Proteomic Characterization of Messenger Ribonucleoprotein Complexes Bound to Nontranslated or Translated Poly(A) mRNAs in the Rat Cerebral Cortex*

Receptor-triggered control of local postsynaptic protein synthesis plays a crucial role for enabling long lasting changes in synaptic functions, but signaling pathways that link receptor stimulation with translational control remain poorly known. Among the putative regulatory factors are mRNA-binding proteins (messenger ribonucleoprotein, mRNP), which control the fate of cytosolic localized mRNAs. Based on the assumption that a subset of mRNA is maintained in an inactive state, mRNP-mRNA complexes were separated into polysome-bound (translated) and polysome-free (nontranslated) fractions by sucrose density centrifugation. Poly(A) mRNA-mRNP complexes were purified from a postmitochondrial extract of rat cerebral cortex by oligo(dT)-cellulose affinity chromatography. The mRNA processing proteins were characterized, from solution, by a nanoflow reverse phase-high pressure liquid chromatography-μ-electrospray ionization mass spectrometry. The majority of detected mRNA-binding proteins was found in both fractions. However, a small number of proteins appeared to be fraction-specific. This subset of proteins is by far the most interesting because the proteins are potentially involved in controlling an activity-dependent onset of translation. They include transducer proteins, kinases, and anchor proteins. This study of the mRNP proteome is the first step in allowing future experimentation to characterize individual proteins responsible for mRNA processing and translation in dendrites.

From the point of their transcription, the fate of every mRNA is controlled by RNA-binding proteins. Inside the nucleus these proteins enable proper transcription, splicing, processing, and their export into the cytosol. Once outside the nucleus, mRNA-binding proteins, also described as messenger ribonucleoproteins (mRNPs), 1 are involved in processes controlling intracellular transport, subcellular localization, translation initiation, translational silencing ("masking"), stability, and degradation of poly(A) mRNAs. Only a subset of mRNAs is transported into distal parts of dendrites (1)(2)(3), and it is conceivable that these mRNAs in particular must be protected from premature translation during their transport through the cell body and dendrites. Furthermore, to ensure localized translation of those mRNAs at activated synapses, the masked mRNAs should become unmasked by receptor-triggered signaling mechanisms, implying dissociation of some mRNPs from the transport complex at the time of translation. Specific mRNA-binding proteins that bind preferentially to either nontranslated or translated poly(A) mRNAs have not been well characterized in neurons. To search for such putative neuronal poly(A) mRNA masking proteins and proteins involved in mechanisms of receptor-stimulated protein translation, we separated and enriched mRNP complexes associated with nontranslated poly(A) mRNAs and those associated with translated poly(A) mRNAs, and we determined the composition of co-purified proteins by mass spectrometry. Here we report that although most proteins were identified in both fractions, there were several examples of proteins only present in silent mRNP complexes and proteins found only in translated mRNP complexes. UNR, STRAP, the RNA-binding protein EWS, RNA, export factor-binding protein 1, and hnRNP-H1 are either specifically or highly enriched in small, Ͻ40 S, mRNP complexes, which are associated with nontranslated mRNAs. Based on their specific localization, these proteins are candidates to play a crucial role in masking poly(A) mRNAs in neurons. On the other hand, proteins such as FMRP, regulator of nonsense transcript 1, protein similar to protein C9orf10, BC010304, protein similar to ubiquitin C-terminal hydrolase-related polypeptide, protein similar to RNA helicase A, and RACK1, which were identified as part of the translated mRNA-mRNP complexes, are potentially involved in processes of translational control. 