Transmission of gametic nuclei through a fertilization tube during mating in a primitive dinoflagellate, Prorocentrum micans Ehr

In 1934 Chatton & Biecheler observed the presence of unquestionable sexual reproduction in the parasitic dinoflagellate Coccidinium mesnilii, during which the two-by-two fusion of slightly dissimilar spores from different individuals gave rise to a mobile zygote (an ookinete) with two pairs of flagella. In the past 15 years many authors have shown that a large number of dinoflagellate species can reproduce sexually when conditions in the medium become unfavourable (Cao Vien, 1967a,b, 1968; Zingmark, 1970; Von Stosch, 1972, 1973; Beam & Himes, 1974; Pfiester, 1975, 1976, 1977, 1984; Tuttle & Loeblich, 1975; Turpin et al. 1978; Anderson & Wall, 1978; Walker & Steidinger, 1979; Pfiester & Skvarla, 1979, 1984; Pfiester & Lynch, 1980; Chapman et al. 1980, 1981; Spector et al. 1981; Yoshimatsu, 1981; Walker, 1982; Anderson et al. 1984;

Sako et al. 1984; Coats et al. 1984; Anderson & Lindquist, 1985).

In Prorocentrum micans Ehr., the subject of the work reported here, aberrant forms have been described by Braarud & Rossavik (1951). Among these forms were vegetative cysts collected from the bottoms of old culture flasks. In 1952 Biecheler reported the existence of ‘nuclear cyclosis’ in P. micans. This phenomenon has since been shown in other species (Von Stosch, 1972) to be connected with meiosis; this was therefore the first observation of a phenomenon related to sexual reproduction in this species. De Sousa e Silva (1959, 1965) reported seeing among natural populations joined individuals of P. micans in lagunar plankton, but did not connect this with a sexual process. Cao Vien (1967a) thought that these linked individuals merely represented a particular method of vegetative multiplication, of which he gave a detailed description.

The appearance of cell types that were different from vegetative forms was induced by changing the medium (from Erdschreiber’s to Provasoli’s). We subsequently observed the development of these cell types and assayed their comparative DNA content by microspectrofluorimetry. Comparing these values with those found for vegetative cells, we conclude that sexual reproduction occurs in this species.

The P. micans strain used, from the Botany School, Cambridge, was routinely grown in the laboratory in Erdschreiber’s medium in an alternating 12 h-light/12 h-dark cycle (see Table 1A). Under these conditions the cell cycle of this species is 5·5 days; divisions of vegetative cells can be seen only during the first 6h of the light phase; the phenomenon is relatively rapid and affects only 17% of the population each day (Bhaud & Soyer-Gobillard, 1986). The appearance of new forms was induced by a change of culture medium. A sample of the strain (7000 cells ml⁻¹) was centrifuged and then resuspended in Provasoli’s medium (see Table IB); it was then subcultured every week in the same medium (1:1 dilution of the culture). After the fourth subculture cell forms different from normal vegetative forms could be observed.

The observations were performed in vivo, either without staining (by differential interference contrast under a Leitz microscope) or under bright-field optics with the nuclei stained with 1% Methyl Green. Photographs were taken with a Leitz Orthomat-Orthoplan microscope with Kodak film, ASA 400.

The relative amounts of DNA were determined on cells lightly fixed by Karnowsky’s (1965) method as modified by Soyer (1977). These measurements were made with a micro-spectrofluorimeter after the cell nuclei had been stained with an intercalating base, ethidium bromide (10 μgml⁻¹ of the fixation buffer).

