Lysophosphatidic acid differentially regulates axonal mRNA translation through 5′ UTR elements

Axonal calreticulin protein levels increase with LPA treatment

We have previously shown that inhibition of the smooth endoplasmic reticulum Ca-ATPase (SERCA) with thapsigargin increases synthesis of calreticulin protein (Wang et al., 2007). Here, we sought to determine the mechanism for this potential translational regulation of axonal calreticulin mRNA, by linking extracellular events with potential axonal mRNA translation. We used LPA as a more physiological stimulus since this has been shown to induce growth cone collapse in DRG neurons and cause release of ER Ca²⁺ through the activation of PLC (Elmes et al., 2004).

To determine if LPA can similarly increase axonal levels of calreticulin protein, we initially used immunostaining to detect axonal calreticulin in axons of cultured DRG neurons. Immunostaining for calreticulin showed an increase in signals in distal axons after 4 h treatment with LPA (Figs. 1A,C and Supplemental Figure S1). In our previous studies, the calreticulin antibody used here proved highly specific (Wang et al., 2007; Willis et al., 2005). Quantifying these signals showed a 2-fold increase in the axonal calreticulin immunoreactivity with the LPA treatment (Fig. 1C). Cell body immunoreactivity for calreticulin also increased over this period but did not reach statistical significance in our studies (Figs. 1B,C). Schwann cells also showed a clear increase in calreticulin protein immunoreactivity with LPA exposure (Fig. 1C). The LPA induced increase in axonal calreticulin immunoreactivity was significantly attenuated by anisomycin (Figs. 1A,C and Supplemental Figure S1), arguing that this is a protein synthesis-dependent event. However, we continued to see an increase in axonal calreticulin immunoreactivity upon LPA treatment in cultures where microtubule-dependent transport was blocked by colchicine (Figs. 1A,C). Together these data suggest that the LPA can trigger an increase in axonal calreticulin levels through protein synthesis-dependent mechanisms.

LPA stimulates translation of calreticulin mRNA in axons

To more directly test if LPA treatment can regulate translation of axonal mRNAs, we asked if axonal calreticulin mRNA shows altered association with polysomes after LPA treatment. As a more global assessment of axonal mRNA translation, we also analyzed distribution of axonal mRNAs encoding the ER chaperone protein Grp78/BiP, the cytoplasmic chaperone protein Hsp70, the mitochondrial chaperone protein Hsp60, and the growth-associated protein GAP-43. For this, 7 d injury-conditioned adult rat DRG neurons were cultured on a porous membrane to allow separation of axonal processes from cell body and non-neuronal cells (Willis et al., 2005). After LPA treatment for 4 h, the axonal compartment was lysed under conditions that maintain mRNA-ribosome association and then processed for fractionation of polysomal and sub-polysomal RNAs. The purity of axonal preparations was verified by the absence of cell body confined transcripts (γ-actin and MAP2) and glial cell-specific mRNAs (GFAP and ErbB3) (Fig. 2A).

RT-qPCR was then used to determine the absolute levels of the individual mRNAs in polysomal vs. sub-polysomal fractions. In untreated cultures, axonal calreticulin mRNA was seen in approximately equal distribution in the polysomal and subpolysomal fractions. However, axonal calreticulin mRNA fractionated predominantly with polysomes after LPA treatment (Fig. 2C). The amount of RNA extracted from these axonal microultracentrifugation fractions was extremely limited, thus we turned to cell body isolates to test the efficacy of polysomal RNA fractionation. We have previously shown that treating DRG lysate preparations with 50 mM EDTA prior to sucrose gradient fractionation effectively dissociates ribosome subunits (Zheng et al., 2001). Adding 50 mM EDTA to DRG cell body lysates prior to fractionation on sucrose gradients shifted nearly all of the polysomal calreticulin mRNA to the sub-polysomal fraction indicating that we had successfully fractionated polysomal mRNAs with these analyses (Fig. 2B). Interestingly, most of the cell body calreticulin mRNA associates with polysomes under basal conditions with minimal response to LPA (Fig. 2B). This suggests different translational regulations for the cell body and axonally localized calreticulin mRNA pools. Similar to calreticulin mRNA, axonal grp78/BiP mRNA shifted to predominantly polysomal with LPA treatment (Fig. 2C). However, axonal GAP-43, Hsp60, and Hsp70 mRNAs showed no significant redistribution between polysomal and subpolysomal fractions with LPA treatment (Fig. 2C). These data suggest that the response to LPA translationally activates axonal mRNAs encoding ER chaperone proteins.

