Genetics, Vol. 150, 675-686, October 1998, Copyright © 1998

Expression and Properties of Wild-Type and Mutant Forms of the Drosophila Sex Comb on Midleg (SCM) Repressor Protein

Douglas Bornemanna, Ellen Millera, and Jeffrey Simona
a Department of Biochemistry and Department of Genetics and Cell Biology, University of Minnesota, St. Paul, Minnesota 55108

Corresponding author: Jeffrey Simon, Department of Biochemistry, University of Minnesota, 1479 Gortner Ave., St. Paul, MN 55108., simon{at}biosci.cbs.umn.edu (E-mail).

Communicating editor: J. A. BIRCHLER


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

The Sex comb on midleg (Scm) gene encodes a transcriptional repressor of the Polycomb group (PcG). Here we show that SCM protein is nuclear and that its expression is widespread during fly development. SCM protein contains a C-terminal domain, termed the SPM domain, which mediates protein-protein interactions. The biochemical function of another domain consisting of two 100-amino-acid-long repeats, termed "mbt" repeats, is unknown. We have determined the molecular lesions of nine Scm mutant alleles, which identify functional requirements for specific domains. The Scm alleles were tested for genetic interactions with mutations in other PcG genes. Intriguingly, three hypomorphic Scm mutations, which map within an mbt repeat, interact with PcG mutations more strongly than do Scm null alleles. The strongest interactions produce partial synthetic lethality that affects doubly heterozygous females more severely than males. We show that mbt repeat alleles produce stable SCM proteins that associate with normal sites in polytene chromosomes. We also analyzed progeny from Scm mutant germline clones to compare the effects of an mbt repeat mutation during embryonic vs. pupal development. We suggest that the mbt repeat alleles produce altered SCM proteins that incorporate into and impair function of PcG protein complexes.

DEVELOPMENT along the anterior-posterior (A-P) body axis in Drosophila is controlled by the homeotic products of the Antennapedia and bithorax complexes (LEWIS 1978 Down; KAUFMAN et al. 1980 Down). The homeotic proteins are expressed in spatially-restricted A-P domains (WHITE and WILCOX 1985 Down; CELNIKER et al. 1989 Down; KARCH et al. 1990 Down), and their proper deployment along this axis is crucial for A-P pattern formation.

The Polycomb group (PcG) genes encode a set of transcriptional repressors that mediate restricted homeotic gene expression. Mutations in PcG genes cause ectopic expression of homeotic proteins in inappropriate positions along the A-P axis (STRUHL and AKAM 1985 Down; MCKEON and BROCK 1991 Down; SIMON et al. 1992 Down). At least 13 Drosophila PcG genes are required for homeotic gene repression (MCKEON and BROCK 1991 Down; SIMON et al. 1992 Down; SOTO et al. 1995 Down; see SIMON 1995 Down and PIRROTTA 1997 Down for reviews). Once PcG proteins become engaged in homeotic gene repression, at ~4 to 5 hr of embryogenesis (JONES and GELBART 1990 Down; SIMON et al. 1992 Down; SOTO et al. 1995 Down), they are continuously required to maintain repression during subsequent stages of development (DUNCAN and LEWIS 1982 Down; WU et al. 1989 Down; JONES and GELBART 1990 Down). In agreement with this, most PcG proteins examined so far are expressed throughout embryonic, larval, and pupal stages (PARO and ZINK 1992 Down; MARTIN and ADLER 1993 Down; DECAMILLIS and BROCK 1994 Down; LONIE et al. 1994 Down; CARRINGTON and JONES 1996 Down).

A likely molecular explanation for the large number of PcG components is that they work together in multiprotein complexes. Thus, loss of one PcG protein could impair the repressive function of the entire complex. Cytological evidence for PcG complexes is provided by identical distributions of the Polycomb (PC), polyhomeotic (PH), and Polycomblike (PCL) proteins at approximately 100 polytene chromosome sites (FRANKE et al. 1992 Down; LONIE et al. 1994 Down). In addition, the Posterior sex combs (PSC) and Sex comb on midleg (SCM) protein distributions overlap extensively with these sites (RASTELLI et al. 1993 Down; PETERSON et al. 1997 Down). Biochemical evidence for PcG protein associations includes coimmunoprecipitation of PC with PH (FRANKE et al. 1992 Down) and extra sex combs (ESC) with Enhancer of zeste [E(Z); JONES et al. 1998] from fly embryo extracts. There is also evidence for complexes of mammalian PcG proteins. The mouse PH and PSC homologs coimmunoprecipitate with the PC homolog (ALKEMA et al. 1997 Down) and the human PH and PSC homologs coimmunoprecipitate and cofractionate on sucrose gradients (GUNSTER et al. 1997 Down).

Despite accumulating data on PcG complexes, the precise biochemical roles of individual PcG members are not known. Although PcG proteins localize to specific chromosomal sites, none of those yet tested exhibits sequence-specific DNA-binding activity in vitro. In addition, none of the PcG proteins yet sequenced contains recognizable catalytic domains. The main functional clues afforded by PcG protein sequences are evolutionarily conserved domains, such as the chromodomain, which mediate PcG chromosome associations or protein interactions (MESSMER et al. 1992 Down; CARRINGTON and JONES 1996 Down; PLATERO et al. 1996 Down; ALKEMA et al. 1997 Down).

The SCM protein plays a key role in PcG repression because embryos that lack both maternal and zygotic SCM die with severe homeotic transformations (BREEN and DUNCAN 1986 Down). SCM protein contains several homology domains, two of which are also present in the PH PcG protein (BORNEMANN et al. 1996 Down). One of these shared domains is a type of zinc finger present in SCM in two copies and in PH in one copy. SCM and PH also share a C-terminal domain of 65 amino acids, termed the SPM domain, which mediates self-binding and cross-binding of these two proteins (PETERSON et al. 1997 Down). SCM is even more similar to another fly protein, the product of the tumor suppressor gene lethal (3) malignant brain tumor [l(3)mbt; WISMAR et al. 1995]. The SCM and L(3)MBT proteins share zinc fingers, the SPM domain, and a third domain consisting of 100-amino-acid long repeats. These repeats, termed mbt repeats (WISMAR et al. 1995 Down; BORNEMANN et al. 1996 Down), are present in two tandem copies in SCM and three copies in L(3)MBT. The biochemical role of mbt repeats is not known.

To investigate the importance of the Scm domains in vivo, we have characterized molecular lesions associated with Scm mutant alleles. This analysis identifies a subset of Scm mutations that maps to the first mbt repeat and that displays especially strong genetic interactions with other PcG mutations. We used polyclonal antibodies to show that SCM protein is nuclear and to determine its temporal and spatial distribution during development. We also assessed the expression and stability of mutant SCM proteins and their accumulation at specific sites on polytene chromosomes.


