Formation in Spermatocytes and Oocytes Ultrastructural Study

Heredity 63 (1989) 97—106 The Genetical Society of Great Britain Received 17 January 1989

Chromosome pairing and chiasma formation in spermatocytes and oocytes of Dendrocoelum lacteum (Turbeflaria, Tricladida); a cytogenetical and ultrastructural study

G. H. Jones and School of Biological Sciences, University of J. A. Croft Birmingham, Birmingham B15 2TT,U.K.

The hermaphroditic flatworm Dendrocoelum lacteum shows a very pronounced sex difference in chiasma frequency; oocytes have 66 per cent more chiasmata than spermatocytes. The ultrastructural basis of this difference was investigated by three-dimensional reconstruction of synaptonemal complexes from ten oocytes and ten spermatocytes at pachytene and smaller numbers of zygotene nuclei. Oocyte nuclei at pachytene are larger and have longer synaptonemal complexes than the equivalent spermatocyte stage. The difference in synaptonemal complex length (60 per cent) corresponds very closely to the difference in chiasma frequency shown by oocytes and spermatocytes and suggests that synaptonemal complex (SC) length may contribute to the regulation of chiasma frequency. Recombination nodules were observed in both spermatocytes and oocytes but their quantitation was unreliable because they are small and differentiation from the unusually dense and prominent central region is difficult. Analysis of the few zygotene nuclei available shows that pairing is predominantly proterminal in both spermatocytes and oocytes, but oocytes may have several additional interstitial initiations. These extra initiations are likely to be necessary for efficient pairing of the longer oocyte chromosomes.

