brd2b transcript variants are differentially regulated during development. (A) Schematic of predicted proteins from the brd2b locus, full-length Brd2b-L and truncated Brd2b-S. Bromodomains BD1 and BD2 (ovals), nuclear localization signal (NLS, triangle), and a C-terminal extraterminal (ET) domain (rectangle). (B) Schematic of the long (brd2b-L) and short (brd2b-S) transcript variants encoded by the brd2b locus, with numbered exons. UTRs of 5′ and 3′ are shown in white; protein-coding regions are shown in orange. The short transcript retains part of intron 5 (shown in yellow), which introduces a premature stop codon (darker yellow shows part of intron that is translated). Primer pairs flanking exon–exon junctions and used in RT-PCR are shown as dark blue arrows. (C) RT-PCR products produced from mRNA of staged oocytes (top gels) and 4, 24, and 48 hpf embryos (bottom gels), using various primer pairs, as shown in B. Lanes labeled with the same number display the PCR product obtained using the same primer pairs, as shown in the key: (1) 1 kb ladder (NEB); (2) ex5/in5 primers (brd2b-S 250 bp predicted product); (3) ex5/6 primers (brd2b-L 400 bp predicted product); (4) ex9/10 primers (brd2b-L 300 bp predicted product); (5) ex11/12 (brd2b-L 404 bp predicted product); (6) ex1/in5 primers (full-length brd2b-S 1447 bp predicted product); (7) ex1/ex12 primers (full-length brd2b-L 4037 bp predicted product); (8) primers for β-actin (+) control (322 bp predicted product); (9) primers internal to intron 1 (not shown in B), as (−) control for genomic DNA (784 bp predicted product). Maternal brd2b-L predicted PCR products are detected in all stages of ooctye tested (blue arrowheads, lanes labeled 3,4,5, top gels), but brd2b-S predicted products are not present (lanes labeled 2, top gels). The bands observed in stage 1, 2 and 3 oocytes in lanes labeled 2 are either smaller than the 250 bp predicted product for brd2b-S (s1, s2 gels), or are very faint products among several faint products (s3 gel), so are likely a spurious. Zygotic brd2b-S predicted product is detected in 24 hpf and 48 hpf embryos (red asterisk, lanes labeled 2, bottom gels), but not in 4 hpf embryos, while zygotic brd2b-L is detected in all stages of embryos tested (blue arrowheads, lanes labeled 3,4,5, bottom gels).

Anti-Brd2b peptide antibody detects major protein products in zebrafish embryos and oocytes, that are reduced in brd2bMO morphants. (A,B) Western blot (A) of total ovary (ov) and staged embryos (4 h, 24 h, 48 h) probed with anti-Brd2b peptide antibody. Blot was prestained (B) to assess total protein and relative loading across lanes. Total ovary samples often show a large amount of high molecular weight staining corresponding to yolk proteins that make up predominant products in this tissue; nevertheless, these do not interfere with the detection of Brd2b. A product over 150 kD is consistently detected in ovaries and 48 hpf embryos and sometimes faintly in 4 and 24 hpf embryos (top arrowheads in A); a product between 50 and 75 kD is also consistently detected in 48 hpf embryos (bottom arrowhead) and sometimes also in 4 hpf and 24 hpf embryos (see Supplementary Figure S1). (C,D) Duplicate blot of A probed with anti-Brd2b peptide antibody in the presence of peptide antigen for peptide competition assay (C). The blot was prestained (D) to assess total protein and relative loading across lanes before antibody + peptide incubation. Peptide competition prevents all antibody signals on blot (C), supporting specificity of antibody for Brd2b epitope. Prestaining of blot using Memcode protein stain (D) assures that lack of anti-Brd2b signal in C is due to peptide competition rather than lack of protein in lanes. (E) Immunohistochemical analysis of staged oocytes (I–IV) and higher magnification stage IV oocyte (IV) using anti-Brd2b peptide antibodies, and showing eventual localization to the micropyle and future animal pole. (F,G) Western blot to assess efficacy of brd2bMO morpholino knockdown. Twenty-four hours post-fertilization embryo treatment groups: uninjected (wt), control 5-base mismatch morpholino-injected (mis5), and brd2bMO1-injected (2bMO), probed with anti-Brd2b peptide antibody (F). The blot was prestained (G) to assess total loading and relative loading across lanes. When relative loading is taken into account, there is a 3-fold reduction in Brd2b (~150 kD band) in brd2bMO-treated embryos (2ng total) as measured by ImageJ (FiJI) using band pixel densities (see Methods). Anti-Brd2B peptide antibody used in all immunoblot experiments in this study was raised against an antigenic epitope between NLS and ET (see Materials and Methods), and detects only the long isoform (Brd2b-L). Bottom edges of Western blots are indicated by horizontal line in (A, C and F). mw = Precision plus dual color molecular weight standard (BioRad); ch = Precision plus Western Chemiluminescent marker (BioRad).

