JOURNAL OF CELLULAR PHYSIOLOGY 1453444449 (1990)

Evidence That the Adenylate Cyclase Secreted From Bordete//a pertussis Does Not Enter Animal Cells by Receptor-Mediated Endocytosis MAURA G. DONOVAN AND DANIEL R. STORM*

Department of Pharmacology, University of Washington, Seattle, Washington 98 I95 Bordetella pertussis, the pathogen responsible for whooping cough, produces a calmodulin-sensitive adenylate cyclase. Several investigators have shown that the partially purified adenylate cyclase is capable of entering animal cells and elevating intracellular cAMP levels (Confer and Eaton: Science 21 7:948-950, 1982; Shattuck and Storm: Biochemistry 24:6323-6328, 1985). However, the mechanism for entry of the catalytic subunit of this adenylate cyclase into animal cells is unknown. It has been reported that the 5. pertussis adenylate cyclase extracted from bacterial cells with urea does not enter animal cells by receptormediated endocytosis. There is, in addition to the cell associated form of the 6 . pertussis adenylate cyclase, a cell-invasive form of the enzyme secreted into the bacterial culture media. The properties of the cell-associated and secreted enzymes are significantly different (Masure and Storm: Biochemistry 28:438-442, 1989). In this study, we report evidence that the secreted form of the B. pertussis adenylate cyclase enters animal cells by a mechanism distinct from receptormediated endocytosis.

Bordetella pertussis is a ram negative bacillus that is the causative agent o whooping cough. Several virulence factors produced by B. pertussis have been identified and characterized (Weiss and Hewlett, 1977). One of these factors is an extracellular adenylate cyclase that has two unusual properties. It is activated by calmodulin (CaM), the eukaryotic regulatory protein (Wolff et al., 1980), and it elevates intracellular cAMP levels in a variety of animal cells (Confer and Eaton, 1982; Hanski and Farfel, 1985; Shattuck and Storm, 1985). There is evidence which suggests that this increase in intracellular cAMP results from the B. pertussis adenylate cyclase invading animal cells and catalyzing the formation of CAMP: The B . pertussis adenylate cyclase elevates intracellular cAMP levels in human erythrocytes which contain no measurable endogenous adenylate cyclase activity (Shattuck and Storm, 1985). The amount of intracellular cAMP formed after adding the B . pertussis adenylate c clase to animal cells is in excess of what is forme with maximal concentrations of forskolin, a potent stimulator of the animal cell membrane adenylate cyclase (Gentile et al., 1988). Furthermore, addition of CaM to the B. pertussis extracellular adenylate cyclase comletely inhibits formation of intracellular cAMP catafyzed by the bacterial enzyme (Shattuck and Storm, 1985). Molecular genetics (Glaser et al., 1988)and biochemical studies (Masure and Storm, 1989) have established that the B. pertussis adenylate cyclase is synthesized as a large molecular weight precursor of approximately 200,000. This polypeptide is proteolytically processed to

P

B

0 1990 WILEY-LISS, INC.

a catalytic subunit of about 45,000 daltons which is secreted into the culture media (Shattuck et al., 1985). Proteolysis of the 200,000 dalton form of the enzyme, which occurs quite readily, leads to a cell invasive enzyme preparation (Masure and Storm, 1989). Furthermore, it has recently been shown that the purified catalytic subunit of the enzyme isolated from culture media cannot enter animal cells (Masure et al., 1988; Donovan et al., 1989). Reconstitution of the purified catalytic subunit with a protein isolated from culture media, termed invasive factor, yields an invasive adenylate cyclase preparation (Donovan et al., 1989). These data suggest that the enzyme is proteolytically processed into an adenylate cyclase catalytic subunit and one or more rotein subunits that facilitate entry of the catalytic su unit into animal cells. Although the mechanism for entry of the catalytic subunit into animal cells is not understood, the data discussed above suggest that this toxin conforms to the eneral A/B model for bacterial toxins proposed by Gil (1978).

g

P

Received April 10, 1990; accepted August 24, 1990. *To whom reprint requestdcorrespondence should be addressed. Abbreviations used: CaM, calmodulin; DMEM, Dulbecco’s modified essential medium; Hepes, 4-(2-hydroxyethyl)-l-piperazineethane sulfonic acid; PAGE, polyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; PMSF, phenylmethylsulfonyl fluoride; QAE, diethyl (2-hydroxypropyl) aminoethyl; SDS, sodium dodecyl sulfate; Tris-HC1, tris (hydroxymethyl) aminomethane hydrochloride.

