Immunoimmobilization of α2-macroglobulin-β-trypsin complexes: A novel approach for the biochemical characterization of modulator-protease interactions

ANALYTICAL

BIOCHEMISTRY

108,

166- 175 (1980)

lmmunoimmobilization of a,-Macroglobulin-/I-Trypsin Complexes: A Novel Approach for the Biochemical Characterization of Modulator-Protease Interactions1 PETER C. HARPEL AND MELVIN Division

of HematologylOncology,

Department New

of Medicine, The New York, New York 10021

B. HAYES York Hospital-Cornell

Medical

Center,

Received April 7, 1980 A new approach to the biochemical characterization of a,-macroglobulin-protease complexes is described based upon our observation that cY,-macroglobulin ((YAM) or oZM-/3trypsin complexes become quantitatively bound to anti-human alM antibody that is covalently coupled to an agarose gel matrix. The immunoimmobilized qM-Ptrypsin complexes hydrolyze the chromogenic tripeptide substrate, N-benzoyl-t-phenylalanylt-valyl-t-arginine-p-nitroanilide HCI and the amidolytic activity of these complexes is proportional to the concentration of bound ptrypsin. The K, and V,,, of the soluble complexes are identical to those of the solid-phase a,M-Strypsin complexes indicating that immunoimmobilization does not alter the hydrolytic capacity of the complex. Optimal pH of the assay was 8-9, at a NaCl concentration of 0.05-0.3 M NaCI. The immobilized enzyme complex was remarkably stable since it retained its amidolytic activity following repeated assays. Varying concentrations of 1311-/3-trypsin were added to plasma containing a trace amount of ‘251-(r,M. The cY,M-Ptrypsin complexes that formed were quantitatively recovered from the plasma by the immunoimmobilization technique as determined by radioactivity and amidolytic activity measurements. Fifty-eight percent of the fi-trypsin added to plasma was bound to (YAMas demonstrated by both the immunoimmobilization assay and by molecular sieve chromatography. Thus a novel approach has been developed for rapid isolation and concentration of cu,M-protease complexes and for the characterization of their catalytic potential.

The a,-macroglobulin (cx,M)~ is a plasma protein that forms a complex with a wide variety of proteolytic enzymes (1,2). In contrast to other naturally occurring protease inhibitors, the cu,M does not bind to the active site of the enzyme with which it reacts. This results in variable inhibition of the activity of the enzyme against high molecular weight protein substrates, and

relatively little inhibition of hydrolytic activity as assayed utilizing low molecular weight, synthetic substrates (3,4). The role of a,M in modulating proteolytic enzymes involved in hemostatic and inflammatory reactions is unclear although it has been established that the azM forms complexes with enzymes of the coagulation, fibrinolytic, and kinin-forming pathways (I ,2). Part of the problem in establishing a physiologic function for this protein is undoubtedly due to difficulties in isolating (YAM from biological fluids and in assaying the catalytic activity of its bound enzyme. The present study details our observations that purified cu,M as well as CQMtrypsin complexes are quantitatively bound

1 This work was supported in part by U. S. Public Health Service Grant HL-18828 (Specialized Center of Research in Thrombosis). * Abbreviations used: a*M, cu,-macroglobulin; BzPhe-Val-Arg-NHNp, N-benzoyl-L-phenylalanyl-Lvalyl-L-arginine-p-nitroanilide HCI; TLCK, a-N-ptosyl-L-lysine chloromethyl ketone HCl; PBS, phosphate-buffered saline. 0003-2697/80/150166-10$02.00/O Copyright All rights

0 1980 by Academic Press, Inc. of reproduction in any form reserved.

166

a,-MACROGLOBULIN-TRYPSINCOMPLEXES

to rabbit anti-human (YAM antibody that is immobilized on a gel matrix. These studies demonstrate that the antibody bound, crzMtrypsin complexes possess amidolytic activity that can be readily assayed while attached to the particulate gel. The enzymatic activity of these insolubilized CQMtrypsin complexes is identical to that of fluid phase cr,M-enzyme complexes. These studies, utilizing purified systems, have been extended to human plasma to which radiolabeled trypsin and a trace amount of radiolabeled a2M are added. Using the insoluble antibody technique developed in this study, a,M-trypsin complexes are quantitatively recovered from plasma as measured both by bound radioactivity and by the capacity of the antibody-bound complex to hydrolyze a synthetic tripeptide chromogenic substrate. Thus a solid-phase immunologic capture technique has been developed that can specifically isolate +Menzyme complexes from the fluid phase and quantitate their hydrolytic potential. MATERIALS Puri’cation

