Caenorhabditis elegans F-Box Protein Promotes Axon Regeneration by Inducing Degradation of the Mad Transcription Factor

In Caenorhabditis elegans, axon regeneration is activated by a signaling cascade through the receptor tyrosine kinase (RTK) SVH-2. Axonal injury induces svh-2 gene expression by degradation of the Mad-like transcription factor MDL-1. In this study, we identify the svh-24/sdz-33 gene encoding a protein containing F-box and F-box-associated domains as a regulator of axon regeneration in motor neurons. We find that sdz-33 is required for axon injury-induced svh-2 expression.


Introduction
The ability of axons to regenerate after damage is a fundamental and conserved property of neurons, modulated by intrinsic processes that regulate axon growth potential. Upon axon severance, regenerative signals are transmitted from the sites of damage to the nucleus, whereupon several transcription factors are upregulated along with the synthesis of proteins participating in neurite outgrowth (Tedeschi and Bradke, 2017). Since manipulation of these signaling processes can improve the likelihood of successful axon regeneration, these processes are potential targets for regenerative therapies. However, these intrinsic signaling mechanisms have yet to be fully elucidated.
The nematode Caenorhabditis elegans has recently emerged as an attractive model to dissect the mechanisms of axon regeneration in the mature nervous system (Yanik et al., 2004). In C. elegans, the JNK MAP kinase (MAPK) cascade plays a crucial role in the initiation of axon regeneration (Nix et al., 2011;Hisamoto and Matsumoto, 2017). This pathway consists of MLK-1 MAPKKK, MEK-1 MAPKK, and KGB-1 JNK, and is inactivated at the KGB-1 activation step by VHP-1, a member of the MAPK phosphatase family (Mizuno et al., 2004). The vhp-1null mutation causes hyperactivation of the JNK pathway, resulting in developmental arrest at an early larval stage. We recently have identified new components functioning in JNK-mediated signaling by using a genome-wide RNAi screen for suppressors of vhp-1 lethality (svh genes; Li et al., 2012). The svh-1 gene encodes a growth factor-like protein homologous to mammalian hepatocyte growth factor (HGF), and the svh-2 gene encodes a homolog of mammalian Met, a receptor for HGF (Li et al., 2012). SVH-2 is a receptor tyrosine kinase (RTK) that activates the JNK pathway via the tyrosine phosphorylation of MLK-1 MAPKKK.
SVH-1ÀSVH-2 signaling specifically regulates axon regeneration, and this specificity is determined by the induction of svh-2 gene expression following axon injury (Li et al., 2012). This upregulation critically involves the physical interaction of the Ets-like transcription factor ETS-4 and the CCAAT/enhancer-binding protein (C/EBP)-like transcription factor CEBP-1 (Li et al., 2015). Upon axon injury, cAMP levels increase in severed neurons, resulting in the activation of cAMP-dependent protein kinase (PKA), which in turn phosphorylates ETS-4. Phosphorylated ETS-4 is able to form a complex with CEBP-1, which then activates svh-2 transcription (Li et al., 2015). Furthermore, we recently identified the Mad-like transcription factor MDL-1, the Max-like transcription factor MXL-1, and TDP2 (tyrosyl-DNA phosphodiesterase 2)-like TDPT-1 as components involved in the regulation of ETS-4 transcriptional activity for the induction of svh-2 gene expression ( Fig. 1A; Sakai et al., 2019). TDPT-1 interacts with and induces SUMOylation of ETS-4, which interferes with PKA-mediated phosphorylation of ETS-4. As a result, Figure 1. Identification of SDZ-33. A, Regulation of svh-2 expression in response to axon injury. MXL-1 forms a complex with MDL-1, and TDPT-1 interacts with ETS-4 to induce its SUMOylation, resulting in the repression of ETS-4 transcriptional activity. Axon injury leads to the degradation of MDL-1, allowing free MXL-1 to interact with TDPT-1. ETS-4 is then de-SUMOylated, and subsequently induces svh-2 expression. B, Structures of SDZ-33 and human UG063H01. The domains shown are the F-box domain (blue) and type 2 FBA (yellow). Percentages of similarities are indicated. The bold line underneath indicates the extent of the deleted region in the tm1210 deletion mutant. C, D, Amino acid alignments of the N-terminal (C) and C-terminal (D) domains. Identical and similar residues are highlighted with dark and pale gray shading, respectively. formation of the ETS-4-CEBP-1 complex is inhibited, and its transcriptional activity thereby repressed. MXL-1 activates svh-2 transcription by interacting with TDPT-1 and relieving inhibition of ETS-4 activity. MDL-1 forms a complex with MXL-1, and this interaction induces the dissociation of TDPT-1 from MXL-1, enabling free TDPT-1 to inhibit ETS-4 transcriptional activity. Thus, TDPT-1 and MDL-1 negatively regulate axonal injuryinduced expression of the svh-2 gene via modulation of ETS-4 ( Fig. 1A). Axon injury leads to the degradation of MDL-1, which is linked to the activation of ETS-4 transcriptional activity. Thus, MDL-1 protein stability is important in the regulation of axon regeneration, but the details of how this stability is modulated are at present unknown.
