cntnap2ab Mutants Display GABAergic Deficits
(A) Cntnap2a and Cntnap2b proteins showing functional domains. DISC, discoidin; EGF, epidermal growth factor; 4.1, 4.1 binding domain; LamG, laminin G; TM, transmembrane.
(B) Western blot analysis of Cntnap2a and Cntnap2b in cntnap2a^Δ25/Δ25cntnap2b^Δ7/Δ7 and cntnap2a^Δ121/Δ121cntnap2b^31i/31i fish.
(C-F and C′-F′) Reporter gene expression in Tg(dlx6a-1.4kbdlx5a/dlx6a:GFP) and Tg(vglut:DsRed) in wild-type (C-F) and cntnap2a^Δ121/Δ121cntnap2b^31i/31i (cntnap2ab) (C′-F′) larvae at 4 dpf. Note the deficit in dlx5a/6a:GFP+ cells in the telencephalon (arrowheads). (C and D, C′ and D′) Lateral views. (E and F, E′ and F′) Ventral views. hyp, hypothalamus; OT, optic tectum; tel, telencephalon.
(G) Total number of dlx5a/6a:GFP+ cells in the forebrain of wild-type (n = 13) and cntnap2ab mutants (n = 14). *p = 3.08 × 10⁷ (one-way ANOVA).
(H) Average activity in response to PTZ (10 mM) in the progeny of incrossed cntnap2a^Δ25/+cntnap2b^Δ7/+ fish at 4 dpf (n = 268; cntnap2ab + PTZ, n = 6; wild-type + PTZ, n = 11; cntnap2ab + water, n = 11; wild-type + water, n = 6; p = 0.0013, two-way ANOVA, genotype × dose interaction).

cntnap2ab Mutants Display Nighttime Hyperactivity
(A) Experimental setup (Prober et al., 2006 and Rihel et al., 2010).
(B) Locomotor activity of cntnap2a^Δ25/Δ25cntnap2b^Δ7/Δ7 (cntnap2ab, red) and wild-type (WT; blue) sibling-matched larvae over 72 hr. Hyperactivity in mutants worsens on successive nights (arrows). The magnified activity profile on night 6 is shown.
(C) Average locomotor activity of cntnap2ab versus wild-type. p = 0.00012 (one-way ANOVA, comparing all genotypes on all nights); p = 0.0193, 0.0236, and 0.0073, nights 4, 5, and 6, respectively.
(D) Hierarchical clustering of the cntnap2ab behavioral fingerprint (red arrow) compared with the fingerprints of wild-type larvae exposed to a panel of 550 psychoactive agents from 4–6 dpf ( Rihel et al., 2010). Each rectangle in the clustergram represents the Z score, or the average value in SDs relative to the behavioral profiles of wild-type exposed to DMSO alone (magenta, higher than DMSO; cyan, lower than DMSO). The cntnap2ab profile correlates with agents that induce nighttime arousal (“Correlating Drugs”).
(E) A magnified section of the clustergram, highlighting compounds that correlate with the cntnap2ab mutant behavioral fingerprint.

Differential Behavioral Responses of cntnap2ab Mutants to Psychoactive Agents
(A) The 14 psychoactive drugs tested in cntnap2a^Δ121/Δ121cntnap2b^Δ31i/Δ31i and wild-type larvae, their biological targets, and the rationale for their selection. The following classes of drugs are shown: correlating (blue), anti-correlating (purple), drugs interacting with the GABA-A receptor (orange), and risperidone (green).
(B and C) Dose-response effects of the NMDA receptor antagonist, (-)-MK-801, on night waking activity at 5 dpf (B) and the non-benzodiazepine GABA-A receptor agonist zolpidem on average night waking activity at 4–6 dpf (C) in wild-type (WT; blue) and cntnap2ab (red) larvae (p = 0.002, (-)-MK-801; p = 0.0003, zolpidem; two-way ANOVA, genotype × drug interaction).
(D and E) Significant enrichment of NMDA receptor antagonists (D) and GABA receptor antagonists (E) in the top ranks of correlating drugs (p = 0.031, NMDA-R antagonists; p = 0.034, GABA-R antagonists; Kolmogorov-Smirnov test).
(F) Hierarchical clustering of the behavioral profiles of wild-type or cntnap2b larvae exposed to 14 psychoactive agents at three doses each. Each rectangle in the clustergram represents the Z score relative to the behavior of wild-type or mutant larvae exposed to DMSO alone (magenta, higher than DMSO; cyan, lower than DMSO).
(G) Magnified sections highlight the behavioral fingerprints of wild-type and cntnap2ab larvae in response to zolpidem (20 μM).
(H) Pairwise Euclidean distances between wild-type and cntnap2b responses to psychoactive agents in the PCA. Note that zolpidem and (±)-baclofen produce the strongest differential responses.