7.5, 1% Nonidet P-40, 50 mM NaCl, 4 mM MgCl 2 , 45 g/ml cycloheximide) and layered on a discontinuous sucrose gradient (4.5 ml of 12% and 4.5 ml of 33.5% sucrose in 20 mM Tris/HCl, pH 9.0, 80 mM NaCl, 3 mM MgCl 2 , and 0.02% ␤-mercaptoethanol). The tubes were centrifuged for 90 min at 130,000 ϫ g in an SW41 rotor (Beckman L8 -70 M centrifuge). To dissociate ribosomes/polysomes and to release poly(A) mRNA-mRNP complexes, the resulting pellet (monosomes and polysomes, see Fig. 1A) was resuspended in a solution (pellet buffer) containing 30 mM EDTA, 0.5% Nonidet P-40, 20 mM Tris/HCl, pH 7.5, and kept on ice for 10 min. The interfaces were recovered and centrifuged again for 20 min at 400,000 ϫ g in a Beckman TL-100 ultracentrifuge. The resulting pellets, which contained mRNPs, free 40 S, and 60 S ribosomal subunits, and a small subfraction of monosomes (Fig. 1A), were resuspended in the pellet buffer and kept for 10 min at 4°C. The suspension was centrifuged for 2 min at 14,000 ϫ g, and the supernatant was used for the oligo(dT)-cellulose binding assay. The KCl concentration was adjusted to 200 mM, and then 40 l of pre-washed oligo(dT)-cellulose (100 g/ml, Sigma) was added. After 90 min of constant rotation at 4°C, the cellulose was washed three times with 1 ml wash buffer (20 mM Tris/HCl, pH 7.5, 100 mM KCl, 5 mM MgCl 2 ). Poly(A) mRNAs and attached mRNPs were eluted with 200 l of 10 mM Tris/HCl, pH 7.5, at 65°C. Either 5 l of the eluate was used for mass spectrometry analysis (see below) or separated in a one-(5-20% acrylamide) or two-dimensional SDS-PAGE (pI 3-10, 8 -18% acrylamide). In a parallel preparation, the three fractions were used to prepare poly(A) mRNAs (microPoly(A)Pure TM , Ambion Inc., Austin TX), and the amount in each fraction was estimated (DNA DipStick TM , Invitrogen).
Northwestern Blot Assay-Oligo(dT)-cellulose bound proteins were released with SDS sample buffer (4% SDS, 250 mM Tris, 50 mM DTT, 3 mM EDTA, 20% glycerol, pH 8.0), separated on a 5-20% polyacrylamide gel, and blotted onto a nitrocellulose membrane. Detection of mRNA-binding proteins was performed as described previously (4). The nitrocellulose membrane was rinsed twice with phosphate-buffered saline, and the proteins were renatured by incubation in buffer A (10 mM Tris/HCl, pH 7.4, 50 mM NaCl, 1 mM EDTA, 1 mM DTT, and 1ϫ Denhardt's (0.2 mg/ml bovine serum albumin, 0.2 mg/ml Ficoll, 0.2 mg/ml polyvinylpyrrolidone) solution) three times for 45 min, followed by a 5-min incubation in a hybridization solution (buffer A including 20 g/ml tRNA and 5 g/ml heparin). Total poly(A) mRNA were prepared from rat cortex according to the manufacturer's protocol (Poly(A)-Pure TM , Ambion, Austin, TX), labeled using [␣-32 P]ATP and poly(A) polymerase (U. S. Biochemical Corp.), precipitated with ethanol sodium acetate, washed, and finally resuspended in 10 mM Tris/HCl. Labeled poly(A) mRNA was added to the hybridization buffer (final concentration about 200,000 cpm/ml), and the nitrocellulose membrane was incubated in this solution for 120 min at room temperature. After a series of washes the blot was dried, and bound radioactivity was detected by PhosphorImaging (FUJIX-BAS1000).
Mass Spectrometry Analysis-Approximately 7.5 g (75 pmol; 100 kDa average molecular mass) of total protein from each prepared sucrose interface was equilibrated in 80 l of 100 mM NH 4 HCO 3 , pH 9. Proteins were reduced (50 mM DTT in 100 mM NH 4 HCO 3 for 1 h at 57°C) and alkylated (100 mM iodoacetamide in 100 mM NH 4 HCO 3 for 1 h in the dark). Modified trypsin (Promega, Madison, WI) was added at a ratio of 1:20 (enzyme:protein) and incubated for 4 h at 37°C, followed by 4 h at room temperature. The digestion was stopped by adding 2 l of glacial acetic acid (Sigma, 99.99% pure). The sample was stored at Ϫ38°C for the duration of its use. A small aliquot of each sample was analyzed using a short HPLC gradient on a ThermoFinnigan LCQ Deca mass spectrometer (San Jose, Ca) to ensure complete digestion.
