|Home | About | Journals | Submit | Contact Us | Français|
Post-transcriptional regulation is an important control mechanism governing gene expression in neurons. We recently demonstrated that VCX-A, a protein implicated in X-linked mental retardation, is an RNA-binding protein that specifically binds the 5´ end of capped mRNAs to prevent their decapping and decay. Previously, expression of VCX-A was reported to be testes-restricted. Consistent with a role in cognitive function, we demonstrate that VCX-A is ubiquitously expressed in human tissues including the brain. Moreover, retinoic acid-induced differentiation of human SH-SY5Y neuroblastoma cells promoted the accumulation of VCX-A in distinct cytoplasmic foci within neurites that colocalize with staufen1-containing RNA granules, suggesting a role in translational suppression and/or mRNA transport. Exogenous expression of VCX-A in rat primary hippocampal neurons, that normally do not express the primate-restricted VCX proteins, promoted neurite arborization and, shRNA-directed knockdown of the VCX genes in SH-SY5Y cells resulted in a reduction of both primary and secondary neurite projections upon differentiation. We propose the cap-binding property of VCX-A reflects a role of this protein in mRNA translational regulation. In support of this hypothesized role, we demonstrate that VCX-A can specifically bind a subset of mRNAs involved in neuritogenesis and is also capable of promoting translational silencing. Thus, VCX-A contains the capacity to modulate the stability and translation of a subset of target mRNAs involved in neuronal differentiation and arborization. It is plausible that defects of these functions in the absence of the VCX genes could contribute to a mental retardation phenotype.
The regulation of gene expression has a fundamental impact on the development, differentiation, and proliferation of cells and can occur at multiple levels including mRNA metabolism. An important property of all mRNAs is their inherent stability and their capacity to be translated into protein. In eukaryotes, the 5´ cap and 3´ polyadenylated (poly(A)) tail are critical structures for mRNA stability and translation. Following removal of their poly(A) tail, mRNAs are primarily degraded by two exonucleolytic pathways involving either continual decay from the 3´ end or a decapping step preceding 5´ to 3´ exonucleolytic decay (Liu and Kiledjian, 2006; Garneau et al., 2007). Removal of the cap renders the mRNA translationally incompetent since the cap is required for ribosomal recruitment and the exposed 5´ end is susceptible to exoribonucleolytic decay.
We recently identified VCX-A as a regulator of mRNA stability and showed that it can function as an inhibitor of the Dcp2 mRNA decapping enzyme (Jiao et al., 2006). VCX-A is a member of a family of proteins encoded by distinct genes located on the X chromosome (Xp22), termed VCX-A, VCX-B, VCX-B1 and VCX-C, as well as two identical copies on the Y chromosome, termed VCY. The proteins encoded by the different VCX/Y genes have a greater than 95% identity at the amino terminal half of the protein and primarily differ in the number of 10-amino acid repeats at their C-termini. Interestingly, the VCX-A gene was identified as the defining interval in patients with X-linked mental retardation (MRX) (Fukami et al., 2000). However, subsequent studies indicated that the potential role of the VCX proteins in MRX is probably more complex (Van Esch et al., 2005; Hosomi et al., 2007; Macarov et al., 2007; Cuevas-Covarrubias and Gonzalez-Huerta, 2008) and likely involves a cumulative VCX dosage rather than reflecting the output of the VCX-A gene alone (Van Esch et al., 2005).
In the mammalian neurons, transcription and translation are often uncoupled in time and space, and this autonomous control of gene expression can play a critical role in the enduring (late-stage) form of synaptic plasticity thought to underlie learning and memory (Frey U, 1988; Nguyen PV, 1994). Various strategies for the regulation of protein translation are used in neurons to control local protein synthesis in dendrites (reviewed in (Sutton MA, 2005)). These regulatory processes are involved in mRNA localization, translation, and degradation. In this context, it is not surprising to find that several RNA-binding proteins are known to contribute to synaptic plasticity through their regulation of specific mRNA subsets. These include FMRP, Cytoplasmic Polyadenylation Element Binding (CPEB )factor and Neuro-oncological ventral antigen (Nova) (Wu et al., 1998; Huber et al., 2002; Si K, 2003; Alarcon JM, 2004; Huang CS, 2005; Koekkoek et al., 2005). In this report, we demonstrate that the VCX-A gene product is a translational inhibitor and is involved in the regulation of neuronal cell morphogenesis and neuritogenesis.
