Cloning of Genes Encoding Extracellular Metalloproteases from ...

3 downloads 0 Views 2MB Size Report
encoding a metalloprotease inhibitor as well as three tandemprt genes encoding metalloproteases. ... were separated from the inhibitor gene by a ca. 4-kilobase ...
JOURNAL

OF

Vol. 172, No. 10

BACTERIOLOGY, OCt. 1990, p. 5803-5815

0021-9193/90/105803-13$02.00/0 Copyright © 1990, American Society for Microbiology

Cloning of Genes Encoding Extracellular Metalloproteases from Erwinia chrysanthemi EC16 G. S. DAHLER, F. BARRAS,t AND N. T. KEEN*

Department of Plant Pathology, University of California, Riverside, California 92521 Received 26 March 1990/Accepted 16 July 1990

A 14-kilobase BamHI-EcoRI DNA fragment cloned from Erwinia chrysanthemi EC16 contained a gene encoding a metalloprotease inhibitor as well as three tandem prt genes encoding metalloproteases. The prt genes were separated from the inhibitor gene by a ca. 4-kilobase region that was necessary for extracellular secretion of the proteases. When individually subcloned downstream from vector promoters, the three prt genes each led to substantial extracellular secretion of the proteases by Escherichia coli cells, provided that the 4-kiobase required region was supplied in cis or trans. One of the protease structural genes, prIC, was sequenced and had high homology to a metalloprotease gene previously described from Serratia species as well as to the prtB gene of E. chrysanthemi B374. Marker exchange mutants of E. chrysanthemi EC16 defective in production of one or all of the extracellular proteases were not impaired in virulence on plant tissue.

Erwinia spp. cause soft rot diseases of various plants and also associated with blight and wilt diseases. Erwinia chrysanthemi produces several pectate lyase proteins which have been associated with virulence in plants (5, 7, 19). However, these bacteria also typically secrete several other degradative enzymes, including proteases. A metalloprotease from Erwinia carotovora may be involved in virulence on potato (1; S. Kyostio and G. Lacy, Annu. Meet. Am. Phytopathol. Soc., Richmond, Va., August, 1989, abstr. no. 518). Three different strains of E. chrysanthemi have also been reported to produce one or more proteases, and the genes have been cloned and expressed in Escherichia coli (3, 44, 45). Significantly, these proteases were exported directly to the medium by E. chrysanthemi and by E. coli cells carrying the cloned genes. Further, the production of these proteases as well as that produced by E. carotovora (Kyostio and Lacy, Annu. Meet. Am. Phytopathol. Soc., 1989) was induced when the bacteria were grown in minimal medium containing various proteins or polygalacturonic acid. In view of these findings, we became interested in further characterizing the protease genes of E. chrysanthemi EC16 (3) and testing their involvement in bacterial virulence on plant tissue.

DNA manipulations. The cloning plasmids used are described in Table 1. Standard techniques (23) were generally used in the construction of the recombinant plasmids shown in Table 1. Subcloning was usually done from soft agarose gels (8). Plasmid DNA was extracted from E. coli DH5a cells by the alkaline lysis method and purified by two rounds of CsCl density gradient centrifugation. Restriction enzymes were generally obtained from New England BioLabs, Inc., (Beverly, Mass.), and digestions were generally done with salts recommended by the supplier or lx KAB salts (a modification of the recipe of McClelland et al. [25] in which potassium glutamate was replaced by the same molarity of potassium acetate, but all other reagents remained the same). Competent E. coli cells were usually prepared by growing cells on L broth at 37°C to ca. 0.4 absorbance unit at 600 nm, cooling the culture to 0°C, and directly adding an equal volume of 2x TSS (6). The cells were then frozen on dry ice and stored at -80°C until needed for transformations. A library of total genomic DNA of E. chrysanthemi UM1005 (a mutant of strain EC16 devoid of four genes encoding pectate lyases [31]) was prepared by Scott Gold in the cosmid vector pLAFR3 (37). The Erwinia DNA was partially restricted with Sau3A and sized on sucrose gradients to yield ca. 30-kilobase (kb) fragments which were ligated with vector arms prepared by the method of Staskawicz et al. (37). Packaged cosmids were transfected into E. coli JA-221, and resultant colonies were maintained on LB plates supplemented with tetracycline at 15 ,ug/ml. Colonies were screened for extracellular protease production by plating on YC gelatin plates. TnS mutants were developed by the use of lambda::TnS mutagenesis done on pGSD1 in E. coli cells by the method of Ruvkun and Ausubel (33). Kanamycin- and ampicillin-resistant colonies were screened for alterations in protease production on gelatin plates. The location of TnS was mapped in several mutants altered in protease secretion as well as several randomly selected mutants that appeared to have normal protease production. Marker exchange mutants. Appropriate cloned DNAs with insertions of a 1.7-kb PvuII fragment containing the Tn9O3 nptII gene (27) were made as described in Table 1. The inserts from these constructs were recloned into pRK415 (Table 1), and the resulting plasmids were conjugated into

