help button home button Genetics Mol Cellular Biology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Genetics, Vol. 168, 199-214, September 2004, Copyright © 2004
doi:10.1534/genetics.104.029439

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Zraly, C. B.
Right arrow Articles by Dingwall, A. K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Zraly, C. B.
Right arrow Articles by Dingwall, A. K.

SNR1 (INI1/SNF5) Mediates Important Cell Growth Functions of the Drosophila Brahma (SWI/SNF) Chromatin Remodeling Complex

Claudia B. Zraly*, Daniel R. Marenda{dagger},1 and Andrew K. Dingwall*,{ddagger},2

* Oncology Institute, Cardinal Bernardin Cancer Center, Stritch School of Medicine, Loyola University of Chicago, Maywood, Illinois 60153
{ddagger} Department of Pathology, Cardinal Bernardin Cancer Center, Stritch School of Medicine, Loyola University of Chicago, Maywood, Illinois 60153
{dagger} Department of Biology, Syracuse University, Syracuse, New York 13244-1270

2 Corresponding author: Oncology Institute and Department of Pathology, Cardinal Bernardin Cancer Center, Room 334, Loyola University Medical Center, 2160 S. First Ave., Maywood, IL 60153.
E-mail: adingwall{at}lumc.edu

Manuscript received March 29, 2004. Accepted for publication May 18, 2004.


   ABSTRACT
TOP
 ABSTRACT
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 LITERATURE CITED
 
SNR1 is an essential subunit of the Drosophila Brahma (Brm) ATP-dependent chromatin remodeling complex, with counterparts in yeast (SNF5) and mammals (INI1). Increased cell growth and wing patterning defects are associated with a conditional snr1 mutant, while loss of INI1 function is directly linked with aggressive cancers, suggesting important roles in development and growth control. The Brm complex is known to function during G1 phase, where it appears to assist in restricting entry into S phase. In Drosophila, the activity of DmcycE/CDK2 is rate limiting for entry into S phase and we previously found that the Brm complex can suppress a reduced growth phenotype associated with a hypomorphic DmcycE mutant. Our results reveal that SNR1 helps mediate associations between the Brm complex and DmcycE/CDK2 both in vitro and in vivo. Further, disrupting snr1 function suppressed DmcycEJP phenotypes, and increased cell growth defects associated with the conditional snr1E1 mutant were suppressed by reducing DmcycE levels. While the snr1E1-dependent increased cell growth did not appear to be directly associated with altered expression of G1 or G2 cyclins, transcription of the G2-M regulator string/cdc25 was reduced. Thus, in addition to important functions of the Brm complex in G1-S control, the complex also appears to be important for transcription of genes required for cell cycle progression.

CHROMATIN modification by ATP-dependent multiprotein complexes is important for developmental regulation of gene expression and cell cycle control. The SWI/SNF complex, originally identified in yeast on the basis of its requirement for transcriptional induction (WINSTON and CARLSON 1992), is among the best characterized (KINGSTON et al. 1996; KINGSTON and NARLIKAR 1999). The SWI/SNF complex contains 11 stable subunits and is required for the expression of a diverse, though limited, set of yeast genes (SUDARSANAM and WINSTON 2000). Complexes highly related to SWI/SNF have been identified and purified in yeast (RSC complex; CAIRNS et al. 1996), flies [Brahma (Brm) complex; PAPOULAS et al. 1998] and mammals (Brg1 and hBrm complexes; IMBALZANO et al. 1994; WANG et al. 1996). While the biochemical properties of the yeast and mammalian SWI/SNF complexes have been studied in considerable detail, their biological function in metazoan development is less well understood.

The yeast SWI/SNF complex, though not essential for growth, is important for both gene activation and repression (DIMOVA et al. 1999; SUDARSANAM et al. 2000). The Drosophila Brm complex, identified on the basis of its requirement for the maintenance of homeotic (HOM) gene expression (KENNISON and TAMKUN 1988; TAMKUN 1995), is essential for proper development as mutations in several Brm complex genes give rise to a broad range of developmental defects. Targeted gene knockouts of Brg1 and hBrm complex components revealed similar essential roles in early murine development (SUMI ICHINOSE et al. 1997; REYES et al. 1998; BULTMAN et al. 2000; KLOCHENDLER-YEIVIN et al. 2000; GUIDI et al. 2001).

Largely on the basis of genetic evidence, the yeast SWI/SNF and RSC complexes have been implicated in aspects of cell cycle regulation (CAO et al. 1997; KREBS et al. 2000). In mammals, the Brg1 and hBrm complexes physically interact with the Retinoblastoma (RB) protein that is essential for both transcription regulation and growth arrest (MUCHARDT and YANIV 1999). Moreover, exit from G1 and S phase has been linked to repressor complexes containing Brg1/hBrm, histone deacetylase (HDAC), and pRB (ZHANG et al. 2000). In addition, the hBrm complex dissociates from mitotic chromosomes, perhaps in response to specific phosphorylation events (MUCHARDT et al. 1996; SIF et al. 1998). Although a specific cell cycle kinase has yet to be identified as a direct effector of Brg1/hBrm complex function, CyclinE/CDK2 has been implicated (SHANAHAN et al. 1999).

Consistent with functions in restricting cellular proliferation, hBrm, Brg1, and INI1/hSNF5 are strongly implicated in a variety of cancers. The first direct genetic evidence for SWI/SNF function in tumor suppression came from studies showing specific inactivating mutations in INI1/hSNF5 were strongly correlated with the majority of malignant rhabdoid tumors (VERSTEEGE et al. 1998; SEVENET et al. 1999b) and they are frequently found in chronic myeloid leukemia (GRAND et al. 1999). In confirmation, chimeric mice harboring INI1/hSNF5 knockout alleles were strongly predisposed to develop nervous system and soft tissue sarcomas that were strikingly similar to the human rhabdoid tumors (KLOCHENDLER-YEIVIN et al. 2000; ROBERTS et al. 2000; GUIDI et al. 2001) and controlled inactivation of INI1/hSNF5 leads to rapid development of aggressive tumors and T cell lymphomas with complete penetrance (ROBERTS et al. 2002). INI1/hSNF5 can be directly recruited to the cyclinD1 promoter where it represses transcription in association with a histone deacetylase (ZHANG et al. 2002). Moreover, Cyclin D1 is overexpressed in atypical teratoid and malignant rhabdoid tumors (AT/RT) linked to loss of heterozygosity at the INI1/hSNF5 locus. Reintroduction of INI1/hSNF5 into AT/RT derived tumor cell lines results in flat cell formation and G1 cell cycle arrest (AE et al. 2002; REINCKE et al. 2003). The growth inhibition is pRB dependent and is associated with decreased expression of a subset of E2F targets (VERSTEEGE et al. 2002) with a corresponding increase in expression of P16INK4a that is typically elevated in senescent cells (BETZ et al. 2002; ORUETXEBARRIA et al. 2004). These tumor cell studies using INI1-deficient cell lines suggested that INI1 acts as a potent regulator of entry into S phase, which may account for its potent tumor suppressor properties. The mammalian hBRM and BRG1 genes are frequently downregulated or mutated in malignant cells derived from a variety of tumors originating in the bladder, lung, and prostate (WONG et al. 2000; REISMAN et al. 2003), as well as breast cancer cell lines (DECRISTOFARO et al. 2001). Moreover, BRG1 physically complexes with BRCA1 involved in most breast cancers (BOCHAR et al. 2000), while ETS-2 and the Brg1 complex form a corepressor to negatively regulate the BRCA1 promoter (BAKER et al. 2003).

