The Interaction of Ca2+ with the Calmodulin-sensitive Adenylate

2 downloads 0 Views 2MB Size Report
Aug 6, 1987 - ited a Ca2+-dependent heat stability with a half-life for loss of enzyme ..... Ca2+-binding proteins including CaM, troponin C, and par- valbumin ...
Vol. 263, No. 14, Issue of May 15, pp. 6933-6940. 1988 Printed in U.S.A.

THEJOURNAL OF‘BIOLOGICAL CHEMISTRY 0 1988 by The American Society for Biochemistryand Molecular Biology, Inc

The Interactionof Ca2+with theCalmodulin-sensitive Adenylate Cyclase from Bordetella pertussis” (Received for publication, August 6, 1987)

H.Robert Masure$, Delene J. Oldenburg, Maura G. Donovan#,Rebecca L. Shattuckll, and Daniel R. Storm11 From the Department of Pharmacology, School of Medicine, University of Washington, Seattle, Washington 98122

Bordetellapertussis, the etiologic agent of whooping SDS gels was 51,000 in the presence of 2 mM CaC12 cough, produces a calmodulin-sensitive adenylate cy- and 45,000 in the presence of 2 mM EGTA. The cataclase which elevates intracellular cAMP in a variety lytic subunit of the enzyme was purified to apparent of eucaryotic cells. Exogenous calmodulin added to the homogeneity by preparative SDS-polyacrylamide gel to SDS gel electrophopartially purified adenylate cyclase has been shown to electrophoresis and resubmitted inhibit invasion of animal cells by this enzyme (Shat- resis in the presence or absence of free Ca2+. The pualso exhibited a Ca2+-dependent tuck, R. L., and Storm, D. R. (1985)Biochemistry 24, rified catalytic subunit 6323-6328). In this study, several properties of the shift in its mobility on SDS gels. These experiments indicatethat thecatalytic subunit calmodulin-sensitive adenylate cyclase are shown to be of the invasiveform of the enzyme is secreted into the influenced by Ca2+in theabsence of calmodulin. The presence or absence of Ca2+during QAE-Seph- culture supernatantas a 45,000-kDa polypeptide. This adex ion exchange chromatographyproduced two dis- polypeptide interacts directly withCa2+which affects tinct chromatographic patterns of adenylate cyclase its physical and enzymatic properties. Furthermore, of the adenylatecyclase was isolated activity. Two different formsof the enzyme ( P k l and the invasive form P~~EG were T Aisolated ) by this procedure. P k l adenyl- in a Ca2+-dependentmanner anddisplayed a reversible ate cyclase readily elevated intracellular cAMP levels Ca2+-dependent increase in itsapparent molecular weight which is probably due to an association with in mouse neuroblastoma cells (NlE-115)while PkSEGTA adenylate cyclase had no effect on cAMP levels in these another component of the culture supernatant. cells. Gel exclusion chromatography of P k l adenylate cyclase gave apparent Stokes radii (Rs)of 43.5 A (2 1.3) in thepresence of 2 mM CaC12 and 33.8 A (f0.94) Bordetella pertussis, the bacteria responsible for whooping in the presence of 2 mM EGTA ([ethylenebis (oxyethylenenitrilo)]tetraacetic acid). These Stokes cough, produces a soluble adenylate cyclase that is stimulated radii are consistent with molecular weights of 104,000 by the eucaryotic regulatory protein calmodulin (CaM).’ This (26,400) and 61,000 (23,600), respectively. P ~ ~ E Genzyme T A invades animal cells and catalyzes the synthesis of adenylate cyclase had an apparent Rs of 33.0 (21.2) cAMP from intracellular ATP (for a review, see Masure et (M. = 60,600 (22,800)) in the presence of Ca2+ or al., 1987). excess EGTA. At 60 OC, P k l adenylate cyclase exhibWolff and Cook (1973) first detected adenylate cyclase a half-life for activity in crude vaccine preparations of B. pertussis, and ited a Ca2+-dependentheat stability with loss of enzyme activity of 10.3 min in 5 mM CaCl, and these investigators subsequently determined that theenzyme a half-life of 2.8 min in the presence of 0.1 p~ CaC1,. is stimulated by CaM (Wolff et aL, 1980). The activation of adenylate TA cyclase was not the enzyme by CaM is unique in that it is stimulated both in The stability of P ~ ~ E G affected by changes in free Ca2+. The adenylatecyclase preparations described above the presence and absence of Ca2+ (Greenlee et al., 1982; were submitted tosodium dodecyl sulfate (SDS)-poly- Kilhoffer et al., 1983). Shattuck and Storm(1985) determined acrylamide gel electrophoresis, and enzyme activity that partially purified adenylate cyclase elevated cAMP levels was recovered from gel slices by extraction with de- in human erythrocytes and mouse neuroblastoma cells (NlEtergent containing buffers. The catalytic subunit iso- 115), and this elevation was blocked by exogenously added lated from SDS-polyacrylamide gels was activated 7- CaM. This confirmed earlier work byConfer and Eaton(1982) fold in the presence of Ca2+ with maximum activity and Hanski and Farfel (1985) who described the ability of observed at 1PM free Ca2+.With both preparations, the impure B. pertussis adenylate cyclase preparations to elevate apparent molecular weight of the catalytic subunit on intracellular cAMP in a variety of eucaryotic cells. The B. pertussis adenylate cyclase is a soluble enzyme which * This work wassupported inpart by National Institutes of Health is released into the culturesupernatant. Recently, several Grant GM 31708. The costs of publication of this article were defrayed laboratories have described significant purifications of the in part by the payment of page charges. This article must therefore enzyme. Shattuck et al. (1985), Kessin and Franke (1986), be hereby marked “udvertisement” in accordance with 18 U.S.C. and Ladant et al. (1986) isolated soluble forms of the enzyme Section 1734 solely to indicate this fact. from culture supernatants. Other investigators have purified $ Supported by a Postdoctoral Fellowship NS-07985. 5 Supported by National Institutes of Health Training Grant GM bacterial cell-associated forms of the enzyme from urea ex07750. W Current address: Howard Hughes Medical Inst., Dept. of Pharmacology, Vanderbilt University, Nashville, T N 37232. 11 To whom correspondence should be addressed Dept. of Pharmacology, SJ-30, School of Medicine, University of Washington, Seattle, WA 98122.

