Multiple interactions are involved in a highly specific association of the Mod(mdg4)-67.2 isoform with the Su(Hw) sites in Drosophila

The best-studied Drosophila insulator complex consists of two BTB-containing proteins, the Mod(mdg4)-67.2 isoform and CP190, which are recruited to the chromatin through interactions with the DNA-binding Su(Hw) protein. It was shown previously that Mod(mdg4)-67.2 is critical for the enhancer-blocking activity of the Su(Hw) insulators and it differs from more than 30 other Mod(mdg4) isoforms by the C-terminal domain required for a specific interaction with Su(Hw) only. The mechanism of the highly specific association between Mod(mdg4)-67.2 and Su(Hw) is not well understood. Therefore, we have performed a detailed analysis of domains involved in the interaction of Mod(mdg4)-67.2 with Su(Hw) and CP190. We found that the N-terminal region of Su(Hw) interacts with the glutamine-rich domain common to all the Mod(mdg4) isoforms. The unique C-terminal part of Mod(mdg4)-67.2 contains the Su(Hw)-interacting domain and the FLYWCH domain that facilitates a specific association between Mod(mdg4)-67.2 and the CP190/Su(Hw) complex. Finally, interaction between the BTB domain of Mod(mdg4)-67.2 and the M domain of CP190 has been demonstrated. By using transgenic lines expressing different protein variants, we have shown that all the newly identified interactions are to a greater or lesser extent redundant, which increases the reliability in the formation of the protein complexes.

The BTB (bric-a-brac, tramtrack and broad complex)/POZ (poxvirus and zinc finger) domain is a conserved protein-protein interaction motif contained in a variety of transcription factors involved in development, chromatin remodelling, insulator activity and carcinogenesis [12,13]. All the well-studied mammalian BTB domains form obligate homodimers and, rarely, tetramers [13]. The BTB domain of Mod(mdg4) belongs to the 'ttk group' that contains several highly conserved sequences not found in other BTB domains [14,15]. The BTB domains of the ttk group can multimerize [14], which was suggested to be essential for the ability of the Mod(mdg4) isoforms to support pairing between the distantly located sites in the chromosomes [16].
& 2017 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution Mutational dissection and differential binding of the Mod(mdg4) isoforms on polytene chromosomes suggest that the variable C-terminal regions encoded by the alternative 3 0 exons determine their functional specificity [1,6,17]. The variable C-terminal regions interact specifically with different proteins [3,18,19]. So far, the functional roles of only two Mod(mdg4) isoforms have been studied in detail. The Mod(mdg4)-56.3/MNM (Modifier of Mdg4 in Meiosis) isoform is required for the homologue conjunction during meiosis [6,20], while the best-studied Mod(mdg4)-67.2 isoform is important for the enhancer-blocking activity of the Su(Hw) insulators [3,21,22]. Twelve repeated binding sites for the Su(Hw) constitute the best-studied Drosophila insulator, which was found at the 5 0 regulatory region of the gypsy retrotransposon [23][24][25]. Insulators in the Drosophila and vertebrate genomes have been identified based on their ability to disrupt the communication between an enhancer and a promoter when inserted between them [26][27][28][29][30][31][32][33][34]. The Drosophila Suppressor of Hairy-wing (Su(Hw)) protein is a classical insulator protein that contains an array of 12 zinc fingers of the C2H2 and C3H types [23,35]. The C2H2 domains, from 5 to 10, specifically recognize an approximately 18 bp site [36]. Later, several other insulator proteins (dCTCF, Zw5, ZIPIC and Pita) with clusters of zinc finger domains have been identified [37][38][39][40][41][42][43].
The best described insulator found at the 5 0 regulatory region of the gypsy retrotransposon has a unique structure because it consists of twelve repeating binding sites for Su(Hw) [23][24][25]. All other genomic regions contain only one or rarely two or three bindings sites for Su(Hw) [36,44]. At the same time, in the transgenic lines only four synthetic Su(Hw)-binding sites can function as an effective insulator [45], but the genomic regulatory elements containing one or two Su(Hw) sites also display strong enhancer-blocking activity [44,[46][47][48]. This discrepancy might be explained by the existence of additional unknown insulator proteins that function in a cooperation with the Su(Hw). Thus, the gypsy insulator is an exceptional example of insulators consisting of the reiterated binding sites for only one protein.
