Regulation of Cell Cycle Synchronization by decapentaplegic During Drosophila Eye Development

See allHide authors and affiliations

Science  10 Jan 1997:
Vol. 275, Issue 5297, pp. 203-206
DOI: 10.1126/science.275.5297.203


In the developing Drosophila eye, differentiation is coordinated with synchronized progression through the cell cycle. Signaling mediated by the transforming growth factor-β-related gene decapentaplegic (dpp) was required for the synchronization of the cell cycle but not for cell fate specification. DPP may affect cell cycle synchronization by promoting cell cycle progression through the G2-M phases. This synchronization is critical for the precise assembly of the eye.

The Drosophila eye has served as a model system for examining the molecular mechanisms that govern the patterning and assembly of a complex tissue. The adult eye develops from an epithelial monolayer known as the eye imaginal disc. Differentiation begins in the posterior end of the disc and progresses anteriorly, marked by an indentation in the disc epithelium called the morphogenetic furrow (MF). Within the MF, unpatterned cells are induced to differentiate into the highly ordered array of retinal cells and nonneural accessory cells that produce the 750 ommatidia of the adult eye (1, 2). Differentiated cells posterior to the MF express the signaling molecule hedgehog (HH), which directs the anterior advancement of the furrow (3, 4) and induces the expression of DPP within the furrow. DPP mediates cell fate determination in the developing wing and leg in response to HH (5, 6).

During the initial stages of sensory unit assembly in the MF, precursor cells exhibit synchronization of the cell division cycle (2, 7). Anterior to the MF, cells divide randomly, but just ahead and within the MF, cell divisions occur more coordinately. The first evidence of cell cycle synchronization is increased amounts of cyclin B (8), a mitotic cyclin required for entry into mitosis (9). As cells enter late G2, cyclin B levels peak and cyclin B then degrades at the metaphase-anaphase transition in mitosis. Loss of cyclin B and the appearance of mitotic figures mark coordination in the M phase (Fig. 1, A through C). After completion of mitosis, cells arrest in G1 in the MF and transcription of dpp is activated (Fig. 1C).

Fig. 1.

Inappropriate expression of cyclin B in the MF by cells lacking the dpp receptor tkv. Confocal images of third instar eye discs stained with antibody to cyclin B (anti-cyclin B) and propidium iodide (A and B), anti-β-galactosidase (anti-β-Gal) and propidium iodide (C), or anti-cyclin B and anti-Myc (D through I) are shown (18). In this figure and subsequent figures, anterior is to the left and posterior is to the right. The MF is visible in cross section (A) or in tangential sections [brackets in (B) through (I) and in Figs. 2 and 3] as an indentation in the epithelium. The first mitotic domain is located anterior to the furrow, and cyclin B staining (red) ends abruptly as cells finish mitosis and enter G1 in the MF (A and B). Cyclin B staining resumes posterior to the furrow in domain 2 (A and B). Propidium iodide (gray) labels the nuclei, and mitotic figures that are apically localized are seen as intensely stained nuclei located anterior to the MF in domain 1 [left arrowhead in (C)] or posterior to the MF in domain 2 [arrowheads in (A) through (C)] but never in the furrow. In (C), anti-β-Gal staining of the dpp reporter line BS3.0 shows that DPP (arrow) is localized in the MF between domains 1 and 2 (all arrowheads). Two independent tkv clones are shown in (D) through (F) and (G) through (I). In (D) and (G), the clones are identified by loss (arrow) of anti-Myc staining (green). tkv mutant clones spanned the MF and inappropriately expressed cyclin B [white arrowhead in (E) and (H)] in the anterior part of the furrow. (F) shows the merge of (D) and (E) and (I) shows the merge of (G) and (H). Cyclin B expression anterior (yellow arrowhead) or posterior (yellow arrow) to the furrow was unaltered in tkv clones (H and I).

Cells respond to DPP through two type I receptors, thick veins (tkv) or saxophone (sax), and a type II receptor, punt (put) (10). Responses to DPP are also attenuated by mutations in the putative transcription factor schnurri (shn) (11). We examined whether DPP was necessary for cell fate specification or cell cycle synchronization by determining whether cells defective for tkv, sax, or shn showed abnormalities in cell division or differentiation. Clones mutant for a null allele of tkv that were anterior or posterior to the furrow had amounts of cyclin B that were indistinguishable from those in surrounding normal cells (Fig. 1, G through I). In contrast, tkv clones that spanned the MF maintained cyclin B expression in the anterior part of the furrow even though the surrounding cells arrested in G1 had no detectable cyclin B (Fig. 1, D through I). Clones defective for sax and shn also showed disruption of cyclin B expression when the clone encompassed the anterior half of the furrow (Fig. 2, A through F). Clones spanning the MF reveal that cyclin B was finally lost in the posterior region of the furrow, despite the absence of DPP signaling. These clones had condensed chromosomes indicative of early stages in mitosis at the interface between the cyclin B-expressing and nonexpressing cells (Fig. 3, A and B).

Fig. 2.

Expression of cyclin B in the anterior part of the MF in cells that are mutant for sax or shn. Discs were labeled with anti-Myc (A and D; green) and anti-cyclin B (B and E; red) as in Fig. 1 (18). (C) and (F) are merges of (A) and (B) and (D) and (E), respectively. A sax clone [arrows in (A) and (C)] spanned the MF and inappropriately expressed cyclin B in the anterior part of the furrow [arrowheads in (B) and (C)]. Three shnIB clones [arrows in (D) and (F)] were located within the MF. Cyclin B was visible as an island within the MF in (E) and (F) (arrowheads). Only the most anteriorly located cells of the two upper clones expressed cyclin B; the bottom clone, which was slightly more posterior, had no cyclin B expression.

