Hydrolysis ofthe Elastase Substrate Succinyltrialanine ......Type II, Sigma Chemical Co., Ltd.)...

8
Hydrolysis of the Elastase Substrate Succinyltrialanine Nitroanilide by a Metal-Dependent Enzyme in Rheumatoid Synovial Fluid JEREMY SAKLATVALA From The Center for Rheutmatic Diseases and Glasgow, Scotland A B S T R A C T A new type of enzyme hydrolyzing the elastase substrate siuccinyl-L-alanyl-L-alanyl-L-ala- nine-4-nitroanilide has been found in cell-free rheuma- toid synovial fluid. Normal plasma and osteoarthritic synovial fluid contained relatively little enzyme. The pH optimum was 8.0. Unexpectedly, the enzyme activity was not due to leukocyte elastase or any proteinase bound to a2-macroglobulin. The enzyme activity was metal-dependent being inhibited by chelating agents but not by di-isopropylfluorophos- phate or thiol-blocking reagents. Gel chromatography showed the enzyme activity was associated with ma- terial of high molecular weight. On Sepharose 4B chromatography two-thirds of the activity eluted in the void volume and one-third in a position of about 106 mol wt. Ultracentrifugation showed that both components were associated with lipid. The buoyant density of the higher molecular weight material was 1.15-1.22 gIml., and that of lower molecular weight material was 1.2-1.33 gIml. No latency of the enzyme was revealed by freezing and thawing or treatment with detergents. The nature of the enzyme is dis- cussed. It is likely to be a proteinase possibly bound to some kind of membrane fragment. INTRODUCTION It has been frequently postulated that tissue destruc- tion in rheumatoid arthritis may be caused by lysosomal proteinases (1, 2). Of particular interest are the proteinases of the neutrophilic polymorphonuclear leukocytes, since large numbers of these cells are found in rheumatoid synovial fluid. Detailed studies Dr. Saklatvala's present address is Strangeways Research Laboratory, Worts' Causeway, Cambridge, England. Received for publication 21 May 1976 and in revised form 9 January 1977. University Department of Medicine, Royal Infirmary, of the action of these proteinases on cartilage proteo- glyean have been reported recently (3, 4). Three pro- teinases from human neutrophils have been described, and all are optimally active near physiological pH: leukocyte elastase (5, 6) a chymotrypsin-like enzyme now called cathepsin G (7, 8) and a collagenase (9, 10). Synovial fluid contains potent inhibitors of these enzymes, the major ones being a,-proteinase inhibitor and a2-macroglobulin (a2M).I Ohlsson (11) has shown by immunological techniques that rheumatoid synovial fluid contains complexes of neutrophil pro- teinases with the inhibitors, which suggests that the proteinases are being released from the cells into the fluid. Complexes of proteinase and a1-proteinase inhibitor are catalytically inactive, but complexes of proteinase with a2M will hydrolyze low molecular weight sub- strates (12, 13). The work to be described arose from an attempt to detect leukocyte elastase bound to a2M in rheumatoid synovial fluid by using the highly specific and stable elastase substrate succinyl-L-alanyl- L-alanyl-L-alanine-4-nitroanilide (Suc-Ala3-NPhNO2) (14). Rheumatoid synovial fluid was found to contain enzymic activity that hydrolyzed this substrate but it was not attributable to leukocyte elastase because inhibition experiments showed that the activity was metal-dependent and did not have the characteristics of a serine enzyme. Nor was the activity associated with a2M, so it was not due to a complex of proteinase with this protein. Finding a metal-dependent enzyme hydrolyzing the elastase substrate was of great interest because it was likely to be a proteinase, and was a hitherto unrecog- I Abbreviations used in this paper: a2M, a2-macroglobulin; Suc-Ala3-NPhNO2, succinyl-L-alanyl-L-alanine-4-nitroanilide; Dip-F, di-isopropylfluorophosphate; Pms-F, phenylmethane- sulfonylfluoride; p, density. The Journal of Clinical Investigation Volume 59 May 1977 794-801 794

Transcript of Hydrolysis ofthe Elastase Substrate Succinyltrialanine ......Type II, Sigma Chemical Co., Ltd.)...

