|Home | About | Journals | Submit | Contact Us | Français|
In order to establish and maintain the dorso-ventrally (DV) or anterior-posteriorly (AP) restricted stripes of homeodomain gene expression in the developing embryo it is necessary to coordinate multiple signaling pathways and the assembly of transcription factors on the regulatory DNA of each gene (Skeath, 1998; Von Ohlen and Doe, 2000) (Yagi et al., 1998; Miller et al., 2001). Regulation of the DV restricted homeodomain proteins Ventral nervous system defective (Vnd) and Intermediate neuroblasts defective (Ind) in the embryonic neurectoderm has been investigated extensively. For both of these genes, multiple enhancer elements have been identified that control different aspects of expression. For example, initiation of vnd expression is controlled by a 348 bp element located in the first intron of the gene, called the vnd NeuroEctodermal Enhancer (NEE) (Markstein et al., 2004). This element contains binding sites for the transcription factors Twist a mesoderm and ventral neurectoderm specific transcription factor, and Dorsal a global regulator of DV patterning. Both Twist and Dorsal are known to activate vnd in the ventral neurectoderm (Mellerick and Nirenberg, 1995; Stathopolous et al., 2002; Markstein et al., 2004). Later aspects of vnd expression, including expression in neuroblasts, are controlled by regulatory elements located upstream of the coding sequence (Saunders et al., 1998; Estes et al., 2001). These elements control activation in neuroblasts as well as repression in the midline (Saunders et al., 1998; Estes et al., 2001; Shao et al., 2002). Likewise, Ind expression is also controlled by multiple regulatory elements. Initiation of ind is associated with an element located downstream of the coding sequence (Weiss et al., 1998; Cowden and Levine, 2003; Stathopolous and Levine, 2005). This element contains binding sites for Dorsal, known to be essential for activation of Ind and Vnd. Also present are Vnd binding sites, which is the mechanism of Vnd repression of ind in the ventral neuroectoderm (Weiss et al., 1998; Von Ohlen and Doe, 2000; Stathopolous and Levine, 2005). An additional Ind regulatory element located upstream of the coding sequence participates in Ind-dependent maintenance of Ind expression (Von Ohlen et al., 2007b). Despite extensive knowledge of how expression of Ind and Vnd are regulated, very little is known about the regulation of muscle segment homeobox (msh) expression and until now, no regulatory DNA has been identified. Interestingly, the regulatory DNA necessary for initiation of both vnd and ind appears to contain elements necessary for both activation of expression in the appropriate domains and repression in inappropriate domains. Because the enhancer and silencer elements for these genes and presumably msh are inseparable, we will refer to these as regulatory elements and not enhancer elements for the remainder of this manuscript.
Our work has been extensively focused on the role of the homeodomain transcription factor Ind in patterning and cell fate specification of the Drosophila embryonic nervous system. In this work we wanted to identify msh regulatory DNA capable of driving reporter gene expression in a pattern consistent with msh expression in the embryonic neuroectoderm, and to demonstrate the requirement for Ind in limiting the ventral boundary of msh expression. We have identified essential regulatory DNA that controls msh expression in the early embryonic neuroectoderm. This regulatory element is not only essential for activation of msh in the lateral neuroectoderm, but it also contains sequences essential for repression by Vnd and more specifically Ind in the ventral and intermediate neuroectoderm. Thus, as with Vnd and Ind, the regulation of msh involves an element that controls both activation in the lateral column and repression in more ventral regions of the neuroectoderm. We also identified an additional element responsible for expression in the lateral mesodermal cells. We believe that this element may be responding to regulation by the mesoderm specific transcription factor Tinman (tin). Finally, we demonstrate that, unlike Vnd and Ind, Msh does not appear to be essential for initiation or maintenance of its own expression.
We used a variety of approaches to identify potentially important msh regulatory elements. Initially we use Evoprinter to identify five regions of highly conserved sequence among several Drosophilid species (Figure 1A, green boxes and Figure S1). (Odenwald et al., 2005) We refer to these regions as msh1-4 from furthest upstream of the coding sequences, and mshI for the piece located in the intron (Figure 1A). We used Cis-analyst and other search programs to search for binding sites of genes known to regulate Msh expression (Berman et al., 2001). Specifically, we looked for Vnd, Ind and Mothers against Dpp (Mad) binding sites, the best-known early regulators of Msh expression (D'Alessio and Frasch, 1996; Von Ohlen and Doe, 2000). The combination of these approaches identified two regions of interest. The first region we chose to focus on, named msh2, is located approximately 6.4 to 8.5kb upstream of the coding sequences. The second region named mshI, is located within an intron closely adjacent to the first exon.
