AJEBAK 59 (Pt. 1) 63-75 (1981)

©SOME PROPERTIES OF NEUTRAL PROTEINASES FROMLYSOSOMES OF RABBIT POLYMORPHONUCLEAR

LEUCOCYTES

by M. L. BRITZ* AND D . A. LOWTHER

(From the Department of Biochcmislry, Monash University, Clayton, Vic, 3168,Australia.)

(Accepted for puhlicaiion November 3, 1980.)

Summary. Neutriil prolsinases capLibIc of degrading proteoglyciin were found inlysosomes of rabbit polymorpbonuclear leucocytes exlriictecl wilb 001 M citricacid. Eslerase uclivily againsi an elast;ise substrate was also present bui chymo-trypsin- and irypsin-like activities were not delected: azocasein-degrading aclivilywas poor. Proteoglycanase activity was stimulated by high concentrations of salts(0.2 M KCI) and divalent cations (Ca, Mg. Mn, Zn) but was inhibited by Cu + +.Elastase activity was also stimulated by high ionic strength buffers and KCI. butnot as much by divalent cations, and was inhibited by Cu+-i-. Proleoglycanasein crude extracts w.is inhibited by EDTA. pbenylmetbanesiilpbonyillnoride(Pms-F). cell cytosol. "i-antitrypsin. gold tbiomalate and N-acetyl-di-l.-alanyl-L-prolyl-L'Valine chloromcthyl ketone (AAAPVCK). Partial inhibition by N-a-p-tosyl-L-Iysine cblorometbyl ketone (TLCK) and L-l-tosylamide-2-phenylethyl chloro-methyl ketone (TPCK) occurred. Elastase adsorbed to CM-cellulose and waseluted by 06-0-7 .M NaCi: a metallo-proteinase failed to adsorb completely butwas retarded by tbe CM-cellulose. (soelectric focusing showed that the majorproteinases had pi's of 5-5, 8-5 and y I; the activity witb pi 8 5 was a melallo-protemase. and the pi 9-1 activity was an elastase. The apparent molecular weightof the elastase, determined on Sephadex CM00. was 8.000 to I ].0(K) daltons.

INTRODUCTIONThere have been relatively few reports describing the properties of

neutral proteinases of rabbit polymorphonuclear leucocytes (PMNL) (re-view by Starkey, 1977). One of the first was that of Davies, Krakauerand Weissmann (1970), who showed that neutral proteinase activity waslocalized exclusively in the azurophil granules. A neutral histonase occurringas the major proteinase here was partially purified by Davies et al. (1971)by Sephadex G-75 chromatography and isoelectric focusing fpl 42-5-2) .Histonase activity in crude extracts was partially inhibited by the chymo-trypsin inhibitor L-l-tosylamide-2-phenylethyl chloromethyl ketone and bysoy bean trypsin inhibitor. Dewald et al. (1975) confirmed that casein- andhistone-degrading proteina.ses occurred in the azurophil granules, and further

• Present address: School of Microbiology, University of Melbourne, Parkville, Vic. 3052,Australia.

64 M. L. BRITZ AND D. A. LOWTHER

demonstrated esterase activity against the elastase substrate, t-butyloxycar-bonyl alanine 4-nitrophenyl ester. The elastase activity followed the samedistribution as the casein- and histone-degrading proteinases within the cell.Previous reports had indicated that rabbit leucocytes contain esterase activityagainst the chymotrypsin substrate N-acetyl-DL-phenylalanine 2-naphthylester (Starkey, 1977) but Dewald et al (1975) found no correlationbetween the distribution of this esterase activity and the proteolytic activity,thus failing to demonstrate proteinase activity corresponding to eathepsin Gof human leucocytes. Since these reports, there have been no further publi-cations that have led to an understanding of the types of proteinases inrabbit PMNL, so that available data reveal no clear relationship of theproteinases described above to the elastase and eathepsin G of human leuco-cytes. Defining such a relationship is of some interest, considering the broaduse of rabbits in experimental models of inflammation and pathologicalprocesses involving ti.ssue destruction, where PMNL proteinases potentiallyhave an important role. This communication attempts, in part, to clarify someof the properties of rabbit PMNL elastase, describing the partial purificationof this enzyme and a lysosomal neutral metallo-proteinase. The properties ofproteinases from lysosomes of rabbit PMNL are compared with those of thehuman, in relation to previously published data on the rabbit enzymes.

