Pits, a protein interacting with Ttk69 and Sin3A, has links to histone deacetylation.

Histone deacetylation plays an important role in transcriptional repression. Previous results showed that the genetic interaction between ttk and rpd3, which encodes a class I histone deacetylase, is required for tll repression. This study investigated the molecular mechanism by which Ttk69 recruits Rpd3. Using yeast two-hybrid screening and datamining, one novel protein was found that weakly interacts with Ttk69 and Sin3A, designated as Protein interacting with Ttk69 and Sin3A (Pits). Pits protein expressed in the early stages of embryos and bound to the region of the tor response element in vivo. Expanded tll expression patterns were observed in embryos lacking maternal pits activity and the expansion was not widened by reducing either maternal ttk or sin3A activity. However, in embryos with simultaneously reduced maternal pits and sin3A activities or maternal pits, sin3A and ttk activities, the proportions of the embryos with expanded tll expression were significantly increased. These results indicate that all three gene activities are involved in tll repression. Level of histone H3 acetylation in the tll proximal region was found to be elevated in embryo with reduced these three gene activities. In conclusion, Ttk69 causes the histone deacetylation-mediated repression of tll via the interaction of Pits and Sin3A.

Eukaryotic cells have evolved extremely sophisticated means of regulating and fine-tuning expression of genes in response to various stimuli. Transcriptional activators and repressors play key roles in these activities to control gene expression. In addition, enzymes catalyse acetylation and deacetylation of the core histones and work closely with these transcription factors, as well as with various co-factors, to dynamically change chromatin status from open to closed and vice versa. Chromatin status correlates well with the activation and repression of transcription. In open chromatin, acetylation of the amino-termini of the histones neutralizes the positive charge of these amino acid residues, which results in loose contact between DNA and the nucleosome. When this occurs, transcriptional activators can easily access the appropriate binding sites, and genes are actively transcribed. In contrast, in closed chromatin, the amino-termini of the histones are hypoacetylated and genes are silenced1. Consistent with this paradigm, HATs are recruited by transcriptional activators to increase the acetylation level of local chromatin, whereas HDACs are recruited by transcriptional repressors to diminish local acetylation23. Both HATs and HDACs are associated with scaffold proteins and form large multiprotein complexes45.

Scaffold proteins associate with various proteins to coordinate their functions in various cellular processes6. Sin3A is one of these scaffold proteins and contains four highly conserved paired amphipathic helix domains, PAH1 to PAH4. The functions of these PAH domains are conserved from yeast to human. For example, a region in PAH3 is known to interact with HDACs. Furthermore, PAH1 and PAH2 bind a variety of transcriptional repressors, as well as co-repressors, and also assist in transcriptional repression in eukaryotes7. Components forming the core of the HDAC/Sin3A complexes include HDAC1 (Rpd3 in yeast and fly), HDAC2, RbAp46, RbAp488, RBP19, and/or p33ING1b10. Sin3A also binds to a number of docking proteins, e.g. SAP30, SAP18, and SAP2511. Because the HDAC/Sin3A complex lacks the ability to bind DNA, it must associate with DNA sequence-specific repressors to function. These repressors include Mad1, E2F-4, MeCP2, ELK1, and KLF1213. The recruitment of the HDAC/Sin3A complex by these repressors triggers transcriptional repression via deacetylation and the remodeling of local chromatin into the closed status in the vicinity of the repressor cognate sites14. We have shown that Ttk69, but not Ttk88 that is an alternatively spliced product of the ttk gene, participates in tll repression15. Additionally, the genetic interaction between ttk and rpd3 is required for the repression16. However, the mechanism by which Rpd3 is recruited is unclear.

Ttk69 is a co-repressor that forms a complex with Hsf and GAGA factor (GAF), and this complex binds to the tor response element (tor-RE) in the tll proximal region1617. Ttk69 contains a BTB domain and a zinc-finger motif at its N-terminus and C-terminus, respectively18. Ttk69 binds to TCCT elements to regulate the spatial and temporal expression of the eve, h, odd, run and tll genes during Drosophila embryogenesis15192021. To investigate how Ttk69 recruits Rpd3, yeast two-hybrid screening and database mining were used to find a novel protein interacting with both Ttk69 and Sin3A. The protein was designated as Protein interacting with Ttk69 and Sin3A (Pits). Mutants deficient in pits expression were generated to reveal its role in tll repression. Dosage-dependent genetic interaction experiments were utilised to determine that the genetic interactions of pits with ttk and sin3A are important for tll repression. Furthermore, chromatin immunoprecipitation (ChIP) was used to show that the level of histone acetylation is increased in the tll proximal region in embryos with reduced pits, sin3A and ttk activities. These results support the possibility that Pits is a novel mediator linking Ttk69 to histone deacetylation via protein-protein interactions between Pits, Sin3A and Ttk69.

