J. Med. Microbiol. - Vol. 45 (1996), 219-225 1996 The Pathological Society of Great Britain and Ireland

BACT E R I A L PATH 0 G E N I C ITY Differentiation of thermolysins and serralysins by monoclonal antibodies

CORA KO01 and PAMELA A. SOKOL

Department of Microbiology and Infectious Diseases, University of Calgary Health Sciences Centre, Calgary, Alberta, Canada TZN 4N1

Two monoclonal antibodies (MAbs) to a 36-kDa extracellular metalloprotease (PSCP) from Burkholderia (Pseudomonas) cepacia were found to react with , Pseudomonas aeruginosa elastase, alkaline (Apr) and LasA, Serratia marcescens protease (SMP), Aeromonas hydrophila protease (AhP), and both the lethal factor (LF) and protective antigen (PA) of Bacillus anthracis on immunoblots. The MAbs were capable of neutralising the proteolytic activity of thermolysin, I? aeruginosa elastase and PSCP but not that of Apr, SMP, and AhP. These results suggest that these MAbs may be able to differentiate between the thermolysin and serralysin family of metalloproteases on the basis of their neutralisation capability and could, therefore, be useful tools in the characterisation of new bacterial .

Introduction homology to thermolysin, l? cholerae HNprotease, and Msp [2, 5, 81. Elastase, HNprotease and Msp are Many bacteria secrete extracellular proteases. Some of all synthesised as large prepropeptides which are these proteases are toxins or factors involved in processed to a mature proteolytically active form [8, virulence, whereas others degrade proteins to produce 13, 141. small peptides and amino acids which can be utilised by the bacterium. Bacteria synthesise numerous The serralysin group of proteases includes P aerugi- proteases belonging to the serine-, cysteine- and nosa alkaline protease and the extracellular metallo- metalloprotease classes [ 11. Bacterial metalloproteases proteases from Serratia rnarcescens and Erwinia can be classified in these groups: the thermolysins and chrysantherni [2, 9, lo]. These three metalloproteases serralysins. Proteases are generally classified in these lack typical signal sequences at their N-termini and groups on the basis of amino-acid sequence homology. are preceded by short propeptides [15]. The S. The most thoroughly characterised metalloprotease is marcescens 50-kDa metalloprotease (SM protease) is thermolysin [2], a heat-stable neutral metalloprot ease thought to be an important virulence factor involved produced by Bacillus therrnoproteolyticus. It contains in both keratitis and pulmonary infections [ 16, 171. E. the HEXXH zinc-binding motif which is characteristic chrysanthemi secretes at least four different metallo- of zinc metalloproteases [3]. Many other thermolysin- proteases [18]. E. chrysantherni B protease shares 60% like bacterial metalloproteases have been identified. homology with the SM protease. Legionella pneumophila secretes a 38-kDa zinc me- talloprotease called the major secretory protein (Msp) r! aeruginosa secretes another protease LasA, that is that may play a role in the pathogenesis of L. synthesised as a 42-kDa protein, which is processed to pneumophila infection [4, 51. Vibrio cholerae secretes a 2 1-kDa active peptide. LasA has both elastolytic and a 33-kDa zinc metalloprotease (HA/protease) [6], staphylolytic activity [ 19,201. Zinc stimulates the which can cleave biologically important substrates such production of LasA [21]; however, LasA activity is as mucin, fibronectin and bactoferritin [7]. Pseudomo- not inhibited by classic metalloprotease inhibitors such nas aeruginosa secretes two well-characterised metal- as EDTA [22]. Aeromonas hydrophila secretes a loproteases, elastase [8] and alkaline protease [9, lo]. P protease (AhP) which has an amino-terminal sequence aeruginosa elastase has been shown to degrade or 69% identical to LasA [23]. AhP is a 19-kDa inactivate several biologically important substrates such metalloprotease that degrades fibrin. as fibrin, collagen, complement, human immuno- globulin G (IgG), transferrin and lactoferrin [ 1 1, 121. Interestingly, it has been found that tetanus toxin and Elastase exhibits considerable amino-acid sequence several of the botulinum toxins are metalloproteases [24-271. Recently, Klimpel et al. [28] reported that the Received 5 Dec. 1995; revised version accepted 9 Feb. 1996. anthrax lethal toxin factor (LF) contains a zinc Corresponding author: Professor P. A. Sokol. metalloprotease consensus sequence. LF may have 220 C. KO01 AND P. A. SOKOL the third zinc ligand on the amino-terminal side of the of PSCP purified from Bur. cepacia Pc715j on days 0, consensus zinc metalloprotease site, making it more 7, 14 and 21. Three days after the final injection, similar to P aeruginosa alkaline protease and the SM spleens were removed and spleen cells were collected protease than to the classical thermolysin-like pro- in a sterile petri dish (Nunc) containing DMEM. Spleen teases. However, all these zinc