Characterization of adenylate cyclase toxin from a mutant of Bordetella ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Bioehemistry and Molecular Biology, Inc.

VOl. 268, No. 11,Iesue of April 15, pp. 7842-7848,1993 Printed in U.S.A.

Characterization ofAdenylate Cyclase Toxin from a Mutant of Bordetella pertussis Defective inthe Activator Gene, cy&* (Received for publication, November 6, 1992)

Erik L. Hewlett$§1[, Mary C. Gray#., IngridE. EhrmannSO, Nancy J. Maloney$$, Angela S. Oteroll, Lloyd Gray**, Margaretta Allietta**, Gabor Szaboll, Alison A. Weiss#.#.,and Eileen M. Barry#.$ From the Departments of $Medicine, §Pharmacology, **Pathology, and 11 Physiology, University of Virginia School of Medicine,

of Microbiology and Immunology, Medical College of Virginia, Churbttesvilk. Virginia 22908 and the iiDeDartment .. Richmond, Virginil23298

Bordetellapertussis adenylate cyclase (AC)toxin has Adenylate cyclase (AC)’ toxin from Bordetellu pertussis is the abilities to 1)enter target cells where it catalyzes a novel moleculewith the abilities to 1)enter targeteukaryotic cyclic AMP production and 2) lyse sheep erythrocytes,cells to catalyze the production of supraphysiologic levels of and these abilities require post-translational modifi- CAMP,2) lyse erythrocytes and possibly other cell types, and cation by the product of an accessory genecyaC (Barry, 3) elicit marker release from multilamellar liposomes (1-6). E. M., Weiss, A. A., Ehrmann, E. E., Gray, M. C., A single protein with the molecular mass of 177 kDa by Hewlett, E.L., and Goodwin, M. St. M. (1991) J. derived amino acid sequence, but apparent mass of 200-216 Bacteriol. 173, 720-726). In the present study, AC kDa by SDS-PAGE, is responsible for these activities (7-10). toxin has been purified from a n organism with a mu- The gene for AC toxin, cyaA, has been cloned and sequenced tation in cyaC, BPDE386, and evaluated forits phys- by Glaser et al. ( l l ) , who noted sequence homology with the ical and functional properties in orderto determine the gene for Escherichia coli hemolysin, hlyA. Subsequent analysis basis for its lack of toxin andhemolytic activities. AC revealed additional downstream genes cyaB, cyaD, and cyaE, toxin from BPDE386 is indistinguishable from wildtwo of which (cyaB and cyaD) are homologous with postulated type toxin in enzymatic activity, migration on SDSsecretion genes (hlyB and hlyD) in E. coli (12). Recently, polyacrylamide gel electrophoresis, ability to bind calcium, and calcium-dependent conformational change. Barry et al. (13) discovered another gene (cyaC) which is Although unableto elicitcAMP accumulation,AC toxin required for production of the active form ofAC toxin. Exfrom BPDE386 exhibits binding to the surface of J u r - tracts from organisms with mutations in the cyaC gene conkat cells whichis comparable to that of wild-type toxin. tained an AC toxin with enzyme activity but no demonstrable This target cell interaction is qualitatively different, toxin or hemolytic activities. The cyaC gene has sequence homology with a gene, hlyC, in E. coli which apparently however, in that 99% of the mutant toxin remains sensitive to trypsin, whereas -20% of cell-associated catalyzes a post-translational activation of the hemolysin, HlyA (14,15). Issartel et al. (16) have demonstrated that an wild-typetoxin enters a trypsin-resistant compartacyl carrierprotein is required for this activity and that ment. To evaluate the ability of this mutant AC toxin to activation is associated with transfer of an acyl group to HlyA. function at its intracellular site of action, the CAMP- There are no data, however, to confirm that acylation is the stimulated L-type calcium current in frogatrial myo- mechanism for activation of AC toxin by CyaC. Characterization of purified AC toxin from wild-type orgacytes was used. Extracellular addition of wild-type toxin results in CAMP-dependent events that include nisms has shown that it is a calcium-binding protein which activation of calcium channels and enhancement of undergoes a conformation change as aresult of its interaction calcium current. In contrast, there is no response to with calcium (17). The calcium-bound form exhibits a shift externally applied toxin from BPDE386. When in- in intrinsic fluorescence, a change in reactivity with monojected into the cell interior, however, the AC toxin clonal antibody 1H6, an alteration in trypsin sensitivity, and from BPDE386 is able to produce increases in the an apparentuncoiling of the molecule as visualized by electron calcium current comparable to those observed with microscopy (18). wild-type toxin. AlthoughAC toxin from BPDE386is In the present work, AC toxin from BPDE386, containing unaffected in its enzymatic activity, calcium binding, an insertional mutation in cyaC, was purified and characterand calcium-dependentconformationalchange, the ized to determine its structural and functional properties. The mutation incyaC does result ina toxin whichis able to molecule is indistinguishable from wild-type (BP338) AC bind to targetcells but unable to elicitcAMP accumu- toxin by SDS-PAGE. It binds calcium, exhibits comparable lation. In that AC toxin from BPDE386 is able to AC enzyme activity, undergoes the calcium-dependent confunction normally when injected artificially to an in- formational change, and binds to target cells as well as wildtracellular site, we conclude that the disruption of cyaC type toxin. It is, however, nonhemolytic and possesses essenproduces a defect in insertionand transmembranede- tially no AC toxin activity. Most importantly, toxin activity livery of the catalyticdomain. can be demonstrated by injection of the mutant toxin into * The costs of publication of this article were defrayed in part by target cells, indicating that itsdefect is in membrane insertion the payment of page charges. This article must therefore be hereby and delivery of the catalytic domain to the cell interior, not marked “advertisement” in accordance with 18 U.S.C. Section 1734 in its ability to function at theintracellular site. solely to indicate this fact. 7 To whom correspondence should be addressed: Box 419 Health Sciences Center, University of Virginia School of Medicine, Charlottesville, VA 22908. Tel.: 804-924-5945; Fax: 804-982-3830.