MATERIALS AND METHODS

Preparation of lysosomal proleinases from rabbit polymorphonuciear leucocytes

Peritoneal-extidate PMNL were induced according to the method of Cohn and Hirsch(1960). Adult rabbits were injected intraperitoneally wilh 150 ml of 0 1 % (w/v) glycogen(Type IL oyster. Sigma) in 0 1 5 M NaCl. Afler 18 h. the peritoneal cavity was washed outwilh 100 ml of 0 1 5 M NaCl and the exudiile collected into flasks on ice. The suspension wasfiltered ihrough gauze and (he cells ( > 9 0 % of which were PMNL) ccnlrifiiged (500 X^', 10min., 4° ) . washed once in 0 3 4 M sucrose, then rcsuspendeil in 0-34 M sucrose, concentratingthe cells 20-fold in comparison with the original exudate. Cells were lysed by mixing on avortex (Baggiolini, Hirsch and deDuve. 1969) and cell debris removed by low speed centri-fugation (500 X,;'. 10 min.. 4 ° ) . Lysosomes were sedimented by high speed centrifugalion(8.0(10 X,;'. 20 min.. 4 ° ) , the post-granule supernatant fluid removed and the lysosomeswashed once with 0 34 M sucrose. Lysosomes were resuspended finally in 0 01 M citric acid(typically, lysosomes from 4(H) m! of exudate were resuspended in 5 to 10 ml citric acid) andgently stirred at 4° for 18 to 24 h (Higuchi, Honda and Hayashi. 1975). Lysosomal debris wasremoved by centrifugation (11.000 Xi'. 20 min.. 4 " ) ; the supernatant fluid was collected andstored at 4° as such, or dialysed (cellulose tubing, 6000-8000 molecular weight cul-ofT) 18 hat 4" against 100 volumes of 0 05 M sodium phosphate buffer. pH 7-2. before storage at 4°.

Detection of proteo^lycan-dei-rtidina activity

{"*''S]-labelled rabbit articular cartilage was the substrate, which WLIS prepared by inject-ing 6-week-old rabbits intravenously with 1 to 2 mCi of {-'̂ 'SJ-sodium sulphate (Amershamt20 h before death. Shavings of articular cartilage were removed from ail joints and washedextensively with 0-15 M NaCI, followed by heating at 60° for 30 min. Heating lessenedautolysis during storage at 20° in 0 15 M NaCl and decreased the 'bleed" of -'"'S from thesubstrate during long-Ierm incubations (Hauser and Vaes, 1978). Assays contained 5 to 10mg of L'artilage, 0 1 M potassium phosphate buffer, pH 7 4, 0 25 M KCl and enzyme in a

RABBIT LYSOSOMAL PROTETNASES 65

tot:il volume of I ml. When incubation periods exceeded 6 h penicillin and streptomycin(Glaxo) were iiuJcd ai lOO fg/ml to minimize microbial attack of the substrate; this level ofantihiotics did not alter PMNL proteinase acliviiy. Incubation mixtures were placed at 37°for 10 min and reactions started hy ihe addition of enzyme; 0 1 ml samples were removed atI, 2, 4, 6 and 18 to 24 h for radioactivity measurement. At the end of the incuhation periodIhe cartilage was washed extensively with 015 M NaCl. then 1 4 ml of O-S M ncetiile buffer.pH 5 7. containing 25 mM EDTA added together with 10 mg cystcine-hydrochloride and0 1 ml pap;iin (5 mg/ml in 0 2 M NaCI). The cartilage was digested at 60" for 16 h. anda sample removed for radioactivity measurement. Activity was calculated as the rate of releaseof !''"'S;-fragments from the cartilage, and expressed as a percentage of the total amount of='-̂ S initially present in the cartilage: one unit released \% of the total radioactivity per hour.All activities were corrected for the release of -'''S occurring in the absence of enzyme. Otherbuffers used in assaying protcoglycanase activity are detailed in 'Results'.

Hydroh.si\ of .yvrUlwlic suhstratv.s

Llastiise-like activity (subsequently termed clastase) was determined by the rate ofliberation of p-nitrophenol from t-hutyIoxycarbonyl-L-al;inine 4-nitrophenyl ester (BOC-AIa-ONp) (Sigma). Assay mixtures contained 0 1 5 M sodium or potassium phosphate buffer. pH7 4. 0-5 M KCl and enzyme in a finul volume of I ml. Reaction was started by the additionof 20 f-l of substrate (10 mg/ml in dimethyl sulphoxide). and increase in iibsorbance at347.5 nm followed for 3 to 5 min at IT. The amount nf /j-nitrophenol released was deter-mined from a standard curve of p-nilropheno! (Sigma) in the assay buffer, and activityexpressed as wmoles p-nitrophenol releasecl/min/m! of enzyme (units/ml). All rates of reac-tion were corrected for the spontaneous release of /i-nitrophenol in assay huffer lackingenzyme; other buffers used in this assay are detailed in 'Results'.