Fig. 3.

Mitotic figures and mislocalized nuclei in tkv and shn clones. Third instar discs were triple-labeled with anti-Myc, anti-cyclin B, and propidium iodide as in Figs. 1 and 2 or with anti-Myc and anti-ATONAL. Clones are indicated by dashed yellow lines in (A) through (D), (G) and (H), and (J) and (K). Nuclei were mislocalized in tkv clones as cells began to enter the MF and were not visible in apical sections [arrow in (A) and (B)]. Mitotic figures were discernible in the MF within and slightly posterior to the ectopic cyclin B expression [arrowhead in (A) and (B)]. In tkv clones that were posterior to the MF, nuclei were underrepresented in apical sections (C) and overrepresented in basal sections (D). In shn clones [arrow in (E)] within the MF, condensed mitotic chromosomes were discernible in some regions that expressed ectopic cyclin B [arrowheads in (F) and (G)], but most of the remaining nuclei were not visible in apical focal planes [(G) and (H)]. In tkv clones (I), atonal expression was still present in the presumptive R8 cell (J and K), although apical sections revealed that the nuclei were mislocalized.

The expression of cyclin B in the mutant cells in the furrow could be interpreted as evidence that DPP inhibits progression through G1 in the furrow. In this case, cells in the clone might be expected to progress through G1 and S and reenter G2, where they would again exhibit cyclin B expression. Failure to arrest in G1 should produce a clone with patchy expression of cyclin B in the anterior half of the furrow as the asynchronously dividing cells continue through the cell cycle. However, cyclin B expression was continuous in the anterior part of the furrow in tkv and sax clones in which we examined serial optical sections (Fig. 1, D through I, and Fig. 2, A through C). Furthermore, mitotic figures were not observed in clones in the anterior half of the MF, which suggests that the cells in the clones were not asynchronously progressing through M and into a new cell cycle. The phenotype observed in the clones is similar to defects caused by mutations in the gene division abnormally delayed (dally), which is required for G2-M progression ahead of the furrow (8). Mutations in dally and dpp display genetic interactions in development of the eye, antennae, and genitalia, which suggests that dally augments dpp function (12). Regulation of cell division in the G2-M transition occurs at several stages in Drosophila development, including in embryonic divisions (13), in the wing disc (14), and in the eye disc (4). The behavior of DPP-receptor mutant clones supports a role for DPP in controlling progression through G2-M as a means of synchronizing the divisions that accompany differentiation in the eye disc.

We next examined whether this defect in synchronization affected patterning or differentiation of cells. For example, mutations in roughex (rux) prevent arrest in G1, cause cells to enter the S phase prematurely, and result in disruptions in cell cycle synchronization and differentiation (7). The assembly of ommatidia and the synchronization of the cell cycle within the furrow are accompanied by changes in the position of nuclei within the epithelium (1, 15). We examined the position of nuclei in tkv, sax, and shn clones as an indicator of events that accompany ommatidial assembly. Clones at the anterior edge of the furrow had cells with mislocalized nuclei; they failed to reach the apical surface where mitosis normally takes place in domain 1 (Fig. 3, A and B). Nuclei were also mislocalized when the clone was within the MF and posterior to it (Fig. 3, C through H). Nuclei were underrepresented in apical sections where mitotic figures are normally present (Fig. 3, A through C and G and H) and were overrepresented in basal sections (Fig. 3D). Although nuclei were mislocalized in these clones, cell fate specification was mostly unaffected, as revealed by the expression of the HH-inducible gene atonal, a proneural gene required for retinal precursor cell 8 (R8) determination (Fig. 3, I through K) (16), and of the antigen recognized by monoclonal antibody 22C10 (17). Because atonal expression was maintained in tkv clones, hh must not act through dpp to induce its expression, and thus dpp only mediates a subset of hh functions in the MF (Fig. 4A). Tangential sections through adult eyes confirmed that some ommatidia were disorganized in tkv clones and lacked the full number of photoreceptor cells (Fig. 4B). Therefore, DPP signaling in the MF appears to be required for cell cycle synchronization and the assembly of ommatidia but not for the specification of cell fate.

Fig. 4.

Defects in cell cycle progression and ommatidial assembly caused by lack of DPP signaling. (A) hh is expressed in differentiated cells and activates dpp transcription in the MF. DPP may diffuse anteriorly and induce cells to enter M ahead of the furrow, possibly by modulating the transcription or activity of a gene involved in G2-M progression. rux prevents cells in G1 from entering the S phase prematurely (7). (B) Tangential section of an adult eye that contained a tkv clone, identified by the lack of pigment cells (19). Normal ommatidia have eight photoreceptor cells, which contain rhabdomeres or rhodopsin-rich apical surfaces, seven of which are visible in each cluster in this section. The clone occurs in the middle of the eye along the dorsal-ventral midline or equator. In the clone, several ommatidia are missing photoreceptor cells (arrowheads). One ommatidia that consists of mostly wild-type tissue also contains abnormalities (arrow).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
View Abstract

Stay Connected to Science

Navigate This Article