Search

The current pertussis vaccines, although efficacious, in some instances produce undesirable side effects. Molecular engineering of pertussis toxin, the major protective antigen, could provide a safer, new generation of vaccines against whooping cough. As a first critical step in the development of such a vaccine, the complete nucleotide sequence of the pertussis toxin gene was determined and the amino acid sequences of the individual subunits were deduced. All five subunits are coded by closely linked cistrons. A promoter-like structure was found in the 5'-flanking region, suggesting that the toxin is expressed through a polycistronic messenger RNA. The order of the cistrons is S1, S2, S4, S5, and S3. All subunits contain signal peptides of variable length. The calculated molecular weights of the mature subunits are 26,024 for S1, 21,924 for S2, 21,873 for S3, 12,058 for S4, and 11,013 for S5. Subunits S2 and S3 share 70% amino acid homology and 75% nucleotide homology. Subunit S1 contains two regions of eight amino acids homologous to analogous regions in the A subunit of both cholera and Escherichia coli heat labile toxins.

Pertussis toxin, a protein composed of five different subunits, is responsible for the pathogenicity of Bordetella pertussis and is the main component of a new vaccine against whooping cough. The genes coding for the five subunits, recently cloned and sequenced, are organized as an operon. We approached the problem of expression of the five genes in Escherichia coli and, although we obtained high levels of transcription of the native pertussis toxin genes, the amount of proteins produced was very low or undetectable. To obtain suitable expression of each of the five subunits, we fused their genes to the gene coding for the DNA polymerase of MS2 in the expression vector pEx31. A total of 5 to 30 mg of purified fusion proteins could be obtained from 1 liter of culture. The purified fusion proteins were used to immunize rabbits to obtain sera against each of the five subunits. These sera, although able to recognize the toxin in an enzyme-linked immunosorbent assay and the corresponding subunits in Western blots, were not able to protect CHO cells from the action of pertussis toxin. Mice immunized with the five subunits were not protected from an intracerebral challenge with B. pertussis. Subunits S2 and S3, which are 67% homologous, were shown to cross-react immunologically. The fused subunit S1 was able to ADP-ribosylate transducin as efficiently as the native pertussis toxin.

Immunization with chemically detoxified pertussis toxin can prevent severe whooping cough with an efficacy similar to that of the cellular pertussis vaccine, which normally gives unwanted side effects. To avoid the reversion to toxicity and the loss of immunogenicity that may follow chemical treatment of pertussis toxin, inactive toxins were constructed by genetic manipulation. A number of genetically engineered alleles of the pertussis toxin genes, constructed by replacing either one or two key amino acids within the enzymatically active S1 subunit, were introduced into the chromosome of strains of Bordetella pertussis, B. parapertussis, and B. bronchiseptica. These strains produce mutant pertussis toxin molecules that are nontoxic and immunogenic and that protect mice from the intracerebral challenge with virulent Bordetella pertussis. Such molecules are ideal for the development of new and safer vaccines against whooping cough.

Different aspects of lethal infection of infant mice with Bordetella pertussis were examined. Mutants deficient in vir-regulated genes were tested for the ability to cause a lethal infection in the infant mouse model. Adenylate cyclase toxin-hemolysin and pertussis toxin were required to cause a lethal infection at low doses. Mixed infection caused by challenging the mice with an equal number of pertussis toxin and adenylate cyclase toxin-hemolysin mutants at a dose at which neither alone was lethal was also unable to cause a lethal infection. Production of the filamentous hemagglutinin and the dermonecrotic toxin was not required to cause a lethal infection. Nine other mutants in vir-regulated genes whose phenotypes have yet to be determined were also tested. Only two of these mutants were impaired in the ability to cause a lethal infection. Expression of fimbriae does not appear to affect the dose required to cause a lethal infection; however, fimbrial expression was correlated with the later stages of a nonlethal, persistent infection. Growth of the bacteria in MgSO4, a condition which reversibly suppresses expression of the genes required for virulence, did not alter the ability of the bacteria to cause a lethal infection. Auxotrophic mutants deficient in leucine biosynthesis were as virulent as the parental strain; however, mutants deficient in methionine biosynthesis were less virulent. A B. parapertussis strain was much less effective in promoting a lethal infection than any of the wild-type B. pertussis strains examined. A persistent infection in the lungs was observed for weeks after challenge for mice given a sublethal dose of B. pertussis, and transmission from infected infants to the mother was never observed.