In addition to Mod(mdg4)-67.2, the CP190 protein interacts directly with Su(Hw) and both are required for the activity of the Su(Hw)-dependent insulators [49,50]. In the genomewide studies [44,51,52], three classes of the Su(Hw)-binding regions have been identified, which are characterized by the binding of the Su(Hw) alone (SBS-O), of both Su(Hw) and CP190 (SBS-C), and all the three proteins (SBS-CM) [44,[53][54][55]. The stand-alone Su(Hw) sites (SBS-O) usually repress transcription [44], while SBS-CM sites display enhancer-blocking activity. In contrast to Mod(mdg4)-67.2, CP190 interacts beside the Su(Hw) with many other known insulator ZF proteins including Pita, dCTCF and ZIPIC [38,40,41,56]. The CP190 protein contains an N-terminal classical BTB/ POZ domain that forms a homodimer that is involved in the interaction with dCTCF and Pita proteins [38,40]. CP190 preferentially binds near the transcription start sites of genes, suggesting a role of this protein in the organization of promoter architecture [44,57,58]. It was shown that CP190 participates in recruiting of the NURF, dREAM and SAGA complexes to chromatin [59][60][61][62], which are critical for the activity of promoters. Transcriptional complexes recruited to chromatin by the Mod(mdg4) isoforms have not been identified yet, but Mod(mdg4)-67.2 is essential for the enhancer-blocking activity of Su(Hw) [11,21,63]. For example, Mod(mdg4)-67.2 blocks the eye-specific enhancer by a direct interaction with Zeste that supports the enhancer-promoter communication of the white gene [22,64].
Here, we have studied how Mod(mdg4)-67.2 is specifically targeted to the Su(Hw)/CP190 complex. While CP190 also interacts with many other DNA-binding proteins, Mod(mdg)-67.2 interacts only with the Su(Hw). Previously, it was suggested that such specificity is dictated by an interaction between the unique part of the Mod(mdg4)-67.2 isoform and the C-terminal region of Su(Hw), between aa 716 and 892, named the Mod(mdg4)-67.2-interacting domain, MID [63,65]. Unexpectedly, we found that the Su(Hw) e7 mutant lacking the MID was still able to recruit Mod(mdg4)-67.2 to the Su(Hw) sites. For this reason, we re-examined the interactions between the insulator proteins and found new domains in these proteins that are essential for the specific recruiting of the Mod(mdg4)-67.2 to the Su(Hw) sites.

Material and methods
The constructs for the yeast two-hybrid assay, GST pull-down assay and transgenic constructs, and details of experimental and analytical procedures, are described in the electronic supplementary material.

Drosophila strains, germ line transformation and genetic crosses
The construct together with P25.7wc, a P element with defective inverted repeats used as a transposase source, was injected into y ac w 1118 preblastoderm embryos as described [66]. All flies were maintained at 258C on the standard yeast medium. The resulting flies were crossed with y ac w 1118 flies, and the transgenic progeny were identified by their eye colour. Chromosome localization of various transgene insertions was determined by crossing the transformants with the y ac w 1118 balancer stock carrying dominant markers, In(2RL),CyO for chromosome 2 and In(3LR)TM3,Sb for chromosome 3. The generation of transgenic lines and construct introduction into the mod(mdg4) u1 or Su(Hw) v /Su(Hw) e04061 background were performed as described [21]. To express transgenes regulated by the UAS promoter, flies homozygous for the construct were crossed with the y1 w*; PfAct5C-GAL4g25FO1/ CyO, yþ driver strain (Bloomington Center #4414). The effects of Mod(mdg4) variants produced from homozygous expression vectors and various mutation combinations were scored by two researchers independently. The level of expression of yellow and cut phenotypes was evaluated in 3-to 5-day-old males developing at 258C. For yellow phenotypes, wild-type expression in the abdominal cuticle, wings and bristles was assigned an arbitrary score of 5, while the absence of yellow expression was scored 1, using as reference the flies in which the y allele was characterized previously. Representative wing forms shown in the figures were selected as 'average' from the series of wings arranged in order of increasing severity of their mutant phenotype. At least 50 flies from each y line were scored.

