M. Bonds and J. Sands contributed equally to this work.
Patterns & Phenotypes
Genetic screen for regulators of ind expression identifies shrew as encoding a novel twisted gastrulation-like protein involved in Dpp signaling
Article first published online: 29 OCT 2007
Copyright © 2007 Wiley-Liss, Inc.
Special Issue: Special Focus on Stem Cells
Volume 236, Issue 12, pages 3524–3531, December 2007
How to Cite
Bonds, M., Sands, J., Poulson, W., Harvey, C. and Von Ohlen, T. (2007), Genetic screen for regulators of ind expression identifies shrew as encoding a novel twisted gastrulation-like protein involved in Dpp signaling. Dev. Dyn., 236: 3524–3531. doi: 10.1002/dvdy.21360
- Issue published online: 14 NOV 2007
- Article first published online: 29 OCT 2007
- Manuscript Accepted: 15 SEP 2007
- National Center for Research Resources (NCRR). Grant Number: P20 RR016475
- Deficiency screen;
- dpp signaling;
Initiation and refinement of expression of the Ind homeodomain protein in the Drosophila embryo is coordinately regulated by global dorsoventral patterning pathways Dorsal, Egfr, and Dpp, and well as by Vnd, which positions the ventral boundary of Ind. Therefore, we set out to look for novel regulators of dorsoventral patterning by screening the Exelixis deficiency collection for modified expression of Ind. Indeed, we found deficiencies that remove components of the known signaling pathways had altered or lost ind expression. These findings included deficiencies that remove screw, dpp, and egfr as well as deficiencies that remove ind itself. In addition, we found several deficiencies that had altered or loss of ind expression. We also observed phenotypes suggestive of dorsoventral patterning defects such as twisting during gastrulation, and defects associated with loss of dorsal specification. These include a pair of overlapping deficiencies that produced ventralized embryos. We find that transheterozygotes of these two deficiencies are also ventralized. There are seven genes common to both deficiencies, including CG11582, which encodes a twisted gastrulation-like protein. These two deficiencies are also allelic with shrew mutations. Here, we present data supporting the conclusion that CG11582 is the gene affected in shrew mutants. Developmental Dynamics 236:3524–3531, 2007. © 2007 Wiley-Liss, Inc.
Dorsoventral (DV) patterning of the Drosophila embryo and central nervous system (CNS) is a two-step process. The first step involves the subdivision of the embryo into specific tissue types along the dorsoventral axis. Three signaling pathways—Dorsal, Decapentaplegic (Dpp), and epidermal growth factor receptor (Egfr)—work in concert to subdivide the embryo into specific tissue types: mesoderm, neuroectoderm, dorsal epidermis, and peripheral nervous system, and amnioserosa. The second step involves the further subdivision of each tissue into more precise DV domains. For example, the neuroectoderm expresses three homeobox genes in adjacent DV columns: the ventral column expresses ventral nervous system defective (vnd), the intermediate column expresses intermediate neuroblasts defective (ind), and the dorsal column expresses muscle segment homeobox (msh; Mellerick and Nirenberg,1995; D'Alessio and Frasch,1996; Isshiki et al.,1997; Mc Donald et al.,1998; Weiss et al.,1998). The expression of these genes is subsequently inherited by the neuroblasts that form in each column. These genes control many aspects of neural stem cell biology, including their formation and types of neurons and glia they produce.