Pertussis toxin, the major virulence factor of Bordetella pertussis, is not produced by the closely related species Bordetella parapertussis and Bordetella bronchiseptica. It is shown here that these two species possess but do not express the complete toxin operon. Nucleotide sequencing of an EcoRI fragment of 5 kilobases comprising the regions homologous to the pertussis toxin genes shows that in this region, B. parapertussis and B. bronchiseptica are 98.5% and 96% homologous, respectively, to B. pertussis. The changes (mostly base pair substitutions) in many cases are identical in B. parapertussis and B. bronchiseptica, suggesting that these two species derive from a common ancestor. Many of the mutations common to B. parapertussis and B. bronchiseptica involve the promoter region, which becomes very inefficient. The S1 subunits of both species, when expressed in Escherichia coli, have the same ADP-ribosylating activity as the S1 subunit from B. pertussis, indicating that the mutations in the S1 gene described here do not affect its function.

Pertussis toxin (PT), the major virulence factor of Bordetella pertussis, is composed of five different subunits whose genes are organized as an operon. We report the mapping of the promoter region of the PT operon and show that this promoter is only weakly active in Escherichia coli. Bordetella parapertussis and Bordetella bronchiseptica, which do not produce any PT, are shown to have a weaker promoter sequence for this operon and not to produce any detectable PT mRNA. We show that transcription of the PT operon in B. pertussis was constant throughout until the late stationary phase, when transcription significantly decreased. Analysis of the transposon Tn5 mutant BP347 showed that the product of the vir locus was required for transcription of the PT operon. Characterization of the Tn5 mutant BP356 showed that subunit S3 was required for the release of PT into the extracellular medium.

The S1 subunit of pertussis toxin possesses two regions (homology boxes), each spanning 8 residues, that are nearly identical in sequence to similarly located regions in the enzymatically active A fragments of two other ADP-ribosylating toxins: cholera toxin and Escherichia coli heat-labile toxin. This observation suggests a functional role for one or both of these regions in enzymatic activity. We have examined the role of one of these regions, located near the amino terminus of the S1 subunit, by using a high-level recombinant expression system and progressive truncation of the gene sequence encoding the amino terminus of the molecule. A series of six truncated, recombinant proteins were produced at high levels in E. coli and examined for their enzymatic and antigenic properties. The three molecules that lacked most or all of the homology box delimited by amino acid residues 8 and 15 lacked detectable enzymatic activity. All of the three molecules in which the box was retained exhibited detectable activity. Only those recombinant molecules that possessed the homology box reacted with a neutralizing and passively protective monoclonal anti-S1 antibody. These findings identify the region of homology located near the amino terminus of S1 as an apparent enzymatic subsite and a potentially important antigenic determinant.

The etiologic agent of whooping cough, Bordetella pertussis, synthesizes several biochemically and biologically active factors believed to contribute to the disease process. They include filamentous hemagglutinin (FHA), pertussis toxin (PTX), adenylate cyclase toxin (AC), and hemolysin (HLY). We have used a genetic approach to evaluate the contribution of these factors to the virulence of B. pertussis. A series of B. pertussis mutants prepared by transposon Tn5 insertion mutagenesis were characterized and determined to be specifically deficient in the elaboration of these factors. The wild type strain and the mutants were tested for their ability to produce a lethal infection in the classical infant mouse intranasal infection model. The dose of organisms required to cause lethal infection in 50% of the animals (LD50) for the wild type strain was 2 X 10(3) bacteria. A mutant deficient in production of the proposed adherence factor, FHA, was equivalent in virulence to the wild type strain (BP353, LD50 = 8 X 10(3)). In contrast, mutants deficient in production of PTX alone (BP357, LD50 = 5 X 10(6)), AC and HLY (BP348, LD50 greater than 3 X 10(7)), or all known toxins and virulence factors (BP347, LD50 greater than 3 X 10(7)) were severely impaired in their ability to cause a lethal infection. Our data indicate the hemolysin and adenylate cyclase as well as pertussis toxin are important virulence factors and provide the basis for a working model of the molecular basis of the disease caused by Bordetella pertussis.