AND METHODS

of a2-Macroglobulin

cr,-Macroglobulin was isolated from fresh human plasma in the presence of soybean trypsin inhibitor as previously described (5). azM was labeled with lz51by the method of McFarlane (6). Carrier-free sodium [lz51]iodide (0.5 mCi) was added to 1.7 mg a,M in 1.0 ml borate buffer (0.2 M), pH 8.0, containing NaCl (0.16 M). Iodine monochloride (0.05 ml containing 0.007 pmol ICI) was added with mixing. After a IO-min incubation at room temperature, the unbound iodide was removed by Sephadex G-25 (Pharmacia) gel filtration chromatography. The specific activity of the radiolabeled a,M was 0.2 FCilpg. Purification

of PTrypsin

/3-Trypsin was prepared from lized, dialyzed salt-free lyophilized

crystalbovine

167

trypsin (Worthington) as described by Yung and Trowbridge (7). The specific activity of the final product was 94% as determined by active site titration with p-nitrophenylp’-guanidinobenzoate HCl (8). This preparation, following reduction, consisted of a single protein band as identified by dodecyl sulfate gel (9% acrylamide) electrophoresis (9) indicating that the single-chain /3 form had been isolated from the original mixture of (Y- and P-trypsin and inactive material. The trypsin was stored at -70°C in HCI (1 mM) containing CaCl, (10 mM) and NaCl (0.1 M). @Trypsin was labeled with 1311 by the method of McFarlane (6). P-Trypsin was dialyzed against borate buffer (0.2 M), pH 8.0, containing benzamidine (0.01 M, Aldrich Chemical Co.), and NaCl (0.16 M). Carrier-free sodium [1311]iodide (0.5 mCi) was added to 1.5 mg /3-trypsin in a total volume of 1.0 ml borate-benzamidine buffer. Iodine monochloride (0.05 ml containing 0.1 pmol ICI) was added with mixing. After a lo-min incubation at room temperature the free iodide and benzamidine were removed by gel filtration chromatography. The labeled preparation was dialyzed against HCl (1 mM) containing NaCl (0.1 M) prior to storage at -70°C. The specific activity of the radiolabeled /3-trypsin was 0.2 $3/pg. Preparation of Rabbit Antihuman CY,M Antisera

New Zealand white rabbits were immunized by intradermal injection of azM mixed with Freund’s adjuvant. The antisera produced one immunoprecipitation arc on double diffusion analysis against human plasma and demonstrated a reaction of identity with the starting antigen. The IgG fraction of the antisera was prepared by chromatography on DEAE-cellulose. This IgG fraction was found to hydrolyze Nbenzoyl-L-phenylalanyl-L-valyl-L-argininep-nitroanilide HCl (BzPhe-Val-Arg-NHNp; S-2160), however, the activity was inhibited by treatment of the IgG material

168

HARPEL

with a-N-p-tosyl-L-lysine chloromethyl ketone HCI (TLCK) (0.01 M) for 3 days at 4°C followed by extensive dialysis. The TLCK-treated IgG was coupled to Bio-Gel A-5m (Bio-Rad Laboratories) by the cyanogen-bromide method as detailed by March et al. (10). Coupling to the activated gel was carried out in citrate buffer (0.2 M), pH 6.5, using 6.0 mg IgG/ml activated gel. Coupling efficiency was greater than 90%. After the coupling procedure, the gel was incubated 1 h in 1.0 M enthanolamine, pH 8.0, to neutralize any remaining protein binding groups. The IgG fractions of normal rabbit serum and rabbit antisera directed against human albumin or human haptoglobin were also prepared and coupled to Bio-Gel A-Sm as detailed above. Studies of the Binding of cllzM to Immobilized Rabbit Anti-Human ILQM Antibody The gel containing the bound LY,M antibody was diluted twofold (v/v) with phosphate buffer (0.05 M), pH 7.2, containing NaCl (0.1 M) (PBS). This coupled antibody was incubated with 1251-(r,M, ‘251-~2M/3-trypsin complexes, or with plasma containing a trace quantity of 1251-~2M and varying concentrations of 1311-p-trypsin as indicated in the figure legends. All incubations were at room temperature, in volumes of 1.0 ml or less, with constant mixing by end-over-end inversion using a Labquake mixer (Labindustries). The incubation was terminated by centrifugation and the pelleted insoluble a2M antibody gel washed repeatedly with l.O-ml portions of PBS. The pellets were counted for associated 1251 or 1311 radioactivity in a Searle 1185 dual channel gamma counter. Assay of the Amidolytic Activity of a,M-PTrypsin Complexes in the Fluid Phase or Bound to Immobilized a2 M Antibody Soluble trypsin or ar,M-trypsin complexes were assayed by methods similar to

