A Screen for Modifiers of decapentaplegic Mutant Phenotypes Identifies lilliputian, the Only Member of the Fragile-X/Burkitt's Lymphoma Family of Transcription Factors in Drosophila melanogaster

Corresponding author: Stuart J. Newfeld, Department of Biology, Arizona State University, Tempe, AZ 85287-1501., newfeld{at}asu.edu (E-mail)

Communicating editor: C.-I WU

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The decapentaplegic (dpp) gene directs numerous developmental events in Drosophila melanogaster. dpp encodes a member of the Transforming Growth Factor-ß family of secreted signaling molecules. At this time, mechanisms of dpp signaling have not yet been fully described. Therefore we conducted a genetic screen for new dpp signaling pathway components. The screen exploited a transvection-dependent dpp phenotype: heldout wings. The screen generated 30 mutations that appear to disrupt transvection at dpp. One of the mutations is a translocation with a recessive lethal breakpoint in cytological region 23C1-2. Genetic analyses identified a number of mutations allelic to this breakpoint. The 23C1-2 complementation group includes several mutations in the newly discovered gene lilliputian (lilli). lilli mutations that disrupt the transvection-dependent dpp phenotype are also dominant maternal enhancers of recessive embryonic lethal alleles of dpp and screw. lilli zygotic mutant embryos exhibit a partially ventralized phenotype similar to dpp embryonic lethal mutations. Phylogenetic analyses revealed that lilli encodes the only Drosophila member of a family of transcription factors that includes the human genes causing Fragile-X mental retardation (FMR2) and Burkitt's Lymphoma (LAF4). Taken together, the genetic and phylogenetic data suggest that lilli may be an activator of dpp expression in embryonic dorsal-ventral patterning and wing development.

THE decapentaplegic (dpp) gene influences many developmental events in Drosophila melanogaster. These include dorsal-ventral patterning in the embryo, larval midgut morphogenesis, and formation of adult appendages (GELBART 1989 ). dpp encodes a member of the highly conserved Transforming Growth Factor-ß (TGF-ß) family of secreted signaling molecules (PADGETT et al. 1987 ). To understand how dpp directs developmental decisions in target cells, mechanisms of dpp activation and signal transduction must be fully described. Genetic screens have been successful in identifying components of Dpp's signal transduction pathway (RAFTERY et al. 1995 ; SEKELSKY et al. 1995 ). These screens exploited recessive embryonic lethal dpp alleles to identify mutations that enhance this phenotype. Mothers against dpp (Mad) and Medea (Med) were identified in these screens. These genes, members of the Smad family, are also highly conserved across species (NEWFELD et al. 1999 ). Smad family members play important roles in mouse development and act as tumor suppressor genes in several human cancers (RIGGINS et al. 1997 ; WALDRIP et al. 1998 ).

Here we report a genetic screen for dpp signaling pathway components that exploits transvection effects at the dpp locus (GELBART 1982 ). Transvection, or pairing-dependent intragenic complementation between two alleles of a gene, is seen at a number of loci (LEWIS 1954 ). As a result of transvection, trans-heterozygous individuals of the genotype dpp^d-ho/dpp^hr4 display wild-type wings. The dpp^d-ho mutation is a small deletion in the 3' cis-regulatory region of dpp. dpp^d-ho homozygous flies have wings that are held out laterally from the body axis (SPENCER et al. 1982 ). The dpp^hr4 mutation is a missense mutation in the protein-coding region of dpp (WHARTON et al. 1996 ). When homozygous, the dpp^hr4 allele is embryonic lethal. When dpp^d-ho and dpp^hr4 are paired, the wild-type regulatory region of the dpp^hr4 allele appears to act in trans on the wild-type coding region of the dpp^d-ho allele to generate viable adults with wild-type wings.

During transvection, the respective regions (regulatory and coding) must be in close physical proximity. A chromosomal rearrangement that physically moves a dpp allele to another part of the chromosome disrupts transvection (GELBART 1982 ). Rather than having wild-type wings, dpp^d-ho/dpp^hr4 flies with chromosomal rearrangements have heldout wings. Analyses of polytene chromosomes from rearrangement genotypes showed asynapsis at the dpp locus. These rearrangements are referred to as normal dpp transvection-disruptors (normal DTDs). Trans-heterozygous dpp^d-ho/dpp^hr4 flies will also display a heldout phenotype if they contain a rearrangement with a breakpoint in a gene required for dpp function (e.g., Mad; SEKELSKY et al. 1995 ). This type of rearrangement, one that generates heldout phenotypes in trans-heterozygous flies without asynapsis at the dpp locus, is referred to as an exceptional DTD (GELBART 1982 ).

To determine if a DTD is normal or exceptional, an unknown DTD is paired with a previously characterized normal DTD. If the unknown DTD is a normal DTD, trans-heterozygous flies will display wild-type wings. Two normal DTDs (even those with very different rearrangements) have the ability to arrange themselves in such a way that synapsis occurs at the dpp locus (GELBART 1982 ). If the unknown DTD is an exceptional DTD, trans-heterozygous flies will display heldout wings. The presence of a normal DTD cannot suppress a heldout phenotype that is due to a mutation in a gene required for dpp function. Mutations that act as exceptional DTDs are therefore candidates for components of the dpp signaling pathway.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila stocks:
dpp^d-ho, dpp^hr4, dpp^e87, dpp^hr56, and dpp^hr92 are described in ST. JOHNSTON et al. 1990 . DTD11, DTD24, Df(2L)DTD16xD42, and Df(2L)DTD51xD52 are described in GELBART 1982 . Med¹, scw^E1, and scw^S12 are described in RAFTERY et al. 1995 . Df(2L)C28, Df(2L)C144, Df(2L)JS17, Df(2L)JS7, Df(2L)DTD62xH7, Mad⁶, Mad¹¹, and Mad¹² are described in SEKELSKY et al. 1995 . sax¹ and tkv⁸ are described in BRUMMEL et al. 1994 . l(2)a16, l(2)k9, l(2)a4, and l(2)a6 were identified in a large screen (>5000 chromosomes) for lethal mutations over Df(2L)JS17 described in SEKELSKY 1993 . gbb¹ is described in WHARTON et al. 1999 . l(2)00632 and l(2)k05431 allelic to lilli, l(2)01361 allelic to toucan, and Df(3R)e-N19 are described in FLYBASE 1999 . lilli^s35 and lilli^xs407 are described in NEUFELD et al. 1998 and REBAY et al. 2000 .

