J. T. DINGLE, A. J. BARRETT AND P. D. WESTON

METHODS

Raising and testing antisera. All animals were bled before immunization to provide control sera, and pooled normal sheep serum was prepared from the blood of 11 ewes. Control sera and antisera were all stored at -20C without any preservative. Sheep and rabbit sera were labelled with the animal number (prefixed 'S' and 'R' respectively) followed by the bleed number, bleed 0 being that taken before the first injection. Initially, straight precipitin lines (Plate lb) were injected as described by Weston (1969); in later work soluble preparations of the enzyme were emulsified with complete Freund's adjuvant (1:1, v/v) and injected intramuscularly into rabbits and sheep on days 0 and 15. The animals were bled on day 25. The methods used to examine the specificity of the antisera were essentially those of Weston (1969). The relative concentrations of precipitating antibodies in the antisera were determined by single radial immunodiffusion, essentially as described by Vaerman, LebacqVerheyden, Scolari & Heremans (1969) with agar or agarose gels containing 2 or 5 units of cathepsin D/ml (Plate la). Preparation of IgG* from serum. Sera were dialysed against 20mm-sodium phosphate buffer, pH 8, and chromatographed on a column of DEAE-cellulose equilibrated with the same buffer. The products were shown by immunoelectrophoresis to be pure IgG2, and IgGI (nomenclature of Feinstein & Hobart, 1969) contaminated with a trace of /-globulin. The calculation of concentration for sheep IgG and univalent antibody fragments was based on an assumption that E% at 280nm is 14.0, by analogy with other species (Little & Donahue, 1968). Purification of anti-(chicken cathepsin D) antibodies on an immunoadsorbent column. Chicken cathepsin D was coupled to Sepharose 4B exactly as described for staphylococcal nuclease by Omenn, Ontjes & Anfinsen (1970), and the adsorbent was used in columns maintained at 40C. After application of the antiserum, the column was washed with 2 bed vol. of phosphate-buffered saline (0.80% NaCl, 0.02% KCI, 0.02% KH2PO4 and 0.12% Na2HPO4) and then the antibody was eluted with 0.1 acetic acid in 1% (w/v) NaCl. Immediately after elution, each antibody pool was neutralized with 0.5m-Na2HPO4, pH 8.9, concentrated to at least 10 mg/ml by ultrafiltration and stored at -20C. Preparation of univalent antibody fragments with papain and pepsin. A papain digest of sheep IgG, made essentially as described by Porter (1959), was separated on Sephadex G-75 into a fraction containing both fragments Fab and Fc, and another containing small peptides, which was discarded. Univalent antibody fragments were prepared from purified antibody by the method of Nisonoff, Wissler, Lipman & Woernley (1960). The pepsin digestion was at pH 5.0 for 17h, and the pepsin was finally inactivated at pH 7.7. Enzyme assay8. Cathepsin D was assayed at pH 3.2 as described by Barrett (1967, 1970). Activity was also measured at pH 5.0 in incubation mixtures (0.50 ml) containing the enzyme sample (0.lOml), serum or diluent