LPA has been shown to increase cytoplasmic Ca²⁺ levels in DRG neurons through release of ER Ca²⁺ stores (Elmes et al., 2004). Release of ER Ca²⁺ is also seen during ER stress, a process known to induce transcription of mRNAs encoding ER chaperone proteins (Harding et al., 2002). A generalized increase in overall levels of calreticulin mRNA with LPA treatment could explain the increased translation seen above, particularly if the newly transcribed calreticulin mRNA selectively trafficked to the axonal compartment. This seemed unlikely since our cultures included the RNA polymerase II inhibitor DRB. Using quantitative PCR methods, we saw no increase in calreticulin or grp78/BiP mRNAs in the cell body following the same LPA treatment even when DRB was excluded from the medium (Supplementary Table 2). However, mRNA levels of calreticulin and grp78/BiP showed a significant increase after treatment with SERCA inhibitor thapsigargin (Supplementary Table 2), which triggers ER stress through depletion of ER Ca²⁺ (Llewellyn et al., 1996). This did not exclude the possibility of a shift in localization of these mRNAs and, we saw that transport of calreticulin mRNA into these axons is increased by neurotrophin 3 (NT3) and myelin-associated glycoprotein (MAG) in a prior work (Vuppalanchi et al., 2010). To determine if LPA can alter transport of calreticulin mRNA from the cell body, we analyzed axonal mRNA levels after LPA treatment. There was no increase in calreticulin or Grp78/BiP mRNA transport following LPA treatment; indeed, the axonal levels of these mRNAs decreased slightly in response to LPA, although this did not reach statistical significance (Supplementary Table 3). Taken together, these data argue that the shift of axonal mRNAs encoding ER chaperone proteins into the polysomal fraction upon LPA treatment reflects an increase in translational efficiency of these axonal mRNAs.

LPA alters the activity of translation initiation factor eIF2α in axons

The above studies point to LPA as a stimulus for activating the translation of specific axonal mRNAs. As noted above, LPA is known to trigger release of ER Ca²⁺ stores. Decrease in ER Ca²⁺ perturbs protein folding and processing in the ER and results in a general translational attenuation in part through phosphorylation of eIF2α on serine 51 (Brostrom and Brostrom, 2003). This phosphorylation converts eIF2α into an inhibitor of the guanine exchange factor activity of eIF2B causing a decrease in global translation (Hinnebusch and Natarajan, 2002). Thus, we asked if LPA treatment causes S51 phosphorylation of eIF2α in cultured DRGs. Immunoblotting lysates from DRGs exposed to LPA for 30 min showed a clear increase in eIF2^PS51 levels compared to vehicle treated cultures (Fig. 3A). To determine if the axonal eIF2α is phosphorylated upon LPA treatment, we determined the eIF2α^PS51 immunoreactivity along axons of injury-conditioned DRG cultures that were exposed to LPA immobilized onto polystyrene microparticles. The axonal eIF2α^PS51 immunoreactivity was significantly increased adjacent to LPA microparticles compared to axons exposed to the control BSA-coated microparticles (Fig. 3B). These data indicate that LPA can induce an inactivating phosphorylation of axonal eIF2α.