Assays of other estenise activities h:ive heen described in detail previously. Assay usingN-acetyl-tri-L-alanine methyl ester was performed according to Janoff and Basch (1971).Hydrolysis of N-benzoyl-L-tyrosine ethyl ester and N-benzoy!-L-arginine ethyl ester wereassayed as described hy Gerber. Carson sind Hadorn (1974). Hydrolysis of N-benzoyl-DL-arginine 4-nitroanilide and N-acetyl-L-phenylalanine 2-naphthyl ester were assayed as describedby Starkey and Barrett (1976).

f'ffeci.y of inliUniors af-tiinsl proteoi-lycanase and elasta.seAppropriate volumes of test agents were added to incubation mixtures of the proteo-

glycanase and elastase :iss;iys and these incubated at 37° for I h before Ihe addition of sub-strate. EDTA (disotlium salt) was used al 5 or 10 mM and 1,10-phenanthroline at 1 mM.Pms-F was prepared as a stock solution of 0-1 M in propanol and used at I mM. TLCK(Sigma) was used at 0-2 mM and L-I-tosylamide-2-phenylethyl chloromethyl ketone (TPCK)(Sigma) was prepared as a stock solution of 20 mM in dimethyl sulphoxide and also used at0-2 mM. AAAPVCK (obtained from A. Sriratana. Dcp:irtment of Biochemistry. MonashUniversity. Melbourne) was dissolved at 0 Olf̂ f (w/v) in methanol and diluted with 0 1 Msodium phosphate buffer. pH 7 4. to 1 mM then used at (M mM. CiolJ thiomalate (May andBaker) was used at K)-- to KH' M. Cell cylosoi. from the post-granule supernatant fluid ofrabbii PMNL. was used :it 100 /'g protein and 7 5 ng of purified at-antitrypsin protein(obt;iincd from M. Nagiishima. Dep:irtment of Biochemistry. Melbourne University) was used.

lon-i.wihani'c ciiroiuato^raphy on CM-ci'lluIose

The CM-celliilose column (]-5 x 30 cm) was equilibrated in 0 05 M potassium phos-phate. 0 2 M NaCi huffer. pH 7 4. Crude citric acid extracts were either dialyscd against thisbuffer or used direclly. m;tking the volume up to 25 ml with column buffer before applicationof sample. The column was washed extensively following application of sample and then alinear gradient of 0 2-1.0 M NaCI in 0 0 5 M potassium phosphate buffer. pH 7.4. applied;the flow rate throughout was 60 ml/b and the procedure was performed at room temperature.The concentration of NaCI in fructions was determined from conductivity measurements.

66 M. L. BRITZ AND D. A. LOWTHER

Isoelectric focu.sin^

This was performed using an LKB 21 17 Multiphor system and was carried out accordingto the instructions supplied by the mantifacturer. Ampholines (pH 3 5-10, LKB) were usedas a S% solution in 100 ml water containing 5% (w/v) Sephadex G-75 (Pharmacia). Anarea of the dry gel was removed from the middle of the bed and mixed with the sample(3 ml containing 5% (v/v) ampholines) and the mixture reapplied to the gel. Electrofocusingwas performed at 400 mV, 20 va.\ initially and cooled by a water flow at 4°. The gel wasfractionated after 16 !o 20 h. then eluted with 3ml of 2 M NaCI. 0 - 1 % (w/v) Brij 35(Sigma) or 0 1 M potassium phosphate bufTer, pH 7-4.

Gel filtration on Sephadex G-lOO

A column of Sephadex G-lOO (Pharmacia) (2 .2 x 90 cm) was equilibrated at 4=' in0.05 M sodium phosphate bufTer, 0 1 5 M KCl. 0 - 1 % Brij 35, pH 7 2 , and flowed at 14 ml/h.The column was calibrated with markers of known molecular weight, blue dextran (Phar-macia), bovine serum albumin (67.000), ovalbumin (43,000), soy bean trypsin inhibitor(21,000) (all from Sigma) and Naiv'^'SO.,.

Other analy.ses

Radioactivity was measured using a Phillips liquid scintillation counter with externalstandardization for quench correction. Aqueous samples were made up to 2 ml with water and8 ml of scintillation fluid ( 0 . 8 % , w/v, 2.5-diphenyloxazole dissolved in 50%. v/v, TritonX-114 in toluene) added. Protein was determined by the method of Lowry et al. (1951),wish bovine serum albumin as the standard.

All chemicals were analytical grade reagents. Trypsin (type III from bovine pancreas)and elasiase were supplied by Sigma. Papain was prepared by J. Sandy (Department of Bio-chemistry, Monash University).

RESULTSStability of proteolytic enzymes

Citric acid extracts of lysosomes fpH 2 5) could be stored at 4° for upto 28 days without significant lo.ss in total proteoglycanase activity. However,holding these extracts at —20° caused a 50% loss of proteoglycanase activityin 15 days and only 20% of the original activity remained after 28 daysat —20\ When the extracts were dialysed against phosphate buffer (0 1 M,pH 7.4), proteoglycanase activity was stable for up to 28 days at 4" or—20°. Protein levels in the extracts tested were 0 3 to 3 mg/ml. Enzymeseluted from CM-cellulose and G-lOO were unstable to storage at 4°, evenwhen 0 1% Brij 35 was included; protein levels in these preparations wereless than 50

