INFECTION AND IMMUNITY, May 1990, p. 1195-1200 Vol. 58, No. 50019-9567/90/051195-06$02.00/0Copyright © 1990, American Society for Microbiology

Synthesis and Secretion of Bordetella pertussis Adenylate Cyclaseas a 200-Kilodalton Protein

JACQUES BELLALOU,* DANIEL LADANT, AND HIROSHI SAKAMOTOUnite de Biochimie des Regulations Cellulaires, Institut Pasteur, 28 rue du Docteur Roux, 75724 Paris Cedex 15, France

Received 26 October 1989/Accepted 17 January 1990

Bordetella pertussis, the etiological agent of whooping cough, synthesizes a calmodulin-sensitive adenylatecyclase that is suspected to play a major role in the virulence of this bacterium. We show that adenylate cyclasesynthesized as a 200-kilodalton protein is the product of the cyaA gene and that various virulent Bordetellaspecies secrete this high-molecular-weight polypeptide without apparent proteolytic processing. When sub-mitted to trypsin digestion, the 200-kilodalton protein was converted to a stable 45- to 50-kilodalton species.This corresponds to the size of the enzyme previously purffied from a culture supernatant. The molecularheterogeneity reported for the various identified forms of adenylate cyclase could therefore result in part fromproteolytic degradation or molecular aggregation of the major 200-kilodalton form of the enzyme.

Bordetella pertussis, the causative agent of whoopingcough, produces a variety of factors responsible for itsvirulence (24). One of them, a calmodulin-activated adenyl-ate cyclase (AC) (26), has been shown to invade eucaryoticcells and to provoke damage by disturbing intracellularcyclic AMP metabolism (4). The cloning and sequencing ofthe structural gene, cyaA, revealed a single open readingframe of 1,706 codons (7) encoding a high-molecular-massprotein of approximately 200 kilodaltons (kDa).Both whole-cell extracts and culture supernatants have

been shown to exhibit AC activity (11). Intra- and extracel-lular enzymes have been purified by several groups andrevealed a variety of forms ranging from 45 to 700 kDa (10,12, 15, 22). Recently, Rogel et al. (20) and Masure and Storm(17) identified in bacterial extracts a 200-kDa protein thatcross-reacted with antibodies obtained by immunizing rab-bits with the low-molecular-mass (45-kDa) form of the en-zyme. It was suggested that the 200-kDa polypeptide largeprecursor of the AC is proteolytically processed during thesecretion of the enzyme into smaller forms of 45 to 50 kDa(17, 20). However, neither the process governing secretionof AC into the external medium nor the relationship betweenthe different molecular forms of the enzyme has been estab-lished. In this study, we report direct evidence that the cyaAgene product is a 200-kDa protein that is secreted into theculture medium with no apparent processing.

MATERIALS AND METHODSChemicals. Tolylsulfonyl phenylalanyl chloromethyl ke-

tone-trypsin, soybean trypsin inhibitor, and calmodulin-agarose were from Sigma Chemical Co. (St. Louis, Mo.).Urea was a product of Schwarz/Mann, Orangeburg, N.Y.Alkaline phosphatase labeled with anti-rabbit immunoglob-ulin was purchased from Promega Biotec Co. (Madison,Wis.). Azidocalmodulin (azido-CaM) was synthesized andiodinated as described previously (14). [1251]azido-CaM wasstored at -20°C, protected from light.

Analytical procedures. AC activity was determined in thepresence of 0.1 ,uM CaM as described previously (15). Oneunit of enzyme activity corresponds to 1 p.mol of cyclic AMPformed per min at 30°C and pH 8. Amino acid analysis wasperformed on a Biotronic LC 5001 amino acid analyzer by

* Corresponding author.

using a single-column procedure. Protein concentrationswere measured as described by Bradford (2). Sodium dode-cyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)was performed as described by Laemmli (16).

