An in Vivo Analysis of the vestigial Gene in Drosophila melanogaster Defines the Domains Required for Vg Function

Corresponding author: John B. Bell, University of Alberta, Edmonton, AB T6G 2E9, Canada., jbell{at}ualberta.ca (E-mail)

Communicating editor: T. SCHÜPBACH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Considerable evidence indicates an obligate partnership of the Drosophila melanogaster Vestigial (VG) and Scalloped (SD) proteins within the context of wing development. These two proteins interact physically and a 56-amino-acid motif within VG is necessary and sufficient for this binding. While the importance of this SD-binding domain has been clearly demonstrated both in vitro and in vivo, the remaining portions of VG have not been examined for in vivo function. Herein, additional regions within VG were tested for possible in vivo functions. The results identify two additional domains that must be present for optimal VG function as measured by the loss of ability to rescue vg mutants, to induce ectopic sd expression, and to perform other normal VG functions when they are deleted. An in vivo study such as this one is fundamentally important because it identifies domains of VG that are necessary in the cellular context in which wing development actually occurs. The results also indicate that an additional large portion of VG, outside of these two domains and the SD-binding domain, is dispensable in the execution of these normal VG functions.

PROPER development of the wing in Drosophila melanogaster involves the coordinate action of several genes. One of the central genes involved in this process is vestigial (vg). The vg gene is a target of both the decapentaplegic (dpp) and Notch (N)/wingless (wg) pathways and patterned vg expression occurs throughout the entire wing field (WILLIAMS et al. 1991 , WILLIAMS et al. 1993 , WILLIAMS et al. 1994 ; KIM et al. 1996 ). In the embryo, vg is expressed transiently in some neural and muscle precursor cells, but the majority of expression occurs within the wing and haltere primordia (WILLIAMS et al. 1991 ). As expected, the pattern and amount of VG found in the wing disc varies between vg mutants (WILLIAMS et al. 1991 ; SIMMONDS et al. 1997 ), and both of these factors determine the severity of the mutant wing phenotype in the adults. For example, the vg^83b27 mutant phenotype is due to loss of the vg boundary enhancer (WILLIAMS et al. 1991 , WILLIAMS et al. 1993 ). In this case, the normal elevation of vg expression does not occur along the dorsal-ventral boundary of the wing disc and wings are absent. The vg¹ mutant is caused by an insertion of a 412-transposon element, which allows more wing tissue to be formed than in vg^83b27 flies (WILLIAMS and BELL 1988 ). The vg^83b27R mutant was derived from the vg^83b27 allele by gamma-ray mutagenesis (WILLIAMS et al. 1990 ) and behaves as a vg null because the 3' exons are deleted. Expression assays using vg transgenes (KIM et al. 1996 ; HALDER et al. 1998 ; SIMMONDS et al. 1998 ) have shown that vg exhibits selector gene characteristics in that it is able to reprogram cells to adopt a wing-like fate when expressed ectopically. Thus, vg is involved specifically with, and required for, wing development. However, the molecular function of the VG protein within the nucleus was only recently elucidated (HALDER et al. 1998 ; PAUMARD-RIGAL et al. 1998 ; SIMMONDS et al. 1998 ).

Genetically, vg is known to interact with scalloped (sd), another important wing-patterning gene. Indeed, vg must act in partnership with sd to manifest selector gene ability (HALDER et al. 1998 ; SIMMONDS et al. 1998 ; GUSS et al. 2001 ; HALDER and CARROLL 2001 ). The expression pattern of sd in the wing imaginal disc (CAMPBELL et al. 1991 , CAMPBELL et al. 1992 ; WILLIAMS et al. 1991 , WILLIAMS et al. 1993 ) is almost identical to that of vg (WILLIAMS et al. 1993 ). However, sd is more widely expressed in other developing tissues compared with the restriction of vg expression to the wing disc. Accordingly, some sd mutants result in phenotypes similar to vg mutants (CAMPBELL et al. 1992 ) while other sd mutants are lethal as homozygotes. In vitro binding experiments have demonstrated a protein-protein interaction between SD and VG, effected via a 56-amino-acid region in the VG protein (SIMMONDS et al. 1998 ). The overlapping expression patterns of sd and vg in wing discs, the nuclear localization of both proteins, and this protein-protein interaction support the idea that VG and SD act in partnership during wing development. HALDER et al. 1998 used ectopic expression of vg to show that the sd gene is upregulated wherever vg is able to successfully induce ectopic wing structures. In sd mutants, the ability of vg to ectopically induce wing tissue is greatly reduced or nonexistent (SIMMONDS et al. 1998 ). As well, when vg is expressed in the absence of SD or when it is deleted for the SD-binding domain, VG is detected primarily in the cytoplasm. In contrast, when SD is present, VG is found primarily within the nucleus. Recently, it has been shown that expression of only the sd TEA domain fused to a full-length vg gene is sufficient to accomplish the nuclear localization of VG (SRIVASTAVA et al. 2002 ). It has also been shown that the association of VG with SD is essential for the activation of target genes, including sd itself (HALDER et al. 1998 ; SIMMONDS et al. 1998 ). Thus, it is evident that SD and VG act together as a transcriptional complex during wing formation, wherein SD provides the DNA-binding activity (CAMPBELL et al. 1992 ; SRIVASTAVA et al. 2002 ) and nuclear localization signal (SRIVASTAVA et al. 2002 ), while VG provides the activation function. Further, other evidence suggests that this complex provides a qualitatively distinct function required to generate a wing-specific response to globally important signaling pathways (GUSS et al. 2001 ) and the presence of VG complexed with SD changes the DNA-binding specificity of SD (HALDER and CARROLL 2001 ).