AND

HAYES

those previously detailed (11). The substrate N-benzoyl-L-phenylalanyl-L-valyl-L-arginine-p-nitroanilide hydrochloride (BzPheVal-Arg-NHNp, S-2160, obtained from Ortho Diagnostics, Inc., or from Vega Biochemicals) was dissolved in distilled water (0.7 mg/ml). /3-Trypsin or a2Mtrypsin complexes were made to a volume of 0.4 ml with Tris-HCI (0.1 M), pH 8.3, containing CaCI, (1.25 mM). Substrate (0.2 ml) was added and the mixtures were incubated at room temperature. The reaction was terminated at varying intervals by the addition of 30% acetic acid (0.2 ml) followed by the addition of 0.6 ml of the Tris-CaCl, buffer to achieve a final volume of 1.4 ml. The absorbance at 405 nm was measured with a Gilford 240 spectrophotometer. The concentration of p-nitroanilide released was determined using a molar absorbancy of 10,500 (12). To determine the amidolytic activity of %M-trypsin complexes bound to gel-coupled rabbit anti-human CQM IgG, Tris-CaCl, buffer (0.4 ml) and BzPhe-ValArg-NHNp (0.2 ml) were added to the immobilized antibody gels that had been incubated with cr,M-trypsin complexes as indicated in the figure legends. After incubation for varying time periods at room temperature, the antibody-containing gel was removed by centrifugation. Acetic acid (30%, 0.2 ml) was added to the supernatant, followed by the addition of 0.6 ml Tris-CaCl, buffer. The absorbance at 405 nm was measured, and the results were expressed (following correction for the substrate blank) either as micromoles of p-nitroanilide released per liter per minute, or when the amount of a2M bound was determined, as micromoles per liter per minute per milligram of (YAM. For the derivation of Michaelis-Menton constants, /3-trypsin (0.25 &ml), a2M (50 &ml), and a trace quantity of 1251-~2M were incubated at 0°C in PBS. Assay of this incubation mixture for free P-trypsin, using the high molecular weight particulate substrate Remazolbrilliant blue hide (4), proved

a,-MACROGLOBULIN-TRYPSIN

--o -

-20 FIG. 1. Lineweaver-Burk solid-phase antibody-bound complexes was determined duplicate and the points are

-10

0

169

COMPLEXES

Fluid-phasea*M-trypsin Solid-phaseazM-trypsin

10 1/[S], mM-’

20

M

40

50

plot of the hydrolysis of BzPhe-Val-Arg-NHNp by fluid-phase or by o,M-Ptrypsin complexes. The hydrolytic activity of (r2M-/3 trypsin as indicated under Materials and Methods. All determinations are in fitted by linear regression analysis.

negative and indicated that all of the ptrypsin had been bound by the c+M. Portions (0.2 ml) of the a,M-/?-trypsin incubation mixture were then incubated in duplicate with rabbit anti-a,M-IgG coupled to BioGel A-5m (10 ~1 packed gel). The duplicate gel incubation mixtures were washed four times each with 1.0 ml PBS and the lz51, a measure of c+.M bound to the gel, was determined. The insolubilized a,M-P-trypsin complexes were then assayed in duplicate for the ability to hydrolyze varying concentrations of BzPhe-Val-Arg-NHNp. Due to the fact that continuous kinetic measurements could not be made with the insolubilized cr,M-enzyme complexes, duplicate assays were terminated at discrete time intervals of 5, 10, and 15 min at each substrate concentration and the concentration of liberated p-nitroanilide was determined. Portions of the original soluble oZM-Ptrypsin mixtures, containing an amount of radioactivity equivalent to that bound to the antibody gels described above were assayed in duplicate at varying substrate concentrations. In this case, production of p-nitroanilide was followed continuously at 405 nm in a Gilford 240 recording spectro-

photometer equipped with a Honeywell 1800 recorder. K, and If,,,,, for the soluble and immobilized a,M-&trypsin complexes were obtained from Lineweaver-Burk linear transformations of the initial reaction velocities and substrate concentrations. Electroimmunoassay The concentration of native cu,M in plasma, or in plasma supernatants following incubation with immobilized CX~Mantibody was determined by the electroimmunoassay, rocket technique described by Laurel1 (13). Linearity of this assay was found to be in the range of 5 to 30 pg/ml afM. RESULTS

Amidolytic Activity of a,M-Trypsin Complexes in the Fluid Phase or Bound to Immobilized Rabbit Anti-Human c+M Antibody The hydrolytic activity of a,M-trypsin complexes was assessed using the chromogenic substrate BzPhe-Val-Arg-NHNp. /3-Trypsin was incubated with a mixture of unlabeled and lz51-labeled a,M at trypsin:

