The Acetyltransferase Activity of CBP Is Required for wingless Activation and H4 Acetylation in Drosophila melanogaster

Plasmid vectors and protein purification.Full-length dCBP cDNA (1) was subcloned into the pMT-V5/His(C) vector (Invitrogen) in frame with the 3′ V5 epitope tag sequence (the V5 epitope originated from paramyxovirus SV5). The dCBP-V5 gene was then subcloned into the pcDNA3 vector (Invitrogen) for expression in COS7 cells. The Q2146A, L2152A, L2158A, Y2160A, and F2161A mutations were introduced into the dCBP sequence with a 300-bp cassette containing the AT domain.

To purify a dCBP protein fragment containing the AT domain, a segment of the dCBP gene (coding amino acids [aa] 1853 to 2339) was subcloned into pET-28a (Novagen) in frame with the 3′ six-His tag sequence. We also introduced a two-Flag tag sequence at the 5′ end of the dCBP segment that resulted in increased protein stability. The Y2160A and F2161A mutations were again introduced into the sequence through a 300-bp cassette. The protein fragments were expressed in bacteria and purified by nickel-nitrilotriacetic acid affinity chromatography (QIAGEN).

HAT enzymatic analysis.HAT assays were performed as previously described (13). In brief, either full-length dCBP-V5 protein (the wild type [wt] or a protein with the F2161A or Y2160A mutation) immunoprecipitated from COS7 cells or 1 μg of a dCBP protein fragment (aa 1853 to 2339; with the wt or the F2161A or Y2160A sequence) was incubated for 1 h at 30°C in the presence of 0.5 μl of [³H]Ac-CoA (5.1 Ci/mmol), 1 μg of fly core histones, and buffer containing final concentrations of 50 mM Tris (pH 7.4), 0.2 mM EDTA, 8% polyethylene glycol 8000, 50 mM NaCl, and 1 mM dithiothreitol in a 30-μl reaction volume. The reaction mixture was applied to cellulose phosphate paper and washed, and the incorporated ³H was assessed by liquid scintillation spectroscopy or subjected to polyacrylamide gel electrophoresis and exposed to autoradiographic film.

Tissue culture, transfection, IP, and Western analysis.The tissue culture techniques employed have been described previously (7). Calcium phosphate transfections were carried out per the instructions of the manufacturer (Gibco BRL). Immunoprecipitation (IP) and Western analysis were performed as previously described (6, 7). V5 antibody was purchased from Invitrogen and used at a 1:5,000 dilution for both IP and Western analysis.

Ac-CoA binding.[³H]Ac-CoA binding assays were performed by incubating a 0.5 μM dCBP fragment (aa 1853 to 2339; with the wt, F2161A, or Y2160A sequence) in the presence of various concentrations of free [³H]Ac-CoA (8 to 512 μM) in 600 μl of the above-mentioned buffer for 60 min at 4°C. The dCBP protein fragment (55 kDa) was then pelleted by centrifugation (28) at 100,000 rpm for 4 h with a Beckman TL-100 ultracentrifuge, washed quickly with cold binding buffer, and dissolved in glacial acetic acid. Control experiments demonstrated that insignificant amounts of [³H]Ac-CoA are pelleted in 4 h. Bound and free [³H]Ac-CoAs were assessed in a liquid scintillation counter. The reaction mixtures also contained a trace amount of [³⁵S]sulfate for assessment of the amount of free fluid trapped in the pellet. Nonspecific counts were determined by simultaneously performing parallel reactions in the presence of excess cold Ac-CoA.

Fly strains, germ line transformation, and whole-mount embryo in situ hybridization and immunostaining.Full-length dCBP-V5 (wt, F2161A, and Y2160A) constructs were cloned into the pUAST vector (4), and transgenic fly lines were generated as described previously (26, 30). At least three independent transformants were generated for each dCBP-V5 (wt, F2161A, or Y2160A) construct. G. Rubin (University of California, Berkeley) kindly provided the GMR-driven Gal4 line. This construct expresses GAL4 under the control of multimerized glass-binding sites, and its gene is referred to as GMR in this paper (33). For overexpression of dCBP in the eye, dCBP-V5 (wt, F2161A, and Y2160A) females were crossed to GMR/CyO males and the eye phenotype of the non-CyO progeny was assessed. Mating the GMR/CyO flies to the yw injection stock generated GMR/+ controls.

