TISSUE & CELL 1979 11 (1) 147-162 Puhlishrri by Longman Group Ltd. Printed in Great Britain
DON W. FAWCETT
and HECTOR
E. CHEMES
CHANGES IN DISTRIBUTION OF NUCLEAR PORES DURING DIFFERENTIATION OF THE MALE GERM CELLS ABSTRACT. Changes in number of nuclear pores in different states of physiologica activity have been reported, but little is known about changing patterns of distribution in the course of cell ditferentiation. Pore distribution in male germ cells was studied in freeze fracture preparations of immature and mature rodent testis. As in other somatic cells, pores were uniformly and apparently randomly distributed in Sertoli cell nuclei. The nucleus of gonocytes and spermatogonia showed varying degrees of pore clustering. Spermatocytes invariably exhibited very striking pore aggregation with close hexagonal packing in pore-rich areas, and large pore-free areas. In early spermatids, pores appeared to be randomly distributed. As the acrosome formed and spread over the apical pole of the nucleus, pores disappeared ahead of its advancing margin and became more concentrated in the post-acrosomal region. The relationship of pore complexes to the chromosomes and the role of the fibrous lamina are discussed. The question as to whether the changing patterns observed involve movement of pores within fluid nuclear membranes, or a dissolution and reformation of new pores remains unanswered.
formation in spermatids have also been described (Fawcett, 1975; Sandoz, 1974). To our knowledge, however, there has been no study concerned specifically with changes in the nuclear envelope throughout the course of a complex process of cell differentiation. The present investigation was undertaken to follow in freeze-fracture replicas the changing pattern of nuclear pore distribution in the germ cells during spermatogenesis in rats, mice and guinea-pigs.
Introduction LITTLE is known about the mechanism of nuclear pore formation or the functional significance of different arrangements of pores in the nuclear envelope. Studies on various somatic cell types have shown that there is an approximate doubling of pores in the interphase nucleus associated with the increase in nuclear volume and replication of DNA prior to cell division (Maul et al., 1972). An increase in pore number has also been reported for cells stimulated to greater metabolic and synthetic activity (Moses, 1964; Maul et al., 1972). In a previous paper from this laboratory, attention was drawn to a non-random distribution of nuclear pores in human spermatogenic cells studied in thin sections (Chemes et al., 1978). Changes in pore distribution associated with acrosome
Materials and Methods The freeze-cleaving method of specimen preparation affords extensive en ,fnce views of the nuclear envelope and is therefore especially useful for studying the number and distribution of nuclear pore complexes. The observations reported here are based upon study of freeze-fracture replicas of the testes of immature (5day and ICday old) and mature guinea pigs of the Duncan Hartley strain: immature mice (2-day and &day old) of the CD-I strain; and pubescent rats (33-day old) of the Sprague-Dawley strain.
Department of Anatomy and Laboratory of Human Reproduction and Reproductive Biology, Harvard Medicai School, Boston, Massachusetts 071 IS. Received 5 October
1978. 147
FAWCETT
148
Adult guinea-pig and 33-day old rat testes were fixed by vascular perfusion either via the abdominal aorta or through the testicular artery. Testes from immature guinea-pigs and mice were fixed by immersion in collidinebuffered (pH 7.4) 5 % glutaraldehyde, or a glutaraldehyde and paraformaldehyde mixture. The results were found to be similar with the two fixatives. After a brief fixation of approximately 30 min duration, small blocks of tissue were washed in buffer, transferred to a 20% solution of glycerin in distilled water for 2 hr, and then in frozen liquid Freon 22 cooled in liquid nitrogen. The blocks were cleaved at - 125” and the exposed surfaces coated with carbon, and platinum shadowed, in a Balzer’s apparatus. The tissue was digested in three successive baths of Clorox, and after careful washing, the replicas were mounted on uncoated copper grids. Samples of the same tissue were left in the original fixative for an additional 2 hr, postfixed in collidine-buffered 1.25% 0~04, dehydrated in ethanol and embedded in an Epon-Araldite mixture. Thick (1 pm) sections were mounted on glass slides and stained with 1 o/0 toluidine blue in borax. Thin sections exhibiting pale gold interference colors were mounted on uncoated copper grids and double stained with uranyl acetate and lead citrate. Both replicas and sections were studied in a Philips EM 200 electron microscope at 60 kV. Results
Testis development and onset of spermatogenesis are much the same in the three mammalian species studied although the time course differs. Gonocytes in the seminiferous cords proliferate throughout fetal life and subsequently move to the periphery of the cords and there give rise to spermatogonia. With the onset of spermatogenesis at puberty, the spermatogonia differentiate into two main categories: those that divide at a slow rate to maintain the stem cell population; and others that proliferate and differentiate to form syncytial clusters of spermatocytes. After a protracted meiotic prophase, the spermatocytes undergo two divisions resulting in the formation of large groups of haploid round spermatids. During spermiogenesis, these are gradually transformed into spermatozoa which are released into the lumen at spermiation.
AND CHEMES
In freeze-fracturing, the path of cleavage preferentially follows the hydrophobic interior of the lipid bilayer in the membranes of cells. Thus the cleavage plane reveals interesting details of internal structure of the membranes both at the surface and in the interior of cells. Where the fracture breaks across the cytoplasmic matrix or nucleoplasm in its path from one membranous organelle to another, the replicas provide little useful information. Identification of the different kinds of cells in the seminiferous epithelium in freeze-fracture preparations depends upon their location in the epithelium and upon recognition of specific membranous cytological features of each cell type, learned from previous electron microscopic study of the testis in thin sections. For the experienced student of spermatogenesis, identification of the several cell types poses no great difficulty. For purposes of comparison, we begin with a description of the nuclei of Sertoli cells-the somatic cells of the seminiferous epithelium. The distribution of nuclear pore complexes of the germ cells will then be described, beginning with gonocytes in the immature testis and proceeding to spermatogonia, spermatocytes, and spermatids of the mature testis. Sertoli cell nucleus
The Sertoli cells of the mature testis have nuclei of elaborate shape, deeply infolded or lobulated. In thin sections, they have a prominent nucleolus, often flanked by two clumps of nucleolus-associated heterochromatin. The remainder of the nucleoplasm is essentially free of heterochromatin. In freezefracture replicas, the nuclear envelope shows many pore complexes. These are uniformly and apparently randomly distributed (Figs. l-3). Maul has shown by analysis of nearest neighbor distances that this kind of distribution is not truly random, but in the absence of any discernible pattern, it will be convenient to describe it as such. The density of pores is much the same in Sertoli cells of all species studied, and shows no obvious variation with stage of the spermatogenic cycle. In no instance was local aggregation of pores observed in Sertoli cell nuclei. Gonocyte nucleus
In the seminiferous
cords of immature mice
NUCLEAR
PORES
IN
MALE
GERM
CELLS
and guinea-pigs, there are many large gonocytes with abundant cytoplasm of low density. The mitochondria tend to congregate with the Golgi complex on one side of the nucleus. Their nuclei are spheroidal and contain little heterochromatin in interphase, but since they are a proliferating population, cells are often encountered that have elongated, loosely organized fibrillogranular areas representing prophase chromosomes in section. In freeze-fracture replicas, the gonocytes are easily identified by their large size. The fracture face of the nucleus is usually smooth contoured, but through distortion in specimen preparation, may have irregular elevations and depressions. The distribution of nuclear pores is rather consistent among gonocytes. They tend to aggregate in clusters of irregular outline consisting of 1040 pores separated by areas of similar size without pores (Figs. 4, 5). A mosaic of small groupings of pores and intervening pore-free areas extends over the entire nuclear envelope. The total number of pores is high and in pore-rich areas, their center-to-center distance may be only 160-200 nm. Occasional areas of very regular hexagonal packing are observed. It is not possible in freeze-cleaved preparations to determine the stage of the cycle of any given nucleus but the consistency of the images observed suggests that this is either the interphase pattern of gonocytes or that pore distribution shows little variation during the cell cycle. Spermutogonial
nucleus
Spermatogonia are the predominant germ cell type in the immature testis, they are usually at the periphery of the seminiferous cords, often in contact with the basal lamina. This facilitates their identification in freezefracture replicas. In some cells, the distribution of pore complexes is apparently random, while in others their pattern resembles that of gonocytes (Fig. 6). But, in general, the pores are less closely aggregated than in gonocytes and regular geometric arrays are rarely observed. In the mature testis, the lower proportional representation of spermatogonia in the germ cell population makes them somewhat more difficult to locate but their proximity to the easily recognized lamina propria is helpful in their identification. As in the immature
149
