Unique features of spermiogenesis in the Musky Rat-kangaroo: reflection of a basal lineage or a distinct fertilization process?


S. Lloyd, School of Veterinary Science, University of Queensland, St Lucia 4072, Qld. Australia. T: +61 733652580; F: +61 733651255; E: Shan.Lloyd@uq.edu.au


Previous research has found the mature spermatozoon of the Musky Rat-kangaroo to share many characteristics with other macopodoids, some phalangeroids and even the dasyurids. While there have been several studies published on the ultrastructure of the mature marsupial spermatozoon, there are only a few detailed studies on marsupial spermatogenesis. Furthermore, there have been no studies undertaken which combine the staging of the epithelial cell cycle with transmission electron microscopy to describe the ultrastructural changes in the developing spermatozoon during these stages. Such studies have the potential to be used in determining the required time taken for certain components of the spermatozoa to develop. During this study, eight stages of the seminiferous epithelium were observed and the ultrastructure of spermatogenesis was divided into nine phases. Maturational processes in the epididymides are also described. Among the features reported are: the formation of a unique acrosomal granule different from those reported in any other marsupial, the absence of contraction of the nuclear ring, a conspicuous acrosomal compaction process despite the almost 100% coverage of the dorsal nuclear surface and the retention of late spermatids within the seminiferous tubules until the early spermatids have developed to the nuclear protrusion phase.


Comparative sperm morphology has often been used successfully in phylogenetic analyses (Rouse & Robson, 1986; Harding, 1987; Harding et al. 1987; Temple-Smith, 1987; Woodall, 1991; Jamieson, 1995; Johnston et al. 1995; Teixeira et al. 1999). A set of identifying characters was suggested for each marsupial group by Hughes (1965) and Harding et al. (1979) and others authors have since found these to be consistent (Harding et al. 1982; Harding, 1987; Kim et al. 1987; Harding et al. 1990; Taggart et al. 1995). Sperm morphology has supported the phylogenetic separation of several species such as Tarsipes rostratus, (Harding et al. 1984) Dromiciops gliroides (Temple-Smith, 1987) and Lagorchestes hirsutus (Johnston et al. 2003). Lloyd et al. (2002) examined the mature sperm ultrastructure of the Musky Rat-kangaroo Hypsiprymnodon moschatus and found there was an amalgam of traits from the marcopodoids, phalangeroids and dasyuroids which they believed reflected the relatedness of the Musky Rat-kangaroo to all of these groups.

Despite the ultrastructure of mature spermatozoa having been described in several marsupial species, there are relatively few detailed ultrastructural studies on marsupial spermatid differentiation, as previously noted by Lin et al. (1997). Though it is really a continuous process, those authors who have assessed the ultrastructural development of the marsupial spermatid in detail have broken down the process into arbitrary phases (Sapsford et al. 1967, 1969; Kim et al. 1987; Lin et al. 1997).

Spermatogenesis occurs in a regular pattern in each mammalian species, with the same cellular associations occurring sequentially and repeatedly. Again, though spermatogenesis is really continuous, for convenience it is classified into identifiable ‘stages’ by most authors (Roosen-Runge & Giesel, 1950; Leblond & Clermont, 1952a; Setchell & Carrick, 1973; Orsi & Ferreira, 1978; Tait & Johnson, 1982; Peirce & Breed, 1987; Patil & Saidapur, 1991; Saidapur & Patil, 1992; Bilaspuri & Kaur, 1994; Lin et al. 2004). Although few marsupial species have been studied, stages appear broadly consistent with eutherians (Setchell & Carrick, 1973; Orsi & Ferreira, 1978; Lin et al. 2004; Ricci & Breed, 2005).

Orsi & Ferreira (1978) found that the discrete acrosome of marsupials did not stain well with the PAS-Haematoxylin technique used by Leblond & Clermont (1952b). Therefore the research reported here utilizes the staging method developed by Roosen-Runge & Giesel (1950), based mainly on the progression of nuclear shape in the spermatids coupled with the stage of development of the other germ cells. This works well for all mammals.

Staging of the seminiferous epithelial cell cycle is usually conducted using light microscopy. Apart from some work on the fibrous sheath by Ricci & Breed (2005), little research has been undertaken to detail the ultrastructural changes which occur in spermatids during successive stages of the seminiferous epithelial cell cycle.

The mature spermatozoa of the Musky Rat-kangaroo displays a distinctive set of traits (Lloyd et al. 2002), including: a particularly streamlined nucleus, an extensive coverage of the nucleus by the acrosome, which includes a button-shaped anterior portion; one of the shortest sperm recorded for marsupials; a midpiece fibre network and mitochondrial arrangement which is similar to the dasyurids; and a mesh-like fibre network surrounding the neck. Many features of mature sperm are variable among marsupial groups but some of the most striking differences actually occur during spermatozoal development, such as the acrosomal process and double rotation of the head in phalangeroids and macropodoids.

This research was undertaken to provide a comprehensive, well-illustrated account of the ultrastructural changes occurring during differentiation and epididymal transit of the Musky Rat-kangaroo spermatozoa in order to compare these with other marsupial groups. The cellular associations within the seminiferous tubules and the frequency with which they occur are described as well as, for the first time in a marsupial, the ultrastructural changes occurring in spermatids during each cellular association. This will provide an indication of the relative duration of ultrastructural changes taking place during spermiogenesis.

Materials and methods


Two male Musky Rat-kangaroos were anaesthetized with isofluorane (Forthane, Abbott, Sydney, Australia) inhalation anaesthetic, and a single testis with epididymis was surgically removed from each animal. The tunica vaginalis was cut at two different sites (one from each pole) to reveal the seminiferous tubules and a tissue sample of 2.0 mm diameter was removed from each site. Wedged-shaped, 2.0-mm tissue samples were also taken at four sites along the epididymides. Biopsied tissues were processed as per Lloyd et al. (2002).

