Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 Jun 12.
Published in final edited form as: Cell Rep. 2014 May 22;7(5):1401–1409. doi: 10.1016/j.celrep.2014.04.050

HuD regulates coding and noncoding RNA to induce APP → Aβ processing

Min-Ju Kang 1, Kotb Abdelmohsen 1, Emmette R Hutchison 2, Sarah J Mitchell 3, Ioannis Grammatikakis 1, Rong Guo 1, Ji Heon Noh 1, Jennifer L Martindale 1, Xiaoling Yang 1, Eun Kyung Lee 1,8, Mohammad A Faghihi 5, Claes Wahlestedt 5, Juan C Troncoso 6, Olga Pletnikova 6, Nora Perrone-Bizzozero 7, Susan M Resnick 4, Rafael de Cabo 3, Mark P Mattson 2, Myriam Gorospe 1,*
PMCID: PMC4074355  NIHMSID: NIHMS593770  PMID: 24857657

Abstract

The primarily neuronal RNA-binding protein HuD is implicated in learning and memory. Here, we report the identification of several HuD target transcripts linked to Alzheimer’s disease (AD) pathogenesis. HuD interacted with the 3’-untranslated regions (UTRs) of APP mRNA (encoding amyloid precursor protein) and BACE1 mRNA (encoding β-site APP-cleaving enzyme) and increased the half-lives of these mRNAs. HuD also associated with and stabilized the long noncoding (lnc)RNA BACE1AS, which partly complements BACE1 mRNA and enhances BACE1 expression. Consistent with HuD promoting production of APP and APP-cleaving enzyme, the levels of APP, BACE1, BACE1AS, and Aβ were higher in the brain of HuD-overexpressing mice. Importantly, cortex (superior temporal gyrus) from AD patients displayed significantly higher levels of HuD, and accordingly elevated APP, BACE1, BACE1AS, and Aβ than did cortical tissue from healthy age-matched individuals. We propose that HuD jointly promotes the production of APP and the cleavage of its amyloidogenic fragment, Aβ.

INTRODUCTION

The post-transcriptional regulation of gene expression underlies many aspects of mammalian physiology and pathology. The two main groups of post-transcriptional regulators, RNA-binding proteins (RBPs) and noncoding RNAs (Moore, 2005; Morris et al., 2010), have been implicated in all steps controlling gene expression after transcription: pre-mRNA splicing and maturation, and mRNA transport, editing, stability, storage, and translation.

Among the noncoding RNAs that control gene expression post-transcriptionally, the best characterized molecules are microRNAs (miRNAs, 22 nt in length), which typically associate with the 3’-untranslated region (UTR) of target mRNAs and repress their translation and/or stability (Fabian et al., 2010). Long noncoding RNAs (lncRNAs) are also gaining recognition as post-transcriptional regulators of gene expression. Through complementary base-pairing, lncRNAs can modulate the turnover and translation rate of target mRNAs; in the absence of complementarity, lncRNAs can suppress precursor mRNA splicing and translation by acting as decoys or competitors of RBPs and miRNAs (Yoon et al., 2012a).

RBPs associate with a wide range of coding and noncoding RNAs and thus modulate many critical functions of the cell, including proliferation, survival, differentiation, motility, senescence, and apoptosis (Glisovic et al., 2008). Among them, the elav (embryonic lethal abnormal vision)/Hu (human antigen) group of proteins, has been implicated primarily in controlling the stability and translation of target mRNAs. The Hu family comprises a ubiquitous member (HuR) and three predominantly neuronal members (HuB, HuC, HuD). Elav/Hu proteins generally bind to U- and AU-rich RNA elements in target transcripts with which they associate via three highly conserved RNA recognition motifs (RRMs 1–3) (Hinman and Lou, 2008; Pascale et al., 2008).

Unlike HuR, which is primarily nuclear, HuD is abundantly present in the cytoplasm. HuD expression is restricted to a few tissues, mainly neurons, gonads, and pancreatic β cells (Good, 1995, Lee et al., 2012). Several lines of evidence indicate that in cultured neurons, HuD promotes neurite outgrowth (Kasashima et al., 1999; Abdelmohsen et al., 2010), but the physiological role of HuD in animals appears to be complex. While adult HuD-knockout (KO) mice do not exhibit morphological defects, HuD KO embryos display transient impairment in cranial nerve development, and neurospheres derived from these mice generate fewer neurons compared to wild-type mice (Akamatsu et al., 2005). At the same time, expression of HuB, HuC, and HuD specifically increases in areas of mouse and rat brain associated with spatial learning, implicating these Hu proteins in learning and memory. In these tissues, elevated HuD is associated with enhanced production of GAP-43 (growth-associated protein-43), encoded by a HuD target mRNA (Anderson et al., 2001; Pascale et al., 2004). The roles of HuD in neuronal development and memory have been reviewed (Deschênes-Furry et al., 2006; Pascale et al., 2008; Perrone-Bizzozero et al., 2011).

