Vol. 175, No. 13JOURNAL OF BACTERIOLOGY, JUly 1993, p. 4008-40150021-9193/93/134008-08$02.00/0Copyright © 1993, American Society for Microbiology

Pseudomonas aeruginosa lasBi Mutants Produce an Elastase,Substituted at Active-Site His-223, That Is Defective in

Activity, Processing, and SecretionKEVIN S. MCIVER,1 JOAN C. OLSON,2 AND DENNIS E. OHMANl*

Department ofMicrobiology and Immunology, University of Tennessee, and Veterans Affairs Medical Center,Memphis, Tennessee 381631 and Department ofPathology and Laboratory Medicine, Medical University of

South Carolina, Charleston, South Carolina 294252

Received 5 January 1993/Accepted 25 April 1993

Pseudomonas aeruginosa secretes elastase in a multistep process which begins with the synthesis of apreproelastase (53.6 kDa) encoded by lasB, is followed by processing to proelastase (51 kDa), and concludes withthe rapid accumulation of mature elastase (33 kDa) in the extracellular environment. In this study, mutants ofP. aeruginosa were constructed by gene replacement which expressed lasBI, an allele altered in vitro at anactive-site His-223-encoding codon. The lasBI allele was exchanged for chromosomal lasB sequences in two strainbackgrounds, FRD2 and PAO1, through a selectable-cassette strategy which placed a downstream Tn501 markernext to lasB1 and provided the selection for homologous recombination with the chromosome. Two lasB1mutants, FRD720 and PDO220, were characterized, and their culture supernatants contained greatly reducedproteolytic (9-fold) and elastolytic (14- to 20-fold) activities compared with their respective parental lasB+ strains.This was primarily due to the effect of His-223 substitution on substrate binding by elastase and thus itsproteolytic activity. However, the concentration of supernatant elastase antigen was also reduced (five- tosevenfold) in the mutant strains compared with the parental strains. An immunoblot analysis of cell extractsshowed a large accumulation of51-kDa proelastase within lasBi mutant cells which was not seen in wild-type cellextracts. A time course study showed that production of extracellular elastase was inefficient in the lasB1 mutantscompared with that ofparental strains. This showed that expression ofan enzymatically defective elastase inhibitsproper processing of proelastase and provides further evidence for autoproteolytic processing of proelastase in P.aeruginosa. Unlike the parental strains, culture supernatants of the lasBi mutants contained two prominentelastase species that were 33 and 36 kDa in size. Extracellular 51-kDa proelastase was barely detectable, eventhough it accumulated to high concentrations within the lasBi mutant cells. These data suggest that productionof an enzymatically defective elastase affects proper secretion because autoproteolytic processing of proelastase isnecessary for efficient localization to the extracellular milieu. The appearance ofreduced amounts of extracellularelastase and their sizes of 33 and 36 kDa suggest that lasB1-encoded elastase was processed by alternate,less-efficient processing mechanisms. Thus, proelastase must be processed by removal of nearly all of the 18-kDapropeptide before elastase is a protein competent for extracellular secretion.

Pseudomonas aeruginosa is an opportunistic pathogenwhich causes a variety of disease manifestations in compro-mised hosts. The ability of P. aeruginosa to secrete severaltoxic and degradative enzymes into the environment is amajor contributor to the pathogenesis of the organism.Elastase is one of several extracellular proteases secreted byP. aeruginosa and is considered a major virulence factor.This is supported by its ability to degrade a number ofbiologically important proteins, including elastin (18), somecollagens (8), immunoglobulins G (2) and A (7), serumal-proteinase inhibitor (19), and complement components(23), and it releases iron bound to transferrin (3).

Elastase is a neutral metalloprotease requiring one zincion per molecule that is essential for its activity and acalcium ion for stability (18). Elastase production and pro-cessing are facilitated by a growth medium containing bothzinc and calcium ions (22). On the basis of the inferred aminoacid sequence (1, 5, 24) and crystallographic structure (25),elastase shares a high degree of sequence and functionalhomology with the zinc metalloprotease thermolysin ofBacillus thermoproteolyticus. These similarities to thermol-

* Corresponding author.

ysin have allowed the prediction of specific residues in-volved in elastase enzymatic activity and substrate binding,as well as zinc and calcium binding.

Kessler and Safrin (9, 10) proposed a model for elastasesecretion, now refined by DNA sequence information, whichinvolves two proteolytic processing steps. Elastase, encodedby lasB, is initially synthesized as a preproelastase with amolecular mass of 53.6 kDa. During translocation throughthe inner membrane, a 2.6-kDa signal sequence is removedto form a 51-kDa proelastase (12). The proelastase is rapidlyprocessed to a 33-kDa mature form by cleavage of an 18-kDaN-terminal propeptide. The model proposes that the propep-tide remains noncovalently associated with a 33-kDaperiplasmic elastase until further processing or dissociationof the complex occurs, which is followed by secretion of themature enzyme through the outer membrane.We have recently shown that overexpression of lasB in

Escherichia coli results in the intracellular accumulation ofprocessed and enzymatically active 33-kDa elastase; how-ever, little 51-kDa proelastase is seen (16). When the codonin lasB encoding His-223, an active-site residue, is changedto encode Asp-223 (lasBl) or Tyr-223 (lasB2), overexpres-sion of these mutant alleles in E. coli results in both loss ofenzymatic activity and accumulation of the unprocessed