testis, the spermatogonia fall into two categories with respect to pore distribution. those with a near random pattern and those with a moderate degree of aggregation. The limited area of nucleus exposed in the cell in Fig. 7 shows little or no pore aggregation. In the cell in Fig. 8 the pores show some tendency to aggregate with shorter center-tocenter distances, but there are no sizeable pore-free areas. Spermatocytes Identification of spermatocytes in freezefracture replicas is facilitated by their large size and their position some distance off the basement membrane. In addition, they have flat microcisternae occurring singly or in rouleaux (Fawcett, 1977). These organelles are found only in spermatocytes. and their presence in fractured cytoplasm permits unambiguous identification of this cell type. The fracture faces of the spermatocyte nucleus shows a highly characteristic distribution of pores. They are closely aggregated in conspicuous clusters separated by large pore-free regions (Figs. 9, IO). The aggregations of pores are much larger and more closely packed than those of gonocytes. One hundred and fifty or more pores form large macular areas of irregular outline or elongated tracts. The pores are closely spaced and often show hexagonal packing, linear arrays and other less geometric patterns with centerto-center spacing of 140-200 nm (Figs. 9, IO). The distinction between pore-rich and porepoor areas is very obvious and the boundaries more sharply demarcated than in any of the earlier germ cells. The conspicuous clustering of nuclear pores appears to persist only through the long prophase of meiosis in the primary spermatocytes. Secondary spermatocytes are present only very transiently. There are no dependable criteria for their identification in freeze-fracture replicas, and we have no observations on their nuclei. Spermatids Spermatids are identified in replicas by their adluminal position and by the presence of portions of the acrosome in the fracture face. In early spermatids the number of pores per unit area is much lower than in gonocytes, spermatogonia or spermatocytes and they appear to be randomly dispersed over
FAWCETT
150
the surface of the nucleus. In thin sections, a transient increase in concentration of pores has been described immediately adjacent to the chromatoid body (Fawcett et al., 1970; Fawcett, 1971) but this has not been observed in freeze-fracture preparations, probably because no spermatid at the appropriate stage was encountered. The random distribution of nuclear pores in early spermatids is soon modified in relation to the developing acrosome. At the site of attachment of the acrosomal vesicle, the outer and inner membranes of the nuclear envelope become very closely associated and the perinuclear space is reduced from some 25 to 5 nm, and pores disappear from this area. These local modifications of the
AND
CHEMES
nuclear envelope take place in anticipation of the attachment of the acrosomal vesicle and thus determine the anterior pole of the sperm nucleus (Sandoz, 1974). Concurrently with the attachment of the acrosomal vesicle, a shallow pore-free depression develops in the nuclear envelope (Fig. 11). This gradually expands as the acrosome enlarges and spreads over the anterior pole of the spermatid nucleus (Fig. 12). In freeze-fracture replicas of nuclear envelopes in general, the plane of cleavage often breaks out of one of the membranes and into the other producing in a sharp offset at the transition. This can be observed in several of the micrographs of spermatogonia and spermatocytes (Figs. 8, 10) and in the post-acrosomal region of
Figs. 1, 2. Freeze-fracture preparations testis. The pore complexes are numerous x 24,300.
of Sertoli cell nuclei from adult guinea-pig and apparently random in their distribution.
Fig. 3. Sertoli cell nucleus of a 5-day-old guinea-pig. The number of pores per unit area is considerably smaller than in the mature testis. This difference is consistent with the greater metabolic activity of the post-pubertal Sertoli cells. x 34,000. Fig. 4. Nucleus of a gonocyte from the testis of a 5-day-old guinea-pig. There is a striking clustering of pores in groups of lo-30 or more separated by pore-free areas. Since gonocytes have little peripheral heterochromatin, it is unlikely that the aggregations of pores are associated with condensed chromatin. x 25,400. Fig. 5. Nucleus of a gonocyte from a 5-day-old guinea-pig. The numerous pores show some degree of aggregation, but less than those of the gonocytes in earlier figures. x 17,000. Fig. 6. Eight-day-old mouse spermatogonium. rather uniformly with only minimal aggregation.
The nuclear x 35,100.
pores
are distributed
Fig. 7. Freeze-fractured adult guinea-pig spermatogonium. The fracture plane includes only a small portion of the nucleus but in the area shown, the center-to-center spacing of the pores is quite uniform and there is no evidence of aggregation. x 13,780. Fig. 8. Adult guinea-pig spermatogonium. The annuli of some of the nuclear pores project above the fracture face giving a spurious appearance of clustering. Pores in the same area that have fractured normally are relatively inconspicuous. These have been outlined with circles to permit more valid assessment of pore distribution. Pore distribution is non-random but pore aggregation is relatively inconspicuous. x 27,500. Fig. 9. Spermatocyte from a mature guinea-pig illustrating the large pore-rich and pore-poor areas characteristic of this stage of germ-cell development. In the pore-rich areas, hundreds of pores are packed with minimal center-to-center spacing. Only the E-face of the outer membrane is seen in this replica. x 22,500. Fig. 10. Another example of an adult guinea-pig spermatocyte nucleus showing large aggregations of closely spaced pores often exhibiting an hexagonal pattern. The relation of the pore-rich areas to the meiotic chromosomes remains uncertain. The fracture plane in this replica reveals the E-face of the outer membrane in some areas and the P-face of the inner membrane in others. x 33,750.
FAWCETT
156
round spermatids (Fig. 12). However, in the subacrosomal region where the two nuclear membranes are in close apposition and devoid of pores, this never occurs. It is always the outer nuclear membrane that is cleaved and it always presents a smooth face uninterrupted by pores or abrupt shifts of the fracture plane from one membrane to the other (Figs. 12-14). Pores continue to be present in the postacrosomal region during expansion of the acrosome and seem to increase in number per unit area. Whether this apparent change in pore density is due to a caudal displacement of pores in advance of the expanding acrosoma1 cap, or to formation of new pores, cannot be ascertained. However, immediately behind the posterior margin of the acrosome, small aggregations of membrane particles are occasionally observed (Figs. 13, inset). These assemblages of closely packed particles occur only in the immediate vicinity of the advancing edge of the acrosomal cap. Their diameter approximates that of nuclear pores and the spacing between aggregations is
AND
CHEMES
comparable to the average distance between pores. It is tempting to speculate that they represent intermediate stages in the formation or dissolution of nuclear pores. Discussion In most somatic cells, the nuclear pores are rather uniformly distributed but in very active cells with large numbers of pores per unit area, non-random patterns appear, the commonest being a regular hexagonal array with a constant center-to-center distance (Maul et al., 1971). Sertoli cells, the somatic cells of the seminiferous epithelium, have abundant pores that are quite evenly spaced and exhibit no definite pattern. The nuclei of the early stages of male germ cells, on the other hand, were all found to show varying degrees of aggregation into pore-rich and pore-poor areas. Clustering of pores was evident in gonocytes of the neonatal testis and in spermatogonia of animals approaching maturity. In adults, two categories of spermatogonial nuclei were found, some showing a
Fig. 11. Freeze-fracture replica of the anterior pole of the nucleus in a guinea-pig spermatid in an early phase of acrosomal development. In spermatids the number of pores per unit area of nuclear envelope is less than in spermatocytes and they are randomly dispersed. The shallow pore-free depression outlined by arrows is the site of attachment of the acrosomal vesicle. x 25,600. Fig. 12. A cap-phase guinea-pig cleaved through acrosome (above) and a lateral hemisphere of the nucleus. The fracture plane in the pore-free subacrosomal region displays the E-face of the outer membrane of the nuclear envelope. At the posterior margin of the acrosome (at the arrows), the fracture plane breaks into the inner membrane exhibiting its P-face, except in a small area at the lower right where the E-face of the outer membrane is seen. x 20,500. Fig. 13. Guinea-pig spermatid in which the posterior portion of the acrosome cap (upper right) is cross fractured. The fracture plane then steps down from the inner acrosomal membrane to the E-face of the outer nuclear membrane. The subacrosomal region is smooth contoured and devoid of pores while the post-acrosomal region has randomly distributed pores in moderate numbers. Immediately behind the linear depression marking the posterior margin of the acrosome are small pore-sized aggregations of intramembrane particles. These are seen at higher magnification in the inset. It is speculated that these may represent sites of dissolution of pores ahead of the advancing margin of the acrosome, or possibly sites of new pore formation. x 30,750. Inset: x 73,000. Fig. 14. In this guinea-pig spermatid, the advanced acrosome (upper right) has been cross fractured and the outer nuclear membrane has been cleaved showing its smooth pore-free subacrosomal region and the pores in the post-acrosomal nuclear envelope. The acrosomal cap in this spermatid is no longer enlarging, and the small aggregations of intramembrane particles illustrated in Fig. 13 are no longer seen. x 30,750.