Abbreviation: IgG, immunoglobulin G.
(0.10ml) and buffered haemoglobin solution [4% (w/v) in 0.25m-sodium acetate buffer, pH5.0; 0.30ml]. After 2h at 45BC the reaction was stopped by addition of 5.0ml of 3.0% (w/v) trichloroacetic acid solution. The procedure was then as described for pH 3.2. The diluent was normal diluting medium consisting of 100lg of bovine serum albumin/ml in 0.1% (v/v) Triton X-100; buffered diluting medium used in some experiments contained in addition 20mM-tris-HCl buffer, pH 7.2. Other enzymes assayed by published techniques were hyaluronidase (Bollet, Bonner & Nance, 1963), arylsulphatase (Roy, 1958), acid phosphatase (Torriani, 1960), cathepsin B (Greenbaum & Fruton, 1957), fl-glucuronidase (de Duve, Pressman, Gianetto, Wattiaux & Appelmans, 1955), f-glucosidase (Conchie & Levvy, 1957), ,B-Nacetylglucosaminidase (Sellinger, Beaufay, Jacques, Doyen & de Duve, 1960) and papain, which was determined by a modification of the method of Anson (1939), casein being used as substrate. Measurement of the inhibitory capacity of antisera. Inhibition was measured at pH 5.0 at constant enzyme concentration and various antiserum concentrations. Immediately before use the sera were adjusted to pH 5.5 with 1 M-H3PO4, a capillary pH electrode being used. They were then de-gassed under reduced pressure, and the pH was readjusted to 5.0+0.1. Then 4 units of cathepsin D (lO,1l) were added to each sample of the antiserum (5-200 ,l) and normal serum was added to make the volume 210,1I. The mixture was incubated for 15min at 37C, and then 300,II of haemoglobin solution (4%, w/v) in 0.25M-sodium acetate buffer, pH5.0, was added. The mixture was incubated for 2 h at 450C before the reaction was stopped and the assay completed as described above. In some experiments inhibition was measured with a constant amount of antiserum (50 pl) and increasing amounts of cathepsin D (1-20 units). Effect of antibody/enzyme ratio on inhibition and precipitation. Antiserum (S467/2) and control normal sheep serum were incubated at 56C for 30 min to inactivate complement, de-gassed at pH 5.5 and adjusted to pH 7.0. The sera were then clarified by centrifuging at lOOOOg for Omin and passed through Millipore membrane filters of 0.45,um pore size. Incubation mixtures at six enzyme/ antibody ratios were set up in triplicate as follows. Into each tube was pipetted 2.5 ml (100 units) of chicken cathepsin D in buffered diluting medium. To each set of tubes was then added, 0, 0.16, 0.31, 0.63, 1.25 or 1.88 ml of antiserum together with sufficient normal serum to make a total of 2.5 ml of serum. The contents of the tubes were mixed and incubated for 15min at 370C, after which samples were removed for assay of enzymic activity at pH 3.2 (0.05 ml) and pH 5.0 (0.20ml) by the usual methods except that 30min incubation was used at pH 3.2. The remainders of the enzyme-antiserum mixtures were allowed to stand at 40C overnight, shaken to resuspend the immunoprecipitates and sampled for assays as before. Duplicate 1.3 ml samples were removed from each mixture to lOml tubes and centrifuged at lOOOOg for 10 min. The supernatant fractions from each pair of tubes were combined and sampled for assays as before. In addition, to detect low concentrations of enzyme the supernatants representing the three highest antiserum concentrations were assayed with larger samples (0.30 ml). The precipitates were washed by resuspension in 5.0 ml

Vol. 123

CATHEPSIN D IMMUNOINHIBITION
of buffered diluting medium and recentrifugation at 10000g. One pellet from each pair was then suspended in 1.3 ml of diluting medium with the aid of a handoperated homogenizer, and assays were made as before. The second pellet from each pair was dissolved in 1.3 ml of 0.20M-acetic acid and the E280 of the solutions was measured. Preparation of 35S-labelled chick cartilage. Embryonic cartilage was obtained in the following way. Six-dayincubated eggs were injected with 100 1d of sterile 0.9% NaCl containing 1 ,Ci of carrier-free 35SO42-. The injection was made into the sub-blastodermic space by the insertion to 1 cm of an intradermal syringe needle. The shell was resealed with paraffin wax and the eggs were incubated until day 10-12; at day 10 the long bones contained approx. 20000 c.p.m./mg dry wt. Cartilage from young animals was obtained from 5-15day-old chickens that had been injected intraperitoneally with 50,uCi of 3"SO42- in sterile 0.9% NaCl containing 0.1 mg of carrier Na2SO4. Older animals (15-30 days) were injected with 100-150juCi. Removal from cartilage of Iy8osomal proteolytic enzymes. Embryonic and young fowl long-bone cartilage labelled with 35S was washed with 100 vol. of 0.1% Triton X- 100 in 0.1 M-sodium phosphate buffer, pH 7.5, for 2h at 4C with constant agitation. The washing medium was replaced by an equal volume of fresh buffer and the washing continued for 18 h. No acid proteinase activity was then detectable in an homogenate of the washed cartilage. Action of cathepsin D on washed cartilage. Washed cartilage (20-50mg wet wt.) was incubated in a shaking water bath at 37C with 0.2 ml of 0.25M buffer, 0.3 ml of serum or 0.9% NaCl and 100,ul of cathepsin D (2-20 units/ml). The buffers used were sodium formate-formic acid, pH 3.0, sodium acetate-acetic acid, pH 4.0 and 5.0, and KH2PO4-Na2HPO4, pH 6.0 and 7.0. At the end of the incubation period (usually 2 h, since a linear rate of release was obtained for up to 6 h) the incubation medium was removed and the 35S-labelled proteoglycans were precipitated, together with 0.5 mg of chondroitin sulphate as carrier, with 10vol. of ethanol containing 0.027% of H2504. The precipitate was centrifuged, washed twice with the acid ethanol and dissolved in 0.5ml of formic acid at 100C for determination of radioactivity in a Packard Tri-Carb liquid-scintillation counter, in the scintillant mixture described by Hall & Cocking (1965). The tissue was rinsed with 0.9% NaCl, and digested in 98% formic acid, at 100C for 1 h. The use of formic acid digests decreased quenching to a minimal level. Autolysis of cartilage. Whole limb bones were dissected from 10-11-day chicken embryos; with older embryos the cartilaginous ends of the long bones were dissected free from the bony shafts, and with chicks and young fowls slices were cut from the articular surface of the long bones. Rabbit ear cartilage was taken from 6-week-old animals. The cartilage (20-50mg wet wt.) was preincubated for 30min at 0C with 100j.l of 0.5M-sodium acetate buffer, pH5.0, containing 0.5% ofTriton X-100; 0-400,u of serum or antiserum was added and the final volume made up to 500,1I with 0.9% NaCl. The tissue was incubated at 37C, usually for 2-3 h, and the released proteoglyeans were measured turbidimetrically (Weston, Barrett & Dingle, 1969), as uronic acid precipitable with cetylpyridinium chloride (Barrett, Sledge & Dingle, 1966) or radiochemi-

