Mutation in Prkra results in cerebellar abnormality and reduced eIF2α phosphorylation in a model of DYT-PRKRA

Abstract

Variants in the PRKRA gene, which encodes PACT, cause the early-onset primary dystonia DYT-PRKRA, a movement disorder associated with disruption of coordinated muscle movements. PACT and its murine homolog RAX activate protein kinase R (PKR; also known as EIF2AK2) by a direct interaction in response to cellular stressors to mediate phosphorylation of the α subunit of eukaryotic translation initiation factor 2 (eIF2α). Mice homozygous for a naturally arisen, recessively inherited frameshift mutation, Prkralear-5J, exhibit progressive dystonia. In the present study, we investigated the biochemical and developmental consequences of the Prkralear-5J mutation. Our results indicated that the truncated PACT/RAX protein retains its ability to interact with PKR but inhibits PKR activation. Mice homozygous for the mutation showed abnormalities in cerebellar development as well as a severe lack of dendritic arborization of Purkinje neurons. Additionally, reduced eIF2α phosphorylation was noted in the cerebellum and Purkinje neurons of the homozygous Prkralear-5J mice. These findings indicate that PACT/RAX-mediated regulation of PKR activity and eIF2α phosphorylation plays a role in cerebellar development and contributes to the dystonia phenotype resulting from the Prkralear-5J mutation.

Keywords: Cerebellum; DYT-PRKRA; Dystonia; EIF2AK2; PACT/RAX; PKR.

Fig. 1.

Schematic representations of the lear-5J frameshift mutation in the Prkra gene. (A) Functional domains of PACT/RAX and lear-5J truncated protein. Orange boxes, conserved dsRBM1 and dsRBM2, which facilitate high-affinity double-stranded RNA (dsRNA) as well as protein-protein interactions; green box, dsRBM3, which does not bind dsRNA but has weak binding affinity to the PBM within the catalytic (kinase) domain of PKR. The frameshift mutation from a single-nucleotide insertion results in the addition of seven novel amino acids represented in red before the stop codon. dsRBM, dsRNA-binding motif; PBM, PACT/RAX-binding motif. (B) Frameshift mutation in Prkra open-reading frame. A single adenine insertion at nucleotide position 534 (underlined) results in a truncated protein with the original 178 amino acids of PACT/RAX followed by seven novel amino acids before a premature stop codon (asterisk) truncating the protein. This truncation occurs within the dsRBM2 functional domain.

Fig. 2.

The lear-5J protein binds dsRNA less efficiently but interacts with PKR similarly to wt PACT/RAX. (A) dsRNA-binding activity of wt PACT/RAX and lear-5J truncated protein was measured by a poly(I):poly(C)-agarose-binding assay with in vitro translated ³⁵S-labeled proteins. T, total input; B, proteins bound to poly(I):poly(C)-agarose. Competition lanes 5 and 6: competition with 100-fold molar excess of single-stranded RNA (ss) or dsRNA (ds). The faint band below the parent PACT/RAX band represents products of in vitro translation from an internal methionine codon in lane 1. (B) Quantification of the dsRNA-binding assay. Bands were quantified by phosphorimaging analysis, and percentage bound was calculated. Error bars: s.d. from three independent experiments. The P-value (0.003) calculated using statistical analyses indicated significant difference between percentage dsRNA-binding of wt (blue bar) and that of lear-5J mutant (red bar). (C) Co-immunoprecipitation of endogenous PKR and Flag-PACT/RAX or Flag-lear-5J overexpressed in HeLa cells. HeLa cells were transfected with Flag wt PACT/RAX or Flag-lear-5J in pCDNA3.1 expression constructs at 40% confluency and harvested 24 h post-transfection. Whole-cell extracts were immunoprecipitated at 4°C overnight using 100 ng anti-PKR antibody per immunoprecipitation. Samples were then analyzed via SDS-PAGE gel electrophoresis and western blot analysis probing for Flag-tagged wt PACT/RAX or lear-5J (co-IP panel) using monoclonal anti-Flag-M2 (Sigma-Aldrich) antibody. To ascertain that an equal amount of protein was immunoprecipitated, blots were re-probed using an anti-PKR antibody (IP panel). Input blots without immunoprecipitation demonstrate that equal amounts of each protein were present prior to immunoprecipitation. IP, immunoprecipitation; mAb, monoclonal antibody; mut, mutant; wt, wild type.

Fig. 3.

(A) Effects of lear-5J protein on PKR kinase activity. (A) PKR kinase activity assay was performed using PKR immunoprecipitated from HeLa cells, recombinant lear-5J and wt PACT/RAX proteins, and 1 μCi [γ-³²P] ATP per reaction. Either pure recombinant lear-5J (left) or wt PACT/RAX (right) protein was added as activator in the amounts indicated above each lane. Phosphorylated PKR was visualized after SDS-PAGE and phosphorimager analysis. (B) The truncated lear-5J protein inhibits PKR activation. PKR immunoprecipitated from HeLa cell extracts was activated with polyI:polyC (lanes 2-5) or 4 ng recombinant pure wt PACT/RAX protein (lanes 6-9). Increasing amounts of recombinant pure lear-5J protein (4-400 ng) were added (lanes 2-9) as indicated to assess the effect on PKR activity. Lane 1, PKR activity without any added activator; lane 2, PKR activity in the presence of polyI:polyC; lanes 3-5, PKR activity in the presence of polyI:polyC and 4 ng, 40 ng or 400 ng lear-5J protein; lane 6, PKR activity in the presence of 4 ng wt PACT/RAX protein; lanes 7-9, PKR activity in the presence of 4 ng wt PACT/RAX and 4 ng, 40 ng or 400 ng lear-5J protein. Phosphorylated proteins were analyzed by SDS-PAGE and phosphorimager analysis. C, control.