4008

on October 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

P. AERUGINOSA lasBi MUTANTS 4009

TABLE 1. Bacterial strains and plasmids

Strain or plasmid Genotype or phenotypea Source or reference

E. coliHB101 proA2 leuB6 thi-1 lacYl hsdR hsdM rec413 supE44 rpsL20 This laboratoryJM109 end4l recA1 gyrA96 thi hsdR17(rk mk+) relA1 supE44 Alac-proAB (F' traD36 proAB Promega

lacIq jM15)

P. aeruginosaPAO1 Prototrophic; lasB+ D. HaasFRD2 Prototrophic; lasB+ 6FRD706 lasB+ TnS01-6 This studyFRD720 lasBi TnS01-6 This studyPDO220 lasBi TnS01-6 This study

PlasmidspBluescript KS- High-copy-number cloning vector; Apr StratagenepKK223-3 Inducible tac promoter expression vector; Apr PharmaciapEMR2 pBR322::TnS cos oriT Apr Kmr 4pEMRZ3 pEMR2 with lacZa multiple cloning DNA at BamHI site; Apr Kmr H. SchweizerRSF1010::TnSOl IncQ Hgr 21pRK2013 ColE1-Tra(RK2)+ Kmr 6pKSM5 pKK223-3 with 2.6-kb P. aeruginosa DNA containing lasBi; Apr 16pKSM15 pLAFR3 (IncP1) with 8-kb EcoRI P. aeruginosa DNA containing lasB; Tcr This studypKSM20 pEMRZ3 with 4.8-kb P. aeruginosa DNA containing lasBi and TnS01-6; Apr Kmr Hgr This studyp720-BAM pEMR2 with 19.2-kb BamHI FRD720 DNA containing lasBi and TnS01-6; Apr Kmr Hgr This study

a Abbreviations for phenotypes: Tcr, tetracycline resistance; Hgr, mercury resistance; Apr, ampicillin resistance; Kmr, kanamycin resistance; Cbr, carbenicillinresistance; Tra, transfer by conjugation.

51-kDa proelastase (16). These results suggest that the rapidprocessing of proelastase to mature elastase is autocatalytic.

In the present study, we developed a gene replacementstrategy to construct defined mutants of P. aeruginosa usinga cloned gene modified in vitro by a single base pair change(e.g., lasB1). To study processing and secretion of anenzymatically defective elastase in the native host, weconstructed mutants of P. aeruginosa with a chromosomallyencoded lasBI allele expressed under its native promoter.These studies showed that modifying the substrate-bindingresidue, His-223, affected not only enzyme activity but alsoproelastase processing and extracellular secretion of elastasein P. aeruginosa. These results indicate that the pathway ofelastase secretion in P. aeruginosa includes autoprocessing.

MATERIALS AND METHODSBacterial strains and plasmids and media. The bacterial

strains and plasmids used in this study are listed in Table 1.Bacteria were cultured in L broth (1% tryptone, 0.5% yeastextract, 0.5% NaCl [pH 7.5]) or a minimal medium (27).Media were solidified with 1.5% Bacto agar (Difco). Unlessotherwise specified, antibiotics were used at the followingconcentrations (per milliliter): ampicillin, 100 ,ug for E. coli;carbenicillin, 300 ,ug for P. aeruginosa; kanamycin, 30 ,ug forE. coli or 500 ,ug for P. aeruginosa; tetracycline, 15 ,ug for E.coli or 100 ,ug for P. aeruginosa; and mercuric chloride, 18,ug for both E. coli and P. aeruginosa. Casein-agar platescontained 1.5% skim milk (Difco) and 0.8% nutrient broth(Difco). Elastin-agar plates contained 0.5% elastin (Sigma)and 0.8% nutrient broth (Difco).DNA manipulations. Routine DNA manipulations and

plasmid extractions were performed as described elsewhere(14). Triparental matings were used to mobilize recombinantplasmids from E. coli to P. aeruginosa as previously de-scribed (6). DNA sequences were determined by the chaintermination technique with Sequenase (U.S. Biochemical) at42°C by using 5'-[a-32P]dCTP (>6,000 Ci/mmol, 10 mCi/ml;

Amersham) and 7-deaza-dGTP. Oligonucleotides used forsequencing primers were synthesized on an Applied Biosys-tems 380B DNA synthesizer in the Molecular ResourcesCenter of the University of Tennessee, Memphis.SDS-PAGE and immunoblotting. Protein samples were

suspended in sodium dodecyl sulfate (SDS) sample buffer (60mM Tris-HCI, 2% SDS, 10% glycerol, 0.1 mg of bromophe-nol blue per ml, 5% 2-mercaptoethanol [pH 6.8]), and loadedonto a 12.5% polyacrylamide gel for polyacrylamide gelelectrophoresis (PAGE) (13). Proteins in polyacrylamidegels were electrotransferred to nitrocellulose in a Trans-Blotapparatus (Bio-Rad) for 2 h at 160 mA and 4°C. Immuno-blotting was performed as previously described (16) withrabbit anti-elastase immunoglobulin G (a gift of E. Kessler)as the primary antibody and then with a goat anti-rabbithorseradish peroxidase conjugate (Sigma).Gene replacement in P. aeruginosa FRD2 with selectable

