Extracellular stimuli can regulate axonal levels and translation of Prenyl-Cdc42 and RhoA mRNAs

We previously found that localized translation of Prenyl-Cdc42 mRNA promotes neurite growth in cultured neurons [4]. Axon growth can be positively or negatively impacted by external stimuli, so we asked if growth-promoting neurotrophins or a growth-inhibiting CSPG might alter axonal transport of Prenyl-Cdc42 mRNA by treating mouse DRG cultures with a neurotrophin cocktail, consisting of NT3, BDNF, and NGF to stimulate all 3 Trk receptors on DRG neuron subpopulations, or aggrecan and assessed axonal Prenyl-Cdc42 mRNA levels by single molecule fluorescent in situ hybridization (smFISH). Since DRG neurons can be cultured from adult mice, this model provides a view of axons regrowing from a mature neuron where we previously showed Prenyl-Cdc42 mRNA localizes. Neurotrophin treatment increased and aggrecan decreased axonal Prenyl-Cdc42 mRNA levels (Fig 1A and 1B). In contrast to CDC42’s promotion of actin filament polymerization, RHOA activation causes actin filament depolymerization and the protein’s activity is increased by growth-inhibiting stimuli [2]. RhoA mRNA also localizes into axons, and this was shown to increase upon exposure to CSPGs [15]; consistent with this, we see that axonal RhoA mRNA levels increase in response to aggrecan but neurotrophin stimulation had no effect on axonal RhoA mRNA levels (Fig 1A and 1C). Together, these data point to reciprocal regulation of axonal RhoA and Prenyl-Cdc42 mRNAs in response to CPSG stimulation but selective increase in axonal Prenyl-Cdc42 mRNA in response to neurotrophins.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Growth-promoting and inhibiting stimuli regulate axonal levels of Prenyl-Cdc42 and RhoA mRNAs.

A) Representative exposure-matched smFISH and IF images for Prenyl-Cdc42 or RhoA mRNA plus neurofilament (NF) protein in adult DRG neuron cultures treated with either 10 ng/ml each NT3, BDNF, plus NGF each (‘neurotrophins’) or 50 ng/ml aggrecan [Scale bar = 10 µm]. B-C) Quantification of axonal Cdc42 (B) or RhoA (C) mRNA smFISH signal intensities shown as mean pixel intensity above background for axons ± SEM (N ≥ 40 neurons across three independent cultures; ** P < 0.01, **** P < 0.001 and NS = not significant for indicated data sets, †††† P < 0.001 vs. control and neurotrophins, and #### P < 0.001 vs. aggrecan by Kruskal-Wallis ANOVA with Dunn post-hoc tests for pair-wise comparisons). D) Representative exposure-matched IF images for Prenyl-CDC42 or RHOA proteins and NF in distal axons of adult DRG neuron cultures treated with either 10 ng/ml neurotrophins or 50 ng/ml aggrecan; see S1A Fig for representative no primary IF images [Scale bar = 10 µm]. E-F) Quantitation of axonal CDC42 (E) or RHOA (F) IF signal intensities shown as mean pixel intensity above background ± SEM; see S1B and S1C Fig for cell body levels of CDC42 and RHOA under these conditions (N ≥ 20 neurons in three independent cultures; * P < 0.05, ** P < 0.01, *** P < 0.005, **** P < 0.001, and NS = not significant between indicated pairs and #### P < 0.001 vs. all other groups by Kruskal-Wallis ANOVA with Dunn post-hoc tests for pair-wise comparisons).

https://doi.org/10.1371/journal.pgen.1011916.g001

To determine if the endogenous Prenyl-CDC42 and RHOA proteins might show similar changes in their axonal levels with these stimuli, we performed immunofluorescence analyses on DRG cultures that had been treated with neurotrophins or aggrecan. Exposure to neurotrophins increased axonal CDC42 protein levels but had no apparent effect on axonal RHOA levels (Fig 1D and 1F). Aggrecan treatment increased RHOA protein levels in the axons but had no apparent effect on axonal CDC42 protein levels (Fig 1D-E). CDC42 protein levels did not change in the neuronal cell bodies or soma with these stimuli (S1A Fig). Soma RHOA protein levels increased with both aggrecan and neurotrophin treatments (S1B Fig). It should be noted that the CDC42 antibody utilized here does not distinguish between Prenyl-CDC42 and Palm-CDC42 proteins. We previously showed that Palm-CDC42 protein is transported into axons [4], so these immunofluorescence data do not distinguish between the two CDC42 isoforms.

To more selectively assess axonal translation of Prenyl-Cdc42 and RhoA mRNAs in response to these stimuli, we visualized axonal signals of diffusion limited GFP^MYR and mCherry^MYR reporters with the 5’ and 3’UTRs of rat Prenyl-Cdc42 and RhoA mRNAs, respectively, as surrogates for local translation of the endogenous axonal mRNAs (GFP^MYR5’/3’prenyl-Cdc42, mCherry^MYR5’/3’RhoA). The 5’ and 3’UTRs were included to ensure that we captured both axonal localizing and any translational control motifs; the MYR tag limits diffusion of the newly synthesized GFP and mCherry proteins in neurites such that fluorescence recovery after photobleaching (FRAP) can be used to visualize sites of reporter mRNA translation [4,16,17]. FRAP analyses in DRG neurons co-expressing GFP^MYR5’/3’prenyl-Cdc42 and mCherry^MYR5’/3’RhoA mRNAs showed fluorescent recovery within 15 min post-bleach that was attenuated by pretreatment with the protein synthesis inhibitor anisomycin, consistent with intra-axonal translation of the reporter mRNAs (Figs 2B, 2C, S2A and S2B). To determine if neurotrophin or aggrecan stimulation might affect translation of the GFP^MYR5’/3’prenyl-Cdc42 and mCherry^MYR5’/3’RhoA mRNAs, we bath applied the neurotrophin cocktail or aggrecan for 30 min before photobleaching. Neurotrophin stimulation increased GFP^MYR5’/3’prenyl-Cdc42 mRNA translation in distal axons but, consistent with the axonal mRNA analyses above, had no effect on axonal mCherry^MYR5’/3’RhoA mRNA translation (Fig 2A-C). In contrast, aggrecan treatment decreased GFP^MYR5’/3’prenyl-Cdc42 and increased mCherry^MYR5’/3’RhoA translation in terminal axons (Fig 2A-C). The recovery of axonal GFP^MYR5’/3’prenyl-Cdc42 fluorescence in the presence of aggrecan was largely indistinguishable from treatments with the protein synthesis inhibitor anisomycin (Figs 2A, 2B, S2A and S2B). Taken together, the data in Figs 1 and 2 are consistent with reciprocal regulation of axonal Prenyl-Cdc42 and RhoA mRNA levels and subsequent intra-axonal translation in response to aggrecan. In contrast, responsiveness to neurotrophins appears to be limited to the axonal Prenyl-Cdc42 mRNA.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Growth-promoting and inhibiting stimuli regulate axonal translation of Prenyl-Cdc42 and RhoA mRNAs.