  • Hydrolysis of the Elastase Substrate SuccinyltrialanineNitroanilide by a Metal-Dependent Enzyme inRheumatoid Synovial Fluid

    JEREMYSAKLATVALA

    From The Center for Rheutmatic Diseases andGlasgow, Scotland

    A B S T RA C T A new type of enzyme hydrolyzing theelastase substrate siuccinyl-L-alanyl-L-alanyl-L-ala-nine-4-nitroanilide has been found in cell-free rheuma-toid synovial fluid. Normal plasma and osteoarthriticsynovial fluid contained relatively little enzyme. ThepH optimum was 8.0. Unexpectedly, the enzymeactivity was not due to leukocyte elastase or anyproteinase bound to a2-macroglobulin. The enzymeactivity was metal-dependent being inhibited bychelating agents but not by di-isopropylfluorophos-phate or thiol-blocking reagents. Gel chromatographyshowed the enzyme activity was associated with ma-terial of high molecular weight. On Sepharose 4Bchromatography two-thirds of the activity eluted inthe void volume and one-third in a position of about106 mol wt. Ultracentrifugation showed that bothcomponents were associated with lipid. The buoyantdensity of the higher molecular weight material was1.15-1.22 gIml., and that of lower molecular weightmaterial was 1.2-1.33 gIml. No latency of the enzymewas revealed by freezing and thawing or treatmentwith detergents. The nature of the enzyme is dis-cussed. It is likely to be a proteinase possibly boundto some kind of membrane fragment.

    INTRODUCTION

    It has been frequently postulated that tissue destruc-tion in rheumatoid arthritis may be caused by lysosomalproteinases (1, 2). Of particular interest are theproteinases of the neutrophilic polymorphonuclearleukocytes, since large numbers of these cells arefound in rheumatoid synovial fluid. Detailed studies

    Dr. Saklatvala's present address is Strangeways ResearchLaboratory, Worts' Causeway, Cambridge, England.

    Received for publication 21 May 1976 and in revised form9 January 1977.

    University Department of Medicine, Royal Infirmary,

    of the action of these proteinases on cartilage proteo-glyean have been reported recently (3, 4). Three pro-teinases from human neutrophils have been described,and all are optimally active near physiological pH:leukocyte elastase (5, 6) a chymotrypsin-like enzymenow called cathepsin G (7, 8) and a collagenase(9, 10). Synovial fluid contains potent inhibitorsof these enzymes, the major ones being a,-proteinaseinhibitor and a2-macroglobulin (a2M).I Ohlsson (11) hasshown by immunological techniques that rheumatoidsynovial fluid contains complexes of neutrophil pro-teinases with the inhibitors, which suggests that theproteinases are being released from the cells intothe fluid.

    Complexes of proteinase and a1-proteinase inhibitorare catalytically inactive, but complexes of proteinasewith a2M will hydrolyze low molecular weight sub-strates (12, 13). The work to be described arosefrom an attempt to detect leukocyte elastase bound toa2M in rheumatoid synovial fluid by using the highlyspecific and stable elastase substrate succinyl-L-alanyl-L-alanyl-L-alanine-4-nitroanilide (Suc-Ala3-NPhNO2)(14). Rheumatoid synovial fluid was found to containenzymic activity that hydrolyzed this substrate butit was not attributable to leukocyte elastase becauseinhibition experiments showed that the activity wasmetal-dependent and did not have the characteristicsof a serine enzyme. Nor was the activity associatedwith a2M, so it was not due to a complex of proteinasewith this protein.

    Finding a metal-dependent enzyme hydrolyzing theelastase substrate was of great interest because it waslikely to be a proteinase, and was a hitherto unrecog-

    I Abbreviations used in this paper: a2M, a2-macroglobulin;Suc-Ala3-NPhNO2, succinyl-L-alanyl-L-alanine-4-nitroanilide;Dip-F, di-isopropylfluorophosphate; Pms-F, phenylmethane-sulfonylfluoride; p, density.