The four samples taken along the epididymides were named according to the area of removal and are as follows: caput, caput–corpus, corpus, and corpus cauda. These sample sites reflect the gross anatomical features of the epididymides rather than a functional distribution within the organ.

Stages of the cycle of the seminiferous epithelium

After the removal of the tissue samples for electron microscopy above, the remaining testicular tissue was fixed in 10% neutral formalin, Carnoy's fixative or Bouin's fixative for a minimum of 24 h. Several pieces of testicular tissue approximately 2.0 mm2 in size were removed and processed, embedded, and sectioned at between 4–6 µm and stained in Mayer's Haematoxylin and Eosin. Each stage or cellular association was determined and described while viewing sections directly under the light microscope. To determine the frequency of each stage, a minimum of 100 round tubular cross-sections from each animal were analysed. Sections used for this analysis were at least 1.0 mm apart and were taken from at least four different areas of the testis. Each tubule was photographed using a digital camera attached to the microscope and stored on compact disc. Stages were then determined directly from the microphotograph of the section while on the computer screen. The frequency of each stage was determined by dividing the number of times a particular stage was recorded by the number of tubules analysed. A third animal, a specimen from the Queensland Museum, was also used for this experiment. However, only two tissue samples were utilized to ensure minimum damage to the specimen. Sections utilized for examination were processed as above and were separated by at least 1 mm.

Determining spermatid ultrastructure during each stage of the seminiferous cell cycle

In order to determine the ultrastructural development of spermatids at different stages of spermatogenesis, semi-thin Araldite sections were viewed under the light microscope (Olympus Optical Co. Ltd, Tokyo, Japan) and the stage determined. The associated ultrastructure of the spermatids in each stage could then be determined from thin sections cut from the same block.

Permits and ethics

This research is covered by Permit Nos F1/000052/97/SAA, F1/000052/97/SAB and F1/000052/99/SAA issued by the Queensland Parks and Wildlife Service and 1009, 1134 and 1324 issued by Department of Natural Resources. This experiment is also covered by The University of Queensland Ethical approval certificate numbers ANAT/282/97/PHD/DEH and VPA/372/98/PHD/DEH.


The ultrastructural observations of spermiogenesis are recorded first. For convenience in presenting these results, the continuous process of spermiogenesis has been categorized into nine phases (Figs 1–9). The changes occurring in spermatids during each phase are described, although there may be some overlap of processes between phases. The phases are named according to the main differentiation processes occurring therein and are as follows:

Figure 1.

Pro-acrosomal vacuole. (a) Section through the early spermatid just after spermatocyte division. Early flagellum arrowed. Inset: intercellular bridge between spermatids. (b) Spermatid in a slightly later stage of development than in (a) showing an active Golgi. A, acrosome; AG, acrosomal granule; An, annulus; B, intercellular bridge; G, Golgi apparatus; Gr, granular material; L, lateral junctional body; LC, longitudinal centriole; N, nucleus; NM, nuclear membrane; PaV, pro-acrosomal vacuole; PJB, proximal junctional body; TC, transverse centriole. Scale bars: (a,b) 1 µm; (a, inset) 500 nm.

Figure 2.

Figure 2.

Acrosomal vacuole. (a) Shows a spermatid at the beginning of this phase. Several small vesicles from Golgi have amalgamated to form the acrosomal vacuole. (b) The vacuole begins to indent the nucleus and a granular material is deposited on the acrosomal membrane adjacent to the nucleus. (c–g) As the acrosome enlarges further, a membranous granule, which takes on a variety of forms as illustrated in these micrographs, is evident. (c, inset) Developing neck structures during this phase. See Fig. 1 for abbreviations. Scale bars: (a,c,f), 500 nm; (b,d,e,g), 1 µm, (c, inset), 100 nm.

Figure 2.

Figure 2.

Acrosomal vacuole. (a) Shows a spermatid at the beginning of this phase. Several small vesicles from Golgi have amalgamated to form the acrosomal vacuole. (b) The vacuole begins to indent the nucleus and a granular material is deposited on the acrosomal membrane adjacent to the nucleus. (c–g) As the acrosome enlarges further, a membranous granule, which takes on a variety of forms as illustrated in these micrographs, is evident. (c, inset) Developing neck structures during this phase. See Fig. 1 for abbreviations. Scale bars: (a,c,f), 500 nm; (b,d,e,g), 1 µm, (c, inset), 100 nm.

Figure 3.

Nuclear protrusion. The acrosome has collapsed and the spermatid has taken on a triangular appearance with the nucleus forming the apex. The nuclear membrane is thickened beneath the acrosome, marked by the white arrow and the nuclear membrane is turned in at the acrosomal edge (black arrow). (Inset) This micrograph clearly demonstrates the nuclear, acrosome, plasma and Sertoli cell membranes as well as the electron-dense narrow band (white arrow) between the nuclear and acrosomal membranes. Scale bars: 1 µm, (inset) 100 nm.

Figure 4.

Nuclear flattening. (a) The nucleus has begun to flatten and the neck is clearly set perpendicular to the plane of flattening. The nuclear membrane is thickened along the future ventral surface, containing pores on the side which will become the anterior aspect of the nucleus. (b) This spermatid is at the end of this phase and the anterior aspect of the nucleus has become evident by the thickening of the anterior portion of the acrosome. Scale bars: (a) 500 nm, (b) 1 µm. F, flagellum; M, manchette; NR, nuclear ring; SS, Sertoli cell spurs; TM, thickened nuclear membrane (without pores); TMP, thickened nuclear membrane with pores. For other abbreviations see Fig. 1.

Figure 5.

