Vol. 262, No. 36,Issue of December 25, pp. 17677-17682,1987 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY

Structural Characterizationof Pertussis Toxin A Subunit* (Received for publication, June 12, 1987)

Drusilla L. Burns$, Sally Z. Hausman, Wolfgang Lindner, Frank A. Robey, and Charles R. Manclark From the Center for Drugs and Biologics, Food and Drug Administration, Bethesda, Marylnnd 20892

The relationship between the structure of the A sub- form a disulfide bond. Small stretches of the sequence near unit of pertussis toxin and its function was analyzed. the NH2-terminal region of the A subunit show striking Limited tryptic digestion of the A subunit converted homology with the sequence of cholera toxin, another ADPthe protein to two stable fragments ( M . = 20,000 and ribosylating toxin (18, 19, 22). Thirty-one of the first 100 18,000). Antibodies raised to synthetic peptides ho- amino acids of pertussis toxin A subunit can be aligned with mologous to regions in the A subunit were used tomap homologous amino acids in the NHz-terminal region of cholthese fragments. Both fragments were shown to con- era toxin A subunit (19). tain the NHz-terminal portion but not theCOOH-terWhile the primary structure of the A subunit is known, minal portion of theA subunit. While these fragments little information is available concerning which regionsof the exhibited NAD glycohydrolase activity, they were un- molecule play a role in the enzymatic reaction catalyzed by able to reassociate with the B oligomer of the toxin. this protein and which regions are important for binding to Thus the COOH-terminal portion of the A subunit does? the B oligomer. We have dissected the purified A subunit by not contain the residues which are required for the NAD glycohydrolase activity of the toxin. However, exposing the molecule to trypsin for limited periods of time. this region of the molecule may be important for main- We found that trypsin cuts this proteinapproximately 40-50 taining the oligomeric structure of the toxin. These residues from the COOH-terminal end. Tryptic fragments results suggest that theA subunit of pertussistoxin is which contain approximately the first 190 residues of the A similar in structure to theA subunit of cholera toxin. subunit retained the NAD glycohydrolase activity of the naIn addition, antibodies raised to a synthetic peptide tive protein. Inaddition, the ability of the proteolyzed protein identical to residues 6-17 of theA subunit of pertussis to interact with the B oligomer of the toxin was altered. These results implicate regions of the molecule which may be imtoxin will bind to the A subunit of choleratoxin. portant for enzymatic activity and regions which play a role in maintaining the oligomeric structure of the toxin. Moreover, the results reported in this paper revealpreviously Bordetella pertussis produces an exotoxin, pertussis toxin, unrecognized similarities between the structures of the A which has been implicated in thepathogenesis of the organism subunits of pertussis toxin and cholera toxin. (1).The toxin interacts with vertebrate cells of many types and interrupts signal transduction in these cells by ADPribosylating a family of GTP-binding regulatory proteins (27). Modification of these regulatory proteins can result in an inability of the eukaryotic cell to respond to a variety of hormones or neurotransmitters (8-14). Pertussis toxin resembles other bacterial toxins such as cholera toxin in that it has an A-B structure (15-17). The B oligomer, composedof 5 subunitsranging in molecular weights from 11,000 to 22,000 (15, 18, 19) is responsible for binding of the toxin to the eukaryotic target cell. The A subunit has a molecular weight of 26,000 and contains the ADP-ribosyltransferase activity of the toxin (18-20). In the absence of a eukaryotic proteinsubstrate, the A subunit catalyzes the hydrolysis of NAD to ADP-ribose and nicotinamide (20, 21). The amino acid sequence of the A subunit hasbeen derived from the nucleotide sequence of the pertussis toxin gene (18, 19). The A subunit has a totalof 234 or 235 amino acids’ and contains 2 cysteine residues at positions 41 and 200 which

