Proc. Natl. Acad. Sci. USAVol. 82, pp. 2637-2641, May 1985Biochemistry

Pertussis toxin inhibition of chemotaxis and the ADP-ribosylationof a membrane protein in a human-mouse hybrid cell line

(fMet-Leu-Phe/adenylate cyclase/GTPase/guanine nucleotide binding proteins)

PETER S. BACKLUND, JR.*, BRUCE D. MEADEt, CHARLES R. MANCLARKt, GIULIO L. CANTONI*,AND ROBERT R. AKSAMIT*t*Laboratory of General and Comparative Biochemistry, National Institute of Mental Health, and tDivision of Bacterial Products, Center for Drugs andBiologics, Food and Drug Administration, Bethesda, MD 20205

Contributed by Giulio L. Cantoni, December 26, 1984

ABSTRACT When WBC264-9C cells are preincubatedwith pertussis toxin, chemotaxis is inhibited and ADP-ribosylation of a membrane protein with a subunit Mr 41,000is observed. Both the inhibition of chemotaxis and the ADP-ribosylation by pertussis toxin display a similar time lag,temperature dependence, and pertussis toxin-concentrationdependence. Although the inhibition of chemotaxis and theADP-ribosylation of the membrane protein are qualitativelycorrelated, nearly complete inhibition of chemotaxis occurswhen there is only partial ADP-ribosylation of the membraneprotein. Pertussis toxin-catalyzed ADP-ribosylation of the Mr41,000 protein in WBC264-9C membranes is stimulated byGDP, GTP, and to a lesser extent by GMP; the nonhydrolyz-able GTP analog guanosine 5'-[fB,-imidoltriphosphate has noeffect. WBC264-9C membranes have a high-affinity GTPaseactivity, which is partially inhibited in membranes frompertussis toxin-treated cells. Neither GTPase activity noradenylate cyclase activity in membranes from WBC264-9Ccells is affected by fMet-Leu-Phe, an attractant for these cells.Our results suggest that a guanine nucleotide binding proteinmay be involved in chemotaxis, but they do not indicate aninvolvement of adenylate cyclase.

In contrast to bacterial chemotaxis, where the use of ad-vanced genetic and biochemical techniques has led to a fairlydetailed understanding of the bacterial sensory mechanism,progress on the biochemical mechanism of mammalian cellchemotaxis has been limited by the availability of experi-mental probes. The evaluation ofthe chemotactic behavior ofcells treated with various compounds that result in decreasedmethylation is one approach that has been applied to bothbacteria and mammalian cells (14). These studies have led tothe conclusion that methylation reactions are involved inchemotaxis in both bacteria and mammalian cells.The discovery by Schiffmann et al. (5) that N-formyl

peptides are attractants for rabbit neutrophils and guinea pigmacrophages provided a probe with which to study detailedbinding ofan attractant to a mammalian cell receptor. Severalstudies of the binding of N-formyl peptides to cells ormembranes have indicated that the receptor may exist in twoaffinity states (6-8) and that the binding can be affected byguanine nucleotides (9, 10). These characteristics are remi-niscent of the regulation of adenylate cyclase by guaninenucleotide binding proteins (11) and may indicate that theseor similar regulatory proteins are involved in chemotaxis.The possible involvement of a guanine nucleotide bindingprotein in chemotaxis is supported by the recent discoverythat pertussis toxin is a potent inhibitor of macrophagechemotaxis (12). Pertussis toxin is known to ADP-ribosylate

at least three nucleotide-binding proteins: transducin, whichincreases the activity of cyclic GMP phosphodiesterase inretina (13); Ni, which inhibits the activity ofadenylate cyclase(11); and No, whose function and relationship to Ni is un-known (14-16).

In this paper, we provide evidence that there is a qualita-tive correlation between the inhibition of chemotaxis bypertussis toxin, the entry of toxin, and ADP-ribosylation ofa single membrane protein of Mr -41,000. We have chosenas an experimental system a hybrid cell line that exhibitschemotaxis to fMet-Leu-Phe.