Results

CG11138 interacts with both Ttk69 and Sin3A

To explore how Ttk69 recruits Rpd3, the yeast-two hybrid system was utilised to identify Ttk69 interacting proteins. Starting with 34 newly isolated interacting proteins, attempts to find a linkage of Ttk69 with Rpd3 failed. Then, marginally interacting proteins that were histidine positive but mostly β-galactosidase negative were picked to mine protein-protein interaction databases. The data in BioGRID showed that CG11138 interacts with not only Ttk69, but also Sin3A (http://thebiogrid.org/).

To verify these protein-protein interactions, pull-down experiments were performed. Full-length of Ttk69 (T-FL) might undergo a conformational change because the N-terminal His-tag on T-FL was detected using an anti-His tag antibody (Fig. 1b), but failed to bind Ni-affinity resin (data not shown). Importance of this putative conformational change is discussed later. To explore this possibility, it was essential to use various fragments of the protein. The amino-acid conservation shown in the UCSC genome browser (http://genome.ucsc.edu/) served as a basis for selecting such fragments of Ttk69 (Fig. 1a). Sin3A is a large protein (Fig. 1a), and Zhao et al. have reported that it is difficult to synthesize the full-length Sin3A in bacteria22. Therefore, using the same rationale, the four PAHs were synthesized. The results from pull-down experiments showed that T-FL weakly interacts with the C-terminus of CG11138 (38-C), but not with full-length CG11138 (38-FL). The N-terminal portion of Ttk69 (T-N) was found to interact with both 38-FL and the middle portion of CG11138 (38-M) (Fig. 1b), supporting the above hypothesis. Furthermore, both 38-FL and 38-M interacted with PAH1. Other portions of CG11138, 38-N and 38-C were found to associate with PAH3 and PAH4 very weakly (Fig. 1c). These results indicated that multiple regions in CG11138 bind weakly to various PAHs and Ttk69. To determine that these three proteins exist in the same protein complex, a co-immunoprecipitation experiment with an anti-Pits antibody was performed. Considering the weak interactions between these three proteins and the relatively low quantity of Pits in the nucleus (Fig. 2i,j), proteins in nuclear extracts were cross-linked by bismaleimidohexane (BMH) before the immunoprecipitation. As shown in Fig. 1d, both Ttk69 and Sin3A were in the same protein complex immunoprecipitated by anti-Pits antibody. These results are consistent with information presented in BioGRID whereby CG11138 directly interacts with both Ttk69 and Sin3A to form a protein complex. Thus, CG11138 was renamed as Pits.

Figure 1

CG11138 interacts with both Ttk69 and Sin3A.

(a) The shaded and black boxes indicate amino acid residues that are identical among Drosophila species and insects, respectively. The dotted box in Sin3A represents the domain that interacts with HDACs. For CG11138 and Ttk69, full-length (FL) and three protein fragments, N, M or C, as indicated by bars below the diagram. The four highly conserved domains in Sin3A, PAH1 to PAH4, as previously described22, were expressed in bacteria. (b,c) Crude extracts containing the GST fusion proteins shown at the top of the panels, were mixed with crude extracts containing one of the S-tag fusion proteins, except for the full-length of Ttk69, which was His-tagged. The proteins were pulled down by glutathione agarose and detected by western blotting using either anti-S tag or anti-His tag antibody. Protein samples, consisting of 10% of the sample used for the pull-down assay, were used as the controls and are designated as Input. (d) Ttk69 and Sin3A in a cross-linked protein complex immunoprecipitated by an anti-Pits antibody. The co-IPed protein complexes were separated in SDS agarose-polyacrylamide gels and proteins in the complexes were detected by western blotting with an anti-Ttk69 antibody. The membrane was then stripped to allow detection of Sin3A.

Figure 2

Pits distribution patterns in Drosophila embryo.

A preabsorbed anti-Pits antibody was used to detect Pits distribution in Drosophila embryos by immunostaining. Embryos at stage 3 (a), stage 4 (b,i,i’), stage 5 (c,j,j’), stage 6 (d), stage 9 (e), stage 10 (f) and stage 13 (g,h) are shown. The anterior is displayed to the left. Except for panel h, which shows a ventral view, all panels show a sagittal view of the embryo. (i,j) To reveal whether Pits is in the nucleus, wild-type embryo was stained with the pre-absorbed anti-Pits (green) and anti-LamC (red) antibodies, and observed using confocal microscopy. Chromosome was labeled with Hoechst 33342 (blue). Punctate staining of Pits was detected in early stage 4 (i,i’) and 5 embryos (j,j’). Dashed yellow circles in panels i’ and j’ represent regions of embryonic nuclei to clearly show Pits nuclear localization. Scale bars are 5 μm.