EXPERIMENTALPROCEDURES

Mater&-Trypsin (~-l-tosylamido-2-phenylethyl chloromethyl ketone-treated), soybean trypsin inhibitor, Staphylococcusaureu V8 protease, chymotrypsin, N-chloroacetylglycylglycine,anisole, thioanisole, Freund’s complete adjuvant, Freund’s incomplete adjuvant, ovalbumin, anti-mouse IgG alkaline phosphatase conjugate, phosphatase substrate (p-nitrophenyl phosphate),and cholera toxin A and B subunits were purchased from Sigma; cholera toxin from List Biological Laboratories; [carbonyl-“CJNAD(30-60 mCi/mmol) from Amersham Corp.; CHAPS’ from Behring Diagnostics; goat anti-mouse IgG horseradish peroxidase conjugate, 4-chloro-1-naphthol, and Tween 20 were from Bio-Ftad; reagents for peptide synthesis including phenylacetamidomethyl resins and t-BOC amino acids were purchased from Applied Biosystems; NNIH(S) micewere obtained from the Division of Research Resources of the National Institutes of Health (Bethesda, MD); pertussis toxin was purchased from the Michigan Department of Public Health. Preparation of Pertussis Toxin Subunits-The A subunit was separated from the B oligomer by the method described previously (23). The A subunit was stored at 4 “C in buffer containing 2 M urea. Gel Electrophoresis-Samples were prepared for SDS-gel electrophoresis by adding dithiothreitol and SDS to give final concentrations of 50 mM and 1%(w/v), respectively. Each sample was then heated a t 100 “C for 2 min. SDS-gel electrophoresis was performed essentially as described by Laemmli (24) using 15% acrylamide gels. The proteins used as standards for molecular weight calculations were bovine serum albumin, ovalbumin, chymotrypsinogen A, and lysozyme. Nondenaturing gel electrophoresis, pH 4, was conducted as

* This work wassupported inpart by Participating Agency Support Agreement BST-5947-P-HI-4265 between the United States Agency for International Development and the United States Public Health Service. The costs of publication of this article were defrayed in part by the payment of page charges. This article must thereforebe hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ To whom correspondence should be addressed Center for Drugs and Biologics, 8800 Rockville Pike, Bethesda, MD 20892. The abbreviations used are: CHAPS, 3-[(3-~holamidopropy1)The two published amino acid sequences for the A subunit differ dimethylammonio]-1-propanesulfonate; SDS, sodium dodecyl sulfate; in the region from amino acids 190-200 (18, 19). In this paper, we PBS, phosphate-buffered saline; ELISA, enzyme-linked immunosoruse the numbering system of Locht and Keith (18). bent assay. 17677

