Fine structure of meiotic prophase chromosomes and modified synaptonemal complexes in diploid and triploid Rhoeo Spathacea

Diploid Rhoeo spathacea (Swartz) Stern (2n = 12) is characterized by chromosomal structural hybridity to such an extent that no 2 chromosomes in the diploid plant are genetically identical. At first meiotic division, the 12 chromosomes are usually arranged into a ring. One, two or three chains are often observed (Lin & Paddock, 1973b). The ring has been interpreted to be a result of extensive reciprocal translocation, so that the 2 arms of each chromosome synapse respectively with an arm of each of 2 other chromosomes. Rhoeo has not been satisfactory for light-microscope studies of synapsis because of the unfavourable and elongated state of chromosomes at zygonema and pachynema (Figs. 5-8). Electron microscopy allows the study of synapsis at an ultrastructural level.

Synaptonemal complexes (SC hereafter) have been shown to be associated with bivalents at meiotic prophase in many species of eukaryotes and have become regarded to be involved in synapsis and perhaps subsequent crossing over and chiasma formation although the exact function is yet not clear (see Moses, 1968; Westergaard & Wettstein, 1972; Gillies, 1975, for review). SC among different species have a remarkable morphological consistency despite great phylogenetic diversity. An SC usually consists of a set of 3 parallel strands, lying in a single plane that curves and twists along the axis of the synapsed chromosomes in which it lies. Two dense lateral elements are held in register by the central region which contains the central element. The central element is generally less dense and is joined to the lateral elements by thin transverse filaments.

Modified lateral elements of SC have been reported. In an allotriploid lily, Moens (1968) observed unusual tube-like structures over a distance of a few micrometres in one or both of the lateral elements in nearly every section of a pachytene nucleus. He speculated that this deformed structure might be associated with the chromosome heterozygosity because this species is an allotriploid derived from hybridization of two or more closely related species. Further studies of SC in an autotetraploid lily (Moens, 1970) revealed no such abnormality. Later a very similar deformity was reported in pollen mother cells of both diploid and triploid Phaedranassa viridiflora (LaCour & Wells, 1973b). LaCour & Wells (1973 a) also studied the SC in a lily hybrid, which had almost complete bivalent formation at metaphase I, and observed rather bizarre irregularities in lateral elements and in whole SC.

In his studies of the SC in diploid Rhoeo, Moens (1972) found coiled SC attaching to the inner membrane of the double nuclear membrane. Therefore, diploid Rhoeo presents another example of a modified SC. The coiled SC, hereafter simply the coil, extends approximately 4 μm from the nuclear membrane and as many as 10 such attachments (of 12 possible attachments) were observed in a small region of the nuclear membrane near the nucleolus (Moens, 1972). Moens concluded that perhaps all the ends of SC in diploid Rhoeo are coiled and attached to the nuclear membrane. This type of coil has not been reported except in a grasshopper, Chloealtis conspersa (Moens, 1974). Later McQuade & Wells (1975) confirmed Moens’ finding of coils in diploid Rhoeo.

Studies on SC have been concentrated mostly in diploid organisms. This study was aimed at investigating the SC and its possible structural modification in triploid Rhoeo, and comparing the SC in the triploid with that in the diploid.

MATERIALS AND METHODS

The process of microsporogenesis and maturation is often synchronous within an anther but is less so among the anthers in a flower of Rhoeo. For accurate identification of the meiotic stages in pollen mother cells (PMC hereafter), an anther in proper size was first cut into 2 halves which were fixed immediately: one half was placed in fixative for squash preparation, and the other was fixed in glutaraldehyde for electron microscopy. If the PMC in squash preparation were found to be in the proper stages, the other half was further prepared for electron-microscope studies. Further stage identification was done with thick sections (o.8 μm) prior to ultrathin sectioning.

The method of making slides for light microscopy has been described (Lin & Paddock, 1973a).

For electron microscopy, excised anthers were fixed immediately for 1 h in 2% glutaraldehyde buffered in o-2M Sorensen’s phosphate buffer at pH 7.2, rinsed in buffer at pH 7.2 for 30 min (2 changes), and postfixed for 1 h in phosphate-buffered 1% osmium tetroxide. Anthers were then dehydrated in a graded ethanol series at 4 °C, and embedded in hard Spurr’s low viscosity resin (Spurr, 1969) at room temperature. Thick sections (o.8 μm) were obtained with an ultramicrotome, then stained with polychrome stain (Sato & Shamato, 1973), and examined with a light microscope for additional stage identification. Only those blocks containing PMC in proper stages were further sectioned into ultrathin sections. The thin sections were mounted on grids, stained with a saturated aqueous solution of uranyl acetate (3 min) followed by lead citrate (3 min) and examined with an RCA-EMU 3G electron microscope.