A) Representative FRAP image sequences for DRG neurons co-transfected with GFP^MYR5’/3’prenyl-Cdc42 and mCherry^MYR5’/3’RhoA are shown (72 h post-transfection). Cultures were treated with either 10 ng/ml neurotrophins or 50 ng/ml aggrecan as in Fig 1. Boxed regions represent the photobleached ROIs; see S2A and S2B Fig for control and anisomycin-treated representative FRAP image sequences [Scale bar = 20 µm]. B-C) Quantitation of FRAP sequences from panel A are shown as average normalized% recovery ± SEM. Note that translation inhibition with anisomycin prior to photobleaching shows that the GFP^MYR5’/3’prenyl-Cdc42 and mCherry^MYR5’/3’RhoA recovery requires protein synthesis (N ≥ 10 neurons over three independent experiments; * P < 0.05, ** P < 0.01, ***P < 0.005, ****P < 0.001 by two-way repeated measures ANOVA with Tukey post-hoc tests for pair-wise comparisons).

https://doi.org/10.1371/journal.pgen.1011916.g002

As noted above, Prenyl-Cdc42 mRNA localizes into axons of sensory neurons and cortical neurons and the 3’UTR encoded by CDC42 exon 7 is necessary and sufficient for the mRNA’s axonal localization [4]. We have previously shown that conservation of 3’UTR regions can have predictive value for identifying functional domains in mRNAs [18,19]. The initial 150 nt of rat Prenyl-Cdc42 3’UTR (nt 764–913; NCBI XM_008764286.3) shows > 85% sequence identity with the 3’UTRs of 26 other vertebrate species (S3 Fig). To determine if this conserved 150 nt region imparts any function for the mRNA, we generated fluorescent reporter constructs containing Prenyl-Cdc42’s nt 764–913 or 914–2164 (i.e., the remainder of the 3’UTR) as the 3’UTR for GFP^MYR cDNA (GFP^MYR3’prenyl-Cdc42^764-913 and GFP^MYR3’prenyl-Cdc42^914-2164, respectively; Fig 3A). smFISH analyses of transfected adult mouse DRG cultures showed that GFP^MYR mRNA only localized into axons of the GFP^MYR3’prenyl-Cdc42^764-913 transfected neurons; GFP^MYR3’prenyl-Cdc42^914-2164 transfected neurons did not show axonal GFP mRNA signal above the scrambled control probe (Figs 3B, 3C, and S4A). Cell body levels of GFP mRNA were not appreciably different between GFP^MYR3’prenyl-Cdc42^764-913 and GFP^MYR3’prenyl-Cdc42^914-2164 transfected neurons (Figs 3B and S4B). These data show that the conserved proximal 150 nt region of Prenyl-Cdc42’s 3’UTR is necessary and sufficient to drive axonal mRNA localization.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. An evolutionarily conserved region of the Prenyl-Cdc42 mRNA 3’ UTR drives its axonal localization.

A) Schematic of the regions of the rat prenyl-Cdc42 3′UTR mRNA that were tested for axonal localizing activity. The yellow-shaded regions show ≥85% sequence identity between available mammalian prenyl-Cdc42 mRNAs (see S3 Fig for sequence alignments across species). GFP^MYR constructs used for testing 3’UTR segment are shown with GFP in green and RNA segment in black. B) Representative exposure-matched smFISH and immunofluorescence (IF) images for GFP^myr mRNA and neurofilament (NF) in adult DRG neuron cultures transfected with GFP^MYR3’prenyl-Cdc42^764-913, and GFP^MYR3’prenyl-Cdc42^914-2164. See S4A Fig for representative images of scrambled smFISH probe [Scale bar = 10 µm]. C) Quantitation of smFISH signal intensities shown as mean ± SEM pixel intensity above background for axons; see S4B Fig for cell body levels under these conditions (N ≥ 45 neurons across three independent cultures; **** P < 0.001 and NS = not significant as indicated comparisons by Kruskal-Wallis ANOVA with Dunn post-hoc tests for pair-wise comparisons). D) Representative exposure-matched smFISH and IF images for GFP mRNA plus NF in adult DRG neuron cultures transfected with GFP^MYR3’prenyl-Cdc42^764-2164, GFP^MYR3’prenyl-Cdc42^764-838, GFP^MYR3’prenyl-Cdc42^801-875, GFP^MYR3’prenyl-Cdc42^839-913, GFP^MYR3’prenyl-Cdc42^764-800 or GFP^MYR3’actg [Scale bar = 10 µm]. E) Quantitation of smFISH signal intensities shown as mean ± SEM pixel intensity above background for axons; see S4C Fig for cell body levels under these conditions (N ≥ 40 neurons across three independent cultures; * P < 0.05, ** P < 0.01, *** P < 0.005, **** P < 0.001 for indicated data sets and †††† P < 0.001 vs. all but 801-875 and 839-913 by Kruskal-Wallis ANOVA with Dunn post-hoc tests for pair-wise comparisons).

https://doi.org/10.1371/journal.pgen.1011916.g003

Prenyl-Cdc42 mRNA’s nt 801–875 contains more than 75% adenine and uridine bases (see Fig 4A), which could contain an AU-rich element (ARE). The ARE in Gap43 mRNA’s 3’UTR is necessary and sufficient for its localization into rat sensory axons and requires the ARE-binding protein HuD [20]. Thus, we asked which sequences within Prenyl-Cdc42 nt 764–913 are sufficient for axonal localization of the mRNA. For this, we generated GFP^MYR reporter constructs containing 3 overlapping 75 nt portions of the 764–913 sequence (GFP^MYR3’prenyl-Cdc42^764-838, GFP^MYR3’prenyl-Cdc42^801-875, and GFP^MYR3’prenyl-Cdc42^839-913; Fig 3A). DRGs expressing the GFP^MYR3’prenyl-Cdc42^764-838 mRNA showed axonal localization of GFP mRNA by smFISH, but axonal GFP mRNA signals in the GFP^MYR3’prenyl-Cdc42^801-875 and GFP^MYR3’prenyl-Cdc42^839-913 expressing neurons were not distinguishable from neurons transfected with GFP^MYR containing a non-localizing 3’UTR (GFP^MYR3’Actg; Fig 3D and 3E). Cell body levels of GFP mRNA were not significantly different between GFP^MYR3’prenyl-Cdc42^764-838, GFP^MYR3’prenyl-Cdc42^801-875, GFP^MYR3’prenyl-Cdc42^839-913, and GFP^MYR3’actg transfected neurons (Figs 3D and S4C).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. ARE-binding protein KHSRP binds to non-localizing conserved region of Prenyl-Cdc42 3’UTR.

