Mechanism and functional role of the interaction between CP190 and the architectural protein Pita in Drosophila melanogaster

Background Pita is required for Drosophila development and binds specifically to a long motif in active promoters and insulators. Pita belongs to the Drosophila family of zinc-finger architectural proteins, which also includes Su(Hw) and the conserved among higher eukaryotes CTCF. The architectural proteins maintain the active state of regulatory elements and the long-distance interactions between them. In particular, Pita is involved in the formation of several boundaries between regulatory domains that controlled the expression of three hox genes in the Bithorax complex (BX-C). The CP190 protein is recruited to chromatin through interaction with the architectural proteins. Results Using in vitro pull-down analysis, we precisely mapped two unstructured regions of Pita that interact with the BTB domain of CP190. Then we constructed transgenic lines expressing the Pita protein of the wild-type and mutant variants lacking CP190-interacting regions. We have demonstrated that CP190-interacting region of the Pita can maintain nucleosome-free open chromatin and is critical for Pita-mediated enhancer blocking activity in BX-C. At the same time, interaction with CP190 is not required for the in vivo function of the mutant Pita protein, which binds to the same regions of the genome as the wild-type protein. Unexpectedly, we found that CP190 was still associated with the most of genome regions bound by the mutant Pita protein, which suggested that other architectural proteins were continuing to recruit CP190 to these regions. Conclusions The results directly demonstrate role of CP190 in insulation and support a model in which the regulatory elements are composed of combinations of binding sites that interact with several architectural proteins with similar functions. Supplementary Information The online version contains supplementary material available at 10.1186/s13072-021-00391-x.


59
The development of modern approaches for the study of genome architecture, including 60 chromosome conformation capture methods, coupled to high-throughput sequencing (Hi-C) and 61 high-resolution microscopy techniques has revealed the hierarchical organization of genome (1,62 2). Chromosomes are composed of discrete sub-megabase domains, called topologically 63 associated domains (TADs) (3)(4)(5). In genomes, regulatory elements, including enhancers, 64 promoters, insulators, and silencers, actively interact with each other, which determines the 65 correct and stable level of gene expression (6, 7). The boundaries between TADs delineate 66 specific genomic regions, and more effective interactions between regulatory elements occur 67 within these regions than between different regions (8). According to the generally accepted 68 model, the cohesin complex, which is retained at CTCF protein binding sites, plays a primary 69 role in the formation of chromatin loops in mammals (9). Auxiliary roles in the organization of 70 specific interactions between enhancers and promoters have been assigned to the proteins LBD1, 71 yin yang 1 (YY1), and ZF143 (10-13). Because the LBD1 protein is the only one of these 72 proteins to contain a well-described homodimerization domain (14), how specific interactions 73 between enhancers and promoters occurs remains unclear. 74 In Drosophila, we suggested the existence of a large family of architectural proteins, which 75 typically contain N-terminal homodimerization domains and arrays of the zinc-finger Cys2-His2 76 (C2H2) domains (15)(16)(17)(18)(19)(20)(21)(22). The specific interactions that occur between the N-terminal domains of 77 architectural proteins can support selective distance interactions between regulatory elements. 78 Pita belongs to a large family of architectural proteins that feature zinc finger-associated domains 79 (ZADs) at the N-terminus (21, 23). Investigations of three architectural proteins, Pita, Zw5, and 80 ZIPIC, showed that the ZAD domains form only homodimers and support specific distance 81 interactions between sites bound by the same architectural protein (17). The 683 aa Pita protein 82 contains an N-terminal ZAD domain (17-93 aa) and a central cluster, consisting of 10 C2H2 5 83 zinc-finger domains (286-562 aa) (24,25). Pita is an essential Drosophila protein, and null pita 84 mutants die during embryogenesis (24,26). 85 Pita binds to a large 15 bp consensus site that is frequently found in gene promoters and 86 intergenic regulatory elements, including boundary/insulator elements in the Bithorax complex 87 (Bx-C) (Maksimenko et al. 2015;Kyrchanova et al. 2017). The Bithorax complex (BX-C) 88 contains three homeotic genes, Ultrabithorax (Ubx), abdominal-A (abd-A), and Abdominal-B 89 (Abd-B), which are responsible for specifying the parasegments (PS5 to PS13) that comprise the 90 posterior two-thirds of the fly segments (27)(28)(29). The expression of each homeotic gene in the 91 appropriate parasegment-specific pattern is controlled by independent cis-regulatory domains 92 that are separated by boundaries. For example, the regulatory domains iab-5, iab-6, and iab-7, 93 determine the expression of Abd-B in the abdominal segments A5, A6, and A7, respectively. The 94 MCP, Fab-6, Fab-7, and Fab-8 boundaries ensure the autonomous function of iab domains (30-95 37). Pita binds to Fab-7 and MCP and is required for their boundary activities (19,20,38). Five 96 Pita binding sites can functionally substitute the Fab-7 boundary that separates the iab-6 and iab-97 7 regulatory domains (19). Previously, Pita was found to interact with CP190 (25), which is also 98 known to bind several other C2H2 architectural proteins, including dCTCF and suppressor of 99 hairy wing [Su(Hw)] (25,(39)(40)(41)(42)(43). 100 Here, we studied the interaction mechanisms between Pita and CP190. Two domains that interact 101 with the BTB domain of CP190 were mapped in Pita. The recruitment of CP190 is required for 102 the chromatin opening and insulator functions of Pita. However, mutant flies that express Pita 103 lacking the CP190 interaction region display normal viability and wild-type (wt) phenotype, 104 demonstrating that these activities are not essential for Pita functions in vivo.

