- Top of page
- Materials and Methods
The most striking finding from the present study is that male gamete structures of Equisetum arvense and E. hyemale (Duckett & Bell, 1977) are remarkably similar. Such conservation in gamete structure parallels that among the leptosporangiate ferns Pteridium, Onoclea, and Ceratopteris, but stands in stark contrast to the immense variation in the Lycopsida (Duckett, 1975; Kotenko, 1990; Maden et al., 1996, 1997; Renzaglia et al., 1998; Renzaglia & Garbary, 2001). The primary difference between the two horsetail gametes is one of size: in E. hyemale the cell extends 3.25 revolutions and the locomotory apparatus occupies 2.5 anterior gyres, compared with a total helix of 2.5 with 1.75 anterior gyres occupied by the locomotory apparatus in Equisetum arvense. The larger cell of E. hyemale is correlated with a greater number of flagella, which in our study was based on maximum counts for both species in SEM images. Spermatozoids of E. hyemale possess at least 80 flagella that are arranged singly at the cell anterior and increase up to six rows of flagella toward the posterior. In comparison, Equisetum arvense gametes contain at least 54 flagella that are similarly arranged around the cell anterior, with up to four rows of flagella near the posterior.
Nuclear shape is much the same in both species, but with a slightly more swollen mid-region in E. hyemale. The general nuclear shape in Equisetum gametes conforms to that of other pteridophytes in that it is narrow at the anterior end, broad in the middle and then abruptly tapers to a narrow posterior. However, as noted above, the mid-nuclear region is broader in Equisetum than in ferns. The nuclei of both horsetail spermatozoids contain peripheral dense inclusions but the aggregation of membrane-bound vesicles in nuclei of E. hyemale were not observed in Equisetum arvense.
The longer and broader spermatozoid nucleus in E. hyemale correlates with larger chromosomes in the subgenus Hippochaete (Manton, 1950). In contrast, published data on amounts of DNA suggest that these are higher in Equisetum arvense (28.4 pg) (Grime et al. 1988) than in E. hyemale (24.5 pg) (Bennett & Leitch, 2001). However, the most recent determinations, using improved and more reliable protocols, indicate a significantly higher amount of DNA in E. hyemale than in Equisetum arvense (M. D. Bennett & I. J. Leitch, unpublished) in line with the karyological and spermatozoid data. Manton (1950) suggests that the larger chromosomes in Hippochaete are indicative of a long-standing subgeneric division, a notion borne out by the occurrence of numerous intrasubgeneric but no genuine intersubgeneric hybrids (Duckett & Page, 1975; Duckett, 1975). However, as with traditional morphology, neither sperm dimensions nor chromosome sizes, nor amounts of DNA shed any new light on which is the more primitive subgenus.
The present study reveals a number of unique membrane-bound organelles associated with the locomotory apparatus and nuclear envelope that were highlighted by the osmium-ferrocyanide fixation and overlooked in earlier observations of E. hyemale. Most striking are the extensive tubular to flattened cisternae containing an electron-opaque material and the similar dense material filling the perinuclear space. Because these membrane arrays are continuous with the nuclear envelope extending on either side of the spline up to the lamellar strip, a likely interpretation is that they are a repository for excess nuclear envelope that is eliminated during nuclear metamorphosis. Careful scrutiny of the published micrographs of E. hyemale also indicates their presence, although they lack the dense contents produced by osmium-ferrocyanide fixation. This protocol also illustrates the lack of pores in the nuclear envelope immediately adjacent to the spline, a feature that appears to be ubiquitous in archegoniate sperm cells (Myles & Hepler, 1977; Myles et al., 1978). The finger-like cisternae associated with the lamellar strip in Equisetum arvense are almost certainly homologous to the sheets of similarly located endoplasmic reticulum (ER) in E. hyemale. During fertilization in ferns, major changes occur in the structures associated with the lamellar strip (Duckett & Bell, 1972; Fasciati et al., 1994a; Fasciati et al., 1994b). The possibility that these organelles in Equisetum contain macromolecules involved in fertilization invites further investigation.