Brd2b knockdown results in reduced hindbrain, ill-defined MHB region, and trunk abnormalities similar to Brd2a morphants, but presents unique circulatory and pronephric defects. Brightfield images of 24 hpf prim 5 embryos (A–F), and larvae at 6 days (H–J). Lateral views of head and trunk of representative embryos from indicated treatment groups: (A) uninjected; (B) brd2bMO1-injected; (C) Brd2bMO1 + HsBrd2RNA co-injected; (D) Brd2MO2-injected; (E) Crispr-Cas9-brd2b-disrupted; and for comparison, (F) paralog brd2aMO-injected. In A) mhb = midbrain-hindbrain; pd = pronephric duct; c = cloaca; pbi = peripheral blood island. Head and trunk defects in morphants and Crispr-Cas9-treated embryos are indicated by black arrows, and include reduced brain and ill-defined MHB region, misformed pronephric duct, and disorganized PBI. These defects are substantially rescued by co-injection of human Brd2 RNA, recapitulated by a non-overlapping anti-brd2b morpholino (MO2), and phenocopied by Crispr-Cas9 disruption, showing gene-specificity of observed effects. (G) Crispr-Cas9 brd2b target validation, showing intact PCR products from wildtype (lane 1, wt) and Crispr-disrupted (lane 2, 2b) brd2b locus; wildtype PCR product treated with mismatch-detecting resolvase (lane 4, wt+) and untreated (lane5, wt−); Crispr-disrupted PCR product treated with resolvase (lane 6, 2b+) and untreated (lane 7, 2b−). Crispr-disrupted brd2b locus treated with resolvase shows cleavage products (asterisk highlighting lane 6), while wildtype locus remains intact, indicating that Crispr-Cas9 targeted designated sequences in the brd2b locus. Pixel density of the 600 bp intact product band is reduced from 980 in untreated (lane 7) to 528 in treated (lane 6) Crispr-disrupted embryos, indicating close to 50% efficacy (see Methods). (H) Six day old morphant larvae injected at 2 cell stage with 1 ng brd2bMO1, or (I,J) 2ng brd2bMO1, illustrating dosage-dependent severe heart edema and trunk deformities, due to nearly complete lack of blood circulation. See Table 1 for population morphology data from these studies. See Supplementary Figure S1 for HsBrd2 RNA control embryos images.

Brd2b knockdown increases cell death in the CNS of prim 5 morphant embryos but reduces cell death in the cloaca of the pronephros. Dark-field images of representative 24 hpf prim5 embryos treated as indicated and after TUNEL assay for apoptotic nuclei. Panels (A–D) (heads and trunks) show treatment groups for testing gene-specificity of apoptotic effects: (A) uninjected; (B) brd2bMO1-injected; (C) brd2bMO1 + HsBrd2RNA-injected; and (D) Crispr-Cas9-brd2b-disrupted embryos. HsBrd2RNA co-injection rescues, while Crispr-Cas9 treatment phenocopies, excess apoptosis in the brain and trunk overall of morphants, showing effects are specific to brd2b. Arrows indicate specific regions of reduced apoptosis at the cloaca (upper arrows) and increased apoptosis in the PBI (lower arrows) in the ventral trunk of morphant and Crispr-Cas9-treated, but not control or rescued embryos. Panels (E–H) (heads and trunks) show treatment groups for testing p53-dependent off-target effects: (E) uninjected; (F) p53MO-injected; (G) brd2bMO1-injected; and (H) brd2bMO1− + p53MO co-injected. p53MO co-injection does not abrogate excess apoptosis in morphants, ruling out off-target effects as the cause of cell death. See Figure 5 for quantitative TUNEL data from these studies. See Supplementary Figure S2 for images on HsBrd2 RNA control embryos.