Bordetella pertussis produces a calmodulin-sensitive adenylate cyclase (AC) which is an essential virulence factor in mammalian pertussis. Here we report the purification and characterization of the toxic form of the enzyme, which penetrates eukaryotic cells and generates high levels of intracellular cAMP. This form was purified from an extract of B.pertussis strain carrying a recombinant plasmid which over-produced both enzymatic and toxic activities of the enzyme. Western blot analysis of the extract using anti-B.pertussis AC antibodies detected only one protein of 200 kd. However, gel filtration of the extract resolved two peaks of enzymatic activity. The first peak of aggregated material contained greater than 70% of the total enzymatic activity, and the second peak contained the majority of the toxic activity. Purification of the enzyme from both peaks yielded proteins of 200 kd, with similar biochemical and immunological properties. Yet only the enzyme purified from the second peak could penetrate human lymphocyte and catalyse the formation of intracellular cAMP. B.pertussis AC gene expressed in Escherichia coli produced a calmodulin-dependent enzyme of 200 kd, which lacked lymphocyte penetration capacity. It is proposed that a post-translational modification that occurs in B.pertussis but not in E.coli confers upon the 200 kd protein of B.pertussis AC the toxic properties.

By introducing a series of six different substitutions at and around position 9, we investigated the structural requirements of the amino-terminal region of the S1 subunit of pertussis toxin for both enzyme activity and immunoreactivity. All mutant S1 analogs with a substitution at this location exhibited severely decreased ADP-ribosyltransferase activity (range, 400- to 2,500-fold). In contrast, alteration of arginine 58 had considerably less effect. The reactivity of the mutant molecules with monoclonal antibody 1B7 varied with the nature of the substitution. These findings indicate an absolute requirement for the presence of an arginine residue at position 9 for the maintenance of efficient ADP-ribosyltransferase activity and illustrate the specific participation of vicinal residues in the formation of the protective epitope.

Previous analysis of the pertussis toxin (PT) promoter has shown that expression of PT requires a trans-activating factor encoded by the vir locus and a 170-base-pair DNA sequence upstream from the transcription start site containing a 21-base-pair direct repeat sequence crucial trans-activation (R. Gross and R. Rappuoli, Proc. Natl. Acad. Sci. USA 85:3913-3917, 1988). In this paper we extend the analysis to the modulative response to environmental stimuli. We show that modulation acts at the transcriptional level and occurs only in phase I bacteria. Modulation also requires a functional vir locus and the same promoter region of 170 base pairs. We show that, in addition to the previously identified direct repeat, even the sequences downstream from position -117 are required for trans-activation and modulation and that the deletion of four cytosine residues at position -31 causes the inactivation of the promoter. The kinetics of the change in transcription show that the PT promoter can be shut off very rapidly by adding 50 mM MgSO4 to the medium, whereas resumption of transcription after removal of the modulative agents from the medium is slow.

Pertussis toxin (PT) is an important virulence determinant of Bordetella pertussis and one of the major protective antigens against whooping cough. The genes coding for PT have recently been cloned, but attempts to express them in Escherichia coli have been unsuccessful. We therefore explored the possibility of expressing these genes in Bacilius subtilis for which efficient vectors are available. The lack of endotoxin in the Gram-positive Bacillus might be an additional advantage for the production of a vaccine component. A DNA fragment coding for S1, one of the subunits of pertussis toxin, was inserted into an alpha-amylase secretion vector and the recombinant plasmid was introduced into B. subtilis. This resulted in high expression of S1, most of which was secreted and therefore found in the culture supernatant. This supernatant had ADP-ribosylating activity similar to that of PT. Western blot with antiserum to B. pertussis holotoxin showed several proteins ranging in size from 28 kDa to 20 kDa reacting in specific manner. About 10% of the protein recognized by the antiserum was of the size expected for native-size S1. The total amount of S1 proteins (full size and truncated) in the culture supernatant was about 100 mg/l. S1 protein made in B. subtilis was partially purified using chromatography with P-cellulose and Blue Sepharose. This preparation was used to immunize rabbits; the immune serum thus obtained recognized subunit S1 of native pertussis toxin.

Pertussis toxin (PT) is the major virulence factor of Bordetella pertussis. The cloning and nucleotide sequencing of the PT genes from B. pertussis, Bordetella parapertussis and Bordetella bronchiseptica has elucidated the evolution of the Bordetella species and allowed considerable advances towards the understanding of their gene expression and the development of safer vaccines against pertussis.

