Fine structural observations of spermiogenesis in sterile mice homozygous for the t-allele tw2 reveal that spermatids develop severe morphological abnormalities during spermiogenesis. In a small proportion of late spermatids the nuclear envelope develops broad areas of discontinuity except at the level of the acrosome. Manchette microtubules, which in normal spermatids form an orderly perinuclear array, appear to undergo premature depolymerization in cells of this type although flagellar structures differentiate normally.

The majority of late spermatids, on the other hand, contain unusually large numbers of disorganized microtubules. Spermatid head shapes likewise become very abnormal. Both acrosomal and post-acrosomal portions of the nucleus exhibit malformations. The presence of disorganized microtubular arrays is believed to contribute to the observed nuclear deformations.

Although condensation of the chromatin occurs in these cells, the spermatids seldom survive to maturity. Most of the cells are phagocytized by Sertoli cells.

Late spermatids from tw2 homozygotes are abnormal in that they fail to maintain a precisely ordered array of microtubules during the concluding stages of spermiogenesis. This defect may reflect an underlying abnormality in the structural or biochemical characteristics of the plasma membrane with which the microtubules become associated.

Male mice homozygous for recessive semi-lethal mutations at the T-locus are completely sterile and almost aspermic (Bennett & Dunn, 1967; Johnston, 1968; Bennett & Dunn, 1971). Yet, male heterozygotes which carry one wild type gene at the T-locus and one semi-lethal, appear to produce normal numbers of sperm, but transmit the mutant gene to a much higher proportion (76–97 %) of their progeny than would be predicted by Mendel’s rules (Dunn, 1960; Yanagisawa, Dunn & Bennett, 1961; Bennett & Dunn, 1967; Bennett & Dunn, 1971). The anomalous situation whereby the heterozygous genotype confers some advantage to sperm carrying the abnormal allele without impairing fertility, while homozygosity for the same gene leads to sterility has not been resolved. Nevertheless, it is clear that these genes must have a significant role in both sperm function and sperm production.

Other work in our laboratory has recently established that genes at the T-locus specify serologically detectable cell surface components on sperm (Yanagisawa et al. 1974a; Yanagisawa et al. 1974b) and it is not unreasonable to assume that these abnormal cell surface antigens may be related to both the transmission ratio distortion and sterility seen in males carrying mutant genes.

We report here an investigation of the fine structure of spermiogenesis in sterile mice homozygous for the mutation tw2, and describe abnormalities of spermatid differentiation which are compatible with the hypothesis that defective spermiogenesis may reflect primary abnormalities of the plasma membrane and of the membranes of the nuclear envelope.

Mice homozygous for tw2 were killed by cervical dislocation, and their testes rapidly dissected and placed into an aldehyde fixative where the tissue was cut into pieces about 3 mm3. Some testes were placed in 2·5 % glutaraldehyde buffered with 0·1 M phosphate buffer, pH 7·4. Other testes were placed in a buffered mixture of glutaraldehyde and acrolein of the following proportions: 2 % glutaraldehyde; 2·5 % acrolein; 0 · 1 M phosphate buffer, pH 7·4. Aldehyde fixation continued for 1 h at room temperature followed by 3 – 4 rinses in 0 · 2 M phosphate buffer, for a total of approximately 10 – 15 min. Postfixation in 1 · 5 % OsO4 in 0 · 1 M phosphate buffer, pH 7 · 4 for 1 h at room temperature was followed by rapid dehydration and embedding in Araldite. Sections were viewed in a Philips 200 electron microscope.

In this report, successive stages of spermatid differentiation have been identified in the light microscope according to the criteria provided by Oakberg (1956), and in the electron microscope according to the present authors’ previously reported guidelines (Dooher & Bennett, 1973).

Mice that are homozygous for tw2 generate a morphologically normal population of new spermatids at the end of stage XII of the cycle of renewal of the seminiferous epithelium. Spermatids in the mutant remain indistinguishable from normal cells in both the light and electron microscopes until spermiogenesis is well advanced. Briefly, the acrosome develops normally and by spermatid stage 6 envelopes the anterior half of the nucleus (Dooher & Bennett, 1973). The formation of the flagellum in the mutant likewise exhibits no remarkable characteristics.