NUCLEAR
PORES
IN
MALE
CiERM
CELLS
random pore distribution and others exhibiting varying degrees of pore aggregation. Whether these correspond to reserve and proliferating spermatogonia, respectively, could not be determined. The most conspicuous aggregation of pores was found in the large nuclei of spermatocytes which present a striking mosaic of pore-rich and pore-poor areas. A lower density and more random pattern is seen again in early spermatids, but this is soon modified in the course of acrosome development. As the acrosome expands over the anterior pole of the nucleus, the pores disappear from the underlying envelope and at the same time increase in density in the posterior hemisphere. After nuclear condensation and differentiation of the post-acrosomal dense lamina, the only remaining nuclear pores are concentrated in redundant folds of the nuclear envelope that extend from the posterior ring back into the neck region of the maturing spermatid. The changing pattern of pore complexes observed in the nuclei of germ cells raises a number of interesting questions. Is the pattern or pores related to the distribution of chromosomes ? If so, are the chromosomes related to the pore-free or the pore-rich areas of the nuclear envelope? Are the changes in pore distribution achieved by elimination of pores in some areas and formation of new ones to increase pore density in other sites? Or, does the fluid nature of the lipid bilayers in the membranes of the nuclear envelope allow the pore complexes to aggregate or disperse in response to changing conditions? If they are free to move, are they simply excluded from areas where high concentrations of chromatin accumulate against the membrane, or are they drawn together during chromosome condensation by virtue of attachment of chromatin fibers to the annuli. The massive clustering of nuclear pores in spermatocytes is associated temporally, if not causally, with prophase rearrangement of the chromosomes. Spermatocyte nuclei have been studied in serial sectioning in a number of invertebrate species (Moens, 1969; Church, 1976). In these species the chromosomes move within the prophase nucleus (Richards, 1975) to establish the familiar ‘bouquet’ arrangement in which they form long loops with their ends all clustered at one pole of the nucleus. The great majority of the
I59
nuclear pores are concentrated in the region where the chromosome ends attach to the nuclear envelope (Church, 1976). The actual attachment sites are small pore-free islands in a region of high pore density. In spermatocytes of these species, the Golgi complex, centrioles and a mass of mitochondria are also concentrated in the cytoplasm adjacent to the region of chromosome attachment, and it is not clear whether the conspicuous pore aggregation is related to prophase chromosomal rearrangement or is functionally related to the concentration of metabolically active organelles in the adjacent cytoplasm. The relationship of the chromosomal bouquet to the pore-rich area of the nuclear envelope in rodent spermatocytes studied here seems to be similar to that found in insects even though the aggregation of mitochondria in the neighboring cytoplasm is less pronounced. The exact relationship of the chromosomes to the nuclear envelope has long been a subject of controversy. In the interphase nuclei of many cell types, heterochromatic regions of the chromosomes are applied to the inner aspect of the nuclear envelope forming the densely staining peripheral clumps observed by light and electron microscopy. There is evidence that their attachment is quite firm since they resist displacement by high-speed centrifugation of living cells (Beams and Mueller, 1970), and after cell fractionation procedures for isolation of the nuclear envelope, clumps of chromatin remain adherent to the membrane. It is not clear whether the chromatin is attached directly to the inner membrane; to a fibrous lamina on its inner aspect; or to the annuli of the pore complexes. It has been reported that chromatin fibers converge upon nuclear pore complexes and attach to the annulus (DuPraw, I965 : Comings and Okada, 1970). It has even been suggested that the nuclear pore complex is essentially a permanent chromosomal element found wherever chromosomes make contact with the developing nuclear envelope in late anaphase and early telophase (Engelhardt and Pusa, 1972). This interpretation of annuli is difficult to bring into accord with the observation of multiple pore complexes present in residues of the nuclear envelope in the cytoplasm during metaphase or with the occurrence of annulate lamellae as a common cytoplasmic organelle. It seems evident that pore complexes can bc
160
assembled at sites remote from the chromosomes. Therefore it is unlikely that the annuli are primarily structural components of chromosomes which become incorporated in the nuclear envelope during its reconstitution. If the chromosomes and annuli are closely attached in the interphase nucleus, their association must take place during or after reconstitution of the nuclear envelope in telophase of mitotic division. Some investigators have suggested that fixation of interphase chromatin fibers to multiple sites on the nuclear envelope may predetermine the arrangement of chromatin during prophase condensation (Comings and Okada, 1970). This speculation assumes a relatively stable pattern of nuclear pore complexes and it is true that to date only relatively minor alterations in pore distribution have been reported during the mitotic cycle of somatic cells. There seems no doubt that there is normally some turnover of pore complexes in somatic cells, and their number can vary in different states of physiological activity (Maul et al., 1973), but their general pattern of distribution is probably maintained by attachment to the fibrous lamina. When isolated nuclear envelopes are extracted with detergent, the normal arrangement of pore complexes persists in the absence of membranes (Aaronson and Blobel, 1975). The peripheral dense lamina therefore seems to provide a more or less coherent fibrous framework that is responsible for the spatial organization of the pore complexes. In the original description of a fibrous lamina associated with the nuclear envelope of vertebrate cells (Fawcett, 1966), it was noted that this component varied in its degree of development in different cell types and it was not detectable in spermatocytes. Failure to visualize it in electron micrographs of thin sections, however, does not necessarily mean that it is absent. Indeed, in hepatic cells from which it was first isolated for chemical analysis, it was not previously described in thin sections. Nevertheless, the remarkable variation in distribution of pore complexes during the course of spermatogenesis and the apparent mobility of pore complexes observed in the present study encourage the speculation that development of this stabilizing component of the nuclear envelope may be lacking or rudimentary in the germ cell line.
