The Drosophila Gene taranis Encodes a Novel Trithorax Group Member Potentially Linked to the Cell Cycle Regulatory Apparatus

Corresponding author: Henri-Marc Bourbon, UMR5547 du CNRS, Université Paul Sabatier, 118 Route de Narbonne, 31062 Toulouse, France., bourbon{at}cict.fr (E-mail)

Communicating editor: T. C. KAUFMAN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Genes of the Drosophila Polycomb and trithorax groups (PcG and trxG, respectively) influence gene expression by modulating chromatin structure. Segmental expression of homeotic loci (HOM) initiated in early embryogenesis is maintained by a balance of antagonistic PcG (repressor) and trxG (activator) activities. Here we identify a novel trxG family member, taranis (tara), on the basis of the following criteria: (i) tara loss-of-function mutations act as genetic antagonists of the PcG genes Polycomb and polyhomeotic and (ii) they enhance the phenotypic effects of mutations in the trxG genes trithorax (trx), brahma (brm), and osa. In addition, reduced tara activity can mimic homeotic loss-of-function phenotypes, as is often the case for trxG genes. tara encodes two closely related 96-kD protein isoforms (TARA-/-ß) derived from broadly expressed alternative promoters. Genetic and phenotypic rescue experiments indicate that the TARA-/-ß proteins are functionally redundant. The TARA proteins share evolutionarily conserved motifs with several recently characterized mammalian nuclear proteins, including the cyclin-dependent kinase regulator TRIP-Br1/p34^SEI-1, the related protein TRIP-Br2/Y127, and RBT1, a partner of replication protein A. These data raise the possibility that TARA-/-ß play a role in integrating chromatin structure with cell cycle regulation.

SPECIFICATION of segmental identities along the anterior-posterior (A-P) body axis in Drosophila is controlled by the spatially restricted expression of the homeotic loci (HOM) of the Antennapedia and bithorax complexes (ANT-C and BX-C; LEWIS 1978 ; AKAM 1987 ; MCGINNIS and KRUMLAUF 1992 ). The importance of precise transcriptional control throughout development is indicated by the spectacular alterations in cell fate resulting from inappropriate HOM expression, such as transformations of antenna to leg or haltere to wing, or by more discrete changes in cell identity, such as the specialized bristles of the male sex comb. The restricted A-P expression domains of the HOM genes initiated during embryogenesis by the localized transient activities of the segmentation genes (DUNCAN 1986 ; HARDING and LEVINE 1988 ; IRISH et al. 1989 ) are subsequently maintained by the antagonistic activities of two families of trans-regulator genes, the Polycomb and trithorax group loci (PcG and trxG, respectively). The PcG genes sustain a repressed state of HOM expression, while trxG genes favor an active state (JURGENS 1985 ; WEDEEN et al. 1986 ; DURA and INGHAM 1988 ; KENNISON and TAMKUN 1992 ; SIMON et al. 1992 ). While the PcG proteins all appear to repress transcription, the trxG products maintain the expression of key developmental control genes, including the homeotic loci, acting at different levels of gene regulation (KENNISON 1995 ; SIMON 1995 ; PIRROTTA 1997 ). Recently, several PcG loci whose mutations behave as enhancers of both trxG and PcG mutations have been grouped as a third family [Enhancers of trxG and PcG: the ETP group (GILDEA et al. 2000 )]. Null mutations in most PcG and trxG genes are recessive lethal, but phenotypes may be observed in heterozygous flies as a consequence of HOM misexpression. Loss-of-function alleles of many trxG genes enhance mutations in other trxG members and suppress homeotic transformations resulting from loss-of-function mutations in bona fide PcG loci (CAPDEVILA and GARCIA-BELLIDO 1981 ; INGHAM 1983 ; SHEARN 1989 ). Hence, most members of the trxG family were identified as dominant suppressors of phenotypes caused by misexpression of homeotic genes (KENNISON and TAMKUN 1988 ).

Given the structural and biochemical diversity observed for trxG genes (FRANCIS and KINGSTON 2001 ), two operational criteria are widely used to define members of this group (SHEARN 1989 ). First, they are functional antagonists of PcG loci. Second, they interact synergistically with other trxG genes. In addition, phenotypic manifestations of these interactions often include homeotic transformations.

Although trxG homologs have been identified in a variety of organisms (GOULD 1997 ; CHAMBERLIN and THOMAS 2000 ), their mechanism of action is not clearly understood. Genetic and molecular studies have suggested that some Drosophila trxG proteins may interact physically at the level of chromatin to regulate the transcription of target genes, including the HOM loci (COLLINS et al. 1999 ; VAZQUEZ et al. 1999 ; KAL et al. 2000 ). One of the best-characterized trxG proteins, BRAHMA (BRM), is a homolog of yeast SWI2/SNF2, a bromodomain-containing DNA-stimulated ATPase (TAMKUN et al. 1992 ). BRM is found within a large (2 MD) dSWI/SNF protein complex that is thought to increase target gene accessibility by using the energy of ATP hydrolysis to overcome the repressive effects of nucleosomal histones and imposing an active chromatin state (PAPOULAS et al. 1998 ; KINGSTON and NARLIKAR 1999 ). How transcriptionally active or inactive chromosomal structures can be stably transmitted through multiple cell divisions in development is largely unknown (for reviews see LYKO and PARO 1999 ; FRANCIS and KINGSTON 2001 ) but represents a question of fundamental interest for better understanding cell differentiation and developmental control (PIRROTTA 1998 ; FARKAS et al. 2000 ). Recent data indicate that the acetylation status of histone H4 is one element of the epigenetic control resulting from PcG/trxG chromosomal imprints (CAVALLI and PARO 1999 ).

Unravelling this problem will be facilitated by the identification of the full complement of trxG proteins, permitting the study of their biochemical roles in the cell. It is not clear just how large the trxG family may be nor the full extent of molecular diversity among its members (KENNISON 1993 ). To identify functional partners of Drosophila HOM loci in directing normal development, we performed a genetic screen for mutations that act as dominant modifiers of phenotypes resulting from ectopic expression of the ANT-C gene proboscipedia [pb; Hox-A2/-B2 homolog (CRIBBS et al. 1995 ; BENASSAYAG et al. 1997 ; BOUBE et al. 1998 )]. One P-element-induced modifier mutation allowed us to isolate a previously unknown gene, taranis (tara), whose genetic properties identify it as a novel trxG member. Molecular analysis indicates that tara encodes two closely related 96-kD proteins (TARA-/-ß) comprising structural hallmarks of nuclear factors, as well as several evolutionarily conserved motifs. Recognizable mammalian counterparts include the transcriptional coactivator proteins TRIP-Br1/p34^SEI-1, a cyclin-dependent kinase regulator that also interacts with PHD zinc fingers and bromodomains (SUGIMOTO et al. 1999 ; HSU et al. 2001 ), and RBT1, a replication protein A-binding protein (CHO et al. 2000 ). We discuss the potential implications of these findings for the epigenetic maintenance of active chromosomal states through successive cell cycles.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Fly stocks and culture:
Drosophila melanogaster fly stocks were maintained on standard cornmeal/yeast/agar medium at 22°. For embryo or cuticle preparations, eggs were collected on apple juice/agar plates. Strains carrying trx^E2/TM6C Sb, osa¹/TM6C Sb, brm²/TM6C Sb, brm² trx^E2/TM6C Sb, Pc¹⁶/TM6 Ubx, or the homozygous viable ph⁴¹⁰ allele were obtained from J.-M. Dura. Unless otherwise noted, all mutations and chromosome aberrations are described in LINDSLEY and ZIMM 1992 . tara¹ is a P[lacW]-induced allele isolated by M.-O. Fauvarque and J.-M. Dura who named the corresponding gene taranis for a Celtic god (FAUVARQUE et al. 2001 ). Df(3R)sbd26 and Df(3R) sbd45 were from the Indiana University Drosophila Stock Center (IUDSC; Bloomington, IN).