Approximately 10 pmol of the solution digests was analyzed by nanoflow-reverse phase HPLC--electrospray ionization mass spectrometry on a LCQ Deca mass spectrometer. Data from an ϳ3-h HPLC gradient (ϳ1500 MS/MS scans per h) were acquired on each sample. A data-dependent analysis was performed in which one MS scan was followed by collisionally activated dissociation MS/MS analysis of the top five most abundant ions present in the MS scan. The five masses were then placed on a dynamic exclusion list and not chosen again for MS/MS analysis for 1 min. This greatly reduced redundancy of data thus allowed for greater sample coverage. Post analysis in-house filtering software was used as follows: 1) to remove low quality spectra and 2) to assign charge state of masses analyzed. Data passing above filters were then subjected to a SEQUEST (ThermoFinnigan, San Jose, Ca) search of a human/rat/mouse protein subset of the nonredundant Gen-Bank TM (National Center for Biotechnology Information (NCBI)) data base. Peptides receiving a cross-correlation score (Xcorr) below 2 were manually confirmed, and in the case where a protein was identified with 3 or fewer peptides, all peptides were confirmed manually. Table  I contains data from two separate MS/MS analyses of the mRNPs bound to translated mRNAs and one MS/MS analysis of the proteins associated with nontranslated poly(A) mRNAs. Multiple MS/MS analyses have been performed on similar sucrose fractions, with similar protein results. However, because the sample preparations in those cases were not identical, a direct comparison or compilation of proteins was not prudent.

RESULTS
To characterize protein complexes that are preferentially associated with nontranslated ("free") or translated ("ribosomebound") mRNAs, we first separated both pools of mRNAs by sucrose density centrifugation as described recently (5). Based on the presence or absence of ribosomal subunits and ribosomes ( Fig. 1, A and B), mRNP complexes enriched from the first interface (on top of the 12% sucrose; fraction 1) were considered to be associated with mRNAs not engaged in translation. The second interface (between 12 and 33.5% sucrose; fraction 2) contained a mixture of nontranslated mRNAs, mRNAs associated with the (pre)initiation complex as well as translated mRNAs, whereas the 33% sucrose pellet (fraction 3) contained mainly translated, polysome-bound mRNAs. From each fraction the poly(A) mRNA-mRNP complexes were bound to oligo(dT)-cellulose, washed, and eluted with 10 mM Tris/HCl (Fig. 1C, Fig. 2). All poly(A) mRNAs were covered by a number of proteins including the poly(A)-binding protein 1 (PABP1). To verify that despite the presence of the PABP1, poly(A) mRNA-mRNP complexes could be purified using oligo(dT)-cellulose, we first performed a Northwestern blot assay. Briefly, oligo(dT)-cellulose-bound proteins were eluted and separated on a 5-20% polyacrylamide gel, blotted on nitrocellulose, renatured, and finally incubated with 32 P-labeled poly(A) mRNAs, which should bind to the mRNA-binding proteins present. A number of poly(A) mRNA-binding proteins were detected, and the amount and variability of these poly(A) mRNA-binding proteins differed considerably between the three fractions studied (Fig. 2B). With this approach, most poly(A) mRNA-binding proteins were found in the fraction corresponding to translated mRNAs. This indicates that the majority of poly(A) mRNAs is already engaged in translation ( Fig. 2A). One of the most abundant proteins in the preparation is the poly(A)-binding protein (PABP1), indicating that the association of this protein with the poly(A)-tail does not prevent binding of poly(A) mRNAs to oligo(dT). Pretreatment of the sample with RNase A (10 g/ml) resulted in an almost complete loss of proteins that could be purified with the method used (Fig. 2C), strongly suggesting that poly(A) mRNA-associated proteins are purified with this method.