FAWCETT
AND
CHEMES
Some degree of pore clustering has been reported in differentiated somatic cells which have a pronounced peripheral accumulation of heterochromatin (Teigler and Baerwald, 1972; Maul, 1975). On the other hand, in lymphocytes and HeLa cells, the clustering disappears when they enter S phase of their mitotic cycle (Markovics et al., 1974). A similar change has been reported in the transition of yeast from stationary cultures to exponential growth (Moore and Muhlenthaler, 1963). It has been suggested therefore that the distribution of chromatin may have an influence on pore distribution. If our speculation, that the germ cell line has a poorly developed stabilizing fibrous framework connecting the pore complexes, is correct, then the dramatic clustering of pores in meiotic prophase may indeed be a reflection of the distribution of the chromosomes in the spermatocyte nucleus. If the reported attachment of chromatin fibers to the annuli is valid, then chromatin condensation might very well result in aggregation of pores that are unrestrained by attachment to a fibrous lamina. The pore-rich areas of the spermatocyte nuclear envelope should then coincide with condensed chromatin in the underlying nucleoplasm. On the other hand, there are observations that favor the opposite interpretation. Pores are absent at sites of termination of the synaptonemal complex on the nuclear envelope (Moses, 1960, 1964) and they are absent adjacent to peripherally placed nucleoli (LaCour and Wells, 1972). They are not observed where cytoplasmic organelles such as the acrosome, or the capitulum of the sperm flagellum attach. An argument could be made that aggregation of pores in spermatocytes may be due to displacement by nuclear or cytoplasmic components attaching to the membrane. The pore-rich areas would then be between chromosomes. This would insure unobstructed access of metabolites to the pores for nucleocytoplasmic exchange. A clear choice between these opposing interpretations cannot be made at this time. If chromosomes are invariably attached to annuli, one would certainly expect pore aggregation to be consistently associated with condensed chromatin. However, the gonocytes and spermatogonia of the present study are cell types with little or no peripheral heterochromatin. Nevertheless they exhibit
NUCLEAR
PORES
IN
MALE
GERM
CELLS
considerable aggregation of their nuclear pores. This finding casts doubt upon the reported correlation of pore distribution with the state of condensation of the chromatin. The formation of nuclear pores is believed to involve local fusion of the outer and inner membranes of the nuclear envelope. We still know very little about the molecular events in membrane fusion. The process has been studied most thoroughly in examples of exocytosis where conflicting interpretations have emerged. In mucocyst discharge in the ciliate Tetrahymena, an annular accumulation of intramembrane particles is observed in the mucocyst membrane. This is in register with a rosette of particles in the overlying plasmalemma (Satir et al., 1973). In exocytosis of secretory granules of mast cells and pancreatic islet cells, on the other hand, freeze-fracture studies reveal a clearance of intramembrane particles from a sharply circumscribed area of both the cell membrane and the membrane limiting the secretory granule (Orci et al., 1977; Chi et al., 1976; Lawson et al., 1977). Thus in the membrane fusion involved in exocytosis, both intramembrane granule aggregation and granule clearance have been reported. In describing the formation of nuclear pores, Maul et al. (1971) published micrographs of thin sections showing small domeshaped convexities projecting from the outer membrane toward the inner, and in freezefracture replicas, particle-free areas were seen which were of about the same diameter as nuclear pores. These were interpreted as presumptive sites of membrane fusion that would result in the formation of a new pore. No similar areas of protrusion and particle clearance were observed in the post-acrosoma1 region of spermatids in the present study. The only ultrastructural features that might be associated with appearance or dissolution of nuclear pores were small aggregations of
161
membrane particles (Fig. 13). The fact that these were found only in the immediate vicinity of the advancing margin of the acrosomal cap suggests that they may represent a transient phase of pore dissolution rather than pore formation. The small size of the clustered particles and their variable number seem to exclude the interpretation that the 8 or 16 particles normally associated with each annulus had simply been incorporated in the membrane during pore dissolution. The subacrosomal region of the nuclear envelope is unique in the consistency of its mode of fracturing. The fracture plane preferentially traversed the outer membrane throughout the area covered by the acrosome. The fracture plane never crossed over from one membrane to the other as it commonly does in unspecialized areas of the nuclear envelope. The two membranes, devoid of pores in this region, are in close apposition and the inner element is supported by a condensation of the underlying nucleoplasm. It is not clear whether consistent and uninterrupted cleavage of one membrane is due to unique features of organization of the hydrophobic zone of the membranes in this region or is simply a mechanical consequence of the close apposition of the membranes and of the stiffening and internal reinforcement of the inner element. The observation seems worth recording for its possible incidental interest to membrane biologists employing the freeze-fracture method on other cell types. Acknowledgements Supported by the Ford Foundation and Grant HD-02344 from the National Institute of Child Health and Human Development (D.W.F.) and recipient of International Research Fellowship 5-FOS-TWO 2223-02 from the United States Public Health Service (H.E.C.).
References AAROKSON, R. P. and BLOBEL, G. 1974. On the attachment of the nuclear pore complex. J. Cell Biof., 62, 746-154. AARONSON,R. P. and BLOBEL, G. 1975. Isolation of nuclear complexes in association with a lamina. Proc. mtn. Acad. Sci. USA, 721, 1007-1011. BEAMS,H. W. and MUELLER, S. 1970. Effects of ultracentrifugation on the interphase nucleus of somatic cells with special reference to the nuclear envelope-chromatin relationship. Z. Zellforsch. mikrosk. Anar., 108, 297T.107. II
162 CHEMES,H. E., FAWCETT, D. W. and DYM, M. 1978. Unusual spermatogenic cells. Anal. Rec. (in press).
features
CHI, E. Y., LAGUNOPF, D. and KOEHI.ER, J. K. 1976. Freeze-fracture
FAWCETT
AND
of the nuclear
envelope
study of mast cell secretion.
CHEMES in human Proc. natn.
Acad. Sci. USA, 73, 2823-2827.
CAURCH, K. 1976. Arrangement of chromosome ends and axial core formation during early meiotic prophase in Brachystola magna by 3D E.M. reconstruction. Chromosoma, 58, 365-376. COMINGS, D. E. 1968. The rationale for an ordered arrangement of chromatin in the interphase nucleus. Am. J. hum. Genet., 20,440-460. COMINGS, D. E. and OKADA, T. 1970. Association Exp. Cell Res., 62, 293-302.
DUPRAW, E. J. 1965. The organization
of chromatin
fibers with the annuli of the nuclear envelope.
of nuclei and of chromosomes
in honey bee embryonic
cells. Proc.
natn. Acad. Sci. USA, 53, 161-166.
ENGLEHARDT,P. and PUSA, K. 1972. Nuclear pore complex: ‘Press stud’ elements of Chromosomes in pairing and control. Nature New Biol., 240, 163-167. FAWCETT, D. W. 1966. On the occurrence of a fibrous lamina on the inner aspect of the nuclear envelope in certain cells of vertebrates. Am. J. Anat., 119, 129-146. FAWCETT, D. W., EDDY, E. M. and PHILLIPS, D. 1970. Observations on the fine structure and origin of the chromatoid body in mammalian spermatogenesis. Biol. Reprod., 2, 129-153. FAWCETT, D. W. 1971. Observations on cell differentiation and organelle continuity in spermatogenesis. In Genetics OJ the Spermatozodn (R. A. Beatty and S. Gliicksohn-Waelch, eds) pp. 37-68, Edinburgh, North-Holland Publishing Co. FAWCETT, D. W. 1975. Morphogenesis of the mammalian sperm acrosome in new perspective. In Functional Anatomy of the Spermatozoon (ed. B. A. Afzelius), pp. 13-36. Pergamon Press, Oxford. FAWCETT, D. W. 1977. Unsolved problems in the morphogenesis of the mammalian spermatozoon. In International Ceil Biology 1976-1977. Rockefeller Press, New York. LACOUR, L. F. and WELLS, B. 1972. Nuclear pores of early meiotic prophase nuclei of plants. Z. Ze/[forsch. mikrosk.
Anat., 123, 178.
LAWSON, D., RAFF, M. C., GOMPERTS, B., FEWTRELL, C. and GILULA, N. B. 1977. Molecular events during membrane fusion. A Study of exocytosis in rat peritoneal mast cells. J. Cell Biol., 72, 242-259. MAIIKOVICS, J., GLASS, L. and MAUL, G. G. 1974. Pore patterns on nuclear membranes. Exp. Cell Res., 85, 443-451. MAUL, G. G., MAUL, H. M., SCOGNA, J. E., LIEBERMAN,M. W., STEIN, G. S., Hsu, B. Y. and BORUN, T. W. 1972. Time sequence of nuclear pore formation in PHA stimulated lymphocytes and in Hela cells during the cell cycle. J. Cell Biol., 55, 433. MAUL, G. G., PRICE, J. W. and LIEBERMAN,M. W. 1971. Formation and distribution of nuclear pore complexes in interphase. J. Cell Biol., 51, 405-418. MAUL, H. M., Hsu, B. Y. L., BORUN, T. M. and MAUL, G. G. 1975. Effect of metabolic inhibitors in nuclear pore formation during the Hela Ss cell cycle. J. Cell Biol., 59, 669-676. MOENS, P. B. 1969. The fine structure of meiotic chromosome polarization and pairing in Locusta migmtoria spermatocytes. Chromosoma, 28, l-25. MOORE, H. and MUHLENTHALER,K. 1963. Fine structure of frozen-etched yeast cells. J. Cell Biol., 17, 609628. MOSES, M. J. 1956. Chromosomal structures in crayfish spermatocytes. J. biophys. hiochem. Cytol., 2, 215218. MOSES, M. J. 1960. Breakdown and reformation of the nuclear envelope at cell division. 4th lnt. Con/: E/e&. Microsc. Proc., Berlin, 1958, pp. 230-233. Springer-Verlag. MOSES, M. J. 1964. The nucleus and chromosomes. In Cytology and Cell Physiology (ed. G. H. Bourne), 3rd edn, pp. 423-558. Academic Press, New York. ORCI, L., PERRELET, A. and FRIEND, D. S. 1977. Freeze-fracture of membrane fusions during exocytosis in pancreatic beta cells. J. Cell Biol., 75, 23-30. RICHARDS, G. K. 1975. Prophase chromosome movements in living house-cricket spermatocytes and their relationship to prophase, anaphase and granule movements. Chromosoma, 49, 407-445. SANDOZ, D. 1974. Modification of the nuclear envelope during spermiogenesis of Discoglosws pectus. J. Submicr.
Cytol., 6, 399-419.
