Hemolytic, but Not Cell-invasive Activity, of Adenylate Cyclase Toxin Is Selectively Affected by Differential Fatty-acylation in Escherichia coli

Abstract

Adenylate cyclase toxin from Bordetella pertussis requires posttranslational acylation of lysine 983 for the ability to deliver its catalytic domain to the target cell interior and produce cyclic adenosine monophosphate (cell-invasive activity) and to form transmembrane channels (hemolytic activity). When the toxin is expressed in Escherichia coli, it has reduced hemolytic activity, but comparable cell-invasive activity to that of adenylate cyclase toxin from B. pertussis. In contrast to the native protein from B. pertussis, which is exclusively palmitoylated, recombinant toxin from E. coli is acylated at lysine 983 with about 87% palmitoylated and the remainder myristoylated. Furthermore, the recombinant toxin contains an additional palmitoylation on approximately two-thirds of the lysines at position 860. These observations suggest that the site and nature of posttranslational fatty-acylation can be dictated by the bacterial host used for expression and can have a significant, but selective, effect on protein function. Adenylate cyclase toxin from Bordetella pertussis requires posttranslational acylation of lysine 983 for the ability to deliver its catalytic domain to the target cell interior and produce cyclic adenosine monophosphate (cell-invasive activity) and to form transmembrane channels (hemolytic activity). When the toxin is expressed in Escherichia coli, it has reduced hemolytic activity, but comparable cell-invasive activity to that of adenylate cyclase toxin from B. pertussis. In contrast to the native protein from B. pertussis, which is exclusively palmitoylated, recombinant toxin from E. coli is acylated at lysine 983 with about 87% palmitoylated and the remainder myristoylated. Furthermore, the recombinant toxin contains an additional palmitoylation on approximately two-thirds of the lysines at position 860. These observations suggest that the site and nature of posttranslational fatty-acylation can be dictated by the bacterial host used for expression and can have a significant, but selective, effect on protein function. INTRODUCTIONAdenylate cyclase (AC) 1The abbreviations used are: ACadenylate cyclaseCADcollision-activated dissociationFmocNα-9-fluorenylmethoxycarbonylHPLChigh performance liquid chromatographyMALDI-TOF MSmatrix-assisted laser desorption/ionization time-of-flight mass spectrometryMSmass spectrometryRSDrelative standard deviationRTXrepeat in toxin. toxin is a bifunctional 177-kDa bacterial protein that contains a calmodulin-activated adenylate cyclase catalytic domain and a pore-forming domain. It exhibits a cell-invasive activity by delivering its catalytic domain to the interior of target cells to elicit supraphysiologic cAMP accumulation(1Confer D.L. Eaton J.W. Science. 1982; 217: 948-950Crossref PubMed Scopus (287) Google Scholar, 2Hanski E. Coote J.G. Alouf J.E. Freer J.H. Sourcebook of Bacterial Toxins. Academic Press, New York1991: 349-366Google Scholar, 3Mock M. Ullmann A. Trends Microbiol. 1993; 1: 187-192Abstract Full Text PDF PubMed Scopus (43) Google Scholar, 4Hewlett E.L. Maloney N.J. Iglewski B. Moss J. Tu A.T. Vaughan M. Handbook of Natural Toxins, Volume 8: Microbial Toxins. Marcel Dekker, New York1994: 425-439Google Scholar). The pore-forming capacity of the molecule is responsible for hemolytic activity against sheep erythrocytes and is entirely independent of the presence and activities of the catalytic domain(5Bellalou J. Sakamoto H. Ladant D. Geoffroy C. Ullmann A. Infect. Immun. 