HuD targets include many mRNAs that encode proteins preferentially expressed in neurons (e.g., (GAP-43, acetylcholinesterase, tau, PSD-95, neuroserpin, musashi-1, and HuD itself), as well as proteins expressed in other tissues (e.g., c-Myc, N-myc, RhoA, c-Fos, VEGF, p21, p27, Bcl-2, NCAM1, and MARCKS) (Deschênes-Furry et al., 2006; Pascale et al., 2008; Abdelmohsen et al., 2010; Bolognani et al., 2010). With the exception of p27 and insulin mRNAs, whose translation is repressed by HuD (Kullmann et al., 2002; Lee et al., 2010), HuD generally promotes the expression of target mRNAs. A recent survey of HuD target transcripts in human neuroblastoma cells (Abdelmohsen et al., 2010) revealed a number of HuD-interacting mRNAs implicated in the synthesis and processing of amyloid precursor protein (APP) into its amyloidogenic fragment, Aβ. HuD binds APP mRNA and BACE1 mRNA, the latter encoding the β-secretase which cleaves APP in the critical first proteolytic processing step that leads to the generation of Aβ. HuD also bound to and increased the abundance of BACE1AS, a lncRNA that stabilizes BACE1 mRNA and promotes BACE1 expression (Faghihi et al., 2008). Our findings indicate that HuD may coordinate the production and cleavage of APP and further suggest that this regulatory paradigm contributes to Alzheimer’s disease pathogenesis, characterized by the accumulation of toxic aggregates of Aβ peptide.

RESULTS

HuD associates mRNAs involved in APP processing

RNAs associated with HuD were identified by immunoprecipitation (IP) of ribonucleoprotein (RNP) complexes using an anti-HuD antibody in parallel with control IgG IP (RIP analysis). The interaction of HuD in the IP material (Fig. 1A) with bound RNAs was assayed by reverse transcription (RT) and subsequent real-time, quantitative (q)PCR amplification. An earlier survey in the human neuroblastoma BE(2)-M17 cells (Abdelmohsen et al., 2010), revealed that APP mRNA was a potential target of HuD. Experiments to investigate this possibility directly revealed that APP mRNA was significantly enriched in HuD IP samples compared with IgG IP samples, and additionally showed that several HuD-bound mRNAs encoded proteases that cleave APP to generate Aβ peptide. Among them, the β-site APP-cleaving enzyme (BACE1) mRNA was also significantly enriched, as previously observed (Bolognani et al., 2010), while HuD associated less prominently with mRNAs encoding components of the γ-secretase complex (PSEN1 and PSEN2 mRNAs, encoding presenilins, and APH1A and APH1B mRNAs, encoding presenilin-stabilization factors) (Fig. 1B). The PEN2 mRNA (encoding presenilin enhancer 2) and the NCSTN mRNA (encoding nicastrin, a component of the γ secretase protein complex) showed no significant enrichment in HuD IP (Fig. 1B). We thus focused on analyzing the interaction of HuD with APP and BACE1 mRNAs in human neuroblastoma SK-N-F1 cells.

Figure 1. HuD binds to APP 3’UTR, enhances APP mRNA and stability and translation.

