WO1990011351A1 - Purification of c. pseudotuberculosis toxin, and cloning and expression of toxin gene - Google Patents

Abstract

Corynebacterium pseudotuberculosis phospholipase D (PLD) toxin in substantially pure form, and a method of preparation thereto from a crude toxin preparation. A vaccine composition comprising toxoided pure PLD toxin is used against caseous lymphadenitis in sheep. Nucleotide and amino acid sequences of the toxin are disclosed, together with the cloning and expression of the PLD toxin gene in E. coli and coryneform bacteria.

Description

PURIFICATION OF C.PSEUDOTUBERCULOSIS TOXIN, AND CLONING AND EXPRESSION OF TOXIN GENE.

This invention relates to the purification of

Corvnebacteriu pseudotuberculosis toxin, to the preparation of toxoid from the purified toxin, and to the cloning and expression of the toxin gene to enable production of toxin in increased yields leading to more efficient vaccine production.

Corvnebacterium pseudotuberculosis is the Gram positive bacterium responsible for the disease in sheep known as caseous lymphadenitis (CLA) or cheesy gland. Carcasses of sheep showing symptoms of the disease are not acceptable for export from Australia. As around 30 to 60% of Australian flocks are infected with C♦pseudotuberculosis (Nairn, 1977, Beveridge, 1983), the disease is of considerable economic significance. The pathogenesis of CLA usually involves entry of

C.pseudotuberculosis through broken skin, following which the bacteria are carried to the local lymph node. Once established, they multiply to cause inflammation, necrosis and abcessation of the node (Beveridge, 1983). An important factor in the dissemination of

C. seudotuberculosis within an infected animal is the production of an exotoxin, phospholipase D (PLD) (Batey, 1986). Sheep can be vaccinated against CLA using a detoxified, PLD preparation obtained from the culture supernatant of C.pseudotuberculosis fermentations (Batey, 1986).

A combined C.pseudotuberculosis - 5 component clostridial vaccine has been used in Australia for the control of caseous lymphadenitis and clostridial diseases in sheep. The CLA component of the vaccine was based on the work of Nairn (1977) and Burrell (1978) but the composition of the combined vaccine was developed in the light of results of a series of field trials, two of which have been recently described (Eggleton et.al. 1987 a,b). These early vaccines were not standardised for toxoid content except by reference to toxin content prior to formalin inactivation. Procedures for an in vitro assay were subsequently developed, based on the work of Barr et.al. (1954) for the assay of clostridial toxoids. The assay permits the precise measurement of C.pseudotuberculosis toxoid concentration.

Information concerning PLD is extremely limited. There have been no previous reports describing toxin purified to homogeneity, and no amino acid sequence data are available. Studies with partially purified PLD toxin indicate that it is a sphingomyelinase with a molecular weight of about 31,000 and a pi of 9.2 to 9.6 (Hsu et.al. 1985; Egen et.al.. 1989).

In accordance with a first aspect of the present invention there is provided C.pseudotuberculosis toxin in substantially pure form. In this aspect, the invention also provides a method for the preparation of the toxin in substantially pure form from a crude toxin preparation which comprises the steps of concentration of the crude toxin by ultrafiltration followed by chromatography of the concentrated crude toxin on a cation exchange resin. The substantially pure form of the toxin which is produced in accordance with this invention is characterised in that it shows a single band of ca 31 kDa when subjected to electrophoresis under reducing conditions on a SDS/polyacrylamide gel and silver-stained (see Figure 1A hereinafter).

This invention also provides a C.pseudotuberculosis toxoid prepared by known toxoiding techniques from toxin in substantially pure form, to a vaccine composition comprising such toxoid, and to a method for immunising sheep against caseous lymphadenitis which comprises administration to the sheep of a vaccine composition comprising this toxoid.

It will also be appreciated that the purified toxin may also be used in immunoassays, such as EIA, for detection of antibody in diagnosis of CLA in sheep flocks.

A vaccine composition in which the corynebacterial toxoid was prepared from chromatographically pure exotoxin has been evaluated in a trial with vaccines containing a range of concentrations of C.pseudotuberculosis toxoid provided by toxoiding unpurified culture supernatant. All vaccines were combined with the 5 clostridial antigens in the commercial 6-component vaccine. Resistance of the sheep to infection with C.pseudotuberculosis was tested at one month, 6 months and 12 months post-vaccination by challenge with pus from ovine lymph glands actively infected with C♦pseudotuberculosis. The outcome was assessed 3 months after challenge by slaughter and inspection of the sheep for caseous lymphadenitis lesions. Protection was demonstrated by a significant reduction in the proportion of immunised sheep exhibiting lesions compared with control sheep and by fewer abscesses in affected immunised sheep than in affected control sheep. A positive correlation was found between amount of unpurified C.pseudotuberculosis toxoid administered and the degree of protection obtained. Chromatographically purified toxoid induced essentially the same protection as the vaccines at the same dose level of toxoid.

In another aspect of the present invention, the production for the first time of C.pseudotuberculosis toxin in substantially pure form has enabled the cloning and expression of the toxin gene, as well as the determination of the complete nucleotide sequence of this gene. Expression of the toxin gene in, for example, E.coli will of course enable efficient production of the toxin in substantial quantities for vaccine production. In addition, the manipulation of the toxin gene to reinsert the gene back into the C.pseudotuberculosis organism will lead to a higher level of production of toxin through over expression of the gene. Furthermore, modification of the PLD gene (using known DNA mutagenesis techniques) may allow production of inactive toxin analogues, thereby abrogating the requirement for chemical detoxification of CLA vaccine.

Recombinant plasmids carrying the Corvnebacterium pseudotuberculosis phospholipase D (PLD) gene have been used to transform Escherichia coli and several species of coryneform bacteria, including Corvnebacterium qlutamicum, Corvnebacterium ulcerans, Brevibacterium lactofermentum and' Brevibacterium flavum. With all hosts except C.ulc _-e'rans. expression of functional PLD was directed by the foreign PLD gene. C.ulcerans. which naturally produces a low level of PLD, produced no additional PLD when transformed with the recombinant gene. In E.coli, two versions of the PLD gene were studied; one in which expression of PLD was under the control of the natural promoter, the other in which expression was under the control of the trc promoter. The relative expression levels indicated that the corynebacterial promoter was five-fold weaker than trc in E.coli. Surprisingly, PLD production in E.coli under control of the trc promoter resulted in substantial amounts of the protein being secreted into the culture medium.

Corynebacteria possess several attractions as host organisms for the production of recombinant proteins. They are readily fermented on an industrial scale for the production of chemicals and vaccines. Like other Gram positive organisms, they are capable of secreting proteins into the extracellular medium. However, unlike Gram positive species such as Bacillus, they do not secrete large levels of extracellular proteases and so are less likely to cause degradation of secreted heterologous proteins. These features have encouraged the recent development of plasmid vectors (Martin, et.al.. 1987; Santamaria, et.al.. 1987) and transformation systems (Dunican and Shivnan, 1989; Haynes and Britz, 1990) for the genetic manipulation of corynebacteria.

According to this aspect of the invention there is provided a recombinant DNA molecule comprising a nucleotide sequence which codes for all or a substantial portion of C.pseudotuberculosis toxin, or a sequence related thereto by base substitution, insertion or deletion, to a cloning vector or vehicle including such a recombinant DNA molecule, and to a host cell transformed with such a vector or vehicle. In this aspect, the invention also includes a synthetic polypeptide in substantially pure form which corresponds to all or a portion of C. seudotuberculosis toxin, which may be produced by expression in a host cell as described above or by chemical synthesis. Such a synthetic polypeptide may be used in a vaccine composition for immunisation of sheep against caseous lymphadenitis or in diagnosis as previously described.

Further details of the various aspects of the present invention will be apparent from the following detailed description which illustrates, by way of example only, the production of the purified toxin, the use of purified toxoid as a vaccine, and the cloning, expression and sequencing of the toxin gene.

In the accompanying drawings: Figure 1 is a restriction map of the phospholipase D (PLD) gene region from C. seudotuberculosis. Hatched box: PLD gene signal sequence. Open box: region encoding mature PLD protein.