The paired (prd)-GAL4 line (38) was kindly provided by D. Kalderon, Columbia University. The dCBP mutation used is null on the basis of RNA and protein expression (1) and has been designated nej³ . For generation of dCBP-null embryos that express the different dCBP constructs in the prd domain, nej³ /Fm7c ftz-LacZ; dCBP-V5 (wt, F2161A, and Y2160A) females were crossed to Fm7c ftz-LacZ; prd-GAL4/TM3 males and the progeny was stained for β-galactosidase, the V5 epitope, and wg transcripts. To visualize the expression of even-skipped (eve) and the levels of acetylation of histone H4 in the nej³ and heterozygous control embryos, nej³ /Fm7c ftz-LacZ females were mated to Fm7c ftz-LacZ males and the progeny was stained for both β-galactosidase and acetylated histone H4 or eve transcripts. The Fm7c ftz-LacZ chromosome was obtained from the Bloomington Stock Center, and the TM3 and CyO chromosomes have been described by Lindsley and Zimm (20).

The dual immunohistochemistry and in situ hybridization of whole-mount embryos were performed as previously described (6). The anti-acetylated H4 (K5, K8, and K12) antibodies were purchased from Upstate Biotechnology Company and were used at a dilution of 1:2,000. Chicken anti-dCBP antibody (3) was used at a 1:800 dilution. Anti-β-galactosidase antibody was used at a dilution of 1:1,000. Larval brains with eye antennal disks attached were dissected from third-instar larvae and subjected to immunostaining with the anti-V5 antibody (Invitrogen) at a 1:5,000 dilution. After being stained, eye antennal disks were further dissected away from larval brains and mounted for light microscopy. Nomarski optics and digital capture were used for the light microscope pictures, and a confocal laser scan microscope (Bio-Rad 1024ES laser and Nikon Eclipse TE300 microscope) was used for confocal microscopy.

Histology and SEM.Histological sections were prepared as described previously (18). In brief, the heads were halved, fixed in 75 mM cacodylate (pH 7.4)-2% paraformaldehyde-2% glutaraldehyde for 4 h at room temperature, and subsequently stained in 1% tannic acid-75 mM cacodylate (pH 7.4) at 4°C for 16 h. After they were washed (in 0.1 M cacodylate), the heads were postfixed (2% OsO₄-0.1 M cacodylate) for 2 h, dehydrated, mounted, and sectioned. To prepare them for scanning electron microscopy (SEM), the flies were treated with ether fumes, coated with SEM mount adhesive, and sputter coated with 20-nm-diameter gold particles. An Amray SEM was used for visualization and imaging.

Generation of Ac-CoA binding mutations in dCBP.To develop a system with which to study the role of the intrinsic AT activity of dCBP in vivo, we generated point mutations in dCBP which were targeted to a region of putative Ac-CoA contact points (Fig. 1a). These mutations were designed to reduce or eliminate AT activity without affecting the overall structure of dCBP. AT activity is known to exist within highly homologous regions of p300 (aa 1195 to 1810) (13) and mCBP (aa 1099 to 1877) (22), and point mutations in this region have been shown to reduce AT activity in vitro (22). Some of these mCBP mutations align in a best-fit model with the known Ac-CoA contact points of Saccharomyces cerevisiae (yeast) HAT1 (yHAT1), as determined from its crystal structure (10, 23) (Fig. 1a). Based on these observations, we generated single alanine substitutions in five sites in dCBP (Q2146A, L2152A, L2158A, Y2160A, and F2161A).

• Open in new tab
• Download powerpoint

FIG. 1.