The pET28a-VCX-A, pET28a-VCX-AΔN40, pcDNA3-FLAG-VCX-A, pcDNA3-FLAG-VCX-AΔN40 and pIRESpuro3-VCX-A plasmids have been described (Jiao et al., 2006), as has the pSV2ALΔ5′ plasmid (de Wet et al., 1987). The pIRESpuro3-VCX-A plasmid expressing myc-tagged VCX-A was used to generate the VCX-A expressing stable cell line. pET28a-UlP1 and pcDNA3-FLAG-IKK plasmids were kindly provided by C.D. Lima (Sloan-Kettering Institute) and G. Xiao (University of Pittsburgh), respectively. The two hairpin-containing VCX sequences were generated according to Paddison et al. (Paddison and Hannon, 2002). The annealed DNA oligonucleotides were inserted into the BseR1 and BamHI sites of the pSHAG-1 vector to generated pSHAG-VCX-A. The pSHAG-1 vector contains a U6 promoter shRNA expression cassette and was kindly provided by G. Hannon (Cold Spring Harbor Laboratory). The generated hairpin RNA targets the VCX mRNA sequence, 5’ CAC TGA GTC AGG AGA GCG AGG TGG AAG AA 3’ (VCX-A 645 to 673bp and 705 to 733bp).
GST fusion and His-tagged proteins were expressed and purified according to the manufacturer’s instructions (GE Healthcare, Piscataway, NJ, USA and Novagen, San Diego, CA, USA, respectively) with minor modification as previously described (Jiao et al., 2006).
Human K562 erythroleukemia cells, Hela epithelial carcinoma cells, 293T embryonic kidney cells, SH-SY5Y neuroblastoma cells, mouse NIH 3T3 embryonic fibroblast cells and MEL mouse erythroleukemia cells were obtained from ATCC (Manassas, VA, USA) and grown according to the supplier. All transfections were carried out using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. 293T expressing a stably integrated VCX-A gene or VCX-AΔN40 mutant were generated by transfecting 293T cells with pIRESpuro3-VCX-A, VCX-AΔN40 mutant or pIRESpuro3 vector plasmids and selected with 3 µg/ml puromycin (Sigma-Aldrich, Louis, MO, USA) as described previously (Jiao et al., 2006). Serial dilutions were carried out to isolate a monoclonal cell line expressing high levels of VCX-A. Transcriptional arrest experiments were carried out by treating 293T embryonic kidney cells stably expressing myc-tagged VCX-A, VCX-AΔN40 mutant, or vector control with 5µg/ml actinomycin D for 0, 0.5, 1, 2, 3 hours.
Neuronal differentiation of the human neuroblastoma SH-SY5Y cell line was induced by the addition of 10 µM Retinoic Acid (RA; Sigma-Aldrich, Louis, MO, USA) to the culture media for 2 days. The differentiation phenotype of SH-SY5Y cells with RA treatment was confirmed by photography under a phase contrast microscope (Zeiss Axiovert 100 M microscope) to follow neurite generation.
Human tissues were obtained from the Cancer Institute of New Jersey. Tissue extracts and cell extracts were prepared by sonication as previously described (Rodgers et al., 2002) in PBS buffer containing protease inhibitors (Complete Protease Inhibitor Cocktail Tablets; Roche, Switzerland).
The 32P-cap-labeled pcP RNA was generated as previously described (Wang et al., 1999; Wang et al., 2002). The Firefly luciferase reporter gene was PCR-amplified from pSV2ALΔ5′ with a 5´ specific reporter gene primer containing a T7 promoter primer adaptor (5´ CGTAATACGACTCACTATAGGGCATTCCGGTACTGTTGGTAAAATGG 3´) and a 3´ specific reporter gene primer containing an A60 tail (5 ´ T60 GCCGCCCACTCAGACTTTATTCAAAGACC 3´). The PCR product was used as template to transcribe RNA with T7 RNA polymerase (Promega, Madison, WI, USA). The cap analogue m7GpppG was added to the in vitro transcription system to generate capped-Firefly luciferase RNA. The generated polyadenlylated luciferase reporter mRNA was used as substrate for protein translation.
The antiserum to VCX was commercially generated (Cocalico Biologicals, Reamstown, PA, USA) by immunizing rabbits with recombinant His-VCX-A. The polyclonal sera was affinity-purified with GST-VCX-A fusion protein consisting of the C-terminal 105 amino acids of VCX-A coupled to an NHS column according to the manufacturer’s instructions (GE Healthcare Life Science, NJ, USA).
Cell or tissue extracts were resolved by 12.5% SDS-PAGE. VCX-A C-terminal affinity purified polyclonal antibody (1:200 dilution), monoclonal anti-eIF4E antibody (1:1500 dilution; Transduction Laboratories, Lexington, KY), polyclonal anti-CaMKI antibody (1:5000 dilution; Abcam Inc., Cambridge, MA), and monoclonal anti-GAPDH antibody (1:2000 dilution; Abcam Inc., Cambridge, MA), were used for Western blot analysis and visualized using secondary antibodies coupled to horse radish peroxidase (Jackson ImmunoResearch, West Grove, PA) and chemiluminescence (ECL; GE Healthcare Life Science, NJ).