17678

Structural Characterization of Pertussis Toxin A Subunit

previously described (25) using 7.5% acrylamide gels. Protein Determination-Protein was measured by the method of Bradford (26) using ovalbuminas the standard. NAD Glycohydrolase Assays-NAD glycohydrolase activity of solutions of the A subunit, trypsin-treated A subunit, and pertussis toxin was measured as previously described (23). The NAD glycohydrolase activity of the A subunit or trypsin-treated A subunit immobilized on nitrocellulose was determined as follows. The A subunit or trypsin-treated A subunit was subjected to SDS-gel electrophoresis and then electrophoretically transferred nitrocellulose to strips (each 6-mm wide) as described previously (27) except that methanol was omitted from the transfer buffer. After transfer, the membrane was soaked in50 mM potassiumphosphate buffer containing 1 mM dithiothreitol for approximately 1 h. The nitrocellulose strips were then cut into 3-mmlengths. These stripswere then added directly to a tube containing 100 pl of 50 mM potassium phosphate buffer, pH 7.5, containing 9.5 p~ [carbonyl-'TINAD, 1%CHAPS, 1 mM dithiothreitol, 1 mg/ml ovalbumin, and0.1 M NaCl for 18 h. The reaction was stopped as previously described (23) and ["C]nicotinamide was isolated (28). Immunoblot Analysis-Proteins separated by SDS-gel electrophoresis were electrophoretically transferred tonitrocellulose membranes by the procedureof Towbin etal. (27). After transfer, the membranes were washed in phosphate-buffered saline, pH 7.0, containing 0.05% Tween 20 (PBS/Tween)and were thenincubatedwith a 1:lOOO dilution of the appropriate mouse serum in PBS/Tween for 1 h. The membranes were washed three times with PBS/Tween and were then incubated for 1 h with goat anti-mouse I& horseradish peroxidase conjugate. The membranes were washed three times with PBS and the bandswere visualized by the addition of the color reagent (30 mg of 4-chloro-1-naphthol in 10 ml of methanol combined with 50 ml of PBS containing 0.015% hydrogen peroxide). ELISA Procedures-The ELISA method used was essentially that described by Manclark et al. (29). Microtiter plates were incubated with the appropriate protein solutiona t a concentration of 5 pg/ml. Nonspecific antibodyadsorption was blocked by incubatingthe coated plates for 1 hwith a solution of 1 mg/ml ovalbumin. Sera were diluted serially (3-fold) with an initial dilutionof 1:30. Allincubations were conducted a t room temperature. Data were plotted as the log,, dilution uersus optical density a t 405 nm. ELISA antibody titerswere calculated as the inverseof the antilogof the x-intercept extrapolated from the linear portionof each curve. Synthesis of Peptides-Peptides were synthesized using an automated solid phase peptide synthesizer (model 430A, Applied Biosystems). In addition to the predetermined amino acid sequence of the desired peptides, N-chloroacetylglycylglycinewas incorporated at the NH, terminus for conjugation to the carrier protein using the following conditions for automation: 2.0 mmol of N-chloroacetylglycylglycine/0.5 mmol of preceding peptide was added to each of two blank synthesizer cartridges and the instrument programmed to perform the same double coupling procedure as that used to couple arginine to a peptide. Because N-chloroacetylglycylglycineis soluble in.N,N'dimethylformamide, the coupling to predetermined peptideswas performed via the active ester formation using dicyclohexylcarbodiimide with 1-hydroxybenzotriazole in N,N'-dimethylformamide. The peptide homologous to amino acids 6-17 was also extendedat theCOOH terminus with cysteine so that it could potentially be conjugated through that residue to a carrier protein. Deprotection and release of the synthetic peptides from the phenylacetamidomethyl resin was performed using H F with 10% anisole or thioanisole a t 0 "Cfor 1-2 h. Following ethyl acetate extractionof the deprotected peptide, the peptide was solubilized in 10% aqueous acetic acid, separated from the resin by filtration, and lyophilized. Structural analysis of the synthesized peptides was done using reversed phase HPLC and aminoacid analysis. Preparation of CarrierProtein-Peptide Conjugates-Conjugates were formed by adding the solid N-chloroacetyl peptide directlyto a freshly prepared solutionof a n iminothiolane-modified carrier protein using Traut's reagent. Priortothepeptideconjugation reaction, excess of the reagent was removed from the protein by gel filtration using Sephadex G-10 or G-25. Typical reaction conditions were the following: 30 mg of carrier protein (bovine serum albumin) were dissolved in 2 ml of 0.1 M NaHC03, pH 8.0. To thissolution 4 mg of solid iminothiolane (Traut's reagent) was added and stirred for 15 min a t room temperature, followed by chromatographic separation on a column (1.5 X 10 cm) containing Sephadex G-25 equilibrated andelutedwith 0.1 M NaHC03. To thefractionscontainingthe modified protein (totalvolume, about 3.5 ml), as judged by absorbance

readings at 280 nm, the N-chloroacetyl-derivatizedpeptide (lo-* M corresponding to about20-50 mg of peptide, depending on itsmolecular weight)was added in solid form and stirred for 3h a t room temperature. The pHwas often adjusted tobetween 7.5 and 8.0 using solid NaHC03. Following conjugation the peptide-proteinconjugates were dialyzedexhaustively against 0.1 M NH,HC03 a t 4 "C and lyophilized. Immunization of Mice-Five NNIH(S) mice were immunized subcutaneously with 50 pg of synthetic peptide conjugated to bovine serum albumin which had been made into anemulsion with Freund's complete adjuvant. After 3 weeks the mice were boosted with 50 pg of synthetic peptide conjugated to bovine serum albumin which had been made into an emulsion with Freund's incomplete adjuvant. After 2 weeks the animals were bled and sera prepared. Equal aliquots of serum from the five mice were pooled and heated at 56 "C for 30 min before use. RESULTS

