Deficiency in the Ubiquitin Conjugating Enzyme UBE2A in Alzheimer's Disease (AD) is Linked to Deficits in a Natural Circular miRNA-7 Sponge (circRNA; ciRS-7)

Abstract

Our understanding of the highly specialized functions for small non-coding single-stranded RNA (ssRNA) in the transcriptome of the human central nervous system (CNS) continues to evolve. Circular RNAs (circRNAs), a recently discovered class of ssRNA enriched in the brain and retina, are extremely stable and intrinsically resilient to degradation by exonuclease. Conventional methods of ssRNA, microRNA (miRNA), or messenger RNA (mRNA) detection and quantitation requiring free ribonucleotide ends may have considerably underestimated the quantity and significance of CNS circRNA in the CNS. Highly-specific small ssRNAs, such as the ~23 nucleotide (nt) Homo sapien microRNA-7 (hsa-miRNA-7; chr 9q21.32), are not only abundant in the human limbic system but are, in addition, associated with a ~1400 nt circRNA for miRNA-7 (ciRS-7) in the same anatomical region. Structurally, ciRS-7 contains about ~70 tandem anti-miRNA-7 sequences and acts as an endogenous, anti-complementary miRNA-7 "sponge" that attracts, binds, and, hence, quenches, natural miRNA-7 functions. Using a combination of DNA and miRNA array technologies, enhanced LED-Northern and Western blot hybridization, and the magnesium-dependent exoribonuclease and circRNA-sensitive probe RNaseR, here we provide evidence of a significantly misregulated ciRS-7-miRNA-7-UBE2A circuit in sporadic Alzheimer's disease (AD) neocortex (Brodmann A22) and hippocampal CA1. Deficits in ciRS-7-mediated "sponging events", resulting in excess ambient miRNA-7 appear to drive the selective down-regulation in the expression of miRNA-7-sensitive mRNA targets, such as that encoding the ubiquitin conjugating enzyme E2A (UBE2A; chr Xq24). UBE2A, which normally serves as a central effector in the ubiquitin-26S proteasome system, coordinates the clearance of amyloid peptides via proteolysis, is known to be depleted in sporadic AD brain and, hence, contributes to amyloid accumulation and the formation of senile plaque deposits. Dysfunction of circRNA-miRNA-mRNA regulatory systems appears to represent another important layer of epigenetic control over pathogenic gene expression programs in the human CNS that are targeted by the sporadic AD process.

Keywords: Alzheimer’s disease (AD); circular RNA (circRNA); genetic control; miRNA-7; proteasome; proteolysis; ubiquitin conjugating enzyme E2A (UBE2A).

Figure 1

A strong miRNA-7-UBE2A mRNA 3’-UTR interaction. (A) Complimentarity map of human miRNA-7 and the UBE2A mRNA 3′-UTR—an “|” represents a hydrogen bond between the miRNA-7 and the UBE2A mRNA 3’-UTR and “:” represents a partial hydrogen bond; “seed” sequences are highlighted in yellow for miRNA-7 and in red for the UBE2A mRNA 3′-UTR; an energy of association (E_A) between the miRNA-7 and the UBE2A mRNA 3′-UTR of −22.9 kcal/mol is very favorable thermodynamically; due to a deficit in ciRS-7 there are insufficient miRNA-7 sponging effects and an increase ambient miRNA-7 levels (Figure 2), resulting in a down-regulation in the expression of miRNA-7 targeted mRNA-3’-UTRs (such as UBE2A mRNA 3’-UTR) which is observed; other circRNAs may be involved; (B) levels of UBE2A protein in Brodmann area A22 and Alzheimer (ALZ) hippocampal CA1 compared to control (CON) brain postmortem samples; as in Figure 2, mean and one standard deviation are shown; UBE2A is observed to be reduced to 3.6- and 2.7-fold of controls, respectively, in Brodmann A22 (the superior temporal neocortex) and the hippocampal CA1 of AD brains and the results are highly significant; insufficient UBE2A might be expected to contribute to deficits in the ubiquitination cycle that helps to clear amyloid peptides via autophagy, thus contributing to amyloidogenesis in the AD brain; in (B) there were no significant differences in age, ApoE allele status, RNA quality (all RIN values were 8.1–9.0) or yield between the control (CON) or Alzheimer (ALZ) groups; control UBE2A levels in the Brodmann A22 (the superior temporal neocortex) and the hippocampal CA1 of AD were arbitrarily set to 10.0; a dashed horizontal line at 10.0 is included for ease of comparison; (* p < 0.001 (ANOVA).