Hydrolysis of synthetic substratesIn an attempt to define the substrate specificity of the proteinases in

the crude lysosomal extract, artificial substrates of known specificity wereexamined for their possible hydrolysis. Neither N-benzoyl-DL-arginine4-nitroanilide or N-bcnzoyl-L-arginine ethyl ester were hydrolysed, althoughsatisfactory hydrolysis of these substrates by trypsin was observed. Tbis isconsistent with the report by Davies et al. (1971), wbo also failed to deteettrypsin-like activity in rabbit PMNL lysosomes. N-Benzoyl-L-tyrosine ethyl

RABBIT LYSOSOMAL PROTEINASES 67

ester was not hydrolysed and N-acetyl-L-phenylalaninc 2-naphthyI esterwas hydrolysed slowly, even when large volumes of extract were used (up to1 mg total protein), implying that only low levels of chymotrypsin-likeesterase aetivity were present in these extraets. Although the elastase substrateN-acetyl-tri-L-alanine methyl ester was readily hydrolysed by commereiallyavailable elastase, the lysosomal extracts showed poor aetivity against thissubstrate. In contrast, elastase-like activity against BOC-Aia-ONp wa.sreadily detected, having an average specific activity of 0079 U/mg protein.Dewald ei al (1975) reported that casein-degrading activity was present inrabbit PMNL lysosomes. but the specific activity was low in comparison withaetivity in human PMNL. Poor azocaseinase activity was found here incitric acid extracted lysosomes; azocasein was not used routinely as the assaywas considered wasteful of enzyme.

Effects of salts and ionic strent^lh of buffers on neutralproteinase and elastase activities

Proteoglycan degrading activity in crude lysosomal extracts was stimu-lated by 0-2 M NaCI or KCl (Table I ) : increasing the concentration of KClto 0 5 or 1 0 M further stimulated the aetivity (200 to 250% of the controlin 0 05 M phosphate) but also caused increased spontaneous release ofradioactivity from the I''''SI-cartilage substrate. Divalent cations, with theexception of copper, stimulated proteoglycanase activity. Magnesium chloridewas the most effective salt tested: even when the concentration of MgCl:-

TABLE iEffects of salts on proteo!"!ycana.se activity from rabbit PMNL ly.w.iomes.

BufTers andconcentrations

0.05 M phosphate

0 05 M Tris

0 05 M Tris,0 2M KCl

Salti-

.KCl (0-2 MlNaCI {"0-2 M)

KCl (0 2 M)MgCl..MgSO4MnCIoZnSO4CiiCUCuCl-.CuSO4

MgCI-.MnCI'.ZnSO,CuCI..

% conlrol activity'^

100179181

287316260196154290

930

12611014625 5

* Activity relative lo that ft)und when the listed buffer was used alone.t All sails were used at 50 mM final concentration, except CaCli.. which was at 10 mM.i-''''S)-labelled cartilage was incubated at 37° in bufTers at pH 7-2, with 30 lo 60 '̂g of

protein from cilric acid extracts of lysosomes. Aliquots were removed after 2. 4 and 24 h.The rate of |''"'SJ-peptide release was determined relative to the total counts originally presentand figures e.xpressed as ii percentage of the appropriate control activity. Each figure is theaverage obtained for four dilTercnt enzyme preparations, each tested in duplicate.

68 M. L. BRTTZ AND D. A. LOWTHER

TABLE 2Effects of salts and ionic strenf^th of phosphate btiffers on ela.stase from rabhit PMNL

Ivso.somes.

Buffers ;indconcentrations

0-05 M phosphate0-1 M phosphate0' 15 M phosphate0-5 M phosphate0.05 M phosphate0 • 1 M phosphate

0 05 M Tris

0 05 MTris.1 5 M KCl

Salt

———

KCl 11 5 M)KCl ( 1 5 M)

KCl ( 1 . 5 M )MgSOitMnCl..ZnS04CaCI..CUSO4

CuSO-i

% control activity

100*189279417150loot19812214285

1101 5

75

* Activity relative to that found in 0-05 M phosphate buffer.t Activity relative to that found in 0.1 M phosphate bufFer.t 50 mM final concentration was used for all salts.All assays were performed at pH 7-2 and 22°. and contained 30 lo 50

extracts of lysosomes.of citric acid

was decreased from 50 mM to 0 5 mM, the proteoglycanase was stimulatedto 2509c of the control activity. Copper ions dramatically inhibited proteo-glycanase activity: the minimum concentration required for 50% inhibitionwas 1 to 5mM CuCl-. The stimulatory effects of the divalent cations andinhibition by CuCl- were abolished, in part, by 0-2 M KCl.

