63:139-147.

EBERLE, J. and J. BANKS, 1996  Genetic interactions among sex-determining genes in the fern Ceratopteris richardii. Genetics 142:973-985[Abstract].

NÄF, U., 1959  Control of antheridium formation in the fern species Anemia phyllitidis. Nature 184:798-800.

NÄF, U., 1979 Antheridiogens and antheridial development, pp. 436–470 in The Experimental Biology of Ferns, edited by A. F. DYER. Academic Press, London.

NÄF, U., K. NAKANISHI, and M. ENDO, 1975  On the physiology and chemistry of fern antheridiogens. Bot. Rev. 41:315-359.

WARNE, T. and L. HICKOK, 1991  Control of sexual development in gametophytes of Ceratopteris richardii: antheridiogen and abscisic acid. Bot. Gaz. 52:148-153.

WARNE, T., L. HICKOK, and R. SCOTT, 1988  Characterization and genetic analysis of antheridiogen-insensitive mutants in the fern Ceratopteris. Bot. J. Linn. Soc. 96:371-379.

YAMANE, H., 1998  Fern antheridiogens. Int. Rev. Cytol. 184:1-32.

This article has been cited by other articles:

				L. G. Quintanilla, L. d. Soto, A. Jimenez, and M. Mendez Do antheridiogens act via gametophyte size? A study of Woodwardia radicans (Blechnaceae) Am. J. Botany, June 1, 2007; 94(6): 986 - 990. [Abstract] [Full Text] [PDF]

				L. Gumaelius, B. Lahner, D. E. Salt, and J. A. Banks Arsenic Hyperaccumulation in Gametophytes of Pteris vittata. A New Model System for Analysis of Arsenic Hyperaccumulation Plant Physiology, October 1, 2004; 136(2): 3198 - 3208. [Abstract] [Full Text] [PDF]

				M. Tanurdzic and J. A. Banks Sex-Determining Mechanisms in Land Plants PLANT CELL, June 1, 2004; 16(suppl_1): S61 - S71. [Full Text] [PDF]