In general, the control of entry into S phase is quite similar between mammals and Drosophila. In flies, cyclin E (DmcycE) partners with CDK2 (DmCdc2c) to regulate the transition from G1 to S phase, with DmcycE being both rate limiting and sufficient (KNOBLICH et al. 1994; RICHARDSON et al. 1995); however, DmcycD does not appear to have an essential role in G1-S progression, but instead promotes cellular growth (DATAR et al. 2000; MEYER et al. 2000). During the early syncytial stage of embryogenesis prior to cellularization, nuclei rapidly cycle through replication and mitosis with no intervening G1 or G2 phases, and (type II) DmcycE/CDK2 is required for progression through these S phases (EDGAR and LEHNER 1996). Following the introduction of a G1 phase into the cell cycle after mitotic cycle 16, DmcycE is downregulated in G1 arrested cells (RICHARDSON et al. 1993; SAUER et al. 1995). A loss-of-function mutation in the Drosophila cdc2c gene (cdk2) enhanced the rough eye phenotype of homozygous DmcycEJP, a hypomorphic allele of cyclin E, due to diminished S-phase cells posterior to the morphogenetic furrow, revealing that functional DmcycE/CDK2 dimers were important for normal S-phase regulation in Drosophila (SECOMBE et al. 1998).

Mutations in Brm complex genes were isolated as dominant suppressors of the S-phase defects associated with DmcycEJP without affecting DmcycE protein levels (BRUMBY et al. 2002). Similarly, genetic screens for modifiers of dE2F1/dDP function identified components of the Brm complex as enhancers of the rough eye phenotype caused by overexpression of dE2F1/dDP in the eye disc without affecting the expression of some known dE2F1/dDP target genes (STAEHLING-HAMPTON et al. 1999). Together, these studies suggested that the Brm complex, in parallel with Drosophila RBF, might restrict entry into S phase through effects on chromatin independent of gene-specific transcription.

Heterozygous null mutations in the Drosophila snr1 gene, which encodes an essential component of the Brm complex and is a direct ortholog of INI1/hSNF5, partially suppressed the rough eye and wing phenotypes associated with DmcycEJP (SECOMBE et al. 1998; BRUMBY et al. 2002). Expression of an SNR1 truncation (SNR1-2) that removed highly conserved C-terminal residues augmented the suppression. These and other data (ZRALY et al. 2003) suggested that the Repeat 2 region of SNR1, highly conserved within all SNF5-family proteins including INI1, might function to help restrain the growth-promoting activities of the Brm complex. Confirmation of this hypothesis has come from analyses of our recently isolated temperature-sensitive conditional allele of snr1 that affects a single amino acid residue in the Repeat 2 region. The snr1E1 allele functions as a weak antimorph, with temperature-dependent phenotypes including ectopic wing veins and increased cellular growth as a consequence of increased cell division (MARENDA et al. 2003). These phenotypes are sensitive to snr1 gene dosage and are enhanced or suppressed by mutations in other Brm complex genes, revealing functions for SNR1 in directly regulating aspects of Brm complex activity.

The tumor suppressor properties of INI1/hSNF5 raised the possibility that the snr1E1 conditional phenotypes reflected disruption of cell cycle control at the G1-S boundary and/or transcriptional misregulation of Brm complex target genes. As the SNR1 and INI1 subunits can mediate contacts between their respective Brm complexes and a variety of cellular factors, including transcription factors and retrovirus-encoded proteins, and mutations in both snr1 and INI1/hSNF5 exhibit cell proliferation defects, it appears likely that SNR1 and INI1 may function within their respective Brm (SWI/SNF) complexes to regulate chromatin remodeling activities that are important for normal cell growth. Our study addresses this possibility by taking advantage of the ability to dominantly enhance and suppress the growth and proliferation phenotypes of snr1E1. We found that mutations in DmcycE and DmcycD could dominantly suppress the snr1E1 proliferation phenotypes, while increasing DmcycD levels with a genomic duplication had an enhancing effect. We also found that SNR1 could physically interact with CDK2, suggesting that it might assist in mediating contacts between DmcycE/CDK2 and the Brm complex during the cell cycle.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 LITERATURE CITED
 
Genetic interaction tests:
Fly stocks and genetic crosses were performed on standard yeast/cornmeal/dextrose at 18° or 29° unless otherwise indicated. The DmcycEJP, cdc2c mutant alleles (SECOMBE et al. 1998) and snr1 mutant alleles (ZRALY et al. 2002; MARENDA et al. 2003; ZRALY et al. 2003) have been described. Hdac3 mutant alleles were generously provided by Jeff Simon and Doug Bornemann (University of Minnesota). Other fly strains are described in FlyBase (http://flybase.bio.indiana.edu).

Wings were dissected in 100% isopropanol, placed in DPX mountant (Fluka Chemical, Buchs, Switzerland) and examined at either x10 or x40 magnification. For SEM analysis flies were dehydrated in acetone, gold coated, and examined at 10 kV/x150 magnification on a Jeol JSM5800LV scanning electron microscope. Biomass analyses were performed as described (MARENDA et al. 2003).

Protein interaction studies:
Co-immunoprecipitation experiments were carried out as previously described (BRUMBY et al. 2002; ZRALY et al. 2002) using extracts prepared from 0- to 24-hr Oregon-R (wild-type) embryos or embryos expressing UAS-Cdk2-myc using a da-GAL4 driver (MEYER et al. 2000). Extracts were precleared with protein G-Sepharose and then incubated with primary affinity-purified rabbit polyclonal {alpha}-SNR1 (ZRALY et al. 2002). Protein complexes were precipitated using protein G-Sepharose beads. Bound and unbound proteins were fractionated by SDS-PAGE and analyzed by Western blotting using rabbit {alpha}-BRM (ZRALY et al. 2003), monoclonal {alpha}-DmcycE (8B10; BRUMBY et al. 2002), {alpha}-MYC (Santa Cruz Biotech), or antibodies directed against a nuclear protein that does not interact with SNR1 ({alpha}-EAR; ZRALY et al. 2002).

GST pulldown assays were carried out using fusions of SNR1 to glutathione-S-transferase in the vector pGEX2TK (Amersham, Arlington Heights, IL). Solubilized fusion proteins were purified from induced bacterial cultures using glutathione-agarose (Sigma, St. Louis) and normalized by Coomassie staining of SDS-PAGE gels. Equivalent amounts of fusion protein bound to glutathione-agarose were incubated with Oregon-R or UAS-Cdk2-myc embryo extracts. Bound proteins were separated by SDS-PAGE and analyzed by Western blotting.

Yeast interaction trap assays were performed using SNR1 fusions to the B42 activation domain in the vector pRF4-5o (gift from R. Finley). Each fusion was tested for galactose induction and stability in yeast by Western blotting (our unpublished data). A Drosophila CDK2 fusion (aa 2–315) to the LexA DNA-binding domain in plasmid pRFHM13 was used as the bait. The LexA-DmCDK2 fusion was carried in yeast strain Y309 and mated to strain RFY231 containing the B42-SNR1 fusions or a control B42-Cdi3 fusion (FINLEY and BRENT 1994; FINLEY et al. 1996; KOLONIN and FINLEY 1998). Diploids were assayed for growth on plates lacking leucine and for LacZ production.

Protein immunolocalization:
Oregon-R embryos (0–4 hr) were collected, fixed, and immunostained with affinity-purified SNR1 antibodies as described in DINGWALL et al. (1995). Histochemical detection was carried out using goat anti-rabbit horseradish peroxidase secondary antibodies (Jackson ImmunoResearch, West Grove, PA). Immunofluorescence analysis of SNR1 localization in fixed embryos was carried out using confocal imaging with Cy3-conjugated donkey anti-rabbit secondary antibodies. Chromatin fluorescence was visualized using the intercalating dye 7-aminoactinomycin D (7-AAD; Molecular Probes, Eugene, OR).

