- THIS ARTICLE
-
Abstract
- Full Text (PDF)
- Alert me when this article is cited
- Alert me if a correction is posted
- SERVICES
- Similar articles in this journal
- Similar articles in PubMed
- Alert me to new issues of the journal
- Download to citation manager
- Reprints & Permissions
- CITING ARTICLES
- Citing Articles via HighWire
- Citing Articles via Google Scholar
- GOOGLE SCHOLAR
- Articles by Vazquez, J.
- Articles by Schedl, P.
- Search for Related Content
- PUBMED
- PubMed Citation
- Articles by Vazquez, J.
- Articles by Schedl, P.
Deletion of an Insulator Element by the Mutation facet-strawberry in Drosophila melanogaster
Julio Vazqueza and Paul Schedlba Department of Biochemistry and Biophysics, University of California, San Francisco, California 94143
b Department of Molecular Biology, Princeton University, Princeton, New Jersey 08544
Corresponding author: Julio Vazquez, Department of Biochemistry and Biophysics, School of Medicine, Box 0448, University of California, San Francisco, CA 94143-0448., vazquez{at}msg.ucsf.edu (E-mail)
Communicating editor: S. HENIKOFF
 | ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Eukaryotic chromosomes are thought to be subdivided into a series of structurally and functionally independent units. Critical to this hypothesis is the identification of insulator or boundary elements that delimit chromosomal domains. The properties of a Notch mutation, facet-strawberry (faswb), suggest that this small deletion disrupts such a boundary element. faswb is located in the interband separating polytene band 3C7, which contains Notch, from the distal band 3C6. The faswb mutation alters the structural organization of the chromosome by deleting the interband and fusing 3C7 with 3C6. Genetic studies also suggest that faswb compromises the functional autonomy of Notch by allowing the locus to become sensitive to chromosomal position effects emanating from distal sequences. In the studies reported here, we show that a DNA fragment spanning the faswb region can insulate reporter transgenes against chromosomal position effects and can block enhancer-promoter interactions. Moreover, we find that insulating activity is dependent on sequences deleted in faswb. These results provide evidence that the element defined by the faswb mutation corresponds to an insulator.
INSULATORS are DNA sequences that confer position-independent expression to reporter genes and prevent or attenuate enhancer-promoter interactions in a position-dependent manner in transgenic assays (
Transgenic assays have been used to define some of the DNA sequences required for insulator function, and a number of proteins required for the activity of insulator elements have been identified. For instance, the gypsy transposable element disrupts gene expression by integrating within the regulatory region of Drosophila genes. The mutagenic effects of gypsy are largely due to the enhancer-blocking activity of the su(Hw) protein bound to a cluster of reiterated sites in the transposon (
Even though much has been learned about the structure and the activity of insulators in transgenic assays and their mutagenic activity when inserted in the regulatory region of genes, comparatively little is known about the normal function of insulators in the genome. For instance, while the mutagenic effects of SU(HW) binding sites in gypsy insertions are well known, the normal chromosomal binding sites for the su(Hw) protein have not been characterized, and their role in gene regulation is unclear. Similarly, no mutations have been recovered in scs or scs' that would shed light on their function at their normal location. One major exception may be provided by a set of mutations at the Drosophila bithorax complex, which disrupt regulation at the locus. One of these mutations, Fab-7, corresponds to a deletion of DNA sequences located between the cis-regulatory regions iab-6 and iab-7, which control expression of the Abd-B gene in parasegments 11 and 12, respectively. The Fab-7 element was shown to possess insulating activity in transgenic assays (
Clearly, it would be of interest to obtain additional evidence linking a genetically defined insulator element to specific defects in gene regulation or chromosome organization. For this reason, we turned our attention to the well-characterized Notch locus of Drosophila. The Notch gene functions in a conserved cell-cell signaling pathway that determines cell fate in a variety of different tissues and cell types (
|
The Notch locus has been the subject of intensive genetic analysis since its discovery in 1919 by Mohr, and a large number of mutations that affect Notch activity have been isolated and characterized. The properties of one of these, the recessive hypomorphic allele faswb, are consistent with those that might be expected for a mutation that disrupts or inactivates a domain boundary (
| MATERIALS AND METHODS |
|---|
| TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Transformation vectors:
Plasmid pRW is a Drosophila transformation vector containing a white gene with ~2 kb of upstream regulatory sequences including two tissue-specific enhancer elements active in the eyes and testes, respectively (
). This vector was used to assay for position effects in the absence of the white enhancers (white "minigene").