cally as described above. Some samples were subjected to diffusion analysis. Diffusion analysis of cartilage autolysates. Weissman & Spilberg (1968), using a specially constructed chamber, demonstrated that diffusion through a Millipore membrane filter could be used to follow the digestion of a proteoglyean; this principal was developed to give a routine method for measuring the digestion of cartilage proteoglycan. The method depends on the equilibration of digestion products through a Millipore membrane under controlled conditions of diffusion. Each Swinnex disposable filter holder (Millipore SXOO 013) was fitted with a small plastic stopper at each end and the membrane (0.10, 0.22 or 0.30nm pore size) was placed in position bounded by two 'O' rings lightly coated with silicone grease. A small glass bead was placed in each side of the chamber, so that on rotation the contents were thoroughly mixed. At the end of the autolysis period 0.5 ml of cold I.OMtris-HCl buffer, pH8.5, was added to the tissue in its incubation medium and the whole was homogenized. The homogenate was centrifuged at 10 OOOg for 5min and 200 ,lI of the supernatant was placed in the small end of the diffusion chamber. In the larger end of the diffusion chamber was placed 600Al of 0.2 M-tris-HCI buffer, pH 8.5, containing 0.02M-EDTA, 0.1% of Triton X-100 and 1% (v/v) of butan-l-ol. Three chambers were normally set up for each tissue sample; a total of 36 chambers could be handled at one time. The chambers were rotated around their long axes at 3rev./min for 24h at 4C; equilibration occurred within 18h. Samples (501l) were taken from both sides of each chamber for scintillation counting and the results are expressed as the percentages of total proteoglyean that was diffusible. The best reproducibility of results was obtained when all the membranes for an experiment were cut from a single large sheet of membrane filter.

RESULTS

Preparation of antisera and antibody fractions Antisera. Cathepsin D purified from human, rabbit and chicken liver was used to immunize rabbits and sheep. It was found that sheep receiving two injections of 3-5mg of chicken cathepsin D gave potent precipitating antisera. The relative concentrations of precipitating antibodies in the sera were measured by reversed radial immunodiffusion essentially as described by Vaerman et al. (1969) (Plate la). Antisera were raised against human and rabbit cathepsin D by injection of straight precipitin lines (Plate lb). A specific antiserum was raised in a sheep by this method by using 250,ug of antigen. All antisera were examined initially by double diffusion against a crude preparation from liver, and those that gave a single precipitin line containing cathepsin D after 24h at room temperature were designated 'specific'. All the sheep injected with preparations of chicken cathepsin D homogeneous by analytical isoelectric focusing (Barrett, 1970) produced specific antisera. Purification of antibodies. Sheep anti-(cathepsin

