Positional information along the dorsal-ventral axis of the Drosophila embryo is acquired through a signal transduction pathway which employs a extracellular protease cascade. The sequential activation of serine protease zymogens results in the ventrally localized production of a ligand in the perivitelline space of the embryo. Snake is one of several serine proteases which function in generating the ventralizing signal. Here, we investigate the biochemical properties of Snake in vivo and in vitro using recombinant forms of the protease. Wild-type Snake zymogen completely rescues embryos from snake null females when microinjected into the perivitelline space. Biochemical evidence for a covalently associated two-chain form of the activated protease is presented. The contribution of the activation peptide region to zymogen activation was addressed using site-directed mutagenesis. The phenotypic rescue properties of an autoactivated form of Snake reveal that the covalently associated proenzyme polypeptide chain suppresses a dominant effect associated with the activated catalytic chain alone. Recombinant active catalytic chain was produced and found to be short lived as a recombinant protein. These results suggest a model in which the proenzyme polypeptide both stabilizes and targets the Snake catalytic chain to a ventrally localized activation complex within the perivitelline space.

Cell fates along the dorsal-ventral axis of the Drosophila melanogaster embryo are established through a signal transduction pathway in which information in the outer egg membrane is passed to the plasma membrane and ultimately to dividing nuclei in the egg cytoplasm (Anderson, 1987; Govind and Steward, 1991; Nuesslein-Volhard and Roth, 1989; Roth, 1994; Rushlow and Arora, 1990). The cytoplasmic part of the pathway is activated through a receptor on the surface of the plasma membrane encoded by the Toll gene and this receptor responds to an extracellular signal (Hashimoto et al., 1988; Schneider et al., 1991). Dorsalventral asymmetry, however, is found upstream within an extracellular compartment of the syncitial blastoderm embryo called the perivitelline space (PVS) in the form of a ventrally restricted spatial distribution of ligand for Toll (Stein et al., 1991). Several of the proteins known to be most intimately involved in the production of the ligand are found in the fluid phase of the perivitelline space. These include the products of the genes snake, easter and spaetzle (Stein and Nuesslein-Volhard, 1992). The protein product of the spaetzle gene appears to be proteolytically processed and a fragment of the Spaetzle protein probably corresponds to the Toll ligand (Morisato and Anderson, 1994; Schneider et al., 1994).

Generation of the ventral ligand distribution requires the activity of the snake and easter and gastrulation defective genes, all of which encode serine proteases (Hecht and Anderson, 1992; Chasen and Anderson, 1989; DeLotto and Spierer, 1986; Konrad and Marsh, 1990; DeLotto, unpublished observations). Snake protease is produced as a preproenzyme. After secretion and cleavage of the signal peptide, it is expected to appear in the perivitelline fluid as an inactive proenzyme or zymogen (see Fig. 1). A specific proteolytic cleavage at the junction between the proenzyme polypeptide chain (PRO) and the protease catalytic chain (CAT) is required for activation of the snake protease. The activated form of Snake is predicted to have a substrate specificity that is trypsinlike, cleaving after positively charged amino acid residues. Conversion of Snake zymogen to the active form, however, requires a chymotrypsin-like protease activity, cleaving between leucine184 and isoleucine185.

Fig. 1.

A schematic diagram of the Snake protease. The Snake zymogen is predicted to be activated by cleavage at the position marked by an arrow (the activation peptide region). The carboxy-terminal portion is the serine protease catalytic chain (CAT). Residues of the tripartite active site are the histidine (H), aspartate (D) and serine (S). The proenzyme polypeptide (PRO) is composed of three domains, an acidic domain (ACIDIC), the disulfide knot (DSN), and a region (CHO) which is glycosylated. The proenzyme polypeptide and catalytic chain are predicted to remain joined by a disulfide bridge after activation (---). Δ corresponds to the position of a factor X-like cleavage site.

Fig. 1.

A schematic diagram of the Snake protease. The Snake zymogen is predicted to be activated by cleavage at the position marked by an arrow (the activation peptide region). The carboxy-terminal portion is the serine protease catalytic chain (CAT). Residues of the tripartite active site are the histidine (H), aspartate (D) and serine (S). The proenzyme polypeptide (PRO) is composed of three domains, an acidic domain (ACIDIC), the disulfide knot (DSN), and a region (CHO) which is glycosylated. The proenzyme polypeptide and catalytic chain are predicted to remain joined by a disulfide bridge after activation (---). Δ corresponds to the position of a factor X-like cleavage site.

SP6 RNA microinjection experiments suggest that activation of the Easter zymogen requires the function of several upstream genes including snake, gastrulation defective, pipe, nudel and windbeutal (Chasen et al., 1992). The activation of the Snake zymogen requires the function of gastrulation defective, pipe, nudel and windbeutal. snkΔne RNA transcripts, which encode a secreted free Snake catalytic chain, ventralize cell fates in a concentration-dependent manner when they are microinjected centrally into the cytoplasm (Smith and DeLotto, 1994). Injection of these RNA transcripts asymmetrically into the cytoplasm will polarize the gastrulation pattern. In contrast, injection of RNA transcripts encoding the wild-type Snake zymogen never produces a dominant ventralizing effect and even at high RNA concentrations the cuticle pattern is normal and polarity is always correct with respect to the egg shape (Smith et al., 1994).

These and other observations support a protease cascade model in which spatially restricted sequential activation of protease zymogens is initiated by a localized cue in the perivitelline space (DeLotto and Spierer, 1986; Chasen et al., 1989; Smith and DeLotto, 1994). The end product of this protease cascade is the restricted production of ligand for the Toll receptor in the space on the ventral side of the embryo (Roth, 1993). One of the major requirements for a protease cascade to generate a localized ligand is exquisite spatial regulation of both the activation of the zymogens and the subsequent interactions of the activated form of the proteases. Exactly how spatial and temporal regulation is achieved in the cascade is presently unknown. In other biochemical pathways such as blood coagulation and the complement system, however, the protease proenzyme polypeptide chain has been shown to play a role (Furie and Furie, 1988).

We have undertaken a biochemical characterization of the Snake protease using in vitro and in vivo assays. Recombinant snake protease, produced using a baculovirus expression system, completely rescues embryos that lack snake function, when microinjected into the perivitelline space. A recombinant Snake protein with its activation peptide altered to one requiring a trypsin-like activator is capable of autoproteolytic activation. The biological properties of this autoactivated form in phenotypic rescue assays are very different from those of the free ‘activated’ catalytic chain, previously described. Production of an active catalytic chain only of snake revealed that it is short lived when produced in the absence of a protease inhibitor. The differences in biological activity of the various forms suggest that zymogen activation cleavage alone is inadequate to completely explain Snake spatial regulation and that activation cleavage must occur physiologically in the context of a spatially localized complex.

