A Metalloprotease Disintegrin That Controls Cell Migration in Caenorhabditis elegans

See allHide authors and affiliations

Science  23 Jun 2000:
Vol. 288, Issue 5474, pp. 2205-2208
DOI: 10.1126/science.288.5474.2205


In Caenorhabditis elegans, the gonad acquires two U-shaped arms by the directed migration of its distal tip cells (DTCs) along the body wall basement membranes. Correct migration of DTCs requires the mig-17 gene, which encodes a member of the metalloprotease-disintegrin protein family. The MIG-17 protein is secreted from muscle cells of the body wall and localizes in the basement membranes of gonad. This localization is dependent on the disintegrin-like domain of MIG-17 and its catalytic activity. These results suggest that the MIG-17 metalloprotease directs migration of DTCs by remodeling the basement membrane.

Many new insights into the molecular mechanisms controlling cell migration have been gained by the genetic analysis of model organisms such as the nematode C. elegans. The shape of the C. elegans hermaphrodite gonad is determined by the migration path of the DTCs during larval development (1). DTCs migrate in a complex trajectory consisting of three linear phases punctuated by two orthogonal turns (Fig. 1, A and B). DTCs are generated at the ventral midbody and migrate in opposite directions along the basement membranes of ventral body wall muscles. Mutation of themig-17 gene alters DTC migration (2): initially, migration along the ventral body wall muscles is normal, but the migration path deviates from normal after the dorsal turn (Fig. 1C). In addition, these DTCs often detach from the dorsal muscles and migrate along the intestine or gonad. As a result of this misdirected migration, mig-17 mutants exhibit morphologically abnormal gonadal arms (3). These results indicate thatmig-17 is not required for DTC migration per se, but rather influences the route of migration.

Figure 1

Morphogenesis in C. elegansgonads. (A) DTC migration in hermaphrodites. A cylindrical projection opened along the dorsal midline. Dorsal and ventral body wall muscles, light blue; lateral epidermis, white; gonad, orange; intestine, dark blue. (B and C) DTC migration in wild-type (B) and mig-17(K174) (C) animals. The photos show posterior gonad arms. Anterior is left and dorsal is top. Arrows indicate DTCs. Morphology of gonadal arms is illustrated below the photo. The DTC and the intestinal nuclei (C, arrowheads) are in the same focal plane. (D) MLC migration in males. Male MLC leads extension into a final shape. Male DTCs do not lead arm extension. (E and F) MLC migration in wild-type (E) and mig-17(K174) (F) animals. Anterior is left and ventral is top. Bar, 20 μm.

A similar defect is observed in the male gonad of mig-17mutant worms. In wild-type worms, the male linker cell (MLC) migrates anteriorly, then reflexes and migrates to the posterior end of the worm (Fig. 1, D and E) (1). In mig-17 mutants, the MLC fails to migrate correctly (Fig. 1F). However, mutation ofmig-17 does not affect the migration of other cell types such as HSN neurons and Q neuroblasts (2). Therefore,mig-17 does not affect cell migration generally, but rather is required specifically for the correct migration of gonadal leader cells.

We cloned the mig-17 gene by positional mapping followed by transformation rescue (4). Fragments containing only the F57B7.4 gene rescued mig-17(k174). Themig-17 gene encodes a single protein of 509 amino acids length (Fig. 2A). The MIG-17 protein is a member of the metalloprotease-disintegrin protein (ADAM) family (5). It contains a signal sequence followed by a pro-domain, a catalytic domain, a disintegrin-like (DI) domain, and a cysteine-rich domain (Fig. 2B). MIG-17 also lacks a transmembrane domain, suggesting that it is secreted. Comparison with the ADAM family members indicates that MIG-17 is most similar to the mouse ADAMTS-1 (6). The metalloprotease and DI domains are relatively well conserved (Fig. 2A). However, the domain organization in the COOH-terminal region following the DI domain of ADAMTS-1 diverges from that of MIG-17 in that ADAMTS-1 possesses the thrombospondin type I motif (6), whereas MIG-17 does not.