To identify the protein composition in each fraction, the eluted proteins were used directly for nano-HPLC--electrospray ionization-MS analysis. Because fraction 2 contained a mixture of free and ribosome-bound mRNA-mRNP complexes, we focused our investigations on fraction 1 (free, nontranslated poly(A) mRNAs) and fraction 3 (ribosome associated, translated poly(A) mRNAs). As listed in Table I, most known poly(A) mRNA-binding proteins identified by mass spectrometry were detected in both fractions, indicating that the majority of mRNPs is associated with mRNAs before and during translation. Although a postmitochondrial supernatant was used as starting material, a number of proteins were detected that were presumed to be present only within the nucleus, such as various splicing factors, and proteins involved in polyadenylation (cleavage stimulation factor and hnRNP-H). Although the total amount of proteins associated with nontranslated poly(A) mRNAs is very small, some of them are specifically or highly enriched in this fraction (Table I). These include the UNR protein (upstream of N-Ras), STRAP (serine/threonine kinase receptor associated protein, also known as UNR-interacting protein), the RNA-binding protein EWS, and phosphofructokinase C. Similarly, calcium/calmodulin-dependent protein kinase II was found primarily in the fraction associated with nontranslated mRNAs. To confirm the finding that these pro-teins are predominantly associated with nontranslated mRNAs, a Western blot assay was performed. Again, STRAP was detectable only in the first fraction, whereas the EWS protein, hnRNP-H, and CamKII were highly enriched in the first fraction (Fig. 3).
As stated above, the majority of mRNA-binding proteins was present in fraction 3, purified from mRNAs co-localized with polysomes. Although a number of them appear to be specifically localized in this fraction (Table I), not all of them are necessarily associated with polysome-bound mRNAs. Thus, it is conceivable that this fraction also contains very large poly(A) mRNA-mRNP complexes (Ͼ Ͼ80 S) as well as poly(A) mRNA-mRNP complexes that are very tightly bound to cytoskeletal elements. The complexes tightly associated to cytoskeletal elements may co-migrate with polysomes in the sucrose gradient, even though they are not linked to translated (polysome-bound) poly(A) mRNAs. Indications for this assumption come from our observation that after a complete dissociation of polysomes and ribosomes, which can be achieved by performing the entire preparation in the absence of magnesium, proteins such as Staufen or FMRP (fragile X mental retardation protein) were still detectable in fraction 3 (Fig. 4). Therefore, any protein we identified to be specifically localized in fraction 3 was not unambiguously a component of an mRNP particle linked to translated (ribosome-bound) poly(A) mRNAs only, but might also be a component of a large, nontranslated mRNP complex.

FIG. 1. Preparation of mRNPs associated with free or translated poly(A) mRNAs.
A, postmitochondrial supernatant was lysed and layered on a discontinuous sucrose gradient (12/33.5%) and centrifuged at 41,000 rpm for 90 min. The resulting pellet and interfaces were resuspended and separated using a continuous sucrose gradient (12-45%). The presence and distribution of RNA (ribosomal RNA) within the sucrose gradient was detected by an ISCO spectrophotometer at a wavelength of 254 nm. In fraction 1 (on top of the 12% sucrose), we could not detect any rRNA, indicating the absence of ribosomal subunits. Fraction 2 (interface between 12% and 33.5% sucrose) contained small and large ribosomal subunits (40 S and 60 S), monosomes (80 S), and 2 and 3 ribosomal complexes, whereas fraction 3 (pellet) contained polysomes and a small amount of the large ribosomal subunit. B, to confirm the absence or presence of the small ribosomal subunit (40 S) in these three fractions, proteins of each fraction were separated by SDS-PAGE and blotted onto nitrocellulose. The small ribosomal subunit protein S6 was then detected in fraction 2 and fraction 3 by immunostaining with a specific antibody (arrow). C, fractions 1 and 3 were used to purify poly(A) mRNA-mRNP complexes. After binding poly(A) mRNA to oligo(dT)-cellulose, the poly(A) mRNA-protein complexes were eluted with 10 mM Tris/HCl, pH 7.5. Thereafter RNA was digested with RNase A, and the remaining proteins were precipitated, solubilized, separated by two-dimensional-PAGE, and finally visualized by silver staining (arrow points to the added RNase A).

