Translational regulation of Chk1 expression by eIF3a via interaction with the RNA-binding protein HuR.

eIF3a is a putative subunit of the eukaryotic translation initiation factor 3 complex. Accumulating evidence suggests that eIF3a may have a translational regulatory function by suppressing translation of a subset of mRNAs while accelerating that of other mRNAs. Albeit the suppression of mRNA translation may derive from eIF3a binding to the 5'-UTRs of target mRNAs, how eIF3a may accelerate mRNA translation remains unknown. In this study, we show that eIF3a up-regulates translation of Chk1 but not Chk2 mRNA by interacting with HuR, which binds directly to the 3'-UTR of Chk1 mRNA. The interaction between eIF3a and HuR occurs at the 10-amino-acid repeat domain of eIF3a and the RNA recognition motif domain of HuR. This interaction may effectively circularize Chk1 mRNA to form an end-to-end complex that has recently been suggested to accelerate mRNA translation. Together with previous findings, we conclude that eIF3a may regulate mRNA translation by directly binding to the 5'-UTR to suppress or interaction with RNA-binding proteins at 3'-UTRs to accelerate mRNA translation.


Introduction
Eukaryotic translation initiation factor (eIF) 3 is the most complex eIF consisting of 13 putative subunits known as eIF3a through eIF3m [1][2][3][4]. It has been reported that different fractions of the purified eIF3 from embryonic chicken red muscle supported the translation of specific mRNAs in vitro [4]. It has also been reported that the existence of two types of eIF3 sub-complexes, eIF3e and eIF3m sub-complexes, in fission yeast and they were found to associate with different sets of mRNAs [5]. These findings suggest that eIF3 complexes of different compositions exist in cells and they may be responsible for regulating the translation of different subsets of mRNAs. eIF3a is the largest putative subunit of eIF3 complex initially purified from rabbit reticulocyte lysate [6]. However, its functional importance in translational control and role in the eIF3 complex is still under investigation. The finding that eIF3a interacts with other subunits of eIF3 [7][8][9], eIF4B [10], and RNA [7,11,12] supports a role of eIF3a in the function of the eIF3 complex and in translation initiation. It has also been observed that eIF3 preparations relatively rich in eIF3a did not differ substantially in specific activity of stimulating formation of pre-initiation complexes from preparations that essentially lacked this protein [13], suggesting that eIF3a may not be essential for initiation of global protein synthesis. Cleavage of the eIF3a by 20S proteasome can differentially alter the translation rate of different mRNAs both in vitro and in vivo [14]. Our previous studies also showed that although a dramatic decrease in eIF3a expression only caused ~25% drop in global protein synthesis in both H1299 and MCF7 cells, the malignant growth phenotype of these two cell lines were significantly reversed [15].
These findings suggest that eIF3a may have functions in regulating mRNA translation and cell growth. Indeed, eIF3a appears to regulate the synthesis of tyrosinated α-tubulin, ribonucleotide reductase M2, p27kip and DNA repairing proteins [15][16][17][18]. Interestingly, the regulation of eIF3a on the synthesis of these proteins is different. While eIF3a increases the synthesis of tyrosinated α-tubulin and ribonucleotide reductase M2, it reduces the synthesis of p27kip and DNA repairing proteins.
Previously, it has been observed that eIF3a expression oscillates with cell cycle and peaks in S-phase [19]. Down regulating eIF3a expression elongated cell doubling time and altered sensitivity of cells to environmental stresses that affect cell cycle, but did not change the cell cycle distribution. It has also been suggested that eIF3a may be essential for G1-S phase transition in yeast and it may be involved in growth control of yeast cells [20]. Therefore, we hypothesize that eIF3a may regulate the synthesis of proteins important for the transition of cells through cell cycle restriction points.
Checkpoint kinases (Chk's) are kinases that are involved in cell cycle control. Two checkpoint kinases, Chk1 and Chk2, have been identified. Chk1 is a central component of genome surveillance pathways and a key regulator of cell cycle and cell survival [21]. Chk1 is required for the initiation of DNA damage checkpoints and has been shown to play a role in normal (unperturbed) cell cycles. Chk1 impacts various stages of the cell cycle including S phase, G2/M transition and M phase [21]. Here we show that eIF3a positively regulates the translation of Chk1 mRNA via HuR (human antigen R) binding to its 3′-UTR. Hence, eIF3a regulation of cell cycle progression is possibly through regulating Chk1 synthesis. analysis were from GE Healthcare (Piscataway, NJ, U.S.A.). Ni-NTA and Sepharose 4B beads were purchased from Qiagen (Germantown, MD, U.S.A.) and Amersham Biosciences (Piscataway, NJ, U.S.A.), respectively. Polyvinylidene difluoride (PVDF) membranes and concentrated protein assay dye reagents were from Bio-Rad (Hercules, CA, U.S.A.). All other reagents of molecular biology grade were obtained from Sigma or Fisher Scientific (Chicago, IL, U.S.A.).

