|Home | About | Journals | Submit | Contact Us | Français|
Amyloid precursor protein (APP) regulates neuronal synapse function and its cleavage product Aβ is linked to Alzheimer’s disease. Here, we present evidence that RNA-binding proteins (RBPs) heterogeneous nuclear ribonucleoprotein (hnRNP) C and fragile-X mental retardation protein (FMRP) associate with the same APP mRNA coding region element and influence APP translation competitively and in opposite directions. Silencing hnRNP C increased FMRP binding to APP mRNA and repressed APP translation, while silencing FMRP enhanced hnRNP C binding and promoted translation. Repression of APP translation was linked to colocalization of FMRP and tagged APP mRNA within processing bodies (PBs); this colocalization was abrogated by hnRNP C overexpression or FMRP silencing. Our findings indicate that FMRP represses translation by recruiting APP mRNA to PBs, while hnRNP C promotes APP translation by displacing FMRP, thereby relieving the translational block.
Amyloid precursor protein (APP) is a transmembrane protein implicated in synapse formation and synaptic plasticity1–3. The secreted extracellular domain of APP (sAPPα) has growth-factor properties and promotes neuritogenesis. Cleavage of APP by β- and γ-secretases releases neurotoxic peptides, including Aβ, whose accumulation is directly linked to the pathogenesis of neurodegenerative disorders such as Alzheimer’s disease (AD). Several studies also support the notion that overproduction of APP underlies AD4–8. Elevated APP mRNA levels can result from altered APP transcription, although the specific transcription factors involved remain elusive9–11. By contrast, there is extensive evidence that APP expression is potently regulated by post-transcriptional mechanisms such as APP mRNA stabilization and APP translation12–18, indicating that the regulation of APP mRNA metabolism is an important event in AD pathophysiology.
Post-transcriptional processes are major mechanisms by which mammalian cells control gene expression19. Changes in mRNA turnover and translation rates are particularly important for altering the levels of expressed proteins20. These events are governed by two major types of trans-binding factors that interact with the mRNA: turnover- and translation-regulatory RNA-binding proteins (TTR RBPs) and noncoding (nc)RNAs such as microRNAs. TTR RBPs and ncRNAs bind to cis elements on the mRNA, frequently at the 5′- and 3′-untranslated regions (UTRs). APP expression was shown to be influenced by 3′UTR cis-elements that constitute the target sites of several microRNAs12,13, as well as by TTR RBPs that increased APP mRNA stability (hnRNP C) or promoted APP mRNA decay (nucleolin)14,15. APP translation was also modulated by 5′UTR cis-elements, including an iron-responsive element (IRE) and an internal ribosome entry site (IRES)16,17. Additionally, APP translation was shown to be modulated by a TTR RBP that associated with a coding region (CR) cis-element and repressed APP translation, the fragile X mental retardation protein (FMRP)18.
In a recent survey, we identified several RBPs that interacted with different regions of human APP mRNA. Among the novel interactions discovered, the RBP hnRNP C was found to bind to the same segment of the APP CR as did FMRP. We present evidence that hnRNP C promotes APP translation while FMRP represses it, and that the two RBPs interact with the APP CR in a competitive fashion. Moreover, repression of APP translation was linked to the colocalization of FMRP and a tagged APP RNA in processing bodies (PBs), where non-translating mRNAs assemble. These results link the repression of APP translation to the recruitment of the FMRP–APP mRNA complex to PBs, and further suggest that hnRNP C promotes APP translation by competing with FMRP, in turn blocking the recruitment of APP mRNA to PBs.
To study the regulation of APP expression by RBPs, we silenced various TTR RBPs by transfecting the human neuroblastoma cell line BE2-M17 with the corresponding siRNAs. This initial survey (Supplementary Fig. 1) revealed that APP abundance was altered after lowering FMRP and hnRNP C. We first tested if FMRP and hnRNP C associated with the APP mRNA by immunoprecipitation of native ribonucleoprotein complexes (RNP IP analysis) followed by detection of the APP mRNA in the RNP complexes by using reverse transcription (RT) and quantitative (q)PCR amplification. As shown in Figure 1a (top) and in the Supplementary Figure 1, the APP mRNA was significantly enriched in both the hnRNP C and FMRP IPs (Fig. 1a, bottom), in keeping with previous reports that hnRNP C and FMRP associated with the APP mRNA14,18. Further studies were then conducted to identify the regions of interaction by testing biotinylated fragments of the APP mRNA (Fig. 1b and Supplementary Fig. 2). hnRNP C had affinity for the 3′UTR (segment G) and, unexpectedly, also for the APP CR (segment C) (Fig. 1a). While hnRNP C is predominantly nuclear, it is also readily detected in the cytoplasm, although in lower abundance (Supplementary Fig. 3).