DISCUSSION
Rapid changes in protein synthesis after receptor stimulation, as observed after induction of long term potentiation, long term depression, and neurotrophin-induced synaptic plasticity (1, 6 -11), could be controlled, at least partly, by a rapid increase in the accessibility of dormant mRNAs to the translation machinery. In addition to stimulus-triggered transport of newly transcribed mRNAs to activated synapses, as described for Arc-mRNA (6), other, already postsynaptically localized, mRNAs may become more readily available for translation initiation as a result of an enhanced polyadenylation (12)(13)(14) or an unmasking process. In particular, mRNAs that are transported into distal parts of dendrites are likely to be protected to avoid premature translation during their transport through the cell soma, where the majority of ribosomes is localized. To search for candidates that are involved in such a process of masking/unmasking and in receptor-triggered control of translation, we have begun to characterize the protein composition of mRNP complexes associated specifically with nontranslated poly(A) mRNAs and specifically associated with translated poly(A) mRNAs in rat cortical tissue. Our approach was intended to identify not only mRNA-binding proteins but also proteins that are part of the entire mRNP complex associated with poly(A) mRNAs. There exists no established method to purify or enrich poly(A) mRNA-mRNP complexes associated with nontranslated poly(A) mRNAs, probably because the size of these protein complexes can vary considerably, and therefore, they may co-migrate in a sucrose density gradient with a number of other large protein complexes. Furthermore, recent results point to a possible rearrangement of RNA-binding proteins and mRNAs subsequent to cell lysis (15), such that detected mRNA-mRNP complexes may be formed during the purification procedure. Therefore, we kept the time between homogenization and separation of nontranslated and translated (polysome-bound) mRNA within the discontinuous sucrose density gradient as short as possible (ϳ15 min). An alternative procedure would be to UV cross-link mRNA and the associated proteins immediately after cell lysis. However, because of the relatively low efficiency of this reaction (16), this FIG. 2. Characterization of the mRNP preparation. The three subcellular fractions containing free or ribosome-bound poly(A) mRNA-mRNP complexes were prepared by sucrose density centrifugation as described in Fig. 1. A, fractions were used first to determine the presence of poly(A) mRNA. For that poly(A) mRNAs were purified using oligo(dT)-cellulose after proteins were solubilized by concentrated guanidinium thiocyanate (Ambion MicroPoly(A)Pure TM ), and the amount was visualized (DNA DipStick TM , Invitrogen). Whereas the majority of poly(A) mRNAs is associated with polysomes (fraction 3), a small quantity can be detected in the ribosome-free fraction 1 indicating that there exists a considerable cytosolic pool of nontranslated poly(A) mRNAs. B, to confirm the possibility also purifying mRNA-binding proteins via poly(A) mRNA oligo(dT)cellulose hybridization, we performed a Northwestern blot assay. For that the three subcellular fractions were incubated with oligo(dT) cellulose without any protein-solubilizing pretreatment. Then oligo(dT)-cellulose-bound proteins were released by 4% SDS, separated by SDS-PAGE, blotted onto nitrocellulose, renatured, and finally incubated with 33 P-labeled total brain poly(A) mRNAs. Comparable with the found distribution of poly(A) mRNAs in these fractions, the majority of poly(A) mRNA-binding proteins was again detectable in fraction 3 (polysome-bound mRNAs). C, because a number of proteins can bind nonspecifically to oligo(dT)-cellulose, poly(A) mRNA-mRNP complexes were subsequently eluted with 10 mM Tris/HCl, which impaired the poly(A) mRNA oligo(dT)-cellulose interaction (5). With this elution buffer the most proteins were again in fraction 3. D, to check if the proteins eluted with 10 mM Tris/HCl are in fact bound via poly(A) mRNA to oligo(dT)-cellulose, we used a fraction that included mRNA-mRNP complexes associated with polysomes (fraction 3). The fraction was divided, and 1 aliquot was incubated without RNase A (control, co), and 1 aliquot was incubated with RNase A for 30 min at 25°C (RNase). Note that even under control conditions the additional incubation time at higher temperature (25 instead of 4°C) results already in a lower amount of mRNA-mRNP complexes that can be purified with this method. No proteins were released from oligo(dT)-cellulose by 10 mM Tris/HCl, pH 7.5, when RNase A (10 g/ml) was added to the oligo(dT)-cellulose binding buffer to digest the poly(A) mRNAs, indicating that the proteins are specifically bound via poly(A) mRNAs to oligo(dT)-cellulose.

