Gene, 104 (1991) 19-24 0
1991 Elsevier
GENE
Science
Publishers
B.V. All rights reserved
19
0378-l 119/91/$03.50
05075
High-level synthesis of active adenylate Escherichia coli system (Recombinant
DNA;
holotoxin
activation;
cyclase
invasiveness;
toxin of Bordetellu
hemolytic
pertussis
in a reconstructed
activity)
Peter Sebo a, Philippe Glaser b, Hiroshi Sakamoto a and Agnes Ullmann a ” Unite de Biochimie des Rt!gulationsCellulaires,and h UmitPde R&ulation de I’Expression GPnPtique,Dipartement de Biochimie et GMtique Mokulaire. Institut Pasteur, 75724 Paris Cedex 15 (France) Received by J.-P. Lecocq: 18 February 1991 Revised/Accepted: 19 April/l 8 May 199 1 Received at publishers: 27 May 1991
SUMMARY
The Bordetellu pertussis adenylate cyclase(Cya) toxin-encoding locus (cya) is composed of five genes. The cyaA gene encodes a virulence factor (CyaA), exhibiting adenylate cyclase, hemolytic and invasive activities. The cyuB, D and E gene products are necessary for CyaA transport, and the cyaC gene product is required to activate CyaA. We reconstructed, in Escherichia coli, the cya locus of B. pertussis by cloning the different genes on appropriate vectors under the control of strong promoters and E. coli-specific translation initiation signals. We show that in the absence of additional gene products, CyaA is synthesized at high levels, is endowed with adenylate cyclase activity, but is devoid of invasive and hemolytic activities. CyaC is sufftcient to confer upon the adenylate cyclase holotoxin full invasive and partial hemolytic activities. Coexpression of the cy&I, D and E genes neither stimulates nor potentiates the activation brought about by CyaC. This reconstructed system should help to elucidate both the mechanism and the structural requirements of holotoxin activation.
INTRODUCTION
The extracellular adenylate cyclase toxin (CyaA) produced by B. pertussis, the causative agent of whooping cough, is now considered as one of the major virulence Correspondence to: Dr. A. Ullmann, Unite de Biochimie des Regulations Institut Pasteur, 28 rue du Docteur Roux, 75724 Paris Cedex
Cellulaires,
15 (France) Tel.: (33-1)45688385;
Fax: (33-1)43069835.
Abbreviations:
A,,,,, absorbance at 600 mm; aa, amino acid(s); Ap, B., Border&z; bp, base pair(s); CAMP, cyclic AMP; Cya,
ampicillin; adenylate
cyclase; cya, gene encoding
isopropyl-p-D-thiogalactopyranoside; nucleotide(s);
oligo,
Cya; Cm, chloramphenicol; kb,
kilobase
oligodeoxyribonucleotide;
or ORF,
IPTG,
1000 bp; open
nt,
reading
frame; ori, origin of DNA replication; PAGE, polyacrylamide-gel electrophoresis; P,,,< , lac promoter; PC,,, hybrid trp-lacpromoter; R, resistant; RBS, ribosome-binding site; rrnB, ribosomal RNA transcriptional unit; SDS, sodium
dodecyl
sulfate;
u, unit(s);
wt, wild type.