Pits protein expresses in the early stages of the embryo and later presents at high levels in midgut and central nervous system (CNS)

The data in FlyExpress show uniform pits mRNA in embryos from stages1 to 5. At stage 9, pits mRNA is enriched in the midgut primordium (http://www.flyexpress.net/). To reveal whether Pits protein follows the same patterns, an anti-Pits antibody was raised and affinity purified. Immunostaining results revealed a high level of Pits in embryos at stages 3 and earlier (Fig. 2a), supporting the possibility that pits mRNA is deposited into the egg during oogenesis. This type of mRNAs encodes regulators to control embryogenesis in the early embryonic stages, known as maternal gene activity. Pits protein gradually declined to an undetectable level from stages 3 to 6 (Fig. 2b–d) and then reappeared at stage 8 (Fig. 2e). Later, from stages 10 to 13 (Fig. 2f,g), Pits protein was detected at high levels in midgut and CNS. The midgut and CNS localizations are consistent with the mRNA patterns shown in FlyExpress and the YFP-tagged protein patterns23.

Because Pits is a candidate co-repressor for tll, it should exist in the nucleus. To reveal whether Pits localizes to the nucleus, immunostaining with confocal microscopy was performed. The results showed 1) punctate patterns uniformly distributed in embryonic cytoplasm at stages 5 and earlier, and 2) lower number of Pits in nuclei, relative to the number in the cytoplasm (Fig. 2i,j). This lower level of Pits in the nucleus opposed its possible role in tll repression. However, two lines of evidence still supported its involvement in this process. The first was the results of ChIP experiments, which indicated that both Pits and Ttk69 associate with the tor-RE (Fig. 3). The second was that Pits protein contains two domains that are highly homologous to the zinc finger in IRF2BP and the RING finger (Supplementary Fig. S1). Proteins containing either domain repress expression of genes2425. Taken together with the interactions between Pits, Ttk69 and Sin3A in vitro, Pits might be involved in tll repression.

Figure 3

Pits and Ttk69 co-localize to the vicinity of the tor-RE in vivo.

At the top of panel a, a diagram shows the relative positions of the transcription unit of tll, the tor-RE and a set of primers that is represented by arrows. Embryos were collected from parents of Oregon-R/w (wt) or pits (pits) (a) or ttk da-GAL4/+ da-GAL4 females crossed with either UAS-GFP (GFP) or UAS-ttk69-HA (ttk-HA) males (b) every 2 hours. Embryos aged for 30 min (a) or 60 min (b) were used in ChIP experiments with anti-Pits (“+”; 6 μg) or anti-HA tag antibody (“+”; 3 μg). Chromatin samples, consisting of 5% of the sample used for ChIP, were used as “input” controls. Mock control, “−”, used the same experimental procedure without antibody. The primer set, tor, was used to reveal whether Pits or Ttk69-HA exists in the vicinity of the tor-RE. Rp is a set of primers to detect the RPII140 gene that serves as a PCR negative control. The PCR products were separated in a 4% agarose gel, followed by ethidium bromide fluorography.

The maternal pits activity is important for tll repression

To test the above hypothesis, pits deletions were generated using the imprecise P-element excision method with EP1313, which is inserted 38 bp downstream of the putative transcription initiation site (Fig. 4a). Two pits alleles, 94 and 240, were obtained that had identical molecular lesions where exon 1 was almost deleted (Fig. 4a). The transcription initiation site remained intact in both deletions. Surprisingly, both alleles are homozygote viable with no observable phenotype. Another pits allele, 64, was found to be semi-lethal. No PCR product was produced when the upstream primers were used (Fig. 4a), suggesting that the lethality results from truncation of one or more genes upstream of pits.

Figure 4

Structure and alleles of the pits gene.

(a) Exons are represented by boxes. Three subregions of Pits, N, M and C as shown in Fig. 1a, are shown by grid, solid and dotted rectangles. The pits gene encodes three putative transcripts, represented by RC, RD and RE. Transcripts C and E are the result of alternative splicing inside exon 2. Furthermore, the 3’ untranslated regions in these three transcripts are different. Using the imprecise P-element excision method and a P-element line, EP1313 inserted 38 bp downstream of the putative transcription initiation site for the transcripts, three pits deletions, 64, 94 and 240, were obtained. The arrows indicate the positions and direction of the primers used for screening the pits deletions. The range of each deletion, from position +39 to +1621, is indicated by brackets. The dashed bracket indicates that the 5′ end of the deletion has not been determined. (b) Pits levels in w, pits, pits and pits embryos from 0 to 4 hours were assessed by western blotting with an anti-Pits antibody. The membrane was then stripped to allow detection of β-Tubulin (β-Tub), which served as the loading control. M represents the protein size marker.