Expression of recombinant Snake protein

Wild-type snake cDNA, cD8, (DeLotto and Spierer, 1986) in M13MP18 was modified to introduce a BamHI site immediately adjacent to the open reading frame, by site directed mutagenesis (Kunkel et al., 1987) using the oligonucleotide 5′ CACAACAGAACTCACGGATCCAAT-GATAATACTTTGG 3′ by the method of Kunkel. RF viral DNA was cleaved with BamHI and KpnI and ligated to pVL941 (Summers and Smith, 1987) cleaved with BglII and KpnI. Recombinant virus was produced by cotransfection of this plasmid with wild-type virus using standard methods and SF9 host cells. Expression of recombinant protein was done in IZD-SF cells which were derived from Mammestra Brassicae IZD-MB0503 (Miltenberger et al., 1976). Significantly enhanced yields of secreted Snake protein were observed in the culture fluid of a cell line IZDMB0503 compared to SF9 cells. IZD-SF cells were produced by gradual weaning of IZD-MB0503 cells into serum-free SF 900 medium (Gibco). IZD-SF cells were infected with recombinant virus and harvested 72 hours post infection. The medium was centrifuged at 900 r.p.m. for 5 minutes to pellet floating cells, and the supernatant was collected. The cleared medium was diluted three fold with water and Snake protein was batch bound to Yellow-3 dye affinity resin (Sigma). Resin was washed with 20 mM Pipes pH 6.8, 1 mM EDTA, 1 mM EGTA, and protein eluted with 20 mM Pipes pH 6.8, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA. Protein was concentrated by centrifugation in centricon-10 microconcentrators (Amicon) and dialyzed into 20 mM Pipes, pH 6.8, 50 mM NaCl, 1 mM EDTA. For microinjection experiments, protein was subsequently dialyzed into isotonic perivitelline buffer (PPS; 54 mM NaCl, 12 mM Na2SO4, 35 mM K2SO4, 4 mM CaCl2, 14 mM MgSO4,2 mM NaH2PO4, 1 mM Na2HPO4; Overall and Jaffe, 1988). Expression of SnkΔn protein was conducted by adding heat sterilize Benzamidine-Sepharose (Pharmacia) 24 hours after infection and expression continued to 72 hours. Affinity resin was collected in a disposible column (Biorad) washed with TN (20 mM Pipes, pH 6.8, 50 mM NaCl), and eluted in TN containing 50 mM Benzamidine.

Monoclonal antibodies

Monoclonal antibody, mAb53, was produced against amino acids 61-436 of Snake using two different bacterial fusion protein constructs. The embryonic cDNA clone cD8 was propagated in a Dam– strain of E. coli, cleaved with BclI and BamHI and subcloned in frame into the carboxy terminus of β-galactosidase using the pEX vectors (Stanley and Luzio, 1984) and into the tryptophan E gene product using the vector, pATH 10 (Klaembt and Schmidt, 1986). Fusion proteins were isolated as insoluble aggregates and checked for purity on SDS poly-acrylamide gels. Polyacrylamide gel purified TrpE-snk fusion protein was injected as a suspension in Freund’s adjuvant, and monoclonal antibodies were prepared as described (Harlow and Lane, 1988). Lines that were positive for Snake were screened using the β-galactosidase-snk fusion protein. The line was subcloned to homogeneity and cell culture supernatants were prepared.

Anticatalytic chain antibody, 3D10, was prepared as follows. A tryptophan E-fusion protein was generated by introducing a PvuI site at the junction between the proenzyme polypeptide and the catalytic chain (amino acids 184-185) in cD8/M13MP18 using site-directed mutagenesis with the oligonucleotide 5′ CCCAGTGTTCCCCC-GATCGTCGGTGGAACCCCC 3′, making RF DNA, digesting with PvuI and PstI. The PvuI site was converted into a Bam HI site by an adapter and the BamHI-PstI catalytic chain fragment was introduced into pATH 10. Fusion protein was made and processed as described above. Positive clones were identified by Elisa using partially purified baculovirus expressed recombinant Snake protein.

RNA phenotypic rescue and site-directed mutagenesis

Site directed mutagenesis and RNA microinjections were performed as previously described (Smith et al., 1994). The mutagenic oligonucleotides used for alterations at the activation peptide were as follows: oligo C2 (5′-CCCAGTGTTCCCGCGATAGTCGGTGGAACCCCC-3′) altered leucine184 to alanine; oligo C3 (5′-CCCAGTGTTCC-CGTGATAGTCGGTGGAACCCCC-3′) altered leucine184 to valine; oligo C4 (5′-CCCAGTGTTCCCTTCATAGTCGGTGGAACCCCC-3′) altered leucine184 to phenylalanine; oligo B2 (5′-CCCAGT-GTTCCCCCGATCGTCGGTGGAACCCCC-3′) altered leucine184 to proline; oligo C5 (5′-CAGTGCGTGCCCAGTGCTCCCCT-GATAGTCGGT-3′) altered valine182 to alanine; oligo C6 (5′-CAGTGCGTGCCCAGTGGTCCCCTGATAGTCGGT-3′) altered valine182 to glycine; oligo C7 (5′-CAGTGCGTGCCCAGTTCTCCC-CTGATAGTC-3′) altered valine182 to serine; oligo C9 (5′-AAGCAGTGCGTGCCCAGTTCTAACCGGATAGTCGGTGG-AACC-3′) altered valine182, proline183 and leucine184 to serine, asparagine and arginine, respectively; oligo C10 (5′-CAGTGCGT-GCCCAGTGTTAAGAAGATAGTCGGTGGAACC-3′) altered both proline183 and leucine184 to lysine; oligo C11 (5′-GTGCCCAGT-GTTCCCAAGATAGTCGGTGGAACC-3′) altered leucine184 to lysine; oligo A4 (5′-GTGCCCAGTGTTCGCCGGATAGT-CGGTGG-3′) altered both proline183 and leucine184 to arginine; oligo Xa-1 (5′-GGGAAGCAGTGCGTGCCCATCGAAGGCCG-GATAGTCGGTGGAACCCCCACT-3′) altered serine181, valine182, proline183 and leucine184 to isoleucine, glutamine, glycine and arginine, respectively.

Protein microinjection into the perivitelline space

Recombinant protein suspended in 1× PPS buffer was microinjected laterally into stage 2 or stage 3 embryos using the method of Stein (Stein and Nuesslein-Volhard, 1992; Stein et al., 1991). Precise measurements of the volume of fluid microinjected are not available, however, fluid was microinjected until the perivitelline space transiently widened by approximately 10-15 μm at the site of injection. Gastrulation patterns of live embryos were observed by transmitted light, and polarity was scored based upon pole cell movement and gastrulation morphology. Larval cuticles were prepared by Hoyer’s mounts as previously described (Van der Meer, 1977).