Chromosomes at leptonema are thin and unsynapsed, and the two sister chromatids are not seen as two distinct strands under the light microscope (Fig. 1). At the ultrastructural level, chromosomes are diffuse and the axial element is elusive in the early stage of leptonema (Fig. 2). At a later stage of leptonema, the medial longitudinal axial elements can be detected occasionally in the space between the sister chromatids (Figs. 3, 4). The axial element has a lower electron density than the chromatin surrounding it. In cross-sections, the axial element can be identified only with ambiguity or cannot be identified at all. Heterochromatic chromocentres (Figs. 2, 3) were observed in leptonema, zygonema and pachynema. The chromocentral chromatins can be readily distinguished from euchromatin in electron micrographs (Figs. 2, 3). The chromocentres are often large and composed of heterochromatins in 2 variable states. Chromosomal attachments to the chromocentre were often seen (Fig. 2). At early zygonema, large portions of the chromosomes are not synapsed. The axial element in the unsynapsed portion of a univalent at zygonema (Fig. 4) is similar in electron density and appearance to that at leptonema (Fig. 3).

In squash preparations, some synapsed chromosome segments can be seen in less-clumped areas of zygotene cells (Figs. 5-7). Part of zygotene chromosome configuration of the cell in Fig. 7 A is interpreted in Fig. 7B-F and a differential segment is identified. A pachytene cell is seen in Fig. 8.

A synapsed bivalent invariably possesses an SC running its central axis (Figs. 10–12). Fig. 4 is an electron micrograph of part of a zygotene cell with SC and axial elements. As synapsis proceeds at zygonema, homologous chromosomes (univalents) with their axial elements come into contact and are held in register at a distance of about 115-0 nm by a central region. The axial element at leptonema is now called the lateral element of the SC in the synapsed bivalent. Each lateral element measures about 46·0 nm in thickness. In the middle of the central region is a central element about 30·0 nm thick.

Both lateral and central elements have the same electron density and are all less extensively stained than the chromatin. These elements do not have evident substructures so they are referred to as amorphous (Westergaard & Wettstein, 1972). The region between the lateral and the central elements appears to be light and loosely textured. The transverse filaments crossing this region are poorly defined (Figs. 10–12). In cross-sections, the tripartite SC appears as a near-circle in outline and lies in the centre like a core in the chromatin of synapsed bivalents. The lateral elements are semicircular in outline and the central element appears to be a rather flat structure (Fig. I2B).

The twists and turns of SC, revealed by serial section reconstruction or whole mount preparations, are generally very gentle (Gillies, 1972; Counce & Meyer, 1973). In the diploid, coiling was observed at the ends of synapsed bivalents (Figs. 9, 13). The coils were always near the nuclear membrane on which the synapsed chromosome ends were attached. The coil in Fig. 13 extends inwards about 3·5 μm from the nuclear membrane and makes one complete turn at a distance of every 0·5 μm. The dimensions of the lateral and the central elements in the coil are the same as those where the complexes are not coiled.

SC were also observed in chromocentres (Fig. 14). The SC in chromocentres and the role of the coil are considered in the Discussion.

Measurements of SC in the diploid are compared with the measurements made by McQuade & Wells (1975) in Table 1: the methods of fixation appear to alter the dimensions of SC. The authors’ measurements of SC in Rhoeo are very close to the average dimensions reported from 1968 to 1971 in various organisms as shown in table I of Westergaard & Wettstein (1972). McQuade & Wells’ measurements of the lateral and central elements in the diploid are apparently much smaller than the average.

Leptotene chromosomes in triploid (Fig. 15) have an appearance similar to those of diploid (Fig. 1). At zygonema, the homologous chromosomes synapse (Fig. 16). A pachytene cell is in Fig. 18. Two-by-two synapsis leaving the third unsynapsed has been observed in a squash preparation (Lin & Paddock, 1978). With the electron microscope, SC were also found to be present in synapsed bivalents in triploid and they were not distinguishable from those in diploid. In Fig. 17 a portion of a zygotene cell with SC distributed over the section is shown. Figs. 19–21 are electron micrographs of SC in various cells. The 2 lateral elements appear as 2 ribbons each about 46-0 nm thick and separated by a region in the centre of which lies a strip of central element about 30-0 nm thick. Fine filaments traversing the central region between the lateral elements are more obvious in Fig. 21. The distance between the lateral elements is about 117·5 nm (Table 1). A cross-sectional view of an SC is in Fig. 20B.

