Department of Biological Sciences
There are two basic morphologies among gametophytes of Ceratopteris richardii, the large meristic hermaphrodite and the small ameristic male. In vitro studies have indicated that the small male morphology is characteristic of multispore cultures, with single-spore cultures resulting in large hermaphrodites. The pheromone antheridiogen is known to cause the induction of the small ameristic male phenotype in C. richardii gametophytes. We investigated the effect of competition for nutrients on gender expression and the timing of the appearance of morphological changes associated with the antheridiogen response. Competition for nutrients had no effect on gender expression in our experiment. Morphological changes associated with the antheridiogen response were apparent upon the emergence of a 2 dimensional gametophyte from the spore wall. If the antheridiogen response is derived from competition for nutrients, there is no apparent remnant of an effect of competition for nutrients on gender expression in C. richardii. The appearance of a size dimorphism that is associated with the antheridiogen response in the earliest stages of gametophyte culture suggests that the factors that trigger the antheridiogen response begin to affect gametophyte development at the earliest stages.
Gender in plants is a result of the expression of the haploid gametophyte genome. In the most widely studied plants, seed plants, the haploid gametophytic phase of the plant life history is greatly reduced and completely dependent upon the diploid sporophyte phase. Analysis of gender determination in seed plants is therefore greatly complicated by sporophytic effects on the expression of the gametophyte genome. For example, Goldberg et al. (1993) have suggested that there are about 25,000 genes expressed in a tobacco anther at the time the microspore nucleus begins to divide, of which about 10,000 are anther specific. Resolving which genes are gametophytic genes whose products are involved in gender determination and which are sporophytic genes whose products affect the gametophytic genes and gene products involved in gender determination becomes a very difficult and complex task.
Homosporous ferns, however, include free-living gametophytic and sporophytic life history phases. Thus gender determination in gametophytes among homosporous ferns can be studied as a phenomenon completely independent of sporophytic effects, allowing for independent analyses of the developmental mechanisms regulating gender determination in the gametophytic phase of the plant life history.
We investigated gender determination in the fern species Ceratopteris richardii Brongniart (water sprite). We chose this plant for our investigations as it has a short life history and is easy to culture (Hickok et al., 1995). Gametophyte development in C. richardii is characterized by the presence of two distinct morphologies, the large meristic hermaphrodite and the small ameristic male with large numbers of antheridia (Banks et al., 1993). All spores have the potential to develop either type of morphology. When grown in isolation, single spores develop into large hermaphrodites (Banks et al., 1993). The small males appear only in multispore cultures. It has been proposed that the intergametophytic interaction responsible for the induction of the small male morphology is regulated by the pheromone antheridiogen (Näf et al., 1975).
Mature C. richardii gametophytes are of the cordate type also characteristic of ferns such as Diplazium and Athyrium species, however, the pattern of development to the cordate morphology differs. When spores of ferns such as Diplazium species germinate, the spore wall is immediately shed, and an exosporic filamentous one-dimensional (1D) stage is formed. The 1D stage is followed by a spatulate (2D) stage when cells begin to divide in 2 dimensions, which may eventually develop into a mature cordate gametophyte with a meristic region and bearing archegonia and possibly also antheridia (Atkinson, 1967). Banks et al. (1993) have demonstrated that when spores of C. richardii germinate, the spore wall is not immediately shed, and an initial multicellular endosporic stage follows spore germination. The spore wall is shed when the multicellular gametophyte grows out of the spore wall, resulting in a multicellular 2D exosporic gametophyte, which may develop into a mature cordate gametophyte.
Schneller (1988) has demonstrated that germination of Athyrium filix-femina spores in the dark in the presence of A. filix-femina antheridiogen result in filamentous exosporic gametophytes that shed the spore wall upon germination, without the initial multicellular endosporic stage. When spores of A. filix-femina were germinated in the dark in the presence of antheridiogen from Dryopteris filix-mas, however, the spore wall was not shed immediately upon germination, and there was an initial multicellular endosporic stage in gametophyte development. Thus A. filix-femina appears to be capable of both types of gametophyte development, with the expression of the two types of gametophyte development possibly being regulated by the chemical characteristics of antheridiogens. At the very least, spores germinated in different environments can develop differently in Athyrium filix-femina.
