The CRMP4, Cav1.2 (a1c, $\alpha$2$\delta$, $\beta$2a), Small Ubiquitin-like Modifier 1 (SUMO1), SUMO2, SUMO3 and Ubc9 genes were cloned from rat cDNA and validated via sequencing. The Glutathione S-transferase (GST)-pulldown assay, CRMP4, Cav1.2 and SUMO1-3 cDNA were inserted into the pGEX-5x-3 vector (Piscataway, NJ, USA). For immunoprecipitation, CRMP4 or the mutant was subcloned into pEGFP-C1 (Takara Bio Europe, Saint-Germain-en-Laye, France), Cav1.2 $\alpha$1c was cloned into pcDNA3.1-V5/His-A vector and SUMO2 into pCMV-HA vectors (Stratagene, Santa Clara, CA, USA). The CRMP4 K374R mutant was constructed by a site-directed mutagenesis kit (Takara). Ubc9 was inserted into a pCMV-FLAG-C vector.

Cortical neurons were separated as previously reported [24]. Briefly, newborn 1-day-old Sprague-Dawley rats from the Animal Center of Jinan University were sacrificed by CO${}_2$ anesthesia. Then the cortices were dissected, cut into small pieces, and digested with 0.05% trypsin (#25300062, Life Technologies Corporation, Grand Island, NY, USA) for 15 min at 37 ${}^{\circ }$C. After dissociation into a single-cell suspension, debris was removed by filtration through Falcon cell strainers. The resultant cells were maintained in a Neurobasal medium with B27 and Glutamax (#17504044, #35050061, Life Technologies Corporation). Human embryonic kidney (HEK) 293 cells (#CRL-1573, American Type Culture Collection, Manassas, VA, USA) were cultured as described previously [25]. Transient transfections of neurons were performed with calcium phosphate, and HEK293 cells were transfected with Lipofectamine 2000 (#11668019, Invitrogen Corporation, Carlsbad, CA, USA).

GFP-overexpressing neurons were immunostained with GFP antibody to boost the GFP signal and reveal the neurites to measure cortical neuron morphology. Images were captured in a blinded manner with an Axio Observer Z1 (#AXIO Observer Z1, Carl Zeiss, Germany). At least 60 neurons in each group were imaged. The total neurite length and branch tip number of each neuron were analyzed using ImageJ [26].

The procedure was performed as previously reported [27]. Briefly, cell lysates were separated by 10% SDS-PAGE and transferred onto a PVDF membrane (Millipore, Sigma, Burlington, MA, USA). The membrane was blocked at room temperature for 1 h with 5% non-fat milk in TBS containing 0.1% Tween-20 and then incubated at 4 ${}^{\circ }$C overnight with antibodies against CRMP4 (#ab244319), Cav1.2 (#ab58552), GFP (#ab290), HA (#ab9110), FLAG (#ab1162), or V5 (#ab27671) (all from Abcam, Cambridge, MA, USA) in TBS buffer containing 3% BSA. GAPDH was used as a loading control. Blots were incubated with secondary antibodies at room temperature for 1 h and then visualized using enhanced chemiluminescence reagents (#WP20005, Thermo Fisher Scientific, Rockford, IL, USA). Results of western blotting are representative of three independent experiments.

GST-CRMP4 and GST-Cav1.2 (a1c) plasmids were transformed into Top10F${}^{\prime }$ E. coli (Invitrogen), and protein expression was induced with 0.1 mM IPTG (Roche Applied Science, Indianapolis, IN, USA). Cells were collected and lysed, and GST-fusion proteins were purified from the supernatant using glutathione–agarose beads (#16101, Pierce Biotechnology, Rockford, IL, USA). The GST-pulldown assay was performed in buffer containing 100 mM NaCl, 20 mM Tris-HCl, 5% glycerol (pH 7.0), 1% Triton X-100, and protease inhibitor cocktail (#539131, Merck, MA, USA), and GST-fusion proteins were incubated with SD rat brain lysates at 4 ${}^{\circ }$C for 8 h. After pulldown, the sedimented proteins were subjected to western blotting.