Exceptional DTD screen:
Homozygous dpp^d-ho males were irradiated and crossed to dpp^hr4/CyO females. All G1 heldout progeny were isolated. These progeny carry DTDs (*). Single G1 heldout males were mated to females carrying a normal DTD (either DTD11 or DTD24). If the G2 progeny was heldout, then the new DTD was an exceptional DTD. The dpp^d-ho * chromosome was then balanced. Gravid G1 heldout females were placed alone in a vial and allowed to produce progeny. Heldout male progeny must be either dpp^d-ho */dpp^d-ho or dpp^d-ho */dpp^hr4. These males were crossed to dpp^hr4/CyO females. Heldout progeny from this cross must bear the genotype dpp^d-ho */dpp^hr4. These males were treated like G1 heldout males and crossed to females carrying a normal DTD. If the resulting progeny was heldout, then the new DTD was an exceptional DTD. The dpp^d-ho * chromosome was then balanced. Wing angle measurements were performed as described (GELBART 1982 ). Polytene chromosome squashes, cuticle preps, maternal enhancement, and stage of lethality tests were performed as described (SEKELSKY et al. 1995 ).

Phylogenetic analysis of Lilli:
Database searches for proteins similar to Lilli were conducted using the National Institutes of Health website: http://www.ncbi.nlm.nih.gov/BLAST. In addition to GenBank, we conducted extensive BLAST searches of the genome databases for D. melanogaster (Berkeley Drosophila Genome Project website: http://fruitfly.berkeley.edu) and Caenorhabditis elegans (Washington University Genome Sequence Center website: http://genome.wustl.edu/gsc). Proteins identified by these searches that showed strong similarity to Lilli (see the legend to Fig 5 for accession numbers) were aligned with MACAW (SCHULER et al. 1991 ). The alignments were refined using CLUSTAL X (JEANMOUGIN et al. 1998 ) and then adjusted manually. Alignments not presented are available upon request. Protein motifs were identified in the alignments using the Kyoto University GenomeNet website: http://www.motif.genome.ad.jp. Pairwise calculations of amino acid identity and similarity were deduced from the alignments using the Baylor College of Medicine website: http://dot.imgen.bcm.tmc.edu:9331.

View larger version (28K): In this window In a new window Download PPT slide   Figure 1. Screen for exceptional DTDs. One version of the screen is shown in which the G1 heldout mutant is male and DTD11 is used to test for exceptional DTDs. See MATERIALS AND METHODS for details.

View larger version (59K): In this window In a new window Download PPT slide   Figure 2. Cytological and genetic mapping of DTD46.4. (A) Polytene chromosomes from larvae heterozygous for DTD46.4 show a T(2;3)23C;93F rearrangement. The 23C and 93F cytological regions are indicated by arrows. (B) A schematic representation of cytological region 23C-D. The cytological locations of several deficiencies are shown. The endpoints of all deficiencies are approximate except that the distal breakpoints of Df(2L)C28 and Df(2L)JS17 have been cloned (SEKELSKY 1993 ). The cytological locations of five complementation groups ordered using the deficiencies are indicated by vertical dashed lines. The EMS-induced mutation l(2)a16 and the P-element insertion lines l(2)00632, l(2)k05431 were used to place lilli . The P-element insertion line l(2)01361 was used to place toucan (toc). The EMS-induced mutations Mad6, Mad11, and Mad12 were used to place Mad. The EMS-induced mutations l(2)k9 and l(2)a4 represent complementation groups not currently assigned to a known gene.

View larger version (27K): In this window In a new window Download PPT slide   Figure 3. lilli mutants are dominant maternal enhancers of dpp and scw recessive embryonic lethality. (A) dpphr4. (B) dpphr92. (C) scwE1. Bars represent the percentage of expected progeny obtained from each mating. The actual value is shown. Solid bars indicate tests for zygotic enhancement of recessive lethality (matings where the father was heterozygous for lilli). Shaded bars indicate tests for maternal enhancement of recessive lethality (matings where the mother was heterozygous for lilli). In control crosses, dppd-ho was utilized in place of the recessive embryonic lethal allele. At least 75 progeny were counted from each mating. In maternal enhancement crosses, adult escaper progeny with the dpp or scw mutant chromosomes had no defects in their eyes, wings, or legs and females were fertile.

View larger version (125K): In this window In a new window Download PPT slide   Figure 4. lilli mutant embryos have a partially ventralized phenotype. (A) Wild-type embryo oriented anterior to the left and dorsal toward the top. The head skeleton at the anterior (thin arrow) and the filzkorper at the posterior (wide arrow) are noted. (B) l(2)a16/Df(2L)C144 embryo. In this example, the head is dysmorphic (thin arrow) and the filzkorper are internalized (wide arrow). The ventral denticle bands are expanded toward the dorsal side and are disorganized. This embryo is similar to a dpphr56 mutant embryo (WHARTON et al. 1993 ). (C) l(2)a16/Df(2L)C144 embryo. In this example, the head skeleton is completely herniated (thin arrow), the embryo is bent into a U shape, and the filzkorper are internalized (wide arrow). This embryo is similar to a scwE1 mutant embryo (RAFTERY et al. 1995 ).

