Morphogenesis of the fibrous sheath in the marsupial spermatozoon

Abstract

The spermatozoon fibrous sheath contains longitudinal columns and circumferential ribs. It surrounds the axoneme of the principal piece of the mammalian sperm tail, and may be important in sperm stability and motility. Here we describe its assembly during spermiogenesis in a marsupial, the brush-tail possum, and compare its structural organization with that of eutherian mammals, birds and reptiles. Transmission electron microscopy showed that possum fibrous sheath assembly is a multistep process extending in a distal-to-proximal direction along the axoneme from steps 4 to 14 of spermiogenesis. For the most part, assembly of the longitudinal columns occurs before that of the circumferential ribs. Immunohistochemical and immunogold labelling showed that fibrous sheath proteins are first present in the spermatid cytoplasm; at least some of the proteins of the sheath precursors differ from those in the mature fibrous sheath. That immunoreactivity develops after initiation of chromatin condensation suggests that fibrous sheath proteins, or their mRNAs, are stored within the spermatid cytoplasmic lobule prior to their assembly along the axoneme. These findings are similar to those in laboratory rats, and thus suggests that the mode of fibrous sheath assembly evolved in a common ancestor over 125 million years ago, prior to the divergence of marsupial and eutherian lineages.

Introduction

In the flagellum of the spermatozoon of eutherian mammals, the fibrous sheath surrounds the axoneme and outer dense fibres of the principal piece. It is composed of two structurally distinct segments, the dorsal and ventral longitudinal columns, and a connecting array of circumferential ribs (Fawcett, 1970, 1975). For many years, the fibrous sheath was viewed as a passive, mechanical component of the sperm flagellum (Fawcett, 1975) that provides elastic rigidity to the sperm tail and/or defines the shape of its beat by placing a constraint on its plane of bending (Phillips, 1972; Fawcett, 1975; Lindemann et al. 1992; Si & Okuno, 1993; Carrera et al. 1994; Jassim, 1995). However, as some of its individual proteins have become isolated and sequenced, it has become increasingly clear that this structure also acts as a scaffold for constituents of signalling cascade events (Carrera et al. 1994, 1996; Fulcher et al. 1995; Vijayaraghavan et al. 1997, 1999; Miki et al. 2002), as well as for glycolytic enzymes (Mori et al. 1992; Welch et al. 1992, 1995). Consequently, in addition to a structural function, the eutherian fibrous sheath appears to have an important role in sperm motility (for a review see Eddy et al. 2003). Furthermore, comparative studies show that a fibrous sheath-like structure is present in sperm tails of reptiles (Harding et al. 1995; Scheltinga et al. 2001) and non-passerine birds, with it becoming lost during the evolution of passerines (Baccetti & Afzelius, 1976; Jones & Lin, 1993; Lin & Jones, 1993). To date, the extent of conservation of its molecular composition, and function, across Amniota is unknown, although differences in ultrastructure suggest that its morphology and perhaps function vary between species (Fawcett, 1970).

The formation of the fibrous sheath in eutherian mammals has, to date, only been described in detail for the laboratory rat (Irons & Clermont, 1982b) in spite of the interspecific differences in its morphology. Transmission electron microscopy and radioautography have shown that its morphogenesis involves a lengthy, multistep process extending from steps 2 to 17 of spermiogenesis. Moreover, in contrast to the formation of the outer dense fibres, in the laboratory rat fibrous sheath morphogenesis has been shown to proceed in a distal-to-proximal direction (Irons & Clermont, 1982a) with its proteins being evident within the spermatid cytoplasm prior to their assembly along the axoneme (Oko & Clermont, 1988; Clermont et al. 1990).

Little is known of fibrous sheath morphogenesis in marsupial sperm. Sapsford et al. (1969, 1970) showed that longitudinal column precursors occur in early spermatids in the long-nosed bandicoot (Perameles nasuta) (Peramelidae), but their morphogenesis does not appear to have been studied in any other marsupial. The bandicoots diverged from the possums (Phalangeridae) early in evolution of the Australian marsupials at least 50 million years ago (Kirsch et al. 1997; Amrine-Madsen et al. 2003; Asher et al. 2004; Cardillo et al. 2004; Nilsson et al. 2004). In the brush-tail possum (Trichosurus vulpecula) the number and molecular weights of the major fibrous sheath proteins have recently been found to differ considerably from those of the laboratory rat (Harris & Rodger, 1998; Ricci & Breed, 2001). Although several ultrastructural studies of possum spermiogenesis have already been carried out (Harding et al. 1976; Lin et al. 2004), little data are available on sperm tail morphogenesis. In this study, we describe the assembly of the fibrous sheath in the brush-tail possum and, using polyclonal antibodies for immunocytochemistry, we have determined the time and site of formation of fibrous sheath proteins during spermiogenesis.

