Background and Prior Art

This invention relates to novel uses for the Pasteurella multocida toxin.

P multocida toxin has been identified as a causative agent of atrophic rhinitis, a disease of growing pigs which results in twisting of the snout and atrophy or loss of the nasal turbinate bones. Intraperitoneal injection of crude preparations of the toxin has been shown to produce turbinate atrophy (Rutter & Mackenzie, 1984). Nasal lesions produced by the toxin included atrophy of mucosal glands, epithelial hyperplasia, osteolysis and proliferation of mesenchy al cells. Other degenerative, obstructive and hyperplastic lesions were observed in the liver, ureter and bladder.

The toxin has since been purified and it has been demonstrated (Chanter et al , 1986b) that the purified toxin could reproduce all the effects noted above. (For a review see Chanter and Rutter, 1989). In 1987, the Institute for Animal Health reported the cloning of the toxin gene (J M Rutter, in "Virulence Mechanisms of Bacterial Pathogens" Ed J A Roth, p 234).

WO 89/09617 (published October 1989) discloses the use of the toxin and analogues thereof, produced by recombinant techniques, in vaccines against atrophic rhinitis.

Summary of the Invention

e have now found, surprisingly, that the toxin causes marked proliferation of cells in culture. We have shown that the toxin is a very powerful mitogen for 3T3 cells and other fibroblasts. Epithelial hyperplasia and proliferation of mesenchymal cells had been observed in vivo after treatment of pigs with P multocida toxin (see above) . However, the mode of action was not known (Chanter et al 1986b) and, in view of the complex situation in vivo, the proliferative effects might have been secondary effects or even a repair response to damaged tissue. Moreover, the toxin is cytotoxic for embryonic bovine lung cells in culture (Rutter and Luther, 1984) and all other in vitro studies had also shown only cytotoxic effects of toxin. Thus, recent papers (Elling et al 1988, Cheville et al 1988) have been entirely in the context of the toxin being necrotic and have noted the similarity between the P multocida toxin and other necrotic bacterial toxins.

One aspect of the invention provides a non- immunogenic pharmaceutically acceptable composition comprising the P mul tocida toxin or a fragment or variant thereof and one or more carriers or diluents. Advantageously the composition is suitable for topical administration to the skin or the cornea.

By "non-immunogenic" we mean that the compositions are not deliberately immunogenic, in other words no component is included in the composition in order to provoke or to help provoke an immune response to the composition. The compositions are not vaccines and are not intended to be used for raising immune sera. Compositions of the invention may, nevertheless, have an incidental degree of immunogenicity if this does not compromise their effective use.

By " P multocida toxin", we mean the toxin produced by a toxigenic strain of P multocida such as strain LFB3. To avoid any possible ambiguity, we have deposited our strain of P mul tocida with the NCIMB, Aberdeen under the terms of the Budapest Treaty (Accession No NCIMB 40158, deposited 26th June 1989). A toxigenic strain (45/78) is also available from the National Collection of Type Cultures, London, as NCTC 12178. The toxin produced by P mul tocida (in the following occasionally abbreviated to

PMT) which, as noted above, is generally believed to be the causative agent of porcine atrophic rhinitis, has in the prior literature been variously termed "dermonecrotic toxin", "osteolytic toxin", "turbinate atrophy toxin" and "heat labile exotoxin", but it would appear to be the same toxin as the amino acid composition, iso-electric point and biological activities of the variously termed toxins show basic similarities, although minor variations in the properties of toxins isolated from different strains of P multocida appear to exist. The estimated amino acid composition of PMT (as deduced from the DNA sequence) is as follows:

Ala is found 76 times - 5.91%

Cys is found 8 times - 0.62%

Asp is found 71 times - 5.53%

Glu is found 100 times - 7.78%

Phe is found 69 times - 5.37%

Gly is found 71 times - 5.53%

His is found 19 times - 1.48% lie is found 92 times - 7.16%

Lys is found 70 times - 5.45%

Leu is found 127 times - 9.88%

Met is found 36 times - 2.80%

Asn is found 73 times - 5.68%

Pro is found 62 times - 4.82%

Gin is found 56 times - 4.36%

Arg is found 58 times - 4.51%

Ser is found 97 times - 7.55%

Thr is found 66 times - 5.14%

Val is found 63 times - 4.90%

Trp is found 18 times - 1.40%

Tyr is found 53 times - 4.12%

The total number of amino acid residues is 1285, and the full-length toxin has a molecular weight of 146.5 kd.

The toxin gene will preferably have the DNA sequence of Figure 6 below or a DNA sequence which encodes the same amino acid sequence and the toxin molecule will preferably have a corresponding amino acid sequence. By "gene" we mean the nucleotide sequence together with its regulatory sequences. The term "variants" is intended to include (but not necessarily to be restricted to) minor variations in amino acid residues (such as molecules lacking one or a few residues, having conservative substitutions or minor insertions of residues, or having minor variations of amino acid structure) which do not reduce to less than 10% (preferably no less than 20%, 30%, 40%, 50%, 60%, 70%, 80% or 90%) the ability of the compound to induce proliferation of 3T3 cells, or DNA synthesis in 3T3 cells, or inhibit cAMP production in 3T3

cells in the assays described below or to potentiate effects on the proliferation or DNA synthesis or cAMP production of another agent, for example a growth factor (eg EGF). Variants may have 80%, preferably 85%, 90%, 95% or 99% homology with the P multocida toxin amino acid sequence depicted in Figure 6.