Although phosphorylation of eIF2α impairs general translation, in non-neuronal cells it is associated with a paradoxical increase in translation of a subset of mRNAs including ER chaperone protein mRNAs (Brostrom and Brostrom, 1998; Brostrom et al., 1996; Kaufman et al., 1989). Since LPA increased the translation of axonal calreticulin mRNA and increased phosphorylation of eIF2α in axons, we asked if expression of phosphomimetic and non-phosphorylatable forms of eIF2α could alter calreticulin protein levels in axons. For this, we transduced DRG cultures with adenovirus expressing eIF2α with serine 51 to aspartate (AV-eIF2α^S51D-GFP) vs. serine 51 to alanine mutations (AV-eIF2α^S51A-GFP) (Lee and Hollenbeck, 2003). Naïve DRG cultures were used in this case to give time for expression of the transduced eIF2α. Over 2–3 DIV, axonal growth from naïve neurons shifts from arborizing to the elongating growth typical of injury-conditioned neurons (Smith and Skene, 1997).

DRGs expressing phosphomimetic eIF2α^S51D, which functions as a constitutively phosphorylated eIF2α, showed significantly higher axonal calreticulin immunoreactivity than those expressing the non-phosphorylatable eIF2α^S51A, which functions as a constitutively active eIF2α (Figs. 4A,C and Supplemental Figure S2). Neurons transduced with AV-GFP showed no changes in axonal or cell body calreticulin immunoreactivity (Fig. 4C) suggesting that the response to AV-eIF2α^S51D-GFP is due to expression of the eIF2α^S51D rather than a response to AV or GFP. Schwann cells and neuronal cell bodies also showed a robust increase in calreticulin immunoreactivity with eIF2α^S51D vs. eIF2α^S51A (Figs. 4B and D). These data are consistent with translation of calreticulin mRNA which have been increased when eIF2α’s activity is inhibited by phosphorylation similar to what has been demonstrated in non-neuronal cells (Llewellyn et al., 1995).

5′UTR of calreticulin contains translational regulatory elements

The 5′UTRs of several mRNAs have been shown to confer translational regulation through secondary structure formation, alternate initiation sites, and direct protein interactions with RNA targets (Spriggs et al., 2008). Since we had previously shown that the 3′ UTR of calreticulin drives transport of the mRNA into axons (Vuppalanchi et al., 2010), we asked if calreticulin’s 5′UTR can confer LPA-dependent translational of calreticulin mRNA in axons. For this, we targeted a diffusion limited reporter, GFP^myr, into axons using the axonally localizing 3′UTR of calreticulin mRNA (Vuppalanchi et al., 2010). The 5′UTR of the original reporter was replaced with the 5′UTRs of calreticulin or β-actin mRNAs (5′Cal-GFP^myr-3′Cal and 5′β-actin-GFP^myr-3′Cal, respectively). After 2 DIV, naïve cultures of DRG neurons expressing 5′Cal-GFP^myr-3′Cal or 5′β-actin-GFP^myr-3′ Cal showed abundant fluorescent signals in the distal axons (Fig. 5) similar to what we have seen for other axonally localizing reporters (Vuppalanchi et al., 2010; Willis et al., 2007; Yudin et al., 2008).

With bath application of LPA, the growth cones of 5′Cal-GFP^myr-3′ Cal expressing DRG neurons showed a significant increase in GFP^myr signals compared to BSA treatment (Fig. 5A). This LPA-dependent increase in axonal 5′Cal-GFP^myr-3′Cal was protein synthesis dependent (Figs. 5A,B). Additionally, this increase in the GFP signals with LPA treatment was not attenuated by pre-treatment with colchicine, arguing that the elevated GFP levels were from axonal protein synthesis rather than transport of reporter mRNA or protein from the cell body via microtubule-dependent mechanisms (Fig. 5B). In contrast, the axons of DRG neurons expressing 5′β-actin-GFP^myr-3′ Cal did not show any change in the GFP^myr signals in LPA treated compared to BSA treated neurons (Fig. 5C). These data suggest that calreticulin’s 5′UTR imparts LPA-dependent translation of the axon-ally localized mRNA.