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

Sequence determination of mutant alleles:
Genomic DNA for Scm mutant alleles and corresponding background chromosomes was amplified by PCR and cloned as described below. The template DNA for sequencing reactions was double-stranded plasmid DNA prepared as a mixture from 8 to 10 independent PCR clones. Sequencing was performed by dideoxy chain termination with Sequenase 2.0 (United States Biochemical, Cleveland).

ScmSu(z)302: Genomic DNA purified from homozygous mutant ScmSu(z)302 pupae was used as template for PCR. The 5' one-third of the Scm gene was amplified using the primers 5'-ACTAATTGTGCGGCTCG-3' and 5'-GAGATTCGCACATGCCC-3', and the product was digested with NgoMI. The resulting 1.2-kb fragment was inserted into pBluescript KSII+. The 3' two-thirds of Scm was amplified using primers 5'-GCTGGATGGAAGTGACT-3' and 5'-GAATCACGAGCAGTTGG-3', the product was digested with SalI and NruI, and the resulting 1.9-kb fragment was inserted into pBluescript.

ScmET50: Template DNA for PCR amplification was prepared from ScmET50/ScmP12 pupae. ScmP12 is a deletion that removes the Scm gene (BORNEMANN et al. 1996 Down). The 5' one-third of Scm was amplified using primers 5'-ACTAATTGTGCGGCTCG-3' and 5'-CGCCAACCATCGAATGT-3', the product was digested with ApoI and XhoI, and the resulting 1.2-kb fragment was inserted into pBluescript. The 3' two-thirds of the gene was amplified using primers 5'-GCTGGATGGAAGTGACT-3' and 5'-GGAAACGCAATTGATAC-3', the product was digested with XhoI and HpaI, and the resulting 1.8-kb fragment was inserted into pBluescript. The ET50 mutant lesion was confirmed by PCR cloning and by sequencing the same region from the st e background chromosome (JURGENS 1985 Down).

ScmD1, ScmD2, ScmH1, ScmM56, ScmM36, ScmR5-13, ScmKM23, and ScmK2: Template DNA for PCR amplification of each allele was obtained from pharate adults of genotype Scmallele/ScmSu(z)302. The transheterozygotes were identified by presence of extra sex combs and absence of the balancer marker, Sb. The 3' two-thirds of the gene was amplified using primers 5'-AGTGCGACAACGTCATC-3' and 5'-GAATCACGAGCAGTTGG-3', the product was digested with BamHI and NruI, and the resulting 2.2-kb fragment was inserted into pBluescript. Mutant lesions were confirmed by PCR cloning and sequencing the corresponding regions from the respective background or isogenic chromosomes: Cbx Ubx for ScmD1 (BREEN and DUNCAN 1986 Down), E(z)61 e for ScmH1 (R. JONES, personal communication), "47.29.1" for ScmR5-13 (M. MULLER and H. GYURKOVICS, personal communication), and ri e for ScmKM23 (K. MATTHEWS, personal communication). Since ScmM56 and ScmM36 were raised on the same kar ry chromosome (G. MARONI, personal communication), they each provided background sequence for the other.

Generation and purification of SCM antibodies:
A 1.9-kb SalI-NruI fragment encoding SCM amino acids 324 to 877 was isolated from the cDNA Sc9 (BORNEMANN et al. 1996 Down) and was inserted into the vector pGEX-BgRP3i (JONES et al. 1998 Down). The resulting glutathione-S-transferase (GST)-SCM fusion protein was prepared from Escherichia coli inclusion bodies as described in LI et al. 1994 Down, purified by preparative SDS gel electrophoresis, and used as immunogen in rabbits. Crude sera that tested positive for immunogen reactivity were affinity-purified against solubilized GST-SCM protein coupled to the Actigel ALD affinity chromatography resin (Sterogene Bioseparations, Inc., Arcadia, CA). Antibodies were bound and eluted from the Actigel column according to the manufacturer's instructions.

Drosophila protein extracts and Western blots:
Protein extracts were prepared by homogenizing tissues in 2x SDS sample buffer (100 mM Tris pH 6.8, 200 mM dithiothreitol, 4% SDS, 0.2% bromophenol blue, 20% glycerol) plus 1 mM phenylmethylsulfonyl fluoride. Homogenates were sonicated for 30 sec and heated to 95° for 5 min. Insoluble material was pelleted by microcentrifugation, and the supernatants were recovered. Unfertilized egg extracts were prepared using 200 eggs per 50 µl of sample buffer. Embryonic extracts were prepared using a 1:2 v:v ratio of embryos to sample buffer. Extracts from pupae and adults were generated using 10 µl of sample buffer per animal. Ovaries and adult heads were dissected from 20 females and homogenized in 40 µl of sample buffer. Extracts from larvae were prepared by first freezing animals in liquid nitrogen, followed by tissue disruption with a mortar and pestle at -20° and resuspension at a 1:2 ratio of tissue volume to sample buffer (first and second instars) or in 10 µl of sample buffer per animal (third instars). We found that mortar and pestle disruption of frozen tissues provided more efficient total protein recovery from larval stages, presumably because of greater disruption of the larval cuticle barrier. Relative concentrations of extracts were gauged by Coomassie Blue staining of proteins after SDS gel electrophoresis.

Immunodetection on Western blots was with affinity-purified rabbit anti-SCM antibody (1:1000) and goat anti-rabbit-HRP secondary antibody (1:2000; Bio-Rad, Hercules, CA). Levels of tubulin, used in some experiments as a control for lane loading, were detected using mouse anti-tubulin primary antibody (1:5000; Amersham, Piscataway, NJ) and goat anti-mouse-HRP secondary antibody (1:5000; Jackson Immunoresearch Labs., Inc., West Grove, PA). Signals were developed using the ECL detection system (Amersham).

Immunostaining of embryos, larval tissues, and polytene chromosomes:
Immunostaining of embryos was performed as described in SIMON et al. 1992 Down using a rabbit polyclonal abdominal-A (ABDA) antibody (KARCH et al. 1990 Down). Whole-mount immunostaining of larval tissues was performed essentially as described in CARRINGTON and JONES 1996 Down using anti-SCM antibody at a 1:35 dilution. Polytene chromosomes were immunostained with anti-SCM as described in PETERSON et al. 1997 Down.

Generation of Scm germline clones:
Scm mutations (M56, D1, Su(z)302, or M36) were recombined onto a third chromosome bearing a centromere-linked FRT site at cytological location 82B (XU and RUBIN 1993 Down). FRT(82B) Scm-/TM3 females were mated to HS-FLP/Y; FRT(82B) ovoD1/TM3 males (CHOU et al. 1993 Down). FLP recombinase was induced by heat shocking the progeny for 1.5 hr at 36° during each of 4 consecutive days starting the fourth day after egg-laying commenced. The FRT(82B) Scm-/FRT(82B) ovoD1 virgin female progeny were collected. To obtain unfertilized Scm- eggs for Western analysis, these females were mock-mated to males carrying the dominant male-sterile ß-tubulin mutation, B2tD (KEMPHUES et al. 1980 Down). To generate embryos and larvae that contain solely the Su(z)302 form of SCM protein from both maternal and zygotic sources, FRT(82B) ScmSu(z)302/FRT(82B) ovoD1 females were mated to ScmSu(z)302/TM6B,Tb males. Embryos produced from this cross were stained with ABDA antibody. Salivary glands from the non-Tb progeny were harvested to obtain ScmSu(z)302 mutant polytene chromosomes.


