INFEcIION AND IMMUNITY, May 1979, p. 411421

Vol. 24, No. 2

0019-9567/79/05-0411/11$02.00/0

Purification and Characterization of a Serratia marcescens Metalloprotease DAVID LYERLY AND ARNOLD KREGER* Department of Microbiology and Immunology, Bowman Gray School of Medicine of Wake Forest University, Winston-Salem, North Carolina 27103

Received for publication 15 January 1979

An extracellular, nonelastolytic, neutral metalloprotease of Serratia marcescens was purified by sequential ammonium sulfate precipitation, hydroxyapatite adsorption chromatography, flat-bed isoelectric focusing, and Sephadex G-100 gel filtration. The protease preparation had a 280/260 nm absorbance ratio of 1.8, was free of detectable amounts of endotoxin, carbohydrate, phosphorus, and other known extracellular enzymes of S. marcescens, and was homogeneous by Ouchterlony double immunodiffusion and Grabar-Williams immunoelectrophoresis. Crossed immunoelectrophoresis, thin-layer electrofocusing in polyacrylamide gel, and polyacrylamide disc gel electrophoresis showed three to four closely migrating, Coomassie blue-staining components in the protease preparation. However, zymogram analyses of the patterns showed that protease activity was associated with each component and that the protease was, therefore, microheterogeneous. The isoelectric point and sedimentation coefficient of the protease were approximately 5.3 to 5.4 and 4.2S, respectively, and the molecular weight estimated by sodium dodecyl sulfate-polyacrylanide gel electrophoresis and by gel filtration was approximately 52,500 and 44,000, respectively. The pH optimum range, with azocasein as the substrate, was 5.5 to 7.5. The enzyme contained a high percentage of acidic amino acids, no cysteine, and 1 g-atom of Zn2' and 7 g-atoms of Ca2+ per mol. Various heavy metal ions and chelating agents and heating at 600C for 15 min inactivated the enzyme. Intracorneal, intratracheal, and intradermal administration of the protease into rabbits elicited rapid and extensive tissue damage. The minimum lethal intravenous dose for mice was approximately 17 mg/kg of

body weight. Serratia marcescens was, for many years, believed to be avirulent but is now known to be an important opportunistic pathogen capable of causing a large variety of infectious diseases in humans (2, 9, 12, 14, 17, 27, 31, 34, 35, 45). In addition to the toxic lipopolysaccharide in its cell wall, the bacterium produces extracellular proteases, and serratia protease preparations of undemonstrated homogeneity have been reported to elicit skin, cornea, and lung damage (8, 25; A. S. Kreger, Abstr. Annu. Meet. Am. Soc. Microbiol., M288, p. 114, 1974) and to induce chemotaxis of polymorphonuclear leukocytes in vitro by generating chemotactic fragments from the C3 and C5 components of complement (52). These studies suggest the possibility that, during serratia-induced diseases, serratia proteases may directly cause tissue damage and/or elicit the release of tissue-damaging enzymes from polymorphonuclear leukocytes. However, at the present time, the role of serratia proteases in the pathogenesis of serratia-induced

diseases has not been critically evaluated or clearly demonstrated. Definitive studies to determine the role of a bacterial product in the development of a bacterial-induced disease process require a preparation of the product which is free of contaminating medium constituents and other biologically active products of the bacterium. Various investigators have described procedures for obtaining purified preparations of serratia proteases (1, 22, 25, 30, 36, 37, 42, 44; E. A. Broussard II, H. D. Braymer, and A. D. Larson, Bacteriol. Proc., p. 129, 1968; A. D. Larson et al., U.S. Patent 3,692,631, 1972); however, the publications did not present data rigorously documenting the purity of the preparations by a battery ofthe currently available, highly resolving, physicochemical, immunological, and biological activity techniques. Our interest in evaluating the possible role of serratia proteases in the pathogenesis of S. marcescens keratitis and pneumonia, therefore, prompted us to develop a purifi411

412

LYERLY AND KREGER

cation scheme for obtaining a S. marcescens protease preparation which was free of contaminating substances detectable by the above-mentioned techniques. This paper presents the results of our studies with an extracellular, tissuedamaging, neutral metalloprotease of S. marcescens, strain BG. MATERIALS AND METHODS Assays. Protease activity against azocasein was assayed as previously described (23) for Pseudomonas aeruginosa proteases. The final purified protease preparation was examined (approximately 100 ,ug of protein tested) for the presence of elastase (39), hexapeptidase (18), phospholipase C (26), alkaline phosphatase (16), deoxyribonuclease (41), lipase (48), esterase (21, 32), and collagenase (40) activities. The methods used are those described in the appropriate references. Reference enzymes and substrates were obtained from the Sigma Chemical Co. (St. Louis, Mo.), Worthington Biochemical Corp. (Freehold, N.J.), and Schwarz/Mann (Orangeburg, N.Y.). The final purified protease preparation (100 ug) was examined for the presence of endotoxin by the limulus amebocyte lysate method with commercially available lysate (Pyrostat; Worthington Biochemical Corp., Freehold, N.J.) as recommended by the manufacturer. The final purified protease preparation (1 mg) was examined for the presence of hexoses (13), pentoses (11), and phosphorus (10), with glucose, ribose, and beta-glycerophosphate, respectively, as standards. The absorbance, at 280 nm and 260 nm, of a 1-mg/ ml solution (in 0.1 M ammonium bicarbonate [AB]) of the purified protease was measured with 10-mm lightpath cuvettes in a Beckman DB-GT spectrophotom-

eter. Protein was estimated by the method of Lowry et al. (29), with crystalline bovine serum albumin as the standard. Purification of protease. Unless otherwise noted, all steps were done at approximately 40C. Stage 1. Culture supernatant fluids. S. marcescens BG, obtained from the culture collection of the Department of Microbiology and Immunology of The Bowman Gray School of Medicine, was cultivated in tryptone-yeast extract-glucose broth (pH 7.2) containing 0.5% tryptone (Difco), 0.25% yeast extract (BBL), and 0.1% glucose. Aliquots (0.1 ml) of a stationaryphase culture of the bacterium were added to 24 2liter flasks containing 200 ml of broth, and the broth was incubated for 20 to 24 h at 300C on a gyratory shaker (model G-25, New Brunswick Scientific Co., New Brunswick, N.J.) operating at 200 to 210 cycles/ min. The culture supernatant fluids were obtained by

centrifugation. Stage 2. Dialyzed concentrate of ammonum sulfate precipitate. Ammonium sulfate (Schwarz/ Mann, enzyme grade) was added slowly to the pooled culture supernatant fluids, with gentle stirring, to a final concentration of 60% saturation (420 g/liter). After 18 to 24 h, the precipitate was recovered by centrifugation, washed once with 60% saturated am-