170

HARPEL

AND HAYES

oZM molar ratios of 0.6, 0.3, and 0.15, well below the 2 mol of /3-trypsin per mole a,M binding capacity of a,M established in a previous study.3 The amidolytic activity of these mixtures was found to be linear with time and proportional to the concentration of /3-trypsin in the system. The p-trypsinazM mixtures were also incubated with the IgG fraction of rabbit anti-human a,M antiserum coupled to Bio-Gel A-5m. After extensive washing the gel pellets were counted in a gamma counter for r*Y and were found to have bound 88% of the 1251a,M radioactivity of the original incubation mixture indicating that the majority of the lz51-o12M in the incubation mixture had been bound to the immobilized antibody. The solid-phase antibody in the absence of oZM-trypsin mixture did not hydrolyze BzPhe-Val-Arg-NHNp, but the solid-phase a,M antibody which was incubated with trypsin-a,M complexes had gained amidolytic activity that was proportional to the amount of trypsin in the original fluidphase incubation mixture. Incubation of the solid-phase antibody with P-trypsin in the absence of a,M did not confer amidolytic activity upon the gel nor did the immobilized IgG fraction derived from normal rabbit serum bind either cu,M or a,M-trypsin complexes. Thus, the activity gained when immobilized anti-human a,M antibody was incubated with the a,M-trypsin mixtures was a reflection of the specific binding of the cr,M-trypsin complexes. Analysis of the kinetics of the hydrolysis of substrate in the fluid-phase by cu2Mtrypsin complexes as compared to the solidphase complexes demonstrated no significant differences (Fig. 1). The Michaelis-Menton constants for the fluid-phase complexes as determined by Lineweaver-Burk plots were K, = 0.079 mM and V,,, = 1707 min-‘, and of the solid-phase a,M-trypsin complexes, K, = 0.078 mM and V,,, = 1700 3 M. Brower, P. Harpel, M. Hayes, 1980, manuscript in preparation.

min-*. This indicates that under the experimental conditions detailed for these assays, the binding of the a,M-trypsin complex to a surface did not affect its amidolytic activity. Effect of pH and Ionic Strength on the Amidolytic Activity of Solid-Phase Antibody-Bound ar,M-Trypsin Complexes

Both soluble a,M-trypsin complexes and solid-phase antibody-bound a,M-trypsin complexes demonstrated a similar peak of amidolytic activity between pH 8.0 and 9.0. At higher pH’s, there was a significant decrease in substrate hydrolysis. In other experiments, the effect of variations in sodium chloride concentration (0 to 0.5 M) on the catalytic activity of the immobilized cu,M-trypsin complex was studied. A broad unchanging peak of activity from 0.05 to 0.3 M NaCl was observed with a decrease in amidolytic activity at both 0 and 0.5 M NaCl. Both soluble and immobilized 02Mtrypsin complexes behaved in a similar manner. of Repetitive Assays on the Amidolytic Activity of Solid-Phase Antibody-Bound a,M-Trypsin Complexes

Effect

In order to assess the stability of the immobilized antibody-bound cr,M-trypsin complexes, four portions of the immobilized a2M antibody were incubated with an a,M-trypsin mixture containing a trace amount of 1251-(r2M. After repeated washings, the gels were counted to determine the amount of bound 02M. The insolubilized antibody a,M-trypsin complex was then assayed for BzPhe-Val-Arg-NHNp hydrolytic activity. This procedure was repeated four times. As illustrated in Table 1, the specific activity of the aZM-trypsin complex, bound to the immobilized antibody, did not change with repeated assays indicating that trypsin itself did not dissoci-