445

SECRETED FORM OF B. PERTUSSIS ADENYLATE CYCLASE

Bacterial toxins frequently enter host cells by receptor mediated endocytosis (Gordon et al., 1988; Morris et al., 1985). Studies conducted with adenylate cyclase isolated from whole cells of B . pertussis by urea extraction suggested that, unlike diphtheria toxin and Pseudomonas exotoxin A, the B. pertussis adenylate cyclase does not enter animal cells by receptor-mediated endocytosis (Gentile et al., 1988; Gordon et al., 1988).In contrast, the CaM-sensitive adenylate cyclase isolated from culture media of Bacillus anthrucis does enter animal cells by rece tor-mediated endocytosis (Gordon et al., 1988). The a enylate cyclase associated with whole B. pertussis cells is distinct from that released into the culture media, as clearly evidenced by differences in their apparent molecular weights (Masure and Storm, 1989).For this reason, and because the effect of urea on the properties of the enzyme have not been been systematically examined, it was of interest to determine if the adenylate cyclase secreted into the culture media enters animal cells by receptor-mediated endocytosis. The data reported in this study indicate that the adenylate cyclase secreted from B. pertussis does not enter animal cells by receptor-mediated endocytosis.

a

MATERIALS AND METHODS Materials QAE-Sephadex was purchased from Pharmacia. Protein kinase and cAMP were from Sigma. Alpha [32Pl ATP and [3H]cAMPwere purchased from International Chemical Nucleus. Dulbecco’s modified essential medium (DMEM) with high glucose content (4.5 g/L) and fetal calf serum were purchased from Grand Island Biological Co. All other reagents were of the finest available grade from commercial sources. Isolation of an invasive calmodulin-sensitive adenylate cyclase preparation B. pertussis (Tohama phase I strain) was rown from a 5% inoculum in supplemented Stainer- cholte medium (Stainer and Scholte, 1971). Invasive preparations of the adenylate cyclase were partially purified from culture supernatants by QAE-Sephadex chromatography as described by Shattuck et al. 1985. Pre arations of the enzyme collected from QAE-Sepha ex chromatography were concentrated by ultrafiltration using Amicon PM-10 membranes and stored at -80°C.

Q

5% fetal calf serum, without antibiotics, in an atmosphere of 10% co2/90%humidified air. The cells were own to 80-90% confluency in plastic tissue culture ishes (60 mm) prior to the start of each experiment. The cells were subcultured weekly, and the culture medium was changed on days 3,5, and daily thereafter. On the day of each experiment, cultures were washed with serum-free DMEM and then preincubated for 20 min at 37°C in 3 mL of DMEM supplemented with 5 mM theophylline and buffered with 10 mM Hepes-HC1, pH 7.4. After preincubation, cells were treated with various preparations of the adenylate cyclase for 20 min at 37°C. Typically, a minimum of 100 nmol/min of enzyme activity was applied per plate of cells. The enzyme-containing solution was removed, and the cells were washed twice with PBS and then lysed with 4 mL of 5% trichloroacetic acid. The protein precipitate was removed by centrifugation and cAMP was isolated from the supernatant by chromatography on 2 mL AG50WX4 (200-400 mesh) Dowex columns. L3H1cAMP was included for determination of column recovery. Each cell entry experiment included a determination of basal cAMP levels because we have observed that parameter varies with the passage of neuroblastoma cells. Aliquots of the eluted samples were analyzed for cAMP

r

”:, =I

160

B

Adenylate cyclase assay Adenylate cyclase was assayed at 30°C by the method of Salomon et al. (1974) using a l ~ h a - [ ~ ~ATP P ] as a substrate and L3H1cAMP to monitor product recovery. Each assay contained 20 mM Tris-HC1 (pH 7.5),1 mM a l ~ h a - [ ~ ATP ~ P I (10 cpm/pmol), 5 mM MgC12, 1 mM EDTA, 1 mM P-mercaptoethanol, 0.1% (wh) BSA, and 2.4 pM CaM in a final volume of 250 ~ 1 Results . are presented as the mean of duplicate assays. Growth of neuroblastoma cells and cell entry assays N1E-115 mouse neuroblastoma cells (passages 1628) were grown at 37°C in DMEM supplemented with