Figure 1

Open in a new tab

(A) Analysis of HuD in SK-N-F1 cells following immunoprecipitation (IP) using IgG and anti-HuD antibodies and Western blot (WB) analysis. HuD (arrowhead), immunoglobulin heavy chain (HC), and molecular weight (MW) markers are indicated. (B) Limited analysis (using RIP) to survey HuD-interacting mRNAs encoding proteins with roles in APP production or processing. Following IP as in (A), mRNA levels in the IP materials were measured by RT-qPCR analysis, normalized to GAPDH mRNA levels in each IP reaction, and represented as ‘Fold enrichment’ relative to IgG IP. (C) Top, schematic of APP mRNA, depicting the 5’UTR, CR, and 3’UTR, and the different biotinylated RNA segments (gray lines) tested for binding to HuD after pulldown using streptavidin beads; HuD was detected by Western blot analysis. Biotinylated GAPDH 3’UTR was included as negative control; ‘Input’, 20 µg of whole-cell lysate. (D,E) 48 h after transfecting Ctrl or HuD siRNAs, the levels of proteins APP, HuD, and loading control β-actin were assessed by Western blot analysis (D), and the reduction in APP mRNA levels by RT-qPCR analysis (normalized to GAPDH mRNA levels) (E). (F) Cells transfected as in (D) were treated with actinomycin D and the levels of APP mRNA and GAPDH mRNA (encoding a housekeeping protein) were assessed by RT-qPCR; the half-lives (t1/2) of APP and GAPDH mRNA were quantified by measuring the time required for transcript levels to decline to 50% of their original abundance relative to time 0 h. (G) Polysomes were prepared from cells transfected as in (D) by fractionating cytoplasmic extracts through sucrose gradients. Arrow: direction of sedimentation; –, no ribosomal components; small (40S) and large (60S) ribosomal subunits and monosomes (80S) in fractions 2–4, and progressively larger polysomes, ranging from low- to high-molecular-weight (LMW, HMW) in fractions 6–11 (right). The distribution of APP and GAPDH mRNAs was studied by RT-qPCR analysis of RNA in gradient fractions, and represented as % of total RNA in the gradient (left). (H) Left, Schematic of reporters prepared to analyze the influence of APP 3’UTR on gene expression; Right, 12 h after transfecting SK-N-F1 cells with Ctrl or HuD siRNAs, cells were further transfected with psiCHECK2 or psiCHECK-APP(3’UTR) and 24 h later, luciferase activity was measured (RL/FL) for each transfected plasmid, and normalized to luciferase activity (RL/FL) in Ctrl siRNA (left). RL mRNA and FL mRNA levels in each transfection group, normalized to GAPDH mRNA, were quantified by RT-qPCR analysis and plotted relative to the changes in the psiCHECK2 group (right). Data are representative of 3 independent experiments; graphs in (B,E,F,H) display the means and S.D. from 3 independent experiments. *, p<0.05.

To identify areas of interaction of HuD with APP mRNA, biotinylated segments spanning the 5’UTR, coding region (CR) and 3’UTR of the APP mRNA were synthesized in vitro and incubated with cytoplasmic lysates of SK-N-F1 cells. After pulldown using streptavidin-coated beads, HuD association with the biotinylated transcripts was assessed by Western blot analysis. As shown in Fig. 1C, several APP 3’UTR segments associated with HuD in vitro, but segments of the APP CR, the APP 5’UTR, or a negative control RNA (GAPDH 3’UTR) did not. Although HuD bound multiple APP 3’UTR segments, the most robust affinity and regulation (Fig. S1) was mapped to segment 3’d. The consequences of the interaction of HuD with the APP mRNA were assessed by silencing HuD; 48 h after transfecting SK-N-F1 cells with a small interfering (si)RNA directed to HuD, the decline in HuD levels caused a marked decrease in the levels of APP protein and APP mRNA, as assessed by WB and RT-qPCR analyses, respectively (Figs. 1D,E).

HuD enhances the stability and translation of APP mRNA

Since HuD was shown to stabilize a number of target mRNAs, we investigated if the loss in APP mRNA after silencing HuD was due at least in part to changes in APP mRNA stability. After silencing HuD expression in SK-N-F1 cells, actinomycin D was used to inhibit de novo transcription; the time needed for APP mRNA to be reduced to 50% of its initial abundance [its half-life (t1/2)] was then calculated by measuring APP mRNA levels using RT-qPCR and normalizing to 18S rRNA levels. As shown in Fig. 1F, APP mRNA half-life in control cells (Ctrl siRNA, >6.0 hours) was much longer than that measured in HuD siRNA cells (~4.0 hours). The half-life of GAPDH mRNA, a stable mRNA that encodes a housekeeping protein, was not shortened by HuD silencing (Fig. 1F). HuD was also shown to promote mRNA translation (Fukao et al., 2009); to investigate if HuD also affected APP translation, we silenced HuD in SK-N-F1 cells and quantified the fraction of APP mRNA associated with the translation machinery in each transfection group. Cytoplasmic extracts from the Ctrl siRNA and HuD siRNA groups were fractionated on sucrose gradients, and the relative abundance of APP mRNA in each fraction indicated the association of APP mRNA with the cellular polysomes and hence its translation status. The polysome fractionation pattern was identical in both groups, suggesting that silencing HuD does not affect translation globally. In Ctrl siRNA cells, APP mRNA levels were very low in non-translating and low-translating fractions of the gradient (fractions 1 to 4, where free RNA and 40S and 60S subunits, as well as 80S monosomes, are found), but it was abundant in the actively translating fractions of the gradient (fractions 5 to 10, spanning low- and high-molecular-weight polysomes) (Fig. 1G) and peaking at fraction 9. However, in HuD siRNA-transfected cells, APP mRNA showed a leftward shift on the gradient, peaking at fraction 8 indicating that APP mRNA formed on average smaller polysomes after silencing HuD. These results agree with a role for HuD both elevating APP mRNA abundance and enhancing the translation of APP mRNA.