Protease digestion of recombinant Snake protein

Human leukocyte elastase (Sigma) and Restriction protease Factor Xa (Boehringer Mannheim) were resuspended at 1 mg/ml in 20 mM Hepes pH 6.8, 50 mM NaCl, 20% glycerol. Enzymes were serially diluted in 20 mM Hepes pH 6.8, 10 mM CaCl2 in a final volume of 10 μl. 50 μl of conditioned cell culture supernatant was added to each dilution and samples were incubated for 90 minutes at 30°C. An equal volume of 2× SDS loading buffer containing 0.2% PMSF was added and one half of the total volume was heated to 100°C for 5 minutes and loaded on a 12.5% acrylamide gel with prestained low molecular mass markers (Biorad). After electrophoresis under reducing conditions the gel was electrophoretically transferred to Immobilon PVDF membrane (Millipore) and snake protein was identified by incubation overnight at 4°C with monoclonal antibody 3D10. Bands were visualized by alkaline phosphatase-conjugated secondary antibody as described by Driever and Nüsslein-Volhard (1988).

Electrophoresis under reducing and nonreducing conditions

SDS polyacrylamide gel electrophoresis was conducted as described by Allore (1984), immunoblotted and probed with 3D10 antibody.

Biochemical properties of recombinant wild-type Snake

We began biochemical analysis by producing recombinant Snake protein using baculovirus vectors and a lepidopteran hemocyte cell line which we call IZD-SF. IZD-SF is more efficient in the production and secretion of recombinant Snake protein into the culture medium than other cell lines which we tested (unpublished observations). To facilitate the purification and subsequent analysis of Snake, monoclonal antibodies were generated against the Snake proenzyme polypeptide and the catalytic chain. Recombinant Snake protein was detected by immunoblotting of conditioned serum-free cell culture medium from cells infected with a recombinant wild-type snake baculovirus construct (Fig. 2A). Wild-type Snake protein appears as a closely spaced doublet with a relative mobility of 52-54×103 Mr, higher than the 48×103 Mr predicted for the zymogen form (DeLotto and Spierer, 1986). In addition, a minor band at 42×103 Mr appears with variable intensity in each protein preparation. The difference in relative mobility of Snake from that predicted from the primary amino acid sequence is due the fact that Snake is glycosylated by both N- and O-linked carbohydrates (data not shown).

Fig. 2.

Expression of recombinant Snake protein and binding to immobilized soybean trypsin inhibitor. An immunoblot of conditioned cell culture fluid probed with monoclonal antibody 3D10. (A) Predominant bands are a strong doublet at 52-54×103Mr and a minor proteolysis band at 42×103Mr. (B) Binding of Snake to soybean trypsin inhibitor agarose: a, starting material; b, flow through; c, wash; d, eluate.

Fig. 2.

Expression of recombinant Snake protein and binding to immobilized soybean trypsin inhibitor. An immunoblot of conditioned cell culture fluid probed with monoclonal antibody 3D10. (A) Predominant bands are a strong doublet at 52-54×103Mr and a minor proteolysis band at 42×103Mr. (B) Binding of Snake to soybean trypsin inhibitor agarose: a, starting material; b, flow through; c, wash; d, eluate.

Microinjection of soybean trypsin inhibitor (Type I) into the perivitelline space of wild-type embryos prior to cellularization dorsalizes, implicating serine proteases of trypsin-like specifity in the production of the ventralizing signal (Stein and Nuesslein-Volhard, 1992). To determine whether the Snake protease might be a target of this inhibitor, recombinant Snake protein was incubated overnight at 4°C with a slurry of soybean trypsin inhibitor (STI) agarose. The resin was transferred to a column, washed with loading buffer, and eluted at low pH (10 mM HCl, pH 2.0). Aliquots from various steps were corrected for volume and analyzed by immunoblotting with a monoclonal antibody which recognizes an epitope in the proenzyme polypeptide, and the results are illustrated in Fig. 2B. A comparison of lane a (starting material) and lane d (eluted) reveals that recombinant Snake protein is quantitatively bound to STI-agarose and eluted by low pH. This result indicates that the Snake zymogen binds to soybean trypsin inhibitor and therefore is a probable target of STI induced dorsalization.

Recombinant Snake zymogen completely rescues embryos lacking snake function when microinjected into the perivitelline space

Activities that partially rescue the snake, easter and spaetzle mutant phenotypes have been detected in the fluid of the perivitelline space (Stein and Nuesslein-Volhard, 1992). To determine directly whether Snake functions in the PVS and to test for biological activity of recombinant Snake, recombinant Snake was partially purified and microinjected into the perivitelline space of embryos produced by snk229/snk229 mothers. Embryos produced by snk229/snk229 mothers are strongly dorsalized and exhibit no dorsoventral (d-v) asymmetry in their gastrulation pattern (Fig. 3A). Phenotypic rescue was scored as the ability of the recombinant protein to restore d-v asymmetry to gastrulation. When protein was injected into the perivitelline space on the ventral side at cleavage stage (n = 77), 46% of the embryos gastrulated with normal d-v asymmetry. When injected at syncitial blastoderm stage (n = 43), 63% showed a normal pattern of gastrulation with respect to the egg shape. Phenotypic rescue was therefore more efficient when Snake zymogen was injected at the syncitial blastoderm stage compared to cleavage stage. Almost all of the embryos that exhibited a normal gastrulation pattern developed cuticles which had ventrolateral pattern elements. Typically, between 2 and 5% of the injected embryos hatched into normal first instar larvae (Fig. 3B). To determine whether rescue is specifically due to recombinant Snake protein, a mutation was introduced in which the serine of the protease active site was altered to alanine. RNA transcript microinjection assays had previously revealed that this alteration completely eliminates Snake rescue activity (Smith et al., 1994). When recombinant Snake (S to A) protein, purified in parallel to wild-type Snake, was microinjected, no phenotypic rescue activity was observed.

Fig. 3.

Phenotypic rescue by microinjection of recombinant Snake zymogen into the perivitelline space. (A) A strongly dorsalized cuticle from an embryo produced by a snk229/snk229 female. (B) A similar embryo which was injected with recombinant Snake protein into the perivitelline space at syncitial blastoderm.

Fig. 3.

Phenotypic rescue by microinjection of recombinant Snake zymogen into the perivitelline space. (A) A strongly dorsalized cuticle from an embryo produced by a snk229/snk229 female. (B) A similar embryo which was injected with recombinant Snake protein into the perivitelline space at syncitial blastoderm.