Construct engineering
Luciferase reporter constructs were engineered by first releasing the luciferase-encoding region plus PolyA signal from pGL3 by digestion with Hind III and Bam HI and subcloning into pcDNA3 vector, resulting in pcLuc. The cDNAs encoding 5′-UTR, 3′-UTR and 3′-UTRM2 of Chk1 were amplified from RNAs extracted from H1299 cells using RT-PCR and primers listed in Table 1. The cDNA encoding 3′-UTRM3 was generated using a transformer site-directed mutagenesis kit (BD Biosciences, Clontech) with primer CH3UM3 and selection primer shown in Table 1 as instructed by the manufacturer. The 3′-UTRM1 construct was generated by digesting the 3′-UTR cDNA of Chk1 with Xba I and Eco RI and sub-cloning into pcLuc. All PCR products were first cloned into pGEM-T-easy vector and subcloned into pcLuc vector for reporter assay.
To engineer Flag-tagged expression constructs, the cDNAs encoding full-length HuR and its mutants were amplified by PCR using primers listed in Table 1 and cloned into pGEM-Teasy vector followed by releasing from pGEM-T-easy by Bgl II and Xho I digestion and subcloning into pCMV-Tag2B vector.
To generate GST-fusion protein-expressing constructs, the sequences encoding the PCI, SPT, and RP domains of eIF3a were amplified using PCR and primers in Table 1 and cloned into pGEM-T-easy vector. The cDNAs encoding these domains were then released from pGEM-T-easy by Eco RI and Xho I digestion and subcloned into the pGEX-4T-1 vector.
To generate His-tagged wild-type and mutant HuR constructs, total RNAs were extracted from H1299 cells and used as a template to amplify the sequences encoding full-length HuR and its mutants using RT-PCR and primers listed in Table 1, followed by cloning into pGEM-T-easy vector. These cDNA fragments were then released from pGEM-T-easy by Bgl II and Xho I digestion and subcloned into pET28a linearized by Bam HI and Xho I. All above constructs were authenticated using double-strand DNA sequencing.

RNA extraction and real-time RT-PCR
RNA extraction and real-time RT-PCR analysis was performed as described previously [22]. Briefly, total RNAs were extracted using RNeasy Mini Kit (Qiagen) and 1 μg of total RNAs were reverse transcribed to synthesize cDNA using iScriptTM cDNA Synthesis Kit (Bio-Rad). PCRs were carried out in ABI Prism@7000 Sequence Detection System (Applied Biosystems) using primers listed in Table 1 and SYBR Green diction according to manufacturer's instructions. The threshold cycle (Ct) was defined as the PCR cycle number at which the reporter fluorescence crosses the threshold reflecting a statistically significant point above the calculated baseline. The Ct of the target product was determined and normalized against that of GAPDH internal control. The relative RNA level = 2 ΔCt .

Sample preparation, Western blot, and co-immunoprecipitation
Sample preparation and Western blot analyses were performed as previously described [16]. Briefly, cells were lysed in TNN-SDS buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5% Nonidet P-40, 50 mM NaF, 1 mM sodium orthovanadate, 1 mM DTT, 0.1% SDS and 2 mM phenylmethylsulfonyl fluoride) at 4°C for 30 min followed by centrifugation at 12,000g for 10 min. The supernatant was collected and protein concentration was measured using Bradford method [22]. Cell lysates were separated by SDS-PAGE, transferred to PVDF membranes and probed with various primary antibodies. The immunoreactivity was detected using secondary antibody-conjugated horseradish peroxidase. Signals were captured using enhanced chemiluminescence and x-ray film.
Immunoprecipitation was performed as previously described [16,23,24]. Briefly, cell lysates (400 μg) were first precleared by incubation with 1 μg of normal IgG at 4°C for 1 h, mixed with 50 μl of protein G agarose beads (50% slurry) and incubated at 4°C for 2 h followed by centrifugation at 500g for 5 min. The cleared supernatants were incubated with 5 μg of anti-FLAG or anti-HuR antibodies at 4°C for 4 h, mixed with 50 μl of 50% protein G-agarose slurry. After overnight incubation at 4°C, the reaction was centrifuged to collect precipitates which were then washed five times with lysis buffer before being subjected to separation by SDS-PAGE and Western blot analysis.