These effects were further examined using heterologous luciferase reporter vectors that expressed renilla luciferase (RL) lacking or containing the APP 3’UTR [psiCHECK2, psiCHECK2-APP(3’UTR), respectively]. The ratio of RL to firefly luciferase (FL) (encoded by an internal control reporter transcript within the same plasmid) was set as 1 for the parent vector (psiCHECK2). The lower RL/FL ratios for HuD siRNA-transfected cells (~48%) relative to those for Ctrl siRNA-transfected cells indicated that the presence of APP 3’UTR reduced luciferase production when HuD was silenced (Fig. 1H, left). As shown in Fig. 1H right, there was a parallel reduction in RL mRNA levels compared with FL mRNA levels in HuD silenced cells (~52%), indicating that accelerated decay of the RL-APP(3’UTR) chimeric mRNA contributed to the decrease in RL expression. In sum, HuD enhanced both the stability and translation of APP mRNA.

HuD stabilizes BACE1 mRNA and BACE1AS RNA

Next, we examined the interaction of HuD with BACE1 mRNA. HuD associated in vitro with partial biotinylated segments of the 3’UTR but not the CR or 5’UTR of BACE1 mRNA, showing preferential binding to a distal BACE1 3’UTR segment (3’b) (Fig. 2A, Fig. S2). The levels of BACE1 mRNA were lower in the HuD siRNA group and this reduction was achieved at least in part through a decline in BACE1 mRNA stability (Fig. 2B,C), just as seen for APP mRNA (Fig. 1F); non-targets of HuD, including mRNAs encoding γ-secretases, did not show this trend (Fig. S2). BACE1 protein levels were correspondingly lower, but the translation of BACE1 mRNA did not appear to be affected by HuD silencing (Fig. 2D,E).

Figure 2. HuD associates with BACE1 mRNA.

Figure 2

Open in a new tab

(A) Top, schematic of BACE1 mRNA, indicating the different biotinylated RNA segments (gray underlines) that were prepared and tested for binding to HuD after pulldown using streptavidin beads; HuD was detected by Western blot analysis as explained in Fig. 1C. (B–E) 48 h after transfecting SK-N-F1 cells with the siRNAs indicated, BACE1 mRNA levels were measured by RT-qPCR analysis (*, p<0.05) (B) and the half-life (t1/2) of BACE1 mRNA (C) was quantified as explained in Fig. 1F. The levels of BACE1 protein were assessed by Western blot analysis (D) and the relative distribution of BACE1 mRNA on polysomes (E) was studied as explained in Fig. 1G.

The stability of BACE1 mRNA was not previously reported to be affected by other RBPs, but it was shown to be stabilized via interaction with the lncRNA BACE1AS (Faghihi et al., 2008). Thus, we asked if the present regulatory paradigm involved BACE1AS. RIP analysis revealed that BACE1AS was robustly present in HuD RNP complexes (Fig. 3A) and that HuD associated in vitro with partial biotinylated segments of BACE1AS RNA; negative controls GAPDH 3’UTR and the lncRNA 7SL were included (Fig. 3B). This interaction likely stabilized BACE1AS, since silencing HuD lowered the steady-state levels (Fig. 3C) and the half-life of BACE1AS (Fig. 3D). Further evidence that BACE1 mRNA and BACE1AS interacted in our study system was obtained by engineering a plasmid that expressed a chimeric BACE1AS tagged with MS2 RNA hairpins (pMS2-BACE1AS). By 36 h after transfecting SK-N-F1 cells with pMS2 or pMS2-BACE1AS into along with plasmid pMS2-YFP, which expressed a fluorescent fusion protein, MS2-YFP (Lee et al., 2010; Fig. 3E), RIP analysis was carried out using anti-YFP antibody to bring down the YFP-MS2 protein bound to MS2-bearing RNAs. As shown in Fig. 3F, BACE1 mRNA, as measured by RT-qPCR analysis in the IP samples, was significantly more abundant in MS2-BACE1AS IP than in control IP (pMS2 transfection group), indicating that BACE1 mRNA and BACE1AS associated physically in SK-N-F1 cells.

Figure 3. LncRNA BACE1AS is the target of HuD and BACE1 mRNA.