factors involved in the pathogenesis (St. Mary Goodwin and Weiss, 1990). The toxin penetrates various mammalian cells and on activation by the intracellular calmodulin, elicits a dramatic increase in the intracellular CAMP concentration resulting in impairment of cellular functions (see Weiss and Hewlett, 1986, for review). The nt sequence of the cyaA gene, encoding the toxin, and the characterization of the gene product revealed that CyaA is a bifunctional protein of 1706 aa exhibiting both invasive adenylate cyclase and hemolysin activities (Glaser et al., 1988a,b; Rogel et al., 1989; Hewlett et al., 1989; Bellalou et al., 1990). The CyaA protein is composed of two regions. The N-terminal 400-aa domain corresponds to the calmodulin-activated adenylate cyclase activity (Ladant et al., 1989); the C-terminal 1306 aa, which share 25% sequence homology with E. coli x-hemolysin, are essential for invasiveness and hemolytic activity of the CyaA toxin (Bellalou et al., 1990). The secretion of the CyaA protein
20
(A)
Plac
pDIA5240
(4.0
oril
kb)
AP R pACT7
(7.9
DPSG I5
(9.8
kb)
pPSGlE
(9.8
kb)
pPSG2E
(I 0.7 kb)
pPSG3CE
(1 I .6b)
(E/N)
pPS4C
(5.4
ori 1
kb)
WN)
i A
kb)
(B) pDlA5240
pPSGlE Fig. 1. Construction in pTZl8
of plasmids
(Pharmacia)
for adenylate
was used as starting
ofcyaA on pDIA5227 ATCAGG
,..
pDIA5240,
form of CyaA consisting
of the two N-terminal
by insertion
and A&II sites around
of pHSG575
Aat _--. II ,!f&&AACAGACCaLgACGTCG-3’.
cyclase toxin expression.
contained
(Takeshita
(CYaS, D, El
of plasmids strand.
The construction
of E. coli fi-galactosidase fragment
in plasmid
codon
in pPSGl5.
(Gilles et al., 1990) bearing the first 399 codons ofc.VuA cloned
for cyaA expression.
In the first step, the 93 nt preceding
ofpDIA5211
pACT7.
fused to the N-terminal
the ATG
the entire fragment
the C-terminal
1437 aa of CyaA, between
of the cyaB, D and E genes we used the unique
BspHI-&I
fragment
of pDlA5211
RBS and start codon (underlined)
of the 6.2-kb AatII-Hind111
fragment
of cyuA.
of the truncated
399 aa of the toxin, In the second step the CJVJA
(Glaser et al., 1988a), encoding
For subcloning
Then, a synthetic
of the cyaB gene, by insertion
was verified by resequencing
of lac.2 fused to the ORF of cyaA; it allowed the overproduction
the start codon of cyaB. First, a 6.2-kb blunt-ended
et al., 1987) resulting
was fused to the second
of the complementary
the RBS and two codons
resulting
*)
(Taylor et al., 1985) using a 34-mer S’-dGAAACAGCTATGACCATGCAGCAATCGC-
mutagenesis
residues
(cyaA
Plasmid pDIA5227
for the construction
of a 4389-bp Bell-BspHI
the BclI and EcoRI sites of pDIA5240, @HI
5’-
were deleted by site-directed
plasmid,
gene was completed
BAACAGCT&f&ACCi&gCAGAA-3’
material
as a primer for the in vitro synthesis
The resulting
5’-
was inserted
sequence
of pPSGl5
and overlapping
into the BarnHI
site
5’-AGGAAACAGACCG
into the blunt-ended
Nrol site of the
pKK233-2 vector (Pharmacia) by the use of a synthetic adaptor S’-ACGT, to preserve the AarII site. This resulted in pPSGIE. Finally, the 7.1-kb EcoRI-ScaI fragment ofpPSGlE was inserted between the EcoRI and Hind111 sites of pHSG575, yielding pPSG2E. The restored EcoRI site of pPSG2E was used for insertion were removed standard
of the 0.9-kb NaeI fragment
from pPSG3CE
protocols
(Sambrook
by an AatII-Hind111
of pDIA5211, deletion,
et al., 1989). (A) Schematic
containing
resulting
the cyaC gene, resulting
in pPS4C.
representation
The recombinant
of the constructed
in pPSG3CE.
In the end, the cyaB, D and E genes
DNA manipulations
plasmids.