Figure 2

Increased miRNA-7 and deficits in ciRS-7 in AD. (A): up-regulated miRNA-7 in AD brain; lack of the miRNA-7 (ciRS-7) sponge by insufficient ciRS-7 suggests the miRNA-7 would be increased in the same brain regions where ciRS-7 is down-regulated as is observed; data in (A) quantified in bar graph format in (B); down-regulation of a family of miRNA-7-sensitive mRNA targets in the sporadic AD brain includes UBE2A; other mRNA targets may be involved; (C): detection of circRNA for miRNA-7 (ciRS-7) in sporadic AD and age-matched control hippocampal CA1 (control (CON) N = 4; AD (ALZ) N = 6]; in (C) the single upper ciRS-7 (~1400 nt) band contains ~70 selectively conserved miRNA-7 binding sites as previously described [8,10,11]; a lower 28S RNA served as an internal loading and reference control; all samples depleted of rRNA were treated with 50 units of RNaseR prior to electrophoresis see [8,10,23]); (D): AD ciRS-7 is significantly reduced to ~0.18-fold of control (CON) in AD (ALZ) hippocampal CA1 (HIPP CA1), to ~0.22 of control in the superior temporal lobe (Brodmann A22; NCTX A22) but not the brain stem (BRAINSTEM); this implicates loss of miRNA-7 sponge effects, and ambient up-regulation of miRNA-7 in at least two anatomical areas targeted by the AD process; there were no significant differences between age for control or AD tissues; (mean +/- one standard deviation (SD) = 71.6 ± 6.3 years (N = 6; control); 73.5 ± 6.1 years (N = 12; AD)); all AD cases were from moderate-to-advanced stages of AD; all postmortem intervals were 2.3 h or less; there were no significant differences in age, ApoE allele status, RNA quality (all RIN values were 8.1–9.0) or yield between the control or AD groups; Figure 1C has been redrawn in part from [10]; values in (B) and (D) are normalized to control levels; a dashed horizontal line at 1.0 or 10.0 is included for ease of comparison; (* p < 0.01, ANOVA); ** p < 0.001 (ANOVA).

Figure 3

Highly schematicized illustration of deficits in the ubiquitination signaling pathway in AD that normally clears waste molecules (such as amyloid peptides) from the brain’s extracellular space: initially, a UBE1 enzyme activates, in an ATP-dependent manner, the 76 amino acid ubiquitin molecule by forming a high-energy thioester bond with ubiquitin; this E1-ubiqutin (UBE1) complex is next transferred to the ubiquitin conjugating enzyme (UBE2A); a ubiquitin ligase (UBE3) next positions the target substrate near the UBE2A enzyme allowing for the sequential transfer of ubiquitin; once a chain of four or more ubiquitin molecules is placed onto the substrate protein, the molecule is then targeted for proteolysis and degradation by the 26S proteasome; large and self-aggregating proteinaceous species (such as amyloids) once formed may be more difficult to ubiquitinate, sequester, and degrade [35]; a ciRS-7-driven miRNA-7-mediated lack of UBE2A near the center of this pathway would impair the trafficking of amyloid peptides and other neurotoxic end-stage molecules to the 26S proteasome complex [3,4,5,6,35,36]; note that in sporadic AD ciRS-7 is down-regulated (Figure 2) and ambient miRNA-7 is elevated, resulting in the down-regulation of UBE2A expression and an impairment at the UBE2A-ub stage (red rectangle) in the normal ubiquitination of end-stage molecules; this eventually leads to dysfunction and deficiency in the 26S proteasome system (downward wavy arrow at right).