|
|
|
DNA constructions:
Plasmid pN35, containing a 4.5-kb EcoRI genomic fragment from the Notch locus (-30 kb to -25.5 kb), was digested with the appropriate restriction enzymes. Restriction fragments were separated on 1% agarose gels, purified, filled in with DNA Polymerase I (Klenow fragment), and cloned in one of the unique restriction sites of pRW or pRW. To generate construct Bgl-XhoHB, the BglII-XhoI fragment was first subcloned in pKS- (Stratagene, La Jolla, CA), digested with HindIII and BssHII, filled in with Klenow, and religated before being transferred to pRW. Construct Bgl-Xho [facet] contained a BglII-XhoI genomic fragment from the mutant line faswb. Two oligonucleotides from the Notch 5' region (5'-gccttatgattcctcgttgggttct-3' and 5'-gcagtgtgaccgcgtcggtgc-3') were used to PCR-amplify a 878-bp fragment [sequences 97 to 975 according to
Germline transformations:
Plasmids were injected at a concentration of 400 µg/ml, together with 100 µg/ml of helper plasmid pUChsD2-3wc ("pTurbo";
Phenotypic analysis of transgenic lines:
Flies were raised at 20°–22°. All estimations of insulating or enhancer-blocking activity were based on at least two sets of eye color determinations by visual inspection under the dissecting microscope, using 2- to 4-day-old females heterozygous for the transgene, or as otherwise indicated. To minimize variability, comparisons were made between flies of the same age grown under similar conditions. To assay for testis pigmentation, testes were dissected from heterozygous or hemizygous 7- to 10-day-old males. Generally, there was a good correlation between the level of pigmentation in eyes and testes (
To determine whether particular constructs (generally a set of test constructs and a control fragment) were similar or not, a two-tailed Student's t-test was performed, and the associated P-values determined (null hypothesis: test fragment is identical to control). This test calculates the probability that the difference between two sets of transgenic lines (test fragment and control) could be due to chance. A low P-value (0.05 or less) indicates that the two samples are very likely to be different.
In the position-effect assay using the white maxigene, eye colors were ranked on a scale of one to four, as described for the enhancer-blocking assay. Insulated transgenes generally give wild-type eyes (category 4), while noninsulated transgenes give a majority of lines with brown or light red eyes (category 3), in addition to wild type. Insulating activity was defined as the percentage of lines with wild-type eyes (category 4/N). The average eye color was calculated for each construct, and a one-tailed Student's t-test was used to determine the significance of the results as described above (a one-tailed test was used since the noninsulated constructs obligatorily have lighter eye pigmentation).
In the position-effect assay using the white minigene, eye colors were ranked on a scale from one to four as follows: 1, pale yellow; 2, yellow; 3, light orange; 4, dark orange or red. Insulated constructs generally give yellow eyes, while noninsulated constructs show a range of phenotypes from very pale yellow to dark orange, or sometimes darker. Insulating activity was defined as the percentage of lines with yellow eyes (category 2/N). When the minigene is used to test for position effects, phenotypic variations in the noninsulated lines deviate from the insulated phenotype in both directions (lighter and darker eye colors). In this case, the differences in the average eye color are minimal, and a test aimed at comparing mean values as above is meaningless. Instead, lines were placed in two categories: (1) yellow eyes and (2) eyes different from yellow, and a two-tailed Student's t-test was performed (null hypothesis: test fragment is identical to control). For all three assays, a comparison of the activities on the basis of either the percentage of blocking or the calculated average eye-color values led to essentially similar conclusions.