To explore Pits levels in pits, pits and pits embryos in the early stages, western blotting was performed. The results revealed a ~75 kDa protein in w embryo with a minor band, just below the major band, also visible (Fig. 4b). These two bands were presumably the products translated from transcripts C and E (see legend of Fig. 4), despite the fact that both proteins were larger than the calculated molecular weights of 60 and 62 kDa. Although no Pits protein was detected in pits or pits embryos (Fig. 4b), perhaps, both pits and pits are not null alleles. The lethality of embryos (38.0%) obtained from pits/Df(X)BSC624 females crossed with pits males was much higher than those from pits or pits parents (approximately 3.5%). The difference might result from that an open reading frame in the truncated pits mRNA encodes a truncated polypeptide containing 29% of 38-M and intact 38-C, which may have residual Pits activity, and that quantity of the truncated Pits is too low to be detected by western blotting. Alternatively, a gene within the deleted region of Df(X)BSC624 may contribute the phenotype, which is similar to the finding of the sad1 deletion26. Residual Pits in pits embryos (Fig. 4b) further supported the possibility that the lethality of pits was due to defects in one or more other gene(s).

To explore whether maternal pits activity is important for tll repression, embryos were collected from pits females crossed with w males. Then, tll expression patterns were investigated by in situ hybridization with digoxigenin-labeled antisense tll RNA probe. In 47% of embryos with reduced maternal pits activity, tll expression patterns were slightly expanded towards the central region of embryos (Figs 5f and 6s). The expanded patterns had gradually returned to normal by stage 8 (Fig. 4h) when compared with wild-type embryos (Fig. 5d). The degree of expansion of the tll expression pattern was similar in embryos from pits parents (data not shown) and embryos from pits/Df(X)BSC624 females crossed with pits males (Figs 5i–l and 6s). These results supported the idea that maternal pits activity is important for tll repression, and suggested that zygotic pits activity contributes less in tll repression, which is consistent with the lack of Pits protein in embryos from late-stage 5 to stage 7 (Fig. 2).

Figure 5

Maternal pits activity is important for tll repression.

The tll expression patterns in embryos from w (a–d), a cross of pits with w (pits/+; e–h) and a cross of pits/Df(X)BSC624 with pit (pits/Df(X); i–l) were determined by in situ hybridization using digoxigenin-labeled antisense tll RNA as the probe. The embryonic stages are indicated at the top of the panels. Embryos are arranged in a sagittal view, with the anterior towards the left.

Figure 6

The genetic interactions of pits with sin3A and ttk are important for tll repression.

To test the maternal effect of the gene activities, embryos from females pits (a,b), pits; sin3A/+ (c,d), pits; ttk/+ (g,h), sin3A/+; ttk/+ (k,l) and pits; sin3A/+; ttk (o,p) crossed with w males were collected to determine tll expression patterns. To test the maternal and zygotic effect of the genes’ activities, females and males that had the same genotypes, as shown at the left, were mated (e,i,m,q). Due to undetectable GFP protein expressed by the ubi-GFP transgene in the balancer chromosome in early embryonic stages, genotypes of embryos homozygous for sin3A or ttk could not be determined. Therefore, RNAi was used. Females carrying the da-GAL4 transgene were mated with RNAi lines to knock down at least two gene activities simultaneously (f,j,n,r). The tll expression patterns in the embryos were revealed by in situ hybridization with digoxigenin-labeled antisense tll RNA. Embryos are arranged in a sagittal view, with the anterior towards the left. (s) A bar graph presents percentages of embryos with expanded tll expression patterns. “M” in brackets represents embryos only with reduced maternal gene activities. Percentage indicates proportion of embryos with expanded tll expression patterns over the total number of stage-4 (solid bars) or stage-5 (open bars) embryos. The left and right numbers beneath each genotype are the total numbers of stage-4 and stage-5 embryos. Df(X) represents Df(X)BSC624. Fisher’s exact test was used to determine statistical significance of proportion of embryos with the expanded tll expression from pits (M) mothers or Df(X)/pits crossed with pits males against those with further reduction of one or two more gene activities (*: increase, #: decrease, p < 0.001).