View larger version (92K): In this window In a new window Download PPT slide   Figure 5. Lilli is a member of the FMR2/LAF4 multigene family. (A) Schematic representation of an FMR2/LAF4 family member. The locations of three conserved domains are shown. (B) Amino acid alignment of the FMR2/LAF4 diagnostic domain of Lilli and the four human family members. Black boxes indicate identical amino acids at that position in at least three sequences. Shaded boxes indicate a similar amino acid at that position in at least three sequences. Gaps in the alignment minimize the number of mutations required to explain all differences between the sequences. Amino acid numbers for each sequence are indicated. Accession numbers are as follows: Lilli, AAF51180; FMR2, AAA99416, AF5–AAF18981, AF4–CAB69660, and LAF4–NP002276.

Phylogenetic trees were generated from the alignments using MEGA (KUMAR et al. 1993 ). First, a Poisson correction distance (NEI and KUMAR 2000 ) was calculated from each pairwise distance to account for multiple substitutions per site. Then the evolutionary divergence (the number of amino acid substitutions per site) between two sequences was calculated from the Poisson correction distance. Trees were then constructed on the basis of the corrected distance matrix using the neighbor-joining method (SAITOU and NEI 1987 ). Trees not presented are available upon request. The degree of confidence for each branchpoint was obtained by the bootstrap method (1000 replications; FELSENSTEIN 1985 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Exceptional DTD screen:
A total of 44,000 dpp^hr4/dpp^d-ho flies were screened (Fig 1) and 321 DTD mutations were isolated. Of these mutations, 30 were exceptional DTDs (Table 1). All exceptional DTDs were cytologically mapped. If an exceptional DTD chromosome appeared cytologically normal, the DTD mutation was mapped by recombination.

All exceptional DTDs were then tested for genetic interactions (enhancement of recessive embryonic lethality) with several classes of mutations affecting the Dpp signaling pathway. First, we tested for interactions with loss-of-function mutations in the Dpp receptors saxophone and thickveins (sax¹ and tkv⁸; BRUMMEL et al. 1994 ) and in the Dpp signal transducers Mad and Med (Mad⁶, Mad¹¹, Mad¹², and Med¹; NEWFELD et al. 1997 ; WISOTZKEY et al. 1998 ). All of these mutations are dominant maternal enhancers of dpp recessive embryonic lethal alleles. Second, we tested for interactions with screw (scw), a gene encoding a TGF-ß family member that augments dpp signaling during dorsal-ventral patterning of the embryo (NGUYEN et al. 1998 ). We used a gain-of-function mutation (scw^E1) that is a dominant zygotic enhancer of dpp recessive embryonic lethal alleles and a loss-of-function mutation (scw^S12) that does not interact with any dpp alleles (RAFTERY et al. 1995 ). Third, we tested for genetic interactions with a loss-of-function allele of glass bottom boat-60A (gbb¹). gbb encodes a TGF-ß family member that cooperates with Dpp to specify positional information in imaginal disks (KHALSA et al. 1998 ). gbb is not involved in embryonic dorsal-ventral patterning (WHARTON et al. 1999 ). To date, no interactions between gbb mutations and dpp recessive embryonic lethal alleles have been reported.

Characterization of DTD46.4:
DTD46.4 is a recessive lethal strain obtained in our screen that has a T(2;3) 23C; 93F rearrangement (Fig 2A). To determine which translocation breakpoint results in the recessive lethality, DTD46.4-bearing flies were mated to flies with deletions spanning one of the two breakpoints. DTD46.4 complemented Df(3R)e-N19, a deletion of 93B-94. DTD-46.4 failed to complement Df(2L)JS17, a deletion spanning cytological region 23C-D that includes Mad. Mad is known to act as a dpp transvection disrupter (SEKELSKY et al. 1995 ), so we suspected that DTD46.4 might be a new allele of Mad. To test this hypothesis we chose to further characterize DTD46.4.

Complementation tests were conducted with a number of deficiencies and other mutations in the 23C-D cytological region (Fig 2B). The DTD46.4 chromosome failed to complement the deficiencies Df(2L)C144, Df(2L)DTD52xD51, and Df(2L)JS17 and an EMS-induced loss-of-function mutation l(2)a16. These five strains are referred to as the 23C complementation group. However, the DTD46.4 chromosome was viable over Mad⁶, Mad¹¹, and Mad¹² and the small deletion Df(2L)C28 that uncovers Mad. These results place the recessive lethality of DTD46.4 distal to Mad in 23C1-2. Polytene in situ hybridization studies utilizing a variety of probes demonstrated that the Drosophila Genome Project P1 clones DS00906 and DS07149 span the 23C1-2 breakpoint (data not shown).

We wanted to determine if the 23C1-2 breakpoint of DTD46.4 was also responsible for disrupting the dpp^d-ho/dpp^hr4 transvection-dependent phenotype. We tested Df(2L)C144 and l(2)a16 for the ability to disrupt this phenotype. Forty-six percent of dpp^d-ho Df(2L) c144 /dpp^hr4 flies had heldout wings; of these flies, 47% were severely heldout. Twenty percent of dpp^d-ho l(2)a16/dpp^hr4 flies had heldout wings; of these flies, 50% were severely heldout. These results are similar to those of DTD46.4. Twenty-six percent of dpp^d-ho DTD46.4/dpp^hr4 flies had heldout wings; of these flies, 53% were severely heldout. We conclude that the site of DTD46.4 recessive lethality in 23C1-2 is also the site that disrupts the dpp^d-ho/dpp^hr4 transvection-dependent phenotype.

