The teflon Gene Is Required for Maintenance of Autosomal Homolog Pairing at Meiosis I in Male Drosophila melanogaster

RESULTS

tef is a male-specific mutation that affects autosome segregation: We identified five EMS-induced mutations in a screen for mutations that increased the frequency of progeny lacking a paternal fourth chromosome. The characterization of four of these mutations is described here. We mapped these mutations by recombination to 80.0 MU on chromosome 2, and between two P elements inserted at salivary gland chromosome bands 53F and 54A (see materials and methods). We have examined these mutations for their effects on meiotic and mitotic chromosome segregation and to determine if they exhibit chromosome and/or sex specificity.

Sex and fourth chromosome nondisjunction were simultaneously monitored by scoring the progeny of Xy/y⁺Y; tef; spa^pol males mated to y w sn³; C(4)EN, ci ey females. From such matings, diplo-XY and nullo-XY sperm resulting from sex chromosome nondisjunction produce y⁺ daughters and y w sn sons, respectively. Fourth chromosome nondisjunction results in diplo-4 sperm, which give rise to spa progeny, and nullo-4 sperm, which produce ci ey progeny. We tested all possible combinations of the tef alleles by such assays. No increases in sex chromosome nondisjunction or loss were observed. In contrast, fourth chromosome exceptions were observed at frequencies between 28.1 and 43.7% (Table 1). For most of the allelic combinations, the frequency of the nullo-4 progeny slightly exceeded that of the diplo-4 progeny, suggesting that tef may cause some loss in addition to nondisjunction.

Fourth chromosome nondisjunction is underestimated by these crosses because we could not distinguish tetra-4 progeny (spa/spa/C(4)EN, ci ey), which result from diplo-4 sperm, from progeny resulting from normal haplo-4 sperm (spa/C(4)EN, ci ey). From independent crosses of C(4)EN spa males and C(4)EN, ci ey females, we found that 40-50% of such tetra-4 progeny survive to adulthood (data not shown). If we adjust our data to account for this by subtracting half the number of diplo-4 progeny from the wild-type class, the true rate of fourth nondisjunction caused by several of the tef alleles approaches 50%. This suggests that fourth chromosome homologs may be segregating at random at meiosis I.

We asked if tef had similar effects on sex and fourth chromosome segregation in females by mating y; tef; spa^pol females to y w sn/y⁺Y; C(4)EN, ci ey males. We observed wild-type frequencies of sex and fourth chromosome nondisjunction in females homozygous for each of the four alleles (<0.5%, data not shown). Thus, all four tef alleles are male specific.

To determine if tef is specific for the fourth chromosome, or if the other autosomes are also affected, we crossed tef males to females bearing compound autosomes. Females carrying C(2)EN, bw sp or C(3)EN, st cu e produce eggs that are nullo- or diplo- for chromosome 2 or chromosome 3, respectively. The only progeny that survive from these females are products of sperm bearing zero or two copies of the major autosome assayed. These assays are qualitative in nature; because euploid sperm do not produce viable offspring we cannot determine absolute nondisjunction frequencies.

For each test, we mated 50 tef or tef/Cy control males to 100 compound-bearing females and counted the adult progeny produced. Males homozygous for any of the four tef alleles produced significantly more progeny than their heterozygous tef/Cy brothers (Table 2). Because the compound autosomes were marked with visible recessive mutations, we could distinguish progeny produced from diplo- vs. nullo-exceptional sperm. In all cases, there were considerably more progeny produced by nullo-exceptional sperm. This difference may reflect meiotic autosome loss or selection for nullo-over diplo-autosomal sperm (e.g., meiotic drive).

To estimate the frequency of nondisjunction of chromosomes 2 and 3, we collected eggs from the compound-bearing females mated to tef ^z5864 males and counted the proportion that hatched. Because embryos aneuploid for either chromosome 2 or 3 do not survive to hatch (Merrillet al. 1988), this number allows estimation of the nondisjunction frequency. From crosses to C(2)EN females, 75/500 (15%) hatched, and from crosses to C(3)EN females, 61/500 (12.2%) hatched. These numbers are not significantly different from the 12.5% egg hatch expected if the autosomes were segregating at random. Together with the data on fourth chromosome disjunction, this suggests that all autosomes may segregate at random at meiosis I in tef ^z5864 males.