’ The abbreviations used are: CaM, calmodulin; Chaps, 3-[(3cholamidopropyl)dimethylaminio]-l-propanesulfonate; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; Hepes, N-2-hydroxyethylpiperazine-N’-2-ethanesulfonicacid; Mops, 3-(N-mOrpho1ino)propanesulfonicacid; SDS, sodium dodecyl sulfate.

6933

6934

B. pertussis Adenylate Cyclase Interaction with

tracts of concentrated cells (Hanski and Farfel,1985; Kessin and Franke, 1986; Friedman, 1987). Several different forms of the enzyme have been reported, although the quarternary structure of the adenylate cyclase has not been determined. There is a small form(s) of the enzyme with reported molecular weights between 43,000 and 70,000 (Hewlett and Wolff, 1976; Hanski and Farfel, 1985; Shattuck et aL, 1985; Kessin and Franke, 1986; Ladant et al., 1986;Friedman, 1987) which is thought to be the catalytic subunit of the enzyme. In this study we report the effect of Ca2+on several properties of this enzyme and propose that Ca2+interacts directly with its catalytic subunit. During the purification and characterization of this enzyme we discovered that its chromatographic behavior on ion exchange columns and gel exclusion columns was affected by the concentration of free Caz+present. Several Ca2+-dependent properties are probably due to the interaction with another component released by the bacteria into the culture supernatant.

Cu2+

(NlE-1151, passages 8-28, weregrown to 80-90% confluency in Dulbecco's modified Eagle's medium with 0.5% fetal calf serum in 60-mm plates as described previously (Shattuck and Storm, 1985). Dulbecco's modifiedEagle's medium wasreplaced with 3 ml of Earle's salt solution (110 mM NaCl, 5.5 mM glucose, 1 mM NaH2P04,1.8 mM CaCl,, and 0.81 mM MgCl,) supplemented with 5 mM theophylline and buffered with 25 mM Hepes-HC1 (pH 7.4). Cells were treated with various preparations of the adenylate cyclase and incubated for 20 min at 37 'C. The enzyme containing solution was removed and the cells washed twice with 2 ml of cold phosphate-buffered saline. The cells were lysed with 5% trichloroacetic acid and the isolated supernatants assayed for cAMP by the method of Gilman (1970). The protein pellets were solubilized in 2 ml of 1 N NaOH and protein determined according to the method of Lowry et al. (1951). In certain instances the free Ca2+concentration of solutions bathing the cells was controlled with a Ca2+-EGTAbuffer system as described above. Free CaZ+was monitored before and after exposure to the adenylate cyclase with a Ca2+-sensitiveelectrode constructed as described by Anker et al. (1981) and Moody and Thomas (1979). RESULTS