Quantitative TUNEL analysis of cell death: knockdown, co-knockdown, and antagonism studies. Cell death levels were measured in brd2bMO morphants and control embryos under various treatments using fluorescence TUNEL assay followed by quantitative laser-scanning confocal microscopy. The number of apoptotic nuclei in multiple optical sections from the brains of three to ten embryos per treatment, depending on experiment, were compared (see Methods for details). (A) RNA rescue treatments: brd2bMO1-injected (2bMO1), human Brd2 RNA-injected (HsRNA), brd2bMO1 + human Brd2 RNA-injected (2bMO+HsRNA), and uninjected (control). Means vary significantly by treatment (p < 0.0001, one-way ANOVA, Tukey’s HSD), with the greatest difference between 2bMO1 and all other treatments and no significant difference between “control” and “2bMO+HsRNA”, indicating effective rescue by exogenous human Brd2 RNA. Cell death numbers for HsBrd2 RNA-injected controls were obtained from embryos from a separate clutch but reflect what we have consistently seen over multiple independent trials. (B) brd2bMO knockdown corroboration treatments: brd2bMO1-injected (2bMO1), brd2bMO2-injected (2bMO2), Crispr-Cas9-brd2b disruption (Crispr2b), and uninjected (control). Means vary significantly by treatment (p < 0.0001, one-way ANOVA, Tukey HSD), with Crispr2b, 2bMO1 and 2bMO2 all showing effects significantly greater than “control”; 2bMO1 and 2bMO2 show equivalent and moderate effects, while Crispr2b shows the greatest effect. Thus, three independent treatments targeting brd2b result in similarly increased cell death levels, showing gene-specificity of effect. (C) p53-dependent off-target effect treatments: brd2bMO1-injected (2bMO1), brd2bMO1 + p53MO co-injected (2bMO1 + p53MO), uninjected (control), and p53MO-injected (p53MO). Means vary significantly by treatment (p < 0.0001, one-way ANOVA, Tukey’s HSD), with no significant difference between 2bMO1 and “2bMO1 + p53MO” or between “control” and p53MO, but significant differences between the two pairs, indicating excess apoptosis is brd2b-specific, and not the result of p53-dependent off-target effects. (D) Co-knockdown treatments: brd2aMO1-injected (2aMO1), brd2bMO1-injected (2bMO1), brd2aMO1 + brd2bMO1 co-injected (2aMO+2bMO), and uninjected (control). Means vary significantly by treatment (p < 0.0001, one-way ANOVA, Tukey’s HSD), with “control” and “2aMO + 2bMO” not significantly different from each other and both significantly different from either 2bMO1 or 2aMO1; 2bMO1 shows a greater effect than 2aMO1. Simultaneous knockdown of both paralogs suppresses excess cell death observed in single knockdowns of either paralog and restores wild type levels of apoptosis. Panels (E–H) Rescue-Enhancement studies using zebrafish brd2a and brd2b-L in vitro-synthesized RNAs (not human Brd2) tested functional antagonism between paralogs at the level of apoptosis. (E): brd2bMO rescue treatments: brd2bMO2 (2bMO2), brd2b-LRNA (2bRNAL), brd2bMO2 + brd2b-LRNA (2bMO2 + 2bRNAL), and uninjected (control). Means vary significantly by treatment (p-0.0286, one-way ANOVA, Tukey’s HSD), with “2bMO2+2bRNAL” closer to “control” in effect (p = 0.8407) than either 2bMO1 (p = 0.1319) or 2bRNAL (p = 0.0295), indicating partial rescue of the brd2bMO morphant defect. (F) brd2bMO enhancement treatments: brd2aRNA-injected (2aRNA), brd2bMO2-injected (2bMO2), brd2aRNA + brd2bMO2 co-injected(2bMO2+2aRNA), and uninjected (control). Means vary significantly by treatment (p = 0.0008, one-way ANOVA, Tukey’s HSD), with “2bMO2+2aRNA” giving significantly greater effect compared to “control” (p = 0.0009) than either 2bMO2 (p = 0.2405) or 2aRNA (p = 0.9151). (G) brd2aMO rescue treatments: brd2aMO1-injected (2aMO1), brd2aRNA-injected (2aRNA), brd2aMO1 + brd2aRNA co-injected (2aMO1 + 2aRNA), and uninjected (control). Means vary significantly be treatment (p = −0.0176, one-way ANOVA, Tukey’s HSD), with “2aMO1 + 2aRNA” closer to “control” in effect (p = 0.9985) than either 2aRNA (p = 0.3286) or 2aMO1 (p = 0.0.0294), indicating partial rescue of the brd2aMO morphant defect. (H) brd2aMO enhancement treatments: brd2aMO1-injected (2aMO1), brd2b-LRNA-injected (2bRNAL), brd2aMO1 + brd2b-LRNA co-injected (2aMO1 + 2bRNAL), and uninjected (control). Means vary significantly by treatment (p = 0.0084, one-way ANOVA, Tukey’s HSD), with “2aMO1+2bRNAL” giving significantly greater effect compared to “control” (p = 0.0085), than either 2aMO (p = 0.3433) or 2bRNAL (0.9342). Note: for experiments in A, B, and D, apoptotic cell counts were obtained from 55 to 60 optical sections from 6 to 10 embryos per treatment; for experiment C, from maximum projection images from 3 embryos per treatment; for experiments (E–H), from 35 optical sections from 3 to 5 embryos per treatment. Green diamonds: confidence interval (95%) for group means with standard error. Black line: grand sample mean. Black circles: comparison circles for absolute differences of group means.

