Repulsive guidance molecule acts in axon branching in Caenorhabditis elegans

Repulsive guidance molecules (RGMs) are evolutionarily conserved proteins implicated in repulsive axon guidance. Here we report the function of the Caenorhabditis elegans ortholog DRAG-1 in axon branching. The axons of hermaphrodite-specific neurons (HSNs) extend dorsal branches at the region abutting the vulval muscles. The drag-1 mutants exhibited defects in HSN axon branching in addition to a small body size phenotype. DRAG-1 expression in the hypodermal cells was required for the branching of the axons. Although DRAG-1 is normally expressed in the ventral hypodermis excepting the vulval region, its ectopic expression in vulval precursor cells was sufficient to induce the branching. The C-terminal glycosylphosphatidylinositol anchor of DRAG-1 was important for its function, suggesting that DRAG-1 should be anchored to the cell surface. Genetic analyses suggested that the membrane receptor UNC-40 acts in the same pathway with DRAG-1 in HSN branching. We propose that DRAG-1 expressed in the ventral hypodermis signals via the UNC-40 receptor expressed in HSNs to elicit branching activity of HSN axons.

Axon branching is a fundamental process for proper axon projection to target tissues and for the formation of correct synapses, both of which are important for development of the functional neural network. Axon branching begins with the formation of an actin-rich filopodium from the existing axon followed by extension of microtubules along the actin filaments 1 . Formation of filopodia and subsequent neurite extension involve various regulators for actin and microtubule polymerization and bundling. The location and the polarity of axon branching are dictated by extracellular cues, along with cytoskeletal activities. Axon guidance molecules are involved in this process.
Netrin-1 is a secreted guidance molecule that induces local filopodial protrusions in the axon shaft, which give rise to branches in the cortical neurons. In contrast, SEMA3A represses cortical axon branching 2 . Ephrins are membrane-bound molecules that abolish branching of thalamic axons 3 . In addition to these well-known guidance molecules, repulsive guidance molecules (RGMs) also repress axon branching in cortical neurons and mossy fibers of the hippocampus 4,5 . In vertebrates, RGMs are glycosylphosphatidylinositol (GPI)-linked membrane proteins that constitute a family with four members: RGMa, RGMb (DRAGON), RGMc (hemojuvelin), and RGMd 6 . RGMa was first discovered as an axon guidance cue that has a repulsive activity to retinal axons. RGMa is expressed in the embryonic tectum in an anterior-to-posterior concentration gradient and functions during the development of the retinotectal projection 7 .
By binding to the membrane receptor neogenin, RGMs function in axon guidance and neuronal survival as well as inhibition of axonal regeneration [8][9][10] . RGMs also bind bone morphogenetic proteins (BMPs) in the regulation of iron homeostasis and endochondral bone development [11][12][13] . Although the function of RGMs in axon guidance as a result of growth cone repulsion has been well studied, their role in axon branching is still elusive. Because of the lethality of knock-out mice and the functional redundancy of RGM proteins, the functions of RGMs have been mostly analyzed using in vitro culture systems 14 .
The nematode C. elegans has a single ortholog of RGM, DRAG-1. Loss-of-function mutations in drag-1 result in a small body size as well as genetic suppression of the mesodermal coelomocyte loss phenotype of sma-9 mutants 15 . drag-1 function in the hypodermal cells is required for the control of body size, while drag-1 function in the mesodermal M cell lineage controls coelomocyte differentiation 15,16 . In the present study, we showed that drag-1 expressed in the hypodermis functions in the formation of branches of hermaphrodite-specific neurons (HSNs), which innervate egg-laying muscles of the vulva 17 . Small body size mutants sma-1 and sma-5 also exhibited HSN axon branching defects although they act in genetic pathways distinct from that of drag-1.