metalloproteases and NS-1 cells (1O:l) were fused by polyethylene exhibit a great deal of homology especially around glycol 4000 (Merck) 50% w/v. Fused cells were the zinc-binding motif [2]. dispensed into 24-well plates at 6 x lo4 input spleen cells/well. The hybridoma cell lines were maintained Burkholderia (Pseudomonas) cepacia has been shown for 14 days in DMEM supplemented with penicillin, to secrete at least two proteases, a 36-kDa protease streptomycin, glutamine, calf supreme 1o%, 100 mM (PSCP) with high protease activity and a 40-kDa hypoxanthine, 1 mM aminopterin and 16 mM thymidine. protease with low activity [29,30]. PSCP was shown This medium was designated HAT. Hybridoma cell to be a metalloprotease that can degrade gelatin, lines producing MAbs to PSCP, detectable by - collagen and hide powder azure but not human linked immunosorbent assay (ELISA) or immunoblot- immunoglobulins [30]. Monoclonal antibodies (MAbs) ting, or both, were transferred to hypoxanthine- to PSCP have been used to demonstrate that the Bur. thymidine (HT) medium and cloned three times by cepacia protease was immunologically related to P limiting dilution in 96-well plates (Nunc). Antibody- aeruginosa elastase and the K cholerae HA/protease producing cells (2 x lo6) were injected into pristane- [29]. Furthermore, MAbs were isolated that could primed BALB/c mice for ascites production. neutralise P aeruginosa elastase and the J! cholerae HNprotease but not P aeruginosa alkaline protease. The isotype of the antibodies in the ascitic fluid was Thus, it was thought that it was possible that these determined with a mouse typer subisotyping kit MAbs could distinguish between two distinct families (BioRad Laboratories) as recommended by the man- of bacterial metalloproteases by their neutralisation ufacturer. capabilities. This study aimed to determine whether two representative MAbs to PSCP could recognise Five MAbs (36-1-5. 36-5-15, 36-6-6, 36-6-8 and 36-9- other metalloproteases such as thermolysin, the SM 19), all of the IgM isotype, were identified which protease, AhP and LasA. Also, the ability of these neutralised PSCP [29]; two (36-6-6 and 36-6-8) were MAbs to neutralise the activity of thermolysin-like chosen for further characterisation in this study. These proteases (eg, B. thermoproteolyticus thermolysin, P antibodies were selected because of their high titre areuginosa elastase, J! cholerae HA/protease) and and strong reaction on immunoblots. For the neutra- members of the serralysin family of metalloproteases lisation experiments, MAbs were purified with the (e.g., the SM protease and P aeruginosa alkaline Pierce IgM purification kit as recommended by the protease) was compared. manufacturer. To determine the ability of these MAbs to react with other proteases, proteins (3-5 pg) were separated by SDS-PAGE [31] and transferred to Materials and methods immobilon membranes by the method of Towbin et al. [32]. After blocking with BSA 5%, the blot was Pro teases incubated with MAb 36-6-6 (1 in 10000) or MAb 36- Elastase and alkaline protease, prepared from P 6-8 (1 in 500) in BSA 5% for 1 h at 37°C. After aeruginosa, were purchased from Nagase Biochemical extensive washing, the blot was incubated with goat Inc. (Osaka, Japan). Thermolysin from B. thermopro- anti-mouse antibody-peroxidase conjugate ( 1 in 2000) teolyticus and the Serratia metalloprotease (SMP) from for 1 h at 37°C. After washing, the blot was developed S. marcescens were purchased from Sigma. The A. with horse-radish peroxidase [29]. hydrophila proteinase (AhP) was generously provided by Dr M. Wiezorek (Cortech Inc., Denver, Colorado, Protease assay USA). Dr E. Kessler (Tel-Aviv University Sakler Faculty of Medicine, Israel) kindly provided P Proteolytic activity was quantified by measuring the aeruginosa LasA. Anthrax toxin lethal factor and hydrolysis of hide powder azure (Sigma) in an assay protective antigen were generously provided by Drs based on the method of Rinderknecht et al. [33]. K. Klimpel and S. Leppla (NIH, Bethesda, MD, USA). Proteases were mixed with 20 mg of hide powder azure Bur. cepacia PSCP and 40-kDa proteases were purified in a final volume of 1.5 ml in 10 mM Tris, pH 8.0. The as previously described [29]. reaction mixtures were incubated for 2 h at 37°C with vigorous shaking. The samples were centrifuged at Monoclonal antibodies 6000 g for 10 min and the A595, of the supernates was measured. To account for non-specific hydrolysis of the The MAbs employed in this study have been described hide powder azure, control samples without protease previously [29]. MAb production was based on the were incubated in parallel and the A595 of the resultant method of Kohler and Milstein [12]. BALB/c mice supernate was subtracted as background. Stocks (1 mg/ were given four intraperitoneal (i.p.) injections of 30 pg ml) of each protease tested were prepared. Dilutions DIFFERENTIATION OF BACTERIAL METALLOPROTEASES 22 1