(B) Sequences
were performed
according
of the newly introduced
to
RBSes
of cyuA and cyaB genes. The start codons of cyaA and cyaB genes are indicated by lower-case underlined characters. Symbols: cyaA *, first 399 codons of evaA; P,,,, I’,,,, promoters; oril, pMBl-derived origin of replication; ori2, pSClOl-derived origin of replication; r, T2, transcription terminators of the rrnB region. A,AotII; by ligation
Bc, BclI; E,EcoRI;
of blunt ends. (A/H).
H, HindIII;
N, NueI; SC, ScaI; Sm, SmnI; Xh,XhoI.
(E/N), (SC/H) are sites erased
by ligation
of blunt ends.
Restriction
sites in parentheses
are those that are destroyed
21 requires the products
of three genes, cyaB, D and E, located
downstream from the cyaA gene (Glaser et al., 1988b). Rogel et al. (1989) showed that the expression of the cyaA
12
kDa
345
C yaA*
- 205
gene in E. coli leads to the production of a catalytically active, 200-kDa CyaA devoid of invasive and hemolytic activities. They proposed that a post-translational moditication, which occurs in B. pertussis but not in E. coli, confers upon the CyaA protein the toxic properties. Recently, a cyaC gene homologous to the hlyC gene, required for the activation of the hlyA gene product in E. cob, has been discovered in B. pertussis and shown to be necessary for hemolytic and toxin activities of the cyaA gene product (Barry et al., 1991). In B. pertussis, however, the secretion genes cyaB, D and E, are coexpressed with cyaC; therefore their involvement in the activation of CyaA protein could not be excluded. The aim of the present work was to produce active CyaA toxin in a reconstructed E. coli system. We show that, whereas the cvaC gene product is sufficient to confer full invasive activity on the CyaA protein, additional and yet unidentified factors are probably required to restore full hemolytic activity.
analysis
DISCUSSION
Ap/ml
toxins. Bacteria
-
66
-
45
of recombinant
Table I were grown at 37°C to an absorbance with
were harvested,
disrupted
by sonication.
proteins
were pelleted
containing
resuspended
cyclase activity were resuspended
were stored at -20°C.
1 ml of 8 M urea extracts washed
at -20°C.
(7.5%) according
1 ml calmodulin-agarose
The proteins
to Laemmli
urea extract B.
columns.
were
separated
(1970). Cell-associated
2 mM EDTA
CyaA protein from
to Bellalou et al. (1990). Lanes: 1,8 M
PS30 and PS40, respectively;
pertussis;M, standards
The col-
by SDS-PAGE
of cell debris of PS40; 2, 3. 4, afftnity-purified
of PSlO,
of the
four times with
with 2 M urea in Buffer A and the CyaA
B. pertussis was purified according extracts
in 1 ml of
For the purification was diluted
were eluted with 8 M urea in Buffer A containing
and stored
CyaA
at 16000 x g. The pellets
HCl pH S.OjO.2 mM CaCI, (Buffer A). At this
Buffer A and passed through proteins
(AhO,, = 2.0 f 0.2) the
and membrane-associated
with the cell debris
60”/, of adenylate
listed in
= 0.2 and induced
in M63 medium (Miller, 1972) and
The aggregated
8 M urea in 50 mM Tris
with 150 ng
strains
of&,,
1mM IPTG. After an additional 4 h ofgrowth
bacteria
were grown in
et al., 1989) supplemented
and 12 pg Cm/ml. The cultures
umns were extensively
(a) Design of an Escherichia coli system for expression of the Bordetella pertussis cya genes The cya locus of B. pertussis is composed of five genes: cyaA encodes the bifunctional CyaA protein, whose secretion requires the expression of the downstream cyaB, D, and E genes (Glaser et al., 1988b). Activation of CyaA requires the expression of a fifth gene, cyaC, located upstream from cyaA (Barry et al., 1991). B. pertussis is a slow-growing organism and the tools for its genetic analysis are still limited. As a first step toward the understanding of the mechanism of activation of the CyaA toxin we developed a system in E. coli that would enable the expression of CyaA endowed with invasive and hemolytic activities. Therefore, we cloned the five known genes of the cya locus in E. coli to assess the contribution of each one to the toxin activity. Preliminary experiments showed that expression of the cyaA gene cloned in E. coli on multicopy plasmids under control of strong inducible promoters was surprisingly low, presumably due to inefficient recognition of the cyaA translation initiation signals. We therefore constructed a plasmid (pACT7, Fig. 1) which contained the cyaA gene expressed under the control of the transcription and translation initiation signals of IucZ. The cyaB, D, E and C genes were cloned in a low copy number vector (five to six copies per cell), pHSG575, compatible with the pMB 1 replicon derived plasmids, such as pACT7. A strong inducible promoter, P,,,. and a synthetic
of purified
liquid 2 x YT medium (Sambrook
stage the urea extracts AND
97.4
- 29 Fig. 2. SDS-PAGE
CyaA proteins, RESULTS
-
are indicated
toxins from
5, toxin purified
from
on the right margin.