Figure 1.
Gene structure and mutant phenotype of drag-1 animals. (a) Structure of drag-1 and mutation sites of the tk81 and tm3773 alleles. The exon and intron regions were determined by sequencing cDNA generated from isolated drag-1 mRNA. Black, yellow, blue, and magenta boxes indicate N-terminal signal peptide, partial von Willebrand factor type D domain, hydrophobic region, and C-terminal GPI-anchor signal sequence, respectively 15 . Bars depict the region of the cDNA used for expressing the antigenic peptide for producing antibodies and the respective mutation sites. tk81 is a 494-bp deletion within exon 3, which is expected to produce a truncated polypeptide that is missing the C-terminal 278 amino acids. tm3773 is an 892-bp deletion spanning from intron 2 to exon 3 (WormBase). (b) Body length phenotype of drag-1(tk81) mutants. Body length of young adult hermaphrodites. tk81 mutants had shorter bodies compared with wild type. Anterior is to the left. Scale bar: 50 μm. (c) Quantification of body length of young adult hermaphrodites for wild-type and drag-1 mutant animals. Significant difference was determined by Student's t-test. ***P < 0.001. n = 60 for each strain.  (Fig. 3a). We found HSN branching defects similar to that observed in the drag-1 mutants in sma-1(e30), sma-5(n678), and sma-8(e2111) (Fig. 3b). sma-1 and sma-5 encode β H -spectrin and MAP kinase, respectively 22,23 . sma-8(e2111) is a dominant mutation for which the causative gene has not yet been identified. We produced double mutants between drag-1 mutants and these sma mutants and found that all double mutants exhibited HSN branching defects stronger than those observed in the respective single mutants (Fig. 3b). Because drag-1(tk81) and drag-1(tm3773) mutants are putative null alleles, these results suggested that sma-1 and sma-5 act in pathways different from that of drag-1 to regulate HSN branching. Because of the dominancy of the sma-8(e2111) mutation, the relationship between drag-1 and sma-8 remains to be determined.

GPI anchoring of DRAG-1 is important for HSN branching.
Because DRAG-1 is thought to be modified by a GPI-anchor, we placed the GFP or Venus coding region right upstream of the putative cleavage site of the C-terminal pro-peptide sequence (drag-1p::drag-1::GFP::GPI or drag-1p::drag-1::Venus::GPI) to keep the GPI-anchor signal intact. These constructs rescued the branching defect of the mutants (Fig. 4a,b). We examined whether the GPI anchoring is important for DRAG-1 to act in axon branching. A construct with the deleted C-terminal GPI-anchor signal sequence (drag-1p::drag-1::GFP::∆GPI) failed to rescue the branching defects of drag-1 mutants (Fig. 4a,b). We also examined a construct in which the C-terminal GPI-anchor signal was replaced by the transmembrane domain of the LIN-12/Notch receptor (drag-1p::drag-1::GFP::lin-12TM). This Branching phenotypes were scored using fluorescence microscopy. drag-1(tk81) and drag-1(tm3773) mutants were compared with those transgenic for drag-1p::drag-1. Significant differences were determined by Fisher's exact test. ***P < 0.001, **P < 0.01. n = 180 for each strain. www.nature.com/scientificreports/ chimeric protein, which is potentially localized to the plasma membrane, failed to rescue the mutant defects ( Fig. 4a,b). These results suggested that DRAG-1 needs to be anchored to the plasma membrane by the GPI anchor rather than being embedded as a transmembrane protein.