INTRODUCTION chromosome behaviour whose significance has only recently been reappreciated. Dendrocoelumlacteum is a common freshwater The hermaphroditic nature of D. lacteum was planarian which is very widely distributed exploited by Pastor and Callan (1952) in a com- throughout Europe in all types of water bodies parative study of chiasmata in later stages of sper- (Ball and Reynoldson, 1981). It has a relatively matogenesis and oogenesis. They found that long but sporadic history of cytogenetical investi- chiasma frequency in oocytes was much higher gation (Matthieson, 1904; Arnold, 1909; Gelei, than in spermatocytes, thus providing one of the 1913, 1921, 1922; Aeppli, 1951; Pastor and Callan, earliest as well as most dramatic examples of sex 1952; Benazzi and Pochini, 1959). Its somatic differences in chiasma frequency. The aim of the chromosome number of 14 was first accurately present study was to investigate the ultrastructural determined by Gelei (1913), but there has not basis of this marked meiotic sexual dimorphism always been agreement on this point (see Benazzi by extending the analysis of meiotic pairing in this and Benazzi-Lentati, 1976 for details). Gelei (1913, species to the ultrastructural investigation of 1921, 1922) also recognised the value of this species synaptonemal complexes (SCs) in oocytes and for the cytogenetical investigation of meiosis and spermatocytes. conducted a detailed analysis of the course of chromosome pairing in oocytes which became a classical study in this field. Gelei's studies not only vindicatedtheparasynapsis(side-by-side) MATERIALS AND METHODS hypothesis of chromosome pairing but also confirmed the homologous nature of this pairing. Adult sexually mature specimens of D. lacteum He also recognised the phenomenon of chromo- were collected from Winterbourne Lake, Edgbas- some interlocking during zygotene, an aspect of ton, Birmingham between November and April, 98 G. H. JONES AND J. A. CROFT when oogenesis and spermatogenesis were most matocytes and oocytes. Entire animals were fixed prevalent (Pastor and Callan, 1952). in 3 : 1 acetic alcohol and then Feulgen stained to facilitate the location of testis and ovary material. Small portions of the animals containing either Electronmicroscopy testis or ovary material were then squashed in Entireanimals (1-2cm long) were fixed in 4 per lactopropionic orcein. cent glutaraldehyde in phosphate buffer at pH 72 for 30 minutes at room temperature. The animals were then cut into small pieces (2-3 mm) and then RESULTS returned to the fixative for a further 3 hours. Some specimens were treated with 2 per cent osmium Light Microscopy tetroxide in phosphate buffer at pH 72 for 2 hours Somatic at room temperature and dehydrated through an karyotype alcohol series. For other specimens the osmium The diploid somatic karyotype of the D. lacteum tetroxide treatment was omitted; instead they were population employed in this study consists of 14 passed directly through an alcohol series and then chromosomes containing seven pairs which can be treated with 1 per cent alcoholic phosphotungstic identified individually on the basis of overall length acid (PTA) for 16 hours at room temperature. Fol- combined with arm ratio (see fig. 1). The somatic lowing these treatments, all specimens were chromosome number of 14 confirms the majority embedded in Epon 812 and polymerised at 60°C of previous investigations and does not accord with for 20 hours. Serial thin sections (lOOnm) were cut somatic chromosome counts of 16 reported by on a Reichert-Jung ultramicrotome equipped with some investigators (see table 2(b) of Benazzi and a diamond knife. Serial ribbons of sections were Benazzi-Lentati, 1976). Published details of the transferred to Formvar coated single slot grids somatic karyotype of D. lacteum are scarce (e.g., using the indirect method of Wells (1974). Sections Dahm, 1961; Melander, 1963) and none of them that had been previously treated with osmium agree completely with the karyotype observed in tetroxide were further stained with 1 per cent the present investigation. Whether this reflects aqueous uranyl acetate at 40°C for 90 minutes and intra-specific chromosome polymorphism remains post stained with lead citrate for 2 minutes at 20°C an open question. in an LKB 2168 Ultrostainer. Sections of PTA treated specimens required no further staining. Meiotic chromosomes and chiasmata Serial sections were examined and photographed using a JEOL 1200EX electron microscope. All stages of spermatocyte meiosis are readily Individual SCs were traced from the series onto observed in Feulgen and orcein-stained squash transparent acetate overlays and SC lengths were preparations of testis material. The chiasmata of measured using a Summagraphics digitising tablet spermatocytes are most conveniently scored in linked to a BBC microcomputer which was pro- nuclei at metaphase I (fig. 2(a))). Most bivalents grammed to include section thickness during the are easily identifiable at this stage, with the excep- reconstruction procedure. tion of chromosomes 5 and 6 which are sometimes indistinguishable. Ovarian squash preparations yield numerous prophase I oocyte nuclei including Lightmicroscopy late stages (diplotene to diakinesis) in which seven Themitotic karyotype was ascertained from relatively uncondensed bivalents are visible (fig. orcein-stained squash preparations of young 2(b)). Because centromeres are not generally dis- embryos. Newly laid coccoons, which contain cernible at these stages, bivalent identification is several embryos, were collected and kept in pond not possible in oocytes. water at 15°C for 6—8 days. At this stage the coc- As part of this study, an extensive survey of coon shells were punctured with a sharp needle chiasma variation in the Winterbourne Lake popu- and the coccoons were treated with 005 per cent lation was undertaken. Several animals were coichicine for 3 hours before fixing in 3: 1 acetic sampled on each of three different collection dates alcohol overnight. Lactopropionic orcein-stained (December, January, February) in order to deter- squash preparations were then made using small mine whether any seasonal variation in chiasma portions of coccoon contents. frequency occurs and also to ascertain the extent Chiasma analysis was carried out on Feulgen of inter-individual chiasma variation (table 1). The and orcein-stained squash preparations of sper- chiasma frequencies of both spermatocytes and MEIOTIC SEX DIFFERENCES IN DENDROCOELUM 99

ILa

Figure1 Colchicine treated embryonic mitotic metaphase of Dendrocoelum kwieum (a) and the mitotic karyotype derived from the same cell (b). (Bar=5 tim). oocytes are slightly higher in collection C matocytes in the same individuals are positively (February) as compared to the other two collec- correlated (fig. 3); the regression of oocyte chiasma tions, but this difference is only significant for frequency on spermatocyte chiasma frequency is oocytes and then only marginally so (P <0.05). significant (F(115) =7402;P < 005). However, by Significant inter-individual variation occurs for far the most striking source of chiasma variation both oocytes and spermatocytes. In agreement with in this study is the difference between oocytes and the earlier finding by Pastor and Callan (1952), spermatocytes. The mean chiasma frequency of the chiasma frequencies of oocytes and sper- oocytes in this population is 1931 (range 1725 to

Figure 2 Meiotic chromosomes of Dendrocoelum lacteum. (a) Spermatocyte first metaphase with seven condensed bivalents containing 12 chiasmata. (Bar =5pm). (b) Oocyte diakinesis with seven relatively uncondensed bivalents containing 18 chiasmata. The nucleolus (N) is associated with two of the oocyte bivalents. (Bar= 10 tam). 100 G. H. JONES AND J. A. CROFT

Table 1 A summary of mean chiasma frequencies () and their standard deviations (s.d.) of spermatocytes and oocytes from Dendrocoelumlacteum individualscollected on three different occasions from Winter- bourne Pool, Edgbaston, Birmingham. A =December,B =January,C =February.n= numberof cells examined

Collections Individuals t n s.d. n s.d.