QAE-Sephudex Chromatography of B. pertussis Adenylate Cyclase in the Presence and Absence of Ca2+"Shattuck et al. Adenylate Cyclase Assay-The production of cAMP was measured by the method of Salomon et al. (1974). Each assay contained 20 mM (1985) reported the isolation of two peaks of adenylate cyclase 5 mM MgCl,, 1 mM EDTA, activity from QAE-Sephadex chromatography. One peak Tris-HC1 (pH 7.4), 1 mM [cY-~~PJATP, and 0.1% bovine serum albumin. Some samples were assayed in the eluted at a conductivity of 7 pMho and thesecond at 10 pMho. presence of 2.4 PM CaM. For assays performed in a controlled divalent We noted variability in the QAE-Sephadex chromatographic cation environment, distilled-deionized H,O was passed over a Bio- profile with different lots of glutamic acid (Aldrich; Lot Nos.: Rad Chelex 100 column to further lower contaminating divalent A P 02825HL; 04722AP; 05006BM), which is a major compocations. Where indicated, a CaZ'-EGTA buffer system was used to control free Ca2+. Calculated free Ca2+ was determined with an nent of Stainer-Scholte medium. Analysis of different lots of iterative program writtenin BASIC computer language using an glutamic acid revealed variation in thelevels of contaminating equilibrium equation with dissociation constants for Ca2+-EGTA, Ca2+.It was determined that variations in the amount of Ca2+ Ca2'-ATP, Ca2'-EDTA, M e - E D T A , M%+-ATP,and other divalent present in the isolated bacterial culture medium, which is cation-ligand complexes taken from Fabiato and Fabiato (1979), Allen applied directly to QAE-Sephadex, had significant effects on et al. (1977), and Dawson et al. (1969). Each assay point was calculated the QAE elution profile. We describe two distinct chromatobased on the assumption that Mg2"ATP was the preferred substrate. graphic patterns of adenylate cyclase elution fromQAEIn divalent cation dose-response assays of adenylate cyclase activity, other divalent cations compete with M%+ for ATP. Mg2"ATP con- Sephadex dependent upon the presence or absence of Ca2+in centration never varied below 0.8 mM. One unit of adenylate cyclase the adsorption and elution of the enzymes from the resin. QAE-Sephadex chromatography was carried out in the catalyzes the synthesis of 1 nmol of cAMP/min. Preparation of Calmodulin-CaM was prepared from bovine brain presence of Ca2+by adjusting culture supernatant isolated by the procedure of Masure et al. (1984). from B. pertussis to 2 mM CaC12 and applying the sample to Purification of the Calmodulin-sensitiveAdenylate Cyclase-E. per- QAE-Sephadex equilibrated in 20 mM Tris-HC1 (pH 7.5), 40 tussis (Tahoma phase 1 strain) was grown from a 5% inoculum in supplemented Stainer-Scholte medium (Stainer and Scholte, 1971). mM NaC1, and 2 mM MgC12 (QAE buffer). The resin was then Culture supernatants were isolated and QAE-Sephadex chromatog- washed with QAE buffer and eluted with a 40 mM to 1 M raphy performed as described by Shattuck et al. (1985) with some NaCl gradient in the same buffer. The concentration of Ca2+ modifications as described below. Preparations of the enzyme col- in QAE buffer was 10 p~ as measured by atomic absorption lected from QAE-Sephadex chromatography were concentrated with spectroscopy. Two peaks of calmodulin-sensitive adenylate Amicon PM-10 or PM-30 membranes and stored a t -80 "C. cyclase activity (Pkl and Pk2) eluted from the column at The catalytic subunit of the CaM-sensitive adenylate cyclase was conductivities of 7 and10 pMho (Fig. lA).This elution profile further purified with SDS-polyacrylamide gel electrophoresis. Concentrated preparations of the adenylate cyclase were equilibrated in is similar to thatreported by Shattuck et al. (1985).Treatment sample buffer (0.0625 M Tris-HC1 (pH 6.9), 0.5% SDS, 1% P-mercap- of culture supernatant with 2 mM CaC12 and elution in QAE toethanol, and 1.0% glycerol) and electrophoresed into 0.25 X 20-cm buffer in the presence of 2 mMCaC1, gave a single peak of slab gels maintained at a constant temperature of 20 'C. Gel lanes activity at a conductivity of 7 pMho. Treatment of the P k l were cut into 0.4-0.5-cm slices and soaked in 20 mM Tris-HC1 (pH adenylate cyclase preparation with 3 mM EGTA and resub7.4), 40 mM NaCI, 2 mM MgCI,, with or without Chaps as indicated. In some instances, to increase recovery of enzyme activity, gel slices mission to QAE-Sephadex chromatography in the absence of Caz+gave one peak of adenylate cyclase activity which eluted were frozen at -20 "C for at least 12 h prior to assay and protein from the column at a conductivity of 10 pMho. Resubmission determination. SDS-Polyacrylamide Gel Electrophoresis-SDS-polyacrylamide gel of P k l in QAE buffer in the presence of 2 mMCaC1, and electrophoresis was performed as described by Laemmli (1970). Pro- elution with QAE buffer containing 2 mM CaC12 gavea single teins were visualized with silver as described by Ansorge (1985). peak of activity at a conductivity of 7 pMho. Resubmission of Stokes Radius Determination-The Stokes radius (Rs) was determined for the different forms of the adenylate cyclase by gelexclusion pooled Pk2 adenylate cyclase activity to QAE-Sephadex chrochromatography with a calibrated Ultrogel AcA 44 (LKB, Uppsaula, matography either in the presence of 2 mMCaC1, or 2 mM Sweden) column (90 X 1.5 cm). Chromatography was performed in a EGTA substituted for CaCL, in all elution buffers gave one buffered solution containing 10 mM Mops-HC1 (pH 7.4), 110 mM peak of enzyme activity which eluted at a conductivity of 10 NaCl, 5 mM MgCl,. This solution was supplemented with either 2 pMhO. mM CaCl, or 2 mM EGTA where indicated. A linear selectivity curve QAE-Sephadex chromatography was carried out inlow free was generated with aldolase (47.4 A), bovine serum albumin (37.0 A), Ca2+by adjusting the culture supernatant to 2 mM EGTA and ovalbumin (27.6 A), and cytochrome c (17.4 A) used as standards. Determination of Intracellular CAMPLevels in Neuroblastoma Cells applying it to QAE-Sephadex as described above. This sample with B. pertussis Adenylate Cyclase-Mouse neuroblastoma cells was treated asdescribed above except all buffers were suppleEXPERIMENTAL PROCEDURES