Alleles of the tara gene:
The P[lacW] insertion mutagenesis was carried out with an X chromosome insertion line (P20) obtained from IUDSC (BIER et al. 1989 ) and isogenized for chromosomes 2 and 3 shortly before initiating the screen. New autosomal insertions were generated by mobilizing the X-linked P20 element with the 2-3 P transposase source (ROBERTSON et al. 1988 ) and selecting male progeny of dysgenic males carrying a transposed transgene copy. These males were crossed with females harboring a pb gain-of-function transgene, HSPB:4d, in search of insertions modifying the effects of ectopic PB (BOUBE et al. 1997 ). The tara^L4 insertion led to deletions of wing veins L4 and L5 specifically in combination with HSPB. In addition to the tara^L4 allele and the independently isolated lethal P[lacW] insertion tara¹ (FAUVARQUE et al. 2001 ), analysis of the recently completed Drosophila genome sequence (ADAMS et al. 2000 ) allowed us to identify 10 other P insertions within the tara locus. The latter contain insertions of an "Enhancer Pirate" (EP) element containing a yeast heterologous UAS promoter that allows modular misexpression (RORTH 1996 ). Their positions within the tara genomic sequence (GenBank accession no. AF227213) are the following: 1686 (EP3427), 5447 (EP1145), 11526 (EP3178), 11573 (EP3095), 17707 (EP1189), 18081 (EP508), 18450 (EP3463), 18466 (EP3294), 18476 (EP3530), and 30901 (EP693). The EP3427 and EP3463 lines contain an appropriately oriented UAS promoter inserted within the tara locus that affords overexpression of a determined TARA isoform in combination with a tissue-specific driver for the yeast transcription factor GAL4 (see ABDELILAH-SEYFRIED et al. 2000 ; FERNANDEZ-FUNEZ et al. 2000 ). To recover excision alleles, the P[lacW]-L4 element was mobilized by generating dysgenic males carrying the P[ry⁺, 2-3]99B Sb chromosome (ROBERTSON et al. 1988 ), crossing them with w^- females, and recovering w^- Sb⁺ animals. Of 120 w^- chromosomes recovered, 8 (R19, R38, R40, R47, R58, R75, R91, and R108) were lethal over Df(3R)sbd26 and tara^L4. Lethal phases were assessed by mating heterozygous tara^-/+ females and males (without balancer chromosomes). The parents were allowed to mate at 22° for several days to minimize the number of unfertilized eggs in the collections. The number of unhatched eggs was recorded 48 hr after the collection period.

Genetic interactions:
Interactions among various combinations of the alleles tara^L4, tara¹, pb⁴, pb¹³, pb⁵, Pc¹⁶, ph⁴¹⁰, trx^E2, osa¹, and brm² were examined after culture at 22°. Labial palp phenotypes of pb¹³/pb⁴ and pb¹³/pb⁴ tara^L4 were compared in independent blind tests by three persons. For the wings-held-out phenotype, individuals with both wings extended were scored as mutant.

Isolation of tara genomic and cDNA clones:
The P[lacW] element allows cloning of adjacent genomic sequences by plasmid rescue (BIER et al. 1989 ). One isolate, pL4, obtained by EcoRI digestion of flies bearing P[lacW]-L4, was ³²P-labeled (Megaprime kit; Amersham Pharmacia Biotech) and used to screen a genomic library of Oregon-R DNA made in charon4 bacteriophage (MANIATIS et al. 1978 ) using standard techniques (SAMBROOK et al. 1989 ). Two independent genomic clones were isolated and characterized by restriction mapping and Southern blot analysis. All EcoRI genomic fragments were subcloned into pBluescript II SK+ (Stratagene, La Jolla, CA). A pool of purified EcoRI inserts was prepared, ³²P-labeled, and used as a probe to screen a Canton-S 8- to 12-hr embryonic cDNA library (BROWN and KAFATOS 1988 ). Thirty positive inserts in the pNB40 plasmid were purified and characterized by restriction mapping. Sequencing was performed using dideoxy chain termination with a Sequenase 2.0 kit (Amersham Pharmacia Biotech). The complete double-stranded sequences of the C12 and C16 cDNAs were determined, using the SP6 and T7 primers as well as specific internal oligonucleotides (Isoprim, Toulouse, France; sequences available upon request). The GenBank accession numbers for the sequences reported in this study are AF227211, AF227212, and AF227213. Sequence similarity searches and clustering were performed at the National Center for Biotechnology Information using BLAST (http://www.ncbi.nlm.nih.gov/BLAST/) and at the Institut de Génétique et de Biologie Moléculaire et Cellulaire (Illkirch, France) using DbClustal (http://igbmc.u-strasbg.fr: 8080/DbClustal/dbclustal.html). The Drosophila protein and annotation databases were searched at the Berkeley Drosophila Genome Project (BDGP; http://www.fruitfly.org/).

In situ hybridization:
In situ hybridizations were carried out with digoxigenin-labeled RNA probes (Roche Molecular Biochemicals, Indianapolis) by the method of TAUTZ and PFEIFLE 1989 . To allow synthesis of antisense riboprobes by transcription from the T3 or T7 promoter, three fragments were each inserted into pBluescript II SK+ cut with the appropriate enzymes: a 2.3-kb XhoI/KpnI genomic fragment (probe C), a 0.75-kb ApaI/EcoRI fragment from C12 cDNA (probe A), or a 1.5-kb EcoRI/KpnI genomic fragment (probe B). Following color development, the embryos or larval tissues were mounted in PBS/50% glycerol, viewed, and photographed with a Zeiss Axiophot microscope.

