Abstract

   Myelination is essential for neuronal function and health. In
   peripheral nerves, >100 causative mutations have been identified that
   cause Charcot-Marie-Tooth disease, a disorder that can affect myelin
   sheaths. Among these, a number of mutations are related to essential
   targets of the posttranslational modification neddylation, although how
   these lead to myelin defects is unclear. Here, we demonstrate that
   inhibiting neddylation leads to a notable absence of peripheral myelin
   and axonal loss both in developing and regenerating mouse nerves. Our
   data indicate that neddylation exerts a global influence on the complex
   transcriptional and posttranscriptional program by simultaneously
   regulating the expression and function of multiple essential
   myelination signals, including the master transcription factor EGR2 and
   the negative regulators c-Jun and Sox2, and inducing global secondary
   changes in downstream pathways, including the mTOR and YAP/TAZ
   signaling pathways. This places neddylation as a critical regulator of
   myelination and delineates the potential pathogenic mechanisms involved
   in CMT mutations related to neddylation.
     __________________________________________________________________

   Neddylation is a potent coordinator of the molecular circuitry that
   promotes Schwann cell myelination and axonal integrity.

INTRODUCTION

   The myelin sheath is essential for neuronal function and health. It
   involves a marked expansion of the plasma membrane that requires a very
   substantial amount of proteins and lipids to be continuously produced.
   In peripheral nerves, this process of myelination is exquisitely
   regulated at multiple levels, including a broad spectrum of extrinsic
   signals, such as neuregulin 1 and laminins, and intracellular inputs
   that include the mechanistic target of rapamycin (mTOR), Wnt/β-catenin,
   and Hippo-Yap signaling pathways, which ultimately converge on an
   intricate network of transcriptional, epigenetic, or
   posttranscriptional activators and repressors ([108]1, [109]2).
   Individually, these extrinsic and intrinsic regulators and effectors
   have been shown to be essential for myelination, as shown by nerve
   defects in genetic mice models and in the >100 genes containing
   causative mutations identified in Charcot-Marie-Tooth disease (CMT)
   ([110]2–[111]4), a group of inherited conditions that can affect myelin
   sheaths in peripheral nerves.

   We found that several of these >100 genes (5 to 6%) are related to the
   posttranslational modification (PTM), neddylation. Neddylation involves
   the conjugation of the ubiquitin-like protein NEDD8 to substrate
   proteins through a sequential, three-step process that is facilitated
   by E1, E2, and E3 enzymes ([112]Fig. 1A). The E1 complex consists of
   NAE1 and UBA3. The E2 enzyme aids a consecutive transthiolation
   reaction of NEDD8. E3 enzymes catalyze the transfer of NEDD8 from the
   E2 enzyme onto the neddylation target. Cullins are the best-known
   targets of neddylation ([113]5). Neddylation is essential to activate
   cullin-RING ligase complexes (CRLs), which are the largest family of
   multi-subunit E3 ligases in eukaryotes that tag proteins for
   degradation via the 26S proteasome ([114]5, [115]6). CRLs are modular
   assemblies typically built around a central cullin scaffold, which
   associates with an E2 interacting RING finger protein, an adaptor
   protein, and a substrate receptor module (SRM), such as F-box,
   suppressor of cytokine signaling, Broad-Complex, Tramtrack, and Bric a
   brac, and DDB1 and CUL4 associated factor (DCAF) proteins ([116]Fig.
   1A). By pairing one of eight cullins to one of 200 SRMs available, CRLs
   are responsible for ubiquitylation and degradation of thousands of
   substrates and are estimated to contribute to about 20% of protein
   turnover in cells ([117]5, [118]6).

Fig. 1. Pharmacological inhibition of neddylation blocks Schwann cell
myelination.

   [119]Fig. 1.
   [120]Open in a new tab

   (A) Graphical representation of the neddylation pathway and its main
   target, Cullin proteins. Mutations in several genes that encode
   critical components of the neddylation-regulated CRL complexes (pink
   boxes) are associated with peripheral neuropathies. (B) Immunoblot
   analyses of NAE1 and neddylated proteins, myelin proteins (MPZ and
   MBP), and positive (EGR2 and ZEB2) and negative (c-Jun and Sox2)
   regulators of myelination in total sciatic nerve extracts at indicated
   ages. (C and D) Nae1 expression in resident cells from P1 sciatic
   nerves in single-cell RNA sequencing (scRNA-seq) dataset [121]GSE138577
   ([122]17) [pericytes and vascular smooth muscle cells (Per/VSMC) and
   fibroblast-related cluster (FbRel*)]. (C) UMAP expression of Nae1.
   Color corresponds to log-normalized expression values scaled to the
   maximum of the gene. UMAP, Uniform Manifold Approximation and
   Projection. (D) Violin plot showing the log-normalized Nae1 transcript
   levels across the three Schwann cell clusters. Wilcoxon rank sum test
   between groups. Prol. SC, proliferating Schwann cells; iSC, immature
   Schwann cells; pmSC, pro-myelinating Schwann cells. (E) Bar plot
   showing Nae1 transcript levels in populations enriched in myelinating
   and non-myelinating cells (mSC and nmSC, respectively) from P5 sciatic
   nerves from public dataset [123]GSE138577 ([124]17). Two-tailed
   unpaired Student’s t test. (F) Representative electron microscopy (EM)
   pictures of P5 sciatic nerve cross sections from mice treated with
   vehicle and MLN4924. Scale bar, 2 μm. Graph shows quantification of
   myelinated axons per sciatic nerve. Data are presented as means ± SEM;
   n = 3 to 4. Two-tailed unpaired Student’s t test. (G) Immunoblot and
   densitometric analyses of MPZ, MBP, and NEDD8 in sciatic nerve extracts
   from vehicle or MLN4924-treated mice. Data are presented as means ±
   SEM; n = 5 vehicle and 6 MLN4924 treatment. Two-tailed unpaired
   Student’s t test. a.u., arbitrary units. (H) Immunoblot analyses of
   MPZ, NAE1, and NEDD8 levels in primary rat Schwann cells cultured under
   basal and myelinogenic conditions [dibutyryl cyclic adenosine
   3′,5′-monophosphate (db cAMP) treatment], in the presence or absence of
   MLN4924. β-Actin is used as loading control for immunoblots in (B),
   (G), and (H).

   The CMT-associated mutations include biallelic mutations in the GAN
   gene, which leads to loss of function of Gigaxonin a substrate adaptor
   of Cul3-E3 ubiquitin ligases and causes giant axonal neuropathy (GAN),
   a progressive childhood neurodegenerative disease that presents as a
   prominent sensorimotor neuropathy and affects both the peripheral
   nervous system and central nervous system ([125]7). Similarly,
   mutations in various other genes encoding core components of CRL
   complexes, including Cullin 4B, FBXO38, or DCAF8 ([126]5, [127]8), or
   proteases possibly involved in removing NEDD8 from protein conjugates,
   such as UCHL1 ([128]8, [129]9), have been identified as causes of
   peripheral neuropathy ([130]10–[131]13). Although, so far, patients
   with mutations have not been reported, genetic ablation of other genes
   related to CRL complexes, including Fbxw7 ([132]14) and Fbxo7, in mice
   models can also lead to nerve defects.

   Furthermore, a recent study identified four individuals with biallelic
   missense variants in the E1 enzyme NAE1 ([133]Fig. 1A), with a range of
   clinical phenotypes, including neurological deficits ([134]15). Brain
   magnetic resonance imaging showed decreased myelination in the four
   individuals. These studies strongly suggest that neddylation and its
   activation of downstream targets, e.g., CRLs, could be important for
   driving myelination, although, according to current literature, there
   are no reports examining the biological and molecular function of
   neddylation in Schwann cells. In this study, we found that neddylation
   inhibition, both genetically and pharmacologically, leads to a block in
   peripheral nerve myelination in developing and regenerating nerves in
   vivo. Notably, this leads to secondary axonal loss. We further
   demonstrate that neddylation has complex regulatory roles by
   fine-tuning the activation or suppression of several of the most
   critical myelination regulators, including the master myelination
   transcription factor early growth response 2 (EGR2), the negative
   transcriptional regulators, c-Jun and Sox2, as well as the mTOR and
   yes-associated protein (YAP)/transcriptional coactivator with
   PDZ-binding motif (TAZ) pathways. Our data thus establish neddylation
   as a potent coordinator of the molecular circuitry that promotes
   myelination and axonal integrity, a finding that provides a strong
   foundation for understanding the pathogenicity of several causative
   mutations associated with the neddylation pathway in patients with CMT.

RESULTS

Neddylation pathway components are expressed during Schwann cell
differentiation

   The role of neddylation in the peripheral nervous system has, so far,
   not been studied. To examine a potential role of neddylation in Schwann
   cell myelination, we analyzed by real-time quantitative polymerase
   chain reaction (RT-qPCR), the levels of the mRNAs encoding both E1-NAE
   subunits (Uba3 and Nae1), the E2 Ubc12, and Nedd8 in sciatic nerves
   from mice at different ages that broadly correspond to the main stages
   of the myelination process: In developing nerves, immature Schwann
   cells envelop groups of axons communally, forming irregular
   axon/Schwann cell columns or families. Large diameter axons then become
   segregated from the immature Schwann cell families to form a 1:1
   relationship with a Schwann cell, a process known as radial sorting,
   giving rise to pro-myelinating Schwann cells. These then subsequently
   elaborate myelin sheaths ([135]1, [136]2, [137]16). Thus, nerves were
   obtained from newborn mice (NB), which are highly enriched for immature
   Schwann cells, postnatal day 7 and 15 (P7 and P15) nerves, which
   contain mostly actively myelinating Schwann cells, and P30 and P90
   nerves, which contain terminally differentiated Schwann cells. We found
   that there was an increase in Nae1 and Uba3 levels with development,
   whereas there were minimal changes for Ubc12 and Nedd8 (fig. S1A).

   To confirm this, we performed immunoblot analyses and found that NAE1
   was highly expressed during the differentiation of Schwann cells, and,
   notably, we found high levels of neddylated proteins at ages,
   corresponding to the onset of myelination (NB to P7), and a progressive
   down-regulation as the actively myelinating Schwann cells complete
   their maturation to terminally differentiated Schwann cells (P15
   onward). Immunoblot analyses also showed elevated levels of myelination
   regulators EGR2 and zinc finger E-box binding homeobox 2 (ZEB2) at
   active stages of myelination, a progressive up-regulation of myelin
   proteins myelin protein zero (MPZ) and myelin basic protein (MBP), and
   sharp down-regulation of negative regulators c-Jun and Sox2, as
   immature Schwann cells undergo the myelination process ([138]Fig. 1B).
   To specifically show that Nae1 was expressed in Schwann cells, we
   examined its expression in single-cell RNA sequencing (scRNA-seq)
   public datasets of the developing sciatic nerve ([139]17). We found
   that Nae1 was expressed in Schwann cells, as well as other cell types
   in peripheral nerves ([140]Fig. 1C) and that its expression was higher
   in myelinating Schwann cells compared to that in immature Schwann cells
   ([141]Fig. 1D). We confirmed this in bulk RNA-seq datasets of purified
   myelinating and non-myelinating Schwann cell populations ([142]Fig. 1E)
   ([143]17). These results show that neddylation pathway components are
   expressed in Schwann cells at the onset and active phases of the
   myelination process.

Pharmacological inhibition of neddylation in vivo blocks Schwann cell
myelination

   Our results above suggested that neddylation could be important for
   Schwann cell myelination. To test this, we examined the effects of
   pharmacological inhibition of neddylation on myelination in vivo. Mice
   pups at 3 days of age (P3) were treated with vehicle or the NAE1
   inihibitor MLN4924 ([144]6) for 2 days, and, then, sciatic nerves were
   analyzed by transmission electron microscopy (TEM) or immunoblot
   analysis. We found that there were fewer myelin sheaths in nerves from
   MLN4924-treated mice ([145]Fig. 1F). This was accompanied by a decrease
   in levels of the myelin proteins MPZ and MBP and neddylated proteins in
   MLN4924-treated nerves ([146]Fig. 1G). We also found that MLN4924
   blocks up-regulation of the myelin protein MPZ in primary rat Schwann
   cells treated with dibutyryl cyclic adenosine 3′,5′-monophosphate
   (cAMP) analogs [dibutyryl (db) cAMP], which can potently up-regulate
   MPZ, as a surrogate in vitro model of Schwann cell differentiation
   ([147]Fig. 1H) ([148]18). These results suggest that neddylation could
   be regulating Schwann cell myelination.