A 1 1850 2 071 1133 15 105 2 1880 10 132 1087 15 106 3 1725 4 150 1080 15 115 4 1871 7 160 1227 15 080 5 2000 8 177 1113 15 092 6 1880 10 215 1240 15 l35

B 1 2100 3 265 1253 15 119 2 1900 5 141 1114 7 107 3 2013 8 136 1220 15 121 4 1860 10 171 1087 15 125 5 1690 10 137 1033 15 111 6 1710 10 160 1147 15 074

C 1 2080 10 249 1200 15 151 2 2275 8 198 1173 15 088 3 1880 5 228 1213 15 155 4 2175 8 128 1247 15 164 5 1940 5 207 1193 15 103

22.75) whilethat of spermatocytes is much less at pro-terminal localisation. Pro-terminal segments 1162 (range 1O33 to 12.53), fully confirming the of bivalents, occupying 40 per cent of total chromo- observations of Pastor and Callan (1952). some length (20 per cent at each end) contain 70 Chiasma distribution is difficult to quantify and 65 per cent, respectively, of all chiasmata in accurately in this species because of the small spermatocytes and oocytes. chromosomes and the lack of centromere visibility in late prophase I oocytes. Nevertheless a general description of chiasma distribution is possible Electron microscopy based on visual assignment of chiasmata to arbitrary sub-divisions of bivalents. In both sper- Pachytene SCs matocytes and oocytes chiasmata are widely dis- The testes of D. lacteum form numerous scattered tributed along bivalents, but with a tendency to islands dispersed along most of the body.

C., . S . 12

cc E I a o . . S a . '0

10 1] 18 19 20 21 22 23 Mean chiasmafrequency Figure 3 Spermatocyte () mean chiasma frequencies plotted against oocyte () mean chiasma frequencies for 17 different individuals of Dendrocoelumlacteum. MEIOTIC SEX DIFFERENCES IN DENDROCOELUM 101

Table 2 Ranked lengths (p.m) of the 7 SCs in 10 spermatocytes and 10 oocytes of Dendrocoelum lacteum at the pachytene stage, together with total SC lengths of nuclei, nuclear volumes and numbers of attached SC ends. A-D refer to four different individual animals

Total Nuclear SC length volume Attached 1 2 3 4 5 6 7 (p.m) (jim3) ends

Spermatocytes 1(A)20-66 18-28 15-46 14-72 14-24 1184 1129 10650 230-38 14 2(A)2023 18-71 15-46 1397 1332 1288 1164 106-21 22936 14 3(B) 19-02 18-32 15-82 14-29 1363 1286 10-49 104-43 18022 14 4(A)20-47 17-63 16-04 1305 12-24 11-67 11-56 10266 223-96 14 5(B) 1922 19-14 1503 1292 11-81 1179 1052 100-43 19029 14 6(A) 1912 1779 1345 13-22 12-20 11-00 10-72 97-50 24171 14 7(B) 16-70 1315 12-50 12-14 1139 1057 991 8636 16628 12 8(B) 16-40 15-98 14-44 11-40 9-67 9-34 7-47 84-69 169-88 12 9(B) 15-69 14-57 12-81 11-55 11-39 10-21 8-19 84-41 152-05 11 10(B) 15-21 14-70 10-98 9-45 8-77 8-48 5-77 7336 112-69 14 Mean94-65

Oocytes 1(C)31-84 31-54 28-48 26-29 22-92 17-01 13-35 171-43 753-00 1 2(C)32-84 29-82 25-27 24-67 21-76 18-48 18-16 171-00 507-08 2 3(C)31-95 27-25 2370 23-43 19-13 19-01 18-41 162-88 582-24 0 4(C)2893 26-97 22-33 22-06 19-10 18-76 16-07 154-23 501-94 11 5(C)28-31 27-61 23-51 21-75 18-09 17-11 16-72 153-09 587-31 8 6(C)28-92 26-56 22-04 17-95 17-33 15-26 1505 143-10 412-72 10 7(C)27-77 25-08 21-80 20-05 16-87 15-50 15-43 142-51 50635 11 8(D)26-36 24-98 23-64 19-79 16-45 15-17 15-01 141-40 401-86 8 9(C)29-46 23-76 22-97 1831 17-80 15-05 12-33 139-68 614-86 6 10(C)25-94 22-33 21-16 19-67 17-80 16-33 14-65 137-87 540-67 6 Mean15172