TEIGLER, D. J. and BAERWALD, R. J. 1972. A freeze-etch 447-456.
study of clustered
nuclear
pores. Tissue & Cell, 4,
DON W. FAWCETT
and HECTOR
E. CHEMES
CHANGES IN DISTRIBUTION OF NUCLEAR PORES DURING DIFFERENTIATION OF THE MALE GERM CELLS ABSTRACT. Changes in number of nuclear pores in different states of physiologica activity have been reported, but little is known about changing patterns of distribution in the course of cell ditferentiation. Pore distribution in male germ cells was studied in freeze fracture preparations of immature and mature rodent testis. As in other somatic cells, pores were uniformly and apparently randomly distributed in Sertoli cell nuclei. The nucleus of gonocytes and spermatogonia showed varying degrees of pore clustering. Spermatocytes invariably exhibited very striking pore aggregation with close hexagonal packing in pore-rich areas, and large pore-free areas. In early spermatids, pores appeared to be randomly distributed. As the acrosome formed and spread over the apical pole of the nucleus, pores disappeared ahead of its advancing margin and became more concentrated in the post-acrosomal region. The relationship of pore complexes to the chromosomes and the role of the fibrous lamina are discussed. The question as to whether the changing patterns observed involve movement of pores within fluid nuclear membranes, or a dissolution and reformation of new pores remains unanswered.
formation in spermatids have also been described (Fawcett, 1975; Sandoz, 1974). To our knowledge, however, there has been no study concerned specifically with changes in the nuclear envelope throughout the course of a complex process of cell differentiation. The present investigation was undertaken to follow in freeze-fracture replicas the changing pattern of nuclear pore distribution in the germ cells during spermatogenesis in rats, mice and guinea-pigs.
Introduction LITTLE is known about the mechanism of nuclear pore formation or the functional significance of different arrangements of pores in the nuclear envelope. Studies on various somatic cell types have shown that there is an approximate doubling of pores in the interphase nucleus associated with the increase in nuclear volume and replication of DNA prior to cell division (Maul et al., 1972). An increase in pore number has also been reported for cells stimulated to greater metabolic and synthetic activity (Moses, 1964; Maul et al., 1972). In a previous paper from this laboratory, attention was drawn to a non-random distribution of nuclear pores in human spermatogenic cells studied in thin sections (Chemes et al., 1978). Changes in pore distribution associated with acrosome
Materials and Methods The freeze-cleaving method of specimen preparation affords extensive en ,fnce views of the nuclear envelope and is therefore especially useful for studying the number and distribution of nuclear pore complexes. The observations reported here are based upon study of freeze-fracture replicas of the testes of immature (5day and ICday old) and mature guinea pigs of the Duncan Hartley strain: immature mice (2-day and &day old) of the CD-I strain; and pubescent rats (33-day old) of the Sprague-Dawley strain.
Department of Anatomy and Laboratory of Human Reproduction and Reproductive Biology, Harvard Medicai School, Boston, Massachusetts 071 IS. Received 5 October
1978. 147
FAWCETT
148
Adult guinea-pig and 33-day old rat testes were fixed by vascular perfusion either via the abdominal aorta or through the testicular artery. Testes from immature guinea-pigs and mice were fixed by immersion in collidinebuffered (pH 7.4) 5 % glutaraldehyde, or a glutaraldehyde and paraformaldehyde mixture. The results were found to be similar with the two fixatives. After a brief fixation of approximately 30 min duration, small blocks of tissue were washed in buffer, transferred to a 20% solution of glycerin in distilled water for 2 hr, and then in frozen liquid Freon 22 cooled in liquid nitrogen. The blocks were cleaved at - 125” and the exposed surfaces coated with carbon, and platinum shadowed, in a Balzer’s apparatus. The tissue was digested in three successive baths of Clorox, and after careful washing, the replicas were mounted on uncoated copper grids. Samples of the same tissue were left in the original fixative for an additional 2 hr, postfixed in collidine-buffered 1.25% 0~04, dehydrated in ethanol and embedded in an Epon-Araldite mixture. Thick (1 pm) sections were mounted on glass slides and stained with 1 o/0 toluidine blue in borax. Thin sections exhibiting pale gold interference colors were mounted on uncoated copper grids and double stained with uranyl acetate and lead citrate. Both replicas and sections were studied in a Philips EM 200 electron microscope at 60 kV. Results
Testis development and onset of spermatogenesis are much the same in the three mammalian species studied although the time course differs. Gonocytes in the seminiferous cords proliferate throughout fetal life and subsequently move to the periphery of the cords and there give rise to spermatogonia. With the onset of spermatogenesis at puberty, the spermatogonia differentiate into two main categories: those that divide at a slow rate to maintain the stem cell population; and others that proliferate and differentiate to form syncytial clusters of spermatocytes. After a protracted meiotic prophase, the spermatocytes undergo two divisions resulting in the formation of large groups of haploid round spermatids. During spermiogenesis, these are gradually transformed into spermatozoa which are released into the lumen at spermiation.
AND CHEMES
In freeze-fracturing, the path of cleavage preferentially follows the hydrophobic interior of the lipid bilayer in the membranes of cells. Thus the cleavage plane reveals interesting details of internal structure of the membranes both at the surface and in the interior of cells. Where the fracture breaks across the cytoplasmic matrix or nucleoplasm in its path from one membranous organelle to another, the replicas provide little useful information. Identification of the different kinds of cells in the seminiferous epithelium in freeze-fracture preparations depends upon their location in the epithelium and upon recognition of specific membranous cytological features of each cell type, learned from previous electron microscopic study of the testis in thin sections. For the experienced student of spermatogenesis, identification of the several cell types poses no great difficulty. For purposes of comparison, we begin with a description of the nuclei of Sertoli cells-the somatic cells of the seminiferous epithelium. The distribution of nuclear pore complexes of the germ cells will then be described, beginning with gonocytes in the immature testis and proceeding to spermatogonia, spermatocytes, and spermatids of the mature testis. Sertoli cell nucleus
The Sertoli cells of the mature testis have nuclei of elaborate shape, deeply infolded or lobulated. In thin sections, they have a prominent nucleolus, often flanked by two clumps of nucleolus-associated heterochromatin. The remainder of the nucleoplasm is essentially free of heterochromatin. In freezefracture replicas, the nuclear envelope shows many pore complexes. These are uniformly and apparently randomly distributed (Figs. l-3). Maul has shown by analysis of nearest neighbor distances that this kind of distribution is not truly random, but in the absence of any discernible pattern, it will be convenient to describe it as such. The density of pores is much the same in Sertoli cells of all species studied, and shows no obvious variation with stage of the spermatogenic cycle. In no instance was local aggregation of pores observed in Sertoli cell nuclei. Gonocyte nucleus
In the seminiferous
cords of immature mice
NUCLEAR
PORES
IN
MALE
GERM
CELLS
and guinea-pigs, there are many large gonocytes with abundant cytoplasm of low density. The mitochondria tend to congregate with the Golgi complex on one side of the nucleus. Their nuclei are spheroidal and contain little heterochromatin in interphase, but since they are a proliferating population, cells are often encountered that have elongated, loosely organized fibrillogranular areas representing prophase chromosomes in section. In freeze-fracture replicas, the gonocytes are easily identified by their large size. The fracture face of the nucleus is usually smooth contoured, but through distortion in specimen preparation, may have irregular elevations and depressions. The distribution of nuclear pores is rather consistent among gonocytes. They tend to aggregate in clusters of irregular outline consisting of 1040 pores separated by areas of similar size without pores (Figs. 4, 5). A mosaic of small groupings of pores and intervening pore-free areas extends over the entire nuclear envelope. The total number of pores is high and in pore-rich areas, their center-to-center distance may be only 160-200 nm. Occasional areas of very regular hexagonal packing are observed. It is not possible in freeze-cleaved preparations to determine the stage of the cycle of any given nucleus but the consistency of the images observed suggests that this is either the interphase pattern of gonocytes or that pore distribution shows little variation during the cell cycle. Spermutogonial
nucleus
Spermatogonia are the predominant germ cell type in the immature testis, they are usually at the periphery of the seminiferous cords, often in contact with the basal lamina. This facilitates their identification in freezefracture replicas. In some cells, the distribution of pore complexes is apparently random, while in others their pattern resembles that of gonocytes (Fig. 6). But, in general, the pores are less closely aggregated than in gonocytes and regular geometric arrays are rarely observed. In the mature testis, the lower proportional representation of spermatogonia in the germ cell population makes them somewhat more difficult to locate but their proximity to the easily recognized lamina propria is helpful in their identification. As in the immature
149