During the course of this study we became aware of a new gene located in cytological region 23C1-2. This gene, lilliputian (lilli), was identified in two screens for Ras/Mitogen-activated protein kinase (MAPK) signal transduction pathway components. In these screens, loss-of-function mutations in lilli were identified as suppressors of gain-of-function phenotypes of seven in absentia (SS2-1; NEUFELD et al. 1998 ) and as suppressors of gain-of-function phenotypes of yan (SY2-1; REBAY et al. 2000 ). Complementation tests showed that both DTD46.4 and l(2)a16 failed to complement either lilli^s35 (NEUFELD et al. 1998 ) or lilli^xs407 (REBAY et al. 2000 ). We conclude that members of our 23C1-2 complementation group are alleles of lilli. In addition, a screen for genes that interact with dRaf, another component of MAPK signaling pathways, identified a locus in 23C1-2 (DICKSON et al. 1996 ). Loss-of-function mutations in Su(Raf)2A suppress gain-of-function dRaf phenotypes. It seems likely that Su(Raf)2A mutations are also allelic to DTD46.4 and lilli.

We tested four lilli alleles for dominant maternal enhancement of dpp recessive embryonic lethality. We excluded Df(2L)JS17 because it uncovers Mad. We tested the lilli alleles with dpp^e87, dpp^hr56, dpp^hr4, and dpp^hr92 (ST. JOHNSTON et al. 1990 ). No genetic interactions were detected with the weak alleles dpp^e87 and dpp^hr56 (data not shown). However, all lilli alleles tested showed significant dominant maternal enhancement of the strong alleles dpp^hr4 (Fig 3A) and dpp^hr92 (Fig 3B). Modest dominant zygotic enhancement of dpp^hr4 was also detected (Fig 3A). Thus, lilli alleles that disrupt a dpp transvection-dependent phenotype are also dominant enhancers of dpp recessive embryonic lethality.

The same alleles of lilli were tested for genetic interactions with other genes that function in dpp signaling. lilli alleles did not enhance the recessive lethality of the loss-of-function mutations Mad¹², Med¹, sax¹, tkv⁸, scw^S12, or gbb¹. However, lilli alleles showed dominant maternal enhancement of the recessive lethality of scw^E1 (Fig 3C). scw^E1 is a gain-of-function allele that is itself a dominant zygotic enhancer of dpp recessive embryonic lethality (RAFTERY et al. 1995 ).

We then determined the stage of lethality for the lilli loss-of-function mutation l(2)a16. We identified lilli mutant individuals [l(2)a16/Df(2L)C144] using the dominant visible marker Black cells (Bc; FLYBASE 1999 ). When l(2)a16/In(2LR)Gla Bc males were mated with Df(2L)C144/In(2LR)Gla Bc females, only Bc larvae were recovered (data not shown). Bc is not visible in first instar larvae, suggesting that lilli mutants were dying as embryos or as first instar larvae. Examination of lilli mutant embryos revealed a partially ventralized phenotype (Fig 4). This phenotype is also seen in zygotic mutant embryos of dpp^hr56 (WHARTON et al. 1993 ) and scw^E1 (RAFTERY et al. 1995 ). Several of the hallmarks of this phenotype are a herniated head, internalized filzkorper, and disorganized/expanded denticle bands. Embryos derived from germline clones of weak Su(Raf)2A mutations (e.g., Su(Raf)2A^161H1) also show this partially ventralized phenotype (DICKSON et al. 1996 ).

Phylogenetic analysis of Lilli:
The sequence of a full-length lilli cDNA has recently been identified (A. TANG, personal communication). A nearly identical protein of 1665 amino acids, except for an 8-amino-acid truncation at the N terminus, was predicted from genomic sequence by the Berkeley Drosophila Genome Project (GenBank accession no. AAF51180; ADAMS et al. 2000 ). BLAST searches using arbitrarily defined segments of the predicted Lilli protein identified very similar regions in four human proteins. These proteins belong to a multigene family called the FMR2/LAF4 family (GECZ et al. 1997 ; NILSON et al. 1997 ).

FMR2 was identified via mutations that result in Fragile-X mental retardation syndrome. Fragile X mental retardation syndrome is the most common form of inherited mental retardation in humans. FMR2 is highly expressed in the fetal brain (reviewed in JIN and WARREN 2000 ). LAF4 was identified via chromosomal translocations that result in Burkitt's lymphoma. Burkitt's lymphoma is associated with highly malignant tumors and is the most common form of childhood cancer. LAF4 is highly expressed in fetal lymphoid tissue, particularly in preB-cells (MA and STAUDT 1996 ). The other family members, AF4 and AF5, were identified via distinct chromosomal translocations that give rise to infant acute lymphoblastic leukemia (ALL). At this time, ALL is resistant to treatment and invariably fatal. AF5 is highly expressed in fetal heart, lung, and brain while AF4 is highly expressed in fetal heart, liver, and brain (LI et al. 1998 ; TAKI et al. 1999 ). These human proteins are nuclear proteins capable of DNA binding and transcriptional activation (LI et al. 1998 ).

Previous studies of this family identified three conserved domains (GECZ et al. 1997 ; Fig 5A). Near the N terminus there is a conserved domain that includes a high mobility group I (HMGI) DNA-binding motif. In the center there is a conserved transcriptional activation domain with no recognizable motif. At the C terminus there is a highly conserved domain diagnostic for the FMR2/LAF4 family with no recognizable motif and unknown function. BLAST searches showed that Lilli contains segments very similar to each of these domains in the proper location.

We conducted an exhaustive analysis of the D. melanogaster genome database using the conserved regions of Lilli and the four human FMR2/LAF4 family sequences. A total of 15 different domains were used as query sequences. We did not identify any additional proteins that contain all 3 conserved domains. Nor did we identify a group of consecutive (mis)predicted proteins that contain the 3 conserved domains in the proper order. We were not able to identify any additional proteins with obvious similarity to only the C-terminal domain diagnostic for the FMR2/LAF4 family. At this time, Lilli appears to be the only D. melanogaster member of this multigene family. We then conducted the same set of exhaustive searches using the C. elegans genome database. We did not identify any proteins with all three domains or any with convincing similarity to the C-terminal diagnostic domain.