A) Schematic showing % adenine-uridine percentages across the conserved Prenyl-Cdc42 mRNA nt 764-913 indicated in Fig 3A with predicted AREs for possible KHSRP binding indicated. B-C) Representative western blot for KHSRP protein (B) using KHSRP vs. control IgG immunoprecipitation from adult DRG cultures. RTddPCR analyses of RNA isolated from KHSRP vs. control IgG immunoprecipitates; co-precipitating Prenyl-Cdc42 mRNA is shown as mean mRNA copies as percentage of input ± SEM (C; N = 3 biological replicates; **** P < 0.001 by Student’s t-test for the indicated data pairs). D) Analyses of GFP mRNA from KHSRP and control IgG immunoprecipitates from DRG neuron cultures transfected with GFP^MYR3’prenyl-Cdc42^764-2164, GFP^MYR3’prenyl-Cdc42^764-800, GFP^MYR3’prenyl-Cdc42^764-838, or GFP^MYR3’prenyl-Cdc42^801-875 is shown as mean of co-precipitating mRNA copies as percentage of input ± SEM (N = 3 biological replicates; **** P ≤ 0.001 for indicated data pairs and #### P ≤ 0.001 for control IgG compared to corresponding KHSRP RIP by ordinary one-way ANOVA with Tukey post-hoc tests for pair-wise comparisons). E) Representative immunoblot analysis for KHSRP protein in RNA affinity pull down using biotinylated oligonucleotides corresponding to nt 764-838, 801-875 and 764-800 of rat Prenyl-Cdc42 mRNA. Scrambled oligonucleotide and biotin alone were used as negative controls.

https://doi.org/10.1371/journal.pgen.1011916.g004

Since nt 764–838 of Prenyl-Cdc42 mRNA contains an AU rich area over nt 801–838 (~76% AU compared to 43% for nt 764–800), we asked if the nt 764–800 has any localizing activity on its own. Thus, we generated a fluorescent reporter construct containing Prenyl-Cdc42’s 3’UTR nt 764–800 (GFP^MYR3’prenyl-Cdc42^764-800; Fig 3A). DRG neurons expressing GFP^MYR3’Prenyl-Cdc42^764-800 mRNA showed robust axonal GFP mRNA FISH comparable to GFP^MYR3’prenyl-Cdc42^764-838 mRNA expressing neurons (Figs 3E and S4C). Thus, the axonal localization motif in Prenyl-Cdc42 mRNA lies in the most proximal 37 nt of its 3’UTR (nt 764–800).

ARE-binding protein KHSRP binds to non-localizing conserved region of Prenyl-Cdc42 3’UTR

The decrease in axonal Prenyl-Cdc42 mRNA in response to aggrecan coupled with the presence of an AU rich nature of the 3’UTR raise the possibility that ARE-binding proteins might impact Prenyl-Cdc42 mRNA levels in axons (Fig 4A). KHSRP binds to ARE-containing mRNAs and promotes their decay by targeting those transcripts to the cytoplasmic exosome [21]. Thus, we tested whether KHSRP might bind to the endogenous Prenyl-Cdc42 mRNA using RNA co-immunoprecipitation (RIP). The Anti-KHSRP antibody was first validated for immunoprecipitation of KHSRP protein by immunoblotting (Fig 4B). RNA isolated from these KHSRP immunoprecipitates showed that approximately 25% of input Prenyl-Cdc42 mRNA co-precipitated with KHSRP by reverse transcriptase-coupled droplet digital PCR (RTddPCR; Fig 4C). Thus, endogenous KHSRP can bind to Prenyl-Cdc42 mRNA in PNS DRGs.

The 3’UTR of GFP^MYR3’prenyl-Cdc42^801-875 mRNA is quite AU-rich and contains predicted ARE binding sites for KHSRP and HuD at nt 818–828 and 853–861, based on previously established consensus sequences [22,23] (Fig 4A). Thus, we asked if Prenyl-Cdc42 mRNA nt 801–875 is bound by KHSRP. For this, we performed RIP analyses from DRG cultures transfected with the GFP^MYR3’prenyl-Cdc42^764-2164, GFP^MYR3’prenyl-Cdc42^764-800, GFP^MYR3’prenyl-Cdc42^764-838, or GFP^MYR3’prenyl-Cdc42^801-875 expression constructs. RTddPCR analyses of the immunoprecipitates showed that GFP^MYR mRNAs containing Prenyl-Cdc42 mRNA nt 764–2164, 764–838, and 801–875, but not 764–800, were precipitated by anti-KHSRP antibodies (Fig 4D). IgG control showed no significant precipitation of GFP^MYR mRNA for any of the DRG transfectants (Fig 4D). The levels of GFP^MYR mRNA coprecipitating with KHSRP in the GFP^MYR3’prenyl-Cdc42^764-838 and GFP^MYR3’prenyl-Cdc42^801-875 mRNA expressing DRG cultures were about half of what was seen in the GFP^MYR3’prenyl-Cdc42^764-2164 mRNA expressing cultures suggesting that KHSRP’s interaction with this 3’UTR region may be affected by sequences downstream of nt 875 (Fig 4D). Consistent with this, ARE-like sequences are present downstream of nt 864 in the Prenyl-Cdc42 mRNA 3’UTR (e.g., nt 994–1003). We used an in vitro RNA affinity pulldown technique where biotinylated RNA oligonucleotides serve as ‘bait’ to test whether endogenous proteins from sciatic nerve axoplasm isolates can bind to the RNA bait [24]. Endogenous KHSRP was clearly detected by immunoblotting in affinity pulldowns with the Prenyl-Cdc42 mRNA nt 764–838 and 801–875 oligonucleotides but not with nt 764–800 or scrambled oligonucleotide affinity pulldowns (Fig 4E). Together, these data show that axonal KHSRP protein can bind to the 3’UTR of Prenyl-Cdc42 mRNA. However, we were not able to show that axonal KHSRP is bound to endogenous Prenyl-Cdc42 mRNA using sciatic nerve axoplasm.

KHSRP deletion increases axonal Prenyl-Cdc42 mRNA

Axonal KHSRP levels increase in rodent sciatic nerve axons after crush injury via localized translation of Khsrp mRNA [13]. Since CDC42 protein activity has been shown to increase axon growth [1,4] and the injury-induced increase in axonal KHSRP levels slows regeneration of peripheral nerves [13], we asked whether KHSRP might post-transcriptionally regulate Prenyl-Cdc42 mRNA within peripheral nerve axons. For this, we compared Prenyl-Cdc42 mRNA levels in axons of Khsrp^-/- and Khsrp^+/+ mice [25] in vivo before and 7 days after a sciatic nerve crush injury. In contrast to axons of the cultured DRG neurons used above, axons in the uninjured sciatic nerve are not growing and we previously found that Prenyl-Cdc42 mRNA only seemed to localize into growing axons of the injured and not the uninjured sciatic nerve in wild type animals [4]. Prenyl-Cdc42 mRNA was not detected in the uninjured Khsrp^+/+ sciatic nerve axons but the mRNA was easily detected in the axons of uninjured Khsrp^-/- sciatic nerves (Figs 5A, 5B, S5A and S5B). 7 day crush injured sciatic nerve axons showed increased axonal Prenyl-Cdc42 mRNA levels in the Khsrp^+/+ mice as anticipated (Figs 5A, 5B and S5B). Remarkably, the regenerating sciatic nerves of the Khsrp^-/- mice showed approximately 5-fold higher axonal Prenyl-Cdc42 mRNA signals compared to those of the Khsrp^+/+ mice (Fig 5B). Thus, KHSRP likely restricts the axonal levels of Prenyl-Cdc42 mRNA in PNS nerves under both naïve and regenerating conditions.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Prenyl-Cdc42 mRNA is increased in KHSRP knockout axons.