Phosphorylated eIF2α regulates translation of calreticulin mRNA through its 5′UTR

Since LPA increased axonal eIF2α^PS51 and introduction of the phosphomimetic eIF2α^S51D increased levels of endogenous calreticulin protein in the axons, we asked if phosphorylation of eIF2α is needed for LPA-dependent translation of the 5′Cal-GFP^myr-3′Cal reporter mRNA in DRG axons. For this, we transduced DRG cultures with AV-eIF2α^S51D and used live cell imaging to monitor axonal reporter mRNA translation in 5′Cal-mCherry^myr-3′Cal transfected DRGs. Considering that the eIF2α mutant could alter basal rates of translation, we used FRAP to test for LPA-dependent translation. The DRG neurons that were expressing eIF2α^S51D-GFP showed significant recovery of axonal mCherry signals after photobleaching in both LPA treated and control cultures (Figs. 6A,B). Pre-treatment with anisomycin attenuated the recovery of mCherry^myr fluorescence showing translation dependence. In contrast, 5′β-actin-mCherry^myr-3′Cal showed no significant recovery after photobleaching in neurons expressing eIF2α^S51D even in the absence of LPA (Figs. 6A,C). Since there were no statistical differences between basal and LPA induced fluorescence recoveries in the 5′Cal-mCherry^myr-3′Cal expressing neurons that were transduced with eIF2α^S51D, these data further support the conclusion that LPA-induced axonal translation through 5′UTR of calreticulin requires phosphorylation of eIF2α. Moreover, recovery of fluorescence in axons for the 5′β-actin-mCherry^myr-3′Cal constructs was the same between LPA and BSA treated cultures when cultures were transduced with AV encoding the non-phosphorylatable eIF2α^S51A (Fig. 6E). Introducing eIF2α^S51A did not affect the LPA induced recovery of 5′ Cal-mCherry^myr-3′Cal (Fig. 6D); however, this is likely due to LPA induced phosphorylation of endogenous eIF2α. Taken together, these data show that axonal translation through 5′UTR of calreticulin occurs at a higher rate when eIF2α is phosphorylated, an event that attenuates translation through the 5′UTR of β-actin.

Cell culture

Neurons were cultured from L4–6 dorsal root ganglia (DRGs) of 150–225 g male Sprague Dawley rats. Dissociated DRG cultures were prepared as previously described (Willis et al., 2005). DRG neurons were plated on substrates coated with poly-L-lysine (Sigma) and laminin (Millipore) and maintained in growth medium. For imaging, neurons were cultured on coated-glass coverslips. For isolation of DRG axons from cell bodies and non-neuronal cells, injury-conditioned dissociated DRGs were plated onto coated polyethylene terephthalate (PET) membrane with 8 µm diameter pores (BD Falcon) (Willis et al., 2005). Naïve DRG cultures were used for expression of transfected or transduced mRNAs and proteins (see below); these cultures were analyzed after 2–3 DIV to allow for expression of the exogenously encoded mRNAs and proteins. Cultures of 7 d injury-conditioned neurons were used to assess endogenous mRNAs and proteins when no transfections or transductions were performed; these cultures were evaluated after ≤1 day in vitro (DIV). For injury-conditioning, sciatic nerve was crushed at mid thigh level 7 d prior to culture using a Dumont 0 forceps (Smith and Skene, 1997).

Axons were isolated from neurons cultured on PET membrane after 16 – 20h in culture by carefully scraping the top membrane surface using a cotton-tipped applicator to remove all cell bodies and non-neuronal cells. The purity of the axonal preparations was tested by reverse transcription coupled polymerase chain reaction (RT-PCR) for microtubule-associated protein 2 (MAP2), γ-actin, β-actin, glial fibrillary acidic protein (GFAP), and erb-b2 erythroblastic leukemia viral oncogene homolog 3 (ErbB3) mRNAs.