Chromatin mapping:
Nuclei were prepared from 0- to 12- and 0- to 24-hr embryos and from Drosophila Kc tissue culture cells and digested with DNase I or micrococcal nuclease as described in
Photography:
Eyes of 2- to 4-day-old transformed females carrying one copy of the reporter construct were photographed using a Nikon HFX-IIA stereomicroscope fitted with an FX-35WA camera and light meter. Illumination was from a Dolan Jenner Fiber Lite A 200 fiber optic light source. Kodak Ektar 100 negative color film was used. Color prints were scanned using a Hewlett-Packard ScanJet IIcx scanner and DeskScanII software on a Power MacIntosh 7100/66. Layout of the figures was done using Macromedia Freehand.
| RESULTS |
|---|
| TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Chromatin structure of the 3C6–3C7 interband region:
Cytological studies have placed the Notch transcription unit in polytene band 3C7, while the promoter region and upstream sequences are located in the 3C6–3C7 interband (
To analyze the chromatin organization of the 3C6–3C7 interband region, we used three probes derived from the 4.6-kb EcoRI fragment. The first probe was an EcoRI-BglII fragment from the distal end of the 4.6-kb fragment (probe a; Fig 2A). In EcoRI restricted chromatin digests, the EcoRI-BglII probe displays the pattern of nuclease cleavage reading from the distal EcoRI site toward the beginning of the Notch transcription unit. The second probe was a BglII-HindIII fragment from the middle of the 4.6-kb EcoRI fragment (probe b). In BglII restricted chromatin digests, this probe displays the nuclease cleavage pattern reading from this distal BglII site, through the faswb region and the Notch promoter, and into the main body of the Notch gene, up to a BglII site located 15 kb downstream. Finally, an XhoI-EcoRI fragment from the proximal end of the 4.6-kb EcoRI fragment (probe c) was used to display the nuclease cleavage products reading from near the beginning of the first Notch intron, through the transcription start sites and the faswb region, up to the distal EcoRI site.
The autoradiograms in Fig 3 show the pattern of DNase I and micrococcal nuclease (MN) cleavage in embryonic nuclei reading across the 3C6–3C7 interband region with probe c (Fig 3A) or probe a (Fig 3B). The chromatin organization of the DNA segment spanning the faswb region is summarized in Fig 2B. Within the Notch transcription unit, there are several prominent DNaseI hypersensitive (HS) regions, which are best visualized using the XhoI-EcoRI and BglII-HindIII fragments for end-labeling (data not shown). Two of these regions, A and B, are located in the first exon, close to the XhoI restriction site, while there are additional DNaseI HS regions in the large first intron. In addition to these internal DNaseI HS regions, a series of five prominent HS regions (sites 1–5) can be observed in the 5' region of Notch (Fig 3A). The Notch promoter region contains two major DNaseI HS sites, 1 and 2, which cover a DNA segment of ~150 bp. The promoter region also contains chromatin-specific MN cleavage products, including a broad region of nuclease sensitivity that appears to correspond to the 150-bp region between the distal and proximal promoters and a relatively strong MN site (site a) just upstream of the 5'-most initiation site (Fig 3A, lanes 2 and 3; Fig 3B, lanes 1 and 2). These chromatin-specific cleavage products are still present after extracting the embryonic nuclei with 0.35 M and 0.55 M NaCl, but largely disappear when the nuclei are extracted with 0.75 M NaCl, a salt concentration that results in extensive nucleosome shuffling (
Upstream of the promoter region are DNaseI HS regions 3, 4, and 5, which together cover a DNA segment of nearly 200 bp. Region 3 spans the proximal faswb breakpoint, while regions 4 and 5 are located entirely within the faswb deletion. Region 5 appears to coincide with the proximal copy of the 47-bp repeat. Although there are chromatin-specific micrococcal nuclease fragments from regions 3 and 4 (sites a and b; compare lanes 2 and 7 in Fig 3A, lanes 1 and 6 in Fig 3B), the two most strongly labeled MN fragments in chromatin digests are located just upstream. The first maps near the proximal copy of the 47-bp repeat (site c), while the second maps near the distal copy (site d). Both fragments also correspond to cleavage sites for the enzyme on naked DNA. However, both sites appear to be enhanced in chromatin. Furthermore, the cleavage for site c coincides with DNaseI HS region 5, suggesting that this micrococcal nuclease fragment is derived from a sequence that is exposed in chromatin. It should also be noted that there is a weak DNaseI site in the distal copy of the 47-bp repeat, corresponding to MN site d (see Fig 3A, lane 1). These results suggest the presence of specific nucleoprotein structures in the promoter region of Notch and extending into the upstream facet-strawberry sequences. In addition to the major DNAseI and MN sites, less prominent chromatin-specific sites appear to be distributed in the distal part of the region in a pattern that suggests the presence of an ordered nucleosomal array. These additional sites, however, are relatively weak and obscured by the presence of nonchromatin-specific cleavage sites, probably reflecting the generally open conformation of chromatin in this interband region.
