RNA editing of microtubule-associated protein tau circular RNAs promotes their translation and tau tangle formation.

Aggregation of the microtubule-associated protein tau characterizes tauopathies, including Alzheimer's disease and frontotemporal lobar degeneration (FTLD-Tau). Gene expression regulation of tau is complex and incompletely understood. Here we report that the human tau gene (MAPT) generates two circular RNAs (circRNAs) through backsplicing of exon 12 to either exon 7 (12→7 circRNA) or exon 10 (12→10 circRNA). Both circRNAs lack stop codons. The 12→7 circRNA contains one start codon and is translated in a rolling circle, generating a protein consisting of multimers of the microtubule-binding repeats R1-R4. For the 12→10 circRNA, a start codon can be introduced by two FTLD-Tau mutations, generating a protein consisting of multimers of the microtubule-binding repeats R2-R4, suggesting that mutations causing FTLD may act in part through tau circRNAs. Adenosine to inosine RNA editing dramatically increases translation of circRNAs and, in the 12→10 circRNA, RNA editing generates a translational start codon by changing AUA to AUI. Circular tau proteins self-aggregate and promote aggregation of linear tau proteins. Our data indicate that adenosine to inosine RNA editing initiates translation of human circular tau RNAs, which may contribute to tauopathies.

Influence of FTDP-17 mutants on circular tau RNAs.

At least 53 mutations in the microtubule associated protein tau gene (MAPT) have been identified that cause frontotemporal dementia. 47 of these mutations are localized between exons 7 and 13. They could thus affect the formation of circular RNAs (circRNAs) from the MAPT gene that occurs through backsplicing from exon 12 to either exon 10 or exon 7. We analyzed representative mutants and found that five FTDP-17 mutations increase the formation of 12➔7 circRNA and three different mutations increase the amount of 12➔10 circRNA. CircRNAs are translated after undergoing adenosine to inosine RNA editing, catalyzed by ADAR enzymes. We found that the interferon induced ADAR1-p150 isoform has the strongest effect on circTau RNA translation. ADAR1-p150 activity had a stronger effect on circTau RNA expression and strongly decreased 12➔7 circRNA, but unexpectedly increased 12➔10 circRNA. In both cases, ADAR-activity strongly promoted translation of circTau RNAs. Unexpectedly, we found that the 12➔7 circTau protein interacts with eukaryotic initiation factor 4B (eIF4B), which is reduced by four FTDP-17 mutations located in the second microtubule domain. These are the first studies of the effect of human mutations on circular RNA formation and translation. They show that point mutations influence circRNA expression levels, likely through changes in pre-mRNA structures. The effect of the mutations is surpassed by editing of the circular RNAs, leading to their translation. Thus, circular RNAs and their editing status should be considered when analyzing FTDP-17 mutations.

Influence of FTDP-17 mutants on circular Tau RNAs.

At least 53 mutations in the microtubule associated protein tau gene (MAPT) have been identified that cause frontotemporal dementia. 47 of these mutations are localized between exons 7 and 13. They could thus affect the formation of circular RNAs (circRNAs) from the MAPT gene that occur through backsplicing from exon 12 to either exon 10 or exon 7. We analyzed representative mutants and found that five FTDP-17 mutations increase the formation of 12➔7 circRNA and three different mutations increase the amount of 12➔10 circRNA. CircRNAs are translated after undergoing adenosine to inosine RNA editing, catalyzed by ADAR enzymes. We found that the interferon induced ADAR1-p150 isoform has the strongest effect on circTau RNA translation. ADAR1-p150 activity had a stronger effect on circTau RNA expression and strongly decreased 12➔7 circRNA, but unexpectedly increased 12➔10 circRNA. In both cases, ADAR-activity strongly promoted translation of circTau RNAs. Unexpectedly, we found that the 12➔7 circTau protein interacts with eukaryotic initiation factor 4B (eIF4B), which is reduced by four FTDP-17 mutations located in the second microtubule domain. These are the first studies of the effect of human mutations on circular RNA formation and translation. They show that point mutations influence circRNA expression levels, likely through changes in the secondary pre-mRNA structures. The effect of the mutations is surpassed by editing of the circular RNAs, leading to their translation. Thus, circular RNAs and their editing status should be considered when analyzing FTDP-17 mutations. Highlights: 47/53 known FTDP-17 mutations are located in regions that could influence generation of circular RNAs from the MAPT geneCircular Tau RNAs are translated after adenosine to inosine RNA editing, most effectively caused by ADAR1-p150FTDP-17 mutations influence both circTau RNA and circTau protein expression levelsCircTau protein expression levels do not correlate with circTau RNA expression levelsCircTau proteins bind to eukaryotic initiation factor 4B, which is antagonized by FTDP-17 mutations in exon 10.