To determine whether the position of microinjection affects the rescue response, recombinant Snake protein was microinjected into the perivitelline space on either the ventral or the dorsal side of the embryo and rescue was analyzed. Whether the protein was introduced on the dorsal (n = 90) or the ventral side (n = 95) the polarity of gastrulation was always correct with respect to the shape of the egg.

Alteration of the activation peptide region of Snake

Conversion of Snake zymogen to its active form requires proteolytic processing. The primary sequence determinants, recognized by the activator of a protease zymogen, make up a region called the activation peptide, a stretch of several amino acids on the amino side of the point of cleavage (Stroud et al.,1977). For Snake, cleavage should occur between leucine184 and isoleucine185, after the sequence VPSVPL. Cleavage after leucine is expected to require a chymotrypsin-like activity.

To determine how changes in the amino acid sequence of the activation peptide region affect Snake rescue activity in vivo, we used an RNA transcript microinjection assay (Smith et al., 1994). The phenotypic rescue activities of various alterations in the activation peptide were compared to those of wildtype Snake. The ability to rescue the dorsal-ventral axis was scored in order of increasing strength by the following criteria: production of filzkorper, ventral denticles and hatching. Experimental results for several alterations of the activation peptide are summarized in Table 1.

Table 1.

Phenotypic rescue by transcripts with alterations in the activation peptide

Phenotypic rescue by transcripts with alterations in the activation peptide
Phenotypic rescue by transcripts with alterations in the activation peptide

Microinjection of wild-type snake SP6 transcripts results in 61% of embryos exhibiting restoration of ventrolateral pattern elements. A conservative change in the P1 position immediately adjacent to the cleavage site, leucine184 to phenylalanine, which should still require a chymotrypsin-like activator, rescues with an efficiency approximately equal to wild type. Alteration of leucine184 to alanine, causes a dramatic change in the rescue phenotypes. Although 48% show some ventrolateral pattern elements, the only structure produced is the filzkorper. Alteration of leucine184 to valine or proline, produces more dramatic reductions in rescue activity with only filzkorper material.

An alteration of valine182 in the P3 position (two amino acid residues from the cleavage site) has different effects on the rescue activity depending upon the nature of the substitution. Valine to alanine, rescues 49% of the embryos with a distribution similar to that of wild type. Valine to serine rescues 57% and also produces a distribution similar to wild type. However, when glycine is placed in this position, rescue is reduced to 34% with the production of essentially only the filzkorper.

RNA transcripts that change the activation peptide to one requiring a trypsin-like activator had the unanticipated effect of rescuing embyros with high efficiency. Alteration of both proline183 and leucine184 to arginine resulted in 49% of the embryos producing ventrolateral pattern elements, although no embryos hatched. Alteration of the four residues that precede cleavage to the sequence recognized by vertebrate blood clotting factor X (IEGR), resulted in 38% of the embryos exhibiting ventrolateral pattern elements and 3% hatching. Conversion of the two residues preceding cleavage to lysine produced no rescue. Changing the three residues which precede cleavage to the sequence found at the analogous position in the Easter proenzyme (SNR) produced 47% rescue. These results indicate that alteration of the activation peptide from a sequence requiring a chymotrypsinlike activator to one requiring a trypsin-like activator produces a significant amount of phenotypic rescue. Furthermore, alteration of the activation peptide to IEGR, the site recognized by clotting factor X, permits hatching. IEGR-Snake transcripts were microinjected into various dorsal group mutant backgrounds to see if the ability to rescue bypassed the requirement for other gene products. IEGR-Snake did not rescue nudel, pipe, gastrulation defective, easter, and spaetzle mutations, indicating that IEGR-Snake does not bypass the requirements for the gene products of these genes.

Autoactivation of recombinant IEGR-Snake

To determine how RNA transcripts encoding IEGR-Snake can rescue embryos lacking snake function, IEGR-Snake was expressed as a recombinant protein. A second recombinant protein was made in which the active site serine was converted to alanine (IEGR-Snake S to A). Cell culture supernatants containing all of the recombinant proteins were immunoblotted and probed with a monoclonal antibody directed against an epitope in the catalytic chain of Snake (Fig. 4). In lane b of Fig. 4, wild-type snake protein appears as a doublet of 52-54×103 Mr with a minor band at 42×103 Mr. Snake (S to A) (lane c) also appears as a doublet of 52-54×103 Mr, but is missing the 42×103 Mr minor band suggesting that production of the 42×103 Mr band is an autoproteolytic event. IEGR-Snake (lane d), appears primarily as a band of 29×103 Mr, close to the size expected for the cleaved catalytic chain of 27×103 Mr. When the active site serine of IEGR-Snake is altered to alanine, a doublet appears at 52-54 ×103 Mr, the same relative mobility as the Snake zymogen without the minor 42×103 Mr band (lane e). These results suggest that IEGR-Snake protein undergoes autoproteolytic cleavage at or near the activation site when it is expressed as a recombinant protein.

Fig. 4.

Autoproteolysis of recombinant IEGR-Snake protein. An immunoblot of conditioned medium using anti-catalytic chain monoclonal antibody 3D10. Lane a, molecular mass markers; 27.5, 32.5, 49.5, 80, and 106×103Mr; b, Wild-type Snake protein; c, Snake (S to A); d, IEGR-Snake; e, IEGR-Snake (S to A). A 29×103Mr band appears in IEGR-Snake.

Fig. 4.

Autoproteolysis of recombinant IEGR-Snake protein. An immunoblot of conditioned medium using anti-catalytic chain monoclonal antibody 3D10. Lane a, molecular mass markers; 27.5, 32.5, 49.5, 80, and 106×103Mr; b, Wild-type Snake protein; c, Snake (S to A); d, IEGR-Snake; e, IEGR-Snake (S to A). A 29×103Mr band appears in IEGR-Snake.

To determine whether the 29×103Mr band corresponds to cleavage at the zymogen activation site, wild-type recombinant Snake protein was digested with increasing concentrations of human leukocyte elastase, a relatively specific enzyme which is predicted to cleave at the wild-type activation peptide cleavage site. After digestion of Snake zymogen with leukocyte elastase, the protein was immunoblotted and probed with the anti-catalytic chain antibody (Fig. 5A). With increasing elastase concentration, the 52-54×103Mr doublet disappears and a major band at 29×103Mr appears (lane f). This band comigrates with the 29×103Mr band spontaneously produced upon expression of recombinant IEGR-Snake (lane g).

Fig. 5.