cassettes. An adjacent Tn5Ol (encoding mercury resistance)was used as a selectable marker to recombine the lasB1mutant allele into the chromosome of P. aeruginosa. Adja-cent to lasB are a 2.2-kb PstI fragment and a 4.3-kb KpnIfragment (Fig. 1). Because TnS01 contains no PstI or KpnIsites, such restriction fragments containing TnS01 can beused as selectable cassettes that can be ligated next to aDNA fragment containing lasBi. pKSM15 (Fig. 1) andplasmids containing fragments of pKSM15 were subjected toTn501 mutagenesis as previously described (21). The relativepositions of insertions, mapped by restriction analysis, areshown on the map of pKSM15 (Fig. 1). The DNA fragmentswith TnS01 insertions were exchanged for chromosomalsequences in FRD2 by a transduction method and withphage F116L as previously described (21), and the strainsconstructed were examined for any defects in proteaseproduction that might occur as a result of the insertion. The2.2-kb PstI and 4.2-kb KpnI fragments containing TnS01-6(8.3 kb) were cloned from pKSM15::TnSOl-6 into pBlue-script KS- to provide a source of these selectable cassettes

VOL. 175, 1993

on October 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

4010 McIVER ET AL.

PLASMID(vector)

pKSM15(pLAFR1)

pBlue/P/Tn6(pBluescript KS-)

pBlue/KiTn6(pBluescript KS-)

pBlue/lasBl(pBluescript KS-)

pBlue/lasBl/iTn6(pBluescript KS-)

pZ3/lasBl(pEMRZ3)

pKSM20(pEMRZ3)

Chromosomal A

R lasB+ X

0 2

p KC f f pI

4 6

PK P2.2 kb+Tn5Ol

K

R lasBI- p

R lasBl

R lasBlcII

R lasBl _b

R A lasB+ I>.v I I

K R

88kb

K.j 4.2 kb+Tn5Ol

2.6kb

P K

P KI 1 2.9 kb

P KI

p

XK-J 7.1 kb+Tn501

R'V

Recombination in FRD2FIG. 1. Restriction maps of P. aeruginosa DNA in plasmids used in this study to generate lasB1 mutants. The lasB1 allele, previously

described (16), contains a single base pair alteration changing the codon for His-223 (CAC) to Asp-223 (GAC). Circles represent Tn501insertions in DNA downstream of lasB. DNA fragments containing lasB1 and the TnS01-6 insertion were ligated together in the genereplacement vector pEMRZ3 to form pKSM20. The bottom of figure illustrates homologous recombination between pKSM20 and thechromosome of FRD2 by selection for mercury resistance (TnS01-6) and coinheritance of lasBi. R, EcoRI; P. PstI; K, KpnI.

(pBlue/P/Tn6 and pBlue/K/Tn6, respectively) (Fig. 1). A2.6-kb EcoRI-PstI fragment containing the lasBi allele wascut from pKSM5 (16) and cloned into pBluescript KS-(pBlue/lasBl). The PstI-TnSOl cassette was cloned in cor-rect orientation into the single PstI site in pBlue/lasB1(pBlue/lasBl/Tn6), from which a 2.9-kb EcoRI-KpnI frag-ment containing lasBi was cloned into the gene replacementvector pEMRZ3 (pZ3/lasBl). The KpnI-TnSOl cassette wascloned in correct orientation into the 1inl site of pZ3/lasBlto form pKSM20 (Fig. 1) for use in the gene replacementprocedures described below.

Wild-type lasB was replaced with the lasBi allele on the P.aeruginosa FRD2 chromosome by the excision marker res-cue method previously described (4), with the followingmodifications. pKSM20 was conjugated into FRD2 by tripa-rental mating, and merodiploid colonies were selected forgrowth on minimal agar containing carbenicillin. Colonieswere pooled, inoculated into 10 ml of L broth containing alow level (3 pg/ml) of HgCl2 to promote expression ofTnS01-encoded mercury resistance, and grown with shakingovernight at 370C to allow for excision of the vector anddiploid sequences by homologous recombination. Dilutionsof the culture were plated onto L agar containing selectablelevels (18 pg/ml) of HgCl2 and incubated at 370C. Colonieswere screened for loss of vector-encoded resistances tokanamycin and carbenicillin. A Southern blot analysis ofdigested genomic DNA obtained from potential mutants,with a 2.6-kb EcoRI-PstI lasB-containing probe, was used toverify single-copy gene replacement (data not shown). Pro-teolytic and elastolytic activities of potential lasBi mutantswere screened on casein and elastin agar plates, respec-tively, and then characterized by more sensitive azocaseinand elastin Congo red assays (described below).

Gene replacement in P. aeruginosa PAO1. A lasBi mutantwas constructed in the PAO strain background as follows.FRD720 (lasBi TnS01-6) genomic DNA was digested withBamHI, ligated into the BamHI site of the gene replacementvector pEMR2, packaged in vitro into X particles (GigapackII X packaging kit; Stratagene), and transduced into E. coliHB101 with selection on L agar containing HgCl2. TnSOJdoes not have a BamHI site. Plasmids from mercury-resis-tant colonies were screened by restriction analysis, and aclone (p720-Bam) was identified containing a single BamHIfragment which included TnS01-6 (8.3 kb) and 19.6 kb of P.aeruginosa DNA. Sequence analysis was used to verify thatthe lasBi mutation and the downstream TnS01-6 werepresent in this clone (data not shown). The p720-Bamconstruct was used to introduce the lasBi mutation into theP. aeruginosa PA01 chromosome by the excision markerrescue method as described above. Gene replacement wasverified through a Southern blot analysis (data not shown).Growth curves and sampling for elastase production. Log-