Figure 1A shows analysis by SDS-PAGE of Cm Sephadex C50 PLD (CLA) toxin containing fractions. Eluted toxin was subjected to electrophoresis under reducing conditions on 15% SDS-PAGE gel. Gel was fixed and stained with silver stain.

Molecular weight markers: soy bean trypsin inhibitor 21,500, carbonic anhydrase 31,000, ovalbumin 45,000, BS

66,200.

Figure 2 shows Western blot analyses of E.coli lysates and PLD purified from C.pseudotuberculosis. Lane 1: lysate of _E.coli harbouring a 2.7 kb clone of the PLD gene region (pCSL33). Lane 2: the plasmid vector without PLD gene insert (pUC12). Lane 3: a 1.5 kb subclone of the PLD gene region (pCSL39). Lane 4: purified PLD from C.pseudotuberculosis. The band beneath the 31 kDa protein in Lane 3 is possibly a degradation product of PLD protein. Figure 3 shows sphingomyelinase activity of PLD produced in E.coli. The assay was performed using 120 nmol TNPAL-sphingomyelin as substrate. Periplasmic extracts from E.coli DH5α harboring pUCllδ (designated CSL163) or pCSL39 (designated CSL170) were used.

Controls included 120 ng of sphingomyelinase purified from S.aureus (SA) and 200 ng of purified PLD (PLD). The samples from CSL163 and CSL170 contained approximately 100 μg of protein. According to enzyme immunoassays approximately 100 ng of PLD protein was present in the CSL170 extract (data not shown).

Figure 4 shows the nucleotide sequence of the PLD gene from C.pseudotuberculosis. Underlined bases correspond to regions with complete or partial homology with the E.coli consensus sequence for -10 and -35 promoter regions. Numbers in brackets are based on Thymine 84 as the transcriptional start site. Putative ribosome binding site: rbs. Underlined amino acid residues are those identified from sequencing purified PLD protein.

Figure 5 shows amino acid sequence homology between C.pseudotuberculosis phospholipase D (PLD) and Laticauda laticaudata phospholipase A2. Underlined residues in PLA2 are involved in calcium binding. In some PLA2 molecules the glycine (G) marked with a line is an alanine (A).

Figure 6 shows the transcriptional initiation site of the PLD gene. Autoradiogram showing results of primer extension reactions used to determine the position of the transcriptional start site. Primer extension products were generated using C.pseudotuberculosis RNA (lane 1) and E.coli RNA (lane 2). Relevant DNA sequence is shown on the right hand side. The arrow marks the position of the proposed start site at position 84 in the DNA sequence. Note that this sequence is complementary to that shown in Fig.4.

Figure 7 shows: A. Southern blot analysis of C.pseudotuberculosis genomiσ DNA using a PLD gene-specific probe. Lanes: 1. Sacl 2. Pstl 3. Sail 4. BamHI 5. Hindlll. B. Northern blot analysis of C.pseudotuberculosis RNA using the same probe.

Figure 8 shows plasmid maps showing the important features of (A) pCSL17 and (B) pCSL58. The sequence shown at the top of pCSL58 indicates the sequence at the fusion point between the trc promoter and the modified coding region of the PLD structural gene.

Figure 9 shows toxicity for mice of the GST-PLD fusion protein compared with wild-type PLD toxin. The figures at right indicate the dose of PLD (CLA) toxin in μg.

Figure 10 shows the antibody response in mice to wild-type PLD toxin and GST-PLD fusion protein. Antibody responses were measured in an ElA using purified PLD as capture antigen. Both wild-type toxin and the GST fusion protein were toxoided using formaldehyde, before being formulated in Syntex Adjuvant Formulatikon (SAF-0) or A1(0H)₃ and injected intraperitoneally. The titres shown are 10 weeks after 2 doses (4 weeks apart) of 0.5 cup of antigen.

EXAMPLE 1 Purification of C.pseudotuberculosis toxin

Culture supernatant from the fermentation of

C. seudotuberculosis was used as the source of crude PLD toxin. Crude toxin was concentrated 100-fold then extensively washed by ultrafiltration using a membrane with a 10,000 M.W. cut-off. Concentrated toxin was loaded on a carboxy methyl cellulose column equilibrated with 0.125M sodium phosphate buffer pH 6.0 and this buffer was used to wash through non toxin materials. The column was eluted successively with 0.1M, 0.2M and 0.4M NaCl in 0.125M phosphate buffer pH 6.3. The fractions eluted with 0.2M NaCl exhibiting the highest toxin activity were pooled. The NaCl was removed by washing the eluted toxin in a stirred cell with a 10 kDa cut-off membrane using an appropriate low salt buffer. For amino acid sequencing studies the buffer used was lOmM NH.HCO^.

The purified material was analysed on silver-stained SDS/2ME polyacrylamide gels to reveal an essentially pure protein with an apparent molecular weight of 31 kDa (Fig.lA). A sample containing approximately 50mg of the purified toxin was used for N-terminal amino acid sequencing on an Applied Biosystems Inc. model 470A protein sequencer operated according to the manufacturer's instructions. An unequivocal sequence was obtained for twenty-three amino acid residues, namely A-P-V-V-H-N-P-A-S-T-A-N-R-P-V-Y- A-I-A-H-R-V-L (using the single-letter amino acid code). Purified toxin was shown to have sphingomyelinase activity when assayed according to the procedure of Gatt et.al. (1978).

EXAMPLE 2

Use of purified toxoid in immunisation against CIA

a. Materials and Methods

Sheep

Romney-Marsh wether lambs, 3 to 5 months of age were obtained from the CLA-free sheep flock maintained at the Commonwealth Serum Laboratories (CSL) Field Station, Woodend. The sheep were divided into treatment groups by random selection at the time of primary vaccination and identified by numbered tags in both ears. The sheep were not maintained in separate groups during the trial. Vaccines

Vaccines were prepared from C.pseudotuberculosis toxoid (CPT) concentrates combined with antigen concentrates of Clostridium perfrinσens type D, C.nowi type B, C.tetani. C.septicum and C.chauvoei. The clostridial and crude CPT antigen concentrates were obtained from the Veterinary Vaccine Production Division of CSL.

The vaccines were prepared with a CPT content of 0.75, 1.5, 3.0, 6.0 and 9..0 combining power units (cpu) per ml for crude toxoid and 3 cpu per ml for purified toxoid. Purified C.pseudotuberculosis toxoid was prepared from purified toxin by incubating the toxin at low protein concentration (about 25mg/mL) for 24 days at room temperature in the presence of 0.067M formaldehyde, 0.0125M lysine and 0.05M sodium carbonate. The antigen content of the toxoid was determined by a combining power assay (see below) using an antitoxin against C♦pseudotuberculosis exotoxin. Vaccination

Vaccines were tested in groups of 20 sheep. All vaccines were administered subcutaneously in the neck, the sheep receiving 2 doses each of 2ml at an interval of 28 days.

Test bleeds Sheep were bled from the jugular vein prior to administration of the first and second doses of vaccine and at appropriate intervals prior to challenge and prior to slaughter.

Challenge The challenge method has been described by

Eggleton et.al. (1987a). Sheep were challenged at 4 separate sites with a total of 4g of pooled pus. Challenges were performed at three separate times during the course of the trial. A group of sheep dosed with vaccine containing 3 cpu (combining power units) of CPT was challenged one month after the second vaccine dose to determine the level of immunity in sheep exhibiting their highest circulating antitoxin levels. Sheep dosed with all 6 vaccines were challenged 6 months after vaccination and a third group of sheep dosed with 3 cpu of CPT was challenged 12 months after vaccination to assess duration of immunity.