Pachytene spermatocytes occur in synchronised while the pale central region contains a prominent groups of eight cells intermingled with other sper- dense central element (fig. 5). matogenic stages in an apparently unordered man- All pachytene nuclei contain seven uninterrup- ner around the central lumen of each testis island. ted SCs, confirming that all seven chromosome The paired spherical ovaries lie anteriorly, directly pairs are fully synapsed at pachytene in both sper- behind the brain. Oocytes evidently show some matocytes and oocytes. The ranges of total SC zonation within the ovaries since there is some lengths in ten spermatocyte and ten oocyte clustering of pachytene oocytes. However a simple pachytene nuclei are shown in table 2, together linear sequence of oocyte development such as with data on nuclear volumes and numbers of occurs in the rhabdocoel Mesostoma ehrenbergii attached SC ends. It is notable that spermatocytes ehrenbergii (Oakley, 1982) is not found in D. show a significant positive linear regression of total lacteum. sc length on nuclear volume (F(18) =21-854,P < Spermatocyte and oocyte nuclei differ con- 0.01) suggesting that both SC length and nuclear siderably in size (table 2) and general appearance volume decrease with the progression of (figs. 4(a) and 4(b)). Spermatocyte nuclei are pachytene. However, for reasons which are not relatively small (6-8 urn diameter) and have fairly immediately apparent, oocytes do not exhibit a evenly dispersed chromatin with some concentra- clear relationship between SC length and nuclear tion around SCs, while oocyte nuclei are consider- volume (F(18)=2.667; P>010). Probably one or ably larger (9-11.5 um) and their chromatin is both of these variables changes non-linearly during heavily concentrated around SCs and separated the progression of oocytes through pachytene. It by large chromatin-free spaces. However SC is known, for instance, that Drosophila oocyte morphology is very similar in spermatocytes and pachytene SCs first decrease in length, then later oocytes and also resembles that found in another increase in length again (Carpenter, 1979a). The platyhelminth worm M. ehrenbergii ehrenbergii difficulty of identifying a clear linear series of (Oakley and Jones, 1982; Croft and Jones, 1989). developing oocytes in Dendrocoelum ovaries hin- The lateral elements are indistinct and poorly dered attempts to identify sequential changes in differentiated from the surrounding chromatin oocyte volume and SC length in this species. 102 G. H. JONES AND J. A. CROFT I t-. — I' .1 I-. C:' 0) 'tm

Figure4 Low power electron micrographs of thin sections through entire spermatocyte (a) and oocyte (b) nuclei of Dendrocoelum lacteum.Arrows indicate synaptonemal complexes. (Bar= 2 rm).

The most striking feature of the SC data is the agree closely with the ratio of chromosome length marked difference in mean SC length between extremes in the mitotic complement. However, the spermatocytes and oocytes. The mean SC length construction of an SC karytype for this species is of spermatocyte nuclei is 9465 p.m (range 7336-- greatly hindered by the lack of consistent and 10649 p.m) while in oocytes the corresponding reproducible centromere structures on SCs. value is 151'72 p.m (range 13787—l7143 p.m),that Although putative centromere structures are some- is 60 per cent longer. The ratios of longest to times observed they do not appear with sufficient shortest SCs are remarkably consistent in oocytes regularity and consistency even to confirm that (1: 1883) and spermatocytes (1: 1876) and also they represent centromeres, let alone serve as an