Proper expression and refinement of ind requires input from all three pathways known to control global DV patterning of the embryo (Von Ohlen and Doe,2000; Stathopolous and Levine,2005). Specifically, initiation of ind expression requires positive inputs from the both the Dorsal and Egfr pathways. The Dpp pathway, on the other hand, has been shown to repress ind expression and may play a minor role in positioning the dorsal boundary of the ind stripe (Von Ohlen and Doe,2000; Mizutani et al.,2005). Additional factors might play a role in positioning the dorsal border of ind (Stathopolous and Levine,2005). To determine whether other genes are involved in DV patterning as well as regulation of Ind expression, we chose to screen the Exelexis deficiency collection for novel regulators of ind expression and for global regulators of DV patterning. These lines were generated by a novel strategy to provide high-resolution deletion coverage of the Drosophila genome. Each of these lines was generated by a process that results in the excision of small segments of the chromosomes (Parks et al.,2004). This effectively produces loss of function mutations for all the genes in the deleted region. This provided a method to rapidly screen for mutations affecting expression of ind as well as DV patterning without the need to generate mutations in each gene independently. In addition, these lines have precisely defined breakpoints. Thus, by using this particular collection of genomic deficiencies, we were able to quickly narrow down the gene product(s) responsible for our observed phenotypes. Because of the design of our screen, we predicted we would be most likely to find components of the zygotically active Egfr and Dpp pathways as opposed to Dorsal pathway, which are predominantly supplied maternally. Indeed, we found that deficiencies lacking genes encoding known components of the Egfr and Dpp pathways gave altered Ind expression or phenotypes associated with dorsoventral patterning defects. These findings include deficiencies that remove the transforming growth factor-beta (TGFβ) ligands Screw and Dpp, Df(2L)Exel6044 and Df(2L)Exel7011, respectively. Both deficiencies produce gastrulation defects associated with ventralized embryos. Furthermore, we find that a deficiency that removes Egfr, Df(2R)Exel6076, shows a loss of ind expression.
In general, we focused on four classes of phenotypes: (1) loss of ind expression, (2) twisted gastrulation, (3) ectopic ind expression, and (4) ventralized embryos with gastrulation defects. We found that the ventralized embryos are primarily due to loss of components of the Dpp pathway. In contrast, the loss of Ind expression was either due to loss of the Egfr pathway required for Ind expression or loss of Ind itself. Here, we describe the characterization of two deficiencies that remove conceptual gene CG11582, a novel Tsg domain containing protein. These two deficiencies both produced gastrulation defects associated with ventralized embryos, similar to those seen in screw or dpp mutants. We also show that transheterozygotes with alleles of shrew, which has not been molecularly characterized, also produce the ventralized phenotype. Finally, we sequenced the CG11582 locus in two shrew alleles and found point mutations in both. Therefore, we conclude that CG11582 is the gene affected in shrew mutants.
Deficiencies Lacking Ind and Egfr Pathway Components Gene Fail to Express ind
In an effort to identify novel regulators of ind expression, as well as general components of the Dorsoventral signaling pathways, we conducted a screen of the Exelexis deficiency collection. We screened for ind message by in situ hybridization. We found four classes of phenotypes: loss of ind expression, ectopic ind expression, twisting during gastrulation, and ventralized embryos. Three of the deficiencies showed a loss of ind expression (Fig. 1). At least two of these deficiencies, Df(3L)Exel6125 and Df(3L)Exel6126, are loss of function for ind itself and are, therefore, not expected to express ind. The third deficiency line that fails to express ind is Df(2R)Exel6076, which removes egfr. Thus, the deficiencies that were found to have loss of Ind expression were those predicted by previous studies.
Several Deficiencies Produced Embryos With Twisted Gastrulation Defects
We predicted that twisting during gastrulation could result from defects in the DV patterning machinery as improper specification of dorsal tissues may cause misguided germ band extension. For example, mutations in twisted gastrulation (tsg) produce twisted embryos. Therefore, we took care to identify deficiencies with a significant amount of twisting during the gastrulation process. Examples of embryos exhibiting the twisted gastrulation phenotype are shown in Figure 2. Although it is possible that twisting can be due to DV patterning defects, we expected that other defects might also be able to produce a twisting phenotype. In general, we found that twisting did not lead to a change in the pattern of ind expression. That is, the stripes of ind expression were spaced normally and of normal width. Therefore, although twisting may be due to defects in DV patterning, there is no direct effect on regulation of ind expression. It is unclear what genes might be responsible for the twisting during germ band extension. One of these deficiencies, Df(3R)Exel6204, takes out the Enhancer of split (E(spl)) complex, which includes the groucho (gro) gene. However, when we examined E(spl) mutant embryos, we did not observe twisting, suggesting that the phenotype was not due to loss of E(spl) genes (data not shown). An additional deficiency, Df(3R)Exel9012, eliminates the pointed (pnt) gene, which is a known component of the Egfr pathway. However, when we examined pnt mutants, we failed to see a similar twisted phenotype (data not shown). Thus, pnt is most likely not the gene responsible for the twisting phenotype.