are

MATERIALS AND METHODS Bacterial strains and culture conditions. The bacterial strains utilized are described in Table 1. E. chrysanthemi and E. coli DHSt cells were grown on LB or M9 medium (23), with thiamine added to M9 medium at 4 ,ug/ml for growth of DH5ca. Glycerol at 0.4% was generally used as the carbon source in M9 medium instead of glucose. YC medium (17) was used with 1% gelatin and 1.5% agar (Difco Laboratories, Detroit, Mich.) added before autoclaving. E. chrysanthemi or E. coli cells carrying various recombinant plasmids were placed onto this medium with toothpicks, and the plates were incubated for 12 or more hours at 37°C. Clear zones denoting extracellular protease activity were present around positive colonies after the plates were flooded with 15% HgCl2-20% HCl. Corresponding author. t Present address: CNRS, Marseilles, France. *

5803

5804

DAHLER ET AL.

J. BACTERIOL. TABLE 1. Bacterial strains, bacteriophage, and plasmids used

Bacterium, phage, or plasmid

Description

Source or reference

Erwinia chrysanthemi EC16 EC16RI UMlOOS M-1 M-2 M-3

Wild type Spontaneous Rif' mutant EC16 pectate lyase mutant nptlI insertion in the required region nptII insertion mutant strain lacking prtC Deletion mutant lacking prtA, prtB, prtC

A. Chatterjee This work 31 This work This work This work

Escherichia coli DH5a

F- lacZA M15 endAl hsdRJ7 supE44 thi-1 gyrA relAl

Bethesda Research Laboratories Inc.

A(lac-proAB) thi supE44 A(srl-recA)306::TnlO (Tet)(F' traD36 proAB lacZAM15) hsdR trpES leuB6 lac Y recA thi F' (lacJq lacZ+ lac rY lacA+ proA+ proB+)

42 24

Helper phage

42

MV1193 JA221

Phage (M13KO9) Plasmids pUC118 and -119 pUC128 and -129 pMTL20 to -23p pRK415 pDSK519 pJRD184 pGSDO-P6 pGSD1 pGSD2N

pGSD2S pGSD2P pGSD2PK

pGSD2PKT pGSD3 pGSD3K pGSD4 pGSD4K pGSD5 pGSD6 pGSD9

pGSD11 pGSD12 pGSD13

pGSD13ARV pGSD14

pGSD15 pGSD16 pGSD17

(Gaithersburg, Md.)