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

Generation and specificity of SCM antibody:
Rabbit polyclonal antibodies were generated against a GST-SCM fusion protein that contains SCM amino acids 324 to 877 (see MATERIALS AND METHODS). The affinity-purified SCM antibodies detect a single major band on Western blots of wild-type fly embryonic extracts (Figure 1A). This reacting species migrates at approximately 100 kD, which is close to the 94-kD predicted size for SCM protein (BORNEMANN et al. 1996 Down).



View larger version (53K):
In this window
In a new window
Download PPT slide
  		Figure 1. Detection of SCM protein in wild-type and mutant extracts. Western blots are shown using affinity-purified polyclonal anti-SCM antibodies. (A) Extract from wild-type (Dfw67c2) fly embryos. Numbers indicate positions of molecular weight markers. (B) Extracts from unfertilized eggs. Lane 1 contains extract from wild-type (Dfw67c2) eggs and lanes 2 through 5 contain extracts from eggs derived from germline clones homozygous for the indicated Scm mutations. The faint species in lane 3 likely represent altered SCM proteins produced from the D1 allele, which is a C-terminal frameshift (see Figure 5). (Bottom) ß-Tubulin detection on the same blot as a control for lane loading.

We wished to determine if this reacting species corresponds to SCM protein by analyzing extracts from Scm mutant embryos. However, embryos that are zygotically mutant for Scm likely contain wild-type SCM protein derived from maternal expression (BREEN and DUNCAN 1986 Down). Therefore, we used the dominant female sterile technique coupled with the FLP recombination system (GOLIC and LINDQUIST 1989 Down; CHOU et al. 1993 Down) to generate females with germlines containing Scm mutant clones. Unfertilized eggs collected from these germline mosaic females should contain solely mutant maternal Scm product.

Animals of genotype HS-FLP/+; FRT(82B) Scm-/FRT(82B) ovoD1 were heat shocked to induce recombination during larval stages (see MATERIALS AND METHODS). To assess production of germline clones, ScmD1/TM3 fathers were mated to mothers containing putative ScmD1/ScmD1 clones, and cuticles from the ScmD1/ScmD1 progeny embryos were examined. We observed embryos with all segments transformed towards eighth abdominal (not shown), which appeared identical to Scm mutant embryos lacking both maternal and zygotic product generated by pole cell transplantation (BREEN and DUNCAN 1986 Down). Similar homeotic phenotypes were seen in hemizygous ScmM56 embryos produced from ScmM56 germline clone females. The severity of these phenotypes, which requires loss of maternal Scm function (BREEN and DUNCAN 1986 Down), confirms that these FRT recombinant chromosomes produce Scm germline clones.

Females containing Scm germline clones were mock-mated to males containing a dominant male sterile ß-tubulin mutation (KEMPHUES et al. 1980 Down), and unfertilized eggs were collected. Protein extracts were prepared from Scm mutant eggs and from wild-type unfertilized eggs collected in parallel. As shown in Figure 1B, lane 1, wild-type eggs contain the ~100-kD species that reacts with anti-SCM antibody. In contrast, little or none of this species is detected in extracts from mutant ScmM56 or ScmD1 unfertilized eggs (lanes 2 and 3). This result indicates that the ~100-kD species is SCM protein. Figure 1B, lanes 4 and 5 show that full-length SCM protein is present in ScmSu(z)302 and ScmM36 unfertilized eggs. The abundance of Su(z)302 protein appears comparable to wild type, whereas the level of M36 protein is reduced.

Expression of SCM protein during development:
Protein extracts were prepared from wild-type animals at different developmental stages, and relative levels of SCM protein were assessed on Western blots. Figure 2A shows that SCM protein is expressed throughout development, with the highest levels detected during embryonic and pupal stages and lower levels seen in the intervening larval stages. The detection of SCM protein in 0- to 2-hr embryos and in ovaries (Figure 2B), together with mRNA expression data (BORNEMANN et al. 1996 Down), is consistent with maternal Scm product furnished as both protein and mRNA. The accumulation of SCM protein in adult males (Figure 2A) and in adult heads (Figure 2B) suggests that SCM might also function in terminally differentiated fly tissues. SCM protein expression in specific tissues was examined by whole-mount staining with anti-SCM antibody. Figure 3 shows that SCM protein accumulates in nuclei and that its distribution appears ubiquitous in larval tissues including the salivary gland, imaginal discs, and the central nervous system. Similar widespread nuclear accumulation has been described for several other PcG proteins (MARTIN and ADLER 1993 Down; DECAMILLIS and BROCK 1994 Down; CARRINGTON and JONES 1996 Down).



View larger version (26K):
In this window
In a new window
Download PPT slide
  		Figure 2. Expression of SCM protein during development. (A) Detection of SCM protein by Western blot in wild-type extracts from indicated embryonic, larval, pupal, and adult stages. Approximately equal amounts of total protein were loaded per lane. (B) Detection of SCM protein in ovaries and heads dissected from females.



View larger version (84K):
In this window
In a new window
Download PPT slide
  		Figure 3. Accumulation of SCM protein in nuclei of larval tissues. Wild-type whole-mount tissues immunostained with anti-SCM antibodies are shown. (A) Salivary gland. (B) Eye-imaginal disc. (C) Wing disc. (D) Larval central nervous system.

Properties of Scm alleles:
Mutations in the Scm gene have been identified in several types of genetic screens, including screens for adults with extra sex combs (JURGENS 1985 Down), for dominant enhancers of Polycomb (KENNISON and TAMKUN 1988 Down), for dominant enhancers of Contrabithorax (BREEN and DUNCAN 1986 Down), for dominant suppressors of zeste1 (WU et al. 1989 Down), and for lethal noncomplementation of deficiencies that remove Scm (K. MATTHEWS, personal communication; G. MARONI, personal communication). The 11 Scm alleles analyzed here, and their sources, are listed in Table 1. We have previously shown that ScmXF24 is a deletion that removes a C-terminal portion of the Scm coding region (BORNEMANN et al. 1996 Down). Each of the other 10 Scm alleles fails to complement the recessive lethality of XF24.