The fluorescence spectra of ethidium bromide (EthBr)-stained nuclei were recorded using a computerized micro-spectrofluorimeter built in the Laboratory of Physical Chemistry, University of Perpignan (Salmon & Viallet, 1978, 1981; Salmon et al. 1981). The fluorescence analysed originates from a precisely defined region of the microscopic field, large enough to take in the whole of the fluorescing nucleus. The size of this region was selected by means of a four-leaf diaphragm used as the entrance slit of the microspectrofluorimeter dispersing system, and adjusted as necessary to the size of the observed nucleus up to microscopic fields of 24μm × 24μm. The fluorescence was excited at 370 nm. For each individual of the population of nuclei, computation of the recorded fluorescence spectra makes it possible: (1) to verify that the fluorescence was always emitted by EthBr; and (2) to evaluate the fluorescence intensity of the fluorophore, which in our experimental conditions was always proportional to the quantity of EthBr. Keeping the overall expenmental conditions constant, we quantified the fluorescence intensity for each nucleus by integrating the fluorescence in the range 596—624 nm. The means of measured values and the corresponding standard errors were calculated for the nucleus of P. micans in the different phases observed. This standard error is the result of the fluctuations of two independent sources: (1) fluctuations in the measurement of the fluorescence intensity quantified by the standard error S.D.F.; and (2) fluctuations resulting from differences in the EthBr content of the nucleus, which can be quantified using the standard error s.D.Q. As these fluctuations are independent, S.D.F. and S.D.Q. combine quadratically to give the calculated standard error S.D. (see Table 2):

During the first three sub-cultures in Provasoli’s medium, the growth curves of the populations of P. micans were similar to those obtained with Erd-schreiber’s medium (Fig. 1, curve A) and the length of the cell cycle was the same: 5·5 days (Bhaud & Soyer-Gobillard, 1986). After the fourth sub-culture in Provasoli’s medium, the samples taken throughout the day contained a high percentage (>50) of cells in the process of vegetative division. This percentage was distinctly higher than that seen during the preceding 3 days of sub-culture. The increase in dividing cells and their consistent presence in the culture suggested a blockage of vegetative divisions. This blockage, resulting in the death of the cells, was confirmed by the sharp drop in population (to 90%) from day 3 after the fourth sub-culture (Fig. 1, curve ß). On day 5, the remaining cells behaved in a way never seen in normal cultures: although morphologically similar to vegetative cells (Fig. 2A), they were grouped in pairs. The two cells of a pair, rotating around each other (Fig. 2B), soon became joined at their anterior ends in the flagellar region, without stopping swimming (Fig. 2C). At the moment of pairing, a communication tube formed in about 6h (Fig. 2D) between the pusular regions of the two individuals, with the flagella and apical spines surrounding this flattened tube, which is about four times wider (Fig. 2E,F) than it is thick (Fig. 2D): the two nuclei were then in a posterior position (Fig. 2C).

After formation of the tube, the nucleus of one of the cells migrated from the posterior part of the cytoplasm towards the anterior part and fitted itself into the tube (Fig. 2E). This migration was observed and photographed repeatedly. The nucleus continued its migration into the cytoplasm of the second cell. Once it had passed through the tube a transverse thickening appeared, partitioning the tube between the recipient cell and the enucleated cell, which then degenerated (Fig. 2F, arrow). In this way a cell with two nuclei and only two flagella (not visible here) was formed (Fig. 2F); the genetic material of the two nuclei fused, and the cell swelled, rounded up, and left the theca by complete separation of the two valves. After the nuclei had fused (Fig. 2G) and before the zygote started growing, it was indistinguishable from a vegetative cell approaching mitosis. Only the fact that there were no longer any normal vegetative divisions in the culture medium at that time suggests that the cells with this morphology were young zygotes. The round, free-swimming zygote with a thin cell wall grew and its nucleus enlarged (to about 53% of the cell area) (Fig. 2H). Two successive divisions (Fig. 2J-M) gave rise to four cells, which for a time remained joined anteriorly, forming a tetrad; it occasionally happened that the second division was not completely synchronous in the two cells formed by the first division (data not shown). Each cell of the tetrad was capable of undergoing normal vegetative division in Provasoli’s medium. There was about a week between the appearance of the paired cells and that of the tetrads. After that, the culture started growing again and we obtained a new growth curve comparable to that with Erdschreiber’s medium (Fig. 1, curve C).