Slaughter and post-mortem examination

Techniques were as described previously (Eggleton et.al. 1987a). Viable Counts Techniques were as described previously (Eggleton et.al. 1987a). Antitoxin titration text

Sera were tested for antitoxin by a modification of the antihaemolysin inhibition test (Zaki 1968). Dilutions of sheep serum were prepared in pH 7.2 buffered saline containing 0.1M N-tris (hydroxymethyl) methyl-2-aminoethanesulphonic acid (TES), 0.1M MgCl₂, 0.15M NaCl and 0.1% w/v bovine serum albumin. The dilutions were dispensed in 0.5ml volumes in 10mm disposable tubes. To each tube was added 8 antihaemolytic doses of C.pseudotuberculosis toxin in 0.5ml volume. After incubation for 2 hours at 37°C, 0.5ml of staphylococcal 0-lysin (8 minimal haemolytic doses) was added to each tube and incubation continued for a further 10 minutes before transferring to an ice bath for not less than 15 minutes. Intact red cells were removed by centrifugation before reading the test. The end point of the test was the first tube exhibiting complete absence of lysis and the titre expressed as the initial serum dilution in that tube. A titre equal to or greater than 1:10 was interpreted as seropositive. Combining power test

Dilutions of toxoid were prepared in pH 7.2 buffered saline and dispensed in 0.5ml volumes in 10mm disposable tubes. To each tube was added 0.5ml of C.pseudotuberculosis antitoxin containing 2 units per ml of anti-exotoxin (arbitrarily defined against a CSL standard serum). After incubation for 2 hours at 37°C then 0.5ml of 3% v/v washed sheep red cells was added to each tube. Following incubation for 30 minutes at 37°C, 0.5ml of staphylococcal β-lysin was added to each tube and incubation contained for a further 10 minutes before transferring to an ice-bath for not less than 15 minutes then centrifuging to remove intact red cells. The end point of the test was taken as the last tube exhibiting complete absence of lysis and the titre expressed in combining power units (cpu) per ml. Calculation of protective capacity of vaccines

The capacity of vaccines to protect animals against challenge when less than 100% of susceptible controls became affected by the challenge procedure can be expressed as a protection percentage using a formula of the following form:

Percentage protection = % infected controls - % infected vaccinates

% infected controls x 100

b: Results Bacterial counts

The samples of pooled pus used to challenge sheep on the 3 separate occasions had viable counts of 6.6 x

10 7, 1.8 x 107 and 2.9 x 107 colony forming units per g respectively. Serology

Primary responses to the Corvnebacterium toxoid were minimal. Of the 20 sheep dosed with the highest level of toxoid - 9 cpu per ml - only 7 had antitoxin titres of 1:10 or higher, while none receiving 1.5 cpu or 0.75 cpu per ml vaccine mounted a detectable primary response (Table 1). The level of the secondary response to crude CPT was also dose associated. Response to purified toxoid however was significantly greater than that to unpurified toxoid. The results in Table 2 show that the highest serum titres together with the most prolonged response to immunisation occurred in sheep dosed with vaccine containing 3 cpu per ml of purified toxoid, followed by those with 9 cpu per ml and then 6 cpu per ml of crude toxoid. Sheep dosed with vaccines containing 3 cpu per ml or less of crude toxoid showed a progressive reduction in level and duration of secondary response. As Table 1 shows, not every sheep in the lower dose groups became seropositive after the second dose of vaccine. The higher antitoxin titres found in vaccinated sheep 1 month after challenge compared with those in control sheep, shows that priming of all vaccinated sheep had occurred, even though some failed to seroconvert. Post-mortem examination

Eleven sheep were lost from the trial either by accidental loss or from acute bacterial infection following the challenge. A further 13 were withdrawn from the trial when serology showed that these sheep were exposed to C.pseudotuberculosis prior to the scheduled challenge procedure, presumably by contact with those sheep challenged 1 month after vaccination. Final group sizes, results of carcase examination after slaughter, and calculation of vaccine protection factor are presented in Table 3.

The numbers in parenthesis in this Table refer to the number of sheep with lesions (either carcase or lung). The majority of vaccinated animals in which lesions were observed at post-mortem were affected at only one of these sites (35 of 40). Conversely, control animals were more frequently infected in both carcase and lungs (15 of 41) and lesion counts were invariably much higher. In two cases (the groups dosed with 9 cpu crude toxoid and 3 cpu purified toxoid) lung lesion counts were high. The former resulted from one sheep with 25 lung lesions and the latter from a single sheep with 21 lung lesions.

c. Discussion Examination of the level of protection afforded by the vaccines containing 0.75, 1.5 and 3 cpu of CPT shows an almost linear increase in protection proportional to the corynebacterial antigen content. Only one sheep was affected in each of the groups dosed with 3 cpu and 6 cpu vaccine suggesting that the optimum antigen concentration in combined vaccines is between 3 and 6 cpu. Both carcase and lung lesion numbers were reduced in affected vaccinated sheep compared with control sheep even in the groups of sheep dosed with the lower concentrations of CPT (Table 3).

Sheep dosed with the vaccine containing 3 cpu per ml of purified CPT and challenged after 6 months exhibited a level of protection between the levels obtained with vaccines containing 1.5 and 3 cpu per ml of crude toxoid, also challenged after 6 months. A similar level of protection was found in sheep dosed with 3 cpu per ml crude toxoid vaccine and challenged after 1 month. As the purified toxin was shown to consist of one component only by PAGE electrophoresis, the ability of the purified toxoi,d to confer a substantial level of protection against challenge shows that antitoxic immunity alone can prevent the development of caseous lymphadenitis following challenge.

At the time of the second challenge, 6 months after vaccination, the only sheep with positive antitoxin titres were 3 animals in the 6 and 9 cpu vaccine groups and 13 of the sheep dosed with purified antigen. However after challenge, 100% of the vaccinated sheep had high levels of circulating antitoxin. Protection against challenge in vaccinated sheep therefore does not depend solely on the presence of circulating antitoxin at the time of challenge, provided the sheep can mount a rapid anamnestic response to the primary skin infection and thus prevent invasion of lymph nodes, lungs and other organs. The vaccinated sheep challenged 12 months after dosing similarly had no circulating antitoxin within one month of challenge however only 45% of these sheep remained free of infection, a level of protection of 38%. The reason for the difference in protection between sheep challenged at 6 months and at 12 months cannot be deduced from the serological data available but is possibly a function of delay in anamnestic response to challenge in the sheep which became affected.

TABLE 1 Antitoxin response to vaccination.

4 weeks after 1 dose 4 weeks after 2 doses Vaccine Median Titre % sero- Median Titre % sero- group titre range conversion titre range conversion

9 cpu <10 <10-40 35 160 20-320 100

6 " <10 <10-20 10 80 20-320 100

3 " <10 <10-80 15 40 <10-160 85

1.5 " <10 All <10 0 20 <10-80 70

0.75" <10 All <10 0 20 <10-80 55

3 cpu* <10 <10-40 15 320 20-640 100

* purified.

TABLE 2 Magnitude and duration of antitoxin responses

Median titres

Vaccine Time , after 2nd dose Time after challenge Group (Weeks) 4 8 12 18 24 (Weeks) 4 12

9 cpu 160 10 <10 <10 <10 3200 1130

6 cpu 80 <10 <10 <10 <10 3200 1130

3 cpu 40 <10 <10 <10 <10 6400 1600

1.5 cpu 20 <10 <10 <10 <10 6400 800

0.75 cpu 20 <10 <10 <10 <10 1600 400

Purified 3cpu 320 40 20 20 20 4000 1600

Controls <10 <10 <10 <10 <10 320 640

TABLE 3 Post-mortem and vaccine protection data

Vaccine Period Sheep Sheep Aver.lesion counts % sheep Prot- group to killed affected Carcase Lung affected ection challenge factor

3 cpu 1 mth 20 5 1.3(3)* 3.0(2) 25 70

Controls II 18 15 3.0(9) 9.0(15) 83

9 cpu 6 mths 20 5 2.0(2) 8.5(4) 25 68

6 " It 20 1 1.0(1) 0 5 93

3 " II 17 1 0 2.0(1) 6 92

1.5" II 15 4 2.0(3) 1.0(1) 27 65

0.75" II 19 10 2.7(4) 2.2(8) 53 31

Purified-

3 cpu tl 18 4 1.5(2) 8.0(3) 22 71

Controls II 13 10 3.5(8) 11.0(5) 77 0

3 cpu 12 mths 18 10 1.0(1) 13.7(10) 55 38

Controls II 18 16 5.3(3) 14.0(16) 89

* Number of sheep with lesions

EXAMPLE 3 Cloning and expression of PLD toxin gene a. Materials and Methods.