    The Journal of Clinical Investigation Volume 59 May 1977 794-801794

  • nized enzyme in rheumatoid synovial fluid which couldbe involved in tissue destruction. Further experi-ments were done to attempt to isolate and charac-terize the enzyme responsible for the activity.

    METHODS

    MaterialsAll chemicals used were analytical reagent grade and ob-tained from British Drug Houses, Ltd., Poole, Dorset,England, except where specified.

    Synovial fluid and plasma. Synovial fluid was aspiratedfrom knee joints of 21 patients with rheumatoid arthritisand put into sterile plastic containers. According to the Ameri-can Rheumatism Association critera, 18 had classical rheuma-toid arthritis and 3 had definite rheumatoid arthritis. Thelatter all had a negative sheep cell agglutination test forrheumatoid factor while the former were all positive. Theages of the patients ranged from 30 to 70 yr. Blood sampleswere taken simultaneously with the synovial fluid from 11patients and put into heparinized tubes, from these theplasma was separated. After leukocyte counts, the leukocyteswere separated from the synovial fluid by centrifugation at3,000 rpm for 20 min at 4°C. Fluid was stored at 40C andassayed for enzymic activity within 24 h. Samples of synovialfluid for chromatography and ultracentrifugation were treatedwith ovine hyaluronidase (hyalase, Fisons Ltd., Lough-borough, England). 10 manufacturer's U were added per 1 mlof synovial fluid and the fluid was incubated at 370Cfor 20 min. The hyaluronidase preparation caused no hy-drolysis of Suc-(Ala)3-NPhNO2 at the concentration used.

    Synovial fluid was also taken from knee joints of fivepatients with osteoarthritis. These were noninflammatory ef-fusions, all with leukocyte counts

  • o p p sE

    50 040

    400

    30-

    20-

    1 0

    0 RA RA normalI QAsynovial plasma plasma synovial

    fluid f luid

    FIGURE 1 Hydrolysis of Suc-(Ala)3-NPhNO2 by synovial fluidand plasma.

    0.1 ml of 10 mMSuc-Ala3-NPhNO2 to 1 ml of the fraction.After incubation for the specified time (1-8 h) the absorbancechange at 410 nm was measured against a control assayto which substrate was added after incubation. For inhibitionof nitroanilide hydrolase activity, inhibitors were prein-cubated with enzyme solution for 20 min at room tempera-ture before addition of substrate. In some experimentsdi-isopropylfluorophosphate (Dip-F) was preincubated withenzyme for up to 10 h. Dip-F, phenylmethane sulfonyl-fluoride (Pms-F), 1, 10 phenanthroline, EDTA, p-chloromer-curibenzoic acid, iodoacetamide, soybean trypsin inhibitor,and Triton X-100 were all from Sigma Chemical Co., Ltd.,London, England. Basic pancreatic trypsin inhibitor wasTrasylol from Bayer AG, Wuppertal-Elberfeld, Germany.

    Inhibition of chymotrypsin activity (Bovine chymotrypsinType II, Sigma Chemical Co., Ltd.) and proteolytic activityof rheumatoid synovial fluid leukocyte extracts by a2M wasmeasured by use of the substrate hide-power azure (Cal-biochem Ltd., Hereford, HR4 9BO, England) as describedelsewhere (18).

    Ultracentrifuge fractions of synovial fluid were assayed forproteolytic activity by use of azocasein substrate (SigmaChemical Co., Ltd.) (19), hide-power azure and Remazolbrilliant blue dyed elastin (15). These assays were donewith 0.1 M Tris-HCl buffer, pH 8.0, containing 0.1% TritonX-100. Reaction mixtures were incubated for periods up to10 h at 370C.