Total RNA was isolated from human tissues and cell lines using Trizol (Invitrogen, Carlsbad, CA, USA) and subjected to RQ1 RNase-Free DNase (Promega, Madison, Wisconsin) treatment to prevent genomic DNA contamination. The RNAs were reverse transcribed into cDNA with Oligo d(T) primer using the Superscript II reverse-transcription kit according to the manufacturer (Invitrogen, Carlsbad, CA). PCR amplification with the forward primer 5' ATG AGT CCA AAG CCG AGA GC 3' and reverse primer 5' TCA CGG CCG CCA TCT TTG TC 3' results in the amplification of all members of the VCX/Y gene family and generates a 187 bp fragment. VCY was amplified by the VCY-specific primers 5' ATG AGT CCA AAG CCG AGA GCC TC 3' and 5' TGA CTC AGT GGT TCT TCC ACC TC 3' to generate a 451 bp fragment. Reverse transcription products (4 µl) were used as PCR templates and amplified for 40 cycles. The PCR products were resolved by a 2.5% agarose gel.
The stability of electroporated firefly luciferase mRNA and endogenous CaMKI mRNA were determined in 293T cells stably expressing myc-tagged VCX-A, VCX-AΔN40 truncation, or vector control by quantitative real-time PCR. Transcription was block by the addition of 5µg/ml actinomycin D to the cell culture medium. GAPDH mRNA was used as an internal control. Real-time PCR was carried out using SYBR green PCR core reagent (Applied Biosystems, Foster City, CA) according to the manufacturer's instructions. Data were analyzed by the comparative Ct method as previously described (Livak and Schmittgen 2001). Sequences for all primer sets are listed in Supplemental Table 1.
Formalin-fixed paraffin-embedded 10 µm human brain sections were obtained from the Alzheimer Disease Research Center at Case Western Reserve University. Immunohistochemistry assay was performed as described previously (Busser et al., 1998). Briefly, deparaffinized sections were treated with citrate buffer to retrieve antigen, and then soaked in 0.3% hydrogen peroxide to remove endogenous peroxidase activity. Affinity purified VCX specific antibody was diluted in blocking solution (1:100) and used to detect VCX in the tissue and detected with biotinylated goat anti-rabbit secondary antibody (1:400), Avidin-biotin complex horseradish peroxidase and DAB substrate (Vector Laboratories, Burlingame, CA). Sections were counter-stained with hematoxylin to visualize nuclei.
SH-SY5Y cells were grown on poly-lysine coated glass coverslips in six-well plates to 70% confluency. The coverslips containing the cells were treated as previously described (Liu et al., 2004). Affinity-purified VCX-A C-terminal rabbit polyclonal antibody was used as primary antibodies to detect endogenous VCX and monoclonal anti-stau1 antibody (1:80 dilution; Sigma-Aldrich, St. Louis, Missouri) was used to detect staufen-1 containing neuronal RNA granules. Monoclonal anti-neuron-specific class III β-tubulin (TuJ1) antibody (Abcam Inc., Cambridge, MA) (1:1000 dilution) was used as a neuronal marker. A goat anti-rabbit or goat anti-mouse secondary antibody conjugated with Texas Red or FITC (Jackson ImmunoResearch, West Grove, PA) were used as noted in the figure legends at 1:200 dilutions. The nuclei were stained with 1µg/ml DAPI for 3 minutes. The images were obtained with a Zeiss Axiovert 100 M microscope.
The in vitro transcribed capped-Firefly luciferase mRNA (1 µg) was electroporated into 293T cells in Opti-MEM® I reduced-serum medium (Invitrogen, Carlsbad, CA) in a 4 mm gap cuvette using a BioRad Genepulser charged at 320 V, 250 µF. All regents and cuvettes were kept on ice prior to electroporation. Following the discharge, the cells were washed once with PBS (25°C) and immediately resuspended in prewarmed D-MEM/10% FBS medium at 37°C and aliquoted into different plates. The cells were harvested at the indicated time points to detect luciferase activity.
Firefly luciferase activity was detected using a Luciferase Assay System (Promega, Madison, WI) according to the manufacturer’s guidelines using GloMax®-Multi Luminescence Module (Promega, Madison, WI, USA).
Anti-Flag monoclonal antibody coupled to M2 agarose beads (Sigma-Aldrich, St. Louis, Missouri) were used for the coimmunopurification assays. FLAG-tagged VCX-A or mDAZL were transiently overexpressed in human SH-SY5Y neuroblastoma cells for 48 hours and differentiated with 10 µM RA. SH-SY5Y cells transfected with the FLAG-tagged empty vector were used as a negative control. Anti-Flag M2 agarose beads (30 µl) were incubated with 1 mg cell extract expressing FLAG-VCX-A or FLAG-mDAZL at 4°C for 1 hour in PBS with moderate shaking in the presence of 1× protease inhibitor cocktail (Roche, Switzerland) and RNase inhibitor (1:50 dilution; Promega, Madison, WI). Following three washes with PBS/0.1% Triton X-100, the coimmunopurified mRNAs were eluted from the M2 agarose beads with TE/1% SDS and oligo d(T) reverse transcribed and PCR amplified with transcript specific primers (listed in supplemental Table 1) and PCR products resolved on a 2.5% agarose gel. SH-SY5Y total RNA RT-PCR was used as a positive control for the PCR reactions.