Pulse and pulse-chase labeling of cells
Pulse and pulse-chase labeling were performed as previously described [16]. Briefly, H1299 cells were incubated for 2 hrs in the methionine-free medium supplemented with 7.5 μCi/ml [ 35 S]methionine followed by washing with PBS and lysate preparation for immunoprecipitation of Chk1. For the chase experiment, the pulse-labeled cells were washed twice with PBS and once with DMEM medium and incubated in DMEM medium supplemented with 100 μg/ml cold methionine. At different time points, the cells were harvested for cell lysate preparation and immunoprecipitation.

RNA pulldown assay
Biotin-labeled RNA probe was prepared using in vitro transcription as previously described [12]. Briefly, DNA templates were linearized using suitable restriction enzymes and used as a template for in vitro transcription using T7 or Sp6 RNA polymerase in the presence of Biotin-11-CTP. The biotin-labeled in vitro transcripts were purified using the Qiagen RNeasy mini kit and used as probes for pulldown assay by incubating with H1299 cell extracts at room temperature for 1 h. The RNA-protein complexes were then isolated using Streptavidin MagneSphere® Paramagnetic Particles and washed three times with wash buffer (10 mM N-2-hydroxyethylpiperazine-N′−2-ethanesulfonic acid, 20 mM KCl, 1 mM MgCl 2 , 1 mM dithiothreitol and 1 mM PMSF). The pulldown materials were then separated by SDS-PAGE and analyzed using Western blot probed by eIF3a or HuR antibodies.

In vitro pulldown assay
His-tagged proteins were expressed in BL21 (DE3) host and purified with Ni-NTA slurry while GST fusion proteins were expressed in DH5α host cells and purified using glutathione-Sepharose 4B according to manufacturer's instructions. Cell lysates were digested with RNase A to remove any RNAs that are potentially bound to eIF3a or HuR before purification.
In vitro pulldown assay was performed as previously described [9], Briefly, purified Histagged proteins was mixed with GST-fusion protein-immobilized beads in binding buffer (20 mM Hepes-KOH, pH 7.9, 200 mM KCl, 0.1% Nonidet P-40, 5 mM MgCl 2 , 1% BSA, 10% glycerol) and incubated at room temperature for 1.5 h. After extensive washes with the same buffer, the pulldown materials were separated by SDS-PAGE, transferred to PVDF membranes, subjected to Western blot analysis using His-tagged antibody.

RNA gel shift assay
About 10 7 cells were lysed in 10 mM Hepes-KOH, pH 7.9, 10 mM KCl, 1.5 mM MgCl 2 , 0.5 mM PMSF, 1 mM DTT by incubation on ice for 15 min and addition of NP-40 to a final concentration of 0.7%, followed by centrifugation at 1500g for 5 min to remove cell debris. The supernatant was collected and subjected to further centrifugation at 16,000g for 10 min. The supernatant was collected and stored in aliquots at −80°C. 32 P-UTP-labelled RNA probe was generated using in vitro transcription method as described above. About 10 5 CPM RNA probe were incubated with cell lysate at room temperature in binding buffer (15 mM Hepes-KOH, pH 7.9, 50 mM KCl, 10% glycerol, 0.2 mM DTT, 5 mM MgCl 2 , 200 μg/ml tRNA) for 30 min, followed by digestion with 100 Units of RNase T1 at 30°C for 15 min. For supershift assay, cell lysates were pre-incubated with HuR antibody for 30 min before incubation with RNA probes. The reaction mixture was then separated on a 5% native polyacrylamide gel and signals were captured by autoradiography using an X-ray film.

RNA-binding protein immunoprecipitation
RNA-binding protein immunoprecipitation was performed using Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore) according to the manufacturer's instructions. Briefly, cells were lysed in lysis buffer containing protease inhibitors, RNase inhibitors, and by freezing at −80°C overnight. The cell lysate was cleared by centrifugation at 12,000g for 10 min.
The magnetic beads-coupled with protein G was pre-incubated with HuR antibody for 30 min at room temperature followed by washing with RIP washing buffer for two times and mixing with cell lysate. Following incubation at 4°C for 6 h with rotation, the beads was washed with RIP wash buffer and the RNAs in the precipitate were extracted and used as a template for RT-PCR analysis.