The A subunit of pertussis toxinwas exposed to trypsin for limited periods of time. The reaction was terminated by the addition of soybean trypsin inhibitor and the reaction products were analyzed by SDS-gel electrophoresis rununder reducing conditions. As shown in Fig. 1, the A subunit was cleaved to a 20,000-Da fragment and an18,000-Da fragment. After prolonged digestion of the A subunit with trypsin, only the 18,000-Da fragment is seen (Fig. 2), suggesting that the 18,000-Da fragment is derived from the 20,000-Da fragment. The A subunit was less susceptible to tryptic digestion when it was complexed with the B oligomer in the holotoxin molecule (Fig. 2). The B oligomer was resistant to proteolysis either alone or when complexed with the A subunit. Other proteases also cut the A subunit into discreet fragments. When the A subunit was exposed to chymotrypsin, a major fragment of 23,000 Da was produced whereas several major fragments of 26,000-21,000Da were produced when the A subunit was exposed to S. aureus V8 protease (Fig. 3). When the A subunit was exposed to eitherof these proteases along with trypsin, the proteolytic pattern was identical to that produced with trypsin alone (Fig. 3). Control experiments demonstrated that chymotrypsin, S. aureus V8 protease, and trypsin could not be detected on the gel at the concentrations used in this experiment (data notshown). Antibodies which had been raised to synthetic peptides which correspond to regions of the A subunit were used to determine from which portions of the molecule the tryptic 66K

-

45K

-

25K

-

14K

-

0

2

4

6

8

1

0

Minutes FIG. 1. T i m e c o u r s eof tryptic hydrolysis of A subunit. The A subunit (2.4 pg) was incubated with trypsin (5pg/ml) a t 35 "C in a total volume of 22 pl of 50 mM sodium phosphate, pH7.5, containing 2 M urea. Reactions were stopped at the indicated times by addition of soybean trypsin inhibitor (10 pg/ml) and samples were subjected to SDS-gel electrophoresis. Positions of molecular weight standards are indicated.

Structural Characterization of Pertussis Toxin A Subunit A subunit

0

15

B Oligomer

25 0

15

45

A

PT

0

15

17679

A'

45

Time (minutes) FIG.2. Trypsin treatment of pertussis toxin and its components. The A subunit (5pg), B oligomer (15pg), or pertussis toxin (20pg) were treated with trypsin (5pg/ml) at 35 "C in a total volume of 32 pl of 50 mM sodium phosphate, pH 7.5, containing 2 M urea. Reactions were terminated at the indicated times by the addition of SDS and sampleswere subjected to electrophoresis.