INFECT. IMMUN. monium sulfate (100 ml), and dissolved in 5 ml of 0.02 M potassium phosphate buffer (pH 7.0). The preparation was dialyzed overnight against 8 liters of phosphate buffer, and a small amount of insoluble residue was removed by centrifugation. Stage 3. Dialyzed concentrate of hydroxyapatite adsorption chromatography pool. The stage 2 preparation was applied to a column (2.6 by 31 cm) of hydroxyapatite (Bio-Gel HTP, Bio-Rad Laboratories, New York, N.Y.) equilibrated with 0.02 M potassium phosphate buffer (pH 7.0). The column was washed, at a flow rate of 20 ml/h or 80 ml/h, with approximately 600 ml of equilibrating buffer, and fractions (10 ml) were collected. After washing the column, a single-step elution was performed, at a flow rate of 20 ml/h, with 0.3 M potassium phosphate buffer (pH 7.0), and fractions (5 ml) were collected. All fractions were assayed for absorbance at 280 nm and for protease activity, and fused rocket immunoelectrophoretic analysis of the fractions was performed. The protease peak fractions were pooled and concentrated to 15 ml by ultrafiltration in a stirred cell equipped with a PM 10 membrane (Amicon Corp., Lexington, Mass.), and the membrane was washed twice with 5 ml of 0.01 M tris(hydroxymethyl)aminomethane (Tris)-hydrochloride buffer (pH 7.5). The pool of the concentrate and the membrane washings was dialyzed overnight against 6 liters of 0.01 M Tris-hydrochloride buffer (pH 7.5). Stage 4. Concentrate of flat-bed isoelectric focusing pool. The stage 3 preparation was subjected to preparative electrofocusing in a flat bed of washed Sephadex G-75, superfine (Ultrodex; LKB Instruments, Inc., Rockville, Md.). The procedure was performed with an LKB 2117 Multiphor electrophoresis apparatus and flat-bed electrofocusing kit, using a modification of the methodology of Winter et al. (54). A 4% (wt/vol) gel slurry containing the stage 3 preparation and the ampholine buffer mixture required to generate a linear pH gradient of 4.0 to 6.5 (2.5 ml each of pH 4 to 6 and 5 to 7 ampholines) was dried, at 10 to 12°C, to the required evaporation limit. Electrofocusing was performed for 18 to 19 h at a constant power of 7 W provided by the LKB 2103 power supply. The pH of odd-numbered gel fractions was determined with a surface glass electrode (Ingold Electrodes, Inc., Lexington, Mass.), and the supernatant fluids derived from mixing each gel fraction with 4 ml of 0.1 M AB were assayed for protease activity and analyzed by fused rocket immunoelectrophoresis. The protease peak fractions were pooled and concentrated to 5 ml by ultrafiltration with a PM 10 membrane, the membrane was washed twice with 2.5 ml of 0.1 M AB, and the concentrate and membrane washings were pooled. Stage 5. Sephadex G-100 gel filtration pool. The stage 4 preparation was applied to a column (2.6 by 96 cm) of Sephadex G-100 (regular) equilibrated with 0.1 M AB (pH 7.8), and was eluted, in the downward flow mode, at a flow rate of 20 ml/h. Fractions (5 ml) were assayed for absorbance at 280 nm and for protease activity, and were analyzed by fused rocket immunoelectrophoresis. Protease peak fractions were pooled and lyophilized, and the lyophilized preparation was stored at 0°C.

VOL. 24, 1979

S. MARCESCENS METALLOPROTEASE

Preparation of antisera. New Zealand white rabbits, weighing approximately 1.4 to 2.3 kg at the start of the vaccination schedule, were injected subcutaneously, every 3 weeks, with 1 ml of an anhydrous oil vaccine (20) containing 5 mg of lyophilized crude protease concentrate (stage 2) per ml of incomplete Freund adjuvant. Rabbits were bled aseptically 10 days after the 4th, 5th, and 6th injections, and the sera were pooled and lyophilized. A semiquantitative titration of the pooled antisera (7) indicated acceptable precipitating antibody titers (optimal proportions zone of 0.05 to 0.1% antigen). Immunological procedures. Ouchterlony double imnunodiffsion, Grabar-Williams immunoelectrophoresis (IEP), crossed IEP, and fused rocket IEP were performed with the LKB 2117 Multiphor apparatus and IEP kit. The general methodology presented in the LKB instruction manuals and application notes 85 and 249 and in the quantitative IEP manual of Axelsen et al. (6) was followed. Ouchterlony double immunodiffusion was performed in a 0.01 M boratesaline buffer system (pH 8.5). Results from preliminary immunological purity monitoring studies performed in the absence of protease inhibition suggested the occurrence of autodigestion, as well as the digestion of contaminating antigens and antigen-antibody complexes by the protease. Therefore, subsequent immunological monitoring was performed in the presence of a protease-inhibiting concentration of 1,10-orthophenanthroline (2 x 10' M OPA). Initially, immunodiffusion and Grabar-Williams IEP gels were kept at 4 and 250C and were examined daily for 4 days. Maintenance at 40C for 2 days gave the best resolution and the largest number of precipitin arcs in both systems. Fused rocket and crossed IEP gels were pressed, washed, stained, and destained as described in the quantitative immunoelectrophoresis manual of Axelsen et al. (6).