Bacterial strains and plasmids. E.coli strain DH5alpha (BRL) was host for plasmids pUC12 and pUCllβ (Vieira and Messing, 1987). The Commonwealth Serum Laboratories production strain of Corvnebacterium pseudotuberculosis (West Australian Department of Agriculture strain 1030) and an isolate from New Zealand (isolate 107) were used in this study. The plasmids pCSL33, 39, and 40 were constructed during this study. pCSL33 is pUC12 containing the PLD gene on a 2.7 kb fragment. pCSL39 and 40 were derived by subcloning a 1.5 kb Sacl fragment (Fig.l) from pCSL33 into pUCllβ. In pCSL39 the PLD gene is transcribed in the opposite direction to the lacZ gene, while in pCSL40 it is transcribed in the same direction.

Media.

E.coli strains were grown in Luria broth (LB: lOg tryptone, 5g yeast extract, lOg NaCl per litre) containing 30/-tg of ampicillin per ml.

C.pseudotuberculosis was grown in nutrient infusion broth (NIB: 500 ml veal extract (CSL), 5g NaCl, and lOg proteose peptone per litre). Toxin purification.

Toxin was purified from C.pseudotuberculosis culture supernatants using cation exchange chromatography (see Example 1). Assay for C.pseudotuberculosis PLD. C.pseudotuberculosis phospholipase D (PLD) activity was detected using a modification of the Zaki assay (Zaki, 1965, Souckova and Soucek, 1972) and an assay to detect sphingomyelinase (Gatt, 1978). For the Zaki assay 1.5ml aliquots of a 15% v/v suspension of sheep red blood cells in Alsevers solution were washed 3 times in 5ml of buffered saline [0.01 M N-tris(hydroxymethyl) methyl-s-aminoethane sulphuric acid (TES), O.OIM MgCl₂, 0.85% w/v NaCl, and 0.1% w/v bovine serum albumin, pH 7.3]. The washed cells were then diluted 10 fold in buffered saline. Doubling dilutions of test samples (lOOyl) in buffered saline, were made in microtitre trays. Samples (lOOμl) of washed sheep red cells were then added to each test sample, and incubated at 37°C for 45 min. β-lysin (CSL special product, 50μl) was then added to each test well and the trays held at 37°C for 30 min and then on ice for 30 min. Microtitre trays were spun in a Beckman GPR bench top centrifuge for 30s at 400Xg. The number of Zaki units present in a given sample was determined from the dilution of PLD test sample that no longer protected the blood cells from lysis by staphylococcal β-haemolysin. Tests were standardised against a stock solution of PLD. The sphingomyelinase assay was performed at pH 7.4 as described by Gatt et.al. (1978) using the chromogenic substrate trinitrophenylaminolauryl sphingomyelin (TNPAL- sphingomyelin, Sigma Chemical Company). Because PLD requires Mg^** for activity, 30 μM MgCl₂ was included in the reaction buffer. N-terminal seguencing of the PLD protein.

The N-terminus of purified PLD (50 μg) was determined using an Applied Biosystems (ABI) 470A protein sequencer.

Preparation of oligonucleotides.

PLD-specific oligonucleotides deduced from N-terminal PLD protein sequence data were synthesised using an Applied Biosystems 380B DNA synthesiser. Oligonucleotides for screening gene libraries were prepared as mixtures to accommodate codon degeneracy or incorporated inosine residues at equivocal positions (Ohtsuka et.al., 1985). Oligonucleotides to be used as sequencing primers were designed from the derived DNA sequence data. Tritylated oligonucleotides were purified with oligonucleotide purification cartridges (ABI). Labelling oligonucleotides.

Oligonucleotides were 5' end-labelled (Maniatis et.al, 1982) using 50 μCi gamma-ATP (3000 Ci per mmol, Amersham) and T4 polynucleotide kinase (5 units) to a specific activity of at least 10⁷ dpm per μg.

Isolation of genomic DNA from C.pseudotuberculosis.

A 10ml overnight culture of C.pseudotuberculosis was used to inoculate 200ml of NIB containing l.Oμg penicillin per ml. After growth overnight at 37°C the culture was centrifuged (14000 x g, lOmin. ) and the pellet suspended in 10ml STET (8% sucrose, 0.5% Triton X-100, 50mM EDTA pH 8.0, lOmM Tris HCl pH 8.0) containing lOmg lysozyme per ml. The mix was held at 37°C for 2h and 3.0 ml of lysis solution (1.0% w/v SDS, 0.2M NaOH) was then added. After a further hour at 55°C the lysate was centrifuged at 47800 x g for 30min. Chromosomal DNA was isolated from the cleared lysates using cesium chloride/ethidium bromide gradients (Maniatis et.al, 1982). Construction and screening of genomic library.

Genomic DNA (22μg) from C.pseudotuberculosis was partially digested with 6 units each of Haelll and Alul. Digests were run on a 1.0% agarose gel and fragments between 2 and 4 kb were excised, purified using Geneclean (Bresatech, Adelaide) and cloned into the Smal site of pUC12. Ligation mixes were transformed into E.coli DH5α. Transformants were screened using a rabbit anti-PLD antiserum and a SuperScreen immunoscreening system kit (Amersham). Expression of PLD gene in E.coli.

Cultures (1 ml) of E.coli harbouring the toxin gene were pelleted and resuspended in 100 μl of lysis buffer (12.5 μg lysozyme per ml STET). The mix was centrifuged and 80 μl of supernatant used in each assay. Lysates were tested for production of PLD using Western blotting and the modified Zaki test. For the sphingomyelinase assays periplasmic extracts were prepared according to Kendall et.al. (1986). Restriction mapping, subcloning and Southern blotting. A restriction map of an antibody reactive clone (pCSL33) was derived using standard procedures (Maniatis et.al.. 1982). Plasmid DNA carrying the putative PLD gene was restricted using Pstl or Sacl and the fragments Southern blotted to nitrocellulose. Filters were hybridised (Meinkoth and Wahl, 1984), at room temperature with a toxin specific end-labelled probe (5'- GTIGTICAC/TAAC/TCCIGC-3' ) for 2 hours in 6 x SSC, 10 x Denhardts solution (Maniatis et.al.. 1982), washed at increasing stringency as necessary (up to 45°C) and exposed to X-ray film (Fuji R ). DNA seguence analysis. M13 clones were sequenced using modified T7 DNA polymerase according to the manufacturer's instructions (United States Biochemicals) using either universal primer or synthetic oligonucleotides. Segments of overlapping DNA sequence were generated using oligonucleotide primers designed from the PLD sequence as it became available. Sequence data were collated and analyzed using the DNASIS software package (Pharmacia LKB). Primer extension. Total RNA was isolated from E.coli and

C.pseudotuberculosis according to Aiba et.al. (1981). 0.1-10 pmol of oligonucleotide primer (5'- TTGAGTGGTTAAAACGCGGTGGGC-3' ) was end-labelled (Maniatis et.al.. 1982), ethanol precipitated then annealed to 20μg RNA. Primer extension was accomplished at 42°C for 1.5 h using 20 μ of AMV reverse transcriptase (Pharmacia) under conditions specified by the manufacturer. Extension products were separated on a 6% polyacrylamide sequencing gel which was fixed, dried and exposed to X-ray film. Analysis of C.Pseudotuberculosis RNA.

Total RNA was isolated from C.pseudotuberculosis as described above. A 950 bp PvuII fragment containing most of the PLD gene plus a small portion of the lacZ gene was isolated from pCSL39 and labelled to a specific activity of 10⁸ dpm/μg using a random primer labelling kit (BRL). RNA was eleσtrophoresed on a 1% formaldehyde agarose gel and blotted to nitrocellulose then hybridised at 68^βC overnight with the gene probe (Maniatis et.al.. 1982). Size estimations were made relative to RNA molecular weight markers (Bethesda Research Laboratories). Analysis of C.Pseudotuberculosis genomic DNA.

Genomic DNA isolated from C.pseudotuberculosis was restricted, electrophoresed on a 1% agarose gel then Southern blotted (Maniatis et.al.. 1982) to a nitrocellulose filter (Amersham). The PLD gene-specific probe was produced as described above. Hybridisation was at 68°C, overnight in 6 x SSC. The filter was then washed in 1 x SSC for 20 min at 68°C and exposed to X-ray film (Fuji RX).

b. Results

Cloning of the PLD structural gene.