Identification of an enhancer-blocking activity in the 3C6–3C7 interband:
Two different transgenic assays, an enhancer-blocking and a position-effect assay, have been used to determine whether particular DNA fragments can function as genetic insulators in Drosophila. In the enhancer-blocking assay, candidate elements are tested for their ability to prevent enhancer-promoter interaction when interposed between an enhancer and a promoter (
|
|
The 3C6–3C7 region insulates transgenes against position effects:
To further establish the existence of an insulator element in the 5' region of Notch, we tested the same BglII-XhoI fragment in a position-effect assay. In this assay, putative insulators are tested for their ability to shield reporter transgenes from chromosomal position effects at the site of insertion. In the absence of insulators, a transgene will show a range of phenotypes due to variable levels of expression in different transgenic lines, while transgenes flanked by a variety of insulator elements will show the same phenotype in all lines (
To test whether the BglII-XhoI fragment is able to insulate a reporter gene against position effects, this fragment was inserted in the XhoI site of pRW, upstream of the enhancer in the white maxigene reporter construct. Eleven independent transgenic lines were isolated. All were found to have identical, wild-type eyes (Table 2 and Fig 6). This result was similar to that obtained when the maxigene was flanked by scs and scs' (9 wild-type lines out of 9; construct scs). On the other hand, a noninsulated reporter construct (random) showed wild-type levels of pigmentation in only 2 out of 8 lines. Therefore, the BglII-XhoI fragment is able to insulate the reporter transgene against position effects.
|
|
We also tested the BglII-XhoI fragment in a position-effect assay using the white minigene. Five transgenic lines were obtained, all of which had yellow eyes similar to lines transformed with the white minigene insulated by scs (Table 3). On the other hand, a noninsulated minigene construct (random) gave a majority of lines with eye colors different from yellow. These results together indicate that the BglII-XhoI fragment is able to insulate the white gene against both positive and negative position effects. Furthermore, the Notch upstream DNA fragment is as effective, in our assays, as the previously characterized scs insulator element.
|
The facet-strawberry deletion impairs insulator function:
The experiments described in the previous section suggest that the BglII-XhoI fragment contains an element that can function as a genetic insulator. An important question is whether the insulating activity detected in these transgenic assays corresponds to the putative boundary element defined by the faswb mutation. To address this question, we generated a set of constructs in which the faswb sequences were deleted from the BglII-XhoI fragment.
