INFECTION AND IMMUNITY, Dec. 1973, p. 1000-1008Copyright 0 1973 American Society for Microbiology

Vol. 8, No. 6Printed in U.S.A.

Purification and Properties of the Cathepsin DType Proteinase from Beef and Rabbit Lung and

Its Identification in MacrophagesOSCAR ROJAS-ESPINOSA, ARTHUR M. DANNENBERG, JR., PATRICK A. MURPHY, PATRICIA A.

STRAAT, P. C. HUANG, AND STEPHEN P. JAMES

Departments of Environmental Medicine, Epidemiology, and Biochemistry and Biophysics, School ofHygiene and Public Health, and Departments of Pathology and Microbiology, School of Medicine,

The Johns Hopkins University, Baltimore, Maryland 21205

Received for publication 2 May 1973

The acid-acting proteinase, cathepsin D (EC 3.4.4.23), was purified fromextracts of homogenized rabbit lung and beef lung by autolysis at acid pH,acetone and ammonium sulfate fractionation, column chromatography, andisoelectric focusing. Four isoenzymes were obtained from each source. With acidhemoglobin as the substrate, the proteinase from rabbit lung had a pH optimumof 3.0 and that from beef lung had a pH optimum of 3.6. Their activity was notaffected by thiol reagents or by Fe2+, Mn2+, or Mg2+. One isoenzyme from rabbitlung was used to immunize a goat, and one from beef lung was used to immunizea rabbit. In immunoelectrophoresis, each resulting antiserum formed a singleprecipitin line with its homologous enzyme. They cross-reacted with the otherthree isoenzymes from the same species, but not with any isoenzyme from theother species. At high concentrations, each antiserum completely inhibited theproteolytic activity of its homologous enzyme. The antiserum against rabbit lungcathepsin D also inhibited the proteolytic activity of rabbit peritoneal andpulmonary macrophages. In limited quantities, this antiserum has now been madecommercially available and is being used with labeled antibody techniques to iden-tify under a microscope the presence of cathepsin D in macrophages and to studyits role in the pathogenesis of tuberculosis.

The lung not only is an organ of respiration,but also a major defense organ, disposing ofinfectious and other inhaled agents. A substan-tial part of the lung parenchyma consists ofalveolar macrophages, which often accomplishthese defense functions by means of hydrolyticenzymes.

In 1955, we partially purified from beef lung aproteinase having a pH optimum between 3 and4, with hemoglobin as the substrate (12, 13). Ithydrolyzed the B chain of insulin at the samesites as pepsin (13) and cathepsin D (7, 19, 21,47), namely adjacent to the aromatic, longchain aliphatic, and dicarboxylic amino acids.Although cathepsin D has also been purifiedfrom liver (4, 5), spleen (19, 21, 33, J. B.Ferguson, J. R. Andrews, I. M. Voynick, and J.S. Fruton, Biochemistry, in press), and uterus(47), its complete purification from lung has notbeen previously accomplished. This report de-scribes such a purification from both rabbit andbeef lungs. The cathepsin D probably comesfrom the macrophages that lung contains.

Antibodies to the purified rabbit and beefproteinases were produced in goats and rabbits,respectively. These antibodies were used (i) toconfirm the purity of each enzyme by immuno-electrophoresis, (ii) to inhibit both the lung andmacrophage enzymes, and (iii) to confirm thepresence of this proteinase in macrophages by alabeled antibody histochemical technique (cf.15, 32, 43). The latter is described in anotherreport (0. Rojas-Espinosa, A. M. Dannenberg,Jr., L. Sternberger, and T. Tsuda, Amer. J.Pathol., in press), which evaluates the role ofthis cathepsin D in the pathogenesis of tuber-culosis.

MATERIALS AND METHODSSources of enzyme. Frozen rabbit lungs, type 3,

were purchased from Pel-Freez Biologicals Inc., Rog-ers, Ark. Fresh beef lungs were purchased from a localslaughter house. All lungs had been trimmed andwashed. Granulomatous rabbit lungs were producedin 25 animals by the intravenous injection of 0.3 ml ofheat-killed, lyophilized tubercle bacilli (BCG) (1.0

1000

on May 15, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

VOL. 8, 1973 CATHEPSIN D TYPE MACROPHAGE PROTEINASE FROM LUNG

mg/ml) in Aquaphor-Marcol 52 (1:4) (27, 30). Four-teen days later, these lungs, containing multipleconfluent granulomata, were removed, trimmed,washed in distilled water, and frozen at -20 C forseveral weeks.

Rabbit peritoneal macrophages (MN) were ob-tained by injecting 35 ml of mineral oil in-traperitoneally into rabbits and, 4 to 6 days later,collecting the MN in citrated saline solution (10, 11).The cell suspension was centrifuged, suspended incitrated saline, counted in a hemacytometer, andquickly frozen at -60 C. Differential counts showed80 to 90% macrophages (10, 11). After at least a weekin the frozen state, the cell suspension was thawed,homogenized with a serological pipette, diluted to theappropriate concentration with 0.9% saline, and usedas a source of proteinase.

Rabbit pulmonary alveolar macrophages (AM)were obtained by perfusing fresh rabbit lungs fourtimes intratracheally with 30 to 40 ml of citratedsaline solution (11, 29). The differential countsshowed 98 to 100% AM (11). The AM were counted,frozen, stored, and used as a source of proteinase inthe manner just described.Assay of protease activity with acid hemoglobin

as the substrate. Macrophage and lung cathepsin Dactivity was assayed with hemoglobin as the substrateby a method (10, 12, 25) adapted from that of Anson(1). Bovine hemoglobin substrate powder (4.0 g)(Worthington Biochemical Corp., Freehold, N.J.) wasdissolved in 100 ml of 0.05 M HCl. The resulting 4.0%solution had a pH of 3.2. No buffer was added. Theenzyme and substrate solutions (each 1.3 ml) weremixed and incubated in a water bath at 38 C. At 0 hand 1 h, 1.0-ml samples were removed, mixed with 1.0ml of 5% trichloroacetic acid, and centrifuged. A 1-mlportion of supernatant fluid was diluted with 2.0 ml ofdistilled water, and the optical density (OD) was readat 280 nm in 1-cm quartz cuvettes. The difference inOD between the 0-h and 1-h samples was linearlyproportional to proteolytic activity up to an OD ofabout 0.350. A 0.100 unit of proteolytic activity wasdefined as that quantity of enzyme that caused adifference of OD of 0.100 under these conditions. Thespecific activity of the enzyme preparation was de-fined as the total units of activity per milligram ofprotein determined by the Lowry procedure (22).