C.pseudotuberculosis genomic DNA was partially digested with Alul and Haelll and fragments approximately 2 kb in size were purified and cloned into the Smal site of pUC12. A library was prepared in DH5α and 15,000 clones were screened using an anti-PLD antiserum. One positive clone was identified.

A 5.6 kb plasmid (pCSL33) was isolated from this clone and mapped using a variety of restriction endonucleases (Fig.l). To confirm that the 2.7 kb insert in pCSL33 contained at least part of the PLD gene, the restricted plasmid DNA was Southern blotted to nitrocellulose and hybridised with a PLD-specific oligonucleotide probe. A 1.5kb Sacl and 200bp Pstl fragment hybridised with this probe (data not shown). Since the hybridising oligonucleotide was designed from protein sequence data derived from the PLD amino terminus, the 5' end of the PLD structural gene must be located on the 200 bp Pstl fragment (Fig.l). Given the size of the mature protein (31 kDa) and hence a gene size of around 850 bp, it is likely that the entire structural gene is contained within the 1.5 kb Sacl fragment (Fig.l). Expression of the PLD gene in E.coli.

The 1.5kb Sacl fragment was subcloned into pUCllβ and recombinants were identified using the PLD-specific antiserum. Restriction analysis of a number of the clones that reacted with the antiserum revealed that the Sacl fragment was present in both orientations. This result suggests that the PLD promoter is functional in E.coli. Two plasmids (pCSL39 and 40) having the 1.5 kb fragment cloned in opposite orientations were chosen for further analysis.

Cell lysates were prepared from strains harbouring pCSL33, 39 or 40 and electrophoresced on a 12% SDS polyacrylamide gel. Proteins were transferred to nitrocellulose by electroblotting, then tested for reactivity with the anti-PLD antiserum. Strains containing either the original clone (pCSL33) or the subclones (pCSL39 and 40) produced a 31 kDa protein that reacted with PLD-specific antibodies (Fig.2). This size is consistent with that of the mature PLD protein, indicating that the complete PLD structural gene is present on the 1.5 kb Sacl fragment.

To determine the subcellular localization of PLD produced in E.coli. lysates of strains containing pCSL 39 were separated into cytoplasmic and periplasmic fractions as described previously (Kendall, 1986). Proteins from these fractions were separated on 12% SDS polyacrylamide gels and analysed by silver staining or Western blotting (data not shown). These experiments demonstrated that

PLD is predominantly located in the periplasm, indicating that the PLD signal sequence functions in E.coli. Activity of PLD produced in E.coli.

PLD produced by C.pseudotuberculosis is commonly assayed using the Zaki test which measures the ability of PLD to block the lysis of sheep erythrocytes by staphylococcal /3-haemolysin. Cell lysates from all PLD clones had demonstrable activity in this assay whilst host cells harbouring plasmids without insert DNA produced no detectable reaction (data not shown). Previous studies have shown that PLD is a sphingomyelinase. In an assay which measures the hydrolysis of a synthetic substrate, TNPAL-sphingomyelin, purified PLD showed similar reactivity to sphingomyelinase prepared from Staphylococcus aureus (Fig.3). Periplasmic extracts prepared from E.coli harbouring the PLD gene were also capable of hydrolysing TNPAL-sphingomyelin whereas no activity was detected from extracts of E.coli cells lacking the PLD gene. These data show that a functional protein is produced when the PLD gene is expressed in E.coli. DNA seguence analysis.

The 1.5kb Sacl fragment was subcloned in both orientations into to M13mpl8. Nucleotide sequence data from both strands were derived for approximately 1100 bp of the Sacl fragment (Fig.4) using a modified dideoxy - chain termination procedure (Tabor and Richardson, 1987). Analysis of the DNA sequence reveals a major open reading frame beginning just before the PvuII site and terminating at the end of the cloned fragment (Fig.l). This open reading frame is capable of coding for a 307 residue protein with a predicted size of 34.1 kDa. To confirm that this protein corresponds to PLD, the predicted polypeptide sequence was compared with the amino terminal sequence data obtained from purified PLD. A stretch of 23 residues identical to the PLD protein sequence was identified 25 amino acids downstream from the predicted translational initiation site. The ___5 preceding 24 amino acid sequence is strongly indicative of a signal sequence (von Heijne, 1985). On this basis the open reading frame encodes a polypeptide which includes a putative signal peptide (2.7 kDa) and mature PLD protein (31.4 kDa). This size of mature PLD predicted from the sequence is therefore in close agreement with that estimated by SDS PAGE (Fig.2).

The DNA and protein sequences of PLD were used to search to the appropriate DNA (GENBANK, EMBL) or protein (NBRF) databases for homologous sequences. Surveys of the nucleotide databases failed to detect sequences showing significant similarity to PLD. Comparison of the PLD protein sequence with those of other phospholipases however, revealed a small region of homology with the phospholipases A2 (Fig.5).

Determination of the transcriptional start site.

Expression studies in E.coli indicated that the E.coli RNA polymerase can use a C.pseudotuberculosis sequence as a promoter. Inspection of the DNA sequence upstream of the PLD structural gene revealed several regions which resemble the consensus E.coli promoter. Primer extension experiments were conducted to determine the location of the PLD promoter. A labelled oligonucleotide designed to anneal to the 5' end of the PLD mRNA was hybridised to total RNA prepared from either C.pseudotuberculosis or E.coli carrying pCSL39, then extended using reverse transcriptase. The extension products were separated on a 6% sequencing gel and compared with a DNA sequence ladder generated using the same oligonucleotide as a primer (Fig.6). These experiments suggested that RNA polymerase in both E.coli and C.pseudotuberculosis initiates transcription at the thymine located at position 84 in the sequence. Based on the E.coli promoter consensus sequence (Reznikoff and Gold, 1986) two overlapping -10 regions and a -35 box were identified upstream of this putative initiation site (Fig.4). Determination of PLD gene copy number in C.pseudotuberculosis.

Separate aliquots of C.pseudotuberculosis genomic DNA were digested with restriction enzymes, electrophoresed on a 1% agarose gel, then transferred to nitrocellulose and hybridised to the same probe as used for the RNA analysis (see above). With the exception of Pstl. the enzymes used in this experiment were chosen because they do not have recognition sites within the PLD coding sequence. Therefore, if the PLD gene is present more than once in the C.pseudotuberculosis genome, under stringent conditions it is likely that one or more of these enzymes will give rise to multiple hybridising bands. Analysis of the PLD restriction map (Fig.l) predicts that the probe should hybridise to 93 bp and 200 bp Pstl - generated fragments as well as to a fragment larger than 600 bp (depending on the position of the next site downstream from the coding sequence). Since the fragments less than 300 bp were not retained on the gel, only one fragment would be detected if there is a single PLD gene.

Single hybridising bands were observed in all the tracks (Fig.7A), suggesting that the PLD gene is present as a single copy in the C.pseudotuberculosis genome. Analysis of C.pseudotuberculosis RNA.

Because the PLD coding sequence finishes within 10 bp of the end of the cloned DNA, the fragment does not contain sequences resembling either transcriptional terminators for the PLD gene or the translational initiation signals for a downstream gene.

To determine the size of the authentic PLD transcript, total RNA was prepared from C.pseudotuberculosis. separated on a 1% denaturing agarose gel then transferred to nitrocellulose and hybridised to a PLD-specific probe labelled by random priming. The size (approx. 1100 bp) of the mRNA species which hybridised to the probe is similar to the size of the PLD gene (Fig.7B). This suggests that the PLD transcript terminates just downstream of the coding sequence. c. Discussion DNA sequence analysis of the C.pseudotuberculosis phospholipase D (PLD) structural gene reveals two possible open reading frames (ORFs). The first, beginning at base 142 (Fig.4), is capable of encoding a polypeptide consisting of 18 amino acids. There is however no evidence to suggest that this polypeptide is synthesised. The predicted translational start site for the second ORF is a methionine codon at -24 (Fig.4) which is preceded by a region similar to the E.coli consensus sequence for a ribosome binding site (Reznikoff and Gold, 1986) and also by two stop codons (Fig.4). The major ORF encodes a 24 residue signal sequence (von Heijne, 1985) and a 259 amino acid (31.4 kDa) polypeptide. This structure is consistent with previous data (Hsu, 1985) identifying PLD as a 31.0 kDa protein secreted by C.pseudotuberculosis.