Point mutations in the AT domain of dCBP specifically reduce AT activity and Ac-CoA binding. (a) Alignment of AT domain motifs A and B of yHAT1, p/CAF, mCBP, and dCBP. The black circles on top indicate the Ac-CoA contact points of yHAT1 (10). The gray circles at the bottom indicate mutated sites in dCBP. (b) HAT assay of various full-length dCBP-V5 constructs after IP from COS7 cells. The mutant and wt dCBP proteins were tagged at the C terminus with the V5 epitope, and IP experiments were performed with V5 antibodies. (c) HAT assay of purified 55-kDa (aa 1232 to 1711) wt, Y2160A, and F2161A dCBP protein fragments. (d) Saturation isotherm of [³H]Ac-CoA binding to wt dCBP protein (prot) fragment. Lines represent total binding (solid), binding in the presence of excess unlabeled Ac-CoA (dotted), and binding to heat-denatured dCBP (dashed). (e) Ac-CoA binding affinities for wt, Y2160A, and F2161A dCBP protein fragments. NSB, no specific binding.

To test the effects of the five point mutations on AT activity, full-length dCBP that was tagged with a C-terminal V5 epitope was transiently transfected into COS7 cells, immunoprecipitated with a V5 antibody, and subjected to AT assays with fly core histones as the substrate (13). As shown in Fig. 1b, dCBP (F2161A) lost all measurable AT activity and the AT activity of dCBP (Y2160A) was reduced by 44% compared to that of the wt. Western blot analysis of the immunoprecipitates confirmed the presence of equal amounts of each full-length dCBP-V5 protein subjected to AT activity analysis.

To confirm that the reductions in the AT activities in the F2161A and Y2160A mutant proteins were due to the loss of intrinsic AT activity rather than to a disruption of dCBP interactions with other AT-containing factors, we expressed 55-kDa fragments (aa 1853 to 2339) of dCBP (with the wt, F2161A, and Y2160A sequences) in bacteria and subjected the purified proteins to AT activity analysis. As shown in Fig. 1c, the AT activities of the dCBP protein fragments containing the F2161A and Y2160A mutations were reduced by 94 and 32%, respectively, compared to that of the wt. All specific activity in this experiment was due to the incorporation of [³H]Ac into histones and not to autoacetylation of the dCBP fragment. This conclusion was evidenced by the lack of acetylation in control HAT reactions in which the histone substrate was omitted and was further confirmed by visualization of ³H-labeled HAT assay products on a polyacrylamide gel (only histones and not the 55-kDa dCBP fragment were acetylated [data not shown]). These findings confirm that the reduction in AT activity of the F2161A and Y2160A mutant proteins was due to the loss of intrinsic AT activity in dCBP.

The 55-kDa dCBP protein fragments (with the wt, F2161A, and Y2160A sequences) were tested for [³H]Ac-CoA binding by using a high-speed sedimentation binding assay. In the presence of increasing concentrations of [³H]Ac-CoA, saturable binding to wt dCBP was obtained (Fig. 1d). To determine specific binding to the dCBP fragments, the assay was performed in the presence of either excess unlabeled Ac-CoA or heat-denatured dCBP; only low levels of nonspecific binding were observed under these conditions. As shown in Fig. 1e, the F2161A mutant protein lacks the ability to specifically bind Ac-CoA, suggesting that the mutation disrupts an Ac-CoA contact point. The Y2160A mutant dCBP only modestly reduced Ac-CoA affinity (K_d = 140 μM compared to 50 μM for wt dCBP), suggesting that its loss in AT activity was due, at least in part, to a reduction in Ac-CoA binding. To ensure that the mutations did not affect the overall protein structure, we determined whether the mutant dCBP proteins could bind the adenoviral transforming factor E1A, which binds the region abutting the AT domain of CBP (3, 16) and is acetylated by CBP (39). Full-length (wt, F2161A, and Y2160A) dCBP-V5 proteins were expressed in COS7 cells and subjected to glutathione S-transferase-E1A pull-down analysis. Equal levels of binding of the wt and mutant proteins to E1A were detected (data not shown). A similar series of experiments confirmed that the changes in the AT activities of the various (wt, F2161A, and Y2160A) dCBP constructs were not due to a disruption of dCBP with glutathione S-transferase-p/CAF, an AT-containing factor which interacts with CBP (37) (data not shown).