0.1 Antiserum (t,l/unit of cathepsin D)
Fig. 1. Inhibition of chicken and rabbit cathepsin D by sheep antisera. Assays for cathepsin D were made on suspensions containing various volumes of antiserum and a fixed volume of enzyme at pH 5.0 (see the Methods section). Anti-(chicken cathepsin D) sera: 0, S467/2; A, S570/2; El, S467/4; *, S467/4 passed through immunoadsorbent column; A, S2/1; *, Sl/l. Anti-(rabbit cathepsin D) serum: V, S522/3.
D) antibodies were isolated from various sera by the use of an immunoadsorbent column containing chicken cathepsin D covalently linked to Sepharose. Thus, when antiserum S467/4 (14ml) was passed through a 14ml bed of the inmnunoadsorbent containing 14mg of cathepsin D, 70mg of antibody was absorbed and subsequently eluted. After passage through the column the serum was free of precipitating anti-(cathepsin D) antibodies, as judged by radial immunodiffusion, and showed no capacity to inhibit the enzyme (see below). The eluted antibody fraction was shown to be pure IgG in immunoelectrophoresis developed with two multivalent anti-(sheep serum) sera. In a subsequent control experiment an equal volume of normal serum was passed through the column, and only a trace of protein was retained to be eluted with dilute acetic acid. Univalent antibody fragmentM. To investigate the relationship between immunoinhibition and precipitation, antibody fragments were prepared from sheep anti-(chicken cathepsin D) IgG. Digestion with either papain or pepsin gave non-precipitating univalent fragments. The fragments produced by papain inhibited precipitation of cathepsin D by antiserum in gel diffusion (cf. Kronvall & Williams, 1969) (Plate lc).
whereas the other sera all produced complete inhibition of the enzyme, the most effective doing so at about 6,u1/unit of cathepsin D. The inhibition of human cathepsin D by a rabbit antiserum is shown in Fig. 6. Sheep anti-(rabbit cathepsin D) (S522/3) caused complete inhibition of rabbit cathepsin D
The amount of precipitating antibodies was not found to give a reliable indication of inhibitory potency. Both IgGl and IgG2 from an immune serum were inhibitory, but the corresponding proteins from normal sheep sera tested at a high concentration (300,ug/unit of cathepsin D) were not. Effect of pH, incubation time and dilution. Inhibition of chicken cathepsin D by antiserum (S467/2) was measured at pH 3.0 and 4.0, as well
EXPLANATION OF PLATE I Immunodiffusion plates. (a) Radial immunodiffusion plate, to which duplicate samples of a pre-immunization serum (S570/0) and five sheep anti-(chicken cathepsin D) sera were applied, together with doubling dilutions of serum S467/4 representing 5.0, 2.5, 1.25, 0.62 and 0.31 ,u. (b) Straight precipitin lines for injection formed between troughs filled alternately with non-specific antiserum and suitably diluted purified isoenzyme solution. (c) Papaindigested antibody inhibiting precipitation. (d) Complete identity ofhuman cathepsin D from rheumatoid synovium and liver, shown by using anti-(human cathepsin D) serum (R18/2) fourfold concentrated. (e) Anti-(chicken cathepsin D y-isoenzyme) serum (S467/2) reacting with chicken isoenzymes. (f) Anti-(chicken cathepsin D ,-isoenzyme) serum (S570/2) reacting with chicken isoenzymes. (g) Anti-(human cathepsin D ,-isoenzyme) serum (R18/1), fourfold concentrated reacting with human isoenzymes. (h) Anti-(rabbit cathepsin D) serum (S522/1) reacting with rabbit isoenzymes.

antiserum or above/unit of cathepsin D caused complete inhibition. At pH 3.2, 9 % inhibition was reached at 18.8,u1/unit of enzyme. The assays made after the suspensions had stood overnight were essentially the same as the first set. The supemnatant fractions obtained after removal of the immunoprecipitates by centrifuging showed very similar patterns of activity at both pH 3.2 and pH 5.0 (Fig. 3). From this it seems that there is no significant amount of inhibitor that is nonprecipitating, since this would have produced high activity at pH 3 relative to pH 5, such as was found before removal of the precipitates. No enzyme activity was detected even in the

X5 0.2 C)