’

The abbreviations used are: AC toxin, adenylate cyclase toxin; PAGE, polyacrylamide gel electrophoresis; Tricine, N-[2-hyroxy-l,lbis(hydroxymethyl)ethyl]glycine.

Characterization of Mutant Adenylute EXPERIMENTALPROCEDURES

Materials-Bordet-Gengou medium, Stainer-Scholte medium, Hanks' balanced salt solution, fetal bovine serum, RPMI 1640, and penicillin-streptomycin were from GIBCO.Phenyl-Sepharose CL-4B and calmodulin-Sepharose 4B were obtained from Pharmacia LKB Biotechnology Inc. ' T a was from Du Pont-New England Nuclear. Polyvinylidene difluoride membranes were obtained from Millipore Corp. (Bedford, MA). All other reagents were obtained from Sigma, unless otherwise indicated. Culture of Organisms and Purification of Adenylate Cyclase Toxins-Wild-type B. pertussis, strain BP338, and strain BPDE386, containing a 4-base pair insertion in the cyaC gene (13), were grown as described previously (7,18). AC toxin was purified by urea extraction, phenyl-Sepharose chromatography, sucrose density gradient centrifugation, and calmodulin-sepharose affinity chromatography as described elsewhere (18). Sample Preparation and Determination of Free Calcium-Purified AC toxin was dialyzed against 10 mM Tricine, 2 mM EDTA, 3 mM MgC12, pH 8 (Buffer A) prior to use in assays. Free calcium concentrations were calculated using the EGTA computer program (19) with substitution of equilibrium constants for EDTA found in Critical Stability Constants (20). Adenylate Cyclase Enzyme Activity-AC activity was measured by to [32P]cAMPas described previously (7, the conversion of [3ZP]ATP 21). Briefly, each assay tube contained 60 mM Tricine, 10 mM MgC12, 1mM ATP with 2 X lo5cpm of [cx-~~PIATP) and 1p~ calmodulin at pH 8.0. The reaction was carried out at 30"Cfor 10 min and terminated by addition of 100 plof a solution containing 1% SDS, 20 mM ATP, and 6.24 mM cAMP (with 1.5 X lo' to 2.0 X lo' cpm of [3H]cAMP). Cyclic AMP was quantitated by the double column method of Salomon et al. (22). Adenylute Cyclase Toxin Activity-Adenylate cyclase toxin activity was determined by quantitation of intracellular cAMP accumulation in Jurkat cells (a human T-helper lymphocyte line). Jurkat cells were maintained a t 37 "C in RPMI 1640 medium supplemented with 15% heat-inactivated fetal bovine serum, 2 mM glutamine, 50 units/ml penicillin, and 50 pg/ml streptomycin in 5% Con.Immediately before assay, 1 X lo6 Jurkat cells/ml were washed three times in Hepes buffer, pH 7.4, containing 140 mM NaCl, 5 mM KCl, 1%glucose, 3 mMMgC12, 2 mM EDTA (Buffer B). AC toxin was then added and the cells incubated for 30 min a t 37 "C. The assay was terminated by washing the cells three times in Buffer B with or without added calcium, as indicated. Intracellular cAMP was extracted with 0.1 N HCl for 30 min at room temperature and measured by radioimmunoassay (23). Protein was extracted with 0.2 N NaOH and quantitated using the method ofLowry et al. (24). Determination of the halfmaximal calcium concentration for toxin activity was determined using nonlinear regression analysis of a sigmoid curve by InPlot (Graphpad). Hemolysis Assay-Hemolysis assays were performed according to Clerc et al. (25) with minor modifications (5). Fifty microliters of washed sheep red cells (Cocalico Biologicals, Reamstown, PA) a t a density of 4 X 10s/mlwere added to each wellof a 96-well flatbottomed microtiter plate followed by addition of 50 pl of sample. After incubation at 37 'C for the indicated time period, the cells were resuspended, and 150 pl of cold PBS were added to each well. After 15 min of centrifugation a t 4 "C and 2200 X g, 100 pl weretransferred from each well to another plate. Hemoglobin absorbance was measured a t 540 nm in a Multiscan Spectrophotometer (Flow Laboratories). Specific hemolysis was determined by: (hemoglobin release by sample - hemoglobin release by PBS)/(total hemoglobuin released by NHIOH) and was expressed as percent of total hemolysis. CalciumBinding-AC toxin was separated on a 5-20% SDSpolyacrylamide gel (26) and transferred to polyvinylidene difluoride membrane for 1.5 h a t 0.6 A. The membrane was incubated with 1 pCi/ml 'Ca and placed on x-ray film for 24 h as described previously (27). Intrinsic Tryptophan Fluorescence-The fluorescence emission spectrum ofAC toxin was measured as described previously (18). Briefly, AC toxin inaquartz cuvette was placed in the sample chamber maintained a t 37 "C of an SLM 8000 spectrofluorimeter (SLM/Aminco, Urbana, IL). Excitation wavelength was 280 nm, and the emission spectrum over the range of 310-370 nm in 1-nm increments was recorded. Bandpass was set a t 4 nm for both excitation and emission monochrometers. From the emission spectrum, the ratio of emission intensities (330 nm/350 nm) was determined. Trypsin Treatment of Cell-associated Adenylate CyclaseToxin-