A model of gametophyte development described by Näf et al. (1975) suggests that the small ameristic male morphology in fern species like C. richardii is due to the secretion of antheridiogens by meristic archegoniate gametophytes. However, Banks et al. (1993) suggested that antheridiogen secretion by gametophytes of C. richardii begins at the onset of the 2D exosporic stage (their stage 4), an ameristic stage. Sayers and Hamilton (1995) observed an induction of the small ameristic male morphology in gametophytes that could not have been exposed to secretions of meristic archegoniate gametophytes. Thus, the model of gametophyte development predicting that the induction of the small ameristic male morphology is due to the secretion of antheridiogens by meristic archegoniate gametophytes does not explain gametophyte development observed in Ceratopteris species.
Secretion of antheridiogens by ameristic gametophytes has also been observed by Hickok and Kiriluk (1984), who noted that a medium supplemented with 10-4M indole-3-butyric acid (IBA) resulted in cultures consisting of exclusively ameristic male gametophytes in cultures of Ceratopteris thalictriodes. Antheridiogens are, however, chemically similar to gibberellins, not auxins (Schraudolf, 1985; Takeno et al. 1989). Assuming that antheridiogens are the cause of the ameristic male morphology and that the only source of antheridiogens in Hickok and Kiriluk's (1984) cultures were the gametophytes themselves, ameristic gametophytes must have been producing antheridiogens in those cultures. Thus, if one assumes that antheridiogens induce the small male morphology, Hickok and Kiriluk's (1984) results indicate environmental factors may induce the secretion of antheridiogens (since the indole-3-butyric acid (IBA) was added to the medium, as was thus a factor present in the environment in Hickok and Kiriluk's (1984) experiments). The results of Hickok and Kiriluk (1984), Banks et al. (1993), and Sayers and Hamilton (1995) all indicate that secretion of antheridiogens begins prior to the onset of the meristic, archegoniate stage of gametophyte development.
Banks et al. (1993) observed that the male morphology was induced only in those gametophytes exposed to antheridiogen from spore germination to the initiation of the 2D exosporic phase. Banks et al. (1993) also suggested that antheridiogen secretion begins at the 2D exosporic stage (their stage 4) of gametophyte development. Sayers and Hamilton (1995) observed that in paired spore cultures of C. richardii, the spores most likely to develop into small males were smaller, later germinating spores. All spores observed by Sayers and Hamilton (1995) germinated over a two-day period, and there were no 2D exosporic gametophytes present in any culture during the period of spore germination. It is thus not possible for the small ameristic males observed to have been exposed to secretions of 2D exosporic gametophytes from time of spore germination to the onset of the 2D exosporic phase, as suggested by Banks et al. (1993). Thus more information is needed to resolve the specific time in gametophyte morphogenesis when antheridiogen secretion begins.
Schneller et al. (1990) predicted an ecological role of antheridiogens in fern populations. Fern spores, including those of Ceratopteris species, are known to have a light requirement for germination (Schraudolf, 1985). Schneller et al. (1990) suggested that antheridiogen functions as a germination signal to buried spores. Antheridiogen induces the expression of antheridia on gametophytes resulting from dark-germinated spores, whose sperms then fertilize the archegoniate gametophytes on the soil surface. In the Schneller et al. (1990) model, antheridiogen is an adaptive response involving an effect on spores prior to germination (the induction of dark germination), and subsequent to germination (induction of the male morphology).