Elastase activity was also stimulated by high concentrations of KCl{Table 2), but this effect was only significant when tested against a back-ground of low ionic strength phosphate buffers or in Tris buffer: changing theconcentration of KCl from 0 1 to 15 M did not alter the elastase activityin 0 1 M phosphate buffer. Increasing the concentration of Tris buffer from0 05 to 0 I M stimulated elastase activity by 3O9r ; further increases in con-centration (up to 0 3 M) failed to enhance this stimulation. In contrast,increasing the concentration of phosphate buffer from 0 05 to 0 5 Mmarkedly stimulated elastase activity (Table 2). High concentrations ofphosphate buffer were not used routinely for assaying elastase, as the rateof spontaneous hydrolysis of BOC-Ala-ONp was also elevated in thesebuffers. Divalent cations did not stimulate elastase to the same extent asseen for proteoglycanase. but copper ions did significantly inhibit elastase.Again, the inhibition by copper was reversed in the presence of KCl.

Effects of selected inhibitors on neutral proteltmse activityProteoglycanase activity in crude citric acid extracts of lysosomes was

inhibited 40 to 60% by 10 mM EDTA or I mM Pms-F, indicating thepresence of metal-dependent and serine proteinase activities. Proteolyticactivity was almost totally inhibited by 0 1 niM AAAPVCK, inhibited by

RABBIT LYSOSOMAL PROTEINASES 69

30-40% by 0 2 mM TPCK, and by 20% by 0 2 mM TLCK. The last twocompounds did not inhibit elastase activity when this was assayed usingBOC-AIa-ONp, whereas AAAPVCK completely abolished elastase activity.Both elastase and proteoglycanase activities were completely inhibited bythe cell cytosol (100 /*g of post-granule supernatant protein used) and «i-antitrypsin (7 5 yg u.sed). Inhibition of rabbit PMNL histonase by the cellcytosol was previously reported by Davies et al (1971). Gold thiomalate(1 mM) inhibited elastase by 80%, and proteoglycanase by 30 to 50%;assays here lacked KCl.

lon-e.xehange chromatography on CM-eelluloseFig. I shows a typical elution profile for a lysosomal extract applied to

CM-celluIose. This particular extract contained 2-83 mg protein, 0 22 unitsof elastase and 220 units of proteoglycanase; 60% of the original elastase

u2 -̂ca

O 2

CQ

4a 4b

/ -

10 20FRACTION

30 40NUMBER

50

o

CO

m

C

3

1 02Q

n0 a

oO

0.6 om2

0 4 ^

o0,2 2

Fig. 1. Ion-exchange chromatography of a citric acid extract of rabbit PMNL lysosomes.The column was equilibrated in 0-05 M potassium phosphate, 0 2 M NaCl buffer. pH 7-4. A6 ml extract of PMNL lyso^iomes was diluted lo 25 ml in the same buffer and applied lo thecolumn and 6 ml fractions collected first as the sample was applied. The column was washeduntil the absorbance at 280 nm was less than 0 02. Elution was performed with a 150 mllinear gradient of 0-2-i-O M NaCl in 0.05 M potassium phosphate buffer. Fractions weretested for absorbance at 2K0 nm ( ). proteoglycanase (PGase) activity (• • ) ,elastase activity ( • • ) and conductivity ( ). Designated fractions were pooled.

M. L. BRITZ .̂ ND D. A. LOWTHER70

and greater than 100% of the proteoglycanase activities applied were eluted.The apparent increase in total proteoglycanase can be accounted for by theelevated levels of NaCI in assays of fractions eluted by the NaCI gradient,where 0 5 ml of the fractions was required to detect activity. The specificactivity of elastase was increased 50-fold by CM-cellulose chromatography.Four peaks of proteoglycanase were usually recovered from CM-cellulose:peak I. comprising here 36% of the elastase recovered, eluted immediatelyafter the extract was applied; peak 2 eluted during the initial stagesof washing, whereas peak 3 eluted much later in the wash, aftermost of the protein had already eluted. The late elution of peak 3suggested that this activity was partially retarded by the CM-cellulose. How-ever, when the gel was equilibrated and washed in 0 05 M phosphate bufferlacking 0 2 M NaCl. activity corresponding to peak 3 still failed to bindand continued to be eiuted late in the wash. Elastase bound to the CM-cellulose and was eluted by 0 6 to 0 7 M NaCl (peak 4) . The proteogly-can-degrading activity associated with this activity normally appeared as abroad peak centred about the elastase activity: such apparent attenuation ofactivity reflects some of the peculiarities of the assay method.

Inhibitor profiles were performed in an attempt to classify the pro-teinases separated on CM-cellulose (Table 3). A worrying point was thatthe elastase inhibitor (AAAPVCK) appeared to significantly inhibit all ofthese activities. However. AAAPVCK failed to inhibit neutral metallo-proteinascs from cultured rabbit cartilage (Britz and Lowther, 1980),implying that this agent was acting selectively and that the dual inhibition ofpeak 3 activity by EDTA and AAAPVCK was a characteristic of thisenzyme. Elastase/proteoglycanase activity (peak 4, a and b) was onlysensitive to inhibition by AAAPVCK. CM-cellulose chromatography couldtherefore separate the two major proteinases in the lysosomal extracts, viz.the metallo-proteinase (peak 3) and the serine-dependent elastase (peak 4) .Activity eluted during application of the extract and initial washing (peaks1 and 2) is made up of any proteinases that failed to adsorb to the gel.Althoush the inhibitor profiles show that these activities are largely metallo-and serine-proteinases. it is impossible to say how many enzymes and what

TABLE 3Sensitivity of protvo\:lycann!:e from CM-cellulose to proteinase inhibitors.