For transfection of sensory neurons, dissociated DRGs were centrifuged at 100×g for 5 min and gently resuspended in ‘nucleofector solution’ (Rat Neuron Nucleofector kit; Lonza, Inc.). 5 µg of each plasmid DNA was electroporated using AMAXA Nucleofector apparatus. After electroporation, the cells were resuspended in culture medium and plated; medium was then replaced 4h later to remove any non-adherent cells and debris.

For transduction with adenovirus (AV), dissociated DRGs were allowed to attach to the culture plate for 4h and then 100 MOI AV-eIF2α^S51D-GFP or AV-eIF2α^S51A-GFP was added to the media. The AV used in these studies encoded an in frame GFP fused to the C-terminus of eIF2α^S51D or eIF2α^S51A (Lee and Hollenbeck, 2003). After 16 h, the culture media were replaced, and the cells were allowed to grow for 48 h.

Expression vectors

The 5′UTR of rat calreticulin was cloned using overlapping oligonucleotides that spanned the entire 5′UTR with 5′ Nhe1 and 3′ BamH1 sites added for subcloning. Oligonucleotides (2.5 µM each) were annealed and extended using HotTaq Master Mix (Qiagen); PCR conditions were 94 °C for 5 min, 94°C×30s, 56°C×30s, 68 °C×30 s for 3 cycles with a final extension of 68 °C×5 min. The 5′UTR of rat β-actin was amplified by standard reverse transcriptase coupled PCR (RT-PCR) with primers to add 5′ AflII and 3′ BamHI restriction sites. The 5′UTR calreticulin and 5′UTR β-actin PCR products were restriction digested and cloned up-stream of GFP^myr-Cal (5′Cal-GFP^myr-3′Cal or 5′β-actin-GFP^myr-3′Cal). Similar to the GFP^myr-Cal constructs, the 5′UTRs of rat calreticulin or β-actin were cloned up-stream of mCherry^myr that included the 3′UTR of calreticulin down-stream of the mCherry^myr coding sequence (5′Cal-mCherry^myr-3′Cal or 5′β-actin-mCherry^myr-3′Cal). All primer sequences for cloning and RNA level analyses (see below) are included in Supplementary Table 1.

Polysome RNA fractionation

Polysomal and subpolysomal RNAs were fractionated on 20% discontinuous sucrose gradients as previously described with minor modifications (Rajasekaran et al., 2004). Briefly, cultured cells were rinsed with cold PBS containing 0.1 mM cycloheximide and then lysed under conditions that maintain mRNA-ribosome association using ‘polysome lysis buffer’ (300 mM KCl, 2 mM MgCl₂, 20 mM Tris HCl [pH 7.4], 2 mM DTT, 0.05% deoxycholate, 0.1 mM cycloheximide, 50 U/ml RNase-In). Insoluble components were cleared from the lysate by centrifugation at 13,000 ×g for 15 min at 4 °C. An aliquot of the lysate was used to test for purity of the axonal isolate (see below). The cleared lysate was layered onto 1 ml 20% sucrose in polysome lysis buffer. Gradients were then fractionated by ultracentrifugation at 105,000 ×g, 4 °C, in a RP55S swinging bucket rotor (Sorvall) for 2h using slow acceleration and deceleration to prevent disruption of the gradient. To test for isolation of polysomal mRNAs, 50 mM EDTA was added to sister lysates to dissociate ribosome sub-units before centrifugation. The upper 80% the gradient was collected as the subpolysomal fraction and the lower 20% was collected as the polysomal fraction. RNA was extracted from each fraction using TRI-ZOL (Ambion) reagent and then precipitated with isopropanol with glycogen-blue (Ambion) as a carrier. The RNA was processed for quantitative RT-PCR (RTqPCR, see below).