In the first set of experiments, we tested the deleted fragment in the enhancer-blocking assay (Fig 5 and Table 1). Two different constructs were used. The first construct carried a genomic BglII-XhoI fragment isolated from a mutant faswb line, inserted between the white enhancer and promoter of pRW. Three transgenic lines were obtained for this construct. As indicated in Table 1, none of the lines showed a phenotype consistent with efficient blocking. Instead, these lines had orange or red eyes (as well as pigmented testes), suggesting that the enhancer-blocking activity of the mutant BglII-XhoI [facet] fragment was severely compromised. The second construct, BglXhoHB, is an "artificial" faswb deletion that was generated by removing sequences between the HindIII and BssHII restriction sites in the BglII-XhoI fragment (see Fig 8). The HindIII restriction site is located 33 bp upstream of the distal faswb breakpoint, while the BssHII restriction site is located 4 bp upstream of the proximal breakpoint (Fig 2). As shown in Table 1 and Fig 5, the enhancer-blocking activity of the deleted BglII-XhoI fragment was also substantially reduced, with 7 out of 10 lines showing higher than expected levels of pigmentation. The level of blocking of this construct (30%) was equivalent to that observed for the faswb fragment and suggests that the loss of enhancer-blocking activity is due to the loss of sequences in the facet-strawberry region. The effect of the mutation faswb on enhancer blocking is illustrated in Fig 7, which shows two representative enhancer-blocking lines carrying the wild type (fly a) or mutated fragment (fly b).
|
|
Even though the mutated fragment was substantially impaired in its ability to block the white enhancer, some residual insulating activity (~30%, compared to ~10% for the negative control) was still present. Since the phenotype of the mutation faswb has been proposed to arise from a position effect affecting Notch expression, we asked whether the mutated fragment would be sufficiently impaired in its ability to insulate a reporter transgene against position effects by testing the faswb BglII-XhoI fragment in the position-effect assay (i.e., when placed distal to the white enhancer; Table 2 and Fig 6). While all the lines (11/11) carrying the wild-type BglII-XhoI fragment had wild-type eye pigmentation, 7 out of 8 transgenic lines carrying the mutated fragment had reduced eye color. This result is virtually identical to that obtained with noninsulated transgenes (random). Therefore, we conclude that the mutated fragment is unable to insulate the white reporter gene against position effects. Taken together with the reduced enhancer-blocking activity of the faswb deletion, these results show that the sequences deleted in the faswb mutation are critical for the insulating activity detected in our transgenic assays.
Localizing the insulating activity:
To further define the sequences that confer enhancer-blocking activity, we tested a number of subfragments from the 3C6–3C7 interband region in the enhancer-blocking assay (Fig 8, Table 4). As shown previously, the BglII-XhoI fragment was able to block the white enhancer in 94% of the lines tested. In contrast, an overlapping EcoRI-HindIII fragment located in the distal part of the interband region showed blocking in less than one-third of the lines, localizing the insulating activity to the promoter-proximal part of the region. We then subdivided the 2.3-kb BglII-XhoI fragment into a distal BglII-HindIII fragment, containing sequences just upstream of the faswb region, and a proximal fragment, HindIII-XhoI, spanning sequences including the faswb region as well as the Notch promoter region. A fragment including an additional 240 bp of 3' sequences, HindIII-HindIII, was also tested. As shown in Fig 8 and Table 4, the BglII-HindIII fragment showed a modest enhancer-blocking activity (33%), similar to the overlapping EcoRI-HindIII fragment. On the other hand, both the HindIII-HindIII and HindIII-XhoI fragments showed enhancer-blocking activity in the majority of the lines (18/20 and 5/5, respectively). In addition, most of the lines had yellow eyes and unpigmented testes, indicating complete or very strong blocking of the white enhancer. The level of blocking of these two fragments was comparable to that of the larger BglII-XhoI fragment, suggesting that the insulating activity of the Notch upstream region is localized within a 1.3-kb HindIII-XhoI fragment including the faswb sequences and the Notch promoter region.