arithmic-phase (optical density at 600 nm [OD6.] of 0.6)cultures of P. aeruginosa strains were used to inoculate(1:100) L broth (250 ml, 1-liter flask), and then they wereincubated at 370C with maximum aeration. Sample sizeswithdrawn each hour were 1 ml for the first 5 h and 10 ml forthe next 13 h. OD600 was used to approximate the celldensity. Samples were prepared for immunoblot analysisand enzyme assays as follows: a 0.5-ml aliquot was centri-fuged (14,000 x g for 2 min at room temperature), and thepellet was resuspended in SDS sample buffer and incubatedat 100'C for 5 min (cell extract fraction). The remainder ofthe sample was centrifuged (8,000 x g for 10 min at 40C), and2 ml of the resulting supernatant was stored at -70'C untilused in assays for enzyme activity and elastase antigen

J. BACTERIOL.

on October 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

P. AERUGINOSA lasBi MUTANTS 4011

(described below). To prevent further nonspecific proteoly-sis, phenylmethylsulfonyl fluoride (1 mM) and EDTA (5mM) were added to the supernatant remaining, which wasthen concentrated 25-fold in a Minicon B15 unit (Amicon).SDS sample buffer (25 ,ul) was added to each, and sampleswere incubated for 6 min at 950C (supernatant fraction). Cellextract and supernatant samples from each strain wereimmunoblotted with antielastase as described above.

Assays of proteolytic and elastolytic activities. Standardizedcultures of P. aeruginosa strains were used for assays ofenzyme activities. L broth was inoculated (1:100) withovernight cultures, grown to an OD6. of 0.6, and thecultures were used to inoculate (1:100) 10 ml of Luria broth,which was incubated at 370C with aeration. Maximal elastaseaccumulated in the extracellular medium by 18 h. Dilutionsof samples were assayed for proteolytic and elastolyticactivities to establish the linear range of the reactions.Proteolytic activity was determined as previously described(11, 15). Elastolytic activity was determined as previouslydescribed (15, 20).

Quantitation of elastase by ELISA. Relative concentrationsof elastase antigen in culture supernatants were measured bya modified direct-binding enzyme-linked immunosorbent as-say (ELISA) as previously described (22). Briefly, samplesof culture supernatants (stored at -70'C) were diluted 1:50in ELISA coating buffer (5 mM sodium carbonate [pH 9.6]),and 100 pl of each diluted supernatant was added to ELISAwells. The ELISA plate was placed in a 100°C water bath for3 min and then placed at 4°C overnight. Under these condi-tions, the proteolytic activity of elastase on immunoglobu-lins was inhibited; however, elastase maintained the poten-tial to be specifically recognized by antibodies as determinedby reproducible quantitation of various dilutions of superna-tant aliquots (22). Elastase was quantified by adding 100 pu ofa 1:250 dilution of rabbit antielastase serum to each well.Addition of a 1:1,000 dilution of peroxidase-conjugated goatanti-rabbit (Cappel Research) and then of the substrate4-chloro-1-naphthol (Sigma) allowed detection. All assayswere performed in triplicate, with the mean ELISA valuerecorded as OD450 per milliliter of culture supernatant.

RESULTS

Construction of a lasBI mutant of P. aeruginosa FRD2utilizing an adjacent selectable marker. To obtain a selectablemarker next to the lasB gene, DNA downstream of lasB wassubjected to TnS01 mutagenesis, and four distinct sites ofinsertion were mapped (Fig. 1). Since these insertions are inunknown loci that could potentially affect elastase produc-tion, each was exchanged for chromosomal sequences bygene replacement in FRD2. None of the transposon inser-tions shown in Fig. 1 had any apparent effect on expressionof lasB or extracellular proteolytic activity (data not shown).Although our preliminary studies (15) suggested that recom-binants with chromosomal insertions at TnSOJ-1 were ad-versely affected in elastase production, further study showedthat this phenotype was related to a spontaneous mutation(lasC) in the strain's background that was not directlyassociated with the insertions. The nature of this mutation isunder investigation.We previously described the lasBI allele which contains a

single base pair mutation that changed the codon for His-223(a substrate-binding residue) to encode Asp-223, a mutationwhich adversely affects proteolytic activity when expressedin E. coli (16). By a selectable cassette strategy, DNAcontaining the lasBI allele was cloned next to DNA contain-

ing Tn501-6 in the gene replacement vector pEMRZ3 (seeMaterials and Methods). The plasmid formed, pKSM20 (Fig.1), was used in a gene replacement procedure to exchangethe chromosomal wild-type allele in FRD2 for the mutantlasB1 allele with the adjacent TnSOl for selection. Amongthe colonies tested, about 4% had undergone gene replace-ment as evidenced by the presence of TnS01 (mercuryresistance) and sensitivity to vector-encoded markers (car-benicillin and kanamycin). All eight of the potential lasB1mutants obtained exhibited a reduction in extracellular pro-teolytic activity and had barely detectable elastolytic activityon plate assays. One of these mutant strains, designatedFRD720, was chosen for further characterization.To verify that FRD720 contained the lasB1 allele, a

BamHI fragment was cloned from FRD720 genomic DNAwhich conferred mercury resistance in E. coli. p720-Bamwas a 19-kb BamHI fragment of P. aeruginosa DNA fromthe FRD720 chromosome in pEMR2 which containedTnS01-6 and the presumptive lasB1 allele. The mutationcausing the His-223 substitution disrupts a restriction siterecognized by SnoI and ApaLI within the lasB gene codingsequence (16). Digestion of p720-Bam with SnoI (or ApaLI)demonstrated the loss of this restriction site and suggestedthe presence of the lasB1 mutant allele (data not shown).This was confirmed by sequence analysis of DNA thatincluded codon 223 on p720-Bam, which showed the pres-ence of the lasB1 mutation (data not shown).