Isoelectric focusing. A modification of the originaltechnique of Svensson (40) was used. A 110-mlcolumn (no. 8102, LKB-Produkter AB, S-161 25Bromma 1, Sweden) was prepared by adding 20 ml of2% monoethanolamine in 60% sucrose to the bottom,the cathode end. Next was added a 100-ml linearsucrose gradient, from 50% sucrose (containing 1%ampholytes, pH 5 to 8) to distilled water (containingthe partially purified enzyme in 1% ampholytes, pH 5to 8). Finally, 10 ml of 2.9% phosphoric acid solutionin water (i.e., 2.9% of the 85.5% commercial solution)was added on top, the anode end. The current wasapplied at 300 V for 24 h and then at 1,200 V for thenext 48 h. The column was cooled to 4 C by waterfrom a constant-temperature bath. The gradient wasremoved by pumping distilled water into the top ofthe column at a rate of about 1 mVmin. Fractions (6ml) were collected and analyzed for OD at 280 nm, forpH, and for cathepsin D actiyity.

Disk electrophoresis. A Buchler vertical gel elec-trophoresis apparatus (Polyanalyst, no. 3-1750,Buchler Instruments Div., Fort Lee, N.J.) was usedwith a modification of the method described byCampbell (9) and Gordon (16). The samples, 0.1 to 0.2ml, contained about 200 gg of protein in 40% sucroseand tris(hydroxymethyl)aminomethane (0.01 M)-gly-cine (0.08 M) buffer (pH 8.2). They were subjected toelectrophoresis on 7% acrylamide at 3 mA per tube for90 min. The gels were stained for 10 min with 0.25%Coomassie brilliant blue R (no. B 0630, Sigma Chemi-cal Co. St. Louis, Mo.) in water-methanol-acetic acid(5:5:1). The excess color was eluted with the samesolvent without dye.Immunization of goats with purified rabbit lung

cathepsin D. High-titer goat anti-rabbit cathepsin Dwas produced by Cappel Laboratories, Inc., Downing-town, Pa., in the following manner. Our most highlypurified cathepsin D preparations, namely the pH 6.6isoenzyme (from normal rabbit lungs [I] and fromgranulomatous rabbit lungs [II]), were shipped toCappel Laboratories in a frozen state. A 9-ml portionof preparation I, containing 0.22 mg of protein per mlof 0.9% NaCl, was emulsified with 9 ml of completeFreund adjuvant (Cappel lot 6031) and injectedsubcutaneously (s.c.) into a healthy goat in multiplesites. Every 2 weeks thereafter, for a 10-week period,2.5 ml of the same preparation was emulsified with 2.5ml of Freund incomplete adjuvant (Cappel Lot 6030)and injected s.c. The goat was bled weekly (150 ml)from 4 weeks on and exsanguinated at 12 weeks.A second healthy goat was similarly immunized

with preparation II containing 0.13 mg of protein perml. A 12-ml portion (instead of 18 ml) was injectedinitially, and three (instead of five) biweekly boosterinjections (5 ml each) were made. The goat was bledweekly from 3 weeks on and exsanguinated at 10weeks.

Samples of these antisera are still available fromCappel Laboratories, as we did not purchase theircomplete supply.Immunization of a rabbit with purified beef lung

cathepsin D. A 2-ml portion of our most purified beefcathepsin D preparation (containing 0.056 mg ofprotein per ml of 0.9% NaCl) was emulsified with 2 mlof complete Freund adjuvant (3113-60-5 Adjuvant,complete with Hs7Ra, Difco Laboratories, Detroit,Mich.) and injected into a 4-kg New Zealand rabbit in40 intradermal sites. Each week thereafter, 20 to 30 mlof blood was removed to determine the antibody titer.The rabbit was injected again, s.c. in several sites,

with antigen every 2 weeks. The first booster injectionwas similar in composition to the initial injection.Each subsequent booster injection consisted of 2.0 mlof the antigen (in saline solution) previously heatedfor 30 min at 65 C to allow its aggregation, and noadjuvants were employed. A good antibody responsewas observed after the third booster injection, and therabbit was exsanguinated 10 days after the fourthbooster injection.

Immunoelectrophoresis. A modification of themethod described by Grabar (17) was employed (seealso 9). Microscope slides were coated with 2.5 ml of1% Special Agar Noble (no. B-142, Difco) or 1%Agarose (no. A 6877, Sigma Chemical Co.) in 0.05 Mbarbital buffer, pH 8.2. Two small holes, made beside

1001

on May 15, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

ROJAS-ESPINOSA ET AL.

the central slot in the agar, were filled three timeswith the antigens at the concentrations described inthe two preceding sections. Electrophoresis was thencarried out in a Lucite cell (no. 3-1021, BuchlerInstruments Div.) filled with the same buffer at aconstant current of 7 mA per slide. Electrophoresiswas allowed to proceed until the tracking dye (0.1%bromophenol blue added after the antigen) reachedthe edge of the plate (about 90 min). The central slotswere then filled with antiserum to purified lungcathepsin D, and the resulting precipitin lines wereallowed to develop for 24 h at room temperature. Theslides were then dialyzed against 0.9% NaCl solution,stained with Coomassie brilliant blue R as in diskelectrophoresis, eluted in solvent, blotted with filterpaper, and then dehydrated further in an incubator at37 C for 2 to 6 h.