At early stage 8, in both normal spermatids and in spermatids from the mutant, an orderly array of cytoplasmic microtubules, the manchette, appears. The manchette forms a cylindrical sheath several μ m in length and approximately 250 nm in width (Fig. 1). Anteriorly, the microtubules closely approach the cell surface where they appear to terminate within an elevation of the plasma membrane immediately behind the acrosomal cap, the perinuclear ring (Fig. 2). Amorphous electron dense material is associated with the microtubules in this region. Spermatid nuclei begin to elongate at this time and the rostrum, which is that portion of the spermatid head containing the acrosome (W. I. Bennett, Gall, Southard & Sidman, 1971) begins to develop a characteristic anterior flexure (Fig. 2).

Fig. 1.

Early stage 8 spermatid from the mutant (tw2 / tw2) illustrating normal morphological features of these cells at the time the manchette (Mt) appears. At this early stage the region of the perinuclear ring (arrowed) is not well defined, × 16500.

Fig. 1.

Early stage 8 spermatid from the mutant (tw2 / tw2) illustrating normal morphological features of these cells at the time the manchette (Mt) appears. At this early stage the region of the perinuclear ring (arrowed) is not well defined, × 16500.

Fig. 2.

Normal ( + / + ) stage 9 spermatid illustrating the organization of the manchette (Mt) including the perinuclear ring (PR) which, at this stage, is well developed. The rostrum has begun to acquire a marked anterior flexure, × 16500.

Fig. 2.

Normal ( + / + ) stage 9 spermatid illustrating the organization of the manchette (Mt) including the perinuclear ring (PR) which, at this stage, is well developed. The rostrum has begun to acquire a marked anterior flexure, × 16500.

Spermatids from the mutant continue to resemble their normal counterparts until late stage 11 or early stage 12. At this time two types of abnormal spermatids with strikingly dissimilar morphological features become evident.

In a small percentage of the spermatids, the nuclear envelope exhibits extensive areas of discontinuity so that, in many places, nucleoplasm and cytoplasm are confluent (Fig. 3 A). Scattered vesicles, most likely remnants of the nuclear envelope, are observed at the nuclear perimeter and also situated deep within the ‘nucleoplasm’. The nuclear envelope appears to retain normal morphology only at the level of the acrosome where it conforms closely to the acrosome and contains electron dense chromatin on its inner leaflet. Although the connecting piece, which develops normally, retains its usual close association with the nuclear envelope (Fig. 3B), discontinuities in the nuclear envelope may also appear at this level. Furthermore, condensed chromatin, normally present on the inner leaflet of the nuclear envelope in this region (Fig. 3C), is not observed.

Fig. 3.

A, Stage 11 spermatid from the mutant. The nuclear envelope is broadly discontinuous and vesicles (arrowed), probably remnants of the nuclear envelope, are visible. Manchette microtubules are not visible in cells of this type. A portion of the connecting piece (CP is included in the lower center of the micrograph), × 12000. B, Higher magnification of the connecting piece illustrated in (A). Although the centriolar derivatives exhibit normal morphology, the nuclear envelope (arrowed) appears disrupted in this region, × 40000. C, Connecting piece from a normal spermatid at stage 11. Note condensed chromatin (arrowed) present on the inner leaflet of the nuclear envelope, × 46000.

Fig. 3.

A, Stage 11 spermatid from the mutant. The nuclear envelope is broadly discontinuous and vesicles (arrowed), probably remnants of the nuclear envelope, are visible. Manchette microtubules are not visible in cells of this type. A portion of the connecting piece (CP is included in the lower center of the micrograph), × 12000. B, Higher magnification of the connecting piece illustrated in (A). Although the centriolar derivatives exhibit normal morphology, the nuclear envelope (arrowed) appears disrupted in this region, × 40000. C, Connecting piece from a normal spermatid at stage 11. Note condensed chromatin (arrowed) present on the inner leaflet of the nuclear envelope, × 46000.

Another remarkable feature of these cells in which the nuclear envelope appears vesiculated, is that they lack the manchette (Fig. 3 A). Apparently either the array of microtubules fails to form in cells of this type, or they undergo premature depolymerization. The fact that cells lacking the manchette are not seen earlier than stage 11 suggests that microtubules did form in these cells and subsequently depolymerized. In normal cells, the microtubules of the manchette persist until late stage 15 (Dooher & Bennett, 1973).