Results eIF3a regulates Chk1 protein level
Hydroxyurea (HU), a ribonucleotide reductase inhibitor that causes DNA replication arrest, has been reported to increase the phosphorylation but reduce protein level of Chk1 [25]. We previously also found that eIF3a expression sensitizes cellular response to HU [19]. To investigate the possible relationship between eIF3a and Chk1, we analyzed the expression of eIF3a and Chk1 in H1299 cells following HU treatment. We also tested Chk2 in parallel. As shown in Figure 1A,B, treatment with 0.4 mM HU for 48 h significantly decreased the total protein level of Chk1 and eIF3a but had little effect on Chk2 protein level. The time-course study showed that HU-induced eIF3a reduction precedes the reduction in Chk1 ( Figure 1C).
To further investigate the possible relationship between eIF3a and Chk1, we first tested if knocking down eIF3a affect Chk1 expression. We again tested Chk2 expression in parallel. As shown in Figure 1D,E, eIF3a knockdown significantly reduced Chk1 expression but had little effect on Chk2. Consistently, overexpressing eIF3a in RIE and NIH3T3 cells increased Chk1 expression ( Figure 1F,G). On the other hand, overexpressing Chk1 or Chk2 and knocking down Chk1 or Chk2 had no effect on eIF3a expression ( Figure 1H-J). These findings suggest that eIF3a may regulate the expression of Chk1 but not Chk2 and that Chk1 and Chk2 unlikely regulate the expression of eIF3a.

eIF3a regulates the synthesis of Chk1 protein
To determine the mechanism of eIF3a regulation of Chk1 expression, we first determined the mRNA level of Chk1 using real-time RT-PCR in H1299 cells following eIF3a knockdown. As shown in Figure 2A, compared with the dramatic reduction in eIF3a mRNA, eIF3a knockdown had no effect on the mRNA level of Chk1. We next performed pulse labeling and pulse-chase experiments to determine if eIF3a regulates the synthesis or degradation of Chk1. As shown in Figure 2B, eIF3a knockdown caused drastic reduction in newly synthesized Chk1. On the hand, eIF3a knockdown had no effect on Chk1 degradation ( Figure 2C,D). Thus, we conclude that eIF3a likely regulates Chk1 expression by regulating its protein synthesis or translation of its mRNA.

Localization of the regulatory cis-elements in Chk1 mRNA
It has previously been suggested that cis-elements in regulating mRNA translation may be localized in the 5′-and/or 3′-untranslated regions (UTRs) [26][27][28]. To identify the potential regulatory cis-element in Chk1 mRNA, we generated chimera luciferase reporter constructs with the reporter open reading frame flanked by the 5′-and 3′-UTRs of Chk1 mRNA ( Figure 3A) and transiently transfected these constructs into cells with eIF3a overexpression or knockdown followed by determination of eIF3a effects on reporter expression. As shown in Figure 3C, eIF3a overexpression significantly up-regulated reporter expression in the presence of 3′-UTR alone or both 5′-and 3′-UTRs. Consistently, eIF3a knockdown significantly reduced reporter expression in the presence of 3′-UTR alone or both 5′-and 3′-UTRs ( Figure 3D). Alteration of eIF3a expression had no effect on the reporter expression in the presence of the 5′-UTR alone ( Figure 3C,D). These findings suggest that the regulatory cis-elements that are responsive to eIF3a may reside in the 3′-UTR of Chk1 mRNA.
We next analyzed the 3′-UTR sequence of Chk1 mRNA using RBPmap, a web server for mapping binding sites of RNA-binding proteins [29] and found three binding sites for HuR, a RNA binding protein. To test the possible involvement of these putative cis-elements on eIF3a regulation of the 3′-UTR activity, we systematically mutated these sites by deletion or mutation in the reporter constructs ( Figure 3B). The wild-type and mutant constructs were then transiently transfected into RIE cells with stable eIF3a overexpression followed by the determination of luciferase activity. As shown in Figure 3E, complete removal of all three binding sites by deletion or combining deletion with mutation abolished eIF3a-stimulated reporter expression. These findings suggest that eIF3a may regulate the translation of Chk1 mRNA via HuR-binding sites in the 3′-UTR.