Two-hybrid and in vitro interactions
Two-hybrid assays were carried out with yeast strain pJ694A using plasmids and protocols from Clontech (Palo Alto, CA). For growth assays, plasmids were transformed into yeast rsob.royalsocietypublishing.org Open Biol. 7: 170150 pJ694A cells by the lithium acetate method, as described by the manufacturer, and plated on media without tryptophan and leucine. After 3 days of growth at 308C, the cells were plated on selective media without tryptophan, leucine, histidine and adenine, and their growth was compared after 2-3 days.

RNA interference (RNAi) treatment and analysis of S2 cells in culture
CP190 cDNA templates were amplified by PCR using the primer pairs 5 0 -ATGGGTGAAGTCAAGTCCGTGAAAG-3 0 and 5 0 -GAATTCCTTAACCTCTTCCAAAC-3 0 , with the 5 0 end of each primer containing the T7 RNA polymerase promoter site. PCR products were purified using the Gel Extraction Kit (Zymo Research) as recommended by the manufacturer. Purified PCR products were used to produce double-stranded RNA (dsRNA) using a Megascript T7 transcription kit (Ambion). The RNA was purified according to the manufacturer's protocol, heated at 658C for 30 min and left to cool at room temperature. Its samples were then resolved in agarose gel to test for the quality of dsRNA. Drosophila embryonic S2 cells were grown in Schneider's insect medium (Sigma) supplemented with 10% fetal calf serum (FCS, HyClone) at 278C. The RNAi treatment and subsequent viable cell count analysis of S2 culture were basically performed as described [67]. To express the pAc5.1Su(Hw) 1-238 -FLAG construct, the S2 cells were transformed using the Effectene Transfection Reagent (Qiagen) as recommended by the manufacturer. Nuclear extracts were prepared and immunoprecipitation experiments were performed as described previously [68].

Protein extract preparation from males and co-IP analysis
The material (about 150-200 mg of adult males, sufficient for four or five independent immunoprecipitations) was homogenized in 5 ml of buffer IP-Sþ (10 mM Tris-HCl ( pH 7.5), 10 mM NaCl; 10 mM MgCl 2 ; 1 mM EDTA; 1 mM EGTA; 1 mM DTT; 250 mM sucrose and PMSF, leupeten, pepstatinA) at þ4C using a Douncer with a type A pestle. Then the homogenate was transferred through the BD Falcon filter to a 50 ml tube and spun down on a centrifuge for 5 min, at 4000g at 48C. The supernatant was discarded. The pellet was resuspended in 3 ml of buffer IP-Sþ and then spun down in the same way. This washing step was repeated 3 times. To the pellet, 0, 5 ml IP-10 bufferþ (10 mM Tris-HCl ( pH 7.5), 10 mM NaCl, 10 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 0,1% NP-40, 10% glycerol and Roche Complete Protease Inhibitor Cocktail) was added and the pellet was homogenized at þ4C using a Douncer with a type B pestle. Equal volume of IP-850 buffer (10 mM Tris-HCl (pH 7.5), 850 mM NaCl, 10 mM

Chromatin preparation and ChIP analysis
Chromatin was prepared from the middle pupa stage as described previously [70]. The resulting chromatin preparation was used for ChIP experiments as described previously [51]. At least three independent biological replicates were made. Primer sequences used in PCR for ChIP analysis are shown in electronic supplementary material, table S1.