In gross architecture, spermatozoids of Equisetum, Psilotum and ferns are readily distinguished from those of lycophytes and bryophytes in that they contain multiple flagella (36 or more) and numerous plastids and mitochondria (Renzaglia & Garbary, 2001). Ultrastructural studies of Psilotum sperm cells reveal a locomotory apparatus similar in many respects to those of Equisetum and ferns (Renzaglia et al., 2001). Comparisons among Equisetum, Psilotum and leptosporangiate ferns uncover common behavioural and structural features of their spermatozoids. In these cells, the longitudinal axis of the coils (a line running down the centre of the coils) is parallel, to the direction in which the cells swim (Bilderback et al., 1973). In contrast, the direction of movement in bryophytes and Selaginella gametes is perpendicular to the longitudinal axis of the coils. Within the Equisetum-fern assemblage, the subcellular basis for this swimming orientation lies in the highly consistent spatial organization of individual components of the locomotory apparatus. The lamellar strip is a long narrow ribbon that wraps around the anterior coils and, unlike in bryophytes, it reaches its maximum structural complexity in the mature cell. Constituent lamellae are aligned more or less parallel to the longitudinal axis of the lamellar strip. The band of parallel microtubules that constitute the spline is oriented at a 25–45° angle to the plates of the LS. This angle determines the pitch of the helix and, together with cell size, fixes the number of revolutions in the fully differentiated cell. Basal bodies overlie the MLS along the anterior coils and are roughly lined up in parallel with spline MTs; their constituent axonemes are directed slightly obliquely toward the cell posterior. This alignment establishes the direction of motility (Bilderback et al., 1973). In addition, the wide spline (150–300 MTs) ensheaths the nucleus (except in Marsilea) and is involved in separation of cellular coils. This latter feature clearly differentiates the male gamete of Psilotum, Equisetum and ferns from those of Lycopodiaceae (Renzaglia & Garbary, 2001; Renzaglia et al., 2001).
Other cellular features unique to Equisetum and ferns include an overlap between a narrow anterior region of the nucleus and the locomotory apparatus. The nucleus extends beyond the locomotory apparatus to the cell posterior, or nearly so. Between the large anterior mitochondrion and nucleus, in the region of overlap, numerous small accessory mitochondria lie beneath the spline MTs. Such accessory mitochondria are absent from Marsilea (Myles & Bell, 1975) and Angiopteris (Renzaglia & Garbary, 2001) spermatozoids. During spermiogenesis, a granular matrix, the so-called amorphous zone (Vaughn et al., 1993), overlies the spline and is interspersed among the basal bodies. This matrix is derived from similar electron-opaque material associated with the reorganizing blepharoplast of the young spermatid. Both matrices contain the calcium-binding protein centrin and both are implicated in nucleation and organization of MTs (Vaughn & Harper, 1998; Klink & Wolniak, 2001). During reorientation and migration of centrioles/basal bodies in the nascent spermatid, the transition region forms on the distal end of the basal bodies. This transition zone contains a stellate pattern that elongates prior to anchoring of the basal body and growth of the flagellar axoneme. As a consequence, unlike the situation in bryophytes and lycophytes, the stellate pattern is positioned internal to a plug of dense material that traverses the base of the flagellar shaft and delimits the boundary of the cell body. The flagella emerge from the cell with a short region of 9 + 0, before they give way to the 9 + 2 axonemal arrangement of MTs (Duckett, 1975).
A variety of electron-opaque bands encircle the anterior rim of multiflagellated gametes of pteridophytes (Renzaglia & Garbary, 2001). Such structures are involved in securing the basal bodies in a fixed position in the cell, thereby determining the angle and direction of divergence of the flagellar axonemes. By nature of their position along the tightly coiled leading edge of these cells, the electron-dense bands ostensibly provide rigid but somewhat elastic structural support that maintains coil integrity at the cell anterior. This idea is supported by the fact that the tighter anterior coils, compared with posterior coils not bordered by such bands, are maintained even after the spermatozoid becomes motile and is released from the confining spherical boundary of the cell in which it develops. In Equisetum and Ophioglossum the coil-delimiting band exhibits a striated substructure (Renzaglia & Garbary, 2001). This microanatomy is reminiscent of the striated roots or rhizoplast of green algal motile cells (Van der Hoek et al., 1995). However, unlike the rhizoplast, which contains centrin (Vaughn & Harper, 1998), the striated band of Equisetum and the osmiophilic crest and ridge of Ceratopteris lack this protein (Vaughn & Renzaglia, 1993; Vaughn et al., 1993; Hoffman & Vaughn, 1995).