A) Representative smFISH and IF images for naïve and 7 day post-crush injured sciatic nerve from Khsrp^+/+ or Khsrp^-/- mice. The upper row of each image set shows the merged confocal XY optical plane; the lower row of each image set shows RNA signal overlapping with NF across individual optical planes that was extracted to a separate channel and projected as XYZ images; see S5A Fig for representative images of scrambled smFISH probe [Scale bar = 5 µm]. B) Quantitation of smFISH signals for RNA probe signals overlapping with NF shown in A as mean ± SEM. See S5B Fig for separate intra-genotype comparison for naïve vs. crush nerves (N = 3 biological replicates; NS = not significant, * P < 0.05, **** P < 0.001 for indicated data pairs by two-way ANOVA with Sidak post-hoc tests for pair-wise comparisons) [Scale bar = 10 µm]. C) Representative exposure-matched smFISH/IF for Cdc42 mRNA in sciatic nerve axons of Khsrp^fl/fl:Syn1-Cre vs. Khsrp^fl/fl mice at 14 d post-crush shown in C. RVG-NPsiRNAs were applied at 7 d post-crush. Merged images show single XY planes and ‘axon only’ images show XYZ project of extracted RNA pixels that overlap with NF in individual Z planes [Scale bar = 5 µm]. D) Quantification of axonal Prenyl-Cdc42 mRNA levels shown in D. See S5E Fig for separate intra-genotype comparison for siCntl vs. siCdc42 treated nerves and S5F and S5G Fig for soma RNA values (N = 5 animals per condition; NS = not significant, ** P < 0.01, *** P < 0.005 for indicated data pairs by two-way mixed effect analysis ANOVA with Sidak post-hoc tests for pair-wise comparisons). E-F) Representative images for SCG10 immunostaining (E) and regeneration indices (F) for Khsrp^fl/fl mice and Khsrp^fl/fl:Syn1-Cre mice at 14 d post-crush with NP-siRNA injection at 7 d post crush with RVG-NPsiRNAs targeting Prenyl-Cdc42 mRNA vs. non-targeting siRNA control (N = 5 animals in Khsrp^fl/fl:Syn1-Cre and N = 3 animals for Khsrp^fl/fl mice; * P < 0.05, ** P < 0.01 by two-way mixed effect analysis ANOVA with Sidak post-hoc tests for pair-wise comparisons) [Scale bar = 100 µm].

https://doi.org/10.1371/journal.pgen.1011916.g005

Both the constitutive Khsrp^-/- mice and Khsrp^fl/fl mice exposed to AAV-Cre show accelerated PNS axon regeneration after traumatic nerve injury [13]. To test for potential functional significance of the elevated Prenyl-Cdc42 mRNA in Khsrp^-/- axons, we used an in vivo siRNA approach to deplete Prenyl-Cdc42 mRNA from sciatic nerve. We reasoned that delivering an siRNA directly to the sciatic nerve might allow us to preferentially deplete the mRNA from sciatic nerve axons. Thus, we packaged siRNAs targeting Prenyl-Cdc42 mRNA or non-targeting siRNA (siCdc42 and siCntl, respectively) into a polymersome nanoparticle [26]; the exterior surface of the nanoparticles was tagged with 29 amino acid rabies virus glycoprotein peptide-9R (RVG) that has been that has previously been used to deliver nanoparticle cargos to neurons [27] and has been shown to bind to NCAM [28]. Specificity and efficacy of the Prenyl-Cdc42 targeting siRNA has been previously published [4]. The siRNA laden RVG-nanoparticles (RVG-NPsiRNA) were initially tested in DRG cultures and showed clear uptake in neuronal soma and axons (S5C Fig). We next tested the RVG-NPsiRNAs in vivo by direct injection into the sciatic nerve of wild type mice that had undergone sciatic nerve crush 7 days previously to increase axonal Prenyl-Cdc42 mRNA. Wild type mice injected with RVG-NPsiCdc42 showed approximately 85% depletion of Prenyl-Cdc42 mRNA based on RTddPCR analyses of sciatic nerve axoplasm compared to siCntl treated nerves (S5D Fig). We next asked whether Prenyl-Cdc42 mRNA depletion from axons of KHSRP-deficient mice could decrease the accelerated regeneration seen in mice lacking neuronal KHSRP. For this, we compared sciatic nerve regeneration in Khsrp^fl/fl crossed to Syn1-Cre (KHSRP^fl/fl:Syn1-Cre) vs. KHSRP^fl/fl mice following injection with RVG-NPsiRNAs. KHSRP^fl/fl:Syn1-Cre mice show altered axon and dendrite growth and nerve injury in adult Khsrp^fl/fl mice where KHSRP was deleted by AAV-Cre show accelerated regeneration [13,29]. Thus, the KHSRP^fl/fl:Syn1-Cre mice allowed us to focus on effects of elevated axonal Prenyl-Cdc42 mRNA rather than potential confounding effects of non-neuronal cells in the nerve. 7 days after sciatic nerve crush, RVG-NPsiCdc42 vs. -siCntl were injected proximal to the injury site and at approximately same level in the contralateral (sham) nerve for KHSRP^fl/fl:Syn1-Cre and KHSRP^fl/fl mice. Axonal Prenyl-Cdc42 mRNA was depleted by approximately 94% in the RVG-NP-siCdc42 vs. -siCntl injected KHSRP^fl/fl:Syn1-Cre mice (Figs 5C, 5D and S5E). Murashov et al. (2007) showed that siRNAs can be retrogradely transported in sciatic nerve [30], and recent studies for rats where sciatic nerves were injected with RVG-NPs showed evidence for limited retrograde transport of these particles to the spinal cord [31]. Consistent with this, the L4-5 spinal motor neuron soma for the mice in Fig 4C showed approximately 40% reduction in Prenyl-Cdc42 mRNA levels by smFISH analysis (S5F and S5G Fig). Thus, although the effect of the RVG-NPsiRNAs is not limited to the site of injection, the axonal Prenyl-Cdc42 mRNA is overwhelmingly much more depleted in the sciatic nerve axons than in the motor neuron soma that projects axons into the sciatic nerve (i.e., 94 vs. 40% reduction). Nerve regeneration was significantly reduced by Prenyl-Cdc42 mRNA depletion from the axons of the Khsrp^fl/fl:Syn1-Cre mice (Fig 5E and 5F). In contrast, depletion of Prenyl-Cdc42 mRNA had no apparent effect on regeneration in Khsrp^fl/fl mice, which have wild type axonal KHSRP levels (Fig 5E and 5F). Taken together, these findings indicate that stabilization of Prenyl-Cdc42 mRNA in sciatic nerve axons contributes to the accelerated regeneration that we previously reported in KHSRP knockout mice.

