A genetic screen reveals dosage-sensitive effects of ASD genes and identifies domino as a regulator of synaptic and behavioral phenotypes

Joint public review:

Summary:

In this study, Stirtz et al., performed a targeted screen of 80 Drosophila strains carrying heterozygous MiMIC insertions in genes that are homologous to human genes that have been linked to autism spectrum disorders (ASD). This is an important and timely topic, as human genetic studies have identified a large number of ASD risk genes, yet the functional characterization of many of these candidates remains limited. The authors identify 48 putative mutants with altered sleep, activity, or social behavior. They then focus on one hit, domino (the orthologue of human SRCAP), for which the heterozygous MiMIC mutants show altered behavior in males but not in females. They show that domino is a candidate regulator of sleep, activity, social behavior, transcriptional programs, and RNA splicing. The authors molecularly validate that the heterozygous MiMIC insertion in domino causes a 50% reduction in gene expression, and use RNA-seq to show that the heterozygous MiMIC males and females have altered gene expression profiles and splicing patterns. Finally, they use immunostaining against the commonly used synaptic marker, Bruchpilot, to show that both males and female heterozygous domino flies express a higher immunosignal compared to the wild-type control.

Strengths:

This work provides potential genetic links between human ASD genes and fly behavioral phenotypes. Overall, it represents an ambitious and technically valuable effort that generates a substantial behavioral dataset across a large number of ASD-associated orthologues and develops quantitative analytical approaches to extract information from complex phenotypes. One strength of this study is its focus on heterozygous mutants, which is more representative of human scenarios. The study also provides a potentially useful resource for the field, particularly through the identification of candidate genes and behavioral signatures that may warrant future mechanistic investigations. The screening experiments and analysis are well conceived, the manuscript is very clearly written and is easily understandable, and the concise, accurate interpretations for each result, aided by clear graphic representation of multiple dimensions in the behaviors tested, allow the reader to understand the paper with ease.

Weaknesses:

The work presents a few important weaknesses, especially with regard to the genetic and molecular validation of the mutants identified.

(1) The authors validate that the MiMIC insertion affects the gene of interest only for the domino gene. The original MiMIC study (PMID: 25824290, eLife) reported that ~8% (5/63) MiMIC lines do not function as strong loss-of-function alleles. Thus, of the 48 hits identified here, one would estimate that ~4 of them may not cause the loss of function of the gene defined by the MiMIC insertion. To strengthen their claim, the authors would need to confirm that all of the MiMIC lines that they consider as hits do indeed significantly reduce the expression of the target genes.

(2) Although the authors document that they validated the phenotype seen in the domino MiMIC line using a second mutant allele (Trojan), these two mutants share the same genetic background because the Trojan line was made from the MiMIC line via recombinase-mediated cassette exchange. Thus, the phenotype seen in the MiMIC and Trojan lines would need to be confirmed using a completely independent mutant in order to demonstrate that the reported behavioral, molecular, and synaptic defects reported can be fully attributed to the partial loss of domino function. Also, while the authors performed an RNA-seq experiment in both the MiMIC and Trojan lines, they do not show whether the Bruchpilot phenotype is also seen in the Trojan allele. Thus, this phenotype would also need to be examined in the Trojan allele or, preferably, in a mutant allele that is independent of the MiMIC line.

(3) The RNA-seq results would benefit from a discussion of potential compensatory or secondary transcriptional effects resulting from the constitutive domino reduction, particularly since the expected global bias toward transcriptional downregulation was not observed. In addition, some neurobiological interpretations appear stronger than currently justified by the literature or the data presented, particularly regarding the Bruchpilot immunoreactivity analyses and their relationship to sleep-regulatory circuits. Additional validation using better-established sleep-related neuronal populations, together with a clearer discussion of sex-specific effects and alternative interpretations of the observed phenotypes, would substantially strengthen the manuscript.

(4) An explanation of the extensive PCA analyses performed would help the naïve reader.