42 Cloning plasmids 18 Cloning plasmids 4 Cloning plasmids 18 Cloning plasmid 18 Cloning plasmid 14 Cloning plasmid pLAFR3 cosmic clone of R. chrysanthemi EC16 (UM1005) DNA directing extracellular This work protease production in E. coli This work Ca. 14-kb BamHI-EcoRI subclone from pGSDO-P6 cloned in pUC119 This work pGSD1 with ca. 2.5-kb internal NsiI fragment deleted and the plasmid religated; extracellular protease negative This work pGSD1 with ca. 1.7-kb internal SmaI fragment deleted and the plasmid religated; extracellular protease positive This work pGSD1 with ca. 3-kb internal SmaI-PmIl fragment deleted and the plasmid religated; extracellular protease negative in E. coli This work pGSD1 with ca. 3-kb internal SmaI-Pmll fragment deleted and replaced with 1.7-kb PvuII fragment from pDSK519 carrying the nptII gene; E. coli cells carrying this plasmid were extracellular protease negative and kanamycin resistant This work BamHI-EcoRI insert fragment of pGSD2PK recloned into the same sites of pRK415 This work Ca. 2.2-kb SspI-EcoRI fragment from pGSD1 was blunted with Si nuclease and recloned in SmaI site of pUC119 such that the coding region of prtC was downstream of the vector promoter This work Insert fragment from pGSD3 was recloned in pDSK519 such that the coding region of prtC was downstream of the vector promoter This work Same as pGSD3, except that the prtC insert was oriented anti- to the vector promoter This work Insert from pGSD4 was recloned into pDSK519 such that prtC was oriented backward to the vector lac promoter; protease positive This work Ca. 9.5-kb BamHI-SmaI fragment from pGSD1 cloned in pUC119; protease negative This work Insert from pGSD5 cloned in pRK415 This work Ca. 7.5-kb BglII-EcoRI fragment from pGSD1 cloned into the same sites of pMTL21p; protease positive This work 1.9-kb ClaI-NsiI fragment from pGSD5 cloned into the same sites of pUC128; directs high protease inhibitor activity This work 1.9-kb ClaI-NsiI fragment from pGSD5 cloned into the same sites of pUC129; directs moderate inhibitor activity This work Ca. 8-kb MluI fragment from pGSD1 cloned into the same site of pMTL23p downstream of the vector promoter; protease positive Partial EcoRV deletion of pGSD13, removing a terminal 1.7-kb EcoRV fragment This work encoding prtB but retaining prtA Same as pGSD13 except opposite insert orientation in pMTL23p; weak protease This work activity Ca. 3.7-kb StuI-MluI fragment from pGSD13 cloned into the same sites of pMTL23p This work such that the prtB gene was oriented downstream of the vector promoter; protease positive in E. coli Same as pGSD15 except opposite orientation relative to the vector promoter of This work pMTL22p; weakly protease positive in E. coli 1.7-kb SspI fragment from pGSD15 cloned into the EcoRV site of pUC128 such that the This work coding sequence was downstream of the vector promoter; protease positive in E. coli Continued on following page

VOL. 172, 1990

CLONING OF METALLOPROTEASE GENES FROM E. CHRYSANTHEMI

5805

TABLE 1-Continued Bacterium, phage, or plasmid

pGSD18

pGSD20 pGSD21 pGSD23

pGSD25 pGSD26

pGSD27 pGSD28

Description Same as pGSD17 except the opposite orientation was cloned in pUC129; weakly protease positive in E. coli Ca. 5.2-kb EcoRV-(SmaI) EcoRI fragment from pGSD5 cloned into the BamHI-EcoRI sites of pUC129; protease negative and relatively unstable Ca. 5.2-kb EcoRV-(SmaI) EcoRI fragment of pGSD5 cloned into the SspI-EcoRI sites of pJRD184; contains the required region but not the protease inhibitor; protease negative Ca. 6.3-kb BglII-MluI fragment from pGSD9 cloned into the same sites of pMTL21p so that the prtA and prtB genes were oriented downstream of the vector promoter 1.0-kb PstI-SmaI fragment from pGSD13ARV cloned into the PstI-SmaI sites of pUC129; protease negative 1.0-kb PstI-SmaI fragment from pGSD13ARV cloned into the PstI-EcoRV sites of pUC129; protease negative 2.2-kb PvuII-PstI fragment from pGSD13 ARV cloned into the EcoRV-PstI sites of pGSD25 such that the reconstructed prtA gene was oriented downstream of the vector lac promoter; protease positive Same as pGSD27 except that the 2.2-kb PvuII-PstI fragment was cloned into the SmaIPstI sites of pGSD26 such that the prtA gene was in the opposite orientation from the vector lac promoter; weakly protease positive