Evidence that FMRP and hnRNP C interacted with fragment C in intact cells was sought by analysis of RNP crosslinking before lysis and IP (CLIP). CLIP analysis revealed the association of each RBP with APP fragment C in the cell (Fig. 1c), indicating that the RNPs detected in Figure 1a occurred in intact cells and did not arise from re-association of FMRP or hnRNP C with APP mRNA after membrane disruption. Since both hnRNP C and FMRP associated with fragment C (while other RBPs tested did not, Supplementary Fig. 2), we hypothesized the existence of a functional link between the two RBPs on the APP CR and set out to study this possibility.
We first investigated whether hnRNP C and FMRP were directly involved in regulating APP expression by silencing hnRNP C or FMRP using small interfering (si)RNA, which effectively reduced hnRNP C and FMRP levels in BE2-M17 cells (Fig. 2a). Under these conditions, APP protein levels were markedly decreased in the hnRNP C siRNA group and strongly upregulated in the FMRP siRNA group (Fig. 2a). These changes did not arise from altered APP mRNA abundance (Fig. 2b) or protein turnover (not shown); instead, we postulated that APP translation could be influenced by these two RBPs.
The rate of APP translation was measured after incubation of BE2-M17 cells with 35S-methionine and 35S-cysteine for 15 min followed by APP IP to detect de novo synthesized APP; as shown, translation was reduced in the hnRNP C group and was elevated in the FMRP group (Fig. 2c). The relative association of APP mRNA with polyribosomes, an indirect measure of its translation, was studied by fractionating the cytoplasmic components on sucrose gradients; representative sucrose gradient profiles are shown in Figure 2d, right (and Supplementary Fig. 4). Fractionation was followed by measurement of the levels of APP mRNA (by RT-qPCR) in each fraction: untranslated (fractions 1 and 2), ribosome subunits and monosomes (fractions 3–5), low-molecular-weight polysomes (fractions 6–8), and high-molecular-weight polysomes (fractions 9–12). Compared with the distribution of APP mRNA in control siRNA cells (peaking at fraction 10), silencing hnRNP C shifted the APP mRNA distribution to lower parts of the gradient, with much of the APP mRNA peaking at fraction 8, in keeping with a reduction in APP translation in the hnRNP C siRNA population. Conversely, silencing FMRP increased the relative abundance of APP mRNA in the highest translating fraction (fraction 10), which contained 33% of the APP mRNA from Ctrl siRNA cells and 41% of APP mRNA from FMRP siRNA cells; these data reveal a greater proportion of large polysomes translating APP after silencing FMRP. FMRP and hnRNP C specifically altered the translational status of the APP mRNA, as the profiles for the mRNA encoding the housekeeping protein β-actin were quite similar among the silencing groups (Fig. 2d, left).
To further test the possibilities that FMRP reduced APP translation while hnRNP C promoted it, each RBP was overexpressed by transfection using plasmid vectors, then APP levels were measured. As shown, hnRNP C overexpression increased APP protein, but not APP mRNA levels (Fig. 2e; Supplementary Fig. 5), while FMRP overexpression decreased APP protein, but not APP mRNA levels (Fig. 2f; Supplementary Fig. 5). Together with the silencing data, these results indicate that FMRP functions as a repressor of APP translation in human neuroblastoma BE2-M17 cells (as previously shown in mouse neurons18) and further suggest that hnRNP C functions as an enhancer of APP translation.