Cyclase Toxin

7843

Jurkat cells were washed three times in Buffer B andresuspended at 1 X lo6 cells/ml in the same buffer. Calcium was added as indicated to achieve a final concentration of 1 mM.AC toxin was added and incubated for 30 min a t 37 "C. Cells were washedthree times, treated with 400 pg/ml acetylated trypsin for 10 min a t 10 "C,and the reaction stopped with 800 pg/ml lima bean trypsin inhibitor. Control samples were treated with 800 pg/ml trypsin inhibitor and then 400 pg/ml trypsin. Cells were washed two times and the pellet resuspended in 50 pl of Hanks' balanced salt solution. Cell pellets were frozen at -70 "C overnight, and adenylate cyclase enzyme activity was measured as described above. Comparable results were obtained when cells werefrozen and thawed twice in liquid nitrogen and AC enzyme activity assayed immediately. ElectrophysiologicStudies-Cellswere dissociated enzymatically from bullfrog atria as described elsewhere (28), and the calcium current (IC.) was recorded using the whole cell configuration of the patch clamp technique (29). Every 1.5 s, cells were depolarized to -5 mV for 250 ms from a holding potential of -85 mV to elicit IC.. IC. was measured as the peak inward current using zero currentas reference and the individually measured currents plotted as afunction of time. Each experiment was performed at 22 "C, and sodium currents were blockedwith tetrodotoxin. The external solution contained 90 mM NaC1,2O mM Hepes, 5 mM MgCl,, 2.5 mM KCl, 2.5 mM CaCh, 3 p~ tetrodotoxin, pH 7.4. The intracellular medium contained 80 mM potassium aspartate, 30 mM KCl, 5 mM Hepes, 1 mM EGTA, 2.5 mM ATP, 0.2 mM GTP, pH 7.4. In some experiments recordings were obtained using K+-free,(%+-containingsolutions (internally: 120 mM CsCl, 1 mM EGTA, 5 mM Hepes, 2.5 mM ATP, 0.2 mM GTP, pH 7.4; externally: 85 mM NaCl, 20 mM Cscl, 10 mM Hepes, 2 mMMgC12, 2.0mM CaC12,3 p~ tetrodotoxin, pH 7.4) to block K' currents, with similar results. Isoproterenol stock solutions contained 1mM ascorbic acid. Bath solutions were exchanged with the aid of a fast flow system. For extracellular applications of toxins, the flow of external solution was halted, and thetoxin preparation was added to thebath; after 36 min the flow was resumed. Intracellular application of toxins was achieved either by passive diffusion or by pressure injection (30). RESULTS

Purification of Mutant Toxin from BPDE386"Initial studies to determine the phenotypesresulting from mutations upstream from c y d were conducted with dialyzed urea extracts of the different organisms (13). Under those conditions, extract from BPDE386 was comparable to that from wildtype (BP338) in enzyme-specific activity but was nonhemolytic and possessed no measurable toxin activity in5774 cells. The AC toxin from BPDE386 has now been purified by the procedure used previously for wild-type toxin, consisting of urea extraction, phenyl-Sepharose chromatography, preparative sucrose gradient centrifugation,and a final affinity chromatography step with calmodulin-Sepharose (7). The band at apparent mass 216 kDarepresenting AC holotoxin from BPDE386 is indistinguishable from that of wild-type toxin (Fig. l), and theprotein has comparable or higher enzymatic 200 -

"

116 -

9266 45 -

31

-

21 14

-

A

B

FIG. 1. Purified AC toxin from BPDE386 has an apparent molecular mass of 216 kDa and is indistinguishable from purified toxin from BP338. Lane A,AC toxin from BP338 (12.5 pg), and lane B,AC toxin from BPDE386 (12.5 pg), were separated on a 5-20% SDS-polyacrylamide gel and stained withCoomassie Blue.

Characterization of Mutant Adenylate

7844

Cyclase Toxin

TABLEI Enzyme, toxin,and hemolytic activities of AC toxin purified from BP338 and BPDE386 AC enzyme activity“ activityb

AC toxin from

AC toxin

200

Hemo,ysisc

96.

%

BP338 (wild type) 1600 11.0 63 BPDE386 (cyaC mutant) 1900 0.0003 0 ‘Adenylate cyclase enzyme activity is expressed as micromoles of cAMP/min/mg toxin. bAdenylate cyclase toxin activity is expressed as micromoles of cAMP/mg Jurkat cell protein/mg toxin. Hemolysis is expressed as percent hemoglobin release with 3.5 pg toxin incubated with 2 X 10’ red blood cells for 18 h a t 37°C.