Materials and methods

Experimental animals

Adult brush-tail possums housed at the Central Animal House of the University of Adelaide were fed a diet of mixed cereals and fruit daily with water available ad libitum. The University of Adelaide Ethics Committee approved all animal experimentation.

Transmission electron microscopy

Possums were anaesthetized with isofluorane (4%, 1.5 L min⁻¹, DBL, Mulgrave, Victoria, Australia), and the testes fixed by vascular perfusion through the abdominal aorta. The vascular system was first rinsed with heparinized physiological saline containing 2.5% polyvinyl pyrolidone and 0.5% procaine hydrochloride until blood outflow had ceased, followed by perfusion with 3% paraformaldehyde/3% glutaraldehyde in 0.1 m phosphate buffer, pH 7.4, containing 2.5% polyvinyl pyrolidone for 5–10 min or until blood vessels had cleared. The testes were then dissected free, cut into small cubes (∼1 mm³) and immersed in the above fixative for 2 h. Immersion and perfusion-fixed tissue was rinsed in two changes of 0.1 m phosphate buffer, pH 7.4, for 10 min each, and post-fixed in 1% osmium tetroxide in PBS, pH 7.4, for 1 h. Tissue was dehydrated by passing through a graded series of ethanols, cleared in two changes of propylene oxide, and infiltrated in a 1 : 1 ratio of propylene oxide/resin overnight. It was then embedded in pure resin (TK3, TAAB Laboratories, Berkshire, UK) and blocks were polymerized at 60 °C for at least 48 h. Thick (0.5–1.0 μm) plastic sections were cut using a Reichert–Jung Ultracut ultramicrotome with a glass knife and stained with 0.25% toluidine blue in 0.5% sodium tetraborate prior to viewing under an Olympus BH-2 light microscope. Areas of interest were selected from thick plastic sections, the blocks trimmed and ultrathin sections of silver/gold interference colours (0.02–0.1 μm) were then cut with a diamond knife (Diatome Ltd, Bienne, Switzerland). Sections were collected onto copper/palladium grids (200 mesh), stained with uranyl acetate and lead citrate, and viewed with a Phillips CM100 transmission electron microscope at 80 kV.

Polyclonal antibody preparation

The polyclonal anti-possum fibrous sheath serum was prepared as previously described (Ricci & Breed 2001). Briefly, the fibrous sheath was isolated by a combination of sonication, incubation in 4.5 m urea and 25 mm DTT, and sucrose density gradient centrifugation. The fibrous sheaths were then solubilized, emulsified in an equal volume of Freund's complete adjuvant (Sigma, St. Louis, MO, USA), and approximately 50 μg of protein was injected either intraperitoneally or subcutaneously into Sprague–Dawley rats. The rats were boosted with 50 μg of solubilized protein in an equal volume of Freund's incomplete adjuvant at 2-week intervals for 6 weeks, and test bleeds were carried out 7 days after each boost. After collection, the serum was stored at −70 °C. Pre-immune sera were collected from all rats prior to immunization.

Light microscope immunohistochemistry

Adult male possums were anaesthetized with isoflurane, and testes and epididymides were fixed by perfusion for 20 min with Bouin's fixative. Testes were excised, cut into 1-m³ cubes and immersion-fixed in Bouin's fixative for 2 h. Then tissue was cut into smaller, 5-mm³ cubes, re-immersed in fixative for 2 h and dehydrated by passing through a graded series of ethanols. Tissue was infiltrated with paraffin wax, and 5-μm sections were cut on a Leica microtome and floated onto slides. The remaining steps in this protocol were adapted from the protocol of Oko & Clermont (1989). Deparaffinization of these sections followed standard procedures except that during hydration (i) sections were immersed in 70% ethanol containing 1% (w/v) lithium carbonate to inactivate residual picric acid, and (ii) endogenous peroxidase activity was eliminated by incubating sections in methanol containing 1% (v/v) hydrogen peroxide for 10 min. Hydrated sections were subsequently washed for 5 min in distilled H₂O containing 300 mm glycine to block any free aldehyde groups, then rinsed in 20 mm Tris/HCl saline (TBS), pH 7.4.

