Index Introduction Materials and Methods Fly strains Collection of pupae and dissection BrdU incorporation Immunohistology Measurement of bristle length and size of the nucleus in the socket cell and the shaft cell Generation of somatic clones X-GAL staining of pharate adult thoraxes Statistical analysis Results DREF is required for external sensory organ development in the adult thorax DREF is not required for SOP differentiation DREF regulates the timings of asymmetric cell division but perhaps plays no direct role in cell differentiation during asymmetric cell division PSC bristle cell lineages still exist underneath the notum of adult flies deprived of DREF DREF is critical for growth and endoreplication in shaft cells DREF regulates pcna expression in endocycling PSC shaft cells DREF is necessary but not sufficient for nuclear growth and protein synthesis in shaft cells Discussion DREF plays critical roles in cell division process, especially in endoreplication process of bristle development, at least in part by regulating the pcna gene expressions Regulation of cell differentiation by DREF during bristle development Regulation of cell division by DREF during bristle development Regulation of replication and cell growth by DREF during bristle development Regulation of protein synthesis by DREF during bristle development Advantages of machrochaete development system to study mechanism of replication and growth in vivo Acknowledgments References

Introduction

The DNA replication related element (DRE) (5'-TATCGATA) is one of the most conserved sequences in core-promoters of many Drosophila genes (Ohler et al., 2002). The zinc-finger transcription factor DREF (DNA replication-related element-binding factor), first identified as dimmers which bind to DRE sequence (Hirose et al., 1993, 1996) to activate various target genes (Matsukage et al., 2008). It has been reported that TRF2 forms complexes with DREF and regulates pcna gene transcription by core-promoter selectivity (Hochheimer et al., 2002). In fact, it is now thought that DREF has dual characteristics as a component of basal transcription machinery and as a transcriptional activator featuring sequence specific binding.

Target genes of DREF are highly-diverse, ranging from examples related to DNA replication, cell cycle regulation, signal transduction, transcriptional regulation, differentiation, protein synthesis and degradation, to those contributing to chromatin structures (Matsukage et al., 2008). Many target genes have been identified using in vitro cultured cell systems, such as e2f1 (Sawado et al., 1998), pcna (Hirose et al., 1993), rfc140 (Tsuchiya et al., 2007), DNA polymerase α 180kD (DNApol-α180) (Hirose et al., 1991), DNA polymerase α 73kD (DNApol-α73) (Takahashi et al., 1996) and others (Matsukage et al., 2008). Our previous studies provided direct evidence for in vivo transcriptional regulation of replication-related genes by demonstrating that DREF regulates DNApol-a180 and e2f1 in salivary glands (Hirose et al., 1999). Hyun et al. also showed that DREF is required for transcription of several cell cycle and replication-related genes in the wing imaginal discs (Hyun et al., 2005). In addition, serial analysis of gene expression (SAGE) revealed that most genes expressed in glass-negative cells anterior to the morphogenetic furrow (MF) have DRE sequences within 1000 bp of their transcription initiation sites (Jasper et al., 2002). In vivo functional analyses of DREF showed that over-expression of DREF in the GAL4/UAS targeted expression system induced ectopic DNA synthesis in post mitotic cells of the eye disc (Hirose et al., 2001). While over-expression of a dominant negative form of DREF caused replication defects in both salivary glands and mitotic bands posterior to the MF (Hirose et al., 1999), indicating it is essential for replication in both mitotic cycles and endocycling. DREF is required for efficient growth and cell cycle progression in Drosophila imaginal discs (Hyun et al., 2005) and a human orthologue of Drosophila DREF (hDREF), identified by BLAST search, plays important roles in proliferation in HeLa cells by regulating transcription of the histone H1 gene (Ohshima et al., 2003) as well as some ribosomal protein genes (Yamashita et al., 2007). However, due to the relative lack of cell biological studies in vivo, the question of how DREF plays a role during a single cell development has yet to be answered in detail. To address this issue, we here focused our attention on the thorax mechanosensory bristle development system.

There are about 200 small bristles (microchaetes) and 22 large bristles (macrochaetes) in the Drosophila adult thorax (Simpson, 1990). Sensory organ precursors (SOPs) of macrochaetes emerge in groups of cells called proneural clusters (PNCs) at the third instar larval stage (Huang et al., 1991). Within PNC, the basic helix-loop-helix (bHLH) transcription factors of Achaete (Ac) and Scute (Sc) encoded by the achaete-scute complex (AS-C) (Ghysen and Dambly-Chaudiere, 1988; Garcia-Bellido and de Celis, 2009) are expressed and gradually accumulate in one of these cells (Cubas et al., 1991; Romani et al., 1989). Finally, the cell with the highest Ac-Sc expression is chosen to become a SOP by lateral inhibition mediated by Notch (N) signaling pathway (Simpson, 1990; Simpson et al., 1999). Each differentiated SOP of a bristle asymmetrically divides to produce a PIIa cell and a PIIb cell. The PIIa cell divides to give rise to a shaft cell and a socket cell (Hartenstein and Posakony, 1989), while the PIIb divides twice to produce a glial cell (Gho et al., 1999; Reddy and Rodrigues, 1999) which undergoes programmed cell death shortly after its birth (Fichelson and Gho, 2003), a sheath cell and a neuron (Hartenstein and Posakony, 1989). After differentiation, both the shaft cell and the socket cell undergo a few rounds of replication (Audibert et al., 2005; Hartenstein and Posakony, 1989). Finally, the shaft cell begins to elongate at the tip by the force of actin polymerization and finally to form a bristle (Lees and Picken, 1944; Tilney et al., 2000; Tilney and DeRosier, 2005).

Bristle development system has been described as an ideal model to study many biological events (Audibert et al., 2005; Furman and Bukharina, 2008; Tilney and DeRosier, 2005). Since the development of bristles is well-characterized in detail. Especially, the position of each macrochaete is fixed (Smith and Sondhi, 1961) and the timings of SOP differentiation and cell division of SOP daughter cells are well-characterized by many researchers (Hartenstein and Posakony, 1989; Huang et al., 1991; Gho et al., 1999; Audibert et al., 2005). Thus, these characteristics were here utilized in combination with the GAL4/UAS targeted expression system to enable us to examine in vivo functional roles of DREF in more detail.

In this study, we specifically reduced DREF expression in posterior scutellar (PSC) macrochaetes SOP lineages and surrounding cells and compared the phenotypes in PSC shaft cell nuclei to those of controls. Our experiments revealed that (1) A PSC bristle deprived of DREF become short and thin or deleted in severe cases; (2) The PSC SOP is correctly differentiated but the timing of asymmetric cell division is retarded without disrupting cell differentiation in most cases; (3) DREF plays critical roles in growth and endoreplication in the shaft cells and possibly in the socket cells; (4) Evident down-regulation of PCNA-GFP expression was observed in the endocycling shaft cell. Based on these observations and other supporting data, we conclude that DREF plays critical roles, especially in endoreplication processes of bristle development, at least in part by regulating pcna transcription.