Conservative substitutions are those where one or more amino acids are substituted for others having similar properties such that one skilled in the art of polypeptide chemistry would expect at least the secondary structure, and preferably the tertiary structure, of the polypeptide to be substantially unchanged. For example, typical such substitutions include alanine or valine for glycine, arginine or asparagine for glutamine, serine for asparagine and histidine for lysine. Variants may alternatively, or as well, lack up to ten (preferably only one or two) amino acid residues in comparison with P multocida toxin; preferably any such omissions occur at the carboxy terminal of the polypeptide. Similarly, up to ten, but preferably only one or two, amino acids may be added, again at the carboxy terminal for preference.

"Fragments" of the toxin are those which display useful cell proliferative properties as defined above in relation to variants of the toxin and which have (or

which comprise regions having) at least 50% homology, preferably 60%, 70%, 80%, 90%, 95% or 99% homology, with the region of the toxin sequence to which they are most similar. The fragments preferably have a molecular weight of at least 10000.

Useful fragments may be selected by reference to the characteristics detailed below. Particular fragments include those of approximate molecular weight 23000, 64000, 74000, 87000 and 138000, obtainable by the methods of Nakai and ume (Nakai & Kume 1987a, 1987b and Kume & Nakai 1985) .

Unless the context indicates otherwise, the term "toxin" is used in this specification to include fragments and variants thereof.

The toxin used in the compositions and methods of the invention may be isolated and purified by known techniques, using a toxigenic strain of P multocida . The Rutter & Mackenzie (1984), Chanter et al (1986b) and Chanter & Rutter (1989) articles are all incorporated herein by reference for this purpose. Alternatively, the toxin may be prepared by recombinant DNA techniques, as are known in the art and as are explained in more detail below and in WO 89/09617 which is incorporated herein by

reference. Variants and fragments may be made by the expression of mutant genes, obtained by site-directed mutagenesis in known ways. Fragments may be obtained by proteolytic (eg enzymatic) degradation of recombinant- produced or naturally occurring material (as is disclosed by the Nakai and Kume references) or by peptide synthesis, using the general method of Marglin and Merrifield (Ann. Rev. Biochem. , 39 , 841-866, 1970) or by the Fmoc-polyamide method of Atherton, Sheppard and their co-workers and by subsequent refinements of these approaches.

The nucleotide sequence may be derived from a P multocida genome by screening for genomic sequences hybridizing to a DNA probe prepared on the basis of the full or partial amino acid sequence of the toxin in accordance with established procedures or by establishing a toxin gene library and screening for toxin-producing clones by means of a toxin-specific antibody (for a more detailed description of this procedure, see WO 89/09617, Example 4) or using an overlay technique as is explained in more detail below. The gene may be prepared more directly by isolating a 15kb Hpall fragment as is described in Lax & Chanter (1990) J Gen Microbiol 136, 81-87, which is incorporated herein by reference.

The nucleotide sequence may also be derived from a bacteriophage infectious for P mul tocida , ie one which has been transferred from one bacterial strain which originally carried the sequence to another strain which did not originally carry the sequence by bacteriophage transfection. Similarly, the nucleotide sequence may be derived from a plasmid or other genetic element transferred from one strain to another by conjugation, transformation or the like.

Furthermore, the nucleotide sequence coding for the toxin may be a synthetic sequence, that is, one prepared according to standard procedures, eg as described in Matthes et al (1984). Finally, the nucleotide sequence may be a mixed genomic and synthetic or mixed cDNA and synthetic sequence prepared by ligating DNA fragments of genomic, cDNA or synthetic origin (as appropriate) which DNA fragments each contain part of the nucleotide sequence encoding the toxin, in accordance with established methods.

In accordance with the explanation given above, the DNA fragment may be one which has been modified by substitution, addition, insertion or deletion of one or

more nucleotides in the sequence with the purpose of establishing a sequence which, when expressed, results in the production of a useful toxin or toxin variant.

In particular, the invention relates to a DNA fragment which comprises a nucleotide sequence substantially as shown in Fig 6 (a)-(j) or a modification thereof as indicated above. The sequence coding for the full-length toxin starts at position 219 (or 213) of the sequence shown in the figure, while the end of the sequence is at position 4073. The DNA sequence shown in Fig 6 (a)-(j) has been established by well-known methods.

The DNA fragment of the invention may further comprise a nucleotide sequence encoding another polypeptide fused to the nucleotide sequence encoding the toxin with the purpose of producing a fused polypeptide, as explained above. A further purpose of preparing a fused polypeptide may be to facilitate purification of the toxin. In this case, the fused sequence may be inserted into an appropriate vector which is transformed to a suitable host microorganism which is grown under conditions ensuring expression of the fused sequence after which the fused polypeptide is recovered from the culture by subjecting the fused polypeptide to affinity chromatography involving an antibody or any other ligand

reacting with the second polypeptide. After purification, the second polypeptide may then be removed, for instance by suitable proteolytic cleavage followed by separation of the two polypeptides.

Generally, production of the toxin may involve:

a) isolating a nucleotide sequence coding for the P multocida toxin,

b) inserting said sequence, optionally in suitably modified form in an expression vector,

c) transforming a suitable host microorganism with the vector produced in step b) ,

d) cultivating the microorganism produced in step c) under suitable conditions for expressing the toxin,

e) harvesting the toxin from the culture, and

f) optionally subjecting the toxin to posttranslational modifications to produce a (further) variant.

In step a) of the method, the nucleotide sequence may for instance be isolated by establishing a P mul tocida gene library and screening for toxin-positive clones in accordance with established methods as indicated above as well as described in detail below.

In step b) of the method, the modification of the sequence optionally carried out may be performed before or after the sequence has been inserted in the vector. The modification may comprise substitution, addition, insertion or deletion of one or more nucleotides in the sequence or a combination thereof, as explained above.