An alignment of the C-terminal domain of Lilli with all of the human family members is shown in Fig 5B. This region of Lilli shows extensive amino acid similarity with all of the human proteins. However, the alignment gives the overall impression that the four human family members are more similar to each other than they are to Lilli. The degree of amino acid identity and similarity, calculated from pairwise comparisons between all five sequences for each of the conserved domains, is shown in Table 2. The comparisons show that there is a significant amount of amino acid similarity (>51%) between Lilli and each human protein in all domains. The human proteins show >63% similarity in all domains with most comparisons >72%.

Data derived from pairwise comparisons were used to construct phylogenetic trees for each domain. A composite tree was also constructed from an alignment consisting of all three domains (Fig 6). Only slight differences were noted between the individual domain trees and the composite tree. The similarity of the trees suggests that the tripartite structure of these proteins predates the divergence of arthropods and vertebrates. The composite tree shows that the human family members are indeed more similar to each other than they are to Lilli. This distinction is 100% supported by the bootstrap analysis. The composite tree contains two clusters of human sequences that are also strongly supported. Sequence clusters with bootstrap values >75% are considered biologically meaningful (NEWFELD et al. 1999 ).

View larger version (10K): In this window In a new window Download PPT slide   Figure 6. Phylogenetic analysis of FMR2/LAF4 family members. Evolutionary relationships between FMR2/LAF4 family members, based on a composite alignment of the three conserved domains, are shown. The length of the alignment was 522 amino acids. The tree is unrooted. The numbers represent the relative incidence of that particular relationship (in percentage) during bootstrap resampling using 1000 replicates. Branch lengths are drawn to scale on the basis of the number of amino acid substitutions per site.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Exceptional DTD screen:
We conducted a genetic screen for new components of the dpp signaling pathway. The screen identified 30 exceptional DTDs. These mutations disrupt transvection at the dpp locus but are not associated with asynapsis at dpp. Mutations were not recovered in genes involved in dpp signaling that act as exceptional DTDs, such as Mad (SEKELSKY et al. 1995 ), suggesting that our screen was not exhaustive.

To determine if any of the exceptional DTDs were associated with mutations in dpp signaling pathway components, we utilized three assays. These are the same tests used in the initial characterization of the Dpp signal transducers Mad and Med (RAFTERY et al. 1995 ; SEKELSKY et al. 1995 ). First, we tested each DTD for genetic interactions with dpp alleles that were not part of the original screen. The original screen exploited dpp's role in adult appendage formation. In this test we examined each DTD for dominant maternal enhancement of dpp alleles that disrupt embryonic dorsal-ventral patterning. Second, we tested each DTD for genetic interactions with other genes that participate in dpp signaling such as sax, scw, and Mad. Third, we looked for similarities between the mutant phenotype of a DTD, or another member of its complementation group, and dpp mutant phenotypes.

Characterization of DTD46.4:
The first mutation we chose to characterize in detail was the 23C1-2 breakpoint of DTD46.4. This breakpoint is allelic to mutations in the newly discovered gene lilli. The results of our genetic tests suggest that lilli is a strong candidate for a new component of the dpp signaling pathway. First, lilli mutations enhance dpp heldout phenotypes and embryonic recessive lethality. The enhancement of dpp embryonic lethality by lilli mutations is not as strong as that of Mad or Med mutations (RAFTERY et al. 1995 ; SEKELSKY et al. 1995 ). Mutations in Mad or Med enhance weak dpp alleles while lilli mutations do not. Second, lilli mutations enhance the recessive embryonic lethality of a gain-of-function allele of the TGF-ß family member scw. scw augments dpp signaling in embryonic dorsal-ventral patterning. To date, tests for interactions between scw^E1 and other dpp pathway components such as Mad or Med have not been reported. lilli mutations do not enhance the recessive lethality of mutations in genes that encode Dpp signal transduction proteins (sax, tkv, Mad, or Med). Third, lilli homozygous mutant embryos have dorsal-ventral patterning defects similar to zygotic mutant embryos of dpp and scw. Utilizing these genetic criteria, lilli has as strong a connection to dpp signaling as Mad and Med.

In addition to our screen, lilli mutations were identified in three other screens. In these screens, lilli mutations suppress dominant phenotypes generated by activated MAPK signaling pathways (DICKSON et al. 1996 ; NEUFELD et al. 1998 ; REBAY et al. 2000 ). MAPK signal transduction is initiated by transmembrane receptor tyrosine kinases. These receptors transmit the signal to transcription factors in the nucleus utilizing a cascade of tyrosine kinases. Alternatively, TGF-ß family members such as Dpp bind to transmembrane receptor serine-threonine kinases. These receptors transmit the Dpp signal via a nonkinase mechanism of nuclear translocation by Mad and Med. The ability of lilli loss-of-function mutations to suppress MAPK signaling gain-of-function phenotypes and to enhance dpp loss-of-function phenotypes is very intriguing. To our knowledge, lilli is one of the first genes involved in MAPK and TGF-ß signaling pathways in a developmental system.

lilli encodes a transcription factor (A. TANG, personal communication). This fact suggests one hypothesis for lilli's role in MAPK signaling and another hypothesis for a role in Dpp signaling. For MAPK signaling, lilli may be a transcriptional effector of MAPK signal transduction pathways. This hypothesis fits the observation that lilli loss-of-function mutations suppress MAPK signaling gain-of-function phenotypes (DICKSON et al. 1996 ; NEUFELD et al. 1998 ; REBAY et al. 2000 ). For Dpp signaling, lilli may be a maternally supplied transcriptional activator of dpp and/or scw during dorsal-ventral patterning. This hypothesis fits three of our observations. First, lilli loss-of-function mutations maternally enhance the recessive lethality of several dorsal-ventral patterning mutations. Second, lilli mutant phenotypes mimic the mutant phenotypes of dorsal-ventral patterning mutations. Third, lilli mutations do not enhance, either maternally or zygotically, the embryonic lethality of genes that encode Dpp signal transduction proteins. Alternatively, lilli could participate in a signaling pathway parallel to the Dpp pathway that is also required for the expression of Dpp target genes.