RT-PCR analyses:
RNA was isolated from appropriately staged Oregon-R and homozygous snr1E1 late third instar larvae or early pupae raised at 18° or 29° using the RNAqueous Midi RNA isolation kit (Ambion, Austin, TX) according to manufacturer protocols. Equal amounts of RNA were treated with DNAse prior to reverse transcription using MMTV reverse transcriptase (RETROscript, Ambion). When possible, primers were selected to span an intron/exon boundary. PCR reactions were carried out using standard protocols with ExTaq DNA polymerase (Takara, Berkeley, CA), as appropriate for each primer pair. Gene-specific primer sequences used are available as supplementary data at http://www.genetics.org/supplemental/.


   RESULTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 ACKNOWLEDGEMENTS
 LITERATURE CITED
 
snr1 genetically collaborates with cell cycle factors and HDACs:
We recently isolated a temperature-sensitive antimorphic allele of snr1 (snr1E1) that exhibits increased cell proliferation and cell differentiation defects (MARENDA et al. 2003). While heterozygous null alleles of snr1 (snr1R3, snr1SR21) appear as wild type, snr1E1 displays dominant dose-dependent ectopic wing veins both anterior and posterior to the L5 longitudinal vein, and overexpression of an SNR1 derivative (SNR1-2) lacking the C-terminal 109 amino acids results in L2 wing vein disruptions (ZRALY et al. 2002), suggesting an important role for snr1 in wing vein patterning (Table 1; Figure 1B).


View this table:
[in this window]
[in a new window]

  		TABLE 1 Reduced DmcycE and cdc2c functions suppress the snr1E1 wing phenotype

 


View larger version (145K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 1.— DmcycEJP suppresses the snr1E1 wing phenotypes. (A) Wild-type wing. Indicated are the positions of the longitudinal veins (L2–5) and the posterior cross vein (PCV). (B) snr1E1/+, 29°. Ectopic wing veins emanate from the PCV and posterior to the L5 vein, shown by the arrow. (C) snr1E1/snr1SR21, 29°. Extensive disruptions of vein patterning are indicated. (D) DmcycEJP/DmcycEJP; snr1E1/+.

 
The SNR1E1 protein is assembled into Brm complexes at both the permissive and the restrictive temperatures and it co-immunoprecipitates with BRM and OSA in protein extracts prepared from early pupae (MARENDA et al. 2003); thus, mutant phenotypes reflect loss of specific SNR1 functions within the Brm complex, rather than complete loss of Brm complex activities, as would likely occur with null alleles that completely removed a subunit (MARENDA et al. 2004). The wing patterning defects associated with snr1E1 reflect impaired snr1 function, as the phenotypes were strongly enhanced in combination with null alleles (Figure 1C) and were rescued by increasing wild-type snr1 with a transgene, indicating that Brm complexes are forming in this mutant background at a relatively late stage of development that coincides with important final wing patterning events (STURTEVANT and BIER 1995; BIER 2000).

Regulators of the cell cycle are important for normal development, including control of cell growth and tissue morphogenesis (EDGAR and LEHNER 1996). As a consequence of reduced Cyclin E (DmcycE) accumulation in the hypomorphic DmcycEJP mutant, fewer cells enter S phase, resulting in rough eyes and mild wing notching (SECOMBE et al. 1998). Mutations in Brm complex genes, including null alleles of snr1, specifically suppress the DmcycEJP phenotypes through restoration of G1-S progression in the eye imaginal disc cells without affecting DmcycE protein levels; thus, the Brm complex appears to function genetically in G1 by helping to regulate entry into S phase downstream of DmcycE (BRUMBY et al. 2002). We found that snr1E1 was able to suppress the DmcycEJP eye and wing phenotypes in a dosage- and temperature-dependent manner (our unpublished data), while both DmcycEJP and a DmcycE deficiency [Df(2L)osp29, referred to herein as Df(2L)cycE] modestly suppressed the snr1E1 wing patterning defects (Table 1; Figure 1D). Since lowering DmcycE levels results in fewer cells entering S phase (SECOMBE et al. 1998; BRUMBY et al. 2002), the ability of DmcycE mutations to suppress the appearance of ectopic veins associated with snr1E1 may be a consequence of reduced numbers of cells completing mitosis.

Cyclin E is the regulatory subunit of the CycE/CDK2 heterodimer, CycE/CDK2 is both necessary and sufficient for entry into S phase in Drosophila, and CDK2 is encoded by the cdc2c gene (LEHNER and O'FARRELL 1990). While CDK2 levels are not normally limiting during the cell cycle (LEHNER and LANE 1997; LANE et al. 2000), consistent with reduced DmcycE function, trans-heterozygous null alleles of cdc2c (cdc2c1, cdc2c3) exhibited a modest suppression of the vein phenotypes associated with snr1E1.

In mammals, HDACs can form corepressor complexes with Retinoblastoma (pRB) and the SWI/SNF-related Brg1 and hBrm complexes to regulate events during the G1-S portion of the cell cycle (ZHANG et al. 2000). The ability of INI1 to repress cyclin D1 transcription in cultured mammalian cells is dependent on recruitment of HDAC1 (ZHANG et al. 2002); thus, it has previously been suggested that the Brm complex may assist in directly repressing DmcycD expression during late G1 phase. In Drosophila, altered expression of DmcycD, using either a small deficiency (Figure 2B) or genomic duplication, revealed no significant wing patterning phenotypes (our unpublished data) and had no detectable effect on the snr1E1 wing phenotype (Table 2; Figure 2, C and D; our unpublished data). We next tested for genetic interactions between snr1E1 and Drosophila Hdac1 and Hdac3 homologs. Flies heterozygous for mutant alleles of Hdac3 or Hdac1/rpd3 showed weak, if any, wing vein defects (Figure 2, E and G), while trans-heterozygous combinations strongly enhanced the snr1E1 ectopic vein phenotype (Figure 2, F and H). Null alleles of Hdac1/rpd3 exhibited modest enhancement of the snr1E1 phenotype, while alleles of Hdac3 strongly enhanced the appearance of ectopic veins (Table 2). Thus, it appears that snr1 may cooperate with HDACs in blocking or restricting the expression of genes involved in vein cell determination. A strong amorphic allele of brm suppressed the interaction between snr1 and Hdac alleles (Table 2), suggesting that the snr1/Hdac interaction phenotype was sensitive to Brm complex levels.



View larger version (132K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 2.— snr1 cooperates with DmcycA and Hdac genes to regulate wing pattern development. Wings were observed at x40 magnification. (A) snr1E1/+. (B) Df(1)cycD/+. (C) Df(1)cycD/+; snr1E1/+. (D) Dp(1;Y)cycD;snr1E1/+. (E) Hdac3N/+. (F) Hdac3N/snr1E1. (G) Rpd304556/+. (H) Rpd304556/snr1E1. (I and J) DmcycA deficiencies. (I) Df(3L)vin5/snr1E1. (J) Df(3L)vin7/snr1E1.