293T VCX-A stable cells and 293T vector control cells were transiently transfected with the firefly luciferase reporter plasmid pSV2ALΔ5′. Cells were cultured at 37°C for 16 hours to approximately 60–70% confluency. Cycloheximide (100 µg/ml) was added to the medium, and cells were cultured for 15 minutes followed by washing with ice cold PBS containing cycloheximide. The cells were harvested and lysed with ice-cold polysome fraction lysis buffer (0.15 M KCl, 15 mM MgOAc, 20 mM Tris-HCl (pH7.5), 1 mg/ml Heparin, 1 mM DTT, 100 µg/ml cycloheximide, 0.5% Triton X-100, 1 × protease inhibitor cocktail (Roche, Switzerland) and RNase inhibitor (1:50 dilution; Promega, Madison, WI)). Cell extracts were spun at 10,000 × g for 10 minutes at 4°C. The supernatant was transferred to a fresh tube and OD260 was measured. Cell extracts (100 × OD260) were layered on 12 ml 10% to 50% sucrose gradients and centrifuged for 4 hours at 23K rpm in a Sorvall Superspin 600 rotor at 4°C. A254 was monitored using a Model 185 density gradient fractionator (ISCO, Inc., Lincoln, NE) and each 1 ml gradient fraction was collected.
Protein aliquots from the gradients were precipitated with trichloroacetic acid (TCA) and resolved by 12.5% SDS-PAGE for Western blot analysis and detected with polycolonal anti-c-Myc antibody (1:500 dilution, Santa Cruz Biotechnology, Santa Cruz, CA). For Northern blot assay, the RNAs were isolated with 1% SDS and 10mM EDTA, precipitated with isopropanol after Phenol/Chloroform extraction and resolved on a 1% formaldehyde denaturing agarose gel. RNA was transferred to a Hybond-N member (GE Healthcare Life Science, NJ), and cross-linked to membrane at 80°C under vacuum for 1 hour. DNA-specific probes were labeled with [α-32P]dCTP by PCR. The hybridization was carried out at 60°C overnight and washed with 0.1% SSC/0.1% SDS. Quantifications were carried out using a Molecular Dynamics PhosphorImager (Storm860) using ImageQuant-5 software.
Primary neuronal cultures at a density of approximately 1800 cells/mm2 were prepared from hippocampi of rat embryos at 18 d gestation and grown as previously described (Firestein et al., 1999; Chen and Firestein, 2007). After 10 days in vitro (DIV), neurons were transfected with plasmids encoding FLAG-tagged VCX-A or VCX-AΔN40 using Effectene (Qiagen, Hilden, Germany). FLAG-tag empty vector plasmid was used as a control. pEGFP-C1 (BD Biosciences, Franklin Lakes, NJ) was co-transfected to enable detection and visualization of transfected cells. Forty-eight hours post-transfection, neurons were fixed in 4% paraformaldehyde in PBS for 15 minutes and immunostained using a monoclonal anti-FLAG tag antibody (Sigma-Aldrich, Louis, Missouri) and polyclonal anti-GFP antibody (Abcam Inc., Cambridge, MA). The transfected neurons were identified and visualized by GFP fluorescence at 400× using Zeiss Axiovert 100 M fluorescence microscope. Rat primary neuron neurite numbers and human SH-SY5Y neuroblastoma cell neurite numbers were counted from digitally obtained images of cells under experimentally blind conditions as previously described (Akum et al., 2004). Only those processes >3 µm were counted as neurites (Akum et al., 2004). Data were analyzed using GraphPad InStat software (Graph Pad, San Diego, CA) with ANOVA followed by the appropriate multiple comparisons test or with Student’s t test when appropriate.
VCX gene expression was reported to be restricted to the testis (Fukami et al., 2000; Lahn and Page, 2000; Zou et al., 2003). Nevertheless, we recently reported the cloning of the VCX-A gene product from human K562 erythroleukemia cells (Jiao et al., 2006), suggesting perhaps a preferential, rather than exclusive, expression in the testis. To determine whether the VCX/Y genes are more ubiquitously expressed, we carried out reverse transcriptase (RT)-PCR of different human and mouse tissues and cell lines using primers that correspond to conserved sequences at the 5´ end of all of the VCX/Y mRNAs. As shown in Figure 1A and 1B, VCX mRNAs can be detected in all human tissues and cell lines tested but not in the mouse erythroleukemia cell mRNA negative control (Figure 1B, lane 4). Interestingly, the use of primers that exclusively detect the Y chromosome VCY gene products reveals that expression of this gene was restricted to testes (Figure 1A lane 5). To further expand upon the expression of the VCX genes, primers that specifically detect VCX, but not VCY, mRNA were used. As expected, VCX mRNA can be detected in all the human tissue and cell lines tested (Figure S1).