Biochanin A Reverses Nighttime Hyperactivity in cntnap2ab Mutants
(A) Rank-sorting of the anti-correlating data set with respect to estrogenic compounds shows significant enrichment of estrogenic agents in the top ranks (p = 0.0003 by random permutation). Black lines indicate drugs defined as having estrogenic activity (25 compounds in total).
(B) Hierarchical clustering of the behavioral fingerprints of cntnap2a^Δ121/Δ121cntnap2b^Δ31i/Δ31i larvae exposed to 14 psychoactive agents at three doses each relative to the wild-type + no drug fingerprint. Each rectangle in the clustergram represents the Z score of drug-exposed mutants relative to untreated wild-type (magenta, higher than wild-type; cyan, lower than wild-type).
(C) Magnified sections of the clustergram show relative suppression of the mutant fingerprint by biochanin A (0.1 μM) compared with risperidone (0.001 μM). The red box highlights parameters that measure nighttime activity.
(D) Pairwise Euclidean distances (“Differential Drug Effects”) between the mutant responses to psychoactive agents compared with untreated wild-type in the PCA (Figure S4G). Biochanin A (0.1 μM) produces the strongest suppression of the mutant phenotype in this assay, with the fewest effects on other behavioral parameters.
(E and F) Dose-response effects of biochanin A (E) and risperidone (F) on nighttime activity at 5 dpf (p = 0.0001, biochanin A; p = 0.0034, risperidone; two-way ANOVA, gene × drug interaction). Although there is some experimental variability in the baseline activity of both wild-type and mutant larvae, nighttime hyperactivity is consistently observed.
(G and H) Effect of the blind addition of biochanin A (0.1–1 μM) or DMSO on activity at night (G) and day (H) in the progeny of incrosses of cntnap2a^Δ25/+cntnap2b^Δ7/+ fish at 5 dpf. p = 0.045 (two-way ANOVA, gene × dose interaction).

Figure S1, Related to Figure 1. Generation of cntnap2ab mutants and characterization of brain structure.

A, B. Fluorescent in situ hybridization of cntnap2a and cntnap2b expression and acetylated tubulin in wild-type embryos at 30 and 48 hpf.

C. Whole mount in situ hybridization of cntnap2a and cntnap2b in wild-type embryos at 24 and 48 hpf. Both paralogs are expressed in the CNS, though cntnap2b shows a more specific expression pattern in the telencephalon and diencephalon (24 hpf), the optic tectum, midbrain-hindbrain boundary, and hindbrain (48 hpf).

D. Genomic sequences of insertion-deletion mutations generated by zinc finger nucleases in each cntnap2 paralog. These include Δ121 and Δ25 in exon 3 of cntnap2a (top) and Δ7 in exon 2 and a 31-nt insertion in exon 4 of cntnap2b (bottom), all of which produce a frameshift and premature stop codon.

E. Alternatively spliced transcripts resulting from the 31-nt allele in exon 4 of cntnap2b identified by RT-PCR of exons 1 to 6 at 5 dpf. These transcripts include: (i) the 31-nt sequence itself, as in the genomic sequence; (ii) Δ10 in exon 4: because the 31-nt insertion occurs close to the intron-exon 4 boundary, an alternate splice site is used (the next “TCAG” in the genomic sequence after the actual intronic splice site), which is 10 nt after the beginning of exon 4; (iii) Δ148 due to splicing of the entire exon 4. All 3 transcripts produce a frameshift and a premature stop codon. We did not identify alternatively spliced transcripts resulting from the other mutations in either gene.

F. Sequence alignment of wild type cntnap2a (top) and cntnap2b (bottom) proteins and the translation of mutant transcripts. All mutant transcripts are predicted to cause premature stop codons in the discoidin domains of both proteins (Figure 1A).

G-I. Acetylated tubulin (AcTub) and synaptic vesicle protein 2 (SV2) labeling in wild-type (G-I) and cntnap2a^Δ121/Δ121cntnap2b^31i/31i (G′-I′) fish at 24, 48, and 120 hpf. While there is a transient delay in the formation of the anterior and post-optic commissures in cntnap2ab mutants at 28 hpf (see K, K′, M), these commissures form by later stages. Brain structure is not otherwise grossly disrupted in cntnap2ab mutants up to 120 hpf.

J-K. Acetylated tubulin and GABA (J, J′) or GFAP (glial cells) (K, K′) in wild-type and cntnap2a^Δ121/Δ121cntnap2b^31i/31i (cntnap2ab) mutants at 28 hpf.

L. There are significantly fewer GABA+ cells in both the telencephalon (arrowheads) and diencephalon (double arrowheads) of cntnap2ab (J′) versus wild-type (J) larvae at 28 hpf. GABA+ cells were counted in z-stack images by three blinded raters. (*p=0.00782, telencephalon; **p=0.0181, diencephalon, two-way ANOVA, genotype factor).

M. The formation of the anterior commissure (AC) and post-optic commissure (POC) is delayed in cntnap2a^Δ121/Δ121cntnap2b^31i/31i (cntnap2ab) (K′) compared to wild-type (K) larvae (*p=0.00428, AC; **p=0.0159, POC, two-way ANOVA, genotype factor.) Three blinded raters quantified the degree of commissure formation based on a qualitative rating scale ranging from no axons crossing (1) to complete formation (7), as previously described (Barresi et al., 2005). The POC appears to be less fasciculated in mutants. Glial bridge formation is largely intact.