RESULTSPurification of the cathepsin D type pro-

teinase of rabbit lung. This acid-acting pro-teinase was purified by a modification of theBarrett method for the purification of rabbitliver cathepsin D (4, 5).Normal rabbit lungs (1,500 g) were homoge-

nized in a Waring blender for 5 min with 1%n-butanol in 1% NaCl (1200 ml) at 2 to 4 C. Theresultant suspension was diluted to 4,000 mlwith the saline-butanol and centrifuged at 7,000rpm (8,000 x g) for 30 min at 2 C. The pH of thesupernatant fluid (a fine suspension) was ad-justed to 3.6 with 5.0 M sodium formate buffer,pH 3.0, and the preparation was incubated at37 C overnight to permit autolysis to occur. Thesuspension was centrifuged at 7,000 rpm for 30min at 2 C, the precipitate was discarded, andthe clear brown supernatant fluid was diluted to4,000 ml with distilled water (DW).When larger amounts of lung were homoge-

nized as above, the protein in the clear brownsupernatant fluid was first concentrated byprecipitation with (NH.)2S04 at 90% satura-tion, dissolved in a minimum amount of DW,and dialyzed against 0.05 M sodium citratebuffer, pH 3.5, until the sulfate ions were absentwhen tested with BaCl2. Its volume was thenbrought to 4,000 ml with DW.The solution was cooled to 2 C in an ice bath.

Then, 3,600 ml of chilled (-10 C) anhydrousacetone (sodium sulfate dried) was added over aperiod of 10 min with continuous stirring. Thestirring was continued for an additional 1 to 2 hin the cold. The fine precipitate was removed bycentrifugation (7,000 rpm for 30 min at 2 C),and the supernatant fluid was treated as beforewith an additional 3,600 ml of chilled anhydrousacetone. The resulting suspension was left over-night at 2 to 4 C.

After centrifugation (7,000 rpm for 30 min at2 C), the precipitate was suspended in a mini-

mum volume of 0.002 M Tris-hydrochloridebuffer, pH 9.0, and dialyzed against four to sixchanges of the 0.002 M Tris-hydrochloridebuffer (each 12,000 ml).The enzyme preparation was then applied to

a column of DEAE-cellulose equilibrated with0.0144 M Tris-hydrochloride buffer (pH 7.8), ina ratio of two volumes of enzyme to 1 volume ofpacked diethylaminoethyl (DEAE)-cellulosebed. The column was then washed with tenvolumes of the buffer, and the enzyme waseluted with ten volumes of a linear gradient of 0to 0.2 M NaCl in the Tris-hydrochloride buffer,pH 7.8. Fractions of about 13.5 ml were col-lected and tested for enzymatic activity. Frac-tions with high specific activity were pooled anddialyzed for 36 h against three or four changes of0.03 M sodium acetate buffer, pH 4.75, (Fig. 1).The enzyme preparation was then applied to

a CM-Sephadex (C-50) column equilibratedwith 0.03 M acetate buffer, pH 4.75, in the ratioof three volumes of enzyme to one volume ofpacked CM-Sephadex. The enzyme was theneluted with pH 5.5 sodium acetate buffer (r/20.2). Fractions of 12 ml were collected andtested for protein and enzyme activity (Fig. 2).The fractions with high specific activity werepooled and dialyzed against 1% glycine in water.In subsequent experiments, dialysis was madeagainst 0.3% glycine.The dialyzed enzyme was placed in an isoe-

lectric focusing column containing a sucrosegradient with ampholytes, pH 5 to 8, (seeMaterials and Methods). The fractions fromthis column were analyzed, and the peaks ofproteinase activity were pooled. There were four

0-

08

.106-

004-

02-

00

TOTAL PROTEIN

20 30 40TUBE NUMBER (13 5,mI/tube)

FIG. 1. Distribution of rabbit lung cathepsin D infractions eluted from a DEAE-cellulose (DE-52) col-umn by a linear gradient of 0 to 0.2 M NaCI inTris-hydrochloride buffer. The proteinase activity ofcathepsin D (represented by the OD increment at 280nm) was determined with hemoglobin as the substrateat pH 3.2. The total protein present was consideredproportional to the OD of the fractions at 280 nm. Theshaded area indicates the fractions pooled for theCM-Sephadex column.

INFECT. IMMUNITY1002

on May 15, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

VOL. 8, 1973 CATHEPSIN D TYPE MACROPHAGE PROTEINASE FROM LUNG

active fractions at pH 5.8, 6.2, 6.6 and 6.8,respectively (Fig. 3), which probably representdifferent isoenzymes (5). The pH 6.6 isoenzymewas used to immunize the goat.Compared with the original saline homoge-

nate, there was approximately 150-fold purifica-tion of rabbit lung cathepsin D with 0.7 to 3%yield (Table 1). Figure 4 shows the results ofelectrophoresis on acrylamide gel during thesevarious stages of purification. Note that thesample used to immunize a goat showed one

main band on gel electrophoresis and one very

minor band, which was, however, not detectedby immunoelectrophoresis. The latter is proba-bly the tail of the pH 6.8 isoenzyme, which didnot fully separate from the pH 6.6 isoenzyme.The broad band near the positive end of the gelsis probably an isoenzyme plus considerableinactive protein.The purification procedure just outlined was

repeated by using 25 rabbit lungs (1,500 g)made granulomatous by the intravenous injec-tion of lyophilized heat-killed BCG in oil. Table1 summarizes the results. The specific activity

20 30 40 50

TUBE NUMBER (12ml/tube)

FIG. 2. Distribution of rabbit lung cathepsin D infractions eluted from a CM-Sephadex (C-50) columnby acetate buffer. The shaded area indicates thefractions pooled for isoelectric focusing. (See Fig. 1 forother details.)

< -52'-

0- 5 P. 66

o 5-,.-68 -A

\ -

3C 35 40 45 50 55 60 65 70 75

TUBE NUMBER (6-l/N

FIG. 3. Distribution of rabbit lung cathepsin Dafter isoelectric focusing in a sucrose gradient contain-ing pH 5 to 8 ampholytes. The shaded area indicatesthe fractions pooled for immunizing a goat.