In vitro cleavage of Snake and IEGR-Snake (S to A) at the activation site. (A) Wild-type Snake was digested with human leukocyte elastase. Lane a, minus enzyme; b, 0.1 ng; c, 1 ng; d, 10 ng; e, 100 ng; f, 1 μg; g, IEGR-snake; h, molecular mass markers. (B) IEGRSnake (S to A) digested with Factor Xa protease. Lane a, minus enzyme; b, 0.1 ng; c, 1 ng; d, 10 ng.; e, 100 ng; f, 1 μg; g, IEGR-Snake; h, molecular mass markers.

Fig. 5.

In vitro cleavage of Snake and IEGR-Snake (S to A) at the activation site. (A) Wild-type Snake was digested with human leukocyte elastase. Lane a, minus enzyme; b, 0.1 ng; c, 1 ng; d, 10 ng; e, 100 ng; f, 1 μg; g, IEGR-snake; h, molecular mass markers. (B) IEGRSnake (S to A) digested with Factor Xa protease. Lane a, minus enzyme; b, 0.1 ng; c, 1 ng; d, 10 ng.; e, 100 ng; f, 1 μg; g, IEGR-Snake; h, molecular mass markers.

To determine whether IEGR-Snake is undergoing autoproteolysis at the predicted site of factor X cleavage, IEGR-Snake (S to A) was digested with increasing concentrations of purified factor Xa protease and immunoblotted as in the previous experiment (Fig. 5B). Again, a band at 29×103 Mr is produced (lane f) and it comigrates with the 29 ×103Mr band seen in IEGR-Snake preparations (lane g). These results suggest that IEGR-Snake becomes autoproteolytically activated at the correct position following the activation peptide.

Snake contains fourteen cystine residues which can be precisely aligned with those of Easter and Tachypleus tridentatus proclotting enzyme. A disulfide bridge exists in proclotting enzyme between a cystine in the proenzyme polypeptide and one in the catalytic chain such that after activation the two chains remain covalently attached (Smith and DeLotto, 1992). It was therefore predicted that after zymogen activation the 29×103 Mr Snake catalytic chain fragment would appear only under reducing conditions. To test this, IEGR-Snake was immunoblotted after electrophoresis in adjacent lanes with a sharp gradient of 2-mercaptoethanol (Allore and Barber, 1984; Fig. 6A). In IEGR-Snake, the 29×103 Mr band splits off only under reducing conditions while under non-reducing conditions the two upper bands migrate more rapidly. The 29×103 Mr band does not appear when wild-type Snake is similarly treated. These results are consistent with a covalent association of the proenzyme polypeptide and 29×103 Mr catalytic chain by a disulfide bond after zymogen activation. A comparison of band intensities in Fig. 6A and B reveals that the extent of conversion to activated form is variable among different protein preparations.

Fig. 6.

IEGR-Snake protein under reducing and non-reducing conditions. (A) Snake or IEGR-Snake was run on a 12.5% polyacrylamide gel with (+) or without (–) β-mercaptoethanol, immunoblotted and probed with an anti-catalytic chain monoclonal antibody 3D10. A 29×103Mr band corresponding to the catalytic chain is released from IEGR-Snake only under reducing conditions. (B) An independent preparation of IEGR-Snake under reducing (lane b) and nonreducing conditions (lane c). Lane a, molecular mass markers.

Fig. 6.

IEGR-Snake protein under reducing and non-reducing conditions. (A) Snake or IEGR-Snake was run on a 12.5% polyacrylamide gel with (+) or without (–) β-mercaptoethanol, immunoblotted and probed with an anti-catalytic chain monoclonal antibody 3D10. A 29×103Mr band corresponding to the catalytic chain is released from IEGR-Snake only under reducing conditions. (B) An independent preparation of IEGR-Snake under reducing (lane b) and nonreducing conditions (lane c). Lane a, molecular mass markers.

IEGR-Snake protein behaves in a spatially regulated fashion

Recombinant IEGR-Snake protein was tested to determine if it could rescue the snake null phenotype when microinjected into the perivitelline space. Table 2 lists the phenotypes observed when the recombinant proteins shown in Fig. 4 were injected into the PVS along the ventral midline. Recombinant IEGR-Snake protein rescued with reduced activity compared to wild-type Snake protein. No embryos hatched, however a high proportion of the embryos developed ventrolateral pattern elements and many had normal width ventral denticle belts. Alteration of the active site serine to alanine completely eliminated rescue for IEGR-Snake just as it does for wild-type Snake protein.

Table 2.

Phenotypic rescue of snake mutants by microinjection of proteins into the perivitelline space

Phenotypic rescue of snake mutants by microinjection of proteins into the perivitelline space
Phenotypic rescue of snake mutants by microinjection of proteins into the perivitelline space

To determine whether the position of IEGR-Snake microin-jection into the PVS could alter the polarity of gastrulation, embryos were injected at the dorsal midline or the ventral midline and scored by their gastrulation pattern (Table 3). Regardless of the position of injection, the polarity of gastrulation was normal with respect to the egg shape. This indicates that IEGR-Snake protein does not orient the polarity of gastrulation like snkΔne RNA which encodes the free catalytic chain (Smith and DeLotto, 1994). Although polarity is normal with respect to the egg shape, the efficiency of rescue is higher when IEGR-Snake protein is introduced into the ventral side of the perivitelline space compared to the dorsal side. This property was also observed in dorsal vs. ventral injections of wild-type Snake zymogen (data not shown). The results suggest that Snake may be preferentially required on the ventral side of the perivitelline space

Table 3.

Position dependance of IEGR-snake protein phenotypic rescue

Position dependance of IEGR-snake protein phenotypic rescue
Position dependance of IEGR-snake protein phenotypic rescue

Recombinant SnkΔn protein is stabilized when produced in the presence of an immobilized protease inhibitor

To determine whether we could produce a recombinant protein, SnkΔn, consisting of the active Snake catalytic chain only, we constructed a baculovirus expression vector which contained a fusion of the Easter signal peptide to the Snake catalytic chain. However, when we examined cell culture fluid by western blotting, no Snake protein was detected.

To test whether recombinant SnkΔn protein is produced but is shorter lived than IEGR-Snake, we attempted to stabilize SnkΔn as it is secreted from infected cells. Benzamidine-Sepharose resin, an immobilized reversible protease inhibitor, was added to the infected cell cultures 24 hours after infection and the culture fluid and resin were harvested 72 hours post infection. The resin was then washed and bound material was eluted with 50 mM Benzamidine (Fig. 7). Only a very weak signal is observed in the starting material and flow-through lanes (b and c). Lanes f through i show the results of expression of SnkΔn in the presence of Benzamidine-Sepharose. Lane i shows a 29×103 Mr protein band corresponding to the Snake catalytic chain which is eluted from the resin by 50 mM Benzamidine. Lanes b through e show the results of a control adding Sepharose. We conclude that immobilized Benzamidine binds SnkΔn protein after secretion and stablizes it over the course of expression. The SnkΔn protein could be eluted by either 50 mM Benzamidine or low pH (pH 2.5), however we found that the protein rapidly disappeared upon either removal of the Benzamidine or after neutralization to pH 6.8 (data not shown). Because of the instability of SnkΔn protein under physiological buffer conditions, we have thus far been unable to reliably assay SnkΔn activity by microinjection assay.