Figure 3

Open in a new tab

(A) RIP analysis of the interaction of BACE1AS RNA with HuD. (B) Schematic of human BACE1AS RNA, depicting the 5’ and 3’ segments (5’S and 3’S) as well as the double-stranded (DS) segment. The interaction of biotinylated BACE1AS RNA with HuD was assayed by biotin pulldown (biotinylated segments assayed are shown, gray underlines). FL, full-length BACE1AS; GAPDH 3’UTR and 7SL were included as negative controls. (C,D) 48 h after transfecting SK-N-F1 cells with the siRNAs indicated, BACE1AS RNA levels were measured by RT-qPCR analysis (C) and the half-life (t1/2) of BACE1AS was quantified as in Fig. 1F (D). (E,F) SK-N-F1 cells were transfected with the plasmids shown (schematic): plasmid pMS2, which expressed 24 repeats of MS2 hairpins, and pMS2-BACE1AS, expressing full-length BACE1AS; plasmid pMS2-YFP, expressing a fusion protein (MS2-YFP) capable of detecting MS2-tagged RNA, was cotransfected (E). By 24 h after transfection of the plasmids in (E), lysates were subjected to RIP analysis using an antibody against YFP. The presence of BACE1 mRNA in IP samples from each transfection group was assessed by RT-qPCR analysis (F). (G) Schematic of proposed influence of HuD upon the expression and β-cleavage of APP: (i) HuD binds to APP mRNA and promotes APP mRNA stabilization and translation, (ii) HuD binds to BACE1 mRNA and stabilizes it, and (iii) HuD binds to BACE1AS and stabilizes it, in turn increasing BACE1 mRNA stability and BACE1 production. Arrowhead: subsequent cleavage by γ-secretase that releases Aβ peptide (blue). In (A,C,D,F) the graphs reflect the means and S.D. from 3 independent experiments; *, p<0.05.

Collectively, these results suggest a model whereby HuD jointly controls three transcripts on the route to generating amyloidogenic Aβ: (i) HuD binds to APP mRNA and enhances APP levels, (ii) HuD binds to BACE1 mRNA and promotes BACE1 production, and (iii) HuD binds to BACE1AS RNA and increases its levels, further augmenting BACE1 production. By acting on functionally related RNAs, HuD serves as a master coordinator of Aβ production (Fig. 3G).

HuD influence on APP→Aβ production in the brain

Based on the model proposed presented in Fig. 3G, we hypothesized that HuD may affect Aβ production in the brain. We tested this possibility using two different models. In transgenic mice overexpressing HuD (HuD-Tg mice) as a tagged protein (Myc-HuD; Fig. 4A), the levels of Aβ40, as measured by ELISA (Materials and Methods) were significantly higher than in WT mice (Fig. 4B) and the levels of Elavl4 mRNA (encoding HuD), App mRNA, Bace1 mRNA, and Bace1as RNA were higher in hippocampus, cortex, and cerebellum, relative to the levels in age-matched WT mice (Fig. 4C).

Figure 4. In mice overexpressing HuD, the levels of App mRNA, Bace1 mRNA, Bace1as RNA, and Aβ are elevated; in Alzheimer’s Disease brains, HuD, HuD target transcripts and encoded proteins, and Aβ are elevated.

Figure 4

Open in a new tab

(A) The levels of tagged HuD in whole-brain lysate from mice, three wild-type, three overexpressing the Myc-HuD transgene (9-month-old females) were assessed by Western blot analysis. Ponceau S staining was used to assess sample loading. (B) ELISA was used to measure soluble Aβ in WT and HuD-Tg (whole brain lysates, 3 mice per genotype). (C) The levels of Elavl4 mRNA (encoding HuD), App mRNA, Bace1 mRNA, and Bace1as in each mouse brain group (n=3 mice per genotype) were measured by RT-qPCR; data were normalized to 18S rRNA. (D,E) Lysates prepared from superior temporal gyrus (STG) from normal subjects (n=10) and from AD subjects (n=10) were assayed for levels of ELAVL4 mRNA, APP mRNA, BACE1 mRNA, and BACE1AS RNA using RT-qPCR (D) and for the encoded proteins [6 normal, 7 AD] by Western blot analysis followed by densitometric analysis (E). RNA levels were normalized to GAPDH mRNA levels, protein levels to β-actin levels. (F) The levels of Aβ40 and Aβ42 were assayed in 10 normal and 10 AD brains (STG) by using ELISA. In (B–F) the graphs reflect the means and S.D. from 3 independent experiments; p values are indicated in (D,E); *, p<0.05.