The transformation in step c) of the method may be carried out by standard procedures, such as disclosed in Maniatis et al (1982).

The cultivation of the host microorganism in step d) of the method may be carried out in a culture medium conventionally used for fermentation purposes, eg Luria Broth medium, and under conditions with respect to pH, temperature, aeration, etc suited to the type of microorganism in question, eg as disclosed in Maniatis et al (1982).

In step e) of the method, the harvesting of the toxin may proceed by well-known methods such as by precipitation, gel filtration, ion exchange or HPLC reverse phase chromatography or immunoaffinity chromatography.

Apart from the toxin as defined above, the compositions of the invention also comprise an acceptable carrier or vehicle. This vehicle may be any vehicle usually employed in the preparation of pharmaceutical compositions, for example a diluent such as isotonic saline or suspending agent. The composition may be prepared by mixing an effective amount of the toxin with the vehicle in an amount resulting in the desired concentration of the toxin in the preparation.

The mitogenic properties of the toxin mean that it may be put to many uses . The toxin may be used to aid wound healing and the invention therefore also provides a method of promoting wound healing in a mammal, for example healing skin ulcers, burns, severe cuts or abrasions, facial incisions resulting from cosmetic surgery, or cuts or burns on the cornea.

The toxin may also be used to accelerate growth of bone marrow following anti-cancer chemotherapy or irradiation. The marrow may be removed from the patient, exposed to the toxin and then returned to the patient. Because of the long-term action of the toxin, marrow with an increased growth rate can be returned to the patient without there being any extracellular toxin and, in at least some cases, without any detectable intracellular toxin. Thus, a further aspect of the invention includes cells, for example bone marrow cells (specifically, stem cells) which have altered characteristics following exposure to the toxin. The "altered characteristics" may include an increased rate of division and/or DNA synthesis in relation to a control sample which was not exposed to the toxin, ' or in relation generally to what would have been expected.

The toxin may also be used in cell fermentations, particularly eukaryotic cell fermentations (including hybridomas), to increase the yield of cells or their products, including viruses or expression products of recombinant nucleotides, or to accelerate the fermentation. A further embodiment of the invention therefore provides a cell fermentation medium comprising the P multocida toxin and being suitable for fermenting cells other than P multocida .

The cells which are most likely to proliferate when exposed to the toxin are those whose growth characteristics are influenced by protein kinase C. However, the person skilled in the art will readily be able to determine whether any given cell type responds to the toxin in a desirable way.

The toxin will function in the absence of serum, although small amounts of serum may still be desirable, for example to encourage adhesion of cells to a support. Thus, the amount of serum can be reduced from, say, 10%, to about 2% or less, for example 1%, 0.5% or 0.1%. This reduces the costs of the medium and frequently aids subsequent purification of the fermentation product.

The toxin has been found to inhibit cAMP production and may therefore be used to treat E coJi-induced diarrhoea, cholera, psoriasis or other conditions characterised by excessive cAMP production, for example the overproduction of some hormones.

The proliferative effects of the toxin may also be used to stimulate specific cell types within the body, a reverse of the so-called "magic bullet". The toxin should preferably be targeted to the specific cell type required. A further embodiment of the invention

therefore provides a conjugate of P multocida toxin or a fragment thereof and means to bind to a specific cell type. The term "conjugate" is used to cover molecules which have been made by merely joining the toxin or fragment thereof to a binding means (for example by methods taught by 0'Sullivan et al (1979) Anal Biochem 100 , 100-108) and also molecules in which the toxin and the binding means have been synthesized integrally by chemical peptide synthesis techniques or recombinant DNA techniques (ie by using fused DNA sequences to express a single polypeptide product) . The "means to bind to a specific cell type" may be any molecule which binds to a specific cell type in the body. Such a molecule may be an antibody, a naturally occurring molecule or fragment thereof or a synthetic peptide.

For example, the cell-binding portion may direct the conjugate to the bone-marrow or the gut epithelium, these being examples of rapidly-dividing tissues which may be damaged by anti-cancer chemotherapy or irradiation and whose growth one might wish to encourage. Thus, the cell-binding portion may be an antibody directed against a component of the gut pili (for example adhesins or colonisation factor antigens).

Alternatively, the cell-binding portion of the conjugate may be low density lipoprotein or a cell- binding portion of human plasma fibronectin, such as the 11.5 kDalton polypeptide described by Ruoslahti __ Pierschbacher in WO 84/00540, the Arg-Gly-Asp-Ser minimal cell-binding part thereof or any intermediate peptide between those two peptides. A nucleotide sequence coding for such a peptide is fused to a nucleotide sequence coding for the P multocida toxin or part thereof (either upstream or downstream thereof) in a known manner, incorporated in an expression vector in a suitable host and used to produce a hybrid polypeptide.

It seems likely that the toxin must be internalised for the mitogenic effect to be obtained. Thus, the conjugate is preferably one which will be internalised.

It appears, however, that the toxin need not be internalised to obtain an inhibition of cAMP production. Thus, conjugates for this purpose need not be internalised.

The toxin or a conjugate of whole toxin or an active part of the toxin with a specific binding compound (eg protein) for osteoclasts or osteoblasts (for example a monoclonal antibody specific to the osteoclast or

osteoblast) can be used to modulate the interaction between these cell types by affecting (increasing or reducing) the production of growth factors, cytokines or paracrine substances by which osteoblasts regulate the function of osteoclasts. In this way the balance of bone resorption verses bone formation can be regulated. Consequently a variety of bone disorders may be treated with different toxin conjugates, some increasing resorption others increasing bone formation. Monoclonal antibody 23C6, which is specific for the vitronectin receptor of osteoclasts is an example of a binding compound [M Horton et al "Monoclonal antibodies to osteoclastomas - definition of osteoclast-specific antigens" (1985) Cancer Res, 45, 5663-5669].