The genes encoding the S1 and S2 subunits of pertussis toxin were expressed in Escherichia coli under lac operon transcription and translation control with pUC8 and pUC18 as the expression vectors. Various versions of the subunits were detected with anti-S1 or anti-S2 monoclonal antibodies. Recombinant S1, but not S2, subunit contained the enzymatic NAD-glycohydrolase and NAD:Gi ADP-ribosyltransferase activities. Both activities were also expressed by a truncated version of the S1 subunit in which the 48 carboxy-terminal amino acid residues, including a predicted Rossman structure and one of the two cysteines, had been deleted. The epitope for an anti-S2 monoclonal antibody was localized to the N-terminal 40-amino-acid region of the S2 subunit. Both the S1 and S2 subunits expressed in E. coli reacted with human hyperimmune serum. The full length and the truncated recombinant S1 subunit also reacted in Western blots with a neutralizing and protective monoclonal anti-S1 antibody. The different versions of S1 and S2 subunits expressed in E. coli are useful for mapping active sites, epitopes, and regions that interact with receptors or the other subunits in the holotoxin. These recombinant subunits will also facilitate the development of a safer, new-generation vaccine against whooping cough.

Bordetella pertussis is the causative agent of whooping cough. The cellular pertussis vaccine introduced in the forties is highly effective and is widely used, but its reactogenicity has led to public concern regarding its safety. The attempts to reduce the side effects associated with pertussis immunization have led to the preparation of acellular B. pertussis products: one composed of detoxified pertussis toxin (PT) and filamentous haemagglutinin (FHA), another one composed only of detoxified PT and a third vaccine composed of detoxified PT, FHA and serotypes 1, 3 of fimbriae. In our laboratories we have approached the study of pertussis toxin, the molecule present in all the proposed acellular pertussis vaccines and one of the main virulence factors of B. Pertussis, with the aim of producing new acellular pertussis vaccines by using recombinant DNA techniques.

A molecular probe for the PT-operon of B. pertussis hybridized with 4.7 Kb EcoRI-fragments of chromosomal DNAs of B. pertussis strain 475 phase I, phase IV, B. parapertussis strains 504 and 17903, B. bronchiseptica strain 214, B. parapertussis strain 17903-convertant of B. pertussis phage 134 but not with phage 134 DNA under stringent conditions of DNA-DNA hybridization. This fact indicates the presence of PT-genes in all Bordetella species. Since there is no production of PT in B. parapertussis and B. bronchiseptica, a presence of regulatory mutations in the PT-operon or absence of the functionally active vir-gene product in these species is suggested.

This paper describes a new method for site-directed mutagenesis which allows mutations by deletion, insertion or substitution of large fragments of DNA with more than 50% efficiency and does not require subcloning in a single-stranded (ss) DNA vehicle. The site of mutagenesis is removed from a linearized plasmid DNA by BAL 31 digestion, ss DNA regions are generated by limited exonuclease treatment and the mutated target site is reconstituted by annealing of the plasmid DNA to a 35-70 nucleotide long mutated ss oligodeoxynucleotide containing the desired mutation. The circularized plasmid is finally used to transform directly Escherichia coli competent cells.

The gene encoding a catalytically active deletion peptide, the C180 peptide, of the S-1 subunit of pertussis toxin was engineered to facilitate mutagenesis at the Trp-26 (wild-type) coding sequence. A synthetic double-stranded oligonucleotide was inserted into the C180 gene such that all possible codons would be introduced into position 26. Seven individual mutants of the C180 peptide which possessed amino acid substitutions at residue 26 (collectively termed C180W26n peptides) were purified from periplasmic extracts of Escherichia coli. Each C180W26n peptide was present as a single major peptide that had an apparent molecular mass of between 20.9 and 24.5 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and each showed similar immunoreactivity relative to the C180 peptide. The C180W26n peptides demonstrated marked reduction of both ADP-ribosyltransferase and NAD glycohydrolase activities at 25 nM and 10 microM NAD, respectively. Kinetic analysis of the two most active mutants, C180W26F and C180W26Y, revealed that the major perturbation of NAD glycohydrolase activity was due to an increase (approximately 20-fold) in the Km for NAD between these mutants and the C180 peptide.

Pertussis toxin is one of the major virulence determinants produced by Bordetella pertussis. The DNA encoding the structural genes for pertussis toxin was cloned in Escherichia coli, and pertussis toxin subunit S4 was expressed under the control of the tac promoter. Mutations were introduced into the cloned toxin genes, and a conjugative shuttle vector system was devised for delivering the mutations from E. coli back into B. pertussis. The mutations were introduced by allelic exchange into the chromosome of B. pertussis resulting in a series of B. pertussis strains which were isogenic except at the loci encoding the structural genes for pertussis toxin. These B. pertussis strains were utilized to study the biogenesis of pertussis toxin. Polar mutations in the S1 gene led to a lack of detectable S2 or S4 subunits in whole-cell lysates, suggesting a polycistronic arrangement for these genes. Mutations in the S5 subunit gene resulted in a truncated S1 subunit, while mutations in the S4 gene resulted in a lack of detectable S2 subunit, suggesting that physical relationships among the toxin subunits are directly reflected in the stable biogenesis of the subunits.