a.- a

-F'S

ft.,'.S L -;Qi-1. '$ a A 1111: $1

Figure 5 Details of short SC stretches containing recombination nodules (arrowed). (Bar=O-5 gm). MEIOTIC SEX DIFFERENCES IN DENDROCOELUM 103 aid to SC identification. Individual SCs cannot be double staining with uranyl acetate and lead citrate identified on the basis of length alone because of RN structures were observed in both spermatocyte the relatively small differences in relative length and oocyte pachytene nuclei. The clearest RNs between adjacent rank positions and the con- seen are small dense spherical structures located sequent risk of SC misclassification due to directly on the central element of the SC (fig. 5). "reversals of order" (Matern and Simak, 1968; These RNs are no more than 50-70 nm in diameter, Croft and Jones, 1986). not much wider than the central element, and The nucleolus organizer (NOR) serves as a against the prominent central element characteris- marker identifying one SC in spermatocytes. The ing this species even the clearest examples are not NOR is subterminally located on the second very obvious structures. In addition to these rela- longest SC in spermatocytes, but although the tively clear examples, there are many more putative mean and modal rank positions of this SC place RNs which are smaller, less densely stained and it as the second longest in the complement, its less distinct against the background of the dense actual rank position in individual nuclei varies central element. These could not be definitely iden- between 1 and 3, which illustrates the "reversal of tified as RNs, although some of them probably order" problem mentioned above. In oocytes the were so. This unforeseen difficulty virtually pre- situation is more complex because additional cluded any meaningful assessment of RN numbers nucleoli and NORs may be present, which make and distributions because the boundary between direct comparisons with spermatocytes difficult. definite and questionable RNs was undefinable. Oocytes may have one or two discrete nucleoli associated with up to 4 NORs occupying sub- terminal locations on up to 3 SCs. Considerable Zygotene observations variation occurs both between different animals Unpaired axial cores (ACs) are not clearly visible and between oocytes from the same animal regard- against the surrounding chromatin in zygotene ing the numbers of expressed NORs in oocytes. nuclei of D. lacteum, reflecting the weak differenti- Spermatocyte SCs show a moderate degree of ation of lateral elements in this species. Con- bouquet polarisation. The SC ends are almost sequently, a full analysis of zygotene nuclei invariably attached to the inner surface of the through paired and unpaired regions is not feas- nuclear envelope and the majority of ends in each ible. Nevertheless partial analyses can be perfor- nucleus are restricted in their distribution to one med, based on the numbers, lengths and distribu- hemisphere of the nucleus. Only 4 per cent (-%) tions of identifiable SC segments present in ofends were observed to be unattached in ten zygotene nuclei. Table 3 gives details of four oocyte reconstructed pachytene spermatocytes. Bouquet and three spermatocyte nuclei at the zygotene stage polarisation is also evident in oocytes, but to a which were reconstructed and partially analysed. lesser extent than in spermatocytes and more SC The total length of SC present in a zygotene ends are unpolarised. A surprising finding is that nucleus, compared to the mean pachytene SC many SC ends ()areunattached to the nuclear length, gives an approximate indication of the envelope in oocytes, although in most of these extent of pairing. Marked SC shortening between cases the ends lie near the nuclear periphery. Individual oocyte SCs may have 0, 1 or 2 SC ends attached to the nuclear envelope. There is consider- Table3The pairing characteristics of 3 spermatocyte and 4 able inter-nuclear variation in the number of oocyte nuclei of Dendrocoelum Iacleurn at the zygotene attached ends (table 2) which appears to bear some stage relation to SC length since the 3 oocyte nuclei with SC Per Fully Paired fewest attached ends (0, 1 or 2) also have the length cent paired chromosome Interstitial longest SCs. (ram) paired SCs ends SCs Spermatocytes Recombination nodules 1 7245 766 2 13 1 2 8336 934 2 13 2 One of the original objectives of this study was to 3 9080 959 4 13 2 carry out a quantitative study of recombination Oocytes nodules (RNs) in spermatocytes and oocytes 1 7328 483 1 5 3 for comparison with chiasmata. Staining with 2 8234 543 0 7 7 phosphotungstic acid failed to reveal any struc- 3 8571 565 5 11 2 4 11724 773 1 12 6 tures which could be candidates for RNs, but after 104 G. H. JONES AND J. A. CROFT zygotene and pachytene would result in consider- more similar in size and shape than those of more able overestimation of pairing extent at zygotene, distantly