P-element-mediated germline transformation and tests for rescuing tara lethality:
The pUbTARA-ß construct used for rescue tests was made as follows. A 3.3-kb MluI/XmnI cDNA fragment from the pNB-C16 plasmid was inserted into the pUbHB1 vector between the MluI and EcoRV polylinker sites to yield the pUbC16 construct. The pUbHB1 plasmid contains a polylinker placed downstream of the ubiquitin-63E promoter [2-kb SalI/BglII fragment (LEE et al. 1988 )] and upstream of the hsp70 3'-untranslated and 3'-flanking genomic regions [0.5-kb BamHI fragment from pCaSpeR-hs (THUMMEL and PIRROTTA 1992 )]. The map of the pUbHB1 vector is available upon request. Finally, the pUbTARA-ß plasmid was generated by insertion of a 5.9-kb XmnI/NotI fragment from pUbC16 into pCaSpeR4 (THUMMEL and PIRROTTA 1992 ) between the StuI and NotI polylinker sites. w¹¹¹⁸ embryos were co-injected with pUbTARA-ß and pUChs2-3 (MULLINS et al. 1989 ) by standard techniques (RUBIN and SPRADLING 1982 ). Emerging adults were crossed individually to w¹¹¹⁸ flies, and w⁺ transformant progeny were identified. Five independent lines were recovered, chromosomal linkage was determined by crosses with a multiple-balancer stock, and homozygous stocks were established. Four transformant lines carrying the P[UbTARA-ß] construct on the second chromosome (lines 2, 3, 4, and 5) were tested for rescue of lethality in tara^L4 homozogotes, in the following way: w¹¹¹⁸; P[UbTARA-ß]/+; tara^L4/TM3, Sb e males were generated and mated to w¹¹¹⁸; tara^L4/TM6B, Hu Tb e females. Rescue was assessed by scoring for surviving progeny with the genotype w¹¹¹⁸; P[UbTARA-ß]/+; tara^L4/tara^L4. All rescued tara mutant individuals (non-Stubble, non-Humeral, e⁺) were mini-w+.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

tara is an essential locus that modulates pb homeotic selector activity:
The homeotic locus pb is a genetic selector required for adult mouthparts (PULTZ et al. 1988 ). Basal ectopic expression of PB protein from a Hsp70-pb (HSPB) transgene, including pb transcriptional regulatory sequences, leads to discrete dose-sensitive adult phenotypes in the wing, the eye, and the prothoracic leg (CRIBBS et al. 1995 ; BOUBE et al. 1997 ). The HSPB minigene thus facilitates genetic screens for dominant modifiers (cf. BOTAS et al. 1982 ; DUNCAN 1982 ) that we sought by insertional mutagenesis with a marked P element [P[lacW]; BIER et al. 1989 ]. Among 5000 independent males tested carrying mutagenized autosomes, one new P[lacW] insertion was associated with partial truncation of the distal part of the longitudinal wing vein L4 in heterozygous combination with the HSPB:4d transgenic line (penetrance of 60%; Fig 1B). This phenotype depends on both HSPB and the P[lacW] insert since it is detected in heterozygous combination with several HSPB lines, but not for HSPB nor for insertion heterozygotes alone (not shown). The chromosome harboring the interacting insertion, P[lacW]-L4, situated on chromosome 3R at 89B13-16, is homozygous lethal. This recessive lethality of the P[lacW]-L4 line and its interaction with ectopic PB were reverted by excision of the marked P element (see below), identifying the interacting locus as an essential gene. This locus, likewise isolated in an independent screen, was tara (FAUVARQUE et al. 2001 ).

View larger version (59K): In this window In a new window Download PPT slide   Figure 1. Adult phenotypes associated with tara mutants. (A and C) Wild-type wing blade and posture, respectively, in w1118 adults. (B) The partial deletion of the L4 wing vein (arrow) in a w1118; HSPB:4d +/+ taraL4 fly. This phenotype is specific to the combination. (D) Held-out wings in a female homozygous for the semiviable allele taraEP3463. (E) The labial palps of this pb4/pb13 fly are nearly normal (with a slight reduction of the pseudotracheal rows). (F) Reduced tara function leads to a discrete, limited transformation of the distal labium to antennal arista (arrows) in a pb4 taraL4/pb13 + adult. This phenotype is typical of reduced pb function. (G and H) Embryonic lethality of taraL4 homozygotes was rescued in flies harboring one (G) or two (H) copies of the transgene P[UbTARA-ß] (line 3; see MATERIALS AND METHODS), regardless of gender. Note the wings-held-out phenotype in G, similar to the tara hypomorph in D. This defect is rectified on increasing TARA-ß, as seen in H. (I and J) Wild-type antenna and haltere, respectively, in w1118 adults. (K and L) Apart from the wings-held-out phenotype mentioned above, other developmental defects were observed in taraL4 homozygotes harboring a single copy of the P[UbTARA-ß]#3 transgene. The thickened antennal arista shown in K resembles a weak antenna-to-leg transformation (arrow), while the haltere defect (L) is interpreted as a nascent transformation to wing margin, as indicated by the row of bristles (arrow).

The initial insertion allele tara^L4 was examined in homozygotes or in hemizygous combination with the chromosomal deficiencies Df(3R)sbd26 and Df(3R)sbd45 covering the 89B13-16 region. Homozygous or hemizygous animals died as late embryos or as first instar larvae without obvious cuticular pattern defects in all examined allelic combinations (Table 1). tara^L4 thus behaves genetically as a strong or complete loss-of-function (lof) mutation. To recover additional alleles by imprecise excision, the P[lacW]-L4 insertion was mobilized utilizing the 2-3 P-transposase source (ROBERTSON et al. 1988 ). Among 120 recovered white^- revertants, 110 were fully viable without apparent adult phenotypes (presumptive precise excisions), 8 were homozygous lethal (see MATERIALS AND METHODS), and 2 (tara^R19 and tara^R51) led to adult escapers with prominent wing phenotypes including balloon and/or held-out wings along with L5 vein truncation (see Table 1 and Fig 1D). The new lethal alleles failed to complement the parental allele, although the lethal phases ranged from first instar larva to pupa (Table 1 and data not shown). No excision alleles were recovered showing fully penetrant embryonic lethality either as homozygotes or as hemizygotes.

To examine whether the observed dose-sensitive interaction of tara^L4 with HSPB in the wing reflects a relationship in the normal lieu of pb action, the adult mouthparts, we tested for a dominant interaction between tara^L4 and different pb alleles. The allelic combination pb⁴/pb¹³ leads to reduced labial pseudotracheae, but without an overt homeotic transformation (CRIBBS et al. 1992 ). In contrast, pb⁴ tara^L4/pb¹³ + flies showed a weak but reliable distal labium-to-antenna transformation (Fig 1, compare E and F), indicating that reduced tara activity diminishes normal pb function. Similarly, the allelic combination pb⁴/pb⁵ transforms distal labium to antennal arista, and distal claws typical of a leg (a more severe loss of function) are not observed in a tara⁺ context. By contrast, claws were observed in nearly half of tara^L4 heterozygotes (not shown). We infer from these interactions that tara acts to positively regulate pb selector activity.

tara suppresses PcG mutations:
An independent genetic screen for suppressors of the Polycomb group gene polyhomeotic (ph) led to the isolation of another recessive lethal P[lacW] insertion allele in taranis (FAUVARQUE et al. 2001 ). This allele, tara¹, was identified on the basis of its dominant suppression of the extra-sex-combs phenotype due to the hypomorphic ph⁴¹⁰ allele (DURA et al. 1985 ). tara¹ also suppresses the dominant extra-sex-combs phenotype caused by the amorphic Polycomb allele Pc¹⁶ (KENNISON and TAMKUN 1988 ). Complementation tests established that tara^L4 and tara¹ are allelic. The former showed a stronger interaction with Pc¹⁶, ph⁴¹⁰ (Table 2), or HSPB:4d (data not shown). Our molecular analysis indicated that the tara^L4 and tara¹ insertions are situated 0.28 kb apart within an intron shared by two nested transcription units (see below and Fig 2).

