Groucho corepressor functions as a cofactor for the Knirps short-range transcriptional repressor

Abstract

Despite the pervasive roles for repressors in transcriptional control, the range of action of these proteins on cis regulatory elements remains poorly understood. Knirps has essential roles in patterning the Drosophila embryo by means of short-range repression, an activity that is essential for proper regulation of complex transcriptional control elements. Short-range repressors function in a local fashion to interfere with the activity of activators or basal promoters within ≈100 bp. In contrast, long-range repressors such as Hairy act over distances >1 kb. The functional distinction between these two classes of repressors has been suggested to stem from the differential recruitment of the CtBP corepressor to short-range repressors and Groucho to long-range repressors. Contrary to this differential recruitment model, we report that Groucho is a functional part of the Knirps short-range repression complex. The corepressor interaction is mediated via an eh-1 like motif present in the N terminus and a conserved region present in the central portion of Knirps. We also show that this interaction is important for the CtBP-independent repression activity of Knirps and is required for regulation of even-skipped. Our study uncovers a previously uncharacterized interaction between proteins previously thought to function in distinct repression pathways, and indicates that the Groucho corepressor can be differentially harnessed to execute short- and long-range repression.

Short-range transcriptional repression has a central role in development, and perhaps nowhere have the molecular workings of eukaryotic developmental gene networks been more extensively analyzed than in the Drosophila blastoderm embryo. Here, both transcriptional activators and repressors transduce temporal and spatial information into characteristic patterns of gene expression essential for development. Repressors have key parts in this process, evidenced by the central position in the hierarchy of genes such as hairy, giant, knirps (kni), and Kruppel, all of which function as dedicated repressors. Transcriptional repressors have been characterized based on their range of action; short-range repressors such as Knirps work over distances of <100 bp to quench activators or basal promoters (1). In contrast, long-range repressors such as Hairy function over distances of 1 kb to silence their target genes in a process suggested to involve extensive spreading of a recruited corepressor, Groucho (2, 3).

Seminal work in this area by the Levine laboratory has prompted the suggestion that the functional differences in the range of action of the two classes of repressors reflect the recruiting of distinct corepressors (4). Short-range repressors such as Knirps, Kruppel, Giant, and Snail associate with the C-terminal binding protein (CtBP) corepressor, whereas Groucho is implicated in mediating long-range repression by Hairy (4, 5). Both evolutionarily conserved corepressors have been linked to chromatin-modifying enzymes, and each associates with sequence-specific DNA binding factors by means of short characteristic motifs in the partner (2, 6, 7). CtBP, a homolog of D2 hydroxy acid dehydrogenases, binds to NAD/NADH, as well as to histone deacetylases and demethylases. The corepressor is recruited by a wide range of transcriptional factors, including short-range repressors in Drosophila, as well as many proteins whose range of action is unknown (4, 7–9).

Groucho is homologous to vertebrate TLE and yeast TUP1 corepressors. The protein possesses a WD40 repeat that permits binding to transcription factors with a C-terminal WRPW motif, as found in Hairy, or alternatively, through a 7-aa “eh1” motif (2, 6, 10, 11). Groucho has been found to interact with the Rpd3 histone deacetylase as well as histones (12, 13). Two Drosophila proteins known to interact with Groucho are Hairy and Dorsal, which are well characterized long-range repressors (14, 15).

Current understanding of short-range repression comes from studies that defined CtBP-dependent and CtBP-independent activities of these proteins, as well as their action on endogenous and synthetic promoters (1, 4, 16–22). Little is known about the actual mechanisms through which the proteins carry out this function; however, our earlier study showed that Knirps is in a large complex (450 kDa) including CtBP and the histone deacetylase Rpd3 (23), indicating that additional components of the Knirps complex remain to be identified.

To gain a greater insight into the short-range repression mechanism and further elucidate the CtBP-independent activity of Knirps, we identified proteins physically interacting with Knirps expressed in the blastoderm embryo. Unexpectedly, Groucho was identified as a part of the Knirps complex. We demonstrate here physical and genetic interactions between Groucho and Knirps, indicating that this corepressor is key to the CtBP-independent activity of Knirps. We provide evidence that this interaction is important for correct expression of eve blastoderm stripes; thereby, establishing the significance of this interaction during Drosophila development.