related species (Carpenter, 1979b). The but this does not seem to be the case. For example, RNs of D. lacteum are small spherical structures, one spermatocyte zygotene nucleus with 908 m not exceeding 70 nm in diameter. They therefore of SC is estimated to be 959 per cent paired and occupy the lower end of the observed range of has four fully paired SCs and only one unpaired RN sizes and are comparable to the rather small end, which seems reasonable. Furthermore, if SC RNs which characterise many fungal species (e.g., shortening were pronounced between zygotene Byersand Goetsch, 1975; HoIm et al., 1981). and pachytene then zygotene nuclei having more Another platyhelminth worm, the rhabdocoel M. SC than the mean pachytene length are expected, ehrenbergii ehrenbergii has, by contrast, large but none were observed. and very obvious RNs (Croft and Jones, 1989), SC segments in zygotene nuclei were classified which are comparable in their size and morph- as terminal if one end was associated with the ology to those of D. melanogaster oocytes. Evi- nuclear envelope, or interstitial if both ends were dently the generalisation that related species have free and unattached. The rather large numbers of similar sized RNs does not hold universally. unattached chromosome ends observed in Quantifying RN populations can be problem- pachytene oocytes are a potential complication to atical when RNs are uniformly small or very vari- this analysis. However, the detailed light micro- able in size. This problem is not unique to D. scopical analyses by Gelei (1921) clearly indicated lacteun and has been encountered, for example, that all chromosome ends are attached to the for the small RNs of yeast (Byers and Goetsch, nuclear envelope and arranged in a bouquet during 1975) ellipsoid nodules of wild-type Drosophila zygotene and early pachytene in oocytes; un- oocytes (Carpenter, 1981), spherical and ellipsoid attached ends are likely, therefore, to be a later RNs of some meiotic mutants of Drosophila (Car- pachytene phenomenon associated with the dissol- penter, 1979b) and zygotene RNs in the fungus ution of the bouquet. This analysis (table 3) indi- Coprinus (Hoim et a!., 1981). In D. lacteum the cates that chromosome ends pair relatively early problem is compounded by the dense central ele- in both spermatocytes and oocytes; all nuclei with ment of the SC which makes it difficult to detect in excess of 55 per cent pairing have most (11—13) with certainty any but the largest and most promi- of their 14 ends paired in SCs. In addition, there nent RNs. is evidence of some interstitial pairing initiation Oocytes and spermatocytes of D. lacteum show in both spermatocytes and oocytes, since separate a striking difference in SC length which almost SC segments which are not attached by either end exactly parallels their mean chiasma frequencies. to the nuclear envelope are observed in both cell Oocyte SCs are on average 60 per cent longer than types. However, spermatocytes show relatively few spermatocyte SCs and the mean chiasma frequency interstitial SC segments, only one or two per of oocytes exceeds that of spermatocytes by 66 per nucleus, while oocytes show many more, between cent. A pronounced sex difference in SC length two and seven per nucleus. This may be partly has also been reported in humans, where oocytes explained by the fact that most of the oocyte nuclei have a mean SC length per nucleus which is twice are at a relatively early stage of zygotene, but the that observed in spermatocytes. There is however difference is also clearly evident in a pair of nuclei some disagreement as to whether or not this having almost identical extents of pairing. One coincides with a sex difference in chiasma and oocyte nucleus shows 773 per cent pairing and recombination frequency (Bojko, 1985; Wallace has six interstitial SCs as well as 12 terminal SCs, and Hulten, 1985). In the mouse, where reliable whereas a spermatocyte nucleus with 766 per cent chiasma data exist for both males and females, a pairing has only one interstitial SC in addition to good correlation of SC length and chiasma 13 terminal SCs. frequency exists. Despite some variation due to age and interstrain differences, chiasma frequencies in male and female mice are very similar DISCUSSION (Speed, 1977) and SC lengths measured from oocytes and spermatocytes are also alike (Speed, RNsare variable in shape and size both between 1982; Moses & Poorman, 1984). In maize, extra and within species (Carpenter, 1979b, 1984; Ras- heterochromatic knobs on chromosomes 10 and 9 mussen and Holm, 1978). From the rather limited have been found to have parallel effects on re- number of species which have been examined, it combination frequencies and SC lengths (Gillies appears that the RNs of closely related species are et a!., 1973; Mogensen, 1977). The study by MEIOTIC SEX DIFFERENCES IN DENDROCOELUM 105