Characterization of dCBP Ac-CoA-binding mutants in vivo: dCBP AT activity is required for a dCBP overexpression phenotype.To test the function of the dCBP mutants in vivo, we cloned full-length wt, F2161A, and Y2160A dCBP-V5 into the pUAST vector (4) and generated transgenic fly lines as previously described (26, 30). This construct places dCBP under GAL4-UAS control. We overexpressed the dCBP constructs in the eye antennal disks of flies containing the GMR enhancer-GAL4 construct. Scanning electron micrographs of the resultant fly eyes are shown in Fig. 2. GMR/+ control eyes have a virtually normal phenotype of regularly arrayed ommatidia and mechanosensory bristles. The eyes from GMR/+; UAS-dCBP-V5/+ flies, in which wt dCBP is overexpressed, have a smooth-eye phenotype, are morphologically atrophic, and completely lack ommatidia and mechanosensory bristles. In contrast, eyes from GMR/+; UAS-dCBP(F2161A)-V5/+ flies, in which the F2161A mutant dCBP is overexpressed, exhibit irregularly arrayed ommatidia and mechanosensory bristles. GMR/+; UAS-dCBP(Y2160A)-V5/+ eyes, in which the Y2160A mutant dCBP is overexpressed, have a phenotype intermediate between the wt and F2161A mutant dCBP overexpression phenotypes.

• Open in new tab
• Download powerpoint

FIG. 2.

The AT activity of dCBP is necessary to maintain the smooth-eye phenotype resulting from dCBP overexpression in adult fly eyes. Scanning electron micrographs of adult Drosophila eyes are shown. Low and high magnifications are shown in the left and right panels, respectively. (i) GMR (GMR enhancer-GAL4)/+ eye in which dCBP is not overexpressed; (ii) GMR/+; UAS-dCBP(wt)-V5/+ eye; (iii) GMR/+; UAS-dCBP(F2161A)-V5/+ eye; (iv) GMR/+; UAS-dCBP(Y2160A)-V5/+ eye.

Histological samples of these four fly lines were prepared to determine the cellular structures of the eyes overexpressing the wt and mutant dCBP proteins (Fig. 3a). Again, the eyes from GMR/+ adults are normal, with ommatidial arrays containing ordered photoreceptors. In the eyes from GMR/+; UAS-dCBP-V5/+ adults, which overexpress wt dCBP, this architecture is completely lost and the photoreceptors are no longer detected. By contrast, the eyes from GMR/+; UAS-dCBP(F2161A)-V5/+ adults, which overexpress the F2161A mutant dCBP, have intact ommatidial arrays, though sometimes an ommatidium may contain an abnormal number of photoreceptors. Again, overexpression of the Y2160A mutant causes an intermediate phenotype. Immunostaining (V5 antibody) and Western analysis (V5 antibody) of the expression levels of each dCBP-V5 construct (with the wt, F2161A, or Y2160A sequence) in the eye antennal disks of third-instar larvae (Fig. 3b and c, respectively) were performed to confirm equal levels of dCBP-V5 overexpression for each transformant. The overexpression of the three proteins does not cause any defect in the development of the eye antennal disk (Fig. 3b). When we stained the disks with the neuron-specific antibody 24B10 (11), we found that the ommatidia and their projections onto the brain were wt (data not shown). Thus, the overexpression of wt dCBP in the eye causes neural degeneration sometime during differentiation. Together, these findings demonstrate that the intrinsic AT domain of dCBP is necessary for the dCBP overexpression activity that leads to neural degeneration.

• Open in new tab
• Download powerpoint

FIG. 3.