layer isoelectric focusing in polyacrylamide gel was performed with the LKB 2117 Multiphor electrophoresis apparatus and commercial PAG plates (pH 3.5 to 9.5) as recommended by the manufacturer, except that samples (15 pl) were applied directly to the gel surface (4 cm from the cathode) and were electrofocused at a maximum fixed wattage of 30 W and a maximum of 1,500 V and 50 mA for approximately 2 h at 40C. Gels were fixed, stained with Coomassie blue R-250, and destained as recommended by the manufacturer, except that the fixed gels were soaked for 15 min in destaining solution before staining. Gels were also examined by a modification of a protease zymogram technique (5). After electrofocusing, the gels were immersed for 5 min in 0.4 M Tris-hydrochloride buffer (pH 7.5), were overlaid with 1.2% molten agarose in 0.1 M Tris-hydrochloride buffer (pH 7.5) containing 1% sodium caseinate, and were incubated at 370C in a humid environment for approximately 1 h before fixing in a solution containing 3.5% sulfosalicylic acid, 11.5% trichloroacetic acid, and 30% (vol/vol) methanol. Molecular weight estimation. (i) Sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The purified protease (stage 5) and various reference proteins were denatured and subjected to electrophoresis in 12% monomer concentration gels containing 0.1% sodium dodecyl sulfate, as previously described for P. aeruginosa proteases (23). The molecular weight of the denatured protease was estimated by the relative mobility method of Weber et al.

Analytical polyacrylamide gel electrophore8i8. Commercially available, precast, 5.5-by-100-mm polyacrylamide gels (Bio-Phore gels; Bio-Rad Laboratories) having 7.5% and 12% monomer concentrations were adjusted to pH 8.9 by electrophoretic introduction, as recommended by the manufacturer, of 0.188 M Tris-glycine buffer. Samples (201d) of protease preparations mixed with Bio-Phore basic tracking dye solution were underlayered onto the surfaces of the gels maintained at 40C. Samples were initially migrated into the gels at a potential of 100 V, and the run was completed at a potential of 225 V. Gels were subsequently fixed, stained with Coomassie blue R250, and destained as described by Fairbanks et al. (15). In addition, the location of protease activity in the 7.5% monomer gels used to analyze the final purified protease preparation (stage 5) was determined by slicing the gels into 1.5-mm sections, placing each section into a protease assay mixture, and assaying for substrate degradation. The location of activity was compared with the location of Coomassie blue-staining bands in gels subjected to electrophoresis under identical conditions. Analytical isoelectric focusing. Analytical thin-

(53).

413

(ii) Sephadex gel filtration. The apparent molecular weight of the native protease was estimated by the method of Andrews (3). Estimation of sedimentation coefficient. The sedimentation coefficient of the purified protease (stage 5) was estimated by rate zonal centrifugation in a linear 5 to 20% (wt/vol) sucrose density gradient (33), as modified for P. aeruginosa proteases (24). Amino acid analysis. Amino acid analyses were performed by the method of Spackman et al. (46). Samples (2 mg) of the purified protease preparation were hydrolyzed for 24, 48, and 72 h at 1100C with 6 N HOl (1 ml) in sealed Pyrex tubes after evacuation and degasg. The hydrolysates were dried under nitrogen, and amino acid analyses were performed with a Durrum D500 amino acid analyzer. The tryptophan content was estimated by the fluorescence method of Pajot (43). Half-cystine was determined as cysteic acid by the performic acid oxidation method described by Moore (38). Metal ion analyses. A lyophilized preparation (75 mg) of the purified protease (stage 5), which was previously dialyzed against deionized-glass distilled water for 24 h at 40C, was digested with concentrated nitric acid (6 ml) and analyzed for Mg2e, Zn2+, C02+, Mn2+, and Fe3` by atomic absorption spectroscopy and for Ca2+ by flame emission spectroscopy using the Instrumentation Laboratories model 151 apparatus (Lexington, Mass.). Determination of pH optimum. Samples (0.2 U in 20 pd) of the purified protease preparation in 0.1 M Tris-hydrochloride buffer (pH 7.5) were added to assay

414

INFECT. IMMUN.

LYERLY AND KREGER

mixtures containing azocasein substrate solution (0.5 ml) and 1 ml of the different test buffers (0.1 M HC1KC1, pH 2; 0.1 M formic acid-NaOH, pH 3 and 3.5; 0.1 M acetate, pH 4, 4.5, 5, and 5.5; 0.1 M phosphate, pH 6 and 6.5; 0.1 M Tris-hydrochloride, pH 7, 7.5, 8, and 8.5; 0.1 M borate-borax, pH 9; 0.1 M cyclohexylaminopropane sulfonic acid-NaOH, pH 10). Protease activity was assayed as previously described. Corrected pH values of the assay mixtures at 370C were determined with a pH meter. Effect of metal ions and inhibitors on protease activity. Samples (0.2 U in 100 pl) of the purified protease preparation in 0.05 M Tris-hydrochloride buffer (pH 7.5) were added to 1 ml of the control buffer (0.1 M Tris-hydrochloride pH 7.5) and to the same buffer supplemented with a metal ion or inhibitor, and the mixtures were incubated at room temperature (250C) for 30 min. Azocasein substrate solution (0.5 ml) then was added to each mixture, and protease activity was assayed as previously described. In addition, the purified protease preparation (2 U of protease/ml) was incubated for 30 min at room temperature in 0.1 M acetate buffer (pH 5) supplemented with Na2 ethylenediaminetetraacetic acid (EDTA). Samples (100 ll) were added to azocasein assay mixtures, and protease activity was assayed as previously described. Reversibility of chelator-induced protease inhibition. Protease preparations inactivated completely with 2 x 10-' M OPA at pH 7.5 and with 5 x 10- M Na2 EDTA at pH 5 were dialyzed for 1 day, at 40C, against 8 liters of 0.1 M ammonium bicarbonate, and the dialyzed preparations were assayed, as described previously, for protease activity. Samples (100 pl) of protease preparations inactivated completely with 1.5 x 10-2 M OPA at pH 7.5 and with 5 x 10' M Na2 EDTA at pH 5 were added to 1 ml of Tris-hydrochloride buffer (pH 7.5) containing a sufficient concentration of metal ions to give a metal ion-to-inhibitor molar ratio, in the final assay mixture, of 2:1. Azocasein substrate solution (0.5 ml) then was added to the mixtures, and protease activity was assayed as described previously. Effect of heat and pH on protease stability. Samples (0.2 U) of the purified protease (stage 5) in 1.1 ml of 0.1 M Tris-hydrochloride buffer (pH 7.5) were incubated at 25, 37, 45, and 600C for 15 min. Azocasein substrate solution (0.5 ml) then was added to each mixture, and the protease activity was assayed as described previously. Solutions (10 U/ml) of the purified protease contained in the same buffers which were used in the pH optimum studies were incubated at room temperature for 1 h. Samples (20 ,il) of each mixture were added to azocasein assay mixtures, and protease activity was determined as described previously. In vivo toxicity studies. All the studies were done with membrane filter-sterilized solutions of the purified protease (stage 5) in 0.1 M AB. Heat-inactivated protease preparations served as controls. New Zealand White rabbits (1.4 to 2.3 kg) were anesthetized with ether and topical application of 0.5% tetracaine HCl and were injected intracorneally with samples (30 tl) of protease solutions. The corneas were examined for gross damage over a 4-h period.