To test the hypothesis that lilli is a maternal activator of dpp and/or scw in dorsal-ventral patterning one would examine dpp and scw expression in embryos derived from lilli mutant germline clones. The prediction is that there would be reduced dpp and/or scw expression in these embryos during dorsal-ventral patterning. At this time, maternal activators of zygotic dorsal-ventral patterning genes such as dpp and scw, as opposed to well-known repressors such as Dorsal (ANDERSON 1998 ), are unknown. It is tempting to speculate that a maternal MAPK signal induces lilli to activate dpp in embryonic dorsal-ventral patterning.

Determining a role for lilli in dpp signaling in adult wings, where lilli mutations enhance the heldout phenotype, is more problematic. There is no a priori reason to believe that lilli plays the same role in dpp signaling during dorsal-ventral patterning and adult appendage formation but it seems a logical place to begin. Thus it is possible that lilli activates dpp expression in wing imaginal disks. This hypothesis fits a report of dpp transcriptional regulation by the heldout cis-regulatory region (HEPKER et al. 1999 ). In this study, two consensus HMGI binding sites (A/TA/TCAAG; VAN DE WETERING et al. 1991 ) are identified as dTcf binding sites in the heldout region. The expression of reporter genes containing the dpp heldout region was disrupted when these putative dTcf sites were mutagenized. In addition, dominant negative forms of dTcf expressed in wing imaginal disks eliminated dpp expression. As a result, the authors conclude that dTcf is required for dpp expression by the heldout cis-regulatory region. However, these data do not preclude the possibility that the HMGI binding sites are actually the target of another HMGI domain protein, such as Lilli. To determine which HMGI domain protein is actually responsible for dpp expression from heldout regulatory sequences, one would examine dpp expression in wing imaginal disks bearing dTcf or lilli somatic clones.

Phylogenetic analysis of Lilli:
Lilli shares three conserved domains with the four human members of the FMR2/LAF4 multigene family of transcription factors. The human family members are all developmental genes with high levels of fetal tissue-specific expression. Mutations in these genes have devastating effects. Mutations in FMR2 lead to mental retardation and mutations in LAF4, AF4, and AF5 lead to treatment-resistant forms of childhood cancer. Our analyses revealed several interesting features of this newly expanded multigene family.

First, BLAST searches demonstrate that Lilli is unique among D. melanogaster genes. No other sequences with convincing similarity to any FMR2/LAF4 family member were found in the D. melanogaster genome. We found this surprising for a gene associated with dpp signaling. To date, all known components of dpp signaling pathways belong to large multigene families with several members in D. melanogaster (NEWFELD et al. 1999 ). Second, pairwise amino acid comparisons suggested that the human genes in the FMR2/LAF4 family are more similar to each other than to Lilli. This suggestion is supported with a 100% bootstrap value by phylogenetic analysis. The basal branch of the tree separates Lilli from the human genes.

Taken together, these two observations strongly support the hypothesis that lilli is the D. melanogaster homolog of the human FMR2/LAF4 family members. We employ the strict evolutionary definition of homology (genes identical by descent from a common ancestor). In this case, we refer to the FMR2/LAF4 family progenitor in the common ancestor of arthropods and chordates. The absence of any FMR2/LAF4 family members in C. elegans suggests that the FMR2/LAF4 family progenitor arose after the split of nematodes and arthropods.

In addition, the appearance of two pairs of sequences for the four human FMR2/LAF4 family members in the phylogenetic tree is compatible with OHNO's (1970) hypothesis that two rounds of genome duplication have occurred in the vertebrate lineage. The original member of each pair of human sequences could have been generated during the first event and the second member of each pair by the second event. Alternatively, the four human sequences could have been generated by three independent gene duplication events. Additional phylogenetic data are needed to distinguish these hypotheses.

In summary, Lilli appears to function in both MAPK and Dpp signaling pathways, suggesting important roles in Drosophila development. Detailed studies of Lilli function in Drosophila will likely shed light on the wild-type function of human FMR2/LAF4 family members. For example, the functional conservation of dpp signaling pathway components suggests that human homologs of Lilli's transcriptional targets are likely to be targets of human FMR2/LAF4 family members. Given that mutations in these human genes lead to mental retardation or childhood cancer and that information on human developmental genes is difficult to gather directly, studies of Lilli are an important weapon in our efforts to combat these human syndromes.

	FOOTNOTES

¹ Present address: Department of Pediatrics, University of California, San Francisco, CA 94143.

	ACKNOWLEDGMENTS

We thank Amy Tang for providing the lilli cDNA sequence prior to publication and Sudhir Kumar for assistance in generating phylogenetic trees. This study was begun in Bill Gelbart's lab and we thank Bill and Jeff Sekelsky for assistance with cytology. We thank Mike O'Connor for valuable discussions. S.J.N. is supported by a Basil O'Connor Starter Scholar Research Award from the March of Dimes.

Manuscript received July 26, 2000; Accepted for publication November 3, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADAMS, M., S. CELNIKER, R. HOLT, C. EVANS, and J. GOCAYNE et al., 2000  The genome sequence of Drosophila melanogaster. Science 287:2185-2194[Abstract/Free Full Text].

ANDERSON, K., 1998  Pinning down positional information: dorsal-ventral polarity in the Drosophila embryo. Cell 95:439-442[Medline].