MHB expression of pax2a, eng2a and krox20 mRNA is normal, but patterning of pax2a (+) spinal interneurons and pronephric cells is disrupted, in brd2b morphants. Panels (A–L): Brightfield images of representative 24 hpf prim5 control and morphant embryos assayed by In situ hybridization for expression of patterning genes at the MHB region (A–I) and in spinal interneurons (J–L). Uninjected embryos (wt; A,D,G,J), embryos injected with control brd2b-5-base mismatch morpholino (mis5MO; B,E,H,K), and embryos injected with brd2bMO (brd2bMO1; C,F,I,L) were assayed for eng2a (A,B,C), krox20 (D,E,F,), and pax2a (G,H,I; J,K,L) expression. No differences in patterning gene expression at the MHB region are detected. Using pax2a(+) as a marker for spinal interneurons reveals mispairing on either side of the midline in morphants (J,K vs. L, gaps between lines). (M) Immunofluorescence with anti-Pax2a antibodies to visualize interneurons on either side of spinal cord, dorsal views of maximum projection confocal images. Representative control uninjected (wt) and brd2bMO1-injected (2bMO) embryos are shown with symmetrically paired interneurons linked by white lines; unpaired interneurons appear solo on one side of the central spinal cord. (N) The number of paired “P” vs. unpaired “U” interneurons in six embryos per treatment group (uninjected controls = none; brd2bMO-injected = 2bMO) was analyzed by chi-square contingency and is shown as a box plot (p < 0.0001). While, on average, only 3% of interneurons are unpaired in wild type, about 65% are unpaired in brd2bMO morphants. Panels (O–R): Analysis of Pax2a(+) pronephric precursor cells in the developing duct in brd2bMO and brd2aMO morphants by quantitative immunofluorescence. (O) Diagram of pronephric duct segments: G, glomerulus; N, neck; PCT, proximal convoluted tube; PST, proximal straight tube; DE, distal early tube; DL, distal late tube; C, cloaca. (P) Pax2a(+) immunofluorescence in uninjected (wt) and brd2bMO morphants (brd2bMO1). (Q) Pax2a(+) immunofluorescence in uninjected (wt) and brd2a morphants (brd2aMO1). Paralogs show opposite effects on distribution of Pax2a(+) pronephric cells (arrows). brd2b morphants show fewer cells in the more distal parts of the tube (PST, DE, DL) but more cells at the very end cloaca (C), compared to wild type, while brd2a morphants show excess cells along the tube, and a normal number of cells at the cloaca. (R) Quantitation of distribution of Pax2a(+) cells in uninjected (none) and brd2bMO1-injected embryos (2bMO), by pronephric segment. Boxplot of ANOVA and paired t-tests shows significant differences in the predicted direction between these two treatment groups in segments PST/DE/DL (p < 0.0001,) and in C (p < 0.0001), but not in G/PCT (p = 0.3971).