Role of the C-terminal domain in Su(Hw) interaction with the Mod(mdg4)-67.2 in vivo
The Su(Hw) e7 mutation was previously characterized and is generated by a C ! T transition at base 3069 that leads to production of a truncated protein lacking the last 223 amino acids that contain the MID region required for the interaction with Mod(mdg4)-67.2 [63,65,71]. The level of Su(Hw) e7 expression is comparable with that of wild-type protein (electronic supplementary material, figure S1a). As shown previously, Mod(mdg4)-67.2 protein completely co-localizes with Su(Hw) on polytene chromosomes [1,3]. rsob.royalsocietypublishing.org Open Biol. 7: 170150 The antibodies raised against the unique C-terminal domain of Mod(mdg4)-67.2 recognized about 200 sites on polytene chromosomes, in particular the sites corresponding to gypsy insertion in the y 2 mutation and the endogenous 1A2 insulator [3,46,48] at the tip of the X chromosome (figure 1a). In the su(Hw)background (su(Hw) v /su(Hw) e04061 ), Su(Hw) proved to still strongly bind to several sites, which could be explained by a weak residual expression of the su(Hw) e04061 allele generated by an insertion of the PiggyBac element near the start codon [72]. In su(Hw)flies, almost no binding of Mod(mdg4)-67.2 to polytene chromosomes was observed, confirming the critical role of Su(Hw) in Mod(mdg4)-67.2 recruitment. Residual staining of Mod(mdg4)-67.2 at a few sites could be explained by a residual binding of the Su(Hw) to the same sites. Unfortunately, we were unable to directly test this point due to inability to independently examine the Su(Hw) and Mod(mdg4)-67.2 binding to the polytene chromosomes. The binding of CP190 was reduced only at a small number of sites, providing additional evidence that many different proteins recruit CP190 to the chromatin.
In su(Hw) v /su(Hw) e7 larvae, the pattern of Su(Hw) binding to polytene chromosomes was the same as in wild-type larvae (figure 1a). Unexpectedly, we also found that a considerable number of Mod(mdg4)-67.2-positive sites coincided with sites for Su(Hw) and CP190. In particular, Mod(mdg4)-67.2 binds to the y 2 site at the tip of the X chromosome. Thus, the deletion of the Su(Hw) C-terminal domain only partially affects the recruitment of Mod(mdg4)-67.2 to the Su(Hw) sites.
To confirm the above results, we used ChIP to study the binding of insulator proteins with chromatin isolated from pupae expressing wild-type Su(Hw)þ (su(Hw) v /TM6,Tb), null for Su(Hw) (su(Hw) v /su(Hw) e04061 ) and Su(Hw) e7 (su(Hw) v / su(Hw) e7 ) (figure 1b; electronic supplementary material, figure S1a). To this end, we used the gypsy and four endogenous insulators that are bound by Su(Hw) in complex with C190 and Mod(mdg4)-67.2 (electronic supplementary material, figure  S2) [47] and antibodies against the N-terminal domain of Su(Hw). In su(Hw)pupae, we still observed residual Su(Hw) binding to strong insulator sites (50A, 62D), which was correlated with Mod(mdg4)-67.2 binding to the same sites. At the same time, the binding of Mod(mdg4)-67.2 and Su(Hw) to the gypsy and 1A2 insulators was almost completely absent. In su(Hw) v /su(Hw) e7 pupae, ChIP analysis revealed almost normal Su(Hw) binding to 62D, 50A and 87E, while its binding     [25,71]. Previously, we found that the C-terminal acidic domain of Su(Hw) partially represses transcription in yeast [11,73], complicating interpretation of the results obtained using this system. For this reason, in most of the experiments we used a truncated version of the Su(Hw) protein lacking the C-terminal domain from  rsob.royalsocietypublishing.org Open Biol. 7: 170150 the 892 to 945 aa region. As shown by using yeast two-hybrid [65] and GST pull-down [63] assays, Mod(mdg4)-67.2 interacts with the C-terminal region of Su(Hw) between aa 716 and 892, including LZ. The Mod(mdg4)-67.2 protein interacts with Su(Hw) through the unique C-terminal domain (aa 453-610) that includes a FLYWCH-type zinc finger domain (between aa 453 and 514) (figure 2a).