Results

Identification of Groucho As a Component of the Knirps Complex.

We sought to identify constituents of the Knirps complex by expressing epitope-tagged Knirps in Drosophila embryos. Previously, we had verified that this Knirps protein is active in regulating bona fide targets of Knirps (22). Proteins from soluble extracts were first purified by metal affinity chromatography, and then by immunoprecipitation with antibody against the C-terminal Flag epitope. The immunoprecipitated sample was then analyzed by MS. In addition to Knirps and CtBP, we identified two peptide fragments corresponding to Groucho; an unanticipated finding, considering the previous association of this corepressor with long-range repressors (data not shown).

To validate the association of Knirps and Groucho, partially purified fractions from the metal affinity chormatography were subjected to DNA affinity purification using Knirps binding sites immobilized on Sepharose beads. Eluted samples were analyzed by Western blotting for Knirps and Groucho (Fig. 1A). Groucho was found to copurify with Knirps, but was not present in affinity-purified samples from uninduced embryo extracts lacking Flag-tagged Knirps. We also directly immunoprecipitated Knirps from crude extracts using antibodies specific for the Flag epitope, and using embryo extracts treated with Benzonase to remove nucleic acids. As shown in Fig. 1B, Groucho was immunoprecipitated only when Knirps was present, and this association does not require DNA.

Fig. 1.

Copurification of Knirps and Groucho. Extracts from 0- to 12-h embryos induced for Knirps expression with heat shock pretreatment (+) or uninduced (−). (A) Flag-tagged Knirps is present in crude lysate, NTA-Ni eluate, and DNA affinity eluate (Upper, lanes 1, 4, and 7). Groucho, present in crude extracts (Lower, lanes 1–3) is found in purified fractions containing Knirps (lanes 4 and 7). (B) Coimmunoprecipitation of Knirps and Groucho. Knirps was induced (+) or not (−), and the Benzonase-treated crude lysate was immunoprecipitated with anti-Flag antibody or IgG (negative control). Antibodies used for Western blotting are indicated.

Genetic Interaction of Groucho with Knirps and eve.

To determine whether the physical association observed above was physiologically relevant, we tested whether Groucho and Knirps interact genetically on an endogenous target gene. The pair-rule gene eve is expressed in a seven stripe blastoderm pattern that is a sensitive measure of Knirps activity. Knirps sets the internal expression boundaries of eve stripes 3,7 and 4,6 by binding to enhancers with different thresholds of repression sensitivity (22, 24). As was previously reported, embryos that are heterozygous for kni⁹ or kni^7G also exhibited defects in the eve pattern observed as fused stripes 4,6 or reduced stripe 5 expression (Fig. 2 B and C; Table 1) (5). Ablation of maternal gro has pleiotropic effects that preclude interpretation of the eve phenotype; therefore, we tested the effects of partial depletion of gro in an otherwise WT or kni mutant background. Depletion of maternal and zygotic gro alone had a measurable effect on eve expression. In a heterozygous background for gro, ≈9–10% of the embryos showed misregulation of eve in the presumptive abdominal region where kni is expressed (Fig. 2D and Table 1). This effect differed from that observed in the kni heterozygote in that fusion of stripes 4–6 or loss of stripe 5 was less frequently observed, rather, a weaker expression of stripe 5. However, the restricted location was consistent with a perturbation of kni function, as gro is expressed throughout the embryo (25). Combining the gro and kni mutations in a double heterozygous background resulted in a more severe disruption (30–46%) in the eve pattern, which includes a greater percentage of embryos showing significant loss of stripe 5 expression (Fig. 2E and Table 1). The kni mRNA expression pattern was not altered in gro mutants (Fig. 2F), suggesting that this effect on eve patterning is not due to altered kni expression. To determine whether Groucho might influence eve expression through changes in other gap genes, we examined the expression of hb, gt, Kr, tll, and hkb in both gro^E48 and kni⁹ heterozygous backgrounds, but we did not observe any discernable changes in their expression patterns (data not shown). These specific perturbations of the eve pattern in response to changes in gro and kni levels provide evidence that gro and kni are involved in the same genetic pathway.

