G-quadruplexes originating from evolutionary conserved L1 elements interfere with neuronal gene expression in Alzheimer’s disease

Abstract

DNA sequences containing consecutive guanines organized in 4-interspaced tandem repeats can form stable single-stranded secondary structures, called G-quadruplexes (G4). Herein, we report that the Polycomb group protein BMI1 is enriched at heterochromatin regions containing putative G4 DNA sequences, and that G4 structures accumulate in cells with reduced BMI1 expression and/or relaxed chromatin, including sporadic Alzheimer’s disease (AD) neurons. In AD neurons, G4 structures preferentially accumulate in lamina-associated domains, and this is rescued by re-establishing chromatin compaction. ChIP-seq analyses reveal that G4 peaks correspond to evolutionary conserved Long Interspersed Element-1 (L1) sequences predicted to be transcriptionally active. Hence, G4 structures co-localize with RNAPII, and inhibition of transcription can reverse the G4 phenotype without affecting chromatin’s state, thus uncoupling both components. Intragenic G4 structures affecting splicing events are furthermore associated with reduced neuronal gene expression in AD. Active L1 sequences are thus at the origin of most G4 structures observed in human neurons.

Transposable elements as new players in neurodegenerative diseases

“…TE promoters could also serve as alternative promoters yielding chimeric transcripts due to the sense and antisense promoters present in the 5′UTR of LINE‐1 initiating bidirectional transcription [179]. It was recently shown that G‐quadruplexes originating from evolutionarily conserved LINE‐1 elements interfere with neuronal gene expression in AD [181]. Some of these mechanisms through which TEs can alter host gene expression are illustrated in Fig.…”

G-Quadruplexes in Neurobiology and Virology: Functional Roles and Potential Therapeutic Approaches

“…This result suggests that G4 structures alter gene expression and are drastically reduced in AD neurons by more tightly compacting heterochromatin. 67 …”

A tensor decomposition-based integrated analysis applicable to multiple gene expression profiles without sample matching

“…This dataset [28], which is denoted as dataset 3 (i.e., k = 3) in this study, was composed of eight samples, four of which were AD cell lines and the other four of which were normal cell lines. Because four AD cell lines were composed of two sets of two samples using two distinct cell lines, in this study, this data set is treated as including three categorical classes when coincidence with singular value vectors or latent variables were investigated; two samples were taken from the first AD cell line, two samples were taken from the second AD cell line, and four samples were taken from normal cell lines.…”