Genetic deletion of Nae1 in vivo leads to development of peripheral
neuropathy

   To further investigate the role of neddylation in Schwann cell
   development, we generated a conditional mouse model lacking NAE1,
   specifically in Schwann cells. For this, we generated Nae1 floxed mice
   and crossed them with the MPZ-Cre mice, a line where Cre recombinase is
   expressed under the control of the endogenous MPZ promoter that Schwann
   cells express since E13.5 (fig. S2A) ([149]19). The offspring of
   MPZ-Cre^+;Nae1^f/f mice are referred to as Nae1 conditional knockout
   (cKO) mice, and their littermates MPZ-Cre^−;Nae1^f/+ or
   MPZ-Cre^−;Nae1^f/f mice were used as controls. PCR and RT-qPCR analyses
   (fig. S2, B and C) confirmed the deletion of Nae1 in P15 mutant nerves.
   Immunoblot analyses of nerve lysates revealed a strong reduction of
   NAE1 levels, which was accompanied by a decrease in total neddylated
   proteins in P30 sciatic nerves from Nae1 cKO mice. Intermediate levels
   of NAE1 were observed in heterozygotes ([150]Fig. 2A).

Fig. 2. Nae1 mutant mice show severe nerve deficits.

   [151]Fig. 2.
   [152]Open in a new tab

   (A) Immunoblot and densitometric quantification of NAE1 and neddylated
   proteins in sciatic nerve extracts from WT, heterozygotes (HET), and
   Nae1 cKO mice. Low and high exposures are shown for NEDD8 immunoblot to
   visualize differences in high (asterisk) and low MW proteins
   (arrowhead) respectively. Data are presented as means ± SEM. One-way
   analysis of variance (ANOVA) with Tukey’s multiple comparisons test.
   Residual expression of NAE1 likely corresponds to other cell types
   present in nerves ([153]Fig. 1C). β-Actin and
   glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is used as loading
   control for immunoblots. (B) Survival curves of WT and Nae1 cKO mice (n
   = 40 WT and 48 Nae1 cKO). P values are from a log-rank test between
   groups. (C) Tail suspension test showing abnormal hindlimb clasping in
   Nae1 cKO mice but not in WT littermate control at P25, a common sign of
   peripheral neuropathy in mice. (D to F) P25 Nae1 cKO show poor
   responses compared to control mice when (D) latency to fall off the
   accelerating rotarod, (E) grip strength of all four limbs or forelimbs,
   and (F) nociceptive responses using the hot-plate test were measured.
   Data are presented as means ± SEM. Each data point represents an
   individual animal. Two-tailed unpaired Student’s t test. (G to K) Nae1
   mutant mice display a marked reduction in conduction velocity. (G)
   Electrophysiological recording of CMAPs from sciatic nerves of Nae1 cKO
   and control mice at P25. Representative traces are shown. S, stimulus;
   R, initiation of CMAP response (red arrows). Graphs show (H) nerve
   conduction velocities, (I) latency, (J) mean peak amplitudes of CMAPs,
   and (K) average durations of CMAPs in sciatic nerves of control,
   heterozygotes, and Nae1 cKO mice at P25. Data are presented as means ±
   SEM. WT (n = 8), HT (n = 6), and cKO (n = 7). One-way ANOVA with
   Tukey’s multiple comparisons test.

   Nae1 cKO mice were born at the expected Mendelian ratio and were
   essentially indistinguishable from littermate controls initially.
   However, Nae1 cKO mice died prematurely or were sacrificed for ethical
   reasons at around P30, although a small proportion could survive till
   P47 ([154]Fig. 2B). Starting from the second week, the mice began to
   develop severe tremor and unsteady gait (movie S1). Upon tail
   suspension, they showed abnormal hindlimb clasping ([155]Fig. 2C) and
   performed poorly in motor and sensory functional tests ([156]Fig. 2, D
   to F).

   Electrophysiological recordings showed a marked reduction in conduction
   velocity in mutant mice ([157]Fig. 2, G and H). The latency and mean
   peak amplitudes and duration of compound muscle action potentials
   (CMAPs) were also severely affected in Nae1 cKO mice ([158]Fig. 2, I to
   K), clearly showing marked conduction blocks, accounting for the defect
   in motor function in these mice. No differences were seen between
   control mice and heterozygous mice, although heterozygous mice
   expressed lower levels of NAE1 ([159]Fig. 2A).

Neddylation is essential for Schwann cell myelination and axonal health

   Our results on motor function and nerve conduction velocity in Nae1 cKO
   suggested major defects in myelination in these mice. To confirm this,
   we performed ultrastructural examination of developing sciatic nerves
   in Nae1 cKO mice by TEM. We did not find any differences in nerve area
   (fig. S2D) but instead found that the Nae1-deficient mice failed to
   myelinate with a notable reduction in number of myelin profiles at all
   ages examined ([160]Fig. 3, A and B). This absence of myelinating
   Schwann cells in the Nae1 cKO mice was accompanied by a robust
   down-regulation of structural myelin proteins CNP, MPZ, and MBP in P28
   nerves ([161]Fig. 3E and fig. S2E).

Fig. 3. NAE1 is essential for Schwann cell myelination.

   [162]Fig. 3.
   [163]Open in a new tab

   (A) Representative EM pictures showing ultrastructure of control and
   Nae1 cKO sciatic nerves, at indicated ages. Arrowheads indicate
   promyelin Schwann cells. (B to D) Graphs show quantification of (B)
   myelinated axons, (C) amyelinated axons (1:1) per nerve, and (D)
   proportion of Schwann cells in a 1:1 relationship with an axon that are
   myelinated or remain unmyelinated, in control and Nae1 cKO mice at
   indicated ages. Data are presented as means ± SEM; n = 3 to 6. Two-way
   ANOVA with Sidak’s multiple comparisons test. (E) Immunoblot analyses
   of myelin proteins (MPZ, CNP, and MBP) in total sciatic nerve lysates
   from control, heterozygote, and Nae1 cKO mice at P28. (F)
   Representative EM pictures showing axonal swellings with organelle
   accumulations, a typical sign of axonal pathology ([164]22). Scale bar,
   1 μm. Graphs shows the quantification of these abnormal axon profiles
   per nerve in control and Nae1 cKO mice at indicated ages. (G and H)
   Graph shows quantification of (G) axons in a 1:1 relationship with a
   Schwann cell, and (H) total axons per nerve in control and Nae1 cKO
   mice at indicated ages. Data are presented as means ± SEM; n = 3 to 6.
   Two-way ANOVA with Sidak’s multiple comparisons test. (I) Immunoblot
   analyses of neurofilament (NF) in total sciatic nerve lysates from
   control and Nae1 cKO mice at P30. Graph shows densitometric
   quantification. Data are presented as means ± SEM; n = 4 to 5.
   Two-tailed unpaired Student’s t test.

   A failure in radial sorting can lead to myelination defects ([165]20),
   but this was not predominantly the case in the Nae1 cKO mice, because
   electron microscopy (EM) analyses showed that most of the large
   diameter axons had segregated from the Schwann cell families and were
   in a 1:1 relationship with Schwann cells ([166]Fig. 3, A and C).
   Instead, we found that myelination was blocked at the pro-myelin stage,
   with these cells failing to elaborate myelin sheets ([167]Fig. 3, A and
   D). While, in P28 wild-type (WT) nerves, almost all sorted axons were
   myelinated, in the Nae1 cKO nerves, a large proportion of large
   diameter axons in a 1:1 relationship with a Schwann cell had still not
   elaborated myelin sheets. Notably, we found that this difference was
   maintained in P47 nerves showing that the myelination defects in these
   mice was due to a failure of pro-myelin Schwann cells to form myelin
   sheets, instead of a delay in development.

   Although radial sorting seemed to proceed normally in general, the
   Remak bundles in the Nae1 cKO mice were also affected. We found that
   the Remak bundles in the Nae1 cKO mice were in general larger in size
   than WT mice, with more axons associated with a single Remak Schwann
   cell (up to 150 in some cases) (fig. S2, F and G). Notably, we also
   found that about 30% of Remak bundles contained one or more large
   diameter axon (>1 μm), especially in P28 Nae1 cKO nerves, whereas these
   were rarely observed in WT nerves (fig. S2H). Furthermore, we also
   found rare events of actively demyelinating Schwann cells in mutant
   nerves (fig. S2I), which possibly represent dysmyelinating profiles of
   abnormally myelinated Schwann cells (fig. S2J), especially in older
   nerves (P28 and P47). The abnormal myelin profiles included one single
   Schwann cell myelinating two axons, similar to mice models with Schwann
   cell–specific ablation of Fbxw7 ([168]14), an E3 ubiquitin ligase that
   forms part of CRL complexes involved in ubiquitination and degradation
   of different substrates, including c-Jun ([169]21). These data
   demonstrate that neddylation is essential for Schwann cell
   differentiation in perinatal nerves, with a major role in driving
   myelination.

   Our electrophysiological recordings above on CMAP amplitude and
   duration ([170]Fig. 2, J and K) suggested an axonal pathology in Nae1
   cKO mice. EM analyses supported this because we found abnormal axonal
   profiles (axonal swellings with organelle accumulations) ([171]22) in
   Nae1 cKO mice at all ages examined ([172]Fig. 3F), fewer large diameter
   sorted axons (myelinated or amyelinated) ([173]Fig. 3G), and a tendency
   toward a reduction in the total number of axons ([174]Fig. 3H) in P28
   nerves. Immunoblot analyses also demonstrated a decrease in
   neurofilament protein levels ([175]Fig. 3I). These data suggest that
   the Schwann cell defects in the Nae1 cKO mice could lead to a
   small-scale secondary axonal degeneration in these mice. Using
   immunohistochemical analyses of sciatic nerves, we also found an
   invasion of Iba1^+ve macrophages in Nae1 cKO nerves, likely in response
   to the axonal and myelin degeneration observed (fig. S2K). In summary,
   our data show that neddylation is essential for peripheral nerve
   development, including the differentiation of myelinating Schwann cells
   and for Schwann cell–mediated promotion of axonal health.

Neddylation is dispensable for maintenance of mature myelin sheaths

   Next, we examined the role of neddylation in mature myelinating Schwann
   cells in adult nerves and, thus, generated an inducible cKO model that
   allowed us to circumvent the problems associated with lethality in our
   constitutive model. For this, we used a tamoxifen-inducible Plp-cre^ERT
   driver line that expresses Cre recombinase under the control of Plp
   promoter following tamoxifen induction ([176]23). This mouse line was
   crossed with Nae1 floxed mice to generate the inducible Nae1 cKO mice
   (Nae1 icKO) and, at 60 days of age, was administered tamoxifen,
   followed by analyses at P100 (fig. S3A). Using RT-qPCR analyses, we
   found that tamoxifen was able to induce highly efficient recombination
   of the Nae1 allele in sciatic nerves (fig. S3B).

   Upon tail suspension, Nae1 icKO mice showed normal hindlimb clasping
   (fig. S3C) and did not show major differences in motor and sensory
   functional tests (fig. S3D) or in electrophysiological recordings (fig.
   S3, E and F). EM analyses of sciatic nerves showed normal myelin
   profiles (fig. S3G) and myelin thickness, as measured by G-ratio
   analyses (fig. S3H). Immunoblot analyses also showed no major
   differences in myelin protein levels in Nae1 icKO nerves in vivo (fig.
   S3I). Similar results were obtained in vitro. Primary Schwann cells
   were cultured for 48 hours with db cAMP to elevate MPZ levels and then
   treated for a further 48 hours with MLN4924. We found that MLN4924 was
   effective in reducing neddylated proteins but had no effect on MPZ
   levels (fig. S3J).

   These data show that neddylation is important for initiating or priming
   Schwann cell myelination but not for the maintenance of myelin sheaths.
   This coincides with the elevated levels of neddylated proteins at early
   stages of Schwann cell development and a progressive decrease at later
   stages when the myelinated Schwann cells have fully differentiated
   ([177]Fig. 1B).