Non-specific binding of IgG was blocked by incubating sections in 10% normal goat serum (NGS) in TBS for 30 min. Sections were then incubated for 1 h with the primary antibody diluted 1 : 100 in TBS-Tween containing 1% NGS, and washed six times for 5 min each in TBS-Tween containing 1% NGS, blocked with 10% NGS, before incubation with biotin-labelled goat anti-rat immunoglobulin (Sigma) diluted 1 : 25 in TBS for 1 h. The sections were washed in TBS-Tween and incubated in streptavidin-alkaline phosphatase (SA-AP) diluted 1 : 3000 in TBS for 1 h, washed again, and antibody binding visualized by incubating sections with diaminobenzidene tetrahydrochloride (Zymed, San Francisco, CA, USA). Finally, sections were lightly counterstained with haematoxylin, dehydrated by passing through a graded series of ethanols and mounted under glass coverslips with DPX mounting medium.

On all control slides pre-immune sera were used in place of primary antibody. Immunostaining was repeated on sections of testes from four different animals for reproducibility.

Immunogold electron microscopy

Possums were anaesthetized with isoflurane, and testes and epididymides were fixed by perfusion for 20 min with 4% paraformaldehyde in 0.2 m PBS, pH 7.4. Tissue was then dissected free and immersed in the same fixative for 2 h, washed in two changes of PBS, and dehydrated by passing through a graded series of ethanols. The tissue was infiltrated in a 2 : 1 ratio of LR White/70% ethanol for 1 h, followed by two further incubations in pure LR White resin for 1 h and then overnight. The resin was again replaced just prior to polymerization in gelatin capsules at 60 °C for 24 h. Ultrathin sections were cut on a Reichert–Jung ultracut microtome and mounted on nickel grids.

Sections on grids were blocked by incubating in 10% NGS in TBS for 30 min, and then incubated with the primary antibody diluted 1 : 100 in TBS-Tween containing 1% NGS for 1 h. They were washed six times for 5 min each in TBS-Tween, and further blocked in 10% NGS for 15 min before being incubated in 10-nm gold-conjugated goat anti-rat immunoglobulin (Sigma), diluted 1 : 20, for 1 h at room temperature. They were then subject to six 5-min washes in TBS-Tween, four washes in double distilled water, before staining with uranyl acetate (3 min) and examination on a Phillips CM100 transmission electron microscope at 80 kV.

Results

Formation of possum fibrous sheath

Possum fibrous sheath morphogenesis is described using the ten stages of the seminiferous epithelium cycle and the 14 steps of spermatid development as described by Lin et al. (2004).

During steps 1–3 of spermiogenesis, the axoneme developed from the distal centriole but a precursor to the fibrous sheath was not yet visible. Sheath assembly first began in step 4 spermatids at the distal end of the flagellum. Electron-dense material accumulated there between the plasmalemma and microtubule doublets 3 and 8 of the axoneme, forming the precursors of the longitudinal columns (Fig. 1a). In step 5 spermatids, the precursors of the longitudinal columns had lengthened in a proximal direction along the flagellum. In step 6 spermatids, considerable electron-dense material had accumulated beneath the plasmalemma in the distal segment of the flagellum attached to the margins of each of the thickened column anlagen. In longitudinal sections, this material presented as a series of parallel, evenly spaced, circumferentially orientated striations and therefore are presumably the precursors of the circumferential ribs (Fig. 1b,c).

Fig. 1.

Transverse (a,f) and longitudinal (b–e,g) sections through possum spermatids showing the development of the fibrous sheath (FS). The anlagen of the longitudinal columns (ALC) first appear in step 4 spermatids adjacent to microtubules 3 and 8 of the axoneme (a), whereas the anlagen of the circumferential ribs (RA) first appear in step 6 spermatids as a series of evenly spaced striations (b,c). The rib anlagen thicken and coalesce during steps 8 (d) and 9 (e) to form mature ribs (R). During step 11, the ribs project away from the axoneme and the longitudinal columns (LC) increase in size (f). The ribs coalesce and thicken further as spermiogenesis progresses, although electron-lucent spaces can be seen in the columns in step 13 spermatids. Scale bars: (a) 0.1 μm, (b) 0.3 μm, (c) 0.03 μm, (d–f) 0.25 μm, (g) 0.3 μm.