|
To determine whether the blocking activity of the HindIII-XhoI fragment was due mainly to sequences in the faswb region or whether the Notch promoter region might also be involved, we further subdivided this fragment into a distal HindIII-BssHII and a proximal BssHII-XhoI fragment. The HindIII restriction site is located ~40 bp upstream of the distal (5') faswb breakpoint, while the BssHII restriction site is located 4 bp upstream of the proximal (3') breakpoint (Fig 2). Therefore, this fragment encompasses all but 4 bp of the faswb region and includes an additional 40 bp of upstream sequences. On the other hand, the BssHII site is located, respectively, ~70 and 190 bp upstream of the two major transcriptional start sites of Notch (
An analysis of transgenic lines carrying the HindIII-BssHII or the BssHII-XhoI fragment indicates that the 0.9-kb upstream fragment was able to block or attenuate enhancer-promoter interactions in 50% of the lines, while the proximal, promoter-containing fragment did not show any significant enhancer-blocking activity (Fig 8 and Table 4). Since nuclease mapping experiments have revealed the presence of nuclease hypersensitive sites at or near the proximal faswb breakpoint (DNAseI HS site 3 as well as the proximal MN site; see Fig 2 and Fig 3A), it is possible that the BssHII site itself might be located in a region important for insulation. Therefore, we used the polymerase chain reaction to generate a DNA fragment, HindIII-BssHII(+), that extends an additional 60 bp downstream from the BssHII site (see MATERIALS AND METHODS). This fragment includes DNAseI HS sites 3–5, but ends ~10 bp upstream of the 5'-most transcription start site (Fig 2C). Two sets of transgenic lines that carried either one or two copies of the PCR-generated fragment (PCR-1 and PCR-2; Table 4 and Fig 8) were obtained. Both constructs showed a substantially increased insulating activity compared to the shorter HindIII-BssHII fragment (>80 vs. 50%). This activity was not significantly different from that of the larger HindIII-XhoI fragment. Therefore, we conclude that the exclusion of the Notch transcriptional start sites and possibly of most, if not all, transcriptional regulatory elements does not substantially impair the insulating activity. We cannot totally exclude, however, that some transcriptional regulatory elements might still be present in the insulating fragment (see DISCUSSION).
The Notch 5' region does not act through promoter competition:
As discussed in the previous section, the promoter-containing fragment (BssHII-XhoI) did not show any significant enhancer-blocking activity in our assay, but instead behaved in a manner similar to that observed with control (noninsulating) DNA fragments (Fig 8 and Table 4), suggesting that the Notch promoter does not compete with white. Since promoters have been shown to compete for regulatory elements (
While the Notch promoter fragment had no discernible effect in the enhancer-blocking assay, the hsp70 promoter constructs showed a weak but significant reduction in the level of expression of the white gene (Table 4). Gene dosage experiments indicate that the hsp70 promoter fragment (p70) causes a reduction of ~50% in the level of expression of white, while the larger p70Z construct showed an even more substantial reduction. It must be noted, however, that none of the lines obtained showed complete suppression of white, as indicated by the absence of flies with yellow eyes or unpigmented testis (Table 4 and data not shown). The greater reduction in white expression observed in p70Z lines is probably due to the increased distance between the white enhancer and white promoter (~5 kb in p70Z vs. 1.5 kb in p70). A similar distance effect has been observed among competing promoters at the human ß-globin locus (HB and Bgl-Hind in Fig 8).
| DISCUSSION |
|---|
| TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
In the studies reported here, we tested DNA fragments from the 3C6–3C7 interband region, corresponding to the distal end of the Notch locus, for insulating activity in two different transgenic assays. This interband contains the Notch promoter, as well as additional 5' sequences including an 880-bp DNA region deleted in the Notch mutation faswb. In the first assay, we found that a BglII-XhoI DNA fragment spanning the faswb region was able to confer position-independent expression to two different white reporter constructs, insulating them against both positive and negative position effects. In the second assay, we found that the same BglII-XhoI DNA fragment was able to block enhancer-promoter interactions when inserted between the enhancer and promoter of the white gene, but did not reduce white expression when located upstream of the enhancer. Therefore, the Notch 5' region contains an activity similar to previously described insulators (
That the sequences defined by the faswb mutation are critical for insulating activity is demonstrated by the analysis of two deletion derivatives of the BglII-XhoI DNA fragment. One derivative, Bgl-Xho, was generated using convenient restriction sites, and it removes most of the sequences deleted in faswb. The other, Bgl-Xho [facet], is a BglII-XhoI fragment isolated from a faswb genomic clone. In both cases, the deletion of the faswb sequences reduced the enhancer-blocking activity of the BglII-XhoI restriction fragment by ~70%. Similarly, when the mutant fragment was tested in a position-effect assay, it was unable to insulate a white reporter construct against chromosomal influences. Instead of the wild-type eye phenotype observed in lines carrying a white maxigene flanked by the BglII-XhoI fragment, lines that carried the mutant fragment generally had reduced expression of white. This result is particularly compelling, since the reduction in the level of expression of white in these lines is similar to the Notch loss-of-function phenotype observed in the faswb mutant.