The amino acid sequence and composition of alpha-subunits of signal transducing G proteins of the same kind appear to vary by no more than 2% from species to species. Here we isolated a human liver cDNA using an oligonucleotide complementary to the sequences encoding the pertussis toxin (PTX) ADP-ribosylation site of the alpha-subunit of the rat brain G protein called Gi. Its open reading frame characterizes it as an alpha i-type cDNA--as opposed to alpha o-type--but predicts an amino acid composition that differs by 7% and 14%, respectively, from two other human alpha i-type molecules. Together with human brain alpha i (type-1) and human monocyte alpha i (type-2), the new human liver alpha i cDNA (type-3) forms parts of a family of alpha i molecules. Type-3 alpha i cDNA hybridizes to a approximately 3.6 kilobase long mRNA and type-2 alpha i cDNA hybridizes to an mRNA species of approximately 2.7 kilobases. This indicates that the human genome has at least three non-allellic genes encoding non-alpha o-type PTX substrates and provides structural evidence for the hypothesis that distinct effector systems are regulated by similar but nevertheless distinct PTX substrates.

The site of interaction of NAD with the isolated S1 subunit of pertussis toxin was investigated by photoaffinity labelling. When S1 was irradiated at 254 nm in the presence of [carbonyl-14C]- or [adenine-14C]NAD, the uptake of radioactivity was equivalent to 0.75 and 0.1 mol/mol respectively, while the NAD glycohydrolase activity was abolished. Inactivation was thus accompanied by crosslinking of the nicotinamide portion of NAD to the protein. Sequence determination of purified radioactive peptides indicated that Glu-129 was a major site of labelling. This residue is therefore closely associated with either NAD binding or hydrolysis.

Pertussis toxin is produced by the causative agent of whooping cough, Bordetella pertussis, and is an adenosine diphosphate (ADP)-ribosyltransferase capable of covalently modifying and thereby inactivating many eukaryotic G proteins involved in cellular metabolism. The toxin is a principal determinant of virulence in whooping cough and is a primary candidate for an acellular pertussis vaccine, yet it is unclear whether the ADP-ribosyltransferase activity is required for both pathogenic and immunoprotective activities. A B. pertussis strain that produced an assembled pertussis holotoxin with only 1 percent of the ADP-ribosyltransferase activity of the native toxin was constructed and was found to be deficient in pathogenic activities associated with B. pertussis including induction of leukocytosis, potentiation of anaphylaxis, and stimulation of histamine sensitivity. Moreover, this mutant strain failed to function as an adjuvant and was less effective in protecting mice from intracerebral challenge infection. These data suggest that the ADP-ribosyltransferase activity is necessary for both pathogenicity and optimum immunoprotection. These findings bear directly on the design of a nontoxic pertussis vaccine.

The S-1 subunit of pertussis toxin was expressed as a fusion protein in a strain of Escherichia coli deficient in protein degradation. The fusion protein reacted with anti-pertussis toxin antibody, and, like authentic pertussis toxin, it ADP-ribosylated a 41,000-molecular-weight membrane protein from human erythrocytes.

Pertussis toxin (PTX) is a major virulence factor in whooping cough and can elicit protective antibodies. Amino acid residues 8 to 15 of PTX subunit S1 are important for the adenosine diphosphate-ribosyltransferase activity associated with the pathobiological effects of PTX. Furthermore, this region contains at least a portion of an epitope that elicits both toxin-neutralizing and protective antibody responses in mice. The gene encoding the S1 subunit was subjected to site-specific mutagenesis in this critical region. A mutant containing a single amino acid substitution (Arg9----Lys) had reduced enzymatic activity (approximately 0.02% of control) while retaining the protective epitope. This analog S1 molecule may provide the basis for a genetically detoxified PTX with potential for use as a component of an acellular vaccine against whooping cough.