A different constellation of abnormalities afflicts the majority of the spermatids by the time they have reached stage 11–12. These cells appear to contain unusually large numbers of long, straight microtubules (Fig. 4). In many cells the microtubules are arranged abnormally with respect both to the nuclear envelope and to the plasma membrane. The nucleus is often grossly distorted and bundles of microtubules are observed within deep invaginations of the nucleus. Furthermore, the microtubules that approach the cell membrane frequently impinge upon large expanses of the plasma membrane flanking the nucleus in an irregular and highly asymmetric fashion instead of being confined to the usual narrow perinuclear band of cortical cytoplasm. Nevertheless, wherever microtubules approach the cell surface, they are associated with flocculent material that resembles the material observed within the perinuclear ring of normal cells. The simplest explanation for these findings is that, in the majority of spermatids from mice homozygous for tw2, an abnormal proliferation of microtubules occurs after the organelles first appear, most likely during stage 11, and that this proliferation is responsible for the bizarre postrostral nuclear shapes that typify some of the cells at this stage.

Fig. 4.

Stage 11 spermatid from the mutant illustrating morphological features shared by the majority of abnormal spermatids at this stage. The acrosome (Ac) is unusual in shape. The cell contains an extensive, disorganized array of microtubules and a severely deformed nucleus. Note electron dense material associated with microtubular endings immediately under the plasma membrane at the anterior end of the cell, × 23000.

Fig. 4.

Stage 11 spermatid from the mutant illustrating morphological features shared by the majority of abnormal spermatids at this stage. The acrosome (Ac) is unusual in shape. The cell contains an extensive, disorganized array of microtubules and a severely deformed nucleus. Note electron dense material associated with microtubular endings immediately under the plasma membrane at the anterior end of the cell, × 23000.

Deformities of the rostrum are also typical of abnormal spermatids (Figs. 4, 5 A, B, 6A). It is possible that these abnormalities, which affect structures anterior to the manchette, may arise independently of the microtubular defect. However, forces imposed upon the cell by the presence of unusually large numbers of microtubules, which are known to be rigid structures, probably contribute to the failure of the rostrum to continue to develop a normal curving profile.

Fig. 5.

A, B, Abnormal stage 13 spermatids illustrating typical examples of rostral deformities and unusual arrangements of manchette microtubules. The nuclei contain electron dense granules (arrowed) scattered within the condensing chromatin. × 14900.

Fig. 5.

A, B, Abnormal stage 13 spermatids illustrating typical examples of rostral deformities and unusual arrangements of manchette microtubules. The nuclei contain electron dense granules (arrowed) scattered within the condensing chromatin. × 14900.

Despite the distorted nuclear shape, condensation of the chromatin appears to occur on schedule, beginning during early stage 12 and continuing until late stage 14 (Figs. 5 A, B, 6 A). This observation demonstrates that chromosomal condensation occurs despite extraordinary abnormalities of nuclear shape and apparently independently of the unusual orientation of the surrounding microtubules. However, an interesting characteristic, occasionally observed in spermatid nuclei undergoing chromosomal condensation, is the presence of scattered, densely staining granules, surrounded by lightly staining halos, features not observed in normal cells (Figs. 5 A, B). Evidently these unusual granular bodies are transitory since they are not detected in nuclei in which condensation has been completed. The significance of these structures remains unexplained.

Spermatids seldom survive to maturity in homozygotes for tw2. The majority of the cells are phagocytized by Sertoli cells shortly before spermiation would normally take place. Evidence for phagocytosis includes our frequent observation of sperm tails which are surrounded by Sertoli cell cytoplasm (Fig. 6B), a condition not encountered in the normal testis. Moreover, occasionally the remains of a late spermatid can be readily recognized, embedded within a Sertoli cell (Fig. 6C). Furthermore, large phagosomes, frequently observed in Sertoli cells, probably represent the remains of engulfed spermatids.

Fig. 6.