Comparison of the PLD protein sequence with other phospholipases reveals some similarity to the phospholipase A2 (PLA2) class (Fig.5). It is interesting to note that this region is highly conserved amongst the PLA2 enzymes and that both the calcium binding and the active sites fall within it (Waite, 1987). This region may be a suitable site for in vitro mutagenesis studies aimed at deriving genetically-toxoided PLD analogues. The sequence data show that the cloned fragment does not contain transcriptional terminators. It is therefore possible that other open reading frames exist on the same transcript downstream from the PLD gene. Analysis of C. seudotuberculosis RNA indicates that this is probably not the case since the mRNA is approximately the same size as the gene (Fig.7B).

There are a number of regions upstream of the beginning of the PLD coding sequence which resemble the consensus sequence for an E.coli promoter (Reznikoff and Gold, 1986). To determine which one is the PLD promoter the position of the transcriptional start site was determined. Primer extension data suggest that transcription of the PLD gene in both

C.pseudotuberculosis and E.coli is initiated at thymine number 84 (Figs. 4 and 6). Initiation at a thymine does not follow the E.coli consensus model which favours purines (McClure, 1985). This result may however, reflect the sensitivity of the primer extension technique.

Regions similar to the consensus sequences for the E.coli promoter -10 and -35 boxes are located close to this putative transcriptional start site (Fig.4). As observed for the Corvnebacterium diphtheriae toxin gene, the PLD promoter has two possible -10 regions (Kaczorek et.al.. 1983). In C.diphtheriae it has been suggested that the TAGGAT box showing optimal spacing (18 bp from the -35 region) would be used in preference to the possible alternative -10 region (TATAAT). Although the latter box is identical to the E.coli consensus sequence it is only 12 bp from the -35 region. Recently both -10 regions were shown to function in E.coli. however, as predicted the optimally spaced -10 region was favoured (Boyd and Murphy, 1988). In the case of the PLD gene, the spacing between the putative -10 (CATAAT) and -35 regions is 18 bp thus fitting the E.coli consensus model. The other putative -10 region for the PLD gene (TTTCAT) does not closely resemble the E.coli consensus -10 box (TATAAT) and is sub-optimally spaced.

EXAMPLE 4 Expression of PLD structural gene in Corvneform bacteria a. Materials and Methods Bacterial strains and plasmids.

E.coli DH5α (Bethesda Research Laboratories, Inc. ) was used as the host for all routine DNA manipulations. Expression studies in E.coli were carried out using strain HB101. Corynebacterial strains used in the expression studies were Corvnebacterium glutamicum strain AS019, Corvnebacterium ulcerans strain 712, Brevibacterium lactofermentum strain BL1 and

Brevibacterium flavum strain BF4. B.flavum BF4 was obtained from Prof. P. Rodgers, University of New South Wales. The origins of other corynebacterial strains have been described previously (Haynes, 1990; Serwold-Davis, et.al.. 1987). The plasmids pCSL33, pCSL39 and pCSL40 are derivatives of pUC vectors containing the PLD gene and have been described previously (see Example 3). The E.coli/Corvnebacterium shuttle plasmid pCSL17 was constructed during this study and is described in the text. The plasmids pCSL41 and pCSL42 are derivatives of pCSL17 which contain either orientation of the PLD gene and its natural promoter as a 1.5-kilobase (kb) Sacl fragment inserted between the unique Sail and Pstl sites in pCSL17. The plasmid pCSL58 is a derivative of pKK233- 2 (Amman, 1985) which contains the PLD structural gene under control of the trc promoter. Its construction is described in the text. DNA manipulations.

Restriction enzyme digestions, DNA ligations, fill-in of cohesive DNA ends with DNA polymerase (either Klenow or T4), and site-directed in vitro mutagenesis were performed according to standard protocols (Sambrook, et.al.. 1989) with commercially available enzymes used according to the manufacturer's instructions. DNA transformation of host cells.

E.coli cells were transformed according to the CaCl₂-induσed competence method (Sambrook, 1989). Coryneform bacteria were transformed using an electroporation protocol involving pregrowth of cells in the presence of glycine and isonicotinic acid hydrazide and plating on an osmotically protective recovery medium (Haynes and Britz, 1990). Media and growth conditions.

All host strains were routinely grown in Luria Broth (Sambrook, et.al.. 1989) supplemented with 0.5% glucose (LBG). E.coli and C.ulcerans were grown at 37°C and 200 rpm. The other corynebacterial strains were grown at 30°C and 200 rpm. Ampicillin (30 μg/ml) or kanamycin (50 μg/ml) were included in the growth medium as appropriate. Expression of PLD gene. Transformed host cells containing the PLD gene with its natural promoter (i.e. plasmids pCSL33, pCSL39, pCSL40, pCSL41 or pCSL42) were grown to stationary phase (A^ 1.5-2.0). E.coli HB101 containing pCSL58 were grown to either early log phase (A^ 0.4) or mid log phase (A^ 0.9-1.0) then induced with 0.1 mM IPTG. Growth was continued after induction until stationary phase (A^ 2.0). Cells were harvested by centrifugation (10,000 x g, 10 min). Cell lysates were prepared by resuspension in STET buffer (Sambrook, et.al.. 1989) containing 50 μg/ml lysozyme followed by incubation at RT, 10 min (E.coli) or 30 min, 37°C (corynebacteria) and then freeze-thawing three times. Periplasmic extracts were prepared according to the method of Kendall et.al. (1986). Culture media, cell lysates and periplasmic extracts were analysed by SDS-polyacrylamide gel electrophoresis and were assayed for the presence of PLD as described below.

Assay for C.Pseudotuberculosis PLD.

PLD activity was assayed by a modification of the Zaki assay as previously described (Example 3). Titres were standardised against a laboratory reference with a nominal titre of 6400 units/ml. Sphingomyelinase activity was measured using the hydrolysis of N-w- trinitrophenylaminolaurylsphingosylphosphorylcholine (Sigma Chemical Co., Mo.) in a modification of the method described by Gatt et.al. (1978). Incubation of enzyme with substrate was performed at 37°C for 30 min. A standard curve was plotted using between 50 and 800 ng of purified PLD. PLD activity in samples was read from the standard curve and converted into μg of PLD present per ml of original culture. Each assay was repeated at least twice. Both the Zaki and sphingomyelinase assays showed excellent correlation with PLD immunoreactivity as estimated by an enzyme immunoassay using a high titre polyclonal sheep antiserum as the capture antibody. Determination of plasmid stability and copy number. For plasmid stability studies, transformed host cells were grown in LBG for 30 generations. Samples were withdrawn periodically for the estimation of total viable counts and counts on selective media (containing kanamycin at 25 μg/ml). Plasmid stability was estimated as the percentage of kanamycin resistant cells remaining after 30 generations. To determine the plasmid copy number, host cell lysates were analysed using a plasmid copy detection kit (Du Pont) as per the manufacturer's instructions.

b. Results and Discussion

Construction and characterisation of the shuttle vector

PCSL17.

The plasmid pCSL17 is a derivative of pBR322 containing a corynebacterial origin of replication and the kanamycin resistance gene from transposon Tn903. Firstly, pBR322 was reduced in size by deleting the 1.4- kb EcoRI-Aval fragment (containing the tetracycline resistance gene) and replacing it with a 48 base pair (bp) linker containing multiple cloning sites. As well, a 0.6-kb Ball-PvuII fragment was deleted. These manipulations yielded a 2.4-kb plasmid designated pCSLlO. Transposon mutagenesis of pCSLlO with Tn903 yielded a 4.1-kb plasmid (pCSL14) containing both ampiσillin and kanamycin selectable markers. The cryptic corynebacterial plasmid pSRl was isolated from plasmid pHY416 Yoshihama et.al..1985) as a 3.1-kb Bell fragment and cloned into the polylinker of pCSL14. This yielded the 7.2-kb plasmid pCSL17. The important features of pCSL17 are shown in Figure 8a.