testis, the spermatogonia fall into two categories with respect to pore distribution. those with a near random pattern and those with a moderate degree of aggregation. The limited area of nucleus exposed in the cell in Fig. 7 shows little or no pore aggregation. In the cell in Fig. 8 the pores show some tendency to aggregate with shorter center-tocenter distances, but there are no sizeable pore-free areas. Spermatocytes Identification of spermatocytes in freezefracture replicas is facilitated by their large size and their position some distance off the basement membrane. In addition, they have flat microcisternae occurring singly or in rouleaux (Fawcett, 1977). These organelles are found only in spermatocytes. and their presence in fractured cytoplasm permits unambiguous identification of this cell type. The fracture faces of the spermatocyte nucleus shows a highly characteristic distribution of pores. They are closely aggregated in conspicuous clusters separated by large pore-free regions (Figs. 9, IO). The aggregations of pores are much larger and more closely packed than those of gonocytes. One hundred and fifty or more pores form large macular areas of irregular outline or elongated tracts. The pores are closely spaced and often show hexagonal packing, linear arrays and other less geometric patterns with centerto-center spacing of 140-200 nm (Figs. 9, IO). The distinction between pore-rich and porepoor areas is very obvious and the boundaries more sharply demarcated than in any of the earlier germ cells. The conspicuous clustering of nuclear pores appears to persist only through the long prophase of meiosis in the primary spermatocytes. Secondary spermatocytes are present only very transiently. There are no dependable criteria for their identification in freeze-fracture replicas, and we have no observations on their nuclei. Spermatids Spermatids are identified in replicas by their adluminal position and by the presence of portions of the acrosome in the fracture face. In early spermatids the number of pores per unit area is much lower than in gonocytes, spermatogonia or spermatocytes and they appear to be randomly dispersed over
FAWCETT
150
the surface of the nucleus. In thin sections, a transient increase in concentration of pores has been described immediately adjacent to the chromatoid body (Fawcett et al., 1970; Fawcett, 1971) but this has not been observed in freeze-fracture preparations, probably because no spermatid at the appropriate stage was encountered. The random distribution of nuclear pores in early spermatids is soon modified in relation to the developing acrosome. At the site of attachment of the acrosomal vesicle, the outer and inner membranes of the nuclear envelope become very closely associated and the perinuclear space is reduced from some 25 to 5 nm, and pores disappear from this area. These local modifications of the
AND
CHEMES
nuclear envelope take place in anticipation of the attachment of the acrosomal vesicle and thus determine the anterior pole of the sperm nucleus (Sandoz, 1974). Concurrently with the attachment of the acrosomal vesicle, a shallow pore-free depression develops in the nuclear envelope (Fig. 11). This gradually expands as the acrosome enlarges and spreads over the anterior pole of the spermatid nucleus (Fig. 12). In freeze-fracture replicas of nuclear envelopes in general, the plane of cleavage often breaks out of one of the membranes and into the other producing in a sharp offset at the transition. This can be observed in several of the micrographs of spermatogonia and spermatocytes (Figs. 8, 10) and in the post-acrosomal region of
Figs. 1, 2. Freeze-fracture preparations testis. The pore complexes are numerous x 24,300.
of Sertoli cell nuclei from adult guinea-pig and apparently random in their distribution.
Fig. 3. Sertoli cell nucleus of a 5-day-old guinea-pig. The number of pores per unit area is considerably smaller than in the mature testis. This difference is consistent with the greater metabolic activity of the post-pubertal Sertoli cells. x 34,000. Fig. 4. Nucleus of a gonocyte from the testis of a 5-day-old guinea-pig. There is a striking clustering of pores in groups of lo-30 or more separated by pore-free areas. Since gonocytes have little peripheral heterochromatin, it is unlikely that the aggregations of pores are associated with condensed chromatin. x 25,400. Fig. 5. Nucleus of a gonocyte from a 5-day-old guinea-pig. The numerous pores show some degree of aggregation, but less than those of the gonocytes in earlier figures. x 17,000. Fig. 6. Eight-day-old mouse spermatogonium. rather uniformly with only minimal aggregation.
The nuclear x 35,100.
pores
are distributed
Fig. 7. Freeze-fractured adult guinea-pig spermatogonium. The fracture plane includes only a small portion of the nucleus but in the area shown, the center-to-center spacing of the pores is quite uniform and there is no evidence of aggregation. x 13,780. Fig. 8. Adult guinea-pig spermatogonium. The annuli of some of the nuclear pores project above the fracture face giving a spurious appearance of clustering. Pores in the same area that have fractured normally are relatively inconspicuous. These have been outlined with circles to permit more valid assessment of pore distribution. Pore distribution is non-random but pore aggregation is relatively inconspicuous. x 27,500. Fig. 9. Spermatocyte from a mature guinea-pig illustrating the large pore-rich and pore-poor areas characteristic of this stage of germ-cell development. In the pore-rich areas, hundreds of pores are packed with minimal center-to-center spacing. Only the E-face of the outer membrane is seen in this replica. x 22,500. Fig. 10. Another example of an adult guinea-pig spermatocyte nucleus showing large aggregations of closely spaced pores often exhibiting an hexagonal pattern. The relation of the pore-rich areas to the meiotic chromosomes remains uncertain. The fracture plane in this replica reveals the E-face of the outer membrane in some areas and the P-face of the inner membrane in others. x 33,750.
FAWCETT
156
round spermatids (Fig. 12). However, in the subacrosomal region where the two nuclear membranes are in close apposition and devoid of pores, this never occurs. It is always the outer nuclear membrane that is cleaved and it always presents a smooth face uninterrupted by pores or abrupt shifts of the fracture plane from one membrane to the other (Figs. 12-14). Pores continue to be present in the postacrosomal region during expansion of the acrosome and seem to increase in number per unit area. Whether this apparent change in pore density is due to a caudal displacement of pores in advance of the expanding acrosoma1 cap, or to formation of new pores, cannot be ascertained. However, immediately behind the posterior margin of the acrosome, small aggregations of membrane particles are occasionally observed (Figs. 13, inset). These assemblages of closely packed particles occur only in the immediate vicinity of the advancing edge of the acrosomal cap. Their diameter approximates that of nuclear pores and the spacing between aggregations is
AND
CHEMES
comparable to the average distance between pores. It is tempting to speculate that they represent intermediate stages in the formation or dissolution of nuclear pores. Discussion In most somatic cells, the nuclear pores are rather uniformly distributed but in very active cells with large numbers of pores per unit area, non-random patterns appear, the commonest being a regular hexagonal array with a constant center-to-center distance (Maul et al., 1971). Sertoli cells, the somatic cells of the seminiferous epithelium, have abundant pores that are quite evenly spaced and exhibit no definite pattern. The nuclei of the early stages of male germ cells, on the other hand, were all found to show varying degrees of aggregation into pore-rich and pore-poor areas. Clustering of pores was evident in gonocytes of the neonatal testis and in spermatogonia of animals approaching maturity. In adults, two categories of spermatogonial nuclei were found, some showing a
Fig. 11. Freeze-fracture replica of the anterior pole of the nucleus in a guinea-pig spermatid in an early phase of acrosomal development. In spermatids the number of pores per unit area of nuclear envelope is less than in spermatocytes and they are randomly dispersed. The shallow pore-free depression outlined by arrows is the site of attachment of the acrosomal vesicle. x 25,600. Fig. 12. A cap-phase guinea-pig cleaved through acrosome (above) and a lateral hemisphere of the nucleus. The fracture plane in the pore-free subacrosomal region displays the E-face of the outer membrane of the nuclear envelope. At the posterior margin of the acrosome (at the arrows), the fracture plane breaks into the inner membrane exhibiting its P-face, except in a small area at the lower right where the E-face of the outer membrane is seen. x 20,500. Fig. 13. Guinea-pig spermatid in which the posterior portion of the acrosome cap (upper right) is cross fractured. The fracture plane then steps down from the inner acrosomal membrane to the E-face of the outer nuclear membrane. The subacrosomal region is smooth contoured and devoid of pores while the post-acrosomal region has randomly distributed pores in moderate numbers. Immediately behind the linear depression marking the posterior margin of the acrosome are small pore-sized aggregations of intramembrane particles. These are seen at higher magnification in the inset. It is speculated that these may represent sites of dissolution of pores ahead of the advancing margin of the acrosome, or possibly sites of new pore formation. x 30,750. Inset: x 73,000. Fig. 14. In this guinea-pig spermatid, the advanced acrosome (upper right) has been cross fractured and the outer nuclear membrane has been cleaved showing its smooth pore-free subacrosomal region and the pores in the post-acrosomal nuclear envelope. The acrosomal cap in this spermatid is no longer enlarging, and the small aggregations of intramembrane particles illustrated in Fig. 13 are no longer seen. x 30,750.