Bacterial strains and growth conditions. Bacterial strainsused in this study are listed in Table 1. Virulent strains of B.pertussis are characterized by the following phenotypes:filamentous hemagglutinin (Fha), pertussis toxin (Ptx) syn-thesis, extracellular AC activity, and hemolytic activity ofisolated bacterial clones on blood agar (24). Bacteria weregrown on Bordet Gengou (Difco Laboratories, Detroit,Mich.) plates supplemented with 15% sheep blood andsubcultured in liquid Stainer and Scholte (SS) medium (23)sterilized on a 0.22-pm pore-size Sterivex filter (MilliporeCorp., Bedford, Mass.). Cells from 20 ml of exponentiallygrowing cultures were used to inoculate 230 ml of SSmedium, and growth was carried out for 20 to 40 h at 36°Cunder aeration. C-mode cells, characterized by a loss ofvirulence factors, were obtained by growth in SS mediumcontaining an excess of nicotinic acid (13).

Plasmids and constructions of strains. The plasmids used inthis work are listed in Table 1, and their construction schemeis shown in Fig. 1. Plasmid pHSP6 is a pLA2917 (1) deriv-ative containing a Sau3A DNA fragment harboring the ampi-cillin resistance gene from pBR322 cloned into the BglII site.Plasmid pDIA5209 is a pLA2917 derivative containing anNsiI DNA fragment harboring the entire cya operon andcloned into the PstI site. Plasmid pDIA5214 is a pLA2917derivative containing the 10-kilobase BamHI fragment as inpDIA5209 cloned into the BglII site. Plasmid pHSP3 is apDIA5214 derivative obtained after internal deletion of the9.5-kilobase EcoRI fragment. Plasmid pHSP5 is a pDIA5214derivative obtained after internal deletion of the 6.5-kilobaseXhoI fragment. All of these plasmids were introduced intostrain B. pertussis 348 by bacterial mating as previouslydescribed (8).

Preparation of bacterial extracts and detection of AC onWestern immunoblots. The cell pellet from a 10-ml culturegrown for 40 h was stirred with 0.8 ml of 5 M urea in bufferA (50 mM Tris hydrochloride [pH 8], 0.22 mM CaCl2, 0.1%Nonidet P-40) for 1 h at room temperature. After centrifu-gation for 20 min at 13,000 x g, the supernatants weretreated with Laemmli sample buffer and run on a 7.5% SDSgel. In some experiments, bacterial samples were directly

1195

1196 BELLALOU ET AL.

TABLE 1. Bacterial strains and plasmids

Bacterial strain or Phenotype Reference or sourceplasmid

B. pertussis18323 AC' Hly+ 1918323M AC- Hly- This studyTohama AC+ Hly+ 25348 AC- Hly- 25

B. bronchiseptica AC+ Hly+ Pasteur Institute collectionB. parapertussis AC' Hly+ Pasteur Institute collection

PlasmidspHSP6 Tcr Apr This studypDIA5209 AC+ Hly+ Tcr This studypHSP3 AC+ Hly- Tcr This studypHSP5 AC' Hly- Tcr This study

equilibrated in SDS sample buffer (100 mM Tris hydrochlo-ride [pH 6.8], 4 mM EDTA, 10% [vol/vol] glycerol, 2%[wt/vol] SDS, 0.1% bromophenol blue) and electrophoresed.When samples of crude supernatants were tested, the pro-teinaceous material present in these supernatants was pre-cipitated with trichloroacetic acid (10% final concentration)at 0°C for 1 h. The precipitated material was harvested bycentrifugation (10,000 x g for 15 min at 4°C), and the pelletwas suspended in sample buffer. Proteins were transferredfrom polyacrylamide gels to nitrocellulose sheets and incu-bated with rabbit polyclonal immune sera directed againstthe 45-kDa secreted AC of B. pertussis (18). The sera weredepleted of contamining antibodies by incubating them withcrude bacterial extract of an avirulent strain of Bordetellabronchiseptica, which is a common pathogen of rabbits. Theimmunochemical detection was performed with 251I-labeledprotein A (18) or with alkaline phosphatase-labeled anti-rabbit immunoglobulins (9).