The toxin, or a conjugate of whole toxin or an active part of the toxin with a specific binding protein for lymphocytes, can be used to reduce the production of pharmacologically active lymphokines, cytokines or paracrine substances from these cells especially when they are over stimulated. The toxin conjugates may be used to treat conditions where there is excessive lymphocyte stimulation, such as that following transplantation which may cause some lymphoid malignancies, that which occurs in the incubation phase of Acquired Immune Deficiency Syndrome and may be

required for the development of disease, and that which occurs during acute infections such as those involving the staphylococcal enterotoxins. The modulation of production and secretion of pharmacologically active substances from stimulated lymphocytes can reduce the consequences of stimulation of these cells .

The P mul tocida toxin itself contains at least one cell binding portion. Many drugs need to be directed to specific cells on which they are intended to act, an example being cytotoxic drugs for tumour chemotherapy. The toxin or a cell-binding portion thereof may be used as the delivery vehicle. A further embodiment of the invention therefore provides a conjugate of P mul tocida toxin or a cell-binding variant or portion thereof and a pharmacologically active compound. Preferably, the variant or portion is not also mitogenic.

The term "conjugate" is used in the same sense as above, and encompasses chemically-joined fragments and fused hybrids produced by automated polypeptide synthesis or expression of fused nucleotides.

The pharmacologically active agent which is conjugated to the toxin or portion thereof may be any cytotoxic or antiviral agent, for example vincristine, vinblastine or methotrexate.

Cell-binding portions of the toxin may be readily identified by radio-labelling a candidate portion and detecting its location on the surface of cells, or by exposing the cells to the candidate portion and then adding a labelled antibody to the portion. In addition, the methods of WO 84/00540 or Yamada (1983) may be used. Target cells include the bladder, liver and spleen.

At least some of the various effects of the toxin and conjugates of all or part thereof may be potentiated by one or more growth factors, such as EGF, PDGF, FGFb or insulin.

Preferred aspects of the invention will now be described in detail in the following non-limiting examples and with reference to the accompanying drawings.

Description of the Drawings

Figure 1 is a restriction map of the recombinant plasmids pAJL12 and pAJLl3. The arrows indicate the direction of readthrough from the tetracycline gene of pAT153. The insert in pAJL12 was not cut by Hpall , Kpol , Pvul , Pvτill , Sail or Sstl . The BamEI site in parentheses was lost upon ligation of the vector BairiEI site to the insert SauIIIA site. Two HiπdHI sites between the Hindi11 sites at the right hand end of the insert were not mapped.

Figure 2 is a representation of an agarose gel of chromosomal DNA from P multocida LFB3; uncut (lane 1) or cut with Hpal (lane 2) or Hpall (lane 3 (cross-hatching representing dark area and open boxes representing light bands) and, alongside, a Southern blot of the gel probed with the large Hpall fragment of pALJ12 (lanes 4-6 as lanes 1-3) .

Figure 3 is a SDS PAGE of purified toxin from P multocida LFB3 (lane 1) and E coli T0X1 (lane 2).

Figure 4 shows the result of crossed i munoelectrophoresis of purified recombinant toxin reacted against gnotobiotic pig antiserum to purified toxin from P multocida LFB3.

Figure 5 shows an immunoblot probed with gnotobiotic pig serum to toxin purified from P multocida LFB3: Lane 1 purified toxin from P mul tocida LFB3, Lane 2 purified toxin from E coli T0X1, Lane 3 whole cell lysate of a recombinant which reacted with the Hpal fragment of the insert in pALJ12, Lane 4 whole cell lysate from E coli HB101 containing pAT153, Lanes 5-8 as Lanes 1-4 without antibody to toxin.

Figure 6 shows DNA sequence data for the PMT gene, taken from WO 89/09617. Our sequence, which is considered to be the more authoritative, differs therefrom in that at their position 460 they have an A whereas we have a C; at 461 they have C and we have T; at 2541, C, G; at 2542, G, C; and at 110, outside the coding region, we have an extra C which therefore alters the numbering by one residue.

Figure 7 shows the PMT amino acid sequence data, deduced from the nucleotide sequence of Figure 6.

Example 1: Preparation of P multocida toxin

Strains and Growth Conditions. P mul tocida strain LFB3 is a toxigenic isolate from a pig with atrophic rhinitis (Rutter, 1983) and E coli HB101 and HB101 harbouring

pAT153 were obtained from Dr J G Williams, Imperial Cancer Research Fund, London. All bacteria were stored as cell suspensions at -70°C in 12% (v/v) glycerol. P multocida were grown in Bacto tryptose broth (Jones and Matthews, 1975) at 37°C with agitation. E coli strains were grown on LB agar or in L broth (Maniatis et al , 1982) .

Chemicals and biochemicals. Restriction enzymes were from BRL or Biolabs, and were used according to the manufacturer's specifications. All other enzymes were from Boehringer Corporation. Ampicillin and tetracycline were from Sigma Ltd, and agarose from BRL. [ S] dATP was from Amersham International. Other chemicals were from BDH.