 

View this table:
[in this window]
[in a new window]

  		TABLE 2 DmcycA and HDAC mutations enhance the snr1E1 phenotype in wings

 
Cyclin A, Cyclin B, and String (Cdc25 phosphatase) are required for G2-M phase and the completion of mitosis. Although DmcycB did not interact with DmcycEJP, deficiencies removing DmcycA and string enhanced the DmcycEJP eye phenotype, suggesting that an inability to complete postmorphogenetic furrow mitoses exacerbated the phenotypic effects of reduced S phases (SECOMBE et al. 1998). While deficiencies removing DmcycA strongly enhanced the snr1E1 wing phenotype (Figure 2, I and J; Table 2), reductions of string, DmcycB, and DmcycC had no significant effect (our unpublished data). Thus, the snr1E1 wing phenotypes appear to be sensitive to specific cyclin levels.

snr1 functions as a growth regulator:
The snr1E1 mutation results in increased body mass and wing size, observed in both snr1E1/+ and snr1E1/snr1E1 flies at 18°. This phenotype is sensitive to snr1 dosage as it is enhanced in an snr1 null mutant background and suppressed by additional copies of the wild-type snr1 gene (MARENDA et al. 2003). Therefore, within the context of growth regulation, the snr1E1 allele represents a loss-of-function phenotype. Both body mass (milligrams per fly) and cell number (cells per unit area; Table 3) were significantly (P < 10–4) greater among the snr1E1 flies (–/– > +/–) compared with parental (red,e) controls (Table 4), suggesting that the snr1E1 growth defect may be a consequence of increased cell division. A developmental delay associated with snr1E1 occurs during late larval growth, and there is significant early pupal stage lethality. To determine whether the growth defect was coincident with this developmental period, wild-type (red,e) and snr1E1 homozygous stage P3 pupae (within 8 hr of the onset of metamorphosis) were selected and weighed. The snr1E1 pupae exhibited significantly increased biomass at both 18° (0.00177 g/fly; N = 90) and 29° (0.00146 g/fly; N = 60) relative to wild-type (red,e) controls at either 18° (0.00156 g/fly; N = 130) or 29° (0.00136 g/fly; N = 63). Biomass differences were also significant when identical genotypes were compared at different growth temperatures. Following larval development, the time spent progressing through metamorphosis to adult emergence was identical between snr1E1 and parental controls (data not shown). Thus, the snr1E1 growth defects coincide with the developmental delay.


View this table:
[in this window]
[in a new window]

  		TABLE 3 Biomass, wing area, and cell counts of select genotypes

 

View this table:
[in this window]
[in a new window]

  		TABLE 4 Statistical significance of extragenic enhancement and suppression of the snr1E1 growth phenotypes

 
To help determine the basis for the snr1E1 growth phenotypes, we looked for dominant genetic interactions (enhancement and suppression) between snr1E1 and mutations in Brm complex genes, as well as known cell cycle/growth regulation genes. Appropriate male flies generated from independent crosses were examined and scored for body mass and cell number (Table 3). Body mass was calculated from at least 100 similarly aged flies of each genotype. Dissected wings were photographed and wing size was determined by measuring the area within the wing margins. Average cell number was scored by counting cells in a defined unit area of the wing, as each wing cell is known to secrete only one hair during development (MEYER et al. 2000). Statistical comparisons were performed to determine the significance of any genetic interactions with snr1E1 (Table 4). The snr1E1 growth phenotypes were suppressed by reducing Brm complex activities (Table 3), as a consequence of either decreased complex formation using a brm amorphic mutation (brm2) or the expression of a dominant-negative brm that allows the complex to form with a defective subunit, but has reduced ATPase activity (brmK804R; ELFRING et al. 1998). Importantly, both types of alleles show comparable suppression of the snr1E1 growth phenotypes, suggesting that brmK804R represents a true loss-of-function effect. Thus, similar to the vein patterning defects (MARENDA et al. 2003, 2004), the increased growth associated with snr1E1 appears to be dependent on otherwise functional Brm complexes, suggesting that the SNR1 subunit may be critically important for proper regulation of some Brm complex functions.

The DmcycEJP mutation causes reduced numbers of eye imaginal disc cells to enter S phase as measured by BrdU incorporation, leading to an overall decrease in cell number (SECOMBE et al. 1998; BRUMBY et al. 2002). We therefore examined the ability of DmcycEJP, as well as cdc2c mutations to modify snr1E1 growth phenotypes. Compared with the heterozygous snr1E1/+ and wild-type controls, wings obtained from DmcycEJP/DmcycEJP;snr1E1/+ flies were significantly reduced in overall size. Similarly, wings from Df(2L)cycE/+;snr1E1/+ were smaller, with reduced cell numbers (Table 3). The homozygous DmcycEJP mutation and heterozygous Df(2L)cycE both resulted in significantly reduced body mass and cell number relative to wild-type controls, consistent with diminished numbers of cells entering S phase. Heterozygous null alleles of cdc2c failed to exhibit any growth defects alone, although they did weakly suppress the snr1E1 growth phenotypes. Thus, reduced DmcycE/CDK2 function suppressed the growth defects associated with snr1E1, suggesting that the increased body mass and cell number were possibly related to increased S phases and sensitive to DmcycE/CDK2 levels.

In mammals, cooperation between the Brg1/hBrm complex and HDAC activities has been associated with repression of both cyclinE and cyclinD expression and controlling entry into S phase (HARBOUR and DEAN 2000; ZHANG et al. 2000, 2002). Reducing Drosophila HDAC function with heterozygous null alleles of Hdac1/Rpd3 or Hdac3 did not significantly impact overall body size (cell growth/biomass) relative to wild-type controls; however, both mutants exhibited increased cell number/unit area in the wing, which in the case of Hdac3N also displayed increased wing size (Table 3). While the Hdac3N mutant did not significantly modify the increased cell growth and wing area phenotypes of snr1E1, it did significantly enhance the increased cell number phenotype (Tables 3 and 4). Thus, our genetic results suggest that while both Hdac1 and Hdac3 functions are important for proper Drosophila wing vein patterning in collaboration with the Brm complex, in our assays only Hdac3 appeared to be rate limiting for regulating cell proliferation in snr1E1 larval wing disc cells.

A duplication of the genomic region containing DmcycD carried on the Y chromosome dominantly enhanced the snr1E1 growth (biomass) and proliferation phenotypes, but did not affect the overall wing size (Table 3). As specific DmcycD mutant alleles are not available, we used a deficiency that removed the surrounding genomic region to determine whether reducing DmcycD dosage would modify the snr1E1 phenotypes. We measured wing area in female flies carrying a heterozygous DmcycD deficiency [Df(1)19], in the absence (1.59 mm2) or presence (1.57 mm2) of snr1E1 and found that wings from these flies were significantly smaller compared to wild type (1.74 mm2) and heterozygous snr1E1 females (1.88 mm2). Consistent with previous reports suggesting that DmcycD influences cell growth (size) but not overall cell proliferation (DATAR et al. 2000; MEYER et al. 2000), we found that cell counts (cells per unit area) were somewhat higher in females harboring the DmcycD deficiency (446), compared to wild-type controls (437), snr1E1/+ (424), or Df, snr1E1 combinations (425). Thus, it appears that the snr1E1 growth and proliferation phenotypes may be sensitive to DmcycD levels.

We next examined mutations in select genes that regulate subsequent cell cycle events, including replication and mitosis (Table 3). Mutations in the dpa gene encoding an MCM-family chromatin binding protein and NTP-hydrolase involved in prereplication complex formation and disc proliferation typically exhibit reduced imaginal disc cell proliferation (FEGER et al. 1995). Surprisingly, dpa1, snr1E1 trans-heterozygotes exhibited significantly increased growth compared to snr1E1 in all three assays (wing area, cell counts, and biomass). A strong amorphic allele of string (stg7B), encoding a cdc25-phosphatase, affects cell size and G2-M phase (EDGAR and O'FARRELL 1990). In combination with snr1E1, both biomass and wing area were increased without affecting cell number, suggesting synergistic nonautonomous effects on cell growth, but not proliferation. A hypermorphic stg9B allele when trans-heterozygous with snr1E1 similarly had no effect on cell number, but wing area and biomass were reduced.

