Arc 3' UTR Splicing Leads to Dual and Antagonistic Effects in Fine-Tuning Arc Expression Upon BDNF Signaling

Abstract

Activity-regulated cytoskeletal associated protein (Arc) is an immediate-early gene critically involved in synaptic plasticity and memory consolidation. Arc mRNA is rapidly induced by synaptic activation and a portion is locally translated in dendrites where it modulates synaptic strength. Being an activity-dependent effector of homeostatic balance, regulation of Arc is uniquely tuned to result in short-lived bursts of expression. Cis-Acting elements that control its transitory expression post-transcriptionally reside primarily in Arc mRNA 3' UTR. These include two conserved introns which distinctively modulate Arc mRNA stability by targeting it for destruction via the nonsense mediated decay pathway. Here, we further investigated how splicing of the Arc mRNA 3' UTR region contributes to modulate Arc expression in cultured neurons. Unexpectedly, upon induction with brain derived neurotrophic factor, translational efficiency of a luciferase reporter construct harboring Arc 3' UTR is significantly upregulated and this effect is dependent on splicing of Arc introns. We find that, eIF2α dephosphorylation, mTOR, ERK, PKC, and PKA activity are key to this process. Additionally, CREB-dependent transcription is required to couple Arc 3' UTR-splicing to its translational upregulation, suggesting the involvement of de novo transcribed trans-acting factors. Overall, splicing of Arc 3' UTR exerts a dual and unique effect in fine-tuning Arc expression upon synaptic signaling: while inducing mRNA decay to limit the time window of Arc expression, it also elicits translation of the decaying mRNA. This antagonistic effect likely contributes to the achievement of a confined yet efficient burst of Arc protein expression, facilitating its role as an effector of synapse-specific plasticity.

Keywords: 3′ UTR; Arc; BDNF; EJC; nonsense mediated mRNA decay; plasticity; post-transcriptional regulation; splicing.

FIGURE 1

Luciferase constructs harboring Arc 3′ UTR are downregulated and subjected to NMD. (A) Schematic representation of the luciferase constructs utilized in this study. Upper panel: schematic Arc gene architecture is depicted, including the general contribution of each region to Arc gene expression. Lower panel: schematic representation of the six Renilla luciferase constructs generated. All constructs are under the control of a Thymidine kinase promoter and include a constitutive intron in the their 5′ UTR. Arc mRNA 3′ UTR was inserted within the pRLTK plasmid (pRLTK) downstream of the Renilla luciferase coding sequence and upstream of SV40 poly(A) sequence (SV40 pA), generating ArcUTR construct. The ArcUTR-noI construct was obtained omitting Arc 3′ UTR introns. SV40 pA was substituted with Arc polyadenylation sequences (Arc pA) to generate Arc UTR-ApA, Arc UTR-noI-ApA, and pRLTK-ApA. (B) Luciferase activity of the indicated constructs tested 24 h post-transfection of cortical neurons (8–9 d.i.v.). The histogram reports Renilla luciferase activity normalized to Firefly activity, expressed as fold change compared to Arc UTR. The results are expressed as the mean ± standard error (SEM) from three independent experiments with four biological replicates each (n = 12). Statistical significance is calculated relative to Arc UTR or Arc UTR-ApA for each set. Student’s t-test: ^∗∗P < 0.01, ^∗∗∗P < 0.001. Multiple comparisons were performed using t-tests with Bonferroni correction based on the number of comparisons.

FIGURE 2

Brain-derived neurotrophic factor induces a significant upregulation of Arc UTR construct, dependent on splicing of its 3′ UTR introns. (A) Luciferase assay of cortical neurons transfected with Arc UTR or Arc UTR-noI constructs and subjected to a time course chronic treatment with BDNF (100 ng/ml). (B) Luciferase assay of cortical neurons transfected with the Arc UTR construct and incubated with BDNF (100 ng/ml) for 5, 10, or 30 min, followed by thorough washes and incubation in media for a total of 4 h. Luciferase activity obtained upon 4 h chronic BDNF treatment is shown for comparison. (C) Luciferase assay of cortical neurons transfected with the Arc UTR construct and treated with the indicated compounds (treatment details, see section “Materials and Methods”). (D) Luciferase assay of cortical and hippocampal neurons transfected with the Arc UTR construct and incubated with BDNF (100 ng/ml) for 4 h. (A–D) Luciferase assays reporting Renilla luciferase activity normalized to Firefly activity, expressed as fold change compared to the “untreated” sample. All the results are expressed as the mean ± standard error (SEM) from at least three independent experiments with two/three biological replicates each (n between 6 and 43). Student’s t-test: ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001, ns, non-significant. Multiple comparisons were performed using t-tests with Bonferroni correction based on the number of comparisons.

