Two Paralogous Tetraspanins TSP-12 and TSP-14 Function with the ADAM10 Metalloprotease SUP-17 to Promote BMP Signaling in Caenorhabditis elegans

The highly conserved bone morphogenetic protein (BMP) signaling pathway regulates many developmental and homeostatic processes. While the core components of the BMP pathway have been well studied, much research is needed for understanding the mechanisms involved in the precise spatiotemporal control of BMP signaling in vivo. Here, we provide evidence that two paralogous and evolutionarily conserved tetraspanins, TSP-12 and TSP-14, function redundantly to promote BMP signaling in C. elegans. We further show that the ADAM10 (a disintegrin and metalloprotease 10) ortholog SUP-17 also functions to promote BMP signaling, and that TSP-12 can bind to and promote the cell surface localization of SUP-17. SUP-17/ADAM10 is known to be involved in the ligand-induced proteolytic processing of the Notch receptor. We have evidence that the function of SUP-17, and of TSP-12/TSP-14 in BMP signaling is independent of their roles in Notch signaling. Furthermore, presenilins, core components of the γ-secretase complex involved in processing Notch, do not appear to play a role in BMP signaling. These studies established a new role of the TSP-12/TSP-14/SUP-17 axis in regulating BMP signaling, in addition to their known function in the Notch signaling pathway. We also provide genetic evidence showing that a known BMP signaling modulator, UNC-40/neogenin/DCC, is one of the substrates of SUP-17/ADAM10 in the BMP signaling pathway.


Introduction
The highly conserved bone morphogenetic protein (BMP) pathway is repeatedly used in metazoan development to regulate multiple distinct processes in different cellular contexts. In canonical BMP signaling, secreted BMP ligands bind to the heteromeric type I/type II receptor complexes and induce the phosphorylation of type I receptors by type II receptors, which in turn phosphorylate the R-Smads. The phosphorylated R-Smads then complex with Co-Smad and enter the nucleus to regulate gene transcription. Increasing evidence has shown that BMP signaling is tightly regulated spatiotemporally and that misregulation of this pathway can cause many different disorders in humans, such as cardiovascular diseases and cancers [1][2][3][4][5]. Recent studies have identified a number of factors that modulate BMP signaling at the level of the ligand-receptor complex [6][7][8][9][10]. In particular, we have previously shown that the small transmembrane tetraspanin proteins are important in promoting BMP signaling in C. elegans [7].
Tetraspanins are a large group of integral membrane proteins with a characteristic protein topology. They contain four transmembrane domains and a number of conserved residues, such as the featured CCG motif in the second extracellular loop [11]. Recent studies have shown that a subgroup of tetraspanins, the TspanC8 family of tetraspanins, functions to promote another highly conserved signaling pathway, the Notch pathway [12][13][14][15]. In particular, the TspanC8 tetraspanins physically associate with ADAM10 and promotes its cell surface localization. ADAM10, in turn, cleaves the Notch receptor in a process called "ectodomain shedding", releasing the intracellular domain of Notch for subsequent activation of downstream gene expression [12][13][14]. Previously, we have shown that in C. elegans, the TspanC8 tetraspanins TSP-12 and TSP-14 function in modulating BMP signaling [7]. However, whether they exert their functions in BMP regulation by acting through the C. elegans ADAM10 protein is not known. Similarly, because of their roles in both BMP signaling and Notch signaling, it is not clear whether TSP-12 and TSP-14 regulate these two pathways independently. In this study, we investigated how TSP-12 and TSP-14 regulate BMP signaling in C. elegans.