Recent identification of transcriptional targets of mesoderm specific factors and DV patterning transcription factors has allowed us to narrow down the potential regions of msh regulatory DNA (Sandmann et al., 2007) (Zeitlinger et al., 2007). The chromatin-IP experiments done by these groups identified a region corresponding to our region 2. One group also identified the msh intron region as a region bound by Twist and other mesoderm specific factors (Sandmann et al., 2007). It is known that msh is expressed in and required for muscle development (Lord et al., 1995; Nose et al., 1998). Since Twist is a major regulator of mesoderm formation, it is entirely possible that the regions identified by these groups are muscle specific enhancers for msh. In order to confirm that potential regions are true regulatory elements we generated reporter constructs with the msh2 and mshI regions in a lacZ reporter vector. Because the other regions did not meet both our criteria we did not investigate them further. Our experiments indicate that the msh2 region contains the regulatory sequences controlling the early aspects of Msh expression (Figure 1B-E). Specifically, we first saw stripes of expression in a pattern similar to the earliest msh expression around stage six (Figure 1B). From stages 7 to 10 the lacZ message was detected in stripes in the presumptive neuroectoderm in a pattern that closely resembles the endogenous msh mRNA pattern (Figure 1B-C). By late stage 11, the LacZ message had largely faded from the neurectoderm, and expression remained in the salivary gland placodes (Figure 1D). Thus, these results indicate that we have identified a regulatory element controlling early DV restricted Msh expression.
Based on the presence of predicted Vnd binding sites, the intron piece (mshI) was also tested for its ability to drive reporter gene expression in the embryo. We found that this reporter construct also expressed in the developing embryo. However, not in the neurectoderm, instead, we observed expression in segmentally repeated spots of cells located laterally in stage 10-11 embryos (Figure 2A). These cells could be the lateral mesodermal precursors, a tissue known to express Msh (D'Alessio and Frasch, 1996). To test this we performed double labels with anti-Msh and anti-βgal antibodies. We found that the expression of Msh and βgal were only partially over-lapping (Figure 2C-E). These data suggested that there might be additional regulatory information required to precisely duplicate the endogenous Msh pattern in the lateral cells. Because Msh is also expressed in the larval imaginal discs during development, we also examined both the Msh2-lacZ and MshI-lacZ reporter lines for expression during larval stages and did not detect expression in third instar larval wing or leg discs for any of the lines tested (Data not shown). This suggests we have successfully identified regulatory DNA controlling expression of msh in the neurectoderm and also an element that partially functions in expression in the lateral mesodermal precursors.
Ind is a direct regulator of msh expression. Thus, we predicted that the msh2 element would contain sequences necessary for repression by Ind. To test this hypothesis, we first performed double label experiments for lacZ message and Ind protein. We found that the domains of lacZ and Ind were mutually exclusive just as the endogenous patterns are (Figure 3C-D). In order to examine the effect of loss of ind on reporter gene expression, we recombined the Msh2-lacZ transgene onto an ind mutant chromosome. This experiment revealed that like the endogenous msh message, the expression of lacZ was expanded ventrally in the ind mutant embryos (Compare Figure 3B to F and G to H). We also tested the ability of ectopically expressed Ind to repress reporter expression in the lateral column. To do this, we built recombinant chromosomes in which the reporter transgene was recombined onto the chromosomes carrying both the Krüppel Gal4 driver transgene and the UAS Ind transgene. Again we found that ectopic expression of Ind efficiently repressed Msh2-lacZ expression (Figure 3I). Taken together, these results strongly support the hypothesis that the region of genomic DNA contained in the Msh2-lacZ reporter construct, not only contains sequences essential for initiation of msh in the lateral neurectoderm, but also the sequences essential for repression of msh in the intermediate column.
Vnd can directly repress Msh expression (Mc Donald et al., 1998). To determine if the Msh2 element we identified is directly repressed by Vnd, we also built vnd; indMsh2-lacZ double mutants to produce double mutant embryos. Msh protein expands ventrally in vnd; ind embryos (Von Ohlen and Doe, 2000). Therefore, we anticipated a similar result if the Msh2 regulatory element also contained sequences necessary for repression by Vnd. We also predicted that ectopic expression of Vnd should be able to repress transcription of the reporter construct. Indeed, in vnd; ind Msh2-lacZ embryos lacZ message was observed throughout the neurectoderm (Figure 3J). Also, ectopic expression of Vnd across the DV axis was sufficient to repress transcription of the reporter construct (Figure 3K). These data suggest that the Msh2 regulatory DNA contains the information necessary for repression by both Vnd and Ind in ventral neurectodermal tissue.
