Adult crayfish, A. astacus, were obtained from a commercial supplier (Keller, Augsburg, Germany) and kept as previously described [14].

To locate the sites of biosynthesis and activation of astacin by immunoassay, 18 crayfish were starved for 1 week. The production of digestive enzymes was then stimulated by artificial emptying of the stomach by inserting a curved glass capillary through the oesophagus. Groups of six specimens were killed either immediately or after 3 or 7 h. Parts of the hepatopancreases from each group were processed for routine light microscopy and immunohistochemistry as described below. Other parts of the same hepatopancreases were used to prepare protein extracts for immunoblotting. In addition, gastric fluid was collected from three animals per group at 0, 3 and 7 h. All samples for immunoblot analysis were directly transferred to SDS gel-loading buffer (50 mm Tris/HCl, pH 6.8, 100 mm dithiothreitol, 2% SDS, 0.1% bromphenol blue, 10% glycerol) which contained 50 mm EDTA and 50 mm EGTA added as metalloprotease inhibitors. The samples were heated to 100 °C, homogenized using 50 strokes of the Microfuge pestle and centrifuged at 13 000 g for 5 min. The supernatant was then decanted and kept at −20 °C until use.

Peptide syntheses were carried out on a modified Applied Biosystems 430A peptide synthesizer using the in situ neutralization protocol for Boc chemistry solid-phase peptide synthesis as described in [15]. Peptides were deprotected and cleaved from the resin with HF, purified by RP-HPLC, and analysed by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) MS. The sequence of the peptides (ProAst-N, ProAst-C) used for preparation of antibodies and the peptides (AstG, AstR, AstR*) used for cleavage assays is given in Fig. 1 and Fig. 4b.

For immunohistochemistry and immunoblotting, antisera against astacin and pro-astacin were used. The astacin antiserum was prepared as previously described [14]. In addition, two different pro-astacin antisera were raised in rabbits, which were immunized with the synthetic peptides ProAst-N and ProAst-C (Fig. 1) conjugated via succinimidyl maleimidobenzoate to keyhole limpet haemocyanin. Booster injections were performed 14, 28, and 56 days after the first immunization. The antisera anti-(ProAst-N) and anti-(ProAst-C) from the final bleed, 5 days after the last boost, yielded the best reactivity and were therefore used in this study. The immunization for the pro-astacin antiserum was carried out by Pineda Antibody Service, Hamburg, Germany. For immunohistochemistry and immunoblotting, fluorescein isothiocyanate-conjugated goat anti-(rabbit IgG) Ig (Dianova, Hamburg, Germany) and alkaline phosphatase-conjugated goat anti-(rabbit IgG) Ig (Sigma, Steinheim, Germany) were used as secondary antibodies, respectively.

For immunohistochemistry and routine light microscopy, the hepatopancreases were fixed in Susa fixative and Bouin’s solution without acetic acid [14], respectively, and then dehydrated in a graded series of ethanol. Finally, the tissues were embedded in paraffin using methylbenzoate as intermedium. Sections (≈ 6-µm thick) were deparaffinized and either stained with Goldner’s stain for routine microscopy or used for immunohistochemistry. In the latter technique, the sections were incubated for 2 h with the first antibody, anti-astacin, anti-(ProAst-N) or anti-(ProAst-C), diluted 1 : 10 with 0.15 m NaCl/P_i (pH 7.2). After three 15-min washes in NaCl/P_i, the sections were incubated for 1 h with fluorescein isothiocyanate-conjugated anti-rabbit IgG from goat diluted 1 : 40 in NaCl/P_i. The sections were then rinsed with NaCl/P_i, stained with 0.1% Evans blue for 5 min, and mounted with NaCl/P_i containing 25 mg·mL⁻¹ diazabicyclo[2,2,2]octane to stabilize the fluorescence. Control screenings were carried out by performing the procedure without the first antibodies (anti-astacin and anti-pro-astacin) or by using anti-(A. astacus haemocyanin) serum (obtained from W. Stöcker, University of Münster, Germany) as first antibody. The results were documented with a Leitz Aristoplan microscope using various Kodak film types.