In the present study, we investigate the role of the svh-24/sdz-33 gene in the regulation of axon regeneration. The sdz-33 gene encodes a protein containing an F-box domain (Robertson et al., 2004), which confers substrate recognition by the SCF [S-phase kinase-associated protein 1 (Skp1)-Cullin1-F-box] E3-ubiquitin (Ub) ligases (Craig and Tyers, 1999). Here, we show that MDL-1 is recognized by SDZ-33, which directs its degradation via the 26S proteasome. Thus, induction of the SDZ-33-mediated MDL-1 degradation pathway following neuron injury is essential for axon regeneration.

Materials and Methods
C. elegans strains. The C. elegans strains used in this study are listed in Table 1. All strains were maintained on nematode growth medium plates and fed with bacteria of the OP50 strain, as described previously (Brenner, 1974).
Microscopy. Standard fluorescent images of transgenic animals were obtained under a 100Â objective on a Nikon ECLIPSE E800 fluorescent microscope and photographed with a Zyla CCD camera. Confocal fluorescent images were taken on a Zeiss LSM-800 confocal laser-scanning microscope with a 63Â objective.
Axotomy. Axotomy and microscopy were performed as described previously (Pastuhov et al., 2012). All animals were subjected to axotomy at the young adult stage. Imaged commissures that had growth cones or small branches present on the proximal fragment were counted as "regenerated." Proximal fragments that showed no change after 24 h were counted as "no regeneration." A minimum of 20 individuals with one to three axotomized commissures were observed for most experiments.
Measuring the length of regenerating axons. The length of regenerating axons for D-type motor neurons was measured using the segmented line tool of ImageJ. Measurements were made from the site of injury to the tip of the longest branch of the regenerating axon. Axons that did not regenerate were excluded from the measurements.
Quantification of VENUS expression. Expression of VENUS fluorescence was quantified using the ImageJ program (NIH). Cell bodies of severed or unsevered D neurons were outlined with closed polygons, and the mean fluorescent intensities of VENUS and CFP were measured to obtain I VENUS and I CFP, respectively. As a control, the cell body next to the cell body of interest was similarly analyzed [I VENUS(c) and I CFP ( Time-lapse imaging of MDL-1::GFP after axotomy. Animals expressing mCherry and MDL-1::GFP in their D-type motor neurons were Figure 2. SDZ-33 is involved in axon regeneration. A, Representative D-type motor neurons in wild-type (WT) and sdz-33 mutant animals 24 h after laser surgery. In wild-type animals, a severed axon has regenerated a growth cone (arrow). In mutants, the proximal end of axon failed to regenerate (arrowhead). Scale bar, 10 mm. B, Percentages of axons that initiated regeneration 24 h after laser surgery. The numbers (n) of axons examined are shown. Error bars indicate 95% confidence intervals. pp , 0.05, ppp , 0.01, as determined by Fisher's exact test. NS, Not significant. Schematic diagrams for Psdz-33::sdz-33::sdz-33 39UTR, Punc-25:: sdz-33 (cDNA)::unc-54 39UTR, Punc-25::sdz-33::sdz-33 39UTR, and Pmec-7::sdz-33::sdz-33 39UTR are shown in the top part. C, Length of regenerating axons 24 h after laser surgery. The data are presented as box plot graph with median (thick line within the box) and interquartile range (edge of box). A white circle indicates an outlier. The numbers (n) of axons examined are shown. Statistical significance was determined using the Wilcoxon rank-sum test; p = 0.0036. D, Expression of the Psdz-33::nls::gfp gene. Fluorescent and differential interference contrast (DIC) images of animals carrying Psdz-33::nls::gfp and Punc-47::mcherry 6 h after cutting are shown. D-type motor neurons are visualized by mCherry under control of the unc-47 promoter. White arrow indicates vulval muscles. Yellow arrowheads indicate severed axons. The fluorescent signals in the pharynx (asterisks) are from the injection marker Pmyo-2::DsRed-monomer. V, Ventral side; D, dorsal side. Scale bar, 20 mm. imaged for 6 h at 15 min intervals beginning shortly after axotomy of selected motor neuron axons.