We wanted to further characterize the action of Ind on the msh regulatory element. To do this, we needed to identify the minimal sequence required for expression in the early neurectoderm. Four additional reporter constructs were built with partially overlapping pieces of the original Msh2 element (Figure 4A). These experiments revealed that the Msh2a fragment positioned furthest from the coding sequences produced a pattern of reporter gene expression largely consistent with endogenous msh expression (Figure 4B-E). The Msh2b element drove reporter gene expression in the dorsal epidermis in a pattern reminiscent of highly responsive Dpp targets (Figure 4F). We failed to detect expression in embryos from elements Msh2c or Msh2d (data not shown). These results demonstrate that we have narrowed the region essential for initiation of msh expression in the lateral neurectoderm and repression in the ventral and intermediate neurectoderm down to the 699bp fragment encompassed by the Msh2a element. In this region we found four putative Ind binding sites and four putative Vnd binding sites. The presence of putative binding sites for Vnd and Ind suggests that the msh2a element contains the sequences necessary for repression of msh in the ventral neuroectoderm.
If Ind is acting to directly repress msh through the msh2a regulatory element, then we predicted that Ind would bind to sequences in the msh2a fragment in a sequence specific manner. To test this hypothesis, we identified four potential Ind/homeodmain binding sites within the sequence of the Msh2a region. For each of these there were pairs of two sites within 20 bp of each other. We designed probes for DNA binding experiments that contained each of the pairs of sites (Figure 5A & D). We tested the ability of Ind to bind to a fragment containing sites one and two. For these experiments, we used the IndHD protein described in (Von Ohlen et al., 2007b), to assess binding. We found that when sites one and two were intact, IndHD protein was able to bind to and alter the mobility of the fragment in a non-denaturing gel (Figure 5B). When we mutated the sequence of each of these sites to sequences that no longer allow Ind binding, we found that mutation of site two resulted in loss of ability of IndHD to bind to the fragment (Figure 5C). These data suggested, that Ind bound specifically to site two but not site one. Next, we examined the ability of Ind to bind sites three and four in the same manner. We found that Ind bound to and altered the mobility of a fragment containing both sites (Figure 5E). However, when we mutated the sequence of the sites, we found that Ind bound to the fragment mutated for site three but not for site four (Figure 5F). This demonstrates that binding is dependent on an intact site four. We conclude that Ind-mediated regulation of msh is via at least two sites in the Msh2a regulatory DNA.
An alignment of the msh2a sequences from six of the Drosophila species used to generate the evo-print reveals interesting details about predicted versus actual Ind binding sites (See supplemental Figure S2). We predicted Ind binding site based on a five out of six base pair match to the homeodomain recognition sequence TAATGG, described in (Zhao et al., 2007). When we compared sequences from the different species, we found that the sequences for site 2, 3 and 4 are highly conserved in at least four of the six sequences. Site one is only conserved in the two most closely related species and does not conserve the TAAT core consensus sequence that is typical for homeodomain proteins. Because site one lacks the TAAT sequence it is not surprising that Ind failed to bind this sequence. Both sites 3 and 4 are conserved in all six species and retain the TAAT core consensus sequence. It is unclear what the difference between these two sites but we hypothesize that additional flanking sequences may play a role in Ind binding. What these analyses lead us to conclude is that Ind binding requires an exact match to the TAAT or ATTA core sequence within the binding site and that additional flanking sequence may be important for efficient binding.
Once we had identified the sites essential for Ind binding to the msh2a element, we wanted to confirm that these were the functional Ind binding sites in vivo. Therefore, we mutated sites two and four in the Msh2a element to sequences that would no longer support Ind binding, and generated reporter transgenic lines to test the effect on expression of the reporter gene lacZ. We then examined expression of the reporter gene with either lacZ antisense probe or anti-βgal antibody in conjunction with anti-Ind antibody. We found that the expression of lacZ in the mutant reporter construct was slightly wider and expanded ventrally compared with expression of the wild-type reporter (Figure 6; compare D & E to A &B). In the embryos carrying the wild-type Msh2a-lacZ, transgene we occasionally observed some over-lap between the expression of βgal and Ind (not shown), suggesting that there are additional Ind binding sites located outside the msh2a region that regulate more complete repression in the intermediate column. However, when we compared the msh2a-mutant lacZ expression pattern to the wild-type reporter, we found a higher degree of overlap between the Ind protein pattern and βgal expression (Figure 6F). These results strongly support the hypothesis that, the Ind binding sites identified above (Figure 5) are at least partially responsible for Ind mediated repression of msh2a reporter gene expression.