Figure 2

The mig-17 gene and protein. (A) MIG-17 sequence alignment with ADAMTS-1 (8). The COOH-terminal 325 amino acids of ADAMTS-1 are omitted. Black boxes, identical amino acids; gray boxes, similar amino acids; black bar, predicted signal peptide. In the metalloprotease domain, the zinc-binding motif (HEXXHXXGXXH) is marked by asterisks. The positions of mig-17 mutations are indicated: D79N (GAC →AAC) in k135; Q111Stop (CAA →TAA) in k174; G292E (GGA → GAA) in k176; R347Stop (CGA →TGA) in k167. The accession number for the MIG-17 sequence is AB044562. (B) Domains of MIG-17. The transgenes are constructed as fusions with GFP. Hatched boxes show deleted regions. Each transgene was injected intomig-17(k174) and wild-type animals to test its ability to rescue DTC migration defects and its expression pattern, respectively. SP, signal peptide; MP, metalloprotease domain; DI, disintegrin-like domain; CR, cysteine-rich domain; BWM, body wall muscle. (C) Immunoblot analysis. Protein lysates from wild-type N2 or N2 expressing MIG-17::GFP were immunoprecipitated with anti-GFP and blotted with anti-GFP.

We sequenced four mig-17 mutant alleles (7). The k135 and k174 alleles involve missense and nonsense mutations, respectively, in the pro-domain. The k167 and k176 mutations occurred in the metalloprotease domain (Fig. 2A). The putative metalloprotease catalytic domain of MIG-17 has a zinc-binding motif similar to the HEXXHXXGXXH motif (5, 8) found in the ADAM family (Fig. 2A). We examined mutant rescue by a MIG-17 transgene bearing a mutation that changes Glu303 to Ala (E303A) within the metalloprotease active site. In the zinc-binding metalloproteases, this mutation abolishes enzymatic activity without altering protein structure or stability (9). The wild-type mig-17 was able to rescue the DTC migration defects of mig-17(k174)mutants, but mig-17(E303A) was not (Fig. 2B). We conclude that MIG-17 is an active metalloprotease and that metalloprotease activity is essential for its function in controlling DTC migration.

In addition to its metalloprotease domain, MIG-17 possesses pro-, DI, and cysteine-rich domains. To determine whether these domains are essential for DTC migration, deletion mutations lacking each domain were generated (Fig. 2B). When these transgenes were introduced into mig-17(k174) mutants and the rescue of DTC migration was scored, none rescued the mutant phenotypes. Thus, the pro-, DI, and cysteine-rich domains are all critical for MIG-17 activity leading to proper DTC migration.

To determine where and when MIG-17 is expressed during development, we constructed a translational fusion betweenmig-17, including the gene promoter, and the green fluorescent protein (GFP), mig-17::GFP(10). This fusion could rescue mig-17 mutant defects (Fig. 2B), suggesting that the fusion protein was functional and expressed in all cells that require mig-17 activity. Protein immunoblotting (11) identified two main bands of 105- and 70-kD size in animals expressing MIG-17::GFP (Fig. 2C). We tentatively assigned the 105-kD band as the precursor form, and the 70-kD band as the processed form generated by cleavage of the prodomain. These sizes are larger than those calculated from the primary structure and most likely reflect glycosylation of the MIG-17::GFP polypeptides. GFP fluorescence was observed on the pseudocoelomic face of body wall muscles, but not on their hypodermal face (Fig. 3, A and B). This expression pattern was first detected in late embryos and continued through the adult stage. Expression of mig-17 was also seen on the surface of gonad, starting when the DTCs migrated over the lateral hypodermis toward the dorsal muscles (Fig. 3, C through F). This timing of mig-17 expression on the gonad is correlated with the stage at which the DTC migration defect is first observed inmig-17 mutants (Fig. 1C). This suggests thatmig-17 expression on the gonad is required for proper migration of DTCs.

Figure 3

Expression of mig-17. (A, C, E, G, andI) GFP fluorescence; (B, D,F, H, and J) Nomarski. Same animal in each left-right pair. Photos show ventral view (A and B) and lateral view of posterior gonad arms (C through J). (A through F)mig-17(k174) with mig- 17::GFP; (G and H)unc-119(e2498) with mig- 17::GFP and pDP#MM016B (unc-119); (I and J)unc-119(e2498) with mig-17EA:GFP and pDP#MM016B. (A and B) A young adult hermaphrodite. GFP is expressed on the pseudocoelomic face of ventral body wall muscles, but not on the ventral hypodermal ridge (arrowheads). (C and D) A mid L3 hermaphrodite. The DTC (arrow) starts turning dorsally. GFP is not expressed on the surface of gonad. The dotted fluorescence results from the auto-fluorescence of gut granules. (E and F) A late L3 hermaphrodite. The DTC migrates toward the dorsal muscle cells. GFP is expressed on the surface of gonad. [Frequency of GFP detection on the gonad: 0% (n = 12) during migration on ventral muscle; 44% (n = 9) during migration on lateral hypodermis; 90% (n = 21) during migration on dorsal muscle]. (G through J) Mid L4 hermaphrodites. MIG-17::GFP is expressed on the surface of the gonad, but MIG-17EA::GFP is not. Bar, 20 μm.