Purification of the 200-kDa form (AC200) of AC. Fourgrams of virulent B. pertussis 18323 were stirred with 25 mlof 5 M urea in buffer A for 1 h at room temperature and thencentrifuged at 13,000 x g for 20 min. The supernatant was

diluted four times with buffer A and adsorbed overnight at4°C onto 3 ml of CaM-agarose. The gel was washed withbuffer A and then with 0.5 M NaCl in buffer A. Boundproteins were eluted with 8 M urea in buffer A. The purifiedenzyme was extensively dialyzed against buffer A and storedat -20°C. About 300 U (corresponding to about 750 ,ug ofprotein) of pure AC200 was obtained, with a mean recoveryof 30%.

Photoaffinity labeling. Purified AC200 (2 U) and 70 nM[1251]azido-CaM in 100 ,ul of buffer A were irradiated at 4°Cfor 2 min with a longwave mercury lamp (mineral lightUVSL 58 without a screen) positioned at 5 cm from thesample.

Limited proteolysis with trypsin. Two units of purifiedintracellular 200-kDa AC in 100 ,ul of buffer A was supple-mented with 1 FxM CaM and then incubated at 4°C with 100ng of tolylsulfonyl phenylalanyl chloromethyl ketone-treatedtrypsin. At different time intervals, 20-,ul samples werewithdrawn and supplemented with 10 ,ug of soybean trypsininhibitor. After the residual AC activity was determined, theremaining samples were subjected to 7.5% SDS-PAGE, andproteins were transferred to nitrocellulose and immunoblot-ted as described above. The ['25I]azido-CaM-labeled en-zyme was trypsinized identically. After electrophoresis theproteins were stained with Coomassie blue, and dried gelswere exposed to Kodak X-Omat AR film at -80°C with anintensifying screen.

RESULTS AND DISCUSSIONCharacterization of the cyaA gene product as a 200-kDa

protein. Several authors (15, 20, 22) purified a 45- to 50-kDaform of AC. However, the predicted amino acid sequence ofthe cyaA gene product indicated that AC was synthesized asa 200-kDa polypeptide (8). We wanted to know whetherantibodies directed against the purified 45-kDa AC couldreact with a high-molecular-mass form of AC. When bacte-rial extracts of virulent strains (18323 and Tohama) weretested in Western blots, these antibodies recognized mainlya 200-kDa polypeptide (Fig. 2). This polypeptide was notdetected in bacterial extracts of the avirulent cya strain B.

82/S B2/SE

pHSP 6

E B-r%IA rnfA I I IpULA UWUU

pDIA5214

E XI l

X B N/P EI I I

A I B I D IE -I

E B6B2 E X x B/B2 EI I I I I I

B/B2E

pHSP 3

B/B2E

E

B-82E X EI,II,

pHSP51 kh

FIG. 1. Construction scheme of plasmids used in this work. Thin bars are B. pertussis chromosomal DNA, hatched boxes are vector DNA,and open boxes are cyaA, -B, -D, and -E. Apr is the pBR322 ampicillin resistance gene. Restriction sites: E, EcoRI; B, BamHI; B2, BglII;N, NsiI; P, PstI; X, XhoI; S, Sau3A. The arrow shows the direction of transcription.

EI I

I 7 I iL-.dK- le .r

a --l -1a i d-'

INFECT. IMMUN.

BORDETELLA PERTUSSIS ADENYLATE CYCLASE 1197

TABLE 2. Amino acid composition of the 200-kDa AC purifiedon CaM-Affi-Gela

No. of the indicated amino acidResidues Deduced from the cyaA

gene DNA sequence

Asp + Asn 194 228Thr 71 79Ser 82 98Glu + Gln 155 156Pro 48 37Gly 197 239Ala 238 226Cys NDb 0Val 127 135Met ND 22Ile 65 69Leu 152 146Tyr 39 41Phe 42 42His 34 34Lys 50 43Arg 98 96Trp ND 15

a The values given are uncorrected and determined after 20 h of hydrolysisin 6 N HCI at 110oC.

b ND, Not determined.

pertussis 348 (resulting from insertion of a Tn5 transposon atthe 5' end of the cyaA gene [25]) and the 18323 phase IV B.pertussis strain and in C-mode cells, which are all devoid ofAC and hemolytic activity (13, 25). In all cases the amount ofthe 200-kDa protein in bacterial extracts paralleled the levelof AC activity. The following three criteria were used toprove that the 200-kDa protein is indeed the cyaA geneproduct.