Maternal pits, sin3A and ttk activities work together to repress tll expression

The results described above indicated that maternal pits activity is involved in tll repression. To test whether maternal pits interacts genetically with maternal sin3A and/or ttk to repress tll expression, the dose dependent genetic interaction experiment was performed. Embryos were collected from females that had activities of two or three of the relevant genes reduced and then crossed with w males. These embryos were used to determine tll expression patterns. The results showed that the percentage of embryos with the expanded tll expression significantly increased from 47% in pits to 71% in pits; sin3A/+ (Fig. 6s). Hereafter, percentage is used to indicate embryos with the expanded tll expression. Although there were reduced percentages of pits; ttk/+ and sin3A/+; ttk/+ embryos (36% and 39%, respectively; Fig. 6h,l,s), unexpectedly, the difference relative to pits94 embryos were not significant (Fig. 6b,s). These results countered the hypothesis that interactions between these genes are involved in tll repression. Nevertheless, a simultaneous reduction of the maternal pits, sin3A and ttk activities significantly increased the percentage of the embryos (82%, Fig. 6p,s), compared to that of pits94 embryos. These results supported the hypothesis that the maternal activities of pits, sin3A and ttk work together to repress tll expression.

Sin3A is a widely distributed factor27. The zygotic activities of the above genes might function in tll repression in spite of the facts that both Pits and Ttk69 are not detected in late stage-5 embryo15. To test this possibility here, tll expression patterns in embryos from parents of pits; sin3A/+, pits; ttk/+, sin3A/+; ttk/+ or pits; sin3A/+; ttk/+ were examined. Because the maternal-zygotic transition occurs at late stage 4, effects on tll expression patterns at either late stage 4 or early stage 5 should result from both maternal and zygotic gene activities. The results showed no significant changes in the degree of expansions in these embryos (Fig. 6e,i,m), compared to those with reduced maternal gene activities (Fig. 6d,h,l). Unexpectedly, the reduction of both maternal and zygotic ttk and sin3A gene activities led to an exceptionally low percentage of embryos (29%, Fig. 6m,s). This may be due to dual functions of Ttk69 and Sin3A (see Discussion). Again, the expanded tll expression in embryos from pits sin3A ttk parents was observed in the middle region of the embryos, compared to those with reduced maternal activities only (Fig. 6p,q). The percentages of embryos significantly increased as 79% (Fig. 6s). In all cases, the expanded tll expression patterns returned to normal by stage 10 (Supplementary Fig. S2). These results indicated that the zygotic activities of these genes contributed less to tll repression.

Because it is difficult to identify embryo genotypes, RNAi was used to verify the results from the genetic interaction experiments. The rationale for this experiment was that maternally provided GAL4 from the da-GAL4 transgene drives the synthesis of double stranded RNA at early stage 4, which consists of the timing of zygotic gene expression and presents the beginning of the stage lasting 50 minutes. Hypothetically, all targeted mRNAs are knocked down by late stage 4, but not early stage 4. Therefore, females carrying the da-GAL4 transgene were crossed with males carrying at least two UAS-RNAi transgenes. tll expression patterns were only observed at late stage 4 and early stage 5. In all cases, degree of the expanded tll expression patterns were increased (Fig. 6f,j,n,r vs. Fig. 6e,i,m,q). Consistently, the percentages of embryos were similar to those with reduced maternal and zygotic gene activities. These data supported the conclusion that interactions between maternal pits, sin3A and ttk gene activities are important for attenuating tll expression.

An increased level of histone H3 acetylation in the tll proximal region is present when there is reduced pits, sin3A and ttk activity

Our previous and current results have indicated that the genetic interactions of ttk69 with rpd3 and sin3A are important for tll repression16, which suggests that Rpd3 is recruited by Ttk69 through Pits and Sin3A to deacetylate histones and results in an attenuation of tll expression. Thus, histone acetylation in the region adjacent to the tor-RE ought to be increased in embryos that have reduced pits, sin3A and ttk activity. To test this possibility, the level of histone acetylation in the tll proximal region was determined by ChIP using an antibody against histone H3 that is acetylated at Lys9 and Lys14. The results showed that the relative amount of acetylated histone H3 in the vicinity of the tor-RE is increased by 4.7-fold in embryos with reduced pits, sin3A and ttk69 activity (22.23-fold; Fig. 7). These results supported the hypothesis that histone deacetylation is involved in tll repression. In conclusion, the above findings show that the Ttk69 co-repressor recruits the Rpd3/Sin3A complex to exert histone deacetylation-mediated transcriptional repression of tll.

Figure 7

Simultaneous reduction of pits, sin3A and ttk activity increases the level of histone acetylation in the tll proximal region.

Embryos were collected from da-GAL4 females crossed with w males (da-GAL4/+) or pits/+; sin3A/+; ttk da-GAL4/daGAL4 females crossed with males carrying multiple transgenes to knock down pits, sin3A and ttk69 mRNA (pits; sin; ttk). Embryos from these crosses were used to determine the level of histone acetylation in region adjacent to the tor-RE by ChIP with an anti-acetyl histone H3 specific antibody (Merck Millipore). A cluster of Pho binding sites, called 5X Pho that is 173 bp upstream the tor-RE, is at the 5′ end of a putative PRE40. Arrows indicate a set of primers used for real-time PCR. Detection of act-5C served as an endogenous control. CT values of act were used to normalize tll, designated as ΔCT. Relative amounts are represented by −ΔΔCT where the ΔCT value for the pits; sin; ttk embryos are subtracted from that of the control embryos. Significance difference was determined by Student’s t-test (*p < 0.05).