Haig and Westoby (1988) and Willson (1981) suggested that antheridiogens may have evolved as an adaptive response to competitive interactions. Differences in size is one of the most common outcomes of competitive interactions among individuals (Keddy, 1989). In plant populations, all else being equal, early germinating individuals enjoy a great competitive advantage over late germinating individuals. Thus it may well be that later germinating spores expressing the male gender as rapidly as possible, sacrificing the possibilty of female gender expression, have an evolutionary advantage over later germinating spores that follow a developmental program leading to the large hermaphrodite.
We sought to investigate the effects of abiotic environmental variation on gametophyte expression to assess the stability of the basic developmental program. Should abiotic environmental variation affect gametophyte development, the role of antheridiogens in gametophyte development would have to be assessed in that light for the given species, C. richardii in this case.
We also sought to determine the time during development when the two basic C. richardii gametophyte morphologies begin to diverge (the small ameristic males and the large, cordate, meristic hermaphrodites. We used a methodology described by Coleman et al. (1994), who suggested quantifying the relationship between age and developmental stage when studying phenotypic variation among plants. One can then observe the effects of different environmental factors (such as antheridiogen) on the relationship between age and developmental stage to generate hypotheses regarding the factors that affect development. We harvested gametophytes from multispore cultures and counted the number of cells in each gametophyte. If all gametophytes were developing towards the same morphology, gametophyte sizes should show a normal distribution at all times throughout development. If gametophytes of C. richardii begin to develop towards the two distinct morphologies (the large hermaphrodite and the small male), the two distinct size classes associated with the two morphologies should appear in gametophyte cultures. We observed the time in development when two distinct size classes appeared among gametophytes in multispore cultures as evidence as to when the initiation of the intergametophytic effects leading to the development of the two morphologies could have begun.
We also tested the prediction that C. richardii gametophytes must be exposed to secretions from 2D exosporic gametophytes by removing gametophytes from multispore cultures prior to the onset of the 2D exosporic stage, and then observing the developmental state several days later. If we observed any small ameristic males, we would challenge the hypothesis (of Banks et al., 1993) that secretions from 2D exosporic gametophytes are necessary for the induction of the small ameristic male morphology.
Material and Methods
Spores of C. richardii (Hn-n) were provided by T. Warne at the University of Tennessee at Knoxville. Gametophytes were cultured in 20mm plastic culture dishes, on a nutrient agar medium described by Klekowski (1969), which has a pH = 4.3, except that 1.5% agar was used to reduce evaporation. This medium was prepared at three different nutrient levels: full-nutrient level, 0.1% and 0.01% of full-nutrient level.
Gametophytes were cultured on a light bench under continuous fluorescent and incandescent light under enclosures constructed of plastic wrap, with each enclosure containing two 100mm diameter culture dishes filled with water, to reduce evaporation. Temperatures on the light bench ranged from 28oC - 31oC throughout all experiments. Fans were placed at each end of the light bench to maintain a lower and more constant temperature across the light bench. Following harvest, gametophytes were cleared with chloral hydrate and observations of gametophytes were made using a Wild light microscope.
For the analysis of the effects of competition, unsterilized spores were imbibed in sterile deionized water in the dark for 2 days at 10oC. Twenty-five full-nutrient agar culture dishes were inoculated with single spores and 25 full-nutrient agar culture dishes were inoculated with paired spores (with spores touching). Spores were sown individually using a sterile micropipettor and sterile tips in a laminar flow hood, as described in Sayers and Hamilton (1995). Similar treatments were prepared with 0.1% and 0.01% of full nutrient agar. Cultures were placed on the light bench for 13 days. Gametophytes were then harvested, with the size and gender of each gametophyte recorded.
We observed two different morphologies among male gametophytes; those with a meristic notch (nM), and the "classic" small male (cM), an ameristic gametophyte with numerous antheridia. The meristic notch is characteristic of the large hermaphrodite morphology, and thus we considered nM gametophytes to be later maturing hermaphrodites based on previous observations that antheridia usually appear prior to archegonia during the development of large hermaphrodites (Banks et al. , 1993). We thus classified hermaphroditic, notched males and female gametophytes as "H" (hermaphrodite), classic small males as "M" (male) and asexual gametophytes as "A".
