Chromosome Alignment and Transvection Are Antagonized by Condensin II

See allHide authors and affiliations

Science  28 Nov 2008:
Vol. 322, Issue 5906, pp. 1384-1387
DOI: 10.1126/science.1164216


Polytene chromosome structure is a characteristic of some polyploid cells where several to thousands of chromatids are closely associated with perfect alignment of homologous DNA sequences. Here, we show that Drosophila condensin II promotes disassembly of polytene structure into chromosomal components. Condensin II also negatively regulates transvection, a process whereby certain alleles are influenced transcriptionally via interallelic physical associations. We propose that condensin II restricts trans-chromosomal interactions that affect transcription through its ability to spatially separate aligned interphase chromosomes.

Interphase chromosomal trans-interactions occur in many species and impact chromosome structure and gene expression (14). As evidenced in Drosophila, trans-interactions can lead to polytene chromosomes, where all maternal and paternal chromatids are aligned in precise register (5). The Drosophila ovarian nurse cells disassemble their polytene chromosomes into unpaired homologs and chromatid fibers during mid-oogenesis (Fig. 1A and fig. S1) (6). We used this system to isolate two noncomplementing mutations in a predicted condensin II subunit, Cap-H2 (79), that cause failure in nurse cell polytene disassembly. Polyteny instead persists in the trans-heterozygous combinations of Cap-H2Z3-0019/Cap-H2Z3-5163 (Fig. 1B and fig. S2A) and when either allele is in trans to a deletion of its genomic locus (Fig. 1, C and D). This was corroborated through fluorescence in situ hybridization (FISH) labeling to a specific locus in stage 7 egg chambers, where wild-type polytenes disassembled, yet 92.9% of mutant nuclei had all maternal and paternal chromatids aligned in register (fig. S3) (7). Polytene persistence in Cap-H2 mutants likely does not occur indirectly through altered cell cycle progression or DNA replication patterns because neither the length of S phase nor ploidy were detectably different in homozygous polytene mutants versus heterozygous controls (figs. S4 and S5) (7). This result instead suggests that Cap-H2 function is necessary to disassemble nurse cell polytene chromosomes.

Fig. 1.

Cap-H2, Cap-D3, and SMC4 are necessary for nurse cell polytene chromosome disassembly. (A to G) DAPI-stained nurse cell nuclei from stage 10 egg chambers. Scale bars, 5 μm. (A) Wild-type (Oregon R). (B) Cap-H2Z3-0019/Cap-H2Z3-5163. (C) Cap-H2Z3-0019/Cap-H2Df(3R)Exel6159. (D) Cap-H2Z3-5163/Cap-H2Df(3R)Exel6159. (E) Cap-D3EY00456/Cap-D3Df(2L)Exel7023. (F) Cap-D3EY00456/Cap-D3Df(2L)Exel7023; Cap-H2Z3-0019/+. (G) SMC4k08819/+; Cap-H2Z3-0019/+ (Table 1). Polyteny in condensin II mutant nurse cells is highly penetrant (n = 150 nuclei for each genotype).

Metazoa have two condensin complexes that are referred to as condensin I and II. Each uses the adenosine triphosphatases SMC2 and SMC4, but forms complexes with different non-SMC subunits Cap-H, Cap-G, and Cap-D2 or Cap-H2, Cap-G2, and Cap-D3, respectively (10, 11). Condensins function in the condensation of chromosomes, facilitate proper anaphase segregation, and in vitro induce and trap DNA positive supercoiling (9, 12, 13). Supercoiling has been proposed to gather chromatin into domains that are then further ordered to assemble metaphase chromosomes (12). Cap-H2 likely acts within a condensin II complex, as other predicted condensin II subunits also regulate nurse cell polytene dispersal. Cap-D3 mutants (fig. S2B) exhibited nurse cell polytene persistence (Fig. 1E) that was enhanced through the introduction of one mutant Cap-H2 copy (Fig. 1F). Furthermore, SMC4/Cap-H2 double-heterozygotes had a loosened, but clear, polytene morphology (Fig. 1G) (7). Consistent with a polytene disassembly function, Cap-H2 protein first becomes enriched within posterior stage 5 and 6 egg chambers, where disassembly is initiated (6), and Cap-H2 is detected in all stage 7 to 10 nuclei (figs. S6 and S7) (7).