We also studied the levels of HuD in brains of Alzheimer’s Disease (AD) patients. Brain sections from the superior temporal gyrus (STG) from 10 persons with AD and 10 age-matched healthy individuals were used for RNA and protein analysis. RNA prepared from these samples was analyzed by RT-qPCR; as shown in Fig. 4D, ELAVL4 mRNA, and BACE1AS RNA were markedly higher in AD brains relative to healthy brains, while APP mRNA and BACE1 mRNA were moderately higher in AD patients. The relative levels of protein followed these general differences (Fig. 4E; Faghihi et al., 2008), with greater than twofold higher levels of HuD and BACE1 in AD brains. In keeping with the higher levels of BACE1, Aβ peptide (Aβ40 and Aβ42) was also more abundant in AD brains (Fig. 4F). In summary, in brains of HuD-overexpressing mice, the levels of APP, BACE1, and Aβ40 correlated with the heightened abundance of HuD, while in human AD brains, HuD levels were significantly elevated and correlated with higher levels of BACE1, BACE1AS, and Aβ.

DISCUSSION

Our results indicate that HuD controls the expression of three mRNAs affecting APP production and processing into Aβ. The three-pronged actions of HuD include binding to APP mRNA, causing it to be more stable and translated, binding to BACE1 mRNA, increasing its stability and hence BACE1 production, and binding to and stabilizing BACE1AS, which further contributes to the production of BACE1. In keeping with this regulatory paradigm, HuD levels were positively linked to Aβ levels in three systems: in neuroblastoma SK-N-F1 cells expressing normal versus silenced HuD levels, in the brains of WT and HuD-overexpressing mice, and in the brains from AD and normal subjects.

The discovery that HuD influences the production of the key pathogenic peptide Aβ raises the immediate question of what determines the levels of HuD in neuronal tissues. Among the possible transcriptional regulators of HuD expression is FoxO1; the repression of ELAVL4 gene transcription in pancreatic β cells was relieved by exposure to high glucose or insulin (Lee et al., 2012). Whether FoxO1 also suppresses ELAVL4 gene transcription in neuronal tissues, and whether this regulation is dependent on glucose/insulin awaits further study. Post-transcriptional regulators of HuD production include the microRNA miR-375, which binds to ELAVL4 mRNA, lowers its stability, and represses HuD translation (Abdelmohsen et al. 2010; Supplemental Text).

The notion that a single RBP, HuD, can jointly control the expression of three transcripts (APP mRNA, BACE1 mRNA, BACE1AS RNA) that are functionally related supports the ‘post-transcriptional operon/regulon’ model put forth by Keene and Tenenbaum (2002). According to this model, a single RBP can associate with and coordinate the expression of multiple mRNAs that share a given RNA sequence with affinity for the RBP and whose encoded gene products are implicated in a specific cellular function. To establish a simple analogy with a ‘transcriptional’ operon/regulon, the RBP functions as the equivalent of a transcription factor which binds to a shared DNA sequence present in the promoter of genes encoding proteins with related functions (as in eukaryotes) or synthesizes a polycistronic RNA whose individual RNA components encode proteins functionally linked. In light of our findings, the traditional post-transcriptional operon/regulon model would help to explain HuD function; we proposed a slightly revised version of the model that includes non-coding RNA.

Our studies did not identify a specific RNA element present in the three transcripts. However, most RBPs do not have strict sequence requirements and instead associate with loosely variable RNA elements, often within the context of a given secondary structure. Approaches such as PAR-CLIP identification of HuD-bound RNAs can give us a more precise view of the subset of transcripts with which HuD interacts. A more complete understanding of the collection of HuD-interacting coding and noncoding RNAs will be particularly informative as we strive to understand the underlying causes of Aβ processing in AD.

EXPERIMENTAL PROCEDURES

Cell culture, siRNA, and plasmids

Human neuroblastoma SK-N-F1 cells were cultured in DMEM (Invitrogen) supplemented with 10% (v/v) FBS and antibiotics. The plasmids and siRNAs used are described in Supplemental Text.

Protein analysis

Western blot analysis and fractionation of polyribosomes are explained in Supplemental Text. Aβ levels (Aβ40 and Aβ42) in SK-N-F1 cells were measured in conditioned media and Aβ in human brain samples was assayed directly from protein lysates; human Aβ was measured by ELISA (Invitrogen KHB3481 and KHB3441) and the manufacturer’s protocol included incubation with guanidine; [methyl(phenyl)-λ3-oxidanyl]formic acid solution. Aβ levels in mouse brain lysates were analyzed by ELISA (Invitrogen).

RNA Analysis

RNA-binding assays RIP (ribonucleoprotein immunoprecipitation), biotin pulldown, and mRNA half-life measurements are explained in Supplemental Text. Trizol (Invitrogen) was used to extract total RNA and acidic phenol (Ambion) was used to extract RNA for RIP analysis (Lee et al., 2010). Reverse transcription (RT) was performed using random hexamers and Maxima reverse transcriptase (Thermo Scientific) and real-time, quantitative PCR (qPCR) using gene-specific primers (Supplemental Text). RT-qPCR was performed using SYBR green master mix (Kapa Biosystems) in an Applied Biosystems 7300 instrument.