We also carried out experiments in which we suddenly put normal vegetative cells (Erdschreiber’s medium) into unfavourable conditions (temperature less than 5°C or more than 25°C); this very rapidly (in less than 6h) provoked the pairing of a large number of cells and the formation of a fertilization tube in more than 50% of the population.

In order to find out whether the various described forms of P. micans represented the various steps of sexual reproduction, we made microspectrofluori-metric measurements of the relative DNA content of all the individuals or groups of individuals induced by culture in Provasoli’s medium. These values were compared with the values for the DNA content of vegetative cells at the end of cytokinesis.

To obtain the mean value of the minimum nuclear DNA content of the P. micans cells, we measured (separately) each nucleus of many vegetative cells grown in Erdschreiber’s medium at the end of cell division (during cytokinesis). We call this mean amount, as measured in the microspectrofluorimeter, q (see calculations in Materials and methods) (Table 2, column 1). In these usual culture conditions (see Materials and methods), the G₁ phase of the cell cycle lasts 5 days, and DNA replication and mitosis do not exceed 12 h (Bhaud & Soyer-Gobillard, 1986). It is therefore normal for most of the vegetative cells of the culture to be in phase G₁. The DNA content of these cells is equal to q (Table 2, column 2).

At a precise time every day, 17% of the population divide (Fig. 5 of Bhaud & Soyer-Gobillard, 1986). Just before mitosis, these cells, which were bigger and less numerous than cells in G₁, had large nuclei. Their DNA content was found to be 2q (Table 2, column 3). These cells had therefore replicated their DNA and were in G₂ phase or at the very beginning of mitosis.

When non-vegetative cells (in Provasoli’s medium) are observed and their nuclear DNA content is measured, the relative values are found to fall into a number of different classes (Table 2, columns 4–7).

Cells that were joined by a fertilization tube (Fig. 2C) contained the same amount of DNA as vegetative cells in G₁ and therefore had not yet replicated their DNA (Table 2, column 4). After the nucleus had moved from one cell into the other, the recipient cell contained 2q DNA (Table 2, column 5). Thus the nucleus moved from one cell to the other before DNA replication in either nucleus (Fig. 2F); the zygote (Fig. 2G) emerging from the theca contained 4q DNA (Table 2, column 6). In the growing zygote, the DNA of the two nuclei replicated in the cytoplasm of the recipient cell, before or after the two nuclei fused (Fig. 2H). After the first division, the two large cells that were still joined at their anterior ends each had 2q DNA (Fig. 2K). After the second division, each cell of a tetrad again contained q DNA (Fig. 2M; Table 2, column 7) and so did cells undergoing vegetative divisions, after adaptation to Provasoli’s medium (Table 2, column 2).

This means that the high signal-to-noise ratio of the fluorescence intensity measurement (≃90) results in only a slight increase in the calculated standard error (S.D. =494 counts as compared to S.D.Q. = 485 counts). Therefore the calculated standard error S.D. can be taken as an estimate of the dispersion of nuclear EthBr content and, by extension, of the DNA content of the nucleus. The calculated standard error represents 5–10% of the measured value, indicating that there is considerable homogeneity within each class of nuclei and furthermore that there are no statistically significant differences between the distributions obtained for the different classes. At least, statistical tests used for comparing the results in Table 2 indicate that there are no significant differences between columns 1, 2, 4 and 7 or between columns 3 and 5.

Among the aberrant forms observed by De Sousa e Silva (1965, Plate IV) were vegetative forms similar to those described by Cao Vien (1967a), and sexual forms (De Sousa e Silva, 1959, Plate I); the latter author classified all these observations as abnormal vegetative divisions due to poor conditions in the medium and did not consider the possibility of sexual reproduction.