Materials and Methods

Fly strains

The following fly stocks were used in this study. UAS-DREF-RNAi¹⁵ corresponds to pUAS-dref-IR (X) strain 15 described previously, which is the strongest RNAi line (Yoshida et al., 2004). UAS-DREF-RNAi¹⁰ corresponds to pUAS-dref-IR (X) strain 10 (moderate RNAi line) and UAS-DREF-RNAi³⁸ corresponds to pUAS-dref-IR (II) strain 38 (mild RNAi line) (Yoshida et al., 2004). Knockdown efficiency of these UAS-RNAi lines was evaluated by Western blot analysis and differences in adult wing and eye morphology induced by en-GAL4 and GMR-GAL4, respectively (Yoshida et al., 2004). UAS-DREF (II) (Hirose et al., 2001), UAS-dMyc (II) (a kind gift from Dr. Corinne Benassayag) (Zaffran et al.,1998), y hs-flp; Act>y>GAL4, UAS-GFP;+ (a kind gift from Dr. Adachi-yamada) (Yoshioka et al., 2008), yw hs-flp122; Act>CD2>GAL4, UAS-GFPnls (a kind gift from Dr. BA Edgar). UAS-RNAi against dMyc (2948), PCNA (108384) were from Vienna Drosophila RNAi Center. P{mus209-EmGFP}T13 (24909), Df(2R)173/SM5, y1 w^*; P{GawB}sca^109-68/CyO, y¹ w¹¹¹⁸; P{GawB}ap^md544/CyO, and other lines were from Bloomington Stock Center or Drosophila Genetic Resource Center.

Collection of pupae and dissection

White female pupae were incubated at 25°C or 28°C until particular collection times then placed on double-faced tape. After peeling off with a tweezer, the thoraxes were dissected out and incubated in Grace’s medium (SIGMA) with 5'-bromo-2'-deoxyuridine (BrdU) or fixed with 4% paraformaldehyde (PFA) for immunohistology.

BrdU incorporation

For BrdU incorporation experiments, dissected thoraxes were incubated with Grace’s medium containing 100 μM of BrdU (5'-Bromo-2'-deoxyuridine Labeling & Detection Kit, Roche) at 25°C for 30 minutes and then washed with PBS. Samples were then incubated with 2N HCl for 15 minutes and neutralized with 0.1 mol/l Na₂B₄O₇ for 5 minutes. After incubation, the samples were washed with PBS containing 0.3% Triton X-100 (PBS-T). For double-labeling experiments using anti-BrdU IgG and anti-LacZ IgG, samples were initially incubated with rabbit anti-lacZ IgG (Cappel, 1:100 dilution) for 20 hours. After washing with PBS-T, samples were incubated with mouse anti-BrdU IgG (5'-Bromo-2'-deoxyuridine Labeling & Detection Kit, Roche, 1:25 dilution) for 2 hours at 25°C After washing with PBS-T, the samples were incubated with the secondary antibody for 2 hours at 25°C.

Immunohistology

Dissected larvae or pupae were fixed in 4% PFA for 30 minutes. After washing with PBS-T, samples were incubated with the following primary antibodies for 20 hours at 4°C or 2 hours at 25°C: mouse anti-lacZ IgG (Developmental Studies Hybridoma Bank, DSHB, 1:500; rabbit anti-lacZ IgG (Cappel, 1:100); mouse anti-Fibrillarin IgG (EnCor Biotechnology, 38F, 1:100); mouse anti-prospero IgG (DSHB, 1:25); mouse anti-achaete IgG (DSHB, 1:100); rat anti-elav IgG (DSHB, 1:200); rabbit anti-GFP IgG (MBL, 1:100); mouse anti-GFP IgG (Invitrogen, 1:250); rabbit anti-DREF IgG (1:100) (Hirose et al., 1996); or rat anti-Su(H) IgG (a kind gift from F Schweisguth, 1:250) (Gho et al., 1996). After washing with PBS-T, samples were incubated with the following secondary antibodies for 2 hours at 25°C: anti-mouse IgG conjugated with Alexa 350; anti-mouse or rat IgG conjugated with Alexa 488; or anti-mouse or rabbit IgG conjugated with Alexa 594 (Molecular Probe, 1:500).

Measurement of bristle length and size of the nucleus in the socket cell and the shaft cell

After each thorax was dissected from the adult fly body and muscles and other attached tissues were removed, the preparations were treated with 2-propanol and put onto slide glasses. The samples were mounted with a small amount of Hoyer’s medium. Then, the samples were incubated at 65°C for 3 days. The length of particular bristles, the area of a socket cell nucleus and a shaft cell nucleus, and the largest area of a shaft cell nucleolus detected with mouse anti-Fibrillarin IgG were measured with Aquacosmos software (Hamamatsu photo).

Generation of somatic clones

For the flip out technique, somatic clones expressing RNAi against dref were generated by heat-shocking y hs-flp/y UAS-DREF-RNAi¹⁵; Act>y>GAL4, UAS-GFP/+; + larvae at early third instar larval stage (Fig. 1H) or yw hs-flp122/UAS-DREF-RNAi¹⁵; Act>CD2> GAL4, UAS-GFPnls/+; + larvae at the first to second instar larval stage (Fig. 1P–R), each for 1.5 hours at 37°C. After the heat-shock, the larvae were raised at 28°C until adult emergence (Fig. 1H) or wandering stage (Fig. 1P–R). The clones were identified by y/y bristle phenotype (Fig. 1H) or GFP (Fig. 1P–R).

View Details

Fig. 1. DREF is required for external sensory organ development in the adult thorax. (A–F) Scanning electron micrographs of an adult thorax. (A)+; sca-GAL4/UAS-lacZ-RNAi. (B) UAS-DREF-RNAi15/+; sca-GAL4/+. (C) UAS-DREF-RNAi15/UAS-DREF-RNAi15; sca-GAL4/+. (D) UAS-DREF-RNAI15/+; sca-GAL4/UAS-DREF. (E) +; ap-GAL4/UAS-lacZ-RNAi. (F) UAS-DREF-RNA10/+; ap-GAL4/+. The arrows in panels A to D indicate PSC macrochaetes. (G) Phenotypic analysis of dref-knockdown effects on macrochaetes. dref-knockdown adult flies (UAS-DREF-RNAi15/+; sca-GAL4/+ and UAS-DREF-RNAi15/UAS-DREF-RNAi15; sca-GAL4/+, N=28) were classified into four phenotypic classes according to the morphology of their PSC macrochaetes into: deletion (macrochaetes deletion); shaft and socket; shaft only; and socket only. The Y-axis indicates percentages of individuals with these phenotypes. (H) The picture shows a part of Stout mechanosensory bristles at the anterior wing margin (AWM). An asterisk indicates a dref knockdown y/y bristle. The others black ones are wild types. (I–L) DREF expression in SOPs of the ASC and the PSC and surrounding cells was reduced in scabrous-GAL4>UAS-DREF-RNAi15 lines (UAS-DREF-RNAi15/+; sca-GAL4/UAS-GFPnls). Larvae were raised at 28°C and third instar larval wing discs were stained with rabbit anti-DREF IgG (magenta for J and L) and mouse anti-GFP IgG (green for K and L). DAPI (cyan for I). (M–O) Expression patterns of scabrous in a wing disc at third instar larval stage (M–O). The images in (M–O) show that scabrous expresses within parts of the posterior scutellar (PSC) and anterior scutellar (ASC) PNC and their SOPs. Larvae or pupae were raised at 25°C and third instar larval wing discs were stained with mouse anti-achaete IgG (magenta for M and O) and rabbit anti-GFP IgG (green for N and O). An arrow indicates PSC SOP. An arrowhead indicates ASC SOP. (P–R) DREF is effectively reduced in somatic clones expressing dsRNAi for dref in the wing disc. After the heat shock described in Material and Methods, wandering larvae were collected and its wing discs were stained with both rabbit anti-DREF IgG (magenta) and mouse anti-GFP IgG. (green).