To better understand the molecular events surrounding the increased cell number phenotype associated with snr1E1, we measured the expression of a set of genes important for cell cycle progression. As controls, snr1 mRNA and protein levels were found to be unaffected in snr1E1 mutants at either the restrictive (29°) or the permissive (18°) temperatures (Figure 3; MARENDA et al. 2003). Both DmcycE and DmcycD transcript accumulations were unaffected in snr1E1 homozygous larvae or pupae at either 18° or 29° as measured by RT-PCR, and DmcycE protein levels remained unchanged (our unpublished data). Other gene products important for cell cycle progression were largely unaffected, including Rbf, DmcycA, DmcycB, DmcycC, dDP, esg, rux, Pol-{alpha}, d-Myc, wg, dpp, RNR2, and dE2F2 (Figure 3; our unpublished data). Although dE2F1 and dDP transcripts were expressed at essentially wild-type levels at both temperatures and the expression of several dE2F1/dDP target genes was unchanged, some genes required for the replication (S) and mitotic (M) portions of the cell cycle showed modest reductions in snr1E1 mutants at both 18° and 29°, including string/cdc25 and dpa, suggesting that only a subset of genes involved in S or G2-M phase were affected. The expression of string did not appear to be reduced in a dominant-negative brmK804R background (data not shown), suggesting that the reduction associated with snr1E1 was not a general consequence of reduced Brm complex ATP-dependent functions. While the snr1E1 growth defects do not appear to be related to altered DmcycE or DmcycD expression per se, they may be related to misregulation of a limited set of DmcycE/CDK2 or dE2F/dDP targets (ISHIDA et al. 2001). Thus, our data suggest that Brm complex activities help to restrict S-phase entry downstream of G1 regulators, including DmcycE and DmcycD.



View larger version (51K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 3.— snr1E1 affects cell division downstream of DmcycE and DmcycD. RT-PCR analysis of transcript accumulation in wild-type and snr1E1 mutant pupae at 18° and 29°.

 
A conserved region in SNR1 mediates contacts with DmcycE/CDK2:
While DmcycE can be found in immunoprecipitated Brm complexes, it was unknown whether the precipitated DmcycE was present as DmcycE/CDK2 heterodimers and which Brm complex subunits were important to facilitate these interactions (BRUMBY et al. 2002). Embryonic extracts prepared from control or transgenic flies overexpressing an MYC-tagged CDK2 (MEYER et al. 2000) were immunoprecipitated with antibodies to SNR1 (Figure 4). While BRM and DmcycE, but not a control protein (EAR; ZRALY et al. 2002), were found in SNR1 immunoprecipitates as expected (BRUMBY et al. 2002), only a portion of the overexpressed CDK2-MYC was found in SNR1 immunoprecipitates. As CDK2-MYC is highly abundant in these extracts, the low efficiency of coprecipitation may result from a large pool of CDK2-MYC that is not present in complexes with DmcycE, resulting in unstable interactions with the Brm complex. It is not possible to perform these experiments using extracts prepared from snr1 null flies, as the gene is essential, it is required for oogenesis (ZRALY et al. 2003), and SNR1is a core component of stable Brm complexes purified from embryos (PAPOULAS et al. 1998).



View larger version (76K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 4.— The Brm complex co-immunoprecipitates with DmcycE and DmCDK2. Embryonic extracts were prepared from flies expressing MYC-tagged DmCDK2 (MEYER et al. 2000), incubated with affinity-purified rabbit {alpha}-SNR1 and precipitated with protein-G Sepharose. E, 100 µg extract; S, supernatant fraction (lane represents 20% of input protein extract); P, pellet fraction (lane represents 50% of co-immunoprecipitated proteins). Input extracts (E), supernatant (S), and pelleted (P) proteins were fractionated by SDS-PAGE and probed with antibodies to BRM, SNR1, DmcycE, MYC (recognizing DmCDK2-MYC), or a control protein (EAR; ZRALY et al. 2002). Antibodies to DmcycE recognize both type I and type II isoforms. The SNR1 protein appears as a doublet (~45 and 47 kD) when the embryonic yolk protein is not present (P).

 
The ability of SNR1 and INI1 to contact a variety of cellular factors together with strong interaction phenotypes observed between snr1E1 and DmcycEJP suggested that SNR1 might be important to assist or mediate contacts between the Brm complex and DmcycE/CDK2. To test this possibility, portions of SNR1 were fused to glutathione S-transferase (GST) and in vivo pulldown experiments were performed (Figure 5). Equivalent amounts of solubilized GST:SNR1 fusions were immobilized on glutathione agarose and incubated with embryonic extracts obtained from flies expressing CDK2-MYC. A monoclonal antibody to the MYC epitope identified CDK2 bound to the GST:SNR1 fusion, but not GST alone. Although CDK2 associated with full-length SNR1, the interaction was strongest between CDK2 and C-terminal SNR1 residues (aa 230–370) that include the Repeat 2 and putative coiled-coil regions. SNR1/CDK2 associations were somewhat reduced relative to full length by deletion of the coiled coil (aa 311–370), while minimal association was observed with N-terminal fusions. Importantly, although DmcycE was not overexpressed in these extracts, it was present in the SNR1/CDK2 complexes with a similar pattern of binding affinity, suggesting that the observed interactions were not a consequence of excess CDK2. As a control, SNR1/Brm complex binding was assayed using the same extracts and fusions, with BRM antibodies for detection. In contrast to SNR1/DmcycE/CDK2, BRM showed significant association only with the full-length SNR1 fusion. These data reveal that sequences within the amino and carboxy terminal regions of SNR1 are important in cis for stable association with the Brm complex and that defined regions of SNR1 (including the highly conserved Repeat 2) are involved in stable contacts with cellular factors, such as DmcycE/CDK2.



View larger version (56K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 5.— A conserved region within SNR1 mediates interactions with DmcycE/CDK2. Portions of SNR1 were fused to glutathione S-transferase (GST). Fusion proteins were solubilized, purified, and quantified as shown in the Coomassie-stained gel. Each fusion was bound to glutathione agarose and incubated with extracts from embryos expressing DmCDK2-MYC. Bound proteins were eluted and the presence of DmCDK2, DmcycE, and BRM was examined by Western blot. I, input protein.

 
Mammalian CDK2 binds directly to a large array of cellular protein targets through divergent sequences (ADAMS et al. 1996). Yeast interaction trap assays were used to help define the SNR1 residues important for interaction with CDK2 (Figure 6). Drosophila CDK2 (cdc2c) fused to the LexA DNA-binding domain (LexA-DB) was used as a bait to test for specific interactions with portions of SNR1 fused to the B42 activation domain (B42-AD). The activation domain fusions were produced at comparable levels following galactose induction and were stable as measured by Western blot using {alpha}-SNR1 antibody (our unpublished data). Independent interaction reporters included growth on media lacking leucine and lacZ activity (FINLEY and BRENT 1994). Consistent with the results using GST:SNR1 fusions, CDK2 strongly interacted in both assays with the full-length SNR1 protein (aa 15–370). Stronger interactions were observed between CDK2 and SNR1 C-terminal residues (aa 240–370), comparable with a control interaction (Figure 6B), while there was no detectable interaction with N-terminal residues (aa 15–240). The G256D substitution affects SNR1E1 function, though not synthesis or stability of Brm complexes in vivo and exhibits modest temperature-dependent interaction with the Drosophila homeotic gene regulator trithorax (TRX) in similar yeast assays (MARENDA et al. 2003). A full-length B42AD:SNR1(G256D) fusion was tested in combination with LexA:CDK2 and displayed only modest reductions in lacZ activity compared with the wild-type SNR1, although growth on media lacking leucine (Gal, –Leu) was not significantly affected at 30°. Thus, our results suggest that residues within the C terminus of SNR1 may be capable of forming direct contacts with Drosophila CDK2. In similar tests, we were unable to detect specific interactions between SNR1 and DmcycE (our unpublished observations).



View larger version (44K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 6.— SNR1 can form direct contacts with DmCDK2. Portions of SNR1 were fused to the B42 activation domain (B42-AD) and expressed in yeast by galactose induction. Yeast strains bearing SNR1 fusions were mated to a strain harboring DmCDK2 fused to the LexA DNA-binding domain to assess interactions. A B42-DmcycD fusion was included as a positive control. Interactions were assessed by production of ß-galactosidase (LacZ) and, independently, growth on media lacking leucine. Yeast harboring both fusions was grown to midlog phase in media containing leucine, washed, and diluted in water at similar cell density. A dilution series was plated onto media lacking leucine and growth was scored after 3 days at 30°.