Human subjects

The sample consists of twenty participants (19 females and 1 male) from the Baltimore Longitudinal Study of Aging (NIA, NIH); cognitive status was determined based on standardized consensus diagnostic procedures for the BLSA (Driscoll et al., 2012) and eventually came to autopsy. Participants were between 55 and 98 (AD) and between 56 and 95 (normal). All studies were approved by the local institutional review boards, and all participants gave written informed consent prior to each assessment. In addition, next of kin or legally designated power of attorney provided consent for autopsy. Superior temporal gyrus (STG) regions of the brain were used for protein and RNA analysis.

Animals

HuD transgenic [HuD-Tg, a kind gift from Dr. N.I. Perrone-Bizzozero] and wildtype mice in the C57BL/6J background were obtained from inhouse breeding colonies of the NIA, Baltimore MD (details in Supplemental Text).

Supplementary Material

01

HIGHLIGHTS.

  • RNA-binding protein HuD is elevated in Alzheimer’s disease brain

  • HuD binds to the 3’UTRs of APP and BACE1 mRNAs and increases APP and BACE1 levels

  • HuD binds to and elevates long noncoding RNA BACE1AS

  • HuD orchestrates APP synthesis and processing into Aβ

ACKNOWLEDGEMENTS

We thank H. Cai for providing reagents and information, J.H. Yoon for advice, and D. Phillips-Boyer, D. Nines, and J. Lucas for exceptional animal care. This work was supported in part by the NIA-IRP, NIH, and by NIDA R01DA034097 to NPB.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

AUTHORS’ CONTRIBUTIONS

MJK, KA, ERH, SJM, IG, RG, JHN, and JLM performed experiments; MJK, KA, ERH, IG, RG, JHN, OP, SMR, analyzed the data; KA, SJM, XY, EKL, MAF, CW, JCT, OP, NPB, RdC, MPM, and MG contributed reagents and expertise; and MJK, SMR, RdC, MPM, and MG wrote the paper.