SDS/PAGE was performed according to standard protocols [16,17] using a polyacrylamide concentration of 15% in the separation gel. For immunoblot determinations, 50 µg total protein and 2 µg purified astacin were loaded, and gels were electroblotted to poly(vinylidene difluoride) membranes (Macherey & Nagel, Düren, Germany) using a semidry blotting apparatus. The protein bands of the size markers were stained with Coomassie Brilliant Blue. The membranes were blocked using 1% blocking powder® (Schleicher & Schuell, Dassel, Germany) dissolved in NaCl/P_i (0.05 m sodium phosphate, pH 7.4, 0.15 m NaCl) overnight at 4 °C and incubated with anti-astacin or anti-(pro-astacin) serum in 0.5% blocking powder in NaCl/P_i for 2 h at room temperature. The blot was washed with NaCl/P_i/0.1% Tween 20. The membranes were then incubated with alkaline phosphatase-conjugated goat anti-(rabbit IgG) Ig for 1 h. After a wash with NaCl/P_i/0.1% Tween 20, staining was carried out using the nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate system according to the manufacturer’s instructions (Boehringer, Mannheim, Germany).

Crayfish trypsin and astacin were prepared as described by Zwilling & Neurath [18], and the lyophilized enzymes were dissolved in 0.05 m Tris/HCl buffer, pH 8.0. Astacin was further purified on an affinity column with Pro-Leu-Gly-hydroxamate as ligand (Bachem, Heidelberg, Germany) coupled to CH-Sepharose 4B (Pharmacia Biotech, Uppsala, Sweden) according to the manufacturer’s recommendations. Affinity chromatography was performed as described in [19]. The cleavage assays were conducted in 0.5-mL plastic tubes containing a total volume of 20 µL, buffered by 0.05 m Tris/HCl/0.01 m CaCl₂, pH 8.0, for trypsin and 0.05 m Tris/HCl, pH 8.0, for astacin. Then 5 µg of the peptides AstG, AstR or AstR* (Fig. 4b) were incubated with 0.5 µg affinity-purified astacin or crayfish trypsin for 2 h at 30 °C. Quantification of the enzymes was based on A₂₈₀ measurements, and substrate concentrations were determined by weight. Release of cleavage products was monitored by MALDI-TOF MS.

MALDI-TOF mass spectra were recorded in the positive-ion mode with delayed extraction on a Reflex II time-of-flight instrument (Bruker-Daltonik GmbH, Bremen, Germany) equipped with a SCOUT multiprobe inlet and a 337-nm nitrogen laser. Ion-acceleration voltage was set to 20.0 kV, the reflector voltage was set to 21.5 kV, and the first extraction plate was set to 15.4 kV. Mass spectra were obtained by averaging 10–100 individual laser shots. Calibration of the spectra was performed externally by a two-point linear fit using angiotensin I (m/z 1269.69) and oxidized insulin &bgr; chain (m/z 3495.65). Samples were prepared by the thin-film protocol of Jensen et al. [20]. Aliquots (0.3 µL) of a nitrocellulose-containing saturated solution of &agr;-cyano-4-hydroxycinnamic acid in acetone were deposited on to individual spots on the target. Subsequently, 0.5 µL of the peptide sample was loaded on top of the thin-film spots and allowed to dry slowly at room temperature. Peptides digested with either trypsin or astacin were diluted 10-fold with 10% formic acid before analysis.

The hepatopancreas of crayfish is composed of a system of blindly ending hepatopancreatic tubules and adjoining collecting ducts (Fig. 2a), which finally terminate in the stomach [14]. The anterior part of the stomach, the cardia, normally includes a brownish gastric fluid rich in digestive enzymes [18]. The hepatopancreatic tubules are composed of a single-layered epithelium, which directly contacts the tubule lumen and the haemolymph space (Fig. 2b). Apart from the embryonic E cells located at the blind tips of the tubules, four cell types are present in the hepatopancreatic epithelium: R (resorptive) cells; F (fibrillar) cells (Fig. 2e); B (blister-like) cells; M (midgut) cells. R cells are responsible for absorption of the nutrients [21], F cells synthesize digestive enzymes [14], and M cells are apparently regulatory [21]. The function of B cells is still not known.