While the importance of the SD-binding domain of VG has been demonstrated clearly both in vitro and in vivo (HALDER et al. 1998 ; SIMMONDS et al. 1998 ), the remaining portions of VG have not been tested for in vivo function. Herein, additional regions within VG have been tested in vivo to assess their respective ability to rescue three vg mutations (vg^83b27, vg^83b27R, and vg¹), to ectopically induce sd expression, and also to perform other functions associated with full-length VG. The identification of vertebrate homologs of vg, tondu (VAUDIN et al. 1999 ), and fondue (HALDER and CARROLL 2001 ) suggested the presence of potential activation domains in VG. Using one-hybrid assays in yeast, potential activation domains were found in two regions, one at the N-terminal and another at the C-terminal region of the protein (VAUDIN et al. 1999 ). Furthermore, in vitro DNA-binding experiments of the VG/SD complex to native and synthetic targets also identify important regions of VG outside the SD-binding domain, which are necessary for efficient and accurate binding of the complex and thus for normal VG function (HALDER and CARROLL 2001 ). However, in vitro assays can identify regions that do not function in vivo (VIDAL and LEGRAIN 1999 ). Thus, we sought to test the functional properties of various VG regions in vivo. The present dissection of the vg open reading frame (ORF) also identifies important functional domains in addition to the SD-binding domain. The results corroborate the presence of these essential domains in Drosophila VG and demonstrate an in vivo role for these regions during VG-mediated wing patterning, including induction of the endogenous sd gene.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Drosophila stocks:
All stocks and crosses were raised at 25°. The vg¹ , vg^83b27, and vg^83b27R alleles were from our local collection. The ptcGal4 and sd-lacZ strains were a gift from S. Carroll.

Construction of vg deletions:
All deletions were constructed using inverse PCR in either pUC19 (VIEIRA and MESSING 1982 ) or pET16b (Novagen). The deletions tested were those from SIMMONDS et al. 1998 , with necessary modifications for in vivo expression of the encoded VG protein. Primer combinations were chosen so as to dissect the vg ORF into various fragments, which were then used in the assays discussed below. The vg constructs were cloned into pUAST (BRAND and PERRIMON 1993 ), using BamHI or BamHI and EcoRI restriction sites in microinjections. Each construct had a SpeI linker at the deletion junction. Thus each final product encoded not only the desired retained portion of the vg ORF, but also two consistent amino acids at the boundaries of each deletion.

Inverse PCR:
The following Taq:Pfu (20:1) PCR protocol with appropriate primer combinations was used: 94° for 5 min (hot start), followed by 35 cycles of 94° for 30 sec, 68° for 7 min, and holds at 72° for 3 min and then 4°. The total reaction mix was 30 µl, and the full-length vg ORF (in pET16b or pUC19) was used as template.

Micro-injections:
Micro-injections were performed as described in RUBIN and SPRADLING 1983 , using vg deletions in pUAST and helper 2-3. Two or more independent transgenic lines of each deletion construct were obtained and used in subsequent assays.

Ectopic expression assay:
The UAS-Gal4 system was utilized (BRAND and PERRIMON 1993 ). To express the UAS-vg constructs ectopically, each was crossed to sd-lacZ; ptcGal4 flies (driven by ptcGal4). The resulting larvae were stained with X-gal to assess the ability of ectopic vg to induce ectopic sd expression (assayed with a sd lacZ reporter). X-Gal staining was performed for 1.5 hr as described in BELLEN et al. 1989 .

Flp-out clones:
Flp-out clones (ITO et al. 1997 ; PAPAYANNOPOULOS et al. 1998 ) were induced in flies trans-heterozygous for an actin > y+ > GAL4(AyGAL4)UAS-GFP chromosome (Y. Hiromi) and the relevant vg deletion-bearing UAS constructs, according to the protocol described in LIU et al. 2000 . Respective larvae with the above genotype also carried a hs-FLP122 construct (STRUHL et al. 1993 ). Briefly, the clones were induced by heat shock and then allowed to grow for 48–72 hr at 25° after induction.

Antibody staining:
Antibody staining was performed as described in WILLIAMS et al. 1991 . VG antibody was obtained from S. Carroll and used at 1:400 dilution. WG antibody was obtained from Developmental Studies Hybridoma Bank and was used at a dilution of 1:10.