DNA isolation and cloning techniques. P multocida DNA was isolated by a modification of the method of Saito and Miura (1963). Cells were collected by centrifugation and washed in an aqueous solution of 0.15M NaCl, O.IM EDTA, pH8.0 and were resuspended in the same buffer at one tenth the original volume of culture. Lysozyme (lmg/ml) was added, and the mixture was incubated at 37°C for 30 minutes and was then rapidly immersed in a dry ice/acetone bath. Eight volumes of lysis solution (O.IM Tris-HCl, 0.1M NaCl, 1% (w/v) SDS and 50μg/ml proteinase

K, pH 9) was added and the mixture was incubated at 60°C for 10 minutes before being rapidly frozen. After thawing at 60°C for 30 minutes, an equal volume of phenol (Maniatis et al , 1982) was added and the suspension was mixed gently for 1-2 hours. The suspension was centrifuged and the aqueous phase was re-extracted with phenol, and after phase separation residual phenol was removed from the aqueous phase by extensive dialysis against an aqueous solution of 50mM Tris-HCl, lOmM EDTA, lOmM NaCl, pH8. The preparation was treated with ribonuclease as described by Maniatis et al (1982), and the purified DNA was stored at 4°C.

Genomic DNA from P multocida was partially digested with SauIIIA to obtain fragments about lOkb. The digest was fractionated on a sucrose gradient as described by Maniatis et al (1982). Fractions were selected containing fragments in the size range 7-12kb. The restricted DNA was ligated overnight at 15°C to pAT153, previously cut with BamEl and treated with phosphatase. Competent HB101 obtained from BRL were transformed. The transformed cells were plated onto L agar containing ampicillin (200μg/ml). After incubation at 37°C overnight, the 6500 bacterial colonies obtained were replica plated onto agar containing tetracycline (12μg/ml). The 2500 Ap ^r Tc ^s colonies were tested for

toxicity for embryonic bovine lung (EBL) cells by the overlay method (Chanter et al 1986a) and stored at -70°C in 12% glycerol in microtitre trays.

Plasmids were isolated according to Ish-Horowitz and Burke (1981). Other molecular biological techniques were as described in Maniatis et al (1982).

Nucleotide Sequencing. DNA sequencing was carried out using an Applied Biosystems 370A Sequencer with the Tag polymerase kit, according to the manufacturer's instructions. The recombinant plasmid from T0X2 was digested with Hpall and the 4.9kb fragment was separated on an agarose gel, and extracted using Geneclean (Stratech Scientific, London). The DNA was digested with Al ul or SauIIIA, and either randomly inserted into M13 (mplO, mpl8 or mpl9) or fragments were gel purified as described above prior to ligation into M13 (mplO, mpl8 or mpl9) .

Protein Sequencing. Protein sequencing was carried out using an Applied Biosystems 477A Gas Liquid Protein Sequencer with on-line 120A PTH Amino Acid analysis according to the manufacturer's instructions.

Cloning the toxin gene. One clone (T0X1) out of the 2500 Ap ^r Tc ^s clones with inserts was positive for toxin production. The plasmid from this clone, pAJL12, was purified and transformed into competent E coli cells; 20 transformed bacteria were selected and all were toxigenic.

Analysis of the Toxin gene. The plasmid pAJL12 contained a 10.7kb insert. The restriction map (Figure 1) was unusual since the insert was not cut by Hpall or Mspl , and only cut three times by Haelll. These three enzymes have a four base pair recognition sequence which contains only G and C, which suggested that the G + C ratio of the DNA might be quite low. The enzymes Oral and Spel , which have a six base pair recognition sequence containing only A and T, cut the insert at least 9 and 11 times respectively (data not shown) . Genomic DNA from P mul tocida was prepared and cut with Hpal and Hpall , and Southern blots of the gels were probed with fragments from the insert. Figure 2 shows that most of the DNA was digested by Hpall to fragments smaller than 4kb, but there were discrete bands of higher molecular weight. A band at about 15kb hybridised to the probe. Three other toxigenic P multocida isolates produced a Hpall fragment

of similar molecular weight which also reacted with the probe. A band of about 6kb in Hpal digests reacted with the probe.

This suggested that a simple method of cloning this gene or variants of it would be to extract the large Hpall fragment and insert it into a suitable vector.

The 2500 colonies in the clone bank were probed with the Hpal fragment from the insert in pAJL12, and 12 colonies hybridised to the probe. None produced toxin when tested with the EBL overlay test. Preliminary analysis of the 12 hybridising colonies showed that they contained plasmids of different size, but none contained the whole toxin gene.

Subcloning. Plasmid pAJL12 was partially digested with SauIIIA, to produce fragments about 5kb, the predicted size of the toxin gene. Fragments of size 5 - 8kb were selected and ligated into pAT153, previously BamEl cut and phosphatased. Recombinants were screened for toxicity by the EBL overlay test, and one (T0X2) contained a 5.Okb insert and produced toxin. The recombinant (pAJL13) was mapped (Figure 1), and was located at one end of the insert in pAJL12, with a duplication of about 0.lkb of the vector sequence. The

insert is in the opposite orientation to pAJL12, which implies that in at least one of the constructs the gene is being read off its own promoter.

Purification and Properties of the Toxin

Toxin Purification. Toxin was purified from a crude extract of the toxigenic recombinants T0X1 or T0X2, grown on L agar containing ampicillin, prepared by the lysis method of Rimler and Brogden (1986). Crude extract, treated with RNase, DNase, benzamidine and phenylmethyl- sulphonylfluoride, was sequentially fractionated by DEAE Sephacel chromatography and preparative polyacrylamide gel electrophoresis (Chanter et al , 1986b). In the final step purified toxin was electroeluted from the polyacryl¬ amide. Quantities of toxin in each fraction were measured using toxicity for EBL cells (Rutter and Luther, 1984) .

Characterisation of toxin purified from recombinant E coli . The homogeneity and molecular weight of the polypeptide(s) in toxin purified from the recombinants were estimated by SDS polyacrylamide gel electrophoresis by the method of Laemmli (1970) and staining of gels with silver as described before (Chanter et al , 1986b).