were made and protease assays were performed as antibody was present. After incubation for 2 h at 37"C, described above. Concentrations of protease that would the plates were washed five times with buffer A. Goat give an A595 of c. 1.0 were used in the neutralisation anti-mouse peroxidase conjugate ( 100 pl, 1 in 2000; assays. The amounts used were SMP 0.4pg, P Kirkegaard and Perry Laboratories, Inc.) was added to aeruginosa elastase 0.01 pg, thermolysin 0.05 pg, each well and incubated at 37°C for 2 h. Wells were PSCP 0.5 pg, P aeruginosa alkaline protease 0.2 pg washed five times with buffer A and aspirated and AhP 0.26 pg. These proteases were pre-incubated thoroughly. ABTS substrate (BioRad) 100 pllwell, was with either Tris-HC1 buffer, control ascitic fluid (1 in applied to each well and, after agitation at room 500) or MAb (1 in 500) in a final volume of 1.0 ml, for temperature for 15 min, the A405 was determined with 2 h at 37°C. The samples were centrifuged briefly and an EL340 BioKinetics Reader (Bio-Tek Instruments). the supernates were transferred to new tubes containing Tris-buffer 0.5 ml and hide powder azure 20mg. Proteolytic activity was measured as described above. Results and discussion

Previously, we reported that MAbs 36-6-6 and 36-6-8 ELISA to measure antibody affinity reacted with Bur: cepacia PSCP and 40-kDa proteases, An ELISA was performed to compare the affinity of P aeruginosa elastase, P aeruginosa alkaline protease MAb 36-6-6 for the various proteases. Immulon 4 96- (Apr) and the I! cholerae HAlprotease on immunoblots well microtitration plates (Dynatech) were coated with [29]. To determine the extent of cross-reactivity of 100 pl of each protease (1 pg/ml) in carbonate coating these MAbs with metalloproteases, the ability of MAb buffer (pH 9.6) for 2 h at 37°C. The plates were washed 36-6-6 and 36-6-8 to react with other metalloproteases with buffer A (phosphate-buffered saline, pH 7.4, BSA was examined on immunoblots. In addition to those 0.05% w/v) and blocked for 2 h at 37°C by the addition proteases previously described, MAb 36-6-6 reacted of 250pl of PBS, BSA 5% to each well. After two with thermolysin (TLN) from B. thermoproteolyticus, washes with buffer A, serial two-fold dilutions starting S. rnarcescens protease (SMP), P aeruginosa LasA and at 1 in 200 of MAb 36-6-6 in buffer A (100pllwell) the A. hydrophila protease (AhP) (Fig. 1). The were applied to the wells. All assays were performed in corresponding Coomassie blue-stained SDS-PAGE gel duplicate and included negative controls in which no of the protease preparations used is shown in Fig. 2.