N-R Head and brain size in wild-type (N-Q) and cntnap2a^Δ121/Δ121cntnap2b^31i/31i (cntnap2ab) (N′-Q′). Representative tracings done in Fiji are shown in yellow. The posterior boundaries of the head and brain were defined by the otic vesicle and the midbrain-hindbrain boundary, respectively.

R. cntnap2ab mutant heads are 6.5% and 9.7% smaller than wild-type at 30 hpf and 96 hpf, respectively, and mutant brains are 18% smaller than wild-type at 96 hpf. (*p=9.71x10-7, 30 hpf head size; **p=0.00513, 96 hpf head size; ***p=3.89x10-6, 96 hpf brain size; one-way ANOVA). In addition, eye and yolk size measurements indicate that mutant fish are smaller overall (Table S1).

(A, B): frontal views; (C, N-O, N′-O′): lateral views; (G-J, G′-J′, P-P′): lateral views of dissected embryos with the eye removed; (K-K′): frontal views of the forebrain; (Q-Q′): ventral views; tel, telencephalon; di, diencephalon; OT, optic tectum; PC, posterior commissure; AC, anterior commissure; POC, post-optic commissure; E, eye; MHB, midbrain-hindbrain boundary; HB, hindbrain.

Related to Figure 1. Inhibitory and excitatory neurons in cntnap2ab mutants. Immunostainings of reporter gene expression in Tg(dlx6a-1.4kbdlx5a/dlx6a:GFP) and Tg(vglut:DsRed) in wild-type

(A-H) and cntnap2a^Δ121/Δ121cntnap2b^31i/31i (A′-H′) larvae at 4 dpf. (A-A′, E-E′): The deficit in GABAergic neurons and precursors (dlx5a6a:GFP+ cells) is evident in the ventral and dorsal telencephalon (tel), hypothalamus (hyp), and cerebellum (CB) (arrowheads in A-A′, E-E′). The reduction in GABAergic neurons by surface area of acetylated tubulin is significant in the forebrain and cerebellum (J).

I-K. dlx5a6a:GFP+ cells, vglut:DsRed and acetylated tubulin stainings were quantified by brain region at 4 dpf: forebrain (small brackets in D-D′, H-H′); hypothalamus (large brackets in D-D′, H-H′); cerebellum (arrowheads in E-E′, H-H′); and optic tectum (arrowheads in E-E′, H-H′). (forebrain and hypothalamus: wild-type, n=9; cntnap2ab, n=9; optic tectum and cerebellum: wild-type, n=7; cntnap2ab, n=8)

I. dlx5a6a:GFP+ cell number normalized to wild-type. (*p<0.001; **p<0.01; ***p<0.02; one-way ANOVA, Bonferroni corrected).

J. dlx5a6a:GFP+ cell number relative to acetylated tubulin surface area normalized to wild-type. Because there are significant reductions in GABAergic neurons relative to area in the forebrain and cerebellum, we concluded that the GABAergic deficits in these regions are not primarily attributable to the reduction in brain size. (*p<0.02; one-way ANOVA, Bonferroni corrected).

K. vglut:DsRed surface area relative to acetylated tubulin surface area normalized to wild-type. Unlike GABAergic neurons, there are no significant reductions in glutamatergic surface area relative to acetylated tubulin surface area across regions, suggesting these differences are due to the overall reduction in brain size as opposed to regional deficits (p=0.225, forebrain; p=0.444, hypothalamus; p=0.408, optic tectum; p=0.490, cerebellum; one-way ANOVA).

L-S. Analysis of the CNS of wild-type (L-S) and cntnap2a^Δ121/Δ121cntnap2b^31i/31i (cntnap2ab) (L′-S′) by in situ hybridization at 48 hpf: (L, L′): glutamate decarboxylase 1b (gad1b); (M, M′): vesicular glutamate transporter or solute carrier family 17, member 6b (vglut2a or slc17a6b); (N, N′): solute carrier family 6 (neurotransmitter transporter, glycine), member 5 (glyt2 or slc6a5); (O, O′): tyrosine hydroxylase (th1 or th); (P, P′): neuronal differentiation 1 (neurod or neurod1); (Q, Q′); transiently expressed axonal glycoprotein (tag1) or contactin 2 (cntn2); (R, R′): TUNEL staining at 28 hpf; (S, S′): phospho-histone H3 (H3P) (green) and acetylated tubulin (blue) immunostaining at 48 hpf. We observed that expression of gad1b was variable. No gross differences in the other neurotransmitter systems, markers of neuronal differentiation, apoptosis, or cell proliferation were observed.

(A-D, A′-D′, E-H, E′-H′) show all channels for the images in Figure 1C-F and 1C′-F′, respectively. (A-D, A′-D′): ventral views; (E-H, E′-H′, L-N, L′-N′; P-P′, R-R′, S-S′):): lateral views; (O-O′, Q-Q′): dorsal views. (tel, telencephalon; hyp, hypothalamus; OT, optic tectum; CB: cerebellum; HB, hindbrain; E, eye.)