The C. elegans BMP-like pathway is called the Sma/Mab pathway. This pathway is known to regulate body size, male tail patterning and mesoderm development [16]. Core members of this pathway include the ligand DBL-1/BMP, the type I and type II receptors, SMA-6/RI and DAF-4/RII, the R-Smads SMA-2 and SMA-3, as well as the Co-Smad SMA-4 [16]. Loss-offunction mutations in any of these core members will result in a small body size, male tail sensory ray formation defects [17], reduced expression of the RAD-SMAD reporter [18], and suppression of the sma-9(cc604) M lineage defect [19]. In particular, mutations in the C. elegans schnurri homolog sma-9 cause a dorsal-to-ventral cell fate transformation in the postembryonic M lineage, thus the loss of two M-derived coelomocytes (CCs). Loss-of-function mutations in any core members of the Sma/Mab pathway can suppress this M lineage defect (Fig 1A and  1B), resulting in the reappearance of the two M-derived CCs. Thus, Sma/Mab pathway mutants exhibit a Susm (suppression of the sma-9 M lineage) phenotype ( [19]; Fig 1A and 1E). We have previously shown that the Susm phenotype can be used as a specific and sensitive assay to screen efficiently for mutants with defects in Sma/Mab signaling [7]. Using the Susm assay, we have identified several highly conserved factors that function to modulate BMP signaling at the level of the ligand-receptor complex. These include the RGM protein DRAG-1 [20], the neogenin/DCC homolog UNC-40 [18], the tetraspanin TSP-21 [7], as well as two redundant tetraspanins, TSP-12 and TSP-14.
Here, we show that TSP-12 and TSP-14 function redundantly in BMP signaling by regulating the cell surface localization of the ADAM10 (a disintegrin and metalloprotease 10) ortholog SUP-17. We demonstrate that the function of SUP-17 in BMP signaling is independent of its well-established role in Notch signaling. Finally, we provide genetic evidence indicating that a known BMP modulator, UNC-40/neogenin/DCC, is one of the substrates of SUP-17/ ADAM10 in the BMP signaling pathway.
TSP-12 is widely expressed and is localized to both the cell surface and the cytoplasm in multiple cell types To determine how TSP-12 functions to promote BMP signaling, we determined the expression and subcellular localization patterns of TSP-12. We tagged the endogenous TSP-12 with GFP at either the N-terminus or at the C-terminus using CRISPR/Cas9-triggered homologous recombination ( [21,22]; see Materials and Methods). Both tagged proteins are fully functional because when introduced into the tsp-14(0) background, neither GFP::TSP-12 nor TSP-12:: GFP caused any defects observed for tsp-12(0); tsp-14(0) double mutants as described above. Additionally, both GFP::TSP-12 and TSP-12::GFP exhibited the same expression and localization pattern. As shown in Fig 2, GFP::TSP-12 is widely expressed from early embryos through larval development to adults in multiple cell types, which include the vulva precursor cells, the hypodermis, and the germline (Fig 2A to 2H). Within a cell, GFP::TSP-12 is both localized to the cell surface and in the cytoplasm in what appears to be small vesicles (Fig 2A, 2E and 2G). Because the fluorescence signal for the endogenously tagged GFP::TSP-12 in somatic cells is faint, we also generated integrated transgenic lines over-expressing a functional, C-terminally tagged TSP-12::GFP transgene (see Materials and Methods). In these animals, we detected TSP-12::GFP in the pharynx, intestine, neurons and M lineage cells (Fig 2I to 2R). The cell surface TSP-12::GFP in the intestinal cells appears to be restricted to the basolateral membrane but absent in the apical membrane (Fig 2K and 2M).