Fig. 2.

Evidence for genetic interaction between kni and gro. (A) WT eve expression pattern. (B and C) Presumptive kni heterozygote showing eve stripe 4–6 fusion and reduced eve stripe 5. (D) Presumptive gro heterozygote showing reduced stripe 4,5. (E) Progeny from gro/kni transheterozygous cross showing absence of stripe 5. (F) The kni expression pattern in presumptive gro heterozygote. Embryos are oriented with dorsal side up, anterior to the left.

Table 1.

The eve expression in gro and kni heterozygous embryos

	Fused eve stripe 4,6 or weak 5,6; %	N
groE48×yw67	10	534
gro 1×yw67	9	215
kni9×yw67	19	211
kni7G×yw67	14	164
(maternal) groE48 × kni9	43	245
(paternal) groE48 × kni9	45	137
groE48×kni7G	30	156
gro1×kni9	38	72
gro1×kni7G	46	58

The eve expression patterns were scored in blastoderm embryos from crosses of heterozygous individuals carrying mutations in gro and/or kni. Heterozygosity at the gro locus for two different alleles had specific effects on eve patterning (9%). As observed previously, heterozygosity at the kni locus led to 14–19% of the embryos exhibiting a fused 4–6 or reduced/missing stripe 5 phenotype. The frequency of this phenotype increased >2-fold in a double heterozygous mutant background. This effect was regardless of whether the gro^E48 allele is maternally or paternally contributed to the double heterozygote. The gro^E48 is a null allele; gro¹ is a hypomorphic allele; kni⁹ is a null allele; kni^{7 g} is a loss of function mutation.

Evidence for Direct Physical Association Between Groucho and Knirps.

We next tested whether Knirps and Groucho proteins directly associate with one another. Groucho binds proteins that contain a short C-terminal tetrapeptide motif WRPW or an internal engrailed homology motif (FXIXXIL) (6, 10). A close inspection of the Knirps protein sequence revealed an eh1-like motif at the N terminus (residues 85–91; Fig. 3A). This motif is conserved in Drosophila as well as kni genes of Apis mellifera and Tribolium castaneum. In Drosophila, this portion of Knirps contains a CtBP-independent repression activity, and previous assays had delimited the CtBP-independent repression domains to residues 75–330 (17). Therefore, we expressed this domain as a GST fusion protein and tested it for the ability to interact with in vitro translated Groucho. As expected, the GST-Hairy protein bound Groucho effectively (Fig. 3B, lane 2), and GST-Hairy lacking the Groucho binding motif was much less effective (Fig. 3B, lane 3). Groucho did not interact with GST alone (Fig. 3B, lane 4) nor did any of the GST proteins interact with in vitro translated CSN4, an unrelated protein. The Knirps 75–330 retained a substantial amount of Groucho (Fig. 3B, lane 5), whereas in contrast, a mutant form of Knirps lacking the N terminus (139–330) displayed a weaker interaction (Fig. 3B, lane 7). To test whether the eh1 motif in particular is essential for this interaction, we made point mutations in four of the conserved residues (FXIXXLL), converting hydrophobic amino acids to alanine. This eh1^mut showed reproducibly weaker interaction with Groucho (Fig. 3B, lane 6). A further truncation (189–330) abolished this interaction (Fig. 3B, lane 8). Interestingly, the ability of these proteins to bind Groucho correlates directly with their in vivo activity as transcriptional repressors. Gal4-Kni 75–330 was highly active against an eve reporter, whereas Gal4-Kni 139–330 was less potent and Gal4-Kni 189–330 was inactive (17, 19) .

Fig. 3.