For immunoprecipitation assays, neuronal extracts were prepared by solubilization in cell lysis buffer (Beyotime Institute of Biotechnology) for 10 min at 4 ${}^{\circ }$C. After a brief sonication, lysates were cleared by centrifugation at 15,000 $\times$ g for 10 min at 4 ${}^{\circ }$C, cell extract was immunoprecipitated with 4 $\mu$g antibodies against CRMP4 (Abcam, #ab244319) or Cav1.2 (Abcam, #ab58552), and the samples were incubated with 60 $\mu$L protein G/A agarose (Calbiochem EMD Millipore) overnight at 4 ${}^{\circ }$C with continuous inversion. Immunocomplexes were pelleted and washed six times, and then the sedimented proteins were subjected to western blotting.

Immunofluorescence assays were performed as previously described [28]. Briefly, neurons cultured on coverslips were fixed in 4% paraformaldehyde for 40 min at 4 ${}^{\circ }$C and blocked with 3% BSA in TBS. Coverslips were then incubated with rabbit anti- GFP (Abcam, #ab290), -CRMP4 (Abcam, # ab244319), and -SUMO2/3 antibodies (Abcam, #ab81371) overnight at 4 ${}^{\circ }$C. After washing with TBST, neurons were labeled with Alexa Fluor 555 donkey anti-goat IgG (H + L) and Alexa Fluor 647 donkey anti-rabbit IgG (H + L; Life Technologies) for 1 h at room temperature. Then neurons were mounted on glass slides using Fluoro Gel II containing DAPI for confocal microscopy studies (LSM 700; Zeiss, GmbH, Germany).

HEK293T cells cultured on coverslips were transfected with DsRed-Cav1.2 ($\alpha$1c, $\alpha$2$\delta$ and $\beta$2a), GFP, GFP-CRMP4, or GFP-CRMP4${}^{K374R}$ plasmids for 24 h. Transfected cells were selected based on the expression of GFP and DsRed. HEK293T cells were patched in an external solution of modified Krebs-Ringer buffer (KRB) containing (in mM): 146 NaCl, 4.7 KCl, 2.5 CaCl${}_2$, 0.6 MgSO${}_4$, 1.6 NaHCO${}_3$, 0.15 NaH${}_2$PO${}_4$, 8 glucose, and 20 HEPES (pH 7.4); the buffer was ~330 mOsm. In transfected cells, I${}_{Ca}$ was recorded using a whole-cell voltage-clamp patch with an internal solution containing (in mM): 125 Cs-methanesulfonate, 10 TEA-Cl, 1 MgCl${}_2$, 0.3 Na${}_2$-GTP, 13 phosphocreatine-(di)Tris, 5 Mg$\cdot$ATP, 5 EGTA, 10 HEPES (adjusted to pH 7.22 with CsOH). Recordings were performed at room temperature in voltage-clamp mode at a holding potential of –70 mV using a MultiClamp 700 B amplifier (Molecular Devices, Sunnyvale, CA, USA) and the Clampex 10.5 software (Axon Instruments, Union City, CA, USA). Current-voltage (I–V) relationships were obtained approximately 3 min after obtaining the whole-cell configuration by subjecting cells to a series of 300 ms depolarizing pulses from the holding potential of –70 mV to test potentials ranging from –60 to +100 mV in 10-mV increments.

The viral vectors, including AAV-GFP, AAV-CRMP4-GFP and AAV-CRMP4-K374R-GFP (serotypes 2, 1 $\times$ 10${}^{12}$ to 5 $\times$ 10${}^{12}$ vg/mL), were purchased from OBIO Technology (Shanghai, China).

Individual rats (male Sprague-Dawley rat, 180–200 g, 8 weeks of age) were placed in test chambers, supported by an elevated platform, with no bottom or cover on a wide-gauge wire mesh. After habituation for 30 min, mechanical pain thresholds were measured, and the 50% paw withdrawal threshold (PWT) was determined as described previously [29]. A heat/light source was positioned under the glass surface, pointing to the plantar surface of the hind paw. Thermal paw withdrawal latencies (PWLs) were measured using a plantar anesthesia tester (Boerni, Tianjin, China).

The experimental data are presented as means $\pm$ SD from at least three experiments. The SPSS 19.0 software (SPSS Software, Chicago, IL, USA) was used for statistical analysis. Student’s t-test was used for comparisons between two groups, and one-way ANOVA was used for multiple comparisons; P 0.05 was considered significantly different.