TSP-12 functions in the signal-receiving cells to promote BMP signaling in regulating body size and mesoderm development It has been well established that the Sma/Mab pathway receptors and downstream Smad proteins function in the hypodermal and M lineage cells to regulate body size and mesoderm development, respectively [19,23,24]. Since tsp-12 is expressed in the hypodermal and M lineage cells, we asked whether it functions in these cells to promote Sma/Mab signaling. We forced the expression of tsp-12 cDNA in specific tissues of the tsp-12(0); tsp-14(0) double mutants using tissue specific promoters. Forced expression of the tsp-12 cDNA in the hypodermis (elt-3p::tsp-12 cDNA and rol-6p::tsp-12 cDNA) rescued the small body size of tsp-12(0); tsp-14(0) double mutants, but forced expression in the pharynx (myo-2p::tsp-12 cDNA), intestine (elt-2p::tsp-12 cDNA) or neurons (rab-3p::tsp-12 cDNA) had no effect (Fig 2S). Forced expression of tsp-12 in the M lineage (hlh-8p::tsp-12 cDNA) partially rescued the Susm phenotype of tsp-12(0); tsp-14(0) double mutants: transgenic tsp-12(0); tsp-14(0) mutants carrying the  The C. elegans ADAM10 ortholog SUP-17 functions in the BMP signaling pathway TSP-12 and TSP-14 orthologs are known to regulate the maturation and trafficking of ADAM10 (a disintegrin and metalloprotease 10) to promote Notch signaling in Drosophila and mammalian cells [13]. We therefore tested whether the C. elegans ADAM10 ortholog, SUP-17 [25], and its close homolog ADM-4 [26], play any role in Sma/Mab signaling using the sma-9 M lineage suppression assay. A deletion allele of adm-4, ok265, failed to suppress the sma-9 M lineage defect; instead, two strong loss-of-function alleles of sup-17, n1258 and n316, both partially suppressed the sma-9 M lineage defect (Fig 3A and Table 2A; sup-17 null mutants are embryonic lethal [25] and cannot be used in this assay). Both sup-17(n1258) and    We then performed double mutant analysis between sup-17(n1258) and null mutations in the Sma/Mab pathway by assaying for their body sizes. We found that sup-17(n1258); dbl-1 (wk70), sup-17(n1258); sma-6(jj1) and sup-17(n1258); sma-3(jj3) double mutants were as small as the corresponding dbl-1(wk70), sma-6(jj1) or sma-3(jj3) single mutants, respectively ( Fig  3F). This lack of enhancement of the small body size phenotype of Sma/Mab pathway mutants by sup-17(n1258) is not because Sma/Mab pathway mutants have reached the minimal body size of C. elegans, because previous studies have shown that mutations in genes regulating body size but not functioning in the Sma/Mab pathway, such as sma-5 (encoding a MAPK, [27] or wts-1 (encoding a serine/threonine kinase, [28]), can make Sma/Mab pathway mutants much smaller. We therefore concluded that SUP-17 functions in the Sma/Mab pathway to regulate body size.

SUP-17 is expressed and functions in the signal-receiving cells to promote BMP signaling in in regulating body size and mesoderm development
To determine how SUP-17 functions in modulating Sma/Mab signaling, we tagged the endogenous SUP-17 by generating SUP-17::GFP knock-in strains using CRISPR/Cas9-triggered homologous recombination (see Materials and Methods). The SUP-17::GFP fusion is functional, because a transgene with the same SUP-17::GFP fusion rescued the small body size and Susm phenotypes of sup-17(n1258) mutants (Fig 4L). Like TSP-12, SUP-17::GFP is also widely expressed from the early embryo through larval development to adults. Cells expressing SUP-17::GFP include the germline, developing vulva, hypodermis, M lineage cells and intestinal cells (Fig 4B to 4M). Also like TSP-12, SUP-17::GFP is localized both at the cell surface and in the cytoplasm in small puncta (Fig 4B, 4D, 4F and 4H). In the early embryo, cell surface localized SUP-17::GFP is enriched in the basolateral, but not the apical, surfaces (Fig 4B). Forced expression of sup-17 cDNA in the hypodermis (rol-6p::sup-17 cDNA), but not in the intestine (elt-2p::sup-17 cDNA), rescued the small body size of sup-17(n1258) mutants (Fig 4N). Similarly, forced expression of sup-17 cDNA in the M lineage (hlh-8p::sup-17 cDNA) partially rescued the Susm phenotype of sup-17(n1258) mutants (Table 2A). These findings indicate that SUP-17 also functions in the signal-receiving cells to modulate Sma/Mab signaling in regulating body size and mesoderm development.