The results confirmed previous data [63,65] that Mod(mdg4)-67. As at most sites, CP190 and Mod(mdg4)-67.2 bind to Su(Hw) together [44], and CP190 also seems to contribute to specific recruiting of Mod(mdg4)-67.2. Indeed, the results of previous studies suggested that Mod(mdg4)-67.2 and CP190 interact with each other [11,50]. It was shown that the BTB domain of Mod(mdg4)-67.2 is required for interaction with CP190 [11]. Therefore, we tested different domains of these proteins in the yeast two-hybrid system in order to reveal and map the domains involved in their interaction (figure 2e). The CP190 protein contains several domains (figure 2a), including the BTB/POZ domain, aspartic acid-rich (D-rich) domain, four C2H2 zinc fingers and C-terminal glutamic acid-rich (E-rich) domain [57,74]. In addition to these motifs, CP190 also contains a centrosomal targeting domain (M) responsible for its localization to centrosomes during mitosis [75]. As a result, we found that the BTB domain of Mod(mdg4)-67. The binding of the insulator proteins in pupae was analysed using ChIP analysis. In addition to the five Su(Hw)/ Mod(mdg4)-67.2/CP190 sites, we tested two stand-alone Su(Hw) sites, two stand-alone CP190 sites and one site in which dCTCF is co-localized with CP190 (figure 3b; electronic supplementary material, figure S2). As it was impossible to detect Su(Hw)DN transgenes with the antibodies to the Su(Hw) N-terminal domain, we used antibodies raised against its C-terminal domains, along with the anti-FLAG antibodies. The binding of Su(Hw)DN to the Su(Hw)/Mod(mdg4)-67.2/ CP190 sites was strongly reduced, comparatively to Su(Hw)þ (figure 3b). Interestingly, we did not observe such difference in the binding between the Su(Hw) variants in the case of the control stand-alone Su(Hw) sites (figure 3b). Thus, the N-terminal domain is essential for preferential recruitment of Su(Hw) only to the CP190/Mod(mdg4)-67.2 sites.

Experiments with a genetic model system confirm the role of multiple interactions between Mod(mdg4)-67.2 and Su(Hw) proteins
To determine the outcomes of mutations, we used gypsyinduced alleles in the yellow and cut loci. In the y 2 mutation (figure 4a), gypsy is inserted between the enhancers controlling yellow expression in the wings and body cuticle and the yellow promoter [24]. As a result, the Su(Hw) insulator blocks the wing and body enhancers, but not the bristle enhancer that is located in the yellow intron [24,76]. We also used four transgenic lines carrying a gypsy insertion between the yellow enhancers and the promoter, all of which displayed a y 2 -like phenotype (electronic supplementary material, figure S4a).  The mod(mdg4) T6 mutation results in the expression of mutant protein that lacks 43 C-terminal residues corresponding to the SID domain alone [65].
In the ct 6 allele (figure 4a), gypsy is between the wing margin enhancer and the cut promoter, which are 85 kb apart [63]. The insulator in ct 6 completely blocks this enhancer, producing a cut wing phenotype. The mod(mdg4) u1 and mod(mdg4) T6 mutations affect the activity of the gypsy insulator inserted in the y 2 and ct 6 alleles (figure 5a). The mod(mdg4) u1 and mod(mdg4) T6 mutations almost completely suppress ct 6 phenotype, suggesting that Mod(mdg4)-67.2 is essential for the enhancer-blocking activity of the gypsy insulator in the case of the ct 6 allele. At the same time, the mod(mdg4) mutations enhance the mutant y 2 phenotype by repressing yellow expression in bristles and inducing a variegated pigmentation in the abdominal segments. Thus, binding of the Mod(mdg4)-67.2 protein prevents direct repression of the yellow promoter by the gypsy insulator in the y 2 allele.