Unanesthetized New Zealand White rabbits were injected intratracheally with samples (2 ml) of protease solutions. The animals were sacrificed with sodium pentobarbital 1 day postinjection, and the lungs were examined for gross damage. Rabbits were injected intradermally with samples (0.1 ml) of protease solutions. Visualization of increased blood vessel permeability was aided by the intravenous injection, 2 h after the intradermal injections, of a 5% solution of Evans blue in saline (2 ml). Random-bred, Dub:(ICR) female mice, weighing approximately 23 g (Flow Laboratories, Dublin, Va.), were injected intravenously with samples (0.25 ml) of protease solutions, and the minimum lethal dose was determined after observing the animals for 2 days postinjection.

RESULTS

Purification of S. marcescens protease. The behavior of the serratia protease during sequential hydroxyapatite adsorption chromatography, flat-bed isoelectric focusing, and gel filtration is shown in Fig. la, 2a, and 3a, respectively. The quantitative results of the purification scheme are summarized in Table 1. The ammonium sulfate precipitation step (stage 2) resulted in a major increase in specific activity. The other steps in the purification scheme did not produce a marked increase in the specific activity of the protease preparations; however, immunological and physicochemical analyses documented the importance of the other steps in the progressive removal of a large number of minor contaminants from the preparations (Fig. lb, 2b, 3b, 4, 5, 6, and 7). The final purified protease preparation (stage 5) was homogeneous by Ouchterlony double immunodiffusion (Fig. 4a) and by Grabar-Williams IEP (Fig. 4b). The more highly resolving techniques of crossed IEP, polyacrylamide gel electrophoresis, and analytical thin-layer isoelectric focusing in polyacrylamide gel revealed three to four closely migrating components in the final purified preparation (Fig. 5d, 6d, and 7d); however, protease zymogram, analyses showed that all the components possessed protease activity. This observation of microheterogeneity in the final purified protease preparation is clearly demonstrated in Fig. 7e. Zymogram analysis (not shown) of the thin-layer isoelectric focusing pattern obtained with the culture supernatant fluids (stage 1) gave results similar to those obtained with the stage 5 preparation, except that an additional protease band was detected close to the anode. Thus, the multiple protease bands in the stage 5 preparation do not appear to be the result of autodigestion occurring during the purification procedures. Elastase, hexapeptidase, phospholipase C, alkaline phosphatase, deoxyribonuclease, lipase,

VOL. 24, 1979

a

.:*~

S. MARCESCENS METALLOPROTEASE

10.0 9.0

700

8.0

600

K 400

70 6.0 J 5.0 z 4.0 Zt 3.0 k

D) 300

2 0 %

Z 500

IR

200

QZ

100

10

415

5 In 0.5

N

0 3M

0.5

1'1

00 200 300 400 500 600 700 800 900 1000 mnl OF EFFLUENT

..b

I

4

it

3)

VI

"

Pool

e...._1..

t7

Jo

0

'PS

tof

FIG. 1. Hydroxyapatite adsorption chromatography and fused rocket IEP analysis of stage 2 protease preparation. (a) Hydroxyapatite adsorption chromatography pattern. Symbols: 0, 280 nm absorbance; 0, protease activity. (b) Antigen distribution profiles obtained by fused rocket IEP of fractions. Samples (18 ul) of odd-numbered fractions were placed, in sequence, into wells (4 mm in diameter, numbered in figure) cut in 1.2% (wt/vol) agarose (Bio-Rad Laboratories) in 0.02 M Tris-barbital buffer (pH 8.6) and were allowed to diffuse, before electrophoresis, for 1 h at 10C. The upperpart of each gel (55.4 cm2 surface area) was composed of a 1.2% agarose gel (9.5 ml) containing 0. 75 ml of antiserum. After diffusion, the samples were subjected to electrophoresis (anode at top ofgel) at 2 V/cm for 18 h at 100C. Similar results were obtained irrespective of whether the column was washed with equilibrating buffer at 20 ml/h or at 80 ml/h. Linear gradient elation with 0.02 M to 0.6 Mphosphate buffers did not separate the protease activity from the contaminating antigens any more effectively than did the shown single-step elation with 0.3 Mphosphate buffer.

esterase, collagenase, endotoxin, hexose, pentose, and phosphorus were not detected in the final purified protease preparation. The 280/260 nm absorbance ratio of the purified protease was approximately 1.8, thus indicating the absence of contaminating nucleic acids. The Ems'.n I ,,was approximately 12.4. Molecular weight determination. Sodium

dodecyl sulfate-polyacrylamide gel electrophoresis of the denatured protease preparation gave a molecular weight of approximately 52,500 (Fig. 8). The apparent molecular weight of the native

enzyme, as estimated by Sephadex G-100 gel filtration, was approximately 44,000. Estimation of sedimentation coefficient. The sedimentation coefficient of the protease, as estimated by rate zonal sucrose density gradient centrifugation, was approximately 4.2S (Fig. 9). Amino acid and metal analyses. The results of the amino acid analysis are presented in Table 2. The protease contained a high percentage of acidic amino acids, no cysteine, and, based on a molecular weight of 52,500, 1 g-atom of Zn2+ and 7 g-atoms of Ca2+ per mol. No significant

416

INFECT. IMMUN.

LYERLY AND KREGER a

b

Pool i 4 A.

R.