X-GAL staining of pharate adult thoraxes

X-Gal staining was performed as described previously (Uemura et al., 1993) with slight modifications. White female pupae were raised at 28°C until the pharate adult stage which corresponds to stage P10 to P11 (Bainbridge and Bownes, 1981). Thoraxes were dissected out and fixed with PBS containing 3.7% formaldehyde for 30 minutes. After washing the samples with PBS-T for several times, the samples were incubated in 37°C pre-warmed Fe/Na incubation buffer containing 3.1 mM K₃Fe(CN)₆, 3.1 mM K₄Fe(CN)₆, 28 mM NaH₂PO₄, 81 mM NaH₂PO₄, 150 mM NaCl, 1 mM MgCl₂, 0.2 % X-Gal and 0.03% Triton-X100 in the dark at 37°C for 2 to 12 hours. The microchaete cell lineages that were not affected by DREF knockdown were used as positive control for the staining.

Statistical analysis

All statistical analyses were performed using EXCEL Toukei Ver.6.0 (ESUMI). For comparison between two groups, the Welch t-test (Fig. 9G) or McNemar’s test (Fig. 9H) were employed. For comparison of more than two groups, the Sheff pos hoc test was used after ANOVA (Fig. 10O, Fig. 11I and J). Significance levels for each test were set at ^*P<0.05,^**P<0.01 and ^***P<0.001. All the data shown are mean±SEM values.

Results

DREF is required for external sensory organ development in the adult thorax

To understand in vivo roles of DREF at the cell biological level, we first confirmed the specificity of the antibody, utilizing flip out technique (Ito et al., 1997) to make somatic clones expressing dsRNAi for dref in the wing disc. The results showed that DREF is effectively reduced in GFP positive cells, as compared to control cells without GFP (Fig. 1P–R), indicating that the anti-DREF antibody is specific to DREF as noted previously (Hirose et al., 1996, 2001). Next, we confirmed specific reduction of DREF expression in the posterior scutellar (PSC) SOP lineage and surrounding cells using scabrous-GAL4 (sca-GAL4) and UAS-DREF-RNAi¹⁵ lines. The immunostaining data with anti-DREF antibody showed that DREF expression is effectively reduced in scabrous expressing cells that are marked with GFP (Fig. 1I–L). We also confirmed that the sca-GAL4 lines express trans-genes in the regions containing the PNCs and the PSC SOP (Fig. 1M–O) that are marked by Achaete accumulation at third instar larval stage (Skeath and Carroll, 1991; Cubas et al., 1991) indicating that DREF is indeed down-regulated in the PNCs and the PSC SOP of dref-knockdown flies (Fig. 1I–O).

Next, we examined the phenotype of flies carrying the sca-GAL4 and UAS-DREF-RNAi¹⁵. SEM analyses revealed that macrochaetes of flies carrying one copy of UAS-DREF-RNAi¹⁵ became thin and short as compared to control flies carrying sca-GAL4 and UAS-lacZ-RNAi (Fig. 1A and B). The data exclude the possibility that the observed phenotype with DREF knockdown is due to the unspecific effects caused by the RNAi machinery. Careful phenotypic analyses revealed that bristles in the posterior scutellar (PSC), anterior scutellar (ASC), posterior dorsocentral (PDC), anterior dorsocentral (ADC), posterior notopleural (PNP) and presutural (PS) areas were strongly affected (Fig. 2). On the other hand, bristles in posterior postalar (PPA), anterior supraalar (ASA), anterior notopleural (ANP) were less influenced (Fig. 2). Next, we utilized a GAL4 enhancer trap line of apterous (ap) to examine the effect of dref-knockdown on the microchaete SOP lineage. In this experiment, we used UAS-DREF-RNAi¹⁰ strain (mild allele) instead of UAS-DREF-RNAi¹⁵ strain (strong allele), since most of the ap-GAL4>UAS-DREF-RNAi¹⁵ flies became lethal. Our data revealed that almost all microchaetes became short and thin (Fig. 1E and F), suggesting that DREF is required for bristle development in the adult thorax. To determine the requirement of DREF itself for bristle development, DREF was over-expressed in sca-GAL4>UAS-DREF-RNAi¹⁵ lines. The data showed that dref-knockdown phenotypes were effectively rescued by over-expression of DREF (Fig. 1D). In flies carrying two copies of UAS-DREF-RNAi¹⁵, almost all macrochaetes were deleted (Fig. 1C and G), suggesting that DREF may play different roles in bristle development depending on the developmental context and dose of DREF. Finally, to determine the requirement of DREF for bristle development in other parts of fly bodies, the flip out technique was used to generate somatic clones. The data showed that stout mechanosensory bristles on the anterior wing margin (AWM) derived from dref-knockdown clones that were marked with yellow also became short and thin (Fig. 1H). Taken together, all these results indicate that dref is required for bristle development in the adult thorax.

View Details

Fig. 2. Effect of DREF knockdown in the thorax. (A) Illustration of 11 macrochaetes in half part of thorax. PSC: posterior scutellar, ASC: anterior scutellar, PDC: posterior dorsocentral, ADC: anterior dorsocentral, PPA: posterior postalar, PPA: anterior postalar, PSA: posterior supraalar, ASA: anterior supraalar,:PNP: posterior notopleural, ANP: anterior notopleural,PC: presutural. (B) Phenotypes of macrochaetes of 33 DREF knockdown flies (UAS-DREF-RNAi15/+; sca-GAL4/+) are classified into seven groups. Deletion (loss of a bristle), extremely short and thin bristle, short and thin bristle, extremely shirt bristle, short bristle, thin bristle, normal bristle. X-axis shows 11 macrochaetes. Y-axis shows percentages of individual with the phenotypes described above.