Construction of a lasBi mutant of P. aeruginosa PAO1 byusing the cloned mutant allele from FRD720. A lasB1 mutantwas also made in the P. aeruginosa PA01 background forcomparison with FRD720 and to control for any strain-dependent phenomena. Gene replacement with pKSM20(lasB1 TnS01-6) in FRD2 was not a frequent event, andlimited attempts to construct lasB1 mutants in PA01 withpKSM20 were unsuccessful. However, p720-Bam (lasB1TnS01-6), described above, was a clone similar to pKSM20and contained another 11 kb of P. aemginosa DNA. Follow-ing conjugation of p720-Bam into PAO1, mercury-resistant(TnS01-6) colonies showed loss of vector-encoded antibioticresistance markers at high frequency, approaching 80%. All80 colonies examined that showed loss of the vector se-quences exhibited reduction in proteolytic activity andbarely detectable elastolytic activity in agar plate assays.One lasB1 mutant of PAO1, designated PDO220, was used inthe subsequent studies.

Proteolytic and elastolytic activities in supernatants oflasB1mutant strains. To determine the consequence of the lasB1mutation on extracellular proteolytic activity in P. aerugi-nosa, supernatants from 18-h standardized cultures wereobtained from the lasB1 mutants (FRD720 and PDO220) andtheir respective wild-type strains (FRD2 and PAO1).FRD706, which contains a chromosomal TnS01-6 insertionand wild-type lasB allele, was also included in this analysisto control for any effects due to the transposon. The hydro-lysis of two substrates was examined: azocasein, for quan-titation of general proteolytic activity, and elastin Congored, for quantitation of elastolytic activity. All three strains(FRD2, FRD706, and PAO1) expressing the wild-type lasBallele exhibited high levels of both proteolytic and elastolyticactivities (data not shown). In contrast, the levels of proteo-lytic activity in the culture supernatants of the two lasB1mutant strains (FRD720 and PDO220) were eight- to ninefoldlower than those observed with their respective wild-typestrains. Compared with the activity of parent strains, elas-tolytic activity was reduced in the lasBI mutants by 14- and20-fold (FRD720 and PDO220, respectively) (data not

VOL. 175, 1993

on October 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

4012 McIVER ET AL.

B.10

o010-

'0To

I-w e70. .,1

.ool

Time (hours)10

Time (hours)FIG. 2. Time course study comparing growth with supernatant elastase concentrations for wild-type and lasBi mutant strains of P.

aeruginosa FRD and PAO. Strains were grown in L broth at 370C with aeration under standardized conditions. Growth (open symbols) isexpressed as culture turbidity determined by A6w. Supernatant elastase concentrations (closed symbols) were determined by ELISA andexpressed as micrograms of elastase per milliliter of supernatant. Time points were taken at hourly intervals from 1 to 18 h for growth and6 to 18 h for supernatant elastase concentrations. (A) Wild-type FRD706 and FRD706 lasBi growth and elastase concentrations; (B) wild-typePA01 and PDO220 lasBi growth and elastase concentrations.

shown). These data supported earlier lasBi expression stud-ies in E. coli which established that the His-223 residue wasimportant for the enzymatic activity of P. aeruginosa elas-tase.Reduced extracellular elastase antigen in lasBi mutant

cultures. The reduction in supernatant proteolytic and elas-tolytic activities observed above with lasBi mutants was notattributable to reduced growth rates. Measurements of cul-ture turbidity, taken at 1-h intervals over an 18-h period ofgrowth, showed no difference between FRD720 (lasBiTnS01-6) and FRD706 (lasB+ TnS01-6) (Fig. 2A [open sym-bols]). Likewise, PDO220 (1asBi TnS01-6) and PAO1 (lasB+)showed the same growth pattern over time (Fig. 2B [opensymbols]) and demonstrated that the TnS01-6 insertion didnot affect growth under these conditions. The aliquots re-moved during the growth analysis were also examined forthe concentration of elastase protein by ELISA. The twowild-type strains, FRD706 (Fig. 2A) and PA01 (Fig. 2B),exhibited a biphasic expression of elastase antigen, with aninitial burst between 6 and 9 h and then a plateau inproduction between 10 and 12 h, which was followed by arapid rise in elastase concentration through the final timepoint at 18 h. In contrast, cultures of the lasBi mutants(FRD720 and PDO220) demonstrated a much slower rise inelastase concentration, and 18-h supernatants containedapproximately five- and sevenfold less elastase antigen,respectively, compared with the amounts of their wild-typestrains (Fig. 2).