Two Deficiencies Show Expanded or Aberrant ind Expression
We predicted that we would find a class of phenotypes that included expanded ind expression. Expansion could occur either ventrally to the midline or dorsally. We find that two deficiencies met these criteria (Fig. 3). Specifically, one deficiency, Df(3R)Exel6150, showed slight dorsal expansion of the ind stripe. This finding was particularly evident in the anterior regions of the embryo, as opposed to the more posterior regions where the stripes appeared more normal (Fig. 3). We find that this deficiency eliminated the gap gene hunchback (hb). When we examined the phenotype of hb mutant embryos, we found similar defects in ind expression (Fig. 3). Therefore, we conclude that the phenotype observed in Df(3R)Exel6150 is most likely due to loss of hb. Df(2L)Exel7071 also results in expanded ind expression. However, in this case rather than dorsal expansion, we observe ventral expansion of ind. These embryos also have highly irregular ind staining, and occasionally, we observe profound midline defects (Fig. 3B).
Absence of the Dpp Signaling Pathway Components Produce Ventralized Embryos
A subset of the deficiency lines tested exhibited gastrulation defects associated with loss of dorsal structures, specifically due to loss of reduction of Dpp/Tgfβ like signaling required to specify the dorsal surface of the embryo. These included Df(2L)Exel7011, which removes dpp itself, and Df(2L)Exel6044 which removes an additional TGFβ-like ligand screw (Fig. 4). In addition, we observed three other deficiencies that resulted in ventralized embryos. These include Df(3R)Exel6151. We do not observe a full 25% of the embryos in Df(3R)Exel6151 exhibiting the ventralized phenotype. So we focused in the overlapping deficiencies, Df(3L)Exel9000 and Df(3L)Exel8098, that produced the ventralized phenotype. We describe these embryos as ventralized based on the presence of gastrulation defects indicative of a failure to carry out complete germband extension, and a partial widening of the distance between the stripes of ind expression. This finding is apparent as a wave or bump in the normally straight stripes (Fig. 4). Transheterozygotes of the two deficiencies also produce the ventralized phenotype (Fig. 4F). This finding suggests that the gene responsible for the observed phenotype will be common to both deficiencies.
Jurgens et al. (1984) identified shrew as one of the original dorsal group genes. That is, mutations that produced ventralized embryos. shrew mutants were originally mapped to 65B12-C12 on chromosome 3 within the Drosophila genome, which placed it in close proximity to the region deleted in each of these deficiencies. Thus, we wanted to test if these deficiencies were allelic with shrew. To do this, we made transheterozygotes with Df(3L)Exel9000 and shrew1 alleles. We found that approximately one fourth of the embryos from a cross of Df(3L)Exel9000 females to shrew1 males were also strongly ventralized and had a phenotype highly comparable to our deficiencies as well as homozygous shrew embryos (Fig. 5B,C). Therefore, we concluded that the gene responsible for the phenotype we observe in our deficiency lines is the same gene that is disrupted in shrew mutations.
shrew Is Encoded by CG11582 and Is Predicted to be a Tsg-like Protein
We examined the set of genes predicted to be absent in each of our lines and found that there were only seven genes common to both deficiencies (Fig. 2a). The previous reports indicating that shrew acted upstream of dpp to potentiate Dpp activity allowed us to quickly narrow down a candidate gene responsible for the phenotype (Ferguson and Anderson,1992). Among the seven genes deleted in both of the deficiency lines was conceptual gene CG11582. CG11582 is predicted to encode a 116 amino acid protein with a single cysteine-rich (CR) domain, which has significant homology to Twisted gastrulation (Vilmos et al.,2001; Fig. 6b). However, the predicted protein, as described in Flybase, is 30 amino acids shorter than the actual protein. The additional amino acids are located amino terminal to the predicted start methionine (Fig. 6B; M.B. O'Connor, personal communication). An additional tsg-like protein in Drosophila is crossveinless (cv), which is known to be involved in regulating Dpp activity in the pupal wing (Shimmi et al.,2005; Vilmos et al.,2005). Our data showing that two deficiencies that remove the CG11582 locus both have gastrulation defects are in contrast to a previously published report (Pilot et al.,2006) claim that CG11582 RNAi resulted in gastrulation defects but that a deficiency that removes CG11582 did not have gastrulation defects. Thus, they dismissed their RNAi results as cross-reaction with tsg. Our data suggest their RNAi results were probably accurate and they had an improperly mapped deficiency.