Expression of DRAG-1.
To understand how DRAG-1 functions, we examined its expression pattern using both transcriptional and translational reporters, as well as immunostaining using anti-DRAG-1 antibodies that we generated. Expression of a transcriptional drag-1p::Venus reporter was detected in the pharynx, intestine and the syncytial hypodermis from late embryos to adult stages. Notably, Venus expression was not detected in hypodermal seam cells and the vulval hypodermis (Fig. 5a). Unlike the transcriptional reporter described above, we could only detect Venus expression in the pharynx using the functional translational fusion construct drag-1p::drag-1::Venus::GPI (Fig. 5b). We therefore raised polyclonal antibodies against a DRAG-1 peptide corresponding to amino acids 130-368 (Fig. 1a). Immunostaining experiments were performed using the L4 to young adult stage animals in which HSN branches are formed 24 . The antibodies detected no signals in nontransgenic wild-type animals, but detected signals in the pharynx, intestinal cells, hypodermal seam cells, and ventral hypodermal cells (except in the vulval region) in animals transgenic for either drag-1p::drag-1 or drag-1p::drag-1::Venus::GPI. (Fig. 5c). The expression patterns were consistent with the observation using anti-GFP in animals expressing drag-1p::drag-1::GFP::GPI 15 . Thus, it is likely that the level of expression of endogenous , sma-6(wk7), and sma-9(wk55) mutants were compared with wild type. (b) drag-1(tk81) and drag-1(tm3773) mutants were compared with sma-1(e30), sma-5(n678), and sma-8(e2111) mutants and with double mutants consisting of drag-1(tk81) or drag-1(tm3773) in combination with individual sma mutations. Significant differences were determined by Fisher's exact test. ***P < 0.001, **P < 0.01, *P < 0.05. NS not significant. n = 180 for each strain. www.nature.com/scientificreports/ DRAG-1 is low. Alternatively, DRAG-1 expression might be upregulated only transiently in HSN branch formation. The hypodermal signals detected by the anti-DRAG-1 antibodies appeared in a granular pattern in the cytoplasm. This could be due to the accumulation of DRAG-1 in the ER or Golgi apparatus as a result of overexpression of drag-1. We detected no DRAG-1 expression in the nervous system including HSNs even in the over-expressed condition.

DRAG-1 functions in hypodermal cells for axon branching.
We have shown that drag-1 is important for axon branching, but drag-1 is not expressed in the nervous system. To determine the tissues in which DRAG-1 expression is important for axon branching, we expressed drag-1 under tissue-specific promoters. We found that hypodermal expression of DRAG-1 using the rol-6 promoter (rol-6p::drag-1) 25 rescued the branching defect, whereas expression in the pharyngeal muscle (myo-2p::drag-1) 26 or in the intestine (elt-2p::drag-1) 27 did not (Fig. 6). These results indicated that DRAG-1 functions in hypodermal cells to induce axon branching of the HSNs. Because we detected DRAG-1 expression in both ventral hypodermis and lateral hypodermal seam cells, we tested whether the expression in the seam cells could rescue the HSN branching defect of drag-1 mutants. drag-1 expression under the seam cell specific SCM promoter 28 failed to rescue the HSN branching defect (Fig. 6). We also attempted to specifically express drag-1 in the ventral hypodermis, but we failed to find an appropriate promoter for this purpose. To circumvent this problem, we drove the expression of drag-1 under the dab-1 promoter, which drives gene expression in the developing vulval precursor cells 29 . Interestingly, the dab-1p::drag-1 construct successfully rescued the branching defect (Fig. 6). Thus even though drag-1 is not normally expressed in the vulval precursor cells, forced expression of drag-1 in these cells is sufficient for HSN branching, again supporting a role of DRAG-1 functioning in hypodermal cells for axon branching.

drag-1 acts in the same pathway as unc-40 in HSN axon branching. Neogenin is a receptor for
RGMa for axonal growth cone guidance 8,30 and UNC-40 is the sole ortholog of neogenin in C. elegans. We examined the HSN axons of unc-40(e271) null allele. The axons of the unc-40 mutants mostly failed to extend to the ventral side and the branches were rarely observed at the vulva or other locations ( Supplementary Fig. S2, Table S1). However, we observed branching defects in unc-40(e271)/+heterozygotes with similar penetrance as in the drag-1 mutants. unc-40(e271)/+heterozygotes exhibited no abnormality in HSN axon trajectory (Supplementary Table S1). The branching defect in unc-40(e271)/+was not enhanced when combined with drag-1 null mutants (Fig. 7). Therefore, UNC-40 acts in the same genetic pathway as DRAG-1 in HSN axon branching.