B. pertussis Adenylate 800

1A

Cyclase InteractionCa2+ with

Pk2

Pkl

c

6935

TABLE I Elevation of intracellular cAMP in muroblustornu cells causedby B. pertussis adenylate c y c h e preparations at several ea2+ concentrations Intracellular CAMP" Enzyme preparationb

Free Ca2'

1.8m M

100 wM

0.1 p M

No enzyme added

8 15 9

48

B

70

92

k2 EGTA

121 (+32) 137 (+24) 150 (+17) Pkl 8404 (+127) 6083 (+117) 1930 (+lo41 P k l + 4 p~ CaM 127 (+31) 142 (+27) 131 (+39) Pk2 158 (+36) 142 (+22) 145 (+27) 136 (+22) 161 (+33) 153 (+15) P~~EGTA Picomoles of cAMP/mg protein. Concentrated adenylate cyclase preparations (200 units/plate) were applied to mouse neuroblastoma cells (NlE-115) and assayed for intracellular cAMP as described under "Experimental Procedures." Each result is an average of triplicate assays. 10,000

-.E

a d

Free Ca 8,000

2% -% & s= o,

0 %

L?

"

+

5mM 100 uM

6,000

4,000

c o

74 2

58

90

Fraction No.

- E

-

a 2.000

FIG.1. Effect of Ca2+on the elution of B. pertussis adenylate cyclase from QAE-Sephadex. A , 18 liters of culture supernatant from B. pertussis was adjusted to a final concentration of 2 mM CaC12,applied to QAE-Sephadex, washed with 3 litersof2 020 0 mM Tris-1 5 0 100 50 Adenylate cyclase activity HCl (pH 7.5), 40mM NaCl, and 2 mM MgClz, and eluted with 40mM to 1 M NaC1, linear gradient. B, culture supernatant was treated the (Ulplateof cells) same as in A except 2 mM EGTA was substituted for CaCL and the FIG. 2. Influence of Ca2+ on the elevation of cAMP in mouse wash and elution buffers were adjusted to a final concentration of 2 mM EGTA. 12-ml fractions were collected and assayed for adenylate neuroblastoma cells (NlE-115) with B. pertussis adenylate cyclase activity in the presence of CaM (0).The conductivity in cyclase. Various concentrations of the P k l adenylate cyclase (unit = nmol/min) were incubated with neuroblastoma cells adjusted to selected fractions was also measured (0). extracellular free Ca2+concentrations of 5 mM (01,100 p M (m), and 0.1 p M (0)for 20 min at 37 "C. Intracellular cAMP levels in neuromented with 2 mM EGTA. A single peak of adenylate cyclase blastoma cells were determined as described under "Experimental activity ( P ~ & G Teluted A ) from the column at a conductivity Procedures." Each point is the average of three determinations. The free Ca2+concentrations in the 100 p M and 0.1 p M solutions were of 10 pMho (Fig. 1B). Elevation of CAMP inNeuroblastoma Cells Stimulated by B. controlled with a Ca2+-EGTAbuffer. The free Ca" concentration of solutions bathing the cells was measured with a Ca2+-selectiveelecpertussis Adenylate Cyclase Preparations-Pkl,Pk2,and trode before and after incubation with the enzyme.