4Q 04

co 0.1

Antiserum (tl/unit of cathepsin D) Fig. 4. Effect of antiserum/enzyme ratio on inhibition and precipitation: assays of immunoprecipitates. The precipitates obtained after centrifuging the mixtures described in Fig. 2 were resuspended and assayed for proteolytic activity at pH 3.2 (e) and pH 5.0 (o) (see the Methods section).
assays made at pH 3.2 with 0.30ml samples of the supernatants rather than 0.05ml, confirming that there was very complete removal of the enzyme from the soluble phase at 6.3 ,l of antiserum per unit and above. The result of assays made with the resuspended immunoprecipitates is recorded in Fig. 4. The values obtained at pH 3.2 show the expected increase in the amount of enzyme precipitated with increasing amounts of antiserum, but the assays at pH 5.0 demonstrate that the enzyme precipitated at 1.6pl of antiserum/unit was largely uninhibited, and activity persisted at 3.1,ul/unit, also, but not at 6.3 ,ul/unit. The finding that precipitates formed at low antiserum/enzyme ratios retained activity raised the question of the stoicheiometry of the enzymeantibody association. From the protein content of the immunoprecipitates (measured as the E280 of their solutions in 0.2m-acetic acid), together with assays of cathepsin D at pH 3.2, it was possible to calculate the molar ratios of antibody to enzyme in the precipitates. In two experiments it was found that precipitates containing 3 or 4 antibody molecules/molecule of enzyme showed only slight inhibition ofthe enzyme they contained, whereas all activity was lost when 6 or 7 molecules of antibody were bound/enzyme molecule. Precipitates containing about 11 molecules of antibody/molecule of cathepsin D were produced in the presence of a large excess of antibody. Inhibition by purified antibody and antibody fragmentM. Purified sheep anti-(chicken cathepsin D) antibody, prepared with an immunoadsorbent column, was an effective inhibitor of the enzyme,

released in 2 h; this concentration of enzyme is comparable with that found in articular cartilage [i.e. 17 units/g at 11 days of age in ovo, 12.5 units/g at 14 days, 8.1 units/g at 17 days and after hatching, 5.9 units/g at 4 days and 5.5 units/g at 18 days of age (these values being the means of triplicate determinations on three animals at each age; J. T. Dingle, unpublished work)]. The pH optimum for the action of pure cathepsin D on cartilage was pH 5 (see Fig. 8), unlike that of cathepsin D on haemoglobin, which has been shown to be pH 3 (Barrett, 1967). The action of pure cathepsin D on articular cartilage was inhibited by antiserum; thus at pH 5.0 50,ul decreased the activity by over 90%, complete inhibition being obtained with increased quantities of antiserum. Inhibition occurred in the pH range 4-8, but as was found with haemoglobin no inhibition occurred at pH 3 (Fig. 8).

Cartilage autoly8i8

The four methods used to determine the rate of release of cartilage proteoglycans during autolysis (see the Methods section) gave similar results. The measurement of autolysis as the release of 35Slabelled ethanol-precipitable material from cartilage had the advantage that reproducible results were obtained from experiments with very small amounts

'q 60 I,

20 O 10 I0

Time (min)

Fig. 7. Action of cathepsin D on cartilage protein-polysaccharide. The viscosity of a 4.3 mg/ml solution of bovine cartilage protein-polysaccharide was followed at pH 5.0 and 370C in the presence of heat-inactivated cathepsin D (A), active chicken cathepsin D (5 units/ml) with 150,u1 of normal rabbit serum (o) and active cathepsin D (5 units/ml) with 150,u1 of rabbit anti-(chicken cathepsin D) serum (R6) (o). The results are expressed as 1rcI. X 100/ r initial.
Fig. 8. Action of cathepsin D on cartilage. Articular cartilage (50 mg) from 4-week-old fowl long bones, labelled with 35S ahd depleted of lysosomal enzymes (see the Methods section) was incubated with 2 units of chicken cathepsin D for 2 h at 37C in a total volume of 0.5 ml containing 501,l of antiserum (S570/2) (o) or normal serum (0). The results, which are the means of three experiments, each in duplicate, are the 35S released into the incubation buffer at each pH (see the Methods section) expressed as percentages of the total 35S.

Table 2. Comparison of the effect of antiwerum and pep8in-degraded antibody on the autolysis of 14-day chick cartilage
Autolysis was carried out in the usual manner for 3h at 37C. The solution of antibody fragments contained 13.5mg of proteinlml. The results are the means of three experiments. Release of 35S (%) Vol. of serum or antibody fragments Pepsin-degraded Sheep antiantibody (chicken cathep(from S467/2) sin D) antibody

(S467/2)

Time (h) Fig. 9. Autolysis of cartilage. Cartilage (100mg) labelled with 35S from chick embryos or young fowls of various ages was autolysed in O.IM-sodium acetate buffer, pH5.0, at 37C in the presence of normal sheep serum or antiserum (S570/2) in a total volume of 1.0 ml. Samples of the incubation medium were taken at various times for the determination of released 35S. *, 10-day embryonic cartilage; 0, 10-day embryonic cartilage plus 501I of antiserum; *, 5-day-old chick cartilage; E, 5-day-old chick cartilage plus 50dl of antiserum; A, 21-day-old fowl cartilage; A, 21-day-old fowl cartilage plus 501u of antiserum.