Immunohistochemistry with antibodies against the mature astacin which also recognize the pro-astacin (Fig. 3a) corroborated our earlier findings [14] that this enzyme is synthesized in the F cells only (Fig. 2d,f). After stimulation of enzyme synthesis by withdrawal of the gastric fluid [14], all F cells stained positive for astacin (Fig. 2d). Extracellular astacin was found in the lumen of all parts of the hepatopancreatic tubule system (Fig. 2c) except where it had been washed out during processing of the tissues. In controls in which anti-haemocyanin was used as first antibody instead of anti-astacin, only the haemolymph space was positive and not the hepatopancreatic epithelium or the tubule lumen (Fig. 2b). Immunohistochemistry with antibodies against ProAst-N or ProAst-C, which specifically recognize the N-terminal or C-terminal moiety of the propeptide produced principally the same results. The immunofluorescence was particularly noticeable in the Golgi bodies (Fig. 2g,h) and also along the microvillous border (Fig. 2g) where these enzymes are released into the tubular lumen. The fluorescing compartments of the F cells have previously been identified by electron microscopy as Golgi bodies [14]. In the period after stimulation of enzyme synthesis, all F cells stained positive with the pro-astacin antibodies (Fig. 2i). Interestingly, pro-astacin-related fluorescence was also found in the lumina of the hepatopancreatic tubules (Fig. 2j,k) and even in the collecting ducts (Fig. 2k). This implies that pro-astacin is activated in the more proximal parts of the hepatopancreatic tubule system on its way to the stomach, where it is exclusively found in the active form (see immunoblotting, Fig. 3a).

Immunoblotting demonstrated that the gastric fluid taken from the cardiac stomach strongly reacts with astacin antibodies, indicating the presence of active astacin in the stomach (Fig. 3a). In contrast, no trace of pro-astacin was detectable in the gastric fluid when anti-(ProAst-N) serum was applied (Fig. 3b). In the hepatopancreas homogenates, astacin antibodies recognized both (Fig. 3a) astacin and its pro form, but anti-(ProAst-N) reacted specifically with the proenzyme (Fig. 3b). Purified active astacin was detected in a control reaction only by anti-astacin, and not by anti-pro-astacin (Fig. 3a,b, lane ‘Ast’). We also noticed that 3 h after stimulation of enzyme synthesis by withdrawal of the gastric fluid, the amount of pro-astacin had considerably increased compared with active astacin (Fig. 3a). This indicates considerable new synthesis of the zymogen, which is in accordance with earlier stimulation experiments [14]. Finally, it must be emphasized that the immunoblots of hepatopancreas homogenates detected no traces of free propeptide of astacin (not shown), while, under the same conditions, pro-astacin can clearly be seen. To prove that the limited size of the propeptide would not prevents its dedection by immunobloting, in preliminary experiments we applied the peptides ProAst-N and ProAst-C to poly(vinylidene difluoride) membranes and showed immunopositive reactions under these conditions. These observations should also rule out any interference of free propeptide with our pro-astacin localization in the hepatopancreas lumen by immunofluorescence (Fig. 2j,k). Furthermore, it is reasonable to assume that intact propeptide is rapidly degraded by trypsin or astacin activity which is present in certain amounts in the hepatopancreas (Fig. 3a).

The alignment of a variety of propeptides from members of the astacin protein family shows that, in most cases, there is a basic amino-acid residue in position −1 where the activation occurs (Fig. 4a). In addition, an accumulation or clustering of arginine or lysine residues N-terminally to position −1 is obvious in many cases (BMP-1, Tolloid, Xolloid, BP-10, SpAN, HCH-1, HMP-1). The conservation of these residues over long periods of evolution suggests a common function, the significance of which is still not known. On the other hand, in several cases, because of the lack of basic residues at position −1, tryptic cleavage is not possible. This is true for nephrosin, PMP-1 and HMP-1, and also for astacin and AEA. For astacin, a basic residue next to the activation site is at position −3, but in AEA there is no basic side chain present at all in the whole propeptide. In HMP-1, the experimentally established N-terminus of the core enzyme is a lysine residue, which also cannot be due to tryptic cleavage (Fig. 4a). These different observations suggest that tryptic activation may play a major role in the processing of the astacins. For instance, it has been shown that human pro-meprin can be activated by the action of trypsin [22]. However, in several other cases in which a basic residue is missing at the activation site, other enzymes must also be involved. Such differences are not surprising given the wide range of occurrence, functional divergence, and tissue specificity of the astacins.