T

L 0

-

P

Fig. 1. Effect of inhibitors of receptor-mediated endocytosis on entry of the B . pertussis adenylate cyclase into neuroblastoma cells. Cells were grown to confluency as described in “Materials and Methods.” On the day of the experiment, cells were washed with serum-free DMEM buffered with 10 mM Hepes-HC1, pH 7.4, and supplemented with 5 mM theophylline. The cells were then preincubated in the same buffered media for 20 min at 37°C 2 inhibitors (100 FM chloroquine, 10 mM NH,CI, or 50 pM monensin). After preincubation, cells were incubated with adenylate cyclase (100 nmoliminiplate of cells) for 20 min at 37°C. The adenylate cyclase preparation (Pklc) used was purified from B. pertussis culture media throu h QAE Sephadex (Shattuck et al., 1985). The cells were then wasted with PBS and assayed for intracellular CAMP.

446

DONOVAN AND STORM

1800

A Blank 500

25oc

37OC

4oc

Wash

I

3TC

100

C

0 Blank

40 c

0 40 c

Trypsin 370 c

using the Gilman assay (Gilman, 1970). Results reported are the mean and range of triplicate samples. Furthermore, each cell entry experiment depicted in the figures was repeated twice.

Fig. 2. Temperature dependency of adenylate cyclase entry into neuroblastoma cells. Cells were washed and preincubated for 20 min at 37°C in serum-free DMEM buffered with 10 mM Hepes-HC1, pH 7.4, and supplemented with 5 mM theophylline. A After preincubation, cells were incubated for 20 min in buffered DMEM containing adenylate cyclase (100 nmol/min/plate of cells) at either 3TC, 25"C, or 4°C as indicated. The cells were then washed with PBS and assayed for intracellular cAMP as described in "Materials and Methods." B:After preincubation, cells were incubated at either 37°C or at 4°C for 20 min in buffered DMEM containing adenylate cyclase (100 nmollminiplate of cells). Following the 20 min incubation, the first two sets of cells (designated 37°C and 4°C) were washed with PBS and assa ed for intracellular CAMP.The third set of cells (designated 4"C!was&37'C) was washed with PBS, incubated for 10 min a t 37°C in buffered DMEM and then washed with PBS and assayed for intracellular CAMP.C After reincubation, cells were incubated at 4°C for 20 min in buffered DMfM containing adenylate cyclase (100 nmol/min/plate of cells). Control cells were incubated in serum-free DMEM for 20 min at 37°C to determine basal cAMP levels under comparable conditions. Following the 20 min incubation, the first two sets of cells were washed with PBS and assayed for intracellular CAMP.The remaining cells were washed with PBS, incubated with 0.01% trypsin for 10 min at 37T, and then washed and assayed for intracellular CAMP.

Isolation of human erythrocytes and cell entry assays Whole blood, collected in 0.15% EDTA, was centrifuged at 200g for 5 min, and the upper layer was

SECRETED FORM OF B. PERTUSSIS ADENYLATE CYCLASE TABLE 1. Temperature dependence of adenylate cyclase activity' Incubation temperature

37oc 25OC 4°C

Enzyme activity'