P~~EG adenylate T A cyclase preparations (Fig. 1)were analyzed for their ability to elevate cAMP in neuroblastoma cells at several different free Ca2+ concentrationsin the solutions bathing the cells. In each case, 200 units of adenylate cyclase activity/plate were incubated with neuroblastoma cells for 20 min in either 1.8 mM, 100 pM, or 0.1 p M free Ca2+ (Table I). Extracellular Ca2+at 100 and 0.1 p~ was controlled by Ca2+EGTA buffers and monitored with a Ca2+-selectiveelectrode before and after incubation with adenylate cyclase preparations. P k l adenylate cyclase significantly increased intracellular CAMP with the highest levels observed at 1.8 mM free Ca2+,and lower accumulations of cAMP were observed at 100 and 0.1 p~ free Ca2+.When P k l adenylate cyclase waspreincubated with 4 p~ CaM, invasion of neuroblastoma cells was TA completely inhibited.NeitherPk2 nor P ~ ~ E Gadenylate cyclase preparations had any affect on intracellular cAMP levels even at levels of Pk2 or Pk2EGTA up to 1000 units of adenylate cyclase activity/plate of cells. Since Pk2and P~~EG adenylate TA cyclase preparations were both noninvasive forms of the enzyme and displayed identical properties,

PkBEGTAwas selected for further characterization with P k l adenylate cyclase. The ability of P k l adenylate cyclase to elevate intracellular cAMP levels in neuroblastoma cells was examined at several different free Ca2+concentrations (Fig. 2). Increases in intracellular cAMP were directly dependent upon the amount of P k l adenylate cyclase activity applied to the neuroblastoma cells. Variation in the concentration of extracellular free Ca2+ affected the increases in cAMP caused by the bacterial enzyme and there was no effect of extracellular Ca2+on basal cAMP levels. These data indicate that invasion of neuroblastoma cells by B. pertussis adenylate cyclase is sensitive to the concentration of extracellular Ca2+. Furthermore, the maximum increase in intracellular cAMP occurred at the highest level of Ca2+ ( 5 mM) examined which most closely approximates physiological extracellular Ca2+concentrations. Gel Exclusion Chromatography of B. pertussis Adenylate Cyclase Preparations-The data described abovesuggested that Ca2+may interact directly with B. pertussis adenylate

B. pertussis Adenylate Cyclase Interaction with Ca2+

6936

cyclase and affects its physical and enzymatic properties. Therefore, we examined the influence of Ca2+on the Stokes radius of the enzyme using gel exclusion chromatography. A cyclase preparations Concentrated P k l or P ~ ~ E G Tadenylate were applied to an Ultrogel AcA 44 column equilibrated in a buffer containing 2 mM CaCl,, 2 mM EGTA, or no additions (Fig. 3). The free Ca2+concentration in the sample applied with no addition as well as the free Ca2+in the elution buffer which produced this chromatogram was 10 p ~In. the presence of 2 mM CaCl,, P k l material elutedas a single peak of activity with an RS of 43.5 A (k1.3). This Stokes radius is consistent with an apparent molecular weight of 104,000 (+6,400). In the presence of 2 mM EGTA, or with no $dditions P k l adenylate cyclase had an apparent Rs of 33.8 A (k1.5) which is consistFnt with an M , of 61,000 (+3,900). Reapplication of the 43.5-A species of P k l adenylate cyclase in the presence of 2 mM EGTA resulted in the e!ution of a single peak of activity with an apparent Rs of 33.8 A. Whereas, reapplication of the 33.8-A species in the presence of 2 mM CaCl, resulted in the elution of a single peak of activity with an identical Rs value. Thus, the apparent size of the invasive form of adenylate cyclase (Pkl) was doubled with the addition of 2 mM Ca2+. This implies that Caz+promotes the reversible association of anotherfactor(s) with the catalyticsubunit.Furthermore, this factor(s) is separated from the catalytic subunit with gel TA exclusion chromatography. Incontrast, P ~ ~ E Gadenylate cyclase eluted from the column with an apparent Rs of 33.0 A (f1.2) in the presence of 2 mM EGTA or 2 mM CaCl,, this value isconsistent with a molecular weight of 60,600 (k2,800). Influence of Ca2+ on the Heat Stability of the E. pertussis Adenylate Cyclase-Adenylatecyclase activity in P k l and P ~ ~ E Gpreparations TA was assessed for heat stability at free caz+concentrations of 5 mM, 100 pM, and 0.1 pM. Concentrated samples of P k l and PkBEGTA were initially isolated in QAE buffer at a free Ca2+ concentration of 10 pM. These samples were adjusted to varying amounts of free Ca2+ and were warmed to 60 "C. Aliquots were removedat various times ADL