Transcriptome profiling shows profound alterations in key differentiation
pathways

   To define the mechanisms by which neddylation exerts such profound
   effects on Schwann cell myelination, we performed RNA-seq profiling of
   P7 control and Nae1 cKO sciatic nerves. We identified distinct
   transcriptomic profiles between the control and mutant nerves, and
   differential expression analysis revealed notable changes in the
   transcriptomic profile. We identified 1389 dysregulated transcripts in
   mutant nerves, 834 of which were increased and 555 of which were
   decreased at a 5% false discovery rate (FDR) and fold change (FC) of 2
   ([178]Fig. 4A).

Fig. 4. Neddylation regulates the Schwann cell differentiation program.

   [179]Fig. 4.
   [180]Open in a new tab

   (A) Volcano plot of transcriptome profiles between control and Nae1 cKO
   P7 sciatic nerves (n = 4). Red and blue dots represent genes
   significantly down-regulated and up-regulated, respectively, in Nae1
   cKO mice [fold change (FC) > 2, adjusted P value < 0.05]. (B) Gene
   Ontology (GO) analysis of up-regulated genes in Nae1 cKO sciatic nerves
   compared to control nerves. Each dot in the connecting lines represents
   the gene count of the corresponding biological function categories. (C)
   Immunofluorescence labeling for proliferative EdU^+ cells (red) in
   sciatic nerves from control and Nae1 cKO at P7. Nuclei were
   counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (blue). Scale
   bar, 100 μm. Graphs shows quantification of EdU^+ cells in sciatic
   nerves of control and Nae1 cKO mice at P1, P7, P15, and P30 (n = 3
   animals per genotype). Data are presented as means ± SEM. Two-way ANOVA
   with Sidak’s multiple comparisons test (between groups). (D) GO
   analysis of down-regulated genes in Nae1 cKO sciatic nerves compared to
   control nerves. Each dot in the connecting lines represents the gene
   count of the corresponding biological function categories. (E)
   ToppCluster plot showing the functional networks among the genes
   down-regulated in Nae1 cKO sciatic nerves.

   Gene Ontology (GO) analysis indicated that there was an enrichment of
   terms associated with regulation of cell proliferation in genes
   up-regulated in Nae1 cKO mice ([181]Fig. 4B). Accordingly, Schwann cell
   proliferation rates, as determined by 5-ethynyl-2′-deoxyuridine (EdU)
   incorporation, showed a general increase in perinatal Nae1 mutant
   sciatic nerves ([182]Fig. 4C). In contrast, GO terms associated with
   the myelination process were enriched in down-regulated genes, as
   expected ([183]Fig. 4D). Furthermore, gene enrichment analysis by
   ToppCluster suite confirmed that the functions of down-regulated genes
   were particularly pertinent to myelin lipid biosynthesis and axonal
   function and additionally identified two key signals that regulate
   Schwann cell myelination: the master myelination transcription factor
   EGR2 and the mTOR pathway ([184]Fig. 4E).

   mTOR, a central signaling hub coordinating cell metabolism, plays a
   complex role in myelinating Schwann cells ([185]24). Before the onset
   of myelination, mTOR suppresses the transition from promyelinating to
   myelinating Schwann cells by transiently blocking expression of
   Krox-20. As the cells differentiate, a decline in mammalian target of
   rapamycin complex 1 (mTORC1) activity releases this block and allows
   myelination to proceed. Genetic mTOR hyperactivation in developing
   Schwann cells in vivo, mediated by Schwann cell–specific ablation of
   the negative regulators of mTOR, TSC1 ([186]25), and Pten ([187]25),
   leads to a block in Schwann cell myelination. Gene set enrichment
   analysis (GSEA) of the transcriptomic profile of Nae1 mutant mice
   showed a close overlap with the gene signatures of peripheral nerves
   from TSC1 ([188]25) and Pten ([189]25) mutant mice, suggesting mTOR
   hyperactivation in Nae1 cKO nerves ([190]Fig. 5A). Immunoblot analysis
   of sciatic nerve lysates from Nae1 cKO mice demonstrated an increase in
   phosphorylation of two well-established mTORC1 targets, S6K and 4EBP1,
   confirming mTORC1 hyperactivation in mutant Schwann cells ([191]Fig.
   5B).

Fig. 5. Neddylation inhibition leads to mTOR hyperactivation in Schwann
cells.

   [192]Fig. 5.
   [193]Open in a new tab

   (A) GSEA plot showing enrichment of TSC1-regulated and PTEN-regulated
   Schwann cell gene signature in Nae1 cKO mice. NES, normalized
   enrichment score. (B) Immunoblot analyses and densitometric
   quantification of mTOR pathway components in sciatic nerve lysates from
   control and Nae1 cKO mice at P28. Data are presented as means ± SEM; n
   = 4 to 7. Two-tailed unpaired Student’s t test, *P < 0.05, **P < 0.01,
   ****P < 0.0001. (C and D) Immunoblot analyses (C) and densitometric
   quantification (D) of MPZ and p-4E-BP1 levels in sciatic nerve lysates
   from control and Nae1 cKO mice, treated with vehicle or rapamycin. Data
   are presented as means ± SEM. Two-way ANOVA with Sidak’s multiple
   comparisons test. (E and F) Representative EM micrographs (E) and
   quantification of myelinated and amyelinated (1:1) axons per nerve (F)
   of P10 sciatic nerves from control and Nae1 cKO mice treated with
   vehicle or rapamycin. Scale bar, 2 μm. Data are presented as means ±
   SEM; n = 3 to 4. Two-way ANOVA with Sidak’s multiple comparisons test.
   (G) Immunoblot analyses of mTOR pathway components in primary rat
   Schwann cells cultured under basal or myelinogenic conditions (db cAMP
   treatment), in the presence or absence of MLN4924. (H) Immunoblot
   analyses showing that inhibiting the hyperactivation of mTOR pathway
   with rapamycin does not rescue the MLN4924-induced suppression of MPZ
   in primary rat Schwann cells cultured under myelinogenic conditions (db
   cAMP treatment). β-Actin is used as loading control for immunoblots in
   (B), (C), (G), and (H).

   The myelin defects seen in mice with artificial mTOR hyperactivation,
   by conditionally ablating TSC1 ([194]25) or TSC2 ([195]26) in Schwann
   cells, can be reversed after treatment with mTORC1-inhibiting drug
   rapamycin in vivo. Here, using a similar approach, we examined whether
   the myelin defects observed in Nae1 cKO mice could also be reversed by
   suppressing mTOR hyperactivation using rapamycin. For this, P5 WT and
   Nae1 cKO mice were treated with rapamycin for 5 days, and, then,
   immunoblot and EM analysis of sciatic nerves were performed. Rapamycin
   was effective in suppressing the elevated levels of eukaryotic
   translation initiation factor 4E-binding protein 1 (4E-BP1)
   phosphorylation in the mutant nerves, although no recovery of MPZ
   levels was seen ([196]Fig. 5, C and D). In line with this, EM analyses
   also showed that rapamycin had no effect on the myelin pathology in
   Nae1 cKO mice ([197]Fig. 5, E and F). We confirmed these findings using
   primary Schwann cell cultures. We found that, similar to mutant mice,
   neddylation inhibition using MLN2924 led to mTOR hyperactivation in
   cells, both in basal and myelinogenic conditions ([198]Fig. 5G) and
   that rapamycin supplementation failed to restore MPZ levels in
   MLN4924-treated cells cultured under myelinogenic conditions ([199]Fig.
   5H).

   These data show that, although Nae1 ablation in Schwann cells induces
   mTOR hyperactivation, this is not sufficient to drive the nerve
   pathology in the mutant mice. We thus examined another key pathway
   involved in Schwann cell myelination, Hippo-YAP signaling
   ([200]27–[201]29), which has previously been shown to be regulated by
   neddylation in cardiomyocytes ([202]30). In this pathway, activation of
   Hippo kinases mammalian STE20 (Mst1/2), large tumor suppressor kinase
   1/2 (Lats1/2), and Mps one binder kinase activator 1A (MOB1) induces
   phosphorylation of YAP, which leads to its cytoplasmic sequestration
   and inhibition of its nuclear translocation, thus inhibiting its
   transcriptional activity ([203]30). Here, immunoblots analysis of
   sciatic nerve lysates from Nae1 cKO mice demonstrated an increase in
   Hippo kinases (MST1, LATS1, and LATS2) and phosphorylation of LATS1/2
   and MOB1. This was accompanied by an increase in phosphorylation of YAP
   ([204]Fig. 6A). Furthermore, RT-qPCR analysis showed that several
   direct YAP targets in Schwann cells ([205]27) were down-regulated in
   P30 sciatic nerves from Nae1 cKO mice, indicating an inactivation of
   YAP signaling ([206]Fig. 6B).

Fig. 6. Neddylation inhibition leads to suppression of the YAP/TAZ pathway in
Schwann cells.

   [207]Fig. 6.
   [208]Open in a new tab

   (A) Immunoblot analyses and densitometric quantification of Hippo-Yap
   signaling pathway components in sciatic nerve lysates from control and
   Nae1 cKO mice at P28. Data are presented as means ± SEM; n = 3 to 8, *P
   < 0.05, **P < 0.01, ****P < 0.0001. Two-tailed unpaired Student’s t
   test. (B) RT-qPCR showing regulation of YAP/TAZ target genes in sciatic
   nerves from Nae1 cKO mice at P28. Data are presented as means ± SEM; n
   = 4 to 12. Two-tailed unpaired Student’s t test. (C and D) Immunoblot
   analyses (C) and densitometric quantification (D) of MPZ and p-YAP
   levels in sciatic nerve lysates from control and Nae1 cKO mice, treated
   with vehicle or XMU-MP-1. Data are presented as means ± SEM. Two-way
   ANOVA with Sidak’s multiple comparisons test. (E and F) Representative
   EM micrographs (E) and quantification of myelinated and amyelinated
   axons (1:1) per nerve (F) of P10 sciatic nerves from control and Nae1
   cKO mice treated with vehicle or XMU-MP-1. Scale bar, 2 μm. Data are
   presented as means ± SEM; n = 3 to 4. Two-way ANOVA with Sidak’s
   multiple comparisons test (between groups). (G) Immunoblot analyses of
   YAP/TAZ pathway components in primary rat Schwann cells cultured under
   basal or myelinogenic conditions (db cAMP treatment), in the presence
   or absence of MLN4924. β-Actin is used as loading control for
   immunoblots. (H) Immunoblot analyses showing that treatment with the
   MST1/2 inhibitor XMU-MP-1 does not rescue the MLN4924-induced
   suppression of MPZ in primary rat Schwann cells cultured under
   myelinogenic conditions (db cAMP treatment). β-Actin is used as loading
   control for immunoblots in (A), (C), (G), and (H).

   Next, we performed rescue experiments using XMU-MP-1, a reversible and
   selective MST1/2 inhibitor ([209]31), similar to rapamycin above, to
   determine whether inhibiting Hippo signaling could rescue the
   dysmyelination in Nae1 cKO mice. We did not observe any recovery in MPZ
   levels ([210]Fig. 6, C and D) or myelinated axons ([211]Fig. 6, E and
   F) after XMU-MP-1 treatment in mutant nerves. In primary Schwann cell
   cultures also, we found that neddylation inhibition with MLN4924 leads
   to an enhanced phosphorylation of YAP and MOB1 both under basal and
   myelinogenic conditions ([212]Fig. 6G), and treatment with XMU-MP-1,
   which very effectively reduced this enhanced phosphorylation, was
   insufficient to recover MPZ levels to the normal levels seen in db
   cAMP–treated cultures alone ([213]Fig. 6H).

   Together, these data show that neddylation fine-tunes mTOR and YAP
   activation in developing Schwann cells, two major pathways involved in
   regulating Schwann cell myelination, although the failure of our rescue
   experiments to reverse the block in myelination suggests a more robust
   role of neddylation in controlling myelination at additional levels.

