Loss of Insulator Activity by Paired Su(Hw) Chromatin Insulators

See allHide authors and affiliations

Science  19 Jan 2001:
Vol. 291, Issue 5503, pp. 495-498
DOI: 10.1126/science.291.5503.495


Chromatin insulators are regulatory elements that block the action of transcriptional enhancers when interposed between enhancer and promoter. The Drosophila Suppressor of Hairy wing [Su(Hw)] protein binds the Su(Hw) insulator and prevents enhancer-promoter interaction by a mechanism that is not understood. We show that when two copies of the Su(Hw) insulator element, instead of a single one, are inserted between enhancer and promoter, insulator activity is neutralized and the enhancer-promoter interaction may instead be facilitated. This paradoxical phenomenon could be explained by interactions between protein complexes bound at the insulators.

The Drosophila gypsy retrotransposon contains a chromatin insulator that consists of cluster of 12 binding sites for the Su(Hw) zinc-finger protein (1–6). In the presence of Su(Hw) protein binding, the insulator blocks the activity of an enhancer separated from the promoter by an Su(Hw) binding region. However, this insulator action fails in certain genetic rearrangements that introduce more than one gypsyretrotransposon in the region of the yellow gene (7). The loss of insulator activity might result from intrachromosomal pairing between the two gypsyretrotransposons, causing chromatin to fold and allowing the enhancer to contact the promoter. Alternatively, interaction between the proteins bound to two Su(Hw) insulator elements might neutralize insulator action. Here, we analyze insulator activity as affected by insulator element copy number and location.

The yellow gene is required for dark pigmentation ofDrosophila larval and adult cuticle and its derivatives. Two upstream enhancers, En-b and En-w, activate expression in the body cuticle and wing blades, respectively (8). When a single Su(Hw) insulator is inserted at position –893 relative to theyellow transcription start, between the enhancers and theyellow promoter (ESY, Fig. 1), enhancer action is blocked, resulting in yellow instead of dark pigmentation of body and wing cuticle. This block is relieved and pigmentation is restored when the construct is tested in asu(Hw) background, confirming that the Su(Hw) protein is responsible (9). In the ESFSY construct (Fig. 1), a fragment bearing two Su(Hw) insulators, separated by a 1.5-kilobase (kb) spacer fragment, was inserted at position –893. The spacer is derived from the second exon of the yellow gene and has no enhancer or insulator activity of its own. In seven ESFSY transgenic lines, yellow expression was higher than that in the control ESY lines, and in three lines it was at wild-type levels. When three of the less pigmented ESFSY were tested in asu(Hw) background, wild-type pigmentation was restored. Thus, the second Su(Hw) insulator partially or completely neutralizes the effect of the first one. Similar results were obtained when the distance between the two Su(Hw) insulators was reduced to 200 base pairs (bp).

Figure 1

Transposon constructs used to test insulator action. The maps of the constructs (20), not drawn to scale, indicate the yellow wing and body enhancers (En-w and En-b, respectively) as partially overlapping white boxes. The Su(Hw) insulator is shown as a black box and the yellow andwhite genes as white boxes with an arrow indicating the direction of transcription. F denotes a spacer fragment. The column to the right summarizes the results, with + indicating that theyellow gene was activated by its enhancers in the majority of the lines.

The body and wing enhancers in these constructs are responsible for wild-type dark pigmentation because, when they were removed from ESFSY, yielding SFSY (Fig. 1), body and wing pigmentation was yellow in all transgenic lines. Similarly, constructs containing theyellow gene alone never result in body or wing pigmentation (7). Increasing the distance between upstream enhancers and the yellow promoter does not weaken insulator activity because lines containing ESFY, bearing a single Su(Hw) insulator 2.4 kb from the yellow transcription start, all had yellow body and wing pigmentation, indicative of the block of wing and body enhancers.