Meiosis I specificity of the tef defect: To address whether tef is required for chromosome disjunction during male germline mitotic, MI, or MII divisions, we performed two experiments. We first asked if tef affected mitotic germline divisions or was specific to meiosis. In testes, stem cells continuously divide throughout adulthood to replenish depleted sperm pools (Hannah-Alava 1965). Thus, if mitotic nondisjunction events were contributing significantly to the overall nondisjunction frequency, then nondisjunction should increase with age (i.e., in proportion to the number of stem cell divisions preceding meiosis). We measured fourth chromosome disjunction in five successive broods from males homozygous for each of the four tef alleles and observed no increase in nondisjunction or loss in later broods (data not shown). This suggests that tef does not affect germline stem cell mitoses, but rather that the defect occurs in meiosis.

To discriminate between a meiosis I and meiosis II defect, we crossed tef males that were heterozygous for the fourth chromosome recessive marker spa^pol to tester C(4)EN, ci ey females. The informative progeny from these crosses are the diplo-4 exceptions. Fourth chromosome nondisjunction at meiosis I in these males would produce diplo-4 sperm heterozygous for the spa^pol mutation and thus produce spa⁺ progeny. Precocious sister chromatid separation at meiosis I or nondisjunction of sister chromatids at meiosis II could produce diplo-4 sperm homozygous for spa^pol, which would result in spa progeny. Males bearing each of the four alleles of tef were tested, and of 2916 total progeny scored, 615 (21.1%) were ci ey, but none were spa. This suggests that meiosis II is normal in tef males and that the defect in chromosome segregation is specific to meiosis I. Furthermore, it argues that tef mutations do not affect sister chromatid cohesion.

tef affects a pairing-dependent aspect of meiosis I chromosome segregation: The events of meiosis I chromosome segregation can be generally classified as pairing dependent or pairing independent. Pairing-dependent processes include partner recognition, establishment and regulation of homolog connections, and orientation of homologs to opposite spindle poles. Pairing-independent events include spindle assembly, kinetochore function, sister chromatid cohesion, and anaphase movements. To ask which general class of events might be affected by tef mutations, we examined the transmission of univalent chromosomes from tef males. Univalent chromosomes lack a pairing partner, and thus their behavior should not be affected by mutations in pairing-dependent processes.

We separately assayed the transmission of two univalent chromosomes, a mini-chromosome derivative of the X, Dp(1;f)1187, y⁺ and the compound-4 C(4)EN. Dp(1;f)1187 is a 1.3-Mb chromosome that retains all sequences required for normal transmission (Karpen and Spradling 1992) but lacks XY pairing sites and assorts randomly with respect to sex chromosomes during meiosis I in males (Parry and Sandler 1974). Although we had failed to see an effect of tef on the segregation of intact sex chromosomes, it remained a possibility that this smaller derivative might show a size-dependent sensitivity.

Transmission of Dp(1;f)1187, y⁺ was measured by the percentage of yellow⁺ progeny produced by Xy/Dp(1;f) 1187, y⁺/Y males mated to y females. The results were similar for each of the four tef alleles tested. In total, males homozygous for tef produced 46.4% Dp-bearing progeny (of 4068 total) whereas control X/Y/Dp; tef/SM1,Cy brothers produced 45.4% Dp-bearing progeny (of 3592 total). Thus the transmission of Dp(1;f)1187 was unaffected by tef mutations.