 
View this table:
In this window
In a new window

  		Table 1. Properties and interactions of Scm alleles

To assess the relative severities of the Scm alleles, homozygous mutant embryos were stained with antibodies against the ABDA homeotic protein. In cases where the homozygous mutants showed grossly abnormal morphology, presumably due to other mutations on the third chromosome, the mutant alleles were instead analyzed as hemizygotes in combination with Df(3R)GB104 (LINDSLEY and ZIMM 1992 Down). In wild-type embryos, ABDA is restricted to a posterior domain encompassing parasegments 7 to 13 (KARCH et al. 1990 Down; Figure 4A). We found that the Scm alleles fall into two broad classes of severity. The first class is exemplified by ScmM56 homozygotes (Figure 4B), which show abundant ABDA misexpression anterior to PS7, primarily in the central nervous system. Similar misexpression patterns are seen in ScmD1 and ScmXF24 homozygotes (SIMON et al. 1992 Down; BORNEMANN et al. 1996 Down), which are null or nearly null alleles according to phenotypic criteria (BREEN and DUNCAN 1986 Down). Based upon these results, and lethal phases during late embryonic or first larval stages (JURGENS 1985 Down; BREEN and DUNCAN 1986 Down), we classify five Scm alleles as null or nearly null (Table 1). The second class of Scm alleles produced little or no ABDA misexpression, including four alleles with patterns indistinguishable from wild type. An example of slight ABDA misexpression is provided by ScmKM23 hemizygotes, which misexpress ABDA in ~30 to 60 cells anterior to PS7 (Figure 4C). Based upon little or no ABDA misexpression, and survival of hemi- or homozygotes to mid-larval or pupal stages, we classify six Scm alleles as hypomorphic (Table 1).



View larger version (69K):
In this window
In a new window
Download PPT slide
  		Figure 4. Expression of ABDA protein in embryos containing null or hypomorphic Scm mutations. Embryos immunostained with ABDA antibody (KARCH et al. 1990 Down) are shown. (A) Wild-type; arrow indicates parasegment 7. (B) ScmM56 homozygote derived from heterozygous parents; arrow indicates misexpression in the central nervous system. (C) ScmKM23/DfGB104 hemizygote derived from heterozygous parents; arrows indicate misexpression in the lateral hypodermis. (D) ScmSu(z)302 homozygote derived from ScmSu(z)302 germline clone mother.

Sequence analysis of Scm alleles:
Genomic DNA was isolated from homozygous ScmSu(z)302 pupae, and the mutant DNA was PCR-amplified, cloned and sequenced (see MATERIALS AND METHODS). We found that the entire Scm open reading frame is wild type in DNA sequence except for a single G to A transition that replaces amino acid D215 with N. This missense mutation maps within the first of the two mbt repeats (Figure 5B and Figure C). The Su(z)302 mutation also fortuitously removes a BamHI site that is unique within the Scm gene (Figure 5A). Southern blot analysis independently verified the loss of this BamHI site in Su(z)302 mutant DNA (data not shown). The loss of this site in Su(z)302 DNA was used to design a strategy for PCR cloning and sequencing the remaining Scm alleles, including the embryonic lethal alleles. We found that Scmx/ScmSu(z)302 animals, where Scmx represents any lethal allele, survive to pharate adulthood with phenotypes similar to Su(z)302 homozygotes. Thus, PCR clones for most of the remaining Scm mutations were obtained by isolating genomic DNA from Scmx/ScmSu(z)302 pupae followed by shotgun cloning of BamHI-digested PCR products (see MATERIALS AND METHODS).



View larger version (32K):
In this window
In a new window
Download PPT slide
  		Figure 5. Molecular lesions of Scm alleles. (A) A restriction map of the genomic region encompassing the Scm transcription unit is shown. Deletions in the M56 and XF24 alleles are indicated below the DNA map. A transcript map is shown with the positions of start and stop codons, an alternative 3' polyadenylation site (BORNEMANN et al. 1996 Down), and introns indicated. The intron sizes are 58, 65, and 61 bp, respectively. Restriction sites: R, EcoRI; C, ClaI; P, PstI; Xb, XbaI; B, BamHI; X, XhoI; S, SalI; H, HindIII; N, NruI; Sa, SacI. (B) A domain map of SCM protein is shown with the SPM domain, two mbt repeats, and three potential zinc fingers indicated. Alterations in SCM protein caused by ten mutant alleles are shown. The allele names are in italic type, and the resulting molecular changes to the protein are indicated in bold type here and are listed in Table 1. The asterisk indicates replacement with a stop codon. M56 is an 11-bp deletion that causes a frameshift and deletes the C-terminal half of SCM. D1 is a frameshift that removes the SPM domain. (C) The amino acid sequence of the first mbt repeat and the effects of four Scm mutant alleles are shown. Underlined residues are conserved positions in an alignment of mbt repeats from SCM and L(3)MBT (BORNEMANN et al. 1996 Down).

The locations of 10 Scm mutant lesions are shown in Figure 5B. There is good correspondence between the strengths of the alleles and their predicted effects upon SCM protein. Each of the molecular lesions for the four null alleles sequenced (Table 1) causes deletion of a substantial portion of SCM protein. ScmH1 is a nonsense mutation located at position 249 (Figure 5B) and thus could produce, at best, a severely truncated protein that lacks several homology domains. Similarly, ScmM56 produces a frameshift that deletes the C-terminal half of SCM protein. These two mutations are the most likely protein nulls among the Scm alleles. Indeed, Western blot analysis fails to detect SCM protein in the ScmM56 mutant (Figure 1B).

The six hypomorphic alleles are mutations that should produce altered forms of essentially full-length SCM protein (Figure 5B and Table 1); five are missense mutations, and the sixth, ScmR5-13, is an in-frame 12-bp deletion that removes four amino acids. Among the hypomorphic alleles, four are clustered in the first mbt repeat (Figure 5B and Figure C). Another hypomorphic allele, ScmM36, produces a cysteine to tyrosine change within a region that might form a zinc-binding domain (Zn3; BORNEMANN et al. 1996 Down). In agreement with their identification as missense mutations, ScmM36 and ScmSu(z)302 produce full-length versions of SCM protein (Figure 1B).

Three Scm alleles affecting the mbt repeat domain show stronger PcG genetic interactions than null Scm alleles:
Animals doubly mutant for two different PcG genes often show phenotypes more extreme than either single mutant alone (JURGENS 1985 Down; ADLER et al. 1989 Down; CHENG et al. 1994 Down). It has been suggested that this phenotypic enhancement results from PcG protein complexes that are more severely impaired by simultaneous reduction or alteration of multiple components. Having defined the molecular lesions and relative strengths of many Scm mutations (Figure 5 and Table 1), we wished to test the Scm alleles for interactions with other PcG mutations. In particular, we wished to compare genetic interactions exhibited by mutations affecting different parts of SCM protein.

