Autoregulation of TDP-43 mRNA levels involves interplay between transcription, splicing, and alternative polyA site selection

Abstract

TDP-43 is a critical RNA-binding factor associated with pre-mRNA splicing in mammals. Its expression is tightly autoregulated, with loss of this regulation implicated in human neuropathology. We demonstrate that TDP-43 overexpression in humans and mice activates a 3' untranslated region (UTR) intron, resulting in excision of the proximal polyA site (PAS) pA(1). This activates a cryptic PAS that prevents TDP-43 expression through a nuclear retention mechanism. Superimposed on this process, overexpression of TDP-43 blocks recognition of pA(1) by competing with CstF-64 for PAS binding. Overall, we uncover complex interplay between transcription, splicing, and 3' end processing to effect autoregulation of TDP-43.

Figure 1.

Cis-acting elements and molecular events in TDP-43 autoregulation. (A) Schematic diagram of TDP-43 illustrating locations of the stop codon (tag), PASs (pA_1–4), the TDPBR, and splicing events (in coding sequences, filled lines; in 3′ UTR, dotted lines). Coding regions (black boxes), untranslated sequences (gray boxes), and introns (connecting black lines) are indicated. RT–PCR primers are indicated. (B, left) Schematic representation of the GFP-3′UTR reporters are shown. (Middle) Western blots of GFP reporter protein in normal (−Tet) and TDP-43 overexpression conditions (+Tet) are shown. (Right) The bar chart shows densitometric quantification of protein expression, normalized against DiGFP expression. Mean values are plotted; error bars indicate the SD from at least three independent experiments. (C) 3′ RACE analysis (−Tet and +Tet) for x7 (left) and x7 Δgt-ag (right) reporters using the specific GFPfw and anchor primers (see A). (Right) The schematic representation denotes amplified mRNA isoforms (bands 1 and 2) confirmed by sequence analysis. (D) 3′ RACE analysis of endogenous TDP-43 (−Tet and +Tet). The forward primer annealed to intron 6 (a), while the anchor was used as the reverse primer (see A). (Right) Amplified mRNA isoforms are indicated.

Figure 2.

Endogenous TDP-43 in mice and humans; mRNA intron 7 splicing, PAS usage, and mRNAs nuclear retention. (A) Schematic representation of human TDP-43. Arrows show the positions of the primers used. 3′ RACE analysis of endogenous TDP-43 before (−Tet) and following (+Tet) TDP-43 transgene induction performed on nuclear (N) and cytoplasmic (C) RNA fractions or following CHX treatment. The forward primer was annealed to intron 6 (a), while the anchor was used as the reverse primer. (Right) The same samples were checked for the amplification of the endogenous variant V2 mRNA using primers b and c as indicated. All PCR products were sequenced to establish their identity. Distribution of p84 and tubulin was analyzed by Western blotting. (B, left) Schematic diagram of mouse TDP-43 showing sequence conservation between humans and mice and detailed alignment of the sequence corresponding to intron 7 splice sites and the various PAS. (Right) 3′ RACE analysis of endogenous murine TDP-43 from control (wild-type [wt]) and transgenic tg A315T mouse brains. The forward primer is annealed on intron 6 (named ma), while the anchor was used as the reverse primer. A representation of amplified mRNA isoforms (bands 1 and 2) is also shown as confirmed by sequencing. HPRT mRNA amplification was used as a control. (Bottom right) Endogenous and transgenic tg A315T protein expression was measured by Western blotting using anti-TDP-43, with anti-tubulin as control.

Figure 3.

Pol II stalling on the TDPBR sequence. (A) Schematic diagram of TDP-43 showing positions of ChIP probes used. (B) BrU nuclear run-on assay measuring nascent transcription in different regions of endogenous TDP-43 under normal conditions (−Tet) and following TDP-43 overexpression (+Tet). (C,D) ChIP analysis using N20 Pol II and H5 Pol II CTDser2P antibodies probing the same regions of TDP-43.

Figure 4.

Competition between TDP-43 and CstF-64 across the pA₁ site. (A) Nucleotide sequence of TDP-43 pA₁ DSE (top) and a hit frequency plot of TDP-43 CLIP data (bottom) (from the University of California at Santa Cruz Genome Browser). (B) RIP analysis of TDP-43 and CstF-64 proteins on endogenous TDP-43 across pA₁ under normal (−Tet) and TDP-43 overexpression (+Tet) conditions. (C) RIP for TDP-43 and CstF-64 proteins across the HPRT PAS used as control.

Figure 5.

Model of TDP-43 autoregulatory mechanism. Under normal conditions, Pol II synthesizes the nascent TDP-43 mRNA from TDP-43, generating two main TDP-43 mRNA isoforms using either pA₁ or pA₄, which results in efficient TDP-43 protein expression. When the levels of nuclear TDP-43 rise, there is an increase in the binding of TDP-43 to the TDPBR. This promotes a dual effect: increased intron 7 splicing with the physical removal of the pA₁ signal, and direct competition with the cleavage polyA factor Cstf-64 over pA₁. Moreover, Pol II stalling may lead to premature termination of transcription and rapid degradation.