In most dinoflagellate species in which sexual reproduction has been demonstrated, sudden or gradual nitrogen depletion in the medium seems to be the main cause of the induction of sexuality (for reviews, see Beam & Himes, 1980; Pfiester, 1984). Most of the species studied that react strongly to such nitrogen depletion are freshwater species. In marine dinoflagellates the action of this factor is less clear. In Gonyaulax tamarensis (Turpin et al. 1978) the transfer of the culture from a normal to a deficient medium induces only a ven’ small measure of sexuality (<1%). Those authors thought that even if the lack of nitrogen influenced the onset of sexual phenomena, it was certainly not the only cause. Turpin et al. (1978) emphasized, however, that in their non-axenic cultures it was difficult to deplete the nitrogen. Anderson et al. (1985) showed that in Gyrodinium uncatenum nitrogen deficiency of the medium was not sufficient to trigger sexuality; an unfavourable temperature was also required. Conversely, too low or too high a temperature did not induce sexuality unless the culture medium was also deficient in nutrients.

In P. micans, sexuality is generally induced by a change in culture medium but also by low or high temperature, as shown above. Although Provasoli’s medium is not nitrogen-deficient (see Materials and methods), it is difficult to compare it with Erdschreiber’s medium, as the latter contains soil extract. When transferred to Provasoli’s medium, P. micans can make three normal vegetative divisions before the appearance of sexuality; this suggests that the element, deficiency of which initiates sexuality, and which is present in Erdschreiber’s medium and absent from or present only in insufficient quantities in Provasoli’s medium, is normally stored in the cell in amounts greatly exceeding the requirements for one division. When these reserves are exhausted, the effects of this deficiency are manifested in the appearance of sexual phenomena. In P. micans, shifting into a sexual phase is not simply a mechanism preserving the species, since in the form of a hypnozygote the cell can await conditions favourable to the resumption of its development. Here sexual reproduction leads very rapidly (in the time taken for a normal vegetative division) to adaptation to new, initially less favourable, environmental conditions; after this passage through a cycle of sexual reproduction, the population growth curve in Provasoli’s medium again becomes comparable to that obtained in Erdschreiber’s medium (see Fig. 1, curves A, C).

Microspectrofluorimetric measurements clearly showed the existence of sexual reproduction with formation of a zygote, which after meiosis gives four normal vegetative cells with q DNA. As in the majority of dinoflagellates, the vegetative cells of P. micans are haploid, with a mean DNA content estimated as 42 pg (Haapala & Soyer, 1974). From this information we were able to calculate the approximate amounts of DNA in the nucleus of the various sexual forms. These values are reported in Table 2. The joined cells function as gametes. In order to find out where they come from, and whether there is a special division that has produced two reproductive cells from a normal vegetative cell, experiments were done in which normal vegetative cells were exposed to unfavourable conditions; as a result, normal pairing of the cells and the beginning of sexual divisions occur in less than 6h. P. micans has a long cell cycle: only 17% of the population divide each day (Bhaud & Soyer-Gobillard, 1986), a percentage well below that of the cells that can pair. The rapidity of this reaction excludes the hypothesis of a division that forms gametes, and suggests that all the vegetative cells have this potential. Such a hypothesis was also proposed by Anderson & Lindquist (1985) for G. tamarensis. However, it must be noted that in P. micans the two joined cells, although morphologically alike, behave differently; one gives up its nuclear material and then degenerates, while the other, the recipient, becomes a zygote. We cannot explain this difference in behaviour.

The absence of fundamental morphological differences between vegetative cells and gametes is common in dinoflagellates (see Table 3); the only details that generally make it possible to distinguish between them are often minute variations of pigmentation and size (see Table 3). Similarly, the differences observed between the gametes when there is anisogamy are differences of size, the small cell being the male gamete and the large cell the female (see Table 3). Such anisogamy is found in Ceratium comutum, C. horridum and Helgolandinium subglobusum (Von Stosch, 1972), in Gyrodinium uncatenum (Coats et al. 1984) and G. tamarensis (Turpin et al. 1978). Despite numerous recent studies (see Table 3), important points remain to be clarified, including how fusion is achieved between two gametes with a theca. Such fusion should lead to the formation of a zygote with, at least temporarily, four flagella. Such a phenomenon has been demonstrated in several species: Peridinium cinctum (Spector et al. 1981), P. cunningtonii (Sako et al. 1984), Gymnodinium pseudopalustre and Wolozynskia apiculata (Von Stosch, 1973), G. tamarensis (Turpin et al. 1978), G. monilata (Walker & Steidinger, 1979), Gymnodinium breve (Walker, 1982) and Gyrodinium uncatenum (Coats et al. 1984). In other cases, the zygote has two flagella; Pfiester (1975) thought that two of the flagella disappeared as the gametes fused. In P. micans, as the zygote is formed from the elements of a single individual, the free-swimming form obtained is biflagellate.