** 66 31 173 * 48

747 597

'Adenylate cyclase was prepared by anion exchange chromatography as described "Materials and Methods" and concentrated with an Amicon PM.10 membrane. CaM-stimulated adenylate cyclase activity was measured by the method of Salomon et al. (1974) as described in "Materials and Methods." Enzyme activity is expressed as nmol cAMP/min/mL

discarded. The erythrocytes were washed twice in PBS containing 5 mM theophylline. The cells were diluted to a hematocrit value of 50%,layered onto Ficoll-Pague, and centrifuged at 200g for 25 min. The lymphoc te layer was removed, the packed cells were washe 2 times with PBS and again diluted to a 50% hematocrit value prior to the start of the experiment. Isolated human erythrocytes were incubated for 15 min at 37°C with preparations of the adenylate cyclase (100 nmoVmin of enzyme activity per 1 mL of diluted erythrocytes). The erythrocytes were centrifu ed through Ficoll-Paque and washed twice with P S. Packed cells were lysed with 4 mL of 5% trichloroacetic acid. The protein precipitate was removed by centrifugation and cAMP was isolated from the extract by chromatography on 2 mL AG50W-X4 (200-400 mesh) Dowex columns. Approximately 15,000 cpm [3HlcAMP was added to the samples for determination of column recovery. Aliquots of the eluted samples were analyzed for cAMP using the Gilman assay (Gilman, 1970). Results represent the mean and range of triplicate samples.

B

%

Determination of protein Protein concentrations were determined by the method of Lowry et al. (1951). Samples were prepared by incubating the precipitated protein with 2 mL of 1N NaOH until they dissolved. Samples were then aliquoted for protein determination. RESULTS AND DISCUSSION Several compounds which inhibit endosome acidification, an event linked to receptor-mediated endocytosis of some proteins, were examined for their effect on entry of the secreted B. pertussis adenylate cyclase into N1E-115 mouse neuroblastoma cells (Fig. 1). Cells were washed with serum-free DMEM and then preincubated in DMEM containing NH,Cl, chloroquine, or monensin. After the preincubation, cells were either incubated with the invasive adenylate cyclase preparation or with control buffer. In this experiment, treatment of neuroblastoma cells with the adenylate cyclase alone caused an %fold increase in intracellular CAMP. Treatment of the cells with NH,Cl, chloroquine, or monensin, without added adenylate cyclase, had no significant effect on intracellular cAMP levels. Preincubation of the neuroblastoma cells with NH,Cl, chloroquine, or monensin did not inhibit increases in intracellular cAMP caused by the B. pertussis adenylate cyclase. In contrast, chloroquine and NH&l have been shown to inhibit increases in intracellular cAMP

447

caused by the adenylate cyclase secreted from B . anthrucis (Gordon et al., 1988). These two drugs did not affect invasion of animal cells by the adenylate cyclase isolated from B. pertussis cells by urea extraction (Gordon et al., 1988). In mammalian cells, incubation at 4°C results in arrest of endocytosis; when the cells are warmed to 37"C, endocytosis resumes in a synchronous fashion (Goldenthal et al., 1984). Therefore, the temperature dependency for entry of the B. pertussis adenylate cyclase into neuroblastoma cells was examined (Fig. 2A). Cell invasion experiments were conducted at 4"C, 25"C, and 37°C. The greatest increase in intracellular levels of cAMP was observed a t 37°C. It was notable, however, that incubation of neuroblastoma cells with the adenylate cyclase at 4°C did result in a 6-fold increase in the intracellular concentration of CAMP compared to basal levels. Thus, B. pertussis adenylate cyclase elevated intracellular cAMP concentrations even when the incubation was conducted at 4°C. The difference between the 17-fold increase in intracellular cAMP concentrations observed at 37°C and the 6-fold increase seen at 4°C is most likely due to decreased kinetic activity of the enzyme at the lower temperature. Therefore, the adenylate cyclase was assayed at 37"C, 25"C, and 4°C (Table 1).At 37°C the activity of the adenylate cyclase was approximately 4 times greater than at 4°C. Based on the results of the precedin experiments, it was hypothesized that the extent o invasion of neuroblastoma cells by the adenylate cyclase was comparable at 37°C and 4°C and that the difference in intracellular cAMP levels at the two temperatures was due primarily to the temperature dependence of enzyme activity. Consistent with this hypothesis was the observation that warming the cells to 37"C, after the 4°C incubation and washing of the cells, resulted in a 13-fold increase in intracellular cAMP concentrations compared to untreated controls (Fig. 2B). This value was comparable to the 17-fold increase in intracellular cAMP obtained by incubating cells with the enzyme at 37°C for the same period of time. In the experiment described above (Fig. 2B), it could be argued that aden late cyclase absorbed to the outside surface of the ce 1 at 4°C and then entered the cell when the temperature was increased to 37°C. To test this possibilit , the cells were treated with 0.01% trypsin after t e 4°C incubation and before raising the temperature to 37°C (Fig. 2 0 . This treatment resulted in a 4-fold increase in intracellular cAMP concentrations over blank levels, a value comparable to the 6-fold obtained by incubating cells with the enzyme at 4°C. From these data, we conclude that the adenylate cyclase is able to enter N1E-115 neuroblastoma cells at both 4°C and at 37"C, although entry may occur more readily at 37°C. There are two additional observations which indicate that invasion of animal cells by the adenylate cyclase secreted by B. pertussis does not occur by rece tormediated endocytosis. First, the enzyme is ab e to invade a variety of animal cell types including mature human erythrocytes (Shattuck and Storm, 1985),which are not thought to undergo receptor-mediated endocytosis (Stahl and Schwartz, 1986). Second, entry into