RT-PCR was used to test for purity of the axonal isolate. For this, RNA was isolated using the RNAqueous Micro kit (Ambion) and quantified by fluorimetry using Ribogreen reagent (Invitrogen). After normalization, RNA was processed for standard RT-PCR using primers for β-actin, γ-actin, MAP2, GFAP, and ErbB2 (Supplemental Table 1). Reactions were analyzed by standard ethidium bromide stained agarose gels.

Axonal mRNA transport analyses

DRGs were cultured on a porous membrane for analyses of axonal mRNA transport as previously described (Willis et al., 2007). To restrict analyses to existing mRNAs, the RNA synthesis inhibitor 5,6-dichlorobenzimidazole riboside (DRB) (80 µM, Sigma) was included in the culture medium. After 16–20h in culture, the neurons were treated with 30 µM LPA (BioMol, reconstituted in 0.1% BSA) for 4h, after which the membranes were rinsed in PBS and the axonal compartment was isolated as described above. RNA was extracted from the fractionated cultures using the RNAqueous Micro kit and quantified by fluorimetry with the Ribogreen reagent (Molecular Probes). To normalize the axon mass between samples, RNA samples were normalized to axonal protein content before RT-PCR analyses. Protein content was measured in the flow-through from the affinity-based RNA isolation using fluorimetry with the NanoOrange reagent (Molecular Probes). For RT-PCR, ∼100 ng of RNA from axonal compartment was used as a template and the RT reaction was performed using the iScript cDNA synthesis kit (Biorad). After purity of axonal isolates was verified by standard RT-PCR (see above), samples were processed for RTqPCR to evaluate axonal mRNA levels (see below).

RNA quantification by RTqPCR

qPCR was performed with the Prism 7900HT sequence detection system (ABI) with 2x SybrGreen Master Mix (Qiagen) with standard cycling parameters. All samples were assayed in quadruplicate over at least three independent experiments. The following primers were used to amplify from reverse transcribed total RNA: Calreticulin, Grp78/BiP, Hsp70, Hsp60, β-actin, GAP-43, and 12S mitochondrial rRNA (see Supplementary Table 1).

Absolute quantification of PCR products was used to analyze mRNA levels from the polysome fractions. For this, RNA concentration was calculated from standard curves for each specific PCR product generated using serial dilutions of cDNA subcloned into pCR2.1TOPO vector for each mRNA. RNA isolated from non-fractionated DRGs was used for standard curves. All standard curves demonstrated a strong linear relationship with a correlation coefficient ≥0.98 between the fractional cycle number and the log of the starting copy number.

Relative quantification of PCR products was used for analyzing potential changes in axonal RNA levels with stimulation (Willis et al., 2005). For this, the levels of each test mRNA were normalized to the 12S mitochondrial rRNA using the comparative threshold method (C_t) method to provide an internal control for RT efficiency. From this, the calculation of ΔΔC_t was determined by subtracting the ΔC_t calibrator value (BSA for LPA and DMSO for thapsigargin). The fold difference was then expressed as 2^−ΔΔCt.

Immunofluorescence

DRG cultures were rinsed in phosphate-buffered saline (PBS), fixed in 4% paraformaldehyde for 20 min, and then permeabilized in PBS plus 0.3% Triton X-100 for 5 min. After permeabilization, samples were incubated in blocking buffer (PBS plus 5% donkey serum) for 1 h. The samples were then incubated overnight at 4 °C in the following primary antibodies in blocking buffer: rabbit anti-calreticulin (1:200; Stressgen), chicken anti-neurofilament (NF) H (1:1000; Millipore), rabbit eIF2α^PS51 (1:100; Cell Signaling). Cultures were rinsed several times in PBS and incubated for 1h in the following secondary antibodies in blocking buffer: FITC-conjugated anti-chicken (1:400; Jackson ImmunoRes.) and Texas-red conjugated anti-rabbit or anti-goat (1:400; Jackson ImmunoRes.). Samples were rinsed in PBS and then mounted with Prolong Gold Antifade reagent (Invitrogen).