-.5 a

40,000-

ea 35.000-

Y

3

u w m

m

-

2

d

2 L

P

30.000-

u L

25.000-

+ 20,000-

x

Y a

5 3 10,000-

E 5.000E

0-

,001

B

C

AC toxin from BP388 (12.5 pg, 58 pmol); lane B, purified AC toxin from BPDE386 (12.5 pg, 58 pmol); and lane C, calmodulin (1 pg, 60 pmol) were separated on a 5-20% SDS-polyacrylamide gel, transferred to a polyvinylidene difluoride membrane, probed with 1 pCi/ ml ‘Ta, and used to expose x-ray film for 24 h. h

.5 12,000

”

Y

8,000

/?

-

d 0

0

A

FIG. 3. AC toxin from BPDE386 binds ‘%a. Lane A, purified

h 10,000

15,000\

v

18 12 -

. ..-..

.o 1

0-*, /O

.1

1

10

AC Toxin (cg)

FIG. 2. Concentration dependence of intoxication by AC toxin from BPDE386 and BP338. Jurkat cells were washed and resuspended in Hanks’ balanced salt solution. AC toxin from BPDE386 (M and )BP338 (M were ) added a t indicated concentrations for 30 min a t 37 “C. Intracellular cAMP levels were determined as described under “Experimental Procedures.”

2,000 10-2

D 2 + 1 (MI FIG. 4. Effect of calcium concentration on intoxication by AC toxin from BPDE386 and BP338. Jurkat cells were washed three times in Buffer B and resuspended at 1 X 106/ml in the same buffer. Calcium was added to achieve the indicated free calcium concentration. Purified toxin (1.2 pg) from BPDE386 ( O ” 0 ) and were ) added and incubated a t 37 “C for 30 min. BP338 (M Intracellular cAMP was determined as described under “Experimental Procedures.”

activity (Table I). Thepurified toxin from BPDE386, likethe crude toxin, is strikingly different from wild-type toxin in its ability to increase cAMP levels in Jurkatcells and lyse sheep erythrocytes (Table I). The production of intracellular cAMP by wild-type toxin is concentration-dependent and exceeds 35,000 pmol cAMP/mg Jurkat cell protein at 30 min in the presence of 1mM calcium (Fig. 2). In contrast, AC toxin from trations of0.1-100 p~ (17, 18). In those studies,culture BPDE386 produces no increase in target cell cAMP under medium was removed from intoxicated cells and cAMP exthose conditions. tracted with 0.1 N HCl without washingthe cells. DemonstraInteraction of Mutant Toxin with Calcium-Wild-type AC tion by Gentile et al. (35) of the release of ATP from cells and toxin from BP338 is a calcium-binding protein (17, 18).This the accumulation of extracellularcAMPuponaddition of finding is not unexpected since the toxin requires calcium for enzymaticallyactive AC toxin led to the realization that its toxin and hemolytic activities, and it possesses 43 copies contamination of the tubeswith residual cAMP might account of a nonameric, glycine-rich repeat which is also present in for the previously observed results. As described under “Exand necessary for the calcium-bindingactivity of E. coli perimental Procedures,” the present studies include washing hemolysin, HlyA (11,31-34). As shown in the autoradiogram of cells three times afterremoval of incubation medium. This in Fig. 3 with calmodulin as a positive control,AC toxin from procedure results inelimination of “low level” intoxication by BPDE386 also binds calcium, indicating that the mutation inwild-type AC toxin at calcium concentrations below 100 p M cyaC does not affect this property of the molecule. and by AC toxin from BPDE386 at all but the highest conThe calcium dependence of the intoxicationprocessis centrations of calcium. illustrated in Fig. 4; no cAMP is produced by wild-type toxin Because AC toxin from BPDE386 binds calcium but does at free-calcium concentrations below 100 p ~ but not possess calcium-dependent toxin activity, the physical , a sharp increase in cAMP occurs with half-maximal response at 432 response of the molecule to calcium was investigated. Purified p M free calcium. The steepslope of the calcium concentration AC toxin from BPDE386 undergoes a calcium-induced condependence and the potentialfor more than one calcium site formational change as shown by a change in intrinsic fluores(18)suggest the possibility of cooperative effects from calcium cence ratio at calcium concentrations of 10 p M to 1 mM (Fig. binding, which is being explored currently. AC toxin from 5) and exhibits anuncoiled structure by electron microscopy BPDE386, on the other hand causes essentially no increase at 1 mM free calcium (data not shown). In the absence of in cAMP levels of Jurkat cells. These data are in contrast to calcium, AC toxin from BPDE386 is extremely sensitive to those reported earlier from this laboratory in which low level trypsin proteolysis with prompt degradation to peptides of intoxication (-10% of maximum) occurredat calcium concen- 55,000 kDa. Inthe presence of 1 mM free calcium, the

Characterization of Mutant Adenylate Cyclase Toxin calcium-induced conformational change, results, however, in a peptide of 60,000 kDa which is resistant to trypsin (data not shown). These properties areindistinguishable from those of wild-type AC toxin (Fig. 5 and Ref. 18). The lack of toxin activity by AC toxin from BPDE386 under these conditions, however, indicates that this calcium-induced conformational change, although it may be necessary, is not sufficient for intoxication and hemolysis by this molecule (18). Interaction of AC Toxins with Target Cells-Because of loss of biological activity during radiolabeling of AC toxin, there has been little work on the interaction of AC toxin with target cells.Cyclic AMP production in Chinese hamster ovary (CHO) cells can be blocked by incubation of AC toxin with a variety of gangliosides and phospholipids (3, 36). There is, however, no evidence that a specific receptor is required for intoxication. Therefore, the term “binding” is used here to reflect the toxin interaction with target Jurkat cells which remains after washing, but not to imply a classical ligandreceptor interaction. Using adenylate cyclase enzymatic activity to detect cell-associated AC toxin, it has been possible to address quantitatively and qualitatively the interaction of toxin with the target cell surface. In theexperiments described here, AC toxin from BP338 or BPDE386 was added to Jurkat cells in the absence or presence of 1 mM free calcium. After incubation for 30 min, the cells were washed three times and some of the cells then exposed to trypsin as described under “Experimental Procedures.” The amount of residual toxin was determined by measurement of AC enzymatic activity in 0.90