To determine whether CG11582 is the gene disrupted in shrew mutants, we sequenced the CG11582 locus in two different shrew mutations shrew1 and shrewB18. For shrew1, we found a single base T>A change that results in a Cys123 to Ser alteration at the amino acid level (Fig. 6B). For shrewB18, we found a C>T change that results in a Gln23 to stop change in the protein sequence. Thus, we conclude that CG11582 is the gene altered in shrew mutations. Furthermore, the Gln23 to stop change in shrewB18 mutants most likely renders this a nonfunctional protein and can be considered a null mutation.
We also examined the expression of CG11582 in wild-type embryos. By in situ hybridization, we were unable to detect the presence of CG11582 message. However, when we performed reverse transcriptase-polymerase chain reaction (RT-PCR) on total RNA isolated from 0–12 hr embryos we were able to demonstrate that the message is present in wild-type embryos (Fig. 6C). Therefore, we suggest that CG11582 message is expressed at low levels in the early embryo, possibly ubiquitously.
shrew Alleles Do Not Reflect a Complete Loss of Dpp Activity
A gradient of Dpp activity is established through the action of Tsg, Sog, and Tld, such that there are very high levels of signaling at the dorsal-most surface of the embryo, in the tissue that gives rise to the amnioserosa. This finding also leads to low levels of signaling at the dorsal epidermis/neuroectoderm boundary. Thus, we characterized the levels of Dpp signaling at various locations along the DV axis by examining the expression of known Dpp target genes. Specifically, high-level targets of Dpp include Race and rhomboid (rho), which are expressed in the presumptive amnioserosa and are dependent on high levels of Dpp activity for expression. Alternatively, Msh is expressed in the neuroectoderm and is a target for repression by low levels of Dpp signaling.
CG11582 is predicted to encode a Tsg-like protein, thus, we hypothesized that shrew mutations would not reflect a complete loss of Dpp signaling, but rather a reduction in the level of Dpp signaling in the dorsal-most tissues. To test this question, we examined the effect of loss of shrew on the expression of Race, rhomboid (rho), and Msh. In our deficiencies, as well as in shrew alleles, we found that the defects led to loss of the dorsal-most expression of Race and severely reduced rho expression (Fig. 7B,E). This finding is consistent with previously published reports that shrew mutants affect formation of the aminoserosa but not other dorsal tissues (Arora and Nusslein-Volhard,1992). The loss of Race and rho expression is not as profound as observed in dpp mutant embryos. In shrew mutants, there is at least some rho expression across the dorsal surface, and Race expression, while absent in the torso, remains expressed in the head regions in a rather broad flat domain (compare Fig. 7B with C).
It has been shown that Msh expression is expanded from its normal domain in the central nervous system across the dorsal surface of the embryo in dpp mutants (D'Alessio and Frasch,1996; Von Ohlen and Doe,2000). However, Msh expression is not expanded dorsally in shrew mutants, indicating the Dpp is still able to repress target genes at the dorsal epidermis/neuroectoderm boundary (Fig. 7H,I). In fact the appearance of the Msh stripe in the shrew mutants was narrower than in wild-type embryos. This finding suggested that Dpp signaling is actually better able to repress transcription at lateral positions in the absence of shrew activity. Taken together, these data support the hypothesis that maximal dpp activity necessary for amnioserosa formation is dependent on shrew activity. However, the lower levels of Dpp activity needed to repress gene expression at the dorsal epidermis/neuroectoderm boundary are not dependent on shrew activity. We conclude that Shrew activity is required for maximal Dpp signaling at the dorsal most surface of the embryo. This conclusion is consistent with the idea that Shrew functions to establish the proper gradient of Dpp signaling in the early embryo.