Rescue assays:
To assess the ability of each vg deletion construct to rescue each of three vg mutants (vg^83b27, vg^83b27R, or vg¹; WILLIAMS et al. 1990 ), the UAS-Gal4 system was used again. All UASvg lines were made homozygous and the corresponding chromosomal location of each insert was determined (data not shown). From these transgenic lines, UASvg; vg lines were generated and crossed to a vgGal4; vg line. Progeny from this cross were scored for the extent of wing tissue restoration. Since the parental lines were homozygous for either vg¹ or vg^83b27 or were heterozygous for vg^83b27R/Balancer, any increase in the amount of wing tissue in the homozygous offspring could be attributed to the function of the respective UASvg constructs.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The deletion constructs used herein are labeled according to the primer combinations that were used in their construction along with a symbol to designate that each is a partial deletion of the vg ORF. A schematic of the deletions made within the vg ORF and the corresponding deleted amino acids is shown in Fig 1. The shaded bars indicate regions that are still present while open bars indicate the regions that are deleted. The previously identified SD-binding motif of the VG protein (amino acids 281–335) is also indicated in Fig 1, and several of the constructs include this motif within the deleted portion.

View larger version (40K): In this window In a new window Download PPT slide   Figure 1. Deletions made within the vg ORF. All deletions are shown with respect to the full-length construct (453 amino acids). Solid bars represent intact regions and open bars designate areas that have been deleted. The corresponding deleted amino acids are indicated to the right of the bars and the construct names are indicated on the left. Construct names refer to the primer sets used in preparation of each construct, and F refers to a full-length vg construct. The known SD-binding domain (amino acids 281–335) is also shown (vertical shaded bar). The results of the ectopic sd expression assays as well as the wing phenotype rescue assays for vg83b27 and vg1 are summarized in the two columns on the right. A minus sign (-) indicates a negative result with respect to both assays, while plus (+) signs indicate positive results presented in a graded fashion, wherein the most positive response (+++) is that produced by wild-type vg constructs in the respective assays. The positions of the functional domains identified herein are indicated by solid bars above the protein map.

The ability of each UASvg transgene to induce sd was assessed by X-gal staining for the product of a sd-lacZ reporter in third instar larvae containing this reporter, the UASvg transgene, and a ptcGal4 driver. A sd-lacZ reporter has been shown previously to be activated by the presence of ectopic transgenic VG (SIMMONDS et al. 1998 ). This was used (Fig 2A) as a positive control for each of the deletion constructs. A negative control, where only the pUAST vector was transformed into flies, was also used (Fig 2B). To ensure that adequate levels of VG were being produced in the wing discs in these ectopic expression assays using the ptc-GAL4 driver, VG antibody staining was performed on larvae from UASvg X sd-lacZ; ptcGal4 crosses. While some constructs were deleted for the anti-VG epitope (4-10, 4-8, and 2-8), all constructs that encoded proteins that were amenable to antibody detection (full-length vg, 5'-5, 5'-6, 5'-7, 1-2, 1-4, and 7-8) showed ectopic expression of the vg transgene. A representative wing disc from this group (1-4) was utilized to detect VG with antibody staining and is shown in Fig 2C. N-terminal deletion constructs 5'-5, 5'-6, and 5'-7 (which delete amino acids 2–65, 2–170, and 2–278, respectively) are still able to induce sd expression, but this ability is very weak (Fig 2D) compared to the level of sd-lacZ staining induced by expression of full-length VG (Fig 2A). C-terminal deletion constructs able to induce ectopic expression activity included 1-2 and 1-4 (amino acids 335–426 and 356–453 deleted, respectively). Wing imaginal discs from flies expressing these constructs (Fig 2E) show a somewhat stronger X-gal stain than that of the N-terminal deletion constructs (Fig 2D), but much less than that produced by full-length UAS-vg. The one internal deletion construct that could induce ectopic expression of sd-lacZ was 7-8 (amino acids 187–278 deleted). Interestingly, in this case, the level of sd-lacZ activation (Fig 2F) appears comparable to that induced by the full-length vg construct. None of the remaining deletion constructs caused activation of sd-lacZ (data not shown) and thus were similar to the negative control staining produced by the pUAST vector alone. The discs presented in Fig 2A–F, were selected for staining in the wing blade portion only and the variability in the endogenous pattern in the notum is not reproducible. The ability of the various deletion constructs to ectopically induce sd is summarized in Fig 1. Multiple transgenic lines (from 2-6) were assayed for each deletion construct and no significant differences were observed between different transgenic lines for the same construct. As expected, all of the UAS-vg constructs that were deleted for the SD-binding domain had no activity in the ectopic sd expression assay (see Fig 1) and these constructs were also negative with respect to showing ectopic production of VG as detected with anti-VG antibody. Thus, the ability to bind SD is a dominant prerequisite in assessing any other aspect of VG function in vivo.