RBS were placed upstream from the ATG start codon of cyaB, and two consecutive transcriptional terminators of the rrnB gene were placed downstream from the cyaE gene (pPSG2E, Fig. 1). The cyaC gene was cloned under the control of the lac promoter, either upstream from the cyaB, D, E genes (pPSG3CE, Fig. l), or separately in a modified pHSG75 vector (pPS4C, Fig. 1). (b) Purification of CyaA proteins expressed in Escherichia coli The recombinant strains harboring different combinations of plasmids are shown in Table I. The expression from the pACT7 plasmid led to high level production of CyaA, which represented about 2% of the E. coli total proteins. Furthermore, over 60% of the adenylate cyclase activity was recovered within the cell debris after sonication. The presence or absence of the cyaC, B, D, and E genes encoding activation and secretion functions did not affect the apparent aggregation or membrane association of the CyaA proteins, produced in E. coli. It is surprising that,
22 TABLE
I
Invasive
and hemolytic
Recombinant
activities
Plasmid
of CyaA proteins
content h
in Escherichia co11
produced
Gene
strains I’
Adenylate activity’
context
cyclase
Hemolytic
(Cya)
Hemolytic
activity d
Internalized
activity Gya activity
( ‘, ) Internalized
Input u/ml
u/ml
PSlO
pACT7
f pHSG575
cyuA
PS20
pACT7
+ pPSG2E
cyuA,B.D,E
PS30
pACT7
+ pPSG3CE
cyaA,C,B,D,E
940
7.3
14
1.91
PS40
pACT7
+ pPS4C
cyaA,C
861
11.2
31
2.76
” The recombinant I~clqZAM15, h Plasmids
strains
are described
‘ Adenylate
(Stratagene)
of erythrocytes 5 The ratios
activities
by transformation
Xl-l
0.05
ND
-
0.07
ND
-
{en&l,
hsdRl7, supE44, thi-I, I
with the indicated
pairs of plasmids,
, recAl, gyrA96, relA1, d(lac-proB)[F’,
described
proAB
’.
in Fig. 1.
in Fig. 1 legend.
cyclase activities
were measured
at 30°C and pH 8. lnternalized d Hemolytic
from E. co/i strain
were derived
TnZO(tetR)]}
885 1400
adenylate
were determined
as previously cyclase
described
activities
by incubation
(Ladant
were determined
of sheep erythrocytes
et al., 1989). One unit (u) corresponds after 30 min incubation,
to 1 nmol of CAMP formed per min
as described
previously
(Bellalou
et al., 1990).