Using bioinformatics tools (SUMPsp, SUMOplot, and GPS-SUMO), we found that CRMP4K374 was a potential SUMOylation site. The GKMD sequence around CRMP4 K374 is consistent with the canonical $\mathrm{\Psi }$KXE/D motif and is evolutionarily conserved (Fig. 1A). To confirm the SUMOylation modification of CRMP4, we performed GST-pulldown, immunofluorescence, and immunoprecipitation (IP) assays. As shown in Fig. 1B, CRMP4 interacted with GST-SUMO1, GST-SUMO2, and GST-SUMO3, with GST-SUMO2 yielding the strongest signal. Because of the homology between SUMO proteins, we selected SUMO2 for subsequent study. Immunofluorescence staining revealed colocalization of CRMP4 and SUMO2/3 in the cell body and branches of neuronal processes, represented as merged yellow signals (Fig. 1C). Furthermore, epitope-tagged GFP-CRMP4 co-immunoprecipitated with HA-SUMO2, under the co-transfection of FLAG-Ubc9 (a sole E2 for SUMOylation) (Fig. 1D); however, when CRMP4 K374 was mutated to arginine (CRMP4${}^{K374R}$), GFP-CRMP4${}^{K374R}$ could not precipitate with HA-SUMO2 (Fig. 1E). These data show that CRMP4 could be SUMOylated at K374.

Fig. 1.

CRMP4 can be SUMOylated at K374. (A) The amino acids around K374 are conserved among species. $\mathrm{\Psi }$KxD is the consensus motif for a SUMOylation site: $\mathrm{\Psi }$ is a hydrophobic amino acid, x is any amino acid, and D is an acidic amino acid. (B) GST-SUMO proteins can pull down CRMP4. The indicated GST-SUMO plasmids were constructed, the corresponding proteins were purified and incubated with brain lysates in pulldown assays, and the sediments were subjected to western blotting. N = 3. (C) Cultured neurons were stained with CRMP4 and SUMO2 antibodies for immunofluorescence. Scale bar, 20 $\mu$m. (D,E) 293T cells were co-transfected GFP-CRMP4 or GFP-CRMP4${}^{K374R}$ and FLAG-Ubc9, with or without plasmids encoding HA-SUMO2. Cell lysates were subjected to western blotting with GFP, HA, or FLAG antibodies. N = 3.

To further explore the detailed mechanism of CRMP4 SUMOylation, we applied interaction proteomics to identify the functional partners of CRMP4. This approach revealed that the L-type calcium channel Cav1.2 interacted with CRMP4 (data no shown). Hence, we performed GST-pulldown and immunoprecipitation experiments to confirm this potential interaction. Purified GST-CRMP4 or GST-Cav1.2 protein was incubated with cortical neuronal lysates, and then the pulldown sediments were subjected to western blotting. As shown in Fig. 2A–B, Cav1.2 was detected in the GST-CRMP4 pulldown sediment and CRMP4 in the GST-Cav1.2 pulldown sediment. Next, we subjected neuronal lysates to co-immunoprecipitation experiments using CRMP4 and Cav1.2 antibodies. The results confirmed that CRMP4 and Cav1.2 proteins were present in each other’s immunoprecipitates (Fig. 2C–D). Together, these data indicate that CRMP4 interacts with Cav1.2. Next, we asked whether CRMP4 SUMOylation affects the interaction between CRMP4 and Cav1.2. To this end, we transfected HEK293 cells with V5-Cav1.2, GFP-CRMP4, or GFP-CRMP4${}^{K374R}$ plasmids together with plasmids encoding SUMO2 and Ubc9. As shown in Fig. 3, GFP-CRMP4 signal was detected in the immunoprecipitates obtained with V5 antibody (Fig. 2E), but GFP-CRMP4${}^{K374R}$ yielded a stronger signal than GFP-CRMP4 (Fig. 2F). These data indicate that CRMP4 interacts with Cav1.2, and SUMOylation of CRMP4 affects this interaction.

Fig. 2.

CRMP4 interacts with Cav1.2. (A,B) Neuronal lysates were incubated with GST or GST-CRMP4 and GST-Cav1.2 for GST-pulldown assays, and the pulldown sediments were subjected to western blotting. (C,D) Neuronal lysates were incubated with IgG or Cav1.2/CRMP4 antibodies for immunoprecipitation assay, and the sediments were subjected to western blotting analysis. (E,F) 293T cells were transfected with the indicated plasmids; 24 h later, cells were harvested for immunoprecipitation with V5 antibody and western blotting with GFP or V5 antibody. N = 3 for all experiments.