--

-

BSA

-

OVA

A A

60

50

70

80

P k l + Ca2+ P k l + EGTA Pkl Pk2 EGTA+ EGTA

90

100

and assayed for adenylate cyclase activity in the absence of CaM. The half-life of enzyme activity was calculated from a linear plot of the log[percentage enzyme activity] versus time. Table I1 summarizes the half-lives obtained for decay of P k l and P ~ ~ E Gadenylate TA cyclase activity at three different free Ca2+ concentrations. CaZf enhanced the stability of P k l adenylate cyclase but had no effect on the stability of P ~ ~ E G T A adenylate cyclase. The half-life for loss of P k l adenylate cyclase activity was 10 min at 5 mM free Ca2+,5.4 min at 100 p M free ca2+,and 2.8 min at 0.1 p~ free Ca2+. Influence of Ca2+on the Mobility of the Catalytic Subunit of B. pertussis Adenylate Cyclase During SDS-Polyacrylamide Gel Electrophoresis-The effects of Ca2+on the properties of B. pertussis adenylate cyclase described above could be due either to direct interaction of Ca2+with the catalytic subunit of the enzyme or to an indirect effect, such as Ca2+binding to another protein subunit that interacts with the catalytic subunit. Although it is currently not technically feasible to obtain sufficient amounts of the catalytic subunit to examine Caz+ binding by equilibrium dialysis or other biophysical techniques, small amounts of homogeneous catalytic subunit can be detected following SDS-polyacrylamide gel electrophoresis (Kessin and Franke, 1986). Furthermore, a number of Ca2+-binding proteinsincluding CaM, troponin C, and parvalbumin exhibit a mobility shift on SDS-polyacrylamide gels in the presence of Ca2+(Head and Perry, 1974; Amphlett et al., 1976; Grab et al., 1979; Klee et al., 1979; Burgess et al., 1980). Therefore, we examined the effect of Ca2+ on the mobility of adenylate cyclase activity in concentrated culture supernatant, Pkl, and P ~ ~ E G Tadenylate A cyclase activity on SDS-polyacrylamide gels. Concentrated culture supernatant and P k l adenylate cyclase was adjusted to either 2 mM CaC12 or 2 mM EGTA equilibrated with sample buffer and applied to a 10% SDSpolyacrylamide gel. The gels were sliced, extracted with QAE buffer supplemented with 2% Chaps, and assayed for adenylate cyclase activity (Fig. 4, A and B ) . For both culture supernatant and Pkl, adenylate cyclase activity migrated with an apparent molecular weight of 51,000 in the presence of Ca2+. When these enzyme preparations were run on a SDS gel in the presence of 2 mM EGTA, it ran with an apparent molecular weight of 45,000 (Fig. 4,A and B ) . These results suggest that thecatalytic subunit of the adenylate cyclase is secreted into the culture supernatant as 45-kDa a polypeptide and that neither QAE-Sephadex chromatography nor Ca2+ leads to degradation of the enzyme. The shift in mobility of the catalytic subunit on SDS gels caused by Ca2+ washighly reproducible and was observed with eight different preparations of adenylate cyclase. The shift in mobility of the catalytic subunit on SDS gels seen in the presence of Ca2+could be dueto a direct interaction of the catalytic with Ca2+or to modification of the catalytic subunit by a Ca2+-dependent enzyme. To rule out either a Ca2+influenced covalent modification of the enzyme or the

Elution (mL)

FIG. 3. Influence of Ca2' on gel exclusion chromatography of the B. pertussis adenylate cyclase.Concentrated P k l adenylate cyclase (475 units) adjusted to either 2 mM CaC1,).( or 2 mM EGTA (O), or no addition (A) was applied to an Ultrogel AcA 44 column and eluted with 10 mM Mops-HC1 (pH 7.4), 110 mM NaC1, 5 mM MgCl, supplemented with either 2 mM CaCl, or 2 mM EGTA, or no addition, respectively. Concentrated P ~ ~ E(455 G Tunits) ~ was adjusted to 2 mM EGTA (A),applied to the column and eluted with a buffered solution containing 10 m M Mops-HC1 (pH 7.4), 110 mM NaCl, 5 mM MgCl,, and 2 mM EGTA. Fractions were collected at a rate of 30 ml/h and assayed for CaM-stimulated adenylate cyclase activity. Aldolase (ADL),bovine serum albumin (BSA),and ovalbumin ( O V A ) and cytochrome c (not shown) were used as standards.

TABLE I1 The influence of Caz+on the enzyme stability of different preparations of B. pertussis adenylate cyclase Half-life" (min) of enzyme activity at 60 "C Enzyme preparation

Free Ca2+

5 mM

100 p M

0.1 p M

Pkl 5.4 (fO.l) 10.3 (k0.2) 2.8 (f0.3) 3.1 (f0.2) 2.9 (k0.4) P ~ ~ E G T A 3.2 (f0.3) a The half-life of adenylate cyclase activity was determined from a linear plot of log[percentage of enzyme activity] uersus time. Each result is an average of triplicate assays.

B. pertussis Adenylate

Cyclase InteractionCa2+ with

6937

kDa 45

17

66

92

200

Sllce No.

Slice No

17

45

kDa

66

92

8

-* Z a^

6

m E 22 O E 0 % E rnc -

z>

4

2 . "

m

B

21

10

Slice

No.