Early nuclear shaping and condensation. (a) This spermatid is at the beginning of this phase. Most of the chromatin has started to condense but areas of uncondensed chromatin appear ventrally. The neck structures are clearly seen in this micrograph. (b) While the shape of the nucleus of this spermatid has now taken on the form of the mature spermatozoa, much of the nucleus, particularly the newly formed tail, is filled with uncondensed chromatin. A vesicle has formed within the acrosome. (b) Inset A, transverse section through the nucleus shows that chromatin is not condensed at the ventral or lateral margins of the nucleus. Scale bars: (a,b) 1 µm, (b, inset) 500 nm.

Figure 6.

Late nuclear shaping and condensation. Longitudinal section through head and neck. (Inset, upper) Transverse section through the anterior head. (Inset, lower) Neck area of Fig. 6 enlarged to show detail of structure at this stage. Note the anterior acrosome is now arched over the nucleus. C, cementum; SA, sub acrosomal space. For other abbreviations see Figs 1 and 4. Scale bars: 1 µm, (Upper Inset) 1 µm, (Lower inset) 200 nm.

Figure 7.

Rotation and lengthening of the neck. Two important processes occur during this phase and are illustrated in this micrograph. The neck has lengthened considerably with the development of the distal portion of the proximal junctional body and is now almost parallel with the nucleus. (Inset) Neck structures enlarged. Scale bars: 500 nm, (inset) 100 nm.

Figure 8.

Mitochondrial sheath formation. Mitochondria have aggregated around the neck and mid-piece. The acrosome has started to collapse down towards the nucleus. Scale bar: 1 µm.

Figure 9.

Spermiation. The tail has rotated again to leave the neck and head at an angle of about 45°. The spermatid is released from the Sertoli cell cytoplasm, a small droplet remaining above the acrosome. DP, distal portion of the proximal junctional body; Gp, gap; Mi, mitochondria; PP, proximal portion of the proximal junctional body; SC, Sertoli cell cytoplasm; SER, smooth endoplasmic reticulum; St, striated sheath. For other abbreviations see Figs 1–6. Scale bar: 1 µm.

  • 1) pro-acrosomal vacuole
  • 2) acrosomal vacuole
  • 3) nuclear protrusion
  • 4) nuclear flattening
  • 5) early nuclear shaping and condensation
  • 6) late nuclear shaping and condensation
  • 7) rotation and lengthening of the neck
  • 8) mitochondrial sheath formation
  • 9) spermiation.

Light microscope observations on the cellular associations occurring during the seminiferous epithelial cell cycle are recorded next. Eight stages were differentiated which corresponded broadly to those found by Roosen-Runge & Giesel (1950) and are simply named Stages 1 to 8. The relevant ultrastructural phases which occur in each of these stages are given in the description of each stage. It is important to note that the term ‘stage’ always refers to the stage of the seminiferous epithelial cell cycle. The term ‘phase’ refers to the ultrastructural differentiation of the spermatids.

Ultrastructural phases of development

Phase 1: Pro-acrosomal vacuole

The irregular nuclei of the early spermatids progressively became more rounded in outline (Fig. 1a) and adjacent spermatids were clearly joined by intercellular bridges (Fig. 1a inset). Initially, the Golgi apparatus was not prominent but by the end of this phase an active Golgi is positioned adjacent to the nucleus (Fig. 1b). Occasionally, the longitudinal centriole and early cytoplasmic canal or the transverse centriole could be seen close to the plasma membrane (Fig. 1a).

Phase 2: Acrosomal vacuole

Initially several small vacuoles coalesced to form the acrosomal vacuole, which moved to the nucleus and indented the surface (Fig. 2a). Granular material formed discrete clumps in association with the inner acrosomal membrane, particularly where it contacted the nucleus (Fig. 2b). Further vacuoles were added to the pro-acrosomal complex enlarging the acrosomal vacuole until it indented the nuclear surface to such a degree that the nucleus often appeared crescent-shaped (Fig. 2c,d).

The granular material eventually condensed and spread out to form a thin layer lining the vacuole in the region adjacent to the nucleus and here the nuclear membrane was thickened (Fig. 2c,d). Membranous material was a frequent feature in large acrosomal vacuoles and towards the end of acrosomal vacuole formation, a single large membranous ‘acrosomal granule’ was prominent in many sections (Fig. 2c–g). Often this membranous granule was positioned to either side of the acrosomal membrane, which gave the appearance that it was moving across this membrane at the time of collection and fixation of the tissue (Fig. 2c,f,g), but it was difficult to determine whether it would have been moving into, or out of, the acrosome. Some sections at the end of this phase and in the early nuclear protrusion phase, exhibited lamellar bodies in the cytoplasm which resembled the membranous acrosomal inclusions.

During acrosomal formation, some neck structures and the flagellar apparatus could occasionally be observed near the nuclear surface at the opposite pole to the acrosome (Fig. 2c inset).

Phase 3: Nuclear protrusion

The acrosomal vacuole commenced to collapse and eventually formed a cap-like structure over the nucleus, which now lay adjacent to the plasma membrane and was still rounded (Fig. 3). Most of the cytoplasm had moved to what could now be considered the posterior end of the now polarized cell. The acrosomal contents were patchily distributed throughout the acrosome, which was frequently thicker anteriorly than it was laterally (Fig. 3). A sub-acrosomal space occurred between the acrosome and the nucleus and this space was bisected by an electron-dense narrow band (Fig. 3 inset). The nuclear membrane remained thickened beneath the acrosome (white arrow Fig. 3) and the nucleus contained a shallow depression at the lateral edges of the acrosome (black arrow Fig. 3). The manchette tubules appeared during this phase forming a sleeve around the nucleus at the acrosomal margins. Neck structures, most frequently the large lateral junctional body, were seen at the posterior pole of the nucleus and the nuclear envelope in this area was becoming thickened.