Effect of lasBI mutation on proelastase processing andsecretion. Expression of wild-type lasB in E. coli resulted intranslocation of some mature (33-kDa), enzymatically activeelastase to the periplasm, although this heterologous host (E.coli) was unable to secrete elastase to the extracellularmedium (16). The product of lasB1 in E. coli was defectivenot only in activity but also in processing and translocation

to the periplasm (16). Here, we examined the potential for asimilar lasBl-mediated defect in P. aeruginosa which af-fected both processing and translocation. A defect in trans-location, in this case, to the extracellular medium wouldexplain the reduced extracellular elastase antigen in culturesof lasBI mutants described above. To test this, cell extractsand supernatant fractions were taken during the time coursestudy described above and analyzed for elastase antigen byimmunoblot analysis. Cell extracts of the lasB+ strains,FRD706 (Fig. 3A) and PAO1 (Fig. 4A), showed mature-size(33-kDa) elastase both within the cell and localized to thesupernatants (Fig. 3B and 4B) at all time points. Thecell-bound elastase species (33 kDa) from both wild-typestrains appeared to form a doublet, which suggests thepossibility of two intracellular species, a feature previouslynoted by Kessler and Safrin (9, 10). In the lasBI mutants,proelastase (51 kDa) was the dominant species in cell ex-tracts of FRD720 (Fig. 3C) and PDO220 (Fig. 4C), indicatinga defect in processing. Both lasBi mutants accumulatedlarge amounts of 51-kDa proelastase within the cell as earlyas 6 h. In these P. aeruginosa lasBi extracts, many sizes ofproelastase breakdown products were observed (Fig. 3C and4C), including a 33-kDa form, which was similar to that seenin our previous E. coli expression studies (16) and whichsuggests that the lasBi 51-kDa proelastase was susceptibleto general proteolytic digestion. Interestingly, the elastaseappearing in the lasB1 mutant culture supernatants was oftwo sizes, a 33-kDa mature-size species and a novel 36-kDaelastase species.

DISCUSSION

The secretion of P. aeruginosa elastase is a multistepprocess which begins with the synthesis of a 53.6-kDapreproelastase and results in the rapid accumulation of

A.

00ToPo

I.-U

U)(A

J. BACTERIOL.

on October 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

P. AERUGINOSA lasBI MUTANTS 4013VOL. 175, 1993

A. FRD706 Cell Extracts6 7 a 9 10 11

C. FRD720 Cell Extracts

51-kDa -

B. FRD706 Supernatants D. FRD720 Supematants

*< 33-kDaX->

FIG. 3. Immunoblots of samples taken in a time course study showing elastase-related proteins in cell extracts and supernatants ofwild-type and lasBi mutant strains of P. aenrginosa FRD. The number above each lane corresponds to the time point (hours) at which eachsample was taken and can be directly compared with Fig. 2. Positions corresponding to the 33-kDa mature elastase and the 51-kDa proelastaseare indicated. (A) FRD706 lasB+ cell extract samples; (B) FRD706 lasB+ supernatant samples; (C) FRD720 lasBi cell extract samples; (D)FRD720 lasBl supernatant samples.

mature 33-kDa elastase in the extracellular environment. Inan effort to better understand this pathway, we altered theactive-site His-223-encoding codon in the structural gene forelastase to form the lasBI allele to study its plasmid-borneexpression in E. coli (16). In the present study, we con-structed P. aeruginosa strains which expressed the lasBIallele from the chromosome to examine the effect of thisenzymatic defect on proelastase processing and secretion inthe native organism. The lasBI allele was exchanged forwild-type lasB sequences in two strain backgrounds, FRD(FRD720) and PAO (PDO220). This was accomplishedthrough a selectable-cassette strategy which placed a down-stream TnS01 marker next to lasBI and provided the selec-tion pressure needed for homologous recombination with thechromosome. In general, the strategy employed here al-lowed the introduction of a defined, single-base pair alter-ation into the chromosome ofP. aeruginosa and should havegeneral application to other genetic studies ofP. aeruginosa.

Supernatants from the two lasBi mutant strains, FRD720and PDO220, contained greatly reduced proteolytic (9-fold)and elastolytic (14- to 20-fold) activities compared with thoseof their respective wild-type parent strains. In part, thiscould be attributed to the low concentration of supernatantelastase, as detected by ELISA, which was reduced by five-to sevenfold in the mutant strains (FRD720 and PDO220,respectively) compared with parental strains grown underthe same conditions. However, reduction in proteolytic andelastolytic activities in the supernatants of lasBi mutantswas primarily due to the His-223 substitution, which affectssubstrate binding by elastase and thus its proteolytic activ-ity. This is in agreement with our previous studies with lasBiexpression in E. coli, which resulted in almost total loss of

both proteolytic and elastolytic activities (16). There areother proteases in the supernatant of P. aeruginosa, includ-ing alkaline protease and LasA, which have been reported toexhibit some elastolytic activity (26, 28), and they mayaccount for the residual elastolytic activity. Determinationof the specific proteolytic activity of lasBi elastase, purifiedfrom mutant culture supernatants, is in progress and willshow whether any residual enzymatic activity remains fol-lowing a substitution at His-223.The five- to sevenfold reduction in the level of elastase