a,-MACROGLOBULIN-TRYPSIN TABLE

171

COMPLEXES

1

serial dilutions of plasma to which 1251-(r,M was added. The radioactivity bound to the insoluble antibody was measured. The native CQM in the starting plasma and the residual azM in the plasma supernatant Specific following absorption was also quantitated activity by electroimmunoassay. The results dem@Mb (/*mol/L/min/ Assay Activity onstrate that the binding of the 1251-~,M No. (~mol/L/min) mg a&f) (CL& added to plasma by the immobilized antibody 6.7 k 0.17c 117.9 2 5.3 1 0.79 k 0.03 parallels the binding of native plasma azM. 2 6.7 4 0.16 0.76 k 0.05 113.4 -c 5.4 Specificity of binding of lz51-(r2M added 3 6.5 + 0.25 0.77 + 0.04 118.5 2 2.4 to plasma by the solid-phase oZM antibody 4 5.1 2 0.65 0.60 T 0.09 117.7 + 6.3 was also tested by incubating plasma cona Portions (0.2 ml) of a mixture of aZM and 1251-aZM taining 1251-a,M with unmodified Bio-Gel (50 &ml) and Ptrypsin (300 ng) were incubated in A-5m, or this same gel to which the purified quadruplicate for 2 h with rabbit anti-human ol,M-IgG IgG fraction of normal rabbit serum, or the coupled to Bio Gel A-5m (10 ~1 packed gel). Following IgG fraction of rabbit antisera against either washing of each sample four times with PBS, it was subsequently assayed for amidolytic activity by adding human albumin or haptoglobin was coupled. EFFECT OF REPETITIVE ASSAYS ON THE AMIDOLYTIC ACTIVITY OF INSOLUBLE ANTIBODY-BOUND (YAM-TRYPSIN COMPLEXES”

Tris-CaC1, buffer, pH 8.3 (0.4 ml), and BzPhe-ValArg-NHNp (0.7 a/ml, 0.2 ml) for 15 min at room temperature with mixing by inversion. After each assay, the samples were washed four times in succession and the amidolytic activity and radioactivity was redetermined. This assay procedure was repeated a total of four times. b The cu,M @g) bound to the beads was calculated as the counts bound/total counts x cu,M (w) in the original incubation mixture. c All values in the table represent the mean of four determinations.

ate from the complex in the presence of the chromogenic substrate. There was a small equivalent loss of both cu,M and amidolytic activity with repeated assays reflecting either loss of gel during the washings or elution of the cz,M-trypsin complex. Studies on the Binding of a2M in Human Plasma to Immobilized Rabbit Anti-Human a2M Antibody

Experiments were designed to test whether the binding of purified LY*M or a,Mtrypsin complexes by solid-phase antihuman czZM antibody could be extended to a complex biological fluid such as plasma. In the study shown in Fig. 2, the solidphase (YAM antibody was incubated with

“01

0

4

”

”

”

50 icn 150 a?-h4acroglohulin illglsysteml

1

2M

FIG. 2. Binding of OL*M in plasma by solid-phase %M antibody. Plasma containing 1.5 mg/ml a,M, as determined by electroimmunoassay, was diluted with PBS. A constant trace quantity of ‘251-~zM was added to each dilution (0.1 pg). The amount of plasma added per incubation system is indicated at the top of the figure. Portions (0.2 ml) were incubated in duplicate for 2 h with solid-phase asM antibody (10 ~1 packed gel). Following separation of the gel by centrifugation the supematant was assayed for cr,M antigen concentration by electroimmunoassay, and the rZ51-(r,M bound to the beads measured. The unlabeled (YAM from the plasma that was bound to the immobilized antibody was determined as that amount of cu,M originally added to the antibody gel minus the remaining ru,M in the supernatant as measured by electroimmunoassay. The counts bound are expressed as a percentage of the total counts originally added to the immobilized antibody.

172

HARPEL

Under conditions in which greater than 95% of the radioactivity became bound to insoluble (YAMantibody, less than 2% of the 1251-~2M radioactivity was associated with these other gel preparations. The effect of time of incubation of plasma with solid-phase CX*M antibody on the binding of (Y*M to the gel was investigated (Fig. 3). The antibody gel was incubated for varying time periods with three different dilutions of plasma which contained 12, 60, and 192 pg a2M, respectively. There was a progressive increase in binding throughout the 6-h incubation period at all concentrations of a2M. At the lowest concentration of cr,M (12 /.&system), 91.7% of the crzM was bound after 6 h. At the highest concentration of a2M, 49% or 94 pg of CY,M was bound. This indicates that 1.0 ml of the a2M antibody gel has the capacity to bind at least 9.4 mg (YAM.

AND HAYES

I

I

I

I

0.8 1.2 Tryprin (W/ml plasma)

I

I

1.6

FIG. 4. Binding of plasma %M-i3iI-trypsin complexes by solid-phase a,M antibody. Varying concentrations of i3’I-trypsin (as indicated on the abscissa) were added to a mixture of 1251-aZM and plasma (diluted l/4). Following a IO-min incubation, 0.2-ml portions were mixed for 2 h with solid-phase (Y*M antibody (50 ~1 packed gel). The pellets were harvested by centrifugation, washed four times with 1.0 ml PBS, and counted for Iz51 and r3rI in a dual channel gamma counter. The gels were then assayed for amidolytic activity utilizing the substrate BzPhe-Val- Arg-NHNp as indicated under Materials and Methods. The activity is corrected for the spontaneous activity associated with the plasma to which no trypsin was added. The points are fitted by linear regression analysis.