View larger version (19K): In this window In a new window Download PPT slide   Figure 2. Molecular analysis of the tara locus. (A) Physical map of the tara region on chromosome 3R, cytological interval 89B13-16 (centromere, left; telomere, right). The restriction map of the P1 bacteriophage DS06428 (A, ApaI; K, KpnI) corresponds to nucleotides 85321–158160 of GenBank accession no. AE003712 containing the tara locus. The positions of the two -bacteriophage used to initiate the structural characterization of the locus (C5 and C13) are shown above the restriction map. The positions of the neighboring P[lacW] insertion mutations taraL4 and tara1 are indicated above the genomic map by inverted triangles. The 5'-to-3' orientation and the length of the two alternative transcription units that define the tara locus are shown below the DNA line, as well as the approximate location (not to scale) and orientation of the neighboring gene bor. tara exons are indicated as solid boxes for coding sequences and open boxes for untranslated mRNA sequences, respectively. The putative transcription initiation sites (arrows) have been inferred from the presence of an additional 5' G in the longest - or ß-class cDNAs. The positions of the P insertions EP3427 and EP3463 allowing overexpression of the TARA isoform or ß, respectively, are indicated by inverted triangles. The approximate positions and sizes of DNA fragments in A–C, employed to generate probes for Northern blots and whole-mount in situ hybridizations, are marked beneath the map. (B) Structure of representative cDNAs derived from the tara region. Shown are tara coding sequences (solid) and positions of the alternative initiator ATGs, as deduced from the apparent full-length cDNAs C12 (-class) and C16 (ß-class).

tara enhances mutations of trxG loci:
Interaction of tara loss-of-function alleles with pb and PcG mutations suggested that tara might encode a new trxG member. Apart from suppression of PcG phenotypes, synergistic interaction with mutations of established trxG genes is an additional criterion for classifying a gene as a trithorax group member (SHEARN 1989 ). Representatives include trithorax (trx), brahma (brm), and osa. The latter two strongly interact to give a fully penetrant wings-held-out phenotype that is thought to result from a failure to properly activate the Antp P2 promoter (VAZQUEZ et al. 1999 ). This phenotype strongly resembles that seen in tara hypomorphs (Fig 1D).

To test whether tara behaves like a member of the trxG family, combinations of the tara^L4 or tara¹ allele with strong or null alleles of trx, brm, or osa were examined specifically for wing phenotypes. Double heterozygotes for tara^L4 with trx^E2, brm², or osa¹ showed a wings-held-out phenotype not observed in single heterozygotes (Table 2). The penetrance of these interactions ranged from 2 to 36%, with the strongest expressivity observed in tara^L4 +/+ osa¹ heterozygotes. Weakly penetrant wing phenotypes were observed in + tara^L4/trx^E2 + or in brm² trx^E2/+ + double heterozygotes (2 and 5%, respectively). By contrast, on examining triple heterozygotes for tara, brm, and trx, 45% of flies of + + tara^L4/brm² trx^E2 + flies displayed held-out wings. Similar synergistic effects were seen with the tara¹ allele (see Table 2), ruling out that the genetic interaction between tara and trxG loci is due to a second-site mutation on the tara^L4 chromosome. Taken together, these results indicate that tara behaves genetically as expected for a member of the trxG family.

tara encodes two classes of mRNA:
The origin of tara^L4 as a P-element insertion allowed us to pursue the molecular analysis of the tara gene. A 1.5-kb fragment of genomic DNA flanking the P[lacW]-L4 element was cloned by plasmid rescue (WILSON et al. 1989 ) and then used to screen an Oregon-R genomic DNA library in the charon4 bacteriophage (see MATERIALS AND METHODS). Among the recombinant phages recovered, two overlapping clones (C5 and C13) spanned 30 kb of genomic sequence (Fig 2A). Subcloned fragments covering this genomic region were used as probes in Northern blot analysis to identify transcribed regions. Two embryonic polyadenylated RNA species of 2 and 4.5 kb were detected (not shown). The 2-kb mRNA, detected only with a probe from one end of the cloned region, corresponds to a novel gene encoding a putative novel member of the AAA family of ATPases that we named belphegor (bor; see Fig 2A; GenBank accession no. AF227209). In contrast, the 4.5-kb mRNA was detected with probes separated by 15 kb (probes B and C) and flanking the tara¹ and tara^L4 insertions (Fig 2A). These results strongly suggest that both P[lacW] insertions reside within an intron of a transcription unit giving rise to the 4.5-kb mRNA species. Consistent with this, a much weaker in situ hybridization signal was observed in tara^L4 homozygous embryos compared to wild type (not shown). Apart from indicating that the 4.5-kb mRNA corresponds to tara, these data also suggested that tara^L4 is not a molecular null allele.

To recover tara cDNAs, relevant C5 and C13 genomic fragments were used to probe a cDNA library from 8- to 12-hr embryos (see MATERIALS AND METHODS). Among 30 positive clones recovered, insert sizes ranged from 0.9 to 4.3 kb. Restriction mapping and partial sequence analysis indicated that all of the cDNA inserts had similar 3' extremities, ending in an A-rich region with oligo-adenylated tails. However, two classes of 5' extremities ( and ß) were detected among the longest cDNAs. All but one cDNA insert shared a common 5' region of the ß-class and differed only in length. The longest of these ß-type inserts, C16, contained a sequence of 4000 bp (Fig 3). The complete sequence of the single -type clone (C12) was 4316 bp, again in reasonable agreement with a fully polyadenylated mRNA of 4500 nucleotides (nt) detected by Northern blots. C12 was identical to C16 for most of its length, sharing a long open reading frame of 2724 bp (see Fig 2B and Fig 3). However, C12 and C16 diverged toward their 5' ends, containing distinct 5' mRNA untranslated regions of 767 and 499 bp, respectively (see Fig 3). The C12 cDNA thus appeared to represent an alternatively spliced mRNA, potentially derived from an independent transcriptional promoter.

View larger version (100K): In this window In a new window Download PPT slide   Figure 3. Sequence of the alternative tara mRNAs and proteins. The nucleotide sequence and deduced amino acid sequence for a putative full-length embryonic cDNA of each class (cf. Fig 2B) is shown. For the unique -class cDNA, only the alternative 5'-untranslated and flanking coding sequence is reported. The position of the alternative intron is indicated by an arrowhead. The running tallies for the nucleotides and aa sequences are shown on the right. Putative initiator ATGs are underlined, as are the four evolutionarily conserved motifs. The GenBank accession numbers for the C12 and C16 tara cDNA sequences are AF227211 and AF227212, respectively.