We performed immunolocalization of these mutant proteins on polytene chromosomes (figure 6a) and analysed them by ChIP with chromatin from mutant pupae (figure 6b). The Mod(mdg4) T6 protein was detected with antibodies raised against the unique C-terminal domain (electronic supplementary material, figure S5), and the Mod(mdg4) u1 protein, with antibodies against the region common to all Mod(mdg4) isoforms.
ChIP analysis of mutant pupae showed that the Mod(mdg4) u1 protein did not bind to the selected Su(Hw) binding regions (figure 6b). In contrast to Mod(mdg4) u1 , ChIP analysis showed that Mod(mdg4) T6 weakly binds to some Su(Hw) sites but not to the gypsy insulator in the y 2 allele. Faint bands of the Mod(mdg4) T6 protein were detected at relatively many sites on polytene chromosomes but not at the tip of the X chromosome corresponding to the y 2 allele (figure 6a). Thus, Mod(mdg4) T6 can weakly bind to the   In our previous study [11], we made a double mutant Mod(mdg4)-67.2 protein, designated ModD33N/H46D, by substituting the most conserved aspartate (33) and histidine (46) in its BTB domain by asparagine and acidic aspartate, respectively. This mutant protein only weakly interacted with CP190 but still bound to the Su(Hw) sites and displayed normal functional activity. As the deletion of the Q domain only slightly affected the binding of the Mod(mdg4)DQ protein, we made a transgenic line expressing double mutant Mod(mdg4)DQ D33N/H46D under control of the UAS promoter (figures 5b and 7a-c; electronic supplementary material, figure S5). The expression of ModDQ D33N/H46D did not complement the mutant mod(mdg4) u1 phenotype (figure 5b). We also observed no binding of ModDQ D33N/H46D to the Su(Hw) sites in pupae analysed by ChIP (figure 7b,c) or to polytene chromosomes ( figure 8). Thus, the combination of two mutations in ModDQ D33N/H46D resulted in the loss of the ability to bind to the Su(Hw) sites.
The binding of Su(Hw) and CP190 in ModDQ D33N/H46D pupae to the Su(Hw)/Mod(mdg4)-67.2/CP190 sites was reduced to the same extent as in the mod(mdg4) u1 background (electronic supplementary material, figure S9). These results confirm that Mod(mdg4)-67.2 facilitates the recruitment of Su(Hw) and CP190 to chromatin.

Discussion
Our results suggest that multiple interactions are required for the formation of the Mod(mdg4)-67.2/CP190/Su(Hw)   Mod-67.2þ is the wild-type protein, ModDFLYWCH lacks the FLYWCH domain, ModDQ lacks the Q-rich domain, and ModDQD33N/H46D is a double mutant that lacks the Q-rich domain and has the most conserved aspartate (33) and histidine (46) in the BTB domain substituted by asparagine and acidic aspartate, respectively (indicated with asterisks). Other designations are as in figure 2a. (b,c) ChIP-qPCR analysis of Mod-67.2, and Mod-com binding in middle pupae of the above transgenic lines. PCR products were amplified from two separate immunoprecipitates of three different chromatin preparations. Error bars indicate the standard deviation of three independent biological replicates. *p 0.05 (Student's t-test), in other cases, p 0.01. The ras64B coding region (Ras) was used as a control devoid of Su(Hw)-binding sites. Other designations are as in figure 1b. Analysis of transgenic lines was performed in the y 2 sc D1 ct 6 ; PfAct5C-GAL4g25FO1/þ; mod(mdg4) u1 /mod(mdg4) u1 background. rsob.royalsocietypublishing.org Open Biol. 7: 170150 complex. It has been shown previously that the unique 567-610 region of the Mod(mdg4)-67.2 isoform interacts with the 693-880 region of Su(Hw), which is necessary for the enhancer-blocking activity [63,65]. However, deletion of the 224 C-terminal residues in Su(Hw) e7 only partially affects the Mod(mdg4)-67.2 recruitment, indicating that other domains may be involved in the interaction of these proteins. Interestingly, the Mod(mdg4) T6 protein lacking the 567-610 region required for interaction with Su(Hw) only weakly binds to the Su(Hw) sites. This suggests that the 567-610 region of Mod(mdg4)-67.2 may bind to an additional domain of Su(Hw). However, we failed to identify such a region in Su(Hw) or CP190. Alternatively, it is also possible that the 567-610 region of Mod(mdg4)-67.2 interacts with an unknown protein that also specifically associates with the Su(Hw). A further study is required to elucidate this question.