TSP-12 can bind to SUP-17 and promote its cell surface localization
The shared expression/localization pattern and function of TSP-12, TSP-14 and SUP-17 in Sma/Mab signaling suggest that TSP-12 and TSP-14 may be functionally linked with SUP-17 in their roles in the Sma/Mab pathway. Using the mating-based split ubiquitin yeast-twohybrid assay for membrane protein interactions (see Materials and Methods), we found that SUP-17 can bind to TSP-12 and TSP-14 ( Fig 5A). We then examined the localization of SUP-17::GFP in tsp-12(0) and tsp-14(0) single mutants, and in tsp-12(0); tsp-14(0) double mutants. We focused on the early embryo for these localization studies because of the large cell sizes and the brighter GFP signals in these embryos. Wild-type embryos expressing the endogenously-tagged SUP-17::GFP showed GFP localization both at the cell surface and in the cytoplasm (Fig 5B and 5F). This localization pattern does not change in tsp-14(0) mutant embryos (Fig 5C and 5F). However, in tsp-12(0) or tsp-12(0); tsp-14(0) mutant embryos, there is a significant reduction of cell surface SUP-17::GFP accompanied by an increase of intracellular SUP-17::GFP (Fig 5D, 5E and 5F) and an increase in the number of bright puncta in the cytoplasm (Fig 5G). These findings indicate that TSP-12 and TSP-14 can bind to SUP-17 and that TSP-12, but not TSP-14, promotes the cell surface localization of SUP-17 in embryos.  The two C. elegans presenilins SEL-12 and HOP-1 do not appear to be required for Sma/Mab signaling The involvement of SUP-17/ADAM10 in Sma/Mab signaling prompted us to test whether the γ-secretase functions in Sma/Mab signaling, because Notch undergoes sequential cleavage by ADAM and γ-secretase [29]. C. elegans has three paralogous genes encoding presenilins, sel-12, hop-1 and spe-4 [30][31][32], and presenilins are major components of the γ-secretase complex. Because spe-4 appears to be specifically expressed and functions in the germline during spermatogenesis, and because sel-12 and hop-1 share redundant functions, in particular in Notch signaling [30][31][32], we tested whether mutations in sel-12 and hop-1 exhibited any Susm phenotypes. Using null alleles for each gene, we found that sel-12(ar171), sel-12(ok2078), hop-1 (ar179) single mutants or hop-1(ar179); sel-12(ar171) and hop-1(ar179); sel-12(ok2087) double mutants did not exhibit any Susm phenotype (Table 2B), suggesting that the γ-secretase is unlikely to be involved in Sma/Mab signaling in C. elegans.
The neogenin homolog UNC-40 is one of the substrates of SUP-17/ ADAM10 in Sma/Mab signaling ADAM10 is known to cleave transmembrane or membrane-associated proteins in a process called ectodomain shedding [33,34]. We have previously shown that the extracellular domain of UNC-40 is sufficient to function in promoting Sma/Mab signaling, suggesting that UNC-40 may be cleaved for its function in Sma/Mab signaling [18]. We therefore tested genetically whether UNC-40 is a SUP-17/ADAM10 substrate. We reasoned that if UNC-40 is a SUP-17 substrate in BMP signaling, then the extracellular domain of UNC-40, which is sufficient to function in promoting Sma/Mab signaling, should be able to, either completely or partially, rescue the Sma/Mab signaling defects of sup-17 mutants. We, therefore, tested whether two unc-40 alleles that lead to the production of truncated UNC-40 proteins lacking the intracellular domain (ICD), ev495 and tr115 [18,35], could rescue the Susm phenotype of sup-17(n1258) mutants. As shown in Table 2C, a null unc-40 allele, e1430, did not rescue the Susm phenotype of sup-17(n1258) mutants, while both unc-40(ev495) and unc-40(tr115) partially suppressed the Susm phenotype of sup-17(n1258) mutants. The similar penetrance of the Susm phenotype shared by unc-40(e1430) and unc-40(e1430) sup-17(n1258) mutants, and the partial suppression of the sup-17(n1258) Susm phenotype by unc-40(ev495) and unc-40(tr115) suggest that: 1) UNC-40 and SUP-17 function in the same pathway to regulate Sma/Mab signaling, and 2) UNC-40 is a substrate of SUP-17 in the Sma/Mab pathway. The incomplete suppression of the Susm phenotype of sup-17(n1258) mutants by unc-40(ev495 or tr115) suggests that additional SUP-17 substrates must exist in the Sma/Mab pathway.