The AT activity of dCBP is necessary to maintain dCBP overexpression. (a) Histology of adult Drosophila eyes. (i) GMR/+ eye; (ii) GMR/+; UAS-dCBP(wt)-V5/+ eye; (iii) GMR/+; UAS-dCBP(F2161A)-V5/+ eye; (iv) GMR/+; UAS-dCBP(Y2160A)-V5/+ eye. (b) Anti-V5 antibody immunostaining of eye disks from third-instar larvae overexpressing different full-length (wt, F2161A, and Y2160A) dCBP-V5 constructs under the control of the GMR-GAL4 construct. The stripe of staining in each panel represents a wt, Y2160A, or F2161A dCBP-V5 construct expressed in the glass expression pattern under the control of the GMR enhancer-GAL4 construct in the developing eye. (c) Western blot analysis (V5 antibody) of eye disks from third-instar larvae (each lane contains protein from 20 eye disks).

dCBP is expressed throughout embryogenesis in wt embryos but not in embryos homozygous for the nej ³ (mutant dCBP) allele.After demonstrating that dCBP AT activity is required for the dCBP overexpression phenotype in the adult eye of Drosophila, we chose to study the role of dCBP AT activity in the expression of endogenous developmentally regulated genes in dCBP-null embryos. For this system, we chose embryos homozygous for the nej ³ allele. This allele is a dCBP molecular null, and nej ³ embryos have no zygotic dCBP expression. Thus, in these embryos the only endogenous source of dCBP is maternal, and dCBP is significantly depleted by stage 9 of embryogenesis. The nej ³ embryos that die at this stage often have a characteristic twisted phenotype. To confirm the absence of dCBP protein in nej ³ homozygotic embryos (from either zygotic or maternal contribution), we used a chicken polyclonal antibody to dCBP (3) and measured the expression of dCBP protein at various stages of embryogenesis in wt and nej ³ embryos. As shown in Fig. 4A and C to E, wt embryos at various stages of embryogenesis express dCBP. This immunostaining is specific to dCBP, since preincubation of the dCBP antibody with antigen reduces staining to background levels (Fig. 4B). In contrast to wt embryos, those homozygous for the nej ³ allele show no specific dCBP binding by stage 11 (Fig. 4F) and are usually not dCBP positive at stage 9.

• Open in new tab
• Download powerpoint

FIG. 4.

Chicken anti-dCBP antibody stains Drosophila embryos throughout embryogenesis. Shown are wt embryos at stages 4 to 5 (a), stage 6 (c), late stage 8 (d), and stage 12 (e) of embryogenesis stained with chicken anti-dCBP antibody; a wt stage 4 embryo stained with the anti-dCBP antibody that had been preincubated with antigen (b); and a dCBP-null (nej ³) embryo (with the twisted phenotype) at about stage 11 stained with anti-dCBP antibody (f).

dCBP AT activity is required for activation of wg expression.The transcription factor cubitus interruptus requires dCBP to activate wg in response to hedgehog signaling (8), and embryos homozygous for nej ³ are no longer able to maintain wg gene expression by stages 9 to 11 (1). To examine the importance of dCBP AT activity for the expression of an endogenous developmentally regulated promoter, we examined the levels of wg gene expression in nej ³ /Y; prd-GAL4/+ embryos (38) in which the three dCBP-V5 constructs were overexpressed in the alternate embryonic segments that express the pair-rule gene prd. As shown in Fig. 5, immunostaining with V5 antibody alone confirmed equal levels of dCBP-V5 overexpression for all constructs (wt, F2161A, and Y2160A). In segments expressing dCBP-V5, the wg gene is expressed in a distribution and intensity nearly identical to those seen in the wt embryo (8). By contrast, segments expressing F2161A dCBP-V5 do not stimulate wg expression. Segments that express Y2160A dCBP-V5 stimulate wg expression at a level that is intermediate between those of the wt and F2161A dCBP-V5 constructs. These results indicate that in vivo, dCBP AT activity is necessary for the induction of wg expression.