    RESULTS

    Detection of hydrolysis of succinyltrialanine nitro-anilide. Rheumatoid synovial fluid contained moreenzymic activity hydrolyzing Suc-Ala3-NPhNO2 thannormal plasma or osteoarthritic synovial fluid (Fig. 1).Two samples of rheumatoid plasma also contained ahigh level of enzymic activity. Apart from one patient,the synovial fluid levels were higher than the plasmalevels. The enzymic activity in synovial fluid correlatedpoorly with the total leukocyte counts and the linearregression coefficient, r = 0.464 was only significant atthe 5% level (Student's t test = 2.29, degrees of free-dom = 15).

    The enzymic activity was independent of ionicstrength (I) in the range I = 0.1-1.0 and the pH de-pendence is shown in Fig. 2. There is a sharp opti-mumat pH 8.0 and no detectable actvity below pH 6.0.The assay could not be used above pH 10.0 becauseof the susceptibility of nitroanilides to hydrolysis atalkaline pH.

    To investigate possible latency of the enzymic ac-tivity in rheumatoid synovial fluid, five samples wereassayed after rapidly freezing and thawing six times,and in the presence of 0.1% Triton X-100. In threefluids 0.1% Triton X-100 caused slight increases ofactivity (up to 20%) in the other two there was noeffect. Freezing and thawing caused no change inthe activity.

    Sephadex gel chromatography. Fractionation onSephadex G-150 of rheumatoid synovial fluid (Fig. 3a)and rheumatoid plasma (Fig. 3b) containing high levelsof nitroanilide hydrolase activity showed that all of theenzyme activity was associated with high molecularweight material eluting in the void volume. The re-covery of activity in these experiments was 90-95%.

    0.10EA00.05

    3 4 5 6 7 8 9 10

    pH

    FIGURE 2 Effect of pH on hydrolysis of Suc-(Ala)3-NPhNO2by rheumatoid synovial fluid. Assay buffers used were pH3-6; 0.1 M citrate, pH 7-9: 0.1 M Tris-HCl, pH 10:0.1 M glycine-NaOH. Enzyme activity is expressed as ab-sorbance change (AA410)/hour per milliliter synovial fluidadded to the assay.

    796 J. Saklatvala

  • t2

    1.0

    0.8

    I1

    Ec

    coC\j

    a')0

    cc

    0

    CD

    0.6

    0.4

    0.2

    1.2

    1.0

    0.8

    0.6

    0.4

    0.2

    0

    A

    a / 4\,/ '11 / \\ II I 1

    I - .1I II

    "

    IOU zuElution volume (ml)

    250 JU3I

    0.3

    Q2

    0.1

    0.3

    0.2

    0.1

    FIGURE 3 Sephadex G150 chromatography of rheumatcsynovial fluid (A) and rheumatoid plasma (B). Column 2x 75 cm; samples: 2 ml synovial fluid, 1 ml plasma. Asorbance at 280 nm ( --- ), hydrolysis of Suc(Ala)3-NPhN1expressed as absorbance change (AA41O)/8 h per ml of fracti(0).

    This result was compatible with the idea that tisubstrate was being hydrolyzed by leukocyte elasta,bound to a2M. Leukocyte elastase is a serine proteinaand is inhibited by compounds such as Dip-F axPms-F which react with the serine residue inactive site (5, 6) so the effect of various proteina,inhibitors on the synovial fluid nitroanilide hydrolaactivity was studied.

    Inhibition of the hydrolysis of succinyltrialanixnitroanilide. The results of inhibition studies ashown in Table 1. The enzymic activity was not Etributable to a serine proteinase since it was ninhibited by Dip-F (preincubated up to 10 h) or Pms-The activity was not significantly inhibited by tithiol blocking agents p-chloromercuribenzoate axiodoacetamide, but it was inhibited by the chelatiiagents EDTAand 1,10 phenanthroline, showing ththe enzymic activity was metal dependent. The failuof soybean trypsin inhibitor or Trasylol to inhibit tiactivity was not surprising since these are seriiproteinase inhibitors, although it is conceivable thif the enzyme were present in complexes with anothmacromolecule it might be sterically hindered froreacting with the higher molecular weight inhibitor