The antigenic similarity of the toxin purified from the recombinants with that purified from P mul tocida was determined using antiserum to toxin purified from P multocida that was produced in a gnotobiotic pig (Chanter et al , 1986b) in an enzyme-linked immunosorbent assay, immunoblotting, a cytotoxin neutralisation test (Rutter and Luther, 1984) and by crossed immunoelectro- phoresis by the method of Moore (1985).

In the ELISA, microtitre plates (Falcon-Becton Dickenson) were coated overnight with lOOμl of different concentrations of toxin in carbonate/bicarbonate buffer, pH 9.6. Plates were washed in phosphate buffered saline containing 0.03% Tween 20 (PBS/Tween) and incubated at 37°C for 1 hour with serial dilutions of pig antiserum in PBS/Tween with 1% dehydrated skimmed milk (Marvel, Cadbury Ltd). Plates were washed again 3 times in PBS/Tween and lOμl of 1:5000 rabbit anti-swine Ig conjugated to horseradish peroxidase (Nordic Immunological Laboratories) were added to each well and incubated at 37°C for 3 hours. Plates were washed again and developed with lOOμl of 0.04% O-phenylenediamine at 0.015% hydrogen peroxide buffered with citrate/hydrogen orthophosphate, pH 5.0 for 15 min when the reaction was stopped with lOμl of 2N sulphuric acid. Control wells

were treated similarly, excluding either the antigen or the pig antiserum. Optical densities of the wells were measured with a Titretec Microplate Reader.

Protein separated by SDS-PAGE were immunoblotted by the method of Towbin et al (1979) using a Transblot apparatus and the recommended protocols of the manufacturer (Biorad) . Toxicity was assayed using EBL cells (Rutter and Luther, 1984) and intraperitoneal injection of gnotobiotic pigs (Rutter and Mackenzie, 1984) and was related to protein content assayed by a Coomassie dye binding method (Biorad) .

The toxin purified from recombinants T0X1 and T0X2 (Figure 3) had the same high molecular weight as that produced by P multocida LFB3. Comparison of the yield and the efficiency of purification indicated that T0X1 produced approximately five fold as much toxin as T0X2 and ten times as much as P multocida . Like P mul tocida both recombinants produced a faint band above the main band.

A chequerboard titration of toxin was used to coat the microtitre plate for an ELISA and dilutions of gnotobiotic pig serum against toxin purified from P multocida were made. The optimum coating concentration

of antigen was Iμg/ml of coating buffer for toxin from P multocida or from either toxigenic recombinant. At this concentration the serum gave an identical titre of 10 ¹ * for all toxin preparations. Control serum did not react with any of the toxin preparations.

In a cytotoxin neutralisation test the gnotobiotic pig serum had a titre of 10^ with 10 cytotoxic units of toxin purified from either P multocida or the recombinants.

Crossed immunoelectrophoresis of toxin from the recombinants or P mul tocida with gnotobiotic pig antiserum to toxin purified from P multocida resulted in a precipitate in the basic pattern of one peak which on closer examination was composed of several closely spaced peaks (Figure 4), although only two of these could be seen reproducibly. The same preparations in SDS-PAGE and stained with silver were composed of only one major band and a fainter band with a slightly higher molecular weight.

In immunoblots of whole cell lysates of the three out of the 12 recombinant which reacted with the Hpal fragment of the insert in pALJ12, there were three polypeptides which reacted with gnotobiotic pig serum to

toxin purified from P multocida LFB3 (Figure 5). These were of approximate molecular weight 76,000, 72,000 and 48,000. The same bands were present in whole cell lysates of the toxigenic recombinants T0X1 and T0X2.

Amino Acid Analysis on hydrolysates: The attached data (Table 1) show zero for glycine and alanine; this is due to the enormous glycine peak derived from the Tris/glycine buffer that the sample was in originally. Native cysteine cannot be estimated by this technique.

We initially used material in Tris/glycine buffer and attempted to remove protein from solution in 80% acetone. The protein remained in the supernatant after 16 hours at -20°C. The first sequence was obtained from material in Tris/glycine but the initial allocation was not possible due to the interference by the glycine.

? ? ? K H F F(P)S D F T V K

To remove the glycine the sample was applied to PVDF membrane (Immobilon from Millipore) as follows: a 1cm square of PVDF membrane was wetted with ethanol then 50% ethanol/H2θ and finally H2O. This was applied to a heating block (at 55°C) covered with aluminium foil. The sample was applied in 3 x 30 μl aliquots (total of 30 μg)

and dried. The membrane was then extensively washed in

HPLC grade water to remove glycine. The centre section containing the protein was excised and placed in the sequencer.

(M) K I K H F F N S D F T V

At this point the sequence ends abruptly, perhaps indicating a modification.

This represents the N-terminal sequence of the toxin but, since this does not match the deduced amino acid sequence (the initial isoleueine is a threonine), there appears to be an error, probably in the amino acid data.

Partial sequence data are as follows.

Single strand DNA inserts found in Sample Nos 2-33, 101- 124, Al-24 (minus A17 & A21), and Bl-24.