Since exposure to the CSPG aggrecan also decreased axonal Prenyl-Cdc42 mRNA (see Figs 1 and 2), we asked if the decrease in Prenyl-Cdc42 mRNA after aggrecan exposure is mediated by KHSRP. Axons of Khsrp^+/+ neurons showed the anticipated decline in axonal Prenyl-Cdc42 smFISH signals following aggrecan exposure; however, there was no change in the axonal Prenyl-Cdc42 mRNA smFISH signals in Khsrp^-/- neurons following aggrecan exposure (Figs 6A, 6B and S6A). Since axonal translation of Khsrp mRNA is activated by increased axoplasmic Ca²⁺ [13] and CSPGs are known to increase axonal Ca²⁺ [32,33], we asked if Ca²⁺ is necessary for the aggrecan-induced decrease in axonal Prenyl-Cdc42 mRNA. Chelation of intracellular Ca²⁺ using BAPTA-AM blocked the aggrecan-induced decrease in axonal Prenyl-Cdc42 mRNA (Figs 6C, 6D and S6B). Aggrecan treatment also selectively increased axonal and not cell body KHSRP levels (Figs 6E, 6F and S6C). Taken together, these studies indicate that KHSRP binds to a sequence within Prenyl-Cdc42 mRNA’s nt 801–875, which is functionally distinct from the mRNA’s axonal localization motif (nt 764–800), and Ca²⁺-dependent elevation of axonal KHSRP promotes decay of axonal Prenyl-Cdc42 mRNA to slow axon growth.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Aggrecan-mediated increase in axonal KHSRP depletes axonal Prenyl-Cdc42 mRNA.

A) Representative exposure-matched smFISH and IF images for Prenyl-Cdc42 mRNA and NF in adult mouse Khsrp^+/+ or Khsrp^-/- DRG neuron cultures treated with 50 ng/ml aggrecan; see. S6A Fig for representative images of scrambled smFISH probe [Scale bar = 10 µm]. B) Quantitation of smFISH signal intensities shown as mean pixel intensity above background ± SEM for axons (N ≥ 40 neurons in three independent cultures; **** P < 0.001 for indicated data pairs and #### P < 0.001 for scramble vs. all but Khsrp^+/+ + aggrecan by two-way ANOVA with Sidak post-hoc tests for pair-wise comparisons). C) Representative exposure-matched smFISH and IF images for Prenyl-Cdc42 mRNA plus NF in adult DRG neuron cultures treated with 50 ng/ml aggrecan ± 3 µM BAPTA-AM; see S6B Fig for representative images of scrambled smFISH probe [Scale bar = 10 µm]. D) Quantification of smFISH signal intensities shown as mean pixel intensity above background for axons ± SEM (N ≥ 20 neurons in three independent cultures; **** P < 0.001 for indicated data pairs, ǂ P < 0.05 for scramble vs. aggrecan and #### P < 0.001 for scramble vs. no treatment and Aggrecan + BAPTA by Kruskal-Wallis ANOVA with Dunn post-hoc tests for pair-wise comparisons). E-F) Representative exposure-matched IF images for KHSRP protein plus NF in adult DRG neuron cultures ± 50 ng/ml aggrecans are shown in E. Quantification of cell body and axonal KHSRP signals from exposure-matched images is shown in F (see S6C Fig for representative IF images for cell bodies; N ≥ 75 neurons over 3 separate cultures; **** P < 0.0001 by student’s t-test) [Scale bars = 10 µm].

https://doi.org/10.1371/journal.pgen.1011916.g006

Ethics statement

All animal work was approved by the Institutional Animal Care and Use Committee at the University of South Carolina (AUP 2633-101765-012023). Recombinant DNA work was approved by the Institutional Biosafety Committee at the University of South Carolina (Protocol # 1-0114-0425).

Key reagents and resources

S1 Table contains details for key resources and source for those resources that were used in this study.

Animal care and use

Institutional Animal Care and Use Committee of the University of South Carolina approved all animal procedures. Sprague Dawley rats (175–250 g) were used for preparing sciatic nerve axoplasm. Male and female wild type C57Bl/6 (Khsrp^+/+), constitutive Khsrp knockout (Khsrp^-/-) [25], and conditional Khsrp^fl/fl [29] mice were used for sciatic nerve injury and DRG culture experiments as indicated in the results. For neuronal specific Khsrp knockout, male Khsrp^fl/fl mice were crossed to female B6.Cg-Tg(Syn1-cre)671Jxm/J (Syn1-Cre; Jackson Laboratories) mice. All animals were euthanized by CO₂ asphyxiation per IACUC guidelines.

For nerve crush surgery, animals were anesthetized with isoflurane by inhalation (5% induction and 2% maintenance). Anesthetized animals were subjected to sciatic nerve crush at mid-thigh level as previously described [51]. Briefly, the nerve was exposed by blunt dissection and then crushed with # 2 fine jeweler’s forceps, twice for 15 sec each; success of the axotomy was monitored by the initial contraction of the hind limb upon applying pressure to the nerve and then lack of hind paw extension during and upon recovery from anesthesia.

For in vivo RNA depletion from sciatic nerve axons, polymersomes with siRNAs (see below) were delivered by injecting 6 µl of polymersome solution in 1x PBS that contained an equivalent of 40 nM siRNAs. Polymersomes were injected into the sciatic nerve of anesthetized mice (see above) at 7 days following nerve crush injury (performed as above) just proximal to the injury site. Delivery of polymersomes was confirmed by RTddPCR for Prenyl-Cdc42 and Gapdh mRNAs, visualization of the polymersomes’ fluorophore in the nerve, and smFISH/IF for Prenyl-Cdc42 mRNA and neurofilament protein.

Mouse genotyping

Genotyping for constitutive KHSRP knockout was performed using PCR with primers spanning the exon 1 to exon 13 deletion of the mouse KHSRP gene or wild type sequence as previously described [29]. For this, DNA was extracted from ear punches taken at weaning. Primers used for genotyping are as follows (5’ to 3’): Khsrp forward P1 – TTCCGAAGCTCTGACTGGTC, Khsrp reverse P2 – CGGTGTTGTAGTCCGACATG, and Khsrp reverse P3 – AAGGGTCCAGGGTTGAAAGG. PCR products were analyzed by agarose gel electrophoresis with SYBRSafe DNA Gel Stain (ThermoFisher).