FIG.4. Immunoblot of A subunit and trypsin-treated A (A*). Samples of A subunit (0.8 pg) or A subunit (1.5pg), A*, which had been digested with trypsin (5 pg/ml for 11 min a t 35 "C; reaction stopped by addition of SDS) were subjected to SDS-gel electrophoresis and subsequently transferred to nitrocellulose strips. Antibodies which had beenraised tosyntheticpeptides homologous tothe indicated regions of the A subunit were tested for their ability to bind to A subunit or A* using immunoblot techniques as described under "Experimental Procedures." oligomer to inhibit theNAD glycohydrolase activity of the A subunit was abolished after treatment of the A subunit with trypsin (Table I). When the proteolyzed A subunit was mixed with B oligomer, no holotoxin formation was observed if the FIG.3. Proteolytic treatment of A subunit. The A subunit (2.5 mixture was analyzed by gel electrophoresis at pH4 (Fig. 6). pg) was treated with trypsin (5pg/ml), chymotrypsin (10 pglml), or Only the B oligomer is seen on the gel as expected if no V8 protease (10pglml) in a total volume of 20 pl of 50 mM sodium reassociation occurred since free A subunit cannot be visualphosphate, pH 7.5,containing 2 M urea. After incubation a t 35 "Cfor ized using this acidic gel system (15). In contrast, holotoxin 7 min, the reactions were stopped by the addition of SDS sample subunit was buffer and then analyzed by SDS-gel electrophoresis. Lane 1 , no wasreadilyformedwhennon-proteolyzedA protease treatment; lane 2, treatment with trypsin; lane 3, treatment mixed with B oligomer. with chymotrypsin; lane 4, treatment with chymotrypsin and trypsin; Since the A subunit of pertussis toxin exhibits aminoacid lane 5, treatment with V8 protease; lane 6, treatment with V8 protease sequence homologies with the A subunit of cholera toxin, and trypsin. antibodies raised to synthetic peptides corresponding to sequences in the A subunit of pertussis toxin were tested for fragments were derived. Immunoblotanalysis shows that their ability to bind to the A subunit of cholera toxin. Imantibodies raised to synthetic peptides withsequences corre- munoblot analysis indicated that antibodies to a synthetic sponding toresidues 6-17, 168-182, and 201-210 all bound to peptide corresponding to residues 6-17 bound to theA subunit the A subunit. In contrast, only antibodiesraised to peptides of cholera toxin (Fig. 7) whereas antibodies to peptides corhomologous to residues 6-17 and 168-182 bound to the trypticresponding toresidues 168-182 or 201-210 did not bind (data fragments of the A subunit (Fig. 4). not shown). Anti-(6-17) antibodies also bound to the native Tryptic Fragments Retain NAD Glycohydrolase Activityform of the A subunit of cholera toxin aswell as theholotoxin The large 20,000-Da and 18,000-Da tryptic fragments were as measured by ELISA (Table 11). separated from smaller peptides by SDS-gel electrophoresis run under reducingconditions. A s shownin Fig. 5, these DISCUSSION fragments exhibited NAD glycohydrolase activity. Tryptic cleavage of the A subunit affected its interactions Several discreet protein fragments are produced when the with the B oligomer. While cleavage of the A subunit did not A subunit of pertussis toxin is exposed to trypsin for short alter its NAD glycohydrolase activity, the ability of the B periods of time. Since the A subunit does not contain lysine

Structural Characterizationof Pertussis Toxin A Subunit

17680 1I 100

75 50

25 n

0 X

z V

100

PT-

75

50

0.3

0.4

0.5 RI

0.6

0.7

FIG.5. NAD glycohydrolase activity of the A subunit and t r y p t i c f r a g m e n t s o f A subunit. The A subunit (12.5 pg) or A subunit (12.5pg) which had been digested with trypsin (5pg/ml for 10 min at 35 “C; reaction stopped by addition of SDS) were subjected to SDS-gel electrophoresis and transferred to nitrocellulose. Nitrocellulose strips (6-mmwide) were cut into3-mm lengths. The protein on the nitrocellulose strips wasassayed forNAD glycohydrolase activity as described under “Experimental Procedures.” Assays were performed in quadruplicate. The amount of [“Clnicotinamide produced during the assay is plotted versus the relative mobility of the protein species as it migrated as compared to the dye front on the SDS gel for the A subunit (panel A ) and trypsin-treated A subunit ( p a n e l B).Duplicate samples which were stained for protein indicated that theA subunit migrated on the SDS gel with an %of 0.33 whereas the two tryptic fragments migrated with Rf values of 0.48 and 0.50.

D

D

TABLE I Effect of B oligomer on the NADglycohydroluse activity of A subunit and trypsin-treated A subunit The A subunit (14.5pg) was digested with trypsin (5pg/ml) for 12 min a t 35 “Cin a volume of 110 pl, and the reaction was stopped with soybean trypsin inhibitor to give a final concentration of 10 pg/ml. FIG.6. Reassociation of subunits. The A subunit (1.5 pg) or Another sample of A subunit (14.5pg) was treated with a complex of trypsin and soybean trypsin inhibitor such that no digestion of A trypsin-treated A subunit (1.5 pg), A*, prepared as described in the subunit occurred. Samples of digested A subunit (A*) or nondigested legend to Table I, was mixed with B oligomer (7.5 pg) in a total A subunit (5pg each) were mixed with B oligomer (25 pg) in a total volume of 20 pl of 50 mM sodium phosphate buffer, pH 7.5,containing volume of 68 pl of 50 mM sodium phosphate buffer, pH 7.5,containing 2 M urea. After incubation a t room temperature for 15 min, the 2 M urea. After incubation at room temperature for 15 min, triplicate preparations were analyzed by nondenaturing gel electrophoresis a t assays for NAD glycohydrolase activity were performed with A sub- pH 4.0.Arrows indicate the positions towhich pertussis toxin andB unit (0.35 pg) or A* (0.35pg) either in the absence or presence of B oligomer migrated. oligomer (1.75pg).