Epifluorescent microscopy with Leica DMRXA2 was used to image cultured neurons. The exposure times, gain and offset were kept constant between samples such that matched images were obtained. Quantitative analysis for protein levels within growth cones was analyzed using fluorescence (pixel) intensities of at least 15 growth cones for each condition over at least three independent experiments. The outlines of the growth cones were traced from DIC images and fluorescence intensity within the growth cone was calculated using ImageJ. Accumulation of eIF2α^PS51 directly at sites of ligand stimulation was tested by immunofluorescence of dissociated DRG cultures exposed overnight to LPA- or control-microparticles (4.5 µm diameter; Polysciences) (Willis et al., 2007). For quantifying signals, the pixel intensity was measured in seven 5 µm length bins along the axon shaft proximally and distally to the center of the microparticle. Background was subtracted from the intensity values, and intensity in each bin was normalized to the mean intensity over the entire 35 µm binned region. Mean normalized intensity of each bin was calculated for ≥25 axons over at least three separate culture experiments.

Immunoblotting

For detection of eIF2α in cell lysates, cultures were lysed in RIPA buffer at 4 °C for 20 min. Lysates were cleared by centrifugation at 13,000×g, 4 °C; cleared lysates were normalized for protein concentration by Bradford assay and then processed for SDS/PAGE. Fractionated proteins were transferred to PVDF membranes, blocked for non-specific binding using 5% non-fat dry milk in Tris-buffered saline (TBS) and then probed with anti-eIF2α (1:1000) or anti-eIF2α^PS51 (1:1000) overnight at 4 °C in blocking buffer. After washing in TBS, blots were incubated in HRP-conjugated anti-rabbit secondary antibodies (1:3000; Cell Signaling) and developed with ECL^Plus (Amersham).

Time-lapse imaging

Naïve DRG cultures transfected with the GFP^myr constructs were used for time-lapse imaging. To test for translation vs. RNA transport dependent recovery, cultures were pre-treated for 30 min prior to time lapse imaging sequences with 50 µM anisomycin or 200 nM col-chicine, respectively. Imaging was performed at 37 °C using a TCS/SP2 confocal microscope with a 40× oil immersion objective. The 488 nm laser line (10% power) was used to excite GFP^myr and the emission was collected at 498–530 nm. Images were collected every 120 s using the LCS confocal software package (Leica). The axon shaft including the growth cones in the distal 50 µm of the axon was traced and the average pixels/µm² within the area was calculated using ImageJ.

Fluorescence recovery after photobleaching

Fluorescence recovery after photobleaching (FRAP) was used to test for localized translation of mCherry^myr mRNAs. Neurons were imaged every 30s for 2 min with 488 nm and 543 nm laser set at 10% power to establish a baseline fluorescence intensity. For photo-bleaching, a region of interest (ROI) comprising the distal 30–45 µm of the axon including growth cone was exposed to 75% power of 488 and 543 laser for 20 frames at 1.6s intervals. GFP and mCherry levels were monitored every 60 s over 30 min by excitation with 10% power of 488 and 543 nm with emissions collected at 498–530 nm and 595–640 nm, respectively. To test for translation dependent recovery, cultures were pre-treated with 50 µM anisomycin for 30 min prior to photobleaching.

ImageJ was used to calculate average pixels/µm² in the ROIs of the raw confocal images. The residual fluorescence signal at t=0 min (post-bleach) was subtracted from all time point intensity values. The corrected values were then normalized to the pre-bleach (set to 100%) and graphed as mean±standard error of the mean.

Statistical analyses

For immunofluorescence and polysomal fractionation studies Student’s Newman–Keul test was used to compare protein and RNA levels in the distal axons and cell bodies; p≤0.05 was considered significant. One way ANOVA was used to determine significance of fluorescence in live cell imaging sequences with Bonferroni post-hoc test; p≤0.01 was considered significant.