(5 x 10s cells/ml) with the toxins at 37°C for 270 min and expressed
in “,,
lysed. ND, not detectable of 14/7.3 and 31/l 1.2.
even though we have preliminary evidence, in a minicell system, that the cyaB, D and E genes were expressed in E. coli, the CyaA protein was not secreted. In addition, when a replicative derivative of pPSG 1E (Fig. 1) was introduced into a secretion-deficient B. pertussis mutant, CyaA secretion was restored to wt level. The aggregation of the CyaA proteins enabled a simple purification procedure by extracting the cell debris with 8 M urea (Glaser et al., 1989), followed by single-step affinity chromatography on calmodulin-agarose. As shown in Fig. 2, the purified preparations of the 200-kDa CyaA protein contained also some lower-M, forms; they correspond to degradation products still bearing the N-terminal calmodulin-binding domain, as assessed by immunoblots (not shown). The yields of CyaA purification ranged between 40 and 50% and about 1 mg of the CyaA proteins, expressed in different genetic backgrounds, could be obtained from lOO-ml cultures with specific activities ranging from 175 to 200 pmol cAMP/min/mg. TABLE
II
Invasive
and hemolytic
Strains”
activities
of the purified
toxins Adenylate
Gene
(c) Invasive activity of CyaA toxin produced in Eschevichia coli We expressed various combinations of c:,la genes in E. coli and in a first set of experiments we used crude extracts of CyaA proteins to assess the invasive adenylate cyclase and hemolytic activities. The results, summarized in Table I, show that CyaA produced alone in E. coli has no invasive or hemolytic activity and that the cyaC gene product is sufftcient to render the CyaA holotoxin invasive and hemolytic in the reconstructed system in E. coli. Coexpression of cyaB, D and E genes in tram to cyaA does not confer invasiveness and hemolytic activity upon the holotoxin, nor does it potentiate the activities brought about by the c_vaC gene product. The nature of the CyaC-mediated activation of the toxin is still unknown. However, the high degree of homology between CyaC and HlyC (Barry et al., 1991) known to activate E. coli x-hemolysin by post-translational modilication (Nicaud et al., 1985; Wagner et al., 1988) strongly sug-
cyclase (Cya) activityb
Hemolytic
Hemolytic activity’
context Input
Internalized
Ui’d
u/ml
Internalized
activity Cya activity
(Y0) ______~
B. pertussis
wild type
PSlO PS30
cyaA
PS40 .’ B. pertussis is strain h See Table I, footnote ‘ See Table I, footnote
cyaA,C,B,D,E cyaA,C 18323 (Pittman,
900
5.9
51
900
0.075
ND
8.64 -
1040 906
5.04 9.48
6 15
1.19 1.58
1984). E. coli strains
c. d; ND, not detectable.
PSlO, PS30, PS40 are described
in Table 1, footnote
a
23 tion by CyaC or to some unknown the cyaB, D, and E products. 60
2
u. 40
0
20
determinants
0
100
TIME
Fig. 3. Time-dependent cytes (5x lo8 cells/ml)
200
OF
hemolytic
toxins (900 u/ml adenylate
strains
activity
are described
(min)
of CyaA toxins. The purified
were incubated
and hemolytic
as described
300
INCUBATION
cyclase)
used for toxin production
B. pertusk)
within the CyaA toxin are involved in invasive
and hemolytic
0
determined
for the
invasive activity, although their hemolytic activity was considerably reduced. This suggests that distinct structural
L
was
by one of
decreased hemolytic activity. Our data confirm the results of Bellalou et al. (1990) that invasive and hemolytic activities of CyaA toxin can be separated. They showed that a truncated toxin, expressed in B. pertussis, was devoid of invasive activity, while still preserving 50% of wt hemolytic activity. Here we show that the CyaC-activated toxins produced in E. coli retained full
u,
is 2 5 p
modification
This could account
activity
previously
with sheep erythro-
(6% of erythrocyte
(Bellalou
and purification
lysis)
et al., 1990). The
activities.
The decreased
hemolytic
activity
of the toxin produced in E. coli could also be accounted for by a difference in the nature of the post-translational modifications taking place in the two organisms. It can, however, not be excluded that an additional factor, present in B. pertussis, is required to confer full hemolytic activity upon the toxin.