Khsrp^fl/fl mice were generated by Biocytogen using CRISPR/EGE-based gene editing to insert loxP sites between exons 1 and 2 and exons 6 and 7 as described [13]. Genotyping for loxP insertion was performed using following primers (5’ to 3’): 5’ LoxP forward – AGTGTTATGTGCTGGTGTGACCTGG, 5’ LoxP reverse – GTGCTTACCCTTGACAGGGAGTGTC, 3’ LoxP forward – CTATGGTGTCACCTCTCAGTGCTGC, and 3’ LoxP reverse – CACGTAGAGGCCAAAGCAAGAGGAC. PCR products were analyzed by agarose gel electrophoresis with SYBRSafe DNA Gel Stain (ThermoFisher). For specific Cre expression in neuronal cells, Khsrp^fl/fl mice were crossed with Syn1-Cre mice [29]. The following primers were used to detect Syn1-Cre-mediated recombination (5’ to 3’): forward transgene – CTCAGCGCTGCCTCAGTCT, reverse transgene – GCATCGACCGGTAATGCA, forward IPC – CAAATGTTGCTTGTCTGGTG, and reverse IPC – GTCAGTCGAGTGCACAGTTT. PCR products were analyzed by agarose gel electrophoresis with SYBRSafe DNA Gel Stain (ThermoFisher).

Primary neuron culture

Dissociated cultures of adult DRGs were prepared as described (Twiss et al., 2000). DRGs were harvested in Hybernate-A medium (BrainBits) and then dissociated with 2,000 units/ml Collagenase type 2 (ThermoFisher) at 37°C, 5% CO₂ for 15 min. Ganglia were triturated using a fire polished Pasteur pipet, diluted into 9 volumes DMEM/F12 (ThermoFisher), and then pelleted at 100 xg for 5 min. After pelleting, dissociated ganglia were washed in DMEM/F12 and then cultured in DMEM/F12, 1 x N1 supplement (Sigma-Aldrich), 10% fetal bovine serum (Hyclone), and 10 µM cytosine arabinoside (Sigma-Aldrich) and plated onto poly-L-lysine (Sigma-Aldrich) and laminin (ThermoFisher)-coated substrates.

For transfections, dissociated ganglia were pelleted at 100 x g for 5 min and resuspended in 100 µl ‘Nucleofector solution’ (Rat Neuron Nucleofector kit; Lonza). 4–6 µg of each plasmid was electroporated using the AMAXA Nucleofector device (G013 program; Lonza) before plating. Dissociated ganglia were then plated as above and analyzed 48–72 h later.

For RVG-NP treatments, dissociated cultures were exposed to RVG-NPsiRNAs (see below) for 2 hours, fixed in buffered 4% PFA, and then directly imaged by confocal microscopy.

Plasmid constructs

GFP^MYR translation reporter originally provided by Dr. Erin Schuman (Max-Plank Inst., Frankfurt) [16]. Mammalian expression plasmids with the coding sequence of the GFP^MYR containing cDNA corresponding to the 5’ and 3’UTRs of rat Prenyl-Cdc42 (GenBank Accession # XM_008764286; Lee et al., 2021) were used as a basis for 3’UTR deletion constructs GFP^MYR3’prenyl-Cdc42^764-913 and GFP^MYR3’prenyl-Cdc42^914-2164. Constructs were produced by digestion with either Not1 and BstX1 for GFP^MYR3’prenyl-Cdc42^764-913, or Bstx1 and EcoR1 for GFP^MYR3’prenyl-Cdc42^914-2164. 3’ overhangs were then filled using Klenow fragment (New England Biolabs) and re-ligated.

To create expression constructs for deletions of the Cdc42 3’UTR nt 764–913, double stranded oligonucleotides corresponding to nt 764–800, 764–838, 801–875, and 839–913 of rat Prenyl-Cdc42 mRNA were custom synthesized by Integrated DNA Technologies (IDT). These 3’UTR segments were engineered with 5’ Not1 and 3’ Xho1 restriction sites and used to replace the 3’UTR in GFP^MYR5’CamK2α/3’Actg plasmid. This plasmid contains the 5’UTR of calcium/calmodulin dependent protein kinase II alpha (CamK2α) that has previously been shown to lack any activity for axonal localization [52].

GFP^MYR5’/3’prenyl-Cdc42 used in our FRAP analyses was previously generated in our lab (Lee et al., 2021). mCherry^MYR5’/3’RhoA was generated by replacing the 5’ and 3’UTR of mCherry^MYR5′/3′Kpnb1 [13] with PCR generated sequences. Primers used for cloning are as follows (5’ to 3’): RhoA 5’UTR HinD3 forward – CCCAAGCTTTGAGTATAAAATAGCAACTCGGTCTTTTATAG, RhoA 5’UTR BamH1 reverse – CGGGATCCCACTTATGAAGGTGCTGAAGAAACTC, RhoA 3’UTR Not1 forward – GGGGCGGCCGCAGCCTTGTGAC, RhoA 3’UTR Xho1 reverse – GGGCTCGAGTTTAGAAAACTGCCT, corresponding to rat RhoA (GenBank Accession # XM_006243699). The 5’UTR was engineered with 5’ Hind3 and 3’ BamH1 restriction sites and the 3’ UTR was engineered with 5’ Not1 and 3’ Xho1 restriction sites and used to replace the 5’ and 3’ UTRs of mCherry^MYR5’/3’Kpnb1.

Polymersome delivery of siRNAs

Synthetic siRNAs were purchased from Dharmacon (Horizon Discovery Biosciences). The Prenyl-Cdc42 and non-targeting control siRNA sequences were previously published [4]. siRNAs were packaged into polymeric nanoparticles called polymersomes that were labeled with a peptide of rabies virus glycoprotein, RVG29 (called RVG herein). Polymersomes were made from solvent injection of a 50:50 mixture of block co-polymer polyethylene glycol (PEG, 1000 kDa)-b-polylactic acid (PLA, 5000 kDa) and PEG(1000)-b-PLA(5000)-maleimide, then lyophilized prior to siRNA encapsulation. Cysteine conjugated RVG is added to solution, enabling a thiol coupling reaction on the polymersome surface to attach the RVG. RVG attachment is confirmed via a shift in surface charge from negative to positive via zeta potential measurements [26]. RVG-tagged polymersomes co-encapsulated 13.6 ± 4.6 µg siRNA/mg polymer with 53 ± 8 µg membranous DiD/mg polymer. RVG-tagged siRNA loaded polymersomes were concentrated to 100 mM prior to injection.

siRNAs were first tested for uptake in primary mouse dissociated DRG cultures based on DiD fluorescence emission and sciatic nerve in vivo by RTddPCR for Prenyl-Cdc42 vs. Gapdh mRNA (see below).