Neddylation controls the protein levels of the negative regulators c-Jun and
Sox2

   Thus, to gain further insight of the impact of neddylation in Schwann
   cell myelination, we performed proteomic analysis of control and Nae1
   cKO Schwann cells by liquid chromatography–tandem mass spectrometry. We
   identified 285 up-regulated and 126 down-regulated proteins (5% FDR and
   FC of 2) of a total of 5189 proteins identified in Nae1 cKO mice
   ([214]Fig. 7A). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways
   analysis revealed an enrichment of a variety of categories, including
   metabolic pathways, cell proliferation, and phosphatidylinositol
   3-kinase–Akt signaling ([215]Fig. 7B), in line with our transcriptomic
   data ([216]Fig. 4). Notably, ubiquitin-mediated proteolysis was one of
   the major categories associated. Among the proteins identified, as
   expected, NAE1 levels were reduced in mutant nerves, but there was an
   up-regulation of several proteins related to protein proteolysis
   ([217]Fig. 7C). Further KEGG analysis showed that, of all the different
   types of E3 ligases responsible for protein degradation, CRLs were the
   most enriched category associated with the dysregulated proteins (fig.
   S4, A and B). Two of these, FBXO38 and DCAF8, are encoded by genes
   containing causative mutations associated with CMT ([218]Fig. 1A)
   ([219]3). These data suggest that CRLs could be involved in protein
   degradation in developing nerves.

Fig. 7. Proteomics analyses show compromised ubiquitin-mediated proteolysis
in Nae1 cKO mice.

   [220]Fig. 7.
   [221]Open in a new tab

   (A) Volcano plot of proteomic analysis in control and Nae1 cKO P7
   sciatic nerves (n = 5). Red and blue dots represent significantly
   down-regulated and up-regulated proteins, respectively, in Nae1 cKO
   mice (FC > 2, adjusted P value < 0.05). (B) Top enriched Kyoto
   Encyclopedia of Genes and Genomes (KEGG) terms of dysregulated
   proteins. PI3K, phosphatidylinositol 3-kinase. (C) Z-score polar plot
   of deregulated proteins associated with ubiquitination and neddylation.
   Green represents down-regulated proteins, and red represents
   up-regulated proteins. (D) Immunoblot analyses of c-Jun and Sox2 levels
   in sciatic nerve lysates from P28 Nae1 cKO mice. (E) RT-qPCR of
   c-Jun–regulated Schwann cell repair genes in sciatic nerves from P28
   Nae1 cKO mice. Data are presented as means ± SEM; n = 4 to 6.
   Two-tailed unpaired Student’s t test. (F) Immunoblot analyses of c-Jun
   and Sox2 levels in primary rat Schwann cells. (G and H) RT-qPCR showing
   expression of (G) c-Jun and (H) Sox2 mRNA levels in primary Schwann
   cells. Data are presented as means ± SEM; n = 3. Two-way ANOVA with
   Sidak’s multiple comparisons test. (I and J) Immunoblot and
   densitometric quantification of (I) c-Jun and (J) Sox2 levels in
   primary Schwann cells. Data are presented as means ± SEM; n = 3.
   Two-way ANOVA with Sidak’s multiple comparisons test. Dotted lines
   represent nonlinear regression analysis (one-phase decay) of c-Jun and
   Sox2 degradation over time after db cAMP treatment. *P < 0.05; **P <
   0.01; ****P < 0.0001. (K and L) Immunoblot and densitometric
   quantification of (K) c-Jun and (L) Sox2 levels in primary Schwann
   cells, treated with cycloheximide (CHX) for the indicated times. Data
   are presented as means ± SEM; n = 8. Two-way ANOVA with Sidak’s
   multiple comparisons test. Dotted lines represent nonlinear regression
   analysis (one-phase decay) of c-Jun and Sox2 degradation over time
   after CHX treatment. (M and N) Immunoblot analyses of MPZ levels after
   (M) c-Jun and (N) Sox2 silencing. β-Actin is used as loading control
   for immunoblots in (D), (F), (H), (I), and (K) to (N).

   The activation of the myelination program in Schwann cells is dependent
   on the down-regulation of key negative regulators of myelination,
   including c-Jun, Sox2, and Notch, and enforced expression of these
   regulators in developing Schwann cells in vivo can inhibit myelination
   ([222]32–[223]34). Because CRLs are estimated to be responsible for
   about 20% of protein turnover in cells and inhibition of their activity
   leads to an accumulation of their target substrates ([224]5, [225]30,
   [226]35), we examined whether they could be responsible for the
   breakdown of these negative regulators. Immunoblot analyses showed
   elevated levels of c-Jun and Sox2 in P28 sciatic nerves from Nae1 cKO
   mice ([227]Fig. 7D and fig. S5A). We found that this effect could be
   observed already in perinatal nerves (fig. S5, B and C). While the
   levels of these regulators were down-regulated in WT mice as from P7
   days of age, their levels were maintained at high levels in Nae1 cKO
   mice. These elevated c-Jun levels were accompanied by an increase in
   expression of c-Jun–regulated target genes ([228]36, [229]37) in nerves
   from Nae1 mutant mice ([230]Fig. 7E).

   Similarly, in primary Schwann cell cultures, neddylation inhibition by
   MLN4924 led to an up-regulation of c-Jun and Sox2 levels under basal
   conditions and could strongly block db cAMP–induced down-regulation of
   c-Jun and Sox2 ([231]Fig. 7F). To further investigate this, we examined
   the effects of db cAMP treatment on both mRNA and protein levels of
   c-Jun and Sox2 in primary Schwann cell cultures over time. In control
   conditions, we found that c-Jun mRNA levels were markedly decreased
   after only 1 hour following db cAMP treatment, while protein levels
   decreased more gradually, requiring up to 8 hours of treatment to
   decrease. Notably, we found that MLN4924 treatment blocked the decrease
   in c-Jun protein levels but did not have any effect on c-Jun mRNA
   levels ([232]Fig. 7, G and I). Similar effects were observed for Sox2
   levels, although, in this case, the effects were retarded. Sox2 mRNA
   levels required about 12 hours to decrease to low levels, whereas
   protein levels required about 24 hours. Similar to c-Jun, MLN4924
   treatment blocked the decrease in Sox2 protein levels, while having no
   effect on Sox2 mRNA levels ([233]Fig. 7, H and J). We further
   challenged this system by treating Schwann cells with MLN4924, 24 hours
   after db cAMP supplementation, when both mRNA and protein levels of
   c-Jun and Sox2 are at the minimum levels. We found that vehicle
   treatment did not alter their protein or mRNA levels. In contrast and
   quite notably, we found that MLN4924 treatment was able to
   progressively increase c-Jun and Sox2 protein levels with time, while
   having no effect on their mRNA levels (fig. S5, D to G).

   These results above show that neddylation can regulate c-Jun and Sox2
   protein expression, at the level of either translation or degradation,
   without affecting their transcription. To distinguish between these two
   possibilities, we performed a cycloheximide (CHX)–based pulse-chase
   assay, which permits visualization of the degradation kinetics of the
   steady state population of a variety of cellular proteins. Here, we
   found that, after blocking protein translation using CHX, c-Jun and
   Sox2 levels rapidly decreased over the next few hours as they are
   degraded. MLN4924 treatment, however, was able to significantly
   increase the stability of c-Jun protein ([234]Fig. 7K). For Sox2, on
   the other hand, MLN4924 had no effect on its half-life after CHX
   treatment, suggesting that neddylation controls Sox2 protein levels at
   the translational level rather than its degradation, as for c-Jun
   ([235]Fig. 7L). Last, we examined whether reducing the elevated c-Jun
   or Sox2 levels could rescue the MPZ levels in Schwann cell cultures
   after neddylation inhibition. We found that neither c-Jun nor Sox2
   silencing using lentiviral vectors could effectively restore the
   MLN4924-induced down-regulation of MPZ after db cAMP treatment
   ([236]Fig. 7, M and N). These results further demonstrate that
   neddylation plays a pleiotropic role in myelination, controlling both
   positive and negative regulators.

Neddylation stabilizes EGR2 protein and antagonizes its ubiquitination and
degradation

   Our transcriptomic and GO analyses of Nae1 cKO mice above showed an
   enrichment of categories associated with regulation by the myelination
   master transcription factor EGR2 ([237]Fig. 4E) ([238]1, [239]2).
   Further GSEA analyses showed that EGR2-regulated gene signatures were
   enriched in WT nerves compared to that in Nae1 cKO nerves ([240]Fig.
   8A). Similarly, we found a close overlap between GSEA profiles from
   Nae1 cKO mice and mutant mice lacking Zeb2, another essential
   transcription factor (TF) regulating Schwann cell myelination
   ([241]Fig. 8B) ([242]38, [243]39). Immunoblot analyses showed a strong
   down-regulation of both EGR2 and ZEB2 in sciatic nerves from Nae1 cKO
   mice ([244]Fig. 8, C and D). Furthermore, we found that, whereas
   MLN4924 had no effect on EGR2 and ZEB2 levels in primary Schwann cells
   cultured under basal conditions, neddylation inhibition was effective
   in reducing db cAMP–mediated up-regulation of EGR2 levels but not of
   ZEB2 ([245]Fig. 8E). We also found strong suppression of EGR2 levels in
   nuclear extracts of MLN4924-treated cultures under myelinogenic
   conditions ([246]Fig. 8F). These data suggest that neddylation could be
   regulating EGR2-mediated effects on Schwann cell myelination.

Fig. 8. Neddylation stabilizes EGR2 protein.

   [247]Fig. 8.
   [248]Open in a new tab

   (A and B) GSEA plots showing suppression of (A) EGR2-regulated and (B)
   Zeb2-regulated Schwann cell gene signature ([249]38) in Nae1 cKO mice.
   (C and D) Immunoblots (C) and densitometric quantification (D) of EGR2
   and ZEB2 in sciatic nerve lysates from P28 WT, HET, and Nae1 cKO mice.
   Data are presented as means ± SEM; n = 5 to 11. One-way ANOVA with
   Tukey’s multiple comparisons test. (E and F) Immunoblot analyses of
   EGR2 and ZEB2 levels in primary rat Schwann cells in (E) whole-cell
   lysates, or (F) cytosolic and nuclear extracts. (G) Graph showing
   positive correlation between EGR2 and NEDD8 protein levels in WT,
   heterozygote, and Nae1 mutant mice (two-sided Pearson’s correlation).
   The dashed line shows the regression line, and the shaded area
   represents the 95% confidence interval. (H) Immunoblot analysis of EGR2
   and ZEB2 levels in HALO-NEDP1 pulldowns and whole-cell lysates (input)
   from primary Schwann cells cultured under myelinogenic conditions (db
   cAMP), in the absence or presence of MLN4924. (I and J)
   Co-immunoprecipitation of (I) endogenous NEDD8 and (J) endogenous EGR2
   from primary Schwann cells. Immunoglobulin G (IgG) heavy chain is
   denoted by an asterisk. (K and L) Immunoblot and densitometric analyses
   of (K) EGR2 and (L) ZEB2 levels in primary Schwann cells, cultured in
   the presence or absence of MLN4924, and treated with CHX (5 μm) for the
   indicated times. Data are presented as means ± SEM; n = 3. Two-way
   ANOVA with Sidak’s multiple comparisons test. Dotted lines represent
   nonlinear regression analysis (one-phase decay) of EGR2 and ZEB2
   degradation over time after CHX treatment. *P < 0.05; ****P < 0.0001.
   (M) Co-immunoprecipitation of endogenous EGR2 from primary Schwann
   cells, followed by immunoblot analysis of whole-cell lysates and
   immunoprecipitates with anti-ubiquitin antibody. (N) Immunoblot
   analyses of MPZ levels in db cAMP and MLN4924-treated primary Schwann
   cells after treatment with MG132. β-Actin is used as loading control
   for immunoblots in (C), (E), (F), and (H) to (N).

   Our immunoblot analyses showed that EGR2 was expressed at intermediate
   levels in heterozygous mice compared to WT or cKO, whereas this was not
   evident for ZEB2 ([250]Fig. 8, C and D). Furthermore, we found a
   positive correlation between NEDD8 and EGR2 protein levels from
   immunoblot analyses of sciatic nerves from WT, heterozygotes, or cKO
   mice ([251]Fig. 8G). These intriguing results of intermediate EGR2
   levels in Nae1 heterozygous mice suggested that EGR2 could be directly
   neddylated. To demonstrate this, we used a NEDD8 affinity resin
   pulldown assay, in which the HALO protein is fused with the
   catalytically active form of NEDP1 (HALO-NEDP1 CA), which can
   efficiently enrich for neddylated proteins in cellular extracts
   ([252]40). When extracts obtained from Schwann cell cultures were
   exposed to this resin, immunoblot assays showed substantial enrichment
   of EGR2 under myelinogenic conditions ([253]Fig. 8H). Notably, this
   enrichment was diminished when a NEDD8 nonbinder mutant with reduced
   binding capacity [HALO-DEP1 DAGC (D29W A98K G99K C163A)] was used
   instead of the HALO-NEDP1 CA resin. Furthermore, treatment with MLN4924
   efficiently abrogated the enrichment of EGR2 and neddylated proteins in
   both the HALO-NEDP1 CA and HALO-DEP1 DAGC resins, indicating the
   specificity of capture of neddylated EGR2. In contrast, no enrichment
   was obtained for ZEB2.