In step 8 spermatids, electron-dense material accumulated between adjacent pairs of the rib precursors (Fig. 1d), which subsequently thickened and began to resemble the ribs in mature sperm. During step 9, these coalescenced further to form the larger, definitive, ribs of the fibrous sheath (Fig. 1e). The number of rib precursors that fused at this stage ranged from 2–3 in the most distal segment of the developing principal piece to 6–12 in the most proximal segment where the developing fibrous sheath was therefore considerably larger. These ribs were united at each of their ends to the longitudinal columns, which simultaneously thickened and grew in a distal-to-proximal direction.

During step 11, the longitudinal columns appeared considerably larger, although the newly formed circumferential ribs were conspicuously electron lucent, perhaps reflecting their early state of development (Fig. 1f), and projected away from the axoneme to give the flagellum an ovoid cross-sectional shape. During steps 12 and 13, as the ribs coalesced and thickened, small, electron-lucent spaces appeared within the longitudinal columns of the fibrous sheath (Fig. 1g). The final step of spermiogenesis was characterized by a gradual increase in size of the fibrous sheath.

Light microscope immunocytochemistry

Immunohistochemical staining of the seminiferous tubules with anti-possum fibrous sheath serum was variable according to the stages, and was step-specific. No reactivity was detected in any of the early, round spermatids (steps 1–6) (Fig. 2a–f), but the cytoplasm and developing flagella of elongating spermatids and spermatozoa were immunoreactive (steps 7–11) (Fig. 2a–c,g–i). Faint immunostaining was first detected within the cytoplasm of step 7/8 spermatids (Fig. 2g); however, labelling rapidly increased in intensity in this region during steps 9 and 10, during which time the flagellum first appeared immunoreactive (Fig. 2h,i, arrows). Immunostaining was generally uniform throughout the cytoplasm of these spermatids. It peaked in step 11 cells (Fig. 2a), whose entire flagellum was stained, to remain elevated during step 12 (Fig. 2b), and diminished rapidly in the remaining step. In comparison, immunostaining of the spermatid flagellum, which began at step 9, increased steadily during steps 12–13 (Fig. 2b,c), to reach a peak in step 14 spermatozoa (Fig. 2d). There was some minor staining of residual material within the seminiferous tubules during Stage IV (Fig. 2d arrow). No immunostaining was detected in any tubules incubated in pre-immune sera.

Fig. 2.

(a–i) Possum testicular sections immunostained with anti-possum fibrous sheath serum. The stages (I–X, top right corner of figures) of the cycle of seminiferous epithelium, and spermatid steps (1–14, top left corner of figures) of spermiogenesis are indicated. Spermatid cytoplasmic staining extends from steps 7 to 14 of spermiogenesis but peaks at step 11. Staining of the spermatid flagellum is first evident during step 9 (arrowheads) and peaks in the final step of spermiogenesis. Round spermatid (*). Scale bars: 12 μm.

Immunogold electron microscopy

No immunogold labelling was detected in step 1–5 spermatids. During steps 6–10 there was no, or only very minor, non-specific labelling over the anlagen of the longitudinal columns and circumferential ribs (Fig. 3a–c). By contrast, an increased labelling over the spermatid cytoplasm during this period reached a peak in step 11 cells (Fig. 3b,d). In step 12 spermatids, there was a marked increase in immunogold labelling over the abaxial cytoplasm, especially in the proximal segment of the flagellum (Fig. 3e). From step 11, specific immunogold labelling was seen over both the longitudinal columns and the circumferential ribs of the fibrous sheath itself, this labelling peaking at step 14 (Fig. 3f,g). No immunogold labelling was evident in any sections incubated in pre-immune sera.

Fig. 3.

Transverse (a,e,g) and longitudinal (b–d,f) sections through possum spermatids treated with anti-possum fibrous sheath serum. These show the development of the fibrous sheath. Immunogold labelling of the flagellum does not occur or occurs very sparsely in step 6 (a,b) and step 8 (c) spermatids. However, the cytoplasm lobule is reactive from step 6 (b) to step 11 (d) (arrowheads). In step 12 spermatids (e) there is intense immunogold labelling beneath the plasmalemma (arrowheads), suggesting that these proteins migrate down the flagellum before becoming incorporated into the fibrous sheath (FS). Immunogold labelling over the fibrous sheath reaches a peak in step 14 spermatids (f,g). Anlagen of the longitudinal columns (ALC), anlagen of the circumferential ribs (RA), immunogold labelling (arrowheads). Scale bars: (a,c) 0.1 μm, (b,d,f) 0.15 μm, (e) 0.06 μm, (g) 0.2 μm.