 
Dissociation of the Brm complex from mitotic chromosomes:
Early precellular blastoderm embryos are characterized by a rapid (~8 min per cycle) series of synchronous S and M phases with no intervening G1 or G2 phases (EDGAR and O'FARRELL 1989; FOE et al. 1993; EDGAR 1995) and DmcycE/CDK2 is required to drive progression through S phase (LEHNER and LANE 1997). Previously, it was shown that BRM and BAP111 (Brm complex components) dissociate from metaphase chromosomes in late embryos (PAPOULAS et al. 2001) and the Brm complex is important for regulating S phases in eye imaginal discs (STAEHLING-HAMPTON et al. 1999; BRUMBY et al. 2002). We therefore used immunolocalizations of SNR1 in early syncytial embryos to determine whether Brm complex associations with chromatin might be regulated in the absence of G1 and G2 phases. Embryos (0–4 hr after egg laying) were collected, fixed, and immunostained with {alpha}-SNR1 antibodies (DINGWALL et al. 1995). SNR1 appeared to be tightly associated with nuclei at the earliest stages (Figure 7A) and uniformly distributed in all cell nuclei along the embryonic AP axis through the extended germ band stage (Figure 7; DINGWALL et al. 1995). However, SNR1 was associated with a subset of nuclei in ~1–3% of precellular blastoderm embryos (Figure 7B). Fluorescent confocal imaging of SNR1 localization in similarly staged embryos showed that the protein was tightly associated with chromatin in the vast majority of nuclei. Occasionally, SNR1 appeared more diffuse in the cytosol and closer examination revealed that the majority of chromosomes were in metaphase. In rare cases, both condensed and decondensed chromosomes were observed within the same blastoderm embryo, and in those cases SNR1 was associated with the decondensed chromosomes (Figure 7, D–F). This suggests a possible coupling between global chromosome condensation and Brm complex localization in the absence of zygotic transcription and G1 phases.



View larger version (70K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 7.— SNR1 dissociates from mitotic chromatin. Fixed embryos of different developmental stages immunostained with {alpha}-SNR1. All embryos are oriented with anterior (a) to the left, posterior (p) to the right, and dorsal at the top. (A) Cleavage stage embryo. SNR1 appears in all nuclei. (B) Stage 3 embryo. SNR1 appears in all nuclei, though punctate in a restricted region near the posterior. (C) Stage 5 embryo (cellular blastoderm). Confocal image showing uniform SNR1 expression. (D–F) Confocal image of a stage 4 embryo. (D) Chromatin is stained with 7-AAD to distinguish S- and M-phase nuclei. Nuclei near both poles are in S phase (diffuse), while those ~20–90% egg length (between arrows) are undergoing mitosis and exhibit condensed chromatin. (E) In S-phase nuclei, SNR1 is punctate, although diffuse and excluded from mitotic chromatin. (F) Merged image.

 

   DISCUSSION
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
ACKNOWLEDGEMENTS
 LITERATURE CITED
 
Our snr1E1 conditional mutant displays wing patterning defects and increased mitotic growth at both the permissive (18°) and the restrictive (29°) temperatures. The mutant phenotypes are sensitive to both temperature of incubation and snr1 gene dosage, indicating that they specifically result from reduced or compromised SNR1 function, rather than from complete disruption of Brm complex activities (MARENDA et al. 2003; ZRALY et al. 2003). In contrast to the use of null alleles that may reduce total complex number by half, snr1E1 produces a stable protein that is assembled into Brm complexes at both temperatures, thus allowing complexes to form and bind their targets, but then are defective in some other function of the complex. This point is critical for our studies, as there are significantly different effects resulting from complete loss of functional Brm complexes or activities as contrasted with impaired functions that result from the incorporation of defective subunits. To help understand the functional roles of SNR1 within the conserved Brm ATP-dependent chromatin remodeling complex during metazoan development, we have taken advantage of these dosage- and temperature-dependent snr1E1 phenotypes, as well as the brmK804R dominant-negative, both of which result in the incorporation of defective subunits.

In this report, we have shown that snr1 can genetically interact with a subset of genes involved in cell cycle control. In addition, co-immunoprecipitation of DmcycE/CDK2 and the Brm complex indicated that stable complexes could form in vivo, while both GST-pulldown and yeast two-hybrid studies suggested that residues within SNR1 might help mediate or stabilize these contacts. SNR1 is strongly conserved with counterparts in yeast (SNF5) and mammals (INI1). The most conserved portions among SNR1-related proteins occur within the ~200-amino-acid C-terminal region comprising two imperfect repeats and a coiled coil. The repeat regions are important for contacts with a variety of cellular factors, including Drosophila Bicoid, the HOX gene regulators TRX and HRX/MLL (ROZENBLATT ROSEN et al. 1998; MARENDA et al. 2003), c-MYC (TAKAYAMA et al. 2000), and Epstein Barr NA2 (WU et al. 1996; CHENG et al. 1999) as well as the viral-encoded proteins HIV integrase and HPV E1 (KALPANA et al. 1994; LEE et al. 1999). In addition, yeast SNF5 is involved in direct associations with the GAL4 transcriptional activator (NEELY et al. 2002). Our recent results have shown that contacts with conserved features of SNR1 were important for recruiting or modulating Drosophila Brm complex functions in vivo (MARENDA et al. 2003). The SNR1/DmCDK2 interaction may also be an important conserved feature, as we observed similar contacts between SNR1 C-terminal residues and mammalian CDK2 using yeast two-hybrid assays (our unpublished data).

Integration of Brm complex functions with developmental and cell cycle signals:
Components of the mammalian Brm complexes, including the hBrm/BRG-1 and BAF155 (MOR) subunits, are phosphorylated prior to the onset of mitosis and this modification may be important for restricting or modulating complex activity (MUCHARDT et al. 1996; SIF et al. 1998). However, the cell cycle kinase involved and specific target residues within Brm complex components have not been identified. On the basis of work from cultured mammalian cells (SHANAHAN et al. 1999) and our results (this study; BRUMBY et al. 2002), CycE/CDK2 appears to be among the likely candidates for important regulatory kinase functions during portions of the cell cycle.

We found that CDK2 is capable of forming contacts with SNR1 through the Repeat 2 and coiled-coil regions. What might be the importance of the SNR1-CDK2 interaction? SNR1 and INI1 do not contain any obvious CDK2 phosphorylation sites (BOULIKAS 1995; ADAMS et al. 1996; LACY and WHYTE 1997; KWON and NORDIN 1998; ADAMS et al. 1999) and SNR1 does not appear to be a phosphoprotein, as assays using a variety of general protein phosphatases produced no detectable change in SNR1 electrophoretic migration on SDS-PAGE gels (our unpublished data). This may be misleading, as other putative phosphoproteins, including Drosophila RBF, do not change electrophoretic mobility when treated with phosphatases (DU et al. 1996). However, yeast SFH1p found in the SWI/SNF-related RSC complex and a close relative of SNR1/INI1/SNF5 appears to be phosphorylated during G1 phase (CAO et al. 1997). Thus, while SNR1 does not appear to be the likely direct target for DmcycE/CDK2 regulation, our genetic results suggest the possibility that contacts between SNR1 and CDK2 may serve to stabilize or regulate interactions between DmcycE/CDK2 and the Brm complex or help to direct kinase activity, targeted either to other components of the Brm complex or to unknown cellular proteins (MUCHARDT et al. 1996; SIF et al. 1998; SHANAHAN et al. 1999).