FIGURE 3

3′ UTR splicing is necessary but not sufficient to confer responsiveness to BDNF. (A) Schematic representation of the Renilla luciferase constructs harboring control 3′ UTR regions. The genomic or the cDNA-derived 3′ UTR region of rat STRN4 gene was cloned downstream of the Renilla luciferase coding region generating respectively an intron-containing (Strn4 UTR) and an intron-less control construct (Strn4 UTR noI). Similarly, construct Cntr UTR and Cntr UTR-noI harbor three exons and two small introns of the Cyc1 gene coding region or the corresponding intron-less cDNA fragment downstream of the Renilla ORF. (B) Luciferase assay of cortical neurons transfected with the indicated constructs and treated for 4 h with BDNF (100 ng/ml) or left untreated. For each set of plasmids, asterisks denote statistical significance compared to the vehicle-treated intron-containing construct. Renilla luciferase values for the Strn4 and Cntr constructs were normalized to Firefly and expressed as fold change compared to Arc UTR “untreated” sample. Values are the mean ± standard error (SEM) from at least two independent experiments with three replicates (n = 6). (C) Luciferase signals from rat cortical neurons transfected with the Arc UTR construct or a plasmid harboring the human Arc UTR sequence downstream of the Renilla luciferase gene (hArc UTR). Neurons were either left untreated or incubated with BDNF (100 ng/ml) for 4 h or for 30 min followed by thorough washes and incubation in media for a total of 4 h. Values obtained with Arc UTR construct are reported for comparison and were obtained as indicated in Figure 2. Histograms represent Renilla luciferase activity normalized to Firefly activity, expressed as fold change compared to the corresponding “untreated” sample. (B,C) Student’s t-test: ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001, ns, non-significant. Multiple comparisons were performed using t-tests with Bonferroni correction based on the number of comparisons.

FIGURE 4

Splicing-dependent upregulation of Arc UTR is not affected by deletion of the primary DTE but is partially inhibited by deletion of a ELAV/miR-19 binding site. (A) Schematic representation of the intron-containing and intron-less reporter deletion mutants, in which either the primary DTE or the putative ELAV/miR-19 binding site were deleted. (B) Schematic representation of the rat Arc 3′ UTR region targeted by rno-miR-19a (nt 2809–2831 of NM_019361.1), as shown on TargetScan, release 7.1. Vertical lines indicate the seed pairing within the 8mer (Lewis et al., 2005). The overlapping consensus sequence for the ELAV/Hu family of proteins is shown in red. (C) Luciferase assay of neurons transfected with the indicated constructs treated for 4 h with BDNF (100 ng/ml) or left untreated. Renilla luciferase values were normalized to Firefly and expressed as fold change compared to Arc UTR “untreated” sample. Bars represent the mean ± standard error (SEM) from at least four independent experiments (n between 9 and 15). Student’s t-test: ^∗∗P < 0.01, ^∗∗∗P < 0.001, ns, non-significant. Multiple comparisons were performed using t-tests with Bonferroni correction based on the number of comparisons.

FIGURE 5

Brain-derived neurotrophic factor promotes translation of Arc UTR reporter while not affecting mRNA stability, splicing pattern or NMD efficiency. (A) qRT-PCR analysis of RNAs extracted from neurons transfected with the indicated constructs and treated with BDNF for 4 h or left untreated. Renilla reporter mRNA levels were normalized to Firefly mRNA levels and are expressed as fold change compared to the untreated ARC UTR sample. Bars represent the mean ± standard error (SEM) from 10 biological replicates and two technical replicates each (n = 20). Asterisks denote statistical significance compared to the untreated intron-containing construct. (B) RT-PCR assay to analyze the splicing pattern of the Arc UTR and Arc UTR-noI reporter mRNAs prior and upon BDNF treatment. (Upper) Schematic representation of the oligonucleotides utilized in the RT-PCR and in the PCR reactions, spanning the Arc UTR intronic region and specific to the Renilla reporter. (Lower) RT-PCR amplification of RNAs extracted from neurons transfected with the Arc UTR (1,2) or the Arc UTR-noI reporter (3,4) and treated with BDNF for 4 h (2,4) or left untreated (1,3). As a size reference of spliced versus unspliced mRNA, RT-PCR products were resolved next to PCR amplifications obtained using Arc UTR (5) or Arc UTR-noI (6) plasmid DNA as template. (C) Luciferase assay of neurons transfected with the Arc UTR construct and treated for 4 h with BDNF (100 ng/ml) or left untreated. The indicated drugs (CHX and Anisomycin) were added to the cells 30 min prior to the treatment. Renilla luciferase values were normalized to Firefly and are expressed as fold change compared to Arc UTR “untreated” sample. Error bars represent the standard error (SEM) from at least four independent experiments (n between 8 and 12). Asterisks denote statistical significance compared to the “no drug” sample treated with BDNF for 4 h. Student’s t-test: ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001, ns, non-significant. Multiple comparisons were performed using t-tests with Bonferroni correction based on the number of comparisons.