Physical association between Groucho and Knirps. (A) Alignment of residues 61–216 and 303–354 of insect Knirps proteins. Amino acid residues 85–91 (eh1-like motif) are conserved among these Knirps homologs. Boxed region between amino acids 187–194 indicates residues resembling an eh1-lke motif that is partially disrupted by the 169–189 deletion. CtBP binding motif PMDLS (331–335), which is conserved only within Drosophila, is also indicated. Asterisks indicate alanine substitutions made in the eh1-like motif. Deletions tested in GST pull-down assays are also indicated. (B) GST- Hairy and Groucho interaction assay. GST fusion proteins were bound to in vitro translated ³⁵S-met Groucho protein, and bound proteins were analyzed by SDS PAGE. The proteins showed differential binding ability to in vitro translated Groucho; strong binding was observed with Kni 75–330 (lane 5), whereas Kni eh1^mut (lane 6) and Kni 139–330 showed weaker interactions (lane 7). No interaction was observed with Kni 189–330 (lane 8). In the context of protein lacking the eh1 motif, deletion of residues 139–149 or 150–169 had little further effect on Groucho binding (lanes 9 and 10), whereas deletion of residues 169–189 or 139–189 strongly reduced Groucho binding (lanes 11 and 12). GST-Hairy and a mutant form of Hairy lacking the C-terminal WRPW motif serve as controls. (C) Coomassie stained gels showing equal amounts of GST fusion proteins used in binding assays.

To further characterize the residual Groucho binding activity observed in the Knirps mutants lacking the eh1 motif, we analyzed an additional series of internal Knirps deletions (Fig. 3A) in proteins in which the eh1-like motif was already removed. Those proteins with deletions between regions 139–149 and 150–169 were able to interact with Groucho at a level similar to that of the intact Kni 75–330 protein lacking the eh 1 motif (Fig. 3B, lanes 9 and 10). However, deletions of residues 169–189 abolished the association with Groucho, as did a larger deletion encompassing this region (139–189) (Fig. 3B, lanes 11 and 12). From these results, we conclude that Groucho can directly interact with Knirps, and that two regions in the N terminus of Knirps contribute to the interaction with Groucho.

Groucho Interaction Motifs Are Essential for Repression Activity in vivo.

An earlier study showed that ectopic expression of the CtBP-independent form of Knirps (Kni 1–330) represses eve; particularly, the 3,7 and 4,6 enhancers (22). To test whether the physical interaction we observed for Knirps and Groucho are functionally relevant, we created and assayed transgenic flies expressing the WT Kni 1–330 protein, as well as mutants lacking solely the eh1-like motif or both the eh1-like motif and the region between 169 and 189 (Kni 169–189^Δ). Expression of the Knirps proteins was induced by heat shock, and in situ hybridization was performed to monitor eve expression (Fig. 4). The expression levels of the various forms of the Knirps protein were quantitated by Western blotting. Under similar induction conditions, the WT and the Kni eh1^mut proteins were expressed at similar levels, whereas the Kni 169–189^Δ protein was expressed at a somewhat lower level (Fig. 4Bi).

Fig. 4.

In vivo assay of Groucho binding mutants. (A) Phenotypes produced by Knirps proteins with impaired Groucho binding activities. (Ai) Pattern of WT endogenous eve expression. (Aii) Embryos expressing Kni 1–330, with intact Groucho binding activity show repression of stripes 3,7 and partial loss of stripe 4. (Aiii) Embryos expressing the Kni eh1^mut, a protein that has reduced Groucho binding activity, show at most some weakening of stripes 3 and 4. (Aiv) Embryos expressing the Kni 169–189^Δ protein with no Groucho binding activity show little or no change to eve patterns. Both WT and mutant proteins were localized to nuclei in embryos (data not shown). A minority of embryos exhibit a slight reduction in stripe 3. (B) Expression of Knirps mutant proteins. (Bi) Western blotting of protein extracts from embryos expressing Kni 1–330 and Kni eh1^mut heat shocked for 20 min and embryos from Kni 169–189^Δ heat shocked for 25 min. (Bii) Kni 169–189^Δ embryos were also heat shocked for 35 min to equalize protein expression levels. M2 α Flag was used to detect the recombinant proteins and signals detected was quantified using a Fuji LAS 3000 imager and Fuji Multigauge image analysis software. Relative levels are indicated below the lane numbers.