Discussion
The TspanC8 tetraspanins and ADAM10 positively promote BMP signaling independent of their roles in Notch signaling In this study, we extended our previous work [7] and provided additional evidence that TSP-12 and TSP-14, two paralogous TspanC8 tetraspanins, function redundantly to promote BMP signaling in C. elegans. We further showed the ADAM10 ortholog SUP-17 also functions in the BMP pathway. Both tsp-12; tsp-14 double mutants and sup-17 single mutants exhibited BMP signaling defects, including a smaller body size, male tail abnormalities, reduced RAD-S-MAD reporter activity, and suppression of the sma-9 mesoderm phenotype (Susm). Both TSP-12 and SUP-17 are expressed and function in the signaling-receiving cells to promote BMP signaling in regulating body size and M lineage development. TSP-12 and TSP-14 can bind to SUP-17. TSP-12 and SUP-17 share similar expression and subcellular localization patterns. Additionally, TSP-12 promotes the cell surface localization of SUP-17 in embryos.
Over-expression and knockdown experiments in Drosophila and mammalian tissue culture cells have shown that TspanC8 tetraspanins promote the cell surface localization of ADAM10 [12][13][14]. Our work demonstrates that this function of the TspanC8 tetraspanins is evolutionarily conserved. How TspanC8 tetraspanins regulate the cell surface localization of ADAM10 requires further investigation. Tetraspanins are known to regulate intracellular trafficking of various cargo proteins via multiple distinct mechanisms [11,41]. Previous work showed that TspanC8 tetraspanins function to promote the exit of ADAM10 from the endoplasmic reticulum (ER) [12][13][14]. We noted in our work that SUP-17::GFP from transgenic lines overexpressing this transgene exhibited a strong perinuclear, ER-like localization pattern with very little cell surface localization (Fig 4J and 4K). However, endogenously-tagged SUP-17::GFP exhibited much more prominent cell surface localization (Figs 4B, 4D, 4F, 4H and 5B). We showed that this cell surface localization is significantly reduced in tsp-12(0) and tsp-12(0); tsp-14(0) mutants; while there was a concomitant increase of bright SUP-17::GFP-positive puncta in these mutant embryos (Fig 5D to 5G). Whether these signals represent SUP-17::GFP trapped in the ER or in other intracellular organelles will need further investigation.
The basolateral plasma membrane localization of TSP-12 in intestinal cells (Fig 2K and 2N) resembles the localization patterns of both SMA-6/RI and DAF-4/RII of the BMP pathway [6]. We have previously shown that both TSP-12 and TSP-14 can bind to SMA-6 in yeast [7]. Whether TSP-12, TSP-14 or both of them are required for the proper localization of SMA-6 and DAF-4 will require further investigation.
Our phenotypic analysis clearly showed that TSP-12 and TSP-14 function redundantly to promote BMP signaling (Fig 1F and 1K). Both proteins can also bind to SUP-17 in yeast ( Fig  5A). However, only tsp-12(0) mutant embryos showed severe defect in SUP-17::GFP localization, and this defect is not enhanced in tsp-12(0); tsp-14(0) double mutant embryos (Fig 5D  to 5G). Due to technical challenges, our SUP-17::GFP localization study was carried out in embryos but not in tissues relevant to Sma/Mab signaling. It is possible that TSP-14 may function similarly as TSP-12 to promote the cell surface localization of SUP-17 in other cell types, but not in the early embryo. Alternatively, it is possible that while TSP-12 and TSP-14 exhibit functional redundancy in regulating BMP signaling and other developmental processes at the organismal level, they may differ in their functions at the sub-cellular and/or molecular level.

The role of SUP-17/ADAM10 in the BMP signaling pathway
In worms, flies and mammals, ADAM proteases are well known for their role in processing the Notch receptor upon ligand binding [42][43][44]. Notch undergoes sequential cleavage, first by the ADAM protease, followed by subsequent cleavage by the γ-secretase. Although ADAM10 has not been previously implicated in BMP signaling in any organisms, the ADAM10 paralog ADAM17 or TACE (tumor necrosis factor-alpha-converting enzyme) has been shown to play a role in TGFβ signaling in mammals [45]. Earlier work by Liu and colleagues indicated that ADAM17/TACE-mediated cleavage of the type I TGFβ receptor (TβR1) causes downregulation of TGFβ signaling in CHO cells [45]. Cleavage of TβR1 by ADAM17/TACE and Presenilin 1, a γ-secretase catalytic core component, has also been found in multiple cancer cell lines, which results in the translocation of the intracellular domain (ICD) of TβR1 into the nucleus and activation of genes involved in tumor invasion [46,47]. More recently, polymorphic variants of the Adam17 gene have been found to differentially regulate TGFβ signaling and influence the severity of Tgfb1-dependent vascular pathology in mice and humans [48]. Because ADAM10 and ADAM17 are known to share a number of common substrates [34], and in C. elegans SUP-17/ADAM10 and ADM-4/ADAM17 share redundant functions [26], our findings raise the possibility that ADAM10 or ADAM17 may also play a role in regulating BMP signaling in mammals. Consistent with this notion, Jackson and colleagues reported that ADAM17/ TACE-null mice exhibit defects in valvulogenesis that are associated with aberrant BMP signaling [49].