4V} 4)

Antiserum (ul)
Fig. 11. Immunoinhibition of autolysis of chicken cartilage. The 11-day embryonic cartilage (100mg) labelled with 35S was allowed to autolyse for 2.5h at pH5 and 370C in the presence of various quantities of antiserum (S570/2) made up to 400 ,lI with normal serum, in a total volume of 1.0 ml. The results, which are the means of three experiments, each done in duplicate, are expressed as the percentages of the total label that were released into the buffer S.E.M.
Fig. 10. Effect of pH on autolysis. The 11-day embryonic chick cartilage (40mg) labelled with 3sS was allowed to autolyse at 370C at various pH values. Incubation was for 2h at pH5, 4h at pH3, 4 and 6, and 6h at pH7, in the presence of 501u of antiserum (S570/2) (e) or normal serum (o), in a total volume of 0.5 ml. The buffers used were as for the action of cathepsin D on washed cartilage (see the Methods section); the results are expressed as c.p.m. of 35S released/h and are the mean of duplicate determinations at each pH.
of tissue (2-10mg dry wt.). Thus in an experiment in which 5-day-old chick cartilage was used the
initial 35 content of the cartilage (S.E.M.) was (12) c.p.m./mg; after 4h autolysis at pH 6 and 37CC 64072 (7) c.p.m./mgwas released whereas
in the presence of antiserum 29032 (8)c.p.m./mg was released. The rate of release of 35S was found to be linear with time up to 6h in both embryonic and young adult cartilage (Fig. 9). The pH optimum

DISCUSSION

(370C)
Normal serum (00C) Inhibition (%)

9.8 1.5 72

specific antiserum to cathepsin D (Fig. 11). Inhibition occurred at pH 5, was diminished at pH 4 and absent at pH 3 (Fig. 10). The antiserum was effective in inhibiting the autolysis of tissues of various ages (Fig. 9). Non-precipitating antibody fragments obtained by pepsin digestion were also capable of inhibiting autolysis (Table 2). The release of proteoglycan, determined as precipitable uronic acid during autolysis of ear
Cathepsin D of chicken, rabbit and man has proved to be a good immunogen in the species we have used, since potent antisera were produced in response to the injection of small amounts of enzyme. All of the precipitating antisera raised against cathepsin D inhibited the enzyme, and several produced complete inhibition in quantities of 6-12,u1/unit of cathepsin D. Incomplete inhibition of cathepsin D and other lysosomal enzymes by an antiserum raised against an extract of lysosomes has been reported by Trouet (1969). In several earlier studies of the immunoinhibition of

Vol. 12:3

other enzymes (Arnon, 1965; Cinader, 1967) incomplete inhibition even in antibody excess has been reported; this phenomenon may be due to the presence of non-inhibitory antibodies that interfere with the binding of inhibitory ones (Cinader & Lafferty, 1963; Arnon & Shapira, 1967), but such antibodies do not seem to be evoked by cathepsin D. The effect of pH on the immunoinhibition of cathepsin D in our experiments was predictable from the known pH-dependence of antigen-antibody reactions (Singer & Campbell, 1955). The dissociation ofthe immune complex of antibody and cathepsin D at pH 3 was illustrated by the elution of antibody from the immunoadsorbent columns by dilute acetic acid, and by the solubility of the immune precipitates in the same solvent, with the accompanying reappearance of enzymic activity. It was convenient to be able to assay total cathepsin D at pH 3 and to determine the extent of immunoinhibition in the same preparation at pH 5, and this method was applied to the study of the characteristics of inhibition at various antiserum/enzyme ratios. Enzyme remaining unprecipitated in the antiserum-enzyme mixtures was not inhibited, so that non-precipitating inhibitory antibodies were absent from the antiserum used. Precipitates formed in antigen excess possessed enzymic activity, so aggregation itself does not necessarily result in inhibition and this was further emphasized by the fact that fragment Fab', the non-precipitating univalent product of antibody degradation by pepsin, was inhibitory. These findings are entirely consistent with the results of Marshall & Cohen (1961) and Cinader & Lafferty (1963), who used univalent antibody to study the inhibition of carbamoyl phosphate synthetase and ribonuclease respectively. The requirement for a ratio of 6 antibody molecules/molecule of cathepsin D for complete inhibition by the antiserum used is compatible with the hypothesis that loss of activity is due to a steric blockade that is only complete when each molecule is surrounded by many immunoglobulin molecules (cf. Fazekas de St Groth, 1963; Cinader & Lafferty, 1964; Arnon, 1968). No effect of dilution or incubation time on the degree ofimmunoinhibition was detected, indicating that the enzyme-antibody complexes are extremely stable, and that the equilibrium concentrations of free reactants must be very low. Also, degradation of the antibodies by the antigen (cf. Arnon, 1965, with papain), which might have destroyed their inhibitory activity, does not seem to have occurred to a significant extent. Since the haemoglobin/ antibody concentration ratio varied tenfold in the dilution experiment, there was apparently no tendency for the substrate to compete with the inhibitor, at least after the initial reaction. Cinader (1967) stated that antibody is a less effective