(PS 10, PS30, PS40,
in Tables I and II.
gests that the CyaC-mediated activation of CyaA is also due to post-translational modification. The fact that this modification is not lost during toxin purification from B. pertussis by a procedure including 8 M urea extraction or SDSPAGE separation (Hewlett et al., 1989) indicates that the modification is covalent. (d) Hemolytic activity of CyaA toxin produced in Escherichia coli To compare the invasive and hemolytic activities of CyaA toxins produced in E. coli and in B. pertussis we used purified proteins from both organisms (Fig. 2). The results are summarized in Table II. It is striking, that for similar invasive adenylate cyclase activities the CyaC-activated proteins produced in E. coli (strains PS30 and PS40) show five times lower hemolytic activity than the toxin produced in B. pertussis. The ratio of hemolytic vs. invasive activity did not change significantly with the purification of the proteins, indicating that the decreased hemolytic activity is not due to some inhibitory factor present in E. cob extracts, but rather reflects an intrinsic property of the toxins produced in E. coli. Two additional arguments support this interpretation: first, the time course of hemolysis is almost linear, regardless of the source of the toxins (Fig. 3) indicating that the stabilities of the proteins purified from E. coli and B. pertussis are similar. Second, some possible differences in the initial conformation of the toxins produced in the two organisms, could be ruled out because the purification procedures involved complete denaturation in 8 M urea. It is, however, possible that the aggregation observed in E. coli might render the protein refractive to full modifica-
(e) Conclusions We have reconstructed in E. coli an expression system, consisting of all known genes of the cya locus of B. pertussis, which enabled production of active CyaA toxin endowed with catalytic, invasive and hemolytic activities. This system should allow us to uncover the mechanism of activation of CyaA by the cyaCgene product and to identify the structural domains of the toxin involved in the invasive and hemolytic activities, without requiring time-consuming genetic manipulation in B. pertussis. We are currently carrying out studies to identify additional factors required to reconstruct a functional system for secretion of CyaA toxin from E. cob.
ACKNOWLEDGEMENTS
We thank D. Ladant for the gift of purified toxin and for helpful advice, E. Krin for performing the site-directed mutagenesis, J. Pidoux for technical assistance, S. Goyard, J. Bellalou and A. Danchin for stimulating discussions, S. Cole for critical reading of the manuscript, J. Lortholary for graphic work and M. Ferrand for secretarial help. This work was supported by the Institut Pasteur and the Centre National de la Recherche Scientifique (URA 1129). P.S. was financed by a fellowship from the Minis&e de la Recherche et de la Technologie.
REFERENCES Barry, E.M., Weiss, A.A., Ehrmann,
I.E., Gray, M.C., Hewlett,
E.L. and
St. Mary Goodwin, M.: Borderella pertussis adenylate cyclase toxin and hemolytic activities require a second gene, cyaC, for activation, J. Bacterial.
173 (1991) 720-726.
24 Bellalou,
I., Sakamoto,
H., Ladant,
D., Geoffroy,
C. and Ullmann,
A.:
Deletions affecting hemolytic and toxin activities of Bordetella pertussis adenylate cyclase. Infect, Immun. 58 (1990) 3242-3247. Gilles, A.M., Munier, Pellecuer,
H., Rose, T., Glaser,
C. and BLrzu,
Bordetella pertussis adenylate Biochemistry Glaser,
P., Krin, E., Danchin,
0.: Intrinsic cyclase
fluorescence
A.,
of a truncated
29 (1990) 8126-8130.
P., Ladant,
D., Sezer, O., Pichot,
A.: The calmodulin-sensitive
F., Ullmann,
A. and Danchin,
cyclase of Bordetella pertussis:
adenylate
in Escherichia coli. Mol. Microbial.
cloning and expression
2 (1988a)
19-30. Glaser,
H., Bellalou,
J., Ullmann,
Secretion
of cyclolysin,
hemolysin
of Bordetella pertussis. EMBO
binding
mutagenesis.