B

i

i

P

448

DONOVAN AND STORM

B

1

I

10

20

30

40

50

Incubation Time (min)

0

10

20

30

40

50

60

70

Incubation Time (min)

Fig. 3. Kinetics for the entry of the B. pertussis adenylate cyclase into animal cells. A Entry into human erythrocytes. Human erythrocytes, prepared as described in “Materials and Methods,” were incubated at 37°C with adenylate cyclase. Samples containing 2.6 x lo9 cells were removed at the indicated times, centrifuged through Ficoll-Paque, and washed with PBS. Packed cells were assayed for intracellular CAMP.B: Entry into N1E-115 mouse neu-

roblastoma cells. Cells were washed with serum-free DMEM, buffered with 10 mM Hepes-HC1 (pH 7.4), and supplemented with 5 mM theophylline, and then preincubated in the same media for 20 min a t 37°C. After preincubation, cells were incubated at 37°C with adenylate cyclase (100 nmoliminiplate of cells) for the indicated periods of time. Cells were then washed with PBS and assayed for intracellular CAMP.

both human erythrocytes and N1E-115 mouse neuroblastoma cells occurs very rapidly (Fig. 3). There was no indication of a lag time for entry of the adenylate cyclase into human erythrocytes (Fig. 3A) and only a 3 to 5 min lag for entry into neuroblastoma cells (Fig. 3B).This contrasts with the entry of diphtheria toxin and pseudomonas exotoxin A into animal cells which proceeds after a 10-30 min lag period (Olsnes and Sandvig, 1983). In cell-free systems these toxins appear to act immediately and it has been sug ested that the lag eriod could be the time required or the toxin to be ta en up by endocytic vesicles and released into the cytoplasm (Refnes et al., 1974). There is now considerable evidence, obtained by a variet of methods, which suggests that the B . pertussis adeny ate cyclase enters animal cells by a mechanism distinct from rece tor-mediated endocytosis. In the present study, we ave used different methods and a different re aration of the bacterial enzyme than was used by 8or:on et al. (1988). The results suggest that the B . pertussis adenylate cyclase, whether isolated from bacterial culture supernatant or obtained by urea extraction of whole cells, utilizes a potentially novel mechanism or mechanisms for membrane translocation.

ACKNOWLEDGMENTS This research was funded by NIH grant GM 31708.