-

0.88 -

f . 3

LL

0.860.84-

LL

0.82 0.80 -Q -e”0

FIG.5. Intrinsic fluorescence of AC toxin from BPDE386 and BP338 in response to calcium concentration. Samples or 97 pg/ml of BP338 containing 82 pg/ml of BPDE386 (c”.) (M in) Buffer A with added calcium to achieve indicated free concentrations were scanned at 37 “C in an SLM 8000 spectrofluorimeter. Excitation was 280 nm and the ratio of emission intensity at 330 nm to that at 350 nm plotted.

7845

cells lysed by freeze-thawing. As shown in Table 11, both AC toxins bind to Jurkat cells in the absence and presence of calcium. In the case of wild-type toxin, cell-associated toxin in the absence of calcium represents 65-70% of that at 1 mM calcium. AC toxin from BPDE386 exhibits greater cell binding than wild type and a4-fold increase in the presence of calcium (compared to the absence of calcium). The important observation, however, is that when the cell interaction occurs without calcium, >98% of toxin (from either BP338 or BPDE336) is susceptible to degradation by trypsin. In the presence of 1 mM calcium, -20% ofAC toxin from BP338 has entered a trypsin-resistant compartment whereas >99% of toxin from BPDE386 is destroyed under the same conditions. This difference is not due to different trypsin sensitivities of the molecules because digestion by trypsin in the absence of target cells results in comparable losses of enzyme activity with or without calcium (data not shown). These data strongly suggest that although both wild-type and mutant toxins have a superficial interaction with the target cell, only wild-type toxin in the presence of calcium is able to undergo an additional step whichremoves it from the reach of a soluble, extracellular protease. Effect of Mutant AC Toxin ut Intracellular Site-The expression of wild-type enzyme activity by BPDE386 suggests, but does not establish, that this mutant toxin should be able to function within an appropriate target cell. To be certain that the mutation in cyaC affects only membrane insertion and delivery and not its capacity to catalyze intracellular cAMP production, AC toxins from BP338 and BPDE386 were compared for their abilities to elicit a CAMPdependent response in single cells from frog atrium. Incardiac myocytes, L-type calcium channels are modulated by CAMPdependent phosphorylation such that an increase in the intracellular concentration of cAMP results in an increase in the whole cell calcium current, IC.(37). The time course of the response of IC. to extracellular application of wild-type AC toxin is shown in Fig. 6A. The onset is rapid ( 4 min from toxin addition) and the calcium current increases 15fold in less than 10 min. The amplitude of ICa remains elevated upon washout of the toxin and there is no evidence of reversibility during the course of these experiments (30-45 min). The magnitude of the response is comparable to thatelicited by maximally effective concentrations of other stimuli. For example, superfusion of toxin-treated cells with the @-adrenergic agonist isoproterenol has no further effect on Ica(data not shown). In addition, the inwardly rectifying potassium currentandthe muscarinic-activated potassium current, which are not regulated by a CAMP-dependent process (38), are not affected by the toxin (data not shown). These data

TABLEI1 Differential twDsinsemitivitv ofAC toxin from BP338 and BPDE386 incubated with Jurkut celk in the Dresence or absenceof calcium AC toxin from:

BP338 (wild type) BP338 (wild type) BP338 (wild type) BP338 (wild type)

Condition

0 Ca2+

+

0 Ca2+ trypsin 1 mM Ca2+ 1 mM Ca2+ trypsin

+

AC enzyme activity“

AC enzyme activity remaining after trypsin treatmentb

pnwl cAMP/lO all10 min

%

776 10.1 1100 217

BPDE386 (cyaC mutation) 1630 0 Ca2+ BPDE386 (CyaC mutation) 0 Caz+ trypsin 24.2 BPDE386 (cyaC mutation) 4650 1 mM Ca2+ BPDE386 (cyaC mutation) 1 mM CaZ+ trypsin 21.2 ‘These data are from a representative experiment with conditions done in duplicate. Percentages represent mean f S.E. from several experiments. N = 8.