Quantitative measures of fluorescence intensity for MDL-1 degradation. Animals expressing mCherry and MDL-1::GFP in their D-type motor neurons were imaged shortly after (0 h) and 6 h after axotomy of selected motor neuron axons. An LSM800 (Zeiss) confocal microscope was used to obtain z-stacks of fluorescent images for mCherry and MDL-1::GFP. The mean intensities of MDL-1::GFP and mCherry in the nuclei of neurons with severed axons were measured by drawing a circular ROI in the middle of the cell and using the measure function of ImageJ. Background intensities were measured near the measured cells. Relative MDL-1::GFP intensity (RI MDL-1 ) was calculated by dividing the background-subtracted value for GFP by the corresponding background-corrected value for mCherry at 6 h postaxotomy divided by the corresponding value at 0 h postaxotomy. The RI MDL-1 values for the wild-type and sdz-33 mutant were plotted and checked for significant differences (Wilcoxon rank-sum exact test) using RStudio.
Homology search, phylogenetic analysis, identification of domains, and alignment of amino acids. A homology search, identification of conserved domains, and alignment of amino acids were performed using the National Center for Biotechnology Information (NCBI) DELTA-BLAST, NCBI CD-search, and GENETYX-MAC programs, respectively.

SVH-24/SDZ-33 is required for efficient axon regeneration
To identify additional components functioning in the JNK pathway regulation of axon regeneration, we previously undertook a genome-wide RNAi screen for suppressors of vhp-1 lethality and isolated 92 svh genes (Table 2). To identify components involved in axonal injury-induced degradation of MDL-1, we asked whether any of the svh genes encode factors that participate in protein degradation. Among these svh genes, we focused on svh-24, which encodes a protein containing an F-box domain.
The svh-24 gene is identical to the sdz-33 gene, which has been identified as a ubiquitously expressed early zygotic transcript (Robertson et al., 2004). The SDZ-33 protein contains Fbox and type 2 F-box-associated (FBA) domains ( Fig. 1B; Thomas, 2006), and has weak homology with the human F-box protein UG063H01 (Fig. 1C,D). The sdz-33(tm1210) deletion is a null allele (Fig. 1B). To characterize the role of sdz-33 in axon regeneration, we performed laser axotomy and assayed regeneration in GABA-releasing D-type motor neurons. In young adult wild-type animals, 62% of the severed axons initiated regeneration within 24 h after axon injury, whereas in sdz-33(tm1210) mutant animals, this frequency was reduced ( Fig. 2A,B, Table 3). Although the morphology of D-type motor neurons was normal in sdz-33 mutants, the size of bulb-like structures in the nonregenerating axons of sdz-33 mutants was larger than that in wildtype animals ( Fig. 2A). We found that the length of regenerating axons in sdz-33 mutants was shorter than that observed in wildtype animals (Fig. 2C). However, sdz-33 mutants had no other detectable developmental or behavioral phenotypes. To verify that the sdz-33 mutation was responsible for this defect in axon regeneration, we generated the transgene Psdz-33::sdz-33::sdz-33 39UTR, which contains the entire genomic sdz-33 region (Fig. 2B). Introduction of Psdz-33::sdz-33::sdz-33 39UTR into sdz-33(tm1210) mutants rescued the defect in axon regeneration ( Fig. 2B; Table 3).
To investigate whether SDZ-33 functions in Dtype motor neurons, we examined the expression pattern of the sdz-33 gene. We constructed a transgene, Psdz-33::nls::gfp, which expresses the fluorescent protein GFP fused to a nuclear localization signal (NLS) under the control of the sdz-33 promoter. In the young adult stage, animals carrying Psdz-33::nls::gfp exhibited weak expression of GFP in vulval muscles but not in D-type motor neurons. GFP expression was still not observed in D neurons after axon injury (Fig. 2D). Therefore, we examined whether SDZ-33 acts in D-type motor neurons. In sdz-33(tm1210) mutants, expression of the sdz-33 cDNA driven by the unc-25 promoter with the unc-54 39UTR could not rescue the sdz-33 (tm1210) defect in axon regeneration (Fig. 2B, Table 3). On the other hand, when sdz-33::sdz-33 39UTR was expressed by the unc-25 promoter in sdz-33 (tm1210) mutants, the regeneration defect was rescued (Fig. 2B, Table 3). These results suggest that the intron or 39UTR region of the sdz-33 gene is required for sdz-33 expression. It is known that axon injury promotes stability and local translation of the mRNA encoding CEBP-1, a homolog of mammalian C/EBP, via its 39UTR (Yan et al., 2009). This suggests that the 39UTR region of the sdz-33 gene may determine the stability and proper localization of the sdz-33 mRNA. In contrast to the expression of sdz-33::sdz-33 39UTR from the unc-25 promoter, expression of the sdz-33::sdz-33 39UTR DNA from the mec-7 promoter in sensory neurons failed to rescue the sdz-33(tm1210) defect (Fig. 2B, Table 3). Thus, SDZ-33 regulates axon regeneration in injured D-type motor neurons after laser axotomy in a cell-autonomous manner.