BRUMMEL, T., V. TWOMBLY, G. MARQUES, J. WRANA, and S. NEWFELD et al., 1994  Characterization and relationship of Dpp receptors encoded by the saxophone and thickveins genes in Drosophila.. Cell 78:251-262[Medline].

DICKSON, B., A. VAN DER STRATEN, M. DOMINGUEZ, and E. HAFEN, 1996  Mutations modulating Raf signaling in Drosophila eye development. Genetics 142:163-171[Abstract].

FELSENSTEIN, J., 1985  Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:783-791.

FLYBASE,, 1999  The FlyBase database of the Drosophila genome projects and community literature. Nucleic Acids Res. 27:85-88[Abstract/Free Full Text].

GECZ, J., S. BIELBY, G. SUTHERLAND, and J. MULLEY, 1997  Gene structure and subcellular localization of FMR2, a member of a new family of putative transcription activators. Genomics 44:201-213[Medline].

GELBART, W., 1982  Synapsis-dependent allelic complementation at the decapentaplegic gene complex in Drosophila melanogaster. Proc. Natl. Acad. Sci. USA 79:2636-2640[Abstract/Free Full Text].

GELBART, W., 1989  The decapentaplegic gene: a TGF-ß homolog controlling pattern formation in Drosophila. Development 107(Suppl.):65-74.

HEPKER, J., R. BLACKMAN, and R. HOLMGREN, 1999  Cubitus interruptus is necessary but not sufficient for direct activation of a wing-specific decapentaplegic enhancer. Development 126:3669-3677[Abstract].

JEANMOUGIN, F., J. THOMPSON, M. GOUY, D. HIGGINS, and T. GIBSON, 1998  Multiple sequence alignment with Clustal X. Trends Biochem. Sci. 23:403-405[Medline].

JIN, P. and S. WARREN, 2000  Understanding the molecular basis of Fragile X syndrome. Hum. Mol. Genet. 9:901-908[Abstract/Free Full Text].

KHALSA, O., J. YOON, S. TORRES-SCHUMANN, and K. WHARTON, 1998  TGF-ß/BMP superfamily members, Gbb-60A and Dpp, cooperate to provide pattern information and establish cell identity in the Drosophila wing. Development 125:2723-2734[Abstract].

KUMAR, S., K. TAMURA and M. NEI, 1993 MEGA: Molecular Evolutionary Genetic Analysis. Pennsylvania State University Press, University Park, PA.

LEWIS, E., 1954  The theory and application of a new method of detecting chromosomal rearrangements in Drosophila melanogaster. Am. Nat. 88:225-239.

LI, Q., J. FRESTEDT, and J. KERSEY, 1998  AF4 encodes a ubiquitous protein that in both native and MLL-AF4 fusion types localizes to subnuclear compartments. Blood 92:3841-3847[Abstract/Free Full Text].

MA, C. and L. STAUDT, 1996  LAF4 encodes a lymphoid nuclear protein with transactivation potential that is homologous to AF4, the gene fused to MLL in t(4;11) leukemias. Blood 87:743-745.

NEI, M., and S. KUMAR, 2000 Molecular Evolution and Phylogenetics. Oxford University Press, New York.

NEUFELD, T., A. TANG, and G. RUBIN, 1998  A genetic screen to identify components of the sina signaling pathway in Drosophila eye development. Genetics 148:277-286[Abstract/Free Full Text].

NEWFELD, S., A. MEHRA, M. SINGER, J. WRANA, and L. ATTISANO et al., 1997  Mad participates in a DPP/TGF-ß responsive serine-threonine kinase signal transduction cascade. Development 124:3167-3176[Abstract].

NEWFELD, S., R. WISOTZKEY, and S. KUMAR, 1999  Molecular evolution of a developmental pathway: phylogenetic analyses of TGF-ß family ligands, receptors and Smad signal transducers. Genetics 152:783-795[Abstract/Free Full Text].

NGUYEN, M., S. PARK, G. MARQUES, and K. ARORA, 1998  Interpretation of a BMP activity gradient in Drosophila embryos depends on synergystic signaling by two Type I receptors, SAX and TKV. Cell 95:495-506[Medline].

NILSON, I., M. REICHEL, M. ENNAS, R. GREIM, and C. KNORR et al., 1997  Exon/intron structure of the human AF4 gene, a member of the AF4/LAF4/FMR2 gene family coding for a nuclear protein with structural alterations in acute leukaemia. Br. J. Haematol. 98:157-169[Medline].

OHNO, S., 1970 Evolution by Gene Duplication. Springer-Verlag, New York.

PADGETT, R., R. ST. JOHNSTON, and W. GELBART, 1987  A transcript from a Drosophila pattern gene predicts a protein homologous to the TGF-ß family. Nature 325:81-84[Medline].

RAFTERY, L., V. TWOMBLY, K. WHARTON, and W. GELBART, 1995  Genetic screens to identify elements of the dpp pathway in Drosophila. Genetics 139:241-254[Abstract].

REBAY, I., F. CHEN, F. HSIAO, P. KOLODZIEJ, and B. KUANG et al., 2000  A genetic screen for novel components of the Ras/Mitogen-activated protein kinase signaling pathway that interact with the yan gene of Drosophila identifies split ends, a new RNA recognition motif-containing protein. Genetics 154:695-712[Abstract/Free Full Text].

RIGGINS, G., K. KINZLER, B. VOGELSTEIN, and S. THIAGALINGAM, 1997  Frequency of Smad gene mutations in human cancers. Cancer Res. 57:2578-2580[Abstract/Free Full Text].

SAITOU, N. and M. NEI, 1987  The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406-425[Abstract].

SCHULER, G., S. ALTSCHUL, and D. LIPMAN, 1991  A workbench for multiple alignment construction and analysis. Proteins 9:180-190[Medline].