To determine whether VCX proteins can be detected in tissues other than testes, we generated antibodies directed against the VCX protein and probed various human tissues for the presence of VCX. Western blot analysis revealed the presence of VCX in human testes extract (Figure 1C, lane 4) as well as in human brain (lane 1 and 2) but not in mouse brain and mouse testes (lanes 3 and 5, respectively). Expression in human brain was further confirmed with immunohistochemistry. Paraffin sections of human autopsy material probed with affinity purified polyclonal VCX-specific antibody revealed a clear presence of the VCX proteins in neuronal cells of human brain (Figure 1E, panels b, d and f). The staining appeared equally robust in both neocortex and hippocampus. The cell bodies were strongly stained, but a significant level of immunoreactivity was found in the neuropil as well (compare 1E panels b, d, and f with panels a, c and e). This is consistent with a function in both the perikaryon and proximal dendrites. The nucleus, though positive, was less darkly stained in most cells.
VCX protein is abundant in human K562 erythroleukemia cells (Figure 1D, lane 1), and although at lower levels, is also detected in the transformed human cell lines tested (lane 2 to 6) but not in mouse MEL cells (lane 7). Due to a greater than 92% identity between the different VCX and VCY proteins (Fukami et al., 2000; Lahn and Page, 2000), we are unable to unambiguously determine the identity of the multiple detected protein bands in testes. However, RT-PCR analysis revealed the expression of multiple VCX genes in all tested human tissues and cell lines (Figure S1).
Collectively, our data are consistent with previous reports indicating that VCX/Y genes are restricted to simian primates but not prosimian or non-primate mammals (Fukami et al., 2000; Lahn and Page, 2000). However, in contrast to these reports, our data indicate that only the VCY genes show testes-specific expression while the VCX proteins are present in all tissues we have tested, including neurons of the central nervous system.
VCX-A was initially identified as a candidate gene implicated in X-linked mental retardation (Fukami et al., 2000), suggesting a specific function in neurons. To determine a potential role for VCX-A in neuronal cells, we used the human SH-SY5Y neuroblastoma cell line to determine the fate of VCX upon neuronal differentiation. Induction of SH-SY5Y cells with retinoic acid (RA) for two days promotes differentiation into a neuronal phenotype (Pahlman et al., 1984) with an increase of neurite projections (Figure 2, compare panel A and D). Although VCX protein levels were not detectably changed upon induced differentiation (data not shown), VCX foci number and size increased (panel E). In particular, VCX foci were prominent within the neurite projections (expanded view in panels C and F), suggesting a function in neuronal cell projections. A similar increase in VCX foci was also observed upon differentiation of human H9 neuronal stem cells (data not shown). These data suggest that VCX could be involved in neuronal process development and/or maintenance. Interestingly, the VCX foci did not colocalize with cytoplasmic Processing (P)-bodies or stress granules (Figure S2), indicating that they are distinct from these two well-characterized cytoplasmic structures, but they do colocalize with staufen1-containing RNA granules (Figure 2G–2I), indicating they may be involved in mRNA transport or mRNA translational suppression (Kiebler and Bassell, 2006).
Normal cognitive function depends on proper connections between neurons. Abnormal morphogenesis and numbers of dendrites are associated with cognitive impairments, such as in Fragile X syndrome (McKinney BC, 2005; Grossman AW, 2006). Association of the VCX-A gene with MRX indicates that VCX-A, and likely other members of this family of proteins, may affect normal synaptogenesis or neurite numbers. The observed increase in accumulation of VCX protein into foci upon differentiation indicates that VCX may affect normal neuronal morphology. Since the VCX/Y genes are only present in primates and absent from lower mammals ((Fukami et al., 2000; Lahn and Page, 2000) and Figure 1), we used primary hippocampal neurons from rats as a model of a VCX-null system. Neurons were isolated from hippocampi of rat embryos at 18 days gestation and co-transfected with a plasmid expressing FLAG-VCX-A and a plasmid expressing green fluorescent protein (GFP) as a marker to detect transfected cells (Akum et al., 2004). Upon expression of FLAG-VCX-A for 48 hours following the initial transfection, the number of primary and secondary neurites both significantly increased compared to the control neurons (1.46 and 1.51 fold respectively; n=25, P<0.0001; Figure 3 A and B). The increase was a consequence of VCX-A since transfection of a construct expressing a mutant VCX-A protein lacking the first 40 amino acids that is unable to bind the 5´ cap (VCX-AΔN40; Jiao et al., 2006) did not affect neurite branching (n=26, p>0.05).