produced by the lasBl mutants was not attributable to anydetectable growth defect, since the two mutants demon-strated growth patterns comparable to those of the wild-typestrains. This was also not due to reduced rates of lasBtranscription in the mutants; a lasB-cat operon fusion,constructed in a low-copy-number plasmid, produced chlor-amphenicol acetyltransferase levels that were almost identi-cal in all strains (17). However, the immunoblot analysis ofcell extracts showed a large accumulation of the 51-kDaproelastase form in lasBI mutant cells which was not seen inwild-type cell extracts. This is further evidence that expres-sion of an enzymatically defective elastase inhibits properprocessing of proelastase. We recently have shown thatcultures of wild-type P. aeruginosa deprived of zinc andcalcium ions, which are required for elastase enzymaticfunction, also result in an accumulation of proelastase (22).It was of interest that the accumulating proelastase in P.aeruginosa lasBI cell extracts generally appeared to degradein a nonspecific fashion, although much of the products wasfound in a stable 33-kDa form. Even 51-kDa lasBiproelastase that was overproduced in E. coli, which alsounderwent nonspecific proteolysis, formed significant

on October 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

4014 McIVER ET AL.

A. PA01 Cell Extracts C. PDO220 Cell Extracts6 7 8 9 10 11

51-kDa -*

*- 33-kDa -*

B. PA01 Supematants D. PDO220 Supernatants6 7 8 910 11 12 14 16 18 6789 01112141618

51-kDa->

if~~~~~~~< 33-kk--

FIG. 4. Immunoblots of samples taken in a time course study showing elastase-related proteins in cell extracts and supernatants ofwild-type and lasBi mutant strains of P. aeruginosa PAO. The number above each lane corresponds to the time point (hours) at which eachsample was taken and can be directly compared with Fig. 2. Positions corresponding to the 33-kDa mature elastase and the 51-kDa proelastaseare indicated. (A) PA01 lasB+ cell extract samples; (B) PAO1 lasB+ supernatant samples; (C) PDO220 lasBi cell extract samples; (D)PDO220 lasBi supernatant samples.

amounts of 33-kDa mature-size elastase (16). These resultssuggest that the lasBi protease may have residual autopro-cessing enzymatic activity or that the normal processing sitein proelastase may be a preferred site for other proteaseswithin the cell, thus generating the 33-kDa elastase. Thesepossibilities are under investigation.

Extracellular secretion of elastase by the wild-type strainswas efficient, with 33-kDa mature elastase appearing inrelatively large amounts by 6 h and continuing to accumulatethrough the 18-h time point. This was consistent with theELISA results, which showed a rapid increase in elastasesupernatant concentration well into the stationary phase(Fig. 2). On the other hand, the mutant strains lagged behindin their ability to secrete elastase across the outer mem-brane, with significant amounts not appearing until 9 h. Inaddition, there were clearly two elastase species found inlasBi mutant supernatant fractions, an approximately 36-kDa form as well as the mature-size 33-kDa form. The33-kDa form was detectable first, and the larger speciesbegan to appear soon afterward. Interestingly, the superna-tant fractions of lasBi mutants contained only barely detect-able amounts of 51-kDa proelastase, even though it accumu-lated to significant amounts within the cell. These datasuggest that production of an enzymatically defective elas-tase perturbs secretion because proelastase must be pro-cessed for efficient extracellular secretion to take place. Theappearance of the extracellular 33- and 36-kDa speciessuggests that nearly all of the propeptide must be removedbefore elastase becomes a protein that is competent forsecretion. Thus, the propeptide may contain sequenceswhich, under most circumstance, prevent exoproteins fromtraversing the membrane. In vitro degradation experimentswere performed with trypsin on lasBi proelastase produced

by E. coli, and a similar 36-kDa protein product was ob-served (data not shown). Thus, there may be a generalprotease cleavage site upstream of the normal maturationsite, at which other proteases in the cell can cleaveproelastase. This may provide a less-efficient secondarypathway for processing and secretion of an inactiveproelastase. Both the smaller 33- and the 36-kDa specieswere secreted across the outer membrane, suggesting thatsecretion through the outer membrane can proceed onceprocessing has occurred by either method.

ACKNOWLEDGMENTS

We thank Efrat Kessler for the kind gift of anti-elastase immuno-globulin and helpful discussions. Excellent technical assistance wasprovided by Iulia Kovari at the Molecular Resources Center,University of Tennessee, Memphis, in oligonucleotide synthesis.

This work was supported by Public Health Service grant AI-26187from the National Institute of Allergy and Infectious Diseases(D.E.O.) and Medical University of South Carolina InstitutionalResearch Funds (J.C.O.).

REFERENCES1. Bever, R. A., and B. H. Iglewski. 1988. Molecular characteriza-

tion and nucleotide sequence of the Pseudomonas aeruginosaelastase structural gene. J. Bacteriol. 170:4309-4314.

2. Doring, G., J. Obernesser, and K Botzenhart. 1981. Extracel-lular toxins of P. aeruginosa. II. Effect of two proteases onhuman immunoglobulins IgG, IgA and secretory IgA. Zentralbl.Bakteriol. Mikrobiol. Hyg. I Abt. Orig. A 249:89-98.

3. Doring, G., M. Pfestorf, K. Botzenhart, and M. Abdallah. 1988.Impact of proteases on iron uptake ofPseudomonas aeruginosapyoverdin from transferrin and lactoferrin. Infect. Immun.56:291-293.

4. Flynn, J. L., and D. E. Ohman. 1988. Use of a gene replacement

J. BACTERIOL.

on October 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/

P. AERUGINOSA lasBl MUTANTS 4015

cosmid vector for cloning alginate conversion genes from mu-coid and nonmucoid Pseudomonas aeruginosa strains: aigScontrols expression of algT. J. Bacteriol. 170:3228-3236.