How SUP-17/ADAM10 functions to modulate BMP signaling is an open question. Our genetic studies suggest that the neogenin/DCC ortholog UNC-40 is one of the substrates of SUP-17. UNC-40, like its vertebrate ortholog neogenin, functions through the RGM (repulsive guidance molecule) proteins to promote BMP signaling [10,18]. We have previously shown that the extracellular domains of UNC-40 and neogenin are sufficient to mediate BMP signaling in C. elegans and in mammalian tissue culture cells, respectively [18]. There are other reports showing that neogenin or DCC can be cleaved by ADAM proteases. Earlier work using ADAM inhibitors and rat dorsal spinal cord explants showed that DCC undergoes ADAMdependent proteolytic processing, which may affect its role in regulating axon migration [50]. Okamura and colleagues showed that ADAM17/TACE can bind to and cleave neogenin, regulating the behavior of rat cortical neurons in response to RGM [51]. Recently, van Erp and colleagues showed that ADAM17/TACE-mediated cleavage of neogenin is regulated by the leucine-rich repeats and immunoglobulin-like domains (Lrig) protein Lrig2, and that this regulation is required for RGM-mediated neurite growth inhibition in vitro and cortical neuron migration in vivo [52]. How SUP-17 cleaves UNC-40 and whether similar types of regulatory interactions also occur in the BMP signaling pathway will require further experimentation. Nevertheless, cleavage of UNC-40 by SUP-17 may not be followed by subsequent cleavage by the γ-secretase, because hop-1(0); sel-12(0) double mutants, which lack two of the redundant C. elegans presenilins, do not exhibit any BMP signaling defects ( Table 2). We are aware that we assayed the Susm phenotype in hop-1(0); sel-12(0) animals produced by hop-1(0)/+; sel-12(0) mothers due to the sterility of the hop-1(0); sel-12(0) double mutants. Thus, although unlikely, we cannot rule out the possibility that residual γ-secretase activity is still present in hop-1(0); sel-12(0) double mutants due to maternally loaded HOP-1 and that this residual activity is sufficient for proper Sma/Mab signaling.
Our genetic studies suggested that UNC-40/neogenin is not the only substrate of SUP-17/ ADAM10 in the BMP pathway. Identifying the additional substrate(s) and determining the functional consequences of ADAM processing of these substrates in vivo will provide insight into the intricate levels of modulation of BMP signaling. Notably, ADAM10 and ADAM17 mutations are associated with many diseases in humans, including a variety of cancers [34,53,54]. We speculate that some of these incidences could be due to abnormal processing of factors involved in BMP signaling caused by altered activities of ADAM10 and/or ADAM17.

C. elegans strains
All strains were maintained at 20˚C using standard culture conditions [55]

Body size measurement and RAD-SMAD reporter assay
Body size measurement and RAD-SMAD reporter assays were performed following the protocols described in Tian et al. [18]. Hermaphrodite worms at the L4 stage (identified based on vulva development) were used for body size measurement. Body sizes were measured using segmented lines in Fiji [56] after images were taken. Synchronized L3 hermaphrodites carrying the RAD-SMAD reporter jjis2437 II [18] were used for imaging. The reporter fluorescence intensity was measured using Fiji.

Oligonucleotides, plasmid constructs and transgenic lines
Oligonucleotides used in this study are listed in SI Appendix, S2 Table. Plasmids generated are listed in SI Appendix, S3 Table. The tsp-12 cDNA was amplified from the Vidal RNAi library [57], while yk23h2 contains the full-length sup-17 cDNA. Each clone contains a point mutation in the respective coding regions of tsp-12 and sup-17, which were respectively fixed via sitedirected mutagenesis, resulting in the generation of pJKL1079 (containing the tsp-12 ORF), and pJKL1140 (containing the sup-17 cDNA). pK3, a rescuing plasmid containing 7kb of the sup-17 genomic region, is a kind gift from Iva Greenwald (Columbia University).