    Polyacrylamide gel electrophoresis. The posebility that the enzymic activity might be due to elatase or some other proteinase bound to a2M was e

    plored further by using electrophoresis. The followingpreliminary experiments showed that polyacrylamidegel electrophoresis was a suitable method for detect-ing a2M/proteinase complexes. The change in theelectrophoretic behaviour of a2M caused by adding aproteinase to it is shown in Fig. 4. Saturation

    , ofa2M with chymotrypsin caused the a2M to run slightlyg ahead of the native a2M position (gels a and b).R Addition of enough chymotrypsin to react with half ofo the a2M (determined from proteolytic inhibition assays)

  • FIGURE 4 Polyacrylamide gel electrophoresis of mix-tures of a2-M and proteinases. (a) 30,ug a2M, (b) a2M saturatedwith chymotrypsin. Equireactant amounts were calculatedfrom proteolytic inhibition. (c) a2M half saturated with chymo-trypsin, (d) a2M saturated with rheumatoid synovial fluidcell extract, (e) a2M half saturated with rheumatoid synovialfluid cell extract. The arrow marks the direction of migra-tion. The line diagram shows the hydrolysis of Suc(Ala)3-NPhNO2 by eluates from sliced gels run with equireac-tant amounts of a2M and rheumatoid synovial fluid leukocyteextract (same as gel d). 0, eluates treated with DFP (1 mM)for 20 min before addition of substrate; 0, not treated withDFP. See under Methods for details of electrophoreticconditions.

    plexes run slightly ahead of free a2M in this electro-phoretic system, and that it was possible to elutethe complexes from the gel to detect their enzymicactivity, and that this activity was inhibited by Dip-Fin the case of a serine proteinase. The purified a2M(gel a) contained a small amount of protein runningahead of the main band and this could representa trace of a2M/proteinase complex, denatured a2M,or impurity.

    To see if the nitroanilide hydrolase activity insynovial fluid was associated with a2M, active frac-tions from Sephadex G- 150 chromatography wereconcentrated by ultrafiltration and run in electro-phoresis. Gels were stained for protein or sliced andprotein eluted; the eluates were assayed for hydrolysisof Suc-Ala2-NPhNO2 (Fig. 5a). 95%of the activity failedto enter the resolving gel, being trapped in the firstslice, but there was a small peak of activity (represent-ing about 5% of the activity) in the region of a2M.Fig. 5b shows a similar experiment with activityfrom rheumatoid plasma after Sephadex G-150 chro-matography; all the activity recovered was in the firstgel slice. The overall recovery of activity from electro-phoresis was about 40%. The recovery of 50-60% ofCa2M (measured as chymotrypsin inhibition) in theseexperiments showed that protein elution was satis-

    798 J. Saklatvala

    factory. An identical distribution of enzymic activitywas found when whole synovial fluid or plasma wasrun on electrophoresis. The enzymic activity that waseluted from the first gel slice was not inhibited byDip-F.

    These results gave no support to the idea that thehydrolysis of Suc-Ala3-NPhNO2 was due to leukocyteelastase or any other proteinase bound to a2M. Ratherthey suggested that the enzyme was of very highmolecular weight or was associated with some highmolecular weight protein other than a2M.

    Ultracentrifugation. These experiments weredone to see if the enzyme activity was associatedwith lipid-containing material rather than with a2M.The synovial fluid was centrifuged at starting densities1.006, 1.063, and 1.225 g/ml. These were chosen be-cause they are standard densities used for flotationof very low density, low density and high densitylipoproteins, respectively (21). Either sucrose or KBrwas used to adjust the density with essentially the

    0.6-

    0Q4 L

    0b I 1

    E;OaL

    .a 14rc 0.6

    0A -

    0.2

    O 5 O 0 15gel slice

    c 1111I1FIGURE S Polyacrylamide gel electrophoresis of the nitro-anilide hydrolase activity from rheumatoid synovial fluid (A)and plasma (B). Samples were from Sephadex G150chromatography void volume peak. Gel (a) (synovial fluid)and gel (c) (plasma), show distribution of protein, gel (b)is pure a,M. Eluates of gel slices were assayed for hydrol-ysis of Suc(Ala)3-NPhNO2 (e), and inhibition of chymotrypsin(O, shown in A only).