In P. micans the only point of attachment between the cells is near the apical spine in the region where the valves of the theca separate easily. When experimental dethecation of the cell is provoked, by the method of Morill (1984), the cell exits from the theca by this same region. Examination of the theca in the scanning electron microscope showed in this region a distinctive differentiation of the apical thecal plates (Soyer et al. 1982, Fig. 7) near the apical spine, below the flagellar pores and near the orifice of the pusule. This thecal depression seems to be near the point where the fertilization tube forms (study in progress). A fertilization tube can be seen between cells in all the paired P. micans cells. An analogous tube was seen in Peridinium cinctum (Spector et al. 1981). There is a ‘fertilization bridge’ in Crypthecodinium cohnii (Pfiester, 1984). Pfiester (1977) and Pfiester & Skvarla (1979) also reported a small protuberance between Pe. gatunense and Pe. volzii cells before fusion. Other differentiated contact zones have been demonstrated (see Table 3), but in every case the process ended with complete fusion of the two gametes. In P. micans, however, the tube remains the sole point of contact between the two cells, and nuclear passage is accomplished through this tube. Our observations in P. micans are somewhat different from those of Von Stosch (1972) in C. cornutum, in which the large female gamete ‘engulfs’ the (small) male gamete through an opening in the region of the sulcus. So in this case there is penetration of nuclear and cytoplasmic material, with only the theca being rejected.

Earlier data on cellular and biochemical aspects (see, e.g., Herzog et al. 1984) suggested that dinoflagellates, which display several ancestral prokaryote-like features, are a primitive group. We have been able to support this assertion by ribosomal DNA sequence comparisons (Maroteaux et al. 1985; Herzog & Maroteaux, 1986; Lenaerset al. 1987), which indicate close relationships between these primitive eukaryotes, ciliates and yeasts.

Prokaryotes, dinoflagellates, yeasts and ciliates share the common features of conjugation. However, in yeasts formation of a conjugation tube occurs between the two conjugants, with passage of one nucleus via the tube first, and then, after enlargement of the tube and fusion of the two sexual cells, formation of a dikaryon (Calleja et al. 1981), as observed in the dinoflagellate Peridinium cincturn (Spector et al. 1981). In ciliates conjugation consists of mutual exchange of genetic material (pronuclei) between the two conjugants via the mouth (for a review, see Raikov, 1982), while in bacteria (e.g. Escherichia coli Hfr (Jacob & Wollman, 1961; Sokatch, 1978)) all the genetic material from a donor (♂) is injected into a recipient cell (♀), as in P. micans. None of the observations on dinoflagellates has so far demonstrated the existence of analogous behaviour in other species. P. micans is unique in having a fertilization tube that permits the transfer of all the genetic material (♂ gametic nucleus) as in bacteria, and in the formation of a partition between the empty sexual cell and the dikaryon. Therefore this species appears extremely primitive and our present observations support the hypothesis of an early branching of the dinoflagellate from the eukaryotic lineage (Loeblich, 1976; Taylor, 1980; Herzog et al. 1984).

We are indebted to Michel Herzog for helpful discussions, to Danielle Saint-Hilaire and Marie Albert for their excellent technical assistance, to Marie-José Bodiou for making the drawing, and to Suzanne Miller (Wigh Scientific London) and S. von Boletsky for correction of the manuscript.