In the digestive tract of the crayfish, two endopeptidases have been described that could be responsible for the activation of pro-astacin: crayfish trypsin and astacin [18]. The different cleavage specificities of these two proteases must be decisive for correct processing of pro-astacin (Fig. 5).

We investigated the effect of crayfish trypsin and astacin on synthetic peptides overlapping the activation site (Fig. 4b) by MS analysis. The amino-acid sequence of the propeptide of pro-astacin was deduced from the cDNA structure [11]. The propeptide extends from position −34 to −1 and is preceded by a presequence or signal sequence, and it thus comprises 34 amino-acid residues. The demarcation of the propeptide was accomplished N-terminally by prediction with the program signalseq (German Cancer Research Centre, Heidelberg, Germany) and C-terminally by direct Edman amino-acid sequencing of active astacin [7] (Fig. 1).

Even with highly purified astacin or crayfish trypsin, minor traces of contaminating proteolytic activity were still detected by MS analysis, which could give rise to misleading interpretations. We therefore suppressed remaining traces of tryptic activity in astacin preparations by the addition of soybean trypsin inhibitor and/or 100 mm phenylmethanesulfonyl fluoride, and remaining traces of astacin activity in trypsin preparations by the addition of Pro-Leu-Gly-hydroxamate [19]. However, even in the presence of these inhibitors, a minor amount of tryptic activity was occasionally observed in astacin preparations, and we had to rule out the possibility that this was the result of intrinsic astacin activity. This problem was solved by shifting the specificity to the substrate itself, substituting l-(–)-arginine for d-(–)-arginine, which precluded any trypsin activity and allowed us to unambiguously distinguish between astacin and trypsin activity. There was no spontaneous hydrolysis of the peptide substrates under the incubation conditions used. The monoisotopic protonated masses for the intact peptides were 1899.0 Da for AstG and AstR*, and 1998.1 Da for AstR. In the unmodified peptide AstG, astacin hydrolyzes only the bond G−1/A+1, which represents the natural activation site (Fig. 6a). Astacin also cleaves at the same position, if Gly-1 is substituted for an arginine residue (Fig. 6b). Crayfish trypsin cleaves as expected at all available basic residues. Unmodified AstG is degraded into three peptides, but the naturally occurring N-terminus of active astacin is not produced in this way (Fig. 6d). Only when an arginine residue is introduced at position −1 (a modification that does not occur in pro-astacin) is this bond −1/+1 also hydrolysed by crayfish trypsin (Fig. 6e). With the modified peptide AstR*, in which all arginine residues are replaced by the isostereomeric d-(–)-arginine to preclude any tryptic activity, astacin still produces the correct N-terminus (Fig. 6c). At the same time, it can be demonstrated that this substrate is not altered at all by crayfish trypsin. (Fig. 6f). These results suggest that crayfish astacin is capable of catalysing its own activation. Any additional activation by trypsin would require the successive action of an aminopeptidase. We have shown that astacin is also capable of hydrolysing an Arg–Ala bond at the activation site. This suggests that for the majority of other astacin family members also, which have a basic amino-acid residue at position −1, this is not a hindrance to autocatalytic activation.

From our studies we conclude that pro-astacin is autocatalytically activated during its progress from the tubular lumen of the hepatopancreas to the cardiac stomach. Any additional contribution from crayfish trypsin would require the subsequent action of an aminopeptidase. Whether the small initial activation can be ascribed to an unknown membrane protease or intrinsic activity of pro-astacin remains to be established.