TABLE 1. Purification of cathepsin D from normaland granulomatous rabbit lungs

Specificactivity Purifica- Recovery

Stepa (units/mg tion ( 'Xo)of (fold)

protein)'

Normal lungI Initial homogenate 0.17 1 100

II Autolysis 0.18 1 40III Acetone 2.1 12 11IV DEAE Cellulose 4.5 26 4V CM Sephadex 10.8 64 1VI Isoelectric focusing 27.0' 160' 0.7"

Granulomatous lungI Initial homogenate 0.46 1 100

II Autolysis 0.33 1 41III Acetone 4.7 10 13IV DEAE Cellulose 12.0 26 8V CMSephadex 37.0 80 5VI Isoelectric focusing 66.0" 144c 3"

a See text.I A unit of proteolytic activity is the amount that should

produce (if linearity held) an OD difference of 1.000 underconditions described in Materials and Methods.

c Value for the isoenzyme used to immunize a goat.Sum of all four isoenzymes.

Robbit Lung Cothepsin D

11 IV V VI

Beef Lung Cothepsin D

IV VFIG. 4. Disk electrophoresis on polyacrylamide gel

at pH 8.2, performed at several stages during thepurification of rabbit lung cathepsin D and at thefinal stages of the purification of beef lung cathepsinD. These sketches were traced from the gel cylindersstained with Coomassie Blue. The isoenzyme used toimmunize the goat and rabbit, respectively, is repre-sented in the last sketch of each series. The numbersindicate the purification steps listed in Tables 1 and2.

of the original homogenate of granulomatouslungs was about twice that of normal lungs. Thepurification steps were similar, and the end

1003

i

C

-iEz

on May 15, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

ROJAS-ESPINOSA ET AL.

product apparently was identical. The pH 6.6isoenzyme was also used to immunize a goat.Purification of the cathepsin D type pro-

teinase of beef lung. This acid-acting protein-ase was purified by continuing the methoddescribed by Dannenberg and Smith (12)through additional steps. Fresh beef lung (2,000g) was blended at 4 C in a Waring blender with4,000 ml of cold (-10 C) 50% acetone in water.(Note that a typographical error was present inthe original directions (12): 50%, not 100%,acetone must be used). The blended mixturewas centrifuged, and the supernatant fluid wasfiltered through gauze. Two volumes of cold100% acetone (-20 C) were added, and theresulting suspension was left overnight in thecold room (2 to 4 C). The sediment was mixedwith three or four volumes of cold 100% acetoneand centrifuged. The precipitate, mixed withthree times its volume of 100% acetone, was

centrifuged again. The sediment was collectedas a paste and stored in a desiccator at - 80 C. Itdid not appear necessary to form an acetonedried powder as previously described (12).About 40 g of the paste was mixed with 1,000

ml of DW and centrifuged. Fe(NH,)2(SO,)2 wasadded to the supernatant fluid to a final concen-tration of 0.004 M, and the pH was adjusted to3.0 with 0.1 M HCl. After the addition of citratebuffer (1.0 M, pH 3.0) (1 ml for each 24 ml), thesuspension was allowed to stand for 3 to 7 daysin the cold (2 to 4 C).The precipitate was discarded, and the super-

natant fluid was further purified by a 0.45 to0.80 saturated ammonium sulfate cut. Afterdialysis against 0.004 M Fe(NH,)2(SO,)2, theproteinase was concentrated fourfold in a lyo-philizer.The partially purified cathepsin was then

chromatographed on CM-Sephadex (C-50) col-umns containing 0.05 M sodium phosphatebuffer, pH 7.4, and eluted with the same buffer.The most active fractions leaving the columnswere pooled and placed in the isoelectric focus-ing apparatus. Similar to the rabbit cathepsin,four active isoenzymes were found (Fig. 5).They had isoelectric points at pH 5.75, 6.4, 6.6,and 7.0. The pH 7.0 isoenzyme was used toimmunize a rabbit.Compared with a saline homogenate of whole

beef lung, there was a 93-fold purification ofbeef lung cathepsin D with a 24% yield (Table2). The sample used to immunize the rabbit hadonly one band in acrylamide gel electrophoresis(Fig. 4).Immunoelectrophoresis. The purified rabbit

and beef lung cathepsin D (used to immunizethe goat and rabbit, respectively) showed a

single precipitin line in immunoelectrophoresis

with their respective antisera (Fig. 6), indicat-ing a high degree of purity. The purified cathep-sin D from both normal and granulomatousrabbit lungs showed an identical single precipi-tin line with antibody against the normal rabbitlung cathepsin D. The antibody preparedagainst rabbit lung cathepsin D showed no linewith purified beef lung cathepsin D. Con-versely, the antibody against beef lung cathep-sin D showed no line with purified rabbit lungcathepsin D. Similar results were observed bythe Barrett group (5) when antibodies to puri-fied chicken, human, and rabbit liver cathepsinD were compared. Sheep antibody against puri-fied rabbit liver cathepsin D, supplied by Bar-rett (5, 15), showed one precipitin line with ourpurified rabbit lung cathepsin D, but no linewith our purified beef preparation. Thus, anti-bodies to these purified proteinases are specificfor the species of origin, but nonspecific for theorgan of origin.The undiluted antisera against the pH 6.6

TUBE NUMBER (6ml/tube)

FIG. 5. Distribution of beef lung cathepsin D afterisoelectric focusing in a sucrose gradient containingpH 5 to 8 ampholytes. The shaded area indicates thefractions pooled for immunizing a goat.

TABLE 2. Purification of cathepsin D from beef lung

Specific

Stepa activity Purification RecoveryStep~~ (units/mg (fold) (%of

protein)"

(Aqueous extract)c 0.18 (1.0)Y (100Y)I Acetonepaste 0.14 1.0 (0.78) 100(30)

II Autolysis 0.60 4.2 (3.3)III Ammonium sulfate 2.0 14.0 (11.0) 73 (22)IV CM Sephadex 3.4 24.0 (19.0) 35 (11)V Isoelectric focusing 13.0 93.0 (72.0) 24_ (7)

a See text.B See Table 1.cThe average of results from four lungs (see Table 4). Note

that the aqueous extract was not a part of the purificationprocedure, as the lungs were homogenized directly in acetone-water. The amount of purification and recovery with respectto the aqueous extract is in parentheses.

