Dosage sensitivity to Pumilio1 variants in the mouse brain reflects distinct molecular mechanisms.

Different mutations in the RNA-binding protein Pumilio1 (PUM1) cause divergent phenotypes whose severity tracks with dosage: a mutation that reduces PUM1 levels by 25% causes late-onset ataxia, whereas haploinsufficiency causes developmental delay and seizures. Yet PUM1 targets are derepressed to equal degrees in both cases, and the more severe mutation does not hinder PUM1's RNA-binding ability. We therefore considered the possibility that the severe mutation might disrupt PUM1 interactions, and identified PUM1 interactors in the murine brain. We find that mild PUM1 loss derepresses PUM1-specific targets, but the severe mutation disrupts interactions with several RNA-binding proteins and the regulation of their targets. In patient-derived cell lines, restoring PUM1 levels restores these interactors and their targets to normal levels. Our results demonstrate that dosage sensitivity does not always signify a linear relationship with protein abundance but can involve distinct mechanisms. We propose that to understand the functions of RNA-binding proteins in a physiological context will require studying their interactions as well as their targets.

Alternative polyadenylation alters protein dosage by switching between intronic and 3'UTR sites.

Alternative polyadenylation (APA) creates distinct transcripts from the same gene by cleaving the pre-mRNA at poly(A) sites that can lie within the 3' untranslated region (3'UTR), introns, or exons. Most studies focus on APA within the 3'UTR; however, here, we show that CPSF6 insufficiency alters protein levels and causes a developmental syndrome by deregulating APA throughout the transcript. In neonatal humans and zebrafish larvae, CPSF6 insufficiency shifts poly(A) site usage between the 3'UTR and internal sites in a pathway-specific manner. Genes associated with neuronal function undergo mostly intronic APA, reducing their expression, while genes associated with heart and skeletal function mostly undergo 3'UTR APA and are up-regulated. This suggests that, under healthy conditions, cells toggle between internal and 3'UTR APA to modulate protein expression.

Dual antibody strategy for high-resolution imaging of murine Purkinje cells and their dendrites across multiple layers.

Despite being among the largest neurons in the mammalian brain, Purkinje cells are difficult to visualize and trace via immunofluorescence because their dendritic arbors extend through several cerebellar layers. This protocol describes a two-antibody strategy we developed to study Purkinje cell morphology in mice. With it, one can reconstruct three-dimensional images of Purkinje cells at single-neuron resolution across multiple layers. The substantially improved image quality reveals subtle defects, enabling more meaningful morphological analysis. For complete details on the use and execution of this protocol, please refer to Gennarino et al. (2015).

Protocol to assess the effect of disease-driving variants on mouse brain morphology and primary hippocampal neurons.

Genetic variants that affect neurological function will often produce changes visible at the level of gross morphology, either of the whole brain or of specific neuronal types. Here we describe how to perfuse and dissect the brain in preparation for Nissl staining. Then we outline steps for culturing mouse primary hippocampal neurons to evaluate dendritic arborization (Sholl analysis). For complete details on the use and execution of this protocol, please refer to Gennarino et al. (2018).

Determining the effects of loss of function mutations in human cell lines.

Quantifying differences in the amount of protein and mRNA caused by missense mutations in a gene of interest can be challenging, especially when using patient-derived primary cells, which are intrinsically variable. In this protocol, we describe how to culture patient-derived lymphoblast and fibroblast cell lines for later mRNA and protein quantification. We also describe the steps to examine variants of PUM1 in HEK293T cells, but the protocol can be applied to other proteins of interest. For complete details on the use and execution of this protocol, please refer to Gennarino et al. (2018).

Challenging Safety and Efficacy of Retinal Gene Therapies by Retinogenesis.

Gene-expression programs modulated by transcription factors (TFs) mediate key developmental events. Here, we show that the synthetic transcriptional repressor (TR; ZF6-DB), designed to treat Rhodopsin-mediated autosomal dominant retinitis pigmentosa (RHO-adRP), does not perturb murine retinal development, while maintaining its ability to block Rho expression transcriptionally. To express ZF6-DB into the developing retina, we pursued two approaches, (i) the retinal delivery (somatic expression) of ZF6-DB by Adeno-associated virus (AAV) vector (AAV-ZF6-DB) gene transfer during retinogenesis and (ii) the generation of a transgenic mouse (germ-line transmission, TR-ZF6-DB). Somatic and transgenic expression of ZF6-DB during retinogenesis does not affect retinal function of wild-type mice. The P347S mouse model of RHO-adRP, subretinally injected with AAV-ZF6-DB, or crossed with TR-ZF6-DB or shows retinal morphological and functional recovery. We propose the use of developmental transitions as an effective mode to challenge the safety of retinal gene therapies operating at genome, transcriptional, and transcript levels.

Targeting and silencing of rhodopsin by ectopic expression of the transcription factor KLF15.

The genome-wide activity of transcription factors (TFs) on multiple regulatory elements precludes their use as gene-specific regulators. Here we show that ectopic expression of a TF in a cell-specific context can be used to silence the expression of a specific gene as a therapeutic approach to regulate gene expression in human disease. We selected the TF Krüppel-like factor 15 (KLF15) based on its putative ability to recognize a specific DNA sequence motif present in the rhodopsin (RHO) promoter and its lack of expression in terminally differentiated rod photoreceptors (the RHO-expressing cells). Adeno-associated virus (AAV) vector-mediated ectopic expression of KLF15 in rod photoreceptors of pigs enables Rho silencing with limited genome-wide transcriptional perturbations. Suppression of a RHO mutant allele by KLF15 corrects the phenotype of a mouse model of retinitis pigmentosa with no observed toxicity. Cell-specific-context conditioning of TF activity may prove a novel mode for somatic gene-targeted manipulation.

Rhodopsin targeted transcriptional silencing by DNA-binding.

Transcription factors (TFs) operate by the combined activity of their DNA-binding domains (DBDs) and effector domains (EDs) enabling the coordination of gene expression on a genomic scale. Here we show that in vivo delivery of an engineered DNA-binding protein uncoupled from the repressor domain can produce efficient and gene-specific transcriptional silencing. To interfere with RHODOPSIN (RHO) gain-of-function mutations we engineered the ZF6-DNA-binding protein (ZF6-DB) that targets 20 base pairs (bp) of a RHOcis-regulatory element (CRE) and demonstrate Rho specific transcriptional silencing upon adeno-associated viral (AAV) vector-mediated expression in photoreceptors. The data show that the 20 bp-long genomic DNA sequence is necessary for RHO expression and that photoreceptor delivery of the corresponding cognate synthetic trans-acting factor ZF6-DB without the intrinsic transcriptional repression properties of the canonical ED blocks Rho expression with negligible genome-wide transcript perturbations. The data support DNA-binding-mediated silencing as a novel mode to treat gain-of-function mutations.