METHODS

Cell Culture and Chemotaxis. The RAW264 (17) and WBC-264-9C cell lines were cultured in minimal essential medium(Eagle) containing 10% heat-inactivated fetal calf serum,penicillin (100 international units/ml), and streptomycin (100,ug/ml) as described previously (18).Chemotaxis of cells across a Nuclepore polycarbonate

filter was determined in multiwell chemotaxis chambers asdescribed (4, 18). Cell migration was assessed by measure-ment of the optical density of filters stained with Diff-Quick(Harleco, Gibbstown, NJ) (unpublished method). The opticaldensity is proportional to cell number. Assays were per-formed in triplicate and the standard error of the mean was<10%.The preparation of endotoxin-activated mouse serum

(EAMS) (12), pertussis toxin (19), and antibody to pertussistoxin (20) has been published previously.

Preparation of WBC264-9C Membranes. WBC264-9C cellswere centrifuged at 250 x g for 5 min and the cell pellet wasfrozen at -70°C. The rest of the steps were all carried out at0°C. The frozen cell pellet was suspended in a hypotonicmembrane buffer (10 mM Tris Cl, pH 7.5/5 mM MgCl2), andthe cells were homogenized with 15 strokes in a Ten Broecktissue grinder. The suspension was centrifuged for 10 min at500 x g to remove nuclei and unbroken cells, and thesupernatant fluid was centrifuged at 40,000 x g for 20 min.The membrane pellet was suspended in membrane buffer at1-5 mg of protein/ml. Protein concentrations were deter-mined by the method of Bradford (21), using bovine y-globulin as the standard.

[32P]ADP-Ribosylation with Pertussis Toxin. ADP-ribosylation by pertussis toxin was determined from theincorporation of 32p from [adenylate-32P]NAD (New EnglandNuclear) into membrane proteins. The standard ADP-ribosylation reaction was carried out in a volume of 0.05 mlcontaining 0.1 M potassium phosphate (pH 7.0), 10 mM

Abbreviations: EAMS, endotoxin-activated mouse serum; Ni, theinhibitory protein of adenylate cyclase.tTo whom reprint requests should be addressed at: NIMH, Bldg. 36,Rm. 3006, 9000 Rockville Pike, Bethesda, MD 20205.

2637

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

2638 Biochemistry: Backlund et al.

thymidine, 1 mM ATP, 0.1 mM GTP, 25 ILM [32P]NAD (sp.activity =1.6 x 103 mCi/mmol; 1 Ci = 37 GBq), activatedpertussis toxin (8 ug/ml), and cell membranes (50 pg ofprotein). Pertussis toxin was activated by incubation of thetoxin for 30 min at room temperature in 50 mM potassiumphosphate, pH 7.0/25 mM dithiothreitol. After addition ofmembranes, the reaction mixture was incubated at 300C for45 min, at which time the reaction had reached completion.Ice-cold membrane buffer (0.5 ml) was added, and themembranes were collected by centrifugation. Under theseconditions, [32P]NAD and pertussis toxin are in excess, andthe Mr 41,000 peptide is the limiting substrate in the reaction.The incorporation of 32p into membrane proteins was

determined by separation ofthe proteins by NaDodSO4/10%oPAGE (22), followed by autoradiography with X-Omat AR-2film (Kodak). In some cases, 32P-labeled membrane proteinswere separated by two-dimensional PAGE according toO'Farrell (23).GTPase and Adenylate Cyclase Assays. High-affinity

GTPase activity in membranes (10 pg of protein) was deter-mined by the release of32P; from [y-32P]GTP in a 10-min assayby the method ofKoski et al. (24), except that no ouabain wasincluded in the reaction mixture (0.1 ml). Corrections fornonspecific GTPase activity were made by subtracting theactivity observed in the presence of 0.1 mM GTP from thetotal GTPase activity. Nonspecific hydrolysis of GTP ac-counted for =40% of the total 32P; released.Adenylate cyclase was determined in membranes by the

conversion of [a-32P]ATP to [32P]cAMP in the presence of 1,uM GTP and 5 mM MgCl2 in a 15-min assay. The labeledproduct was separated by column chromatography on AG50W-X4 (Bio-Rad) and then on alumina (25).