Studies on the Binding of a,M-Trypsin Complexes in Plasma by Solid-Phase a2M Antibody

-0

1

2

3 tkurs

4

5

6

FIG. 3. Effect of time of incubation on the bindittg of CQM in plasma by solid-phase CQM antibody. Plasma containing 1.5 mg/ml cu,M was diluted to 4, 20, and 64% with PBS containing a trace amount of rZ51-a2M (0.1 &system). Portions (0.2 ml) were incubated in duplicate with immobilized a,M antibody (10 ~1 packed gel) and the reaction stopped by centrifugation at the intervals indicated on the abscissa. After washing four times in 1.0 ml PBS, the pellets were counted for **51 in a gamma counter and the counts bound expressed as a percentage of the total counts originally added to the insoluble antibody. The amount of a,M contained in the plasma that was added to each incubation mixture is indicated.

The demonstration that a,M in plasma could be quantitatively removed by immobilized 02M antibody made it possible to study the binding of exogenous 131I-ptrypsin to azM in a plasma system. 1311-ptrypsin was added to plasma that contained 1.5 mg/ml CQM to give trypsin concentrations of 0.10 to 1.65 E.Lg/ml plasma. This represents a trypsin/cu,M molar ratio of 0.002 to 0.035. The plasma also contained a trace amount of 1251-azM. Portions of plasma were incubated with solid-phase a,M antibody for 2 h. The pellets were washed, counted for both lz51 and 1311 radioactivity, and the amidolytic activity was measured (Fig. 4). Whereas 86.7% of the 1251-(r2M in the incubation mixture was bound to the antibody containing gel, only 50.3% of ‘311-p-trypsin was similarly bound suggesting that trypsin also binds to other

oZ-MACROGLOBULIN-TRYPSIN

plasma proteins. The percentage binding of both aPM and trypsin was unaffected by the varying concentrations of trypsin added to the plasma. The 1311radioactivity bound to the solid-phase antibody was proportional to the amount of 1311-trypsin in the original mixture. To control for endogenous amidolytic activity, the solid-phase azM antibody was incubated with a plasma-1251-(r2M mixture to which no 1311-trypsin had been added. The resulting CY~Mantibody gel hydrolyzed 2.4 pmol substrate/liter/min/mg (YAM corresponding to the addition to plasma of approximately 142 ng .@ypsin/ml. This background activity was subtracted from the amidolytic activity of the gels incubated with the plasma-trypsin mixtures. The amidolytic activity of the a2M antibody gels incubated with the 1311-trypsin-plasma mixtures was proportional to the initial concentration of trypsin added, and parallel to the 1311counts bound to the gel (Fig. 4). Analysis of a 1311-PTrypsin-Plasma Mixture by Molecular Sieve Chromatography

173

COMPLEXES

1.0 I 0; 0.8 x L Y 0.6 Iio.4 F;

Fraction

FIG. 5. Gel filtration chromatography of a mixture of 1311-ptrypsin iz51-o,M, and plasma. To 1.0 ml plasma, 1251-a2M (36 pg) and 1311-/3-trypsin (13 pg) were added and incubated 5 min at room temperature. The mixture was applied to a 1.4 x 70-cm column of Sephacryl S-208 gel filtration material (Pharmacia) maintained at 4°C in 0.05 M Tris-HCI, pH 8.0, containing 0.16 M citrate and 0.1 M NaCl. Fractions (1.6 ml) were collected, counted for lz51 and 13*1in a gamma counter, and the absorbance at 280 nm was measured.

To further confirm the specificity of the insoluble a,M antibody, a,M-trypsin interaction and to identify the nature of the From a comparison of the binding of 1311-trypsin peaks from the gel filtration (YAMto the solid-phase (YAM antibody in the column, 0.6 ml of each peak fraction of experiment illustrated in Fig. 4 (86.7%), 1311-trypsin was incubated with the solidand the binding of 1311-trypsin (50.3%), it phase a2M antibody (0.1 ml packed gel). can be calculated that the total a2M in Following an 18-h incubation, 99% of the plasma should have bound 58.0% of the 1251-a2M and 94% of the 1311-trypsin of the trypsin added to the plasma. Thus it seemed a2M peak fraction 36 was bound to the likely that approximately 42% of the trypsin immobilized antibody. Less than 5% of may have been complexed to other plasma 1311-trypsin was bound to the antibody gel proteins and therefore unavailable to the in the incubation mixture of the second a2M. To verify this possibility, a mixture *3*I-trypsin peak (fraction 46). Therefore of 1311-trypsin, *251-~ZM, and plasma was most of the 1311-trypsin in the first high fractionated by molecular sieve chromamolecular weight peak was complexed with tography (Fig. 5). The major 1311-trypsin a,M. The failure of 1311-trypsin in the second peak, containing 57.4% of the total 1311 radioactive peak to bind to the a,M antibody counts, coeluted with the 1251-(r2M in the gel provides further evidence that trypsin first, high molecular weight protein peak. had reacted with a protein other than a2M. The second 1311-trypsin peak contained 40.3% of the total 1311 counts and eluted DISCUSSION earlier than the albumin peak at a position This report presents a new approach consistent with complex formation between to the study of the physiologic function of trypsin and a plasma protein.