E.L., Gordon,
adenylate
J. 7 (1988b)
A., Krin, E., Ladant, of residues essential EMBO
A.:
cyclase-
3997-4004.
Pittman,
M.: Genus
I.B.: Characteriza-
and secretion
Bordetella Moreno-Lopez J.G. (Eds.),
Vol. I. teriology, pp. 388-393. Rogel, A., Schultz,
of hemolysin
characterization
In: Krieg,
1952 178”.
Bergey’s
Manual
of Systematic
Bac-
and
Wilkins,
Baltimore.
1984,
Williams
J.E., Brownlie,
R.M., Coote, J.G.G.,
E.: Bordetella pertussis adenylate
Hanski,
cyclase:
of the toxic form of the enzyme.
cyclase
by
.I. 8 (1989) 967-972.
V.M., McCaffery,
J.D., Sutherland,
Sambrook,
J., Fritsch,
Laboratory
St. Mary Goodwin,
Parton,
R. and
purification
and
EMBO J. 8 (1989)
for colonization
BordeteNa pertussis in Takeshita,
S., Sato, M.,Toba,
of the holotoxin
lacZa-complementation
J. Biol. Chem. 264
(1989) 19379-19384.
resistance
B&u,
S., Sarfati,
0.: Characterization
catalytic
domains
R., Gilles, A.M., Predeleanu, of the calmodulin-binding
of Bordetella pertussis adenylate
R. and
and of the
cyclase.
J. Biol.
of structural
the head of bacteriophage Miller, J.H.: Experiments Laboratory,
proteins
T4. Nature
in Molecular
Cold Spring Harbor,
Taylor,
during
the assembly
of
227 (1970) 680-685. Genetics.
NY, 1972.
cyclase toxin IS critical
toxin is critical
infant
mice.
for lethal infection
Infect.
Immun.
Cold Spring
Harbor
M., Masahashi,
58
by
(1990)
and
selection.
thioate-modified W., Kuhn,
hemolysin
W. and Hashimoto-Gotoh,
and low-copy-number
plasmid
chloramphenicol-
or
vectors
for
kanamycin-
Gene 61 (1987) 63-74.
J.W., Ott, J. and Eckstein,
nuclcotide-directed Wagner,
Chem. 264 (1989) 4015-4020. U.K.: Cleavage
A
Press.
3445-3441. W.M. and
cation
D., Michelson,
Cloning.
Laboratory
NY, 1989.
and pertussis
T.: High-copy-number
molecule.
T.: Molecular
2nd ed. Cold Spring Harbor
M. and Weiss, A.: Adenylate
cyclase toxin from Bordetellapertussis: identili-
and purification
E.F. and Maniatis,
Manual,
Gray, M.C.: Adenylate
Laemmli,
L. and Holland,
of activation
from E. coli 2001. FEBS Lett. 187 (1985) 339-344.
Cold Spring Harbor,
D., Bbrzu, 0. and for catalysis and
in Bordetella pertussis adenylate
of calmodulin
site-directed
A. and Danchin,
the calmodulin-sensitive
Glaser, P., Elmaoglou-Lazaridou, Danchin, A.: Identification
Ladant,
N., Gray,
2755-2760.
P., Sakamoto,
Hewlett,
J.M., Mackman,
tion of HlyC and mechanism
N.R. and Holt,
in Escherichia coli.
expressed
Nicaud,
mutations
F.: The rapid
DNA. Nucleic M. and Goebel,
Annu. Rev. Microbial.
of oligo-
using phosphoro-
Acids Res. 13 (19X5) 8765-8785. W.: Active and inactive
(HlyA) from Escherichia coli. Hoppe
369 (1988) 39-46. Weiss, A.A. and Hewlett,
generation
at high frequency
E.L.: Virulence
factors
40 (1986) 661-686.
forms of
Seyler’s. Biol. Chem. of BordeieNa pertussis.