5. Fukushima, J., S. Yamamoto, K. Morihara, Y. Atsumi, H.Takeuchi, S. Kawamoto, and K. Okuda. 1989. Structural geneand complete amino acid sequence ofPseudomonas aeruginosaIFO 3455 elastase. J. Bacteriol. 171:1698-1704.

6. Goldberg, J. B., and D. E. Ohman. 1984. Cloning and expressionin Pseudomonas aeruginosa of a gene involved in the produc-tion of alginate. J. Bacteriol. 158:1115-1121.

7. Heck, L. W., K. Morihara, and D. R. Abrahamson. 1986.Degradation of soluble laminin and depletion of tissue-associ-ated basement membrane laminin by Pseudomonas aeruginosaelastase and alkaline protease. Infect. Immun. 54:149-153.

8. Heck, L. W., K. Morihara, W. B. McRae, and E. J. Miller. 1986.Specific cleavage of human type III and IV collagens byPseudomonas aeruginosa elastase. Infect. Immun. 51:115-118.

9. Kessler, E., and M. Safrin. 1988. Partial purification and char-acterization of an inactive precursor of Pseudomonas aerugi-nosa elastase. J. Bacteriol. 170:1215-1219.

10. Kessler, E., and M. Safrin. 1988. Synthesis, processing, andtransport of Pseudomonas aeruginosa elastase. J. Bacteriol.170:5241-5247.

11. Kessler, E., M. Safrin, N. Landshman, A. Chechick, and S.Blumberg. 1982. In vitro inhibition ofPseudomonas aeruginosaelastase by metal-chelating peptide derivatives. Infect. Immun.38:716-723.

12. Kessler, E., M. Safrin, M. Peretz, and Y. Burstein. 1992.Identification of cleavage sites involved in proteolytic process-ing of Pseudomonas aeruginosa preproelastase. FEBS Lett.299:291-293.

13. Laemmli, U. K. 1970. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature (London)227:680-685.

14. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

15. McIver, K. S., E. Kessler, and D. E. Ohman. 1992. Expressionof elastase in Pseudomonas aeruginosa by chromosomal lasBand mutant alleles requires the downstream gene lasC. Abstr.Annu. Meet. Am. Soc. Microbiol. 1992 D-50, p. 104.

16. McIver, K. S., E. Kessler, and D. E. Ohman. 1991. Substitu-tion of active-site His-223 in Pseudomonas aeruginosa elastaseand expression of the mutated lasB alleles in Escherichia coli

show evidence for autoproteolytic processing of proelastase. J.Bacteriol. 173:7781-7789.

17. McIver, K. S., and D. E. Ohman. 1992. Unpublished data.18. Morihara, K., and J. Y. Homma. 1985. Pseudomonas proteases,

p. 41-79. In I. A. Holder (ed.), Bacterial enzymes and virulence.CRC Press, Inc., Boca Raton, Fla.

19. Morihara, K., H. Tsuzuki, and K. Oda. 1979. Protease andelastase of Pseudomonas aeruginosa: inactivation of humana1-proteinase inhibitor. Infect. Immun. 24:188-193.

20. Ohman, D. E., S. J. Cryz, and B. H. Iglewski. 1980. Isolationand characterization of a Pseudomonas aeruginosa PAO mutantthat produces altered elastase. J. Bacteriol. 142:836-842.

21. Ohman, D. E., M. A. West, J. L. Flynn, and J. B. Goldberg.1985. Method for gene replacement in Pseudomonas aeruginosaused in construction of recA mutants: recA-independent insta-bility of alginate production. J. Bacteriol. 162:1068-1074.

22. Olson, J. C., and D. E. Ohman. 1992. Efficient production andprocessing of elastase and LasA by Pseudomonas aeruginosarequire zinc and calcium ions. J. Bacteriol. 174:4140-4147.

23. Schultz, D. R., and K. D. Miller. 1974. Elastase ofPseudomonasaeruginosa: inactivation of complement components and com-plement-derived chemotactic and phagocytic factors. Infect.Immun. 10:128-135.

24. Tanaka, E., S. Kawamoto, J. Fukushima, K. Hamajima, H.Onishi, Y. Miyagi, S. Inami, K. Morihara, and K. Okuda. 1991.Detection of elastase production in Escherichia coli with theelastase structural gene from several non-elastase-producingstrains of Pseudomonas aeruginosa. J. Bacteriol. 173:6153-6158.

25. Thayer, M. M., K. M. Flaherty, and D. B. McKay. 1991.Three-dimensional structure of the elastase of Pseudomonasaeruginosa at 1.5-A resolution. J. Biol. Chem. 266:2864-2871.

26. Todor, D. S., M. J. Gambello, and B. H. Iglewski. 1991.Pseudomonas aeruginosa LasA: a second elastase under thetranscriptional control of lasR. Mol. Microbiol. 5:2003-2010.

27. Vogel, H. J., and D. M. Bonner. 1956. Acetylornithinase inEscherichia coli: partial purification and some properties. J.Biol. Chem. 218:97-106.

28. Wolz, C., E. Hellstern, M. Haug, D. R. Galloway, M. L. Vasil,and G. Doring. 1991. Pseudomonas aeruginosa LasB mutantconstructed by insertional mutagenesis reveals elastolytic activ-ity due to alkaline proteinase and the LasA fragment. Mol.Microbiol. 5:2125-2131.

VOL. 175, 1993

on October 13, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

http://jb.asm.org/