Phase 4: Nuclear flattening

The nucleus began to flatten in a plane perpendicular to the flagellum. Depending on the plane of section, the neck structures were seen to be composed of the proximal junctional body, below which were the transverse and then longitudinal centrioles. Surrounding the longitudinal centriole, the striated sheath was developing, whilst to one side of the longitudinal centriole lay the lateral junctional body. The anlagen of the annulus was present at the anterior edge of the flagellar canal and axoneme (Fig. 4a).

The ventral nuclear membrane thickened and contained pores in the section anterior to the implantation fossa (Fig. 4a,b). By the end of this phase, orientation was well established, as the acrosome now consisted of a thickened dorso-cranial portion and a thin flattened plate along the rest of the dorsal nuclear surface (Fig. 4b).

Sertoli cell spurs arose at the margin of the acrosome and a diffuse Sertoli cell reaction arose. The nuclear ring, crescent shaped in cross-section, developed beneath the plasma membrane, not far below the margin of the acrosome, and anterior to the end of the tubules of the manchette (Fig. 4a,b).

Phase 5: Early nuclear shaping and condensation

After initial flattening, spermatid heads rapidly elongated and underwent shaping. The anterior nucleus became dished beneath the thickened section of acrosome and the distal nucleus developed a curve posteriorly (Fig. 5b). The implantation fossa was at an angle and lay anterior, rather than dorsal, to the structures of the neck whilst the nucleus was shaped around the proximal junctional body prior to and during this phase (Fig. 2c inset, and  Fig.5a). The angle between the distal sperm head and flagellum decreased to 45º during this stage (Fig. 5a,b), but full rotation, with the head and flagellum becoming parallel, does not occur until later. Chromatin began to condense as the nucleus started to elongate but not at the anterior ventral or curved distal sections, or at the lateral edges of the nucleus (arrowed in Fig. 5b).

The acrosome, which had developed a large vesicle, extended along most of the dorsal nuclear surface, ending just anterior to the point where the Sertoli cell spur, the edge of the condensed chromatin and the nuclear ring all occurred (Fig. 5b). The nuclear membrane was thickened where it was in contact with the condensed chromatin dorsally, posterio-ventrally and at the implantation fossa. Further posterior to the thickened ventral section of the nuclear membrane, where the chromatin was not condensed, the nuclear membrane was also thickened but was penetrated by pores (Fig. 5b and inset).

Neck structures consisted of the small round proximal junctional body, immediately below which were the transverse and then longitudinal centrioles and the striated sheath; this was surrounded by a new structure, the distal junctional body. The lateral junctional body remained anterior to the other structures and a bar of electron-dense material could sometimes be seen beneath this body (Fig. 5a). Towards the end of this stage, cementum appeared to form between the neck structures and the thickened nuclear membrane.

Phase 6: Late nuclear shaping and condensation

The nucleus continued to elongate and condense whereas the angle of the long axis of the head to the flagellar apparatus remained at 45°. The Sertoli cell spurs and the nuclear ring, now circular in cross-section, reached the posterior and anterior margins of the nucleus. The acrosome extended to the distal edge of the nucleus as a flat plate and anteriorly was projected forward, over the concave tip of the nucleus, leaving a very large subacrosomal space (Fig. 6 and upper inset). The nuclear membrane remained thickened at the implantation fossa and for some distance distally, where pores became evident. The neck structures remained unchanged and cementum lay between both junctional bodies and the thickened nuclear membrane lying dorsally and anteriorly to these structures (Fig. 6 and lower inset).

Phase 7: Rotation and the lengthening of the neck

The condensed nucleus rotated so that the long axes of the head and flagellum became parallel. The shaft of the proximal junctional body and the striated sheath lengthened and the annulus and the flagellar apparatus became distally displaced (Fig. 7). As the proximal junctional body elongated, it appeared to form two parts, the original rounded piece contacting the nucleus and a second longer, narrower distal portion. Further cementum, which now took on a finely banded appearance, formed between the proximal junctional body and adjacent nuclear surfaces and also filled the area between the proximal and distal portions. At the implantation fossa, the outer nuclear membrane was thicker than the inner. The transverse centriole and lateral junctional body ceased to be identifiable during this stage (Fig. 7 inset).

The acrosome remained high over the anterior concave portion of the nucleus, leaving a gap between them. The lateral acrosomal edges were swollen and projected downwards to the nuclear edge as did the anterior portion of the acrosome. Dorsally, the acrosome formed a thin plate-like structure extending between the lateral and anterior swollen edges. Though less pronounced, the swelling of the lateral margins of the acrosome extended distally past the concave section of the nucleus. Areas of uncondensed chromatin remained at the lateral and anterio-ventral edges of the nucleus (Fig. 7).

Phase 8: Mitochondrial sheath formation

Many rounded mitochondria with a fine mostly electron-lucent matrix, now moved into the space surrounding the anterior flagellum below the nucleus. Some mitochondria were found as far forward as the neck structures (Fig. 8). In transverse sections of the anterior mitochondrial sheath, the rounded, dense outer fibres closely overlay the axonemal doublets. Throughout most of this stage the Sertoli cell reaction remained obvious, but the Sertoli cell spurs were less well defined and by the end of this stage neither spurs nor reaction was evident. Without any noticeable contraction, the nuclear ring lost its identity during this stage, usually before the Sertoli cell reaction subsided. The acrosome commenced to collapse back into the concave portion of the nucleus and the swellings bulged out laterally rather than projecting downwards. The acrosome projected forwards over the anterior of the nucleus (Fig. 8).

Phase 9: Spermiation

As the Sertoli cell released the spermatid during spermiation, the head rotated again to form a 45° angle between the distal head and the flagellum. Cytoplasm surrounded the neck and ventral surface of the nucleus but a large vacuole formed beneath most of the nucleus. The transverse centriole had presumably degenerated and was now absent, leaving a gap in the distal portion of the proximal junctional body. The anterior acrosome had settled back down into the concave portion of the nucleus but the edges remained swollen, so that in the concave portion the acrosome was now cup-shaped (Fig. 9).