   To further demonstrate this, we performed co-immunoprecipitation assays
   with anti-NEDD8 antibodies and immunoblotted for EGR2. Similar to the
   previous experiment, we found strong enrichment of EGR2 in db
   cAMP–treated cultures, which was reduced in MLN2924-treated cells
   ([254]Fig. 8I). Last, we then immunoprecipitated EGR2 from primary
   Schwann cell extracts instead and immunoblotted for NEDD8. We found
   that EGR2 was neddylated in Schwann cells after db cAMP treatment and
   was abolished in cultures treated with MLN4924 ([255]Fig. 8J),
   demonstrating that EGR2 could be possibly directly neddylated.

   Next, to examine the functional consequence of EGR2 neddylation, we
   examined the stability of EGR2 protein. Primary Schwann cells were
   cultured for 48 hours under myelinogenic conditions (db cAMP treatment)
   to elevate levels of EGR2 and ZEB2 and then treated with CHX (5 μm) for
   the indicated times, in the absence or presence of MLN4924, and
   immunoblot analyses of EGR2 and ZEB2 were performed. We found that
   MLN4924 strongly reduced the stability of EGR2. No such effect was seen
   for ZEB2 ([256]Fig. 8, K and L). Immunoprecipitation assays further
   showed that neddylation inhibition by MLN4924 robustly increased the
   polyubiquitination of EGR2 in db cAMP–treated Schwann cells ([257]Fig.
   8M). This was observed both under basal conditions or after inhibition
   of the proteasome using MG132, which leads to an accumulation of
   ubiquitinated proteins. Last, we examined whether blocking the
   MLN4924-induced degradation of EGR2 using MG132 could recover MPZ
   levels in Schwann cell cultures after db cAMP treatment. While MG132
   was efficient in recovering EGR2 levels, MPZ levels, however, could not
   be restored to the elevated levels seen with db cAMP treatment alone
   ([258]Fig. 8N). In summary, we demonstrate that, during Schwann cell
   myelination, EGR2, unlike ZEB2, could be a direct neddylation target
   and that this antagonizes its ubiquitination and degradation.

Neddylation regulates several nonredundant parallel myelination pathways

   Our data above show that neddylation is required for regulating several
   key pathways in Schwann cell myelination. However, all our experiments
   to rescue the myelination defects, induced by neddylation inhibition,
   for each of these individual pathways were unsuccessful. This suggested
   that targeting individual pathways in our rescue experiments could be
   having no effect on the other pathways.

   To show this, we examined the regulation of each of the molecular
   signals identified (mTOR, YAP, c-Jun, Sox2, and EGR2) in each rescue
   experiment. We found that rapamycin treatment, which strongly
   suppressed mTOR hyperactivation in Schwann cells ([259]Fig. 5H), had no
   effect on MLN4924-mediated up-regulation of c-Jun and Sox2, inhibition
   of the YAP pathway or down-regulation of EGR2 levels (fig. S6A).
   Similarly, we found that Hippo-YAP inhibition using XMU-MP-1 (fig.
   S6B), silencing of c-Jun (fig. S6C) or Sox2 (fig. S6D), and treatment
   with MG132 (fig. S6E) did not have, on the whole, any effect on the
   other pathways. These results indicate that neddylation is a master
   controller of Schwann cell myelination because it regulates multiple
   nonredundant positive and negative regulators that individually play an
   essential role in driving the myelination program.

Neddylation is required for axonal regeneration and remyelination after
injury

   Peripheral nerves have a notable regenerative potential. Nerve injury
   triggers the reprograming of mature Schwann cells to adopt a repair
   Schwann cell phenotype specialized to support regeneration. Repair
   Schwann cells provide metabolic and trophic support and serve as
   guidance cues for regrowing axons. Eventually, as axons reconnect to
   their targets, the Schwann cells re-differentiate and remyelinate
   regrown axons ([260]41).

   To examine whether neddylation was involved in nerve regeneration and
   remyelination, we performed sciatic nerve crushes in inducible
   conditional Nae1 knockout (Nae1 icKO) P60 mice, 40 days after the first
   tamoxifen injection, and monitored functional recovery ([261]Fig. 9A).
   Control mice groups functionally recovered at 40 days after crush, as
   measured by the sciatic functional index (SFI), whereas mutant mice
   remained severely impaired ([262]Fig. 9B). Electrophysiological
   recordings showed a marked reduction in conduction velocity, latency,
   and mean peak amplitudes and duration of CMAPs Nae1 icKO ([263]Fig. 9,
   C and D). When we analyzed remyelination by EM, we found that Nae1 icKO
   Schwann cells were blocked at the promyelinating stage and failed to
   remyelinate, as shown by a notable reduction in number of myelinated
   axons and concomitant elevated numbers of amyelinated axons (1:1),
   although the total number of Schwan cells were unchanged ([264]Fig. 9,
   E and F). Western blot analyses of regenerated nerves also showed
   decreased levels of the myelin proteins MBP and MPZ in mutant mice
   ([265]Fig. 9G and fig. S6F). Both EM counts of sorted axons ([266]Fig.
   9F) and Western blot analyses ([267]Fig. 9G and fig. S6F) showed fewer
   axons in regenerated nerves in Nae1 icKO mice. Similar to developmental
   myelination, neddylation inhibition also led to a decrease in EGR2
   levels, hyperactivation of mTOR pathway, suppression of YAP signaling,
   and up-regulation in c-Jun levels in regenerated mutant nerves
   ([268]Fig. 9G and fig. S6F). These data show that the onset of Schwann
   cell differentiation during both normal myelination and in
   remyelination following nerve injury is dependent on
   neddylation-mediated regulation of key molecular signals.

Fig. 9. Neddylation inhibition blocks peripheral nerve generation after
injury.

   [269]Fig. 9.
   [270]Open in a new tab

   (A) Schematic diagram depicting the strategy for nerve regeneration
   paradigm. Sixty-day-old Nae1 icKO mice were treated with tamoxifen,
   followed by crush injury 40 days later, and sacrifice at P140. (B)
   Quantification of sensory-motor function using sciatic functional index
   (SFI) in WT and Nae1 icKO mice. Mutants show a substantial block in
   regeneration, although mutant and WT SFIs are similar before and
   immediately after injury. Two-way ANOVA with Sidak’s multiple
   comparisons test (*P < 0.05; **P < 0.01; ****P < 0.0001). (C and D)
   Nae1 mutant mice display a marked reduction in conduction velocity. (C)
   Electrophysiological recording of CMAPs from regenerated sciatic nerves
   40 days after crush. Representative traces are shown. S, stimulus; R,
   initiation of CMAP response (red arrows). (D) Graphs show nerve
   conduction velocities, latency, mean peak amplitudes of CMAPs, and
   average durations of CMAPs. Data are presented as means ± SEM. WT (n =
   7) and icKO (n = 6). Two-tailed unpaired Student’s t test. (E)
   Representative EM pictures showing ultrastructure of regenerated
   sciatic nerves 40 days after crush. Scale bar, 5 μm. (F) Graphs show
   quantification of myelinated axons, amyelinated axons (1:1), Schwann
   cell nuclei and total axons per nerve. Two-tailed unpaired Student’s t
   test. (G) Immunoblot analyses of myelin proteins (MPZ and MBP), and
   axonal marker NF in regenerated sciatic nerves 40 days after crush. (H)
   Graphical abstract showing role of neddylation in regulating levels of
   key molecular signals during myelination.

DISCUSSION

   In this study, we demonstrate that neddylation is essential for Schwann
   cell myelination and axonal integrity both in developing and
   regenerating nerves. Genetic or pharmacological inhibition of
   neddylation leads to an arrest of Schwann cells at the pro-myelin
   stage, and they are not able to proceed with the myelination process.
   In these differentiating Schwann cells, neddylation is necessary for
   fine-tuning the expression/activity of multiple regulators that play a
   critical role in myelin differentiation.

   The profound defects in myelination led to severe gait abnormalities,
   muscle weakness, and hindlimb clasping, a typical presentation of
   neuromuscular dysfunction very early after birth. Consistent with the
   defective myelination in the Nae1 cKO mutants, the motor unit function
   of mutant mice was severely impaired, as reflected by a marked
   reduction in conduction velocity. Most mice did not survive past 3
   weeks of age. This could be possibly due to respiratory failure caused
   by a similar dysmyelination in phrenic nerves that innervate the
   diaphragm leading to its paralysis, as well as other contributory
   factors including weakness of the intercostal and other accessory
   muscles of respiration, a typical complication of the demyelinating
   disease Guillain-Barré syndrome ([271]42). In addition, we found that
   the defective Schwann cells cannot provide adequate support to maintain
   axonal integrity, similar to other models of metabolic dysfunction in
   Schwann cells ([272]43–[273]45), suggesting that neddylation induces a
   metabolic imbalance in Schwann cells that affects their axonal
   partners.

   The onset of myelination is a highly orchestrated process that requires
   a timely and precise coordination in expression and function of various
   molecular regulators, and studies in the past decades have shown that
   this program involves a balance between positive and negative
   regulators ([274]1, [275]2, [276]20, [277]46). We demonstrated that the
   profound nerve defects in the mutant mice were caused by the disruption
   of this molecular network ([278]Fig. 9H). We found that neddylation
   could directly regulate the activity or expression of several critical
   regulators of myelination that are essential components of this
   molecular network, including EGR2, which is considered the master
   transcription factor of Schwann cell myelination. EGR2, in cooperation
   with other TFs, drives the transcription of structural proteins and
   lipid synthesis genes needed for myelin formation, and its ablation in
   Schwann cells leads to severe peripheral myelination deficits,
   including an arrest at the pro-myelin stage ([279]47), similar to the
   Nae1 cKO mice. Here, we found EGR2 levels were decreased in mutant
   nerves, and our data indicate that EGR2 could be a direct neddylation
   target and that this antagonizes its ubiquitination and degradation to
   maintain high expression levels in myelinating Schwann cells. This
   uncovers a previously unidentified mechanism of regulation of
   expression of this critical factor in Schwann cells that complements
   the established transcriptional regulatory mechanisms, mediated by
   upstream factors, including Sox10, Oct6, or YY1 ([280]1, [281]2), and
   long noncoding RNAs ([282]48). Notably, we did not find any regulation
   in expression of Zeb2, another key transcriptional regulator of
   myelination ([283]38, [284]39), showing that the function of
   neddylation in regulating positive regulators of myelination is
   specific to some regulators only.

   As mentioned, several negative regulators, including c-Jun ([285]32),
   Sox2 ([286]33), and Notch ([287]34), have been identified that oppose
   this myelination program during development. These regulators are
   highly expressed in immature Schwann cells, and their levels gradually
   decrease as the cells undergo differentiation. Artificially preventing
   the decrease in the level of these negative regulators in vivo can
   induce a block in myelination, suggesting that their down-regulation is
   an essential prerequisite of the myelination program. Here, we found
   that neddylation inhibition maintains high protein levels of c-Jun and
   Sox2 in mature nerves, without affecting their transcription. In the
   case of Sox2, we showed that this was likely to be due to regulation at
   the translation level, whereas, in the case of c-Jun, neddylation was
   likely responsible for its proteasomal degradation.