Transmission of the compound-4 was measured by the percentage of ci ey progeny produced from matings of tef; C(4)EN, ci ey or control tef/SM1, Cy; C(4)EN, ci ey males to C(4)EN, spa^pol females. (Note that the C(4)EN/C(4)EN progeny also received the paternal C(4)EN chromosome, but were excluded from consideration here because their survival is highly variable from cross to cross.) In tests of all four tef alleles, the transmission of the C(4)EN chromosome from tef homozygous males was similar to that of their tef/SM1, Cy brothers. From tef males, a combined total of 36.7% of the progeny were ci ey (1907 ci ey and 3292 spa F₁). Control tef/SMI males produced 32.7% ci ey progeny (1615 ci ey and 3323 spa F₁). These results suggest that the tef gene product is involved in a process that is either required for or dependent on homolog pairing.

tef influences the behavior of translocations between a sex chromosome and autosome: We reasoned that if tef specifically affects a process related to autosome pairing, then it might modify the meiotic behavior of translocations between a sex chromosome and autosome. In such translocations, homologous centromeres can undergo nondisjunction resulting in adjacent II segregation. Elimination of pairing between the autosomal portions of such a translocation might be expected to increase the frequency of proper disjunction of the sex chromosomes.

We tested this premise using T(Y;2h; 4)CB25, a translocation in which the fourth chromosome is attached to the tip of the long arm of the Y (Ault and Lyttle 1988). This chromosome regularly pairs with both the X and the free fourth chromosome. Otherwise wild-type males bearing this translocation produce 10% XY exceptions and 15-20% fourth chromosome exceptions (Ault and Lyttle 1988). This translocation appears to retain both the Y and fourth centromeres, and it has been postulated that nondisjunction results from dicentric behavior (Aultet al. 1982). However, in the majority of meioses examined, only a single centromere region stains with antibodies to the essential kinetochore component ZW10 (Williamset al. 1996). Thus, although both centromeres on this translocation may be capable of functioning, it appears that nondisjunction may be a consequence of its pairing configuration rather than the opposing activities of two centromeres.

The frequency of sex chromosome and fourth chromosome nondisjunction was monitored by matings of T(Y;2h;4)CB25-bearing males to y w sn; C(4)EN, ci ey females. We could not detect progeny produced from diplo-4 sperm from these crosses, because the translocation was wild type for fourth chromosome visible markers. However, fourth chromosome nondisjunction could be estimated by the frequency of nullo-4 exceptions and was elevated by tef as expected. In contrast, sex chromosome nondisjunction was more than fivefold lower in tef males compared to their tef/Cy brothers (Table 3). We suggest that these results reflect a decrease in the ability of the free fourth chromosome to direct the orientation of its homolog in T(Y;2h;4)CB25; tef males. The orientation of the translocation on the meiotic spindle may be dictated by the relative strength of the pairing interactions between the sex chromosomes compared to that of the autosomal portions. If so, then the absence of stable interactions between the fourth chromosomes may increase the frequency of orientation of the translocation opposite the X chromosome. This interpretation implies that the tef gene product normally plays a role in orientation of the fourth chromosome bivalent at meiosis I either directly or indirectly through the establishment or maintenance of homolog connections.

We examined the segregation of two additional translocations in tef males to determine if the tef effect was general or specific to CB25. We previously isolated two X;4 translocations in which the fourth chromosome is appended to the short right arm of the X. Both of these translocations (called mscd1 and mscd2 for male sex chromosome disjunction) cause nondisjunction of the sex chromosomes and can pair in a trivalent with the free fourth chromosome and the Y at meiosis I (Briscoe and Tomkiel 2000). In homozygous tef ^z5864 males, each mscd translocation segregated from the Y chromosome with higher fidelity than in tef ^z5864/Cy siblings (Table 4). These results are consistent with the hypothesis that mscd sex chromosome nondisjunction results at least in part from trivalent formation with the fourth chromosome and that tef mutations alter this pairing configuration or its consequences on orientation.