EC16RI. Cells were grown in 15 ml of phosphate starvation medium (32) or in liquid LB medium for 48 h at 28°C to cure plasmids. Cells from the cultures were then plated on L agar plates supplemented with kanamycin, and the resultant colonies were screened for those that were sensitive to tetracycline, indicating loss of the recombinant pRK415 plasmids and putative replacement of the wild-type DNA by homologous recombination. Total genomic DNA was prepared from wild-type EC16RI (Table 1) and from the various putative marker exchange mutants as previously described (40). These DNAs were restricted with appropriate enzymes and run out on agarose gels. The DNA was then blotted onto nylon sheets which were probed with desired DNA fragments. Protease assays and gel electrophoresis. Protease activity produced by E. chrysanthemi EC16 or by E. coli cells carrying various cloned DNA fragments was routinely determined by growing the bacteria on YC gelatin plates for 16 to 48 h and then flooding the plates with the mercuric chloride-hydrochloride solution noted above. Diameters of halos surrounding the colonies were determined as a semiquantitative measure of the protease activity produced. The protease activity of bacterial culture fluids (after adjustment to pH 7.5) was determined by using hide powder azure (Calbiochem-Behring, San Diego, Calif.). Suitably diluted culture fluid (0.6 ml) was mixed with 50 mM potassium phosphate (pH 7.5) (0.6 ml), and 0.3 mg of hide powder azure was added. Reactions were run in standard microcentrifuge tubes, whose contents were mixed and incubated at 37°C. At various time intervals, the contents of the tubes were mixed and the insoluble substrate was allowed to settle for 1 min before determination of absorbance of the supernatant fluids at 595 nm. After a delay of 20 to 40 min in tubes with relatively low activity, the release of dye was linear, and data are reported as the rate of change in A595 per hour per milliliter of bacterial culture fluid assayed. Bacterial culture fluids were passed through 0.22-p,mpore-size filters and then concentrated from 10 to 40 times with Centricon 10 microconcentrators (Amicon Corp., Lexington, Mass.). These preparations were examined by denaturing (2) or nondenaturing (3) polyacrylamide gel electrophoresis. The gels were either stained with Coomassie brilliant blue R250 (20) or overlaid with a layer of 1%

Source or reference

This work

This work This work This work

This work This work

This work This work

agarose-1% gelatin in 10 mM Tris hydrochloride (pH 7.4)

and incubated at 28°C for 1 to 4 h. Bands of activity in the overlays were visualized by flooding with HgCl2-HCl as described above. Proteins in denaturing gels were renatured before assay by washing the gels in 2.5% Triton X-100-10 mM Tris hydrochloride (pH 7.4) for 2 h to remove sodium dodecyl sulfate (35). Periplasmic fractions from bacteria were prepared by the method of Witholt et al. (47) and dialyzed against 0.01 M Tris hydrochloride (pH 8.0). Pelleted spheroplasts were lysed by gentle sonication in 0.1 M Tris hydrochloride (pH 7.5) (25 W for 10 s) and dialyzed as for periplasmic fractions. These preparations were assayed for protease activity, which was compared with that of the corresponding culture fluids. DNA sequencing. The insert DNA of plasmids pGSD3 and pGSD4 (Table 1), containing the cloned prtC gene, was sequenced by the methods reported by Tamaki et al. (38). Appropriate exonuclease III deletions were prepared, and single-stranded DNA was isolated after transformation of deletion plasmids into E. coli MV1193 and superinfection with M13K07 (42). These templates were then sequenced by the dideoxy method (34) with the E. coli DNA polymerase I large fragment or the Sequenase kit (United States Biochemical Corp., Cleveland, Ohio). Both strands were entirely sequenced, and data were analyzed by using the computer program of Pustell and Kafatos (30). Pathogenicity tests. Cylinders of potato tuber tissue (cv. Russet Burbank) were removed with a no. 5 cork borer, and 5-mm-thick slices were prepared. Ten slices were placed into sterile petri dishes containing filter paper disks and 18 ml of distilled water. Bacterial culture fluids or log-phase cultures growing on LB medium (ca. 0.4 absorbance unit at 600 nm) were applied (20 ,ul) to the surface of the tuber disks, and they were incubated for 20 h or more at 280C. Macerated tissue was removed with a spatula and quantitated by weighing. In some experiments, bacteria in the macerated tissue were suspended in LB medium and populations were determined by dilution plating on LB agar supplemented with appropriate antibiotics. Unless otherwise indicated, Dendranthema grandiflora Tzvelev. (= D. morifolia Ramat. = Chrysanthemum morifolia Ramat.) plants were grown at 30°C with continuous illumination in a peat moss-sandy loam mixture. Bacteria for