Fig. 4.

Tunicamycin-induced apoptosis is reduced in Prkra^lear-5J MEFs. (A) DNA fragmentation analysis. Mouse embryonic fibroblasts (MEFs) established from wt (BTBR T⁺ Itpr^tf/J) and Prkra^lear-5J mice were treated with 0.5 µg/ml tunicamycin for the indicated times. Lane 1, 100 bp marker ladder (M); lanes 2-4, wt (BTBR T⁺ Itpr^tf/J) MEFs; lanes 5-7, Prkra^lear-5J MEFs. Lanes 2 and 5, untreated cells; lanes 3, 4, 6 and 7, tunicamycin-treated cells. (B) Caspase-Glo 3/7 assay. MEFs established from WT (BTBR T⁺ Itpr^tf/J) mice and Prkra^lear-5J mice were treated with 0.5 µg/ml tunicamycin for the indicated times, and caspase 3/7 activities were measured. Blue bars, wt (BTBR T⁺ Itpr^tf/J) MEFs; red bars, Prkra^lear-5J MEFs. The P-values that were significant are indicated.

Fig. 5.

Lear-5J truncated mutant protein is present in mouse brains but not in MEFs. (A) Western blot analysis using brain extracts prepared from wt and Prkra^lear-5J brain samples. Blots were probed for PACT/RAX using a polyclonal antibody; the best of three representative blots is shown. The position of the full-length PACT/RAX and the truncated lear-5J protein is indicated by arrows. The blots were probed with β-actin antibody, which indicates equal loading. (B) Reverse transcriptase PCR (RT-PCR) using total RNA from the brains of the indicated Prkra^lear-5J genotypes using ribosomal protein S15 as the positive control to ascertain presence of an equal amount of total RNA in all samples. (C) Western blot analysis of cell extracts derived from MEFs of the indicated Prkra^lear-5J genotypes and HeLa cells overexpressing Flag-lear-5J protein. Blots were probed for PACT/RAX using a polyclonal antibody; the best of three representative blots is shown. The position of the full-length PACT/RAX and the truncated lear-5J protein is indicated by arrows. (D) RT-PCR using total RNA isolated from wt and homozygous Prkra^lear-5J MEFs. Ribosomal protein S15 was used as a positive control to ascertain that an equal amount of total RNA was analyzed in each sample.

Fig. 6.

PACT/RAX protein is abundantly expressed in mouse cerebellum and especially in Purkinje neurons. (A) Immunohistochemistry on day 28 sagittal section of wt C57BL/6 cerebellum using anti-PACT/RAX antibody. Brown staining indicates the presence of PACT/RAX protein. (B) Immunohistochemistry of the tissue described in A, showing colocalization of PACT/RAX with a Purkinje neuron-specific marker, calbindin. Red, PACT/RAX; blue, DAPI; green, calbindin.

Fig. 7.

The Prkra^lear-5J mutation affects cerebellar development and reduces arborization in Purkinje neurons. (A,B) Hematoxylin and Eosin staining on day 28 sagittal sections of wt (BTBR T⁺ Itpr^tf/J) and Prkra^lear-5J cerebellum. We examined six Prkra^lear-5J mouse cerebellums, which exhibited varying degrees of reduced foliation. (C-F) Immunohistochemistry of day 28 sagittal sections of mouse cerebellar tissue stained for the Purkinje neuron marker, calbindin (green), and nuclear stain DAPI (blue). Dashed line boxes in C and E indicate areas magnified in D and F, respectively. Scale bars: 100 μm (C,E) and 20 μm (D,F).

Fig. 8.

Prkra^lear-5J cerebellum exhibits dysregulation of eIF2α phosphorylation. (A) Immunohistochemistry of day 28 sagittal sections of wt (BTBR T⁺ Itpr^tf/J) (a,c,e) and Prkra^lear-5J (b,d,f) mouse cerebellum stained with the nuclear marker DAPI (blue) and for the protein of interest (green). a and b, DAPI (blue) and phosphorylated eIF2α (green); c and d: DAPI (blue) and CreP (green); e and f, DAPI (blue) and ATF4 (green). (B) Western blot analysis of wt (BTBR T⁺ Itpr^tf/J) and Prkra^lear-5J cerebellar extracts. Cerebellar extracts from two wt (BTBR T⁺ Itpr^tf/J) (lanes 1 and 2), one Prkra^lear-5J/+- heterozygous (lane 3) and four Prkra^{lear-5J/lear5-J} homozygous (lanes 4-7) mice collected on day 28 were analyzed with antibodies against phospho-eIF2α (p-eIF2α), total eIF2α, CreP, ATF4 and β-actin. The best of three representative blots for each is shown.