RESULTSWBC264-9C is a hybrid cell line isolated from a fusion ofhuman peripheral leukocytes with a thioguanine-resistantvariant of the mouse macrophage cell line RAW264. Theisolation and characterization ofthe WBC264-9C cell line willbe described elsewhere. The distinctive feature of the WBC-264-9C cell line is that it exhibits chemotaxis to fMet-Leu-Phe, whereas the RAW264 mouse macrophage line does not.

x

co

E

4-5c0

0)

2

0.

120

100

80

60

40

20

O' 10 100

Pertussis Toxin, ng/ml

FIG. 1. Inhibition of chemotaxis by pertussis toxin. Pertussistoxin was added to a suspension of WBC264-9C or RAW264 cells,and the cells were immediately added to chemotaxis chamberscontaining either 20 nM fMet-Leu-Phe (FMLP) or a 1:100 dilution ofEAMS as an attractant. After 4 hr at 37°C, migration of cells across

the filter was measured. o, Cell migration in the absence ofattractant.

0

E080

C6060

PertussisToxin, ng/mlAM

0

40-20

0 0.1 1 10 100

Pertussis Toxin, ng/mI

FIG. 2. Inhibition of chemotaxis after preincubation ofWBC264-9C cells with pertussis toxin. WBC264-9C cells were preincubatedwith pertussis toxin at either 0°C or 37°C for 1 hr and then were addedto chemotaxis chambers containing either 20 nM fMet-Leu-Phe(FMLP) or a 1:100 dilution of EAMS as an attractant. After 4 hrat 3PC, migration of cells across the filter was measured. o, Cellmigration in the absence of attractant.

Both the WBC264-9C and RAW264 cell lines exhibitchemotaxis to EAMS, an unpurified attractant generated inmouse serum. When pertussis toxin and WBC264-9C orRAW264 cells were mixed and immediately placed inchemotaxis chambers, inhibition of chemotaxis to EAMSwas observed for both cell lines, although RAW264chemotaxis was more sensitive to inhibition (Fig. 1). Inhibi-tion of WBC264-9C chemotaxis to fMet-Leu-Phe was notdetected. However, after a 1-hr preincubation ofWBC264-9Ccells with pertussis toxin at 37°C, chemotaxis both to fMet-Leu-Phe and to EAMS was inhibited with a sensitivitycomparable to that observed for RAW264 to EAMS (Fig. 2).The time dependence and attractant specificity of the

inhibition of chemotaxis can be explained by two sets ofobservations. First, the attractant-dependent kinetics ofmigration of RAW264 and WBC264-9C cells are different. Aprevious study (18) has shown that <5% of RAW264 cellscross the chemotaxis membrane in response to EAMS duringthe first 2 hr; and we show here that in 1 hr 45% of theWBC264-9C cells migrated to EAMS, whereas migration tofMet-Leu-Phe was essentially complete (Table 1). The com-plete migration of WBC264-9C to fMet-Leu-Phe in 1 hr wasin accord with the observation that, after preincubation withpertussis toxin, WBC264-9C chemotaxis to fMet-Leu-Phewas inhibited to the same extent in either 1-hr (data notshown) or 4-hr chemotaxis assays (Fig. 2). Second, as we willshow below, entry of the toxin was required for inhibition ofchemotaxis, and it has been shown in other studies that toxinentry and activation require approximately an hour (26, 27).Therefore, during the time required for the entry and activa-tion of pertussis toxin, a larger percentage of the totalmigrating WBC264-9C cells will cross the filter and be scored

Table 1. Time required for WBC264-9C cells to migrate toEAMS or fMet-Leu-Phe

Cell migration, Percent ofoptical density* cells migrated

Attractant 1 hr 4 hr after 1 hr

EAMS (1:100) 0.167 0.372 44.9fMet-Leu-Phe (20 nM) 0.129 0.136 94.9

*See Methods.

W-9C264-9CFMLP)

WBC264-9C(EAMS)

RAW264MEAMS)....I . . I . .. ...