5806

DAHLER ET AL.

inoculation were taken from LB plates incubated at 28°C for 48 h or less. Cuttings were surface sterilized in 3% Clorox and, after several rinses in sterile water, were trimmed to approximately 6 cm in length. The tip of a toothpick was infested with bacterial cells and inserted several millimeters into the base of each cutting. The cuttings were then placed into a test tube of water and enclosed in a plastic bag at 26°C. Whole plants were inoculated by placing a mass of bacterial cells on the stem below the youngest fully expanded leaf or equivalent distance from the apex. The stem was then pierced with the tip of a 26-guage needle to penetrate the vascular bundle. Plants were placed in plastic bags and incubated at 30°C. Disease symptoms in cuttings and whole plants were rated by splitting the stems open with a razor blade and measuring the extent of visible pith and vascular necrosis (29). RESULTS Cloning of EC16 DNA encoding protease activity. The EC16 UM1005 cosmid library in E. coli JA221 was screened for extracellular protease activity on gelatin plates, and 9 positive clones which all exhibited clear halos of ca. 1-cm diameter surrounding the colonies were isolated from ca. 2,000 colonies screened. All the positive clones were found to contain a ca. 16-kb EcoRI fragment, presumed to be the same as previously observed to direct protease production (3). One of these EcoRI fragments was further subcloned as a ca. 14-kb BamHI-EcoRI fragment, resulting in pGSD1 (Table 1; Fig. 1). The insert fragment of this plasmid was mapped with several restriction enzymes to facilitate further subcloning (Fig. 1). pGSD1 and pGSD5 yielded abnormally low amounts of plasmid DNA from E. coli DH5a, indicating that these plasmids were unstable or occurred at reduced copy number. Plasmid yield in these cases was improved by the addition of 0.3% glucose to LB medium for growth of DH5a cells before plasmid extraction and purification. Two TnS insertions into the insert DNA of pGSD1 destroyed extracellular protease production in E. coli; both of them mapped to the central portion of the insert (Fig. 1). One TnS insertion which mapped ca. 3.2 kb from the BamHI end of the insert did not affect activity. A recombinant clone in which the nptIl gene of Tn9O3 was cloned into one of the PvuII sites of pGSD1 also completely eliminated activity (Fig. 1). In addition, deletion of an internal NsiI fragment from pGSD1 resulted in a plasmid (pGSD2N) that did not lead to detectable extracellular protease production (Fig. 2). However, deletion of the internal SmaI fragment from pGSD1 to yield pGSD2S yielded full protease activity in plate tests (Fig. 2). Deletion of a 3.0-kb fragment extending from the leftmost SmaI site to the unique PmlI site of pGSD1 (Fig. 1), to generate pGSD2P, entirely abolished protease production. These results suggested that two or more protease structural genes occurred at the right end of the BamHIEcoRI fragment in pGSD1 (Fig. 1) but that DNA in the central portion of the insert was required for maximal protease production and/or excretion. This region is referred to hereafter as the "required region." The construction of the various subclones (Fig. 1; Table 1) disclosed the presence of three different protease structural genes in pGSD5, all of which appeared to read from left to right (Fig. 1), based on levels of activity produced by E. coli cells harboring both insert orientations in the pUC or pMTL plasmids. Some of the structural gene subclones such as pGSD17 were relatively unstable when maintained in E. coli in the absence of glucose to repress the vector lac promoter.