Multiple strains of Bordetella parapertussis and B. bronchiseptica were examined for the presence of nucleotide sequences which hybridized with a cloned 4.5-kilobase (kb) fragment of B. pertussis DNA containing the genes responsible for pertussis toxin expression. All six B. parapertussis strains tested had nucleic acid sequences that hybridized with the cloned 4.5-kb fragment in Southern blot analyses. When the B. parapertussis DNA was digested with restriction endonuclease PstI, the pattern of hybridization was identical to that obtained with B. pertussis. Only five of the seven B. bronchiseptica strains tested had sequences that hybridized with the 4.5-kb fragment. Three of these B. bronchiseptica strains had a hybridization pattern identical to B. pertussis upon PstI digestion and Southern blot analyses. Two B. bronchiseptica strains were shown to lack a PstI cleavage site downstream from the region analogous to that coding for the pertussis toxin structural genes. Monoclonal antibody analyses were unable to detect pertussis toxin subunits S1 and S2 in Western blots with cellular material or culture supernatant from several B. bronchiseptica and B. parapertussis strains that possessed the DNA homologies. In addition, preliminary Northern hybridizations with RNA isolated from B. bronchiseptica and B. parapertussis strains suggested that the homologous regions were not transcribed. The data show that the gene coding for the toxic component of B. pertussis is common in other Bordetella species, though the gene probably is not expressed.

We have cloned a 4.5 kb EcoRI/BamHI DNA fragment from Bordetella pertussis which contains at least two genes responsible for expression of pertussis toxin. The S4 subunit of the toxin was isolated by high pressure liquid chromatography and the NH2-terminal amino acid sequence determined. Using a mixed synthetic oligonucleotide probe designed by reverse translation of a portion of the protein sequence, a cloned DNA fragment was identified which contains the coding information for at least the S4 structural subunit of the toxin. Sequence analyses indicate that the mature protein is derived by proteolytic cleavage of a precursor molecule. Southern blot analyses of Tn5-induced B. pertussis toxin-deficient mutants show that the Tn5 DNA is inserted 1.3 kb downstream from the S4 subunit gene. This second gene could code for another subunit required for assembly of the mature toxin or a non-structural transport protein, possibly in the same polycistronic operon. The molecular cloning of pertussis toxin genes provides the basis for development of a safer recombinant "new generation" vaccine for whooping cough.

The cytotoxicity of pertussis toxin, a multisubunit exotoxin produced by Bordetella pertussis, is believed to be due to the ADP-ribosyltransferase activity of the S1 subunit. We have previously described the recombinant expression of each of the five individual pertussis toxin subunits in Escherichia coli and the production of an enzymatically deficient form of the S1 subunit by site-directed mutagenesis. We now report the in vitro assembly of holotoxin from native pertussis toxin B oligomer and recombinant S1 subunits, the latter purified and refolded from insoluble inclusion bodies. Holotoxin assembled with recombinant S1 of authentic amino acid sequence was indistinguishable from native pertussis toxin in its electrophoretic migration and ability to elicit a cytopathic response in cultured Chinese hamster ovary cells; in contrast, holotoxin assembled with the genetically deactivated analog of recombinant S1 displayed greatly diminished cytopathicity. These results verify that the in vitro cytopathic effects of pertussis toxin are the result of the enzymatic activity of the S1 subunit and illustrate the potential for constructing complex quaternary protein structures in vitro from insoluble, unfolded polypeptides derived from expression in recombinant systems.

Monoclonal antibodies with specificity for pertussis toxin subunits S1, S2, and S4 were used in Western blots to show that the subunits were not secreted into culture medium from Tn5 insertion mutants. The mutants are deficient in toxin biological activities due to an insertion in the S3 subunit structural gene. The Western blots demonstrated that each of the respective subunits was exported in a wild-type strain. Anti-S1 and anti-S2 monoclonal antibodies were capable of detecting subunits in solubilized whole-cell material from a wild-type strain and from the Tn5 mutants lacking only in biologically active toxin (Tox-). Another Tn5 insertion mutant, lacking all known B. pertussis virulence factors (Vir-), did not produce any of the subunits either in whole cellular extracts or in culture supernatants. The data demonstrate that Tn5 Tox- insertion mutants, though defective in toxin activity, synthesize some toxin subunits. The presence of the S3 subunit is most likely a necessity for transport of the toxin from cells. Alternatively, a nonstructural gene coding for a protein involved in transport of the toxin across the membrane may be affected by the Tn5 mutation.

Although the genus Bordetella contains several closely related species, pertussis toxin (PT) is produced only by phase I Bordetella pertussis. In this work we have studied the regulation of expression of the PT operon and investigated why PT is produced by phase I and not by phase III B. pertussis despite the presence of the PT genes. We have constructed a vector for Bordetella species that contains the PT promoter fused to the coding region of the chloramphenicol acetyltransferase (CAT) gene, and we have used it to identify the regulatory elements involved in the transcription of the PT operon. Efficient transcription of these genes requires at least two features: (i) the 170-base-pair DNA sequence upstream from the start site of transcription and (ii) a trans-activating factor encoded by the vir locus. Bordetella parapertussis and Bordetella bronchiseptica, although endowed with a functional trans-activating system, do not produce PT because of mutations within their PT promoter regions. In contrast, phase III Bordetella species do not show any trans activity.