Alzheimer's disease pathogenetic progression is associated with changes in regulated retained introns and editing of circular RNAs.

Introduction: The molecular changes leading to Alzheimer's disease (AD) progression are poorly understood. A decisive factor in the disease occurs when neurofibrillary tangles (NFT) composed of microtubule associated protein tau (MAPT) form in the entorhinal cortex and then spread throughout the brain. Methods: We therefore determined mRNA and circular RNA changes during AD progression, comparing Braak NFT stages I-VI. Total RNA was isolated from human brain (entorhinal and frontotemporal cortex). Poly(A)+ RNA was subjected to Nanopore sequencing, and total RNA was analyzed by standard Illumina sequencing. Circular RNAs were sequenced from RNase R treated and rRNA depleted total RNA. The sequences were analyzed using different bioinformatic tools, and expression constructs for circRNAs were analyzed in transfection experiments. Results: We detected 11,873 circRNAs of which 276 correlated with Braak NFT stages. Adenosine to inosine RNA editing increased about threefold in circRNAs during AD progression. Importantly, this correlation cannot be detected with mRNAs. CircMAN2A1 expression correlated with AD progression and transfection experiments indicated that RNA editing promoted its translation using start codons out of frame with linear mRNAs, which generates novel proteins. Discussion: Thus, we identified novel regulated retained introns that correlate with NFT Braak stages and provide evidence for a role of translated circRNAs in AD development.

Chicken cognin interactome and the memory of filial imprinting.

Visual imprinting is a learning process whereby young animals come to prefer a visual stimulus after exposure to it (training). The intermediate medial mesopallium in the domestic chick forebrain is critical for visual imprinting and contributes to molecular regulation of memory formation. Criteria used to infer that a change following training is learning-related have been formulated and published. Cognin (protein disulphide isomerase) is one of several identified plasma membrane and mitochondrial proteins that are upregulated in a learning-related way 24 hours after training. Since virtually nothing is known about the cognin interactome, we have used immunoaffinity chromatography and mass spectrometry to identify proteins that interact with cognin in the cytoplasmic and plasma membrane-mitochondrial fractions. As the learning-related upregulation of cognin has been shown to occur in the plasma membrane-mitochondrial fraction and not in the cytoplasmic fraction, we studied the effect of training on three cognin-interacting partners in the plasma membrane-mitochondrial fraction: the b5 subunit of mitochondrial ATP synthase and the alpha-2 and alpha-3 subunits of sodium-potassium ATPase. Learning-related upregulation was found in the left intermediate medial mesopallium 24 hours after training for the b5 subunit of mitochondrial ATP synthase and the alpha-2 subunit of sodium-potassium ATPase. The hemispheric asymmetry revealed here is consistent with the predominance of many other learning-related effects in the left intermediate medial mesopallium. The alpha-2 subunit of sodium-potassium ATPase is mainly expressed in astrocytes, supporting a role for these glial cells in memory.

Long-Term Effects of Myoinositol on Behavioural Seizures and Biochemical Changes Evoked by Kainic Acid Induced Epileptogenesis.

Epilepsy is one of the most devastating neurological diseases and despite significant efforts there is no cure available. Occurrence of spontaneous seizures in epilepsy is preceded by numerous functional and structural pathophysiological reorganizations in the brain-a process called epileptogenesis. Treatment strategies targeting this process may be efficient for preventing spontaneous recurrent seizures (SRS) in epilepsy, or for modification of disease progression. We have previously shown that (i) myoinositol (MI) pretreatment significantly decreases severity of acute seizures (status epilepticus: SE) induced by kainic acid (KA) in experimental animals and (ii) that daily post-SE administration of MI for 4 weeks prevents certain biochemical changes triggered by SE. However it was not established whether such MI treatment also exerts long-term effects on the frequency of SRS. In the present study we have shown that, in KA-induced post-SE epilepsy model in rats, MI treatment for 28 days reduces frequency and duration of behavioural SRS not only during the treatment, but also after its termination for the following 4 weeks. Moreover, MI has significant effects on molecular changes in the hippocampus, including mi-RNA expression spectrum, as well as mRNA levels of sodium-MI transporter and LRRC8A subunit of the volume regulated anionic channel. Taken together, these data suggest that molecular changes induced by MI treatment may counteract epileptogenesis. Thus, here we provide data indicating antiepileptogenic properties of MI, which further supports the idea of developing new antiepileptogenic and disease modifying drug that targets MI system.