Mogensen (1977) on the effects of a heterochro- chromosome ends is in broad agreement with the matic knob on the short arm of chromosome 9 is classical light microscopical observations (Gelei, especially pertinent since higher female and lower 1921). Limited data are presented which indicate male crossover frequencies in the short arm of that interstitial pairing initiations may be more chromosome 9 were accompanied by longer and frequent in oocytes than in spermatocytes. Their shorter SCs, respectively, for that arm but not function is presumably to facilitate efficient pairing elsewhere in the genome. Finally, a difference in of the relatively longer chromosomes of oocytes. SC length which parallels a chiasma frequency difference has been observed in spermatocytes of Acknowledgement We thank Dr Adelaide T. C. Carpenter two locust species, Locusta migratoria and Schis- for reading the manuscript and making several perceptive Comments and helpful suggestions. This work was supported tocerca gregaria (croft and Jones, 1986) but this by a research grant (GR/C/98368) from the Science and case is rather different to the previous sex- Engineering Research Council. difference cases since the SC length and chiasma frequency differences are associated with a 49 per cent difference in genomic DNA amount between REFERENCES these species. AEPPLI,E. 1951. Die Chromosomenverhaitnisse bei Dendro- Correlations between SC length and recombi- coelum infernale (Steinmann). Em Beitrage zur Polyploidie nation frequency suggest, but do not prove that in Tierreich. Rev. Suisse Zool., 58, 5 11-518. they are causally related. The assertion by ARNOLD,G. 1909.The prophase in the oogenesis and the Mogensen (1977) that "different lengths of the SC spermatogenesis of Planaria lactea O.F.M. (Dendrocoelum lacteum Oerst.). Arch. Zellforsch., 3, 431—446. are responsible for the sex differences in crossover BALL, I. R. AND REYNOLDSON, T. B. 1981. British Planarians. frequencies" involves a further assumption about Synopses of the British Fauna, 19. Edited by D. M. Kermack the direction of causality. This assumption may be and R. S. K. Barnes. Cambridge University Press. judged to be reasonable in the light of our present BENAZZI, M. AND BENAZZI LENTATI, G. 1976. Animal knowledge of the timing of recombination in rela- Cytogenetics. Edited by B. John, H. Bauer, S. Brown, H. Kayano, A. Levan, M. White. Vol. 1: Platyhelminthes. tion to SC formation. It is certainly more difficult Gebruder Borntraeger, Berlin. to imagine the reverse situation, that is an effect BENAZZI, M, AND POCHINI, N. 1959. Alcune osservazioni of recombination frequency on SC length, but that citologische sulla planaria Dendrocoelum lacteum (0. F. possibility cannot be excluded. The suggestion by Muller). Boll. Zool., 26,435-444. BOJKO,M.1985. Human meiosis. IX. Crossing over and Mogensen (1977), based on an earlier proposal by chiasma formation in oocytes. Carlsberg Res. Commun., Stern et a!. (1975), that longer SCs are likely to 50,43-72. entrap more "preselected" DNA sequences which BYERS, B. AND GOETSCH, L. 1975. Electron microscopic in turn will enhance the probability of recombina- observations on the meiotic karyotype of diploid and tetra- tion occurring in that region is interesting and ploid Saccharomyces cerevisiae. Proc. Nat! Acad. Sci. USA, 72, 5056-5060. reasonably plausible but necessarily conjectural. CARPENTER, A. T. c. 1979a. Synaptonemal complex and re- This points to the need for a better understanding combination nodules in wild-type Drosophila melanogaster of the precise role of the SC in mediating and females. Genetics, 92, 511—541. regulating recombination events. CARPENTER, A. T. C. 1979b. Recombination nodules and synap- Information on the organization of chromo- tonemal complex in recombination-defective females of Drosophila melanogaster. Chromosoma, 75,259-272. some pairing at zygotene, in terms of the numbers CARPENTER A. T. C. 1981. EM autoradiographic evidence that and distributions of initiations and the progression DNA synthesis occurs at recombination nodules during of synapsis, is still relatively sparse. Among animal meiosis in Drosophila melanogaster females. Chromosoma, species, small chromosomes tend to initiate pairing 83, 59-80. CARPENTER, A. T. C. 1984. Recombination nodules and the nearly exclusively at the ends (e.g., Rasmussen and mechanism of crossing-over in Drosophila. Controlling Holm, 1978, 1982) while larger chromosomes may, Events in Meiosis. Society for Experimental Biology Sym- in addition, have one or more interstitial initiations posium 28. Edited by C. W. Evans and H. G. Dickinson. (Jones and Croft, 1986). It is reasoned that addi- Company of Biologists Ltd., Cambridge. CROFT, J. A. AND JONES, G. H. 1986. Surface spreading of tional interstitial initiations are an adaptation to synaptonemal complexes in locusts. I. Pachytene observa- facilitate the rapid and efficient pairing of larger tions. Chromosoma, 93, 483-488. chromosomes. Although oocytes and spermato- CROFT. J. A. AND JONES, G. H. 1989. Meiosis in Mesostoma cytes of D. lacteum have identical DNA amounts, ehrenbergii ehrenbergii. IV. Recombination nodules of sper- the chromosome material is organized differently matocytes; a test of the correspondence of late recombination nodules and chiasmata. Genetics, 121, 255-262. to give much longer SCs in oocytes. The inference DAHM, A. G. 1961. Cytotaxonomical analyses of four Den- from the analysis of reconstructed zygotene nuclei drocoelum species. Lunds Universitets Arsskrft N.F. Avd. that early pairing occurs predominantly from 2, 571-44. 106 G. H. JONES AND J. A. CROFT