1004 INFECT. IMMUNITY

0D

z

c)

Q.

on May 15, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

VOL. 8, 1973 CATHEPSIN D TYPE MACROPHAGE PROTEINASE FROM LUNG

Rabbit Cathepsin D

Antirobbit ( )

Robbit Coahepsin D(granulomotous)

Robbit Cothepsin D

Antirobbit D (

Beef Cothepsin D

Beef Cothepsin CD

The antiserum against rabbit cathepsin Donly partially inhibited the proteolytic activityof rabbit AM and MN when the enzyme-anti-body precipitate was not removed: a 1:2 dilu-tion of the antiserum inhibited 88% of the AMand MN activity, and a 1:4 dilution of theantiserum inhibited 50% of their activity. Evi-dently at pH 3.2, the antiserum only partiallyblocks the enzyme's active site or readily dis-sociates from it.

Characterization of purified rabbit andbeef lung cathepsin D. These studies em-

ployed rabbit and beef cathepsin D purifiedthrough the CM-Sephadex step. The rabbitproteinase had a pH optimum of 3.0, and thebeef proteinase had an optimum of 3.6 (Fig. 7).

Antibeef D .j~

Rabbit Cothepsin D

FIG. 6. Immunoelectrophoresis of purified rabbitand beef lung cathepsin D. Depicted are the precipi-tin arcs from the reaction ofgoat and rabbit antiserumwith the pH 6.6 rabbit isoenzyme and the pH 7.0 beefisoenzyme, respectively. Slots were filled with anti-body against purified rabbit lung cathepsin D or puri-fied beef lung cathepsin D.

rabbit isoenzyme and the pH 7.0 beef isoenzymeprecipitated in capillary tubes the four isoen-zymes of their own species. They did not precip-itate the isoenzymes of the other species. Theseresults were confirmed in immunoelectrophore-sis with the four rabbit isoenzymes and three ofthe beef isoenzymes. (Our supply of the pH 5.8beef isoenzyme had been exhausted.)

Specificity of antiserum. Goat antiserumagainst purified rabbit lung cathepsin D (0.1ml) was incubated with the purified rabbitenzyme (0.1 ml) in 0.9% NaCl in small testtubes (10 by 50 mm) for 30 min at 37 C, chilledin an ice bath for 10 min, and centrifuged in thecold to remove the precipitate. The supernatantfluids (0.1 ml) were assayed for proteinaseactivity at pH 3.2 with hemoglobin as thesubstrate. Similar procedures were performedwith homogenates of rabbit pulmonary AM andrabbit oil-induced MN. Also, rabbit antiserumagainst purified beef lung cathepsin D was

tested against the purified beef enzyme. Theresults are listed in Table 3. All of the protein-ase activity of these preparations was removedby undiluted goat antiserum against rabbitcathepsin D (listed as 1: 2 final concentration).More dilute antiserum was less effective. About75% of the proteinase activity of purified beeflung cathepsin D was removed by its specificantiserum.

TABLE 3. Effect of antibody on proteinase activity ofpurified rabbit and beef lung cathepsin D and rabbit

peritoneal and pulmonary macrophagesa

Dilution ofantibody (aftermixing with

Source of proteinase enzyme) andpercentage of

original proteo-lytic activity

Rabbit lung cathepsin (0.30 units)' 1:2 0%1:4 101:8 541:16 761:32 100

Rabbit pulmonary macrophages 1:2 0%(0.6 x 10" cells 0.30 units)c 1:4 4

1:8 341:16 441:32 100

Beef lung cathepsin (0.30 units) 1:2 26%1:4 651:8 821:16 911:32 100

Rabbit peritoneal macrophages 1:2 1%(5 x 105 cells 0.36 units)c 1:4 24

1:8 401:16 581:32 93

a Antibody was mixed with its respective enzymepreparation for 15 min at 37 C and centrifuged. Theproteolytic activity remaining in the supernatantfluid is listed (see Text). Antirabbit cathepsin Dantibody (produced in goats) was mixed with therabbit enzyme preparations; antibeef cathepsin Dantibody (produced in a rabbit) was mixed with thebeef enzyme preparation.

b See Table 1.cThe results listed represent an average of four

preparations.

1005

on May 15, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

ROJAS-ESPINOSA ET AL.

a)(, 0 30

w020-RABBIT LUNG BEEF LUNGPROTEINASE PROTEINASE

8010-

2 3 4 5 6 7

pH

FIG. 7. Hydrolysis of hemoglobin by rabbit andbeef lung cathepsin D as a function of pH. The pHoptimum for each enzyme was 3.0 and 3.6, respec-

tively. Both enzymes had been purified throughCM-Sephadex columns, but had not been elec-trofocused.

The proteolytic activity of both enzymes was a

linear function of enzyme concentration up toan OD increment of about 0.350. Then therewas a gradual decrease in slope.

Various compounds were tested for the abilityto inhibit or activate these cathepsins (see 25).At twice the final concentration, they were

preincubated with the enzyme (0.3 units) for 15min at 37 C before the hemoglobin substratewas added. With both proteinases, cysteine hadno effect, nor did any of the f'ollowing com-

pounds: iodoacetamide, p-chloromercuribenzo-ate, disodium ethylenediaminetetraacetate, Fe(NH4)2(SO4)2, Mg (CH3COO)2, and MnCl2.The concentrations of each were 0.0005 and0.005 M after the addition of substrate.Pb(N03)2 and HgCl2 (both 0.005 M) inhibitedthe rabbit enzyme about 15%. The HgCl2 inhib-ited the beef enzyme about 20%.Sodium Pepstatin, a highly specific cathepsin