5

FIG. 2. Preparative flat-bed electrofocusing and fused rocket IEP analysis of stage 3 protease preparation. (a) Electrofocusing pattern of protease preparation. Symbols: 0, pH of odd-numbered fractions measured at 40C; 0, protease activity offractions. Trough application of the stage 3 preparation and the incorporation of sucrose (10%, wt/vol) into the gel before electrofocusing did not improve enzyme recovery. (b) Antigen distribution profiles obtained by fused rocket IEP of fractions. Fused rocket IEP was performed as described in the legend of Fig. 1.

a

o O.e iWPooi..

200

4 C

X

4.

ZC00L IR

k

A0

Q!J -i

0

i0o CLk

Cv

ct

6 4

i0.04

2

0.02 130

170

210

250

290

mi

OF

330

370

410

450

490

§30-

EFFLUENT

b Pool l

f,

-i

.A

Il

if

P my

it

9

-i

FIG. 3. Sephadex G-100 gel filtration and fused rocket IEP analysis of stage 4 protease preparation. (a) Gel filtration pattern of protease preparation. Symbols: 0, 280 nm absorbance; 0, protease activity. (b) Antigen distribution profiles obtained by fused rocket IEP of fractions. Fused rocket IEP was performed as described in the legend for Fig. 1.

VOL. 24, 1979

S. MARCESCENS METALLOPROTEASE

417

TABLE 1. Purification of S. marcescens BG protease Purification stage

Vol (Ml)

Protein

Total (mg/ |prote | Aciiy (U/mi) (mg) ~~~mI) 2.2 7 9,240

1. Culture supernatant fluids 4,200 377 1,350 2. Dialyzed concentrate of ammo12 31.4 nium sulfate precipitate 234 320 3. Dialyzed concentrate of Bio-Gel 32 7.4 HTP chromatography pool -a 565 4. Concentrate of flat-bed isoelectric 10 focusing pool 94b 5. Sephadex G-100 gel filtration pool 39 2.4 120 a , Ampholine interference prevented Lowry protein determination. b 85

Total

activity

Recov-

ery

(U)

Sp act (U/mg protein)

29,400

16,200

100 55

3.2 43

10,240

35

43

6,000

20

-

4,680

16

5OC

mg (dry weight).

C 55 U/mg (dry weight).

amount of Fe3+, C02+, Mn2+, and Mg2+ was detected. Determination of pH optimum. Protease activity against azocasein was optimal from pH 5.5 to 7.5. Effect of metallic ions and inhibitors on protease activity. The effect of various metallic ions and enzyme inhibitors on protease activity is shown in Table 3. Various heavy metal cations, oxidizing agents, chelating agents, and reducing agents inactivated the serratia protease. Sulfhydryl-inactivating agents and trypsin inhibitors had no significant effect on activity. Exposure of the protease to EDTA at acidic pH inactivated the enzyme more effectively than did exposure to EDTA at alkaline pH. Dialysis of OPA- and EDTA-inactivated protease preparations resulted in complete restoration and 60% restoration of activity, respectively. The addition of various metal ions to OPA- and EDTA-inactivated protease preparations (metal ion-to-inhibitor molar ratio of 2:1) gave the following results: (i) partial reactivation of OPAtreated protease by Mg2+ (25%) and Co3+ (50%); (ii) total reactivation of EDTA-treated protease by Mn2+, Co', and Zn2+; and (iii) partial reactivation of EDTA-treated protease by Ca2" (50%) and Cu2+ (30%). Effect of heat and pH on protease stability. Heating for 15 min, in 0.1 M Tris-hydrochloride buffer (pH 7.5), at 25, 37, and 450C did not affect activity; however, heating at 600C eliminated detectable protease activity. Enzyme activity was stable at room temperature for 60 min at pH 5 to 10; activity was destroyed at pH 2 to 4. The enzyme was stable in deionized glassdistilled water (5 mg/ml) for at least 1 day at 40C. In vivo toxicity studies. Intracorneal injection of 0.5 pIg of stage 5 protease caused destruction of rabbit corneas by 4 h postinjection. Rabbits injected intratracheally with 50 pIg of protease exhibited extensive pulmonary hemor-

rhage by 1 day postinjection. Intradermal injection of 3 or more micrograms of protease increased capillary permeability, and hemorrhagic lesions were noted 30 min after the administration of 20 pLg of protease. The minimum lethal intravenous dose of the protease for mice was approximately 400 pug or 17 mg/kg of body weight.

DISCUSSION This paper is the first full-length publication to describe a purification scheme for obtaining large amounts of a tissue-damaging, S. marcescens neutral metalloprotease preparation free of contamination with medium constituents and other biologically active products of the bacterium. Other investigators (1, 36, 37) have reported previously on the isolation and properties of a serratia metalloprotease, and the results of their characterization studies agree, for the most part, with the data presented in this paper. However, the preparations studied by other workers were not rigorously examined for purity, as described in this paper, by a battery of the currently available, highly resolving, physicochemical, immunological, and biological activity techniques. Thus, it should now be possible to perform definitive studies to evaluate the possible role of serratia protease production in the pathogenesis of serratia-induced diseases. Our observation of the microheterogeneous nature of the purified serratia protease is not unique. McQuade and Crewther (30) also reported on the molecular microheterogeneity of a S. marcescens protease preparation examined by isoelectric focusing in polyacrylamide gel. In addition, preparations of crystalline bovine trypsin have been found to contain variable quantities of at least five molecular weight species, apparently resulting from partial autodigestion (51). Furthermore, purified P. aeruginosa elastases have been reported to exhibit microheter-

418

INFECT. IMMUN.

LYERLY AND KREGER

fluids of Serratia sp. produced localized hemorrhage and necrosis; however, the preparations were not examined for protease activity, and their heterogeneous nature made it impossible to determine which of the various serratia prodb

a

iW ...

