Abnormal expression of GABAA receptor sub-units and hypomotility upon loss of gabra1 in zebrafish

We used whole exome sequencing (WES) to determine the genetic etiology of a patient with a multi-system disorder characterized by a seizure phenotype. WES identified a heterozygous de novo missense mutation in the GABRA1 gene (c.875C>T). GABRA1 encodes the alpha subunit of the Gamma-Aminobutyric Acid receptor A (GABAAR). The GABAAR is a ligand gated ion channel that mediates the fast inhibitory signals of the nervous system and mutations in the sub-units that compose the GABAAR have been previously associated with human disease. To understand the mechanisms by which GABRA1 regulates brain development, we developed a zebrafish model of gabra1 deficiency. gabra1 expression is restricted to the nervous system and behavioral analysis of morpholino injected larvae suggests that the knockdown of gabra1 results in hypoactivity and defects in the expression of other sub-units of the GABAAR. Expression the human GABRA1 protein in morphants partially restored the hypomotility phenotype. In contrast, the expression of the c.875C>T variant did not restore these behavioral deficits. Collectively, these results represent a functional approach to understand the mechanisms by which loss of function alleles cause disease.


INTRODUCTION 0 5
and therefore, additional rescue experiments with higher concentrations could not be attempted. 2 0 6 2 0 7 The c.875C>T GABRA1 variant does not restore the hypomotility phenotype in 2 0 8 morphants.

0 9
The functional consequences of the c.875C>T variant are currently unknown. Therefore, 2 1 0 we asked whether expression of the c.875C>T variant was sufficient to restore the hypomotility 2 1 1 induced by knockdown of gabra1. Embryos were injected at the single cell stage with RC 2 1 2 morpholinos, tbMO morpholinos, GABRA1 c.875C>T mRNA (SDM), or a combination of 2 1 3 GABRA1 mRNA with tbMO morpholinos. Consistent with previous experiments, injection of the 2 1 4 tbMO morpholino caused a significant reduction (p=0.0153) in the total distance swam relative 2 1 5 to embryos injected with the RC (Fig. 4C). Interestingly, the co-injection of the mRNA encoding 2 1 6 the c.875C>T (SDM) and the tbMO was unable to restore the total distance traveled to control 2 1 7 levels (Fig. 4C). Importantly, the injection of the GABRA1 c.875C>T variant (SDM) at 2 1 8 1000pg/embryo had no significant effects on the total distance swam (Fig. 4C).

1 9
The expression of gabrb2 and gabrg2 are decreased in gabra1 morphants.

0
Previous studies suggest that approximately 60% of all GABA A Rs consist of two α 1, two 2 2 1 β 2, and one γ 2 subunits (Sigel and Steinmann, 2012). We hypothesized that the knockdown of 2 2 2 gabra1, which encodes the α 1 sub-unit would alter the sub-unit composition of the GABA A R. To  2  2  3 begin to test this, we analyzed the expression of the genes that encode the β 2 and γ 2 sub-units.

4
As shown in Figure 5A, knockdown of gabra1 caused a decrease in the expression of gabrb2 2 2 5 (β2) and gabrg2 (γ2). We next measured the expression of other alpha sub-units in gabra1 2 2 6 morphants. As shown in Figure 5A, injection of the tbMO was associated with increased 2 2 7 expression of gabra6a and gabra6b, but only gabra6b was statistically significant across 2 2 8 biological triplicates. A similar expression pattern of gabra6a and gabra6b was observed upon 2 2 9 injection of the sMO (Fig. S1D), with both genes demonstrating a statistically significant 2 3 0 increase in expression. We did not detect a statistical change in the expression of any other α 2 3 1 sub-unit across either the tbMO or the sMO (Figs 5A, S1D).

3 2
We sought to build upon these data by determining whether morphant larvae had an 2 3 3 intact receptor capable of responding to pentylenetetrazol (PTZ), an antagonist of the GABA A R. 2 3 4 Non-injected wildtype embryos treated with 10mM PTZ exhibit short convulsions and a whirlpool 2 3 5 swimming behavior with a 2-fold increase in swim speed (p= 6.74E-06) and an approximate 6-2 3 6 fold increase in total distance swam (p= 2.14E-05) (Fig. 5B,C). These phenotypes were 2 3 7 consistently observed in larvae injected with RC morpholinos as the RC morpholino had no 2 3 8 affect on larval behavior or their response to PTZ. Interestingly, gabra1 morphants (tbMO) 2 3 9 responded normally to PTZ according to both distance and speed measurements (Fig. 5B,C).