1990; 58: 3242-3247Crossref PubMed Google Scholar, 6Ehrmann I. Gray M. Gordon V. Gray L. Hewlett E.L. FEBS Lett. 1991; 278: 79-83Crossref PubMed Scopus (76) Google Scholar, 7Rogel A. Meller R. Hanski E. J. Biol. Chem. 1991; 266: 3154-3161Abstract Full Text PDF PubMed Google Scholar, 8Szabo G. Gray M.C. Hewlett E.L. J. Biol. Chem. 1994; 269: 22496-22499Abstract Full Text PDF PubMed Google Scholar, 9Benz R. Maier E. Ladant D. Ullmann A. Sebo P. J. Biochem. (Tokyo). 1994; 269: 22239-27231Google Scholar).The AC toxin gene (cyaA) has been cloned and sequenced by Glaser et al.(10Glaser P. Ladant D. Sezer O. Pichot F. Ullmann A. Danchin A. Mol. Microbiol. 1988; 2: 19-30Crossref PubMed Scopus (206) Google Scholar, 11Glaser P. Sakamoto H. Bellalou J. Ullmann A. Danchin A. EMBO J. 1988; 7: 3997-4004Crossref PubMed Scopus (277) Google Scholar) and the AC toxin locus was noted to have homologies with the Escherichia coli hemolysin operon and that of other RTX toxins. These toxins are characterized by a set of glycine- and aspartate-rich nonapeptide repeats(12Welch R. Mol. Microbiol. 1991; 5: 521-528Crossref PubMed Scopus (437) Google Scholar, 13Coote J.G. FEMS Microbiol. Rev. 1992; 88: 137-162Crossref Google Scholar). The formation of biologically active AC toxin requires an accessory protein expressed from an upstream gene, cyaC(14Barry E.M. Weiss A.A. Ehrmann I.E. Gray M.C. Hewlett E.L. Goodwin M.S. J. Bacteriol. 1991; 173: 720-726Crossref PubMed Google Scholar). That protein, CyaC, appears to be involved in posttranslational activation of AC toxin. Recently, tandem mass spectrometry was used to determine that the modification on native AC toxin from Bordetella pertussis (Bp-CyaA) consists of palmitoylation on the ɛ-amino group of Lys983 (15Hackett M. Guo L. Shabanowitz J. Hunt D. Hewlett E. Science. 1994; 266: 433-435Crossref PubMed Scopus (193) Google Scholar). In vitro random chemical acylation has also been shown to confer limited cell-invasive and hemolytic activities on CyaA(16Heveker N. Bonnaffe D. Ullmann A. J. Biol. Chem. 1994; 269: 32844-32847Abstract Full Text PDF PubMed Google Scholar).It has been observed previously that recombinant AC toxin expressed in E. coli exhibits cell-invasive activity identical to the native toxin, but hemolytic activity which is severalfold reduced(17Sebo P. Glaser P. Sakamoto H. Ullmann A. Gene (Amst.). 1991; 104: 19-24Crossref PubMed Scopus (68) Google Scholar). In order to determine whether the nature of the acylation could be the basis of this functional difference, wild type AC toxin from B. pertussis and recombinant AC toxin from E. coli were analyzed by mass spectrometry. We report here that despite the presence of the same requisite CyaC, the pattern and chemical nature of the acylation are different when AC toxin (CyaA) is expressed in these two organisms.MATERIALS AND METHODSPurification of ToxinsThe four proteins evaluated in this study are designated as follows: Bp-CyaA, the native CyaC-activated AC toxin from B. pertussis; Bp-proCyaA, the non-activated AC protoxin from strain BPDE386 lacking functional CyaC; r-CyaA, the recombinant CyaC-activated AC toxin produced in E. coli; and r-proCyaA, the recombinant non-activated AC protoxin produced in E. coli lacking CyaC. The E. coli strain XL1-Blue (Stratagene) was used for expression of r-CyaA from the pCACT3 and for expression of r-proCyaA from pACT7(17Sebo P. Glaser P. Sakamoto H. Ullmann A. Gene (Amst.). 1991; 104: 19-24Crossref PubMed Scopus (68) Google Scholar, 18Betsou F. Sebo P. Guiso N. Infect. Immun. 