TABLE I Summary of proteins detected by nanoflow HPLC-ESI-MS analysis
After sucrose density centrifugation, the fractions corresponding to nontranslated poly(A)-mRNA/mRNP complexes (fraction 1) and to polysomebound poly(A)-mRNA/mRNP complexes (fraction 3) were resuspended and incubated in the presence of oligo(dT)-cellulose for 60 min at 4°C. Poly(A)-mRNA/mRNP complexes were released from the oligo(dT)-cellulose beads by 10 mm Tris/HCl (pH 7.5), and the present proteins were identified by mass spectrometry analysis. Listed are the identified protein sequences with the highest x corr and the number of different independent hits for the same protein. X Similar to splicing coactivator subunit Srm300 6 gi͉34852270͉ X Similar to splicing factor, arginine/serine-rich 3 6 gi͉35493817͉ X RNA-binding region containing protein 2 isoform c 3 gi͉19526824͉ X U2 small nuclear ribonucleoprotein auxiliary factor 2 63 3 gi͉6756033͉ X Y box protein 1 55 5 approach seems only advisable when one interaction partner is known. To minimize the problem of potential contamination by other protein complexes, we first purified poly(A) mRNAs by oligo(dT) affinity chromatography and then eluted the bound protein complexes. Although the effect of a pretreatment with RNase A points to a specific enrichment of RNP complexes (Fig.  2), we also co-purified a number of cytoskeletal proteins such as actin and tubulin. This confirms the previously described close association of actin/tubulin with poly(A) RNAs (17) and may also explain the abundant presence of cytoskeletal associated proteins like the MAP2b and MAP1b in our preparation. On the other hand, no indications for the presence of the highly abundant axonal localized microtubule-associated protein tau were found in our preparation. This may indicate that the dentritically lo-calized MAP2b could also be involved in mechanisms of targeting mRNP complexes within the dendritic compartment. In addition to proteins containing a known RNA-binding domain, such as RNA recognition motif, double-stranded RNAbinding motif, K homology RNA-binding domain, and cold shock domain, we identified a number of proteins without any known RNA-binding domains. This may point to the fact that we in fact purified complex mRNP particles rather than simply mRNA-binding proteins. This is supported by the fact that among the identified proteins, three proteins (RACK1, STRAP, and cleavage stimulation factor subunit 1) were present that consist almost entirely of WD repeat (Trp-Asp) domains that are thought to mediate various protein-protein interactions FIG. 3. Association of different mRNPs with nontranslated and polysome-bound poly(A) mRNAs. Poly(A) mRNA-mRNP complexes were purified from fraction 1 (corresponding to nontranslated mRNAs), fraction 2 (small and large ribosomal subunits and monosomes), and fraction 3 (polysome-bound, translated mRNAs) and the presence of different mRNPs quantified by Western blot assay. The amount of PABP, which should bind to most poly(A) mRNAs, correlates well with the observed total quantities of mRNA-mRNP complexes as shown in Fig. 2. Although the protein content in fraction 3 is much higher (see Figs. 1 and 2), STRAP, EWS, and hnRNP-H1 are highly enriched in fraction 1, indicating that these proteins are predominantly associated with nontranslated mRNAs. In contrast, RACK1, FMRP, G3BP2a, and Staufen are only detectable in fraction 3.   (18,19). Furthermore, we identified a number of so far uncharacterized proteins, such as BC010304, protein similar KIAA1096, protein similar to C9orf10, protein similar to expressed sequence Al256361, protein similar to ubiquitin Cterminal hydrolase, and caprin-1. Caprin-1, formerly known by the misleading name glycosylphosphatidylinositol-anchored protein p137, appears to be very abundant in the mRNP fraction. Caprin-1, a protein that is highly conserved throughout vertebrate evolution, is a cytosolic localized phosphoprotein and has potential sites for binding of SH2 and phosphotyrosinebinding domains (20). This protein, although it contained no RNA-binding domains, was also found to be part of a protein complex that linked the receptor for activated C kinase (RACK1) to poly(A) mRNAs (5) and to change the degree of phosphorylation during mitosis in Xenopus embryos (21); thus it becomes tempting to speculate that Caprin-1, as part of an mRNP complex, is able to link second messenger-dependent phosphorylation systems with the mechanism of translational control. Of special interest is also the repeated detection of protein similar to ubiquitin C-terminal hydrolase within mRNP complexes, which is likely involved in the processes of de-ubiquitination. Whether this protein is part of a specific signaling pathway or an element controlling the ubiquitination and function of ribosomal proteins, such as L40, S27a P1, translation factors, or AUF1 (22)(23)(24)(25), remains to be investigated.