As expected, the CtBP-independent Kni 1–330 protein potently effected repression of eve stripes 3,7 and 4,6; 88% of embryos showed this phenotype (Fig. 4Aii and Table 2). The eh1^mut was markedly less effective, repressing mainly eve stripe 3; this phenotype was observed in only 24% of embryos (Fig. 4Aiii and Table 2). The eve stripe 4 was affected to a lesser extent (6%), whereas stripe 6 was virtually unaffected. These observations suggest that the presence of eh1-like motif is crucial for Knirps activity. A further mutation that removed the second set of residues important for Groucho interaction (169–189^Δ) resulted in a protein that only weakly affected eve stripes 3 and 4 in a small percentage of the embryos (Fig. 4Aiv and Table 2). When induced for equal amounts of time, this protein was expressed at somewhat lower levels than the Kni 1–330 and Kni eh1^mut; therefore, we heat shocked the embryos for longer periods of time (35 min) to achieve higher levels of expression (Fig. 4Bii). Even after induction of higher levels of protein, only a small fraction of the embryos showed a reduction in eve stripes 3 and 4 (Table 2). Nontransgenic yw embryos heat shocked for similar periods of time did not show similar misreguation of the eve expression pattern (data not shown). Together, these results strongly suggest that interaction with Groucho is essential for the CtBP-independent repression activity of Knirps, and that this corepressor contributes significantly to the total repression potential of Knirps.

Table 2.

Percentage of transgenic embryos showing complete or partial loss of eve stripe after heat shock

	Heat shock duration
	20 min		25 min	35 min
eve stripe	1–330	Eh1mut	169–189Δ	169–189Δ
1	0	0	0	0
2	0	0	0	0
3	88	24	5	2.4
4	16	6.4	3	10.7
5	0	0	0	<1
6	16	3	0	0
7	88	3	1.7	1
N	441	514	423	540

Role of Groucho in Knirps Specific Repression in S2 Cells.

To study the importance of Groucho for Knirps mediated repression in another context, we used a Tet-Knirps chimeric repressor (Knirps amino acids 75–330) previously demonstrated to possess specific short-range repression activity in S2 cells (21). Luciferase activity was assayed in untreated cells or those depleted of Groucho by dsRNA treatment.

We were able to efficiently deplete cells of Groucho using dsRNA treatment (Fig. 5B). In the absence of RNAi, the Tet-Knirps repressor was capable of inhibiting reporter activity 2.5-fold, and this repression activity was significantly reduced on depletion of Groucho by RNAi, but not by control lacZ RNAi treatment (Fig. 5A). Also, depletion of Groucho did not have significant effect on the repression activity of an unrelated fusion protein, Tet-CtBP (data not shown; see ref. 21). We noted that the Tet-Knirps protein retained some activity in Groucho depleted cells. Removal of the Groucho interacting motifs reduced, but did not eliminate this residual activity (data not shown). Previous analysis of a Gal4-Knirps fusion protein suggested that residues 189–254 can mediate repression in S2 cells; this portion of Knirps is contained within our Tet-Knirps fusion protein and may explain the residual Groucho-independent activity, which may function through other, as yet unidentified factors (26).

Fig. 5.

Reduction of Knirps repression activity in Groucho depleted S2 cells. (A) Luciferase activity of reporter gene containing dual Tet binding sites at −75 bp in S2 cells transfected with reporter alone, reporter + Tet-Knirps 75–330 (CtBP-independent repression domain of Knirps), or Tet-Stop (contains only the Tet DNA binding domain). In parallel, cells were pretreated with dsGroucho or dslacZ RNA. Luciferase activity normalized to Renilla luciferase activity was expressed as fold change relative to appropriate controls. Error bars represent SD; n = 11 for Knirps activity in WT and Groucho-depleted cells, n = 3 for lacZ RNAi and other controls; ***, P < 0.001; **, P < 0.01 using a two tailed Mann–Whitney–Wilcoxon method. (B) Western blotting showing levels of Groucho and Tubulin in S2 cells treated as indicated. Proteins were assayed from samples of cells used in transfection assays.