IAH SAMPLE No 6

TCTTATTAGCGCTTCATATAGANGGGCTGTGGANGTACAGATTAGATCANATNCA

GAAGCTACCCGTGACTATGATCAAAAAAATACAANGCTAAAAGAAAAATTGCA

ACAATTAGAGGAGCNGTTANAGGAAGCTATTGAAAGAGGACAGAGGGCAATCTC

TCAAGCTCAATTAACAAAATCAGGACATGTTTATGTAATCAGTAATATTGGTTC

ATTTGGT - 222 bases

IAH SAMPLE Nos 2, 3, 8, 10, 11, 14, 16, 17, 19, 23 CCTAGTTTCGATTTTAAGGCTTTTGAGAC - 29 bases

IAH SAMPLE Nos 4, 5, 7, 9, 12, 15, 18, 20, 21, 24 GCTTGAAGATAGTGATGTACAGATTAGAT - 29 bases

IAH SAMPLE No 110, 118, 119, 120, 121, 123, B3 Ca.TCTGTATTAAGAAATTCAACAAATGGTTGTGTTCCTCCAGGAATGTAA AGTA - 54 bases

IAH SAMPLE No BIO

TTGCATGCCTGCAGGTCGACTCTAGAGGATCAGAANCAGGAGNNTNGNNAANNTT

ACGTGTGGCAAAGGCCTGCT(___J^AATCCTNNNNCACCTTGTNCTTCTTGANGCA NAC

TNNNTAAAGTGAAAGGNTATNAATTCATAAGGCCCTCCACCCTGGTTTGAATATG

GAGAAATNCTGGTAAAACATTTCTTCATAGCTATNGCCCAGTATAAAGGCTATT

TTCTATGATCACGAC^CTTACNTTANNNGCCCCTTGCTTTTCGATNC^CGTAACAA

NCCAGGNCAGNAATGCGCTATAACTCCCTNCGTGCCAGCNTGCANAAGCAGGGAAG

TCTCC^GTAANNNTNTTNGTNTTTTACTCATATACAGGANATTTTTNGCTNTTAC

ATATTTTACCAGGGATATCTTCANTTGCTNCCNTGTNACGCANCCNAGCNTTAGGC

CGCGATATGGGCGCCGGCANTNATCNG - 470 bases

Total of 1018 bases

Table 1

AMINO ACID ANALYSIS DATA

MOLAR 1

Enter pM of selected AA in F7 117.0 ERIAH2

Enter lowest No for AA in F8 90.0 91 92 93_ pMOLES

ASP 225 173.1 ### ### 178.85

76.9 77.78 78.83 79.49

119.2 ### ### 123.21

177.7 ### ### 183.62

61.5 62.22 62.91 63.59

0.0 0.00 0.00 0.00

0.0 0.00 0.00 0.00

0.0 0.00 0.00 0.00

63.8 64.56 65.26 • 65.97

63.1 63.78 64.48 65.18

90.0 91.00 92.00 93.00

153.1 ### ### 158.18 50.8 51.33 51.90 52.46

75.4 76.22 77.06 77.90

61.5 62.22 62.91 63.59

106.2 ■### ### 109.69 44.6 45.11 45.61 46.10

Total AA's 1316.92. MW based on column F: 157361. Total sample in pMoles: 38.89 or 0.04 nM. Sample in pMoles: 1.50. Total Wt: 6.12μg

These data have been largely superseded by the deduced sequence data of Figure 7. Anomalies therewith can be explained by the fact that the analyser was zeroed for glycine and alanine (glycine being present in the bufferV does not detect cysteine and does not distinguish between Asp and Asn.

Example 2: Mitogenesis in 3T3 Cells

To determine whether native Pasteurella multocida toxin (PMT) can modulate the mitogenic response of murine Swiss 3T3 cells, confluent and quiescent cultures of these cells were washed and transferred to medium containing increasing concentrations of PMT. Cumulative ( )thymidine incorporation was measured after 40 h of incubation. We showed that PMT was an extremely potent inducer of DNA synthesis in Swiss 3T3 cells. Half- maximum effect was obtained at a concentration as low as 0.32 ng/ml (approximately 0.2 pM) . The maximum effect, obtained at 1.25 ng/ml, was equivalent to the stimulation of DNA synthesis induced by medium containing 10% fetal bovine serum. Thus, in contrast to most known mitogens

for Swiss 3T3 cells, PMT at picomolar concentrations induces maximal DNA synthesis in the absence of any other synergistic factor.

We also found that rPMT stimulated DNA synthesis in Swiss 3T3 cells as potently as native toxin. The half maximal effect was achieved at 0.15 ng/ml (0.1 pM) and maximum effect at 1.25 ng/ml (0.83 pM) . In contrast, no mitogenic activity could be demonstrated with bacterial extracts derived from E coli transformed with plasmid without the PMT gene. All subsequent experiments were performed using the recombinant toxin.

Effect on DNA synthesis of a brief incubation with rPMT

Many bacterial toxins bind to the external surface of the plasma membrane, enter the cells and then cannot be removed by extensive washing or neutralized by antibodies. To test whether rPMT acts in this manner to stimulate mitogenesis, quiescent 3T3 cells were incubated with rPMT at 1, 5 or 20 ng/ml for various times. When the cells were exposed for 30 min to 20 ng/ml rPMT, washed extensively and incubated in medium without the toxin, the level of ( ~U )thymidine incorporation was similar to that induced in the cultures incubated continuously with rPMT. Lower concentrations of rPMT

required longer preincubation times to induce maximal DNA synthesis after removal of unbound toxin. Thus, the potent mitogenic effect of rPMT persists after the medium containing it has been removed. In other experiments, the cells were preincubated with 5 ng/ml rPMT for various times and then transferred to media in the absence or presence of PMT antiserum. When the cells were treated with 5 ng/ml rPMT for 1 h, subsequent DNA synthesis was markedly inhibited in cultures transferred to medium containing antiserum. In contrast, after 3 h of incubation with rPMT the mitogenic effect of the toxin is no longer blocked by the addition of antiserum. Thus it is plausible that the rPMT undergoes a time-dependent internalization into a compartment not accessible to external antibodies.