Here, we have found that the BTB and Q-rich domains of the Mod(mdg4)-67.2 (common to all its isoforms) interact with the M domain of CP190 and the N-terminal region of Su(Hw), respectively (figure 9a). As shown previously, the retention of the original Mod(mdg4) BTB domain in the Mod(mdg4)-67.2 isoform is critical for the specific recruitment of this protein to the Su(Hw)/CP190 sites [11]. For example, a chimeric Mod(mdg4)Gaf protein containing the GAF BTB domain can interact with Su(Hw) in vitro but completely loses its ability to associate with the Su(Hw)-binding regions [11]. Partially inactive BTB D33N/H46D still shows a weak interaction with CP190, and ModD33N/H46D binds to the chromatin, similar to the wild-type protein [11]. However, here we have found that the double mutant carrying also the deletion of the Q domain fails to bind to the Su(Hw) sites. Thus, the Q-rich domain has a partially redundant role in recruiting Mod(mdg4)-67.2 to the chromatin. According to the genome-wide studies, all the Mod(mdg4)-67.2/Su(Hw) sites contain also the CP190 [44], suggesting that CP190 is important for the recruitment of Mod(mdg4)-67.2 to the Su(Hw) sites.
Our results also suggest a role for the FLYWCH domain in the specific Mod(mdg4)-67.2 recruitment to the Su(Hw)/ CP190 sites. However, the mechanism of such an activity of the FLYWCH domain is still unknown. The results of the yeast two-hybrid assay show only that this domain improves the interaction between the BTB domain of Mod(mdg4)-67.2 and the M domain of CP190. Further analysis is required to resolve the mechanistic role of the FLYWCH domain in the functionality of Mod(mdg4)-67.2, taking into account that different variants of FLYWCH are present at the specific C-termini of the majority of the Mod(mdg4) isoforms [1].
According to the current model [70,78], the insulator bodies help to form protein complexes that subsequently bind to the regulatory elements such as insulators and promoters. It could be possible that the Su(Hw)/CP190/Mod(mdg4)-67.2 complexes are performed in the insulator bodies, and after this are recruited to the chromatin. CP190 and Mod(mdg4)-67.2 might determine the recruitment of the insulator complexes to the specific sites, due to the assembly of the multiple protein -protein interactions. In accordance with this model, we found that the interaction of the Su(Hw) with CP190 and Mod(mdg4)-67.2 is essential for the recruitment of the insulator complex to SBS-CM (the Su(Hw)/ CP190/Mod(mdg4)-67.2 sites) (figure 9b).
Taken together, it seems likely that the recruitment of the Mod(mdg4)-67.2, CP190 and Su(Hw) proteins to SBS-CM is mutually dependent. The specificity of the Mod(mgd)4-67.2 recruitment is achieved through complex interactions of The question remains unresolved as to why the other Mod(mdg4) isoforms do not bind to the Su(Hw) complex even though their common BTB and Q domains interact with the CP190 and Su(Hw) proteins, respectively. It seems likely that each Mod(mdg4) isoform specifically interacts with one or several DNA-binding transcription factors, as does Mod(mdg4)-67.2 with Su(Hw). If so, all the Mod(mdg4) isoforms prefer to interact with their specific protein complexes but not with the Su(Hw)-CP190 complex.
In summary, our results provide evidence for the high complexity of interactions between insulator proteins that are required to form the specific Su(Hw) insulator complex. Deletion of a single domain involved in the protein-protein interactions in either the Su(Hw) or the Mod(mdg4)-67.2 only partially disturbs its formation, indicating that the stability of the complex is ensured by the multiplicity/redundancy of such interactions. rsob.royalsocietypublishing.org Open Biol. 7: 170150