Discussion

Pits plays an important role in Ttk69-mediated recruitment of the Rpd3/Sin3A complex to attenuate tll expression. When initiating tll repression at early stage 4, Ttk69 acts as a co-repressor that increases the ability of the GAF/Hsf heterodimer to bind to the tor-RE16. The Rpd3/Sin3A complex is then brought to the tor-RE via the interaction of Pits with both Ttk69 and Sin3A. As a result, the level of acetylated histone H3 in the vicinity of the tor-RE decreases in the middle region of embryo (top panel in Fig. 8). This hypoacetylation is consistent with the data in modENCODE (http://gbrowse.modencode.org/fgb2/gbrowse/fly/), which shows no acetylation or low level on Lys9 and Lys27 of histone H3 from position −550 to +160 in the tll gene. In addition, Rpd3, HDAC3 and HDAC6 are detected in this region in vivo (modENCODE). This raises the question as to which enzyme executes the deacetylation. The data in the modENCODE development RNA-Seq database show that the level of rpd3 mRNA is significantly higher than those of HDAC3 and DHAC6 mRNAs in 0–4 hr embryos28. Therefore, in the early stages of Drosophila embryo, Rpd3 is the major enzyme to execute histone deacetylation, which results in the chromatin status of the tll gene switching to the closed configuration. In turn, this conformational change results in attenuation of tll expression.

Figure 8

A model of Pits associating with Rpd3 and Mi-2/NuRD complexes with and without stimulation of the active tor pathway.

Previous results showed that GAF, Hsf and Ttk69 form a protein complex that binds to the tor-RE (grid rectangle). tll expression is attenuated in the middle of embryo where tor is inactive16. Here, Pits serves as a mediator to recruit the Rpd3/Sin3A complex. In addition, based on information from the literature, GAF likely recruits the Mi-2/NuRD complex (indicated by shaded diagrams)30. Both Rpd3 and NuRD complexes associate with Ttk69 and GAF to inhibit histone acetylation (Ac) in the tll locus. At both poles of Drosophila embryo, Erk activated by the active tor pathway phosphorylates Ttk69, Sin3A and Hsf, indicated by asterisks. The phosphorylated Hsf becomes an activator, whereas the phosphorylated Ttk69 is released from the protein complex, leading to the disruption of the association of the Rpd3 and NuRD complex with the tor-RE, and also converts to an activator that binds to TC5. Previous work has shown that Sin3A is phosphorylated by Erk and converts it to an activator34. Therefore, phosphorylated Sin3A may associate with Mi-2/MBD-like proteins to activate tll expression.

In addition to Pits, Ttk69 together with GAF interacts with other co-repressors that can recruit Rpd3. First, GAF interacts with SAP18, a docking protein in the Rpd3/Sin3A complex29. GAF, SAP18, Sin3A and Rpd3 co-localize to a Polycomb repression element, Fab-7. Reduction of these gene activities results in histone hypoacetylation on Fab-7 that reduces its function in gene silencing30. Secondly, GAF and Ttk69 associate with the Mi-2/NuRD chromatin remodeler complex (NuRD) that also contains Rpd331 (top panel in Fig. 8). Compensation by these redundant factors may therefore result in less expanded tll expression patterns indicated by lower levels of involvement of the zygotic activities in tll repression.

An opposite effect was observed in which the percentage of sin3A/+; ttk/+ embryos is significantly lower than others. Regardlessly, this may be similar to what occurs with stg expression that is up- or down-regulated by ttk69 overexpression or by sin3A knock-down. These may result from indirect inhibition of a stg activator32. Here, it is difficult to fully elucidate the opposite effect coming from indirect inhibition to expression of a tll activator. When considering NuRD involved in tll repression, it provides a better scenario to explain the opposite effect. Previous work demonstrated that SUMOylated LIN-1 acts as a repressor through NuRD, whereas LIN-1 phosphorylated by Erk becomes an activator and that NuRD is able to activate expression of certain genes3334. In addition, Erk converts Sin3A to an activator by phosphorylation35. Although Ttk69 is a transcription repressor, similarly, it also cooperates with REPO to activate expression of the M84 marker36. Thus, at both poles of embryo, Erk activated by the active tor pathway phosphorylates Ttk69 and Sin3A. Phosphorylated Ttk69 binds to the cognate binding sites (TC5) flanking the tor-RE1516 to activate tll expression and phosphorylated Sin3A retained by Mi-2/MBD-like activates tll expression (bottom panel in Fig. 8). When both Ttk69 and Sin3A are reduced, the repressive activity of Rpd3 and NuRD complexes may be slightly affected because Pits can hold on both complexes associated with the tor-RE. A low level of tll activation occurs in sin3A/+; ttk/+ embryos, leading to the lowest percentage. When Pits is also absent, the repressive activity of NuRD may greatly be affected, causing a high level of tll expression. Therefore, the embryo shows the highest percentages (Fig. 6q,s; bottom panel in Fig. 8). The involvement of Pits in the Rpd3 and NuRD repressive mechanisms provides insights into how transcription co-repressors fine-tune the expression levels of genes responding to stimuli, such as cellular signals34.