0:R

I:

d

i:

.C,

FIG. 5. Crossed IEP of serratia protease preparations. The gels and the samples contained 2 x 10-3 M OPA. (First dimension) Samples (18 ILI) were placed into wells (4 mm in diameter, indicated by circles) cut in gels composed of 1.2% (wt/vol) agarose (Bio-Rad Laboratories) in 0.02 M Tris-barbital buffer (pH 8.6) and were subjected to electrophoresis (anode to right) at 9 to 10 V/cm for 90 min at 10'C. (Second dimension) Upper part of each gel (62.2 cm2) was composed of a 1.2% agarose gel (11 ml) containing 0.8 ml of and FIG. 4. Ouchterlony double immunodiffusion Electrophoresis (anode at top of gel) was Grabar- Williams IEP analyses of serratia protease antiserum. at 2 V/cm for 18 h at 10'C. (a) Stage 2 (200 preparations. Samples (18 ,ul) examined by immuno- performed g; (b) stage 3 (100 pg); (c) stage 4 (4.7 U); (d) stage mm wells the into (4 peripheral were placed diffusion in diameter) cut in gels composed of 1.2% (wt/vol) 5 (100 agarose (Bio-Rad Laboratories) in 0.01 M boratesaline buffer (pH 8.5), and the center well (4 mm in diameter) was filled with 2x concentrated antiserum (18 ull). Distance from the peripheral wells to the center well is approximately 4.5 mm, edge to edge. Samples (18 ,ul) examined by Grabar- Williams IEP were placed into wells (4 mm in diameter) cut in gels composed of 1.2% agarose in 0.04 M barbital buffer (pH 8.6), and were subjected to electrophoresis (anode at right) at 9 V/cm for 65 min at 10 C; 2x concen4IF trated antiserum (100 Iul) was added to each trough. Immunodiffusion and IEP gels and samples contained a protease-inhibiting concentration of OPA (2 10' M). Results shown in both patterns are after 2 days of diffusion at 40C. (a) Immunodiffusion patterns. (Well 1) Stage 2 (200 pg); (well 2) stage 3 (100 pg); (well 3) stage 4 (4.7 U); (well 4) stage 5 (100 ptg); (wells 5 and 6) empty. (b) Immunoelectrophoresis patterns. (Well 1) Stage 2 (200 WJg); (well 2) stage 3 (100 aJ; (well 3) stage 4 (4.7 U); (well 4) stage 5 (100 ag). g.

4"

x

ogeneity (23), and multiple forms of a variety of different bacterial toxins and mammalian enzymes have been revealed by isoelectric focusing (4, 47). Liu (28) observed previously that the intradermal injection of sterile culture supernatant

FIG. 6. Polyacrylamide gel electrophoresis of serratia protease preparations. Anode at bottom of gels. Patterns shown were obtained with 7.5% monomer gels. (a) Stage 2 (163 pg); (b) stage 3 (83 pg); (c) stage 4 (3.7 U); (d) stage 5 (83 pfg); (e) blank control gel.

VOL. 24, 1979

S. MARCESCENS METALLOPROTEASE

b

a

d

c

e

5

419

r Bovine serum olbumin

4

Ovolbumin

S. morcescens proteose

activity

Cup

I

%-* _ . %. 0 _0

0

3

CY

In !

_ 0111 _

_

II

I

I

I

I

II

18 19 to 21 22 23 24 25 26 27 DISTANCE OF PEAK FROM MENISCUS (mm) 17

FIG. 9. Estimation of serratia protease sedimentation coefficient by sucrose density gradient centrifugation. Results shown are mean values obtained by centrifugation of triplicate samples. TABLE 2. Amino acid analysis of S. marcescens BG protease No. residues/ Amino acid

FIG. 7. Analytical thin-layer isoelectric focusing of serratia protease preparations. Anode at top of gels. (a) Stage 2 (163 pg); (b) stage 3 (83 pg); (c) stage 4 (3.7 U); (d) stage 5 (83 pg). Gels in (a) to (d) are stained with Coomassie blue R-250. (e) Protease zymogram of stage 5 preparation (20 pg, 1 U). The protease bands appear as white zones on a dark background. The white line at the top of the gel is an artifact caused by the acidic anode solution. 12% Monomer Gel Bovie Serum - Albumin - S. marcecens eProteose

70.000 60,000

50,000

Ovalbumin

40.000

Pepin

30.0001 Chymotrypsinogen A 20,000

[

Myoglobin

ymprotease

mol of pro-

teasem

Aspartic acid Threonineb

1.30 0.54

Serineb

0.58

31

Glutamic acid Proline

0.68 0.18 1.01 0.81 0.45 0.007 0.38 0.50 0.29 0.50 0.26 0.19 0.15 0.00 0.13 0.13

36 9 53 43 23 0 20 26 15 26 14 10 8 0 7 7

Glycmie

Alanine Valinec Methionine Isoleucinec Leucinec Tyrosine Phenylalanine

Lysine Histidine Arginine

68 28

Half-cystined Tryptophane NH3 aBased on a molecular weight of 52,500. bCalculated by extrapolation to zero hydrolysis time. 'Calculated from the 72-h hydrolysis samples. d Determined as cysteic acid. 'Determined by the method of Pajot (43).

ucts in the preparations caused the observed tissue damage. The present study demonstrates Cytocrome C that the intradermal injection of a S. marcescens metalloprotease preparation of well-docuI I I I I I I I I in Im mented purity elicits a similar response. The 05 0 0.1 .7 08 0.9 0.3 0.6 A2 Q4 ,UV a small amount of serratia protease that finding RELATIVE MOBILITY FIG. 8. Estimation of serratia protease molecular increases capillary permeability agrees with the weight by sodium dodecyl sulfate-polyacrylamide gel observation of Conroy et al. (8) that intradermal electrophoresis. Results shown are mean values ob- injection of 2 1ug of a serratia protease preparation of poorly documented purity increased captained by electrophoresis of duplicate samples.

420

INFECT. IMMUN.

LYERLY AND KREGER

TABLE 3. Effect of various metallic cations and enzyme inhibitors on activity of S. marcescens BG protease

Reagent'

Residua`l activity

(%)

Mg2+ Mn2+ Ca2+ Ba2+ Co3+ Cu2+ Fe2+ Hg2+

100 100 100 100 65 20 20 20

Pb2+

Ag2O

15 0

Zn2+

0

Residual

Reagenta Sodium cyanide Sodium azide Sodium thioglycolate

Cysteine HOl

Dithiothreitol Monoiodoacetic acid Sodium arsenite p-Hydroxymercuribenzoate N-ethyl maleimide

activity

(%)

ACKNOWLEDGMENTS

100 100 85 60 30 100 100 100

This investigation was supported by Public Health Service grants EY-01104 and HL-16769 from the National Eye Institute and the National Heart, Lung, and Blood Institute, re-

Phenylmethylsulfo-

90 100

Soybean trypsin in-

100

nyl fluoride

KMnO4 N-bromosuccinimide Na2 EDTA OPA

Nitrilotriace-

55 0

80 (O) 0

hibitor Lima bean trypsin inhibitor Egg white trypsin inhibitor Human a-l-antitrypsin Pancreatic trypsin inhibitor

points (50), and electrophoretic mobilities in agarose gel (19). Additional studies are needed to isolate, characterize, and compare the different proteases of S. marcescens.