inhibitor when combining with enzyme in the presence than in the absence of substrate and that the degree of inhibition is commonly affected by the order of mixing of enzyme, antibody and substrate; we could not detect this phenomenon in our experiments. To be of use in the elucidation of the physiological role of cathepsin D in connective tissue metabolism, it is clear that the antisera must be specific; we therefore supplemented the evidence from immunodiffusion, which demonstrated the specificity of the precipitating activity of the sera, with experiments on the specificity of inhibition. As expected from the known specificity of immunological reactions, the antisera to chicken cathepsin D did not inhibit a number of other enzymes that might contribute to connective-tissue turnover. It has been shown that antisera to human and rabbit cathepsin D, raised in rabbit and sheep respectively, do not react with cathepsin D from a number of other mammals, whereas antisera to chicken cathepsin D raised in rabbits reacted with cathepsin D from a variety of different species of bird (P. D. Weston, unpublished work). Although we are primarily concerned with the function of cathepsin D in cartilage, this was not a practicable source of the enzyme and it was necessary to raise antisera against cathepsin D prepared from liver. It was possible that antisera raised against the liver enzyme might fail to react with that of cartilage, for several enzymes are known to occur in organ-specific forms that are immunologically distinct (Sussman, Small & Cotlove, 1968; Henion & Sutherland, 1957; Sherwin, Karpati & Bulcke, 1969). However, Weston (1969) reported that cathepsin D of chicken liver, heart, kidney, spleen, testis, brain and embryonic bone rudiments are indistinguishable in immunoprecipitation; we have now shown similar tissue non-specific imnmunoinhibition of acid proteinase activity of extracts made from liver, cartilage or kidney of 6-day-old chicks. Also, cathepsin D from human rheumatoid synovium was indistinguishable in immunodiffusion from the human liver enzyme. The finding that the major isoenzymes of chicken liver cathepsin D are immunologically identical, both in precipitation and in inhibition, is not unexpected, since the molecules probably differ in only a few amino acid residues (Barrett, 1971). The work reported in the first part of this paper formed the basis for an immunoenzymic investigation of the role of cathepsin D in connective-tissue metabolism. Thus in the second part of the work it has proved possible to inhibit the action of cathepsin D on purified bovine nasal protein-polysaccharide and on cartilage that had been depleted of its own proteinases. More important was the finding that the degradation of young chicken and rabbit