1993; 61: 3583-3589Crossref PubMed Google Scholar). Strains BP338 and BPDE386 were used for production of Bp-CyaA and Bp-proCyaA, respectively(14Barry E.M. Weiss A.A. Ehrmann I.E. Gray M.C. Hewlett E.L. Goodwin M.S. J. Bacteriol. 1991; 173: 720-726Crossref PubMed Google Scholar).The fermentor cultures of E. coli strains were supplied by the Service of Fermentations of Institut Pasteur (Paris). The r-CyaA and r-proCyaA were extracted from E. coli cell debris after French press disruption, as described previously(17Sebo P. Glaser P. Sakamoto H. Ullmann A. Gene (Amst.). 1991; 104: 19-24Crossref PubMed Scopus (68) Google Scholar). Bp-CyaA was extracted from washed B. pertussis cells with 4 M urea as described previously(5Bellalou J. Sakamoto H. Ladant D. Geoffroy C. Ullmann A. Infect. Immun. 1990; 58: 3242-3247Crossref PubMed Google Scholar). The different CyaAs were purified close to homogeneity by a combination of ion-exchange chromatography (19Sakamoto H. Bellalou J. Sebo P. Ladant D. J. Biol. Chem. 1992; 267: 13598-13602Abstract Full Text PDF PubMed Google Scholar) and affinity chromatography(17Sebo P. Glaser P. Sakamoto H. Ullmann A. Gene (Amst.). 1991; 104: 19-24Crossref PubMed Scopus (68) Google Scholar). In the final step CyaAs were eluted from calmodulin-agarose columns with 8 M urea, 50 mM Tris-HCl, pH 8.0, 2 mM EDTA, and frozen at −70°C.Assays of Adenylate Cyclase Enzymatic, Invasive, and Hemolytic ActivitiesAdenylate cyclase enzymatic activity was measured as described previously(20Ladant D. Michelson S. Sarfati R. Gilles A. Predelenau R. Barzu O. J. Biol. Chem. 1989; 264: 4015-4020Abstract Full Text PDF PubMed Google Scholar). One unit of activity corresponds to 1 μmol of cAMP formed in 1 min at 30°C, pH 8. Hemolytic activity and the cell-invasive activity, which has been shown to be a good measure of intoxicating activity of CyaAs, were both determined using sheep erythrocytes (5 × 108/ml) essentially as described(5Bellalou J. Sakamoto H. Ladant D. Geoffroy C. Ullmann A. Infect. Immun. 1990; 58: 3242-3247Crossref PubMed Google Scholar).Enzymatic DigestionsRecombinant and B. pertussis expressed AC toxin preparations were digested with sequencing grade modified trypsin (Promega) or sequencing grade Asp-N (Boehringer Mannheim) at an enzyme:substrate ratio of approximately 1:50 (w/w) in Tris adjusted to pH 8.5 for trypsin and 8.0 for Asp-N, ~0.2 ml total volume. All digests were carried out for 8 h at 37°C, adjusted in volume to 0.5 ml with 10% acetonitrile, acidified with trifluoroacetic acid, and frozen at −40°C until HPLC fractionation.Microbore HPLC and Screening by MALDI-TOF Mass SpectrometryAll r-CyaA digests and Asp-N digests of Bp-CyaA were separated with a Polymer Laboratories PLRP-S 8 μm 300 Å 2.1 × 250 mm column using a reversed-phase binary gradient(21Gulden P.H. Hackett M. Addona T. Guo L. Walker C.B. Sherman N.E. Shabanowitz J. Hewlett E.L. Hunt D.F. Burlingame A.L. Carr S.A. Mass Spectrometry in the Biological Sciences. Humana, Clifton, NJ1995: 281-306Google Scholar). HPLC conditions for the fractionation of the tryptic fragments derived from Bp-CyaA have also been described previously(15Hackett M. Guo L. Shabanowitz J. Hunt D. Hewlett E. Science. 1994; 266: 433-435Crossref PubMed Scopus (193) Google Scholar). Each fraction was screened for the molecular mass of the fragments using MALDI-TOF mass spectrometry with an instrument built in-house(31Arnott D. Shabanowitz J. Hunt D.F. Proceedings of the 40th American Society for Mass Spectrometry Conference on Mass Spectrometry and Allied Topics, Washington, D. C., May 31 to June 5. American Society for Mass Spectrometry, Santa Fe, NM1992: 328-329Google Scholar), as described previously(15Hackett M. Guo L. Shabanowitz J. Hunt D. Hewlett E. Science. 