Gametophyte size was estimated by measuring the width and length of the gametophyte and using these dimensions to calculate the area of an ellipse. The one-way analysis of variance in PC-Statistics v2.5 (Marvasti, 1991) was used to compare the sizes of "H" gametophytes among treatments for each of the paired- and single-spore cultures. Treatment means were compared using the least significant difference (LSD) as described by Steel and Torrie (1980). For all comparisons of means we used the critical value for p < 0.01 to calculate the LSD.
For the investigation of the timing of morphological divergence, 10 full-nutrient agar culture dishes were inoculated in a laminar flow hood with dry unsterilized spores by gently tapping a vial containing spores such that a thin layer of spores could be observed through a dissecting microscope at 10X magnification. Cultures were placed on the light bench and observed daily. Spore germination was observed as the opening of the spore wall and the protrusion of the first rhizoid. Ten randomly chosen gametophytes were harvested from each of 5 randomly chosen cultures beginning with the day 2D exosporic gametophytes first appeared and ending with the day that multiple archegoniate gametophytes appeared within cultures. Gametophytes were chosen by placing the point of a heat-sterilized dissecting needle into the culture using the naked eye, and then selecting the 10 gametophytes nearest to that point, except for the last day of sampling, where each plate was sampled for 5 small gametophytes and 5 large gametophytes nearest the point. A change in sampling strategy was required for the last day as it proved impossible to determine the 10 gametophytes nearest the point due to the increased area covered by thallus tissue and the resultant overlapping of gametophytes.
Gametophyte size was recorded as numbers of cells. The number of cells for each gametophyte was counted twice, and if the two counts were within an error of 10%, the first count was used as an estimate of the number of cells. When the two counts diverged more than 10%, the observer recounted.
Euclidean distance was used as a proximity measure for the UPGMA clustering algorithm in NTSYS-pc v. 1.8 (Rolfe, 1993) as a simple a priori method of resolving any size differences among gametophytes. A separate cluster analysis was completed for each of the days that the above-mentioned gametophyte harvests were performed.
Effects of Competition
In the investigation of the effects of competition, M gametophytes occurred once in the 0.1% and twice in the 0.01% of full-nutrient level treatment in single-spore cultures (Table 1). In paired-spore cultures there were HM pairs in 12/21 cultures at 0.01% of full-nutrient, compared with 7/22 HM pairs at 0.1% of full-nutrient and 8/21 HM pairs at full-nutrient levels (Table 1). There were HH pairs in 8/21 cultures at 0.01% compared with 14/22 HH pairs at 0.1% of full-nutrient and 13/22 HH pairs at full-nutrient levels (Table 1).
Among H gametophytes the ANOVA indicated significant size differences among treatments in both paired- and single-spore cultures (F = 4.22, p < 0.05 for single-spore, F = 5.38, p < 0.05 for paired-spore cultures). In paired-spore cultures, the mean size of H gametophytes for the full-nutrient treatment was 0.727 mm2, which was significantly larger than in both the 0.1% of full-nutrient (0.516 mm2) and the 0.01% of full-nutrient (0.525 mm2) treatments. In single-spore cultures, the mean size of H gametophytes in the full-nutrient treatment was 0.683mm2, which was significantly larger than in 0.1% of full-nutrient treatment (0.351mm2), but not significantly different from 0.01% of full-nutrient treatment (0.527mm2).
In the investigation of morphological divergence, opening of spore walls was observed 3 days following inoculation and spore germination was first observed 4 days following inoculation. Multicellular endosporic gametophytes first appeared 6 days following inoculation. Exosporic 2D gametophytes first appeared 7 days following inoculation. Antheridia began to appear on larger gametophytes sampled 8 days following inoculation, while archegonia began to appear on larger gametophytes sampled 9 days following inoculation. Developing antheridia were observed on small ameristic gametophytes sampled 9 days following inoculation. Ten days following inoculation, all large gametophytes sampled were hermaphroditic and small ameristic gametophytes with numerous antheridia were present.