CRMP4 promotes neurite growth in cultured hippocampal neurons [14, 30]. However, the role of CRMP4 SUMOylation remains to be explored. To further evaluate the effect of CRMP SUMOylation on neurite growth, we transfected cortical neurons with GFP control, GFP-CRMP4, and GFP-CRMP4${}^{K374R}$ along with plasmids encoding Ubc9 and SUMO2. Because GFP-CRMP4${}^{K374R}$ cannot be SUMOylated, this mutant form of CRMP4 mimics the deSUMOylated state of CRMP4. As shown in Fig. 3A, successfully transfected neurons were stained with GFP to reveal neurite growth, and then the length and number of branches were measured. As shown in Fig. 3B–C, transfection of CRMP4 significantly increased both length and tip number. However, overexpression of the deSUMOylated form of CRMP4 (GFP-CRMP4${}^{K374R}$) further increased the growth of the neuronal process relative to GFP-CRMP4, which could be SUMOylated. These data suggest that deSUMOylation of CRMP4 further promotes the neurite growth of cortical neurons.

Fig. 3.

The deSUMOylated form of CRMP4 promotes neurite growth. (A) DIV3 Cultured neurons were transfected with plasmids encoding Ubc9, SUMO2, GFP, GFP-CRMP4, or GFP-CRMP4K374R, then harvested DIV5 neuronal morphological analysis. Representative images of neurons are shown in (A), and length per neuron (B) and the number of tips per neuron (C) were measured in each group. Each group was enrolled in over 30 neurons. N = 3; Scale bar, 50 $\mu$m. * denotes P 0.05.

The interaction between CRMP4 with Cav1.2 suggests that CRMP4 is involved in calcium influx and neuronal signal transmission. To confirm this, we transfected HEK293 cells with a plasmid encoding GFP, GFP-CRMP4, or GFP-CRMP4${}^{K374R}$ and Cav1.2 and then conducted a whole-cell patch-clamp to record the changes in the calcium current through Cav1.2. The peak current-voltage curves for the various transfections of –70 mV are shown in Fig. 4A. The results revealed that GFP-CRMP4 significantly increased the current density of Cav1.2, whereas GFP-CRMP4${}^{K374R}$ further promoted the effect (Fig. 4B). The same trends were apparent in I–V curves from different groups (Fig. 4C). Together, these data suggest that CRMP4 deSUMOylation promotes calcium influx through Cav1.2.

Fig. 4.

CRMP4 SUMOylation regulates calcium influx via Cav1.2. (A) 293T cells were transfected with Cav1.2 ($\alpha$1c, $\alpha$2$\delta$ and $\beta$2a), together with plasmids encoding GFP, GFP-CRMP4, or GFP-CRMP4${}^{K374R}$, and then whole-cell patch-clamp was applied to record the calcium current via Cav1.2. (B) I–V curves of GFP control, GFP-CRMP4 (green) and GFP-CRMP4${}^{K374R}$ (red). (C) Current density values at +10 mV. * denotes P 0.05 vs. GFP; # denotes P 0.05 vs. GFP-CRMP4. N = 4, each group was tested over 20 cells.

Because Cav1.2 is involved in pain signaling [10], we asked whether CRMP4 and SUMOylation regulate pain sensation. To determine this, we used a rat model of thermal pain sensitivity. CRMP4-overexpressing AAV (AAV-CRMP4 and AAV-CRMP4${}^{K374R}$) viruses were prepared and injected into the sciatic nerve. As shown in Fig. 5A, CRMP4-overexpressing rats exhibited thermal pain sensitivity in the injected hind paws, and deSUMOylated CRMP4${}^{K374R}$ further decreased the threshold of thermal pain sensation (Fig. 5B). These data suggest that deSUMOylation of CRMP4 promotes thermal pain sensitivity in rats.

Fig. 5.

CRMP4 SUMOylation regulates thermal pain sensitivity in rats. AAV viruses of Control, CRMP4, or CRMP4${}^{K374R}$ were injected into the hind paws of rats, and the rats were tested for thermal pain sensitivity. (A) Thermal pain thresholds were determined using a plantar tester. (B) Effects of CRMP4 overexpression on thermal pain thresholds of rats. * denotes P 0.05.