  • same results. After 20 h centrifugation at 50,000 rpmthe enzymic activity sedimented to the bottom of thetube at p = 1.006 and 1.063 g/ml, but 75% of it floatedin the top fraction at p = 1.225 g/ml (Fig. 6). Fivesamples of synovial fluid gave similar results and theoverall recovery was 70-80%. Evidently a major partof the enzymic activity was associated with lipid.

    The material buoyant at 1.225 g/ml was also assayedfor proteinase activity. The sample from one synovialfluid hydrolyzed azocasein in the presence of 0.1%Triton X-100. No activity was observed without deter-gent and the activity was inhibited by Dip-F andnot EDTA. The samples from the other four synovialfluids did not hydrolyze azocasein. None of the sampleshydrolyzed hide-powder azure or Remazol brilliantblue-elastin. It was concluded that there was no detect-able proteinase activity attributable to the enzymehydrolysing Suc-(Ala)3-NPhNO2. The proteinase ac-tivity found in the one sample was due to a serineproteinase.

    Sepharose 4B gel chromatography. The concen-trate from the Sephadex G-150 column (Fig. 3a) wasfractionated further on Sepharose 4B (Fig. 7). 65%of the enzymic activity recovered was in the voidvolume and the remainder in the first part of the mainprotein peak. The main peak corresponds to a2M asshown by chymotrypsin inhibition: note that thesecond enzyme peak does not correspond to the a2M

    p =1.0630.6[

    0.4[

    c0

    0.6

    -1

    0.4

    0.2

    0.2[

    i- ~ ~~p=1.225

    _ ,0 1 2 3

    Fraction4 5

    FIGURE 6 Ultracentrifugation of rheumatoid synovial fluid.The histograms show the distribution of enzyme hydrolys-ing Suc-(Ala)3-NPhNO2. Fraction 1 = top of tube, fraction 5= bottom of tube. 3 ml of rheumatoid synovial fluid werecentrifuged in total volume of 10 ml. Density adjustedwith KBr. Centrifugation at 50,000 rpm, 20 h, 15°C.

    Ecm

    8C

    0

    .0

    Cal.0

    0 20 40Elut ion volume (ml)

    10116 E14 -

    12 fllo.*E

    ._

    8 i6

    4 >

    2 C)a)

    60

    FIGURE 7 Sepharose 4B chromatography of void volumepeak from Sephadex G150 chromatography or rheumatoidsynovial fluid. Column 15 x 30 cm. Absorbance at 280 nm(- -), hydrolysis of Suc-(Ala)3-NPhNO2 (0), inhibition ofchymotrypsin (0).

    peak. A similar distribution was obtained on chro-matography of whole synovial fluid and in both ex-periments the overall recovery of activity was about80%. On rechromatography the active peaks elutedin their original positions. Inhibition studies on theenzymic activity in the two peaks showed that itwas inhibited by EDTAand not Dip-F.

    Because ultracentrifugation had shown that a majorpart of the enzymic activity was of buoyant density

  • 75-0.0

    E 50

    ~25-

    0/1.1 1.2 1.3

    density g/mI

    FIGURE 8 Ultracentrifugal flotation of nitroanilide hydrolaseactivity fractionated from rheumatoid synovial fluid. Thepercentage of enzyme buoyant (in the top third of the tube)is plotted against the density at which centrifugation wasperforned. 0, the void volume peak from Sepharose 4Bfractionation of rheumatoid synovial fluid (see Fig. 8) and0, the second enzyme peak from Sepharose 4B. 10-ml tubeswere centrifuged at 50,000 rpm for 20 h at 15°C.

    part (95%) of the nitroanilide hydrolase activity wasunlikely to be due to leukocyte elastase or indeedany proteinase bound to a2M. Although inhibitionstudies of crude material such as synovial fluid canbe subject to error, the failure of Dip-F to causeany significant inhibition after preincubations up to10 h, as well as its failure to inhibit the enzymeactivity in either peak from Sepharose 4B chromatog-raphy, was very strong evidence against the activitybeing due to any serine proteinase such as leukocyteelastase. Furthermore Dip-F failed to inhibit the en-zyme eluted from the top of the polyacrylamidegel after electrophoresis.