A cytoplasmic sleeve remained around the mitochondrial sheath but the mitochondria were now arranged in a more ordered fashion. A principal piece fibre network was clearly evident.

Stages of the seminiferous epithelial cell cycle

The following account of the stages of the seminiferous epithelial cell cycle is based on light microscopy. However, in addition to the stage descriptions, the relevant ultrastructural observations for the spermatids pertaining to that specific stage are included.

Stage 1

A generation of spermatids had recently left the tubule (see Stage 8), although large residual bodies remained in the apical cytoplasm of the Sertoli cells. The single generation of remaining spermatids were triangular shaped, with an eccentric, round, partially condensed nucleus which later became slightly elongated. These spermatids were in ‘nuclear protrusion phase’, but towards the end of this stage many had entered ‘early nuclear flattening’. Spermatids usually flattened in a plane parallel to the tubule wall and tails developed in the direction of the lumen. Two distinct layers of spermatocytes occurred, an inner, ‘older’ layer, in pachytene and an outer, ‘younger’ basal layer, which were in either pre-leptotene or leptotene (Fig. 10a).

Figure 10.

Stage 1–4 of the seminiferous epithelial cell cycle. (a) Stage 1, (b) Stage 2, (c) Stage 3, (d) Stage 4. Esp, early spermatocytes; Est, early spermatids; Lsp, late spermatocytes; Lst, late spermatids; M, spermatocytes undergoing meiotic division; S, Sertoli cells; Sp, only layer of spermatocytes; St, only layer of spermatids; R, residual bodies. Scale bar: 40 µm.

Stage 2

Residual bodies were no longer evident and spermatid nuclei were becoming more condensed and elongated. The long axis of the nucleus was usually parallel to the tubule wall. These spermatids were in ‘nuclear flattening phase’. Older spermatocytes entered diplotene and the younger leptotene spermatocytes reached zygotene by the end of Stage 2 (Fig. 10b).

Stage 3

This stage commenced when the older spermatocytes began to divide and ended once all secondary spermatocytes had undergone the second meiotic division to form the earliest undifferentiated form of the spermatid. Spermatids from Stage 2 are further elongated with heads remaining parallel to the tubule wall with their long axis perpendicular to the long axis of the tail. These spermatids were in ‘early nuclear shaping and condensation phase’, thus by light microscopy they began to resemble spermatozoa, taking on the mature sperm shape and becoming more deeply basophilic. The younger spermatocytes remained in zygotene (Fig. 10c).

Stage 4

The spermatids of Stage 3 will now be referred to as ‘late’ spermatids to distinguish them from recently formed ‘early’ spermatids. The late spermatids appeared disorganized, with some heads remaining parallel to the tubule wall while others had formed, or were forming, bunches with their heads perpendicular to the tubule wall and embedded into the Sertoli cell cytoplasm. Early in this stage, spermatids went from ‘early nuclear shaping and condensation phase’ to ‘late nuclear shaping and condensation phase’ and by the end may have developed further to ‘Rotation and lengthening of the neck’. Early spermatids were unspecialized and formed a layer usually four deep. The nuclei were homogeneous and round, whereas the cytoplasm was distributed less evenly so that the cell was not always rounded in outline. Light microscopy did not permit detection of an acrosomal vacuole; however, ultrastructural studies and semi-thin sections revealed that towards the later part of this stage the early spermatids moved from the ‘Pro-acrosomal phase’ to the ‘Acrosomal phase’. There was a single layer of spermatocytes which were either in late zygotene or had developed to pachytene, as well as a few spermatogonia which were large and oval (Fig. 10d).

Stage 5

The heads of the late spermatids had become further elongated and condensed and all now lay perpendicular to the tubule wall. They had formed bunches which had penetrated the surrounding layer of early spermatids and were in ‘Rotation and lengthening of the neck phase’ and entering ‘Mitochondrial sheath formation phase’. The acrosomal vacuoles of the early spermatids could now be seen readily by light microscopy (see above) and were indenting the nuclear surface. The acrosomal granule was visible in electron micrographs and semi-thin sections during this stage. All primary spermatocytes were now in pachytene (Fig. 10e).

Figure 10.

Stage 5–8 of the seminiferous epithelial cell cycle. (e) Stage 5, (f) Stage 6, (g) Stage 7, (h) Stage 8. Est, early spermatids; Lst, late spermatids; R, residual bodies; S, Sertoli cells; Sp, only layer of spermatocytes. Scale bar: 40 µm.

Stage 6

The late spermatids were now seen to line the lumen of the tubule, being fairly evenly distributed rather than inserted in bunches between the early spermatids as in Stage 5. The cytoplasm was mostly distributed posterior to the head, and thus appeared to form a neat rim around the lumen of the tubule. These spermatids were all in the ‘Mitochondrial sheath formation phase’. The early spermatids were still round, with a large acrosome indenting the nucleus and a prominent acrosomal granule visible in electron micrographs. Primary spermatocytes remained in pachytene whilst spermatogonia were usually oval and in some sections were undergoing mitosis (Fig. 10f).

Stage 7

Although the late spermatids still lined the lumen neatly, they again appeared partially embedded among the early spermatid layer similarly to Stage 5 but not as deeply; they remained in the ‘Mitochondrial sheath formation phase’. The tails were discernible in the lumen and the cytoplasm, which was now less voluminous, contained fine, granular clumps of debris. The acrosome of the early spermatids was either unchanged or was beginning to be smaller, flatter and less rounded. Ultrastructural observations verified that the acrosome of early spermatids was collapsing and by the end of this stage some spermatids were entering the ‘Nuclear protrusion phase’. The spermatocytes still remained in pachytene (Fig. 10g).