174

HARPEL

human plasma q-macroglobulin. The method detailed in this study makes it possible to further characterize the protease binding capacity of a,M in both purified systems and in complex biological fluids. Our findings are based on the ability of azM to bind with proteolytic enzymes without inhibiting the catalytic potential of the enzymes’ active site (3,4,14). We have established that soluble cvzM and (Y*M-Ptrypsin complexes bind to rabbit antihuman 02M antibody that is immobilized on agarose gel. The solid-phase antibody-bound cuzMPtrypsin complexes retain their ability to hydrolyze the tripeptide chromogenic substrate BzPhe-Val-Arg-NHNp with Michaelis-Menton constants identical to those of the soluble complexes. This indicates that the binding of the a,M-trypsin complex to an immobilized antibody does not appreciably effect the amidolytic activity of the bound protease. The antibody-bound cu,M-trypsin complexes proved to be remarkably stable since they remain associated with the agarose gel matrix through repeated washings and assays of amidolytic activity. The finding that trypsin remains bound to azM in the presence of substrate provides further evidence that trypsin binds to aZ2Mat a site different from its catalytic site. Studies of the conditions required to optimize substrate hydrolysis showed that the optimum pH of the reaction for both the soluble a,M-trypsin complex as well as the solid-phase, antibody-bound complex was between 8 and 9. In contrast, Rindernecht et al. (15), using a different chromogenic substrate, N-carbobenzoxyglycylglycyl-L - arginine - 2 - naphthylamide HCl, found that catalysis was greatest at pH 10.0. They postulated that this degree of alkalinity was necessary to alter the microenvironment of the trypsin active site within the a,M-trypsin complex so that the substrate could be hydrolyzed at a rate similar to that of free trypsin. It is not clear whether the difference in pH optimum between their study and ours reflects the behavior of the

AND

HAYES

purified /3-trypsin used in the present study, the difference in substrate, or other factors. We have demonstrated that the binding of cll,M-trypsin complexes by immobilized CX~M antibody also occurs when the assay system is applied to human plasma. Prior studies by Haverback and colleagues (16) demonstrated that when trypsin was added to serum a complex was formed between trypsin and a serum protein that possessed an a,-electrophoretic mobility. Mehl and co-workers (14) provided direct evidence that asM was the protein responsible for this enzyme-binding activity. In the present study, when plasma containing a trace quantity of 1251-ar2M was incubated with immobilized a2M antibody, both radiolabeled and native a,M were bound. The electroimmunoassay technique of Laurel1 was used to establish that the binding of 1251-a2M to the solid-phase a2M antibody was proportional to the amount of unlabeled native (YAM bound. The addition of the radiolabeled antigen to plasma made it possible, therefore, to quantitate the binding of plasma a2M to the solid-phase antibody. The binding of plasma cr2M to the insoluble gel containing the antibody was specific since other immobilized antibodies such as anti-albumin, anti-haptoglobin, or the IgG fraction of normal rabbit serum did not bind significant amounts of plasma cr2M. To establish that the solid-phase antibody technique detailed in this study could detect a,M-enzyme complexes in a biological fluid such as plasma, varying concentrations of 13%p-trypsin were added to plasma that contained a trace quantity of 1251-a2M. Subsequent assay demonstrated that /3trypsin was bound to the immobilized (YAM antibody and that the amount of 1311-trypsin bound and the amidolytic activity associated with the solid-phase antibody was linearly related to the concentration of trypsin originally added to the plasma. Since control studies established that 1311/3-trypsin itself did not bind to the antibody gel, the trypsin activity that we observed