HuR binding to the 3′UTR of Chk1 mRNA
To determine the possible involvement of HuR in eIF3a regulation of Chk1 mRNA translation, we first performed RNA mobility shift assay to determine if HuR binds to the 3′-UTR of Chk1. As shown in Figure 4A, the mobility of the 3′-UTR probe shifted by a binding protein that was identified as HuR using HuR antibody in a supershift assay. This binding was abolished when cold Chk1 3′-UTR RNA competitor was added. However, the non-specific RNA competitor, the 5′-UTR of RPA mRNA without HuR binding sequence and the 5′-UTR of Chk1, had no effect on HuR binding to the 3′-UTR probe.
To ensure that the binding occurs at the HuR-binding site in the 3′-UTR of Chk1, we performed RNA mobility shift assay using the 3′-UTR containing mutations of the HuRbinding site in the 3′-UTR of Chk1 (see Figure 3B). As shown in Figure 4B, only the wild type 3′-UTR was able to bind to HuR.
To further validate HuR binding to the 3′-UTR of Chk1, we performed RNA immunoprecipitation (RIP) followed by quantification using real-time RT-PCR. Figure 4C,D show that HuR antibody was able to pull down the 3′-UTR of Chk1 likely via HuR. Thus, the putative HuR-binding sites in the 3′-UTR of Chk1 mRNA are active sites for HuR to bind.

HuR regulates Chk1 expression
The binding of HuR to the 3′-UTR of Chk1 mRNA suggests that it may regulate HuR expression. To test this hypothesis, we determined the effect of HuR knockdown on Chk1 expression and its relationship with eIF3a. As shown in Figure 5A,B, eIF3a and HuR knockdowns both decreased Chk1 expression. However, eIF3a knockdown did not significantly affect HuR expression and vice versa ( Figure 5C). Thus, it is unlikely that eIF3a and HuR regulate Chk1 expression via regulating each other's expression. We next performed a pulldown assay using the 3′-UTR of Chk1 mRNA as a bait. As expected, the 3′-UTR of Chk1 mRNA successfully pulled down HuR protein ( Figure 5D). Interestingly, it also pulled down eIF3a. The mutant 3′-UTR without the HuR binding sites were unable to pulldown either HuR or eIF3a while the mutant with a single HuR binding site pulled down significantly less eIF3a and HuR compared with the wild type 3′-UTR. These findings suggest that eIF3a may also bind to the HuR-binding sites in the 3′-UTR of Chk1 mRNA possibly by interacting with HuR.

eIF3a interaction with HuR
To determine if eIF3a may interact with HuR, we performed co-immunoprecipitation assay using HuR antibody and lysates from H1299 and CaCo-2 cells. As shown in Figure 6A, IP with HuR antibody pulled down endogenous eIF3a in both H1299 cells and CaCo-2 cells, suggesting that HuR interacts with eIF3a. To further determine if this interaction is direct and to map the eIF3a binding site in HuR, we first generated a set of truncation mutant HuR with Flag tag (Figure 6B), which were transiently transfected into H1299 cells followed by immunoprecipitation using Flag antibody. Figure 6C shows that the wild-type and mutant HuR's are equally expressed and that the mutant with the third RRM (RNA recognition motif) deleted maintains minimum activity in binding to eIF3a. Further deletion to remove the hinge region completely abolished the binding activity of the HuR to eIF3a. Thus, the third RRM of HuR including its upstream hinge region likely contributes to HuR interaction with eIF3a.
Next, we mapped the domain in eIF3a responsible for interaction with HuR. For this purpose, we took advantage of the GST fusion protein containing three different domains of eIF3a ( Figure 6D), which were generated in our previous study [9], and used them as baits for pulldown assay of purified His-tagged HuR. As shown in Figure 6E, only the purified RP (10-amino acid repeat) domain of eIF3a was able to pull down the purified wild-type Histagged HuR. These findings suggest that eIF3a likely interacts with HuR directly via its RP domain.
To validate above findings, we generated His-tagged HuR mutants with different deletions as shown in Figure 6B, which were used as target for pulldown assay using GST-tagged domains of eIF3a. Figure 6E shows that the mutant HuR with the deletion of the third RRM retains minimal activity to interact with the RP domain of eIF3a. Further deletion of the hinge domain completely abolished the binding activity of HuR to the RP domain of eIF3a. These findings are consistent with the above observations of co-immunoprecipitation and indicate that the direct interaction between HuR and eIF3a is likely via interaction between the third RRM of HuR and the RP domain of eIF3a.