How might interactions between the Brm chromatin remodeling complex and DmcycE/CDK2 contribute to appropriate cell cycle regulation? A growing body of evidence strongly suggests that ATP-dependent chromatin remodeling complexes perform essential functions in controlling normal mitotic cell cycles (HARBOUR and DEAN 2000; NEELY and WORKMAN 2002). For example, the SWI/SNF complex is important for the expression of mitotic genes and DNA replication in yeast (FLANAGAN and PETERSON 1999; KREBS et al. 1999). In mammals, the Brm-related complexes functionally interact with histone deacetylases and pRB to block entry into S phase (DUNAIEF et al. 1994; STROBER et al. 1996; MUCHARDT and YANIV 1999; STROBECK et al. 2000; ZHANG et al. 2000). As a consequence of losing or misregulating chromatin remodeling activities, normal cell cycle control is disrupted. Specifically, loss of INI1 is associated with aggressive cancers, leads to the rapid development of tumors in knockout mice (ROBERTS et al. 2002), and results in G1-specific defects (AE et al. 2002). Further, overexpression of Cyclin E can abrogate cell cycle arrest caused by the introduction of BRG1 into SW13 adenocarcinoma cells (SHANAHAN et al. 1999).

The requirements for ATP-dependent chromatin remodeling activities during the cell cycle are likely to be quite complex, perhaps involving known functions in controlling gene transcription (activation and repression) and/or regulating aspects of chromosome replication. In cultured mammalian cells, INI1 was shown to repress cyclinD1 transcription in G1 phase through collaboration with HDAC1 (AE et al. 2002; VERSTEEGE et al. 2002; ZHANG et al. 2002; REINCKE et al. 2003). Unlike mammalian cyclinD, Drosophila DmcycD is not required for entry into S phase, but has been proposed to function during G1 to regulate cell growth (DATAR et al. 2000; MEYER et al. 2000). While our snr1E1 mutant phenotypes are sensitive to Cyclin D levels, the expression of DmcycD is unaffected in our mutant, consistent with our view that the snr1E1 growth defects are likely due to misregulation of genes downstream of DmcycE, possibly involving targets of E2F regulation.

In addition to demonstrating Brm complex regulation of gene expression during the S and G2 phases, our results also suggest RNA PolII-independent roles in restricting S-phase entry. For example, SNR1 is excluded from mitotic chromatin during the early embryonic nuclear divisions in the absence of zygotic transcription or G1-G2 phases. During these early divisions, type II DmcycE is a potent inducer of S phase (CRACK et al. 2002) and this form exhibits strong in vivo associations with SNR1. One scenario is that the Brm complex is recruited to specific chromosomal sites by sequence-specific repressors where the complex might act to stabilize binding of the repressor and/or remodel nucleosomes in an ATP-dependent manner, thereby establishing a repressive environment to restrict replication initiation (Figure 8). The cellular proteins involved in potentially recruiting the Brm complex to specific loci involved in replication initiation are not presently known, but may include transcription factors, such as RBF/E2F or ORC (BRUMBY et al. 2002; our unpublished data). Recruitment of CycE/CDK2 to replication origins (FURSTENTHAL et al. 2001) and interaction with SNR1 might then allow for inactivation of Brm activity and release of the complex from chromatin through phosphorylation of specific subunits. The SNR1E1 mutant protein likely compromises one or more of these interactions, reducing the effective recruitment of the Brm complex to targets that are normally repressed by Brm complex activities. This could possibly lead to compromised S-phase restriction, partly relieving the requirement for DmcycE/CDK2 activity to allow progression into S phase.



View larger version (34K):
[in this window]
[in a new window]
[Download PPT slide]
  		FIGURE 8.— SNR1 functions to regulate activities of the Brm (SWI/SNF) chromatin remodeling complex. The conserved SNR1 (INI1/SNF5) subunit helps to mediate contacts between the complex and cellular factors including Cyclin E/CDK2 (this study) or transcriptional repressor proteins bound to specific chromatin sites (MARENDA et al. 2004). The SNR1 subunit assists in regulating the activities of the complex, perhaps through associations with cellular proteins that restrict the ATP-dependent chromatin remodeling activities. During the G1 phase of the cell cycle, the Brm (SWI/SNF) complex likely functions not only in gene activation, but also to restrict cell cycle progression downstream of the activities of Cyclin E/CDK2 and in parallel with pRBF (BRUMBY et al. 2002). Following phosphorylation by cell cycle kinases, the complex dissociates from chromatin, perhaps facilitating the start of S phase. Subsequent dephosphorylation by an unknown phosphatase would then allow the complex to assist some gene activation functions of E2F to promote cell cycle progression (this study; HENDRICKS et al. 2004).

 
While the molecular mechanisms or targets for Brm regulation of the cell cycle remain largely unresolved, several observations support our model. First, mutations in Brm complex genes can suppress the DmcycEJP S-phase defects without affecting DmcycE protein or transcript accumulation or other G1-phase regulators. Conversely, reduced DmcycE function can suppress the mitotic defects associated with snr1E1, suggesting that snr1E1 perhaps affects events related to the G1-S transition. Second, reduction of Brm complex functions, by either reduced complex abundance or disrupted ATPase activity with dominant-negative BRMK804R, can similarly suppress the snr1E1 and DmcycEJP defects. In this scenario, if the Brm complex cannot be recruited to specific loci or appropriately function in chromatin remodeling, then not only could it bypass strict requirements for DmcycE function at promoting S-phase entry, but also it could similarly suppress the enhanced growth functions associated with snr1E1. This interpretation is consistent with the Brm complex acting in parallel with RBF (BRUMBY et al. 2002) and findings that brm and mor reductions enhanced the overexpression phenotypes of dE2F/dDP in developing eye discs (STAEHLING-HAMPTON et al. 1999). While the expression of some late dE2F1/dDP targets are reduced in snr1E1, G1-S-phase regulators including DmcycE are unaffected. Factors that act at chromosomal loci to affect replication timing or specificity, such as ORC2, ORC5 (LOUPART et al. 2000; PFLUMM and BOTCHAN 2001), dE2F1/dDP (CAYIRLIOGLU and DURONIO 2001), DmMyb (BEALL et al. 2002), or nonhistone chromosomal proteins such as HP1 (NIELSEN et al. 2002), may therefore collaborate with the Brm complex to modify chromatin in a cell cycle-dependent manner. It is interesting to note that E2F1/DP is important for localizing the origin recognition complex to assist in activating replication (BOSCO et al. 2001). Mounting evidence has, in fact, linked ATP hydrolysis and early replication events (LEE and BELL 2000). The Brm complex also functions coordinately with the chromatin assembly factor ASF1 that may help to target the complex to newly replicated and assembled chromatin in S phase (MOSHKIN et al. 2002), and assist in the transcription of S-phase-specific genes (SUTTON et al. 2001).

Reduced expression of some cell cycle genes in an snr1E1 mutant suggests that the Brm complex may be important for DmcycE/CDK2 or dE2F-dependent gene regulation. Targets of Brm complex activity may include dpa and stg, which are important for S- and G2-M-phase events. Both the dpa and the stg transcripts are reduced in snr1E1, and mutations in both genes dominantly enhance the growth phenotypes of snr1E1. However, while reduced stg affects only cell size and not cell number, dpa affects the growth and proliferation phenotypes of snr1E1. The DPA protein is a member of the MCM family of replication licensing factors (MCM4), with intrinsic DNA-dependent ATPase and helicase activities and is a stable component of prereplication complexes (FEGER et al. 1995; SU et al. 1996). As phosphorylation of the mammalian hBrm complex is associated with entry into M phase, one possibility is that the fly complex directly or indirectly restricts expression of dpa and string through collaboration with RBF and HDAC. In this scenario, DmcycE/CDK2 would be important to promote cell cycle progression and subsequent transcriptional program through interactions with the Brm complex, perhaps mediated or stabilized by contacts with the SNR1 subunit. It is also plausible, perhaps quite likely, that the Brm complex might be targeted by another cyclin-cdk complex at the G2-M transition.