To further examine the expression pattern of mig-17, we constructed another fusion gene (mig-17ΔSP::GFP), which lacks a predicted signal peptide of MIG-17 (Fig. 2B). Transgenic animals carryingmig-17ΔSP::GFP exhibited GFP fluorescence in the cytoplasm of the body wall muscle cells but not on the gonadal cells. This raises the possibility that the MIG-17 protein is produced in the body wall muscles and secreted. To test this possibility, theunc-54 promoter (12) was used to express the wild-type mig-17 fused to GFP in body wall muscles. When this unc54p::mig-17::GFP transgene (10) was introduced into mig-17(k174)mutant animals, it was able to rescue the DTC migration defects (Fig. 4B). Next, to examine whether expression of mig-17 on the gonad is sufficient to direct DTC migration, we expressed mig-17 under the control of thelag-2 promoter (13), which drives expression in the hermaphrodite DTC. Transgenic mig-17(k174)mutants harboring thelag-2p::mig-17::GFP transgene (10) showed normal DTC migration (Fig. 4C). In these animals, GFP fluorescence was limited to the DTCs but not observed on the body wall muscles. Thus, expression of mig-17 on the gonad is important for its effect on DTC migration. Taken together, these results suggest that MIG-17 is secreted from muscle cells and functions in the basement membrane of the gonad to guide DTC migration.

Figure 4

Morphology of gonadal arms. Differential interference contrast microscopy of adult hermaphrodites. The photos show posterior gonad arms. Bar, 20 μm. Morphology of the gonadal arms is illustrated below the photo. (A) mig-17(k174). (B) mig-17(k174) withunc-54p::mig-17::GFP expressed in body wall muscle. (C) mig-17(k174) withlag-2p::mig-17::GFP expressed in DTC. (D) mig-17(k174); unc-6(e78). [Frequency of the ventral-to-dorsal migration defect in which posterior DTC fails to migrate from ventral to dorsal: 3% (n = 60) in mig-17(k174); 10% (n = 60) in unc-6(e78); 45% (n = 60) in mig-17(k174);unc-6(e78)].

How does MIG-17 localize to the surface of the gonad after it is secreted from the body wall muscle? It is believed that the DI domains of ADAMs may mediate interactions with the extracellular matrix or cell surface receptors (5). Similarly, the DI domain in MIG-17 could be important for its localization. Indeed, we observed that in animals expressing the reporter constructmig-17ΔDI::GFP, which lacks the DI domain of MIG-17, GFP fluorescence was present on the body wall muscles but not on the surface of the gonad (Fig. 2B). Even when the MIG-17ΔDI protein was expressed in the DTCs using thelag-2 promoter in mig-17(k174) mutants, it failed to rescue the DTC migration defects. In addition, a catalytically inactive MIG-17(E303A) and a MIG-17ΔPro lacking the pro-domain of MIG-17 were unable to localize on the surface of the gonad (Figs. 2B and 3, G through J). In contrast, the cysteine-rich domain is not required for the localization of MIG-17 on the gonad (Fig. 2B). Metalloprotease activity is therefore essential for the localization of MIG-17 on the gonad, and the DI domain may anchor MIG-17 to the gonad basement membrane.

DTCs migrate on a defined pathway while adhering to basement membrane during larval development. In mig-17 mutants, the initial migration of DTCs along the ventral body wall muscles is normal, indicating that the MIG-17 metalloprotease is not required for DTC migration itself. However, subsequent dorsal migration is defective. Another set of C. elegans gene products, netrin UNC-6 and its receptors, UNC-5 and UNC-40, are critical for dorsal DTC migration, with UNC-6 acting as a guidance molecule in the basement membranes of the body wall and UNC-5 as a receptor for UNC-6 in the DTCs (14). We observed that introduction ofmig-17(k174) enhanced the defect in dorsal DTC migration in unc-6(e78) mutants (Fig. 4D), suggesting that MIG-17 may be required for the processing or effecting of guidance cues provided by the UNC-5–UNC-6 system. For example, the MIG-17 metalloprotease may function in the basement membrane to modify the extracellular matrix, thus allowing the DTCs to attach to the body wall and be guided along the correct migration pathway. Another extracellular metalloprotease, GON-1, also controls DTC migration (15). Mutations in gon-1 block DTC migration completely. GON-1 and MIG-17, although both metalloproteases, support different aspects of cell migration in C. elegans(16). Analysis of C. elegansmetalloproteases should uncover structural changes in the extracellular matrix necessary for cell migration.

  • * To whom correspondence should be addressed. E-mail address: nishiwak{at}


View Abstract

Navigate This Article