Stage 8

Several changes occurred during this stage but because the changes in one layer were not consistently seen to occur in synchrony with changes in other layers it was difficult to subdivide this stage further. Late spermatids now lined the lumen and were no longer embedded in deeper layers. Initially, spermatozoal heads were parallel with the tails but later they rotated, becoming perpendicular to the tail and parallel to the tubule wall. The cytoplasmic debris initially found posterior to the sperm heads in Stage 7 moved anterior to the heads where it aggregated, eventually forming several large clumps prior to rotation of the head. These late spermatids moved from the ‘Mitochondrial sheath formation phase’ to ‘Spermiation phase’ during this stage. The acrosome of the early spermatids continued to collapse and the nucleus became slightly smaller, more condensed and eccentrically located within the cytoplasm, causing the spermatid to be triangular in shape, with the nucleus forming the apical point. Electron microscopy confirmed that spermatids were continuing to enter ‘Nuclear protrusion phase’ while the spermatocytes remained in pachytene. In some sections towards the end of Stage 8, many spermatogonia were in mitosis and by the end of this stage and at the beginning of Stage 1 there were many pre-leptotene primary spermatocytes present in tubular cross-sections (Fig. 10h).

Percentage frequency of stages of seminiferous epithelial cell cycle

The total number of tubules recorded in each of the eight stages of the seminiferous epithelial cell cycle for all three males is tabulated in Table 1. The frequency with which each stage occurred is also recorded as a percentage which reflects the relative proportion of time spent in each stage of development.

Table 1.  Number of sections recorded and resulting frequency of stages of the cycle of the seminiferous epithelium
No. of sections in this stage3829236785351252341
Per cent11.1 8.5 6.719.624.910.3 3.515.2 

Ultrastructural development of spermatozoan during epididymal transport

Changes to sperm ultrastructure in the epididymides were less synchronized among spermatozoa than those that occurred in the testis, causing sperm from any one section of epididymidis to be in slightly different stages of maturation. Nonetheless, the majority of spermatozoa in a single section were sufficiently similar to describe the processes occurring in each section of epididymidis.

Caput epididymidis

Within the caput region the angle between the long axes of the neck and head remained at about 45º (Fig. 11a,d). The mitochondrial sheath was curved so the spermatozoa were not particularly streamlined and the mitochondrial cristae were not distinct. There was no evidence of a mid-piece fibre network (Fig. 12i). Spermatozoa were sparse compared to other regions. In some spermatozoa, there were peripheral nuclear extensions which ballooned from the ventral and lateral edges of the nucleus (Figs 11a, 12b). These extensions had not been evident in sperm within the testis except at spermiation.

Figure 11.

Changes in spermatozoan head and neck structures during transit through the epididymides. (a) Longitudinal section through a spermatozoan head in the caput epididymides. Note the nuclear extensions, angle of the neck and head, vacuole, position of the cytoplasmic droplet, inner core of smooth endoplasmic reticulum in the cytoplasmic droplet and the residual Sertoli cell cytoplasm above the acrosome. (b) Longitudinal section through a spermatozoan head in the corpus epididymides. Note the forward contraction of the cytoplasmic droplet, a reduction in the angle between the distal head and the neck, development of the mid piece fibre network and the movement of the smooth endoplasmic reticulum to the outer portion of the cytoplasmic droplet. (c) Longitudinal section through a spermatozoan head in the cauda epididymides. The head and neck are now parallel and the cytoplasmic droplet is lost. (d) SEM of a caput epididymidal spermatozoa. Note the acrosomal extensions. (e) SEM of a cauda epididymidal spermatozoa. The acrosome has settled back into the nuclear indentation. A, acrosome; CD, cytoplasmic droplet; M, mitochondria; MFN, mid-piece fibre network; N, nucleus; NE, nuclear extensions; SC, Sertoli cell cytoplasm; SER, smooth endoplasmic reticulum; V, vacuole. Scale bar: 1 µm.

Figure 12.

Changes occurring in the acrosome, nuclear extensions and mitochondrial sheath during epididymal transport. (a) Typical appearance of the acrosome in transverse section in the caput epididymides. Note the Sertoli cell cytoplasm within the cup-shaped acrosome. (b–e) As the spermatozoa moves through the caput to the corpus epididymides the lateral swollen extensions of the acrosome fold up and over, often causing vacuoles to form within the acrosome. Note the nuclear extensions in (c). (f) Once the spermatozoa have reached the corpus most of the acrosomal folding processes are complete with only a minor settling of contents to occur. Tubules of membranous material can be seen above the acrosome in this section. (g, h) Nuclear extensions frequently occur in spermatozoa between the caput and corpus. They may appear constricted at the base as in g, or may appear moving across the plasma membrane as in (h). (i) Transverse section of a mitochondrial sheath in the caput epididymides. Note that at this stage there is no mid-piece fibre network or distinct mitochondrial cristae, both of which develop throughout the course of the epididymides. DOF, dense outer fibres; T, tubules. For other abbreviations see Fig. 11. Scale bars: (a,b,d,h) 200 nm, (c,e,i) 100 nm, (f) 500 nm, (g) 400 nm.

The cytoplasmic droplet surrounded the head and neck both dorsally and ventrally (Fig. 11a,d). There was usually a large vacuole between the ventral surface of the nucleus and the residual cytoplasm; the latter contained an inner core of smooth endoplasmic reticulum (Fig. 11a).

The acrosome in the concave portion of the nucleus was cup-like and the lateral edges were still swollen and projected upwards and sometimes inwards, so that they almost met above the dorsal surface of the acrosome (Fig. 12a–c). The most anterior portion extended beyond the nucleus as a balloon-like extension (Fig. 11a,d). Distal to the concave portion of the nucleus, the lateral edges of the acrosome were also slightly swollen (Fig. 11d). A droplet of cytoplasm, delimited by a separate membrane to that of the plasma membrane of the spermatozoa and, therefore, presumably of Sertoli cell origin, often lay above the cup-like section of the acrosome (Figs 11a, 12a).