• Open in new tab
• Download powerpoint

FIG. 5.

The AT activity of dCBP is required for dCBP-induced wg gene expression in dCBP-null Drosophila embryos. nej ³/Y; UAS-dCBP(wt, F2161A, and Y2160A)-V5/+; prd-GAL4/+ embryos (stage 9) were immunostained with V5 antibody alone (left panels) or double stained with both V5 antibody and wg antisense RNA probe (right panels). In each panel, dCBP -V5 staining is apparent in alternate segments that express prd, but the V5 antibody staining in the right panels was intentionally left faint so as not to obscure the levels of wg staining in the segments overexpressing the exogenous dCBP proteins. The right panels show magnifications of embryos that show only three segments each. Segments of dCBP-null embryos overexpressing wt dCBP in the prd domains (arrow, top right panel) have significant wg expression, while the dCBP-null segment (between the prd segments) has minimal wg expression. Segments of dCBP-null embryos overexpressing the F2161A mutant dCBP protein (arrow, middle right panel) do not activate wg expression (all three segments have low levels of wg expressed). Segments of dCBP-null embryos overexpressing the Y2160A mutant protein (arrow, bottom right panel) have wg expression that is intermediate between the wt and F2161A mutant dCBP phenotypes.

dCBP AT activity is not required for the activation of eve expression.To determine whether dCBP AT activity at the wg gene has a direct, rather than a secondary, effect on alterations in global acetylation, we compared the levels of eve expression in wt and dCBP-null embryos. eve does not require dCBP for gene activation (S. M. Smolik, unpublished observation). As shown in Fig. 6a, eve mRNA is expressed in a normal alternating segment pattern in wt embryos. Similarly, as shown in Fig. 6b, eve is also expressed in a wt pattern and intensity in dCBP-null embryos.

• Open in new tab
• Download powerpoint

FIG. 6.

The AT activity of dCBP is not required for expression of eve in dCBP-null Drosophila embryos. The embryos were double stained with both the anti-β-galactosidase antibody and eve antisense RNA. (a) wt embryo (stage 9). The positive β-galactosidase staining (brown) indicates the presence of the Fm7c balancer chromosome that contains the β-galactosidase and wt dCBP genes. eve staining (blue) is present in a normal alternating segment pattern. (b) dCBP-null embryo (stage 9). The absence of β-galactosidase staining indicates that the embryo does not contain the Fm7c balancer chromosome and is therefore dCBP null (nej ³ /Y). In the absence of dCBP, eve staining (blue) is still present in a normal alternating segment pattern and at a wt level.

Acetylation levels of histone H4 lysine 5, 8, and 12 are differentially regulated during embryogenesis.In order to study the role of the intrinsic AT domain of dCBP on histone acetylation, we first measured various anti-acetylated histone H4 lysine-specific antibodies in wt and dCBP-null embryos (Fig. 7). Anti-acetylated H4 (K8) staining in wt and dCBP-null embryos revealed that acetylation of H4 (K8) is ubiquitous on chromatin and requires the presence of dCBP for normal expression levels (Fig. 7a and b and 8a). The levels of H4 (K8) acetylation were constant throughout embryogenesis (data not shown). By contrast, anti-acetylated H4 (K12) staining in wt (Fig. 7c) embryos revealed that the distribution of acetylated K12 changes differentially in the germ layers during development; the ectodermal nuclei exhibited a strong punctate staining while the mesodermal nuclei exhibited a weaker, more uniform staining. This staining also requires the presence of dCBP (Fig. 7d). Anti-acetylated H4 (K5) staining was most developmentally regulated. wt embryos at the syncytial and cellular blastoderm stages exhibited anti-acetylated H4 (K5) staining in a nuclear distribution (Fig. 7e). By the time the ventral furrow was formed, acetylated K5 of H4 was no longer detectable (Fig. 7f). Furthermore, staining with the anti-acetylated K5 antibody was not detected in nuclei during the remainder of embryogenesis. Because we did not have markers that would allow us to detect dCBP-null embryos at this early stage, we could not determine whether dCBP affects the early acetylation of H4 (K5). In Drosophila, acetylated lysine 16 of histone H4 is found only on the hyperactive X chromosome of the male and not the autosomes of either the male or female (35). Because these experiments could not distinguish control from mutant dCBP embryos, we did not assess the effect of dCBP on the acetylation of histone H4 (K16).