   CRLs, which comprise the largest family of E3 ubiquitin ligases in
   eukaryotic cells, are responsible for ubiquitylation of ~20% of
   cellular proteins ([288]5, [289]49). At least three CRL–E3 ubiquitin
   ligase complexes target c-Jun for polyubiquitination and proteasomal
   degradation, including F-Box And WD Repeat Domain Containing 7 (FBXW7),
   COP1 E3 ubiquitin ligase (COP1), and cereblon (CRBN) in different
   cellular sytems ([290]21, [291]50, [292]51). FBXW7 was shown to
   specifically bind to the phosphorylated motifs (phosphodegron) of
   c-Jun, which is then targeted for proteasomal degradation ([293]21).
   Schwann cell–specific ablation of Fbxw7 in mice leads to elevated
   levels of c-Jun ([294]14), similar to our data on genetic or
   pharmacological inhibition of neddylation in Schwann cells. This
   suggests that c-Jun degradation could be regulated by a CRL-FBXW7 axis
   activated by neddylation, which needs to be determined in future
   studies together with a potential role for other E3 ubiquitin ligases
   such as COP1 and CRBN. Our proteomics data also revealed that several
   members of CRLs complexes, including Cullins and SRMs, were
   up-regulated in Nae1 cKO nerves. This is consistent with studies
   showing that neddylation inhibition traps CRLs in an inactive state,
   and, as a result, both the CRL substrates as well as SRMs, which are
   often auto-ubiquitinated by their respective CRL ([295]52–[296]54), are
   protected from degradation. This raises the possibility that within
   differentiating Schwann cells, these dysregulated CRL components could
   form part of different active CRL complexes that could be responsible
   for the turnover of many substrate proteins, in addition to c-Jun.

   In addition to the direct interference to key regulators, such as EGR2
   and c-Jun above, we also found that neddylation disruption led to
   global secondary changes in downstream pathways that include the mTOR
   and YAP/TAZ pathway. Neddylation inhibition led to an inactivation of
   the YAP pathway in vivo, by inducing an accumulation of the Hippo
   kinases LATS1 and MOB, possibly due to an impairment in their
   proteasomal degradation via CRLs, in line with previous studies showing
   that neddylation is essential for heart maturation through repression
   of Hippo signaling by CRL-mediated degradation of the Hippo kinases
   Mst1 and LATS1/2 ([297]30). The YAP/TAZ pathway has been previously
   shown to be critically important for Schwann cell myelination by
   several laboratories ([298]27–[299]29), and our studies show that
   neddylation could be important for regulating activation of the YAP/TAZ
   pathway during this process. Similarly, the mTOR pathway plays complex
   roles in the metabolic control of Schwann cell myelination. In immature
   Schwann cells, mTOR hyperactivation can lead to an arrest or delay in
   the onset of myelination, whereas, in myelinating Schwann cells, on the
   other hand, mTOR promotes the production and expansion of myelin
   sheaths ([300]24). Here, we show that nerves from Nae1 cKO mice show
   hyperactivation of the mTOR pathway, which coincided with a suppression
   of Krox20 levels. These results suggest that neddylation is required
   for suppression of mTOR activity at the onset of myelination, although
   the exact mechanisms still need to be uncovered.

   Unexpectedly, we were not able to correct the myelination defects after
   neddylation inhibition in the rescue experiments for either of the five
   deregulated pathways/signals that we studied. Each of these have been
   shown to individually have profound effects on myelination, affecting
   discrete aspects of this complex process. Here, by correcting one of
   these, the remaining four still remained dysregulated, likely
   explaining the lack of effect in these rescue experiments. These
   results also demonstrate the profound nerve defects observed in mutant
   mice after neddylation inhibition is because of this capacity of
   neddylation to regulate several essential nonredundant pathways
   simultaneously, ranking neddylation as one of the global master
   regulators of the molecular network that drives the myelination
   program. It is also likely that neddylation could be regulating the
   function of various other molecular regulators of Schwann cell
   myelination, including β-catenin ([301]55), HuR ([302]56), or Srebf1
   ([303]57) among others, which have been shown to be targets of
   neddylation in other cellular systems ([304]35, [305]58, [306]59).

   PTMs play an essential role in regulating the activity and viability of
   thousands of proteins in cells. They drive substantial changes in the
   cellular proteome in a short period of time, helping the cells to
   respond to different stimuli and environmental changes. We found that
   neddylation plays complex roles in Schwann cells by regulating multiple
   control mechanisms, including protein stability and degradation and
   possibly translation, through either possible direct PTM of substrates
   or through activation of CRL complexes, that ultimately affect a wide
   variety of critical myelination regulators. This places neddylation as
   central regulatory hub driving Schwann cell myelination. This work
   opens previously unidentified avenues to understand human nerve
   pathology, particularly inherited demyelinating disorders. There is a
   clear indication that mutations in several genes that encode critical
   components of the neddylation-regulated CRL complexes are associated
   with human nerve pathology ([307]3), and our study provides a platform
   to understand the pathogenetic mechanisms that could be involved in
   these cases.

MATERIALS AND METHODS

Animals

   To generate the constitutive Nae1 knockout mice (Nae1 cKO), mice
   homozygous for the floxed allele of Nae1 mice were crossed with an
   MPZ-cre line ([308]19), which expresses Cre recombinase specifically in
   Schwann cells from embryonic day E14.5 (strain no. 017927, the Jackson
   Laboratory). We generated Nae1 floxed by crossing a transgenic mouse
   line bearing a NAE1 “knockout-first” allele (Nae1^tm1a(EUCOMM)Wtsi)
   [European Conditional Mouse Mutagenesis Program (EUCOMM), no.
   EPD0441_1_C08] with the ROSA26:FLPe knock in line(strain no. 003946,
   the Jackson Laboratory) to remove Frt-flanked neo cassette. To generate
   tamoxifen-inducible Nae1 knockout mice (Nae1 icKO), Nae1 floxed mice
   were crossed with Plp1-CreERT mice (strain no. 005975, the Jackson
   Laboratory) ([309]23), followed by tamoxifen administration.

   Mice were bred at the Center for Cooperative Research in Biosciences
   (CIC bioGUNE) Association for Assessment and Accreditation of
   Laboratory Animal Care–accredited animal facility and at the Centro de
   Biomedicina Experimental-University of Santiago de Compostela animal
   facility. All animal procedures were approved by the CIC bioGUNE
   Institutional Animal Care and Use Committee, by the Faculty Animal
   Committee at the University of Santiago de Compostela, the Country
   Council of Bizkaia, and the Xunta de Galicia, and the experiments were
   performed in agreement with the Rules of Laboratory Animal Care and
   International Law on Animal Experimentation. Mice were maintained in
   groups of a maximum of five mice per cage, containing woodchip bedding,
   on a 12-hour light (8:00 to 20:00)/12-hour dark cycle, under controlled
   temperature (20° to 22°C) and humidity of 45 ± 10%. Mice were
   maintained with ad libitum access to water and to a rodent maintenance
   diet (2914 Teklad global 14% protein or the 2918 Teklad global 18%
   protein rodent maintenance diets, Envigo).

Tamoxifen induction of gene deletion

   Tamoxifen (T5648, Sigma-Aldrich) was dissolved to a stock concentration
   of 20 mg/ml in a vehicle of ethanol and corn oil (C8267,
   Sigma-Aldrich). The mix was incubated 2 hours at 65°C in agitation (700
   rpm) protected from sunlight.

   For adult mice (P60) treatment, 100 μl was administered by
   intraperitoneal injection once daily for four consecutive days,
   followed by a 2-day rest period, after which the mice were again
   treated for 4 days. Control mice (Plp-CreERT^−; Nae 1 ^fl/fl) were
   treated identically.

Sciatic nerve crush surgery

   Sciatic nerve crush was performed under deep surgical anesthesia
   (isoflurane; 5% induction, 2% maintenance). The hindlimbs and lower
   back were shaved and sterilized with iodine. An incision was made on
   the skin and the sciatic nerve exposed. The sciatic nerve was
   compressed once for 30 s using a pair of delicate forceps (Fine Science
   Instruments, catalog no. 11063-07) and again for 30 s at the same site
   but orthogonal to the initial crush. The incisure was sutured, and all
   animals undergoing surgery were given appropriate postoperative
   analgesia and monitored daily.

In vivo drug administration

   For neddylation inhibition, C57 mice pups were administered
   intraperitoneally with MLN4924 (HY-70062, MedChemExpress) (60 mg/kg),
   dissolved in a solution of 10% dimethyl sulfoxide (DMSO; D2650,
   Sigma-Aldrich), 40% PEG300 (8074841000, Sigma-Aldrich), 5% Tween 80
   (P1754, Sigma-Aldrich), and 45% saline solution (616003.9, KabiPac) at
   P3 of age for 2 days, followed by sacrifice at P5 and EM or immunoblot
   analysis.

   Rapamycin (AY22989, Selleck Chemicals) was dissolved in a solution of
   4% ethanol (1.08543.0250, EMD Millipore), 5% PEG300, 5% Tween 80, and
   85% ultrapure water, and 5 μg/g of body weight (or only carrier
   solution as control) was administered intraperitoneally once a day from
   P5 to P10, followed by sacrifice at P10 and EM or immunoblot analysis.

   XMU-MP-1 (HY-100526, MedChemExpress) was dissolved in a solution of 10%
   DMSO, 40% PEG300, 5% Tween 80, and 4% saline solution, and 5 mg/kg of
   body weight (or only carrier solution as control) was administered
   intraperitoneally once a day from P5 to P10, followed by sacrifice at
   P10 and EM or immunoblot analysis.

Rat primary Schwann cell culture

   Rat Schwann cells were isolated and purified, as described before
   ([310]60). In brief, sciatic nerves were obtained from P3 to P5 Wistar
   rats of either sex and digested in 0.3% trypsin (Glibco, 27250-018),
   0.6% type 2 collagenase (Worthington, [311]LS004174) in Hanks’ balanced
   salt solution (Gibco, 14170112). They were then cultured for 3 days in
   Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal bovine serum
   (FBS) containing AraC (10^−3 M) (Sigma-Aldrich, C6645-25MG). After 3
   days in culture, Schwann cells were immunopanned to remove remaining
   fibroblasts and expanded in expansion medium [defined medium, 0.5% FBS,
   2 μM forskolin, and neuregulin 1 (10 μg/m)]. Only the first seven
   passages were used. To induce differentiation, Schwann cells were
   cultured overnight in defined medium (DMEM/Ham’sF-12,
   antibiotic/antimycotic, GlutaMAX, B27 supplement, SATO supplement, T3,
   Hi insulin solution)/0.5% FBS and, the next day, switched to
   differentiation medium [defined medium/0.5% FBS and 1 mM dibutyryl cAMP
   (Sigma-Aldrich, D0627-1G)]. Sister cultures without db cAMP treatment
   were considered as basal conditions.

In vitro drug treatment

   For in vitro drug treatments, the following compounds were used for
   indicated times: XMU-MP-1 (2.5 μM; MedChemExpress, HY-100526); MLN4924
   (2.5 μM; MedChemExpress, HY-70062; CHX (5 μM; Carl Roth, 8682),
   rapamycin (100 ng/ml; Selleck Chemicals, AY22989), and MG132 (1 μM;
   Sigma-Aldrich, M7449-1ML). All drugs were dissolved in DMSO, except for
   rapamycin that was resuspended in ethanol. For control treatment, an
   equal volume of vehicle (DMSO or ethanol) was added to cultures.

Immunohistochemistry

   Sciatic nerves were dissected out, fixed in 4% paraformaldehyde
   overnight, washed three times with 1× phosphate-buffered saline (PBS)
   and then cryoprotected by incubating in 30% sucrose solution (in PBS).
   The nerves were embedded with optimal cutting temperature compound then
   frozen over dry ice and kept at −80°C. The samples were cut in a
   cryotome (Leica) at 5-μm thickness, and the sections placed onto
   Superfrost slides (Thermo Fisher Scientific).

   For immunostaining, slides were permeabilized with 0.5% Triton X-100 in
   PBS for 10 min, blocked with 1× PBS with 1% bovine serum albumin (BSA)
   and 0.1% Triton X-100 and incubated with the primary antibody anti-Iba1
   (Cell Signaling Technology, no. 17198S) at 1/100 concentration
   overnight at 4°C in a dark chamber. After washing three times with PBS,
   secondary antibodies at 1/500 concentration were added to the samples
   and incubated for 1 hour at room temperature (secondary antibodies
   linked with an Alexa Fluor 488 fluorophore). After washing three times
   with 1× PBS, samples were incubated with 4′,6-diamidino-2-phenylindole
   (DAPI; 50 ng/ml) (Merck) for 30 min. Last, slides were washed with PBS
   and mounted with Dako mounting medium (Agilent technologies). For EdU
   pulse labeling, an intraperitoneal injection of EdU (100 mg/kg;
   Invitrogen [312]C10338) was administered to mice before sciatic nerve
   collections. To label EdU^+ cells, the Click-it EdU proliferation assay
   kit for imaging (Thermo Fisher Scientific) was used, following the
   manufacturer’s instructions. DAPI (50 ng/ml; Merck) dye was used for 30
   min at room temperature to stain DNA. Digital images were acquired with
   an Axio imager A1 microscope (Carl Zeiss AG). The sections were
   analyzed in a blinded manner.