View larger version (70K): In this window In a new window Download PPT slide   Figure 2. Ectopic expression assay results. (A) sd-lacZ expression in a UASvg; ptcGal4;sd-lacZ background. Superimposed on the normal sd-lacZ pattern is the vertical stripe (arrow) indicating ectopic sd-lacZ expression. The UASvg is a full-length vg, which serves as a positive control for the activity of the various vg deletion constructs. (B) No ectopic sd-lacZ expression is present in a pUAST; ptcGal4; sd-lacZ background. In this case only the normal endogenous sd pattern is seen, and this serves as a negative control for the activity of the vg deletion constructs. (C) VG is expressed in the UASvg constructs and can be detected with an anti-VG antibody. The white arrow identifies the ectopic VG vertical stripe in a UASvg1-4; ptcGal4 background while the yellow arrow identifies endogenous VG expression. For the ectopic sd assays (D–F), only representative results from those constructs able to promote ectopic sd expression (black arrows) are presented. (D) 5'-5 (amino acids 2–65 deleted). (E) 1-2 (amino acids 335–426 deleted). (F) 7-8 (amino acids 187–278 deleted).

The ectopic expression of vg provides another assay of VG function in vivo. It has been established that expressing VG in tissues where it is not normally present, but where SD is present, results in the production of ectopic wing tissue (KIM et al. 1996 ). Thus, loss of the ability to produce this ectopic tissue indicates the loss of VG function. An eyelessGal4 driver was used to express various UASvg constructs. A representative from each vg deletion group that was able to induce at least some ectopic sd expression was compared to a full-length UASvg construct with respect to being able to produce ectopic wing tissue in the eyes. As expected, full-length vg did produce ectopic tissue and UASvg7-8 also exhibited this ability at a similar level (results not shown). However, the loss of either the 5' putative activation motif (using UASvg5-5) or the 3' activation motif (using UASvg1-4) abolished the ability to produce the ectopic outgrowths in the eye, thus indicating the importance of having both of these regions for normal function. Normal VG function may also be visualized in mitotic clones produced using Flp-out technology (ITO et al. 1997 ). Mitotic clones expressing wild-type VG or a mutant truncated version of VG were marked by coexpression of green fluorescent protein (GFP) in the same clones. Clones expressing only a marker GFP served as a negative control (not shown) and exactly resembled the clones in Fig 3A and Fig B, which express GFP and a nonfunctional VG. Further, ectopic expression of wild-type vg is also known to induce ectopic wg (LIU et al. 2000 ). Thus, when the various vg constructs are used to induce ectopic wg expression, loss of either the 3' (using UASvg1-4) or the 5' putative activation domain (using UASvg5-6) results in the inability to induce ectopic wg (Fig 3, A–F). However, the UASvg7-8 construct behaves like full-length vg in retaining this ability (Fig 3, G–L). Deletion of either of the putative activation domains results in clones that fail to grow (smaller compared to wild-type vg), have an irregular shape (demonstrating the existence of adhesive differences within the clones), and are unable to induce wg bordering the clones that overlap the hinge region (Fig 3, A–F). The opposite behavior is exhibited by clones from a UASvg7-8 or a UASvg full-length genetic constitution (Fig 3, G–L). Not only are the clones larger with smoother borders, but also functional VG-expressing clones can induce wg expression non-cell autonomously in some clones that overlap the hinge region of the disc (see arrowheads in Fig 3H, Fig I, Fig K, and Fig L). This assay again emphasizes the functional importance of the two putative activation domains in vivo in the fulfillment of normal VG functions.

View larger version (69K): In this window In a new window Download PPT slide   Figure 3. The putative activation domains are required for adhesive properties of the cells in which VG is expressed and for the induction of wg in the hinge as judged by Flp-out clonal analysis. Red channel: Wing discs from late third instar larvae allowed to grow for between 48 and 72 hr after clone induction and stained for WG. (B, E, H, and K) The VG-expressing clones are marked by the coexpression of GFP from a UAS-GFP construct and (C, F, J, and L) the overlay of VG and WG expression is shown. (A–C) Flp-out clones from the construct deleting the 3' putative activation domain of vg (UASvg1-4). These clones resemble control GFP clones with characteristic irregular shape, small size, and inability to induce WG in the hinge region. (D–F) Flp-out clones from the construct deleting the 5' functional domain of vg (UAS vg5-6). These clones also resemble control GFP clones as in A–C. (G–I) Clones expressing a UASvg7-8 construct retain both putative activation domains. Note that the clones are larger, circular in shape with smooth edges, and now are able to induce WG non-cell autonomously in cells surrounding the clones overlapping the hinge region of the disc. See arrowheads indicating circular patterns of WG induction in the red channels that surround VG-expressing clones in the green channels for clones overlapping the hinge region. (J–L) Flp-out clones from a full-length vg construct resemble those in G–I.