Peak number

1234a4b

% inhibition

TPCK

492628

010

TLCK

600000

AAPVCK

9834

10010095

EDTA

6996

10070

TPCK and TLCK were used ill a final concenlration of 0 2 mM, AAAPVCK at 0 1 mMand EDTA ;il 5 mM. Figures are expressed as percentiige inhibition relative to activity in theabsence of inhibitor and in the presence of the solvent used for the inhibitor.

RABBIT LYSOSOMAL PROTEINASES 71

types are present here. There is no evidence at this stage to indicate ifelastase activities in peaks 1 and 4 are different, or if peak 1 corresponds tounbound peak 4 elastase activity. Peak I was usually 25-30% of the totalelastase eluted.

Isoelectric focusing (lEF) of lysosomal proteinasesAfter several attempts at isoelectric focusing of lysosomal extracts, it

became apparent that the same enzymes were not present in the same pro-portions when difl'erent preparations were used. It also became obvious that,if low ionic strength buffers were used to elute the proteinases followingfocusing, then the elastase activity was not detected. Recovery of proteinaseactivity was poor, being 20 to 50% of that applied, and the highest yield ofelastase activity was 10%. It is therefore dillicult to say if the profile shownin Fig. 2 represents all of the proteinases originally present, and if the relativeproportions reflect the original preparation. This profile, however, shows theenzymes typically recovered, in the proportions normally seen. The crudeextract applied here contained 0 22 units of elastase, 190.H units of proteo-glycanase and 2 25 mg of protein; 10%' of the elastase and 50% of theproteoglycanase were recovered. Three major peaks of protcoglycan-dcgrad-

ooo

X

"i 1-5

1.0

0-5

TO

o—1mOOI—

O

pHFig . 2 . Isoelectric focusing of rahbi t P M N L lysosomal proteinsises in a 3 -5 to 10 pH gra-d ient . T h e crude^ e n z y m e p repa ra t i on in I ) (H M citr ic acid w;is d ia lysed against lOO vo lumesof 0 -05 M sod ium p h o s p h a t e . O l ' ^ r Brij 35 bufTer. pH 7 2 ;ind 2 X 5 ml upplicd to iheS e p h a d e x G - 7 5 - a m p h o l i n e gel. A vol tage of 4()t) m V was appl ied , the cu r ren t chang ing f rom20 to 2 m A over 18 h. Thu gel was divided into 30 f ract ions wh ich were e lu ted with 2 x 1 5ml of 2 M N a C l . 0 ' \'','r Brij 35. Aficr record ing the p H of each f ract ion, these were diiilysedagainst 2 I of 0 -05 sod ium p h o s p h a t e . 0 I ' c Brij 35 buffer. p H 7 - 2 . for 8 h a t 4 " and testedfor p ro t eog lycanase ( • • ) and e las lase ( • • ) act ivi ty.

72 M. L. BRITZ AND D. A. LOWTHER

30 40 50FRACTION

60 70NUMBER

80

Fig. 3. Gel chromatography on Sephadex G-100. Five ml of a crude enzyme preparation wasdialyzed against 100 volumes of 0-05 M sodium phosphate, 0-15 M KCl. 0 - 1 % Brij 35 buffer,pH 7-2. then applied to a column of Sephade.x G-IOO equilibrated at 4° in the same buffer.Fractions (3-6 rnl) were collected and assayed for proteoglycanase (• • ) and elastase( • • ) activities, and absorbance at 280 nm recorded ( ). The column was cali-brated using blue dextran (V,,). bovine serum albumin (BSA), ovalbumin (Ov), soy beantrypsin inhibitor (SBTI) and Naj-'^'SO., ( V T ) .

ing activity were seen (pH 5 5. 8 5 and 9 1 ) plus several minor peaks(pH 4 2. 6 2. 7-2 and 7-7). The proteinases focusing at pH 5 5 and 8 5could be readily eluted from the gel by low ionic strength buffers and werefrequently recovered as the major proteinases. The proteinase focusing atpH 5-5 was partially inhibited by TPCK, AAAPVCK and TLCK. but notby EDTA: it is apparently a serine-proteinase but is sensitive to a range ofdifferential inhibitors. One would not expect this proteinase to bind to CM-cellulose. The proteinase focusing at pH 8 5 was completely inhibited by10 mM EDTA. I mM 1. 10-phenanthroline and AAAPVCK. and partiallyinhibited by TLCK. This activity probably corresponds to peak 3 from CM-cellulosc. Elastase focused in a single peak coincident with proteoglycanaseactivity at pH 9.1. No other metallo-proteinases were recovered from the geland the minor peaks were all partially inhibited by the serine-proteinaseinhibitors. All of the proteinases above also degraded histone (Sriratana,personal communication).