Electron microscopy

   Proximal sciatic nerves (before they branch out to sural, peroneal, and
   tibial nerves) were fixed with 2% glutaraldehyde (prepared in 0.1 M
   sodium phosphate buffer) at 4°C overnight and then postfixed with 1%
   osmium tetroxide at 4°C overnight. Following en bloc staining with 4%
   uranyl acetate solution during 45 min at 4°C, the nerves were then
   dehydrated in graded alcohol, infiltrated with propylene oxide, and
   embedded in Epoxy resin (Agar Scientific). Semithin sections (1 μm) and
   ultrathin (50 nm) sections were cut with a Leica ultramicrotome (Leica,
   Germany). Semithin sections were stained with 0.1% toluidine blue in
   ethanol for 15 s and visualized using a Nikon Optiphot-2 microscope.
   Ultathin sections were stained with lead citrate and visualized under a
   transmission electron microscope FEI Tecnai G2 Spirit BioTwin (Thermo
   Fisher Scientific company, Oregon, USA). All images were acquired with
   a Xarosa digital camera (EMSIS GmbH, Münster, Germany) controlled by
   Radius software (version 2.1). Morphometric measurements (1:1
   promyelinating cells, myelinated fibers, Schwann cell nuclei, and
   number of axons) were performed using electron micrographs of ultrathin
   sections by manually counting all of the figures present in the sciatic
   nerve sections using the cell counter plugin from NIH ImageJ software
   ([313]http://rsb.info.nih.gov/ij/). Five pictures of each experimental
   group (at least three biological replicates per group) were quantified,
   and the counts of the quantified area were extrapolated to the total
   sciatic nerve area. Total axon counts on P47 nerves could not be
   performed because of advanced stages of tissue damage (collapsed small
   diameter axons) in Nae1 cKO mice. For G-ratio analysis, the axons
   (without myelin) and fiber areas (including myelin) of randomly chosen
   fibers were measured, their diameters were estimated, and the G-ratio
   was calculated as axon diameter/fiber diameter. At least 100 nerve
   fibers per animal (n = 3 for each group) were analyzed.

Lentiviral infections

   Lentivirus was produced by transfection of lentiviral constructs
   purchased from Sigma-Aldrich (table S1) into human embryonic kidney
   293T cells with the packaging vectors pMDL, pREV, and pVSV-G using
   Turbofect transfection reagent (Thermo Fisher Scientific, no. R0531),
   according to the manufacturer’s instructions. The supernatant was
   collected at 24 and 48 hours after transfection, concentrated with
   Lenti-X concentrator (Takara, no. 631231). For knockdown, Schwann cells
   were treated with short-hairpin lentiviral particles in the presence of
   hexadimethrine bromide (8 μg/ml). Twenty-four hours after transduction,
   the cells were selected using puromycin (0.5 μg/ml) (Sigma-Aldrich,
   P7255-250MG), and puromycin-resistant cell clones were grown, analyzed,
   and frozen for future use.

Cytosol/nucleus extraction

   The cytosol and nuclear fractions were extracted following the protocol
   from ProteoExtract Subcellular Proteome Extraction Kit (Sigma-Aldrich,
   no. 539790) according to the manufacturer’s instructions.

Immunoblotting

   For immunoblotting, sciatic nerves were dissected, and the epineurium
   was stripped before snap freezing and storage at −80°C. Nerves or rat
   Schwann cells were lysed in radioimmunoprecipitation assay (RIPA)
   buffer [1.6 mM NaH[2]PO[4], 8.4 mM NaHPO[4], 0.1% Triton X-100, 0.1 M
   NaCl, 0.1% SDS, 0.5% sodium azide, 20 mM indole-3-acetic acid (IAA)
   (Sigma-Aldrich, no. 8047440025), 10 mM N-ethylmaleimide (NEM)
   (Sigma-Aldrich, no. E3876-5G), phosphatase (PhosSTOP EASYpack,
   Sigma-Aldrich 04906837001), and protease (Complete Mini Protease
   Inhibitor Cocktail, Sigma-Aldrich, no. 04693124001) inhibitors]. Cells
   were scraped directly from the plate, and sciatic nerves were processed
   in a 1.5-ml Eppendorf tubes containing zirconium beads (Next Advance,
   ZrOB05) using Bullet Blender Homogenizer (Next Advance, BBX24-CE) for
   homogenizing. Samples were centrifuged at 13,000 rpm for 20 min at 4°C,
   and supernatants were quantified with the Pierce Micro BCA Assay Kit
   (Thermo Fisher Scientific, no. 23235). Protein samples were separated
   by SDS–polyacrylamide gel electrophoresis in 8 to 15% acrylamide gels
   using a Mini-PROTEAN Tetra Cell electrophoresis system (Bio-Rad) or 4
   to 12% NuPAGE Bis-Tris (Thermo Fisher Scientific, WG1403BOX,). Blue
   Star Protein marker from Nippon genetics was used as molecular weight
   marker. Proteins were transferred from gels into 0.2-μm–pore size
   nitrocellulose membranes (GE Healthcare, no. 10600001) using a
   Trans-Blot Cell system electroblotting or using a Bio-Rad Trans-blot SD
   Semi-dry Transfer Cell (Bio-Rad). The presence of total protein was
   detected by Ponceau S solution (Sigma-Aldrich, no. P7170) staining.
   Nonspecific binding was blocked by incubation of the membranes with
   0.1% (v/v) Tween 20–tris-buffered saline (TBS), solution containing 5%
   (w/v) skimmed milk powder or 2% (w/v) BSA (Sigma-Aldrich, no. A3912)
   for 1 hour at room temperature, before addition of the primary antibody
   (table S2). After washing the unbound primary antibody with 0.1% (v/v)
   Tween 20–TBS three times, membranes were incubated with the
   corresponding horseradish peroxidase (HRP)–linked secondary antibody
   for 1 hour at room temperature. Membranes were washed three times to
   remove the excess of secondary antibody and the Western Lightning
   Plus-ECL HRP substrate (PerkinElmer, no. NEL104001EA) was subsequently
   added. The chemiluminescent signal from immunoreactive proteins was
   captured using an iBrightTM Imaging System (model no. FL1500) or by
   exposition to Super RX Fuji Medical X-Ray films (Fujifilm, no.
   06-SHRGH1824B), which were further developed in a Curix 60 processor
   (AGFA Healthcare). Films were digitalized and transformed into 8-bit
   images. Bands were quantified by densitometric analysis using the
   open-source image processing program ImageJ
   ([314]http://rsbweb.nih.gov/ij) and normalized to
   glyceraldehyde-3-phosphate dehydrogenase or β-actin housekeeping
   protein expression.

RNA isolation and RT-qPCR

   Total RNA was isolated from Schwann cells or sciatic nerves with TRIzol
   reagent (Sigma-Aldrich, no. T9424-200ML) according to the
   manufacturer’s instructions. Cells were scraped directly from the
   plate, and sciatic nerves were processed in a 1.5-ml Eppendorf tube
   containing zirconium beads (Next Advance, ZrOB05) using Bullet Blender
   Homogenizer (Next Advance, BBX24-CE) for homogenizing. RNA
   concentration was determined using NanoDrop 2000 spectrophotometer
   (Thermo Fisher Scientific). A minimum of 200 ng of total RNA and up to
   1 μg was reverse-transcribed in a 20 μl of reaction containing
   deoxynucleotide triphosphates (Thermo Fisher Scientific, no. R0182),
   random hexamers (Thermo Fisher Scientific, no. 48190011), 0.1 M
   dithiothreitol (Invitrogen, no. Y00147), RNaseOUT (Thermo Fisher
   Scientific, no. 100000840), and 5× Buffer and M-MLV Reverse
   Transcriptase (Thermo Fisher Scientific, no. 28025013) according to the
   manufacturer’s instruction. The resulting cDNA was diluted 1/5 with 80
   μl of RNase free H[2]O (Sigma-Aldrich). RT-qPCR was performed in a 5-μl
   reaction mix containing 2.5 μl of Applied Biosystems SYBR Select Master
   Mix (Thermo Fisher Scientific, no. 44-729-19), 0.5 μl of 10 μM primers,
   and 2 μl of cDNA. All reactions were performed in triplicates in a QS6
   Real-Time PCR System (Applied Biosystems). Forty cycles with a melting
   temperature of 60°C and 30 s of each step were used. Specific primers
   were designed with Primer Blast database
   ([315]www.ncbi.nlm.nih.gov/tools/primer-blast/) (table S3) and
   synthesized by Sigma-Aldrich. Quantification was performed using the
   ∆∆CT method and data was normalized using GADPH mRNA as a standard.

Proteomics

   Sciatic nerves were dissected from 15-day-old mice and homogenized with
   a Fastprep machine in RIPA buffer supplemented with deoxycholate, IAA,
   NEM, phosphatase, and protease inhibitors. Filter aided sample
   preparation protocol ([316]61) was used for sample processing and
   digestion. Trypsin was added to a trypsin:protein ratio of 1:50, and
   the mixture was incubated overnight at 37°C, dried out in a RVC2 25
   speedvac concentrator (Christ) and resuspended in 0.1% formic acid
   (FA). Peptides were desalted and resuspended in 0.1% FA using C18 stage
   tips (Merck Millipore). Peptides were analyzed in a hybrid trapped ion
   mobility spectrometry–quadrupole time-of-flight mass spectrometer
   (timsTOF Pro with PASEF, Bruker Daltonics) coupled online to a
   nanoElute liquid chromatograph (Bruker). Protein identification and
   quantification were carried out using PEAKS software (Bioinformatics
   solutions). The polar plot showing the z-scores of proteins related to
   ubiquitination and neddylation processes was performed with Tbtools
   v1.120. KEGG pathway enrichment analysis of significantly deregulated
   proteins with a FC > 1.5 was performed on the ShinyGO V0.76 platform
   ([317]http://bioinformatics.sdstate.edu/go/). The 22 most significant
   KEGG terms (lowest FDR) were plotted on a Ballon Plot in GraphPad Prism
   v9. The mass spectrometry proteomics data have been deposited to the
   ProteomeXchange Consortium via the PRIDE ([318]62) partner repository
   with the dataset identifier PXD043917 and 10.6019/PXD043917.

RNA-seq and data analysis

   RNA was isolated from 7-day-old mice sciatic nerves, as above. Total
   RNA was treated with deoxyribonuclease I (Invitrogen) for 30 min at
   37°C. Then, the samples were cleaned with Genejet RNA cleanup and
   Concentration kit (Thermo Fisher Scientific) (the columns were eluted
   with 20 to 30 μl of ultrapure water).

   Their concentration and integrity were determined with a Qubit RNA HS
   Assay kit (Thermo Fisher Scientific) and Agilent RNA 6000 pico Chips
   (Agilent Technologies), respectively. The “Truseq Stranded Total RNA
   with Ribo-Zero Globin” kit (Illumina Inc.) and “TruSeq RNA CD Index
   Plate” (Illumina) were used to prepare the sequencing libraries. The
   protocol followed was “TruSeq Stranded Total RNA Sample Prep-guide.” In
   brief, an amount of 250 ng of total RNA was used for the cDNA
   preparation. SuperScript-II Reverse Transcriptase kit (Thermo Fisher
   Scientific) was used for the first cDNA strand synthesis, and Illumina
   specific reagents were used for the second cDNA strand synthesis. The
   next step was the ligation of adaptors and A-tails followed by an
   enrichment PCR (30 s at 98°C; 15 cycles of 10 s at 98°C, 30 s at 60°C,
   30 s at 72°C; and 5 min at 72°C; and it was paused at 4°C). Libraries
   were quantified using the Qubit dsDNA HS DNA Kit (Thermo Fisher
   Scientific, catalog no. [319]Q32854) and visualized on an Agilent 2100
   Bioanalyzer using an Agilent High Sensitivity DNA kit (Agilent
   Technologies, catalog no. 5067-4626) and sequenced in an Illumina Inc.
   instrument. RNA-seq reads were aligned to mm10 using TopHat with
   default settings ([320]http://tophat.cbcb.umd.edu/). We used Cuff-diff
   to (i) estimate fragments per kilobase of transcript per million mapped
   reads (FPKM) values for known transcripts and to (ii) analyze
   differentially expressed transcripts. In all differential expression
   tests, a difference was considered significant if the q value was less
   than 0.05 (Cuff-diff default). A heatmap of gene expression was
   generated using the R language (version 3.2.1) and was generated on the
   basis of log[2] (FPKM). RNA-seq data are available under Gene
   Expression Omnibus (GEO) accession [321]GSE241269.