While the ectopic sd expression assay or ectopic vg assay provides a quick method to map potential activation domains within VG, the ability of these deleted proteins to substitute for normal VG function needed to be examined as well. Thus, each of the deleted transgenes was also tested for the ability to rescue the phenotypes of vg^83b27 and vg¹ homozygous flies. The vg^83b27 mutant stems from a deletion within intron 2 that includes the vg boundary enhancer (vgBE), and flies homozygous for this allele have virtually no wings. The vg¹ mutant phenotype is less severe and is caused by the insertion of a 412 transposon into intron 3. However, since these two vg alleles retain some vg function, the vg^83b27R null allele was also included in the rescue assay. The ability of each deletion construct to rescue each of these three vg mutants was assessed by comparing the progeny from a cross of UASvg; vg x vgGal4;vg to the phenotypes of flies homozygous for the respective vg tester allele. In the case of vg^83b27R, the homozygous vg progeny were selected from a cross of vg^83b27R vgGAL4/Cy x vg^83b27R UASvg/Cy. The rescue of vg^83b27 by the full-length UAS-vg construct is shown in Fig 4A and that of vg¹ in Fig 5A, and these results are also summarized in Fig 1. The same negative control (pUAST construct alone) was used in these rescue experiments as in the ectopic sd assay. No wing tissue restoration was observed in the control flies (results not shown). In each case, a correlation was observed between the ability to induce ectopic sd expression and the ability to rescue each vg mutant phenotype. For vg^83b27, deletion constructs 5'-5, 5'-6, and 5'-7 were able to produce only a partial rescue, as shown in Fig 4B for 5'-5. The rescue ability of deletion constructs 1-2 and 1-4 was also only partial (Fig 4C for 1-4) while that of 7-8 was virtually complete (Fig 4D). The vg¹ mutant phenotype was not rescued as completely as vg^83b27, but the level is still substantial. Moreover, in each case, the relative ability of the deletion constructs to rescue vg¹ correlates with that observed for vg^83b27 (Fig 5, B–D vs. Fig 4, B–D), so all rescue results are summarized in Fig 1. Again, there were no notable differences in the degree of rescue when comparing multiple transgenic lines for a particular deletion construct. Selected rescue data for the vg^83b27R mutant are portrayed in Fig 6 and these data agree with those for the two leaky vg alleles. Only the constructs representing regions that exhibited partial rescue activity for the other two alleles are portrayed for vg^83b27R, since all other constructs were completely negative. As expected, even the constructs that gave partial rescue of vg^BG and vg^83b27 did not rescue vg^83b27R as well (Fig 6, A–C vs. D–I). The UASvg7-8 was unable to produce a full rescue of vg^83b27R (Fig 6J and Fig K) but a full-length vestigial construct is also unable to rescue this allele fully (results not shown). Since vg7-8 and wild type still behave similarly, this partial rescue is likely a reflection of the relative strength of the vgGAL4 driver. Note that there is considerable rescue of the anterior bristles using vg7-8, which is not as apparent with vg1-4 and not observed with vg5-5 (Fig 6E vs. H and K).

View larger version (97K): In this window In a new window Download PPT slide   Figure 4. Wing rescue assay for vg83b27. The rescue levels exhibited by representative UASvg constructs are portrayed. (A) Full-length vg (no amino acids deleted). (B) 5'-5 (amino acids 2–65 deleted; domain 1). (C) 1-4 (amino acids 356–453 deleted; domain 2). (D) 7-8 (amino acids 187–278 deleted). The rescue assay was done by crossing UASvg; vg83b27 flies to vgGal4;vg83b27 flies and comparing the progeny to homozygous vg83b27 flies.

View larger version (67K): In this window In a new window Download PPT slide   Figure 5. Wing rescue assay for vg1. The rescue levels exhibited by representative UASvg constructs are portrayed. (A) Full-length vg (no amino acids deleted). (B) 5'-5 (amino acids 2–65 deleted; domain 1). (C) 1-2 (amino acids 335–426 deleted; domain 2). (D) 7-8 (amino acids 187–278 deleted). The rescue assay was done as in Fig 4 except that the vg1 allele was used.

View larger version (93K): In this window In a new window Download PPT slide   Figure 6. Wing rescue assay for vg83b27R. (A–C) Wild-type wings. (D–F) Wings from flies with the genotype UASvg7-8/+; vg83b27R/vg83b27R vgGAL4. (G–I) Wings from flies with the genotype UASvg1-4 vg83b27R/vgGAL4 vg83b27R. (J–L) Wings from flies with genotype UASvg5-5 vg83b27R/vgGAL4 vg83b27R. (B, E, H, and K) Higher magnification of the anterior wing margins from the panels showing the full wings. (C, F, I, and L) Higher magnification of the posterior wing margins of these wings. A, D, G, and J are all at the same magnification. Anterior is up.