DREF is not required for SOP differentiation

Disruption of regulators for the ASC complex and/or N signaling results in extra SOPs formation or deletion of SOPs. (Held, 2002; Furman and Bukharina, 2008) Considering the observation that strong dref-knockdown in SOPs lineage caused deletion of bristles (Fig. 1G), we speculated that missing bristles may result from defects in proper SOP formation. To test the possibility, we crossed dref-knockdown lines with a lacZ enhancer trap line of neuralized, A101-lacZ (Huang et al., 1991) to visualize a SOP of each macrochaete. SOPs of both control flies and dref-knockdown flies were correctly formed (Fig. 3A–D), suggesting that defects in SOPs formation cannot explain the observed missing bristle phenotype. To exclude the possibility that the result occurred due to specificity of the sca-GAL4 expression pattern, we used another GAL4 driver; the ac-GAL4 line. We first confirmed that the ac-GAL4 lines express GFP-reporter genes within PNCs and SOPs of the PDC and the ADC (Fig. 4D–F). We also confirmed that DREF protein was expressed within the PNCs in control flies (Fig. 4G–I) and its level was effectively reduced in the PNCs and the SOPs of the ac-GAL4>UAS-DREF-RNAi¹⁵ flies (Fig. 4J–L). However, the PDC and the ADC bristles were apparently normal in the adult thoraxes (Fig. 4A and B) even under the conditions of extensive reduction of DREF within the PNCs, suggesting that DREF is not involved in SOP formation process within ac expressing cells. In addition, further experiments revealed that over-expression of DREF in sca-GAL4>UAS-DREF lines and ac-GAL4>UAS-DREF lines neither induce ectopic SOPs (data not shown) nor macrochaetes in the adult thorax (Fig. 4A, C and Fig. 5) as well as ectopic campaniform sensillum on the L3 wing vein (data not shown), confirming that DREF does not play decisive roles in macrochaetes SOPs formation processes. Taken together, all these results indicate that DREF is not required for SOPs differentiation at least within ac and sca expressing cells.

View Details

Fig. 3. Analyses of roles of DREF in cell differentiation during SOP formation process (A–D) and asymmetric cell division (E–R). (A and C) +; scabrous-GAL4/+; A101-lacZ/+. (B, D, E–R) UAS-DREF-RNAi15/UAS-DREF-RNAi15; sca-GAL4/+; A101-lacZ/+. (A–D) DREF is not required for SOP differentiation. Larvae were raised at 28°C and third instar larval wing discs at 0 hour APF were collected and stained with rabbit anti-lacZ IgG (green). An arrow indicates a SOP of an ASC macrochaete. An arrowhead indicates a SOP of a PSC macrochaete. (A and B) Normarski images of the third instar larval wing discs. (E-R) DREF plays no direct role in cell differentiation during asymmetric cell division. Larvae were raised at 28°C and pupae at 21 to 24 hours APF were collected and stained with rabbit anti-lacZ IgG (green) (E–J, Q and R), mouse anti-lacZ IgG (green) (K–P, Q and R), rat anti-elav IgG (magenta) (E–G, N–P, Q and R), mouse anti-prospero IgG (cyan) (E–G, Q and R) and rat anti-Su(H) antibodies (green) (H–M, Q and R). An arrow indicates a PSC shaft cell nucleus. An arrowhead indicates a PSC socket cell nucleus. (Q–R) Statistical analysis of the thoraxes of dref-knockdown flies stained with specific combinations of antibodies. Double-immunostaining was employed to obtain accurate semi-quantitative data. The thoraxes were stained with mouse anti-lacZ IgG and rat anti-elav IgG or mouse anti-lacZ IgG and rat anti-Su(H) IgG or rabbit anti-lacZ IgG and mouse anti-prospero IgG. (Q) The number of lacZ positive cells was counted and were classified into four phenotypic classes according to the sizes and the number of the lacZ positive cells and types of differentiated cells. Y-axis indicates a percentage of the lacZ positive cells. (K–P) The images show examples of abnormal cell differentiation caused by dref-knockdown. (K–M) One lacZ and Su (H) positive cell. (N–P) The two lacZ positive cells, which consist of a Su (H) positive larger one and a Elav positive smaller one. Table (R) shows raw data used to make Q graph. The number indicates a percentage of the lacZ positive cells. The number in parenthesis indicates sample number.

View Details

Fig. 4. DREF plays no apparent role in SOP differentiation within ac expressing cells. (A–C) Scanning electron micrographs of an adult thorax. (A) y1 ac1 w1118/+;UAS-lacZ-RNAi/+; ac-GAL4/+. (B) y1 ac1 w1118/UAS-DREF-RNAi15; UAS-DREF-RNAi38/+; ac-GAL4/+. (C) y1 ac1 w1118/+; UAS-DREF/+; ac-GAL4/+. All larvae were raised at 28°C. An arrow and an arrowhead indicate a PSC bristle and a PDC bristle, respectively. (D–L) Expression levels of DREF are reduced in cells within PNCs and SOPs of ac-GAL4>UAS-DREF-RNAi lines. (D–I) y1 ac1 w1118/+; UAS-GFPnls/+; ac-GAL4/+. (J–L) y1 ac1 w1118/UAS-DREF-RNAi15; UAS-GFPnls/+; ac-GAL4/+. Larvae were raised at 25°C (D–F) or 28°C (G–L) and third instar larval wing discs at wandering stage were collected and stained with rabbit anti-GFP IgG (green) and mouse anti-achaete IgG (magenta) (D–F) or mouse anti-GFP IgG (green) and rabbit anti-DREF IgG (magenta) (G–L). (D–F) ac-GAL4 driver expresses GAL4 within PNCs and SOPs of a PDC and a ADC at third instar larval stage. An arrowhead indicates a SOP of PDC. Arrows indicates two candidates with high accumulation of AC either of which is destined to become ADC SOP. (G–I) Expression levels of DREF are almost ubiquitous in cells within ac expressing cells of control lines. (J–L) Expression levels of DREF are reduced in cells within ac expressing cells of ac-GAL4>UAS-DREF-RNAi lines.

View Details

Fig. 5. Extopic expression of DREF in macrochaete cell lineages did not induce extra macrochaetes in the adult thoraxes. (A–C) Scanning electron micrographs of adult thoraxes. (A) +; sca-GAL4/+. (B) +; sca-GAL4/UAS-EGFR.1A887T12-4. (C) +; sca-GAL4/UAS-DREF. All crossings were raised at 28°C. Arrowheads indicate examples of extra macrochaetes

DREF regulates the timings of asymmetric cell division but perhaps plays no direct role in cell differentiation during asymmetric cell division

