Bordetella pertussis Produces. Ion Conductance across Artificial. Lipid Bilayers in a Calcium- and. Polarity-dependent Manner*. (Received for publication, May ...
Communication
T H E JOWNAL OF BIOL~CICAL CHEMISTRY Vol. 269, No.36, Issue of September 9, pp. 22496-22499, 1994
0 1994 hy The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.
cy&, has been cloned and sequenced by Glaser et al. (11, 12), Adenylate Cyclase Toxin from who noted sequence homology with the gene for Escherichia Bordetella pertussisProduces coli hemolysin, hZyA. In addition, Barryet al. (13)discovered a gene ( c y a C )that is requiredfor production of the active formof Ion Conductance across Artificial AC toxin. Purified AC toxin from an organism witha mutation Lipid Bilayers in a Calcium- and in cyaC (BPDE386) possesses enzyme, but no toxin or hemolytic activity(14). This gene has sequence homology with a gene Polarity-dependent Manner* to catalyzethe in E. coli (hZyC)encoding aprotein that appears
post-translational acylation of the hemolysin (15, 16). Thesignificance of these similarities identified on the basis of amino acid sequence analysis is supported by immunologic cross-reGabor SzaboS, Mary C. Gray§, and activity between AC toxin and E . coli hemolysin (5). Erik L. Hewlett§nll AC toxin is a calcium-binding protein, and interactionof the From the Departmentsof $Molecular Physiology and toxin molecule with calcium results in a major conformational Biological Physics, §Medicine, and nPharmacology, change (17). This calcium-inducedconformational change is Uniuersity of Virginia School of Medicine, probably necessary but not sufficient for insertion and toxin Charlottesuille, Virginia 22908 delivery, since a mutant AC toxin (BPDE386), demonstrating Adenylate cyclase toxin (AC toxin) from Bordetella no measurable toxin or hemolytic activity, undergoes the conpertussis enters target cells to produce supraphysiologic formational change(14). However, both the delivery of the catalevels of CAMPand, by a CAMP-independent process, is lytic domain and the process that results in the hemolytic event hemolytic. In the present study, we show for the first are calcium-dependent (5, 17-21). time thatthis toxin also produces ion-permeable,cationThere are at least three fundamental issues that remain selective pores in phospholipid bilayers. The resulting unresolved concerning AC toxin function: 1)the mechanismby membrane conductance is absolutely calcium-depend- which the toxin causes hemolysis, 2) the mechanism by which ent, as are the intoxication and hemolytic activities. It is the toxin inserts into thecell membrane to deliver its catalytic strongly affected bythe polarity and magnitude of the domain to the cell interior, and3) whether these two activities membrane potential and enhanced by the presence of reflect the same or different processes (14). Studies demonnegatively charged phospholipid. AC toxins from two mutants, BPDE386 and BPD377, which are defectivein strating osmotic protection against hemolysis (5) suggest that lysis, but toxin activity, producelittle or no conductance. Finally, the toxin may produce a pore resulting in erythrocyte there were, until now, no other data to support this hypothesis. evaluation of the current-voltage relationships andthe In the present study, we have addressed the first question concentration dependenceof pore formation andof hemolysis reveal a greater than 3rd power dependence, defined above by employing a phospholipid bilayer system to suggesting that a multimer of AC toxin, probably con- determine whether AC toxin is capable of producing an ionsisting of three or more holotoxin molecules, is involved conducting pathway across the membrane. The results indicate that conductance that is characteristic of such a transmemin pore formation. brane pore is demonstrable with AC toxin, but only in the presence of calcium. Furthermore, the conductance is much greater when the bilayer contains negatively charged phosphoAdenylate cyclase toxin (AC’ toxin) is a single polypeptide lipid. There is also voltage and polarity dependence to this molecule produced by Bordetella pertussis and present on the activity, and the stoichiometric data suggestthat these events bacterium in a predominantly extracytoplasmic location (1). are mediated by an oligomer of toxin molecules. This unusualtoxin has the ability to enter target cells to catalyze the production of supraphysiologic levels ofCAMP. The EXPERIMENTALPROCEDURES consequence of this intoxication is suppressionof normal funcMaterials-Phenyl-Sepharose CL-4B and calmodulin-Sepharose 4B tions carried outby phagocytes, such as phagocytosis, develop- were obtained fromPharmacia Biotech Inc. Dulbecco’s phosphate-buffment of an oxidative burst, and killing of bacteria (2-41, and ered salinewas obtained from Life