SDZ-33 is involved in axotomy-induced svh-2 expression in injured D-type motor neurons
Our RNAi screen for svh genes was originally designed to identify components functioning in the JNK pathway (Li et al., 2012;Pastuhov et al., 2015). We next investigated where in this pathway SDZ-33 functions during axon regeneration. Activation of the JNK cascade following axonal injury is mediated by the SVH-2 Met-like RTK (Li et al., 2012). We confirmed that SDZ-33 functions in the same pathway as SVH-2 because the phenotype of sdz-33(tm1210); svh-2(tm737) double mutants was indistinguishable from that of either single mutant (Fig. 3A, Table 3). We examined whether overexpression of svh-2 might reverse the defect in axon regeneration observed in sdz-33 mutants. We found that this was indeed the case (Fig. 3A, Table 3). These results suggest that SDZ-33 functions at or upstream of SVH-2 in the SVH-2-JNK pathway.
Our genetic analysis of mdl-1 and sdz-33 above raised the possibility that SDZ-33 could act as a specific subunit in an SCF complex to mediate ubiquitylation of MDL-1, leading to its degradation. Since an F-box protein functions to target substrate proteins, we first examined whether SDZ-33 interacts with MDL-1 by cotransfecting mammalian HEK293 cells with Flag-tagged SDZ-33 and Myc-tagged MDL-1. Coimmunoprecipitation experiments revealed that Flag-SDZ-33 associated with Myc-MDL-1 (Fig. 4C). We next determined whether MDL-1 can be degraded by the SDZ-33-mediated 26S proteasome pathway in mammalian COS-7 cells. For this purpose, we coexpressed Myc-MDL-1 with HA-tagged Ub in COS-7 cells. Cell lysates were immunoprecipitated with anti-Myc antibody and immunoblotted with anti-HA antibody. We found that MDL-1 was poly-ubiquitylated ( Fig. 4D), suggesting that endogenous mammalian E3-Ub ligases are capable of ubiquitylating MDL-1. We then asked whether SDZ-33 expression could cause ubiquitylation of MDL-1. COS-7 cells were cotransfected with Myc-MDL-1, HA-Ub, and Flag-SDZ-33. This assay assumes that C. elegans SDZ-33 interacts with the endogenous COS-7 SCF complex and provides substrate specificity for MDL-1. We found that coexpression of Flag-SDZ-33 decreased the levels of poly-ubiquitylated MDL-1 protein (Fig. 4D), suggesting that SDZ-33 promotes the degradation of ubiquitylated MDL-1. Consistent with this possibility, when cells were treated with MG132, a specific inhibitor of the 26S proteasome, the amounts of poly-ubiquitylated MDL-1 protein clearly increased (Fig. 4D). Thus, proteasome-mediated degradation contributes to the control of MDL-1 protein levels by SDZ-33. Together, the above results indicate that SDZ-33 is an F-box protein that directly targets MDL-1 for Ub-mediated degradation.
We next investigated whether SDZ-33 regulates MDL-1 protein levels in animals expressing GFP-tagged MDL-1 from the unc-25 promoter. As observed previously , in wild-type animals MDL-1::GFP was found predominantly in the nucleus of D neurons, and its fluorescence intensity significantly decreased following axon injury (Fig. 5A,B). In contrast, we found that the sdz-33(tm1210) mutation resulted in significant stabilization of nuclear MDL-1::GFP levels, which persisted even 6 h after axotomy (Fig. 5A,  B). These results provide evidence that SDZ-33 is involved in axon injuryinduced degradation of MDL-1.