Discussion

In the current study, the morphogenesis of the fibrous sheath is described for sperm of the brush-tail possum. Transmission electron microscopy has shown that this sheath is morphologically similar to that of the long-nosed bandicoot (Sapsford et al. 1969, 1970) and laboratory rat (Irons & Clermont, 1982b) in spite of differences in its final form. In these three species it has been found that (1) the precursors of the longitudinal columns of the fibrous sheath first assemble adjacent to axoneme microtubules 3 and 8 in early spermatids, (2) the precursors of the ribs appear later as circumferential striations that gradually enlarge and coalesce, (3) growth of the columns and ribs is largely independent of each other except for late in spermiogenesis when their development occurs concurrently, and (4) fibrous sheath assembly takes place in a distal-to-proximal direction along the sperm tail.

One aim of the current study was to differentiate, by immunohistochemistry, the time and site of formation of the components of the fibrous sheath using a possum fibrous sheath polyclonal antibody. This antiserum has previously been shown, by Western blotting, to label specifically the major possum fibrous sheath proteins, but not any others, including those of the outer dense fibres, thus suggesting that the sheath is composed of a unique suite of proteins (Ricci & Breed, 2001). Furthermore, this serum also recognizes fibrous sheath proteins from species in several other marsupial families (e.g. tammar wallaby: Macropodidae; koala: Phascolarctidae; and fat-tailed dunnart: Dasyuridae), which diverged from the Phalangeridae at least 40–50 million years ago (Kirsch et al. 1997; Asher et al. 2004; Nilsson et al. 2004), thus suggesting an early origin of some of the fibrous sheath proteins. In addition, the antibody also stained some proteins in the fibrous sheath of rat sperm (Ricci & Breed, 2001), indicating their conservation across marsupial and eutherian lineages. The immunolabelling data of the current study indicate that the timing of fibrous sheath formation is, like its method of assembly, similar in marsupials and eutherians [for data on the laboratory rat see Oko (1988), Oko & Clermont (1989) and Clermont et al. (1990)]. In the possum, peak cytoplasmic immunoreactivity occurs in step 10 spermatids, well after chromatin condensation has become initiated (Lin et al. 2004). Thus, as in rats (Oko, 1988; Oko & Clermont, 1989; Clermont et al. 1990; El-Alfy et al. 1999), some of the fibrous sheath proteins are translationally regulated. The fact that there is marked immunolabelling of the periaxonemal cytoplasm of step 11 possum spermatids makes it likely that fibrous sheath proteins are transferred from the spermatid cytoplasmic lobule to the developing fibrous sheath late in spermiogenesis.

The proteins of the anlagen of the longitudinal columns and circumferential ribs of possum spermatids are not labelled with the anti-fibrous sheath serum and therefore are presumably dissimilar to those of the mature fibrous sheath. Either the precursor structure is composed of a different suite of proteins from those of the mature fibrous sheath or these proteins are processed during morphogenesis. Immunocytochemical analysis of fibrous sheath formation in the laboratory rat (Oko & Clermont, 1989; Clermont et al. 1990), mouse (Sakai et al. 1986; Fenderson et al. 1988; Brown et al. 2003), human (Jassim et al. 1991) and the cockerel (Bozkurt & Holley, 1995) has yielded similar results, suggesting that the primitive column and rib anlagen might somehow act as organizers, or triggers, of fibrous sheath assembly (Oko, 1988; Clermont et al. 1990) in both mammals and birds. In the mouse, it has been suggested that a fibrous sheath protein A-Kinase Anchoring Protein 3 (AKAP3), which is synthesized in round spermatids and incorporated into the fibrous sheath concurrently with formation of the rib anlagen, may be responsible for organizing the basic structure of the fibrous sheath (Brown et al. 2003).