(i) The enzyme was purified by CaM-agarose chromatog-raphy from extracts of the B. pertussis 18323 strain asdescribed in Materials and Methods. The purified prepara-tion (specific activity, 390 U/mg of protein at saturating CaMconcentration) exhibited a single band of 200 kDa on SDS-PAGE (see Fig. 4C, lane 5). Its amino acid composition wasin agreement with the amino acid composition deduced fromthe nucleotide sequence of the cyaA gene (Table 2).

(ii) It has been recently reported that AC and hemolyticactivities could be restored in B. pertussis 348 (AC- Hly-) bycomplementation with plasmids carrying the complete cyaAoperon (3, 8). We extended these findings by showing thatcomplementation with plasmid pDIA5209 resulted in thesynthesis of a 200-kDa polypeptide (Fig. 3, lane 2).

(iii) Deletions of the cyaA gene at its 3' terminus by 700and 850 codons were expected to yield truncated AC prod-ucts in the recombinant strains after transformation of B.pertussis 348 with the corresponding plasmids pHSP5 andpHSP3 (Fig. 1). Bacterial extracts of the recombinant strainsexhibited AC activity, and Western blot analysis revealedtwo polypeptides that cross-reacted with anti-AC antibodiesof 120 kDA [B. pertussis 348(pHSP5)] and 95 kDa [B.pertussis 348(pHSP3)], respectively (Fig. 3, lanes 3 and 4).These results showed a parallel decrease in size of the200-kDa AC corresponding to the size of the deletionsgenerated in the cyaA gene.

Conversion by limited proteolysis AC200 into a 45- to50-kDa species. The purified intracellular AC200 was ex-posed for different time intervals to trypsin and subsequentlyanalysed by SDS-PAGE and Western blotting. The 200-kDaprotein was cleaved in less than 1 min in a 43-kDa immuno-

1 2 3 4 5 6

200 _

98-

67-

45-

29-

AC activity 5.2 0 0 1.3 0

Hly phenotype + - - +

FIG. 2. Identification of an intracellular 200-kDa form of AC. B.pertussis urea extracts were obtained as described in Materials andMethods, and subjected to 7.5% SDS-PAGE. The proteins werethen transferred to nitrocellulose and incubated with immune serumdirected against the B. pertussis 45-kDa AC. The immunochemicaldetection was performed with 1251I-labeled protein A. The positionsof molecular mass standards (kilodaltons) are indicated on the leftside of the figure. Lanes: 1, purified 45-kDa AC; 2, urea extract ofB.pertussis 18323 (AC+ Hly'); 3, B. pertussis 18323M, phase IVavirulent strain (AC- Hly-); 4, B. pertussis 18323 C mode (AC-Hly-); 5, B. pertussis Tohama (AC' Hly+); 6, B. pertussis 348 (AC-Hly-). The specific AC activity of the bacterial extracts (expressedin units per milligram of protein) and hemolytic phenotypes ofstrains are indicated on the bottom of each lane. The standards wererabbit muscle myosin (200 kDa), rabbit muscle phosphorylase b (98kDa), bovine serum albumin (67 kDa), ovalbumin (47 kDa), andbovine erythrocyte carbonic anhydrase (29 kDa).

reactive fragment with no loss of enzymatic activity (Fig.4A). After further 10 min of incubation with trypsin, this43-kDa residual fragment seemed to be resistant to proteo-lysis, since no appreciable degradation occurred (Fig. 4A).In a second set of experiments the purified AC200 wascross-linked with [125I]azido-CaM and then submitted toproteolysis by trypsin. Irradiation of a mixture of AC200 and[125I]azido-CaM yielded radioactive species with an appar-ent molecular mass of 210 to 220 kDa (Fig. 4B, lane 5).Exposure of the [125I]azido-CaM-labeled AC200 to trypsindigestion resulted in the conversion of the radioactive cross-linked product to species of 63 kDa (Fig. 4B). Most likely the63-kDa band corresponds to a cross-linked product of[1251I]azido-CaM with the 43-kDa fragment, which was evi-denced by Coomassie blue staining (Fig. 4C). Recently,Gilboa-Ron et al. also reported that the 200-kDa form of ACcould be cleaved by lymphocyte lysates into an active42-kDa fragment (6).