Fig. 7.

Expression of SnkΔn protein in the presence of Benzamidine Sepharose. SnkΔn protein was expressed in the absence or presence of Benzamidine-Sepharose resin and immunoblotted and probed with 3D10 antibody. Lane a, molecular mass markers. Lanes b-e, without Benzamidine-Sepharose: b, culture fluid; c, flow-through; d, wash; e, 50 mM Benzamidine elution. Lanes f-i, with Benzamidine-Sepharose: f, culture fluid; g, flow-through; h, wash; i, 50 mM Benzamidine-Sepharose. A 29×103Mr Snake catalytic chain is bound to and eluted from Benzamidine-Sepharose when the resin is present in the cell culture fluid.

Fig. 7.

Expression of SnkΔn protein in the presence of Benzamidine Sepharose. SnkΔn protein was expressed in the absence or presence of Benzamidine-Sepharose resin and immunoblotted and probed with 3D10 antibody. Lane a, molecular mass markers. Lanes b-e, without Benzamidine-Sepharose: b, culture fluid; c, flow-through; d, wash; e, 50 mM Benzamidine elution. Lanes f-i, with Benzamidine-Sepharose: f, culture fluid; g, flow-through; h, wash; i, 50 mM Benzamidine-Sepharose. A 29×103Mr Snake catalytic chain is bound to and eluted from Benzamidine-Sepharose when the resin is present in the cell culture fluid.

The biochemical properties of recombinant Snake protein are consistent with it being a secreted serine protease zymogen. Recombinant Snake migrates in SDS protein gel electrophoresis as a closely spaced doublet of higher than predicted molecular mass. Data indicate that the difference in molecular mass between that predicted and observed can be accounted for by both N- and O-linked glycosylation. The ability to rescue snake null phenotypes by microinjection of recombinant Snake zymogen into the perivitelline space indicates that the site of action of Snake is clearly within the perivitelline space and that the conversion of Snake from zymogen to active form occurs there. The global rescue which results from local microinjection indicates that Snake is capable of diffusing from the site of injection to other parts of the PVS. Phenotypic rescue is more efficient at the syncitial blastoderm stage compared to earlier cleavage stages. This suggests that Snake protease activity is only required shortly before cellularization and Snake protein may not be extremely stable in the perivitelline space if it is present too long before the syncitial to cellular blastoderm transition.

Recombinant Snake zymogen binds to immobilized soybean trypsin inhibitor. This suggests that Snake is a probable target of soybean trypsin inhibitor induced dorsalization in the microinjection experiments of Stein (Stein and Nuesslein-Volhard, 1992). Naturally, this does not preclude other proteases within the signal transduction pathway from also being inhibited by STI.

In contrast to results obtained with a free active catalytic chain produced by microinjection of snkΔne RNA, neither the Snake zymogen nor the already cleaved IEGR-Snake protein exhibits the ability to polarize the gastrulation pattern. The degree of phenotypic rescue is higher when recombinant Snake is injected into the ventral side rather than the dorsal side of the perivitelline space. This suggests that Snake protease activity may be required preferentially or exclusively on the ventral side of the space. Upon injection of snkΔne RNA transcripts, normal gastrulation polarity is only observed when the RNA is introduced locally in the cortical cytoplasm on the ventral side of the egg (Smith and DeLotto, 1994). These results argue that during normal development, the Snake zymogen becomes activated preferentially or exclusively on the ventral side of the embryo.

The nature of IEGR-Snake rescue activity

The data presented here suggest that IEGR-Snake RNA transcripts rescue the phenotype because the IEGR-Snake protein autoactivates in the embryo. When expressed as a recombinant protein, IEGR-Snake becomes proteolytically cleaved in a process which requires the integrity of its own active site serine. This would require that the substrate specificity of Snake is somewhat close to that of blood clotting factor X. The Easter zymogen, which is the next identified protease in the cascade downstream of Snake, requires a cleavage after the amino acid residues SNR (Chasen and Anderson, 1989). SNR was tested in the context of the snake activation peptide sequence and it was also active in phenotypic rescue (see Table 2). Recent results from coexpression experiments indicate that active Snake can activate the zymogen form of Easter (unpublished data).

The activated protease and the role of the proenzyme polypeptide

IEGR-Snake protein functions through the normal signaling pathway since it requires the activities of the other dorsal group genes for phenotypic rescue. In contrast, microinjection of snkΔne RNA transcripts polarizes the gastrulation pattern and the RNA concentration defines dorsal-ventral cell fates in a concentration dependent manner. Furthermore, snkΔne RNA transcripts produce a dominant effect in wild-type embryos and embryos from females homozygous for strong alleles of pipe, nudel, windbeutal and gastrulation defective.

The ability of IEGR-Snake to autoactivate suggests that the Snake zymogen normally has some low level of proteolytic activity due to spontaneous activation. Conversion of Snake’s activation peptide to IEGR enables Snake to cleave itself at the activation peptide sequence. Spontaneous activation of one molecule of IEGR-Snake leads to a conversion of other IEGR-Snake molecules to their active forms. The proportion of IEGR-Snake which is found in the cleaved form relative to the zymogen form is variable depending upon the individual protein preparation. This interpretation is also supported by the observation that in all recombinant wild-type Snake preparations a 42×103 Mr band appears with variable intensity. This band is absent in Snake (S to A). The 42×103 Mr band is consistent with autoproteolytic cleavage in a trypsin-like site that maps between the acidic and DSN domains (see Fig. 1). There is a precedent for the zymogen form of coagulation factor VII having protease activity (Nemerson, 1988).

The primary difference between SnkΔn and IEGR-Snake is that the SnkDn produces an activated catalytic chain which is missing the proenzyme polypeptide, while IEGR-Snake produces an active form in which the proenzyme polypeptide is covalently associated with the catalytic chain. The proenzyme polypeptides of many serine proteases contain various functional domains which serve regulatory functions and modulate the activity of the catalytic chain (Furie and Furie, 1992). Site directed mutagenesis of Snake has revealed that the proenzyme polypeptide has at least two functional domains, an acidic region (ACID) and the disulfide knot (DSN) (see Fig. 1). While, the acidic domain is not required for the generation of dorsal-ventral asymmetry, the remainder including the disulfide knot is essential for normal spatial regulation of Snake (Smith et al., 1994).