+ i IJI,

Protein

NAD glycohydrolase activity

nmolfrninfny!A subunit

A A+B A*” A* + B a Trypsin-treated A subunit.

1.7 0.1

2.0 1.9

residues (18,19), these fragments must generated be by cleavage at arginine residues. The two major fragments observed after proteolysishave molecular weights of approximately 20,000 and 18,000 as determined by SDS-gel electrophoresis

run under reducing conditions. Since both the20,000-Da and 18,000-Da fragments bindantibodies which were raised against syntheticpeptides identical to residues 6-17 and 168182 of the native protein but not to antibodies raised against residues 201-210, trypsinmustcutthe molecule approximately 40-50 residues from its COOH-terminal end. Because of discrepancies in thetwo published amino acid sequences of the A subunit in the region from residues 190-200 (18, 19), the exactlocations of tryptic cleavage in this area are difficult to determine. The COOH-terminal end of the molecule appears to be the portion of the molecule which is most susceptible to proteolysis. In addition to trypsin,chymotrypsin and V8 protease (a

Structural Characterization of Pertussis Toxin

PT

CT

A Subunit

17681

protect it from proteolysis, or the B oligomer may alter the conformation of the A subunit such that the COOH-terminal portion is less exposed and therefore less susceptible to proteolysis. Proteolytic cleavage at the COOH-terminalend of the molecule affects the abilityof the A subunit to interactwith the B oligomer. Thus, it seems possible that the COOHterminal endof the A subunit may be directlyinvolved in the binding of the A subunit to the B oligomer. Alternatively, integrity of the COOH-terminal endof the A subunit may be necessary for maintaining the conformation of the A subunit which readily interacts with theB oligomer. While the ability of the A subunit to interact with the B oligomer is altered after proteolysis withtrypsin, themolecule still exhibits NAD glycohydrolase activity (Table I, Fig. 6). Even after theproteolyzed A subunit was subjected to SDSgel electrophoresis in order to separate thelarge 18,000- and 20,000-Da fragments from smaller peptides, the large fragments retained NAD glycohydrolase activity (Fig. 5). These results demonstrate that the residues which are necessary for NAD glycohydrolase activity are not located in the 40-50 amino acids at theCOOH-terminal end of the molecule. The results described suggest that the structure of the A subunit of pertussis toxin is similar to that of the A subunit of cholera toxin. The A subunit of cholera toxin is also an ADP-ribosyltransferase (30, 31), but the targets for the two toxins differ. Pertussis toxin ADP-ribosylates a cysteine residue on transducin (32), whereas cholera toxin ADP-ribosyFIG. 7. Immunoblot of pertussis toxin (PT)and cholera toxin ( C T ) . Samples of pertussistoxin (4.3 pg) or cholera toxin (2.5 lates an arginineresidue (33). The A subunit of cholera toxin pg) were subjected to SDS-gel electrophoresis and then transferred is synthesized as a singlepolypeptide chain and is subseto nitrocellulose strips. The strips were either stained for protein with quently cleaved by a protease to form the A, (residues 1-194) for the and A2 (residues 195-240) chains which are connected by a Amido Black or were examined using immunoblot techniques ability of antibodies, which had been raised to a