In some sections, an aggregation of a granular material occurred beneath the proximal junctional body. The point of flexion that formed the 45° angle between the long axes of the head and tail was clearly between the anterior and posterior proximal junctional bodies.

Caput–corpus epididymides

There was little difference to the sperm contained within this region from those in the caput epididymides. The sleeve of cytoplasm surrounding the mitochondrial sheath beyond the cytoplasmic droplet was reduced and contained a granular outer layer. In some spermatozoa, the mitochondria of the mitochondrial sheath were becoming more orderly in arrangement and brick-like in appearance. The dorsal cup-like extensions of the acrosome had now moved sufficiently inwards to cause them to merge with one another over the dorsal surface of the nucleus, but the anterior portion still projected forward over the nuclear tip (Fig. 12d,e). Within the acrosome lay pockets of membrane and Sertoli cell material, apparently engulfed during the merging of the extensions (Fig. 12e). In some spermatozoa, the angle between the head and flagellum was slightly reduced. Some of the dense outer fibres of the mitochondrial sheath were displaced outwardly from their corresponding axoneme doublets.

Corpus epididymides

The acrosomes of most spermatozoa had settled into the concave portion of their nuclei, with remnants from the merging acrosomal processes remaining in many. Others were still undergoing the acrosomal condensation process. Membranous tubules often formed either between the acrosome and the plasma membrane (Fig. 12f) or above the plasma membrane. Both the neck and mid-piece fibre networks had formed and the excess cytoplasm had contracted to produce a rounded droplet which hung ventrally from the neck and anterior nucleus. Spermatozoa were becoming more densely packed and the long axes of the neck and the nucleus were approaching parallel. In the sperm that maintained nuclear extensions, these were now constricted at their base and were smaller and more electron-dense than in previous sections (Fig. 12g). Sometimes, these extensions were free from the nuclear surface and appeared to have been moving out of the spermatozoan or towards the cytoplasmic droplet (Fig. 12h). However, the presence of these nuclear extensions in the medium surrounding the spermatozoa or within the body of the cytoplasmic droplet was not encountered, thus the fate of these extensions is unclear.

The segmented sheath was distinguishable as nine columns and the cytoplasmic sleeve surrounding the mitochondrial sheath remained thickened in some sperm. The mitochondria were becoming increasingly organized and brick-like in structure, though in some sperm still remained rounded. The smooth endoplasmic reticulum of the cytoplasmic droplet had frequently moved to the outer portion so that it now lay close to the plasma membrane rather than forming an inner core, as in the caput region (Fig. 11b).

Corpus–cauda epididymides

On reaching the corpus–cauda region, the spermatozoa had lost the cytoplasmic droplet. The nuclear extensions were no longer present and the long axes of the sperm head and tail were now parallel. The granular material below the neck was no longer evident and the spermatozoa resembled the mature sperm from the cauda epididymides. The acrosomal condensation process was almost complete, with a minor settling of the acrosomal contents at the posterior edge of the thickened portion being the only development needed for the acrosome to resemble the mature cauda epididymal sperm described in Lloyd et al. (2002).


Spermiogenesis and sperm maturation in the Musky Rat-kangaroo are similar to other marsupials (Sapsford et al. 1967, 1969; Phillips, 1970; Harding et al. 1976a,b, 1982, 1984; Kim et al. 1987; Lin et al. 1997; Setiadi et al. 1997; Johnston et al. 2004). However, there are some features in the Musky Rat-kangaroo which are infrequently reported in other macropodoids or in some instances even in the general marsupial literature and these will be the focus of this discussion.

Acrosomal granules in marsupials

Though acrosomal granules are rarely reported in marsupials, when they do form the process is similar to that of eutherians, via coalescence of smaller vesicles containing granules [American Wooly Opossum Caluromys philander (Phillips, 1970); Koala Phascolarctos cinereus (Harding et al. 1987); Rufous Hare-wallaby Lagorchestes hirsutus (Johnston et al. 2004)]. However, in the present study the granule has a different structure and appears to form from a condensation of excess membrane either from smaller vesicles joining the acrosome and/or from the main acrosomal membrane as it collapses at the start of nuclear protrusion. Lamellar bodies similar to these granules were occasionally found in the cytoplasm after acrosomal collapse which may be indicative of the fate of the acrosomal granule in this species.

Nuclear ring contraction

In most Australian marsupials studied to date (Sapsford et al. 1969; Harding et al. 1976a, 1984; Harding, 1977; Kim et al. 1987; Lin et al. 1997), the nuclear ring undergoes an anterior contraction which has been suggested to define the distal margin of the acrosome (Harding, 1977). Like the petaurids and dasyurids (Harding, 1977) the Musky Rat-kangaroo does not undergo nuclear ring contraction and all three groups have an extensive acrosomal coverage, which provides additional support for Harding's theory.

The contraction of the nuclear ring was also proposed to be responsible for producing a large sub-acrosomal space during spermiogenesis in the Brush-tail Possum (Trichosurus vulpecula) and the Honey-possum (Tarsipes rostratus) (Harding et al. 1976a, 1984). However, Lin et al. (1997) found this space to arise prior to nuclear ring contraction in the Tammar Wallaby (Macropus eugenii) and suggested a vesicle forming in the acrosome during early spermiogenesis develops into the sub-acrosomal space. As there is no nuclear ring contraction in the Musky Rat-kangaroo and a vesicle appears to give rise to the sub-acrosomal space similarly to the Tammar Wallaby Lin et al.'s suggestion is supported.

Harding et al. (1983) are of the opinion that epididymal maturation might be responsible for compacting the acrosome into the reduced area defined by nuclear ring contraction. A pronounced acrosomal maturation process is usually restricted to those marsupials with small acrosomes and is either lacking or reduced in species with more extensive acrosomes. However, the process is substantial in the Musky Rat-kangaroo despite the extensive acrosome and the lack of nuclear ring contraction. The epididymal maturation process may be a continuation of the collapse of the acrosome into the sub-acrosomal space which commences during late spermiogenesis.