Next, we sought to identify the regions of APP mRNA implicated in this translational control using reporter constructs. As hnRNP C associated prominently with the APP 3′UTR segment G (Fig. 1b), we first tested the contribution of this RNA region by preparing a construct that expressed a chimeric RNA containing APP fragment G at the 3′UTR (EGFP + 3′ UTR(G); Fig. 3a). When control and hnRNP C siRNA-transfected cells were compared, expression of the control EGFP reporter mRNA was unchanged but expression of EGFP + 3′ UTR(G) reporter mRNA was reduced in the hnRNP C siRNA group, as assessed both by Western blot analysis (Fig. 3b) and by fluorescence microscopy (Fig. 3c). This reduction was not due to a decrease in EGFP + 3′ UTR(G) mRNA abundance, since these levels remained unchanged, as measured using RT-qPCR (Fig. 3d). Instead, the altered EGFP expression appeared to be due to modest upregulation of EGFP translation in the presence of hnRNP C through the 3′UTR(G) sequence.
Since both hnRNP C and FMRP associated with segment C of the APP CR, we tested the contribution of their interaction at this RNA region by using a reporter construct in which region C was inserted in frame into the CR of EGFP (EGFP + CR(C)) (Fig. 4a). Expression of the control EGFP reporter protein was unchanged among the transfection groups. By contrast, expression of EGFP + CR(C) protein was lower in the hnRNP C siRNA group and was higher in the FMRP siRNA group, as tested both by Western blot analysis (Fig. 4b) and by fluorescence microscopy (Fig. 4c). As seen with the 3′UTR reporter (Fig. 3), the changes in EGFP + CR(C) protein levels were not due to altered EGFP + CR(C) mRNA abundance (Fig. 4d) and were instead attributed to altered translation rates in the hnRNP C and FMRP siRNA transfection groups. This notion was tested further by overexpressing hnRNP C, which enhanced reporter EGFP protein levels without significantly altering the levels of EGFP or EGFP + CR(C) mRNAs (Fig. 4e). On the other hand, FMRP overexpression lowered reporter EGFP protein levels but did not significantly alter the levels of EGFP or EGFP + CR(C) mRNAs (Fig. 4f). In sum, FMRP can reduce the translation of a reporter construct through APP CR(C) while hnRNP C can enhance translation strongly through the APP CR(C) and modestly through the APP 3′UTR(G).
As both hnRNP C and FMRP could associate with APP CR(C) but had opposite effects on APP mRNA levels, we postulated that hnRNP C and FMRP might compete for interaction with the APP mRNA. To test this hypothesis, we studied the binding of hnRNP C in cells expressing normal (control siRNA) or reduced (FMRP siRNA) levels of FMRP; as shown (Fig. 5a), silencing FMRP resulted in greater levels of the RNP complex (hnRNP C–APP mRNA) than were seen in cells with normal FMRP levels. Conversely, a comparison of cells expressing normal (control siRNA) or reduced levels of hnRNP C (hnRNP C siRNA) showed that the RNP complex (FMRP–APP mRNA) was higher in cells with silenced hnRNP C than in cells with normal hnRNP C levels (Fig. 5b). These findings were recapitulated using EGFP reporters: EGFP + CR(C) mRNA was significantly more enriched in hnRNP C IP samples after silencing FMRP (Fig. 5c), while the same reporter mRNA was more prominently associated with FMRP after silencing hnRNP C (Fig. 5d). To study the competition between FMRP and hnRNP C in a neurological disease model, we used mice lacking FMR1 (Fmr1 KO)21. As shown in Figure 5e, the steady-state abundance of mouse APP (mAPP) mRNA in whole-brain RNA was comparable between wild-type (WT) and Fmr1 KO mice. However, the interaction of hnRNP C with mouse mAPP mRNA was significantly more abundant in Fmr1 KO brain lysates than in WT brain lysates (Fig. 5f), associated with significantly higher expression of APP protein in the Fmr1 KO mice (Fig. 5g, in agreement with an earlier findings18). Collectively, these observations support the view that hnRNP C and FMRP compete for association with the APP CR and further indicate that translation increases if APP CR associates with hnRNP C, while it decreases if APP CR associates with FMRP.