Comparison of the complete C12 and C16 cDNA sequences with newly available BDGP genomic sequences (P1 bacteriophage DS06428) allowed us to establish the following three points: (i) Both cDNAs are likely to be full length, since each contains a G nucleotide at its 5' end lacking in the genomic sequence that might correspond to a 7-methyl-G cap added post-transcriptionally (HIRZMANN et al. 1993 ); (ii) the sequences shared by C12 and C16 correspond to a single large 3' exon of 3.5 kb (annotated as CG6889 by the complete D. melanogaster genome database), ending with a putative polyadenylation signal (ATTAAA; PROUDFOOT and BROWNLEE 1976 ) in the immediate vicinity of the A-rich 3'-terminal region found in all cDNA clones (not shown); and (iii) this 3' exon is linked to alternative 5' exons of 790 and 511 bp situated 30 and 13.5 kb distant and corresponding to C12 and C16 cDNAs, respectively (see Fig 2B). The most direct interpretation of these data is that the - and ß-type cDNA inserts correspond to alternatively spliced 5' exons (termed 1A and 1B, respectively) initiated from distinct transcriptional start sites (see Fig 2A and Fig B).

The exons 1A and 1B each include an in-frame initiator ATG in a favorable context compared to the Drosophila consensus start site (CAVENER 1987 ; see Fig 3). The expected 4.5- and 4.2-kb polyadenylated mRNA species thus encode two distinct proteins (hereafter referred to as TARA- and -ß isoforms) with predicted molecular weights of 96,274 and 95,730 D, respectively. These isoforms differ only in their amino (N)-terminal sequences, sharing 908 residues of 916 () or 912 (ß; Fig 3).

Genetic evidence for partially independent tara promoters:
In addition to tara^L4 and tara¹ (see above), analysis of the recently completed Drosophila genomic sequence (ADAMS et al. 2000 ) allowed us to identify 10 other P insertions distributed throughout the tara locus (see MATERIALS AND METHODS), including four (EP3294, EP3427, EP3463, and EP3530) that are located within or near either one of the two alternative 5' exons. The 7 examined insertions (EP508, EP693, EP1189, EP3294, EP3427, EP3463, and EP3530) are all homozygous viable or subviable (see Table 1 for representative EP alleles). When those insertions were tested in combination with themselves, with each other, with tara^L4 or the tara^- deficiency Df(3R)sbd26, various adult phenotypes were observed including held-out wings, defects of wing veins L2 and/or L5, and abnormally located bristles on the ventral abdomen (Table 1, Fig 1D, and data not shown). Interestingly, two insertions located within exon 1A and close to exon 1B (EP3427 and EP3463, respectively; see Fig 2A) complement each other. This observation suggests that transcription of the two alternative mRNAs may depend on distinct promoter sequences and give rise to functionally redundant TARA-/-ß isoforms.

tara is broadly expressed throughout development:
The spatial expression of the tara gene during various developmental stages was examined by whole-mount tissue in situ hybridization. Three antisense RNA probes were prepared, permitting detection of overall gene expression and discrimination between the - and ß-type mRNA species (for details, see Fig 2A and MATERIALS AND METHODS).

tara mRNAs are detected at all embryonic stages with a common probe derived from exon 2 (probe C). Weak but reproducible signal is obtained in blastoderm embryos (not shown), presumably reflecting maternally contributed mRNA. By the onset of gastrulation (stage 6), stronger staining is evident in invaginating cells in the cephalic furrow, the prospective mesoderm, and posterior midgut (Fig 4C). This heightened local accumulation is transient and broadens to encompass the entire germ band at full extension. By stage 13, and until stage 17, tara mRNAs are present at uniformly low levels, except for the ventral nerve cord, brain, and visceral mesoderm (Fig 4F and Fig I). As in embryos, generalized tara expression is detected in late third instar larval tissues including the imaginal discs, the brain, and ventral nerve cord (not shown).

View larger version (62K): In this window In a new window Download PPT slide   Figure 4. Spatial distribution of tara mRNAs during embryonic development. Whole-mount preparations of wild-type (w1118) embryos at indicated stages were hybridized with antisense tara riboprobes corresponding to exon 1A () (A, D, and G), exon 1B (ß) (B, E, and H), or exon 2 (C, F, and I). Anterior is to the left; dorsal is uppermost. - and ß-type mRNAs are broadly expressed throughout embryogenesis. In stage 6 embryos note the higher mRNA accumulation in invaginating tissues, in particular the ventral mesodermal cells. By stage 13, also note the higher accumulation of ß-type mRNA species throughout the gut.

To reveal mRNA species specific for the TARA- or -ß isoform, probes were derived from the 5' part of the C12 cDNA insert corresponding to exon 1A (probe A) and from a genomic fragment overlapping the exon 1B (probe B), respectively (see Fig 2A). The spatial and temporal expression patterns observed with the -specific probe closely resemble that obtained with the common probe (above). In contrast, although a low-level uniform staining is detected, the ß-specific probe revealed mRNA accumulation in the developing gut throughout embryogenesis. As a whole, these data indicate that tara is broadly expressed throughout development. However, the difference in accumulation between and ß mRNAs again suggests that their expression depends on distinct regulatory sequences.

Ubiquitous expression of the TARA-ß isoform is sufficient for viability:
To address the relative contributions of the - and ß-transcription units to tara activity, we performed phenotypic rescue experiments using a transgene coding for only the ß isoform. A ß-coding cDNA fragment was subcloned into the P-transformation vector pCaSpeR4 (THUMMEL and PIRROTTA 1992 ), downstream of a ubiquitously active promoter from the ubiquitin-63E gene (LEE et al. 1988 ; see MATERIALS AND METHODS for details). Four independent transformant lines carrying the resulting P[UbTARA-ß] construct were tested for their capacity to rescue the embryonic/larval lethality of tara^L4 homozygote animals. In all four cases P[UbTARA-ß]/+; tara^L4/tara^L4 animals were fully viable (Table 3), indicating that TARA-ß protein possesses tara genetic function. However, these rescued flies showed poor fertility and a fully penetrant wings-held-out phenotype (Fig 1G). Furthermore, a small number of these incompletely rescued adults showed clear homeotic phenotypes, including partial arista-to-tarsa and haltere-to-wing transformations (Fig 1, compare K and L with I and J, respectively), supporting our previous conclusion that tara is a bona fide trxG gene since the homeotic selectors are sensitive to trxG function. Two simple models (not mutually exclusive) might readily explain the incomplete rescue observed. The level of transgene expression might be inadequate for complete rescue of some specific tara functions. Alternatively, the isoform might be required in dorsal mesothoracic cells. To distinguish between these two possibilities, we asked whether the rescue could be improved by increasing the dose for the same isoform. Many animals carrying a second transgene copy (P[UbTARA-ß]/P[UbTARA-ß]; tara^L4/tara^L4) now showed normal wing posture (Fig 1H) and increased fertility allowing stable stocks (see Table 3). These data, as well as overexpression experiments (not shown), therefore suggest (i) that tara function is dose sensitive in specific tissues and (ii) that the two TARA protein isoforms perform largely or fully redundant functions.