Nuclear extensions

Although only recently reported in the spermatozoa of the Rufous Hare-wallaby (Lagorchestes hirsutus) (Johnston et al. 2003, 2004), the nuclear ‘bleb-like’ extensions occurring in spermatozoa from the proximal parts of the Musky Rat-kangaroo epididymides, can also be identified in micrographs of the Tammar Wallaby (see Figs 8a and 10c Setiadi et al. 1997). Similar nuclear projections have also been reported in spermatozoa within the cauda epididymides of dasyurids (Harding et al. 1979). Although the fate of these nuclear extensions is uncertain in the Musky Rat-kangaroo, they are usually absent by the time spermatozoa reach the corpus–cauda. Sapsford et al. (1969) suggest that uncondensed chromatin in the Long-nosed Bandicoot (Perameles nasuta) sperm leaves the nucleus during spermiogenesis via adjacent nuclear pores and this may also occur in the Musky Rat-kangaroo. However, as nuclear blebs first appear around the time of spermiation and subsequently disappear around the time the cytoplasmic droplet is shed, it is possible these nuclear extensions contain the residual nuclear material.

Post-acrosomal complex

Though the post-acrosomal complex was first reported in the Tammar Wallaby by Lin et al. (1997) it appears in the micrographs of several other species (Phillips, 1970; Harding, 1977). Lin et al. (1997) report that the position of the post-acrosomal complex is close to the area where Taggart et al. (1993) found sperm-egg fusion to occur and that the post-acrosomal complex structure is ‘... remarkably different from the rest of the nuclear membrane’. While the post-acrosomal complex is close to the posterior edge of the acrosome in the Musky Rat-kangaroo, it clearly remains on the ventral nuclear surface where it could be expected to make difficult any role in sperm-egg fusion in this species.

Nuclear rotation

There is little information in the literature on the possible mechanisms or functions of sperm head rotation in marsupials. During the second part of the initial nuclear rotation, the cementum takes on a banded appearance and the distal portion of the proximal junctional body extends. The appearance of banding within the cementum and the lengthening of the distal portion of the proximal junctional body occur at a time that may suggest these structures play a role in the process of rotation.

Cycle of the seminiferous epithelium

Spermatogenesis was categorized into eight stages, though these differed slightly from those reported in other mammals analysed using the Roosen-Runge & Giesel (1950) method (Roosen-Runge & Giesel, 1950; Setchell & Carrick, 1973; Tait & Johnson, 1982; Peirce & Breed, 1987; Wrobel et al. 1993; Kerlin, 1999).

The nuclei of early spermatids in other marsupials appear rounded and uncondensed prior to the release of the spermatozoa (Setchell & Carrick, 1973; Mason & Blackshaw, 1973; Orsi & Ferreira, 1978; Lin et al. 2004). In contrast, the early spermatids have become triangular with condensed nuclei (nuclear protrusion) before the spermatozoa leave the tubule in the Musky Rat-kangaroo; indicating that either spermatozoa are retained for a proportionately longer period of time (modification of the extensive acrosome may require additional time) or that differentiation of the early spermatid is more rapid in the Musky Rat-kangaroo. Because spermatozoa remained in the lumen throughout part of nuclear protrusion, Stage 8 in the Musky Rat-kangaroo is extended into what would ordinarily be considered Stage 1 in other species. This would have reduced the number of stages in the Musky Rat-kangaroo to seven had it not been for the extensive acrosomal vacuole and the elongate head of the spermatozoa, assisting in demarcating stages, and thus an extra post-divisional stage (Stage 7), not reported by other authors was observed.

Timing of sperm development

The frequency of stages indicates the proportion of time germ cells remain in each phase of development. For instance, elongation and nuclear shaping occur during Stage 3, which occupies only 6.7% of the cycle, suggesting this process is rapid. For the Musky Rat-kangaroo greater proportions of time are spent in acrosomal development and mitochondrial sheath formation phases, which is perhaps indicative of the complexity of these developments, or a factor in development which is rate limiting. It is tempting to speculate that the slow development of the acrosomal complex reflects a rate limiting effect of the Golgi apparatus and its capacity to form vesicles. If this were true, then it could be expected that those species with more discrete acrosomes would spend proportionately less time in the acrosomal phase.

Part of this research involved the testicular injection of [3H] thymidine to quantify the duration of the cycle and the timing of ultrastructural development. However, testicular tissue taken between 9 and 17 days after administration of isofluorane anaesthetic and testicular injection, had disruptions to the seminiferous cycle within varying cellular cohorts depending on time since injection. Inhalation anaesthetics in rabbits disrupts spermatogenesis (Ceyhan et al. 2005) and this may also be true for the Musky Rat-kangaroo.

Concluding comments

Spermiogenesis and sperm maturation in the Musky Rat-kangaroo is similar to that reported in other marsupials. However, there are some processes and features not found elsewhere among the macropodoids or the phalangeroids, the groups to which it is most closely related. These are the occurrence of a non-eutherian type acrosomal granule, the lack of nuclear ring contraction, the occurrence of an acrosomal maturation process despite the extent of acrosomal coverage (though this is also found in Tarsipes), and a delay in spermiation. This is the first time that the changing ultrastructure of spermatids has been described for each stage of the seminiferous epithelial cell cycle.


Special thanks must go to the many field assistants that assisted with this project in trapping Musky Rat-kangaroos. Our gratitude goes to Darren Storch for his advice on trapping and the use of traps and to the dedicated technical staff for the long hours in assisting with tissue processing and section cutting; Rick Webb, Rob Gould, Tina Chau, Lina Daddow and Anthony Chang.