rPMT stimulation of cell proliferation

The mitogenic activity of rPMT in cultures of Swiss 3T3 cells also could be readily shown when cell number (rather than ( )thymidine incorporation into acid- precipitable material) was monitored over a period of several days either in confluent or subconfluent cells. Addition of rPMT at 10 ng/ml to the medium in which confluent and quiescent 3T3 cells were grown (depleted medium) resulted in loss of density-dependent inhibition

of growth: after 3 days, the cell number was about twice that of the control. rPMT stimulated reinitiation of cell proliferation in a concentration-dependent manner. In other experiments, addition of rPMT to subconfluent 3T3 cells resulted in a striking increase in cell proliferation; the final saturation density increased 6- fold. In view of the results, we tested whether a transient exposure to rPMT would be sufficient to stimulate cell proliferation in toxin-free medium. When quiescent cultures were exposed to 10 ng/ml rPMT for 24 h and then washed, trypsinized and replated in the absence of the toxin, the subsequent growth of rPMT-pretreated cells was markedly enhanced.

Example 3: Effect of rPMT on DNA synthesis and cell division in other cells

Several urine cell lines including Balb/c 3T3 cells, NIH-3T3 or 3T6 which were rendered quiescent by growth to confluency (Balb/c and NIH-3T3) or by incubation in 0.5% serum (3T6) responded to rPMT with a striking increase in cell proliferation. The dose response curves for stimulation of DNA synthesis were similar to those obtained with Swiss 3T3 cells. Interestingly, the toxin also stimulated DNA synthesis and cell proliferation in tertiary cultures of mouse

embryo cells. The growth-promoting effects of rPMT are not confined to murine cells. Table 2 shows that rPMT stimulated ( H)thymidine incorporation in quiescent cultures of human fibroblasts. The toxin was more effective than PDGF, EGF or FGF in these cells. The effect of rPMT was markedly enhanced in the presence of these factors (Table 2).

TABLE 2

Incorporation of [^H]thymidine Without rPMT With rPMT

0 9 60

Insulin 16 75 rFGFb 28 82

EGF 38 104

PDGF 33 119

Numbers expressed as cpm x 10"^, means of two determinations. The concentrations used were as follows: rPMT (10 ng/ml); insulin (1 μg/ml); rFGFb (5 ng/ml); EGF (5 ng/ml) and PDGF (10 ng/ml). 10% FBS gave

incorporations of 139 x 10^ cpm with rPMT. Human foreskin fibroblasts were used at the 18th passage and rendered quiescent by incubation in 0.5% FBS for 4d.

Abbreviations used:

PMT native P multocida toxin rPMT recombinant P multocida toxin

PDGF platelet-derived growth factor

EGF epidermal growth factor rFGFb recombinant fibroblast growth factor

(basic)

PDBu phorbol 12,13 dibutyrate

FBS foetal bovine serum

DMEM Dulbecco's modified Eagle's medium

References

Chanter, N, Rutter, J M and Luther, P D (1986a). Rapid detection of toxigenic Pasteurella multocida by an agar overlay method. Veterinary Record 119, 629-630.

Chanter, N, Rutter, J M and Mackenzie, A (1986b).

Partial purification of an osteolytic toxin from

Pasteurella multocida . Journal of General Microbiology 132, 1089-1097.

Chanter, N & Rutter, J M (1989) "Pasteurella and Pasteurellosis" 161-195.

Cheville, N F et al (1988) Vet Pathol 25, 518-520.

Elling, F et al (1988) APMIS 96, 50-55.

Ish-Horowitz, D and Burke, J F (1981). Rapid and efficient cosmid cloning. Nucleic Acids Research 9 , 2989 -2998.

Jones, P W and Matthews, P R J (1975). Examination of slurry from cattle for pathogenic bacteria. Journal of Hygiene, 74, 57-64.

Kume K, & Nakai T (1985) Jpn J Vet Sci 47(5), 829-833.

Laemmli, ϋ K (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 681-685.

Maniatis, T, Fritsch, E F and Sambrook, J (1982). "Molecular Cloning: A Laboratory Manual" Cold Spring Harbor Laboratory, Cold Spring Harbor, NY (and 2nd Edition thereof) .

Ma thes, H W D, Zenke, W H, Grundstrδm, T, Stauls, A, Wintzerith, M and Chambon, P (1984) EMBO J 3, 801-805.

Nakai, T & Kume, K (1987a) FEMS Microbiol Lett 44, 259- 265.

Nakai, T & Kume, K (1987b) Res Vet Sci 42, 232-237.

Moore, L J & Rutter, J M (1987). Antigenic analysis of fimbrial proteins from Moraxella bovis. J Clinical

Microbiology 25, 2063-2070.

Ri ler, R B & Brogden, K A (1986). American Journal of Veterinary Research, 47, 730-737.

Ruoslahti, E I & Pierschbacher, M D. WO 84/00540

Rutter, J M (1983). Virulence of P multocida in atrophic rhinitis of gnotobiotic pigs infected with B bronchiseptica , Research in Veterinary Science, 34, 287-295.

Rutter, J M & Luther P D (1984). Cell culture assay for toxigenic Pasteurella multocida from atrophic rhinitis of pigs. Veterinary Record 114, 393-396.

Rutter, J M and Mackenzie, A (1984). Pathogenesis of atrophic rhinitis in pigs: A new perspective. Veterinary Record 114, 89-90.

Saito, H and Miura, K (1963). Preparation of transforming deoxyribonucleic acid by phenol treatment. Biochimica et Biophysica Acta 72, 619-629.

Yamada, K M (1983) Ann. Rev. Biochem. 52, 761-99