Chromocentres were also observed at prophase I in triploid (Figs. 22, 23). Diffuse chromosomes and chromocentres of a leptotene cell can be seen in Fig. 22 ; in some sections, some chromocentres did have a suggestion of SC (Fig. 23).

Coiling at the ends of synapsed bivalents was also observed in triploid (Fig. 24). In Fig. 24, the end of the chromosome is associated with a nucleolus and the nuclear membrane. It appears then that a nucleolar organizer region is located at the tip of this chromosome.

No SC were found inside nucleoli in this study although central region materials have been found inside nucleoli of a fungus, Neottiella (Westergaard & Wettstein, 1970). Triple chromosome synapsis as reported in triploid chickens (Comings & Okada, 1971) was not observed in triploid Rhoeo.

No abnormal tubular lateral elements such as those described in triploid lily (Moens, 1968) and triploid Phaedranassa (LaCour & Wells, 19736) were observed either in triploid or in diploid Rhoeo in this study. Therefore the deformed lateral elements in these 2 instances may not be due to the triploidy.

The association of chromosome ends with the nuclear membrane at prophase I has been found consistently in many animals (insects: Moens, 1969; Church, 1976; Wettstein & Sotelo, 1967; mammals: Baker & Franchi, 1967; birds: Ford & Woolam, 1964; snail: Gall, 1961 ; and rat: Esponda & Giménez-Martin, 1972). Occasionally, the association was observed in higher plants (Figs. 9, 13, 24 in Rhoeo;Gillies, 1973, in maize). In the ascomycete fungi Neottiella (Westergaard & Wettstein, 1970) and Neurospora crassa (Gillies, 1972) the ends of chromosomes are also connected with the nuclear membrane. Both ends of the same chromosomes attaching to the nuclear membrane were observed in some organisms (Ford & Woollam, 1964; Wettstein & Sotelo, 1967; Moens, 1969).

The attachment points may be polarized to a small region near the centrioles making a bouquet configuration in Locusta migratoria and synapsis was found to be initiated near the attachments and proceeded away from the nuclear membrane in this organism (Moens, 1969).

The polarization of chromosome ends at prophase I is probably important to the process of synapsis. It may provide a firm association of chromosome ends with each other in the beginning of synapsis and promote the chance of the synaptic partners meeting.

The attachment of chromosome ends with nuclear membrane has been noted only rarely in higher plants (Westergaard & Wettstein, 1972). In my study, both diploid and triploid Rhoeo were found to have not only the attachments but also coils (Figs. 9, 13, 24). From a large number of serial sections, as many as 10 such attachments were observed in a cell by Moens (1972) and 11 by McQuade & Wells (1975). The 3 nuclei examined with serial sections by Moens had all the attachments polarized on a small area of the nuclear membrane. Three coils together in a small region of nucleus can be seen clearly in a single thin section (my Fig. 9). Similarly to Moens’ (1969) observation, McQuade & Wells’ (1975) data indicate that the initiation of synapsis in diploid Rhoeo probably begins near the nuclear membrane. Synapsis in triploid Rhoeo perhaps proceeds in a similar fashion.

The existence of coils has been reported only in Rhoeo and a grasshopper, Chloealtis conspersa (Moens, 1974). The coils in C. conspersa are located in the terminal portion of some of the bivalents.

There is a striking correlation between such modified ends of SC and terminal chiasmata. In diploid Rhoeo, the 12 meiotic chromosomes at diakinesis are joined by terminal chiasmata and form a ring. In C. conspersa, there are 3 pairs of large chromosomes with terminal chiasmata which always form rings at meiosis (Moens, 1974). Another species of grasshopper, Chorthippus longicomis, has a karyotype similar to C. conspersa but there are no terminal chiasmata, and coils have not been found in this species (Moens, 1974). Such correlations suggest that the modified SC are probably involved in terminal chiasma formation.