Gel chromatography on Sephadex G-100Gel chromatography was performed on part of the same preparation

used above for electrofocusing: 0-27 units of elastase. 231 units of proteo-glycanase and 2-74 mg of protein were applied. Fig. 3 shows the elution

RABBIT LYSOSOMAL PROTEINASES 73

profile, where 20% of the elastase and approximately 30% of the proteo-glycanase were recovered. Total activity recovered from this column wasnormally low, with elastase as the major activity eluted; recovery wasenhanced by 0-1% Brij 35. Although the protein associated with the elastasewas low (<20 iig/m\), the loss in total activity meant that the specificactivity did not increase. The two minor peak.s of proteoglycan-dcgradingactivity had molecular weights of approximately 80.000 and 35.000; activitywas low and did not survive storage at 4° for 48 h, so that inhibitor profileswere not determined. The molecular weight of the elastase was 8.000 to10,000. The chromatography was repeated in 1 5 M NaCl in 0 05 M sodiumphosphate, 01%i Brij 35 buffer, pH 7 2. in an attempt to eliminate possiblenon-specific adsorption to the gel: molecular weight of the elastase here wasabout 11,000.

DISCUSSIONThe major neutral proteinases of rabbit PMNL lyso.somes are two serine

proteinases and one metallo-proteinase. all of which can degrade rabbitcartilage proteoglycan and histone. These activities could be separated byisoelectric focusing (IEF) and CM-cellulose chromatography and differen-tiated on the basis of sensitivity to inhibitors. The proteinase of pi 8 5 wasa metallo-proteinase which failed to adsorb completely to CM-cellulose. butwas retarded by this to elute after the majority of protein had washed through.Elastase was also a cationic protein of pi 9 1, and could be partially purifiedby CM-cellulose and Sephadex G-100 chromatography. The molecular weightdetermined by gel chromatography was 8.000 to 11,000 daltons. However,it must be pointed out that this figure is only an indication of the true mole-cular weight of the enzyme, as elastases have been shown to be anomalouslyretarded on Sephadex gels even in the presence of high salt concentrations(Starkey and Barrett, 1976). The rabbit elastase indeed proved to be 'sticky',requiring high salt concentrations to elute it following IEF. Even so, themolecular weight of the rabbit PMNL elastase is substantially lower thanpreviously reported figures for human neutrophil elastases, which range from22,000 to 37,000 daltons (Starkey, 1977). The protein levels associatedwith the elastase following partial purification were too low to allow alterna-tive determinations of size by methods such as amino acid analysis. Thethird major proteinase was a serine-proteinase with a pi about 5 5. Severalminor activities were also present, none of which hydrolysed synthetic sub-strates for trypsin or chymotrypsin and were not characterized beyond IEFbecause of low activity.

It is interesting to compare the spectrum of proteinases reported herewith those .seen by Davies et al (1971) and Dewald et al (1975). Thelatter reported that rabbit PMNL azurophil granules contained, on the basisof electrophoretic patterns of Triton X-lOO extracts, one major elastase plus3 or 4 very minor activities, but no chymotrypsin-like proteinase activity.We also found only one major elastase and only low chymotrypsin-like

74 M. L. BRITZ AND D. A. LOWTHHR

esterase activity in citric acid extracts of lysosomes. The major neutralhistonase seen by Davies et al (1971) had a pi of 4 2 - 5 2 and, if oneassumes that the major enzyme eluting from Sephadex G-75 was the sameactivity, a molecular weight of 30.000 to 50,000 (calculated from the datapresented). Starkey (1977) suggested that the pi estimated by Davies et al(1971) was aberrantly low because it was determined in the presence ofheparin. However, any ionic complex between the histonase and heparinwould probably dissociate during IEF, so that the pi 4 2-5 2 histonase ofDavies et al. (1971) may correspond to the pT 5 5 activity reported in thepresent communication. This is contrary to the further prediction of Starkey(1977) that the pi 4 2-5 2 histonase activity was the rabbit PMNL elastase.Davies et al. (1971) did not look at the inhibitor profiles of the focusedenzymes or activities eluted from the Sephadex G-75. and only assayed forhistonase activitity after showing that histonase was located exclusively inthe azurophil granules. If one re-examines the data presented here and com-pares it with data in the present communication, it becomes, apparent thatDavies et al (1971) were detecting the same range of enzymes but indifferent proportions. Besides the major histonase activity, they showedseveral minor proteolytic activities focusing at pH 6 3. 7.0. 8 5 and 9 0,but. because elastase was not assayed and inhibitor profiles were not per-formed, it is diOicult to correlate these activities to our elastase (pi 9 1 )and metallo-proteinase (pi 8 5). Furthermore, the Sephadex G-75 profileof Davies et al. (1971) shows a minor peak of histonase eluting aftercytochrome C and lysozyme which probably is the elastase/proteoglycanase/histonase (MW 8.000-fl.OOO) seen on our Sephadex G-100 profiles. Thedifferences in the proportions of proteinases reported here and those seen byDavies et al (1971) may reflect the different methods of preparing crudeextracts of lysosomes.