Micro-RNAs, their target proteins, predispositions and the memory of filial imprinting.

Visual imprinting is a learning process whereby young animals come to prefer a visual stimulus after exposure to it (training). The intermediate medial mesopallium (IMM) in the domestic chick forebrain is critical for visual imprinting and contributes to molecular regulation of memory formation. We investigated the role of micro-RNAs (miRNAs) in such regulation. Twenty-four hours after training, miRNA spectra in the left IMM were compared between chicks with high preference scores (strong memory for imprinting stimulus), and chicks with low preference scores (weak memory for imprinting stimulus). Using criteria of significance and expression level, we chose gga-miR-130b-3p for further study and found that down-regulation correlated with learning strength. No effect was detected in posterior nidopallium, a region not involved in imprinting. We studied two targets of gga-miR-130b-3p, cytoplasmic polyadenylation element binding proteins 1 (CPEB-1) and 3 (CPEB-3), in two subcellular fractions (P2 membrane-mitochondrial and cytoplasmic) of IMM and posterior nidopallium. Only in the left IMM was a learning-related effect observed, in membrane CPEB-3. Variances from the regression with preference score and untrained chicks suggest that, in the IMM, gga-miR-130b-3p level reflects a predisposition, i.e. capacity to learn, whereas P2 membrane-mitochondrial CPEB-3 is up-regulated in a learning-specific way.

Mitochondrial fusion and fission proteins and the recognition memory of imprinting in domestic chicks.

Visual imprinting is a learning process through which young, visually naive animals come to recognize a visual stimulus by being exposed to it (training) and subsequently approach the stimulus in preference to others. A large body of evidence indicates that a restricted part of the forebrain, the intermediate medial mesopallium (IMM), is a memory region for visual imprinting in the domestic chick. Previous studies have shown learning-related up-regulation of several mitochondrial proteins in the IMM 24 h after training. Learning-related increases in transcription factors involved in mitochondrial biogenesis were found without significant change in mitochondrial DNA copy number, but the issue of whether mitochondrial fusion or fission processes change with learning was unresolved. The present study enquired whether proteins involved in mitochondrial fusion and fission contribute to memory following imprinting. Tissue was sampled from the left and right IMM, and the left and right posterior pole of the nidopallium (a control brain region not involved in imprinting). The amounts of the following proteins were measured by Western immunoblotting 24 h after training: mitochondrial mitofusin-1 (MTF-1, as indicator of mitochondrial fusion), membrane dynamin-related protein-1 (DRP-1, as indicator of mitochondrial fission) and cytoplasmic DRP-1. Learning-related increases in MTF-1 and DRP-1 were observed bilaterally in the IMM, but not in either side of the posterior pole of the nidopallium. Cytoplasmic DRP-1 was not changed significantly in any region studied. The results implicate increased, balanced levels of mitochondrial fusion and fission in memory formation up to 24 h after training.Supplementary Video Abstract (Supplemental digital content 1, http://links.lww.com/WNR/A446).

A Proteomic Study of Memory After Imprinting in the Domestic Chick.

The intermediate and medial mesopallium (IMM) of the domestic chick forebrain has previously been shown to be a memory system for visual imprinting. Learning-related changes occur in certain plasma membrane and mitochondrial proteins in the IMM. Two-dimensional gel electrophoresis/mass spectrometry has been employed to identify more comprehensively learning-related expression of proteins in the membrane-mitochondrial fraction of the IMM 24 h after training. We inquired whether amounts of these proteins in the IMM and a control region (posterior pole of the nidopallium, PPN) are correlated with a behavioral estimate of memory for the imprinting stimulus. Learning-related increases in amounts of the following proteins were found in the left IMM, but not the right IMM or the left or right PPN: (i) membrane cognin; (ii) a protein resembling the P32 subunit of splicing factor SF2; (iii) voltage-dependent anionic channel-1; (iv) dynamin-1; (v) heterogeneous nuclear ribonucleoprotein A2/B1. Learning-related increases in some transcription factors involved in mitochondrial biogenesis were also found, without significant change in mitochondrial DNA copy number. The results indicate that the molecular processes involved in learning and memory underlying imprinting include protein stabilization, increased mRNA trafficking, synaptic vesicle recycling, and specific changes in the mitochondrial proteome.