43

FIG. 4. Ca2+induced mobility shift ofthe B. pertussis adenylate cyclase submitted to SDS-polyacrylamide gel electrophoresis. A, 200-microliter samples of fresh concentrated culture supernatant (65 units) were adjusted to either 2 mM CaC1, (0)or 2 mM EGTA (W), equilibrated with sample buffer (0.0625 M Tris-HC1 (pH 6.9), 1.0% SDS, 1%8-mercaptoethanol, and 1%glycerol), boiled for 2 min, and electrophoresed into a 10% SDSpolyacrylamide gel. Each gel lane was cut into 0.45-cm slices, and each slice was placed into 1 ml of a solution containing 20 mM Tris-HCl (pH 7.5), 40 mM NaCl, 2 mM MgCl,, and 2% Chaps. Samples were assayed for adenylate cyclase activity in the presence of 2.4 PM CaM. B , 200-4 samples of concentrated Pkl adenylate cyclase (170units) were adjusted to either 2 mM CaCIZ(0)or 2 mM EGTA (W), treated as described in A, and electrophoresed into a 10% SDS-polyacrylamide gel. This is a representative example of eight separate experiments. C, the peak fraction of adenylate cyclase activity from Fig. 4B run in the presence of 2 mM CaClZwas divided in half and precipitated with 10%trichloroacetic acid. Each sample was resuspended in sample buffer containing either 2mM CaCL (0)or 2 mM EGTA (W). These samples were electrophoresed into a 12.5% SDS-polyacrylamide gel; each lane was sliced and treated asdescribed above.

selective isolation of two different forms of the enzyme, the in the mobility of the catalytic subunit caused by Ca2+was 51,000 and 45,000 forms of the catalytic subunit from P k l reversible, and that it occurred with the homogeneous cataadenylate cyclase were purified to homogeneity by preparative lytic subunit. Effect of the Catalytic Subunit Isolated from SDS Gels on SDS gel electrophoresis and rerun on SDS-polyacrylamide gels in the presence or absence of Ca2+.In the presence of 2 Intracellular CAMPLevels of Neuroblastoma Cells-The catmM CaC12,the 51,000 form of the enzyme migrated as a single alytic subunit of B. pertussis adenylate cyclase from a P k l peak of activity with an apparent molecular weight of 51,000 adenylate cyclase preparation was purified to apparent hoon a SDS gel (Fig. 4C). With 2 mM EGTA present, this mogeneity by preparative SDS gel electrophoresis. 1.2 pg of polypeptide ran with an apparent molecular weight of 45,000 total protein was obtained with a specific activity of 333 pmol (Fig. 4C). The 45,000 form of adenylate cyclase exhibited of cAMP/min/mg. The SDS gel elution profile for purification of the catalytic subunit and asilver-stained gel obtained when similar behavior when it was rerun on SDS gels. In other words, it migrated with apparent molecular weight on SDS the purified subunit was rerun on a SDS gel i s depicted in gels of 51,000 in thepresence of Ca" and 45,000 in theabsence Fig. 5. The preparation contained only one polypeptide ( M r This material was assayed for its ability to elevate T 45,000). A of Ca2+. Similarresults were obtained when the P ~ ~ E G = adenylate cyclase preparation was submitted to SDS gel elec- cAMP levels in neuroblastoma cells (Fig. 6). Intracellular trophoresis. This polypeptide ran with apparent molecular cAMP was increased approximately 9-fold relative to unweights of 51,000 in the presence of Caz+ and 45,000 in the treated cells when treated with P k l adenylate cyclase. The absence of Caz+and the shift in mobility caused by Ca2+was catalytic subunit isolated from SDS gels, containing an equivreversible. This would imply that thecatalytic subunitin P k l alent amountof adenylate cyclase activity, had no measurable ~ indistinguishable T A by gel electrophoresis and effect on neuroblastoma cAMP levels. This experiment indiand P ~ ~ Eare are probably identical. These data illustrate that the change cates that the catalytic subunit purified by SDS gel electro-

B. pertussis Adenylate

6938

1

10

Cyclase Interaction with ea2'

20

Slim No.

FIG. 5. Purification of the catalytic subunit of the B. pertuesis adenylate cyclase. 300 pl of concentrated P k l adenylate cyclase (330 units) was equilibrated with sample buffer (defined in Fig. 4) and electrophoresed into a 7% SDS-polyacrylamide gel. The gel was cut into 0.45-cm slices, and each slice was placed into 250 pl of a solution containing 20 mM Tris-HC1 (pH 7.5), 40 mM NaCl, and 2 mM MgCI,. The samples were frozen for 12 h a t -20 "C, thawed,