Discussion
In this study, we have shown that the sole C. elegans repulsive guidance molecule DRAG-1 functions in hypodermal cells to regulate axon branching of the HSNs that form synapses with the vulval egg-laying muscles in hermaphrodites. We further showed that DRAG-1 functions in the same genetic pathway as the neogenin homolog UNC-40 in regulating axon branching. DRAG-1 has been previously shown to function in the BMP signaling pathway to regulate body size and mesoderm development 15 . sma-6 (BMP type I receptor); sma-2, -3, and -4 (Smads); and sma-9 (BMP antagonist schnurri) are components of the BMP signaling pathway in C. elegans. Mutations in these genes result in a small body size (Sma) phenotype. However, none of the sma mutants in the BMP pathway affected HSN axon branching. Instead, we observed HSN axon branching defects in sma-1(e30) and sma-5(n678) mutants, which are not involved in the BMP signaling. These results suggest that BMP signaling does not regulate HSN axon branching and that the function of DRAG-1 in regulating HSN axon branching is independent of BMP signaling.  -1(tk81) and drag-1(tm3773) mutants were compared with those transgenic for rol-6p::drag-1, myo-2p::drag-1, elt-2p::drag-1, SCMp::drag-1 and dab-1p::drag-1. Two independently isolated transgenic lines (#1 and #2) were examined for each transgene. Significant differences were determined by Fisher's exact test relative to the corresponding drag-1 mutants. **P < 0.01, *P < 0.05. NS not significant. n = 180 for each strain. www.nature.com/scientificreports/ Unlike BMP signaling, two genes known to function in regulating body size, sma-1 and sma-5, also play a role in regulating HSN axon branching. We have shown that SMA-1 and SMA-5 appear to act in pathways parallel to that of DRAG-1. sma-1 encodes β H -spectrin, which is a very large spectrin found in invertebrates such as C. elegans and Drosophila 22 . The submembrane skeletal network is primarily formed from α2β2 spectrin tetramers, each composed of two α-spectrin and two β-spectrin subunits 31 . The spectrin network interacts with peripheral actin filaments to act in synapse function, muscle sarcomere structure, and axonal outgrowth 32,33 . Although SMA-1 function in shaping cells in the hypodermis and pharyngeal muscles has been reported 34,35 , its function in neuronal cells is unknown. SMA-1 may function in HSNs for branch formation by regulating the actin filaments. SMA-5/MAP kinase is specifically expressed in the intestine to control intestinal tube stability and body size 23 . Because the intestine has no direct contact with HSNs, it is possible that SMA-5 indirectly affects branching of the neuron. Because 60% of HSNs produce at least one branch in the drag-1 null mutant background and 40-50% of HSNs still make branches even in drag-1; sma-1 or drag-1; sma-5 double mutants, additional mechanisms must exist to control branch formation in HSNs. These additional factors/pathways include KAL-1/anosmin-1 and EGL-17/fibroblast growth factor (FGF), which act through SAX-7/L1CAM and EGL-15/FGF receptor to regulate HSN branching 36 . Additionally, the immunoglobulin superfamily protein SYG-1, which is expressed in HSN, is required for HSN branching and appropriate synaptogenesis with egg-laying muscles 37 .
How does DRAG-1 functions to regulate HSN axon branching? We believe that hypodermally expressed DRAG-1 acts through the neogenin receptor UNC-40 that is expressed in the HSNs to regulate HSN axon branching. We found via immunostaining that DRAG-1 is expressed in hypodermal seam cells and in the ventral hypodermis with the exception of the vulval epithelium. Furthermore, the defective axon branching of drag-1 mutants can be rescued by expressing drag-1 under the hypodermis specific rol-6 promoter or under the vulval epithelial precursor cell-specific dab-1 promoter, but not the hypodermal seam cell-specific SCM promoter. The dab-1 promoter drives gene expression in the descendants of vulval precursor cells P5.p, P6.p and P7.p during the Pn.pxx stage and in the descendants of P5.p and P7.p during the