synthetic peptide disulfidebond (34-36). The A, chaincontainstheADPhomologousto residues 6-17 ofthe pertussis toxin A subunit, to bind ribosyltransferase activity of the toxin (36) whereas the A2 to the proteins immobilizedon the strip. chain which is required for reassembly of cholera toxin from its subunits hasbeen postulated to anchor the A, chain to the TABLEI1 B oligomer (35).The datadescribed in this papersuggest that Binding of anti-(6-17) antibodies to the A subunits of pertussis and just asin the A subunit of cholera toxin, a region exists in the cholera toxins The ability of antibodies raisedto a synthetic peptide homologous A subunit of pertussis toxin near the COOH-terminal end to residues 6-17 of the A subunit of pertussis toxin to bind to toxins which is susceptible to proteolysis. The COOH-terminal ends and their components was analyzed byELISA as described under of both toxinA subunits seem to be important for maintaining "Experimental Procedures." the oligomeric structure of the toxins. Moreover, approxiProtein ELISA titer mately the first 200 residues of the A subunit of pertussis toxin exhibitNAD glydohydrolase activity and therefore conPertussis toxin 9,540 A subunit 8,240 tain the NAD-binding site as well as residues which play a ND" B oligomer role in this enzymatic reaction. Thus the NH2-terminal region of the A subunit of pertussis toxin is similar in this respect Cholera toxin 6,327 to theA, polypeptide chain of cholera toxin. A subunit 10,070 In addition to these gross structural similarities, the A ND B oligomer subunits of pertussis toxin and cholera toxin have been shown a ND, not detectable. to have sequence homology (18, 19,22). Oneof these regions protease which cleaves after acidicresidues) digest the A of homology is shown in Fig. 8. Seven of the 12 residues in subunit into discreet fragments. Since combination of these the region corresponding toresidues 6-17 of the A subunit of region correspondproteases with trypsin gives the digestion pattern observed pertussis toxin areidentical to those in the ing to residues 4-15 of the A subunit of cholera toxin. We with trypsin alone, all proteases must be cleaving the same have found that antibodies raised to amino acids 6-17 of the end of the protein. This interpretation is supported by the A subunit of pertussis toxin bind to both the potentially finding that anti-(6-17) antibodies bind to the major frag(as measments generatedby chymotrypsin andV8 protease suggesting denatured (as measuredby immunoblot) and native ured by ELISA) forms of the A subunits of pertussis and that the large fragments contain the NH2-terminalregion of the molecule (data not shown). Therefore, the COOH-termi- cholera toxins. Thus, both toxinsmay have similar structures nal endof the A subunit islikely the mostexposed portion of PT (6-17) Thr-Val-Tyr-Arg-Tyr-Asp-Ser-Arg-Pro-Pro-Glu-Asp the molecule. I I I I I I I CT (4-15) Lys-Leu-Tyr-Arg-Ala-Asp-Ser-Arg-Pro-Pro-Asp-Glu The isolated A subunit is digested by trypsin more rapidly than the A subunit which is associated with the B oligomer. FIG. 8. Sequence homology between residues 6-17 of the A These resultssuggest that theB oligomer may either interact subunit of pertussis toxin (PT)and residues 4-15 of the A directly with the COOH-terminal portionof the A subunit to subunit of cholera toxin(CT)(from Refs. 18 and19).