Fluorescence in situ hybridization and immunofluorescence

Custom Stellaris oligonucleotide probes with 5′ Quasar 570 or 670 labels (BioSearch Tech.; see S1 Table) were used for smFISH combined with IF to detect Prenyl-Cdc42, RhoA and GFP mRNAs and Neurofilament (NF). Scrambled probes were used as control for specificity. For cultured neurons, coverslips with dissociated DRGs were briefly rinsed in phosphate buffered saline (PBS), fixed for 15 min in 2% PFA in PBS, and then processed for pre-hybridization and hybridization as described [53]. Primary antibodies to detect neurofilament (NF) consisted of a cocktail of mouse RT97 (1:500, Devel. Studies Hybridoma Bank) and SMI312 (1:250; BioLegend). FITC-conjugated donkey anti-mouse (1:200, Jackson ImmunoRes.) was used as the secondary antibody.

smFISH/IF was performed on sciatic nerve and spinal cord cryosections as described previously [53]. Briefly, tissues were immersion fixed overnight in 2% PFA in PBS at 4°C and then cryoprotected overnight in 30% sucrose in 1x PBS at 4°C. Samples were processed for cryosectioning (10 µm thickness), sections were adhered to Superfrost^plus glass slides (Fisher) and then stored at -20°C until use. Slides were dried at 37°C for 1 h and then brought to room temperature. All subsequent steps performed at room temperature unless indicated otherwise. Sections were washed for 10 min in PBS, then 10 min in 20 mM glycine three times followed by three 5 min incubations in fresh 0.25 M NaBH₄. Sections were rinsed in 0.1 M triethanolamine (TEA), incubated for 10 min in 0.25% acetic anhydride in 0.1 M TEA, and washed twice with 2x saline-sodium citrate (SSC) buffer. Sections were then dehydrated through graded ethanol solutions (70, 95, and 100% for 3 min each) followed by delipidation in chloroform for 5 min. Sections were rehydrated in 100 and 95% ethanol for 3 min each, equilibrated in 2x SSC and then incubated at 37°C in a humidified chamber in hybridization buffer (50% dextran sulphate, 10 μg/ml E. coli tRNA, 10 mM ribonucleoside vanadyl complex [Millipore Sigma], 80 μg BSA, and 10% formamide in 2 × SSC) for 5 min. Hybridization and immunolabeling were performed overnight in a humidified chamber in hybridization buffer containing 7 µM of each probe. Sections were then washed twice in 2x SSC plus 10% formamide at 37°C for 30 min and once in 2x SSC for 5 min. After permeabilization in PBS plus 1% Triton X-100 (PBST) for 5 min, nerve sections were incubated for 1 h in a cocktail of mouse RT97 (1:500, Devel. Studies Hybridoma Bank) and SMI312 (1:250; BioLegend) primary antibodies. FITC-conjugated donkey anti-mouse-IgG antibody (1:200) in 0.3% Triton X-100 supplemented with 1 × blocking buffer (Roche). Spinal cord samples were incubated in 1:400 dilution fluorescent Neurotrace 640/660 (ThermoFisher) for 1 h after FISH. After washing in PBS for 5 min, sections were post-fixed in 2% PFA in PBS for 15 min [54], washed in PBS three times for 5 min, and then rinsed in DEPC-treated water. Both the cultured neurons and tissues were mounted using Prolong Gold Antifade (Invitrogen).

All smFISH/IF performed on tissue sections was imaged by confocal microscopy using a Leica SP8X or Leica Stellaris confocal microscopes with HyD detectors and matched post-processing measures to distinguish axonal signals from non-neuronal signals. Scrambled probes were used to assign maximum acquisition parameters to limit any nonspecific signal from the probes.

Standard IF was performed as previously described [41] with all steps at room temperature unless specified otherwise. Cultures were fixed with 4% PFA in PBS for 15 min and washed 3 times in PBS. PBS washed cultures were permeabilized with 0.3% Triton X-100 in PBS for 15 min and then blocked in PBST plus 5% BSA for 1 h. Tissues were immersion fixed in 4% PFA in PBS for 4 h, cryoprotected overnight in 30% sucrose at 4°C, and then processed for cryosectioning at 25 µm thickness. Sections were stored at -80°C and then warmed to RT, and permeabilized for 15 min in 0.3% Triton X-100. Sections were washed in PBS wash 3 times and then blocked for 1 h in PBST containing 10% normal donkey serum. Tissue and culture samples were incubated with primary antibodies overnight at 4°C. Primary antibodies for cultured neurons consisted of rabbit anti-CDC42 (1:500; Abcam), mouse anti-RhoA (1:25; Abcam), and chicken anti-NF (1:500, NFM, NFL and NFH; Aves Labs). Primary antibodies for tissues consisted of rabbit anti-SCG10 (1:100; Novus Biologicals). After washing in PBS, coverslips were incubated with combination of FITC-conjugated donkey anti-chicken and Cy5-conjugated donkey anti-rabbit or anti-chicken (all at 1:500; Jackson ImmunoRes.) as secondary antibodies for 1 h. After 1 h, samples were washed 3 times in PBS, rinsed with distilled H₂O, and mounted with Prolong Gold Antifade with DAPI.

Fluorescence recovery after photobleaching

FRAP analyses were performed as published [18], with minor modifications. DRG neurons were transfected with GFP^MYR5’/3’prenyl-Cdc42 and mCherry^MYR5’/3’RhoA as above. Cells were maintained at 37°C, 5% CO₂ during imaging. 488 nm and 587 nm laser lines on a Leica SP8X confocal microscope with HyD detectors were used to bleach GFP and mCherry signals, respectively (argon laser at 70% power, pulsed every 0.82 sec for 80 frames). Leica 63x/1.4 NA oil immersion objective was used with the confocal pinhole set to 3 Airy units to ensure full thickness bleaching and acquisition (Yudin et al., 2008). 488 nm and 587 nm laser lines on the Leica SP8X confocal microscope with HyD detectors were used to bleach GFP and mCherry signals, respectively (argon laser at 70% power, pulsed every 0.82 sec for 80 frames). Regions of interest (ROI) for bleaching and analyses consisted of at least 50 µm of terminal axon length separated from the soma by at least 250 µm. Prior to photobleaching, two frames were acquired at 60 sec intervals to determine baseline fluorescence in the ROI (15% power with pulsed white light laser; 498–530 nm for GFP; 597–630 nm for mCherry). The same excitation and emission parameters were used to assess recovery over 15 min post-bleach with images acquired every 30 sec. For some experiments, 50 ng/ml of aggrecan (Sigma-Aldrich) or a neurotrophin cocktail consisting of 10 ng/ml each of NT3 (Alamone Labs) + BDNF (Alamone Labs) + NGF (Inotiv) was bath applied immediately prior to imaging. To test whether any fluorescence recovery in axons was due to translation, DRG cultures were treated with 100 µM anisomycin (Sigma) for 30 min prior to photobleaching.