Cytological examination of meioses in tef males: We examined DAPI-stained meiotic chromosome spreads from squashes of testes from tef males and their heterozygous tef/+ brothers for visible defects in chromosome behavior. Males homozygous for each of the four alleles were examined, as well as trans-heterozygotes for the tef^z4169 and tef^z5864 alleles. Phenotypes were similar in all tef allelic combinations. The earliest abnormality was detected at the S6 stage of spermatocyte growth (Cenciet al. 1994). At these stages in wild-type males, the two major autosomes and sex chromosome bivalents appear as three discrete DAPI-positive bodies. The fourth chromosomes can occasionally be visualized as a small spot, but because of their small size are not always visible. This stage is considered to be late meiotic prophase, although prophase in Drosophila spermatocytes differs from that of most organisms in that the progressive condensation and synapsis of homologs that can be seen in most organisms does not occur. Rather, the chromatin remains in three large clumps and a small spot. Real-time observations of chromosome condensation and movements in spermatocytes expressing the green fluorescent protein-labeled histone H2AvDGFP (Clarkson and Saint 1999) confirm that these clumps each contain a single bivalent (J. Tomkiel, unpublished results). Whether homologs are paired in these clumps or are merely associated as a vestige of mitotic somatic pairing is unknown. The end of this stage is characterized by the disassembly of the network of microtubules surrounding the nucleus and the appearance of asters, which begin to migrate to opposite sides of the spermatocyte nucleus (Cenciet al. 1994). In tef males, the homologs of either one or both of the autosomal bivalents were separate in 10% of the S6 stage meiocytes examined (143/1400, Figure 1). This phenotype was never observed in over 1000 tef/+ and wild-type meiocytes examined at this stage.

In squashed preparations of wildtype or tef/+ metaphase I cells, the three major bivalents and the fourth chromosome bivalent often appear as three large and one small DAPI-positive body (Figure 2A). Thus, the forces placed on the cell during squashing often disperse chromosomes from the metaphase plate, but very rarely disrupt homolog associations. The only homologs that are typically separated in wild-type squashes are the fourth chromosomes. The fourth chromosomes often disjoin slightly before the other bivalents (Linet al. 1981). In contrast, in tef meiocytes, unpaired autosomes were frequently observed, often at or near the spindle poles, whereas the sex bivalent was intact and often positioned medially between the poles (Figure 2, B-D). In addition, the precocious separation of the fourth bivalent appeared exaggerated in tef cells, as the fourth chromosomes had often reached the spindle poles before the sex chromosomes had visibly separated (Figure 2D). These phenotypes suggest that tef either weakens autosomal homolog associations or alters the timing of their dissolution.

To avoid the possibility of artifacts introduced by squashing, we also examined meiocytes in whole mounts of dissected or intact testes using confocal microscopy. Meiocytes from tef/+ and tef males were fixed and stained with antitubulin antibodies and DAPI, and spindle morphology and chromosome distributions were compared. No defects in meiotic spindles could be detected in tef mutants. In 50 tef/+ cells, all chromosomes were observed together in a single cluster on the metaphase plate at stages M2 and M3. These stages correspond to the end of prometaphase, in which chromosomes have nearly completed (M2) or completed (M3) congression to the metaphase plate (Cenciet al. 1994). In tef/+ control males, the chromosomes were usually clustered such that boundaries of individual chromosomes could not be discerned, but in some cells single bivalents were observed slightly separated from the plate (Figure 3, top). These were presumed to be in M2 and still in the process of congression. In 10/50 tef/+ cells one or both fourth chromosomes were observed slightly separated from the other chromosomes. In contrast, in stage M2 and M3 tef cells the chromosomes were dispersed along the pole-to-pole axis. In 72 metaphase tef cells in which the fourth chromosomes could be seen, they appeared unpaired (Figure 3, middle and bottom). In 48/96 metaphase cells, homologs of at least one of the major autosomes were also unpaired. In an additional 22 cells, the autosomal bivalents had an unusual dumbbell-like shape or were separated by a slight gap, suggesting that they were not as tightly associated as in wild type (Figure 3, middle). These bivalents were also often displaced from the metaphase plate, perhaps also indicating a problem in orientation. In contrast, in all tef cells in which the sex chromosomes could be identified, they were closely apposed in a configuration not discernibly different from wild type. These observations suggest that pairing and/or subsequent conjunction, specifically between autosomal bivalents, is defective in tef mutants and that subsequent failure of unpaired or weakly paired autosomes to orient to opposite poles results in nondisjunction.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

—DAPI-stained primary spermatocytes from tef ^z5864 and sibling tef ^z5864/Cy males, showing normal appearance of chromatin and microtubules in tef males during the S5 stage of spermatocyte growth. In contrast, a separation between homologs of the autosomal bivalents is observed at late prophase, stage S6, in tef meiocytes.