Discussion
In this study, we show that eIF3a positively regulates the expression of Chk1 at the translational level via interaction with HuR that binds to the 3′-UTR of Chk1 mRNA. The interaction between eIF3a and HuR occurs at the RP domain of eIF3a and the third RRM of HuR. The finding of the positive regulation of Chk1 synthesis by eIF3a is consistent with our previous finding that eIF3a regulates cell cycle progression but not distribution [19] due to Chk1's function in regulating all cell cycle checkpoints [21]. eIF3a has three putative domains including PCI, spectrin and RP domains (33). While we have shown previously that the spectrin domain is the docking site for eIF3b, g and i subunits to form the a:b:i:g subcomplex [9], the exact function of the PCI domain, which exists in proteasome, COP9, and initiation factor 3, remains unknown. In this study, we showed that the RP domain of eIF3a is responsible for its interaction with HuR in translational regulation of Chk1. Because the HuR-binding site is in the 3′-UTR of Chk1 mRNA and eIF3a may bind to the 5′-UTRs together with other initiation factors, it is possible that eIF3a may play a crucial role in accelerating Chk1 synthesis via circularizing its mRNA by bringing the 5′-and 3′-ends together to form an end-to-end complex ( Figure  7) as previously suggested for the PABP-eIF4G complex (see below). This possible mechanism of translational regulation may also occur with other mRNAs that has a HuRbinding sequence in the 3′-UTR. One example is p53, which has been shown to be translationally regulated by HuR binding to the 3′-UTR [30].
3′-UTR is a well-known cis-element that plays important roles in regulating mRNA stability, intracellular localization and translation [31][32][33]. Interaction of polyadenylationelement-binding protein (PABP) in the 3′-UTR with eIF4G [34,35] and eIF4B [36,37] in the 5′-UTR circularizes mRNAs, resulting in enhanced translation of the mRNA [38,39]. Most recently, the binding between eIF3h and a RNA-binding protein, METTL3, circularizes a subset of mRNAs and enhances their translations, which may contribute to oncogenesis [40]. Together with our findings, these observations suggest that different eIF3 subunits may interact with different RNA-binding proteins to circularize different mRNAs for enhanced translation. It remains to be determined, however, if eIF3a regulation of protein synthesis via HuR contributes to oncogenesis and drug sensitivity.
Previously, it has been shown that eIF3a regulates mRNA translation via binding to the internal ribosome entry site sequence in 5′-UTRs [12]. In this case, eIF3a binding suppresses the translation of the target mRNA and may contribute to cellular drug sensitivity. The translation of several additional mRNAs have also been shown to be suppressed by eIF3a although the mechanism of the eIF3a suppression of these mRNA translations is not yet known [16][17][18].
eIF3a interaction with HuR that binds to the 3′-UTR of target mRNAs to accelerate their translation may represent a new mechanism of eIF3a regulation of protein synthesis. Although we showed that all three putative HuR-binding sites in the 3′-UTR of Chk1 mRNA may contributes to eIF3a regulation, whether they have synergy and co-ordinate HuR interactions with eIF3a remains unresolved. Mutations of each binding site individually or in different combinations may be needed to delineate the specific role of each binding site. It also remains unknown if translation of all mRNAs containing a HuR-binding site in their 3′-UTRs are regulated by eIF3a and if and how eIF3a regulates translation of mRNAs containing HuR-binding site in their 5′-UTRs. Whether eIF3a potentially interacts with other RNA-binding proteins such as PABP to circularize mRNAs in accelerating general protein synthesis also remains to be addressed. We are currently working to answer these questions.
The role of HuR in regulating protein synthesis has been shown previously by stimulating mRNA stability and translational efficiency [41][42][43][44]. Recent evidence suggests that HuR regulation of mRNA stability may be via inhibiting the function of microRNAs [45][46][47]. However, it has also been shown that HuR interacts with the 3′-UTR of the p53 mRNA and increases p53 protein synthesis rate without affecting the mRNA level [30]. Our findings here show that HuR regulation of protein synthesis may be via its interaction with eIF3a to accelerate mRNA translation initiation. It is, however, noteworthy that the third RRM of HuR may be responsible for interaction with eIF3a. Thus, it remains unclear if the binding of eIF3a to the third RRM of HuR possibly inhibits HuR recognition of and binding to the 3′-UTR of target mRNAs such as Chk1 or p53. Although this possibility appears to be contradictory to the finding that HuR binding to the 3′-UTR increases translation of the mRNA, future studies will be needed to evaluate this possibility.

Funding
This work was supported in part by a grant R01 CA211904 (JTZ) from the National Cancer Institute/NIH.       Schematic model of eIF3a regulation in Chk1 mRNA translation via binding to HuR to circularize Chk1 mRNA.