RNA isolation and analyses

RNA was isolated from mouse DRG culture lysates and affinity pull down samples using the RNeasy Microisolation kit (Qiagen). Fluorimetry with Ribogreen (Invitrogen) was used for RNA quantification for total RNA isolates. For analyses of total RNA levels and inputs from immunoprecipitates, RNA yields were normalized for mass across samples prior to reverse transcription (RT) using Superscript IV Vilo (ThermoFisher). For co-immunoprecipitation of RNA, samples were processed based on equivalent proportions of the precipitate rather than normalizing to RNA mass. Droplet digital PCR (ddPCR) was performed using Taqman probe sets (IDT) and QX200 droplet reader (Bio-Rad). Primer/probe sets were as follows (5’ to 3’): prenyl-Cdc42 sense primer – CGTTTGTGGGGATTTGCGTT, prenyl-Cdc42 antisense primer – GACAGACGACCTGCACCTAC, prenyl-Cdc42 Probe -/56-FAM/GCCCCCTTG/ZEN/CCCTTCCGGTA/3IABkFQ/, GFP sense primer – CTGCTGCCCGACAACCAC, GFP antisense primer – TCACGAACTCCAGCAGGAC, and GFP probe -/56-FAM/CCAGTCCGC/ZEN/CCTGAGCAAAGACC/3IABkFQ/.

Affinity isolation of RNA-interacting proteins

RNA-Protein pull-down was performed as described [24]. We used axoplasm from sciatic nerve as a source of proteins. To obtain enriched axonal contents, approximately 3 cm segments of rat sciatic nerve were dissected and axoplasm was extruded into 20 mM HEPES [pH 7.3], 110 mM potassium acetate, and 5 mM magnesium acetate (‘nuclear transport buffer’) supplemented with 1x protease/phosphatase inhibitor cocktail (Roche) and 40 U/µl RNasin Plus (Promega) as previously described [55]. After clearing by centrifugation at 20,000 xg at 4°C for 30 min, supernatants were mixed with 5’ biotin-conjugated RNA oligonucleotides (IDT), which had been adsorbed to streptavidin (SA) dynabeads (ThermoFisher) [19], and incubated for 4 h at 4˚C. Beads were precipitated using a magnetic rack and then washed extensively with 10 mM HEPES (pH 7.4), 3 mM MgCl₂, 250 mM NaCl, 1 mM DTT, and 5% glycerol. Bound proteins were eluted by treating with 50 µg/ml RNase A (Sigma-Aldrich) in wash buffer for 15 min at 37˚C [19]. Proteins were denatured by boiling at 95˚C in Laemmli sample buffer for 5 min, fractioned by SDS/PAGE, and then transferred to nitrocellulose membranes for Immunoblotting.

RNA co-immunoprecipitation

For co-precipitating RNAs with proteins, DRG cultures were lysed in 100 mM KCl, 5 mM MgCl₂, 10 mM HEPES [pH 7.4], 1 mM DTT, and 0.5% NP-40 (‘RIP buffer’) supplemented with 1x protease inhibitor cocktail and RNasin Plus. Lysates were passed through 25 Ga needle 5–7 times and then cleared by centrifugation at 12,000 xg for 20 min. Cleared lysates were incubated with Protein G-Dynabeads (ThermoFisher) for 30 min to reduce non-specific binding. After collection, supernatants were then incubated with rabbit anti-KHSRP (5 μg, Novus) or rabbit IgG (5 μg, Jackson ImmunoRes.) for 3 h at 4˚C with rotation. Immunocomplexes were incubated with Protein G-Dynabeads for an additional 2 h at 4°C with rotation. Beads were washed six times with cold RIP buffer. An aliquot was reserved for validating protein precipitation (see below), and bound RNAs were isolated by addition of RNeasy Microisolation kit lysis buffer and analyzed by RTddPCR (see above).

For validation of pull-down of protein, the reserved aliquot from the immunoprecipitates was resuspended in 1 x Laemmli sample buffer and denatured by boiling at 95°C x 5 min. Supernatants were then processed for immunoblotting.

Protein electrophoresis and immunoblotting

Protein concentrations were determined by BCA assay. Cell lysates and axoplasm were normalized for concentration prior to electrophoresis or immunoprecipitation. Lysates, immunoprecipitates, and RNA affinity isolates were denatured by boiling in Laemmli sample buffer, fractionated by SDS-PAGE, and electrophoretically transferred to nitrocellulose membranes. Blots were incubated for 1 h at room temperature in blocking buffer (5% non-fat dry milk in Tris-buffered saline with 0.1% Tween 20 [TBST]). Membranes were incubated overnight incubation at 4°C with rocking in mouse anti-KHSRP (1:1000; Abcam) diluted in TBST plus 5% BSA. After washing in TBST, blots were incubated HRP-conjugated anti-mouse IgG antibodies (1:5000; Jackson ImmunoRes.) diluted in blocking buffer for 1 h at room temperature. After washing in TBST (3 times), signals were detected using Clarity Western ECL Substrate (Bio-Rad).

Quantitation and statistical details

All imaging experiments included at least three technical replicates for each culture, and each experiment was replicated across at least three separate culture preparations. For molecular studies using transfected cultures, analyses were performed across at least three separate culture preparations.

For smFISH on tissue sections the Z stacks of XY optical planes were captured at two locations along each nerve section. The Colocalization Plug-in for NIH ImageJ (https://imagej.nih.gov/ij/ plugins/colocalization.html) was used to extract RNA signals from smFISH probes in each optical plane that overlapped with NF signals as an ‘axon only’ mRNA signal. All smFISH signal quantifications for axonal mRNA signals from tissue sections were generated by analysis of pixel intensity across each XY plane of the extracted ‘axon only’ channels for the image sequences using ImageJ. These smFISH signal intensities across the individual XY planes were then normalized to the area of NF immunoreactivity in each XY plane and averaged across the image stack [53]. The relative mRNA signal intensity was then determined from the average in each biological replicate.

For FRAP assays, fluorescent intensities in the bleached region of interest (ROI) were calculated using Leica LASX software. For normalizing across experiments, fluorescence intensity value for the ROI at t = 0 min post-bleach from each image sequence was set as 0%. The relative fluorescence recovery at each time point after photobleaching was then calculated by normalizing to the pre-bleach fluorescence intensity of the ROI (set at 100%). All bleached ROIs were at least 250 µm from the cell soma, so any significant fluorescence recovery occurring in less than 15 min post-bleach interval can be attributed to local protein synthesis rather than anterograde transport of reporter protein into the ROI if recovery was significantly attenuated by protein synthesis inhibitors.

Quantitative data are reported as mean ± SEM. GraphPad Prism 9 software was used for all statistical analyses. Data outliers were removed using Prism’s ROUT with Q set at 1% (i.e., maximum allowable false discovery rate). Normality of data sets was assessed, if all samples pass normality test for Gaussian distribution, an ordinary ANOVA was performed with Tukey multiple comparisons test. Data that failed Gaussian distribution were then analyzed as nonparametric distribution with a Kruskal-Wallis test with Dunn’s multiple comparisons test. Statistical comparisons across experimental conditions in addition to genotype were analyzed either by two-way repeated measures ANOVA with Tukey post-hoc tests for pair-wise comparisons for data sets with equal numbers per group or two-way mixed effect analysis ANOVA with Sidak post-hoc tests for pair-wise comparisons for data sets with unequal number in each group. Pairwise comparisons were performed by either Student or Welch’s t-test as indicated.