J. BACTERIOL.

Detection of protease inhibitor. E. coli colonies carrying pGSD5 or pGSD2N did not produce protease activity on gelatin plates but inhibited the halos elaborated by cells carrying pGSD1 or other protease-positive clones (Fig. 2). This suggested that pGSD5 and pGSD2N carry a gene for an inhibitor protein and afforded a mechanism for localizing the inhibitor gene on pGSD1 by growing E. coli cells carrying various cloned DNA fragments adjacent to cells expressing one or more structural genes. Further subcloning of pGSD5 disclosed that the putative inhibitor gene occurred between the leftmost ClaI site and the leftmost NsiI site of pGSD1 (Fig. 1). Subclones involving the SphI or EcoRV sites in this region resulted in clones devoid of inhibitor activity, suggesting that the inhibitor gene resides in the central portion of the 1.9-kb ClaI-NsiI fragment in pGSD11 and pGSD12 and reads from left to right (Fig. 1; Table 1). Characterization of required region. E. coli transformants of pGSD5 containing the required region, as well as pGSD3K and pGSD4K, carrying the prtC structural gene in pDSK519, gave much larger halos on gelatin plates than bacteria carrying only the structural gene clones. This suggested that the required region in pGSD5 contained transacting factors important for expression and/or secretion of protease. Two constructs (pGSD20 and pGSD21) contained the required region on a ca. 5-kb DNA fragment but lacked the protease inhibitor gene. In conjunction with plasmids carrying any of the three protease genes, pGSD21 greatly increased the halo diameter surrounding E. coli DH5a cells on gelatin plates (Table 2). Characterization of proteases. As previously reported (3), wild-type strain EC16 produced considerable extracellular protease activity in culture medium (Table 2). However, production in M9 glycerol medium was 10% or less than in M9 glycerol cultures supplemented with 0.5% tryptone. EC16 cells grown on LB medium produced about the same extracellular protease activity as in M9 tryptone cultures. EC16 as well as E. coli DH5a cells carrying cloned protease structural genes all produced significant halos on YC gelatin plates. E. coli cells carrying pGSD6 in addition to the prt structural genes generally produced larger and more rapidly appearing halos than cells carrying the structural gene constructs only. Significantly, plasmid constructs carrying only the prtA, prtB, or prtC structural gene did not lead to detectable extracellular protease activity when E. coli cells were grown in liquid medium (Table 2). However, cells that contained pGSD6 or pGSD21 (carrying the required region) in addition to the structural gene constructs produced significant levels of extracellular protease activity. In experiments comparing the protease activity of culture fluids with periplasmic fractions and cell lysates of E. coli cells carrying pGSD6 and the various structural gene plasmids, virtually all the activity was found in the extracellular fraction (data not shown). The extracellular protease activity of E. chrysanthemi EC16 was categorized by Barras et al. (3) as a metalloprotease on the basis of sensitivity to EDTA and insensitivity to other protease inhibitors. The extracellular protease activities released from E. coli DH5at carrying pGSD1, pGSD6 and pGSD3, pGSD6 and pGSD17, or pGSD6 and pGSD27 were also sensitive to EDTA (data not shown). This suggests that prtA, prtB, and prtC all encode metalloproteases. When dialyzed culture fluids from E. coli DH5a carrying both pGSD3 and pGSD6 were examined, it was observed that the divalent cations Mg2+ and Ca2+ increased protease activity significantly -above the control value but that Rb2+ and Zn2+ were not effective (data not shown).

|>|r | " 0pGSD

CLONING OF METALLOPROTEASE GENES FROM E. CHRYSANTHEMI

VOL. 172, 1990

npt

Tn5

Tn5

Tn5

5807

npt

4--j

:3

3U

4U

-

fi

-

>

cr

E

'

;; E 8 LcnO E 8 X

m I

IcIc

II

1

I

1 8

I

2

6

4

10

"required realon"

inh >

8 -

E

12

14

I-

FI

I

I

I I

I

I.

I I

I I

I. I.

.4 -.4

I.