Discussion

Groucho mediates the CtBP-independent repression activity of Knirps. The essential logic of Drosophila blastoderm transcription cascade is reliant on the short range of gap repressors proteins such as Knirps, Kruppel, and Giant acting on modular enhancers. Thus, the functional features of these repressors, which set them apart from long-range acting proteins such as Hairy, have been of special interest. Earlier studies suggested that the distinction between these classes of repressors may be attributed to differential recruitment of the CtBP corepressor to short-range repressor and Groucho to long-range repressors (2, 4, 10). The genetic and physical interactions of CtBP and Hairy were contradictory to this simple model, but further work has indicated that CtBP may not in fact serve as a Hairy corepressor, but as an antagonist of Groucho (27, 28). In another case, the Brinker transcription factor can interact with Groucho and CtBP in vitro, but appears to rely on Groucho for repression of many target genes, whereas CtBP has a minor role (29). Significantly, the range of Brinker repression has never been elucidated. Foreshadowing our study, Andrioli et al. (30) showed that Slp1 acts as a gap type regulator of pair-rule genes in the early embryo. The short-range nature of this regulation was apparent on eve, hairy, run, ftz, prd, and odd when the Slp1 protein was expressed in a ventral pattern. Consistent with a role for Groucho, a mutant form of the Slp1 protein lacking the eh1 motif was reported to be inactive, but this assay was complicated by the brief temporal window of repression. We show here that the well characterized short-range repressor Knirps physically and functionally interacts with Groucho, and this interaction is pivotal for the CtBP-independent repression potential of Knirps. These findings are definitely not consistent with the differential recruitment model of short- and long-range repression. Instead our results suggest an alternative model, that Groucho functions distinctly in the context of short- and long-range repression.

One possible explanation for the diverse function of Groucho may involve oligomerization. Recent studies have shown that Groucho and its homolog can form oligomeric structures that have been proposed to spread along DNA (2, 3, 31). Mutations that block Groucho oligomerization in vitro compromise the activity of this protein in vivo in the imaginal disc (2, 32). Thus, Groucho oligomerization has been assumed to be critical for its function and potentially related to the long-range activity of repressors such as Hairy. However, it seems likely that in the context of Knirps repressor complex, Groucho does not spread, because repression effects are clearly short range. Possibly the mode of recruitment dictates whether Groucho oligomerizes or not. We hypothesize that the distinct eh1-like repression motifs in Knirps interact with Groucho in a unique conformation to restrain Groucho from spreading and, thus, from mediating long-range repression. Crystal structures of the WD domain of human Groucho homolog TLE1 bound to either WRPW or eh1 peptide revealed that these peptides adopt different conformations on the corepressor binding surface (33). Such differences may affect the ability of Groucho to oligomerize. Other components in the Knirps corepressor complex may also control Groucho oligomerization. An “optional oligomerization” by Groucho model may explain earlier studies that found Hairy does not always cause dominant silencing of nearby enhancers (34). Also, hypomorphic alleles of Groucho have been identified that appear to compromise oligomerization but still retain some activity (35).

What role might Groucho have in Knirps-mediated repression? As shown previously, the CtBP-independent repression activity of Knirps is critical for full activity on some endogenous enhancers, underscoring the importance of Knirps-Groucho association (22). The histone deacetylase Rpd3 is recruited by Groucho, and is also a part of the Knirps repression complex. CtBP proteins are known to interact with histone deacetylases; thus, both CtBP and Groucho may recruit Rpd3 cooperatively (13, 36). The deacetylase activity may then augment Groucho-histone interactions, bringing about local modification of the chromatin, resulting in enhanced repressor output. Consistent with the cooperative recruitment of Rpd3, our purification of Knirps complexes indicated that Rpd3 associates preferentially with the full-length protein, and not the CtBP-independent domain alone (23). Therefore, in the context of Kni 1–330, Groucho may use another HDAC protein or rely on its HDAC-independent repression activity (13). The functional importance of this association may be to achieve quantitatively correct levels of Knirps activity, suggesting a similarity of function of these two corepressors. For example, in the context of the composite eve promoter, both of these activities can have roles in repressing enhancers of differential sensitivity.

In conclusion, our study provides compelling evidence that Groucho can mediate short-range repression; thus, the long- and short-range effects of transcriptional repression do not appear to be a simple function of differential recruitment of distinct corepressors. Not only does this change the perspective of Groucho, it changes the perspective of different repressor proteins. It appears that long-range repressors such as Hairy and short-range repressors such as Knirps may function as modulators of the repression range of common machinery.