    The ultracentrifugal flotation experiments showedthat the two forms of the enzymic activity separatedby Sepharose 4B chromatography were both associatedwith some lipid. The larger form contained more lipidthan the smaller form since it had a lower buoyantdensity. It is not known whether the two forms con-tained the same enzyme. The large form was unlikelyto be a simple aggregate of the small form becauseit had a different buoyant density. It is possiblethat the small form may have arisen from break-down of the large form. The interaction with lipidmay not have been fortuitous since it was not disruptedby high salt concentrations, which usually prevent non-specific association between protein and lipid. Al-though it was possible that the enzyme may have beenforming aggregates with plasma lipoproteins presentin synovial fluid, the buoyant densities and chromatog-raphic characteristics of both forms did not correspondto those of plasma lipoproteins. The large form behavedlike very low density lipoproteins or chylomicrons on

    Sepharose 4B chromatography but had a much higherdensity than either of these. The small form had ahigher density than high density lipoprotein and elutedslightly earlier from the Sepharose column. The largeform could well have been some type of membranefragment, perhaps arising from cell breakdown; itsbuoyant density was consistent with this suggestion.The small form could have been a low molecularweight enzyme aggregated with lipid or even a highmolecular weight lipoprotein enzyme.

    The activities of nitroanilide hydrolase were, withone exception, higher in the synovial fluid than inthe plasma so the enzymic activity probably arosewithin the joint cavity. Studies are in progress tosee whether similar activity can be found in thesynovial membrane or synovial fluid leukocytes.

    Amino acid and peptide nitroanilide substrates aregenerally hydrolyzed by proteinases and aminopep-tidases. The synovial fluid enzyme is unlikely to bean aminopeptidase since the terminal amino group ofSuc-(Ala)3-NPhNO2 is blocked. Therefore the en-zyme responsible is very likely to be a proteinase.No metal-dependent proteolytic activity was detectedin the enzyme-rich fractions prepared by ultracentri-fugal flotation, but this does not mean that the synovialfluid enzyme was not a proteinase. The enzyme mayhave been hindered from attacking a high molecularweight or powder substrate because of its associationwith other material or the substrate specificity of theenzyme may have been such that the substratesused were unsuitable.

    Identification of the enzyme as well as determinationof any role it might play in tissue destruction awaitits purification.

    ACKNOWLEDGMENTSI wish to thank Professor W. Watson Buchanan for allowingme to study his patients, Dr. J.-P Mathieu for supplyingme with so many samples of synovial fluid, and Dr. A. J.Barrett of the Strangeways Research Laboratory, Cambridge,for his help and criticism during revision of the manuscript.

    I am grateful to CIBA Research Laboratories Horsham,Sussex, England for their financial support.

    REFERENCES1. Dingle, J. T. 1962. Lysosomal enzymes and the degrada-

    tion of cartilage matrix. Proc. R. Soc. Med. 55: 109-111.2. Weissman, G. 1972. Lysosomal mechanisms of tissue in-

    jury in arthritis. N. Engl. J. Med. 286: 141-147.3. Janoff, A., G. Feinstein, C. J. Malemud, and J. M. Elias.

    1976. Degradation of cartilage proteoglycan by humanleukocyte granule neutral proteases-A model of jointinjury. I. Penetration of enzyme into rabbit articularcartilage and release of 35SO4-labeled material from thetissue. J. Clin. Invest. 57: 615-624.

    4. Keiser, H., R. A. Greenwald, G. Feinstein, and A. Janoff.1976. Degradation of cartilage proteoglycan by humanleukocyte granule neutral proteases-A model of joint

    800 J. Saklatvala

  • injury. II. Degradation of isolated bovine nasal cartilageproteoglycan.J. Clin. Invest. 57: 625-632.