NUCLEAR
PORES
IN
MALE
CiERM
CELLS
random pore distribution and others exhibiting varying degrees of pore aggregation. Whether these correspond to reserve and proliferating spermatogonia, respectively, could not be determined. The most conspicuous aggregation of pores was found in the large nuclei of spermatocytes which present a striking mosaic of pore-rich and pore-poor areas. A lower density and more random pattern is seen again in early spermatids, but this is soon modified in the course of acrosome development. As the acrosome expands over the anterior pole of the nucleus, the pores disappear from the underlying envelope and at the same time increase in density in the posterior hemisphere. After nuclear condensation and differentiation of the post-acrosomal dense lamina, the only remaining nuclear pores are concentrated in redundant folds of the nuclear envelope that extend from the posterior ring back into the neck region of the maturing spermatid. The changing pattern of pore complexes observed in the nuclei of germ cells raises a number of interesting questions. Is the pattern or pores related to the distribution of chromosomes ? If so, are the chromosomes related to the pore-free or the pore-rich areas of the nuclear envelope? Are the changes in pore distribution achieved by elimination of pores in some areas and formation of new ones to increase pore density in other sites? Or, does the fluid nature of the lipid bilayers in the membranes of the nuclear envelope allow the pore complexes to aggregate or disperse in response to changing conditions? If they are free to move, are they simply excluded from areas where high concentrations of chromatin accumulate against the membrane, or are they drawn together during chromosome condensation by virtue of attachment of chromatin fibers to the annuli. The massive clustering of nuclear pores in spermatocytes is associated temporally, if not causally, with prophase rearrangement of the chromosomes. Spermatocyte nuclei have been studied in serial sectioning in a number of invertebrate species (Moens, 1969; Church, 1976). In these species the chromosomes move within the prophase nucleus (Richards, 1975) to establish the familiar ‘bouquet’ arrangement in which they form long loops with their ends all clustered at one pole of the nucleus. The great majority of the
I59
nuclear pores are concentrated in the region where the chromosome ends attach to the nuclear envelope (Church, 1976). The actual attachment sites are small pore-free islands in a region of high pore density. In spermatocytes of these species, the Golgi complex, centrioles and a mass of mitochondria are also concentrated in the cytoplasm adjacent to the region of chromosome attachment, and it is not clear whether the conspicuous pore aggregation is related to prophase chromosomal rearrangement or is functionally related to the concentration of metabolically active organelles in the adjacent cytoplasm. The relationship of the chromosomal bouquet to the pore-rich area of the nuclear envelope in rodent spermatocytes studied here seems to be similar to that found in insects even though the aggregation of mitochondria in the neighboring cytoplasm is less pronounced. The exact relationship of the chromosomes to the nuclear envelope has long been a subject of controversy. In the interphase nuclei of many cell types, heterochromatic regions of the chromosomes are applied to the inner aspect of the nuclear envelope forming the densely staining peripheral clumps observed by light and electron microscopy. There is evidence that their attachment is quite firm since they resist displacement by high-speed centrifugation of living cells (Beams and Mueller, 1970), and after cell fractionation procedures for isolation of the nuclear envelope, clumps of chromatin remain adherent to the membrane. It is not clear whether the chromatin is attached directly to the inner membrane; to a fibrous lamina on its inner aspect; or to the annuli of the pore complexes. It has been reported that chromatin fibers converge upon nuclear pore complexes and attach to the annulus (DuPraw, I965 : Comings and Okada, 1970). It has even been suggested that the nuclear pore complex is essentially a permanent chromosomal element found wherever chromosomes make contact with the developing nuclear envelope in late anaphase and early telophase (Engelhardt and Pusa, 1972). This interpretation of annuli is difficult to bring into accord with the observation of multiple pore complexes present in residues of the nuclear envelope in the cytoplasm during metaphase or with the occurrence of annulate lamellae as a common cytoplasmic organelle. It seems evident that pore complexes can bc
160
assembled at sites remote from the chromosomes. Therefore it is unlikely that the annuli are primarily structural components of chromosomes which become incorporated in the nuclear envelope during its reconstitution. If the chromosomes and annuli are closely attached in the interphase nucleus, their association must take place during or after reconstitution of the nuclear envelope in telophase of mitotic division. Some investigators have suggested that fixation of interphase chromatin fibers to multiple sites on the nuclear envelope may predetermine the arrangement of chromatin during prophase condensation (Comings and Okada, 1970). This speculation assumes a relatively stable pattern of nuclear pore complexes and it is true that to date only relatively minor alterations in pore distribution have been reported during the mitotic cycle of somatic cells. There seems no doubt that there is normally some turnover of pore complexes in somatic cells, and their number can vary in different states of physiological activity (Maul et al., 1973), but their general pattern of distribution is probably maintained by attachment to the fibrous lamina. When isolated nuclear envelopes are extracted with detergent, the normal arrangement of pore complexes persists in the absence of membranes (Aaronson and Blobel, 1975). The peripheral dense lamina therefore seems to provide a more or less coherent fibrous framework that is responsible for the spatial organization of the pore complexes. In the original description of a fibrous lamina associated with the nuclear envelope of vertebrate cells (Fawcett, 1966), it was noted that this component varied in its degree of development in different cell types and it was not detectable in spermatocytes. Failure to visualize it in electron micrographs of thin sections, however, does not necessarily mean that it is absent. Indeed, in hepatic cells from which it was first isolated for chemical analysis, it was not previously described in thin sections. Nevertheless, the remarkable variation in distribution of pore complexes during the course of spermatogenesis and the apparent mobility of pore complexes observed in the present study encourage the speculation that development of this stabilizing component of the nuclear envelope may be lacking or rudimentary in the germ cell line.