Generating endogenously tagged TSP-12::GFP or SUP-17::GFP strains using CRISPR/Cas9-mediated homologous recombination We tagged the endogenous TSP-12 at either the N-terminus or the C-terminus using CRISPR/ Cas9. To generate these TSP-12 knock-ins, we followed the strategy described in Arribere et al. [21] to generate two sgRNA expression plasmids for each tagging strategy (see S2 and S3 Tables for oligo and plasmid information). We then used the method described in Dickinson et al. [22] to generate the homologous repair templates, and to subsequently obtain stable knock-in strains that carry either GFP::3×FLAG::TSP-12 (jj181 and jj182) or TSP-12:: GFP::3×FLAG (jj194 and jj196). We first confirmed the success of knock-ins by genotyping, and then tested jj181 and jj194 for functionality by crossing each of them into the tsp-14(0) background and examined the strains for viability and fertility.

Microscopy
To examine the male tail, the developing vulva or developing embryos, animals were visualized under a Leica DMRA2 compound microscope equipped with a Hamamatsu Orca-ER camera using the iVision software (Biovision technology, Inc.). An inverted Zeiss LSM880 confocal microscope with Airyscan was used to examine the expression and subcellular localization patterns of GFP::3×FLAG::TSP-12 or SUP-17::GFP. Fluorescence was detected with an Airyscan detector using the 40× objective (NA 1.4, WD 0.13 mm) and in the super-resolution mode (pinhole is around 2.25 AU). The laser line for GFP excitation is at 488 nm, and for mCherry is at 561 nm. Collected images were subsequently processed by ZEN (Carl Zeiss), and further quantification of the fluorescence signals was performed using Fiji.
To quantify the fluorescence intensity and bright intracellular puncta of SUP-17::GFP, live embryos from worms of different genotypes that were grown under the same condition were collected on the same day and imaged using the same settings as described above. Images of stage-matched embryos were further processed in the same way using Fiji. Average SUP-17:: GFP fluorescence intensity was calculated by drawing a straight line of fixed length (for cell surface signal) or by drawing an oval of fixed width and height (for intracellular signal). Only cells with focal plane going through the nucleus were used for this purpose. Bright puncta were defined as those signals that are above a threshold of brightness (3,500 pixels) and size (the area is above 0.07 μm 2 ). Signals on top of the cell membrane were excluded from the counting.

Split ubiquitin yeast two-hybrid assay
The split-ubiquitin yeast two-hybrid experiments were carried out by following the protocol described in Grefen et al. [61]. The bait CubPLV and prey NubG fusion constructs used are listed in SI Appendix, S2 Table. After transforming into yeast, interactions among each pair of bait and prey constructs were visualized by streaking diploid cells on SC-Trp, -Leu, -Ade, -His, -Ura, -Met plates that were supplemented with four different concentrations of methionine: 0mM, 0.075mM, 0.150mM, 0.300mM, respectively. Growth was monitored for 2-9 days at 30˚C. The plasmids KAT-1-Cub-PLV, KAT-1-NubG (in XN21 vector), empty NubG vector, and the vector expressing soluble wild-type Nub (NubWT) were used as controls [7]. Confirmation of expression of each fusion protein was by western blot analysis using rabbit polyclonal anti-VP16 antibodies (ab4808, Abcam, for CubPLV fusions) and monoclonal anti-HA antibodies (Clone 12CA5, Sigma, for NubG fusions).

Statistical analyses
Statistical analyses were conducted with Prism 5 (GraphPad), and the test for significance for groups of two were analyzed using the unpaired Student's t-test. Two-tailed p values were calculated. Data was shown as mean ± 95% confidence interval (CI).  Research Infrastructure Programs (P40 OD010440). The confocal imaging data was acquired through the Cornell University Biotechnology Resource Center, with NYSTEM (CO29155) and NIH (S10OD018516) funding for the shared Zeiss LSM880 confocal/multiphoton microscope.