An additional readout of Dpp signaling activity is the presence of phosphorylated Mothers against dpp (pMAD; Dorfman and Shilo,2001). In blastoderm stage Drosophila embryos, the highest levels of pMAD can be detected in a stripe at the dorsal surface of the embryo (Fig. 8A), this corresponds to the highest levels of Dpp signaling activity. When we examined the expression of pMAD in shrew mutant embryos, we found no obvious peak of pMAD expression (Fig. 8B). We conclude from this that shrew activity is required for maximal Dpp signaling at the dorsal most surface of the embryo.
We screened the Exelexis deficiency collection for potential modifiers and regulators of Ind expression as well as potentially novel regulators of DV polarity. One advantage to this group of deficiencies is that they are relatively small deletions and have precisely mapped endpoints. This resulted in the identification of several genes known to be involved in DV patterning, as well as regulation of ind expression, including Egfr, dpp, and ind. However, one disadvantage is, the collection only uncovers approximately 56% of the euchromatic genome and, therefore, many potential components of the known DV pathways may not be included in this collection. This is evident in the fact that several genes known to regulate either ind expression or DV patterning were not identified in our screen. Examples of genes not represented in the Exelexis collection included vnd, tsg, sog, and tld.
Here, we present the identification of a pair of overlapping deficiencies that remove function of the conceptual gene CG11582. CG11585 encodes a putative Tsg-like protein and is, therefore, predicted to be involved in regulation of Dpp activity. These deficiencies are allelic with each other and mutant alleles of shrew (Jurgens et al.,1984). In addition, we have sequenced the CG11582 locus in two independent shrew mutations and in both cases have found point mutations leading to altered amino acid sequence. Our results represent the first evidence that shrew encodes a twisted gastrulation-like protein that is essential for maximal Dpp signaling at the dorsal surface of the embryo. It has been previously suggested that CG11582 is the gene affected in shrew mutants. However, this has not been definitively shown until this work (Vilmos et al.,2001; Oelgeschlager et al.,2004). We screened the Exilexis deficiency collection for potential modifiers and regulators of Ind expression. One advantage to this group of deficiencies is that they are relatively small deletions and have precisely mapped endpoints. However, one disadvantage is that, at this point, the collection only represents approximately 56% of the genome and, therefore, many potential regulators of the known DV pathways may not be included in this collection. This is evident in the fact that several known components of the signaling pathways were not identified, including vnd, tsg, sog, and tld.
During the early stages of embryonic development, Dpp signaling serves to pattern the dorsal surface of the embryo and to subdivide the domain into two tissues amnioserosa and dorsal epidermis. While dpp message is expressed uniformly across the dorsal surface, one readout of Dpp signaling, phosphorylated mothers against Dpp (pMad), suggests that there is a stepwise gradient of Dpp activity. The presence of pMad is at its highest concentrations at the dorsal most point where aminoserosa is forming, and lower levels in more lateral regions that give rise to dorsal epidermis (Dorfman and Shilo,2001; Wharton et al.,1993). Dpp activity also serves to set the dorsal boundary of the CNS by positioning the dorsal limit of Msh expression (D'Alessio and Frasch,1996; Von Ohlen and Doe,2000). The gradient of Dpp activity is accomplished by the action of the at least three extra cellular molecules that act to transport and modulate Dpp activity. The combinatorial action of Short gastrulation (sog), Twisted gastrulation (tsg), and Tolloid (tld) serves to redistribute the Dpp ligand to the dorsal surface of the embryo where the highest levels of Dpp activity are detected (reviewed in Raftery and Sutherland  and O'Connor et al. ). Sog, which is expressed in the ventral neurectoderm, serves to sequester and inactivate Dpp. Diffusion of Sog dorsally from the source of expression carries Dpp toward the dorsal-most region of the embryo. At the dorsal surface of the embryo Sog is cleaved by the protease Tld. This results in the release of Dpp from the complex and facilitated binding to the Dpp receptors. Although we have not definitively shown that Shrew interacts with this complex, our data suggests a role for Shrew in directing formation of the Dpp gradient. Thus, we would like to propose that Shrew is in fact capable of complexing with these proteins.