Cluster analysis (figure 1) indicates two main size clusters for each of days 7, 8, 9 and 10, with the magnitude of the differences separating the two clusters increasing from day 7 to day 10. The specific gametophytes constituting each of the main size clusters and the size of each gametophyte is recorded in table 2.
There was no contamination observed in any of the gametophyte cultures.
There was no effect of nutrient level on the number of small males in single spore cultures. If competition for nutrients among neighboring gametophytes could induce the small male morphology, a switch from almost exclusively hermaphroditic to almost exclusively male cultures would have been expected at some nutrient level in single spore cultures.
The lack of substantial differences in the expression of the small male morphology in double spore cultures confirms the result from single spore cultures. If competition for nutrients were the factor that induced the small male morphology, then reduction in nutrient level should have substantially increased the frequency of H/M pairs in double spore cultures. While we did observe and increase from 8/21 H/M cultures at full nutrient to 12/21 H/M cultures at 0.01% of full nutrient, the lack of a clear effect in single spore cultures prevents any conclusion that competition for nutrients is major factor regulating gender expression in C. richardii.
We observed two distinct size clusters at the initiation of the 2D exosporic stage (stage 4 of Banks et al., 1993). Banks et al. (1993) were unable to observe morphological differences among gametophytes at this stage. It is likely that the differences in size among gametophytes at stage 4 are not detectable in the absence of a quantitative measure of gametophyte size and the application of statistical methods such as cluster analysis, and the differences were in fact present in the cultures analyzed by Banks et al. (1993).
Since spores began to open 3 days following inoculation and the first 2D exosporic gametophytes were sampled 7 days following inoculation, it took more than 3 and less than 4 days following the opening of spores for gametophytes to reach the 2D exosporic stage. The only gametophytes that could have possibly been exposed to secretions of 2D exosporic gametophytes from the opening of the spore wall to the onset of the 2D exosporic stage would have been those gametophytes harvested 10 days following inoculation (A spore that germinates 7 days following inoculation and reaches the 2D exosporic stage 3 days following germination would have been exposed to the secretions of 2D exosporic gametophytes from germination till the onset of the 2D exosporic stage). The observation that the small ameristic male morphology first appeared 10 days following inoculation is consistent with the Banks et al. (1993) model, which suggests that the small male morphology is induced in response to exposure to antheridiogens from spore germination to the onset of the 2D exosporic stage.
Hamilton and Sayers (1995) observed the induction of the small male morphology in paired spore cultures where both spores germinated within a day or two of each other. C. richardii gametophytes have never been observed to grow to the 2D stage within 3 days on the light bench used for both this investigation and that of Hamilton and Sayers (1995). Gametophytes in such cultures thus could not have been exposed to antheridiogen from the germination of the spore until the onset of the 2D exosporic stage if antheridiogens are secreted only by 2D exosporic gametophytes, as no 2D exosporic gametophytes were present at the time of spore germination. More detailed analyses of spore germination time and the time of appearance of developmental stages within paired spore cultures are needed to determine if either spore in paired spore cultures where an antheridiogen response is observed could be exposed to secretions from a 2D exosporic gametophyte from germination until the onset of the 2D exosporic stage.
Antheridiogens in our media could have affected our results, however if any antheridiogens were present in our media, the effects would have been observed in single spores cultures during the analysis of the effects of competition, as the media used for all cultures in the analysis of morphological was identical to the full nutrient media used in the analysis of the effects of competition.
Our results do clearly show that significant size differences appear in gametophyte cultures prior to the appearance of any other morphological differences. In fact, size differences are apparent with the first emergence of gametophytes from the spore wall. If this size dimorphism is a consequence of the antheridiogen response, then the factors that cause the antheridiogen response begin their action at the earliest stages of gametophyte development. Our results are thus consistent with the prediction of Banks et al. (1993) that gametophytes must be exposed to antheridiogen from spore germination until the onset of the 2D exosporic stage. We did, however, observe morphological characteristics associated with gametophyte development at earlier stages than have been previously observed. However our results do not explain the results of Hamilton and Sayers (1995) who observed an antheridiogen effect in gametophytes that were not exposed to the secretions of 2D exosporic gametophytes from spore germination until the onset of the 2D exosporic stage.