A, Abnormal spermatid at late stage 14. Although the spermatid head is extremely distorted in shape, chromosomal condensation is well advanced. Manchette microtubules are highly disorganized in this cell. Note that the section includes two disconnected profiles through the nucleus (N). × 14900. B, Sperm tails, embedded in Sertoli cell cytoplasm which provide evidence that, in the mutant, late spermatids are phagocytized by Sertoli cells before reaching maturity, × 34500. C, The remains of a late spermatid, recognized on the bases of the nucleus (N) and remnants of the connecting piece (CP), embedded within a Sertoli cell and situated near the basement membrane (BM) of the seminiferous tubule. Portions of two Sertoli cell nuclei (SN) are included in the micrograph, × 19600.

Fig. 6.

A, Abnormal spermatid at late stage 14. Although the spermatid head is extremely distorted in shape, chromosomal condensation is well advanced. Manchette microtubules are highly disorganized in this cell. Note that the section includes two disconnected profiles through the nucleus (N). × 14900. B, Sperm tails, embedded in Sertoli cell cytoplasm which provide evidence that, in the mutant, late spermatids are phagocytized by Sertoli cells before reaching maturity, × 34500. C, The remains of a late spermatid, recognized on the bases of the nucleus (N) and remnants of the connecting piece (CP), embedded within a Sertoli cell and situated near the basement membrane (BM) of the seminiferous tubule. Portions of two Sertoli cell nuclei (SN) are included in the micrograph, × 19600.

It has been known for some time that male mice homozygous for tw2 have very few sperm in the vas deferens or in ejaculates; those sperm that are found are abnormal in shape, non motile, and are either not capable of reaching the fallopian tube or of remaining in it (Johnston, 1968; Bennett & Dunn, 1971). Our present studies on these animals demonstrate that the late stages of spermiogenesis are severely impaired, and that most abnormal sperm cells are eventually phagocytized by Sertoli cells; this observation accounts for the very low numbers of sperm observed in the terminal ducts and in ejaculates.

The morphological defects we have observed in electron micrographs of the testis appear to be confined exclusively to the sperm head; these defects become evident first in late stage 11 or early stage 12 spermatids when condensation of the chromatin normally occurs and the shaping of the sperm head enters its final phase.

The formation of the structurally complex connecting piece appears to be normal, and likewise the flagella seem to develop normally. The normal flagellar pattern of microtubules which we have seen regularly in testicular spermatids of these sterile males is surprising in view of two reports that mice carrying various t-allele combinations show dramatic distortions of the normal 9 + 2 arrangement of microtubules within the axial complex in sperm found in the epididymis or vas deferens, as well as in ejaculates (Yanigisawa, 1965; Olds, 1971). In both of these studies the presence of extra microtubules was observed among the outer doublets in some axial filaments, whereas in other flagella a variable number of microtubules appeared to be missing. The most likely explanation for the normal pattern of microtubules we find within the flagella of testicular sperm and the abnormal patterns described by others in more mature stages is an instability of the polymerized structure of the microtubules in these abnormal cells. It may be that the absence of manchette microtubules in a small proportion of the late spermatids we have observed, likewise reflects an instability of the polymerized microtubular protein in these cells.

With respect to nuclear morphology, our observations suggest that factors responsible for the formation and maintenance of the precisely arranged microtubules of the manchette are impaired in spermatids from mice homozygous for tw2. In normal mammalian spermatids the microtubules of the manchette establish a close structural relationship with the plasma membrane at the level of the perinuclear ring (Burgos & Fawcett, 1955; Fawcett, Anderson & Phillips, 1971 ; Rattner & Brinkley, 1972) and it has been proposed that the perinuclear ring contributes to the structural rigidity of the manchette and to the maintenance of the precisely ordered array of the microtubules (Fawcett et al. 1971; Rattner, 1972). It has further been suggested that the perinuclear ring may function in the initial organization of the microtubular array (Rattner, 1972; Rattner & Olson, 1973). Additional evidence that microtubules may bear a functional relationship with respect to the plasma membrane comes from biochemical evidence that microtubular protein may play a role in the control of the distribution and mobility of various kinds of receptor sites on the surfaces of cells, including receptor sites for the plant lectin Concanavalin A (Berlin & Ukena, 1972; Yin, Ukena & Berlin, 1972; Edelman, Yahana & Wang, 1973; Ukena, Borysenko, Karnovsky & Berlin, 1974). Thus it is reasonable to assume that the plasma membrane is specialized in some way at the level of the perinuclear ring so that the appropriate association of microtubules and the plasma membrane occurs.