A number of genes are required for asymmetric cell division (Abdelilah-Seyfried et al., 2000) and their defects cause a disorder of cell differentiation termed cellular transformation. We therefore considered the possibility that missing bristles caused by strong dref-knockdown might result from defects in cell differentiation. To test the possibility, we monitored the PSC SOP lineage by immunostaining with anti-Su(H) IgG (a marker for socket cells) (Gho et al., 1996), anti-prospero IgG (marker for sheath cells) (Manning and Doe, 1999) and anti-elav IgG (a marker for neurons) (Bier et al., 1988) together with anti-lacZ IgG (a marker for the SOP lineage in the A101-lacZ line). The data revealed that in 87 % of the cases, four lacZ positive cells were detected in the dref-knockdown SOP lineage (Fig. 3Q and R). Among these four lacZ positive cells, the two smaller cells were a sheath cell and a neuron (Fig. 3E–G), and we observed only a single Su(H) positive cell among the other two larger lacZ positive cells (Fig. 3H–J). In addition, a tiny shaft cell and a socket cell in the adult thorax were formed in PDC macrochaete of dref-knockdown lines (Fig. 6). The data suggest that in most cases, cell differentiation correctly occurred even under the conditions of severe reduction of dref expression. However, in remaining 13% of the cases, the number of lacZ positive cells became one or two or three (Fig. 3Q and R). Among these one or two or three lacZ positive cells, the single lacZ positive cell was always Su(H) positive (Fig. 3K–M). The two lacZ positive cells consist of one larger cell that is Su(H) positive and one smaller cell that is Elav positive (Fig. 3N–P). The three lacZ positive cells consist of one larger cell that is Su(H) positive and two smaller cells, one of them is either Elav positive or Prospero positive (data not shown), suggesting that failures in cell differentiation occurred in the PSC SOP linage of dref-knockdown flies. Considering the observation that total cellular transformation such as shaft to socket, neuron to sheath and vice versa were never observed in this case, the data suggest that the dref-knockdown affected cell differentiation indirectly via other mechanisms.

View Details

Fig. 6. Comparison of bristle phenotypes between wild type and dref-knockdown line at high magnification. (A–B) Scanning electron micrographs of adult thoraxes of (A) +/+ (Canton S) and (B) UAS-DREF-RNAi15/UAS-DREF-RNAi15; sca-GAL4/+ at 1000-fold magnification. A white asterisk indicates a socket cell in PDC macrochaetes. A white arrow indicates a shaft cell in a PDC macrochaete. A black asterisk indicates an example of a socket cell in a microchaete. A black arrow indicates an example of a shaft cell in a microchaete.

To obtain more insights into the other possibility, we focused on cell division defects, since previous studies reported that DREF is required for efficient proliferation in vivo in the wing discs (Hyun et al., 2005) and inactivation of cdc2 delayed division and produces similar abnormal sets of SOP daughter cells in microchaetes lineages (Fichelson and Gho., 2004). To test the possibility, we examined the timing of asymmetric cell division in PSC SOP daughter cells of dref-knockdown flies at 0 hour APF and 3 hours APF and compared them to those of control flies. At 0 hour APF, two cells consisting of a PIIa cell in the posterior side and a PIIb cell in the anterior side were observed in 90% of the cells in control flies (Fig. 7A and E), However, in dref-knockdwon flies, only 38% of the PSC SOPs completed the first asymmetric cell division and the other cells remained in the SOP state (Fig. 7C and E), suggesting that dref-knockdown delayed the onset timing of asymmetric cell division. At 3 hours APF in control, 80% of the cells were 4 or 5 cell state, in which at least the PIIa division was completed (Fig. 7B and E). However, at the same time in dref-knockdown flies, 52% of the cells were 4 or 5 cell state and most of the other cells remained in 3 cell state that possibly consists of a PIIIb, a PIIb and a glial cell as previously described in microchaete lineage (Fig. 7D and E), suggesting that dref-knockdown delayed the timing of last two round of asymmetric cell division. Taken together, all these results suggest that DREF does not play a direct role in cellular differentiation process during asymmetric cell division. DREF rather regulates the timing of asymmetric cell division that may have resulted in the rarely observed apparent defects in cell differentiation during asymmetric cell division.

View Details

Fig. 7. DREF regulates the timing of asymmetric cell division. (A, C and E) +; sca-GAL4/+; A101-lacZ/+. (B, D and E) UAS-DREF-RNAi15/UAS-DREF-RNAi15; sca-GAL4/+; A101-lacZ/+. Larvae and pupae were raised at 25°C until 0 hour or 3 hours APF and their wing discs were stained with mouse anti-lacZ IgG (green). Types of cells were identified by differences in nuclear size. The position in the notum and the number of lacZ positive cells according to the previous report (Gho et al., 1999). An arrow in (A–D) indicates a ASC SOP lineage. Arrowheads in (A–D) indicate PSC SOP daughter cells. (E) DREF regulates timing of the onset and last two round (0 hour APF and 3 hours APF, respectively) of asymmetric cell division. The number of cells detected with anti-lacZ was counted and were classified into four groups. One cell (a SOP), two cells (a PIIa and a PIIb), three cells (a PIIIb, a PIIb and a glial cell [g]), four cells or (a socket cell [so], a shaft cell [sf] plus a pIIIb cell and a glial cell or a sheath cell [s] and a neuron [n]) and five cells (a socket cell and a shaft cell plus a sheath cell, a neuron and a glial cell).

PSC bristle cell lineages still exist underneath the notum of adult flies deprived of DREF

Our studies so far suggest that the deleted PSC bristles deprived of DREF (Fig. 1C, G) did not result from cell differentiation defects during either SOP formation (Fig. 3A–D) or asymmetric cell division (Fig. 3Q and R). The data raise a question of whether the SOP cell lineage died in the course of development. To clarify the question, X-GAL staining of pharate adult flies was carried out. Our observations revealed that in about 60% of the cases, light and smaller but apparent lacZ signals were present in the PSC regions of the notum of dref knockdown flies (Fig. 8B and C). and brighter and bigger lacZ signals were present in that of control (Fig. 8A and C). The data suggest the possibility that the PSC bristle cell lineages exist underneath the notum of adult flies deprived of DREF.

View Details

Fig. 8. PSC bristle cell lineages still exist underneath the notum of adult flies deprived of DREF. (A and C) +; sca-GAL4/+; A101-lacZ. (B and C) UAS-DREF-RNAi15/UAS-DREF-RNAi15; sca-GAL4/+; A101-lacZ/+. Pupae raised at 28°C were collected and X-GAL staining was performed as described in Material and Methods. An arrow indicates lacZ signal which corresponds to a ASC bristle cell lineage, and an arrowhead indicates lacZ signal which corresponds to a PSC bristle cell lineage. (C) Quantitative analysis of the X-GAL staining. X-axis indicates percentage of lacZ-positive cells in PSC SOP lineage detected by the staining.

DREF is critical for growth and endoreplication in shaft cells

Cell size for a given cell type is generally proportional to the amount of nuclear DNA, endoreplication in which cell cycle lack all visible aspects of mitosis constitutes an effective strategy of cell growth (Edgar and Orr-Weaver, 2001). Shaft cells and socket cells in the microchaete cell lineage undergo endoreplication a few times after the end of asymmetric cell division (Audibert et al., 2005). Previous studies reported that DREF is required for endoreplication in salivary glands (Hirose et al., 2001). Considering the observation that macrochaetes as well as microchaetes became smaller on reduction of DREF expression levels (Fig. 1B, C and F) and most of the SOP daughter cells correctly differentiated (Fig. 3E–J, Q and R), we considered the possibility that missing bristles as well as the smaller bristle phenotype result from DNA endoreplication defects after asymmetric cell division. To test the possibility, we measured nuclear sizes of the PSC shaft cell and the PSC socket cell of control flies together with those of dref-knockdown flies. The sizes became smaller than those of flies only expressing GAL4 (Fig. 9A and D). At 30 hours APF, half reduction was noted for the shaft cell whereas the socket cell nuclei of dref-knockdown flies became 70% of the size of their counterparts in control flies (Fig. 9H).