When pCSL17 was transformed into either E.coli or corynebacterial hosts, transformants were obtained which were resistant to kanamycin at levels of 1000 μg/ml (E.coli) or 150 μg/ml (C.glutamicum). This showed that the Tn903 kanamycin resistance gene was functional in the coryneform background. Copy number studies indicated that pCSL17 was present at 10-20 copies per cell in E.coli and 2-4 copies per cell in the different corynebacterial species (data not shown). The relative copy numbers of pCSL17 in E.coli versus corynebacteria are consistent with the relative levels of kanamycin resistance noted above. Similar differences (4 to 10- fold) in levels of expression of antibiotic resistance between E.coli and C.glutamicum transformed with the same plasmid have been observed before (Ozaki, et.al.. 1984). However, in this case plasmid copy numbers were equivalent, so that differences in transcriptional and translational efficiencies were proposed as the cause.

Intact pCSL17 could be recovered from each of the transformed species, indicating that the plasmid was stable in both E.coli and corynebacterial backgrounds. Furthermore, growth for over 30 generations in the absence kanamycin selective pressure resulted in 100% retention of pCSL17 in both backgrounds. Thus, pCSL17 appeared to be a suitable vector for use in corynebacterial hosts. Expression of PLD in corynebacterial host cells.

Plasmids pCSL41 and pCSL42 were constructed from pCSL17 by insertion of the PLD gene as a 1.5-kb Sacl fragment between the unique Pstl and Sail sites. The plasmids differ only in the orientation of the PLD gene. Both confer the phenotype Amp^s, Kan^R and PLD^*.

C.glutamicum AS019, C.ulcerans 712, B.lactofermentum BL1 and B.flavum BF4 were transformed with pCSL41 or pCSL42. Stationary phase cultures of the transformed cells were analysed for the presence of PLD activity using the modified Zaki assay or a sphingomyelinase assay (see Materials and Methods). Minimal PLD activity was detectable in cell lysates. However, significant levels of PLD were found in the culture media (Table 1). Only C.ulcerans did not show an increase in PLD production as a result of the presence of pCSL41 or pCSL42. Although each of the other recombinant corynebacterial strains showed PLD production, the levels were disappointingly low when compared to those produced by non-recombinant C.pseudotuberculosis. Orientation of the PLD gene (i.e. ρCSL41 versus pCSL42) did not seem to have any effect on the levels of expression. Construction of pCSL58.

With the aim of achieving high level expression of PLD in E♦coli. the PLD structural gene was placed under control of the strong E.coli promoter, trc. The inactivated tetracycline resistance gene in plasmid pKK233-2 (Amman, 1985) was removed by digestion with EcoRI and PyuII, followed by Klenow "fill-in". In it place was ligated the 1.7 -kb EcoRI fragment (Klenow filled) from pMC9 (Calos, et.al.. 1983) containing the lacl and lacZ' genes. The resultant plasmid was designated pCSL53.

The PLD gene in pCSL39 was subjected to site- directed mutagenesis around the translation initiation codon to create a Sphl site. The modified PLD gene was removed as a 1.5-kb Sacl fragment, filled with T4 DNA polymerase, and ligated into the Pstl (T4 filled) site of pCSL53. The resultant plasmid was cut with Ncol and Sphl. filled with T4 polymerase, and religated to give pCSL58. Figure 8b shows the important features of pCSL58, including the DNA sequence around the translation initiation site.

Expression of PLD in E.coli.

E.coli HB101 was transformed with several plasmids containing the PLD gene. In each case, cultures of the transformed cells yielded substantial levels of PLD, as assessed by sphingomyelinase activity (Table 2). Where the PLD gene was under the control of its own promoter (i.e. in pCSL33, pCSL40 and pCSL41) PLD expression was around 5-fold lower than when under control of the trc promoter in pCSL58. Assuming equivalent copy numbers of the plasmids (since each was derived from pBR322) the difference in PLD expression must largely reflect the difference in efficiency between the two promoters. In all but one case, around 90% of the PLD produced in E.coli was localised in the periplasm of the transformed cells. This has been observed previously (Example 3) and was expected since the PLD gene encodes a typical signal sequence. The exception to the above observation was in the case of cells transformed with pCSL58 and induced in early log phase. These cells produced the highest levels of PLD in E♦coli. with around 70% being secreted into the culture medium. Viable counts taken at stationary phase indicated that release of PLD into the medium was not due to release of the periplasmic contents through cell lysis (results not shown).

When E.coli transformed with pCSL33, pCSL40, pCSL41 or pCSL58 were plated onto media containing sheep red blood cells and filtered Rhodococcus egui culture supernatant, zones of haemolysis were clearly visible around the bacterial colonies. The synergistic activities of C.Pseudotuberculosis PLD and R.egui exotoxin have previously been noted, and form the basis of a synergistic haemolytic assay (SHA) (Egen, et.al.. 1989). Using a plate screening procedure based upon the SHA it should be possible to select inactive forms of PLD resulting from mutagenesis of the PLD gene. This may enable the derivation of inactive PLD analogues which could form the basis of a genetically-toxoided vaccine. TABLE 1 PLD production in different corynebacterial species.

PLD activity in medium^a(%)

Species Plasmid Zaki titre Sphingomyelinase (μ/ml) activity (μg/ml)

C. pseudotuberculosis 6400 (100) 20.0 (100) C. ulcerans 256 (4.0) 1.1 (5.5) pCSL41 256 (4.0) 1.7 (8.5)

C. glutamicum <2 (0.0) N.D.^C pCSL41 16 (0.3) N.D.

B. lactofermentum <2 (0.0) N.D. pCSL41 32 (0.5) N.D. pCSL42 32 (0.5) N.D.

B. flavum <2 (0.0) N.D. pCSL41 128 (2.0) N.D. pCSL42 256 (4.0) 1.2 (6.0)

PLD activity in the culture medium was measured using the modified Zaki assay or a colorimetric sphingomyelinase assay (see Materials and Methods).

Percentage of activity detected relative to non-recombinant C.pseudotuberculosis.

Not determined.

TABLE 2 PLD production in E.coli transformed with various plasmids.

Sphingomyelinase activity⁸ (j_tg/ml) Percentage"

Plasmid Periplasm Medium Total activity

pCSL33 1.9 0.18 2.1 10.5 pCSL40 2.3 0.16 2.5 12.5 pCSL41 1.6 0.09 1.7 8.5

PCSL58 (0. .4 ind.^c) 4.4 9.6 14.0 70.0

PCSL58 (1. .0 ind.^d) 8.0 0.6 8.6 43.0

^a Sphingomyelinase activity was estimated in periplasmic extracts and culture supernatants using a colorimetric assay (see Materials and Methods). ^b Percentage of total sphingomyelinase activity relative to culture supernatant from non-recombinant C.pseudotuberculosis (20.0 μg/ml; see Table 1). ^c Induced at A^ 0.4. ^d Induced at A^,, 1.0.

EXAMPLE 5

In order to increase the yield of PLD that could be achieved in E.coli, the gene was cloned into the plasmid expression vector pGEX-1 (Smith and Johnson, 1988).

The pGEX vectors are constructed so that foreign polypeptides are made as fusions with -COOH terminus of a 26 kDa glutathione-S-transferase (GST) protein. These are claimed to be usually soluble or partially soluble in aqueous solution. The protein is made under the control -of a strong IPTG inducible taσ promoter so large quantities can be produced. After harvesting, the E.coli cells are sonicated and the fusion protein purified from crude lysates by affinity chromatography on immobilised glutathione agarose. The GST-wild type toxin (GST-PLD) was found to bind well to the glutathione agarose column and had an increased toxicity compared with wild-type toxin (Fig.9). The antibody response was also measured in mice injected with PLD toxin (CLA) and GST-PLD (CLA-GST). Fig.10 shows a greatly increased antibody titre in mice injected with the fusion protein. This result suggests that E.coli- derived GST-PLD fusion protein may be useful as an immunogen for eliciting anti-PLD immunity.