While the sequences deleted in faswb are clearly important for insulation in the white transgene assays, they are not on their own sufficient to completely reproduce the strong enhancer-blocking activity of the BglII-XhoI fragment. We found that a HindIII-BssHII DNA fragment that contains all but 4 bp of faswb sequences has only an intermediate level of activity (50% blocking). On the other hand, a PCR-generated DNA fragment encompassing the faswb region and including an additional 60 bp immediately 3' of the proximal faswb breakpoint showed 80% of the activity present in the BglII-XhoI fragment. The sequences around the proximal faswb breakpoint are characterized by a set of nuclease hypersensitive sites, a feature that has also been observed in other insulator elements. Such sites may represent targets for boundary-associated proteins implicated in the insulating activity (
Does the Notch promoter contribute to the insulating activity?
On the basis of our analysis of different DNA fragments from the 5' region of the Notch locus, it is clear that most of the insulating activity (~80%) is located within a region that includes the faswb sequences and the 60-bp region immediately downstream. Additional sequences may also contribute, to a lesser extent, to enhancer-blocking activity. Since the Notch promoter region is located just 3' of the faswb proximal breakpoint, the strong blocking activity of the BglII-XhoI fragment (or the HindIII-HindIII and HindIII-XhoI fragments; see Fig 8) could be due to the combined effects of an insulator (corresponding approximately to the faswb sequences) and promoter competition. By providing an alternative target for the white enhancer, the Notch promoter could cause a decrease in the level of expression of white in our enhancer-blocking assay, a result that could be interpreted as evidence for an insulator. Such a mechanism is not uncommon. For instance, at the human ß-globin gene locus, different promoters compete for regulation by a unique locus control region in a distance-dependent manner (
We were unable, however, to obtain evidence that the Notch promoter is capable of competing with the white promoter. Drosophila core promoter elements generally consist of one or more of three conserved elements: a TATA element located at around -30, an initiator element (Inr) located near the transcription start site, and a downstream element (Dpe) located at ~+30 (
When the BssHII-XhoI fragment containing the putative promoter elements was placed (in either orientation) between the white enhancer and the mini-white reporter gene, transgenic lines showed levels of expression of white identical to the control lines (Table 4), suggesting that the Notch promoter region did not significantly compete with the white promoter or interfere with enhancer-promoter communication. That the Notch promoter region has no detectable effects on white expression is also supported by the analysis of the deletion derivatives of the BglII-XhoI fragment (Bgl-Xho [facet] and Bgl-Xho). Both fragments are deleted for the faswb sequences, but retain the Notch promoter region as well as sequences in the BglII-HindIII fragment 5' of the deletion. In the enhancer-blocking assay, these fragments showed insulating activity identical to that of the upstream BglII-HindIII fragment alone (Fig 8 and Table 4). In agreement with these results, when the large BssHII-XhoI fragment, which includes the facet-strawberry sequences as well as the Notch promoter region, was placed in the 5'-3' orientation upstream of the white enhancer (i.e., with the promoter region immediately adjacent to the white enhancer), no reduction in the level of expression of white was detected, as we would expect if the Notch promoter were competing for the white enhancer (Fig 4 and Table 2). Instead, all lines showed a uniform, wild-type level of expression consistent with insulation. It is interesting to note that an activity required for promoter competition associated with the eve promoter required the TATA element (
Association between promoters and insulator elements:
Our results suggest that the insulator element defined by the facet-strawberry mutation is located immediately upstream of, or perhaps partially overlaps with, the Notch promoter. A close association between insulator elements and promoters such as the one observed here is not uncommon. For instance,