The SNR1 subunit helps to constrain Brm complex activities
The mammalian homolog of the snr1 gene, INI1/hSNF5, has been directly linked to the majority of childhood aggressive rhabdoid tumors (MRT; VERSTEEGE et al. 1998; SEVENET et al. 1999a,b) and the onset of tumorigenesis in transgenic knockout mice (KLOCHENDLER-YEIVIN et al. 2000; ROBERTS et al. 2000; GUIDI et al. 2001). Further, conditional removal of murine INI1/hSNF5 leads to a rapid (~11 weeks) onset of lymphomas or rhabdoid-type tumors with 100% penetrance (ROBERTS et al. 2002) and ~45% of human CD8+ T cell lymphomas are associated with loss of genetic material at the INI1 locus (22q11.2). Overexpression of INI1 in INI1-deficient cells leads to growth arrest in G1 and apoptosis (AE et al. 2002). Although the exact mechanism of growth arrest in mammalian cells is unknown, it may involve increased expression of P16INK4a or activation of pRB (BETZ et al. 2002; ORUETXEBARRIA et al. 2004). Thus INI1/hSNF5 and, by analogy, snr1 may be formally classified as potent tumor suppressors, a property that appears to be unique among genes encoding metazoan SWI/SNF complex subunits.

Brm complex functions are essential throughout development (TAMKUN 1995; SIMON and TAMKUN 2002). In addition to convincing roles for the Brm complex in regulating the HOM genes, genetic analyses and ectopic expression of dominant-negative BRMK804R, OSA, and SNR1 have revealed striking phenotypes. These include significant reductions in eye and wing size as well as wing vein and abdominal defects (ELFRING et al. 1998; COLLINS et al. 1999; PAPOULAS et al. 2001; ZRALY et al. 2003), suggesting important functions of the complex in both gene activation and repression. Moreover, snr1 can genetically function as a suppressor of variegation (PEV), suggesting possible direct roles in restricting gene expression through chromatin effects (ZRALY et al. 2003). While little is known about the individual subunit roles or specific functional relationships among Brm complex subunits, our results suggest that SNR1 can serve under certain circumstances to constrain Brm complex activities at in vivo targets. In support of this view, a mutation that disrupts the Brm-associated ATPase activity (BrmK804R) without affecting complex assembly or stability not only results in reduced growth (our results; ELFRING et al. 1998), but also suppresses the snr1E1 mitotic and wing phenotypes. Thus, it may be that the increased growth associated with snr1E1 (and possibly INI1-associated tumors) is a consequence of misregulated Brm complex ATPase-dependent activities. HDACs are important for gene silencing, including critical roles in regulating genes required for cell cycle progression (HARBOUR and DEAN 2000; ZHANG et al. 2000, 2002). There are four identified HDAC genes in Drosophila, with rpd3 most similar to Hdac1. While Hdac mutants showed only modest effects on the snr1E1 growth phenotype, both Hdac1/rpd3 and Hdac3 mutations significantly enhanced the snr1E1 ectopic wing veins. These interactions were suppressed by reducing Brm function, supporting our view that SNR1 may assist in gene repression through restraints on Brm complex chromatin remodeling activities.


   ACKNOWLEDGEMENTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
LITERATURE CITED
 
We gratefully acknowledge Helena Richardson for DmcycE antibodies, fly strains, advice, and sharing unpublished information; Jeff Simon and Doug Bornemann for Hdac3 alleles; Christian Lehner for the UAS-Cdk2-myc strain; Russ Finley for yeast strains and plasmids; Kaye Suyama and Matthew Scott for confocal imaging; David Amberg for advice on yeast interaction assays; and Mark Davino for technical assistance. This work was supported in part by grants from the March of Dimes (5-FY97-702) and the National Science Foundation (MCB-0221563).


   FOOTNOTES
 
1 Present address: Department of Cell Biology, Emory University School of Medicine, Atlanta, GA 30322. Back


   LITERATURE CITED
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGEMENTS
 LITERATURE CITED
 

ADAMS, P., W. SELLERS, S. SHARMA, A. WU, C. NALIN et al., 1996 Identification of a cyclin-cdk2 recognition motif present in substrates and p21-like cyclin-dependent kinase inhibitors. Mol. Cell. Biol. 16: 6623–6633.[Abstract]

ADAMS, P. D., X. LI, W. R. SELLERS, K. B. BAKER, X. LENG et al., 1999 Retinoblastoma protein contains a C-terminal motif that targets it for phosphorylation by cyclin-cdk complexes. Mol. Cell. Biol. 19: 1068–1080.[Abstract/Free Full Text]

AE, K., N. KOBAYASHI, R. SAKUMA, T. OGATA, H. KURODA et al., 2002 Chromatin remodeling factor encoded by ini1 induces G1 arrest and apoptosis in ini1-deficient cells. Oncogene 21: 3112–3120.[CrossRef][Medline]

BAKER, K. M., G. WEI, A. E. SCHAFFNER and M. C. OSTROWSKI, 2003 Ets-2 and components of mammalian SWI/SNF form a repressor complex that negatively regulates the BRCA1 promoter. J. Biol. Chem. 278: 17876–17884.[Abstract/Free Full Text]

BEALL, E. L., J. R. MANAK, S. ZHOU, M. BELL, J. S. LIPSICK et al., 2002 Role for a Drosophila Myb-containing protein complex in site-specific DNA replication. Nature 420: 833–837.[CrossRef][Medline]

BETZ, B. L., M. W. STROBECK, D. N. REISMAN, E. S. KNUDSEN and B. E. WEISSMAN, 2002 Re-expression of hSNF5/INI1/BAF47 in pediatric tumor cells leads to G1 arrest associated with induction of p16INK4a and activation of RB. Oncogene 21: 5193–5203.[CrossRef][Medline]

BIER, E., 2000 Drawing lines in the Drosophila wing: initiation of wing vein development. Curr. Opin. Genet. Dev. 10: 393–398.[CrossRef][Medline]

BOCHAR, D. A., L. WANG, H. BENIYA, A. KINEV, Y. XUE et al., 2000 BRCA1 is associated with a human SWI/SNF-related complex: linking chromatin remodeling to breast cancer. Cell 102: 257–265.[CrossRef][Medline]

BOSCO, G., W. DU and T. L. ORR-WEAVER, 2001 DNA replication control through interaction of E2F-RB and the origin recognition complex. Nat. Cell Biol. 3: 289–295.[CrossRef][Medline]

BOULIKAS, T., 1995 Phosphorylation of transcription factors and control of the cell cycle. Crit. Rev. Eukaryotic Gene Expression 5: 1–77.[Medline]

BRUMBY, A. M., C. B. ZRALY, J. A. HORSFIELD, J. SECOMBE, R. SAINT et al., 2002 Drosophila cyclin E interacts with components of the Brahma complex. EMBO J. 21: 3377–3389.[CrossRef][Medline]

BULTMAN, S., T. GEBUHR, D. YEE, C. LA MANTIA, J. NICHOLSON et al., 2000 A Brg1 null mutation in the mouse reveals functional differences among mammalian SWI/SNF complexes. Mol. Cell 6: 1287–1295.[CrossRef][Medline]

CAIRNS, B. R., Y. LORCH, Y. LI, M. ZHANG, L. LACOMIS et al., 1996 RSC, an essential, abundant chromatin-remodeling complex. Cell 87: 1249–1260.[CrossRef][Medline]

CAO, Y., B. R. CAIRNS, R. D. KORNBERG and B. C. LAURENT, 1997 Sfh1p, a component of a novel chromatin-remodeling complex, is required for cell cycle progression. Mol. Cell. Biol. 17: 3323–3334.[Abstract]