+

+

1.61 f 0.16 19.3 f 0.09 1.43 f 0.09 0.81 f 0.14

Characterization of Mutant Adenylate

7846

Cyclase Toxin

domain to the target cell interior, but not its activity at the intracellular site. DISCUSSION

Adenylate cyclase toxinfrom a mutant ofB. pertussis bearing a mutation in cyaC (BPDE386) has wild-type level enzymatic activity, but is inactive in producing intracellular ‘9 cAMP accumulation or hemolysis (Table I). The purpose of this work was to review the multiple-stepprocess involved in these activities of AC toxin (Fig. 7) and determine the step at which the mutant is defective. Toxin from BPDE386 is indisTime (min) tinguishable from wild-type toxin by SDS-PAGE (Fig. 1)and B it exhibits comparable calcium binding activity (Fig. 3). Al1000 BPDE386 though AC toxin from BPDE386 has enzyme activity comparable to wild type, it is unable to intoxicate Jurkat cells (Fig. 2 ) or hemolyze sheep erythrocytes (Table I). AC toxin 600 from BPDE386 appears to undergo a calcium-dependent conformational change (Fig. 5 ) equivalent to that of toxin from BP338 (18).This altered conformationin response tocalcium 200 was, however, unexpected since the conformational change was thought to benecessary for intoxication (18).These data 0 5 10 15 can be interpreted in one of two ways regarding the relationship of the conformational change to toxin and hemolytic Time (min) activities. The structural change which occurs as a function of calcium concentration could be an epi-phenomenon, neither required for nor involved in intoxication. Alternatively, these striking changesmay be essential, but notsufficient for intoxication andhemolysis. The relationshipsof calcium concentrations to structural changes and to functional activities (i.e. intoxication and hemolysis) favor the latter possibility (18).In either case, the activation mediatedby the product of cyaC is certainly necessary for intoxication and hemolysis, as 0 shown in Table I and illustrated diagrammaticallyin Fig. 7. 0 5 10 15 The next step in this sequence is the interaction ofAC Time (min) toxin with the targetcell (Fig. 7). Although we do not believe FIG. 6. External application of AC toxin from BP338, but that this process represents a classical receptor-ligand internot fromBPDE386 increases IC.in atrial myocytes; when AC action, we have chosento use interchangeablytheterms toxin from BPDE386 is injected into the cell, however, it “binding” and “cell association.” The data presented herein produces an increase in IC,. Panel A , I C . was recorded from a single atrial myocyte as described under “Experimental Procedures.” (Table 11) and elsewhere’ indicate that the targetcell associtoxin from ation of AC toxin is not impaired quantitatively in the absence The bar indicates the period during which 1 pgofAC BP338 was present in the bath. Zero time refers to the point where of cyaC; binding of toxin from BPDE386 to Jurkat cells is contact was established between the patch pipette and the cell inte- comparable toor greater than that of toxin from BP338 (Table rior. Panel B , conditions were as in panel A , except that 1.6 pg of AC 11).The interactionof wild-type AC toxin with target cells in toxin from BPDE386 were added to the bath for the period defined the absence of calcium, and the interaction of mutant AC by the bar, and thenremoved. After washout was judged complete, 1 toxin with target cells in the absence or presenceof calcium pg of AC toxin from BP338 was applied for the indicated time. Panel C , an atrial myocyte was perfused internally with standard internal are, however, clearly nonproductive, yielding neither intoxisolution until the time indicated by the arrow. At this point, pressure cation nor hemolysis. The productive binding which occurs (30 kilopascals) was applied to thereservoir containing 0.12 pg of AC onlywhenwild-type toxin is added to target cells in the toxin from BPDE386. The pressure was released 5 min later when presence of calcium reflects a different type of interaction the solution in the reservoir containing the AC toxin from BPDE386 which results in the development of a poolof trypsin-resistant had all been injected. AC toxin representing insertion and delivery of the catalytic domain to thecell interior (Table 11). Therefore, the effectof indicate that AC toxin is able to elicit a response in cardiac the modification produced by CyaC must be either 1)to confer myocytes which reflects its ability to increase intracellular an orientation or conformation on the molecule to allow it to cAMP levels. enter the membraneor 2) to participate directly in the memIn contrast to the wild-type toxin, toxin from BPDE386 brane insertion event. Additional work with these and other has noeffect on IC.when applied to the exteriorof atrial cells mutants will be required to distinguish between these possi(Fig. 6B). The rapid response of the same cells, however, to bilities. wild-type toxin (Fig. 6B) or isoproterenol (data not shown) Rogel and Hanski have addressed the initial interactionof documents that the sequence of events leading to activation AC toxin with target erythrocytes ( 2 , 6, 39). Most recently, of the calcium channels, including wild-type toxin binding, is using AC toxin from BP348pRMB1, they demonstrated that not directly inhibited by the mutant toxin. Importantly, direct 3.9% of toxin added to sheep red blood cells in the presence application of toxin from BPDE386 to the cell interior (Fig. of EGTA remained cell-associated following a series of washes 6C) produces a n increase in IC. comparable to that which with EGTA and then sodium carbonate. They described the occurs when wild-type toxin is added externally. These data establish that the defect in AC toxin from BPDE386 affects N. C. Maloney, M. C. Gray, and E. L. Hewlett, manuscript in its ability to insert into the membrane deliver andthe catalytic preparation.

i

’“1

7

,J

7847

Characterization of Mutant AdenylateCyclase Toxin

/

AC Toxin (active)

FIG. 7. Sequence of events involved in the biological activities of AC toxins from B. pertussis. Boxed Toxin’

step reflects the termination of each pathway under the conditions indicated.