Recently, several studies have shown that FMRP co-localizes with PBs, where non-translating mRNAs accumulate and can be sorted for transient storage or degradation22–25. First, we tested whether Argonaute (Ago) proteins, which function as translational repressors, were functionally linked to the inhibitory activity of FMRP. As shown in Figure 6a, FMRP overexpression increased significantly the association of APP mRNA with HA-tagged Ago1 and Ago2 proteins, which are PB-resident proteins. In keeping with the suppression of gene expression by Ago proteins, overexpression of HA-Ago1 or HA-Ago2 each reduced basal APP abundance, although it did not prevent the increase in APP levels that ensued silencing FMRP (Fig. 6b). FMRP and Ago associated by RNA-independent protein–protein interaction, as determined by IP followed by Western blot analysis in the presence of RNases (Fig. 6c). This analysis also revealed the interaction of FMRP with RCK, a component of PBs implicated in translational repression26 (Fig. 6c). It is important to note that FMRP did not repress APP translation in the absence of RCK, Ago1 or Ago2 (Supplementary Fig. 6). The distribution of FMRP and PBs was further studied by immunofluorescence, which revealed extensive colocalization of FMRP with Ago2, as well as with the PB markers Dcp1 and Rck (Fig. 6d). These data indicate that in human neuroblastoma BE2-M17 cells, FMRP is highly abundant in PBs.
To investigate directly whether the localization of FMRP in PBs is implicated in FMRP’s repression of APP translation, we studied the subcellular localization of APP mRNA. First, we tested the presence of APP mRNA in PBs by performing anti-RCK RNP IP followed by RT-qPCR to detect APP mRNA. When FMRP is overexpressed, APP mRNA was significantly more abundant in RCK IP samples, suggesting that FMRP enhances the association of APP mRNA with PBs (Supplementary Fig. 7a). However, since not all RCK may be present in PBs, we sought to study the subcytoplasmic localization of APP mRNA in intact cells. To this end, we prepared reporter construct pMS2-APP (details in Methods), which expressed a chimeric RNA (MS2-APP) comprising the APP CR(C) segment and 24 tandem MS2 RNA hairpins (Fig. 7a). Co-transfection of pMS2-APP together with plasmid pMS2-YFP, which expressed the chimeric fluorescent protein MS2-YFP with a nuclear localization signal (NLS; see Methods), allowed us to track the subcellular localization of the chimeric MS2-APP RNA (as the MS2-YFP–MS2-APP complex) as well as the control MS2 RNA (as the MS2-YFP–MS2 complex) by confocal microscopy. Despite potential artifacts, the MS2 system has been used successfully to track the subcellular localization of RNAs27 As shown in Figure 7b (left), the control MS2 RNA appeared to be exclusively nuclear in all of the transfected cells, due to the presence of the NLS (Fig. 7a). By contrast, some MS2-APP RNA was retained the cytoplasm with a punctate pattern, colocalizing to some extent (but not exclusively) with RCK signals; colocalization results in yellow signals in the merged images (Fig. 7b, right, arrowheads, and Supplementary Fig. 7). The colocalization of MS2-YFP–MS2-APP and RCK signals was lost in the transfected cells when hnRNP C was overexpressed (Fig. 7c right) but was seen in the corresponding control cells (Fig. 7c left). Similarly, the colocalization of MS2-YFP–MS2-APP and RCK signals was lost when FMRP was silenced (Fig. 7d middle) but was seen in the corresponding control cells (Fig. 7d left) and in cells with silenced hnRNP C (Fig. 7d right). Together, these results support the view that FMRP represses APP mRNA translation at least in part by recruiting the transcript to PBs. According to this paradigm, hnRNP C promotes translation by competing for interaction of FMRP with the APP CR, thereby preventing the localization of APP mRNA at PBs (Fig. 8).
hnRNP C and FMRP were found to compete for binding to the CR of APP mRNA in a competitive manner and modulated APP translation in opposite directions: hnRNP C enhanced APP translation, while FMRP repressed it. These conclusions were reached by analyzing both the endogenous APP mRNA and reporter constructs bearing the ~120-nt APP CR fragment C where binding by these RBPs was mapped. As shown, silencing FMRP promoted binding of hnRNP C to APP mRNA and enhanced APP translation, while silencing hnRNP C increased FMRP binding to APP mRNA and lowered APP translation. Accordingly, FMRP overexpression inhibited APP translation and hnRNP C overexpression increased it (Figs. 2, ,4,4, ,5).5). Our results confirm and expand upon the findings of Westmark and Malter, who showed that FMRP bound to a G-rich (G-quartet-like) CR segment in the APP CR and repressed APP translation in mouse neurons18. The authors further showed that the inhibitory interaction of FMRP with the APP CR was relieved after treatment with DHPG, an agonist of the metabotropic glutamate receptor (mGluR)18.