D inhibitor (2, 3, 6, 18, 20, 25, 28, 39, 41, 46),inhibited the beef and rabbit lung proteinases100% at 0.05 qM concentration and about 50%at 0.005 ,uM (25). The beef enzyme (specificactivity of 1.43 units per mg of protein) had beenpurified through the CM-Sephadex step. Therabbit enzyme was more crude (specific activity0.54). A CM-Sephadex purified rabbit protein-ase (specific activity 10.8) was also tested andfound to be inhibited by one-tenth of the abovePepstatin concentrations. The proteolytic activ-ity of rabbit pulmonary and peritoneal macro-

phages was inhibited 100% by 0.05 gM Pep-statin and about 50% by 0.005 ,uM Pepstatin(25).Acid proteinase activity of rabbit and beef

organs. Beef and rabbit liver, spleen, lung, andkidney were assayed for proteinase activity at

pH 3.2. Four animals of each species wereevaluated. A 1-g portion of each organ washomogenized with five times its weight of' DW ina chilled Potter-Elvehjem homogenizer. Thecells were further disrupted by three cycles offreezing and thawing in a dry ice-acetone bathand a 37 C water bath, respectively. The sus-pensions were then centrifuged at 4 C for 30 minat 3,000 rpm (2225 x g). The supernatant fluidswere diluted 10-fold with 0.05 M sodium citratebuffer, pH 3.2, and their proteinase activity wasdetermined. The method of Lowry was used todetermine their protein concentration.Table 4 shows the results. The extractable

proteinase per gram of' tissue was highest forspleen, as was the specific activity. Rabbit lungextracts had the next highest specific activitv.Extracts of rabbit liver and kidney showed lessspecific activity than extracts of beef liver andkidney. At present, we have no truly satisfac-tory explanations of these findings.

DISCUSSIONCathepsin D is a major lysosomal enzyme of

many cells: macrophages, lymphocytes (8), fi-broblasts, epithelial and synovial cells, chron-drocytes, and liver parenchymal cells (see 32).Cells, like macrophages, that specialize indigestive functions are particularly rich in thisacid-acting proteinase (10, 31, 32; 0. Rojas-Espinosa, A. M. Dannenberg, Jr., L. Stern-berger, and T. Tsuda, Amer. J. Pathol., inpress). A brief discussion of how this proteinasehydrolyzes its substrates appears in our reviewof inhibitors of pepsin and cathepsin D (25).Woessner (unpublished data) tested our par.

tially purified beef lung cathepsin D (specificactivity 2.6) on the hexapeptide Glu-Ala-Leu-Tyr-Leu-Val and found that it split the Leu-Tyrand Tyr-Leu bonds. With Phe-Phe-NH2, thesplit occurred at the Phe-Phe bond and withPhe-Tyr-NH2 at the Phe-Tyr bond (unpub-

TABLE 4. Acid proteinase activity of rabbit and beeforgansa

Rabbit Beef

Extractable Specific Extractable SpecificOrgan proteinase activity of proteinase activity of

per gm of extract per gm of extracttissue (units/mg tissue (units/mg(units)b of protein) (units) of protein)

Lung 11.7 ± 3.3 0.20 ± 0.07 5.3 ± 0.9 0.18 0.03Liver 10.1 ± 2.1 0.08 ± 0.01 12.7 3.2 0.17- 0.02Spleen 16.4 1.2 0.38 ± 0.04 16.8 ± 4.1 0.28 0.08Kidney 5.1 + 0.2 0.07 ± 0.01 11.9 + 0.8 0.19 0.10

The means (and their standard errors) of results from atleast four animals are listed.

h See Table 1.

INFECT . IMMUNITY1006

on May 15, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

VOL.8, 1973 CATHEPSIN D TYPE MACROPHAGE PROTEINASE FROM LUNG

lished data). Cleavage of the hexapeptide andPhe-Tyr-NH2 was blocked by 2 Ag of Pepstatin.

Partially purified preparations of beef lungcathepsin D polymerized aromatic amino acidesters (13). Purified beef lung cathepsin D didnot (0. Rojas-Espinosa, P. Arce-Paredes, A. M.Dannenberg, Jr., and R. L. Kamenetz, manu-script in preparation). In fact, the polymerizingactivity of the impure preparations was found tobe associated with the chymotrypsin-like pro-teinase that hydrolyzed N-benzoyl-D, L-phenylalanine-B-naphthol ester (10, 12, 0. Ro-jas-Espinosa et al., manuscript in preparation).Purification and characterization of this newenzyme (a monoaminopeptidase) will be thesubject of another report (0. Rojas-Espinosa etal., manuscript in preparation).The lung cathepsin D described herein proba-

bly comes from the lung macrophages for thefollowing reasons: (i) the lung is rich in macro-phages; (ii) labeled-antibody and substrate filmtechniques showed that lung macrophages con-tain more cathepsin D than any other cellpresent (0. Rojas-Espinosa et al., Amer. J.Pathol., in press, cf. 32); (iii) granulomatouslungs contained more cathepsin D and moremacrophages (Table 1); (iv) the pH optima oflung proteinase and macrophage proteinase areidentical (10, 12); (v) Pepstatin, a highly spe-cific cathepsin D inhibitor (2, 3, 6, 18, 20, 25, 28,39, 41, 46), readily inhibits the proteinase fromboth sources (25); and (vi) the antibody topurified lung cathepsin D also inhibits macro-phage cathepsin D (Table 3).The cathepsin D present in macrophages

throughout the body undoubtedly has manybiological functions. It aids these cells in digest-ing the bacteria, fungi, and viruses that theyingest. It also aids them in digesting otherproteins, e.g., the hemoglobin from effete eryth-rocytes and many antigens. In the latter case,the antigenic dose is thereby regulated to avoidan excess that might induce tolerance.

Split-protein products released from macro-phages probably contribute to the vasodilationand exudation of the inflammatory response(26, 38, 45). C5a can be produced from the CScomponent of complement by lung and macro-phage cathepsin D (36). This fragment is anextremely active chemotactic factor for bothPMN and MN (35, 36, 37, 42). Cathepsin Dshould therefore increase the number of cells(and their proteases) at the site of chronicinflammation. Because it can also digest carti-lage in vitro (14, 15, 44, 46), it may play a role inthe development of rheumatoid arthritis. Mac-rophage cathepsin D may also be a mediator ofthe fibrotic response, because macrophage infil-

tration frequently precedes fibroblast infiltra-tion.