We began by generating animals doubly heterozygous for each of the Scm alleles in Table 1 and for a lethal allele of Polycomb, Pc3. We found that transheterozygous Pc3/Scm adults display more severe homeotic phenotypes than Pc3/+ adults, and that this enhancement is seen with all Scm alleles tested. These phenotypes include transformations of wing to haltere, antenna to leg, second and third leg to first leg, and fourth abdominal segment to fifth. However, three Scm alleles, Su(z)302, R5-13, and ET50, produced much stronger interactions with Pc3 than did other Scm alleles. These three were the only Scm alleles to cause partial lethality in combination with Pc3. As shown in Figure 6A, the transheterozygous progeny classes for Su(z)302, R5-13, and ET50 are reduced to about one-third that expected for full viability. In contrast, the Scm null alleles H1 and M56 are fully viable with Pc3. The surviving Su(z)302, R5-13, and ET50 transheterozygous progeny also exhibited more severe homeotic phenotypes than did other Pc3/Scm combinations. We observed a marked male sex bias among these survivors (Figure 6A). In the most severe case, only about 5% of the surviving Pc3/ScmSu(z)302 progeny were female. Similarly, partial lethality and a male sex bias were seen with the reciprocal crosses consisting of Su(z)302, R5-13, or ET50 females mated to Pc3 males (data not shown). These interactions likely result from the Scm lesions rather than mutations at other loci because these three Scm alleles produce similar phenotypic effects and were isolated independently on different genetic backgrounds (JURGENS 1985 Down; WU et al. 1989 Down; M. MULLER and H. GYURKOVICS, personal communication). Each of the three alleles maps to the first mbt repeat in SCM protein (Figure 5B and Figure C). Thus, Su(z)302, R5-13, and ET50 are hypomorphic mutations based upon their behavior as homozygotes (Table 1), yet they interact with Pc3 more strongly than do null Scm alleles (Figure 6A).



View larger version (39K):
In this window
In a new window
Download PPT slide
  		Figure 6. Survival of adults heterozygous for Scm mutations and for mutations in other PcG genes. Bar graphs depict the percentage of total progeny from the indicated crosses that are doubly mutant for an Scm allele and another PcG allele. (A) Survival of Scm/Pc3 adults. (B) Survival of Scm/SceD1 adults. (C) Survival of ph409; Scm/+ adults. Full viability of the double mutant progeny class corresponds to frequencies of 33% for the crosses in A and B and 50% for the cross in C. The Scm alleles used and the numbers of total progeny scored are indicated below each graph. Scmx represents any Scm allele used. In A and B, the proportions of surviving double mutant progeny that are male or female are represented by the black or grey portions, respectively, of the bars. Numbers within these bars indicate the percentage of surviving progeny that are male. Due to X linkage of ph, only data for the male progeny are depicted in C.

We wondered if the strong Su(z)302, R5-13, and ET50 interactions reflect a general property of these mbt repeat alleles. In particular, since Pc3 is an antimorphic allele (DUNCAN and LEWIS 1982 Down), we were concerned that it might show unusual interaction patterns with Scm alleles. To address this, we examined genetic interactions of Scm alleles with mutations in other PcG genes and with other mutant alleles of the Pc gene.

Figure 6B shows data for interaction of several Scm alleles with SceD1, an allele of the Sex comb extra (Sce) gene (BREEN and DUNCAN 1986 Down). This is the single existing allele of the uncloned Sce gene; it is homozygous lethal and its precise nature has not been characterized. As with Pc3, there was partial lethality and male sex bias among SceD1/Scm transheterozygotes when the Scm mbt repeat alleles were used. In the most extreme example, only 1 out of 60 surviving SceD1/ScmET50 adults was female. Once again, transheterozygotes with the Scm null alleles H1 and M56 were fully viable and produced progeny in the expected sex ratios (Figure 6B).

The ph409 mutation disrupts one of the two tandem copies of the polyhomeotic gene located on the X chromosome (DURA et al. 1987 Down). Although ph409 is a hypomorphic, homozygous viable allele, CHENG et al. 1994 Down have shown that it is lethal or near-lethal in combination with other PcG mutations, including alleles of Scm. To assess ph interaction, we crossed homozygous ph409 females to males bearing Scm mutations and scored viability among the double mutant male progeny. We found that all Scm alleles (Table 1) cause partial lethality among ph409; Scm/+ males. However, Su(z)302, R5-13, and ET50 showed much higher lethality in this combination than did H1 and M56 (Figure 6C).

Finally, we tested a subset of the Scm alleles for interaction with additional Pc alleles, Pc2 and PcXT109. Pc2 is a frameshift near the C terminus that produces detectable PC protein of about the normal size (FRANKE et al. 1995 Down). PcXT109 is associated with a 2-kb deletion, and PC protein is not detected in PcXT109 mutant embryos (FRANKE et al. 1995 Down). We found that interactions with Pc2 were similar to interactions with Pc3; the three mbt repeat alleles produced partial lethality in combination with Pc2, whereas ScmH1/Pc2 and ScmM56/Pc2 animals were fully viable (data not shown). The Scm mutations, as a group, showed less severe enhancement with PcXT109 than with Pc2 or Pc3. However, the same trend was observed; combinations with the three mbt repeat alleles produced more severe wing-to-haltere and extra sex comb transformations compared to the H1 and M56 null alleles. Thus, in tests for genetic enhancement employing different PcG genes and different Pc alleles, we found that three mbt repeat alleles of Scm consistently produced the strongest interactions.

Interactions of Scm alleles with zeste:
Mutations in a subset of PcG genes, including Scm, have been shown to modify eye color in flies bearing the zeste1 (z1) mutation (WU et al. 1989 Down; JONES and GELBART 1990 Down). Indeed, these workers isolated the ScmSu(z)302 allele as a dominant suppressor of z1. The ScmXF24 null allele and a deficiency for the Scm locus are also dominant suppressors of z1 eye color (WU et al. 1989 Down; BORNEMANN et al. 1996 Down), although their effects are weaker than seen with Su(z)302. To further this analysis, we tested each of the additional Scm alleles (Table 1) in combination with the z1 wis tester chromosome (WU et al. 1989 Down). We found that, except for Su(z)302, heterozygosity for each Scm allele converts eye color in z1 wis males from light-orange to red-orange (Table 1). Su(z)302 is unique in producing much stronger suppression, which is manifested by dark red eye color in this combination (WU et al. 1989 Down). Since Su(z)302 is a missense mutation in mbt repeat 1 (Figure 5), this difference may reflect a role for mbt repeats in molecular interactions that contribute to zeste1 suppression.

mbt repeat mutants produce stable SCM protein that associates with normal chromosomal target loci:
The Su(z)302, R5-13, and ET50 molecular lesions (Figure 5) suggest that these strongly-interacting alleles encode altered SCM proteins that are essentially full-length. To test for production of these mutant proteins, extracts were prepared from Scm mutant pupae and Western blots with anti-SCM antibody were performed. Figure 7 shows that homozygous Su(z)302, homozygous R5-13, and hemizygous ET50 pupae express full-length versions of SCM protein at levels similar to wild type (lanes 1, 2, 4, and 5). Since these mutant pupae are derived from heterozygous Scm mutant mothers, it is conceivable that the signals could reflect maternal protein that has perdured to pupal stages. However, the increase in wild-type SCM levels during development from larval to pupal stages (Figure 2), indicates that a substantial portion of pupal SCM is newly synthesized protein. In addition, reduced signal was reproducibly seen with extracts from homozygous M36 mutant pupae (lane 3); this indicates that maternal product, if present, does not compromise detection of reduced zygotic SCM levels at this stage. Thus, the Su(z)302, R5-13, and ET50 mutants accumulate SCM proteins at levels comparable to wild type.