In these assays, the SD-binding domain (amino acids 281–335) as well as N-terminal amino acids 1–65 (called domain 1) and C-terminal amino acids 335–453 (called domain 2) are required for optimal VG activity. The general location of the these domains identified by in vivo experiments correlates well with the domains located by in vitro experiments (VAUDIN et al. 1999 ) and with those implicated as important in the DNA-binding ability of the VG/SD complex (HALDER and CARROLL 2001 ). Although the deletion construct that left both domains 1 and 2 intact (7-8) exhibited virtually the same levels of ectopic sd expression and rescue ability of the leaky vg alleles as those of full-length vg, neither it nor wild-type could fully rescue vg^83b27R. Further, since the ectopic behavior of vg7-8 is also concordant with other activities associated with ectopic expression of wild-type vg, such as wing outgrowths in eye tissue or induction of wg, the failure to rescue vg^83b27R is not likely due to the nature of the vg7-8 deletion. When full-length VG was compared to the protein produced by vg7-8 with respect to folding ability using a computer-based analysis (ROST and SANDER 1993 , ROST and SANDER 1994 ; ROST 1996 ), the result was that this 91-amino-acid portion was not predicted to possess any significant secondary structure. Thus its deletion would not impair proper folding of the remainder of the protein. Computer analysis of structural motifs (BAIROCH et al. 1997 ) predicts a putative nuclear localization sequence within the region deleted in vg7-8. However, since the protein encoded by this deletion construct behaves with virtual wild-type activity, one can infer that this motif is either nonfunctional or not required for full-length VG function. Furthermore, it is likely that VG enters the nucleus via its ability to bind to SD and utilizes the nuclear localization signal therein (SRIVASTAVA et al. 2002 ).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The results presented herein are consistent with an in vivo role of VG acting as a transcriptional activator in Drosophila cells in conjunction with binding to SD. It was already known that ectopic expression of vg induces the ectopic expression of sd in the same pattern. Further, it was also known that an expressed vg construct lacking the ability to bind SD was also unable to induce ectopic sd expression (SIMMONDS et al. 1998 ). However, it was not known which portions of the vg ORF were necessary to carry out these functions of VG in vivo. The data presented herein are necessary and novel in that they identify functional roles for regions of the VG protein in the cellular context within which wing development actually occurs. Previous biochemical studies, although useful, gained their conclusions from artificial situations. We demonstrate for the first time that three regions of VG are necessary for full physiological function. VAUDIN et al. 1999 identified the same N-terminal and C-terminal domains as herein by using a yeast one-hybrid assay capable of providing an activation function. However, these results were merely suggestive until confirmed by in vivo results. HALDER and CARROLL 2001 showed that these domains are necessary for binding a VG/SD complex to a consensus SD-binding site in vitro, but a direct extrapolation to an in vivo situation is difficult. Both sets of in vitro data leave open the possibility of other important functional motifs in vivo. A more definitive way to assay for functional motifs for the VG protein is by an in vivo analysis such as presented herein.

From our results using ectopic sd-lacZ induction, the ability to rescue vg mutations, and the ability to carry out other functions associated with normal vg, it can be discerned that certain portions of the vg ORF, in addition to the SD-binding domain, are necessary to accomplish normal VG function. These appear to be the critical regions, as other portions can be deleted without effect. More specifically, the N-terminal amino acids (approximately the first 65) and C-terminal residues from 335 to 453 seem to play an important role in the induction of sd-lacZ. When the N-terminal deletion 5'-5 (deleting amino acids 2–65) is assayed, the ectopic expression ability is reduced markedly compared to that seen with the full-length vg construct, although it is not eliminated completely. Moreover, the larger N-terminal deletions (amino acids 2–170 and 2–278, respectively) do not further lower the ability to express sd. Thus, it seems that the fundamentally important region is already removed with the 5'-5 construct. For C-terminal deletions 1-4 and 1-2 (amino acids 356–453 and 335–426, respectively), the ability to ectopically express sd is much less than that produced by full-length vg but somewhat stronger than that produced when the N-terminal deletion constructs are assayed. Deletions 5'-5, 5'-6, and 5'-7 retain the encoded amino acids missing from 1-4 and 1-2 and vice versa. Taken together, these data suggest the presence of two important functional domains for VG: one within amino acids 1–65 (domain 1) and the other within amino acids 336–453 (domain 2). See Fig 1 for the position of these domains within the VG protein. Although the precise boundaries of these domains have not yet been determined, domain 1 is very likely within the first 65 amino acids (deleted in vg5'-5) since this is the region most highly conserved between D. melanogaster and the mosquito Aedes egyptii (HALDER and CARROLL 2001 ). There is 82% identity over the first 66 amino acids, but over the next 20 amino acids the identity drops to 35% and even further beyond that. In agreement with this notion, the extent of "functional" loss in UASvg 5'-6 and 5'-7 is no stronger than that exhibited by UASvg 5'-5, which deletes the first 65 amino acids only. The activity of domain 2 appears to be weaker, since domain 1 deletions produce a slightly more drastic impairment of VG function than do domain 2 deletions (amino acids 356–453 or 335–426). However, homology between Drosophila and mosquito VG is also high within the SD-binding domain of VG and, in fact, remains strong to the carboxyl terminus of VG (82% identity from residue 335 to 453; HALDER and CARROLL 2001 ). The data herein define the presence of two necessary functional domains for the VG protein in vivo. These domains correlate well with data that predict two activation regions using in vitro experiments, including yeast one-hybrid assays (VAUDIN et al. 1999 ). The regions identified herein also complement more recent in vitro data, implicating these regions of VG as necessary for binding of the VG/SD complex to target genes (HALDER and CARROLL 2001 ). However, in VAUDIN et al. 1999 the N-terminal region was not as active as the C-terminal region, while in the present in vivo study the opposite may be the case.

The 7-8 construct (deleting amino acids 187–278) behaves similarly to full-length vg in all assays reported herein even though the encoded protein lacks 91 amino acids present in normal VG. Thus, a significant portion of the VG protein can be absent without any obvious impairment of functions associated with this protein. This portion of the protein has no homology to other known proteins, including vertebrate VG homologs, and is not predicted to contain any well-defined structural motifs. Thus, the function of this segment, comprising 20% of normal Vg, remains unknown at this time.