The major neutral proteinases of human neutrophil leucocytes are theserine proteinases. elastase and eathepsin G, which are described in somedetail by Starkey (1977). The rabbit PMNL neutral proteinases differ inseveral ways from the human enzymes, tbe most obvious disparity being thelack of a eathepsin G analogue in the rabbit PMNL, plus the presence of acationic neutral metallo-proteinase. Furthermore, elastase in the humanPMNL is a composite of three closely related activities together with severalminor elastases. whereas the rabbit cells have one major elastase activity(data presented here and by Dewald et al, 1975). The rabbit and humanelastases share certain properties, such as similar isoelectric points, stimulationby high salt concentrations, and a tendency to adsorb non-specifically toinsoluble surfaces at low ionic strength. The inhibition of rabbit PMNLelastase by copper ions has not been reported for human elastase. A majordifference in the elastases from rabbit and human PMNL may be in the sizeof the molecules. This is of some interest regarding the potential physiologicalroles of these enzymes in the two species, as the molecular weight mayinfluence the ability of the molecules to penetrate macromolecular matrices.To clarify this point, further purification of the rabbit elastase would be

RABBIT LYSOSOMAL PROTEINASES 75

required. This is made difficult by the relatively small starting levels ofenzyme (10- to 30-fold lower than in human leucocytes; Dewald et al,1975) plus significant losses in activity at several stages during purificationwhich was compounded by the instability of the partially purified enzymes(also seen by Davies (?r fl/., 1971).

REFERENCESBAGGIOLINI. M.. HIRSCH. J. G.. and

deDuvn, C. (1969): 'Resolution ofgranules from rabbit heterophil leuco-cytes inio distinct populations by zonalsedimenlalion.' J. Cell. Biol., 40, 529-541.

BR[TZ, M. L., and LOWTHEK. D . A. (1980):

'Effects of medium constituents on secre-tion of latent collagenase and proteogly-canase by cultured rabbit articular car-tilage.' Proc. Au.st. Biochem. Soc. 13. 19.

COHN, Z. A., and HIRSCH. J- G. (1960):

'1 he isolation and properties of thespecific cylopliismic granules of rabbitpolymorphonuclear leucocytes.' /. Exp.Med.. 112. 983-1004.

DAVIHS. P,, KRAKAUER. K.. and WEISSMANN,

G. (1970): 'Subcellular distribution ofneutral protease and peptidases in rabbitpolymorphonuclear leucocytes.' Nature.Lond., 228, 761-762.

DAVIES, P., RITA, G . A.. KRAKAULR, K., and

WEISSMANN, G . (1971): 'Characterizationof a neutral protease from lysosomes ofrabbit polymorphonuclear leucocytes.'Biochem. /.. 123, 559-569.

D F . W A L D , B . , RlNULER-LUDWKl, R.. B R E T Z .

U.. and BAGGIOLINI, M. (1975): 'Subcel-lular localization and heterogeneity ofneutral proteases in neutrophilic polymor-phonuclear leucocytes.' J. E.xp. Med.,141, 709-723.

GERBER, A. C , CARSON. J. H., and HADORN,

B. (1974): 'Parlial purification andcharacterization of a chymolrypsin-like

enzyme from human neutrophil leuco-cytes". Biochim. Biophys. Acta. 364. 103-112.

H/vusrR, P.. and VAES. G. (1978): 'De-gradation of cartilage proteoglycans by aneutral proteinase secreted by rabbitbone-marrow macrophages in culture.'Biochem. /., 173. 275-284.

HioucHi, Y.. HONDA, M.. and HAYASHI,

H. (1975): 'Produclion of chemotacticfaclor for lymphocytes by neutral SH-dependent protease of rabbit PMN leu-cocytes from immunoglobulins, especially]gM.' Cellular Immunol., 15, 102-108.

JANOEF. A., and BASCH, R. S. (1971):

'Further studies on elastase-like esterasesin human leucocyte granules.' Proc. Soc.Exp. Biol. Med., 136. 1045-1049.

l-owRv, O. H.. RoSENBROuriH, N. J., FARR,A. L.. and RANDALL, R. J. (1951): 'Pro-tein measurement with the Folin phenolreagent". /. Biol. Chem., 193. 265-275.

STARKEY. P. M. (1977): 'Elastase andeathepsin G; the serine proleinases ofhuman neutrophil leucocytes and spleen.'In "'Proteinases in Mammalian Cells andTissues", A. J. Barret (ed.). NorthHolland Publishing Company, Amster-dam, pp. 57-89.

STARKrv. P. M., and BARRETT, A. J. (1976):•Neutral proteinases of human spleen.Purificalion and crileria for homogeneityof elastase and eathepsin G." Biochem.J.. 155, 255-263.