Another inconsistent result was observed in this study. The BTB domain (T-N) weakly interacts with 38-FL and 38-M, whereas T-FL weakly interacts with only 38-C. There is a two-part explanation for these results. First, Pits can contact the PAH1 domain in Sin3A to open its C-terminus; then, the free C-terminus of Pits can bind to Ttk69 and cause the release of the N-terminus of Ttk69 from the wrapping. The BTB domain of Ttk69 is an important domain for protein-protein interactions. Aside from interacting with dCtBP3 and SMRT37, it also strongly interacts with the BTB domains of other members of the Ttk BTB subfamily3839 and forms a protein complex. These co-repressor complexes may then function in tll repression.

In the proposed model, transcriptional co-repressors play important roles to silence gene expression by remodeling chromatin into a closed status that prevents transcription in eukaryotes. The pleiotropic roles of Sin3A in transcriptional repression have been shown. Sin3A forms the Rpd3/Sin3A complex to deacetylate the histones in the vicinity of a promoter7. Furthermore, the Rpd3/Sin3A complex can be expanded by adding extra catalytic modules onto the platform. In addition to NuRD described above, transcriptional modulators include enzymes that catalyze methylation; still others catalyze the O-linkage of monosaccharide N-acetylglucoseamine to histones and various other chromatin remodeling enzymes7. These chromatin remodelers play pivotal roles in epigenetic regulation40. In the tll gene, there is a putative Polycomb repression element (PRE) adjacent to the tor-RE41. In the present study, it was found that histone deacetylation occurs in the region overlapping the 5′ end of the PRE (Fig. 7). Thus, it is likely that chromatin structure in the tll gene is remodeled by the Rpd3/Sin3A complex to facilitate the silencing of posterior tll expression and that this occurs by Polycomb epigenetic repression (PER) after stage 6. The finding of Pits/Rpd3/Sin3A in initiating tll repression provides a clue for further investigation into how PER is established. This newly elucidated mechanism could increase our understanding of various aspects of metazoan biology, such as development, diseases and ageing.

Methods

Fly lines and genetics

Lines ttk/TM3 Sb42 and sin3A/CyO27, generously provided by Drs. Y.-N. Jan and D.-H. Huang, were used in the genetic studies, described below. Line UAS-ttk69-HA provided by Dr. L.C. Lai3 was used in ChIP to show the in vivo binding of Ttk69 to the tll locus. Three RNAi lines, w; P{GD7212}v18154, w; P{GD4387}v10808/CyO and w; P{GD4414}v10855 that knock down activities of pits, sin3A and ttk69, were obtained from Vienna Drosophila RNAi Center. Because the w; p{GD4387}v10808/CyO line has a few escapers, p{GD4387} was mobilized to obtain lines that are homozygous viable43.

P{EP}CG11138 from Bloomington Stock Center was used to generate pits deletions using imprecise excision method43 and polymerase chain reaction (PCR) (locations of primers are shown in Fig. 3a). Molecular lesions were revealed by DNA sequencing the PCR amplified DNA fragments. A deletion mutant, Df(X)BSC624 w/Binsinscy from Bloomington Stock Center, was used to test whether the newly generated pits alleles were null alleles.

Yeast two-hybrid screening

The coding region of ttk69 was amplified using PCR with a set of primers and pNB40844 as template DNA and inserted into of pAS2-1 (Clontech. Inc. Palo Alto, CA). The resulting plasmid DNA was transformed into a yeast strain Y187. A cDNA library, made from 0–12 hour of Drosophila embryo, was generously provided by Dr. C. Chein and transformed into a yeast strain, CG1945. Transformants were pooled to the final concentration of 50 OD600. Nine 1-ml aliquots were mated with the ttk69 transformants45. The cDNA inserts were amplified from the histidine positives by PCR and then digested with HaeIII for grouping. One from each group was randomly picked to confirm by analyzing β-galactosidase activity to eliminate false positives, compared to that of a negative control, pAS-Laminin. DNA sequences at the 5’end of the cDNA inserts were then determined and used to identify genes by BLAST.