Identification of a high-molecular-weight form of AC inculture supernatants of B. pertussis. It has previously beensuggested that the 45-kDa AC form secreted by B. pertussisis derived from proteolytic processing of a large intracellularprecursor (17, 20). This model was based only on theidentification of low-molecular-weight species of AC, inpartially concentrated or purified culture supernatants. Wetherefore focused our attention on the molecular weightforms ofAC present in crude supernatants of early exponen-tial phase cultures. A Western blot analysis of mid-log-phase

VOL. 58, 1990

1198 BELLALOU ET AL.

20z0

98

67

29

1 2 3 4 5FIG. 3. Immunoblots of bacterial extracts from recombinant

strains expressing the product of partially deleted cyaA gene.Bacterial extracts were prepared as described in Materials andMethods and subjected to 7.5% SDS-PAGE. Immunoblotting ofresolved proteins was performed as defined in the legend to Fig. 1.The molecular weight standards are the same as in Fig. 2. Lanes: 1,B. pertussis 348 with plasmid pHSP6 (no insert); 2, B. pertussis 348with plasmid pDIA5209; 3, B. pertussis 348 with plasmid pHSP5; 4,B. pertussis 348 with plasmid pHSP3; 5, purified 45-kDa AC.

culture supernatants of virulent B. pertussis strains is shownin Fig. SA. The virulent strains B. pertussis 18323 (lane 2)and Tohama (lane 3) but not the control strain phase IV B.pertussis 18323M (lane 4) displayed a major polypeptide of200 kDa that cross-reacted with anti-AC antibodies. Undersimilar experimental conditions, a 200-kDa form of AC wasalso detected in culture supernatant of B. parapertussis andB. bronchiseptica species (Fig. 5A, lanes 5 and 6). Theamount of extracellular 200-kDa polypeptide released inculture supernatants differed widely from one Bordetella

A

100 17 E07 98

200-

98 -

67 -

45

29 -

1 2 3 4

B

1!f)o r79 !fJt3 109

200-

98 -

67 ft740G45

5 6 7 8

200-

98 -

67 -

45 -

5 6 7 8

FIG. 4. Proteolytic conversion of AC200 to AC45. (A) PurifiedAC200, complexed with CaM, was exposed to digestion by trypsinat 4°C for 0 (lane 1), 1 (lane 2), 5 (lane 3), or 10 (lane 4) min. Residualactivity (expressed as a percentage of the initial activity on the topof each lanes) was determined for each sample. The remainingsamples were subjected to 7.5% SDS-PAGE, transferred to nitro-cellulose, and immunoblotted with anti-AC antibodies as describedin Materials and Methods. (B and C) Purified AC200 was labeledwith [125I]azido-CaM as described in Materials and Methods andexposed to digestion by trypsin at 4°C for 0 (lane 5), 1 (lane 6), 5(lane 7), or 10 (lane 8 ) min. Residual AC activity was determinedand expressed as above. The remaining samples were subjected to7.5% SDS-PAGE, and the gels were stained with Coomassie blue(C) and exposed for autoradiography (B). The molecular weightmarkers as in Fig. 2.