View larger version (66K):
In this window
In a new window
Download PPT slide
  		Figure 7. Accumulation of SCM mutant proteins in pupae. (Top) Western blot using anti-SCM antibodies. Lanes contain protein extracts from ScmR5-13 homozygous (lane 1), ScmSu(z)302 homozygous (lane 2), ScmM36 homozygous (lane 3), ScmET50 hemizygous (lane 4), or wild-type (lane 5) pupae. ScmP12 is a deletion that removes Scm (BORNEMANN et al. 1996 Down). (Bottom) Immunodetection of ß-tubulin on the same blot as a control for lane loading.

To assess whether these SCM mutant proteins associate with target sites in vivo, we used anti-SCM antibodies to stain polytene chromosomes from larvae homozygous for the Su(z)302 or R5-13 alleles. Figure 8A shows the wild-type distribution of SCM protein on a portion of the third chromosome that includes the bithorax complex (BX-C; arrow). Five sites of SCM protein accumulation are apparent. Figure 8B and Figure C show that these sites still accumulate SCM protein encoded by the Su(z)302 and R5-13 alleles. The number of staining sites per genome and the signal intensities were similar for these two mutants and wild type. Figure 8D shows that the same result is obtained if both the maternal and zygotic contributions consist of Su(z)302 mutant SCM protein. Thus, the signals are not due to perdurance of wild-type maternal SCM and must reflect the chromosome-binding properties of the mutant protein.



View larger version (94K):
In this window
In a new window
Download PPT slide
  		Figure 8. Accumulation of SCM mutant proteins at sites on chromosomes. Portions of polytene chromosomes spanning the bithorax complex (BX-C, arrows) and immunostained with anti-SCM antibodies are shown. Chromosomes were obtained from (A) wild-type larvae, (B) homozygous ScmR5-13 larvae from heterozygous mothers, (C) homozygous ScmSu(z)302 larvae from heterozygous mothers, or (D) homozygous ScmSu(z)302 larvae from ScmSu(z)302 germline clone mothers. Homozygous mutant larvae were recognized as non-Tb progeny of heterozygous parents containing a Tb-marked balancer.

Su(z)302 mutant SCM protein is sufficient for embryonic but not pupal development:
The homeotic gene misexpression and embryonic lethality seen with Scm null alleles (Figure 4B and Table 1) show that SCM protein is required during embryogenesis. In contrast, expression of ABDA protein in homozygous mutant ScmSu(z)302 embryos appears normal (Table 1), and these animals survive most of pupal development to die as pharate adults with homeotic phenotypes (WU et al. 1989 Down). Thus, Su(z)302 mutant protein appears defective in homeotic repression in pupae. Whether Su(z)302 protein is also defective in embryonic processes is unclear because maternally provided, wild-type SCM product is present in homozygous Su(z)302 embryos. To address this, we used the dominant female sterile technique and FLP recombination system (see MATERIALS AND METHODS) to generate Su(z)302 mutant embryos from Su(z)302 germline clone mothers. As shown in Figure 4D, embryos that express solely the Su(z)302 form of SCM protein from maternal and zygotic sources still show normal patterns of ABDA expression. Like Su(z)302 homozygotes from Su(z)302/+ mothers, these animals survive to pupal stages. We conclude that Su(z)302 mutant protein provides sufficient Scm function for apparently normal embryogenesis but that it is compromised in requirements for pupal development.


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

Functional domains in SCM protein:
The SCM protein contains multiple homology domains, including an N-terminal pair of zinc fingers, two copies of the mbt repeat, and the C-terminal SPM domain (Figure 5B; BORNEMANN et al. 1996 Down). The SPM domain has been shown to mediate protein interaction in vitro between the SCM and PH proteins and also between SCM and itself (PETERSON et al. 1997 Down). Although it is not yet clear how SPM domain interactions contribute to PcG function in vivo, these properties suggest a role in PcG complex assembly or stabilization.

Little is known about the biochemical role of mbt repeats either in SCM or in the L(3)MBT fly tumor suppressor protein (WISMAR et al. 1995 Down). However, the fact that four Scm alleles cause alterations within the first mbt repeat (Figure 5B and Table 1) shows that this domain contributes to Scm function in vivo. Several of these hypomorphic alleles produce wild-type levels of SCM protein (Figure 7) that associates with normal sites in polytene chromosomes (Figure 8). This suggests that mbt repeats are not crucial for targeting SCM protein to specific loci but, rather, that they provide a biochemical activity required at the normal location in chromatin. The repeats could provide catalytic activity or a protein interaction surface needed to localize or bind another partner protein. The conservation of the two mbt repeats with 69% identity in a mouse SCM homolog (F. RANDAZZO, personal communication) implies that their biochemical role is key for Scm function. Thus, the SPM domain and the mbt repeats are two distinct functional domains in SCM protein.

Our analysis of Scm alleles did not identify molecular lesions that affect the N-terminal SCM zinc fingers. Their possible role in Scm function is unclear because they do not belong to canonical DNA-binding zinc finger classes (BORNEMANN et al. 1996 Down) and in vitro assays have not detected sequence-specific DNA-binding by SCM protein (D. BORNEMANN and J. SIMON, unpublished data). Further studies, including site-directed mutagenesis, will be needed to assess the role of the N-terminal fingers. In contrast, we did find a mutation within a third potential zinc finger of SCM. This mutation, ScmM36, changes cysteine-511 to tyrosine (Figure 5B) and it appears to compromise protein stability, because M36 protein levels are reduced compared to wild type (Figure 1 and Figure 7).

Genetic interactions and PcG complexes:
Mutations in different PcG genes often produce phenotypic enhancement in double mutant combination (JURGENS 1985 Down; ADLER et al. 1989 Down; CHENG et al. 1994 Down). These genetic interactions have been interpreted in the context of PcG repressors that work together in multiprotein complexes. In this study, we compared how heterozygosity for different Scm mutations affects PcG function in a situation already compromised by heterozygosity for another PcG gene. Thus, the concentration or effectiveness of PcG complexes is brought near some critical threshold by the first PcG mutation, and we have assessed the impact of additional damage inflicted by changes in the SCM component.

The enhancement could result from two types of further alterations to PcG complexes. First, the Scm mutations could simply eliminate half of the SCM available to form complexes, resulting in haploinsufficiency for PcG complex function. Second, the mutations could produce stable, defective SCM proteins that incorporate into complexes and impair their activities. This type of dominant negative mechanism of inhibition has been described for many proteins that act in complexes (HERSKOWITZ 1987 Down; KHAVARI et al. 1993 Down). Consistent with their molecular lesions (Figure 5), we found that some Scm mutations, such as M56, fail to produce detectable protein (Figure 1). Because these Scm null mutations do enhance mutations in other PcG genes (Figure 6), the enhancement mechanism in these cases appears to involve a haploinsufficiency effect.