1994; 266: 433-435Crossref PubMed Scopus (193) Google Scholar, 21Gulden P.H. Hackett M. Addona T. Guo L. Walker C.B. Sherman N.E. Shabanowitz J. Hewlett E.L. Hunt D.F. Burlingame A.L. Carr S.A. Mass Spectrometry in the Biological Sciences. Humana, Clifton, NJ1995: 281-306Google Scholar). Calculations for predicting peptide mass values were performed using Mac ProMass. 2S. Vemuri and T. D. Lee(1989) City of Hope, Duarte, CA. This software is currently marketed by Perkin-Elmer Sciex under the name Mac Biospec. Combined Microcapillary HPLC Electrospray Ionization Mass SpectrometryHPLC Fractions from the tryptic digests which were shown by MALDI-TOF MS to contain peptides which might possess a posttranslational modification were analyzed further for amino acid sequence (22Hunt D.F. Yates J.R. Shabanowitz J. Winston S. Hauer C.R. Proc. Natl. Acad. Sci. U. S. A. 1986; 83: 6233-6237Crossref PubMed Scopus (1073) Google Scholar, 23Hunt D.F. Alexander J.E. McCormack A.L. Martino P.A. Michel H. Shabanowitz J. Sherman N. Moseley M.A. Jorgenson J.W. Tomer K.B. Villafranca J.J. Techniques in Protein Chemistry II. Academic Press, New York1991: 441-454Google Scholar) by tandem quadrupole mass spectrometry using an upgraded TSQ70 instrument with a Finnigan electrospray source, as described(21Gulden P.H. Hackett M. Addona T. Guo L. Walker C.B. Sherman N.E. Shabanowitz J. Hewlett E.L. Hunt D.F. Burlingame A.L. Carr S.A. Mass Spectrometry in the Biological Sciences. Humana, Clifton, NJ1995: 281-306Google Scholar). The molecular masses of the Asp-N fragments of interest were confirmed by electrospray. To obtain semi-quantitative estimates of the relative amounts of palmitoylation versus myristoylation at Lys983, and palmitoylation versus unmodified peptide at Lys860, the following procedures were used. For the two tryptic acyl peptides containing Lys983, the relevant HPLC fractions were adjusted to 100 μl, a 10-μl aliquot removed from each fraction, and the aliquots pooled into two 1.5-ml microcentrifuge tubes, one containing the total myristoylated peptide, the other the total palmitoylated peptide. Each tube was adjusted in volume to 50 μl using a solution of 10% acetonitrile, 0.1% trifluoroacetic acid. Three replicate 1-μl injections were made using the microcapillary HPLC apparatus on-line with the TSQ70, as described previously(15Hackett M. Guo L. Shabanowitz J. Hunt D. Hewlett E. Science. 1994; 266: 433-435Crossref PubMed Scopus (193) Google Scholar, 21Gulden P.H. Hackett M. Addona T. Guo L. Walker C.B. Sherman N.E. Shabanowitz J. Hewlett E.L. Hunt D.F. Burlingame A.L. Carr S.A. Mass Spectrometry in the Biological Sciences. Humana, Clifton, NJ1995: 281-306Google Scholar, 23Hunt D.F. Alexander J.E. McCormack A.L. Martino P.A. Michel H. Shabanowitz J. Sherman N. Moseley M.A. Jorgenson J.W. Tomer K.B. Villafranca J.J. Techniques in Protein Chemistry II. Academic Press, New York1991: 441-454Google Scholar). For each determination of the myristoyl peptide, the centroid signals from ions with +3, +2, and +1 charge states were summed, m/z 532, 797, and 1593. Similarly for the palmitoyl peptide, m/z and were The of the was and the by a pooled for New Scholar). The same was used for the peptide, the charge states for the unmodified Asp-N peptide at m/z to For the palmitoylated Asp-N at m/z the charge states from to were of were using J. Protein 1990; PubMed Scopus Google Scholar) with a and purified by reversed-phase to a P.H. Hackett M. Addona T. Guo L. Walker C.B. Sherman N.E. Shabanowitz J. Hewlett E.L. Hunt D.F. Burlingame A.L. Carr S.A. Mass Spectrometry in the Biological Sciences. Humana, Clifton, NJ1995: 281-306Google Scholar). The modified peptides were acylated to the at the ɛ-amino group of the lysine by the of Scholar). The purified peptides were used to molecular sequence and HPLC AND recombinant AC toxin was produced in E. P. Glaser P. Sakamoto H. Ullmann A. Gene (Amst.). 