The above data indicate that overexpression of VCX-A in a cell that normally does not contain this protein can promote neurite arborization, suggesting that VCX-A is a positive regulator of neurite outgrowth. To further expand upon the role of VCX proteins, VCX levels were reduced in SH-SY5Y cells using a VCX-specific shRNA-directed knockdown (Figures 4A and Figure S3). The consequence on neurite outgrowth in SH-SY5Y cells upon VCX knockdown is shown in Figure 4B. Following transfection of either a control or VCX-directed shRNA expression plasmid into SH-SY5Y cells, the cells were induced with 10 µM RA for two days. A two-fold reduction in both primary (2.12) and secondary (2.17) neurite numbers was observed in the VCX knockdown cells relative to the control knockdown cells (Figure 4B). A reduction in neurite numbers was also detected upon expression of the VCX-AΔN40 truncated protein in SH-SY5Y cells, suggesting the mutant VCX-A might function as a trans-dominant negative (Figure 4C). However, expression of wild type VCX-A into SH-SY5Y cells, which already express endogenous VCX proteins, had no appreciable consequence on differentiation with RA. Interestingly, irrespective of changes in neurite numbers with the knockdown of VCX-A or expression of wild type or truncated VCX-A, neurite length remained unchanged (Figure S4), indicating that VCX-A is involved in the regulation of neurite number but not neurite length. Collectively, these data support a role for the VCX proteins as modulators of the number of neuronal cell projections.
VCX-A is a non-canonical cap binding protein that is unable to bind the cap directly but only binds the cap when it is linked to an RNA (Jiao et al., 2006). This preferential binding to the 5′ end of an mRNA can lead to inhibition of decapping and stabilization of the mRNA (Jiao et al., 2006). The requirement for accessibility of the 5´ cap for mRNA translation in addition to its presence in staufen1-containing RNA granules (Figure 2G–2H) also suggests that VCX-A could be a modulator of mRNA translation. Consistent with this hypothesis, addition of increasing concentrations of recombinant VCX-A to an in vitro mRNA translation reaction resulted in a dramatic inhibition of translation as determined by the level of luciferase activity (Figure 5). In contrast, the VCX-A N-terminal truncation protein VCX-AΔN40, which is unable to bind the 5´ cap (Jiao et al., 2006) and an unrelated control protein UlP1 were unable to inhibit translation (Figure 5A). We next tested whether VCX-A also inhibits protein translation in cells by assessing translation derived from a capped and polyadenylated firefly luciferase mRNA transfected into 293T cells overexpressing either the FLAG-tagged VCX-A protein or control proteins. As shown in Figure 5B, translation was delayed in cells expressing FLAG-tagged VCX-A compared to cells overexpressing the FLAG-VCX-AΔN40 mutant protein or a control unrelated FLAG-IKK protein. The decrease in luciferase activity was not due to a decrease in luciferase mRNA levels since, as expected, the stability of this mRNA increased in the presence of VCX-A in these assays (Figure 5C). The above data indicate that VCX-A can function as a translational repressor both in vitro and in cells.
To further address the mechanism by which VCX-A can inhibit translation, we sought to identify cognate mRNAs that VCX-A might bind to and potentially regulate. The RNA-binding property of VCX-A (Jiao et al., 2006) indicates that it may specifically bind a subset of mRNA and regulate their expression. FLAG-tagged VCX-A expressed in human SH-SY5Y neuroblastoma cells was immunopurified along with bound mRNA using a FLAG-M2 column. mRNAs co-purifying with FLAG-VCX-A were isolated and reverse transcribed into cDNAs. A small subset of mRNAs known to be modulators of neuritogenesis were tested for their potential selective binding by VCX-A. The mRNA encoding two regulators of dendrite branching, glutamate receptor interacting protein 1 (GRIP1) and Calcium/calmodulin-dependent protein kinase I (CaMKI) mRNA were bound by FLAG-VCX-A but not the germ line-specific mDAZL control protein (Figure 6A). Three other mRNAs encoding proteins involved in neuritogenesis, RhoA, EphB2 and Sema3A, were not specifically bound by FLAG-VCX-A (Figure 6A, compare lane 2 to 3). These data demonstrate that VCX-A can specifically bind CaMKI and GRIP1 mRNAs. The binding to CaMKI is noteworthy as this is a low abundance mRNA that is not detected in the total RNA lane but is nonetheless selectively enriched by its association with FLAG-VCX-A.