Pn.pxxx stage 29 . It is during this period that the HSN axons contact with some of these vulval precursor cells, defasciculate from the ventral nerve cord and form branches 24,37,38 . Thus, it is likely that DRAG-1 ectopically expressed in the vulval precursor cells signals the HSN axon to induce branching. In wild type animals DRAG-1 is expressed in the ventral hypodermis, which is directly adjacent to the ventral nerve cord, and the HSN axons fasciculate with the ventral nerve cord twice at regions posterior and anterior to the vulva 21 . Thus, DRAG-1 from the hypodermal cells can bind to the neogenin receptor UNC-40 that is expressed in the HSNs to induce HSN branching in a non-cell-autonomous fashion. This is consistent with UNC-40 being an ortholog of vertebrate neogenin, which acts as a receptor for RGM proteins 30,39 . Moreover, DRAG-1 physically interacts with the extracellular domain of UNC-40 in C. elegans 16 , similar to the interaction observed between human RGMc and neogenin 40 . Neither DRAG-1 with its GPI-anchor sequence deleted (therefore as a potential secreted form) nor DRAG-1 fused with the LIN-12 transmembrane domain (therefore as a potential membrane-anchored form) rescued the axon branching defects of drag-1 mutants. The latter was unexpected because the same construct significantly rescues the drag-1 defect in the control of mesodermal cell differentiation 15 . With respect to mesodermal cell differentiation, DRAG-1 and UNC-40 are expressed and function in the same cells to promote BMP signaling. In HSN branching, however, they are expressed in different cells that contact each other. Thus, it is possible that the modes of signaling from DRAG-1 to UNC-40 may differ depending on the cellular context. Studies of vertebrate RGM and neogenin proteins have yielded similar findings showing that they either function in the same cells or in different cells to regulate different processes 14 .
In summary, we provide in vivo evidence that RGM proteins function to promote axon branching in C. elegans. Our finding contrasts with the observation that RGM proteins suppress the branching of axons in the mammalian brain 4,5 . RGMs may function in both ways depending on the tissues or the phases of organogenesis. Further research is needed to understand the precise function of RGMs in axon branching.
Production of transgenic animals. We  Production of antibodies. The RNA sample extracted from wild-type worms was treated with SuperScript III reverse transcriptase (Invitrogen) using a primer 5′-TCA GCA TAA CAA TGA TAA AAG AGC -3′ designed to anneal at the 3′-end of the coding region, and single-strand cDNA was produced. The double-strand cDNA was amplified by PCR using a primer designed to anneal at the SL1 splice leader sequence 5′-GGT TTA ATT ACC CAA GTT TGAG-3′ and the 3′-end primer 48 . The region coding for DRAG-1 peptide from I131 to E368 was amplified using primers 5′-GTC ACA TAT GAT AAT GTT CAA TGG CTC CGT GC-3′ and 5′-GTC ACT CGA GTT CTT TCT GGA ACC GAG CAT G-3′, digested with NdeI and XhoI, and ligated into the pET-19b vector using the NdeI and XhoI sites. The resulting antigenic peptide of DRAG-1 was expressed as a histidine-tagged fusion protein in Escherichia coli and was used to immunize rabbits. The generated antibody was affinity purified.
Immunohistochemistry. Immunohistochemistry was performed as described 49 . The DRAG-1 antibody was used as the primary antibody at 4 μg/ml. Alexa 594-labeled donkey anti-rabbit IgG (Life Technologies) was used as the secondary antibody at a dilution of 1:500.
Microscopy. Nomarski and fluorescence microscopy was performed using an Axioplan 2 microscope equipped with Axiocam CCD camera (Zeiss). Confocal laser scanning microscopy was conducted with LSM5 (Zeiss) equipped with a C-apochromat 63× (water immersion; NA, 1.2) lens controlled by PASCAL version 3.2 SP2 or ZEN software (Zeiss).

Data availability
Strains and plasmids are available upon request. The authors state that all data necessary for confirming the conclusions presented in the article are represented fully within the article.