However, the strongest genetic interactions were consistently seen with three Scm alleles that are not nulls. These alleles, Su(z)302, ET50, and R5-13, alter the first mbt repeat and produce stable SCM proteins. The accumulation of protein encoded by two of these alleles at normal sites in chromatin (Figure 8) strongly suggests that the altered proteins incorporate into PcG complexes. Assembly of the mutant proteins into complexes could occur through protein contacts supplied by the SPM domain (PETERSON et al. 1997 Down), which remains intact in the mbt repeat alleles (Figure 5). In contrast to the null alleles, we suggest that the mbt repeat mutations produce especially strong interactions by poisoning PcG complexes. A dominant negative effect is also consistent with phenotypes reported for these alleles in heterozygous single mutants; WU et al. 1989 Down found that more than 85% of Su(z)302/+ or ET50/+ adults bear ectopic sex comb teeth as compared to only 13% of adults heterozygous for the XF24 null allele.

The strong interactions seen in animals heterozygous for mbt repeat alleles contrast with the relatively mild loss-of-function seen in mbt repeat allele homozygotes (Figure 4 and Table 1). Most strikingly, embryos possessing solely the Su(z)302 form of SCM protein appear to develop normally (Figure 4D), and they survive to pupal stages. This indicates that PcG complexes containing Su(z)302 protein retain substantial function, at least in embryos. We suggest that the mbt repeat alleles encode partially functional SCM proteins that are sufficiently compromised that they exert dominant negative effects under conditions that are particularly sensitive to PcG function. We note that Su(z)302, ET50, and R5-13 were isolated for dominant effects upon adult phenotypes that are sensitive indicators of PcG function (JURGENS 1985 Down; WU et al. 1989 Down; M. MULLER and H. GYURKOVICS, personal communication). Thus, the special properties of these mbt repeat alleles probably reflect phenotypic requirements imposed by the screens employed.

Regulatory targets of SCM protein:
The accumulation of PcG proteins, including SCM, at approximately 100 sites on polytene chromosomes (FRANKE et al. 1992 Down; RASTELLI et al. 1993 Down; LONIE et al. 1994 Down; PETERSON et al. 1997 Down) reveals a much larger number of PcG targets than just the two homeotic loci. Indeed, roles for PcG proteins in control of segmentation gene expression (MOAZED and O'FARRELL 1992 Down; PELEGRI and LEHMANN 1994 Down) and in dorsal-ventral patterning (TIONG et al. 1995 Down) have been described. The presence of SCM protein in adult heads (Figure 2B) presumably also reflects a role besides control of homeotic genes.

The synthetic lethality seen with strongly interacting Scm-PcG double mutant combinations (Figure 6) could be explained by effects upon either homeotic or nonhomeotic target genes. We found that surviving adults from the synthetic lethal combinations also showed the most extreme homeotic transformations. However, the lethal frequencies are comparable in Scm/Pc3 and Scm/Pc3 DfUbx109 animals (J. SIMON, unpublished results). If lethality were due to misexpression of either Ubx or abdA, which are removed by DfUbx109 (KARCH et al. 1985 Down), then the lower dosage of these products might be expected to reduce the lethal frequency. This suggests that the lethality is either associated with a homeotic gene besides Ubx or abdA, or with a nonhomeotic PcG target.

A striking outcome of the Scm-PcG double mutant analysis was that the frequency of synthetic lethality was much higher among female transheterozygotes than among males (Figure 6). Similarly, expressivity of antenna-to-leg and wing-to-haltere transformations among Pc3/+ and Pc3/+; Pcl1/+ adults is greater in females than in males (DUNCAN 1982 Down). We envision several possible explanations for more severe effects of PcG loss-of-function in females vs. males. First, the extra euchromatin present in XX vs. XY animals might sequester PcG proteins away from one or several of the normal euchromatic target sites. Under conditions where PcG function is already compromised, the need to assemble PcG proteins at sites on two X chromosomes rather than one might tip the balance below a critical threshold for function at other sites in the genome. In this titration model, sex-biased lethality could result from defects in homeotic gene control or in control of other target loci. A second possible explanation is that PcG proteins might function directly in dosage compensation to repress global expression from the X chromosome in females. This seems unlikely, however, because PcG proteins including SCM are only present at about 10 sites on polytene X chromosomes (FRANKE et al. 1992 Down; PETERSON et al. 1997 Down) and because Drosophila dosage compensation relies upon transcriptional up-regulation in males rather than down-regulation in females (BASHAW and BAKER 1996 Down for review). A third possibility is that the sex bias reflects events at one or more specific X-linked loci. In this scenario, the lethality could result from misexpression of a specific X-linked gene that occurs when PcG repressors are compromised. In females, the extra copy of this locus might increase misexpression to levels that more frequently exceed a threshold leading to inviability. In the context of this discussion, we note that phenotypes in Caenorhabditis elegans mutant for PcG homologs are more severe in XX than in XO animals. GARVIN et al. 1998 Down have shown that this difference depends on X chromosome dosage rather than on sexual phenotype. Further studies should identify which PcG target loci are responsible for differential consequences of PcG loss-of-function between the two sexes.


*  ACKNOWLEDGMENTS

We have benefited from the generosity of many who have isolated Scm mutations and shared them with us; we thank Ting Wu, Ian Duncan, Henrik Gyurkovics, Rick Jones, Gerd Jürgens, Jim Kennison, Gustavo Maroni, Kathy Matthews, and Martin Muller for these contributions. We are especially grateful to Ting Wu for sharing information and providing advice about several of the Scm alleles. We thank Hugh Brock, Jim Kennison, Martha Soto, and Kathy Matthews at the Indiana Stock Center for sending fly stocks and Tom Hays for suggestions on preparing protein extracts from larvae. We thank Susan Strome for communicating results before publication. We thank Bob Herman, Mike O'Connor, and Mike Simmons for comments on the manuscript. We thank Genelle Belmas for suggestions on figure design. This work was supported by National Institutes of Health (NIH) grant GM49850 to J.S., and D.B. was supported in part by NIH training grant HD07480.

Manuscript received February 13, 1998; Accepted for publication June 19, 1998.


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

ADLER, P. N., J. CHARLTON, and B. BRUNK, 1989  Genetic interactions of the Suppressor 2 of zeste region genes. Dev. Genet. 10:249-260[Medline].

ALKEMA, M. J., M. BRONK, E. VERHOEVEN, A. OTTE, and L. J. VAN'T VEER et al., 1997  Identification of Bmi1-interacting proteins as constituents of a multimeric mammalian Polycomb complex. Genes Dev. 11:226-240[Abstract/Free Full Text].

BASHAW, G. J. and B. S. BAKER, 1996  Dosage compensation and chromatin structure in Drosophila. Curr. Opin. Genet. Dev. 6:496-501[Medline].

BORNEMANN, D., E. MILLER, and J. SIMON, 1996  The Drosophila Polycomb group gene Sex comb on midleg Scm encodes a zinc finger protein with similarity to polyhomeotic protein. Development 122:1621-1630[Abstract].