The proenzyme polypeptide most likely regulates the activity of the catalytic chain by directly binding to it much as a protease inhibitor does. Recently, it has become clear that proenzyme polypeptides can function as strong inhibitors of the mature protease (Baker et al., 1993). Proenzyme kringle domains, a three disulfide, triple loop sequence of about 80 to 85 residues mediate protein-protein interactions and in the case of Thrombin inhibit activity of the catalytic chain (Patthy et al., 1984; Tulinsky, 1991; Wu et al., 1994). The disulfide knot domain of Snake is similar in size and disulfide arrangement to a kringle and thus may function similarly (Smith and DeLotto, 1992).

This leads us to propose a model in which the disulfide knot domain alternately mediates interaction with the activation complex or binds intramolecularly to the catalytic chain inhibiting its protease activity. In this model, the disulfide knot binds to an exosite on the catalytic chain of the zymogen and blocks substrate access (Fig. 8). During physiological activation (Fig. 8A), Snake zymogen docks via the disulfide knot with the activation complex and subsequently becomes cleaved. With the autoactivated (IEGR-Snake) form (Fig. 8B), although the protease is already cleaved, the catalytic chain remains inhibited by the binding of the disulfide knot preventing ectopic Snake activity. IEGR-Snake can be recruited by a vacant ventrally localized activation complex reconstituting the normal Snake activation complex configuration. This model explains how the magnitude and polarity of the ventralizing signal is normal even if excess IEGR-Snake is introduced ectopically within the perivitelline space.

Fig. 8.

A model for intramolecular inhibition of the Snake catalytic chain by the proenzyme polypeptide. (A) Normal Snake activation. Step 1: Snake zymogen is recruited to a localized activation complex. The proenzyme polypeptide (PRO) which was bound to the catalytic chain (CAT) is released and binds to a receptor in the activation complex. Step 2: cleavage and zymogen activation of Snake occur. (B) IEGR-Snake rescue. Step 1: IEGR-Snake becomes autoactivated due to low levels of spontaneous activation and autoproteolysis, however the proenzyme polypeptide blocks substrate binding to activated IEGR-Snake catalytic chain in solution. Step 2: IEGR-Snake is recruited to a waiting activation complex via a receptor which binds to the proenzyme polypeptide. IEGR-Snake is now activated and bound to a correctly localized activation complex. Proenzyme receptor interactions are proposed to occur through the disulfide knot domain (DSN).

Fig. 8.

A model for intramolecular inhibition of the Snake catalytic chain by the proenzyme polypeptide. (A) Normal Snake activation. Step 1: Snake zymogen is recruited to a localized activation complex. The proenzyme polypeptide (PRO) which was bound to the catalytic chain (CAT) is released and binds to a receptor in the activation complex. Step 2: cleavage and zymogen activation of Snake occur. (B) IEGR-Snake rescue. Step 1: IEGR-Snake becomes autoactivated due to low levels of spontaneous activation and autoproteolysis, however the proenzyme polypeptide blocks substrate binding to activated IEGR-Snake catalytic chain in solution. Step 2: IEGR-Snake is recruited to a waiting activation complex via a receptor which binds to the proenzyme polypeptide. IEGR-Snake is now activated and bound to a correctly localized activation complex. Proenzyme receptor interactions are proposed to occur through the disulfide knot domain (DSN).

The results presented here indicate that there exist spatial constraints on the activity of Snake in the embryo. Regulation of protease activity in the blood coagulation cascade is achieved at multiple levels (Esmon, 1993; Mann et al., 1990). Circulating protease inhibitors remove the active forms from the blood and reduce the undesirable effects of spontaneous activation. There, protease activation always occurs in the context of a membrane bound complex which includes a cofactor (Mann et al., 1988, 1982). Our results suggest that activation of Snake also occurs in the context of a ventrally localized activation complex. An elucidation of both the components and the architecture of this activation complex will be necessary to understand exactly how the embryo generates a spatially graded distribution of ligand.

We would like to thank Paul Schedl for assistance with monoclonal antibody production, Yvonne DeLotto for technical assistance in microinjection, David Stein, Siegfried Roth and Charles Craik for helpful discussions and Carl Hashimoto for communicating results prior to publication. This work was supported by grant IBN-9405619 from the National Science Foundation to R. D.