4 0
Collectively, these data demonstrate that knockdown of gabra1 alters the expression of unique 2 4 1 GABA A R subunits, although, morphants continue to respond to PTZ treatment. Recent work in zebrafish has demonstrated that the homozygous nonsense mutation of gabra1, 3 2 0 does not disrupt overall brain structure or the total number of GABAergic cells, but does 3 2 1 influence the brain transcriptome (Samarut et al., 2018). Collectively, these data raise the 3 2 2 possibility that other alpha sub-units may compensate for the loss of gabra1, ultimately 3 2 3 producing unique compositions of the GABA A R. The function of these receptors is unknown. But 3 2 4 it is conceivable that the production of GABA A R with unique sub-unit composition in incorrect 3 2 5 regions of the brain might underlie the impaired synapse formation observed in zebrafish 3 2 6 harboring germline mutations in the gabra1 gene (Samarut et al., 2018). 3 2 7 We provide strong evidence that heterozygous de novo mutation of GABRA1 is 3 2 8 associated with a multi-system disorder characterized by severe seizures. We further 3 2 9 characterized the developmental and behavioral defects associated with knockdown of gabra1 3 3 0 in zebrafish. Behaviorally, morphant animals present with hypomotility at 5 DPF measured by 3 3 1 reduced swim speed and total distance traveled. These deficits coincide with significant 3 3 2 changes in the expression of GABA A R sub-units and cannot be restored by the de novo 3 3 3 c.875C>T allele. Although a zebrafish harboring a mutation in the gabra1 gene has recently 3 3 4 been created, detailed behavioral analysis was performed at the juvenile stage (weeks post 3 3 5 fertilization). Here we complement previous studies using a morpholino mediated knockdown 3 3 6 approach, as the homozygous deletion of gabra1 was lethal (Samarut et al., 2018). Our 3 3 7 behavioral study is the first to our knowledge that comprehensively characterizes the phenotype 3 3 8 of gabra1 deletion during early development (DPF as opposed to weeks post fertilization). We 3 3 9 observed hypomotility consistent with previous studies in zebrafish and our study likely informs 3 4 0 about specific types of mutations, those of which result in loss of function alleles. Importantly, 3 4 1 our restoration experiments with the c.875C>T allele suggest that this allele is in fact a loss of 3 4 2 function allele. They were maintained and bred in groups of two females and two to four males. The collected 3 5 0 zebrafish embryos were kept in egg water consisting of 0.03% Instant Ocean (Aquaneering, 3 5 1 San Diego, CA) in D.I. water at 28°C. For morpholino injections, a translational blocking morpholino (tbMO) 4 0 5 (TCTTCCACCCCACATCATTCTCCGA) and a splice site inhibiting morpholino (sMO) 4 0 6 (ACACGCTCTGTTGAAGCAAGAAATT) targeting gabra1 were designed. The efficiency of 4 0 7 knockdown for the sMO was performed with primers flanking the target site (Fwd: 4 0 8 GACAGCCTCCTCGATGGTTA and Rev: GCAGAGTCCCTTCCTCTGTG). Each morpholino 4 0 9 was injected independently at the single cell stage at a concentration of 1.6 ng/embryo. An 4 1 0 equivalent concentration of randomized control morpholinos (25-N) was injected as a control. 4 1 1 Final concentration of morpholino was determined empirically after an injection gradient was 4 1 2 performed to determine optimal survival. For rescue experiments, the human GABRA1  4  1  3 complete open reading frame was purchased from TransOMIC Technologies (Huntsville, AL Embryos injected with random control morpholinos, tbMO, sMO, GABRA1 mRNA, 4 4 3 GABRA1 (c.875C>T), or a combination as indicated in the figure legends were raised to 5 DPF. 4 4 4 Behavioral analysis was performed using the Zebrabox (ViewPoint Behavior Technology, 4 4 5 Montréal, Canada). Larvae were individually tracked for swim speed and total distance swam in 4 4 6 a 96 well plate. The behavioral protocol (adapted from (Afrikanova et al., 2013)) was a total of 4 4 7 15 minutes divided into 5 minute intervals of dark/light/dark conditions. All larvae were 4 4 8 acclimated to the dish and housing conditions for 1 hour prior to analysis. Settings for the 4 4 9 program include a threshold of 16 and integration period of 300 seconds. Data was measured 4 5 0 as total distance traveled (mm) and total swim speed (mm/sec) (