Notwithstanding the numerous structural commonalties that unify male gametes of Equisetum and leptosporangiate fern gametes, there are highly distinctive traits that define both groups. The most notable are the general dimensions of the cell, and especially the nuclear shape. Spermatozoids of leptosporangiate ferns are more streamlined than those of Equisetum, that is, in cross-sectional profile, the coils are flattened. In the mid- to posterior region of Equisetum spermatozoids, the prominent nucleus is quite broad. The anterior coils, although they exhibit a similar arrangement of organelles along the spline, contain numerous mitochondria and plastids. In motile gametes, this abundance of cytoplasm ‘fuses’ the anterior coils and prevents complete extension compared with the more protracted posterior coils. Moreover, the present observations clearly demonstrate that, on release of the gamete from the antheridium, organelles along the inner regions of the nucleus compact down along the ventral nuclear surface and become integral components of the motile gamete. In fact, male gametes of Equisetum pass down the archegonial necks into the venter in their entirety (J. G. Duckett, unpublished). These features of Equisetum gametes are even more exaggerated in Psilotum sperm cells, where successive coils are inflexible owing to the occurrence of abundant organelles, especially the massive, central nucleus (Renzaglia et al., 2001). Similarly the compact organization of lycophyte spermatozoids (Renzaglia et al., 2000; Renzaglia & Garbary, 2001) indicates that their entire organelle complement is almost certainly retained during motility. In leptosporangiate ferns, including Marsilea, virtually all extraneous cytoplasm is eliminated in mass during some phase of sperm motility and at the latest at the mouth of the archegonial neck (Bell et al., 1971; Sears, 1980; Fasciati et al., 1994a, 1994b). The anterior coils of fern gametes are therefore completely individualized and, as such, they enable the flexible helix to elongate along its entire length. An accessory band of up to 42 microtubules is a prominent structure that is restricted to leptosporangiate ferns and Angiopteris, and whose presence may be associated with the full-scale elimination of cytoplasm and separation of coils at the cell anterior. This microtubular band overlies the MLS at the cell anterior and provides a rigid framework in which to anchor components of the locomotory apparatus and to maintain integrity of the narrow, ribbon-shaped anterior coils.
Turning to bryophytes, in Sphagnum the spermatozoid plastid is jettisoned during motility (Manton, 1957) and in the liverwort Sphaerocarpos (Diers, 1967) the posterior mitochondrion and the plastid are both lost by the time the male gametes enter the archegonial neck. However, similar loss of organelles in other liverworts and bryalean mosses is unlikely since here the plastid and posterior mitochondrion are compressed against the spline and nucleus, respectively. Regardless of the fate of the posterior mitochondria and spermatozoid plastids, entry of the anterior mitochondrion, which is firmly anchored to the spline, into the egg is almost certainly universal in archegoniates.
These considerations of the fate of spermatozoid organelles during motility have major implications for the possible inheritance of paternal plastid and mitochondrial genomes. In turn, phylogenetic reconstructions based on organellar molecular data must take account of whether the inheritance of chloroplast and mitochondrial genomes in a particular group is material or biparental. In ferns Gastony & Yatskievych’s (1992) elegant demonstration of purely maternal inheritance of the chloroplast and mitochondrial genomes from sequence data is completely in line with cytological studies showing exclusion of the sperm plastids in the archegonial necks and rapid digestion of all the paternal mitochondria in the egg cytoplasm (Duckett & Bell, 1971, 1972; Fasciati et al., 1994a,b). However, Gastony and Yakskievych do not rule out the possibility of some leakiness in maternal inheritance, as is well documented in angiosperms (Sears, 1980; Medgyesy et al., 1986). It now becomes pertinent to explore paternal organelle inheritance in groups with (Equisetum, Psilotum, lycophytes, bryalean mosses and hornworts) and without (Sphagnum, hepatics and ferns) a complete male gamete complement of mitochondria and plastids. For such studies, Equisetum is particularly inviting since intergametophytic crossing is widespread in nature (Soltis et al., 1988) and can be readily performed in culture (Duckett, 1979).