Interestingly, Knirps protein sequences from different insect genomes indicate that Groucho binding by Knirps may be an ancestral trait, because the Groucho-binding eh1-like motif is present in Drosophila species as well as Tribolium and Apis (Fig. 3A). In contrast, the critical CtBP-interacting residues are present only in Drosophila, suggesting that the acquisition of an additional corepressor may be a derived trait, possibly as a part of the remodeling of embryonic gene circuitry associated with the unique syncytial environment.

Materials and Methods

Antibodies and Flies.

Flag M2 monoclonal antibody (Sigma F 3165) was used at 1:10,000 dilution. Rabbit polyclonal antiserum against fly CtBP (23) was generated against full-length CtBP and used at 1:10,000 dilution. Groucho monoclonal antibody was obtained from the Developmental Studies Hybridoma Bank and used at 1:20 dilution. Transgenic flies expressing double tagged (hexa-His at N terminus and flag at C terminus) Knirps (Kni 1–429 and Kni 1–330) were previously described (22). The Kni 1–330 with eh1 substitutions and deletions were generated by Quick Change mutagenesis (Stratagene) following the manufacturer's protocol.

Heat Shock.

To induce expression of recombinant full-length Knirps protein (1–429), 0- to 12-h transgenic embryos were collected on apple juice plates at room temperature, and were incubated at 37 °C for 30 min in a water bath. After induction, the embryos were dechorionated, weighed, and were either frozen at −80 °C until further use or were used to prepare lysate for subsequent purification steps. To induce expressions of Kni 75–330, Kni eh1 ^mut and Kni 169–189 ^Δ, 2- to 4-h embryos were collected on apple juice plates at room temperature, incubated for 20 min (Kni 75–330, Kni eh1 ^mut) and 25 or 35 min (Kni 169–189 ^Δ) at 38 °C in a water bath to ensure rapid and even heating. After induction, embryos were allowed to recover in a water bath at room temperature for 30 min before fixation for in situ hybridization experiments. In situ hybridization was performed using digoxigenin-UTP labeled antisense RNA probe to eve.

Purification of Knirps Complex.

Extracts were prepared from 0.5 g of heat shocked or nonheat shocked embryos after suspending and sonicating (four cycles, 12–15 pulses per cycle, output 4, duty cycle 60%, 1 min on ice between cycles) in 5 mL of lysis buffer containing 300 mM NaCl/50 mM Hepes, pH 7.9/10% glycerol/10 mM imidazole/10 mM β mercaptoethanol/1 mM PMSF/1 mM sodium metabisulfite/1 mM benzamidine/10 mM pepstatin A. Lysates were then cleared by centrifugation (20 min at 14,000 rpm, twice), and to it 0.5 mL of preequilibrated Ni-NTA beads (His select, Sigma 6611) were added. Incubations were carried out at 4 °C for 30–60 min, and the beads were washed several times in lysis buffer supplemented with 20 mM imidazole. The Ni-NTA bound His-tagged protein was then eluted twice by increasing the concentration of imidazole to 150 mM. The elutions were pooled and to them preequilibrated protein G coupled flag M2 antibody (200 μL of coupled beads to 500 μL of Ni elution) were added. Preequlibration was done in binding buffer containing 300 mM NaCl/50 mM Hepes, pH7.9/12.5 mM MgCl₂/10% glycerol/0.1 mM EDTA/2 mM DTT/1 mM sodium metabisulfite/1 mM benzamidine/1 mM PMSF. Incubations were then carried out at 4 °C for 3 h with end to end shaking. The beads were washed thrice with binding buffer and twice with the same buffer without MgCl₂. Protein elutions were carried out in binding buffer without MgCl₂, but supplemented with 0.2% sarkosyl or with 0.5 mg/mL of 3× flag peptide (Sigma F4799). The eluted proteins were subject to SDS/PAGE, digested in-gel with trypsin, chromatographed on C-18 peptide trap columns, and analyzed by MS using ThermoFisher LTQ Linear Ion trap mass spectrometer. Spectral assignments were validated using Scaffold software.

DNA Affinity Purification.