The presence of two enzymes involved in glycolysis (phosphofructokinase C and glyceraldehyde-3-phosphate dehydrogenase) in an mRNP preparation was not expected but may support the assumption that there exists a link between glycolysis and the protein synthesis machinery (26). Thus the product of the phosphofructokinase, fructose 1,6-bisphosphate, acts as a stimulant of mRNA translation (27), whereas the glyceraldehyde-3-phosphate dehydrogenase may affect mRNA translation via its know ability to bind viral mRNAs (28,29).
One of the aims of this study was to identify those proteins within the mRNP complex that may have served as a link between the receptor-triggered signaling pathway and translational control. Based on our findings the following proteins are candidates for such a function. (i) For the STRAP-UNR complex, both proteins interact with each other in an mRNA-independent manner (28), and although UNR binds internal ribosomal entry site-containing mRNAs (30), STRAP can interact with the serine/threonine receptor kinase, transforming growth factor-␤ (31), and thus this complex may be involved in mechanisms of receptor-triggered translation of internal ribosomal entry site-containing mRNAs. (ii) For the RNA-binding protein EWS (preferentially bound to nontranslated mRNAs), this protein contains an IQ domain, known as a calmodulin-binding domain and, most interestingly, can be phosphorylated by protein kinase C (32,33). Because phosphorylation by protein kinase C reduces the ability of EWS to bind mRNAs (33), a hypothetical masking effect of EWS might be reduced by activated protein kinase C. (iii) For calcium/calmodulin-dependent protein kinase II, this protein was the only protein kinase we could identify among the proteins associated with poly(A) mRNAs. However, in order to describe calcium/ calmodulin-dependent protein kinase II unambiguously as an mRNP-associated kinase, the interacting protein(s) that link the kinase to poly(A) mRNAs must be identified. (iv) For G3BP2a, a protein that contains several SH3 domain-binding consensus sequences and an endoribonuclease activity is controlled by phosphorylation (34,35). (v) For RACK1, a receptor for activated protein kinase C has recently been described in the mechanisms of translational control (5, 36 -38). (vi) The protein activator of the interferon-induced protein kinase (PACT) is the only known cellular protein that binds and thereby activates protein kinase R (39). This in turn inhibits translation. The binding of PACT to protein kinase R is again controlled by the phosphorylation degree of PACT (39), and thus a possible receptor-mediated phosphorylation of PACT may control the translation efficiency.
In summary, our proteomic approach to characterize proteins associated with poly(A) mRNAs led to the identification of UNR/STRAP and the RNA-binding protein EWS, both components of complex signaling pathways, as putative members of the mRNA-masking machinery, based upon their selective association with nontranslated mRNAs (versus those undergoing translation). To what extent these proteins are also involved in processes of receptor-triggered unmasking of dendritic mRNAs can begin to be examined when the mRNAs bound to these proteins are identified. Although the UNR-STRAP complex has properties compatible with control of translation initiation of mRNAs with an internal ribosomal entry site, no subset of mRNAs has yet been described that binds specifically to the EWS protein. Furthermore, the identification of a protein likely involved in processes of de-ubiquitination (protein similar to ubiquitin C-terminal hydrolase) may point to an additional regulatory mechanism for translational control besides protein phosphorylation. The characterization of the so far unknown proteins, such as KIAA1096, similar to expressed sequence Al256361, similar to protein C9orf10, and caprin 1 may reveal additional control mechanisms for mRNA metabolism in the brain.