1991; 104: 19-24Crossref PubMed Scopus (68) Google Scholar, 18Betsou F. Sebo P. Guiso N. Infect. Immun. 1993; 61: 3583-3589Crossref PubMed Google Scholar) and with Bp-CyaA extracted from the B. pertussis. It was that the purified r-CyaA about hemolytic activity on sheep erythrocytes and pore-forming activity in Bp-CyaA R. Maier E. Ladant D. Ullmann A. Sebo P. J. Biochem. (Tokyo). 1994; 269: 22239-27231Google Scholar), the cell-invasive activity of both proteins was In the r-CyaA and Bp-CyaA were at different as to a total of 1 and the cell-invasive and hemolytic activities of the were The cell-invasive AC activity was of the ratio of the proteins in the In the hemolytic activity of the with the of the Bp-CyaA in the an both cell-invasive and hemolytic activities on the posttranslational fatty-acylation of Lys983 of E.M. Weiss A.A. Ehrmann I.E. Gray M.C. Hewlett E.L. Goodwin M.S. J. Bacteriol. 1991; 173: 720-726Crossref PubMed Google Scholar, M. Guo L. Shabanowitz J. Hunt D. Hewlett E. Science. 1994; 266: 433-435Crossref PubMed Scopus (193) Google Scholar), whether or the in the hemolytic activities of r-CyaA and Bp-CyaA could be for by in the chemical nature of posttranslational characterized by mass both native Bp-CyaA and r-CyaA the amino acid sequence from the sequence of the gene P. Ladant D. Sezer O. Pichot F. Ullmann A. Danchin A. Mol. Microbiol. 1988; 2: 19-30Crossref PubMed Scopus (206) Google Scholar, P. Glaser P. Sakamoto H. Ullmann A. Gene (Amst.). 1991; 104: 19-24Crossref PubMed Scopus (68) Google Scholar) and the following posttranslational Bp-CyaA has previously been shown to a modification of a palmitoylation at the ɛ-amino group of M. Guo L. Shabanowitz J. Hunt D. Hewlett E. Science. 1994; 266: 433-435Crossref PubMed Scopus (193) Google Scholar). this acylation was observed when the toxin was produced in the presence of a functional protein and was observed in M. Guo L. Shabanowitz J. Hunt D. Hewlett E. Science. 1994; 266: 433-435Crossref PubMed Scopus (193) Google Scholar). The pattern of acylation in the r-CyaA, from that observed in the B. palmitoylation and myristoylation were observed at Lys983 of the r-CyaA were to a modification at Lys860, entirely of with the unmodified at was also of of r-CyaA are of the tryptic of r-CyaA to and containing palmitoylation at The was m/z For of and 1990; PubMed Scopus Google Scholar). MALDI-TOF MS of the HPLC fraction containing palmitoylated r-CyaA, observed at m/z and to with a mass of MALDI-TOF MS of the fraction containing unmodified at m/z The mass was The observed in the at m/z and are from which was to the as an standard for mass Lys983 and of in RTX toxins. acylation of the two lysine of the E. coli has been observed in P. V. C. Science. 1994; 266: PubMed Scopus Google Scholar), the pattern of in acylation of the these to the that both Lys983 and of might be P. V. C. Science. 1994; 266: PubMed Scopus Google Scholar). appears to be the for r-CyaA produced in E. coli and for the Bp-CyaA produced by B. pertussis. tryptic to of Bp-CyaA, was by MALDI-TOF MS and sequenced by tandem MS This be in tryptic digests was the tryptic The tryptic peptide containing unmodified and to was by HPLC M. Guo L. Shabanowitz J. Hunt D. Hewlett E. Science. 