GELEI, .r. 1913. Uber die Ovogenese von Dendrocoelum lacteum. OAKLEY, H, A. 1982. Meiosis in Mesostoma ehrenbergii ehren- Arch. Zellforsch., 11, 51—150. bergii (Turbellaria, Rhabdocoela) 11. Synaptonemal com- GELEI,.1921.Die Langskonjugation der Chromosomen. Arch. plexes, chromosome pairing and disjunction in achiasmate Zellforsch., 16, 52—130. oogenesis. Chromosoma, 87,133-147. GELEI, j.1922.Die Konjugationsfrage der Chromosomen in OAKLEY, H. A. AND JONES, ci. H. 1982. Meiosis in Mesostoma der Literatur und meine Befunde. Arch. Zellforsch., 16, ehrenbergii ehrenbergii (Tubellaria, Rhabdocoela)I. 300—370. Chromosome pairing, synaptonemal complexes and GILLIES, C. B., RASMUSSEN, S. W. AND WETTSTEIN D. VON. chiasma localisation in spermatogenesis. Chromosoma, 85, 1973. The synaptinemal complex in homologous and non- 3 11-322. homologous pairing of chromosomes. Cold Spring Harbor PASTOR, J. B. AND CALLAN, H. G. 1952. Chiasma formation in Symp. Quant. Biol., 38, 117-122. spermatocytes and oocytes of the Turbellarian Dendro- HOLM, P. B., RASMUSSEN, S. W., ZICKLER, D., LU, B. C. AND coelum lacteum. I Genet., 50,449-454. SAGE, .i.1981. Chromosome pairing, recombination RASMUSSEN, S. W. AND HOLM, p 1978.Human meiosis. II. nodules and chiasma formation in the Basidiomycete Cop- Chromosome pairing and recombination nodules in human rinus cinereu,c. Carlsberg Res. Commun., 46, 305-346. spermatocytes. Carlsberg Res. Commun., 43, 275—327. JONES.G. H. AND CROFT, J. A. 1986. Surface spreading of RASMUSSEN, S. W. AND HOLM, P. B. 1982. The meiotic prophase synaptonemal complexes in locusts. II. Zygotene pairing in Bombyx mon. Insect ultrastructure. Edited by R. C. King behaviour. Chromosoma, 93, 489-495. and H. Akai. Plenum, New York, London, pp. 61-85. MATERN, R AND SIMAK, M. 1968. Statistical problems in kary- SPEED, R. M. 1977. The effects of ageing on the meiotic chromo- otype analysis. Hereditas, 59, 280—288. somes of male and female mice, Chromosoma, 64,241-254. MATTHIESON, E. 1904. Em Beitrag zur Embryologie der Suss- SPEEL), R. M. 1982. Meiosis in the foetal mouse ovary. I. An wasserdendrocoelen. Z. wiss. Zoo!., 77274-361. analysis at the light microscope level using surface spread- MELANDER, y1963.Cytogenetic aspects of embryogenesis in ing. Chromosoma, 85, 427—437. Paludiocola, Tricladida. Hereditas, 49, 119—166. STERN, H., WESTERGAARD. M. AND WETrSTEIN, D. VON. 1975. MOGENSEN, H. L. 1977. Ultrastructural analysis of female Presynaptic events in meiocytes of Lilium longiflorum and pachynema and the relationship between synaptonemal their relation to crossing-over. A preselection hypothesis. complex length and crossing-over in Zea mays. Carlsberg Proc. Nat! Acad. Sci. USA, 72, 961-965. Res. Commun., 42, 475-497. WALLACE. B. M. N. AND HULTEN, M. 1985. Meiotic chromo- MOSES, M. J. AND POORMAN, P. A. 1984. Synapsis, synaptic some pairing in the normal human female. Ann. Hum. adjustment and DNA synthesis in mouse oocytes. Chromo- Genet,, 49, 215—226. somes Today 8. Proceedings of the8th InternationalChromo- WELLS, B. 1974. A convenient technique for the collection of some Conference. Edited by M. D. Bennett, A. Gropp and ultra-thin serial sections. Micron, 5, 79—81. U. Wolf, George Allen and Unwin, London, pp. 90-103.