Our results suggest that Shrew activity is required for maximal activation of Dpp signaling at the dorsal-most surface of the embryo but not at the dorsal ectoderm/ventral nervous system boundary. An additional Twisted gastrulation like protein, Crossveinless (cv) has also been shown to be required for proper formation of the Dpp gradient in pupal wings (Shimmi et al.,2005; Vilmos et al.,2005). How Shrew functions in conjunction with this pathway remains unclear. However, while tsg and cv both have two CR domains, Shrew has only one. The CR domain found in Shrew bears the closest sequence similarity to the second domain in Tsg and Cv (Fig. 2b). Interestingly it has been suggested that the second CR domain of Tsg in Xenopus, is the domain responsible for the pro–bone morphogenetic protein activity (Oelgeschlager et al.,2003). In fact, the observation that the Msh stripe is narrower in shrew mutants than in wild-type embryos suggests that there is either more Dpp available for signaling at the neuroectoderm boundary or that it is better able to signal. Thus, we predict a role for Shrew in localization or activation of the Dpp ligand at the dorsal surface of the embryo.
Embryos from yw flies were used as wild-type with respect to ind expression. The Exelexis deficiency collection was obtained from the Bloomington Stock Center (Parks et al.,2004). This collection includes approximately 56% coverage of the euchromatic genome. Other fly stocks Df (3R)Espl22, tx2/TM6B, Tb1; hbb1rsd1/TM3, Sb1; shrewB18pp/TM3ThB8lacZ Sb Ser, shrew1st1eb1/TM3 were obtained from the Bloomington stock center. Df(3L)Exel9000 females were crossed to shrew1 males, we found that 17 of 69 (24.6%) progeny embryos from this cross also produced the ventralized phenotype.
In Situ Hybridizations and Antibody Stains
In situ hybridizations were done according to standard methods using digoxigenin-labeled antisense ind probe (Tautz and Pfeiffle,1989). Initially, 0–16 hr overnight collections were fixed for in situ hybridization and stained with antisense ind probe. Any line exhibiting any unusual pattern of ind expression was then set aside for rescreening. A total of 198 lines were re-screened for various reasons. From these, 0–8 hr collections from each stock were fixed and stained for ind by in situ hybridization. Additional probes, including Race and rho, were made from expressed sequence tag clones obtained from the Drosophila Genome resource center (DRGC). Rabbit Anti-Msh was used at a concentration of 1:500 (Isshiki et al.,1997). Mouse anti-pMAD was used at 1:500, this was a gift from Chris Q Doe.
Cloning of CG11582
We cloned CG11582 from genomic DNA isolated from yw flies. The following PCR primers were used: CG11582 forward ccgtctcgagttgctatcctttggagag and CG11582 reverse ctactctagaccgataattagcctctag. The 418-bp product was cloned in to pBluescript KS between the XhoI and XbaI sites.
Detection of CG11582 Message
RT-PCR was done with Access RT-PCR kit (Promega) according to the manufacturer's instructions. Total RNA was isolated from 0–12 hr yw embryos using TRIZOL reagent (Invitrogen). Primers for CG11582 are as described above. Control primers for amplification of ind message are IndHD forward atatcatctagactgatcaacgattacgcc; and IndHD reverse tattcaggtaccggtgattgattctacgcc.
Sequencing of Shrew Mutants
shrewB18pp/TM3ThB8lacZ Sb Ser; shrew1st1eb1/TM3 lines were balanced over the TM3 Kr green fluorescent protein (GFP) balancer chromosomes. The CG11582 locus was amplified by PCR using the primers described above and subjected to direct sequence of PCR products. DNA from heterozygous flies for each allele was sequence twice from each direction for a total of four independent PCR and sequencing reactions for each allele. The primers used for sequencing of CG11582 are Foward tcc ttt gga gag tta ggc and Reverse cct cta gct gct gac aag. Direct sequencing of PCR products was performed by the facility in the department of Plant Pathology at Kansas State University.