110
To understand the interaction mechanism between the architectural protein Pita and the BTB 111 domain of CP190, we attempted to precisely map the interaction regions in Pita. Previously, we 112 found that the BTB domain of CP190 interacted with the 95-302 aa region of Pita, which was 113 mapped between the ZAD and the C2H2 clusters (25). We used bacteria to express overlapping 114 glutathione S-transferase (GST)-fusion peptides that covered the 95-302 aa region of Pita. The 115 borders of the deletion derivatives were set according to conserved blocks of amino acids in Pita 116 protein from various Drosophila species. The obtained GST-peptides were tested for interactions 117 with the CP190 BTB domain, fused with 6×His, in a pull-down assay (Fig. 1A). This process 118 allowed us to map two binding regions between 95-165 aa and 220-232 aa (Fig. 1B, C).

119
Interestingly, the deletion of 220-232 aa, which was defined as a 13 aa core, resulted in the 120 complete loss of interaction between the 95-302 fragment and BTB in a pull-down assay, even 121 though this protein fragment still contained the second binding region. The 13 aa core was 122 predicted to be unstructured, but it contains several conserved hydrophobic residues (Fig. 1D).

123
Taken together, these results showed that the BTB domain interacts with the 95-165 aa region 124 and the 13 aa core, whose sequences have no obvious homology. The 95-165 aa region appeared 125 to stabilize the interaction between the BTB domain and the 13 aa core.

126
To better understand the functional significance of the interaction between Pita and CP190, we 127 deleted the 13 aa core that is necessary for Pita to bind with CP190 in vivo (Pita ΔCP1 ). The Pita wt 128 and Pita ΔCP1 proteins were tagged with 3×FLAG ( Fig. 2A) and co-expressed with CP190 in S2 The CP190 interacting domain in Pita is not essential for its role in Drosophila development 134 To understand the functional roles of the 13 aa core (CP1) and the 95-165 aa regions (CP2) in 135 Pita, we used previously described null mutations in the pita/spdk gene: pita 02132 and pita k05606 136 (Bloomington stock numbers 11179 and 10390, respectively). Pita protein is essential for early 137 Drosophila development and mitoses, and homozygotes bearing the null mutation died during 138 the embryonic stage (24,26). Transgenes expressing Pita wt -FLAG, Pita CP1 -FLAG, or Pita CP1+2 -139 FLAG under control of the Ubi promoter (Ubi-Pita wt , Ubi-Pita CP1 , and Ubi-Pita CP1+2 ) were 140 inserted into the same 86Fb region on the third chromosome, using a φC31 integrase-based ChIP-seq signal values were estimated in the set of Flag peaks reproduced in Pita wt and Pita ΔCP1+2 154 embryos. We found 5,023 such FLAG peaks (Fig. 3B). Then, we defined 1,029 peaks that 155 overlapped with the Pita motif site obtained from previously published data (17). From among 156 these 1,029 peaks, we selected 44 peaks that demonstrated an enhanced signal in Pita wt embryos 8 158 intersect with Pita motif sites, we found only 10 peaks with enhanced signals in Pita wt compared 159 with Pita ΔCP1+2 . As a result, the Pita ΔCP1+2 binding efficiency was only significantly reduced in a 160 minor proportion of the binding sites. Thus, CP190 binding is not essential for Pita binding to 161 most chromatin sites.

162
All Pita peaks were divided into three groups. In group 1, we included Pita motif site peaks with 163 at least a 2-fold decrease in the average signal for Pita ΔCP1+2 embryos compared with that in 164 Pita wt embryos (Fig. 3C). Group 2 consisted of peaks with Pita motif sites in which no significant 165 changes in the FLAG signals were observed when comparing the results of Pita wt and Pita ΔCP1+2 166 embryos (Fig. 3D). All FLAG peaks that did not intersect with a Pita motif were included in 167 Group 3 (Fig. S1A).