REFERENCES:

Aiba et.al. (1981), J.Biol.Chem. 256:11905-11910.

Amman, E. and Brosius, J. (1985). Gene 40:183-190.

Barr, M., Glenny, A.T.and Stevens, M.F. (1954). J.Hyg. 52: 379.

Batey, R.G. (1986). Aust. Vet Journal 63:269-272.

Beveridge, W.I.B (1983). Animal Health in Australia, Vol.4. Bacterial diseases of cattle, sheep and goats. Australian Government Publishing Service, Canberra.

Boyd, J. and Murphy, J.R. (1988). J.Bacteriol. 170:5949- 5952.

Burrell, D.H. (1978). Proc. 55 Ann.Conf. Aust.Vet.Assoc. : 79.

Calos, M.P., Lebkowski, J.S. and Botchan, M.R. (1983). Proc. Natl. Acad. Sci. (USA). 80:3015-3019.

Dunican, L.K. and Shivnan, E. (1989). Biotechnol. 7:1067-1070.

Egen, N.B., Cueras, W.A., McNamara, P.J., Sammons, D.W., Humphreys, R. and Songer, J.G. (1989). Am. J. Vet. Res. 50:1319-1322.

Eggleton, D.G., Middleton, H.D., Doidge, C.V. and Minty, D.W. (1987a). Aust.Vet.J.. submitted for publication.

Eggleton, D.G., Doidge, C.V., Middleton, H.D. and Minty, D.W. (1987b). Aust.Vet.J.. submitted for publication.

Gatt, S., Dinur, T. and Barenholz, Y. (1978). Biochim.Biophvs.Acta 530:503-507.

Haynes, J.A. and Britz, M.L. (1990). J. Gen. Microbiol. (in press).

Hsu, T.Y. , Renshaw, H.W., Livingston, C.W., Augustine,

J.L., Zink, D.L. and Gauer, B.B. (1985). Am. .Vet.Res. 46:1206-1211.

Kaczorek, M., Delpeyroux, F., Chenciner, N., Streek, R.E., Murphy, J.R., Boquet, P.and Tiollais, P. (1983). Science 221:855-858.

Kendall, D.A., Bock, S.K. and Kaiser, E.T. (1986). Nature 321:706-708. Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982). Molecular cloning - a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

Martin, J.F., Santamaria, R. , Sandoval, H., del Real, G. , Mateos, L.M., Gil, J.A. and Aguilar, A. (1987). Biotechnol. 5_:137-146.

McClure, W.R. (1985). Ann. Rev. Biochem. 54:171-204.

Meinkoth, J. and Wahl, G. (1984). Anal.Biochem. 138:267-284.

Nairn, M.E., Robertson, J.P. and McQuade, N.C. (1977). Proc.54 Ann.Conf.Aust.Vet.Assoc. : 159.

Ohtsuka, E., Matsuki, S., Ikehara, M., Takahashi, Y. and Matsubara, K. (1985). J.Biol.Chem. 260:2605-2608.

Ozaki, A., Katsumata, R., Oka, T. and Furuya, A. (1984). Mol. Gen. Genet. 196:175-178.

Reznikoff, W. and Gold, L. (1986). Maximising gene expression, Butterworths, Boston.

Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989). Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, New York.

Santamaria, R.I., Martin, J.F. and Gil, J.A. (1987). Gene 56:199-208.

Serwold-Davis, T.M., Groman, J. and Rabin, M. (1987). Proc. Natl. Acad. Sci. (USA). 84:4964-4968.

Smith, D.B. and Johnson, K.S. (1988). Gene 67:31-40.

Souckova, A. and Soucek, A. (1972). Toxicon 10: 501-509.

Tabor, S. and Richardson, C.C. (1987). Proc. Natl. Acad. Sci. (USA). 84:4767-4771.

Vieira, J. and Messing, J. (1987). Methods Enzymol. 154:3-11. von Heijne, G. (1985). J.Mol Biol. 184:99-105.

Waite, M. (1987). Handbook of lipid research 5. The phospholipases. Plenum Press, New York.

Yoshihama, M., Higashiro, K. , Rao, E.A., Akedo, M., Shanabruch, W.G., Follettie, M.T., Walker, G.C. and Shinskey, A.J. 91985). J. Bacteriol. 162:591-597. Zaki, M.M. (1965). Nature. 205:928-929. Zaki, M.M. (1968). Res. Vet. Sci. £ : 489.

Claims

CLAIMS:

1. Corvnebacterium pseudotuberculosis phospholipase D (PLD) toxin in substantially pure form.

2. A method for the preparation of

C.pseudotuberculosis PLD toxin in substantially pure form which comprises the steps of:

(i) concentration of a crude PLD toxin preparation by ultrafiltration, followed by (ii) chromatography of the concentrated PLD crude toxin on a cation exchange resin.

3. A method according to claim 2 wherein said ultrafiltration step is performed with a membrane having a 10,000 MW cut-off.

4. A method according to claim 2 wherein said chromatography step is performed on a carboxymethyl cellulose column.

5. A method according to claim 4, wherein said column is eluted with a 0.1M NaCl/phosphate buffer pH 6.3.

6. C.pseudotuberculosis PLD toxoid prepared from toxin according to claim 1.

7. A method for the preparation of the PLD toxoid of claim 6, which comprises incubating C.pseudotuberculosis PLD toxin in substantially pure form with a toxoiding agent.

8. A method according to claim 7 wherein said toxoiding agent is formaldehyde.

9. A vaccine composition for use against caseous lymphadenitis (CLA) in sheep, which comprises a C.pseudotuberculosis PLD toxoid according to claim 6, and optionally an adjuvant.

10. A vaccine composition according to claim 9, further comprising clostridial antigens.

11. A method of immunising sheep against CLA, which comprises administration of an effective amount of a vaccine composition according to claim 9.

12. A method for the diagnosis of CLA in sheep, characterised in that a sample taken from the animal is subjected to immunoassay for antibody to

C.pseudotuberculosis using C.pseudotuberculosis PLD toxin in substantially pure form as antigen.

13. A recombinant DNA molecule comprising a nucleotide sequence which codes for all or a substantial portion of C.pseudotuberculosis PLD toxin, or a sequence related thereto by base substitution, insertion or deletion.

14. A DNA molecule according to claim 13 comprising a nucleotide sequence which codes for a polypeptide with the amino acid sequence 1 to 283 shown in Figure 4 or a sequence related thereto by amino acid substitution, insertion or deletion, or for parts thereof with the antigenicity of C.pseudotuberculosis PLD toxin.

15. A DNA molecule according to claim 13 comprising a nucleotide sequence corresponding to all or part of the nucleotide sequence of the C.pseudotuberculosis PLD gene shown in Figure 4, or the degenerate forms thereof, or a sequence related thereto by base substitution, insertion or deletion.

16. A recombinant cloning vector containing a DNA molecule according to claim 13.

17. A replicable expression vector capable of expressing in a transformed host cell a nucleotide sequence which codes for all or a substantial portion of C.pseudotuberculosis PLD toxin, or a sequence related thereto by base substitution, insertion or deletion.

18. A transformed host cell capable of expressing a nucleotide sequence which codes for all or a substantial portion of C.pseudotuberculosis PLD toxin, or a sequence related thereto by base substitution, insertion or deletion.

19. A transformed host cell according to claim 18, said host cell being E.coli.

20. A transformed host cell according to claim 19, wherein expression of said nucleotide sequence is under the control of the trc promoter.

21. A transformed host cell according to claim 18, said host cell being a coryneform bacterium.

22. A synthetic polypeptide coded by a DNA molecule according to any one of claims 13 to 15, or by a replicable expression vector according to claim 17, having the antigenicity of C.pseudotuberculosis PLD toxin.

23. A synthetic polypeptide with the amino acid sequence 1 to 283 shown in Figure 4 or a sequence related thereto by amino acid substitution, insertion or deletion, or parts thereof with the antigenicity of

C.pseudotuberculosis PLD toxin.

24. A process for preparing a synthetic polypeptide according to claim 22 or claim 23, which comprises culturing a transformed host cell according to any one of claims 18 to 21 under suitable conditions, and isolating and purifying the expressed polypeptide.