In the abnormal cells that contain disorganized microtubular arrays, microtubules nevertheless appear to establish structural association with regions of the plasma membrane upon which they impinge. What is abnormal in these cells is that unusually large numbers of microtubules, and, consequently, abnormally extensive areas of the plasma membrane become involved in the phenomenon. The plasma membrane may be unusual in possessing a relatively unrestricted capacity to become structurally bound to microtubules as spermiogenesis proceeds.

The possibility that a mutant gene may be responsible for the presence of abnormalities in spermatid membranes and resulting development of morphologically defective spermatids has previously been proposed (W. I. Bennett et al. 1971). Electron microscopy of spermiogenesis in mice homozygous for the mutant gene quaking, qk, reveals that the heads of late spermatids become distorted in shape. Although these head defects strikingly resemble those seen in spermatids from mice homozygous for tw2, abnormalities of manchette formation apparently do not occur in quaking spermatids.

A fine structural study of embryos homozygous for the lethal allele t9 has revealed that cells of the primitive streak fail to migrate normally and therefore the embryos fail to form normal mesodermal structures (Spiegelman & Bennett, 1974). In the mutant, cells deriving from the primitive streak fail to form normal numbers of filopodia with their associated intercellular junctions, observations that have been interpreted to explain the failure of these cells to undertake normal migratory behaviour. These results are also compatible with the hypothesis that t-alleles influence properties of the plasma membrane.

Finally, serological data has convincingly demonstrated on the surface of spermatozoa and of embryonic cells, the presence of components specified by genes at the T-locus (Bennett, Goldberg, Dunn & Boyse, 1972; Yanagisawa et al. 1974a, b; Artzt, Bennett & Jacob, 1974). These results add strong support to the hypothesis that the T-locus is a site with a major role in the genetic control of functional properties of the plasma membrane of spermatozoa and of embryos. Mutations at this critical locus may thus produce their effects during spermiogenesis and embryogenesis by disrupting critical events of differentiation in which the plasma membrane plays a significant role.

A summary of this work was presented in preliminary form at the 13th International Congress of Genetics at Berkeley, California in August 1973. The authors thank Victoria Neufeld for her expert technical assistance. We extend thanks to Dr Roy C. Swan for critical reading of the manuscript. This work was supported by National Science Foundation Grant BG 33804X.