Fig. 1. Immunoblot of proteases with MAb 36-6-6. Proteins (3-5 pg) were separated by SDS-PAGE and blotted on to immobilon. Lane 1, mol. wt markers, 2, I? aeruginosa Las A; 3, B. thermoproteolyticus thermolysin (TLN); 4, A. hydrophila protease (AhP); 5, PSCP; 6, Bur. cepacia 40-kDa protease; 7, S. rnarcescens major metalloprotease (SMP); 8, blank; 9, I? aeruginosa alkaline protease (Apr). 222 C. KO01 AND P. A. SOKOL

Fig. 2. SDS-PAGE of (a) protease preparations; (b) B. anthracis LF and PA preparations used in this study. a lane 1, mol. wt markers; 2, F! aeruginosa Las A; 3, F! aeruginosa elastase; 4, TLN; 5, AhP; 6, PSCP; 7, Bur. cepacia 40-kDa protease; 8, SMP; 9, blank; 10, I? aeruginosa Apr. b, lane 1, B. anthracis LF; 2, B. anthracis PA; 3, PSCP. The mol. wt markers are indicated on the right.

Similar results were obtained with MAb 36-6-8 (data protease and TLN but not Apr and SMP (Table 1). not shown). This result suggests that the epitope(s) Therefore, although these metalloproteases have con- recognised by these MAbs are conserved among all served regions which are recognised by these MAbs, these metalloproteases. MAb 36-6-6 was shown in the case of the TLN-like metalloproteases antibody previously not to react with chymotrypsin [29] which binding leads to neutralisation but, with the serraly- indicates that these antibodies do not recognise serine sins, binding of the MAb does not lead to neutralisa- proteases. Anthrax lethal factor (LF) contains a zinc tion. The difference in the ability of these MAbs to metalloprotease consensus sequence; however, no neutralise may be due to differences in amino-acid substrate for proteolytic activity has yet been described composition of the epitope or slight changes in [28]. MAbs 36-6-6 and 36-6-8 also recognised anthrax tertiary configuration. MAb 36-6-8 did not neutralise LF weakly on immunoblots (Fig. 3 and data not AhP activity (Table l), suggesting that this protease shown). The faint reaction was detectable only after a may be more similar to SMP than to TLN and may be significantly longer development time, even though an a member of the metzincin family of metalloproteases equivalent amount of LF was loaded on the gel as other to which the serralysins belong [lS]. LasA does not proteins which reacted strongly, suggesting that either utilise hide powder azure as a substrate. It has been the site(s) recognised by these MAbs is not as well reported to cleave elastin, p-casein and to have conserved in anthrax LF, or that SDS-PAGE affects the staphylolytic activity [20,22]. Recently LasA has been antigenic site. Curiously, MAb 36-6-6 reacted with the shown not to be inhibited by EDTA and o-phenanthro- anthrax PA, which has not been reported to be a line, and, therefore, may not be a metalloprotease [22], metalloprotease, more strongly than with the LF (Fig. although zinc has been shown to enhance LasA 3). Comparison of the P aeruginosa elastase and activity [21]. A proteolytic substrate for LF has not anthrax PA amino-acid sequences reveals very little yet been reported. Therefore, the ability of the MAbs homology. At present the reason for the cross-reactivity to neutralise the activity of LF and LasA was not is unknown. examined.