Gels-The catalyticsubunit of theadenylate cyclase was purified to apparenthomogeneity by SDS-polyacrylamidegel electrophoresis as describedabove, andthesensitivity of adenylate cyclase activity to Ca2+ and several divalent cations was examined. The catalytic subunit of adenylate cyclase isolated from a SDS-polyacrylamide gel was stimulated 7-fold at a n optimal free Ca2+ concentration of 1 p~ (Fig. 7 A ) . Increasing the concentrations of free Ca2+ up to 5 mM inhibited enzyme activity. This stimulation is specific for Ca2+. Neither Mn2+ nor Sr2+ stimulated P k l adenylate cyclase in the concentrationrange between 0.1 p~ and 1 mM (Fig. 7 B ) . Enzyme activity was stimulated by Zn", however, with maximal activity observed at 7.5 p~ Znz+. Although activation of adenylate cyclase by Zn2+is interesting, the concentrations required are quite high and probably not physiologically significant. The optimal stimulationby Ca2+ did occur a t intracellular concentrations of the ion. DISCUSSION

Ca2+influencedseveral properties of the CaM-sensitive adenylate cyclase from B. pertussis including its chromatographicbehavior on QAE-Sephadexchromatography, the Stokes radius of the enzyme, the mobility of the catalytic subunit on SDS gels, enzymestability at 60 "C, catalytic activity, and the effectiveness of the enzyme in elevating cAMP levels in neuroblastoma cells. None of these effects of 2000 Ca2+required the presenceof CaM. The datareported in this study strongly suggest a direct interaction between Ca2+and the catalytic subunit of the enzyme since Ca2+ affected the enzyme activity and SDS gel mobility of the homogeneous catalytic subunit. We believe that interactions of Ca2+ with the catalytic subunit are physiologically relevant because both enzyme activity andcell invasiveness were Ca2+ sensitiveand the effects of Ca2+ on the properties of the enzyme were observed a t concentrations of the ion that arephysiologically relevant. The direct interaction of Ca2+ with the catalytic subunit of the adenylatecyclase probably promotes the association of the catalytic subunit with another factor present in the culture supernatant. This interaction may account for the different propertiesobserved with anionexchange chromatography, gel exclusion chromatography, heat lability, and cell invasion. Previous studies have described the influence of Ca2+ on B A C D E the activationof B. pertussis adenylate cyclase in thepresence Adenylate cyclase preparation of CaM (Wolff et al., 1980; Greenlee et al., 1982; and Kilhoffer FIG. 6. Effect of the purified catalytic subunit on cAMP et al., 1983). Ca2+ enhancedtheaffinity of CaM for the levels of mouse neuroblastomacells (NlE-116). Adenylate cy- enzyme, presumably by the interaction of the cation with clase preparations were incubated with plates of neuroblastoma cells CaM, but there was significant enzyme stimulation by CaM for 20 min a t 37 "C as described under "Experimental Procedures." Samples are: A, 180 units/plate of cells of the catalytic subunit of even at low free Ca2+ (Greenlee et al., 1982). This current P k l adenylate cyclase purified with preparative SDS gel electropho- study documents a specific Ca2+ stimulation of the enzyme resis as described in Fig. 5; B , 180units/plate of cells of P k l adenylate independent of CaM which was optimal a t free Ca2+concencyclase: C, 180 units/plate of cells of pK1 adenylate cyclase which trations comparable to intracellular levels, i.e. micromolar was incubated with a polyacrylamide gel slice:D, a solution containing free Ca2+.Inhibition of enzyme activity at millimolar concen20 mM Tris-HC1 (pH 7.5), 40 mM NaCl, and 2 mM MgCl, was incubated with a polyacrylamide gel slice containing no enzyme trations of Ca2+most likely reflects competition between Ca2+ percentage of ATP complexed to activity. 150 pl of this solution was added to each plate of cells. E, no and Mg2+ for ATP since the Caz+ is significanta t higher Ca2+ concentrations. Under the addition. These results are an average of triplicate assay points. assay conditionsof this study, itwas calculated that approxphoresis does not invade neuroblastoma cells. The inability imately 20% of the ATPwould be complexed to Ca2+at 1mM a substrate for the enzyme. It of the isolated subunit to enter neuroblastomacells may have free Ca2+,and Ca2+-ATP is not been due to the loss of one or more additional subunits that is noteworthy that invasion of neuroblastoma cells by the facilitate its entry into animal cells. Alternatively modifica- enzyme was most effective at millimolar concentrations of Ca2+, which approximates the extracellular concentrationof tion of the catalytic subunit may have occurred during its Ca2+surrounding animal cells. purification on SDS-polyacrylamide gels. The effect of Ca2+on the mobility of the catalytic subunit Divalent Cation Activation of the B. pertussis Adenylate Cyclase Catalytic Subunit Isolated from SDS-Polyacrylamide of adenylate cyclase on SDS-polyacrylamide gels is similar to and thenassayed for adenylate cyclase activity inthe presence of 2.4 p~ CaM and 0.08% Chaps. The migration of molecular weight standards is indicated on the top axis.The inset shows a silver-stained 10% SDS-polyacrylamidegel of50 pl of the peak adenylate cyclase fraction recovered from the 7% polyacrylamide gel.

r

h

B. pertussis Adenylate Cyclase Interaction with Ca2'

6939

80

--

1

.->

I

0 0

E

6o

-3 ,