View Details

Fig. 9. DREF is critical for endoreplication in shaft cells and possibly in socket cells. (A, B, C, G and H) +; sca-GAL4/+; A101-lacZ/+. (D, E, F, G and H) UAS-DREF-RNAi15/UAS-DREF-RNAi15; sca-GAL4/+; A101-lacZ/+ (A–G). BrdU assay of PSC shaft cell nuclei. Pupae were raised at 25°C until 28 hours APF and their thoraxes were stained with both rabbit anti-lacZ IgG (green) and mouse anti-BrdU IgG (magenta). Numbers of BrdU positive signals in the PSC shaft cell nucleus of each individual of dref-knockdown flies were counted and compared to those of controls. Statistical analysis (McNemar’s test **P<0.01) was conducted between BrdU positive ratio of shaft cell nuclei in control flies and that of shaft cell nuclei in dref knockdown flies. An arrowhead indicates a nucleus of a PSC socket cell and an arrow indicates a nucleus of a PSC shaft cell. The nucleus of the PSC shaft cell and the socket cell was identified by differences in nuclear size and/or the position in the notum. The bigger one corresponds to a shaft cell nucleus and the socket cell is located posterior to the shaft cell (Gho et al., 1999). (H) Dynamic change in growth of a PSC shaft cell and a socket cell nucleus of both control and DREF knockdown flies. Pupae were raised at 25°C until 18, 21, 24, 27 and 30 hours APF and their thoraxes were stained with rabbit anti-lacZ IgG and measured the area of a nucleus of the PSC shaft cell and that of the socket cell. Blue lines indicate control flies. Red lines indicate dref-knockdown flies. Solid lines indicate shaft cell nuclei. Doted-lines indicate socket cell nuclei. Statistical analysis (Welch t-test, **P<0.01, ***P<0.001) was conducted between area of shaft cell nuclei or socket cell nuclei in control flies (n=25, 20, 28, 31, 28 corresponds to sample numbers at 18, 21, 24, 27, 30 hours APF) and that of shaft cell nuclei or socket cell nuclei in dref knockdown flies (n=26, 16, 28, 18, 18 corresponds to sample numbers at 18, 21, 24, 27, 30 hours APF) at each stage of development.

To obtain more insights into whether effects on nuclei size result from defects in endoreplication, we next carried out 5'-bromo-2'-deoxyuridine (BrdU) incorporation experiments and counted numbers of BrdU-positive PSC shaft cell nuclei at 28 hours APF (Fig. 9A–G). Rates for BrdU-positive nuclei in the PSC shaft cell of dref-knockdown flies were lower than those of control flies (Fig. 9G). In addition the data suggest that rates of BrdU incorporation in the PSC socket cell is lower than those of control (data not shown). These results support the idea that DREF is critical for endoreplication in shaft cells and possibly in socket cells.

DREF regulates pcna expression in endocycling PSC shaft cells

It has been suggested that DREF regulates many cell cycle and DNA replication-related genes (Matsukage et al., 2008) and our previous studies suggested that pcna is one of the DREF target genes in vitro and in vivo (Hirose et al., 1993; Ida et al., 2007; Seto et al., 2006). To examine whether DREF regulates transcriptions of the pcna gene in bristle development, we used PCNA-EmGFP lines (Thacker et al., 2003) which expresses GFP under the control of 100-bp enhancer element containing the two E2F binding sites as well as DRE site. As reported previously, the 100-bp enhancer element accurately reproduce S phase-associated, E2F-dependent pcna expression at many stages of Drosophila development (Thacker et al., 2003). Our observations revealed that PCNA-EmGFP expression was decreased in the endocycling PSC shaft cell of dref-knockdown flies (Fig. 10A–F), suggesting that DREF regulates pcna expression in endocycling shaft cells. To support the observation, we checked genetic interaction between dref and pcna by crossing sca-GAL4>UAS-DREF-RNAi¹⁵ lines with a line deficient for pcna (Df(2R)173). The resulting PSC bristles in UAS-DREF-RNAi¹⁵/+; sca-GAL4/Df(2R)173 lines became shorter than those in +; sca-GAL4/Df(2R)173. (Fig. 10I-O), suggesting that a genetic interaction between dref and pcna exists in the shaft cell. Finally, we examined the requirement of the pcna gene itself for bristle development by crossing sca-GAL4 lines with UAS-PCNA-RNAi lines. We observed that PSC bristles of pcna knockdown flies were deleted and other macrochaetes became as short and thin as those of dref-knockdown flies (Fig. 10G and H). All these data are compatible with DREF as an upstream regulator of the pcna gene in shaft cells.

View Details

Fig. 10. DREF regulates pcna expression in the endocycling PSC shaft cell. (A–C) +; sca-GAL4/+; PCNA-EmGFP/A101-lacZ. (D–F) UAS-DREF-RNAi15/UAS-DREF-RNAi15; sca-GAL4/+; PCNA-EmGFP/A101-lacZ. Pupae were raised at 28°C until 21 h APF and their thoraxes were stained with rabbit anti-GFP IgG (green) and DAPI (magenta). The nucleus of the PSC shaft cell and the socket cell was identified by differences in nuclear size and/or the position in the notum. The bigger one corresponds to a shaft cell nucleus and the socket cell is located posterior to the shaft cell (Gho et al., 1999). The immunostaining experiments in two different lines were carried out under the same condition in parallel experiments and the images (B and E) were taken under the condition of the some exposure time. An arrow indicates expression of PCNA-GFP in cytoplasm of the PSC shaft cell. An arrowhead indicates a nucleus of the PSC shaft cell (G–H) Scanning electron micrographs of the adult thorax of a dref-knockdown fly (G) (UAS-DREF-RNAi15/UAS-DREF-RNAi15; sca-GAL4/+) and a pcna-knockdown fly (H) (+; sca-GAL4/UAS-PCNA-RNAi. (I–N) Scaning electron micrographs of an adult notum. (I) +; sca-GAL4/+. (J) +; sca-GAL4/Df(2R)173. (K) UAS-DREF-RNAi15/+; sca-GAL4/+. (L) UAS-DREF-RNAi15/+; sca-GAL4/Df(2R)173. (M) UAS-DREF-RNAi15/UAS-DREF-RNAi15; sca-GAL4/+. (N) +; sca-GAL4/UAS-PCNA-RNAi. An arrowhead indicates a PSC macrochaete. (I, J, K, and L) half dose reduction of pcna enhanced short and thin PSC bristle phenotypes caused by dref-kcnokdown. (O) Quantitative analyses of the length of the PSC bristle of flies with different genotypes. Statistical analysis (sheff-test,**P<0.01) was conducted among groups. Significant differences are indicated by letters above bars.