FAWCETT
AND
CHEMES
Some degree of pore clustering has been reported in differentiated somatic cells which have a pronounced peripheral accumulation of heterochromatin (Teigler and Baerwald, 1972; Maul, 1975). On the other hand, in lymphocytes and HeLa cells, the clustering disappears when they enter S phase of their mitotic cycle (Markovics et al., 1974). A similar change has been reported in the transition of yeast from stationary cultures to exponential growth (Moore and Muhlenthaler, 1963). It has been suggested therefore that the distribution of chromatin may have an influence on pore distribution. If our speculation, that the germ cell line has a poorly developed stabilizing fibrous framework connecting the pore complexes, is correct, then the dramatic clustering of pores in meiotic prophase may indeed be a reflection of the distribution of the chromosomes in the spermatocyte nucleus. If the reported attachment of chromatin fibers to the annuli is valid, then chromatin condensation might very well result in aggregation of pores that are unrestrained by attachment to a fibrous lamina. The pore-rich areas of the spermatocyte nuclear envelope should then coincide with condensed chromatin in the underlying nucleoplasm. On the other hand, there are observations that favor the opposite interpretation. Pores are absent at sites of termination of the synaptonemal complex on the nuclear envelope (Moses, 1960, 1964) and they are absent adjacent to peripherally placed nucleoli (LaCour and Wells, 1972). They are not observed where cytoplasmic organelles such as the acrosome, or the capitulum of the sperm flagellum attach. An argument could be made that aggregation of pores in spermatocytes may be due to displacement by nuclear or cytoplasmic components attaching to the membrane. The pore-rich areas would then be between chromosomes. This would insure unobstructed access of metabolites to the pores for nucleocytoplasmic exchange. A clear choice between these opposing interpretations cannot be made at this time. If chromosomes are invariably attached to annuli, one would certainly expect pore aggregation to be consistently associated with condensed chromatin. However, the gonocytes and spermatogonia of the present study are cell types with little or no peripheral heterochromatin. Nevertheless they exhibit
NUCLEAR
PORES
IN
MALE
GERM
CELLS
considerable aggregation of their nuclear pores. This finding casts doubt upon the reported correlation of pore distribution with the state of condensation of the chromatin. The formation of nuclear pores is believed to involve local fusion of the outer and inner membranes of the nuclear envelope. We still know very little about the molecular events in membrane fusion. The process has been studied most thoroughly in examples of exocytosis where conflicting interpretations have emerged. In mucocyst discharge in the ciliate Tetrahymena, an annular accumulation of intramembrane particles is observed in the mucocyst membrane. This is in register with a rosette of particles in the overlying plasmalemma (Satir et al., 1973). In exocytosis of secretory granules of mast cells and pancreatic islet cells, on the other hand, freeze-fracture studies reveal a clearance of intramembrane particles from a sharply circumscribed area of both the cell membrane and the membrane limiting the secretory granule (Orci et al., 1977; Chi et al., 1976; Lawson et al., 1977). Thus in the membrane fusion involved in exocytosis, both intramembrane granule aggregation and granule clearance have been reported. In describing the formation of nuclear pores, Maul et al. (1971) published micrographs of thin sections showing small domeshaped convexities projecting from the outer membrane toward the inner, and in freezefracture replicas, particle-free areas were seen which were of about the same diameter as nuclear pores. These were interpreted as presumptive sites of membrane fusion that would result in the formation of a new pore. No similar areas of protrusion and particle clearance were observed in the post-acrosoma1 region of spermatids in the present study. The only ultrastructural features that might be associated with appearance or dissolution of nuclear pores were small aggregations of
161
membrane particles (Fig. 13). The fact that these were found only in the immediate vicinity of the advancing margin of the acrosomal cap suggests that they may represent a transient phase of pore dissolution rather than pore formation. The small size of the clustered particles and their variable number seem to exclude the interpretation that the 8 or 16 particles normally associated with each annulus had simply been incorporated in the membrane during pore dissolution. The subacrosomal region of the nuclear envelope is unique in the consistency of its mode of fracturing. The fracture plane preferentially traversed the outer membrane throughout the area covered by the acrosome. The fracture plane never crossed over from one membrane to the other as it commonly does in unspecialized areas of the nuclear envelope. The two membranes, devoid of pores in this region, are in close apposition and the inner element is supported by a condensation of the underlying nucleoplasm. It is not clear whether consistent and uninterrupted cleavage of one membrane is due to unique features of organization of the hydrophobic zone of the membranes in this region or is simply a mechanical consequence of the close apposition of the membranes and of the stiffening and internal reinforcement of the inner element. The observation seems worth recording for its possible incidental interest to membrane biologists employing the freeze-fracture method on other cell types. Acknowledgements Supported by the Ford Foundation and Grant HD-02344 from the National Institute of Child Health and Human Development (D.W.F.) and recipient of International Research Fellowship 5-FOS-TWO 2223-02 from the United States Public Health Service (H.E.C.).
References AAROKSON, R. P. and BLOBEL, G. 1974. On the attachment of the nuclear pore complex. J. Cell Biof., 62, 746-154. AARONSON,R. P. and BLOBEL, G. 1975. Isolation of nuclear complexes in association with a lamina. Proc. mtn. Acad. Sci. USA, 721, 1007-1011. BEAMS,H. W. and MUELLER, S. 1970. Effects of ultracentrifugation on the interphase nucleus of somatic cells with special reference to the nuclear envelope-chromatin relationship. Z. Zellforsch. mikrosk. Anar., 108, 297T.107. II
162 CHEMES,H. E., FAWCETT, D. W. and DYM, M. 1978. Unusual spermatogenic cells. Anal. Rec. (in press).
features
CHI, E. Y., LAGUNOPF, D. and KOEHI.ER, J. K. 1976. Freeze-fracture
FAWCETT
AND
of the nuclear
envelope
study of mast cell secretion.
CHEMES in human Proc. natn.
Acad. Sci. USA, 73, 2823-2827.
CAURCH, K. 1976. Arrangement of chromosome ends and axial core formation during early meiotic prophase in Brachystola magna by 3D E.M. reconstruction. Chromosoma, 58, 365-376. COMINGS, D. E. 1968. The rationale for an ordered arrangement of chromatin in the interphase nucleus. Am. J. hum. Genet., 20,440-460. COMINGS, D. E. and OKADA, T. 1970. Association Exp. Cell Res., 62, 293-302.
DUPRAW, E. J. 1965. The organization
of chromatin
fibers with the annuli of the nuclear envelope.
of nuclei and of chromosomes
in honey bee embryonic
cells. Proc.
natn. Acad. Sci. USA, 53, 161-166.
ENGLEHARDT,P. and PUSA, K. 1972. Nuclear pore complex: ‘Press stud’ elements of Chromosomes in pairing and control. Nature New Biol., 240, 163-167. FAWCETT, D. W. 1966. On the occurrence of a fibrous lamina on the inner aspect of the nuclear envelope in certain cells of vertebrates. Am. J. Anat., 119, 129-146. FAWCETT, D. W., EDDY, E. M. and PHILLIPS, D. 1970. Observations on the fine structure and origin of the chromatoid body in mammalian spermatogenesis. Biol. Reprod., 2, 129-153. FAWCETT, D. W. 1971. Observations on cell differentiation and organelle continuity in spermatogenesis. In Genetics OJ the Spermatozodn (R. A. Beatty and S. Gliicksohn-Waelch, eds) pp. 37-68, Edinburgh, North-Holland Publishing Co. FAWCETT, D. W. 1975. Morphogenesis of the mammalian sperm acrosome in new perspective. In Functional Anatomy of the Spermatozoon (ed. B. A. Afzelius), pp. 13-36. Pergamon Press, Oxford. FAWCETT, D. W. 1977. Unsolved problems in the morphogenesis of the mammalian spermatozoon. In International Ceil Biology 1976-1977. Rockefeller Press, New York. LACOUR, L. F. and WELLS, B. 1972. Nuclear pores of early meiotic prophase nuclei of plants. Z. Ze/[forsch. mikrosk.
Anat., 123, 178.
LAWSON, D., RAFF, M. C., GOMPERTS, B., FEWTRELL, C. and GILULA, N. B. 1977. Molecular events during membrane fusion. A Study of exocytosis in rat peritoneal mast cells. J. Cell Biol., 72, 242-259. MAIIKOVICS, J., GLASS, L. and MAUL, G. G. 1974. Pore patterns on nuclear membranes. Exp. Cell Res., 85, 443-451. MAUL, G. G., MAUL, H. M., SCOGNA, J. E., LIEBERMAN,M. W., STEIN, G. S., Hsu, B. Y. and BORUN, T. W. 1972. Time sequence of nuclear pore formation in PHA stimulated lymphocytes and in Hela cells during the cell cycle. J. Cell Biol., 55, 433. MAUL, G. G., PRICE, J. W. and LIEBERMAN,M. W. 1971. Formation and distribution of nuclear pore complexes in interphase. J. Cell Biol., 51, 405-418. MAUL, H. M., Hsu, B. Y. L., BORUN, T. M. and MAUL, G. G. 1975. Effect of metabolic inhibitors in nuclear pore formation during the Hela Ss cell cycle. J. Cell Biol., 59, 669-676. MOENS, P. B. 1969. The fine structure of meiotic chromosome polarization and pairing in Locusta migmtoria spermatocytes. Chromosoma, 28, l-25. MOORE, H. and MUHLENTHALER,K. 1963. Fine structure of frozen-etched yeast cells. J. Cell Biol., 17, 609628. MOSES, M. J. 1956. Chromosomal structures in crayfish spermatocytes. J. biophys. hiochem. Cytol., 2, 215218. MOSES, M. J. 1960. Breakdown and reformation of the nuclear envelope at cell division. 4th lnt. Con/: E/e&. Microsc. Proc., Berlin, 1958, pp. 230-233. Springer-Verlag. MOSES, M. J. 1964. The nucleus and chromosomes. In Cytology and Cell Physiology (ed. G. H. Bourne), 3rd edn, pp. 423-558. Academic Press, New York. ORCI, L., PERRELET, A. and FRIEND, D. S. 1977. Freeze-fracture of membrane fusions during exocytosis in pancreatic beta cells. J. Cell Biol., 75, 23-30. RICHARDS, G. K. 1975. Prophase chromosome movements in living house-cricket spermatocytes and their relationship to prophase, anaphase and granule movements. Chromosoma, 49, 407-445. SANDOZ, D. 1974. Modification of the nuclear envelope during spermiogenesis of Discoglosws pectus. J. Submicr.
Cytol., 6, 399-419.
TEIGLER, D. J. and BAERWALD, R. J. 1972. A freeze-etch 447-456.
study of clustered
nuclear
pores. Tissue & Cell, 4,