Studies in our laboratory (T. Tsuda, A. M.Dannenberg, Jr., M. Ando, and 0. Rojas-Espinosa, manuscript in preparation) and Ya-mamura's (25, cf. 48) indicate that macrophagecathepsin D is involved in the liquefaction ofthe caseous focus in tuberculosis. Liquefactionresults in tremendous extracellular multiplica-tion of tubercle bacilli, with frequent develop-ment of antibiotic-resistant mutants and spreadof the disease throughout the bronchial tree andto other people (23, 24, 34). It is theoreticallypossible that the use of effective protease (andother) inhibitors might arrest the process ofliquefaction and aid in the control of tuberculo-SiS.

ACKNOWLEDGMENTS

This investigation was supported by Public Health Servicegrants AI-08876 and AI-03772 from the National Institute ofAllergy and Infectious Diseases under the auspices of theUnited States-Japan Cooperative Medical Science Program,by grant HE-14153 from the National Heart and LungInstitute, U.S. Public Health Service, for The Johns HopkinsSpecialized Research Center on Lung, by contract DADA17-72-C-2187 with the U.S. Army Medical Research andDevelopment Command, and by grant GB 8058 from theNational Science Foundation.

Dr. 0. Rojas-Espinosa is on leave of absence from theDepartamento de Immunologia, Escuela Nacional deCiencias Biol6gicas del Instituto Politecnico Nacional, Mex-ico 17, D.F., Mexico, and was partially supported during thisstudy by a fellowship from the C.O.F.A.A.-S.E.D.I.C.T.,Mexico.

The technical assistance of Donna K. Class and Delores R.Fleming in parts of this work is gratefully acknowledged.

LITERATURE CITED1. Anson, M. L. 1939. The estimation of pepsin, trypsin,

papain, and cathepsin with hemoglobin. J. Gen. Phys-iol. 22:79-89.

2. Aoyagi, T., S. Kunimoto, H. Morishima, T. Takeuchi,and H. Umezawa. 1971. Effect of Pepstatin on acidproteases. J. Antibiot. (Tokyo) 24:687-694.

3. Aoyagi, T., H. Morishima, R. Nishizawa, S. Kunimoto,T. Takeuchi, H. Umezawa, and H. Ikezawa. 1972.Biological activity of pepstatins, pepstatone A andpartial peptides on pepsin, cathepsin D and renin. J.Antibiot. 25:689-694.

4. Barrett, A. J. 1967. Lysosomal acid proteinase of rabbitliver. Biochem. J. 104:601-608.

5. Barrett, A. J. 1971. Purification and properties of cathep-sin D from liver of chicken, rabbit and man, p. 109-133.In A. J. Barrett and J. T. Dingle (ed.), Tissue Pro-teinases. North Holland Publishing Co., Amsterdam.

6. Barrett, A. J., and J. T. Dingle. 1972. The inhibition oftissue acid proteinases by Pepstatin. Biochem. J.127:439-441.

7. Bohley, P., H. Kirschke, J. Langner, S. Ansorge, B.Wiederanders, and H. Hanson. 1971. Intracellularprotein breakdown, p. 187-219. In A. J. Barrett and J.T. Dingle (ed.), Tissue Proteinases. North HollandPublishing Co. Amsterdam.

8. Bowers, W. E. 1972. Lysosomes in rat thoracic ductlymphocytes. J. Exp. Med. 136:1394-1403.

9. Campbell, D. H., J. S. Garvey, N. E. Cremer, and D. H.

1007

on May 15, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

1008 ROJAS-ESPINOSA ET AL.

Sussdorf. 1970. Methods in immunology, 2nd ed., p.94-104, 260-267. W. A. Benjamin, Inc., New York.

10. Dannenberg, A. M., Jr., and W. E. Bennett. 1964.Hydrolytic enzymes of rabbit mononuclear exudatecells. I. Quantitative assay and properties of certainproteases, non-specific esterases, and lipases of mono-nuclear and polymorphonuclear cells and erythrocytes.J. Cell Biol. 21:1-13.

11. Dannenberg, A. M., Jr., M. S. Burstone, P. C. Walter,and J. W. Kinsley. 1963. A histochemical study ofphagocytic and enzymatic functions of rabbit mononu-clear and polymorphonuclear exudate cells and alveo-lar macrophages. I. Survey and quantitation of en-zymes and states of activation. J. Cell Biol. 17:465-486.

12. Dannenberg, A. M., Jr., and E. L. Smith. 1955. Proteo-lytic enzymes of lung. J. Biol. Chem. 215:45-54.

13. Dannenberg, A. M., Jr., and E. L. Smith. 1955. Action ofProteinase I of bovine lung. Hydrolysis of the oxidizedB chain of insulin; polymer formation from amino acidesters. J. Biol. Chem. 215:55-66.

14. Dingle, J. T., A. J. Barrett, A. R. Poole, and P. Stoivin.1972. Inhibition by Pepstatin of human cartilage degra-dation. Biochem. J. 127:443-444.

15. Dingle, J. T., A. J. Barrett, P. D. Weston. 1971. Cathep-sin D. Characteristics of immunoinhibition and theconfirmation of a role in cartilage breakdown. Bio-chem. J. 123:1-13.

16. Gordon, A. H. 1969. Electrophoresis of proteins in polya-crylamide and starch gels, p. 34-72. North HollandPublishing Co., Amsterdam.

17. Grabar, P. 1964. Immunoelectrophoretic analysis, p.79-91. In J. F. Ackroyd (ed.), Immunological methods.F. A. Davis Co., Philadelphia.

18. Ikezawa, H., T. Aoyagi, T. Takeuchi, and H. Umezawa.1971. Effect of protease inhibitors of actinomycetes onlysosomal peptide hydrolases from swine liver. J. Anti-biot. 24:488-490.