100

100

100 100

75

tate

60 a,a'-dipyridyl a Final concentration in assay mixtures (1.6 ml) was 2 x 10-3 M, with the exception of (i) the mixture containing the enzyme exposed to EDTA at pH 5 (which contained 3 x 10' M EDTA) and (ii) the mixtures containing the naturally occurring trypsin inhibitors (which contained 0.5 mg of the reagent per assay mixture). Residual protease activity was determined as described in the text. b Protease preparation (2 U of protease/ml) exposed for 30 min at 250C to 5 x 10' M EDTA at pH 5.0 before addition of sample (0.1 ml) to assay mixture.

illary permeability. In addition, the rapid and extensive cornea and lung damage elicited by our purified protease preparation corroborates previous observations made with partially purified serratia protease preparations (25; Kreger, Abstr. Annu. Meet. Am. Soc. Microbiol., M288, p. 114, 1974). The results of our previous fractionation experiments (25) and observations made during the current hydroxyapatite adsorption chromatography (Fig. la) and protease zymogram analysis of the stage 2 and stage 1 preparations, respectively, suggest that the S. marcescens strain used in our studies can elaborate at least two different extracellular proteases. Examinations of other S. marcescens strains, by four other groups of investigators, support the idea that many strains ofS. marcescens can elaborate several extracellular proteases which differ from one another with regard to their substrate specificities (49), pH optima (22), sensitivities to inactivation by EDTA and DFP (22), isoelectric

spectively. LITERATURE CITED 1. Aiyappa, P. S., and J. 0. Harris. 1976. The extracellular metalloprotease of Serratia marcescens. 1. Purification and characterization. Mol. Cell. Biochem. 13:95-100. 2. Altemeier, W. A., W. R. Culbertson, W. D. Fullen, and J. J. McDonough. 1969. Serratia marcescens septicemia. Arch. Surg. 99:232-238. 3. Andrews, P. 1964. Estimation of the molecular weights of proteins by Sephadex gel-filtration. Biochem. J. 91: 222-233. 4. Arbuthnott, J. P., A. C. McNiven, and C. J. Smyth. 1975. Multiple forms of bacterial toxins in preparative isoelectric focusing, p. 212-239. In J. P. Arbuthnott and J. Beeley (ed.), Isoelectric focusing. Butterworth and Co. Ltd., London. 5. Arvidson, S., and T. Wadstrom. 1973. Detection of proteolytic activity after isoelectric focusing in polyacrylamide gel. Biochim. Biophys. Acta 310:418-420. 6. Axelsen, N. H., J. Kroll, and B. Weeke (ed.). 1973. A manual of quantitative immunoelectrophoresis. Meth-

ods and applications. Universitetsforlaget, Oslo. 7. Campbell, D. H., J. S. Garvey, N. E. Cremer, and D. H. Sussdorf. 1970. Methods in immunology, 2nd ed., p. 245-246. W. A. Benjamin, Inc., New York. 8. Conroy, M. C., N. H. Bander, and I. H. Lepow. 1975. Effects in the rat of intradermal injection of purified proteinases from Streptococcus and Serratia marcescens. Proc. Soc. Exp. Biol. Med. 150:801-806. 9. Davis, J. T., E. Foltz, and W. S. Blakemore. 1970. Serratia marcescens: a pathogen of increasing clinical importance. J. Am. Med. Assoc. 214:2190-2192. 10. De Siervo, A. J. 1969. Alterations in the phospholipid composition of Escherichia coli B during growth at

different temperatures. J. Bacteriol. 100:1342-1349. 11. Dische, Z. 1949. Spectrophotometric method for the determination of free pentose and pentose in nucleotides. J. Biol. Chem. 181:379-392. 12. Dorwart, B. B., E. Abrutyn, and H. R. Schumacher. 1975. Serratia arthritis: medical eradication of infection in a patient with rheumatoid arthritis. J. Am. Med. Assoc. 225:1642-1643. 13. Dubois, M., K. A. Gilles, J. K. Hamilton, P. A. Rebers, and F. Smith. 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem.

28:350-356. 14. Epstein, E. E., and T. E. Carson. 1973. Serratia granuloma. J. Am. Med. Assoc. 223:670-671. 15. Fairbanks, G., T. L. Steck, and D. F. H. Wallach. 1971. Electrophoretic analysis of the major polypeptides of the human erythrocyte membrane. Biochemistry 10:

2606-2617. 16. Garen, A., and C. Levinthal. 1960. A fine-structure genetic and chemical study of the enzyme alkaline phosphatase of Escherichia coli. I. Purification and characterization of alkaline phosphatase. Biochim. Biophys. Acta 38:470-483. 17. Graber, C. D., L. S. Higgins, and J. S. Davis. 1965. Seldom-encountered agents of bacterial meningitis. J. Am. Med. Assoc. 192:114-118.