   GO Term Enrichment using PANTHER ([322]www.pantherdb.org/) and GSEA
   ([323]www.broadinstitute.org/gsea/index.jsp) were performed with GSEA
   4.1.0 using the hallmark gene set (MSigDB database v7.4) as gene set
   database. We also performed GSEA analysis of custom gene sets from
   public RNA-seq datasets in murine models of our interest: TSC1 cKO
   ([324]25) and PTEN cKO ([325]25), Zeb2 cKO ([326]38), and
   LE_EGR2_TARGETS_DN MsigDB dataset. We used ToppCluster
   ([327]https://toppcluster.cchmc.org/) to construct the network of genes
   belonging to overrepresented GO term categories.

   Public datasets from the Sciatic Nerve Atlas project
   ([328]https://snat.ethz.ch/) were used to analyze Nae1 expression. Bulk
   RNA-seq: Raw counts were obtained from a bulk RNA-seq composed of
   mixed, myelinating, and not-myelinating Schwann cells extracted from P5
   mouse sciatic nerves ([329]GSE137947). Differential expression analysis
   between groups was performed with DESeq2 (1.38.3, R version 4.2.2). FDR
   was obtained from DESeq2 contrast with Benjamini-Hochberg correction.
   scRNA-seq: The Seurat scRNA-seq object of mouse cells selected from
   sciatic nerves at P1 was obtained directly from the atlas website and
   corresponds to dataset [330]GSE138577. Differential expression analysis
   between Schwann cell clusters was performed with Seurat’s FindMarkers
   function (4.3.0) using a Wilcoxon rank sum test.

Electrophysiological measurements

   Nae1 cKO mice were anesthetized with sodium pentobarbital (25 mg/kg
   intraperitoneally). At these low doses, the anesthetic does not block
   nerve conduction. To be sure of the location of the electrodes, their
   placement was performed under a surgical microscope. A thermostatic pad
   at 34° to 36°C maintained the temperature of the animals.

   For motor evoked potentials (MEPs), stimulation was performed through a
   concentric needle electrode placed near to the sciatic notch with
   square pulses (100-μs duration, 0.2-Hz frequency) delivered through a
   constant current stimulus isolator (Stimulus Isolator A365; World
   Precision Instruments, USA). MEPs were recorded using monopolar steel
   needles electrodes placed in the palmar interosseous muscle located at
   the third metatarsal space, using a Neurolog System (UK). Signals were
   amplified, filtered (5 kHz), digitized in a PC, and each five sweeps
   were averaged using PClamp software and Digidata 1322A (Molecular
   Devices, USA). For each recording, the minimum intensity to reach 100%
   deflection, the latency from the stimulus to the first negative wave
   and the distance in millimeters from the stimulation and recording
   electrodes were measured.

   Alternatively, for Nae1 icKO mice, the UltraPro S100 EMG/NCS/EP
   Neurodiagnostic System (Natus) was used using a protocol adapted from
   ([331]63). In brief, Nae1 icKO mice were anesthetized with isoflurane
   (5% induction, 2% maintenance). Mice were positioned in prone position,
   and hindlimbs were extended with tape. Needle electrodes (27 gauge)
   were used to record hindlimb CMAP measurements. Stimulating electrodes
   were placed subcutaneously on both sides of the sciatic notch with a
   2-cm distance, approximately (anode = 1 and cathode = 2). The recording
   electrode was also placed subcutaneously aligned with the gastrocnemius
   muscle. The reference electrode was inserted subcutaneously next to the
   Achilles tendon in a 30° angle. The ground electrode was placed
   subcutaneously on the other side of the mouse. First supramaximal
   stimulation was searched gradually increasing the stimulation, until a
   perfect defined curve was obtained. Afterward, stimulations were
   recorded at 10.2, 15.3, and 20.0 mA. The distance to calculate NCV was
   the distance between cathode stimulation electrode and recording
   electrode.

Sensory-motor functional tests

Rotarod

   Before analyses, mice were trained on the rotating rod (UgoBasile, no.
   47650) (5 rpm for 60 s) for a total of three trials separated by 10-min
   intertrial intervals. For rotarod, animals were tested in one daily
   session for three consecutive days on an accelerating rotarod (4 to 40
   rpm for 5 min). Each session included three trials, and the average
   time on the rod per session was reported.

Grip strength

   The protocol was adapted from Biomedical And Obesity Research Core
   ([332]https://cehs.unl.edu/borc/grip-strength-system/). Forelimb
   measurement: Mice were lowered over the grid (UgoBasile, no. 47200)
   allowing only its forepaws to attach to the grid before any
   measurements are taken. They were gently pulled back by their tail
   ensuring that they grip the top portion of the grid keeping the torso
   horizontal. The maximal grip strength value displayed on the screen was
   recorded. Three forelimb grip strength measurements in total were
   obtained. Forelimb and hindlimb measurement: Mice were lowered over the
   grid keeping the torso parallel with the grid and allowing both their
   forepaws and hind paws to attach to the grid before any measurements
   are taken. Mice were gently pulled back by their tail ensuring that the
   torso remained parallel with the grid and the maximal grip strength
   value of the mouse that is displayed on the screen was recorded. Three
   forelimb/hindlimb grip strength measurements in total were obtained.

Hot plate

   The Hot Plate apparatus (UgoBasile, no. 35150) was set to a temperature
   of 55° ± 0.2°C. On the testing day, mice were placed on the surface of
   the hot plate and covered by a glass transparent cylinder, (25-cm high
   and 12-cm diameter). A 30-s cutoff time was assigned in this protocol.
   A remote foot-switch pad wass used to control the start/stop/reset
   functions. The latency to response was recorded when the first hind paw
   lick or jump occurs.

Sciatic functional index

   The mice were allowed to walk on a 50-cm walking track, and the
   footprints were recorded and digitalized. The distance between toes 1
   and 5 and the length of the print were measured using the FOOTPRINTS
   program ([333]64), and the SFI was calculated according to Inserra et
   al. ([334]65). The observer was blinded regarding the genotype of the
   mice.

Immunoprecipitation

   Cells were lysed in NP-40 lysis buffer [50 mM tris HCl (pH 8.5), 150 mM
   NaCl, 5 mM EDTA, 1% NP-40 Substitute (Sigma-Aldrich, no. 74385),
   ultrapure water, 20 mM IAA, 10 mM NEM, phosphatase (PhosSTOP EASYpack,
   Sigma-Aldrich, no. 04906837001), and protease (Complete Mini Protease
   Inhibitor Cocktail, Sigma-Aldrich, no. 04693124001) inhibitors]. They
   were centrifuged at 13,000 rpm for 2 min at 4°C, and the supernatant
   was quantified with the Pierce Micro BCA Assay Kit (Thermo Fisher
   Scientific, no. 23235).

   For immunoprecipitation with Protein A Sepharose beads, antibodies were
   cross-linked with Protein A Sepharose 4B, Fast Flow from Staphylococcus
   aureus, aqueous ethanol suspension (Sigma-Aldrich, P9424-5ML). Protein
   (100 μg) from cell lysates were incubated with antibodies cross-linked
   with Protein A Sepharose overnight at 4°C in rotation. The next day,
   three washes with the lysis buffer were performed by centrifugation at
   5000 rpm for 5 min at 4°C. Last, the beads were resuspended in loading
   buffer, and two gels were run and transferred to a nitrocellulose
   membrane. Negative control was performed with a pool of all of the
   samples and immunoglobulin G (IgG) antibody.

   For immunoprecipitation with magnetic beads, Dynabeads M-280 Sheep
   Anti-Rabbit IgG (Thermo Fisher Scientific, ref. 11203D) were washed
   twice with PBS and incubated for 10 min with BSA (0.1 mg/ml) in PBS.
   They were incubated with the antibody at the manufacturer recommended
   dilution for 1 hour at 4°C. Control beads are incubated with the
   corresponding IgG (BD Biosciences, no. 550875) at the same
   concentration as the antibody used. The bead-antibody cross-link was
   performed with dimethyl pimelimidate (6.5 mg/ml; Sigma-Aldrich, no.
   D8388) in PBS for three cycles of 30 min incubation and 5 min washes
   with 0.2 M triethanolamine. Quenching was performed with 50 mM
   ethanolamine in PBS for 5 min at room temperature twice. Beads were
   then washed twice for 10 min at room temperature with 1 M glycine (pH
   3) to remove the excess of antibody, washed with PBS, and resuspended
   in the same protein extract buffer to perform the incubation with the
   protein extracts (see above) at 4°C for 16 hours. Beads are washed
   three times with NP-40 buffer, eluted in loading buffer, and boiled 5
   min at 4°C before immunoblotting.

Halo-Nedp1 capture

   The protocol was adapted from Keuss et al. ([335]40). The following
   plasmids were used: his6-HALO NEDP1(C163A) DSTT DU28042 [Nedp1 CA
   (capturing condition)] and his6-HALO NEDP1 (C163A D29W A98K G99K) DSTT
   DU28008 [Nepd1 DAGC (non-capturing condition)]. pET28a 6His-HALO NEDP1
   was transformed into BL21 Rosseta cells and purified with Ni-NTA
   Agarose (QIAGEN, no. 30210). The recombinant proteins were then
   conjugated with HALO link agarose resin (Promega, no. G1912).

   For HALO pulldown, rat Schwann cells were cultured under three
   different conditions: control, cAMP (1 mM) and DMSO, and cAMP (1 mM)
   and MLN4924 (2.5 μM). They were scraped in ice-cold NP-40 lysis buffer
   (as above) containing 3 mM Phenantroline (Sigma-Aldrich, no.
   131377-2.5G). They were collected and subjected to one cycle of
   freezing/thawing. Samples were centrifuged at 13,000 rpm for 30 min,
   and the supernatant was quantified with the Pierce Micro BCA Assay Kit
   (Thermo Fisher Scientific, no. 23235). Supernatants were incubated with
   HALO-Nedp1 CA or DAGC in rotation at 4°C for 2.5 hours. They were
   centrifuged at 500g for 2 min and washed three times with lysis buffer.
   The beads were then resuspended in loading buffer, heated at 95°C for 5
   min, and centrifuged at 500g for 2 min at room temperature. The
   supernatant was then loaded on 12% Bis-Tris gel and transferred to
   nitrocellulose membrane (Amersham Protran, no. GE10600001). The
   membranes were probed with NEDD8 antibody to test the efficiency of the
   capture and with EGR2 or ZEB2 antibody to detect their presence in the
   capture.

Genotyping and recombination

   Genomic DNA was obtained from tail snips after digestion with 25 mM
   NaOH and 0.2 mM EDTA (pH 12) for 30 min at 95°C and then buffered with
   40 mM tris-HCl (pH 5). DreamTaq Green PCR Master Mix (2×) (Thermo
   Fisher Scientific, K1081) was used for PCRs. Primers used for PCR
   reactions are detailed in table S3. For MPZ-cre, PCR conditions were
   94°C for 3 min, (94°C for 30 s, 60°C for 30 s, and 72°C for 60 s) for
   35 cycles and 72°C for 10 min. For Nae1-floxed and Plp-CreERT, PCR
   conditions were 94°C for 3 min (94°C for 30 s, 62°C for 30 s and 72°C
   for 60 s) for 35 cycles and 72°C for 10 min.

Statistical analysis

   Statistical analysis was performed using GraphPad Prism 8 software.
   Data were represented as means ± SEM for each experimental group.
   Statistical significance was determined by Student’s t test or one-way
   analysis of variance (ANOVA) and two-way ANOVA followed by Sidak’s or
   Tukey’s post hoc tests. P < 0.05 was considered as statistically
   significant.

Acknowledgments