REFERENCES

Ali, S. Y. (1964). Biochem. J. 93, 611. Ali, S. Y., Evans, L., Stainthorpe, E. & Lack, C. H. (1967). Biochem. J. 105, 549. Anson, M. L. (1939). J. gen. Phy8iol. 22, 79. Arnon, R. (1965). Immunochemi8try, 2, 107,
Arnon, R. (1968). Eur. J. Biochem. 5, 583. Arnon, R. & Shapira, E. (1967). Biochemi8try, Ea8ton, 6, 3942. Barrett, A. J. (1967). Biochem. J. 104, 601. Barrett, A. J. (1968). In Comprehensive Biochemi8try, vol. 26B, p. 425. Ed. by Florkin, M. & Stotz, E. H. Amsterdam: Elsevier Publishing Co. Barrett, A. J. (1969). Biochem. J. 115, 36r. Barrett, A. J. (1970). Biochem. J. 117, 601. Barrett, A. J. (1971). In Tissue Proteinases, p. 109. Ed. by Barrett, A. J. & Dingle, J. T. Amsterdam: NorthHolland Publishing Co. (in the Press). Barrett, A. J., Sledge, C. B. & Dingle, J. T. (1966). Nature, Lond., 211, 83. Bollett, A. J., Bonner, W. M. & Nance, J. L. (1963). J. biol. Chem. 238, 3522. Bollett, A. J., Handy, J. R. & Sturgill, B. C. (1963). J. clin. Inve8t. 42, 853. Cinader, B. (1967). In Antibodies to Biologically Active Molecule8, Proc. FEBS 2nd Meet., War8aw, vol. 1, p. 85. Cinader, B. & Lafferty, K. J. (1963). Ann. N.Y. Acad. Sci. 103, 653. Cinader, B. & Lafferty, K. J. (1964). Immunology, 7, 342. Conchie, J. & Levvy, G. A. (1957). Biochem. J. 65, 389. de Duve, C., Pressman, B. C., Gianetto, R., Wattiaux, R. & Appelmans, F. (1955). Biochem. J. 60, 604. Dingle, J. T. (1962). Proc. B. Soc. Med. 55, 109. Dingle, J. T. (1968). In Rheumatic Di8eases, p. 80. Ed. by Duthie, J. J. R. & Alexander, W. R. M. Edinburgh University Press. Dingle, J. T. (1969). In Lyoo0omes in Biology and Pathology, vol. 2, p. 421. Ed. by Dingle, J. T. & Fell, H. B. Amsterdam: North-Holland Publishing Co. Fazekas de St Groth, S. (1963). Ann. N.Y. Acad. Sci. 103, 674. Feinstein, A. & Hobart, M. J. (1969). Nature, Lond., 223, 950. Fisher, D., Higham, M., Kent, P. W. & Pritchard, P. (1966). Biochem. J. 98, 46P. Fisher, D., Whitehouse, M. W. & Kent, P. W. (1967). Nature, Lond., 213, 204. Gerber, B. R., Franklin, F. C. & Schubert, M. (1960). J. biol. Chem. 235, 2870. Greenbaum, L. M. & Fruton, J. S. (1957). J. biol. Chem. 226, 173. Hall, T. C. & Cocking, E. C. (1965). Biochem. J. 96, 626. Hamerman, D. & Barland, P. (1966). Bull. rheum. Di8. 16, 396. Henion, W. F. & Sutherland, E. W. (1957). J. biol. Chem. 224, 477. Kronvall, G. & Williams, R. C. (1969). J. Immun. 103, 828. Little, J. R. & Donahue, H. (1968). In Methods in Immunology and Immunochemi8try, vol. 2, p. 346. Ed. by Williams, C. A. & Chase, M. W. New York: Academic Press Inc. Lucy, J. A., Dingle, J. T. & Fell, H. B. (1961). Biochem. J. 79, 500. Marshall, M. & Cohen, P. P. (1961). J. biol. Chem. 236,
Morrison, R. I. G. (1970). In Chemistry and Molecular Biology of the Intercellular Matrix, p. 1683. Ed. by Balazs, E. A. New York: Academic Press Inc.

Nisonoff, A., Wissler, F. C., Lipman, L. N. & Woernley, D. L. (1960). Arch8 Biochem. Biophys. 89, 230. Omenn, G. S., Ontjes, D. A. & Anfinsen, C. B. (1970). Nature, Lond., 225, 189. Page-Thomas, D. P. (1969). In Ly8o8omes in Biology and Pathology, vol.2, p. 87. Ed. by Dingle, J. T. & Fell, H. B. Amsterdam: North-Holland Publishing Co. Porter, R. R. (1959). Biochem. J. 73, 119. Roy, A. B. (1958). Biochem. J. 68, 519. Sellinger, 0. Z., Beaufay, H., Jacques, P., Doyen, A. & de Duve, C. (1960). Biochem. J. 74, 450. Sherwin, A. L., Karpati, G. & Bulcke, J. A. (1969). Proc. natn. Acad. Sci. U.S.A. 64, 171. Singer, S. J. & Campbell, D. H. (1955). J. Am. chem. Soc. 77, 3504.
Sussman, H. H., Small, P. A. & Cotlove, E. (1968). J. biol. Chem. 243, 160. Torriani, A. M. (1960). Biochim. biophy8. Acta, 38, 460. Trouet, A. (1969). Caracteristiques et Propriete8 Antigenique de8 Ly8osome8 du Foie. Louvain: Vander. Vaerman, J. P., Lebacq-Verheyden, A. M., Scolari, L. & Heremans, J. F. (1969). Immunochemistry, 6, 287. Weissman, G. & Spilberg, I. (1968). Arthr. Rheum. 11, 162. Weston, P. D. (1969). Immunology, 17, 421. Weston, P. D., Barrett, A. J. & Dingle, J. T. (1969). Nature, Lond., 222, 285. Woessner, J. F. (1967). Fedn Proc. Fedn Am. SoC8 exp. Biol. 26, 607.