If the loss of insulator activity is due to a steric constraint imposed by a physical interaction between the two insulators, flanking either the enhancers or the target gene with insulators might have the same effect. This was tested with the SFESY construct in which two Su(Hw) insulators frame the wing and body enhancers. Flies from nine SFESY transgenic lines exhibited yellow wing and body pigmentation. When two of these lines were crossed into a su(Hw) background, wild-type levels of pigmentation were restored, confirming that the proximal Su(Hw) element retained insulator activity.

In the ES(–893)YS construct, the yellow gene is flanked by Su(Hw) insulators, one at position –893 and the other downstream of the yellow gene (Fig. 1). In nine ES(–893)YS transgenic lines, yellow expression in the body and wings was blocked. When two of these lines were crossed into asu(Hw) background, wild-type pigmentation of wings and body was restored. Thus, the second insulator, downstream ofyellow, does not prevent the insulating function of the first.

Next, we tested a different enhancer-promoter combination. Thewhite gene is required for eye pigmentation and is regulated by its eye-specific enhancer. Roseman et al. (5) found that interposing the Su(Hw) insulator between the eye enhancer and white promoter completely blocked enhancer activity, whereas bracketing the mini-white gene between two Su(Hw) insulators protected white expression from position effects. In the EyeSYW construct, the eye enhancer was inserted between theyellow wing and body enhancers and was flanked by Flp recognition target (FRT) sites to permit its excision from transgenic flies (10). The three enhancers are separated from theyellow and white genes by a Su(Hw) insulator. Flies of 20 EyeSYW lines displayed eye pigmentation levels like those produced by an enhancerless white transposon, that is, ranging from pale yellow to red, depending on the insertion site. In two red-eyed EyeSYW lines, the deletion of the eye enhancer by Flp-dependent excision did not influence eye color, implying that in these two lines the white gene was activated by some genomic enhancer element. Thus, one Su(Hw) insulator interposed between eye enhancer and white gene blocks enhancer-promoter communication. The body and wing enhancers of the yellowgene were also blocked in these lines, indicating that the insulator functioned normally. Similarly, if the insulator was placed in front of the white gene, to give EyeYSW, the 14 transgenic lines obtained had eye colors in the range expected in the absence of eye enhancer. Deletion of the eye enhancer in five dark orange–eyed lines did not change eye pigmentation. Thus, the eye enhancer is blocked by one copy of the Su(Hw) element inserted either near or far from thewhite promoter.

The EyeSYSW construct uses the same enhancer configuration described above and contains one Su(Hw) insulator at position –893 relative to the yellow transcription start and another inserted between the yellow gene and the mini-white promoter (Fig. 2). Therefore, just one insulator intervenes between the enhancers and yellow but two insulators between the enhancers and white. In 19 of 21 transgenic EyeSYSW lines, wing and body pigmentation were yellow, indicating that theyellow enhancers were blocked, whereas whiteexpression was stronger than in lines bearing the mini-whitegene without eye enhancer. To demonstrate that the eye enhancer stimulates white expression in these lines, we excised it by Flp-induced recombination between FRT sites. In nine ΔEyeSYSW lines tested, the deletion of the eye enhancer strongly diminished eye pigmentation, indicating that the enhancer can activate thewhite gene despite the two intervening insulators. Therefore, also in this case, two insulators between enhancer and promoter neutralize one another. However, interaction between the two insulators does not simply inactivate them, because the upstream insulator can still block the activation of the yellow gene.

Figure 2

Transposon constructs to test whiteenhancer action. The white box (Eye) indicates the eye enhancer of thewhite gene, and the thick arrows marked FRT represent the target sites of the Flp recombinase. The other symbols are the same as in Fig. 1. The two columns on the right summarize the results, with + indicating that the yellow or whitegenes were activated by their respective enhancers in the majority of the lines.

In the same EyeSYSW lines, white expression was studied in asu(Hw) background. In five lines, the absence of Su(Hw) protein reduced white expression, implying that the Su(Hw) protein actually has a positive role, facilitating enhancer-promoter interactions. In four other lines, the absence of Su(Hw) protein had no effect. Thus, the stimulating effect of the two Su(Hw) insulators may depend on genomic context and/or local chromatin structure. To show that the Su(Hw) protein does not by itself activate white expression, we crossed five lines bearing the EyeSYSW transposon with deleted eye enhancer (ΔEyeSYSW) into asu(Hw) background. The absence of Su(Hw) protein did not influence white expression.