REFERENCES

  1. Abdelmohsen K, Hutchison ER, Lee EK, Kuwano Y, Kim MM, Masuda K, Srikantan S, Subaran SS, Marasa BS, Mattson MP, Gorospe M. miR-375 inhibits differentiation of neurites by lowering HuD levels. Mol. Cell. Biol. 2010;30:4197–4210. doi: 10.1128/MCB.00316-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Akamatsu W, Fujihara H, Mitsuhashi T, Yano M, Shibata S, Hayakawa Y, Okano HJ, Sakakibara S, Takano H, Takano T, et al. The RNA-binding protein HuD regulates neuronal cell identity and maturation. Proc. Natl. Acad. Sci. USA. 2005;102:4625–4630. doi: 10.1073/pnas.0407523102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Anderson KD, Sengupta J, Morin M, Neve RL, Valenzuela CF, Perrone-Bizzozero NI. Overexpression of HuD accelerates neurite outgrowth and increases GAP-43 mRNA expression in cortical neurons and retinoic acid-induced embryonic stem cells in vitro. Exp. Neurol. 2001;168:250–258. doi: 10.1006/exnr.2000.7599. [DOI] [PubMed] [Google Scholar]
  4. Bolognani F, Contente-Cuomo T, Perrone-Bizzozero NI. Novel recognition motifs and biological functions of the RNA-binding protein HuD revealed by genome-wide identification of its targets. Nucleic Acids Res. 2010;38:117–130. doi: 10.1093/nar/gkp863. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Deschênes-Furry J, Perrone-Bizzozero NI, Jasmin BJ. The RNA-binding protein HuD: a regulator of neuronal differentiation, maintenance and plasticity. Bioessays. 2006;28:822–833. doi: 10.1002/bies.20449. [DOI] [PubMed] [Google Scholar]
  6. Driscoll I, Troncoso JC, Rudow G, Sojkova J, Pletnikova O, Zhou Y, Kraut MA, Ferrucci L, Mathis CA, Klunk WE, O'Brien RJ, Davatzikos C, Wong DF, Resnick SM. Correspondence between in vivo (11)C-PiB-PET amyloid imaging and postmortem, region-matched assessment of plaques. Acta Neuropathol. 2012;124:823–831. doi: 10.1007/s00401-012-1025-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Fabian MR, Sonenberg N, Filipowicz W. Regulation of mRNA translation and stability by microRNAs. Annu. Rev. Biochem. 2010;79:351–379. doi: 10.1146/annurev-biochem-060308-103103. [DOI] [PubMed] [Google Scholar]
  8. Faghihi MA, Modarresi F, Khalil AM, Wood DE, Sahagan BG, Morgan TE, Finch CE, St Laurent G, 3rd, Kenny PJ, Wahlestedt C. Expression of a noncoding RNA is elevated in Alzheimer's disease and drives rapid feed-forward regulation of beta-secretase. Nat Med. 2008;14:723–730. doi: 10.1038/nm1784. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Fukao A, Sasano Y, Imataka H, Inoue K, Sakamoto H, Sonenberg N, Thoma C, Fujiwara T. The ELAV protein HuD stimulates cap-dependent translation in a Poly(A)- and eIF4A-dependent manner. Mol. Cell. 2009;36:1007–1017. doi: 10.1016/j.molcel.2009.11.013. [DOI] [PubMed] [Google Scholar]
  10. Glisovic T, Bachorik JL, Yong J, Dreyfuss G. RNA-binding proteins and post-transcriptional gene regulation. FEBS Lett. 2008;582:1977–1986. doi: 10.1016/j.febslet.2008.03.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Good PJ. A conserved family of elav-like genes in vertebrates. Proc. Natl. Acad. Sci. U.S.A. 1995;92:4557–4561. doi: 10.1073/pnas.92.10.4557. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Hinman MN, Lou H. Diverse molecular functions of Hu proteins. Cell Mol. Life Sci. 2008;65:3168–3181. doi: 10.1007/s00018-008-8252-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kasashima K, Terashima K, Yamamoto K, Sakashita E, Sakamoto H. Cytoplasmic localization is required for the mammalian ELAV-like protein HuD to induce neuronal differentiation. Genes Cells. 1999;4:667–683. doi: 10.1046/j.1365-2443.1999.00292.x. [DOI] [PubMed] [Google Scholar]
  14. Keene JD, Tenenbaum SA. Eukaryotic mRNPs may represent posttranscriptional operons. Mol. Cell. 2002;9:1161–1167. doi: 10.1016/s1097-2765(02)00559-2. [DOI] [PubMed] [Google Scholar]
  15. Kullmann M, Göpfert U, Siewe B, Hengst L. ELAV/Hu proteins inhibit p27 translation via an IRES element in the p27 5'UTR. Genes Dev. 2002;16:3087–3099. doi: 10.1101/gad.248902. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Lee EK, Kim W, Tominaga K, Martindale JL, Yang X, Subaran SS, Carlson OD, Mercken EM, Kulkarni RN, Akamatsu W, Okano H, Perrone-Bizzozero NI, de Cabo R, Egan JM, Gorospe M. RNA-binding protein HuD controls insulin translation. Mol. Cell. 2012;45:826–835. doi: 10.1016/j.molcel.2012.01.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Lee EK, Kim HH, Kuwano Y, Abdelmohsen K, Srikantan S, Subaran SS, Gleichmann M, Mughal MR, Martindale JL, Yang X, Worley PF, Mattson MP, Gorospe M. hnRNP C promotes APP translation by competing with FMRP for APP mRNA recruitment to P bodies. Nat. Struct. Mol. Biol. 2010;17:732–739. doi: 10.1038/nsmb.1815. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Moore MJ. From birth to death: the complex lives of eukaryotic mRNAs. Science. 2005;309:1514–1518. doi: 10.1126/science.1111443. [DOI] [PubMed] [Google Scholar]
  19. Morris AR, Mukherjee N, Keene JD. Systematic analysis of posttranscriptional gene expression. Wiley Interdiscip. Rev. Syst. Biol. Med. 2010;2:162–180. doi: 10.1002/wsbm.54. [DOI] [PubMed] [Google Scholar]
  20. Pascale A, Amadio M, Quattrone A. Defining a neuron: neuronal ELAV proteins. Cell. Mol. Life Sci. 2008;65:128–140. doi: 10.1007/s00018-007-7017-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Pascale A, Gusev PA, Amadio M, Dottorini T, Govoni S, Alkon DL, Quattrone A. Increase of the RNA-binding protein HuD and posttranscriptional up-regulation of the GAP-43 gene during spatial memory. Proc. Natl. Acad. Sci. USA. 2004;101:1217–1222. doi: 10.1073/pnas.0307674100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Perrone-Bizzozero NI, Tanner DC, Mounce J, Bolognani F. Increased expression of axogenesis-related genes and mossy fibre length in dentate granule cells from adult HuD overexpressor mice. ASN Neuro. 2011;3:259–270. doi: 10.1042/AN20110015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Yoon JH, Abdelmohsen K, Gorospe M. Posttranscriptional Gene Regulation by Long Noncoding RNA. J Mol Biol. 2013;425:3723–3730. doi: 10.1016/j.jmb.2012.11.024. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS593770-supplement-01.pdf (2.1MB, pdf)

RESOURCES