As MAbs 36-6-6 and 36-6-8 were shown previously to It is possible that differences in neutralisation may neutralise the activity of P aeruginosa elastase, the I? reflect a difference in affinity of the MAbs for the cholerae HNprotease and PSCP but not Apr [29], we various proteases. Therefore, the affinity of MAb 36- hypothesised that these MAbs would be able to 6-8 for the proteases was estimated by comparing the distinguish between two families of metalloproteases dose-response curves of the MAb with the same [ 15,341, the thermolysins and the serralysins, by their concentration of each protease in an ELISA [35]. As neutralisation capabilities. MAbs 36-6-6 and 36-6-8 shown in Fig. 4, there did not appear to be a were found to neutralise elastase, the I? cholerae HA/ correlation between the MAb binding curves for the DIFFERENTIATION OF BACTERIAL METALLOPROTEASES 223

difference between the end-point titration of any of the proteases. When the direct absorbance reading was compared at the same antibody dilution, the difference in absorbance was not more than two-fold at most dilutions. At concentrations of antibody excess, there was some difference in the maximum absorbance values. The protease which would appear to have the lowest affinity for the MAb, however, was PSCP, which was the antigen used to produce the MAb. The differences observed may only reflect slight inaccura- cies in quantifying the small amounts of protein used in the assay. Differences in the ability of the MAbs to neutralise thermolysin and serralysin proteases do not appear to be due to marked differences in the affinity of the MAbs for these proteases.

Although the serralysins and thermolysins share a common zinc-binding motif (XXHEXXHX) and the upper domains are topologically similar, there are differences in the active sites [15]. For example, the helix of thermolysin extends beyond the HEXXH motif, with a polypeptide chain that loops back to the active site and is responsible for the third zinc-binding residue (E166). In the case of the metzincins (, serralysins, snake venom and matrix metalloproteases), the zinc-binding motif has been extended to HEXXHXXGXXH. Thermolysins have a consensus sequence NEXXSD (El66 for TLN) that is not found in the serralysins. At this approximate region, the serralysins (and other metzin- cins) have a conserved methionine which is part of the methionine-containing turn, which may play a role in the structural integrity of the active site. This site is not present in the thermolysins. Thus, it is possible that a MAb may bind to a site near the active site of both thermolysins and serralysins, but only neutralise the members of the thermolysin family. Fig. 3. Immunoblot of B. anthracis PA and LF with MAb 36-6-6. Lane 1, mol. wt markers; 2, PSCP; 3, PA; 4, LF. It is also possible that the epitope(s) recognised by these two MAbs are located near the active site in the different proteases and the ability of the MAb to thermolysin group of proteases but the MAbs recog- neutralise that protease. For example, the binding nize a different epitope in another region of the curve for alkaline protease was almost identical to that serralysin . These epitope(s) may have a for elastase and HNprotease. There was no significant similar amino-acid sequence to the epitope, which

Table 1. Ability of MAbs to Bur: cepacia PSCP to neutralise metalloprotease activity Mean percentage (SD) proteolytic activity* with

Protease MAb 36-6-6 MAb 36-6-8 MCA-6t

PSCP 0.3 (7)$ 11.8 (5): 75 (5) Elastase 0 (1): 0 (3)f 98 (5) Thermol y sin 3 (6)$ 0 (1411 87 (7) Serratia protease 131 (9) 71 (15) 129 (8) Alkaline protease 106 (2) 108 (5) 108 (2) AhP ND 137 (16) ND ND, not done. "Proteolytic activity was expressed as a percentage of the control (100%) which contained no antibody. Values represent mean (SD) of three replicates. tMCA-6 is a purified antibody to P ueruginosu FBP. $Significantly different (p < 0.00 1 ) than reactions containing no antibody with Bonferroni multiple comparisons test. 224 C. KO01 AND P. A. SOKOL