19. Keilova, H. 1971. On the specificity and inhibition ofcathepsins D and B, p. 45-68. In A. J. Barrett and J. T.Dingle (ed.), Tissue Proteinases. North Holland Pub-lishing Co., Amsterdam.

20. Kunimoto, S., T. Aoyagi, H. Morishima, T. Takeuchi,and H. Umezawa. 1972. Mechanisms of inhibition ofpepsin by pepstatin. J. Antibiot. 25:251-255.

21. Lapresle, C. 1971. Rabbit cathepsins D and E, p. 135-155.In A. J. Barrett and J. T. Dingle (ed.), Tissue Pro-teinases. North Holland Publishing Co., Amsterdam.

22. Lowry, H. O., N. J. Rosebrough, A. L. Farr, and R. J.Randall. 1965. Protein measurement with the Folin-phenol reagent. J. Biol. Chem. 193:265-275.

23. Lurie, M. B. 1941. Heredity, constitution and tuberculo-sis. An experimental study. Amer. Rev. Tuberc. 44(Suppl. 3):1-125.

24. Lurie, M. B., and A. M. Dannenberg, Jr. 1965. Macro-phage function in infectious disease with inbred rab-bits. Bacteriol. Rev. 29:466-476.

25. McAdoo, M. H., A. M. Dannenberg, Jr., C. J. Hayes, S.P. James, and J. H. Sanner. 1973. Inhibition of thecathepsin D-type proteinase of macrophages by Pep-statin, a specific pepsin inhibitor, and other substances.Infect. Immunity 7:655-665.

26. Menkin, V. 1950. Newer concepts of inflammation.Charles C. Thomas, Publisher. Springfield, Ill.

27. Meyer, 0. T., and A. M. Dannenberg, Jr. 1970. Radia-tion, infection and macrophage function. II. The effectof whole body radiation on the number of pulmonaryalveolar macrophages and their levels of hydrolyticenzymes. J. Reticuloendothel. Soc. 7:79-90.

28. Morishima, H., T. Takita, T. Aoyagi, T. Takeuchi, and

INFECT. IMMUNITY

H. Umezawa. 1970. The structure of Pepstatin. J.Antibiot. 23:263-265.

29. Myrvik, Q. N., E. Soto-Leake, and B. Fariss. 1961.Studies on pulmonary alveolar macrophages from thenormal rabbit: a technique to procure them in a highstate of purity. J. Immunol. 86:128-132.

30. Myrvik, Q. N., E. Soto-Leake, and S. Oshima. 1962. Astudy of macrophages and epithelioid-like cells fromgranulomatous (BCG-induced) lungs of rabbits. J.Immunol. 89:745-751.

31. Opie, E. L. 1906. The enzymes in phagocytic cells ofinflammatory exudates. J. Exp. Med. 8:410-436.

32. Poole, A. R., J. T. Dingle, and A. J. Barrett. 1972. Theimmunocytochemical demonstration of cathepsin D. J.Histochem. Cytochem. 20:261-265.

33. Press, E. M., R. R. Porter, and J. Cebra. 1960. Theisolation and properties of a proteolytic enzyme, cath-epsin D, from bovine spleen. Biochem. J. 74:501-514.

34. Rich, A. R. 1951. The pathogenesis of tuberculosis, 2nded. Charles C. Thomas, Publisher, Springfield, Ill.

35. Shin, H. S., R. Snyderman, E. Friedman, A. Mellors, andM. M. Mayer. 1968. Chemotactic and anaphylatoxicfragment cleaved from the fifth component of guineapig complement. Science 162:361-363.

36. Snyderman, R., H. S. Shin, and A. M. Dannenberg, Jr.1972. Macrophage proteinase and inflammation: theproduction of chemotactic activity from the fifth com-ponent of complement by macrophage proteinase. J.Immunol. 109:896-898.

37. Snyderman, R., H. S. Shin, and M. H. Hausman. 1971. Achemotactic factor for mononuclear leukocytes. Proc.Soc. Exp. Biol. 138:387-390.

38. Spector, W. G., and D. A. Willoughby. 1968. Thepharmacology of inflammation, p. 55. Grune andStratton, Inc., New York.

39. Suda, H., T. Aoyagi, M. Hamada, T. Takeuchi, and H.Umezawa. 1972. Antipain, a new protease inhibitorisolated from actinomycetes. J. Antibiot. 25:263-265.

40. Svensson, H. 1962. Isoelectric fractionation, analysis, andcharacterization of ampholytes in natural pH gradi-ents. III. Description of apparatus for electrolysis incolumns stabilized by density gradients and directdetermination of isoelectric points. Arch. Biochem.Biophys. Suppl. 1:132-138.

41. Umezawa, H., T. Aoyagi, H. Morishima, M. Matsuzaki,M. Hamada, and T. Takeuchi. 1970. Pepstatin, a newpepsin inhibitor produced by actinomycetes. J. Anti-biot. 23:259-262.

42. Ward, P. A., and L. J. Newman. 1969. A neutrophilchemotactic factor from human C'5. J. Immunol.102:93-99.

43. Weston, P. D. 1969. A specific antiserum to lysosomalcathepsin D. Immunology 17:421-428.

44. Weston, P. D., A. J. Barrett, and J. T. Dingle. 1969.Specific inhibition of cartilage breakdown. Nature(London) 222:285-286.

45. Wilhelm, D. L. 1965. Chemical mediators. I. p. 389-425. InB. W. Zweifach, L. Grant, and R. T. McCluskey (ed.),The inflammatory process. Academic Press Inc., NewYork.

46. Woessner, J. F., Jr. 1972. Pepstatin inhibits the digestionof hemoglobin and protein-polysaccharide complex bycathepsin D. Biochem. Biophys. Res. Commun.47:965-970.

47. Woessner, J. F., Jr., and R. J. Shamberger. 1971. Purifi-cation and properties of cathepsin D from bovineuterus. J. Biol. Chem. 246:1951-1960.

48. Yamamura, Y. 1958. The pathogenesis of tuberculouscavities. Advan. Tuber. Res. 9:13-37.

on May 15, 2020 by guest

http://iai.asm.org/

Dow

nloaded from