The TARA proteins belong to a novel evolutionarily conserved family apparently restricted to higher eukaryotes:
The common primary sequence of TARA-/-ß proteins contains several regions with a strong amino acid (aa) bias. These include a long acidic stretch (residues 200–217 in TARA-ß) as well as several runs of alanine and glutamine residues (see Fig 3). The carboxy(C)-terminal third of the predicted TARA-/-ß proteins is markedly rich in alanine, serine, and threonine residues (58% over 264 amino acids). Finally, they share a potential nuclear localization signal (ATKRKH, positions 16–21 of TARA-ß) near the N terminus. These are all frequently encountered structural features of nuclear regulatory components.

Apart from the preceding general traits, initial analyses of the TARA twin isoforms showed no motif diagnostic of established biochemical or functional properties. However, more detailed sequence database searches revealed several mammalian proteins structurally related to TARA-/-ß (Fig 5). Based on DbClustal analysis (THOMPSON et al. 2000 ) the highest identity score (E value of 0.098) was with human Y127 (NAGASE et al. 1995 ), a hypothetical 34-kD protein. In turn, searches for protein homology with Y127 as the query sequence revealed at least three human paralogs and presumptive mouse orthologs of the four Y127 family members (not shown). TARA-/-ß and the four mammalian Y127 family proteins (hereafter referred to as TRIP-Br family members; see below) show four distinct regions of clear similarity (see Fig 5A). The four evolutionarily conserved regions comprise (i) an N-terminal basic cyclin A-binding motif homologous to that of the cell cycle regulatory transcription factors E2F1–3 (KREK et al. 1994 ; ADAMS et al. 1996 ); (ii) a novel motif that we designate SERTA (for SEI-1, RBT1, and TARA), corresponding to the largest conserved region among TRIP-Br proteins; (iii) a PHD-bromo interaction domain (HSU et al. 2001 ); and (iv) a conserved C-terminal (C-ter) motif of unknown biochemical function. Each of the five proteins contains all four motifs, organized in the same order in the linear sequence (see Fig 5B). These considerations led us to conclude that TARA and mouse/human TRIP-Br proteins are members of a novel evolutionarily conserved family. The TARA- related human proteins include (a) TRIP-Br1/p34^SEI-1, recently identified both as a cyclin-dependent kinase regulator (SUGIMOTO et al. 1999 ) and as a transcriptional regulator interacting with PHD and bromodomains (HSU et al. 2001 ), two motifs widely found in chromosomal proteins (AASLAND et al. 1995 ; WINSTON and ALLIS 1999 ); (b) RBT1, a potent transcriptional coactivator that interacts with the second subunit of replication protein A (CHO et al. 2000 ); and (c) HEPP, a novel protein expressed preferentially in hematopoietic progenitors and mature blood cells (ABDULLAH et al. 2001 ). As shown in Fig 5A, a careful analysis of protein databases revealed a SERTA region conserved in predicted human (GenBank accession no. CAB81635) and Drosophila proteins [Drosophila Genome Annotation Database (GadFly) identifier CG2865 (ADAMS et al. 2000 )]. No conservation with TRIP-Br proteins was detected outside of the SERTA motif. In particular, no C-ter motif was detected in CG2865, nor in any other predicted Drosophila protein. Interestingly, no SERTA-containing proteins were discernible in the yeast or Caenorhabditis elegans genomes, suggesting that functional roles for TRIP-Br proteins may be restricted to higher eukaryotes. Taken together, these data support the notion that tara encodes the Drosophila homologs of the mammalian TRIP-Br proteins. The implications of this observation, in particular for the epigenetic maintenance of active chromosomal states with regard to the cell cycle, are discussed below.

View larger version (47K): In this window In a new window Download PPT slide   Figure 5. Constitution of a family of TARA-related proteins. (A) The sequences of the conserved portions of TARA-ß are shown aligned with homologous motifs from the related human TRIP-Br proteins, human transcription factor E2F-1, as well as predicted human (Hs) and Drosophila (Dm) proteins. Numbers flanking the amino acid sequences indicate positions in the complete protein sequences. Shown on a solid or shaded background are amino acids that are identical or similar, respectively, to TARA proteins. The following amino acids were treated as similar: R/K/H; S/T; E/D/Q/N; A/G/P; and I/L/V/M/C/F/Y/W. For the Cyclin A-binding homology motif the equivalent region of E2F-1 is shown below. For the SERTA motif homologous regions from the predicted proteins HsdJ667H12.2.1 and DmCG2865 (without discernible Cyclin A-binding, PHD-bromo-binding, or C-ter motifs) are shown below with consensus positions displayed farther below. In these positions, similar or identical amino acids were found in at least six of the seven aligned sequences. Note that only four positions were identical in all nine SERTA motifs compared (marked by an asterisk). The percentage of amino acid identity with respect to the TARA-/-ß SERTA motif is indicated on the right. The GenBank accession numbers are hsY127/TRIP-Br2 (BAA09476); HsSEI-1/TRIP-Br1 (AAF08349); HsRBT1 (AAF05761); HsHEPP (AAK31075); HsdJ667H12.2.1 (CAB81635); DmCG2865 (AAF45770); and HsE2F-1 (Q01094). (B) Comparison of the motif structure of TARA-/-ß and human TRIP-Br proteins. For each alignment the overlying box refers to the corresponding motif in A. Note the equivalent order of the four conserved motifs in TARA and the mammalian TRIP-Br proteins. The minimal region of TRIP-Br1/SEI-1 required for interaction with CDK4 is indicated by brackets (aa 44–161).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We have identified a novel member of the trxG family, taranis, based notably on the dose-sensitive effects of reduced tara⁺ function (i) in suppression of the extra-sex-combs phenotype resulting from mutations in the PcG genes Polycomb and polyhomeotic and (ii) in enhancement of a specific wing phenotype in combination with mutations of the trxG genes trithorax, brahma, and osa. Further, the developmental transformations in incompletely rescued tara mutants resembling homeotic phenotypes (Fig 1), and genetic interactions with a HOM gene (here, proboscipedia), correspond to often-encountered properties of trxG members.

As for most other trxG genes, tara is required for viability. Reduced tara activity during larval/pupal development leads to a wings-held-out phenotype resembling certain mutations of Antp. One potential function of tara, as for brm and osa, is thus to ensure proper transcriptional expression of Antp in cells of developing wing imaginal discs. In support of this possibility, a genetic interaction was detected with certain Antp gain-of-function alleles, which was associated with detectably reduced accumulation of ANTP protein (data not shown). However, the wing vein defects and diminished fertility observed on reducing normal function indicate that tara is in all likelihood also required for the expression of other target genes in addition to homeotic genes. Zygotic lethality was not accompanied by a detected cuticular phenotype for the alleles used in this study. The isolation of a null mutation of tara may reveal additional roles in embryonic and adult development.