FMRP was previously shown to suppress strongly and specifically the translation of several mRNAs in rabbit reticulocyte lysates and in microinjected oocytes28. En masse analysis of FMRP RNP complexes revealed altered translational profiles for many target transcripts29; FMRP was generally found to bind target mRNAs in the 5′- and 3′-UTRs29,30. The precise mechanisms of translational repression by FMRP are unclear, but the RBP appears to prevent the assembly of 80S on target mRNAs, and its phosphorylation correlates with its presence in stalled polyribosomes27,31–33. Earlier studies also provided evidence that FMRP associated with the RNAi effector complex RISC, suggesting a link between FMRP and the inhibition of translation through the microRNA pathway34–37. More recently, FMRP was implicated in the assembly of stress granules (SGs), which form transiently in response to cellular damage; this finding is potentially relevant to the translational repression by FMRP, since SGs are believed to contain untranslated mRNAs that are subject to RNP remodelling to modulate their subsequent turnover and translation rates38. Our findings support the view that FMRP represses APP translation at least in part by reducing the rate of translation initiation (since a larger population of heavy polyribosomes is seen after silencing FMRP, Fig. 2d), by reducing nascent APP translation (Fig. 2c), and translation of a reporter RNA (Fig. 4) and by recruiting a tagged APP RNA (containing CR(C)) to PBs (Figs. 6d and and7),7), where non-translating mRNAs accumulate22–25. It remains to be seen whether FMRP also inhibits APP translation by stalling preinitiation complexes or by recruiting the APP mRNA onto SGs, RISC, or other cellular machineries.
The mechanisms whereby hnRNP C promotes APP translation are also unclear at present. hnRNP C1/C2 was previously suggested to promote polyadenylation and enhanced the initiation of Unr (upstream of N-Ras) translation during mitosis39,40. The promotion of Unr translation by hnRNP C was shown to be mediated by an internal ribosome entry site (IRES), antagonizing the polypyrimidine tract-binding protein (PTB). Interestingly, the APP mRNA also has a functional IRES in the 5′UTR17. Although it lies several hundred nucleotides away from the G-rich region of association with hnRNP C, it will be interesting to study if the promotion of APP translation by hnRNP C is related to the APP IRES. Nonetheless, our results indicate that hnRNP C overexpression increases the initiation of APP mRNA translation (Fig. 2d), the overall de novo APP translation (Fig. 2c), the translation of a reporter mRNA (Fig. 4), and the recruitment of tagged APP RNA to PBs (Fig. 7). In light of our results that hnRNP C promotes APP translation by competing with FMRP for binding to the APP CR, it will also be important to study whether hnRNP C competes with FMRP in the binding of other FMRP target mRNAs identified by Brown and coworkers29, and whether hnRNP C stimulates their translation.
How PBs repress the translation of resident mRNAs is not fully understood. However, PBs contain many translational repressors, including decapping enzymes (DCP1, DCP2), mRNA deadenylation factors (e.g. the CCR4–CAF-1–Not complex), activators of decapping (Dhh1/RCK/p54, Pat1, Scd6/RAP55, Edc3, the Lsm1-7 complex), and exonucleases (e.g. XRN-1)41–43. FMRP interacts with Ago proteins and the microRNA pathway suggesting that microRNAs could also participate in controlling APP translation. MicroRNAs, such as miR-106a/b, miR-520c, miR-20a and miR-17-5p, associate with the APP 3′UTR and contribute to repressing its translation12,13. It is not known at this time whether such interactions are functionally linked to the actions of FMRP or hnRNP C, although these microRNAs interact with regions outside of CR segment C. Also awaiting experimental analysis is whether stresses can alter the binding of any of the factors (RBPs, microRNAs) that interact with the APP mRNA.
In studies that examined the APP 3′UTR, hnRNP C and nucleolin were shown to bind to a 29-nt sequence in the APP 3′UTR and increased APP mRNA stability in rabbit reticulocyte lysates14. In agreement with these findings, we also observed extensive binding of hnRNP C to the APP 3′UTR (Fig. 1); however, we did not observe significant differences in APP mRNA levels or stability under the conditions of our study, perhaps because the two cell types differ in this respect (Fig. 3 and data not shown). Several other RBPs were also found to influence APP expression. HuD bound the APP mRNA at the 3′UTR (Supplementary Fig. 1) and silencing lowered APP mRNA and protein levels (Supplementary Fig. 1 and data not shown). However, given the long half-life of APP mRNA (t1/2 >12 h, (ref. 44 and data not shown)), it seems that HuD may regulate APP expression indirectly, perhaps by affecting the expression of a transcription factor that controls APP gene transcription.