Our results also show variation in nutrient levels has very little effect of gametophyte development, if any at all. IF the antheridiogen response is a modification of an adaptation to the effects of competition, we observed no remnant of that competitive interaction for the case of competition for nutrients. Any effect of competition would likely involve competition for light.
Atkinson, L. 1968. The gametophyte of Diplazium. Phytomorphology 17: 99-109.
Banks, J., L. Hickok and M. Webb. 1993. The programming of sexual phenotype in the homosporous fern Ceratopteris richardii. International Journal of Plant Sciences 154: 522-534.
Coleman, J., K. McConnaughay and D. Ackerly. 1994. Interpreting phenotypic variation in plants. Trends in Ecology and Evolution 9: 187-191.
Goldberg, R. B., T. Beals and P. Saunders. 1993. Anther development: basic principles and practical applications. The Plant Cell 5: 1217-1229.
Haig, D. and M. Westoby. 1988. Sex expression in homosporous ferns: An evolutionary perspective. Evolutionary Trends in Plants 2: 111-119.
Hamilton, R. G. and R. M. Lloyd. 1991. Antheridiogen in the wild: the development of fern gametophyte communities. Functional Ecology 5: 804-809.
Hickok, L. and R. Kiriluk. 1984. Effects of auxins on gametophyte development and sexual differentiation in the fern Ceratopteris thalictriodes (L.) Brongn. Botanical Gazette 145: 37-42.
Hickok, L., T. R. Warne and R. Fribourg. 1995. The biology of the fern Ceratopteris and its use as a model system. International Journal of Plant Sciences 156: 332-345.
Klekowski, E. J. 1969. Reproductive biology of the Pteridophyta III. A study of the Blechnaceae. Botanical Journal of the Linnean Society 62: 361-377.
Keddy, P. A. 1989. Competition. Chapman and Hall, New York.
Marvasti, F. 1991. PC-Statistics, v 2.5. Franklin, Beadle and Associates, Wilsonville Oregon.
Näf, U., K. Nakashini and M. Endo. 1975. On the physiology and chemistry of fern antheridiogens. Botanical Review 41: 315-359.
Rolfe, F. J. 1993. NTSYS-pc version 1.8. Applied Biostatistics Inc. Setauket, NY.
Sayers, A, and R. Hamilton. 1995. The effect of neighbors on gametophyte development in Ceratopteris richardii. American Fern Journal 85: 47-53.
Schneller, J. J. 1988. Spore bank, dark germination and gender determination in Athyrium and Dryopteris. Results and implications for population biology of Pteridophyta. Botanica Helvetica 98: 77-86.
Schneller, J. J., C. H. Haufler and T. A. Ranker. 1990. Antheridiogen and natural gametophyte populations. American Fern Journal 80: 143-152.
Schraudolf, H. 1985. Phytohormones and Filicanae: Chemical signals triggering morphogenesis in Schizaeceae. Pages 270-274 in M. Bopp, ed., Plant Growth Substances. Springer-Verlag, Berlin.
Steel, G. D. and J. H. Torrie. 1980. Principles and Procedures of Statistics, A Biometrical Approach, 2nd Edition. McGraw Hill, Toronto
Takeno, K. H. Yamane, T. Yamauchi, N. Takahashi, M. Furber and L. Mander. 1989. Biological activities of the methyl ester of gibberellin A73, a novel and principle antheridiogen in Lygodium japonicum. Plant Cell Physiology 30:201-205.
Willson, M. R. 1981. Sex expression in fern gametophytes: some evolutionary possibilities. Journal of Theoretical Biology 93: 403-409.