BREEN, T. R. and I. M. DUNCAN, 1986  Maternal expression of genes that regulate the bithorax complex of Drosophila melanogaster. Dev. Biol. 118:442-456[Medline].

CARRINGTON, E. C. and R. S. JONES, 1996  The Drosophila Enhancer of zeste gene encodes a chromosomal protein: examination of wild-type and mutant protein distribution. Development 122:4073-4083[Abstract].

CELNIKER, S. E., D. J. KEELAN, and E. B. LEWIS, 1989  The molecular genetics of the bithorax complex of Drosophila: characterization of the products of the Abdominal-B domain. Genes Dev. 3:1424-1436[Abstract/Free Full Text].

CHENG, N. N., D. A. R. SINCLAIR, R. B. CAMPBELL, and H. W. BROCK, 1994  Interactions of polyhomeotic with Polycomb group genes of Drosophila melanogaster. Genetics 138:1151-1162[Abstract].

CHOU, T.-B., E. NOLL, and N. PERRIMON, 1993  Autosomal P[ovoD1] dominant female-sterile insertions in Drosophila and their use in generating germ-line chimeras. Development 119:1359-1369[Abstract].

DECAMILLIS, M. and H. W. BROCK, 1994  Expression of the polyhomeotic locus in development of Drosophila melanogaster. Roux's Arch. Dev. Biol. 203:429-438.

DUNCAN, I., 1982  Polycomblike: a gene that appears to be required for the normal expression of the bithorax and Antennapedia gene complexes of Drosophila melanogaster. Genetics 102:49-70[Abstract/Free Full Text].

DUNCAN, I., and E. B. LEWIS, 1982 Genetic control of body segment differentiation in Drosophila, pp. 533–554 in Developmental Order: Its Origin and Regulation, edited by S. SUBTELNY. Alan R. Liss, Inc., New York.

DURA, J.-M., N. B. RANDSHOLT, J. DEATRICK, I. ERK, and P. SANTAMARIA et al., 1987  A complex genetic locus, polyhomeotic, is required for segmental specification and epidermal development in D. melanogaster. Cell 51:829-839[Medline].

FRANKE, A., M. DECAMILLIS, D. ZINK, N. CHENG, and H. W. BROCK et al., 1992  Polycomb and polyhomeotic are constitutents of a multimeric protein complex in chromatin of Drosophila melanogaster. EMBO J. 11:2941-2950[Medline].

FRANKE, A., S. MESSMER, and R. PARO, 1995  Mapping functional domains of the Polycomb protein of Drosophila melanogaster. Chromosome Res. 3:351-360[Medline].

GARVIN, C., R. HOLDEMAN, and S. STROME, 1998  The phenotype of mes-2, mes-3, mes-4, and mes-6, maternal-effect genes required for survival of the germline in C. elegans, is sensitive to chromosome dosage. Genetics 148:167-185[Abstract/Free Full Text].

GOLIC, K. and S. LINDQUIST, 1989  The FLP recombinase of yeast catalyzes site-specific recombination in the Drosophila genome. Cell 59:499-509[Medline].

GUNSTER, M. J., D. P. E. SATIJIN, K. M. HAMER, J. L. DEN BLAAUWEN, and D. DE BRUIJN et al., 1997  Identification and characterization of interactions between the vertebrate Polycomb-group protein bmi1 and human homologs of polyhomeotic. Mol. Cell. Biol. 17:2326-2335[Abstract].

HERSKOWITZ, I., 1987  Functional inactivation of genes by dominant negative mutations. Nature 329:219-222[Medline].

JONES, R. S. and W. M. GELBART, 1990  Genetic analysis of the Enhancer of zeste locus and its role in gene regulation in Drosophila melanogaster. Genetics 126:185-199[Abstract].

JONES, C. A., J. NG, A. J. PETERSON, K. MORGAN, J. SIMON, and R. S. JONES, 1998  The Drosophila esc and Ez proteins are direct partners in Polycomb-group-mediated repression. Mol. Cell. Biol. 18:2825-2834[Abstract/Free Full Text].

JÜRGENS, G., 1985  A group of genes controlling the spatial expression of the bithorax complex in Drosophila. Nature 316:153-155.

KARCH, F., B. WEIFFENBACH, M. PEIFER, W. BENDER, and I. DUNCAN et al., 1985  The abdominal region of the bithorax complex. Cell 43:81-96[Medline].

KARCH, F., W. BENDER, and B. WEIFFENBACH, 1990  abdA expression in Drosophila embryos. Genes Dev. 4:1573-1587[Abstract/Free Full Text].

KAUFMAN, T. C., R. A. LEWIS, and B. T. WAKIMOTO, 1980  Cytogenetic analysis of chromosome 3 in Drosophila melanogaster: the homeotic gene complex in polytene chromosome interval 84A-B. Genetics 94:115-133[Abstract/Free Full Text].

KEMPHUES, K. J., E. C. RAFF, R. A. RAFF, and T. C. KAUFMAN, 1980  Mutation in a testis-specific ß-tubulin in Drosophila: analysis of its effects on meiosis and map location of the gene. Cell 21:445-451[Medline].

KENNISON, J. A. and J. W. TAMKUN, 1988  Dosage-dependent modifiers of Polycomb and Antennapedia mutations in Drosophila. Proc. Natl. Acad. Sci. USA 85:8136-8140[Abstract/Free Full Text].

KHAVARI, P. A., C. L. PETERSON, J. W. TAMKUN, D. B. MENDEL, and D. B. CRABTREE, 1993  BRG1 contains a conserved domain of the SWI2/SNF2 family necessary for normal mitotic growth and transcription. Nature 366:170-174[Medline].

LEWIS, E. B., 1978  A gene complex controlling segmentation in Drosophila. Nature 276:565-570[Medline].

LI, M.-G., M. MCGRAIL, M. SERR, and T. S. HAYS, 1994  Drosophila cytoplasmic dynein, a microtubule motor that is asymmetrically localized in the oocyte. J. Cell Biol. 126:1475-1494[Abstract/Free Full Text].

LINDSLEY, D. L., and G. G. ZIMM, 1992 The Genome of Drosophila melanogaster. Academic Press, San Diego.

LONIE, A., R. D'ANDREA, R. PARO, and R. SAINT, 1994  Molecular characterization of the Polycomblike gene of Drosophila melanogaster, a trans-acting negative regulator of homeotic gene expression. Development 120:2629-2636[Abstract/Free Full Text].

MARTIN, E. C. and P. N. ADLER, 1993  The Polycomb group gene Posterior sex combs encodes a chromosomal protein. Development 117:641-655[Abstract].

MCKEON, J. and H. W. BROCK, 1991  Interactions of the Polycomb group of genes with homeotic loci of Drosophila. Roux's Arch. Dev. Biol. 199:387-396.

MESSMER, S., A. FRANKE, and R. PARO, 1992  Analysis of the functional role of the Polycomb chromodomain in Drosophila melanogaster. Genes Dev. 6:1241-