Pertussis toxin is a principal determinant of virulence produced by Bordetella pertussis in the disease whooping cough and is the primary toxinogenic component of the pertussis vaccine. This toxin is not produced by the closely related species Bordetella parapertussis. Toxinogenic strains of B. parapertussis were constructed by the conjugative introduction of cloned pertussis toxin genes. These and analogous toxinogenic and nontoxinogenic strains of B. pertussis were assayed for their toxicity and reactogenic activities. Expression of active toxin by either Bordetella sp. correlated with the induction of leukocytosis, anaphylaxis, and histamine sensitivity.

Pertussis toxin is produced only by strains of Bordetella pertussis. Cloned genes encoding pertussis toxin from B. pertussis were transferred into Bordetella bronchiseptica and Bordetella parapertussis by conjugation. These transconjugants expressed pertussis toxin at levels comparable to those expressed by B. pertussis. The toxin made by these strains was biologically active in the Chinese hamster cell clumping assay, contained all five subunits, and was mostly periplasmic. Toxin expression appeared to be modulated in the same way as are the vir-regulated genes of B. pertussis. Introduction of these plasmids into B. pertussis failed to produce hypertoxigenic strains. Instead, these transconjugants underwent plasmid loss, gene deletions, or conversion to the avirulent phase.

The toxicity of pertussis toxin is mediated by the ADP-ribosyltransferase activity of subunit S1. To understand the structure-function relationship of subunit S1 and guide the construction of nontoxic molecules suitable for vaccines, we constructed and expressed in Escherichia coli a series of amino-terminal and carboxyl-terminal deletion mutants as well as a number of molecules containing amino acid substitutions. The shortest peptide still retaining enzymatic activity contains amino acids 2-179. Within this region we identified three mutants in which amino acid substitutions abolish the enzymatic activity. Mutation of amino acids 8 and 9 or 50 and 53, located within the region of the S1 subunit of pertussis toxin homologous to cholera toxin, causes loss of enzymatic activity. Outside this homology region, substitution of Glu-129 with glycine or aspartic acid also eliminates the enzymatic activity of the S1 subunit. In this respect, Glu-129 resembles the glutamic acid that is crucial for the catalytic activity of diphtheria and Pseudomonas toxins. Once introduced into the Bordetella pertussis chromosome, the above mutations should lead to the synthesis of nontoxic pertussis toxin molecules suitable for vaccine production.

The enzymatic ADP-ribosyltransferase activity associated with the S1 subunit of pertussis toxin is considered to be responsible for its biological effects. Although pertussis toxin has no significant homology to other ADP-ribosylating toxins such as diphtheria toxin and Pseudomonas aeruginosa exotoxin A, the results presented in this paper show that, as for diphtheria toxin and exotoxin A, tryptophan and glutamic acid residues are essential for the enzymatic activities of pertussis toxin. Moreover, a structural motif can be identified around the critical glutamic acid residue. Chemical modification or site-directed deletion or replacement of Trp-26 abolishes ADP-ribosyltransferase and the associated NAD glycohydrolase activities. Both enzymatic activities are also abolished when Glu-129 is deleted or replaced by aspartic acid. Mutations at the Glu-106 position do not significantly reduce the enzymatic activities of the S1 subunit. The mutations do not affect the ability of the different S1 forms to be recognized by a variety of monoclonal antibodies, including neutralizing antibodies. Pertussis toxin containing a deletion or replacement of Trp-26, Glu-129, or both in the S1 subunit should thus be devoid of toxic activities without losing its reactivity with protective antibodies and, therefore, could be safely included in new generation vaccines against whooping cough.

The adenylate cyclase toxin of the prokaryote Bordetella pertussis is stimulated by the eukaryotic regulatory protein, calmodulin. A general strategy, using the adenylate-cyclase-calmodulin interaction as a tool, has permitted cloning and expression of the toxin in Escherichia coli in the absence of any B. pertussis trans-activating factor. We show that the protein is synthesized in a large precursor form composed of 1706 amino acids. The calmodulin-stimulated catalytic activity resides in the amino-terminal 450 amino acids of the adenylate cyclase. The enzyme expressed in E. coli is recognized in Western blots by antibodies directed against purified B. pertussis adenylate cyclase, and its activity is inhibited by these antibodies.