B

i

9

K

LITERATURE CITED Confer, D.L., and Eaton, J.W. (1982) Phagocyte impotence caused by an invasive bacterial adenylate cyclase. Science, 217:948-950. Donovan, M.G., Masure, H.R., and Storm, D.R. (1989) Isolation of a protein factor from Bordetella pertussis that facilitates entry of the calmodulin-sensitive adenylate cyclase into animal cells. Biochemistry, 28:8124-8129. Gentile, F., Raptis, A., Knipling, L.A., and Wolff J. (1988) Bordetella pertussis adenylate cyclase penetration into host cells. Eur. J. Biochem., 175:447453. Gill, D.M. (1978) Seven toxic peptides that cross cell membranes. p. 291332. In: Bacterial Toxins and Cell Membranes. J. Jeljaszewicz and T. Wadstrom, ed. Academic Press, Inc., New York, pp. 291-332. Gilman, A.G. (1970)A protein binding assay for adenosine 3’:5’-cyclic monophosphate. Proc. Natl. Acad. Sci. U.S.A., 67:305-312. Glaser, P., Ladant, D., Sezer, O., Pichot, F., Ullman, A., and Danchin A. (1988) The calmodulin sensitive adenylate cyclase of Bordetella pertusis: cloning and expression in E . coli. Mol. Microbiol., 2:19-30. Goldenthal, K.L., Pastan, I., and Willingham, M.C. (1984) Initial steps in receptor-mediated endocytosis: The influence of temperature on the shape and distribution of plasma membrane clathrin coated pits in cultured mammalian cells. Exp. Cell Res., 152:558-564. and Hewlett, E.L. (1988) Inhibitors of Gordon, V.M., Leppla, S.H., receptor mediated endocytosis block the entry of Bacillus anthracis adenylate cyclase toxin but not that of Bordetella pertussis adenylate cyclase toxin. Infee. Immun., 56:1066-1069.

SECRETED FORM OF B . PERTUSSIS ADENYLATE CYCLASE Hanski, E., and Farfel, Z. (1985) Bordetella pertussis invasive adenylate cyclase. J. Biol. Chem., 260:552&5532. Jiang, G., Solow, R., and Hu, V. (1989) Characterization of diphtheria toxin induced lesions in liposomal membranes. J. Biol. Chem., 264:1342&13429.

Leppla, S. (1982) Anthrax toxin edema factor: a bacterial adenylate cyclase that increases CAMP concentrations in eukaryotic cells. Proc. Natl. Acad. Sci. U.S.A., 79:3162-3166. Lowry, O.H., Rosebrough, N.H., Farr, A.L., and Randall, R.J. (1951) Protein measurement with the Folin phenol reagent. J. Biol. Chem., 193:265-275.

Masure, H.R., Oldenburg, D.J., Donovan, M.G., Shattuck, R.L., and Storm,D.R. (1988) The interaction of calcium with the calmodulin sensitive adenylate cyclase from Bordetella pertussis. J . Biol. Chem., 263:6933-6940. Masure, H.R., and Storm, D.R. (1989) Characterization of the bacterial cell associated calmodulin sensitive adenylate cyclase from Bordetella pertussis. Biochemistry, 28:43EM42. Morris, R.E., Gersten, AS., Bonventre, P.F., and Saelinger, C.B. (1985) Receptor mediated entry of diphtheria toxin into monkey kidney (Vero) cells: Electron microscopic evaluation. Infect. Immun., 50:721-727. Olsnes, S., and Sandvig, K. (1983) Entry of toxic proteins into cells. In

449

receptors and recognition. series B, Vol. 15. P. Cuatrecasas and T. Roth, eds. Chapman and Hall, London, pp 187-236. Refnes, K., Olsnes, S., and Pihl, A. (1974) On the toxic proteins abrin and ricin. J. Biol. Chem., 249:3557-3562. Salomon, Y., Londos, D., and Rodbell M. (1974) A highly sensitive adenylate cyclase assay. Anal. Biochem., 58541-548. Shattuck, R.L., Oldenburg, D.J., and Storm, D.R. (1985) Purification and characterization of a calmodulin sensitive adenylate cyclase from Bordetella pertussis. Biochemistry, 24:63566362. Shattuck, R.L., and Storm, D.R. (1985) Entry of Bordetella pertussis adenylate cyclase into human erythrocytes. Biochemistry, 24:63236328.

Stahl, P., and Schwartz, A.L. (1986) Receptor mediated endocytosis. J. Clin. Invest., 77:657-662. Stainer, D., and Scholte, M. (1971) A simple chemically defined medium for the production of Phase I Bordetella pertussis. J . Gen. Microbiol., 63:211-220. Weiss, A.A., and Hewlett, E.L. (1977) Virulence factors of Bordetella pertussis. Annu. Rev. Microbiol., 40:661-674. Wolff, J., Cook, G.H., Goldhammer, A.R., and Berkowitz, S.A. (1980) Calmodulin activates prokaryotic adenylate cyclase. Proc. Natl. Acad. Sci. U.S.A., 77:3840-3844.