Artzt
,
K.
,
Bennett
,
D.
&
Jacob
,
F.
(
1974
).
Primitive teratocarcinoma cells express a differentiation antigen specified by a gene at the T-locus in the mouse
.
Proc. natn. Acad. Sci. U.S.A
.
71
,
811
814
.
Bennett
,
D.
&
Dunn
,
L. C.
(
1967
).
Studies of effects of r-alleles in the house mouse on spermatozoa. 1. Male sterility effects
.
J. Reprod. Fert
.
13
,
421
428
.
Bennett
,
D.
&
Dunn
,
L. C.
(
1971
).
Transmission ratio distorting genes on chromosome IX and their interactions
.
Proc. Symp. Immunogenetics of the H-2 System
, pp.
90
103
.
Basel
:
Karger
.
Bennett
,
D.
,
Goldberg
,
E.
,
Dunn
,
L. C.
&
Boyse
,
E. A.
(
1972
).
Serological detection of a cell-surface antigen specified by the T(Brachyury) mutant gene in the house mouse
.
Proc. natn. Acad. Sci. U.S.A
.
69
,
2076
2080
.
Bennett
,
W. I.
,
Gall
,
A. M.
,
Southard
,
J. L.
&
Sidman
,
R. L.
(
1971
).
Abnormal spermio-genesis in quaking, a myelin-deficient mutant mouse
.
Biol. Reprod
.
5
,
30
58
.
Berlin
,
R. D.
&
Ukena
,
T. E.
(
1972
).
Effect of colchicine and vinblastine on the agglutination of polymorphonuclear leucocytes by concanavalin A
.
Nature New Biol. Lond.
238
,
120
121
.
Burgos
,
M. H.
&
Fawcett
,
D. W.
(
1955
).
Studies on the fine structure of the mammalian testis. I. Differentiation of the spermatids of the cat
.
J. biophys. biochem. CytoL
1
,
286
300
.
Dooher
,
G. B.
&
Bennett
,
D.
(
1973
).
Fine structural observations on the development of the sperm head in the mouse
.
Am. J. Anat
.
136
,
339
362
.
Dunn
,
L. C.
(
1960
).
Variations in the transmission ratios of alleles through eggs and sperm in Mus musculus
.
Am. Nat
.
94
,
385
393
.
Edelman
,
G. M.
,
Yahana
,
I.
&
Wang
,
J. L.
(
1973
).
Receptor mobility and receptor-cytoplasmic interactions in lymphocytes
.
Proc. natn. Acad. Sci. U.S.A
.
70
,
1442
1446
.
Fawcett
,
D. W.
,
Anderson
,
W. A.
&
Phillips
,
D. M.
(
1971
).
Morphogenetic factors influencing the shape of the sperm head
.
Devi Biol
.
26
,
220
251
.
Johnston
,
P. G.
(
1968
).
Male sterility in mice homozygous for the tw2 allele
.
Aust. J. biol. Sci
.
21
,
947
951
.
Oakberg
,
E. F.
(
1956
).
A description of spermiogenesis in the mouse and its use in analysis of the cycle of the seminiferous epithelium and germ cell renewal
.
Am. J. Anat
.
99
,
391
413
.
Olds
,
P. J.
(
1971
).
Effect of the T locus on sperm ultrastructure in the house mouse
.
J. Anat.
109
,
31
37
.
Rattner
,
J. B.
(
1972
).
Nuclear shaping in marsupial spermatids
.
J. Ultrastruct. Res
.
40
,
498
512
.
Rattner
,
J. B.
&
Brinkley
,
B. R.
(
1972
).
Ultrastructure of mammalian spermiogenesis. III. The organization and morphogenesis of the manchette during rodent spermiogenesis
.
J. Ultrastruct. Res
.
41
,
209
218
.
Rattner
,
J. B.
&
Olson
,
G.
(
1973
).
Observations on the fine structure of the nuclear ring of the mammalian spermatid
.
J. Ultrastruct. Res
.
43
,
438
444
.
Spiegelman
,
M.
&
Bennett
,
D.
(
1974
).
Fine structural study of cell migration in the early mesoderm of normal and mutant mouse embryos (T-locus: t°lt°)
:
J. Embryol. exp. Morph.
32
,
723
738
.
Ukena
,
T. E.
&
Berlin
,
R. D.
(
1972
).
Effect of colchicine and vinblastine on the topo-graphic separation of membrane functions
.
J. exp. Med
.
136
,
1
7
.
Ukena
,
T. E.
,
Borysenko
,
J. Z.
,
Karnovsky
,
M. J.
&
Berlin
,
R. D.
(
1974
).
Effects of colchicine, cytochalasin B, and 2-deoxyglucose on the topographical organization of surface-bound concanavalin A in normal and transformed fibroblasts
.
J. Cell Biol
.
61
,
70
82
.
Yanagisawa
,
K.
(
1965
).
Studies on the mechanism of abnormal transmission ratios at the T locus in the house mouse. IV. Some morphological studies on the mature sperm in males heterozygous for r-alleles
.
Jap. J. Genet
.
40
,
97
104
.
Yanagisawa
,
K.
,
Bennet
,
D.
,
Boyse
,
E. A.
,
Dunn
,
L. C.
&
Dimeo
,
A.
(
1974a
).
Serological identification of sperm antigens specified by lethal r-alleles in the mouse
.
Immunogenetics
1
,
57
67
.
Yanagisawa
,
K.
,
Dunn
,
L. C.
&
Bennett
,
D.
(
1961
).
On the mechanism of abnormal transmission ratios at T locus in the house mouse
.
Genetics, Princeton
46
,
1635
1644
.
Yanagisawa
,
K.
,
Pollard
,
D. R.
,
Bennett
,
D.
,
Dunn
,
L. C.
&
Boyse
,
E. A.
(
1974b
).
Transmission ratio distortion at the T-locus: serological identification of two sperm populations in r-heterozygotes
.
Immunogenetics
1
,
91
96
.
Yin
,
H. H.
,
Ukena
,
T. E.
&
Berlin
,
R. D.
(
1972
).
Effect of colchicine, colcemid and vin-blastine on the agglutination, by concanavalin A, of transformed cells
.
Science, N.Y.
178
,
867
868
.