2.00

1.75

1.50

1.25

a0

(d z 1.00 a 0.75

0.50

0.25

0.00 2 3 4 5 6 Log,o dilution of MAb 36-6-6

Fig. 4. Comparison of ELISA binding curves obtained with MAb 36-6-6 and metalloproteases coated on plates at a concentration of 0.1 pg well. Plates were processed as described in the Materials and methods. PSCP; A elastase; V thermolysin; + HA/protease; 0 40-kDa protease; 0 alkaline protease; A SMP; V AhP. may be located near the active site of the thermo- 4. Moffat JF, Edelstein PH, Regula DP, Cirillo JD, Tompkins LS. Effects of an isogenic Zn-metalloprotease-deficient mutant of lysins. Experiments are currently underway in this Legionella pneumophila in a guinea-pig pneumonia model. laboratory to map the epitopes recognised by the Mol Microbiol 1994; 12: 693-705. MAbs. 5. Black WJ, Quinn FD, Tompkins LS. Legionella pneumophila zinc metalloprotease is structurally and functionally homo- logous to Pseudomonas aeruginosa elastase. J Bacteriol 1990; The finding that MAbs to the PSCP neutralise 172: 2608-2613. elastase, TLN, the l? cholerae HNprotease as well 6. Booth BA, Boesman-Finkelstein M, Finkelstein R. A Vibrio cholerae soluble hemagglutinidprotease is a metalloenzyme. as PSCP but not Apr, SMP or AhP suggests that the Infect Immun 1983; 42: 639-644. PSCP belongs to the thermolysin metalloprotease 7. Finkelstein RA, Boesman-Finkelstein M, Holt P. Vibrio family and that AhP may belong to the serralysin cholerae hemagglutinidlectidprotease hydrolyzes fibronectin and ovomucin: F. M. Burnet revisited. Proc Nut1 Acad Sci USA family of metalloproteases. It also appears that these 1983; 80: 1092-1095. MAbs may be able to differentiate between the 8. Bever RA, Iglewski BH. Molecular characterization and serralysins and thermolysins on the basis of their nucleotide sequence of the Pseudomonas aeruginosa elastase neutralisation capability. Therefore, these antibodies structural gene. J Bacteriol 1988; 170: 43094314. 9. Okuda K, Morihara K, Atsumi Y et al. Complete nucleotide may be useful in the initial classification of proteases sequence of the structural gene for alkaline protease from when amino-acid sequence information may not be Pseudomonas aeruginosa IF0 3455. Infect Immun 1990; 58: available. 40834088. 10. Duong F, Lazdunski A, Cami B, Murgier M. Sequence of a cluster of genes controlling synthesis and secretion of alkaline This study was supported by the Canadian Bacterial Diseases protease in Pseudomonas aeruginosa: relationships to other Network of Centres of Excellence. secretory pathways. Gene 1992; 121: 47-54. 11. Britigan BE, Hayek MB, Doebbeling BN, Fick RB. Transferrin and lactoferrin undergo proteolytic cleavage in Pseudomonas References aeruginosa-infected lungs of patients with cystic fibrosis. Infect Immun 1993; 61: 5049-5055. 12. Wretlind B, Pavlovskis OR. Pseudomonas aeruginosa elastase 1. Bond JS, Butler PE. Intracellular proteases. Annu Rev Biochem and its role in Pseudomonas infections. Rev Infect Dis 1979; 5 1987; 56: 333-364. Suppl 5: S998-Sl004. 2. Hase CC, Finkelstein RA. Bacterial extracellular zinc-contain- 13. Hase CC, Finkelstein RA. Cloning and nucleotide sequence of ing metalloproteases. Microbiol Rev 1993; 57: 823-837. the Vibrio cholerae hemagglutinidprotease (HNprotease) gene 3. Jongeneel CV, Bouvier J, Bairoch A. A unique signature and construction of an HNprotease-negative strain. J Bacteriol identifies a family of zinc-dependent metallopeptidases. FEBS 1991; 173: 3311-3317. Lett 1989; 242: 21 1-214. 14. Moffat JF, Black WJ, Tompkins LS. Further molecular DIFFERENTIATION OF BACTERIAL METALLOPROTEASES 225