While the ectopic sd-lacZ assay illustrates the ability of each region to activate transcription of sd, the ability to rescue the vg mutants (vg^83b27, vg^83b27R, and vg¹ ) was also assessed to confirm that these regions are required for the majority of VG-mediated activation of downstream genes. All deletion constructs that were able to rescue vg^83b27 mutants were also able to rescue vg¹ mutants, and these same constructs were the ones that retained the ability to ectopically induce sd. As expected, the constructs that could accomplish some rescue of the leaky vg alleles were able to rescue vg^83b27R to a lesser extent, but this merely strengthens the conclusion that the respective deleted regions are essential for in vivo VG function. Furthermore, the relative ability of the constructs to induce the ectopic expression of sd correlates with their relative ability to rescue the wing phenotypes of the tester vg alleles. The full-length vg construct gives almost complete rescue of vg^83b27 but rescues vg¹ to a lesser extent and vg^83b27R to an even lesser extent. The most likely explanation for incomplete rescue of vg¹ and vg^83b27R is the vgGAL4 line that was used in these experiments. It was constructed (S. MORIMURA and M. HOFFMANN, unpublished results) by excising the vg gene intron 2 boundary enhancer region (WILLIAMS et al. 1991 , WILLIAMS et al. 1994 ) from psub26 and cloning this into pGawB (BRAND and PERRIMON 1993 ). Because only the vg intron 2 region was used, it is not surprising that a complete vg expression pattern is not produced by this driver.

The SD-binding domain (residues 278–335) has already been shown to be necessary and sufficient for binding of SD and TEF-1 (SIMMONDS et al. 1998 ). Further, the deletion of residues 187–278 (vg7-8) does not result in any obvious impairment of VG function. Although the removal of either putative activation domain does not completely destroy the ability of VG to execute its functions, these functions are impaired considerably. This could be interpreted as a partial functional redundancy or, more likely, it could be interpreted that both domains are required for full normal function. Future efforts will be directed toward identifying a possible functional role for the portion of the VG protein that appears to be dispensable from the results of the present study. Studies aimed at identifying additional protein-protein interactions with VG may be able to implicate this region in other aspects of VG function.

	ACKNOWLEDGMENTS

We thank Shelagh Campbell for suggestions and a critical reading of the manuscript and Sean Carroll for reagents. We also thank Sandra O'Keefe for help in the preparation of the figures. This research was supported by a NSERC Canada grant to J.B.

Manuscript received October 16, 2002; Accepted for publication January 3, 2003.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BAIROCH, A., P. BUCHER, and K. HOFFMAN, 1997  The PROSITE database: its status. Nucleic Acids Res. 25:217-221.[Abstract/Free Full Text]

BELLEN, H. J., C. J. O'KANE, C. WILSON, U. GROSSNIKAUS, and R. K. PEARSON et al., 1989  P element-mediated enhancer detection: a versatile method to study development in Drosophila. Genes Dev. 3:1288-1300.[Abstract/Free Full Text]

BRAND, A. H. and N. PERRIMON, 1993  Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 118:401-415.[Abstract]

CAMPBELL, S. D., A. DUTTAROY, A. L. KATZEN, and A. CHOVNICK, 1991  Cloning and characterization of the scalloped region of Drosophila melanogaster.. Genetics 127:367-380.[Abstract]

CAMPBELL, S. D., M. INAMDAR, V. RODRIGUES, V. RAGHAVAN, and M. PALAZZOLO et al., 1992  The scalloped gene encodes a novel, evolutionarily conserved transcription factor required for sensory organ differentiation in Drosophila.. Genes Dev. 6:367-379.[Abstract/Free Full Text]

GUSS, K., C. E. NELSON, A. HUDSON, M. KRAUS, and S. B. CARROLL, 2001  Control of a genetic regulatory network by a selector gene. Science 292:1164-1167.[Abstract/Free Full Text]

HALDER, G. and S. B. CARROLL, 2001  Binding of the Vestigial co-factor switches the DNA-target selectivity of the Scalloped selector protein. Development 128:3295-3305.[Abstract/Free Full Text]

HALDER, G., P. POLACZYK, M. E. KRAUS, A. HUDSON, and J. KIM et al., 1998  The Vestigial and Scalloped proteins act together to directly regulate wing-specific gene expression in Drosophila.. Genes Dev. 12:3900-3909.[Abstract/Free Full Text]

ITO, K., W. AWANO, K. SUZUKI, Y. HIROMI, and D. YAMAMOTO, 1997  The Drosophila mushroom body is a quadruple structure of clonal units each of which contains a virtually identical set of neurons and glial cells. Development 124:761-771.[Abstract]

KIM, J., A. SEBRING, J. J. ESCH, M. E. KRAUS, and K. VORWERK et al., 1996  Integration of positional signals and regulation of wing formation and identity by Drosophila vestigial gene. Nature 382:133-138.[Medline]