o,-MACROGLOBULIN-TRYPSIN

was a specific measure of a,M-trypsin complexes. At each concentration of trypsin added to plasma, 58% was bound to azM, a finding explained by gel filtration studies that demonstrated that the remainder of the trypsin was associated with another protein, possibly cY,-antitrypsin. These results parallel the findings of Ganrot (17) who demonstrated that the affinity of trypsin for azM was greater than that for other trypsin inhibitors in plasma. a,M is a unique plasma protease inhibitor in that it forms a complex with an exceptionally wide variety of proteases including serine, thiol, carboxyl, and metalloproteases derived from plasma, cells, or from microorganisms (1,2). Complexes between QM and several different proteases have been identified in plasma (3,18), in peritoneal (19)) pleural (20,2 l), and in synovial effusions (22-24), however, a convenient method suitable for the quantitative analysis of these complexes has not previously been available. The solid-phase a,M antibody technique described in this study facilitates the rapid isolation and concentration of cY,M-protease complexes and the measurement of their catalytic capacity in biological fluids in various physiologic and pathologic states. Studies in progress indicate that modifications of the assay conditions will permit the measurement of the activity of physiologically important proteases such as thrombin, plasmin, and plasma kallikrein when bound to a,M.

COMPLEXES

3. 4.

5. 6. 7. 8.

9.

10. 11. 12. 13. 14. 15. 16. 17. 18.

ACKNOWLEDGMENTS

19.

The studies in this report were performed with the expert technical assistance of Mr. T. S. Chang and Mrs. Maureen M. Kraics.

20.

REFERENCES

21.

1. Harpel, P. C., and Rosenberg, R. D. (1976) Progress in Hemostasis and Thrombosis (Spaet, T. H., ed.), Vol. 3, pp. 145-189, Grune & Stratton, New York. 2. Starkey, P. M., and Barrett, A. J. (1977) Proteinases in Mammalian Cells and Tissues (Barrett,

22. 23. 24.

175

A. J., ed.), pp. 663-696, Elsevier/North-Holland, Amsterdam. Harpel, P. C., and Mosesson, M. W. (1973) J. Clin. Invest. 52,2175-2184. Hayes, M. B., and Harpel, P. C. (1979) The Physiological Inhibitors of Blood Coagulation and Fibrinolysis (Collen, D., Wiman, B., and Verstraete, M., eds.), pp. 273-280, Elsevieri North-Holland, Amsterdam. Harpel, P. C. (1976) in Methods in Enzymology (Lorand, L., ed.), Vol. 45, pp. 639-652, Academic Press, New York. McFarlane, A. S. (1956) Biochem. .Z. (London) 62, 135- 143. Yung, B. Y., and Trowbridge, C. G. (1975) Biothem. Biophys. Res. Commun. 65,927-930. Chase, T., Jr., and Shaw, E. (1970) in Methods in Enzymology (Perlmann, G., and Lorand, L., eds.), Vol. 19, pp. 20-27, Academic Press, New York. Weber, K., Pringle, J. R., and Osbom, M. (1972) in Methods in Enzymology (Hirs, C. H. W., and Timashiff, eds.), Vol. 26, pp. 3-27, Academic Press, New York. March, S. C., Parikh, I., and Cuatrecasas, P. (1974) Anal. Biochem. 60, 149-152. Svendsen, L., Blomback, B., Blomback, M., and Olsson, P. I. (1972) Thromb. Res. 1, 267-278. Aurell, ‘L., Friberger, P., Karlsson, G., and Claeson, G. (1977) Thromb. Res. 11, 595-609. Laurel], C. B. (1972) &and. J. C&z. Lab. Invest. 29, (Suppl. 124), 21-37. Mehl, J. W., O’Connell, W., and DeGroot, .I. (1964) Science 145, 821-822. Rinderknecht, H., Fleming, R. M., and Geokas, M. C. (1975) Biochem. Biophys. Acta 377, 158- 165. Haverback, B. J., Dyce, B., Bundy, H. F., and Wirtschafter, S. K. (1962) J. Clin. Invest. 41, 972-980. Ganrot, P. 0. (1%7) Ark. Kemi Bd. 26(50), 577-582. Nilehn, J. E., and Ganrot, P. 0. (1967) Stand. J. C/in. Lab. Invest. 20, 113-121. Ohlsson, K. (1974) Bayer Symposium V, Proteinase Inhibitors (Fritz, H., Tschesche, H., Green, L. J., and Truscheit, E., eds.), pp. 96- 105, Springer-Verlag, Berlin. Bieth, J., and Metais, P. (1%8) Clin. Chim. Acta 22,639-642. Bieth, J., Barth, J-G., and Metais, P. (1970) Clin. Chim. Acta. 30, 621-626. Shtacher, G., Maayan, R., and Feinstein, G. (1973) Biochim. Biophys. Acta 303, 138-147. Abe, S., and Nagai, Y. (1973) .I. Biochem. (Tokyo) 71, 919-922. Abe, S., and Nagai, Y. (1973) J. Biochem. (Tokyo) 73, 897-900.