Co-knockdown of Brd2a and Brd2b suppresses morphant defects in both brain and pronephros of morphant embryos at 24 hpf. Double knockdown was used to test for genetic interaction between paralogs. Panels (A–D): Brightfield images of head and trunk of representative 24 hpf prim5 embryos from indicated treatment groups assessed for morphology: (A) uninjected, (B) brd2bMO-injected, (C) brd2aMO-injected, and (D) brd2bMO + brd2aMO co-injected. Panels (E–H): Darkfield images of representative embryos from the same treatment groups assessed by TUNEL assay for apoptotic nuclei: (E) uninjected, (F) brd2bMO-injected, (G) brd2aMO-injected, and (H) brd2bMO + brd2aMO co-injected. Co-injection of MOs from each paralog suppresses morphant defects in both morphology and cell death in the brain and trunk, indicating genetic interaction with functional antagonism between paralogs. Black arrows indicate reduced brain, ill-formed MHB region, disorganized PBI common to both morphants (B,C), and plug of excess cells at the cloaca unique to brd2bMO morphants (upper arrow in B, trunk). White arrows show increased apoptosis in brain and PBI common to both morphants (F,G, head; G, trunk), while lack of apoptosis is seen at the cloaca only in brd2bMO morphants (F, trunk). See Table 1 for population morphology data and Figure 5 for quantitative TUNEL data. Panels (I–L): Darkfield images of pronephric duct in trunks of representative 24 hpf prim 5 embryos from the indicated treatment groups assessed for Pax2a(+) cells by immunofluorescence: (I) uninjected, (J) brd2bMO-injected, (K) brd2aMO-injected, and (L) brd2bMO + brd2aMO co-injected. Cloacal opening is indicated by white arrow in left panels. Distribution along the ductal tube of Pax2a(+) pronephric precursor cells is highlighted between white arrows in right panels. Co-knockdown suppresses morphant defects of each paralog, bringing distribution of cells along the duct and at the cloaca closer to wildtype. (see also Figure 6O,P,Q). Panels (M–P) Darkfield images of spinal interneurons along dorsal trunks of representative 24 hpf embryos from the indicated treatments groups assessed for Pax2a(+) cells by immunofluorescence: (M) uninjected, (N) brd2bMO-injected, (O) brd2aMO-injected, and (P) brd2bMO + brd2aMO co-injected. Co-knockdown suppresses the morphant mis-pairing defect of each paralog, and restores interneuron number and pairing to near wildtype levels. Lines connect symmetrically paired neurons; unpaired neurons appear solo on one or the other side of the central nerve cord.

Enhancement of morphant brain phenotypes by injection of paralogous RNA corroborates genetic antagonism between brd2a and brd2b. Rescue-Enhancement studies using zebrafish brd2a and brd2b-L in vitro-synthesized RNAs (rather than human Brd2) were used to test functional antagonism between paralogs implied by double knockdown suppression. Panels (A–I): Brightfield images of representative 24 hpf prim 5 embryos from indicated treatment groups. For controls: (A) uninjected wild type, (B) brd2aRNA-injected, and (C) brd2b-LRNA-injected. For brd2bMO rescue-enhancement: (D) brd2bMO-injected, (E) brd 2bMO + brd2b-LRNA co-injected (rescue), and (F) brd2bMO + brd2aRNA co-injection (enhancement). For brd2aMO rescue-enhancement: (G) brd2aMO-injected, (H) brd2aMO + brd2aRNA co-injected (rescue), and (I) brd2aMO + brd2b-LRNA co-injection (enhancement). Arrows indicate rescue (E,H) and enhancement (F,I) of morphant brain defects by co-injection of cognate or paralogous RNA, respectively. See Table 2 for population morphology data. Panels (J–R): Darkfield images of representative 24 hpf prim 5 embryos from the same treatment groups after TUNEL assay: (J) uninjected wildtype, (K) brd2aRNA-injected, and (L) brd2b-LRNA-injected; (M) brd2bMO-injected, (N) brd2bMO + brd2b-LRNA co-injected (rescue), and (O) brd2bMO + brd2aRNA co-injection (enhancement); (P) brd2aMO-injected, (Q) brd2aMO + brd2aRNA co-injected (rescue), and (R) brd2aMO + brd2bRNA co-injection (enhancement). Note rescue (N,Q) and enhancement (O,R) of morphant excess apoptosis by co-injection of cognate or paralogous RNA, respectively, supporting the idea of functional antagonism between paralogs. See Figure 5 for quantitative TUNEL data.