DREF is necessary but not sufficient for nuclear growth and protein synthesis in shaft cells

Our results showed that DREF is required for growth of shaft cell nuclei and socket cell nuclei (Fig. 9). It is reported that DREF over-expression is sufficient to promote tissue growth during larval development (Hyun et al., 2005). Our previous studies reported that DREF is involved in regulation of protein synthesis by regulating eIF4A (Ida et al., 2007). Thus, it can be considered that DREF may play a role in promoting growth of shaft cell nuclei by regulating protein synthesis. To test the possibility, first, DREF or growth regulator dMyc (Schreiber-Agus et al., 1997) was over-expressed or down-regulated using sca-GAL4 and sizes of PSC shaft cell nuclei were compared with those of controls. The nuclear size of the dMyc over-expressed shaft cell became larger (Fig. 11E, G and I) and that of the dMyc-knockdown shaft cells became smaller than those of control (Fig. 11D, E and I). In contrast, while the nuclear size of the DREF over-expressed shaft cells did not change (Fig. 11E, F and I), with knockdown it became smaller (Fig. 11E, H and I), suggesting that DREF is necessary but not sufficient for nuclear growth. Next, to determine involvement of DREF in regulation of protein synthesis in shaft cells, we monitored the sizes of nucleoli marked with anti-fibrillarin IgG (Fig. 11A, B and C) (Aris and Blobel, 1988). In dMyc over-expressing shaft cell nuclei, the nucleolar size became larger than that of control (Fig. 11E, G and J). While no such change was evident with DREF over-expression (Fig. 11E, F and J), the nucleolar sizes of DREF or dMyc down-regulated shaft cells were reduced (Fig. 11D, E, H, J). Taken together, all these results support the idea that DREF is necessary but not sufficient for nuclear growth and protein synthesis in shaft cells.

View Details

Fig. 11. DREF is necessary but not sufficient for nuclear growth and protein synthesis in the PSC shaft cell. (A–C) In these experiments Fibrillarin was used as a maker for nucleoli in the PSC shaft cell. Pupae were raised at 28°C until 21 hours APF and their thoraxes were stained with mouse anti-Fibrillarin IgG (green) and DAPI (magenta). (A–C) Canton S. (D–J) Over-expression of DREF is not sufficient enough to promote growth of the nucleolus and the nucleus in the PSC shaft cell. Pupae were raised at 28°C until 21 hours APF and their thoraxes were stained with mouse anti-Fibrillarin IgG (not shown) and DAPI (cyan). (D) +; sca-GAL4/UAS-dMyc-RNAi. (E) +; sca-GAL4/+. (F) +; sca-GAL4/UAS-DREF. (G) +; sca-GAL4/UAS-dMyc. (H) UAS-DREF-RNAi15/UAS-DREF-RNAi15; sca-GAL4/+. (A–H) An arrowhead shows a nuclear of a PSC shaft cell. An arrow shows nucleoli of a PSC shaft cell. (I) Comparison of the area of nucleus in a PSC shaft cell among flies shown in panels D–H. The Y-axis indicates the area of nucleus in a PSC shaft cell. (J) Comparison of the area of nucleolus in a PSC shaft cell among flies shown in panels D–H. The Y-axis indicates the area of nucleolus in a PSC shaft cell. (I–J) Statistical analyses (sheff-test,**P<0.01) were conducted among groups. Significant differences are indicated by letters above bars.

Discussion

DREF plays critical roles in cell division process, especially in endoreplication process of bristle development, at least in part by regulating the pcna gene expressions

It has been suggested that DREF is an important transcription factor required for efficient proliferation in vitro and in vivo. However, the relative lack of cell biological studies have hampered progression of in vivo characterization of DREF. The present study characterized in vivo roles of DREF utilizing the cell lineage of macrochaete development system. Our detailed analyses revealed some new characteristics of DREF in vivo: (1) DREF is critical for mehcanosesory bristle development in the adult thorax (Fig. 1 and Fig. 2). (2) DREF plays no apparent role in SOP formation process (Fig. 3, Fig. 4 and Fig. 5). (3) DREF regulates the timing of asymmetric cell division but perhaps it plays no direct role in cell differentiation process during asymmetric cell division (Fig. 3, Fig. 6 and Fig. 7). (4) Most importantly, DREF affects endoreplication process by regulating a candidate target gene, pcna, in shaft cells (Fig. 9 and Fig. 10). So that target genes are specifically influenced depending on the developmental context. (5) DREF is necessary for but does not promote growth and protein synthesis in shaft cells (Fig. 11). (6) PSC bristle cell lineages deprived of DREF still exist underneath the notum of adult flies (Fig. 8). Based on these data, we therefore conclude that DREF plays critical roles in cell division, especially in endoreplication process of bristle development, at least in part by regulating the pcna gene expressions (Fig. 12).

View Details

Fig. 12. Models of DREF function during macrochaete development. (A) DREF is critical for macrochaete development by regulating endoreplication in shaft cells. Shaft cell without DREF cannot grow efficiently due to retarded replication caused by reduction of transcriptions of replication-related genes such as pcna. Therefore the shaft cell may not be able to accumulate protein needed for proper elongation of the shaft. so; socket cell, sf; shaft cell, n; neuron, s; sheath cell. (B) Possible target genes of DREF during macrochaete development. Judging from the series of observations of the phenotypes caused by dref-knockdown, we conclude that DREF does not play a role in regulating target genes required for SOP differentiation and possibly cell differentiation during asymmetric cell division either. DREF may rather regulate target genes required for normal asymmetric division. DREF also plays critical roles in growth and endoreplication probably by regulating target genes related to replication, growth and protein synthesis during shaft cell development. The size of arrows indicates the strength of transcriptional activity by DREF.

Regulation of cell differentiation by DREF during bristle development

DREF regulates a wide variety of genes with multiple functions, including a cell differentiation related gene raf (Yoshida et al., 2004; Matsukage et al., 2008) which encodes a protein kinase that binds to RAS-GDP and transmit signals within canonical receptor tyrosine kinase (RT9K)–Ras-mitogen-Activated protein kinase (MAPK)-extracellular signal-regulated kinase (ERK) pathways (RTK-Ras-ERK pathway) (Wassarman et al., 1995; Sundaram, 2005). Our previous study revealed that DREF thereby contributes to wing vein development (Yoshida et al., 2004). However, it is reported that DREF is not required for normal photoreceptor differentiation (Hyun et al., 2005). The present study provided lines of evidence indicating that DREF plays no apparent role in regulating SOP differentiation at least within ac and sca expressing cells (Fig. 3A–D and Fig. 4). In addition, our immuno-staining analysis detected apparent DREF signals in PNCs and a SOP of PDC and ADC (Fig. 4D-I) and other SOPs (data not shown). The data lead to the suggestion that DREF regulates other candidate genes, but not differentiation-related genes during SOP formation processes.