In closing, RBPs such as FMRP and hnRNP C together with microRNAs that interact with the APP mRNA are emerging as pivotal post-transcriptional regulators of APP production. These factors help to ensure that APP is expressed in the correct abundance, as dictated by the developmental and metabolic state of the cell. Given the multiplicity of factors controlling the turnover and translation of APP mRNA, further studies are warranted to elucidate their complex interactions. Although the physiologic function of APP is not understood completely, the levels of APP directly impact upon the levels of processed Aβ. Thus, a thorough knowledge of the control of APP levels is critical in order to understand how AD arises and to develop effective AD interventions.
We cultured human neuroblastoma BE2-M17 cells in Opti-MEM and Dulbecco’s modified essential medium (Invitrogen) supplemented with 10% FBS. We transfected small interfering (si)RNAs targeting hnRNP C and FMRP (sc-35577 and sc-36870, Santa Cruz Biotechnology) and control (Ctrl) siRNA (Qiagen), comprising three pooled siRNAs without known off-target effects, at 20 nM final concentration using Oligofectamine (Invitrogen) and analyzed cells 48 h later. We constructed reporter plasmids by inserting fragments from the APP 3′UTR (2232–2635) and CR (901–1020) into plasmid pEGFP-C1 (BD Bioscience)45. To overexpress hnRNP C, we prepared an expression vector by amplifying the hnRNP C CR (NM_031314.2) using PC and ligating it at BamHI and XhoI sites of plasmid pcDNA3 using primers ACTTAGGATCCATGGCCAGCAACGTTACC and ACTCATCTCGAGTTAAGAGTCATCCTCGCCATTG. Dr. R. Willemsen (Erasmus MC, The Netherlands) kindly provided pEGFP-FMRP. We obtained HA-Ago1 and HA-Ago2 from Addgene. We prepared pMS2-APP from plasmid pSL-MS2(24X). Dr. R. H. Singer generously provided plasmids pSL-MS2 and pMS2-YFP25. We inserted MS2-YFP cDNA into plasmid pcDNA3 to increase expression levels (see below). We ligated the MS2 hairpin sequence (24 repeats) from pSL-MS2 at EcoRI and EcoRV sites and APP CR(C) at the XhoI site of pcDNA3. We used Lipofectamine 2000 (Invitrogen) for plasmid transfections.
We used RIPA buffer to prepare whole-cell lysates, separated them by electrophoresis in SDS-containing polyacrylamide gels, and transferred them onto PVDF membranes (Millipore). We used primary antibodies that recognized APP (Calbiochem), GFP (Santa Cruz Biotechnology), FMRP (Chemicon) or β-actin (Abcam), incubated the blots with the appropriate secondary antibodies conjugated with HRP (GE Healthcare) and detected the protein signals using enhanced luminescence (GE Healthcare).
We used Triazol (Invitrogen) to prepare total RNA directly from cells or after immunoprecipitation (IP) from cellular RNA–protein complexes obtained by IP (using anti-FMRP (Abcam), anti-hnRNP C (Santa Cruz Biotechnology) or IgG antibodies), as described below and in Ref. 46. After reverse transcription (RT) using random hexamers and SSII reverse transcriptase (Invitrogen), we assayed the abundance of transcripts by real-time, quantitative PCR (qPCR) analysis using SYBR Green PCR master mix (Applied Biosystems) and gene-specific primer sets: GCCAAAGAGACATGCAGTGA and AGTCATCCTCCTCCGCATC for APP mRNA, TGCACCACCAACTGCTTAGC and GGCATGGACTGTGGTCATGAG for GADPH mRNA, and GGACTTCGAGCAAGAGATGG and AGCACTGTGTTGGCGTACAG for β-actin mRNA.