1994; 266: 433-435Crossref PubMed Scopus (193) Google Scholar, 21Gulden P.H. Hackett M. Addona T. Guo L. Walker C.B. Sherman N.E. Shabanowitz J. Hewlett E.L. Hunt D.F. Burlingame A.L. Carr S.A. Mass Spectrometry in the Biological Sciences. Humana, Clifton, NJ1995: 281-306Google Scholar). at m/z was observed by in an HPLC fraction to the of the tryptic palmitoylated peptide. The mass for this peptide was The HPLC was on the of a acyl peptide standard as that shown in and the tryptic acyl peptide from the site in r-CyaA The from Bp-CyaA was to a one Asp-N was from this observed at m/z with unmodified the sequence in the for on the that Bp-CyaA palmitoylated at of the total of Bp-CyaA is essentially of the unmodified from a tryptic of Bp-CyaA, which could from at an unmodified This to The was m/z of its in other RTX M. Guo L. Shabanowitz J. Hunt D. Hewlett E. Science. 1994; 266: 433-435Crossref PubMed Scopus (193) Google Scholar, P. V. C. Science. 1994; 266: PubMed Scopus Google Scholar), the site of r-CyaA was also with to was for M. Guo L. Shabanowitz J. Hunt D. Hewlett E. Science. 1994; 266: 433-435Crossref PubMed Scopus (193) Google Scholar), unmodified was in the presence of the same protein, the is acylated in E. coli and in B. pertussis. One is that acylation of AC toxin is by but that it is by an which as and as a This might have a different in E. coli and in B. pertussis. the in vitro acylation the and the that is a J. C. 1991; PubMed Scopus Google Scholar, E. C. V. Mol. Microbiol. 1991; 5: PubMed Scopus Google Scholar). In amounts of were for activation of and could be in vitro by a acyl a was It is that the might have been as a in the purified preparations of of E. coli, used for the in vitro acylation is that be the acylation of and when expressed in E. coli, its be by in the of the r-proCyaA effect of acylation on toxin a in hemolytic activity, but effect on activity, against the that the involved in hemolytic activity has a in of the catalytic R. Maier E. Ladant D. Ullmann A. Sebo P. J. Biochem. (Tokyo). 1994; 269: 22239-27231Google Scholar, A. Hanski E. J. Biol. Chem. 1992; 267: Full Text PDF PubMed Google Scholar). it appears from activities that Bp-CyaA and r-CyaA are comparable in to into the the in acylation a step in toxin The for this functional is the of to form the hemolytic of target cells as a of AC toxin F. Sebo P. Guiso N. Infect. Immun. 1993; 61: 3583-3589Crossref PubMed Google Scholar), a for of the catalytic domain. B. M. C. G. and E. L. In and hemolytic activities are a of toxin a of a toxin G. Gray M.C. Hewlett E.L. J. Biol. Chem. 1994; 269: 22496-22499Abstract Full Text PDF PubMed Google Scholar, 18Betsou F. Sebo P. Guiso N. Infect. Immun. 1993; 61: 3583-3589Crossref PubMed Google Scholar). the basis of these that in different states of but can deliver catalytic into of AC toxin is a for formation of hemolytic that r-CyaA is in the formation of the channels that for the of with the the acylation of also be involved in of the and formation of It to be determined the acylation on of r-CyaA can its this to of the of fatty-acylation in and INTRODUCTIONAdenylate cyclase (AC) 1The abbreviations used are: ACadenylate cyclaseCADcollision-activated dissociationFmocNα-9-fluorenylmethoxycarbonylHPLChigh performance liquid chromatographyMALDI-TOF MSmatrix-assisted laser desorption/ionization time-of-flight mass spectrometryMSmass spectrometryRSDrelative standard deviationRTXrepeat in toxin. toxin is a bifunctional 177-kDa bacterial protein that contains a calmodulin-activated adenylate cyclase catalytic domain and a pore-forming domain. It exhibits a cell-invasive activity by delivering its catalytic domain to the interior of target cells to elicit supraphysiologic cAMP accumulation(1Confer D.L. Eaton J.W. Science. 