Both Vnd and Ind are known to be essential for maintaining their own expression (Saunders et al., 1998; Uhler et al., 2007; Von Ohlen et al., 2007b). Thus, we tested the hypothesis that the Msh homeodomain protein served a similar role in maintaining msh expression. To do this, we took two approaches: First we tested for the presence or absence of msh message in embryos that were homozygous for loss of function alleles of msh. For these experiments we used the mshSOD54 allele described by (Mozer, 2001). We chose this allele because it was generated by EMS mutagenesis and had the potential to produce transcript. Other described alleles were generated by P-element excision and are known to take out the transcription start site (Isshiki et al., 1997). We found that msh mRNA was detectable in mshSOD54 homozygous mutant embryos in the normal pattern throughout the early stages of neurogenesis (Figure 7D-F). Second, we recombined the Msh2-lacZ reporter transgene onto the mshΔ68 mutant chromosome. When we examined reporter gene expression we found that lacZ message was expressed in the embryos in a pattern similar to the wild type expression pattern (Figure 7G-I). Thus, Msh activity was not exclusively necessary for the maintenance of either endogenous message or reporter gene expression. Other genes, perhaps those required for initiation of msh expression, are also essential for maintenance of the pattern. Furthermore, additional sequences not identified here may also be required to confer msh dependent expression.
In Drosophila and in vertebrates expression of proteins in precisely delimited stripes is critical for proper development if the embryo and differentiation of specific cell types. Although these stripes can be in different orientations and tissues we have focused on the formation DV restricted stripes in the embryonic neuroectoderm. The mechanism of regulation of msh expression in the neurectoderm of Drosophila embryos has not been studied at the molecular level.
This article represents the first description of regulatory DNA essential for expression of msh in the early neuroectoderm of the embryo and direct repression by Ind. We identified two regions of genomic DNA that potentially could function as neuroectodermal regulatory DNA. One of these regions (mshI) is located in the intron between the first and second exons. This element appears to be partially able to reproduce msh expression in the lateral mesodermal precursors. The second region (msh2) located 6.5 to 8.7kb upstream of the coding sequences, is capable of driving reporter gene expression in a pattern consistent with the endogenous msh pattern in stripe in the early embryonic neuroectoderm. Further analysis of this element identified a 699bp fragment (msh2a) that functions as a minimal regulatory element. This element was directly bound by the Ind homeodomain in a sequence specific manner. We identified four putative binding sites for Ind in the minimal element. We found that, Ind efficiently bound two of these sites and that this interaction was required for maximal repression of a reporter gene in the intermediate column. Furthermore, the identified regulatory elements are each required for specific subsets of the embryonic expression pattern. Thus far, we have examined expression of reporter constructs for these two conserved regions of potential regulatory DNA, in the embryo. We also tested for expression in larval tissues such as imaginal discs and brain, and failed to detect any reporter gene expression in these tissues. However, the pattern of msh expression in the embryo and throughout development is quite dynamic. Thus, the other regions of conserved sequence may include regulatory information essential for other aspects of msh expression including expression in the muscle precursors and larval tissues. We also demonstrated that, Msh function was not required for maintenance of msh expression in the embryonic neuroectoderm. This is interesting because both Ind and Vnd are required for maintenance of their own expression (Shao et al., 2002; Von Ohlen et al., 2007b). Our results point to a different mechanism of maintenance for msh that does not require an auto-regulatory interaction.
msh regulatory DNA was identified based on multiple criteria including; conservation of sequence among Drosophilid species and presence of binding sites for known regulators of msh expression. The msh2 regulatory region and the mshI regulatory region were both found to contain exact matches to the known Vnd binding site (Gruschus et al., 1997). Although the putative Vnd binding sites in the mshi region meet the criteria for Vnd binding sites they are more likely to be Tin-binding sites instead. Tin and Vnd have identical or highly similar DNA binding specificities (Chen and Schwartz, 1995; Gruschus et al., 1997; Weiss et al., 1998). Although we have not tested this hypothesis directly, we believe that because the mshI-lacZ reporter expresses in the mesodermal precursors, and Tin is a master regulator of mesoderm specification, that the sites we identified as Vnd sites in the mshI region, are more likely Tin sites.