The current study indicates that the morphogenesis of the fibrous sheath has been conserved across the two major extant mammalian subclasses, despite their divergence over 125 million years ago (Cifelli & Davis, 2003; Woodburne et al. 2003; Asher et al. 2004; Nilsson et al. 2004). Taken together with our previous conclusion that the 62- and 76-kDa fibrous sheath proteins have been conserved between sperm of possum, koala, wallaby and dunnart, as well as laboratory rat (Ricci & Breed, 2001), this suggests that an increased complexity of the sperm flagellum evolved in a common ancestor prior to the divergence of the eutherian and marsupial lineages.

In mammals, the thickness of the sheath's circumferential ribs and the prominence of its longitudinal columns varies between species. For example, in a bat species, Myotis lucifugus, the ribs are very broad and flat and the longitudinal columns thin, whereas the Chinese hamster Cricetulus griseus sperm tail displays thin ribs but broad columns (Fawcett, 1970). In both the American opossum, Didelphis virginiana (Fawcett, 1970), and brush-tailed possum, the longitudinal columns are relatively narrow, but the ribs are greatly expanded adjacent to the columns, and there is a pyramidal electron-lucent region adjacent to the longitudinal columns. In the monotreme Ornithorhynchus anatinus, the columns are poorly developed with the sheath being almost entirely a series of helically orientated dense loops (Carrick & Hughes, 1982; Lin & Jones, 2000). An accessory sperm tail structure, closely resembling the mammalian fibrous sheath, also occurs in reptiles (Furieri, 1970; Harding et al. 1995; Ishmail & Dehlawi, 1995; Scheltinga et al. 2001) and non-passerine birds (Baccetti & Afzelius, 1976; Burgess et al. 1991; Jones & Lin, 1993; Lin & Jones, 1993). In non-passerine birds, this structure has been described as an amorphous, irregularly beaded, filamentous mesh (Baccetti & Afzelius, 1976; Thurston & Hess, 1987; Burgess et al. 1991; Jones & Lin, 1993; Lin & Jones, 1993; Jamieson et al. 1995; Jamieson, 1995), whereas in squamates, a thick, rib-like coat extends into the midpiece of the flagellum (Jamieson & Healy, 1992; Harding et al. 1995; Ishmail & Dehlawi, 1995; Scheltinga et al. 2001; Ferreira & Dolder, 2002). The presence in at least some bird and reptile spermatozoa of a sheath-like structure, or an amorphous sheath, suggests a somewhat similar structure to that of the mammalian fibrous sheath. However, whether it is biochemically and/or functionally comparable is unknown. Although there are no specific studies of bird and reptile fibrous sheath morphogenesis, this structure appears early in spermiogenesis in both the quail (Lin & Jones, 1993), and a species of neotropical lizard, Tropidurus torquantas (Vieira et al. 2001), and it assembles in a distal-to-proximal direction along the axoneme in the common lizard Lacerta vivipara (Courtens & Depeiges, 1985).

The presence of a fibrous sheath, or a fibrous sheath-like structure, in spermatozoa from reptiles, birds and mammals, coupled with the current finding that its morphogenesis is similar in eutherian and marsupial mammals, suggests an important function(s) for it in sperm of all amniotes. In the rat and mouse, the most abundant fibrous sheath protein is an A-kinase anchoring protein, AKAP4 (Carrera et al. 1994, 1996; Fulcher et al. 1995) which may anchor cAMP-dependent protein kinase A (Vijayaraghavan et al. 1997; Colledge & Scott, 1999; Miki et al. 2002; for a review see Eddy et al. 2003). The molecular weight of AKAP4 is 76 kDa and, because possum fibrous sheath extracts have a major protein of similar molecular weight (Ricci & Breed, 2001), it too may contain much AKAP4. In addition, using both immunofluorescence and immunogold microscopy, we have recently found that an antibody raised against mouse spermatogenic glycolytic pathway enzyme glyceraldehyde 3-phosphate dehydrogenase (GAPDS) (Fenderson et al. 1988; Welch et al. 1992; Bunch et al. 1998) binds intensely to the possum fibrous sheath by both immunofluorescence and immunogold microscopy (M.R., M. Eddy and W.G.B., unpublished observations). Thus, in the possum GAPDS also appears to be a major fibrous sheath protein and may well provide its sperm with an energy source. A more detailed investigation of the molecular biology of the marsupial fibrous sheath may shed light on the proteins present in this sperm tail cytoskeletal structure. Such findings will further extend our understanding of the evolution of this accessory cytoskeletal structure of the sperm tail that appears to have evolved around the time of the evolution of internal fertilization and viviparity in early amniotes.