    5. Janoff, A., and J. Scherer. 1968. Mediators of inflamma-tion in leukocyte lysosomes. IX. Elastinolytic activityin granules of human polymorphonuclear leukocytes.

    J. Exp. Med. 128: 1137-1155.6. Feinstein, G., and A. Janoff. 1975. A rapid method for

    purification of human granulocyte cationic neutral pro-teases; purification and further characterisation of humangranulocyte elastase. Biochim. Biophys. Acta. 403: 493-505.

    7. Gerber, A. C., J. H. Carson, and B. Hadorn. 1974.Partial purification and characterisation of a chymo-trypsin-like enzyme from human neutrophil leukocytes.Biochim. Biophys. Acta. 364: 103-112.

    8. Feinstein, G., and A. Janoff. 1975. A rapid methodfor purification of human granulocyte cationic neutralproteases: purification and characterisation of humangranulocyte chymotrypsin-like enzyme. Biochim. Bio-phys. Acta. 403: 477-492.

    9. Lazaris, G. S., J. R. Daniels, R. S. Brown, H. A. Bladen,and H. M. Fullmer. 1968. Degradation of collagen bya human granulocyte collagenolytic system. J. Clin. In-vest. 47: 2622-2629.

    10. Ohlsson, K., and I. Olsson. 1973. The neutral proteinasesof human granulocytes. Isolation and partial characterisa-tion of two granulocyte collagenases. Eur. J. Biochem.36: 473-481.

    11. Ohlsson, K. 1975. a,-antitrypsin and a2-macroglobulin in-teractions with human neutrophil collagenase and elas-tase. Ann. N.Y. Acad. Sci. 256: 409-419.

    12. Mehl, J. W., W. O'Connell, and J. DeGroot. 1964.Macroglobulin from human plasma which forms an enzy-matically active compound with trypsin. Science (Wash.D. C.). 145: 821-822.

    13. Barrett, A. J., and P. M. Starkey. 1973. The interac-tion of a2-macroglobulin with proteinases. Characteristicsand specificity of the reaction, and a hypothesis concern-ing its molecular mechanism. Biochem. J. 133: 709-724.

    14. Bieth, J., B. Spiess, and C. G. Wermuth. 1974. Thesynthesis and analytical use of a highly sensitive andconvenient substrate of elastase. Biochem. Med. 11:350-357.

    15. Pryce-Jones, R. H., J. Saklatvala, and G. C. Wood. 1974.Neutral proteinase from the polymorphonuclear leuco-cytes of human rheumatoid synovial fluid. Clin. Sci.Mol. Med. 47: 403-414.

    16. Saklatvala, J. 1975. The proteinase inhibitors of humanplasma. Ph.D. Thesis. University of Strathclyde, Glas-gow, Scotland.

    17. Erlanger, B. F., N. Kokowsky, and W. Cohen. 1961.The preparation and properties of two new chromogenicsubstrates of trypsin. Arch. Biochem. Biophys. 95: 271-278.

    18. Saklatvala, J., G. C. Wood, and D. D. White. 1976.Isolation and characterisation of human a,-proteinase in-hibitor and a conformational study of its interactionwith proteinases. Biochem. J. 157: 339-351.

    19. Chamey, J., and R. M. Tomarelli. 1947. A colorimetricmethod for the determination of the proteolytic activityof duodenal juice. J. Biol. Chem. 171: 501-505.

    20. Pryce-Jones, R. H., and G. C. Wood. 1975. Purificationof granulocyte neutral proteinase from human blood andrheumatoid synovial fluid. Biochim. Biophys. Acta. 397:449-458.

    21. Havel, R. J., H. A. Eder, and J. H. Bragdon. 1955.The distribution and chemical composition of ultra-centrifugally separated lipoproteins in human serum. J.Clin. Invest. 34: 1345-1353.

    Hydrolysis of the Elastase Substrate Suiccinyltrialanine Nitroanilide 801