strain to another (Table 3). B. pertussis 18323 produced asmuch as 0.25 to 0.5 ,ug of the 200-kDa protein per ml ofculture supernatant. It is unlikely that this amount of extra-cellular protein could result from cell lysis, because itrepresents about 20 to 50% of the total AC synthesized bythe bacteria. It can therefore be concluded that the largeunprocessed form of AC is secreted into the external me-dium. No difference in molecular weight was observedbetween the intracellular enzyme and that identified in theculture supernatants (Fig. SA). When a supernatant of thevirulent 18323 B. pertussis strain was submitted to limitedproteolysis with trypsin, the 200-kDa polypeptide was con-verted into a low-molecular-mass immunoreactive 43- to50-kDa protein with no loss of AC activity (Fig. 5B).We propose that secretion of B. pertussis AC does not

necessarily require proteolytic cleavage of the 200-kDaintracellular polypeptide. It seems likely that the previouslypurified extracellular 45-kDa AC is a degraded form of thesecreted 200-kDa AC found during culture growth or resultssimply from the enzyme purification procedures used. Infact, growing B. pertussis in filtered sterile SS mediuminstead of the classical autoclaved medium (23) enhancesbacterial growth and significantly delays proteolytic degra-dation of the 200-kDa AC form into 45- to 50-kDa ACspecies. This indicates that the composition of the bacterialgrowth medium could modify the sensitivity of the secreted200-kDa protein to proteolytic degradation. Moreover, whenbovine serum albumin was added to the culture medium, ahigher yield of secreted 200-kDa polypeptide was obtained(data not shown).

Concluding remarks. In the present work, we provideddirect evidence that the cyaA gene product is a 200-kDaprotein that corresponds to the major intracellular form ofthe enzyme AC. We also showed that the 200-kDa polypep-tide was secreted into the external medium of virulent strainsof B. pertussis without any detectable proteolytic process-ing. The secretion of a large, unprocessed form of AC wasalso observed for B. parapertussis and B. bronchisepticastrains and is likely to be common feature of the Bordetellagenus. It is probable that the molecular heterogeneity ofACpurified from culture supernatants results in part from pro-teolytic degradation into catalytically active fragments, ashas been proposed by several authors (17, 20). Nevertheless,proteolytic degradation does not seem to be a relevantphysiological feature, since the toxic form of AC that pene-trates eucaryotic cells has been shown to correspond to the200-kDa protein (21) (data not shown). The finding that AC isnot only synthesized but also secreted as a 200-kDa poly-peptide would therefore account for the main physiologicalrole of this molecular species.

TABLE 3. Whole cells and extracellular AC activities producedby various Bordetella strainsa

AC activity (nmollmin per ml)Strain

Whole cells Supernatant

B. pertussis18323 100 100-15018323M 0 0Tohama 100 2-6

B. bronchiseptica 70 60-70B. parapertussis 70 12-15

a Mid-log-phase cultures were centrifuged. Cell pellets were washed andsuspended in the same volume of saline buffer. AC activity was determined onsonicated bacteria and culture supernatants.

INFECT. IMMUN.

BORDETELLA PERTUSSIS ADENYLATE CYCLASE 1199

A200

go -

67 -

1 2 3 4 5 5

B trypsin proteolysisof secretedAC200

-rw-#wr W- -

KDa

200

98

_67

___ "~-45

1 2 3 4 5 6

100 100 115 108 98

FIG. 5. (A) Identification of a 200-kDa form of AC in culturesupernatants of various Bordetella species. For lane 1, 1 ml ofwhole-cell culture of B. pertussis 18323 was centrifuged, and thepellet was suspended in 100 Rl of 10% SDS sample buffer; 50 [lI ofthe SDS bacterial extract was then subjected to 7.5% SDS-PAGEand immunoblotted as defined in the legend Fig. 2. For lanes 2 to 6,culture supernatants of various Bordetella species were precipitatedwith trichloroacetic acid (10% final concentration) at 4°C for 1 h. Theproteinaceous pellet was washed once with cold acetone and sus-

pended in 10%o SDS sample buffer, and a sample containing between0.03 and 0.05 U of AC activity was subjected to 7.5% SDS-PAGEand immunoblotted. Lanes: 2, B. pertussis 18323; 3, B. pertussisTohama; 4, B. pertussis 18323M (no AC activity, same amount ofprotein as in lane 2); 5, B. parapertussis; 6, B. bronchiseptica. (B)Immunoblots after limited proteolysis of the secreted 200-kDa AC.A 1-ml sample of crude culture supernatant of B. pertussis 18323containing 0.15 U ofAC activity was supplemented with CaM (1 JIMfinal concentration) and incubated with 10 ng of tolylsulfonyl phe-nylalanyl chloromethyl ketone-treated trypsin at 4°C for 0 (lane 1), 1(lane 2), 5 (lane 3), 10 (lane 4), or 20 (lane 5) min. Proteolysis was