Unlike nurse cells, polyteny is persistent in the nuclei of the larval salivary glands (Fig. 2, A and B). Cap-H2 overexpression induced drastic separation of salivary gland polytene chromosomal components, as visualized through green fluorescent protein (GFP) labeling of a second chromosome locus (Fig. 2, C and D, and figs. S8 to S10, movies S1 and S2). In the wild-type, the GFP locus had a width 15.9 ± 1.6% (SEM) of the nuclear radius (Fig. 2, A and B), yet individual foci reached distances 110.8 ± 9.1% (SEM) of the nuclear radius after Cap-H2 induction (Fig. 2, C and D, and fig. S9) (7).

Fig. 2.

Cap-H2 is sufficient to disassemble salivary gland polytene chromosomes and cooperates with Cap-D3. (A and B) Wild-type controls not overexpressing Cap-H2. (C) Nucleus overexpressing Cap-H2. (D) Stacked z series of the same nucleus shown in (C) to demonstrate the full extent of polytene disassembly. Color scale is in microns of the z axis. (E) Cap-D3EY00456/+ control overexpressing Cap-H2. (F) Cap-D3EY00456/Cap-D3Df(2L)Exel7023 mutant overexpressing Cap-H2. In comparison with (E), note Cap-D3′s ability to rescue Cap-H2–induced polytene disassembly. Scale bars, 5 μm.

In the salivary gland, Cap-H2–induced polytene disassembly occurs only 6 hours after Cap-H2 overexpression in fully developed late larvae (fig. S10), which makes it unlikely that disassembly is an indirect consequence of altered larval development. It is also improbable that disassembly occurs through the creation of large-scale chromosomal breaks, because this was not detected after Cap-H2 overexpression, and that the creation of DNA breaks with γ-radiation did not alter polytene alignment (fig. S11) (7). Rather, the ability of Cap-H2 overexpression to induce polytene disassembly indicates that polytene alignment of chromatids is constrained with wild-type Cap-H2 levels. Providing excess Cap-H2 may induce polytene disassembly because its dosage is limiting to other condensin II subunits in salivary glands and/or it acts as a catalytic subunit that promotes condensin II activity. Cap-H2 does rely on Cap-D3 to induce polytene disassembly as all salivary gland nuclei from a Cap-D3 mutant background overexpressing Cap-H2 had polytenes that appeared like the wild-type (n = 195 nuclei). This contrasted to Cap-D3 heterozygous controls, where only 24.1 ± 8.5% (SEM) nuclei per gland contained wild-type polytenes (n = 182) (Fig. 2, E and F) (7).

Because of Cap-H2's ability to transform aligned polytene structure into chromosomal components, we predicted it would function in a similar manner to disrupt aligned loci within diploid somatic cells. We therefore investigated whether it regulates diploid trans-chromosomal interactions by studying its role in transvection, a phenomenon whereby certain mutant alleles are influenced transcriptionally via association with their homologous locus (1). It is inferred from transvection phenomena that somatic homolog pairing also plays a role in regulating wild-type loci.

The first transvection system we utilized involves the gain-of-function mutation UbxCbx-1, which causes misexpression of Ubx in the imaginal wing disc and leads to a partial wing-to-haltere transformation (1416). Awing transformation occurs even in flies where Ubx of the UbxCbx-1 allele is rendered null through the introduction of a second mutation (UbxCbx-1 Ubx1) (Fig. 3A, compare class A with class D). This UbxCbx-1 Ubx1/++ phenotype suggested that the Cbx1 lesion is capable of transcriptionally activating the wild-type Ubx on the homologous chromosome through a trans physical association (14, 17). This was supported by the ability of chromosomal rearrangements (R) that disrupt homolog pairing at Ubx to suppress transvection (Fig. 3A, compare class D with class B) (14, 17, 18). Consistent with a role for Cap-H2 in antagonizing homolog pairing, the UbxCbx-1 Ubx1/++ phenotype was dominantly enhanced by Cap-H2 mutations [(Fig. 3A), compare class D with class E (Fig. 3C and table S1)]. Conversely, Cap-H2 overexpression suppressed the UbxCbx-1 Ubx1/++ wing phenotype closer to wild-type [Fig. 3A, compare class D with class B (Fig. 3C and tables S3 to S5)].

Fig. 3.