168
Then we compared the CP190 signal in these three groups of peaks. CP190 binding falls  construct. To express the chimeric protein, flies were maintained at 29°C from the embryonic to 198 pupal stages, as described in (50). 199 We used a previously described transgenic line (50) are lost during morphogenesis (Fig. 6B). In the absence of a boundary between these two 223 domains (Fab-7 attP50 mutant males), iab-7 is ectopically activated in all A6 (PS11) cells, and they 224 assume an A7 (PS12) identity. These males lack both the A6 and A7 segments (Fig. 6B). The 225 insertion of the Pita ×5 sites blocks the cross-talk between the iab-6 and iab-7 domains but does 226 not allow for communications between the iab-6 enhancers and the Abd-B promoter. As a result, 227 the iab-5 enhancers stimulate the Abd-B transcription in A6, which results in the conversion of 228 the A6 segment into one that resembles the A5 segment (Fig. 6B). Decreasing the protein level 229 by half due to the introduction of the Pita mutation leads to the loss of the insulating function of 230 the Pita ×5 boundary in some cells, which is reflected by the reduction and deformation of the A6 231 sternite (Fig. 6B).

232
Boundaries typically function better when present on both homologous chromosomes, which is 233 likely because homologous pairing improves the binding of proteins to boundaries.

234
Heterozygous Pita ×5 /+ males display a very weak A6  A5 transformation, suggesting that, 235 even in one copy, Pita ×5 can block the cross-talk between the iab-6 and iab-7 regulatory domains 236 (Fig. 6B). However, Pita ×5 /+ males that also carry heterozygous cp2/+ (or cp3/+) display the 237 partial transformation of A6 into a copy of A7 (Fig. 6B). The equally high sensitivity to 238 mutations in genes encoding both Pita and CP190 suggests that CP190 acts as a key factor in the 239 organization of the Pita-mediated boundary.

240
Next, we combined one copy of the Ubi-Pita wt or Ubi-Pita CP1 with Pita ×5 (Fig. 6C). In contrast 241 with Pita wt -FLAG, the overexpression of Pita CP1 -FLAG led to a partial transformation of A6 242 towards A7 (Fig. 6C). To test changes in the binding of Pita variants and CP190 with the Pita ×5 243 region, we used the quantitative analysis of ChIP (ChIP-qPCR) performed in extracts obtained 244 from adult three-day-old males (Fig. 6D). Anti-FLAG antibodies were used to test the over-245 expressed Pita variants. The ChIP study showed that Pita wt -FLAG and Pita CP1 -FLAG bound 246 with similar efficiency to the Pita ×5 region. In contrast, the binding of CP190 to the Pita ×5 region 247 was reduced in a transgenic line expressing Pita CP1 . Thus, boundary activity mediated by Pita ×5 248 was closely correlated with the efficiency of attracting CP190 to this region.

249
To directly demonstrate the role played by the CP190-Pita interaction during boundary activity, results confirmed that the 13 aa core is essential for the binding between CP190 and the Pita sites 257 and that CP190 is essential for the boundary activity of Pita. involved in the regulation of the abd-A and Abd-B genes (19, 61). The binding of dCTCF to 307 MCP is highly dependent on the presence of the Pita site, suggesting that Pita may function to 308 assist the binding between other architectural proteins and regulatory elements. The inability of 309 Pita to interact with CP190 is likely compensated for by other architectural proteins that 310 cooperate with Pita in the organization of the same regulatory regions. Indeed, we observed that 311 CP190 still binds to most genomic sites associated with the Pita CP1+2 protein in embryos. In 312 many cases, these sites are associated with proteins that are known to be able to recruit CP190 313 (25, 39-41, 62-65). Such functional redundancy creates a stable and reliable architecture of 314 regulatory elements, which is necessary for the correct regulation of genes during development.  (17). Anti-CP190 antibodies and rat IgG were used for co-immunoprecipitations.

352
The results were analysed by Western blotting. Proteins were detected using the ECL Plus  The enrichment of specific DNA fragments was analysed by real-time PCR using a QuantStudio

398
ChIP-seq analysis was performed for 4 samples (Flag and CP190 in Pita wt and Pita ΔCP1+2 lines); 399 two biological replicates were obtained for each sample. Paired-end sequencing technology was 400 applied, with an average read length of 101. Adapters, poly-N, and poly-A read ends were 401 removed using cutadapt software (67). Cutadapt was also used to trim low-quality ends (quality 402 threshold was set to 20 and reads with lengths less than 20 bp after trimming were discarded).

403
The remaining reads were aligned against version dm6 of the Drosophila melanogaster genome 404 using Bowtie version 2 (68). Only reads that aligned concordantly exactly one time were passed 18 405 for further analysis. The average insert size between mates was 156 bps. and the Self-consistency Ratio (SR) was less than 2, see Table S1)]