Oligonucleotides containing Knirps binding sites (5′ GCA TCT GAT CTA GTT TGT ACT CTG ATC TAG TTT 3′) were ligated and coupled to CNBr activated Sepharose beads following the protocol described previously (37). In brief, for affinity purification 250 μL of the beads were added to 500 μL of Ni elution, and the total binding volume was made up to 1 mL with buffer containing 25 mM Hepes/7% glycerol/0.2 mM ZnSO₄/0.2 mg BSA/4 mM DTT. Incubations were carried out for 3 h at room temperature, and the DNA bound Knirps was eluted using 700 mM NaCl. The salt was removed by dialysis, and the sample was then lyophilized and used for Western blotting after resuspending in SDS buffer.

Immunoprecipitation.

Embryos expressing Kni 1–429 were heat shocked for 30 min at 38 °C in a water bath, and the protein extract was prepared by sonication in lysis buffer [150 mM NaCl/50 mM Hepes, pH 7.9/10% glycerol and complete protease inhibitor Tablets (Roche)]. Knirps protein was immunoprecipitated using monoclonal M2 Flag antibody. The sample was separated on a 10% SDS PAGE gel and transferred to a PVDF membrane. The Groucho monoclonal antibody was used for immunoblotting and was detected with HRP coupled secondary antibody (Pierce).

GST Pull Down.

Appropriate expression plasmid was transformed in Escherichia coli BL21 and was grown at 37 °C for 3 h. The Knirps-GST fusion proteins were induced with IPTG for 17 h at 25 °C, whereas the Hairy fusion proteins (plasmids were a kind gift of G. Jiménez) were induced for 3 h at 30 °C. The proteins were purified on glutathione Sepharose beads (GE Healthcare). S³⁵methionine labeled Groucho was synthesized in vitro from pET Groucho (kind gift of G. Jiménez) using TNT-coupled reticulocyte lysate system (Promega). In vitro translated protein was bound to preincubated immobilized GST proteins, and the mixture was incubated for 1 h at 4 °C. Equal amounts of fusion proteins were used in each experiment. The beads were washed four times with 1 mL PBS/1 mM EDTA/0.2% Nonidet P-40. Bound protein was released by boiling in gel sample buffer and analyzed by SDS page and autoradiography.

Genetic Interaction Assay.

To test for a genetic interaction between kni and gro, transheterozygous flies for kni and gro were generated by crossing the gro heterozygous mutant females to the kni heterozygous mutant males and vice versa, and the expression pattern of eve was monitored by in situ hybridization. The kni⁹ (Bloomington stock 3332) carries a null mutation. kni^7G (Tübingen stock Z334) is a loss of function mutation. The gro E48 l(3)DE FRT^82B/TM6B, kindly provided by Z. Paroush, is a presumptive null. The gro¹ (Bloomington stock 511) carries a hypomorphic mutation. Heterozygous phenotypes for each kni and gro allele were noted after crossing balanced lines to yw ⁶⁷.

Cell Culture, RNAi, and Transient Transfections.

The luciferase reporter and the Tet repressor plasmids were generated as described previously in Ryu and Arnosti (21). For transient transfections, Drosophila S2 cells were grown at 24 °C in Schneider Drosophila medium (Gibco/BRL) containing 10% FBS and pencillin-streptomycin. For RNAi, gro cDNA was used to amplify the coding sequence using T7 tagged primers (38). The PCR product was then transcribed using the megascript kit (Ambien). The primers used to amplify the lacZ region from the C4 PLZ plasmid is as follows:

• Forward: 5′TTAATACGACTCACTATAGGGAGGCGTCGTTTAGAGCAGCAGAG 3′

• Reverse: 5′ TTAATACGACTCACTATAGGGAGTGGGATAGGTTACGTTGGTGT 3′.

S2 cells were incubated for 72 h with dsRNA (soaking) at a concentration of 15 μg/10⁶ cells for gro and lacZ as negative control. After this step, the cells were cotransfected with luciferase reporter alone (100 ng) or reporter (100 ng) and Tet-Kni 75–330 (1 ng) or Tet-Stop (1 ng) plasmid using Effectene (Qiagen) according to manufacturers instructions. Renilla luciferase plasmid (250 ng) was also cotransfected to normalize transfection efficiencies. Approximately 48 h later, cells were collected for Western blotting and luciferase activity measurements. The luciferase activity was measured by using the Dual glo Luciferase assay system (Promega). Immunoblotting was done to assess the efficiency of gro knock down.