To assess whether there is a functional significance for the binding of VCX-A, polysome profiles were analyzed in cells overexpressing VCX-A. Extract obtained from a 293T cell line containing a stably integrated VCX-A gene or a control vector were generated and used. A clear increase in 80S monosomes and a corresponding decrease in polysomes are apparent in cells overexpressing VCX-A relative to control cells (Figure 6B, panel a), consistent with a translational inhibition property for VCX-A. An increase in the 80S population is indicative of a block in translation initiation. Western blot analysis revealed that exogenous myc-VCX-A protein was predominantly at the top of the polysome profile gradient within the free messenger ribonucleoprotein particle (mRNP) fractions (Figure 6B, panel b, lanes 13 and 14) and residual amounts within the polysome fractions. We next tested the mRNA distribution of an exogenously transfected firefly luciferase mRNA whose translation was inhibited by VCX-A in vitro (Figure 5) as well as the endogenous CaMKI mRNAs within the polysome distribution. Northern blot analysis revealed that the silenced exogenous firefly luciferase mRNA was shifted from the polysomal fractions in the control cells to the mRNP fractions in the presence of myc-VCX-A (Figure 6B, panel c, compare lanes 7 – 12 to 19 – 24 and 1–2 to 13–14). The re-distribution of mRNAs from polysomes towards monosomes was also observed with the CaMKI mRNA detected by RT-PCR in the presence of overexpressed VCX-A protein (panel f). The effect of myc-VCX-A on translational silencing is relatively specific as shown by the fact that the distribution of RhoA mRNA was not detectably altered with an increase in VCX-A protein (panel d). These data suggest that VCX-A may function to modulate mRNA translation.
To directly test the stability and translation of the CaMKI mRNA in cells, 293T cells were treated with actinomycin D to block transcription and CaMKI protein and mRNA levels were followed over time. Endogenous CaMKI mRNA has an approximate 80-minute half-life in 293T cells, which remained constant in the presence of overexpressed VCX-A or VCX-AΔN40 truncated protein (Figure 6C) as did the steady state level of CaMKI mRNA in the different cells (data not shown). This is in contrast to the increased stability of the transfected luciferase mRNA upon overexpression of VCX-A (Figure 5C), suggesting a potential transcript specific role of VCX-A in mRNA stability. However, consistent with the polysome profile, CaMKI protein levels were lower in cells overexpressing VCX-A compared to the controls (Figure 6D). Collectively, our data support a role for VCX-A in proper neuritogenesis and provide insight into a possible role for these proteins in cognitive function by the post-transcriptional control of gene expression.
VCX-A was recently shown to be a member of a growing list of noncanonical cap binding proteins capable of simultaneously binding the 5´ cap and RNA, and as such, can modulate mRNA decapping (Jiao et al., 2006). We now demonstrate that VCX-A is also a regulator of mRNA translation and its capacity to bind the 5´ end of an mRNA promotes inhibition of mRNA translation initiation. These properties of VCX-A suggest that it has the potential to bind a subset of mRNAs and stabilize and silence their expression. The positive effect of VCX-A in neurite outgrowth combined with the translational silencing by VCX-A suggests that it could modulate a subset of mRNAs encoding regulators of neuritogenesis.
VCX-A is implicated in X-linked mental retardation (Fukami et al., 2000) but was reported to be testis-specific (Fukami et al., 2000; Lahn and Page, 2000). Indeed, by RT-PCR, the Y chromosome localized VCY genes are exclusively and robustly expressed in male germ cells (Figure 1A). Despite the reported testis germ cell restriction of the VCX genes (Fukami et al., 2000; Lahn and Page, 2000; Zou et al., 2003), we initially cloned the cDNA of VCX-A from human K562 erythroleukemia cells (Jiao et al., 2006). The presence of the VCX-A mRNA in erythroleukemia cells indicated that expression of this gene is not restricted to testis. Western blot and RT-PCR analyses demonstrated that the VCX genes are expressed in all human tissues and cell lines we have tested thus far, including human brain (Figure 1 and Figure S1). VCX proteins are also expressed in human neurons as determined by immunohistochemistry of postmortem human brain hippocampal and cortical sections (Figure 1). Collectively, these data strongly indicate that the VCX genes are ubiquitously expressed in human tissues, and notably in neurons, consistent with a role in cognitive function. As the VCX/Y genes are only present in simian primates (Fukami et al., 2000; Lahn and Page, 2000) it is plausible that they appeared more recently in evolution. We have also failed to identify a homologous protein or nucleic acid sequence in mouse or rat via the mining of currently available databases. Although an obvious homolog is not detectable, we cannot rule out the presence of a functional homolog in other mammalian organisms.
The presence of VCX proteins in discrete foci that colocalize with staufen1-containing RNA granules suggests that these might be sites of VCX proteins in transit with silenced mRNAs. The preferential binding property of VCX-A to the 5′ end of mRNAs can both lead to stabilization of an mRNA (Figure 5C) as well as inhibition of translation (Figure 5B and Figure 6C) are consistent with this possibility. Interestingly, although VCX-A has the potential to stabilize an mRNA (Figure 5C), this function appears to be transcript-specific as it did not stabilize the CaMKI mRNA (Figure 6C) suggesting a complex array of functions. At present, whether the mRNA stabilizing or translational silencing functions of VCX more significantly contribute to neuritogenesis is not known, but it is likely that a combination of both properties will be involved. The simultaneous function of both of these events could lead to assembly of mRNAs into stabilized and translationally silenced mRNPs.