Several trxG proteins function as chromatin modifiers:
The PcG and trxG genes are believed to encode proteins that play out a direct functional antagonism at the level of nucleosomal and higher-order chromatin structures. Most PcG members encode chromosomal proteins belonging to common complexes that repress gene transcription, although the mechanism of this repression remains poorly understood (for a recent review, see FRANCIS and KINGSTON 2001 ). The recent purification and analysis of one such complex, PRC1, containing PC and PH proteins, suggested that PRC1 and SWI/SNF complexes might compete for the nucleosomal template in vivo (SHAO et al. 1999 ). However, only 4 of the 17 cloned trxG genes (BRM, MOR, SNR1, and OSA) correspond to dSWI/SNF subunits (PAPOULAS et al. 1998 ; COLLINS et al. 1999 ; CROSBY et al. 1999 ; KAL et al. 2000 ). Hence, most trxG proteins are likely to function via other mechanisms. Among these, ASH1, ASH2, and possibly KIS, have been shown to belong to multiprotein chromatin-remodeling complexes distinct from the BRM complex and of unknown functions (PAPOULAS et al. 1998 ; DAUBRESSE et al. 1999 ).

A major issue in the field is how the assembly and activity of these diverse chromatin-modifying complexes are regulated to control transcription of specific target genes in a mitotically stable manner (for a recent review, see FARKAS et al. 2000 ). Several Drosophila trxG proteins that directly interact with regulatory DNA sequences (TRX, GAGA/TRL, and ZESTE) are thought to recruit those complexes to overcome the repressive effects of nucleosomal histones and higher-order chromatin organization (ROZOVSKAIA et al. 1999 ; KAL et al. 2000 ). GAGA/TRL is of particular interest among the trxG proteins since, contrary to most transcription factors (see MARTINEZ-BALBAS et al. 1995 ), it remains associated with chromosomes during mitosis (RAFF et al. 1994 ; O'BRIEN et al. 1995 ). Characterization of other trxG proteins should help to better understand how the activities of the diverse chromatin-modifying machines are regulated throughout development.

A role for TARA family proteins in linking chromatin-remodeling complexes to cell cycle regulators?
The tara gene structure suggests a possible molecular mechanism of action. Our molecular analysis of the twin TARA proteins revealed significant homologies with members of a novel family of mammalian proteins including human p34^SEI-1/TRIP-Br1 (Fig 5), a potent transcriptional activator reported to regulate the cyclin D1-CDK4 pair implicated in control of the G1 phase of the cell cycle (SUGIMOTO et al. 1999 ; HSU et al. 2001 ). p34^SEI-1/TRIP-Br1 has been reported to favor cyclin D1-CDK4 complex formation by antagonizing the inhibitory action of the tumor suppressor p16^INK4a. The positive effect of p34^SEI-1/TRIP-Br1 on the assembly and activation of cyclin D1-CDK4 complex is thought to involve a specific interaction of TRIP-Br1 with the kinase domain of CDK4 (SUGIMOTO et al. 1999 ). Hence, it may be significant that the CDK4-interacting segment of p34^SEI-1 (amino acid residues 44–161) includes most of the SERTA motif (see Fig 5A).

Taken together, these data suggest that TARA-/-ß proteins might participate in a cell memory process that couples chromatin structure to cell cycle progression. The observed detrimental consequences of under- and overexpression indicate a probable stoichiometric role for tara. Mitotic recombination experiments employing the strong hypomorphic tara^L4 allele indicate that tara⁺ is required for viability and/or proliferation of imaginal cells, since clones of mutant cells were strongly reduced compared to reference wild-type clones (data not shown). Conversely, the reduction or loss of adult structures noted in overexpression experiments suggests that TARA protein quantity must be finely weighed in normal development of most examined adult external tissues (not shown). Apart from CDK4, other CDKs identified in Drosophila include CDK1/CDC2 and CDK2 (the catalytic partners of cyclins A and E, respectively), as well as several other CDKs for which no function has yet been assigned (SAUER et al. 1996 ; FOLLETTE and O'FARRELL 1997 ). Interestingly, the wing vein defects observed for viable tara alleles resemble those observed with hypomorphic alleles of cyclin E (DURONIO et al. 1998 ), raising the possibility of a shared molecular role that it will be of interest to examine.

Some hints emerge from recent results of molecular and genetic interaction screens. CDKs are thought to initiate and coordinate cell division processes by sequentially phosphorylating key protein targets (DYNLACHT 1997 ). Three potential targets are human trxG homologs of the dSWI/SNF proteins BRM and MOIRA. Physical and functional interactions detected between components of the human SWI/SNF complex (hSWI/SNF) and Cyclin E (SHANAHAN et al. 1999 ), and the phosphorylation of these proteins detected during mitosis (SIF et al. 1998 ), may be relevant to a cyclin-mediated role in modulating hSWI/SNF chromatin-remodeling activity. A potential link between TARA and the cell cycle machinery is also suggested by the ability of related mammalian TRIP-Br proteins to interact in vitro with Cyclin A (HSU et al. 2001 ). The Cyclin A-interacting domain of TRIP-Br1 is conserved among all TRIP-Br family members including TARA (Fig 5). Hence, one prediction is that the TRIP-Br proteins should serve as specific substrates for phosphorylation by one or several CDK complex(es).

Additional connections have been established between chromatin-remodeling complexes and E2F proteins, key transcription factors coupling the transcriptional program to cell cycle progression (HARBOUR and DEAN 2000 ) and to the retinoblastoma tumor-suppressor protein pRB, itself a target of Cyclin-D1 complexes (DYSON 1998 ). Recent genetic interaction screens in Drosophila establish functional ties between dSWI/SNF, E2F, and pRB (STAEHLING-HAMPTON et al. 1999 ; GILDEA et al. 2000 ). Interestingly, the TRIP-Br1 and TRIP-Br2 proteins interact in vitro with the E2F1 partner DP1 and stimulate transcriptional activity of the E2F1/DP1 heterodimer in a manner that is regulated by RB (HSU et al. 2001 ). The ability of TRIP-Br proteins to specifically contact PHD and bromodomains raises the clear possibility that they might function as molecular integrators coupling the cell cycle machinery to transcriptional activity. Interestingly, the genetic screen for dominant suppressors of polyhomeotic that yielded tara¹ also allowed us to identify toutatis, a novel trxG gene encoding a PHD zinc-finger- and bromodomain-containing protein (FAUVARQUE et al. 2001 ). As BRM and TRX also contain a bromo and a PHD domain, respectively, it will be important to ascertain which pairs or groups of trxG molecules are capable of direct physical interaction.