The nucleotide sequence of the pertussis toxin operon of Bordetella pertussis, Bordetella parapertussis and Bordetella bronchiseptica, has shown that the last two species contain many common mutations and are likely to derive from a common ancestor (Aricò and Rappuoli, 1987). To elucidate further the evolutionary relationships between the Bordetella species, we have cloned and sequenced the promoter region and the gene coding for the S1 subunit of pertussis toxin from additional B. pertussis strains, such as the type strain BP 18323 and two recent clinical isolates, namely strain BP 13456 from Sweden and strain BP SA1 from Italy. While the strains BP SA1 and BP 13456 are shown to differ from the published B. pertussis sequences by only one base pair, the type strain BP 18323 contains a total of 11 base-pair substitutions. Remarkably, 9 of the 11 substitutions found in BP 18323 are also common to B. parapertussis and B. bronchiseptica, strongly suggesting that this strain derives from the same ancestor as B. parapertussis and B. bronchiseptica. Computer analysis of the sequence data allows the construction of an evolutionary 'tree' showing that the B. pertussis strains are very homogeneous and significantly distant from B. parapertussis and B. bronchiseptica. Therefore the proposed conversion from B. parapertussis to B. pertussis appears highly improbable.

IS481v1 and IS481v2 are two copies of a Bordetella pertussis insertion sequence element. We have shown that IS481v1 is located within 3 kbp of the start of the adenylate cyclase gene whilst IS481v2 is immediately adjacent to the end of the agglutinogen 2 gene and provides the stop codon for that gene. In addition, IS481v1 and IS481v2 were present at these two specific sites in nine strains of B. pertussis, including two Phase IV strains which expressed neither adenylate cyclase nor agglutinogen 2 and three Phase I strains which did not express agglutinogen 2. The loss of expression in these strains is not the result of DNA rearrangements at the sites of IS481v1 or IS481v2.

The ADP-ribosyltransferase activity of pertussis toxin resides within the S-1 subunit of the toxin. Deletion mapping of a recombinant S-1 subunit produced in Escherichia coli showed that amino acids 2 through 180 are required for ADP-ribosylation of Gi protein. Mutants of the S-1 subunit which lacked either amino acids 2 through 22 or amino acids 153 through 180 failed to express enzyme activity, implicating a functional or structural role for these residues in catalysis. The catalytic carboxy-terminal S-1 deletion, C-180, was found to be more soluble than the recombinant S-1 subunit, making it a useful construct for future structure-function studies on enzyme catalysis. Four independent single-amino-acid substitutions which decreased ADP-ribosyltransferase activity were constructed in the recombinant S-1 subunit. Substitution of Asp-11 by Ser, Arg-13 by Leu, or Trp-26 by Ile decreased enzyme activity to below detectable levels (less than 1.0% of that of the recombinant S-1 subunit). The Glu-139-to-Ser substitution reduced ADP-ribosyltransferase activity to 15% of that of the recombinant S-1 subunit. Both the oxidized and reduced forms of the recombinant S-1 subunit and recombinant S-1 subunits containing single-amino-acid substitutions were degraded through identical immunoreactive tryptic peptides, suggesting that the conformations of the mutants are similar to that of the recombinant S-1 subunit. Identification of noncatalytic forms of the S-1 subunit of pertussis toxin which have conserved protein structure is an initial step in the generation of a recombinant noncatalytic form of pertussis toxin which may be tested as a candidate for an acellular vaccine against Bordetella pertussis.

We isolated and characterized human genomic clones encompassing the gene for the alpha-subunit of Gi2, a GTP-binding signal transduction protein abundantly expressed in myeloid cells. The gene is divided into 9 exons and spans 23.5 kb. Exons 2, 6 and 7 encode putative guanine nucleotide-binding domains that are highly conserved among GTP-binding proteins. A polyadenylation signal located within exon 9 predicts an mRNA size (approximately 2.3 kb) several hundred bases longer than that of published cDNAs, and consistent with the size seen on RNA blot hybridization. Primer extension and S1 nuclease analysis determined a major and several minor transcriptional start sites. The first exon and 5' flanking region are highly G + C rich, contain several GC boxes (SP1 transcription factor binding sites), a CAAT box, and lack a TATA box. The presumptive promoter region is thus similar to that of ras and other widely expressed genes.