Allore
,
R.
and
Barber
,
B.
(
1984
).
A recommendation for visualizing disulfide bonding by one-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis
.
Anal. Biochem
.
137
,
523
527
.
Anderson
,
K.
(
1987
).
Dorsal-ventral pattern embryonic pattern genes of Drosophila
.
Trends Genet
.
3
,
91
97
.
Baker
,
D.
,
Shiau.
A.
and
Agard
,
D.
, (
1993
).
The role of pro regions in protein folding
.
Curr. Opin. Cell Biol
.
5
,
966
970
.
Chasen
,
R.
and
Anderson
,
K. V.
(
1989
).
The role of easter, an apparent serine protease, in organizing the dorsal-ventral axis of the Drosophila embryo
.
Cell
56
,
391
400
.
Chasen
,
R.
,
Jin
,
Y.
and
Anderson
,
K.
(
1992
).
Activation of the easter zymogen is regulated by five other genes to define dorsal-ventral polarity in the Drosophila embryo
.
Development
115
,
607
616
.
DeLotto
,
R.
and
Spierer
,
P.
(
1986
).
A gene required for the specification of dorsal-ventral pattern in Drosophila appears to encode a serine protease
.
Nature
323
,
688
692
.
Driever
,
W.
and
Nüsslein-Volhard
,
C.
(
1988
).
A gradient of bicoid protein in Drosophila embryos
.
Cell
54
,
83
93
.
Esmon
,
C.
(
1993
).
Cell mediated events that control blood coagulation and vascular injury
.
Ann. Rev. Cell Biol
.
9
,
1
26
.
Furie
,
B.
and
Furie
,
B.
(
1992
).
Molecular and cellular biology of blood coagulation
.
New England J. Med
.
326
,
800
806
.
Govind
,
S.
and
Steward
,
R.
(
1991
).
Dorsoventral pattern formation in Drosophila
.
Trends Genet
.
7
,
119
125
.
Harlow
,
E.
and
Lane
,
D.
(
1988
).
Antibodies: A Laboratory Manual
.
New York
:
Cold Spring Harbor Laboratory Press
.
Hashimoto
,
C.
,
Hudson
,
K.
and
Anderson
,
K.
(
1988
).
The Toll gene of Drosophila, required for dorsal-ventral embryonic polarity, appears to encode a transmembrane protein
.
Cell
52
,
269
279
.
Hecht
,
P.
and
Anderson
,
K.
(
1992
).
Extracellular proteases and embryonic pattern formation
.
Trends Cell Biol
.
2
,
197
202
.
Klaembt
,
C.
and
Schmidt
,
O.
(
1986
).
Developmental expression and tissue distribution of the lethal (2) giant larvae protein of Drosophila melanogaster
.
EMBO J
.
5
,
2955
2961
.
Konrad
,
K.
and
Marsh
,
L.
(
1990
)
gastrulation defective encodes a serine protease In Abstracts of the 31st Annual Drosophila Conference Asilomar, CA
.
Kunkel
,
T.
,
Roberts
,
J.
and
Zakour
,
R.
(
1987
).
Rapid and efficient site-specific mutagenesis without phenotypic selection
.
Meth. Enzymol
.
154
,
367
382
.
Mann
,
K.
,
Jenny
,
R.
and
Krisnaswamy
,
S.
(
1988
).
Cofactor proteins in the assembly and expression of blood clotting enzyme complexes
.
Ann. Rev. Biochem
.
57
,
915
956
.
Mann
,
K.
,
Nesheim
,
M.
,
Church
,
W.
,
Haley
,
P.
and
Krishnaswamy
,
S.
(
1990
).
Surface-dependent reactions of the vitamin K-dependent enzyme complexes
.
Blood
76
,
1
16
.
Mann
,
K.
,
Niesheim
,
M.
,
Tracy
,
P.
,
Hibbard
,
L.
and
Bloom
,
J.
(
1982
).
Assembly of the prothrombinase complex
.
Biophys. J
.
37
,
106
107
.
Miltenburger
,
H.
,
David
,
P.
and
Wulf
,
H.
(
1976
).
In Proceedings of the 1st International Colloquium of Invertebrate Physiology
, pp.
393
394
.
Queen’s University, Kingston
,
Ontario
.
Morisato
,
D.
and
Anderson
,
K.
(
1994
).
The spaetzle gene encodes a component of the extracellular signalling pathway establishing the dorsal- ventral pattern of the Drosophila embryo
.
Cell
76
,
677
688
.
Nemerson
,
Y.
(
1988
).
Tissue factor and hemostasis
.
Blood
71
,
1
8
.
Nuesslein-Volhard
,
C.
and
Roth
,
S.
(
1989
).
Axis determination in insect embryos
. In
Cellular Basis of Morphogenesis
, Vol.
144
. pp.
37
-
55
.
New York
:
John Wiley and Sons
.
Overall
,
R.
and
Jaffe
,
L.
(
1988
).
Patterns of ionic currents through Drosophila follicles and eggs
.
Dev Biol
.
108
,
102
119
.
Patthy
,
L.
,
Trexler
,
M.
,
Vali
,
Z.
,
Banyai
,
L.
and
Varadi
,
A.
(
1984
).
Kringles: modules specialized for protein binding
.
FEBS Lett
.
171
,
131
136
.
Roth
,
S.
(
1993
).
Mechanisms of dorsal-ventral axis determination in Drosophila embryos revealed by cytoplasmic transplatations
.
Development
117
,
1385
1396
.
Roth
,
S.
(
1994
).
Proteolytic generation of a morphogen
.
Curr. Biol
.
4
,
755
757
.
Rushlow
,
C.
and
Arora
,
K.
(
1990
).
Dorsal ventral polarity and pattern formation in the Drosophila embryo
.
Semin. Cell Biol
.
1
,
137
149
.
Schneider
,
D.
,
Hudson
,
K.
,
Lin
,
T.-Y.
and
Anderson
,
K.
(
1991
).
Dominant and recessive mutations define functional domains of Toll, a transmembrane protein required for dorsal-ventral polarity in the Drosophila embyro
.
Genes Dev
.
5
,
797
807
.
Schneider
,
D.
,
Jin
,
Y.
,
Morisato
,
D.
and
Anderson
,
K.
(
1994
).
A processed form of the Spaetzle protein defines dorsal-ventral polarity in the Drosophila embryo
.
Development
120
,
1243
1250
.
Smith
,
C.
and
DeLotto
,
R.
(
1992
).
A common domain within the proenzyme regions of the Drosophila snake and easter proteins and Tachypleus proclotting enzyme defines an new subfamily of serine proteases
.
Protein Science
1
,
1225
1226
.
Smith
,
C.
and
DeLotto
,
R.
(
1994
).
Ventralizing signal determined by protease activation in Drosophila embryogenesis
.
Nature
368
,
548
551
.
Smith
,
C.
,
Giordano
,
H.
and
DeLotto
,
R.
(
1994
).
Mutational analysis of the Drosophila snake protease: an essential role for domains within the proenzyme polypeptide chain
.
Genetics
136
,
1355
1365
.
Stanley
,
K.
and
Luzio
,
J.
(
1984
).
Construction of a new family of high efficiency bacterial expression vectors: identification of cDNA clones coding for human liver proteins
.
EMBO J
.
3
,
1429
1435
.
Stein
,
D.
and
Nuesslein-Volhard
,
C.
(
1992
).
Multiple extracellular activities in Drosophila egg perivitelline fluid are required for establishment of embryonic dorsal-ventral polarity
.
Cell
68
,
429
440
.
Stein
,
D.
,
Roth
,
S.
,
Vogelsang
,
E.
and
Nuesslein-Volhard
,
C.
(
1991
).
The polarity of the dorsoventral axis in the Drosophila embryo in defined by an extracellular signal
.
Cell
65
,
725
735
.
Stroud
,
R.
,
Kossiakoff
,
A.
and
Chambers
,
J.
(
1977
).
Mechanisms of zymogen activation
.
Ann. Rev. Biophys. Bioeng
.
6
,
177
193
.
Summers
,
M.
and
Smith
,
G.
(
1987
).
A manual of methods for Baculovirus vectors and insect cell culture proceedures
.
Texas Agricultural Station Bulletin No. 1555
.
Tulinsky
,
A.
(
1991
).
The structures of domains of blood proteins
.
Thromb. Haemostasis
66
,
16
31
.
Van der Meer
,
J.
(
1977
).
Optical clean and permanent whole mount preparations for phase-contrast microscopy of cuticular structures of insect larvae
.
Dros. Info. Serv
.
52
,
160
.
Wu
,
Q.
,
Picard
,
V.
,
Aiach
,
M.
and
Sadler
,
J. E.
(
1994
).
Activation-induced exposure of the thrombin anion-binding exosite
.
J. Biol. Chem
.
269
,
3725
3730
.