FIGURE 6

Brain-derived neurotrophic factor treatment increases the association of ARC UTR reporter with translating polyribosomes in a splicing-dependent manner. (A; upper panel) Representative polysome profiles of rat cortical neurons co-transfected with the Firefly Luciferase construct and with Arc UTR or Arc UTR-noI Renilla constructs and incubated with BDNF for 4 h or left untreated. Cytosolic lysates were separated on a 15–50% sucrose gradient to separate mRNPs, monosome and polysome fractions. Twelve fractions were collected with UV monitoring of RNA levels at A254. Fraction numbers are indicated at the top of the panel. RNAs from each fraction were extracted together with a spike-in RNA transcript to control and normalize for recovery efficiency. RNAs were then subjected to real time RT-PCR to detect the relative distribution on the polysome gradient of our Renilla reporter mRNA as a consequence of splicing and upon BDNF treatment. (A; lower panel) qRT-PCR of the indicated reporter mRNAs recovered from each fraction was resolved on agarose gel to visualize their distribution along the sucrose gradients before and after BDNF treatment. The bottom lane shows a representative qRT-PCR of the spike-in RNA recovered from each polysomal fraction and utilized to normalize qRT-PCR results. (B) Quantitative RT-PCR was used to measure Arc UTR, Arc UTR-noI, and Firefly mRNA levels from each fraction of the polysome gradients, as described in Section “Materials and Methods.” Profiles from neurons treated with BDNF (light gray) or left untreated (dark gray) are superimposed. Data are plotted as a fraction of the total recovered from the gradient, and are normalized for spike-in RNA recovery from each fraction. Error bars represent the standard error (SEM) from two independent polysomes gradients for each transfected reporter and three technical replicates (n = 6).

FIGURE 7

Splicing-dependent translational upregulation of ARC UTR constructs requires the activity of ERK, PKA and PKC kinases, the mTOR/S6K1 pathway, eIF2α dephosphorylation and CREB-dependent transcription. (A,B) Luciferase assay of neurons transfected with the Arc UTR construct and treated for 4 h with BDNF (100 ng/ml) or left untreated. Inhibitors for the indicated kinases or receptors or transcription factors were added to the cells 30 min prior to treatment with BDNF or with vehicle. Treatment detail in Section “Materials and Methods.” Renilla luciferase values were normalized to Firefly and are expressed as fold change compared to Arc UTR “no drug – untreated” sample. Error bars represent the standard error (SEM) from at least three independent experiments (n between 6 and 14). Asterisks denote statistical significance compared to the “no drug” sample treated with BDNF for 4 h. Student’s t-test: ^∗∗P < 0.01, ^∗∗∗P < 0.001, ns, non-significant. Multiple comparisons were performed using t-tests with Bonferroni correction based on the number of comparisons.

FIGURE 8

Suggested model depicting how 3′ UTR splicing may participate in fine-tuning Arc activity-dependent expression. Induction with BDNF (or NT3) leads to CREB activation via PKA, PKC and ERK signaling and eIF2α dephosphorylation. CREB, in turn, induces transcription of target genes, including Arc and a putative trans-acting factor required to couple Arc 3′ UTR splicing and EJC deposition with translational activation. Following export to the cytoplasm and translational de-repression, Arc mRNA undergoes a first round of translation, eliciting NMD of the mRNA (splicing-dependent mRNA decay). A pool of Arc mRNA “escapes” NMD and the 3′ UTR-bound EJCs mediate translational activation of these transcripts. This process may be directly modulated by mTOR, eIF2α dephosphorylation, PKA, PKC, and ERK and likely requires a CREB-dependent trans-acting factor (splicing-dependent translational activation; see section “Discussion”).