1-H1

1

-I pGSD 2N -4 pGSD 2S -4 pGSD 2P -pGSD 3/4 pGSD 5/6 -I pGSD 9 pGSD 11/12 pGSD 13/14 pGSD 13ARV pGSD 15/16 pGSD 17/18 pGSD 20/21 pGSD 23 pGSD 25/26 pGSD 27/28

FIG. 1. Restriction map of the 14.5-kb BamHI-EcoRI DNA fragment from E. chrysanthemi EC16 cloned in pGSD1. Vertical arrows above the map denote TnS or nptII insertions which did not affect (lightface arrows) or abolished (boldface arrows) extracellular protease production by E. chrysanthemi EC16 or by E. coli DH5a carrying pGSD1. The bars below the restriction map denote various subcloned DNA fragments described in the text and Table 1 that were used to localize regions of the DNA fragment. Horizontal arrows below the restriction map denote the approximate positions of genes as determined by subcloning and sequencing data. The reading directions were deduced by levels of expression when the appropriate DNA fragments were cloned in both orientations relative to the lac promoter of pUC or pMTL vectors. The SspI sites shown are not unique; no sites were observed for the following enzymes: Hindlll, KpnI, NheI, Sacl, SplI, XbaI, XhoI.

Culture fluids from E. coli DH5oa cells carrying pGSD6 in addition to pUC129, pGSD3, pGSD17, or pGSD27 were concentrated and examined by sodium dodecyl sulfate-gel electrophoresis. Cells carrying the three cloned protease structural genes yielded unique bands, assumed to be the proteases, that were not present in fluids from control cells harboring pGSD6 and pUC129 (Fig. 3). These proteins had apparent masses of 50 kilodaltons (kDa) (protease C), 51.5 kDa (protease A), and 52.5 kDa (protease B) (Fig. 3). Renaturation of replicate unstained gels and gelatin overlays indicated that protease activity was associated with these unique bands but was not present in the lane from control cells carrying pUC129 (data not shown). Culture fluids from

E. chrysanthemi EC16 cells grown on LB medium contained only one protease band detectable by gelatin overlays or Coomassie brilliant blue staining (data not shown). The protease activity migrated at ca. 50 kDa. This is about the position expected for the prtC gene product, but the identity of the major protease in EC16 culture fluids was not established. Nevertheless, it appears that the three proteases are not produced in equal amounts by EC16 cells grown on LB medium. Concentrated bacterial culture supernatants were also examined by native gel electrophoresis. The preparations from both EC16 and E. coli DH5a cells carrying pGSD1 revealed single bands of activity that migrated close to the

5808

DAHLER ET AL.

J. BACTERIOL.

4,

-.f1z% *.

._ -66

a4 ;. --45 FIG. 2. Gelatin plate assays of extracellular protease production by E. coli DH5a cells harboring various plasmids. Rows: 1, pGSD1; 2, pGSD2N; 3, pGSD2S; 4, pGSD3; 5, pGSD4. The inhibition of halos formed by rows 1 and 3 was presumed to result from a protease inhibitor produced by cells carrying pGSD2N in row 2.

TABLE 2. Production of extracellular protease activity by E. chrysanthemi EC16 and by E. coli DH5a carrying various plasmids Bacterium'

EC16; M9 EC16; M9 tryptone EC16; LB DH5a(pUC129) DH5a(pGSD1) DH5a(pGSD2N)

DHSa(pGSD2P) DH5a(pGSD3) DH5a(pGSD3K) DHSa(pGSD4K) DH5a(pGSD17) DHSa(pGSD18) DHcS(pGSD27) DH5a(pGSD28)

prt

geneb A, B, C A, B, C A, B, C None A, B, C A, B, C None C C C B B A A

DH5a(pGSD6)(pUC129) None C DHSa(pGSD6)(pGSD3) C DH5a(pGSD6)(pGSD4) DH5a(pGSD6)(pGSD3K) C DHSc(pGSD6)(pGSD4K) C DH5a(pGSD6)(pGSD17) B DHSa(pGSD6)(pGSD18) B DH5a(pGSD6)(pGSD27) A DH5a(pGSD6)(pGSD28) A

Diameter of plate halo (cm)c

1.0 None 1.4 None None 0.7 1.5 0.6 1.0

Very weak Weak Very weak None 1.1 0.9 1.5 1.3 1.7 1.0 1.5 0.8

Protease

activityd

0.036 0.63 0.56