AC Toxin’

t-”

e ~

AC ~

~

bv CvaC

AC

Toxin

I“ ’ AC (inactive)

SI

C~II-AS~O~~W

CelMssociated

\

~

i

CAMP AcavnuYion and/or Hemolysis ~

’

I-

BP338

1

‘Calduminducad conformational change

residual, cell-associated toxin as “inserted” on the basis of its comparable to wild-type hemolysin, but is unable to bind to resistance to washing, but that material was completely re- or lyse erythrocytes. These data differ from those presented here for AC toxin from BPDE386, which is able to bind to moved by trypsin and appears to represent the same material target cells despite the mutation in cyaC and its inability to which we refer to ascell-associated or bound to Jurkatcells. In order to defineprecisely the functional status of the intoxicate and hemolyze target cells. These and other submutant toxin from BPDE386, was it necessary to evaluate its stantial differences, such as thetime course of hemolysis ability to act at its intended intracellularsite. The use of an caused by AC toxin (42, 46) indicate that caution should be of electrophysiological approach measuring a CAMP-dependent used in making any assumptions about the mechanism process in atrialmyocytes provided an opportunity to address activation mediated by CyaC. In summary, AC toxin from a mutant of B. pertussis conseveral important issues. The modulation of cardiac L-type calcium channels by CAMP-dependent phosphorylation is taining a lesion in cyaC has been purified and characterized. As with wild-type toxin, the mutant molecule binds calcium well characterized (37) and may underlie the chronotropic and undergoes a calcium-dependent conformational change effect ofAC toxinoncultured chick heart cellsobserved hemolytic activityand virtually notoxin previously by Selfe et al. (40). Wholecell patch clamp record- (18). Ithasno ing of ZCa in cardiac myocytes provides a rapid and sensitive activity. It does, however, bind to target cells in a manner wild-type AC toxin, but indicator of increases in intracellular CAMP. The measured which is quantitatively comparable to events are characteristic of calcium currents and cannot in- is obviously different, as illustrated by the inability to enter volve sodium currents because of the block of sodium channels the trypsin-resistant compartmentwhich is occupied by wildof calcium. That the toxin functions by tetrodotoxin. The rapid onset( 4 min) of the toxin effect type toxin in the presence normally when conveyed via patch pipette to thecell interior of AC toxin acting with essenis consistent with the concept tially no lag (2-4, 41, 42). There was no evidence of intoxica- demonstrates that thelesion induced by the absence of CyaC tion by externally added mutant toxin and thatmolecule did is in the membrane insertion/catalytic domain delivery (Fig. not block the action of subsequently added wild-type toxin 7). A better understanding of this process of insertion and (Fig. 6B), supporting the hypothesis that there is no specific delivery will have broad implications for the mechanisms by and saturable receptor system through which this toxin acts. which bacterial toxins and other molecules enter eukaryotic cells. Although thesestudies were not designed todetermine whether wild-type AC toxin produces an ion-permeable pore REFERENCES in the target cell membrane, as has been described for other 1. Confer, D. L., and Eaton, J. W. (1982) Science 217,948-950 pore-forming toxins such as E. coli hemolysin and staphylo2. Hanski, E., and Farfel, 2. (1985) J. Biol. Chem. 260,5526-5532 3. Gordon, V. M., Young, W. W., Lechler, S. M., Leppla, S. H., and Hewlett, coccal a toxin (43-45), one would expect conductivityof such E. L. (1989) J. Biol. Chem. 264, 14792-14796 a structure to be recognized under the conditions of these 4. Hanski, E. (1989) Trends Biochem. Sci. 14,459-463 I. E., Gray, M. C., Gordon, V. M., Gray, L. S., and Hewlett, E. experiments and none was detected. This illustrates the dis- 5. Ehrmann, L. (1991) FEBS Lett. 2 7 8 , 79-83 sociation between intoxication and cytolytic activities noted 6. Rogel, A., Miller, R., and Hanski, E. (1991) J. Biol. Chem. 266,3154-3161 7. Hewlett, E. L., Gordon, V. M., McCaffery, J. D., Sutherland, W. M., and previously by several investigators (5, 41). Gray, M. C. (1989) J. Biol. Chem. 264, 19379-19384 8. Rogel, A., Schultz, J. E.. Brownlie. R. M. Coote. J. G.. Panton. R.. and As noted above, the entire cya operon shares significant Hanski, E. (1989) EMbO J. 8, 2755-2760 sequence homology with the hly operon of E. coli hemolysin, 9. Gentile, F., Knlpllng, L. G., Sackett, D. L., and Wolff, J. (1990) J . B ~ O L . C h n.... n ”_, 26K 1-7. l.._ including hlyC which encodes the protein required for postJ., Ladant, D., and Sakamoto, H.(1990) Infect. Immun. 58, 1195translational activation (11-13). AC toxin and E. coli hemo- 10. Bellalou, 1200 lysin are alsoimmunologically cross-reactive (42, 46). Re- 11. Glaser, p., Ladant, D., Sezer, O., Pichot, F., Ullmann, A,, and Danchin, A. (1988) Mol. Microbiol. 2, 19-30 cently, Issartel et al. (16) reported that the HlyC-dependent 12. Glaser, p., Sakamoto, H., Bellalou, J., Ullmann, A,, and Danchin, A. (1988) EMBO J . 7,3997-4004 activation requires a charged acyl carrier protein which ap13. B a w , E. M., Weiss, A.A.. Ehrmann. E. E.. Grav. M. C.. Hewlett. E. L.. pearsto be contributing a fatty acyl group for thepostand Goodwin, M. St. M. (1991) J. Bacteriol. 173, 720-726 ’ translational event. Myristoylated carrier protein was most 14. Nicaud, J. M., Mackman, N., Gray, L., and Holland, I. B. (1985) FEBS T A t . 187. .1.19-.144 --effective, but other fatty acids also yielded active hemolysin. 15. Wagner, W., Kuhn, M., and Goebel, W. (1988) Biol. Chem. Hoppe-Seyler 369,39-46 These and subsequent datafrom the same laboratory (16,47) 16. Issartel, J. P., Koronakis, V., and Hughes, C. (1991) Nature 351, 759-761 strongly support the concept that HlyC is actively involved 17. Hewlett, E. L., Ehrrnann, I., Gray, L., Fremgen, P., Allietta, M., Gordon, V. M., and Gray, M. (1990) in Proceedings of the Sixth Internutional in the fatty acylation of HlyA and that event is required for 90-1164) (ManSymposium on Pertussis(DHHSPublication(FDA) its hemolytic activity. Boehm et al. (33,34) have demonstrated Clark, C. R., ed) pp. 13-19, Department of Health and Human Services, United States Public Health Service, Bethesda, MD that a mutant of E. coli from which a portion of hlyC has been 18. Hewlett, E. L. Gray, L., Allietta, M., Ehrmann, I. E. Gordon, V. M., and deleted produces hemolysin which binds calcium in a manner Gray, M. C.’(1991) J . Biol. Chem. 266, 17503-17568 .