stopped by the addition of soybean trypsin inhibitor, and residualAC activity (expressed as a percentage of initial activity andindicated at the bottom of each lane) was determined. The differentsamples were precipitated with trichloroacetic acid. The proteinpellets were run on a 7.5% acrylamide gel and immunoblotted asdescribed in Fig. 1. Purified 45-kDa AC was run as a control (lane 6).The molecular weight standards are as in Fig. 2.

Recent studies revealed that the secretion of B. pertussisAC requires the expression of three additional genes (cyaB,cyaD, and cyaE) contiguous to the structural cyaA gene (8).These four genes appear to form a single operon with astructure similar to that of the E. coli. hly operon. Twoproducts of the secretion genes cyaB and cyaD show a largeextent of similarity with the products of the secretion genesof E. coli alpha-hemolysin hlyB and hlyD (8). Interestingly,E. coli alpha-hemolysin is secreted as a 110-kDa polypeptidewithout apparent proteolytic processing but was shown to berapidly degraded in the culture supernatant (5). A similarobservation is reported here for B. pertussis AC and sup-ports the hypothesis of a common secretion mechanism forthis enzyme and the E. coli alpha-hemolysin.

ACKNOWLEDGMENTS

This work was supported by grants from the Centre National de laRecherche Scientifique (URA D1129), the Pasteur-Weizmann jointResearch Program (Hanski/Ullmann), and Pasteur Vaccins (contract23843).We are indebted to Agnes Ullmann for constant interest and

guidance during this work and for precious help in the preparation ofthe manuscript. We thank Octavian Barzu for helpful discussionsand critical readings of the manuscript, Brid Laoide for linguisticediting, Philippe Glaser for his kind gift of plasmid pDIA5209, andAnne-Marie Gilles for amino acid determination. We are grateful toMireille Ferrand for expert secretarial help.

LITERATURE CITED1. Allen, L. N., and R. S. Hanson. 1985. Construction of broad-

host-range cosmid cloning vectors: identification of genes nec-essary for growth of Methylobacterium organophilum on meth-anol. J. Bacteriol. 161:955-962.

2. Bradford, M. 1976. A rapid and sensitive method for thequantitation of microgram quantities of protein, utilizing theprinciple of protein-dye binding. Anal. Biochem. 72:248-257.

3. Brownlie, R. M., J. G. Coote, R. Parton, J. E. Schultz, A. Rogel,and E. Hanski. 1988. Cloning of the adenylate cyclase geneticdeterminant of Bordetella pertussis and its expression in Esch-erichia coli and B. pertussis. Microb. Pathogen. 4:335-344.

4. Confer, D. L., and J. W. Eaton. 1982. Phagocyte impotencecaused by an invasive bacterial adenylate cyclase. Science217:948-950.

5. Felmlee, T., S. Pellett, E.-Y. Lee, and R. A. Welch. 1985.Escherichia coli hemolysin is released extracellularly withoutcleavage of a signal peptide. J. Bacteriol. 163:88-93.

6. Gilboa-Ron, A., A. Rogel, and E. Hanski. 1989. Bordetellapertussis adenylate cyclase inactivation by the host cell. Bio-chem. J. 262:25-31.

7. Glaser, P., D. Ladant, 0. Sezer, F. Pichot, A. Ulimann, and A.Danchin. 1988. The calmodulin-sensitive adenylate cyclase ofBordetella pertussis: cloning and expression in Escherichia coli.Mol. Microbiol. 2:19-30.

8. Glaser, P., H. Sakamoto, J. Bellalou, A. Ullmann, and A.Danchin. 1988. Secretion of cyclolysin, the calmodulin-sensitiveadenylate cyclase haemolysin bifunctional protein of Bordetellapertussis. EMBO J. 7:3997-4004.