Vnd and Ind interact with additional co-factors such as sox domain proteins Dichaete and SoxNeuro (soxN) (Buescher et al., 2002; Overton et al., 2002) (Zhao and Skeath, 2002) (Zhao et al., 2007). For example, Ind and Dichaete interact on the achaete regulatory DNA, and this interaction is necessary for effective repression of achaete in the intermediate column (Zhao et al., 2007). Because we believe these types of interactions are highly variable and dependent on which regulatory DNA the homeodomain proteins are interacting with, we explored how msh regulatory DNA is affected by these interactions. Dichaete interaction has also been implicated in Ind and Vnd mediated repression of Msh (Zhao and Skeath, 2002). However, our data suggest that interaction with Dichaete is not essential for regulation of Msh by Ind. Our reasoning for this is multi-fold, first in the absence of Dichaete function, msh expression appears completely normal and not partially expanded ventrally (Buescher et al., 2002; Overton et al., 2002). Second, we have carefully examined the sequence of the msh2a regulatory element defined here, and cannot identify any potential Sox domain binding sites. Finally, ectopically expressed Ind was an efficient repressor of msh in the lateral column, despite of the fact that Dichaete is only expressed in the ventral and intermediate columns (Von Ohlen et al., 2007a). Thus, association between Ind and Dichaete is not essential for Ind to repress msh expression.
UASInd (Von Ohlen et al., 2007b), KrGal4, yw, tin346/TM3ftzlacZ was obtained from Rolf Bodmer (San Diego); ind 16.2/TM3ftzlacZ (Weiss et al., 1998); mshS0D54/TM3ftzlacZ was obtained from Brian Mozer (NIH) (Mozer, 2001); mshΔ68/TM3ftzlacZ (Isshiki et al., 1997); UASvnd and vnd6/FM7 (Skeath et al., 1994; Mc Donald et al., 1998).
An msh evoprint was generated with the help of Thomas Brody at the NIH (Odenwald et al., 2005). msh2 and mshI fragments were amplified by PCR from genomic DNA isolated from yw flies, using the following primers: msh2F: agttgccccgaggtctggagtt and msh2R: gctggctttcgttggctttctgc. MshIF ccgctcgagatttagagccttaacatc and mshIR: ccgctcgaggttaaattatgaaggcac.
Msh2a-d constructs were generated by the same approach using the following primer sets:
In all cases PCR products were cloned using the pCR-TOPO4 cloning kit from Invitrogen, and subsequently cloned into the pCaSpeR Hsp43lacZ transformation vector at the EcoRI site. Transformation was done at the Non-mammalian model systems unit at Duke University. For each reporter constructs several independed insertion lines were tested. Figures show representative examples of the expression patterns observed.
Electrophoretic mobility shift assays were performed using the Histagged Ind Homeodomain protein (His6-IndHD) as described in (Von Ohlen et al., 2007b). WT and mutant oligos were annealed to form dsDNA molecules. EMSA assay was performed using the Dig non-radioactive Gel shift kit (second generation, Roche). The DNA/protein complexes were resolved on a 5% non-denaturing polyacrylamide gel. The probes used are as follows: Fragment 1
|Mut 1(TTATGG)||Mut 2 (GCATTA)|
|Fragment 2:||Mut 1(TTATGG)||Mut 2 (GCATTA)|
Mutated sequence is show in italics actual sequence for each site is shown in parenthesis above.
Putative Ind binding sites in the GN4 enhancer element we mutagenized using the Quick Change Multi-site directed mutagenesis kit (Stratagene). The following mutagenic primers were used to convert the core ATTA sequence to sequences that no longer support Ind binding. The Mutagenic oligos used are:
Italics indicate altered base pairs.
In situ hybriziations were performed according to standard procedures as described in: (Tautz and Pfeiffle, 1989). In all experiments shown we used a digoxygenin labeled anti-sense lacZ probe.
The antibody stains used rabbit anti-Ind (Von Ohlen and Moses, 2009) at a 1:2000 dillution and mouse anti-βGal at a 1:500 dillution (Promega). Secondary antibodies were biotin labeled anti-rabbit (Vector labs) was used in conjunction with the Vectastain Elite streptavidin-Horse radish peroxidase kit to produce a brown precipitate. Alkaline Phosphatase-conjugated anti-mouse (Cappel) was used to produce purple precipitate.
We are grateful to Dr. S. Keith Chapes for his help editing the manuscript. We are also grateful to Dr Tom Brody for help with generating and interpreting the Evoprint. This publication was in part, made possible by an Innovative Research Award, to TVO, from the Terry C. Johnson Center for Basic Cancer Research at Kansas State University. Additional funding was provided by: Grant Number P20 RR016475, to TVO, from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH). The contents of this publication are solely the responsibility of the author and do not necessarily represent the official views of NCRR or NIH.