LIU, X., M. GRAMMONT, and K. D. IRVINE, 2000  Roles for scalloped and vestigial in regulating cell affinity and interactions between the wing blade and the wing hinge. Dev. Biol. 228:287-303.[Medline]

PAPAYANNOPOULOS, V., A. TOMLINSON, V. M. PANIN, C. RAUSKOLB, and K. IRVINE, 1998  Dorsal-ventral signaling in the Drosophila eye. Science 281:2031-2034.[Abstract/Free Full Text]

PAUMARD-RIGAL, S., A. ZIDER, P. VAUDIN, and J. SILBER, 1998  Specific interactions between vestigial and scalloped are required to promote wing tissue proliferation in Drosophila melanogaster.. Dev. Genes Evol. 208:440-446.[Medline]

ROST, B., 1996  PHD: predicting one-dimensional protein structure by profile based neural networks. Methods Enzymol. 266:525-539.[Medline]

ROST, B. and C. SANDER, 1993  Prediction of protein secondary structure at better than 70% accuracy. J. Mol. Biol. 232:584-599.[Medline]

ROST, B. and C. SANDER, 1994  Combining evolutionary information and neural networks to predict protein secondary structure. Proteins 19:55-77.[Medline]

RUBIN, G. M. and A. C. SPRADLING, 1983  Vectors for P-element mediated gene transfer in Drosophila. Nucleic Acids Res. 11:6341-6351.[Abstract/Free Full Text]

SIMMONDS, A. J., S. HUGHES, S. COCQUYT, J. TSE, and J. B. BELL, 1997  The effect of dominant vestigial alleles upon vestigial-mediated wing patterning during development of Drosophila melanogaster.. Mech. Dev. 67:17-33.[Medline]

SIMMONDS, A. J., X. LIU, K. H. SOANES, H. M. KRAUSE, and K. D. IRVINE et al., 1998  Molecular interactions between Vestigial and Scalloped promote wing formation in Drosophila.. Genes Dev. 12:3815-3820.[Abstract/Free Full Text]

SRIVASTAVA, A., J. MACKAY, and J. B. BELL, 2002  A Vestigial:Scalloped TEA domain chimera rescues the wing phenotype of a scalloped mutation in Drosophila melanogaster.. Genesis 33:40-47.[Medline]

STRUHL, G., K. FITZGERALD, and I. GREENWALD, 1993  Intrinsic activity of the Lin-12 and Notch intracellular domains in vivo. Cell 74:331-345.[Medline]

VAUDIN, P., R. DELANOUE, I. DAVIDSON, J. SILBER, and A. ZIDER, 1999  TONDU (TDU), a novel human protein related to the product of vestigial (vg) gene of Drosophila melanogaster interacts with vertebrate TEF factors and substitutes for Vg function in wing formation. Development 126:4807-4816.[Abstract]

VIDAL, M. and P. LEGRAIN, 1999  Yeast forward and reverse ‘n’-hybrid systems. Nucleic Acids Res. 27:919-929.[Abstract/Free Full Text]

VIEIRA, J. and J. MESSING, 1982  The pUC plasmids, an M13mp7-derived system for insertion mutagenesis and sequencing with synthetic universal primers. Gene 19:259-268.[Medline]

WILLIAMS, J. A. and J. B. BELL, 1988  Molecular organization of the vestigial region in Drosophila melanogaster.. EMBO J. 7:1355-1363.[Medline]

WILLIAMS, J. A., I. M. SCOTT, A. L. ATKIN, W. J. BROOK, and M. A. RUSSELL et al., 1990  Genetic and molecular analysis of vgU and vgW: two dominant vg alleles associated with gene fusions in Drosophila melanogaster.. Genetics 125:833-844.[Abstract]

WILLIAMS, J. A., J. B. BELL, and S. B. CARROLL, 1991  Control of Drosophila wing and haltere development by the nuclear vestigial gene product. Genes Dev. 5:2481-2495.[Abstract/Free Full Text]

WILLIAMS, J. A., S. W. PADDOCK, and S. B. CARROLL, 1993  Pattern formation in a secondary field: a hierarchy of regulatory genes subdivides the developing Drosophila wing disc into discrete subregions. Development 117:571-584.[Abstract]

WILLIAMS, J. A., S. W. PADDOCK, K. VORWERK, and S. B. CARROLL, 1994  Organization of wing formation and induction of a wing-patterning gene at the dorsal/ventral compartment boundary. Nature 368:299-305.[Medline]

This article has been cited by other articles:

				A. Garg, A. Srivastava, M. M. Davis, S. L. O'Keefe, L. Chow, and J. B. Bell Antagonizing Scalloped With a Novel Vestigial Construct Reveals an Important Role for Scalloped in Drosophila melanogaster Leg, Eye and Optic Lobe Development Genetics, February 1, 2007; 175(2): 659 - 669. [Abstract] [Full Text] [PDF]

				A. Srivastava, A. J. Simmonds, A. Garg, L. Fossheim, S. D. Campbell, and J. B. Bell Molecular and Functional Analysis of scalloped Recessive Lethal Alleles in Drosophila melanogaster Genetics, April 1, 2004; 166(4): 1833 - 1843. [Abstract] [Full Text] [PDF]