Technologies, Inc. Diphytanoyl phosphatidylcholine,phosphatidylserine,andphosphatidylethanolamine this is hypothesized to be its role in clinical pertussis. AC toxin also has the ability to lyse erythrocytes (5-8) and were purchased from Avanti Polar Lipids, Inc. (Albaster,AL). HEPES (San Diego, CAI. Allother reagentswere elicit marker releasefrom multilamellar liposomes (9). Weiss et was obtained from Calbiochem obtained from Sigma, unless otherwise indicated. al. (10) showed that transposon mutagenesis resultingloss in of Purification of Adenylate Cyclase Toxin-Wildtype B. pertussis, hemolytic activity of B. pertussis was associated with loss of strain BP338, strain BPDE386, containinga 4-base pairinsert in cyaC adenylate cyclase activity. The structural gene for AC toxin, (13), and BPD377, with a 157-amino acid deletion in a hydrophobic region of the structural gene (131, were grown as described previously (22).AC toxin was purified by urea extraction, phenyl-Sepharose chro* The costs of publication of this article were defrayed in part by the matography, sucrose density gradient centrifugation, and calmodulinpayment of page charges. Thisarticle must thereforebe hereby marked Sepharose chromatography (14).Analysis of SDS-polyacrylamide gels “aduertisement”inaccordancewith18U.S.C.Section1734solelyto by laser densitometry reveals that AC toxin purified by this method is indicate this fact. 11 To whom correspondence should be addressed: Universityof Virginia > 85% pure(14,22). AC toxin, storedat -70 “C in 10 mM Tricine, 0.5 mM Health Sciences Center, Box 419, Charlottesville,VA 22908. Tel.: 804- EDTA, 0.5 mM EGTA, 8 M urea, pH 8.0, was used for bilayer experiments. The residual urea (200-400PM) present in this preparation did 924-5945; Fax: 804-982-3830. notaffect the elicitedmembraneconductance.Enzymeactivitywas The abbreviations used are: AC, adenylate cyclase; Tricine, in a cell-free measured by the conversion of [a-32PlATP to [32PlcAMP h”tris(hydroxymethy1)methylglycine; PC, phosphatidylcholine; PS, assay (14).The CAMP formed was isolatedas described by Salomon et phosphatidylserine; PE, phosphatidylethanolamine. (Received for publication, May 23, 1994, and in revised form, June 29, 1994)
22496
22497
Adenylate Cyclase Toxin Pore Formation
A
CaZ' (2.5 mM. as)
02
00
I
-50mV lo
III
8
1
t
6
I
4 2 -20
0
20
40
60
Time (s)
FIG.1.Calcium is required for the membrane current elicited by wild type AC toxin. The bathing solution in both sides of the chamber contained 20 mM Tris pH 7.4, 200 mM KCI. After the bilayer was formed, wild type toxin (4 pg/ml) was added to both sides of the chamber and incubated for 10 min while monitoring the membrane current with 50 mV applied to the cis compartment. At t = 0 calcium (2.5 mM) was added t o the cis side while stirring. The addition and presence of calcium is indicated by a solid line above the tracing. The membrane current was suppressedon changing the polarityof the applied potential (-50 mV, indicated by solid line). Positive deflections indicate cis to trans currents.
f
-4
1
B '0 1
c
al. (23). Enzyme specific activities for BP338, BPDE386, and BPD377 ranged from 0.8 to 1.2 mmol of cAMP/min/mg of toxin. Toxin activity was determinedby quantitation of intracellular CAMPaccumulation in Jurkat cells exposed to AC toxin for 30 min at 37 "C (14).Toxin specific activityforBP338was 12-15 pmol of cAMP/mg of Jurkat cell proteinimg of toxin;neitherBPDE386orBPD377elicited CAMP accumulation. Hemolytic Activity of Adenylate Cyclase Tbxin-Purified AC toxin -60 4 0 -20 20 40 8060 to dialyzed intoDulbecco's phosphate-buffered saline, pH 7.4, was used V (mv) measure hemolytic activity, as described previously (24), with minor -2 modifications (5). Sheep erythrocytes were incubated with AC toxin for FIG.2. The current elicited by wild type AC toxin is voltage3.5 h a t 37 "C, and hemoglobin absorbance was measured at 540 nm and polarity-dependent. The bathing solution in both sides of the using a multiscan spectrophotometer. chamber contained 10 mM Tricine, pH8.0,lOO mM NaCl, 20 mM KC], 0.5 Planar Bilayer-Bilayer experiments were performed as described previously (25) with minor modifications, usinga Teflon chamber with mM EDTA, 0.5 mM EGTA, 2.5 mM CaC1,. Panel A, bilayer was formed and wild type toxin (3.3 pg/ml) was added to both sides of the chamber a 1.5-ml reservoir on each side of the bilayer. Unless otherwise indiand stirred. After15 min, indicated voltages were applied and the rate cated, bilayers were prepared from PS:diphytanoyl PC (1:l) indecane of increase of the current wasmeasured. The membranepotential was (20 mg/ml). The lipid bilayer was prepared by drawing a film of the held at 0 mV between pulses. The solid line shows the fit to Equation 1 lipids across a 0.5-mm aperture using a n air bubble at the tip of a with parameters given under "Results Discussion." and Panel B , bilayer freshly cleaned Pasteur pipette. Experiments were performed a t room was formed and wild type toxin (3.3 pg/ml) was addedto the cis sideof temperature using two Ag-AgC1 electrodes that madeelectrical contact the chamber and stirred. Approximately 20 min later, indicated voltages with the solutions in the chamber. In the ion-selectivity experiments werethe applied and the rate of increase of the membrane current measAg-AgC1 electrodes wereplaced in 1M KC1 agar bridges. Bath solutions ured. The solid line is fit a of Equation 1with c = 0, as indicated under were made with reagent grade chemicals and filtered with a detergent"Results and Discussion." free membrane (Type HA, Millipore). The trans side was held at a virtual ground (reference), while potentials were applied tocisthe side. Membrane currents were recorded using a custom-designed current that the pore created by AC toxin does not remain functional amplifier using a fast, low noise, high input impedance amplifier and a when thepolarity is reversed and second, that theAC toxin on 1-gigaohm feedback resistor. The output was amplified and filtered at the trans side of the membrane cannot function to support 300 Hz (Bessel, 48 dB/octave). Ion selectivitywasdetermined from zero-current potential measurements in asymmetric ionic solutions us- conductance, even with the proper polarity, because of the abing the Goldman-Hodgkin-Katz constant field formalism (26) and are sence of calcium in that compartment. Conductance develops with a relatively slow time course afexpressed as permeability ratios.
J
RESULTS AND DISCUSSION
AC toxin-induced intoxication and hemolysis are dependent on the presence of extracellular calcium (5,17-21). As shown in Fig. 1, AC toxin is able to produce transmembrane conductance in thelipid bilayer system, butonly in thepresence of calcium. In this experiment, AC toxin is present on both sides of the bilayer, initially without calcium in either chamber. Under these conditions, there is no conductance even after extended incubation (20 min or more) with thetoxin. When calcium (2.5 mM) is added to the cis side of the membrane (presence of calcium indicated by solid line above tracing), conductance develops within 30 s but is polarity-dependent, as illustrated by the loss of conductance when thepolarity is reversed by making the cis side negative. These data illustrate two features: first,
ter the application of calcium. When AC toxin is added in the presence of calcium, conductance also develops slowly. The inordinately long times that would be required to reach steady state, especially at low AC toxin concentrations, together with considerations of membrane stability, led us t o use the rate of development of membrane current (AIlAt) as a convenient and reproducible measure of AC toxin-induced pore formation. With AC toxin and calcium on both sides of the bilayer, conductance develops in eitherdirection, in anapproximately symmetrical manner with respect to membranepotential. Furthermore, the conductance is stronglyvoltage-dependent,with increasing values as the voltage is raisedincrementally. Thisis illustrated by the activation-voltage relationship of Fig. 2A that can be fit by Equation 1, with (AZ/AtI0= 0.048 PA, c = 0.66, and p = 0.088 when V is expressed in mV.
22498
Adenylate Cyclase Toxin Pore Formation AUAt = (AI/At)o (exp(pV)- c exp(-pV))(l)
(Eq. 1)
A
A reasonable interpretationis that the first term represents ’1 ion flow through cis-inserted channels while the second term corresponds t o trans-inserted channels with c indicating the trandcis activity ratio. If this interpretation is correct, then c should benear zero when AC toxin is omittedon the transside. That this is indeed observed is shown in Fig. 2B where essentially no conductancedevelops when thetoxin-free trans side is made positive. This resultis corroborated by the excellent fitof the theoretical line drawn to Equation 1with (AZ/At)o= 0.030, p = 0.073, and c = 0. The presence of calcium and the resultant calcium-induced conformational change are necessary but not sufficient for production of the ion-permeable pore, as illustrated by studies 0 I using two mutants ofAC toxin that possess enzyme activity I 10 100 and undergo the calcium-dependentconformational change, AC TOXIN (pg/ml) but demonstrate no toxin or hemolytic activity (Ref. 14, and data not shown). BPDE386 contains a insert in cyaC and thus is unable to carry out the post-translational activationof the B structural protein of AC toxin (CyaA) (13). This modification, / which has recently been demonstrated to consist of acylation of an internallysine residue, is absent from AC toxin producedby BPDE386 (27). In thelipid bilayer,AC toxin from BPDE386,a t twice the concentration of wild type, elicits only a small conductance (