SEKELSKY, J., 1993 Mothers against dpp: a gene required for decapentaplegic function in Drosophila melanogaster. Ph.D. Thesis, Harvard University, Cambridge, MA.

SEKELSKY, J., S. NEWFELD, L. RAFTERY, E. CHARTOFF, and W. GELBART, 1995  Genetic characterization and cloning of Mad, a gene required for dpp function in Drosophila. Genetics 139:1347-1358[Abstract].

SMITH, R. and T. SMITH, 1990  Automated generation of primary sequence patterns from sets of related protein sequences. Proc. Natl. Acad. Sci. USA 87:118-122[Abstract/Free Full Text].

SPENCER, F., F. HOFFMANN, and W. GELBART, 1982  decapentaplegic: a gene complex affecting morphogenesis in Drosophila melanogaster. Cell 28:451-461[Medline].

ST. JOHNSTON, R., F. HOFFMANN, R. BLACKMAN, D. SEGAL, and R. GRIMAILA et al., 1990  Molecular organization of the dpp gene in Drosophila. Genes Dev. 4:1114-1127[Abstract/Free Full Text].

TAKI, T., H. KANO, M. TANIWAKI, M. SAKO, and M. YANAGISAWA et al., 1999  AF5q31, a newly identified AF4 related gene, is fused to MLL in infant acute lymphoblastic leukemia with ins (5;11)(q31;q13q23). Proc. Natl. Acad. Sci. USA 96:14535-14540[Abstract/Free Full Text].

VAN DE WETERING, M., M. OOSTERWEGEL, D. DOOIJES, and H. CLEVERS, 1991  Identification and cloning of TCF1, a T lymphocyte-specific transcription factor containing a sequence specific HMG box. EMBO J. 10:123-132[Medline].

WALDRIP, W., E. BIKOFF, P. HOODLESS, J. WRANA, and E. ROBERTSON, 1998  Smad2 signaling in extraembryonic tissues determines anterior-posterior polarity of the early mouse embryo. Cell 92:797-808[Medline].

WHARTON, K., R. RAY, and W. GELBART, 1993  An activity gradient of decapentaplegic is necessary for the specification of dorsal pattern elements in the Drosophila embryo. Development 117:807-822[Abstract].

WHARTON, K., R. RAY, S. FINDLEY, H. DUNCAN, and W. GELBART, 1996  Molecular lesions associated with alleles of decapentaplegic identify residues necessary for TGF-ß/BMP cell signaling in Drosophila melanogaster. Genetics 142:493-505[Abstract].

WHARTON, K., J. COOK, S. TORRES-SCHULMANN, K. DE CASTRO, and E. BOROD et al., 1999  Genetic analysis of the Bone Morphogenetic Protein-related gene, gbb, identified multiple requirements during Drosophila development. Genetics 152:629-640[Abstract/Free Full Text].

WISOTZKEY, R., A. MEHRA, D. SUTHERLAND, L. DOBENS, and X. LIU et al., 1998  Medea is a Drosophila Smad4 homolog that is differentially required to potentiate Dpp responses. Development 125:1433-1445[Abstract].

This article has been cited by other articles:

				A. Khokhar, N. Chen, J.-P. Yuan, Y. Li, G. N. Landis, G. Beaulieu, H. Kaur, and J. Tower Conditional Switches for Extracellular Matrix Patterning in Drosophila melanogaster Genetics, March 1, 2008; 178(3): 1283 - 1293. [Abstract] [Full Text] [PDF]

				F. Bejarano, C. M. Luque, H. Herranz, G. Sorrosal, N. Rafel, T. T. Pham, and M. Milan A Gain-of-Function Suppressor Screen for Genes Involved in Dorsal-Ventral Boundary Formation in the Drosophila Wing Genetics, January 1, 2008; 178(1): 307 - 323. [Abstract] [Full Text] [PDF]

				K. L. Pepple, A. E. Anderson, B. J. Frankfort, and G. Mardon A Genetic Screen in Drosophila for Genes Interacting With senseless During Neuronal Development Identifies the Importin moleskin Genetics, January 1, 2007; 175(1): 125 - 141. [Abstract] [Full Text] [PDF]

				N. T. Takaesu, C. Hyman-Walsh, Y. Ye, R. G. Wisotzkey, M. J. Stinchfield, M. B. O'Connor, D. Wotton, and S. J. Newfeld dSno Facilitates Baboon Signaling in the Drosophila Brain by Switching the Affinity of Medea Away From Mad and Toward dSmad2 Genetics, November 1, 2006; 174(3): 1299 - 1313. [Abstract] [Full Text] [PDF]

				N. T. Takaesu, E. Herbig, D. Zhitomersky, M. B. O'Connor, and S. J. Newfeld DNA-binding domain mutations in SMAD genes yield dominant-negative proteins or a neomorphic protein that can activate WG target genes in Drosophila Development, November 1, 2005; 132(21): 4883 - 4894. [Abstract] [Full Text] [PDF]

				R. DasGupta, A. Kaykas, R. T. Moon, and N. Perrimon Functional Genomic Analysis of the Wnt-Wingless Signaling Pathway Science, May 6, 2005; 308(5723): 826 - 833. [Abstract] [Full Text] [PDF]

				S. Luschnig, B. Moussian, J. Krauss, I. Desjeux, J. Perkovic, and C. Nusslein-Volhard An F1 Genetic Screen for Maternal-Effect Mutations Affecting Embryonic Pattern Formation in Drosophila melanogaster Genetics, May 1, 2004; 167(1): 325 - 342. [Abstract] [Full Text] [PDF]

				S. Szuplewski, B. Kottler, and R. Terracol The Drosophila bZIP transcription factor Vrille is involved in hair and cell growth Development, August 15, 2003; 130(16): 3651 - 3662. [Abstract] [Full Text] [PDF]