To determine what configuration of two insulators neutralizes their insulator activity, we constructed EyeSYWS, in which the two Su(Hw) insulators frame the yellow and white genes. Fourteen EyeSYWS lines displayed weak expression of bothwhite and yellow, indicating that all three enhancers upstream of the interposed Su(Hw) insulator were blocked. However, when the mini-white gene flanked by two Su(Hw) insulators was inserted at position –893 relative to theyellow transcription start site (EyeSWFSY, Fig. 2), theyellow gene was expressed in the body and wings. In asu(Hw) background, yellowexpression decreased in three lines and did not change in one line, showing that the activation of the yellow promoter by distant yellow enhancers is improved by an interposed insulator pair. Thus, for both white and yellow, the insertion of two Su(Hw) insulators between the respective enhancers and promoters may facilitate their interaction instead of blocking it. When the Su(Hw) insulator between white andyellow genes was removed, yielding EyeSFWY (Fig. 2),yellow expression in the body and wings was suppressed, showing again that a single insulator blocks the wing and body enhancers. Two copies of Su(Hw) do not simply neutralize one another by an exclusive binary interaction. In the EyeSFSYSW construct, three insulator copies intervene between eye enhancer and whitegene and two copies are between the yellow enhancers and theyellow gene. In 12 of 16 lines carrying this transposon, both yellow and white are activated, producing flies with strongly pigmented eyes and wing and body cuticle.

In summary then, when two or more Su(Hw) insulators are introduced between enhancer and promoter, their enhancer-blocking effect is neutralized in most cases and enhancer-promoter communication is often improved. Entirely similar results, using different promoter and enhancer combinations have been obtained by Cai et al.(11). The implication is that two insulators interact, probably through the protein complexes bound to them. This interaction by itself does not neutralize the blocking action, because when the insulators frame the enhancers or the target gene, the block still occurs. A possible explanation is that the “looping out” of the sequences separating enhancer and promoter displaces the insulators out of the way and, by bringing the enhancer and promoter closer, may even stimulate expression (Fig. 3). This may explain why the stimulating effect increases with the distance between enhancers and promoter.

Figure 3

Model of the double insulator bypass. (A) A single insulator blocks enhancer-promoter interaction. (B) Two insulators may interact with one another through the protein complexes bound to them, forming a loop and bringing the enhancers closer to the promoter.

These effects may have a bearing on the mechanism of insulator action. A possible way to envision how the insulator interferes with the access of the enhancer to the promoter is by associating with the nearest Su(Hw)-related complexes in the nucleus (12, 13). The effect of this association would be to tether loops containing members of an enhancer-promoter pair, thereby interfering with the interaction of the enhancer on one loop with the promoter on another loop. When two Su(Hw) elements are placed between enhancer and promoter, the loop would form preferentially between the two neighboring Su(Hw) elements, thereby shortening the distance between enhancer and promoter rather than inhibiting their interaction. This type of mechanism may also help to explain the role of boundary elements in the Drosophila bithorax complex (14). In the Abd-B regulatory region, boundary elements like Fab-7 and Fab-8 flank theiab enhancer regions, insulating them from the silencing or activating effects of adjacent regulatory regions (15–17). However, as insulators, the boundary elements would also block activation of the Abd-Bpromoter by more distant iab enhancers, thus defeating the purpose of these enhancers. Although other explanations are possible, our results with insulator pairs may account for this discrepancy. Interaction between boundary elements flanking each enhancer may effectively protect the iab enhancers from outside repressing effects and facilitate, instead of blocking, enhancer-promoter communication. It is possible, in fact, that one role of certain kinds of insulator is to promote the interaction between distant enhancers and promoters.

  • * To whom correspondence should be addressed. E-mail: pirrotta{at}


View Abstract

Navigate This Article