We thank Dr. Chris Doe and S. Keith Chapes for helpful comments on the manuscript and M.B. O'Connor for sharing unpublished information. This publication was made possible by a grant from the NCRR, a component of the NIH (Grant # P2-RR016475). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of NCRR or NIH.
- 1992. Altered mitotic domains reveal fate map changes in Drosophila embryos mutant for zygotic dorsoventral patterning genes. Development 114: 1003–1024. , .
- 1996. msh may play a conserved role in dorsoventral patterning of the neuroectoderm and mesoderm. Mech Dev 58: 217–231. , .
- 2001. Biphasic activation of the BMP pathway patterns the Drosophila embryonic dorsal region. Development 128: 965–972. , .
- 1992. Localized enhancement and repression of the activity of the TFG-β family member, decapentaplegic, is necessary for dorsal-ventral pattern formation in the Drosophila embryo. Development 114: 583–597. , .
- 1997. The role of the msh homeobox gene during Drosophila neurogenesis: implication for the dorsoventral specification of the neuroectoderm. Development 124: 3099–3109. , , .
- 1984. Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster. Rouxs Arch Dev Biol 193: 283–295. , , , .
- 1998. Dorso-ventral patterning in the Drosophila CNS: the vnd homeobox gene specifies ventral column identity. Genes Dev 12: 3606–3612. , , , , , .
- 1995. Dorsal-Ventral patterning genes restrict NK-2 homeobox gene expression to the ventral half of the central nervous system of Drosophila embryos. Dev Biol 171: 306–316. , .
- 2005. Formation of the BMP activity gradient in the Drosophila embryo. Dev Cell 8: 915–924. , , , , , , , , .
- 2006. Shaping BMP morphogen gradients in the Drosophila embryo and pupal wing. Development 133: 183–193. , , , .
- 2003. The pro-BMP activity of Twisted gastrulation is independent of BMP binding. Development 130: 4047–4056. , , , , , .
- 2004. Identification of a second Xenopus Twisted Gastrulation gene. Ind J Dev Biol 48: 57–61. , , , .
- 2004. Systematic generation of high-resolution deletion coverage of the Drosophila melanogastar genome. Nat Genet 36: 288–292. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , .
- 2006. Developmental control of nuclear morphogenesis and anchoring by charleston, identified in a functional genomic screen of Drosophila cellularization. Development 133: 711–723. , , , , .
- 2003. Gradients and thresholds: BMP response gradients unveiled in Drosophila embryos. Trends Genet 19: 701–707. , .
- 2005. The crossveinless gene encodes a new member of the Twisted gastrulation family of BMP-binding proteins which, with Short gastrulation, promotes BMP signaling in the crossveins of the Drosophila wing. Dev Biol 282: 70–83. , , , .
- 2005. Localized repressors delineate the neurogenic ectoderm in the early Drosophila embryo. Dev Bio 280: 482–493. , .
- 1989. A non-radioactive in situ hybridization method for the localization of specific RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback. Chromosoma 98: 81–85. , .
- 2001. The Twisted gastrulation family of proteins, together with the IGFBP and CNN families comprise the TIC superfamily of cysteine rich secreted factors. J Clin Pathol 54: 317–323. , , , .
- 2005. crossveinless defines a new family of twisted-gastrulation-like modulators of bone morphogenetic protein signaling. EMBO Rep 6: 262–267. , , , .
- 2000. Convergence of Dorsal, Dpp and Egfr signaling pathways subdivides the Drosophila neuroectoderm into three dorsal-ventral columns. Dev Biol 224: 362–372. , .
- 1998. Dorsoventral patterning in the Drosophila central nervous system: the intermediate neuroblasts defective Homeobox gene specifies intermediate column identity. Genes Dev 12: 3591–3602. , , , , , .
- 1993. An activity gradient of decapentaplegic is necessary for the specification of dorsal pattern elements in the Drosophila embryo. Development 117: 807–822. , , .