Other features that differentiate the sperm cells of Equisetum from those of ferns are more subtle, subcellular details. These include the complex interconnected endomembrane system in Equisetum discussed earlier, and the prominent nuclear envelope, with a distinct perinuclear matrix. Unlike mature gametes of most archegoniates, but similar to those of Lycopodiaceae, basal body substructure is not modified during development in Equisetum. In ferns, the central cartwheels disappear and the basal bodies become occluded with an electron-opaque substance similar to the material of the amorphous zone (Vaughn et al., 1993). Finally, large, abundant nuclear inclusions, as occur in Equisetum, have not been described in ferns but find counterparts in the nuclei of Psilotum (Renzaglia et al., 2001) and Lycopodiaceae (Maden et al., 1996, 1997). These inclusions are also absent in bryophytes, Selaginella and Isoëtes, and as far as we know have not been described in animal sperm. Further studies are now required to ascertain whether these inclusions reflect unusual forms of DNA packing perhaps related to the high chromosome numbers in these plants.
It has long been recognized that the angle between the LS plates and longitudinal axes of the spline MTs is instrumental in establishing the layout of the cellular coils (Bell et al., 1971). In ferns, this angle typically ranges from 35 to 45° (Duckett, 1975), in bryophytes it is a constant 45° (Renzaglia & Duckett, 1988) and in lycophytes the angle is highly variable, ranging from 45° to 90° (Maden et al., 1996, 1997; Renzaglia & Garbary, 2001). Based on the present study of Equisetum arvense and after careful evaluation of the micrographs in Duckett (1973), it appears that the spline–LS angle in Equisetum deviates from those reported in all other archegoniates in that it ranges from 23° to 43°. Among archegoniates, such acute angles have been reported only in Psilotum and Equisetum and may well be diagnostic of spermatozoids of basal groups in the fern clade (Renzaglia et al., 2001).
Consistent with numerous phylogenetic analyses based on spermatogenesis (Garbary et al., 1993; Maden et al., 1997; Renzaglia et al., 2000; Renzaglia & Garbary, 2001), the present study supports a sphenopsid-fern clade (Moniliformoses of Kenrick & Crane, 1997). This relationship was strongly reinforced by a series of molecular analyses (Duff & Nickrent, 1999; Nickrent et al., 2000; Renzaglia et al., 2000), culminating in the most recent combined multigene and morphological analysis by Pryer et al. (2001). Conservation of the detailed architectural features among male gametes in Equisetum and ferns points to the utility of characters relating to spermatogenesis in inferring phylogenetic relationships among plants.
We now propose that comparative data from spermatogenesis can provide additional insights into phylogenetic events during land-plant evolution. Diversity among male gamete structure in lycophytes may be interpreted as indicative of deep (most likely Palaeozoic) divergences in this group and between homosporous and heterosporous ferns. By contrast, the remarkable similarities in male gamete architecture between the two subgenera of Equisetum suggest a much more recent subgeneric split, though the genus itself is almost certainly Palaeozoic. This is clearly in line with the northern hemisphere rather than Gondwanaland distribution of Equisetum. Bryophytes present a similar picture. Here highly divergent blepharoplast architecture within the Metzgeriales, a group anchored in the southern hemisphere (Renzaglia, 1982), underlines wide, and most likely very ancient, generic discontinuities, whereas uniformity in the blepharoplasts of hornworts (again a most ancient group) suggests recent origins for the genera (Ranzaglia & Duckett, 1991). These broad summations clearly call for an expansion and new directions in morphological and molecular analyses to include sampling of multiple taxa in key plant groups.