17682

Structural Characterization of Pertussis ToxinA Subunit

17. Moss, J., and Vaughan, M. (1979)Annu. Rev. Biochem. 48,581600 18. Locht, C., and Keith, J. (1986)Science 232, 1258-1264 19. Nicosia, A., Perugini, M., Franzini, C., Casagli, M., Borri, M., Antoni, G., Almoni, M., Neri, P., Ratti, G., and Rappuoli, R. (1986)Proc. Natl. Acad. Sci. U. S. A. 83, 4631-4635 20. Katada, T., Tamura, M., and Ui, M. (1983)Arch. Biochem. Biophys. 224,290-298 21. Moss, J., Stanley, S. J., Burns, D.L., Hsia, J. A., Yost, D. A., Myers, G. A., and Hewlett, E. L. (1983)J. Biol. Chem. 268, 11879-11882 REFERENCES 22. Capiau, C., Petre, J., Van Damme, J., Puype, M., and Vandek1. Manclark, C. R., and Cowell, J. L. (1984)in Bacterial Vaccines erckhove, J. (1986)FEBS Lett. 204,336-340 (Germanier, R., ed) pp. 69-106,Academic Press, New York 23. Burns, D. L., Kenimer, J. G., and Manclark, C. R. (1987)Infect. 2. Katada, T., and Ui, M. (1982)Proc. Natl. Acad. Sci. U. S. A. 79, Immunol. 65,24-28 3129-3133 24. Laemmli, U. K. (1970)Nature 227,680-685 3. Katada, T., and Ui, M. (1982)J. Bwl. Chem. 267,7210-7216 25. Gabriel, 0. (1971)Methods Enzymnl. 22, 565-578 4. Gilman, A. G. (1984)Cell 36,577-579 26. Bradford, M. M. (1976)A d . Biochem. 72,248-254 5. Bokoch, G. M., Katada, T., Northup, J. K., Ui, M., and Gilman, 27. Towbin, H., Stachelin, T., and Gordon, J. (1979)Proc. Natl. Acad. A. G. (1984)J. Biol. Chem. 269, 3560-3567 Sci. U. S. A . 76,4350-4354 6. Manning, D. R., and Gilman, A. G. (1983)J. Biol. Chem. 258, 28. Carroll, S. R., Lory, S., and Collier, R. J. (1980)J. Bwl. Chem. 7059-7063 268,1435-1438 7. Sternweis, P. C., and Robishaw, J. D. (1984)J. Biol. Chem. 269, 29. Manclark, C.R., Meade, B.D., and Burstyn, D. G. (1986)in 13806-13813 Manual of Clinical Laboratory Immunology (Rose, N. R., Fried8. Hazeki, O., and Ui, M. (1981)J. Bwl. Chem. 266, 2856-2862 man, H., and Fahey, J. L., eds) 3rd Ed., American Society for 9. Burns, D. L., Hewlett, E. L., Moss, J., and Vaughan, M. (1983) Microbiology, Washington, D. C. J. Bwl. Chem. 268,1435-1438 30. Gill, D. M., and Meren, R. (1978)Proc. Natl. Acad. Sci. U. S. A . 10. Okajima, F., and Ui, M. (1984)J. Bwl. Chem. 269, 13863-13871 76,3050-3054 11. Nakamura, R., and Ui, M. (1985)J. Biol. Chem. 260,3584-3593 31. Cassel, D., and Pfeuffer, T. (1978)Proc. Natl. Acad. Sci.U. S. A. 12. Bokoch, G., and Gilman, A. (1984)Cell 301-308 39, 76,2669-2673 13. Kikuchi, A., Kozawa, O., Kaibuchi, K., Katada, T., Ui, M., and 32. West, R. E., Jr., Moss, J., Vaughan, M., Liu, R., and Liu, T.-Y. Takai, Y. (1986)J. Biol. Chem. 261, 11558-11562 (1985)J. Bbl. Chem. 260,14428-14430 14. Asano, T., Ui, M., and Ogasawara, N. (1985)J. Biol. Chem. 260, 33. Van Dop, C., Tsubokawa, M., Bourne, H., and Ramachandran, 12653-12658 J. (1984)J. Biol. Chem. 259,696-698 15. Tamura, M., Nogimori, K., Murai, S., Yajima, M., Ito, K., Katada, 34. Mekalanos, J. J., Swartz, D. J., Pearson, D. N., Harford, N., T., Ui, M., and Ishii, S. (1982)Biochemistry 21,5516-5522 Groyne, F., and deWilde, M. (1983)Nature 306,551-557 16. Gill, D. M. (1978)in Bacterial Toxins and Cell Membranes (Jel- 35. Gill, D. M. (1976)Biochemistry 15,1242-1248 jaszewica, J., and Wadstrom, T., eds) pp. 291-332, Academic 36. Mekalanos, J. J., Collier, R. J., and Romig, W. R. (1979)J. Biol. Chem. 264,5855-5861 Press, New York

in this region. Since both toxins share a common substrate, NAD, it ispossible that thehomologous regions may be those which interact with NAD. These resultssuggest that these two ADP-ribosylating toxins may have more structuralsimilarities than previously thought. Information obtained concerning the structure of one of these toxins may therefore shed light on the structure of the other.