The indication from light-microscope studies that synapsis in diploid Rhoeo is limited to short terminal regions of chromosomes (Lin & Paddock, 1973b) is in conflict with the observation that synapsis (SC formation) occurs all over in the interior of the nucleus (McQuade & Wells, 1975). This can be explained, however, by the fact that SC does not always lead to chiasma formation. The role of the modified SC, coils, perhaps is in enhancing the chance of crossing-over at terminal regions. Light-microscope studies showed the frequency of terminal chiasmata tends to be high in diploid Rhoeo meiotic cells (Lin & Paddock, 1973b). Only a very small percentage of cells had less than 9 terminal chiasmata. On the other hand, prophase I cells seem to have a high number of coils in each cell because the 2 zygotene and 1 pachytene nuclei studied had 8, 10 and 11 attachments (coils), respectively (McQuade & Wells, 1975). Thus, the number of terminal chiasmata probably corresponds to the number of coils. The chiasma failure might be due simply to the failure in coil formation at the attachment during synapsis. An alternative hypothesis is that the chromosome ends are held together by nothing more than the coils themselves, i.e. chiasmata are not involved in the ring formation at all.

Heterochromatic chromocentres are often found at prophase I in Rhoeo, and exist in 2 states of condensation (Figs. 3, 22). Chromosome-chromocentre attachment can be observed in Fig. 2. The chromocentres in diploid and triploid cells are indistinguishable (compare Fig. 3 with 22 and 23). In squash preparations of prophase I cells, the chromocentres and nucleolus usually do not appear as distinct entities among the chromosomal materials and are thus difficult to detect (Figs. 6, 7,16). However, the nucleolus is obvious in Fig. 8. Natarajan & Natarajan (1972) were able to show clumping of chromocentres at pachynema in diploid. The heterochromatins in diploid are situated around the centromeres (pericentromeric) in all the 12 chromosomes and are late replicating (Natarajan & Natarajan, 1972). They are constitutive heterochromatins and may be considered to contain highly repetitive DNA sequences (Yunis & Yasmineh, 1971). Aggregation of heterochromatin has been shown in many organisms (Yunis & Yasmineh, 1971). In a recent report (Godin & Stack, 1975), association of telomeric heterochromatins was shown in rye, Secale cereale. In Rhoeo, the chromocentre formation thus appears to be from the aggregation of at least part of the pericentromeric heterochromatins. It can be conceived as a result of extensive reciprocal translocation in the following way. It is generally accepted that the structural hybridity in Rhoeo has resulted from a series of translocations. As shown in Fig. 25 A if the break points were at or near the centromeres, then heterochromatins in the 4 chromosomes could be brought together at pachynema and a large mass of heterochromatic chromocentre formed. A ring-enlarging translocation, if at centromeric vicinities, will give rise to an even larger chromocentre by clumping together of centromeric heterochromatins in 6 chromosomes (Fig. 25 B).

The indefinite shapes of chromocentres may reflect that the various reciprocal translocations involved in building the ring did not occur equally distant from their respective centromeres.

Short stretches of synaptonemal complexes were occasionally found in chromocentres in both diploid and triploid. Those in Figs. 14 and 23 and in McQuade & Wells’ (1975) figs. 11 and 12 are all in the less-condensed area which has a similar appearance to that of chromosome. It is possible but doubtful that SC in Rhoeo do not occur in the highly condensed region of the chromocentres.

Electron-microscope studies of meiosis in Fritillaria lanceolata (LaCour & Wells, 1970) also showed the presence of SC in the chromocentres.

The view that heterochromatin as such may aid in initial alignment prior to synapsis was criticized by Maguire (1972) who considers that a direct functional role of heterochromatin in synapsis still lacks sound evidence. In Rhoeo, chromocentres are probably a result of synapsis causing association of pericentromeric heterochromatins.

Based on considerations in the foregoing discussion and some observations in the study, theoretical configurations of diploid pachytene cells are presented in Figs. 26 and 27. Only lateral elements of SC are presented. As shown in Fig. 26 G, 5 successive translocations among non-homologous chromosomes occurring 2 in centromeric regions and 3 in non-centromeric regions result in a branched 2-dimensional pachytene configuration containing 12 chromosomes that can become a ring at diakinesis. A corresponding 3-dimensional pachytene configuration is in Fig. 27. A chromocentre forms wherever more than 2 centromeric regions aggregate. Where translocation has not occurred at a centromeric region, such as in Fig. 26B, D and E, no chromocentre is expected to be present at that point. The speculation that break points are not necessarily localized at centromeric regions is based on the interpretation in Fig. 7B-F in which neither of 2 adjacent translocation points is associated with the chromocentres (see also Fig. 27).

This paper represents part of a dissertation submitted for a Ph.D. degree at The Ohio State University, Columbus, Ohio, U.S.A. I am thankful to Dr Elton F. Paddock for his guidance, patience and encouragement during the course of this study.