In summary, these and other related data hint at regulated activities of trxG proteins, including TARA, as cells traverse the mitotic cycle. Available genetic data and structural conservation of identified functional domains of related TRIP-Br proteins noted above suggest that the TARA-/-ß proteins might intervene as integrators linking key cell cycle regulators to chromatin-remodeling complexes. The availability of cloned cyclins and CDKs and corresponding loss-of-function alleles should allow us to test for specific molecular relations of TARA proteins with CDK complexes and chromatin components in establishing and maintaining cellular "memory" of active transcriptional states.

	ACKNOWLEDGMENTS

We thank Kathy Matthews and the Indiana University Drosophila Stock Center for providing most stocks used in this work. We thank A. Lepage for generating the transgenic lines used in this work and other members of our laboratory for their support throughout the course of this work. The manuscript benefitted from the critical readings of Drs. A. Vincent and M. Crozatier. We particularly thank Drs. M.-O. Fauvarque and J.-M. Dura who generously shared mutants and useful information. This investigation was supported by recurrent funding from the French Centre National de la Recherche Scientifique (CNRS) as well as grants to D.C. from the French Ministère de l'Education Nationale et de la Recherche (ACC program) and the Association pour la Recherche sur le Cancer (ARC). S.C. and M.B. were supported by fellowships from the French Ministère de la Recherche et de la Technologie and the ARC.

Manuscript received December 11, 2000; Accepted for publication November 13, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AASLAND, R., T. J. GIBSON, and A. F. STEWART, 1995  The PHD finger: implications for chromatin-mediated transcriptional regulation. Trends Biochem. Sci. 20:56-59[Medline].

ABDELILAH-SEYFRIED, S., Y. M. CHAN, C. ZENG, N. J. JUSTICE, and S. YOUNGER-SHEPHERD et al., 2000  A gain-of-function screen for genes that affect the development of the Drosophila adult external sensory organ. Genetics 155:733-752[Abstract/Free Full Text].

ABDULLAH, J. M., X. JING, D. S. SPASSOV, R. G. NACHTMAN, and R. JURECIC, 2001  Cloning and characterization of hepp, a novel gene expressed preferentially in hematopoietic progenitors and mature blood cells. Blood Cells Mol. Dis. 27:667-676[Medline].

ADAMS, M. D., S. E. CELNIKER, R. A. HOLT, C. A. EVANS, and J. D. GOCAYNE et al., 2000  The genome sequence of Drosophila melanogaster. Science 287:2185-2195[Abstract/Free Full Text].

ADAMS, P. D., W. R. SELLERS, S. K. SHARMA, A. D. WU, and C. M. 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].

AKAM, M., 1987  The molecular basis for metameric pattern in the Drosophila embryo. Development 101:1-22[Abstract].

BENASSAYAG, C., M. BOUBE, L. SEROUDE, and D. L. CRIBBS, 1997  Point mutations within and outside the homeodomain identify sequences required for proboscipedia homeotic function in Drosophila. Genetics 146:939-949[Abstract].

BIER, E., H. VAESSIN, S. SHEPHERD, K. LEE, and K. MCCALL et al., 1989  Searching for pattern and mutation in the Drosophila genome with a P-lacZ vector. Genes Dev. 3:1273-1287[Abstract/Free Full Text].

BOTAS, J., J. MOSCOSO DEL PRADO, and A. GARCIA-BELLIDO, 1982  Gene-dose titration analysis in the search of trans-regulatory genes in Drosophila. EMBO J. 1:307-310[Medline].

BOUBE, M., C. BENASSAYAG, L. SEROUDE, and D. L. CRIBBS, 1997  Ras1-mediated modulation of Drosophila homeotic function in cell and segment identity. Genetics 146:619-628[Abstract].

BOUBE, M., L. SEROUDE, and D. L. CRIBBS, 1998  Homeotic proboscipedia cell identity functions respond to cell signaling pathways along the proximo-distal axis. Int. J. Dev. Biol. 42:431-436[Medline].

BROWN, N. H. and F. C. KAFATOS, 1988  Functional cDNA libraries from Drosophila embryos. J. Mol. Biol. 203:425-437[Medline].

CAPDEVILA, M. P. and A. GARCIA-BELLIDO, 1981  Genes involved in the activation of the bithorax complex of Drosophila. Roux's Arch. Dev. Biol. 190:339-350.

CAVALLI, G. and R. PARO, 1999  Epigenetic inheritance of active chromatin after removal of the main transactivator. Science 286:955-958[Abstract/Free Full Text].

CAVENER, D. R., 1987  Comparison of the consensus sequence flanking translational start sites in Drosophila and vertebrates. Nucleic Acids Res. 15:1353-1361[Abstract/Free Full Text].

CHAMBERLIN, H. M. and J. H. THOMAS, 2000  The bromodomain protein LIN-49 and trithorax-related protein LIN-59 affect development and gene expression in Caenorhabditis elegans. Development 127:713-723[Abstract].

CHO, J. M., D. J. SONG, J. BERGERON, N. BENLIMAME, and M. S. WOLD et al., 2000  RBT1, a novel transcriptional co-activator, binds the second subunit of replication protein A. Nucleic Acids Res. 28:3478-3485[Abstract/Free Full Text].

COLLINS, R. T., T. FURUKAWA, N. TANESE, and J. E. TREISMAN, 1999  Osa associates with the Brahma chromatin remodeling complex and promotes the activation of some target genes. EMBO J. 18:7029-7040[Medline].

CRIBBS, D. L., M. A. PULTZ, D. JOHNSON, M. MAZZULLA, and T. C. KAUFMAN, 1992  Structural complexity and evolutionary conservation of the Drosophila homeotic gene proboscipedia. EMBO J. 11:1437-1449[Medline].

CRIBBS, D. L., C. BENASSAYAG, F. M. RANDAZZO, and T. C. KAUFMAN, 1995  Levels of homeotic protein function can determine developmental identity: evidence from low-level expression of the Drosophila homeotic gene proboscipedia under Hsp70 control. EMBO J. 14:767-778[Medline].

CROSBY, M. A., C. MILLER, T. ALON, K. L. WATSON, and C. P. VERRIJZER et al., 1999  The trithorax group gene moira encodes a brahma-associated putative chromatin-remodeling factor in Drosophila melanogaster. Mol. Cell. Biol. 19:1159-1170[Abstract/Free Full Text].

DAUBRESSE, G., R. DEURING, L. MOORE, O. PAPOULAS, and I. ZAKRAJSEK et al., 1999  The Drosophila kismet gene is related to chromatin-remodeling factors and is required for both segmentation and segment identity. Development 126:1175-1187[Abstract].

DUNCAN, I. M., 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., 1986  Control of bithorax complex functions by the segmentation gene fushi tarazu of D. melanogaster. Cell 47:297-309[Medline].

DURA, J. M. and P. INGHAM, 1988  Tissue- and stage-specific control of homeotic and segmentation gene expression in Drosophila embryos by the polyhomeotic gene. Development 103:733-741[Abstract/Free Full Text].