characterization of the cloned Legionella pneumophila zinc Montecucco C. Tetanus toxin is a zinc protein and its metalloprotease. Infect Immun 1994; 62: 75 1-753. inhibition of neurotransmitter release and protease activity 15. Bode W, Gomis-Ruth F-X, Stocker W. Astacins, serralysins, depend on zinc. EMBO J 1992; 11: 3577-3583. snake venom and matrix metalloproteases exhibit identical 26. Schiavo G, Rossetto 0, Santucci A, DasGupta BR, Montecucco zinc-binding environments (HEXXHXXGXXH and met-turn) C. Botulinum neurotoxins are zinc proteins. J Biol Chem 1992; and topologies and should be grouped into a common family, 267: 23479-23483. the ‘metaincins’. FEBS letters 1993; 331: 134140. 27. Schiavo G, Shone CC, Rossetto 0, Alexander FCG, Mon- 16. Lylerly D, Gray L, Kreger A. Characterization of rabbit corneal tecucco C. Botulinum neurotoxin serotype F is a zinc damage produced by serratia keratitis and by a serratia specific for VAMPIsynaptobrevin. J Biol Chem protease. Infect Immun 1981 ; 33: 927-932. 1993; 268: 11516-11519. 17. Lylerly DM, Kreger AS. Importance of serratia protease in the 28. Klimpel KR, Aurora N, Leppla SH. Anthrax toxin lethal factor pathogenesis of experimental Serratia marcescens pneumonia. contains a zinc metalloprotease consensus sequence which is Infect Immun 1983; 40: 113-119. required for lethal toxin activity. Mol Microbiol 1994; 13: 18. Ghigo J-M, Wandersman C. A fourth metalloprotease gene in 1093-1 100. Erwinia chrysanthemi. Res Microbiol 1992; 143: 857-867. 29. Kooi C, Cox A, Darling P, Sokol PA. Neutralizing monoclonal 19. Toder DS, Ferrell SJ, Nezezon JL, Rust L, Iglewski BH. lasA antibodies to an extracellular Pseudomonas cepacia protease. and lasB genes of Pseudomonas aeruginosa: analysis of Infect Immun 1994; 62: 2811-2817. transcription and gene product activity. Infect Immun 1994; 62: 30. McKevitt AI, Bajaksouzian S, Klinger JD, Woods DE. 1320-1327. Purification and characterization of an extracellular protease Kessler E, Safrin M, Olson JC, Ohman DE. Secreted LasA of 20. from Pseudomonas cepacia. Infect Immun 1989; 57: 771-778. Pseudomonas aeruginosa is a staphylolytic protease. J Biol 31. Laemmli UK. Cleavage of structural proteins during the Chem 1993; 268: 7503-7508. assembly of the head of bacteriophage T4. Nature 1970; 21. Olson JC, Ohman DE. Efficient production and processing of 227: 680-685. elastase and LasA by Pseudomonas aeruginosa require zinc and calcium ions. J Bacteriol 1992; 174: 4140-4147. 32. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of 22. Peters JE, Park SJ, Darzins A et al. Further studies on proteins from polyacrylamide gels to nitrocellulose sheets: Pseudomonas aeruginosa LasA: analysis of specificity. Mol procedure and some applications. Proc Nut1 Acad Sci USA Microbiol 1992; 6: 1155-1 162. 1979; 76: 4350-4354. 23. Loewy AG, Santer UV, Wieczorek M, Blodgett JK, Jones SW, 33. Rinderknecht H, Geokas MC, Silverman P, Haverback BJ. A Cheronis JC. Purification and characterization of a novel zinc- new ultra-sensitive method for determination of proteolytic proteinase from cultures of Aeromonas hydrophila. J Biol activity. Clin Chim Acta 1968; 21: 197-203. Chem 1993; 268: 9071-9078. 34. Jiang W, Bond JS. Families of and their 24. Bairoch A. The PROSITE dictionary of sites and patterns in relationships. FEBS Lett 1992; 312: 11CL114. proteins, its current status. Nucleic Acids Res 1993; 21: 3097- 35. Steward MW, Lew AM. The importance of antibody affinity in -3 107. the performance of immunoassays for antibody. J Immunol 9< LJ. Schiavo G, Poulain B, Rossetto 0, Benfenati F, Tauc L, Methods 1985; 78: 173-190. .