Three important observations demonstrate a role for the VCX proteins in neurite number and provide insight into its potential role in cognitive function. First, the enhanced presence of VCX-foci within neurites following induction of SH-SY5Y cells by RA (Figure 2) suggests a function during neurite development. Second, a reduction in neurite outgrowth in RA induced SH-SY5Y cells with reduced VCX protein levels (Figure 4B) underscores the significance of VCX proteins in this process. Lastly, overexpression of VCX-A in primary rat hippocampal neurons, that normally do not express the VCX genes, results in an increase in neurite number (Figure 3) also demonstrating a direct correlation between the VCX proteins and neuronal processes.
The stimulation of neuronal processes by VCX implies that VCX either positively or negatively regulates mRNAs encoding proteins involved in neurite number and complexity. Considering the ability of VCX-A to stabilize mRNA (Figure 5C and (Jiao et al., 2006)) and silence mRNA translation (Figure 5 and Figure 6), either or both of these properties could contribute to neuritogenesis. At present, it is not clear whether the VCX-A-mediated regulation of neurite outgrowth involves one specific or multiple substrate mRNAs and a global analyses of VCX-A target substrates will be required to directly address this question. However, the preferential binding by VCX-A and the potential silencing of the mRNA encoding CaMKI (Figure 6), a positive regulator of dendrite outgrowth (Kamata et al., 2007) and its colocailzation with staufen1-containing RNA granules, suggests VCX-A could be involved in the transport of a subset of translationally silenced mRNAs that promote neuronal projections upon induction. Several RNA-binding proteins have been shown to regulate the translation of specific mRNAs that contribute to neuronal plasticity (Wu et al., 1998; Huber et al., 2002; Si et al., 2003; Alarcon JM, 2004; Huang CS, 2005; Koekkoek et al., 2005). Aberrant translation of mRNAs in neuronal cells can contribute to disease states, as is the case with Fragile X syndrome. This disorder is caused by the absence of, or a mutation in, the FMRP RNA binding protein, which leads to mis-regulation of translation at synapses (Zalfa et al., 2006). Based on the modest expression level of the VCX proteins in brain, it is possible that this family of proteins function in transcript-specific regulation of mRNA that can contribute to neuritogenesis.
The ubiquitous presence of VCX in all tissues tested indicates that VCX also contributes to an mRNA stability and translational regulatory role in non-neuronal cells. Identification of mRNAs specifically bound by VCX in neuronal and non-neuronal cells will begin to address this model. An interesting question remains regarding regulation of the VCX association at the 5´ end of an mRNA. One possible regulatory role could be analogous to the Y-box protein, YB-1. YB-1, like VCX, is a noncanonical cap binding protein that preferentially binds the 5´ end of an mRNA to stabilize and silence an mRNA (Evdokimova et al., 2001). Its association with the 5´ cap is modulated by Akt-mediated phosphorylation, which reduces the affinity of YB-1 for the 5´ cap, resulting in the release of translational suppression (Evdokimova et al., 2006). Whether a similar mechanism is involved in the regulation of VCX binding to the 5´ cap remains to be determined.
An interesting observation from our studies is, translation of some, but not all, mRNAs tested were affected by VCX overexpression. It appears unlikely that all the affected RNAs have a specific primary sequence that VCX binds to. A more likely possibility is that VCX preferentially binds to a higher order structure at the mRNA 5´-end analogous to the Dcp2 decapping enzyme, a protein that similar to VCX, preferentially binds the 5´ end of mRNAs (Li et al., 2008). Collectively, our studies demonstrate that VCX-A, and most likely all members of the VCX/Y family of proteins, are modulators of mRNA stability and translation. They can preferentially bind the 5´ cap of an mRNA and potentially sequester the mRNA into a silenced mRNP complex to prevent its expression into protein. A decrease in VCX protein expression in neurons results in the reduction of neurite numbers but not length, suggesting that VCX is essential for the number of neuronal projections rather than their extension, by selectively modulating the expression of mRNAs involved in neurite outgrowth. Our data indicates that the origins of at least part of the implicated mental retardation that accompanies VCX-A deficiency could be a consequence of aberrantly translation of a subset of mRNAs. Studies are under way to determine what mRNAs are specifically regulated by VCX and how this regulation contributes to neuronal differentiation.
We thank Terri G. Kinzy and Anthony M. Esposito for assistance with the polysome gradient experiments. These studies were supported by NIH grants AG024494 and NS20591 to K.H., NSF grant IBN-0548543, March of Dimes 1-FY08-464 and a NARSAD Toulmin Independent Investigator Award to B.L.F., and NIH grant GM67005 to M.K.