1

7848

Characterization of Mutant AdenylateCyclase Toxin

19. Stockbridge, N. (1987)Comput. Biol. Med. 17,299-304 20. Martell, A. E., and Smith, R. M. (1974)Critical StabilityConstants, Vol. I . Amino Acids, Plenum Press,New York 21. Hewlett, E. L., Urban, M. A., Manclark, C. R., and Wolff J. (1976)Proc. N ~ t lAcad. . Sci. U. S. A. 73,1926-1930 22. Salomon, Y.,Londos C., AND Rodhell, M. (1977)A d . Biochem. 68,541548 23. Brooker, G., Terasaki, A., and Price, M. G. (1970)Science 194,270-276 24. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951)J. Biol. Chem. 193,265-275 25. Clerc, P., Baudry, B., and Sansonetti,P. J. (1986)Ann. Inst. Pasteur 137A,

-". -.OG7-9711

26. Laemmli, U. K. (1970)Nature 227,680-685 27. Mawama, K.,Mikawa, T., and Ebashi, S. (1984)J. Biochem. 95, 511519 28. Otero, A. S., Breitwieser, G. E., and Szabo, G . (1988)Science 242, 443445 29. Hamill, 0.P., Marty, A., Neher, E., Sakmann, B., and Sigworth, F. J. (1981)Pfluegers Arch. 391,85-100 30. Lapoint, J. Y., and Szabo, G. (1987)Pfluegers Arch. 410, 212-216 31. Welch, R. A., Felmlee, T., Pellett, S., and Chenoweth, D. (1986)in Proteincarbohydrate Interaction in Biobgkal Systems(Lark, D. L., Normark, s., Uhlin, B. E., and Wolfwartz, H., eds) pp. 433-438,Academic Press, New

34. Boehm, D. F., Welch, R. A., and Snyder, I. S. (1990)Infect. Immun. 58, 1959-1964 35. Gentile, F. Raptis, A., Knipling, L., and Wolff, J. (1988)Biochim. Biophys. Acta 97i, 63-71 36. Gordon, V.M. (1989)A Comparative Study of the Calmodulin-actiuated Adenylate Cyclase Toxins rom Bordetellapertclssis and Bacillus anthracis: Interactions with Taraet iells, Ph.D. dissertation, University of Virginia,

Charlottesville, VA 37. Trautwein, W., and Hescheler, J. (1990)Annu. Reu. Ph siol 52,257-274 38. Nar eot J Nerbonne J. M Engels J and Lester, &. A. (1983)Proc. dtl.hc&. Sci. U. Sl A. 80; 2395-299; 39. Rogel, A,, and Hanski, E.(1992)J. Biol. Chem. 267,22599-22605 40. Selfe, S., Hunter, D. D., Shattuck, R. L., Nathanson, N. M., and Storm,D. R. (1987)MOL Phrmacol. 31,529-534 41. Ro el, A., Miller, R., and Hanski, E. (1990)in Proceedings of the Sixth sium on Pertussis (DHHS Publication (FDA) 90fnternational S 1164)(Manclarg,yR., ed)pp. 20-29,Department of Health and Human Services, United States Publx Health Service, Bethesda, MD 42. Ehrmann, I. E. (1991)Hemolytic Activity of Adenylate Cyclase Toxin rom Bordetella rtussis, Ph.D. dissertation, University of Virginia, dharlottesville 43. Bhakdi, S., hackman, N., Nicaud, J.-M., and Holland, I. B. (1986)Infect. Immun 62 63-69 44. Menestrina, 6.,Machman, N., Holland, I., and Bhakdi, S. (1987)Biochim. Biophys. Acta 906,109-117 45. Thelestam, M., and Blomquist L. (1988)Toxicon 26,51-65 46. Ehrmann, I. E., Weiss, A. A., hoodwin, M. S., Gray, M. C., Barry, E., and Hewlett E. L. (1992)FEBS Lett. 304,51-56 47. Hardie, K.'R., Issartel, J.-P.,Koronakis, E., Hughes, C., and Koronakis, V. (1991)Mol. Mtcmbwl. 5,1669-1679

FA