VOL. 24, 1979

S. MARCESCENS METALLOPROTEASE

18. Grassman, W., and A. Nordwig. 1960. Quantitativer Kolorimetrischer Test auf Kollagenase. Hoppe Seyler's Z. Physiol. Chem. 322:267-272. 19. Grimont, P. A. D., F. Grimont, and H. L. C. Dulong de Rosnay. 1977. Characterization of Serratia marcescens, S. liquefaciens, S. plymuthica, and S. marinorubra by electrophoresis of their proteinases. J. Gen. Microbiol. 99:301-310. 20. Herbert, W. J. 1978. Mineral-oil adjuvants and the immunization of laboratory animals, appendix 3, p. A3.1. In D. M. Weir (ed.), Handbook of experimental immunology, 3rd ed. Blackwell Scientific Publications, Oxford. 21. Huggins, C., and J. Lapides. 1947. Chromogenic substrates. IV. Acyl esters of p-nitrophenol as substrates for the colorimetric determination of esterase. J. Biol. Chem. 170:467-482. 22. Kaska, M., 0. Lysenko, and J. Chaloupka. 1976. Exocellular proteases of Serratia marcescens and their toxicity to larvae of GaUeria mellonella. Folia Microbiol. 21:465-473. 23. Kreger, A. S., and L D. Gray. 1978. Purification of Pseudomonas aeruginosa proteases and microscopic characterization of pseudomonal protease-induced rabbit corneal damage. Infect. Immun. 19:630-648. 24. Kreger, A. S., and O. K. Griffin. 1974. Physicochemical fractionation of extracellular cornea-damaging proteases of Pseudomonas aeruginosa. Infect. Immun. 9: 828-834. 25. Kreger, A. S., and O. K. Griffin. 1975. Cornea-damaging proteases of Serratia marcescens. Invest. Ophthalmol. 14:190-198. 26. Kurioka, S., and P. V. Liu. 1967. Improved assay for phospholipase C. Appl. Microbiol. 15: 551-555. 27. Lazachek, G. W., G. L. Boyle, A. L. Schwartz, and J. H. Leopold. 1971. Serratia marcescens, an ocular pathogen. Arch. Ophthalmol. 86:599-603. 28. Liu, P. V. 1961. Observations on the specificities of extracellular antigens of the genera Aeromonas and Serratia. J. Gen. Microbiol. 24:145-153. 29. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275. 30. McQuade, A. B., and W. G. Crewther. 1969. Activity against a synthetic substrate by a preparation of extracellular proteinase from Serratia marcescens. Biochim. Biophys. Acta 191:762-764. 31. Maki, D. G., C. G. Hennekens, C. W. Phillips, W. V. Shaw, and J. V. Bennett. 1973. Nosocomial urinary tract infection with Serratia marcescens: an epidemiological study. J. Infect. Dis. 128:579-587. 32. Marks, J. 1952. Recognition of pathogenic staphylococci: with notes on non-specific staphylococcal haemolysin. J. Pathol. Bacteriol. 64:175-186. 33. Martin, R. G., and B. N. Ames. 1961. A method for determining the sedimentation behavior of enzymes: application to protein mixtures. J. Biol. Chem. 236: 1372-1379. 34. Meltz, D. J., and M. H. Grieco. 1973. Characteristics of Serratia marcescens pneumonia. Arch. Intern. Med. 132:359-364. 35. Mills, J. 1976. Serratia marcescens endocarditis: a regional illness associated with intravenous drug abuse. Ann. Intern. Med. 84:29-35.

36. Miyata, K., K. Maejima, K. Tomoda, and M. Isono. 1970. Serratia protease. Part I. Purification and general properties of the enzyme. Agric. Biol. Chem. 34:310318. 37. Miyata, K., K. Tomoda, and M. Isono. 1971. Serratia protease. Part III. Characteristics of the enzyme as a metalloprotease. Agric. Biol. Chem. 35:460-467. 38. Moore, S. 1963. On the determination of cystine as cysteic acid. J. Biol. Chem. 238:235-237. 39. Morihara, K., H. Tsuzuki, T. Oka, H. Inoue, and M. Ebata. 1965. Pseudomonas aeruginosa elastase. Isolation, crystallization, and preliminary characterization. J. Biol. Chem. 240:3295-3304. 40. Nagai, Y., C. M. Lapiere, and J. Gross. 1966. Tadpole collagenase: preparation and purification. Biochemistry 5:3123-3130. 41. Nestle, M., and W. K. Roberts. 1969. An extracellular nuclease from Serratia marcescens. J. Biol. Chem. 244: 5213-5218. 42. Ovenfors, C., M. Markland, E. Johnson, and G. Lundblad. 1973. Investigation of the protease-forming ability of Serratia marcescens and one of its mutant strains. Acta Chim. Scand. 27:3791-3800. 43. Pajot, P. 1976. Fluorescence of proteins in 6-M guanidine hydrochloride. A. Method for the quantitative determination of tryptophan. Eur. J. Biochem. 63:263-269. 44. Ryden, A. C., and B. von Hofsten. 1968. Some properties of the extracellular proteinase and the cell-bound peptidase of Serratia. Acta Chim. Scand. 22:2803-2808. 45. Salceda, S. R., J. Lapuz, and R. Vizconde. 1973. Serratia marcescens endophthalmitis. Arch. Ophthalmol. 89:163-166. 46. Spackman, D. H., W. H. Stein, and S. Moore. 1958. Automatic recording apparatus for use in the chromatography of amino acids. Anal. Chem. 30:1190-1206. 47. Susor, W. A., M. Kochman, and W. J. Rutter. 1969. Heterogeneity of presumably homogeneous protein preparations. Science 165:1260-1262. 48. Tietz, N. W., T. Borden, and J. D. Stapleton. 1959. An improved method for the determination of lipase in serum. Am. J. Clin. Pathol. 31:148-154. 49. Traub, W. H., and I. Kleber. 1974. Detection of fibrinolytic and elastase activity among clinical isolates of Serratia marcescens. Pathol. Microbiol. 41:334-340. 50. Wadstrom, T., and C. J. Smyth. 1975. Isoelectric focusing in polyacrylamide gel for analysis of cell proteins in bacterial taxonomy: a methodological study, chapter 13. In P. G. Righetti (ed.), Progress in isoelectric focusing and isotachophoresis. North-Holland Publishing Co., Amsterdam. 51. Walsh, K. A. 1970. Trypsinogens and trypsins of various species. Methods Enzymol. 19:41-63. 52. Ward, P. A., J. Chapitas, M. C. Conroy, and I. H. Lepow. 1973. Generation by bacterial proteinases of leukotactic factors from human serum, and human C3 and C5. J. Immunol. 110:1003-1009. 53. Weber, K., J. R. Pringle, and M. Osborn. 1972. Measurement of molecular weights by electrophoresis on SDS-acrylamide gel. Methods Enzymol. 26:3-27. 54. Winter, A., H. Perlmutter, and H. Davies. 1975. Preparative flat-bed electrofocusing in a granulated gel with the LKB 2117 Multiphor. LKB Application Note 198. LKB-Produkter AB, Stockholm.

421