We prepared whole-cell lysates by incubating cells in RIPA buffer for 10 min on ice followed by centrifugation at 10,000 × g for 15 min at 4°C. We incubated the supernatants with protein A-Sepharose beads coated with primary antibody or control IgG (Santa Cruz Biotechnology) with or without RNaseT (Invitrogen) for 16 h. After washing the beads with RIPA buffer, we assayed the complexes by western blot analysis as described above.
Using whole-cell extracts, we performed IP of native RNP complexes (RNP IP analysis) as described46 using primary antibodies (anti-hnRNP C or control IgG, Santa Cruz Biotechnology); after washes and digestion with DNase I and Proteinase K, we analyzed the RNA in the IP samples by RT-qPCR using the primers described above. We performed IP of crosslinked RNP complexes as described47. After RNase T1 digestion, we isolated RNA and analyzed it by RT-qPCR using primers that amplified segment C of the coding region (below).
We harvested whole brains from WT (4 female, 1 male) or Fmr1 KO (5 female) FvB mice21, 4–5 months of age. To prepare brain homogentes, we used PEB buffer containing RNAse-OUT and 1X protease inhibitor cocktail, followed by centrifugation at 13,000 rpm for 30 min at 4 °C. We incubated the lysates (2 mg aliquots) with beads that were pre-coated with antibody (15 μg anti-hnRNP C or IgG) for 2 h at 4 °C. Subsequent steps are as described above.
We prepared PCR templates to synthesize biotinylated transcripts spanning the APP mRNA (NM_201414). Forward primers contained the T7 RNA polymerase promoter sequence (CCAAGCTTCTAATACGACTCACTATAGGGAGA [T7]):
We tested biotinylated transcripts as previously explained46.
After silencing hnRNP C or FMRP for 48 h, we preincubated cells with cycloheximide (100 μg ml−1, 15 min) and lysed them with PEB (polysome extraction buffer) containing 20 mM Tris-HCl at pH 7.5, 100 mM KCl, 5 mM MgCl2 and 0.5% NP-40. We fractionated the cytoplasmic lysates by ultracentrifugation through 10–50% linear sucrose gradients and obtained 12 fractions for RNA extraction and RT-qPCR analysis, as described48.
We studied nascent translation of APP and GAPDH as described48. After incubation of BE2-M17 cells with 1 mCi L-[35S] methionine and L-[35S]cysteine (Easy Tag ™EXPRESS, NEN/Perkin Elmer, Boston, MA) per 60-mm plate for 15 min, cells were lysed in RIPA buffer (10 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1% NP-40, 1 mM EDTA, 0.1% SDS, and 1 mM DTT). Following IP with anti-IgG1 (BD Pharmingen), anti-APP (Calbiochem) or anti-GAPDH antibodies (Santa Cruz Biotechnology), we washed the reaction beads in RIPA buffer, resolved the IP material by SDS-PAGE, transferred onto PVDF filters, and visualized and quantified it with a PhosphorImager (Molecular Dynamics).
After transfection of plasmids or siRNAs, we fixed cells with 2% formaldehyde, permeabilized them with 0.2% Triton X-100, blocked with 5% BSA, and incubated them with primary antibodies recognizing Dcp1a (Abcam), RCK, EGFP or HA (Santa Cruz Biotechnology). We then used Alexa 488- or Alexa 568-conjugated secondary antibodies (Invitrogen) to detect primary antibody–antigen complexes with different color combinations as needed. We acquired the images using Axio Observer microscope (ZEISS) with AxioVision 4.7 Zeiss image processing software or with LSM 510 Meta (ZEISS). We acquired confocal microscopy images with Z-sectioning mode with 15 slices and 0.4 μm spacing and merged them using maximum intensity.
This research was supported by the National Institute on Aging-Intramural Research Program, National Institutes of Health. P.F.W. is suppported by DA00266. We thank F.E. Indig (Confocal Imaging Facility, NIA) and M.H. Dehoff (Johns Hopkins University School of Medicine) for assistance with experiments.
AUTHOR CONTRIBUTIONSE.K.L., H.H.K., K.A., J.L.M., X.Y., M.P.M., and M. Gorospe designed the study E.K.L., H.H.K., Y. K., K.A., S.S., S.S.S., J.L.M., and X.Y. performed the experiments E.K.L., Y. K., M. Gleichmann, M.R.M., X.Y., P.F.W, and M.P.M. contributed key reagents E.K.L., M.P.M., and M.G. wrote the paper