9. Goyard, S., C. Orlando, J.-M. Sabatier, E. Labruyere, J.d'Alayer, G. Fontan, J. van Rietschoten, M. Mock, A. Danchin,A. Ullmann, and A. Monneron. 1989. Identification of a commondomain in calmodulin-activated eukaryotic and bacterial adenyl-ate cyclases. Biochemistry 28:1964-1967.

10. Hanski, E., and Z. Farfel. 1985. Bordetella pertussis invasiveadenylate cyclase: partial resolution and properties of its cellu-lar penetration. J. Biol. Chem. 290:5526-5532.

11. Hewlett, E., and J. Wolff. 1976. Soluble adenylate cyclase fromthe culture medium of Bordetella pertussis: purification andcharacterization. J. Bacteriol. 127:890-898.

12. Kessin, R. H., and J. Franke. 1986. Secreted adenylate cyclaseof Bordetella pertussis: calmodulin requirements and partial

VOL. 58, 1990

INFECT. IMMUN.

purification of two forms. J. Bacteriol. 166:290-296.13. Lacey, B. W. 1960. Antigenic modulation of Bordetella pertus-

sis. J. Hyg. 58:57-93.14. Ladant, D. 1988. Interaction of Bordetella pertussis adenylate

cyclase with calmodulin. Identification of two separated calmod-ulin-binding domains. J. Biol. Chem. 263:2612-2618.

15. Ladant, D., C. Brezin, J.-M. Alonso, I. Crenon, and N. Guiso.1986. Bordetella pertussis adenylate cyclase. J. Biol. Chem.261:16264-16269.

16. Laemmli, U. K. 1970. Cleavage of structural protein during theassembly of the head of bacteriophage T4. Nature (London)277:680-685.

17. Masure, H. R., and D. R. Storm. 1989. Characterization of thebacterial cell associated calmodulin-sensitive adenylate cyclasefrom Bordetella pertussis. Biochemistry 28:438-447.

18. Monneron, A., D. Ladant, J. d'Alayer, J. Bellalou, 0. Barzu,and A. Ullmann. 1988. Immunological relatedness betweenBordetella pertussis and rat brain adenylyl cyclases. Biochem-istry 27:536-539.

19. Pittman, M. 1984. Genus Bordetella Moreno-Lopez 1952 178AL,p. 388-393. In N. R. Krieg and J. G. Holt (ed.), Bergey's manualof systematic bacteriology, vol. 1. The Williams & Wilkins Co.,Baltimore.

20. Rogel, A., Z. Farfel, S. Goldschmidt, J. Shiloah, and E. Hanski.1988. Bordetella pertussis adenylate cyclase. Identification ofmultiple forms of the enzyme by antibodies. J. Biol. Chem.263:13310-13316.

21. Rogel, A., J. Schultz, R. M. Brownlie, J. G. Coote, R. Parton,and E. Hanski. 1989. Bordetella pertussis adenylate cyclase:purification and characterization of the toxic form of the en-zyme. EMBO J. 8:2755-2760.

22. Shattuck, R. L., D. J. Oldenburg, and D. R. Storm. 1985.Purification and characterization of a calmodulin-sensitive ade-nylate cyclase from Bordetella pertussis. Biochemistry 24:6356-6362.

23. Stainer, D. W., and M. J. Scholte. 1971. A simple chemicallydefined medium for the production of phase I Bordetella pertus-sis. J. Gen. Microbiol. 63:211-220.

24. Weiss, A. A., and E. L. Hewlett. 1986. Virulence factors ofBordetella pertussis. Annu. Rev. Microbiol. 40:661-686.

25. Weiss, A. A., E. L. Hewlett, G. A. Myers, and S. Falkow. 1983.Tn5-induced mutations affecting virulence factors of Bordetellapertussis. Infect. Immun. 42:33-41.

26. Wolff, J., G. H. Cook, A. R. Goldhammer, and S. A. Berkovitz.1980. Calmodulin activates prokaryotic adenylate cyclase. Proc.Natl. Acad. Sci. USA 77:3841-3844.

1200 BELLALOU ET AL.