1982; 217: 948-950Crossref PubMed Scopus (287) Google Scholar, 2Hanski E. Coote J.G. Alouf J.E. Freer J.H. Sourcebook of Bacterial Toxins. Academic Press, New York1991: 349-366Google Scholar, 3Mock M. Ullmann A. Trends Microbiol. 1993; 1: 187-192Abstract Full Text PDF PubMed Scopus (43) Google Scholar, 4Hewlett E.L. Maloney N.J. Iglewski B. Moss J. Tu A.T. Vaughan M. Handbook of Natural Toxins, Volume 8: Microbial Toxins. Marcel Dekker, New York1994: 425-439Google Scholar). The pore-forming capacity of the molecule is responsible for hemolytic activity against sheep erythrocytes and is entirely independent of the presence and activities of the catalytic domain(5Bellalou J. Sakamoto H. Ladant D. Geoffroy C. Ullmann A. Infect. Immun. 1990; 58: 3242-3247Crossref PubMed Google Scholar, 6Ehrmann I. Gray M. Gordon V. Gray L. Hewlett E.L. FEBS Lett. 1991; 278: 79-83Crossref PubMed Scopus (76) Google Scholar, 7Rogel A. Meller R. Hanski E. J. Biol. Chem. 1991; 266: 3154-3161Abstract Full Text PDF PubMed Google Scholar, 8Szabo G. Gray M.C. Hewlett E.L. J. Biol. Chem. 1994; 269: 22496-22499Abstract Full Text PDF PubMed Google Scholar, 9Benz R. Maier E. Ladant D. Ullmann A. Sebo P. J. Biochem. (Tokyo). 1994; 269: 22239-27231Google Scholar).The AC toxin gene (cyaA) has been cloned and sequenced by Glaser et al.(10Glaser P. Ladant D. Sezer O. Pichot F. Ullmann A. Danchin A. Mol. Microbiol. 1988; 2: 19-30Crossref PubMed Scopus (206) Google Scholar, 11Glaser P. Sakamoto H. Bellalou J. Ullmann A. Danchin A. EMBO J. 1988; 7: 3997-4004Crossref PubMed Scopus (277) Google Scholar) and the AC toxin locus was noted to have homologies with the Escherichia coli hemolysin operon and that of other RTX toxins. These toxins are characterized by a set of glycine- and aspartate-rich nonapeptide repeats(12Welch R. Mol. Microbiol. 1991; 5: 521-528Crossref PubMed Scopus (437) Google Scholar, 13Coote J.G. FEMS Microbiol. Rev. 1992; 88: 137-162Crossref Google Scholar). The formation of biologically active AC toxin requires an accessory protein expressed from an upstream gene, cyaC(14Barry E.M. Weiss A.A. Ehrmann I.E. Gray M.C. Hewlett E.L. Goodwin M.S. J. Bacteriol. 1991; 173: 720-726Crossref PubMed Google Scholar). That protein, CyaC, appears to be involved in posttranslational activation of AC toxin. Recently, tandem mass spectrometry was used to determine that the modification on native AC toxin from Bordetella pertussis (Bp-CyaA) consists of palmitoylation on the ɛ-amino group of Lys983 (15Hackett M. Guo L. Shabanowitz J. Hunt D. Hewlett E. Science. 1994; 266: 433-435Crossref PubMed Scopus (193) Google Scholar). In vitro random chemical acylation has also been shown to confer limited cell-invasive and hemolytic activities on CyaA(16Heveker N. Bonnaffe D. Ullmann A. J. Biol. Chem. 1994; 269: 32844-32847Abstract Full Text PDF PubMed Google Scholar).It has been observed previously that recombinant AC toxin expressed in E. coli exhibits cell-invasive activity identical to the native toxin, but hemolytic activity which is severalfold reduced(17Sebo P. Glaser P. Sakamoto H. Ullmann A. Gene (Amst.). 1991; 104: 19-24Crossref PubMed Scopus (68) Google Scholar). In order to determine whether the nature of the acylation could be the basis of this functional difference, wild type AC toxin from B. pertussis and recombinant AC toxin from E. coli were analyzed by mass spectrometry. We report here that despite the presence of the same requisite CyaC, the pattern and chemical nature of the acylation are different when AC toxin (CyaA) is expressed in these two