Discussion
In C. elegans, svh-2 expression is upregulated in neurons following axon injury. This gene induction is mediated by a transcription factor complex composed of ETS-4 and CEBP-1 (Li et al.,  2015). We have recently found that TDPT-1 inhibits axon injury-induced svh-2 expression by inducing the SUMOylation of ETS-4, which inhibits formation of an ETS-4-CEBP-1 complex . We have additionally identified a transcription factor, MXL-1, that acts as a positive regulator of svh-2 expression in response to axon injury. However, in contrast to MXL-1, its partner transcription factor MDL-1 negatively regulates svh-2 induction. Based on these findings, we propose the following molecular mechanism of how axon injury activates svh-2 expression (Fig. 6). In the absence of axon damage, TDPT-1 maintains the SUMOylation of ETS-4, resulting in the repression of its transcriptional activity. Axon injury leads to the degradation of MDL-1, allowing free MXL-1 to interact with TDPT-1. This causes the release of ETS-4 from TDPT-1, resulting in ETS-4 deSUMOylation and consequent complex formation with CEBP-1. This ETS-4-CEBP-1 complex then induces svh-2 expression. By this model, there must be some mechanism by which axon injury induces the degradation of MDL-1. Here, we show that the sdz-33 gene encoding an F-box protein is required for induction of svh-2 gene expression following axon injury. F-box proteins serve as substrate recruitment factors for the SCF E3-Ub ligase (Craig and Tyers, 1999). SDZ-33 promotes poly-ubiquitylation of MDL-1, leading to its subsequent destruction by the 26S proteasome (Fig. 6). Deletion of sdz-33 stabilizes MDL-1, which accumulates and is able to repress svh-2 gene expression in response to axon injury. The ubiquitylation of target proteins is orchestrated by a Ub-activating enzyme (E1), a Ub-conjugating enzyme (E2), and an E3-Ub ligase, the latter of which is responsible for substrate recognition and Ub transfer. One E3-Ub ligase is the SCF complex, which is composed of Skp1, Cullin1, Rbx1, and an F-box protein (Craig and Tyers, 1999). Cullin functions as a scaffold that interacts with Skp1 and Rbx1 to form a complex generating the core E3-Ub ligase activity. F-box proteins selectively recruit target proteins and thereby mediate Ub interactions with their substrates. Thus, the F-box protein determines substrate specificity. F-box proteins contain an N-terminal F-box domain that binds to Skp1 to enable ubiquitylation of the target. F-box proteins often contain other motifs such as a WD40 repeat and a leucine-rich repeat, both of which are involved in substrate binding (Craig and Tyers, 1999). Identifying target substrates for F-box proteins would enhance our understanding of how SCF E3-Ub ligase mediates multiple biological functions. The C. elegans genome contains ;700 genes encoding F-box proteins (Thomas, 2006), but only a few have been successfully assigned a function (Clifford et al., 2000;Kipreos et al., 2000;Liao et al., 2004;Fielenbach et al., 2007;Bounoutas et al., 2009;Wu et al., 2017). Here, we identify SDZ-33 as an F-box E3-Ub ligase that promotes the degradation of MDL-1. SDZ-33 lacks WD40 or leucine-rich repeats, but it does contain an FBA domain, which may be involved in the interaction with MDL-1. The C. elegans genome encodes .20 Skp1-related proteins and 6 Cullin proteins (Nayak et al., 2002). It would be interesting to determine which Skp1-related proteins or Cullin proteins are involved in SDZ-33-mediated degradation of MDL-1.
In the mammalian Myc/Mad/Max family of transcription factors, Myc forms a heterodimer with its partner Max. This Myc-Max complex then binds to the promoter region of target genes to activate their transcription. Mad1 suppresses Myc transcriptional activity by interacting with Max and inhibiting Myc-Max complex formation (Amati and Land, 1994). The C. elegans mxl-1 and mdl-1 genes encode homologs of vertebrate Max and Mad, respectively, but C. elegans lacks an obvious homolog of Myc (Pickett et al., 2007). MXL-1 has been shown to form a complex with MDL-1 and TDPT-1, and MDL-1 protein levels are tightly regulated in response to axon injury . Therefore, the availability of MXL-1 to TDPT-1 SUMOylation is dependent on MDL-1 protein expression levels. The association of MXL-1 with TDPT-1 redirects the MXL-1-MDL-1 complex to regulate ETS-4. The results presented here suggest that Ubmediated degradation regulates the Max-Mad network by promoting Mad degradation (Fig. 6). This work thus reveals an important module of this transcriptional regulatory system and illustrates the complex nature of the Max transcription factor network.