During asymmetric cell division, cell differentiation was affected in 15% of dref-knockdown PSC SOPs (Fig. 3Q and R). As consequences, one to three abnormal SOP daughter cells, which did not undergo cellular transformation were produced (Fig. 3K–R). The previous studies reported that inactivation of cdc2, which is required during neuroblast mitosis for asymmetric protein localization (Tio et al., 2001; Chia et al., 2008) caused similar abnormal SOPs daughter cells and leads to mother to daughter cell transformation in microchaetes SOP lineages (Fichelson and Gho., 2004), Therefore, the apparent failures in cell differentiation may have occurred indirectly via defects in mitosis and/or S phase transition during asymmetric cell division. However, we cannot exclude the other possibility that DREF plays a direct role in cell differentiation by regulating genes required for asymmetric protein localization. Since the appearance ratio of the abnormal cells was quite low, perhaps due to knockdown efficiency of UAS lines and/or the strength of sca-GAL4 driver, we could not thoroughly conduct detailed analysis of distributions of those proteins in dref-knockdown PSC SOP lineage. Further experiments, such as clonal analysis utilizing loss of function allele of dref will be required to clarify the issues in the future study, although loss of function allele of dref is not available at the moment.

Regulation of cell division by DREF during bristle development

Our observations showed that dref-knockdown delayed the timing of the mitotic cycles in PSC SOP lineage (Fig. 7), suggesting the possibility that DREF regulates transcription of target genes required for normal asymmetric cell division (Fig. 12). One of the candidate genes could be Cyclin E (CycE) of which expression proceeds SOP division and is required for transition to the S-phase after SOP mitosis in the microchaete bristle lineage (Audibert et al., 2005) and a previous study reported that DREF is required for transcription of many proliferation genes including cycE in the wing discs (Hyun et al., 2005). A previous study also reported that misexpression of Dacapo, which is a major inhibiter for CycE produced loss of external cells of scutellar and dorsocentral macrochaetae in the adult thorax which were detected by X-GAL staining (Abdelilah-Seyfried et al., 2000). However our results revealed that almost all the SOP lineages were correctly differentiated (Fig. 3Q and R) and in about 60% of the cases (Fig. 8C), PSC bristle cell lineages deprived of DREF still exist underneath the notum of adult flies (Fig. 8), suggesting that CycE activity which is possibly required for normal cell differentiation during SOP formation and/or asymmetric cell division may not be mediated by DRE/DREF pathway. Therefore DREF may regulate the timing of asymmetric cell division by regulating other proliferation and replication-related genes required for S-phase transition.

Regulation of replication and cell growth by DREF during bristle development

It is known that the size of a cell is proportional to its ploidy (Nurse, 1985; Conlon and Raff, 1999). Many cells in a Drosophila larva use an exaggerated form of the same strategy to grow very large, going through repetitive rounds of DNA synthesis without intervening mitoses (Conlon and Raff, 1999). This ‘endocycling’ is seen in many tissues/cells in Drosophila (Edgar and Orr-Weaver, 2001), including shaft cells and socket cells of microchaetes and macrochaetes lineages (Hartenstein and Posakony, 1989).

The present study demonstrated that DREF is required for growth and endoreplication in shaft cells of mechano-sensory bristles (Fig. 9). Our previous studies showed that DREF is required not only for growth and endoreplication in salivary glands but also for expression of several replication related genes (Hirose et al., 1999), suggesting that roles of DREF are conserved in various endocycling tissues/cells in Drosophila. Hyun et al. (2005) showed that cells lacking DREF activity accumulate in G1 and continue to grow to larger than their counterparts in wild-type wing discs (Hyun et al., 2005). Our additional data showed that unlike dMyc, over-expression of DREF using GMR-GAL4 could not promote growth of a facet in adult compounds eyes (data not shown) suggesting that DREF plays a less important role in the mitotic cycle in terms of cell growth. Similarly, we revealed that unlike dMyc, over-expression of DREF itself here did not promote growth of shaft cell nuclei (Fig. 11F and J), Thus, we concluded that DREF is not sufficient for promotion of cell growth in both mitotic and endocycling tissues/cells.

Regulation of protein synthesis by DREF during bristle development

The production of each bristle require a very high rate of protein synthesis in a single cell during a short developmental period (Marygold et al., 2007). When mitosis ends, the shaft cell is packed with rough endoplasmic reticulum as protein synthesis starts for shaft elongation and cuticle deposition (Held, 2002). It is well known that a half dose reduction of many ribosomal proteins genes causes a haplo-insufficient minute phenotype characterized by short and slender bristles (Lambertsson, 1998). It is reported that hDREF regulates some ribosomal protein genes in HeLa cells (Yamashita et al., 2007). Also, in the Drosophila melanogastor genome, DRE sequences are found in promoter regions of many ribosomal protein genes (Yamashita et al., 2007) and our studies suggest that DREF is necessary for protein synthesis in shaft cells (Fig. 11H and I). However, our additional experiments showed that over-expression of DREF could not rescue these phenotypes (data not shown), suggesting that rp genes are not likely to be target genes of DREF in shaft cells.

Advantages of machrochaete development system to study mechanism of replication and growth in vivo

Drosophila are thought as useful organisms to investigate the regulation and mechanism of polytenization (Andrew et al., 2000). Salivary glands or follicle cells in the adult ovary are described as the two major targets most appropriate for that purpose. In the case of salivary glands, a salivary cell ceases dividing once specification ends (Bradley et al., 2001). Throughout the larval stage, it undergoes a number of endoreplications, finally reaching 1024 to 2048C values (Rudkin, 1972). Thus, it’s simplicity has advantages for analyses. In the case of the ovary system, follicle cells in adults offer an excellent system in which to study the mitotic-to-endocycle transition (Calvi and Spradling, 1999; Deng et al., 2001) and have also been used to study mechanisms of gene amplification during choriongenesis (Calvi and Spradling, 1999). In the present study, macrochaete bristle development system proved an advantageous model to study cell growth and replication. Unlike salivary glands and ovary system, effects of reduction of particular gene expression on replication could be easily quantitated with change in length of adult bristles as an index of genetic interaction between particular genes. In this way, it could be applied to genome wide screening of replication-related genes. In addition, cell lineage analysis can be utilized to examine in vivo functions of particular genes in more detail. Finally, since the position of a particular macrochaete is fixed, time-lapse live imaging can be readily applied to monitor expression patterns of replication related-proteins limited to a single cell during development. For these reasons, in the future, this bristle developmental system could help us approach to mechanisms of replication and growth from different points of view.

Acknowledgments

We thank B.A. Edgar, R.N. Eisenman, C. Benassayag and F. Schweisguth separately for antibodies and fly stocks, Adachi-Yamada for fly stocks, J.P. Aris, F. Hirose and H. Yoshida for technical advice, M. Moore for advice with English usage. This study was supported by Grants-in-Aid from JST, JSPS and the Ministry of Education, Culture, Sports, Science and Technology of Japan.