• Open in new tab
• Download powerpoint

FIG. 7.

Anti-acetylated histone antibodies (K5, K8, and K12) have specific patterns of expression during embryogenesis in Drosophila. The red stripes identify β-galactosidase expression from the FM7c chromosome. The green staining identifies the different acetylated lysines of histone H4. Shown are wt (a) and dCBP-null (b) embryos (stage 11 and 9, respectively) stained with anti-acetylated H4 (K8) antibody; wt (c) and dCBP-null (d) embryos (stage 9) stained with anti-acetylated H4 (K12) antibody; and syncytial blastoderm (e) and early-gastrulation (f) wt embryos stained with anti-acetylated H4 (K5) antibody.

• Open in new tab
• Download powerpoint

FIG. 8.

The AT activity of dCBP is required for dCBP-induced histone H4 (K8) acetylation in dCBP-null Drosophila embryos. The embryos were immunostained with anti-V5 antibody (Texas Red secondary antibody) and acetylated-H4 (K8) antibody (fluorescein secondary antibody) and analyzed by laser confocal microscopy. All embryos shown were from the same experimental population and were visualized with the same fluorescence settings to allow for the comparison of relative expression levels. (a) wt (left) and dCBP-null (right) embryos labeled with anti-acetylated H4 (K8) antibody [αAcH4(K8)]. (b) nej ³/Y; UAS-dCBP(wt, F2161A, and Y2160A)-V5/+; prd-GAL4/+ embryos. The red stripes show staining for the expressed dCBP-V5 constructs (left), and the green stripes show staining for acetylated histone H4 (K8) (right). In each panel, a relatively small section (containing only a few segments) on the surface of the whole embryo was visualized.

Mutations in the dCBP AT domain affect acetylation levels of histone H4 lysine 8.Because lysine 8 of H4 is the only one of the four H4 acetylation targets that has consistent levels and patterns of acetylation throughout embryogenesis (Fig. 7), we chose to monitor it in our experiments on the effect of dCBP AT mutations on histone acetylation. wt embryos labeled with anti-acetylated H4 (K8) antibodies revealed global histone H4 (K8) acetylation in the nuclei of developing embryos. By contrast, nej ³/Y embryos from the same population as that of the wt embryos had significantly reduced levels of staining with the anti-acetylated H4 (K8) antibodies. To determine whether the intrinsic AT domain of dCBP is necessary to maintain wt levels of histone H4 (K8) acetylation, nej ³/Y; prd-GAL4/+ embryos expressing the wt, F2161A, and Y2160A dCBP-V5 constructs were double-labeled with antibodies to the V5 (red) and acetylated-H4 (K8-specific; green) epitopes (Fig. 8b). The levels of overexpression for the wt, F2161A, and Y2160A dCBP-V5 constructs were equal. In those segments overexpressing wt dCBP, H4 (K8) acetylation was higher. By contrast, segments expressing F2161A dCBP-V5 did not show augmented H4 (K8) acetylation. The segments expressing Y2160A dCBP-V5 affected H4 (K8) acetylation at levels intermediate between those of the wt and F2161A dCBP-V5 constructs. Although we have not ruled out the possibility that the intrinsic AT domain of dCBP is required for the synthesis of H4, we find it unlikely because basal levels of histone H4 (K8) acetylation are visualized in the nuclei of the dCBP-null embryo (Fig. 8a) and dCBP-null segments (Fig. 8b). These findings suggest that the intrinsic AT activity of CBP is necessary for wt levels of histone H4 (K8) acetylation throughout the developing organism.