Plasmid construction and expression of proteins in Escherichia coli

Amino acids conserved among Drosophila species and other insects served as bases for subcloning various fragments of Pits, Sin3A and Ttk69 into modified pGEX2T (GE Bioscience) and pET29a (Novagen) (Fig. 1a). Primers are listed in Supplementary Table S1 and used to amplify full-length and various portions of cDNAs in RE41430, LD13852 or pNB408. The PCR amplified DNA fragments were then cloned into the bacterial expression vectors. The GST- or S-tag-fusion proteins were expressed in either E. coli DH5α pG-tf2 or E. coli BL21 (DE3) pG-tf2 using procedures described by Ausubel et al.46. The expression of the proteins was monitored by the predicted size in SDS polyacrylamide gel46.

The GST-fused 38C was purified using glutathione agarose chromatography (GE Bioscience) and used to raise anti-Pits antibody. Production and affinity purification of anti-Pits antibody were performed by LTK BioLaboratories, Taiwan.

GST pull-down assays and western blotting

The detection of interactions of Pits with Sin3A or Ttk69 was carried out as previously described47.

Protein extracts from Drosophila embryos from various parents were prepared and used to determine Pits levels in these embryos. Western blotting was used to detect Pits and the bacterially expressed proteins during pull-down experiments48.

Co-immunoprecipitation

The co-IP procedure was carried out as previously described49 with modifications. Nuclei were isolated from 0.5-3-hr embryos17, resuspended in a lysis buffer containing 1.5 mM NaH2PO4, 8 mM Na2HPO4, 145 mM NaCl, 1 mM MgCl2, 10% glycerol, and protease inhibitors (EDTA free, Roche Applied Science). The nuclear debris was removed by centrifugation. Proteins in nuclear extracts were cross-linked by 25 μM of BMH (ThermoFisher Scientific). The cross-linked protein complexes were immunoprecipitated by an anti-Pits antibody described by Liu et al.49. The co-IPed proteins were separated in 0.5% agarose/3.4% SDS polyacrylamide gel50 and detected proteins in the co-IPed protein complexes by western blotting with anti-Ttk69 (affinity purified by GeneTex, Inc. Taiwan) and anti-Sin3A antibodies (Santa Cruz Biotechnology, Inc).

Immunostaining and in situ hybridization

Embryos were collected every 12 hours and fixed by 5% paraformaldehyde in 1X PBS and then incubated with anti-Pits antibody (1:2000)51. Because several proteins in w embryonic extract were weakly detected by the anti-Pits antibody (Fig. 4b), the antibody was preadsorbed by pits embryos to minimize background of the immunostaining. Pits localizations were revealed by incubating with an anti-rabbit IgG conjugated with alkaline phosphatase and colorimetric substances (5-bromo-4-chloro-3-indolyl phosphate and nitroblue tetrazolium, Roche) or anti-rabbit IgG conjugated with Cy351. in situ hybridization was carried out as described previously52. Patterns of Pits protein and of tll mRNA in embryos were viewed using DIC light microscope (Leica Model DMR) or confocal microscope (Olympus Model FV10).

Chromatin immunoprecipitation

The ChIP procedure was carried out as previously described53 with modifications. Staged embryos were collected to isolate paraformaldehyde-fixed and fragmented chromatin52. The fixed nuclei were pelleted by centrifugation at 2000× g. Background of protein-A Dynabeads (Invitrogen) was reduced by pre-treatment and blocking as previously described54.

A highly acetylated region in the act-5C gene presented in modENCODE was selected to serve as an endogenous control. Primers actin-F and actin-R (Supplementary Table S1) were designed to amplify the acetylated region, position from −190 to −314 in act-5C. To reveal whether a reduced level of pits, sin3A and ttk activity was correlated to an increased level of histone acetylation in the tll proximal region, primers tll-F and tll-R (Supplementary Table S1) were used to amplify the cis-regulatory regions of the tll gene. The relative quantity of ChIPed chromatin from tll, compared to act-5C, was determined using StepOne software v2.1 (ABI Biosystems) and SYBR qPCR kit (2x master mix, KAPA Biosystems), represented by ΔCT values. The ΔCT values from pits, sin3A and ttk embryos were subtracted from those of the control embryos to obtain ΔΔCT values.

To reveal whether Pits or Ttk69 binds to the vicinity of the tor-RE, primers tor-RE F and tor-RE R (Supplementary Table S1) were used.

Additional Information

How to cite this article: Liaw, G.-J. Pits, a protein interacting with Ttk69 and Sin3A, has links to histone deacetylation. Sci. Rep. 6, 33388; doi: 10.1038/srep33388 (2016).