Cap-H2 negatively regulates transvection at Ubx and yellow loci. (A) Wing phenotypes depicting Cap-H2 modification to UbxCbx-1 Ubx1 transvection. Class A: wild-type. Class B: typical R(UbxCbx-1 Ubx1)/++ wing. Notice slight loss of tissue at the posterior. Class C: enhancement of the R(UbxCbx-1 Ubx1)/++ wing phenotype found through introduction of a Cap-H2 mutation. Notice withering at the posterior. Class D: typical UbxCbx-1 Ubx1/++ wing. Class E: enhancement of the UbxCbx-1 Ubx1 wing phenotype through introduction of a Cap-H2 mutation. (B) Flies depicting Cap-H2 enhancement of y1#8/y82f29 transvection. (Left) Wild-type. (Middle) y1#8/y82f29; Cap-H2 Z3-0019/+ control illustrating “light” stripes. (Right) y1#8/y82f29; Cap-H2Z3-0019/Cap-H2Df(3R)Exel6159 with “dark” stripes. (C) UbxCbx-1 Ubx1 transvection data (n = 50 wings) (see tables S1 to S5 for complete data set). (D) Quantification of y1#8/y82f29 transvection enhancement by Cap-H2 mutants. Darker stripes were consistently found for y1#8/y82f29; Cap-H2Z3-0019/Cap-H2Df(3R)Exel6159 flies relative to the two heterozygous controls (bars indicate SEM, n = 97, 95 (controls), and 43 (trans-heterozygote) flies, P < 0.005, two-tailed t test assuming unequal variances).

Cap-H2 mutant enhancement of the UbxCbx-1Ubx1 phenotype was suppressed in a chromosomal rearrangement background [R(UbxCbx-1 Ubx1)/++] that is thought to disrupt allelic associations between UbxCbx-1 Ubx1 and wild-type Ubx [(Fig. 3A), compare class E with class C (Fig. 3C, table S2)] (7). The UbxCbx-1 Ubx1/Cap-H2 and R(UbxCbx-1 Ubx1)/Cap-H2 flies only vary by the reciprocal translocation that moves 3R bearing UbxCbx-1 Ubx1 to 2R and vice versa. This suggests that Cap-H2 enhancement of the UbxCbx-1 Ubx1 phenotype is through increasing the association of homologous loci. Alternatively, Cap-H2 function may follow trans-chromosomal interactions, for example, acting locally to enable enhancer interactions in trans or as a general transcriptional repressor. Although either is formally possible, Cap-H2's ability to globally disrupt aligned polytene structure suggests it carries out a related function in diploid cells to antagonize trans-chromosomal interactions.

Cap-H2 was tested in a second transvection system involving mutant alleles of the gene yellow (y). In y82f29/y82f29 and y1#8/y1#8 flies, there is minimal cuticle pigmentation, yet when placed in trans to one another (y82f29/y1#8) complementation occurs with partial restoration of pigment nearer to wild-type levels. The y1#8 allele is a deletion of the yellow promoter and the y82f29 allele a deletion of upstream enhancer elements. It is thought that partial complementation occurs in y82f29/y1#8 through the ability of y1#8's enhancers to act in trans, to associate with the intact promoter of y82f29, and then to activate yellow transcription (19, 20). As are UbxCbx-1 Ubx1, transvection of y82f29/y1#8 is enhanced in a Cap-H2 mutant background, which leads to darker pigmentation of the abdominal stripes relative to controls (Fig. 3, B and D) (7).

Transvection can be enhanced by slowing the rate of cell division (21). The percent of Cap-H2 homozygous mutant cells specifically in mitosis was cytologically found to be greater relative to heterozygous controls, but this was statistically insignificant [1.37% (n = 1318) versus 0.95% (n = 3375), P > 0.05 (χ2)]. Furthermore, with flow cytometry, homozygotes and heterozygotes did not vary significantly in the percentage of cells in G1, S, and G2/M (fig. S12). Although these data do not rule out a cell cycle delay leading to enhanced transvection, they also do not support a major regulatory role for Cap-H2 in cell cycle progression. Cap-H2's ability to disassemble the aligned structure of polytene chromosomes instead suggests that it antagonizes transvection by inhibiting homology-dependent chromosomal interactions in diploid somatic cells.

Just as condensin-mediated supercoiling has been proposed to initiate chromosome condensation (12), we speculate that supercoiling activity also exists in interphase nuclei and can disrupt chromosome alignment. This may be through providing a force that physically disrupts interchromosomal associations and/or favors intrachromosomal higher-order structures that make inaccessible regions prone to trans-associate (fig. S13). This condensin activity may be a crucial aspect of gene regulation by disrupting trans-communication of allelic regulatory elements.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S13

Tables S1 to S5


Movies S1 and S2

References and Notes

Table 1.

Percentage of polytene nurse cells in condensin II mutant backgrounds. The number of polytene nurse cells per stage 10 egg chamber was quantified (n = 10 egg chambers and 150 total nuclei) (±SEM).

View this table:
View Abstract

Navigate This Article