Proc. Natl. Acad. Sci. USA 82 (1985)

1

Proc. NatL. Acad. Sci. USA 82 (1985) 2639

as chemotactic when the attractant is fMet-Leu-Phe thanwhen the attractant is EAMS. Preincubation ofRAW264 cellswith pertussis toxin is not required to observe inhibition ofchemotaxis, since the lag before migration of RAW264 cellsis approximately the same as the time for entry of the toxin.The temperature dependence of the inhibition of

chemotaxis by pertussis toxin indicates that entry ofthe toxinis required. Pertussis toxin was mixed with WBC264-9C cellsand preincubated at either0C or 370C for 1 hr before the cellswere placed in chemotxis chambers. WBC264-9C cells pre-incubated with pertussis toxin at 370C showed increasedsensitivity to inhibition of chemotaxis when either EAMS orfMet-Leu-Phe was the attractant (Fig. 2). In contrast, afterpreincubation at 0C, which should inhibit entry of the toxinbut not binding, the sensitivity of WBC264-9C cells toinhibition of chemotaxis by pertussis toxin was not increasedbut, in fact, was similar to that of cells that had not beenpreincubated (Figs. 1 and 2). Experiments with antibodiesagainst pertussis toxin provided another indication that theentry of pertussis toxin was required. When antibodies aremixed with the toxin and then added to cells, inhibition ofchemotaxis is prevented (12). However, if antibody wasadded at various times during a 1-hr preincubation withWBC264-9C cells and pertussis toxin at 37°C, the antibodywas effective in preventing inhibition of chemotaxis onlywhen added during the first 15 min (data not shown).

After entry into cells, pertussis toxin is known to catalyzeADP-ribosylation of the a subunit of guanine nucleotidebinding proteins, such as transducin (13) and the inhibitoryprotein of adenylate cyclase, N; (28, 29). When membranesof WBC264-9C cells were incubated with [32PJNAD in thepresence of pertussis toxin, a single protein was labeledwhich migrated upon NaDodSO4/PAGE in the presence of2-mercaptoethanol as a Mr 41,000 peptide (Fig. 3). This is thesame Mr reported for the a subunit of N; from a number ofsources (28, 30-35). When [32P]ADP-ribosylated membraneswere separated by two-dimensional gel electrophoresis, this

1 2 3 4 5

Mr41,000O

FIG. 3. Effect of guanine nucleotides on ADP-ribosylation ofWBC264-9 membranes by pertussis toxin. Membranes from WBC-264-9C cells were incubated with [32P]NAD and pertussis toxin for30 min. Membrane proteins (40 ug) were separated by NaDod-S04/PAGE, and the incorporation of 32p into membrane proteinswas detected by autoradiography (lane 1). The effect of guaninenucleotides on ADP-ribosylation was determined by incubation ofmembranes and [32P]NAD with 100 AM GTP (lane 2), guanosine5'-[,8,yimido]triphosphate (lane 3), GDP (lane 4), or GMP (lane 5).The apparent M, of the major labeled protein (41,000) was calculatedfrom its mobility relative to standards. Pertussis toxin-independentlabeling of a M, 100,000 protein was also observed as a minor band.

high-resolution separation method again showed only a singletoxin-dependent radioactive spot with Mr 41,000 (data notshown). Fig. 3 also shows that the labeling of the Mr 41,000peptide was greatly altered by the presence of guaninenucleotides. GTP greatly stimulated the labeling of thepeptide, whereas the nonhydrolyzable GTP analog guanosine5'-[P,y-imido]triphosphate had no effect on the ADP-ribosylation. GDP produced a greater stimulation of ADP-ribosylation than GTP, and GMP also stimulated the re-action, but to a lesser extent than GDP. The effects ofguaninenucleotides on chemotaxis were not determined becauseintact cells are not permeable to nucleotides.

If ADP-ribosylation of Ni or a similar protein causesinhibition of chemotaxis, then the enzymatic activity of thetoxin must occur in intact cells under conditions similar tothose which inhibit chemotaxis. To examine this question, weincubated intact cells with the toxin at various concentrationsand time intervals to determine the extent of ADP-ribosylation of the Mr 41,000 peptide. Since [32P]NAD doesnot enter intact cells, ADP-ribosylation in intact cells, whichis due to endogenous NAD, was determined indirectly inmembranes from pertussis toxin-treated cells by measuringthe decrease in the Mr 41,000 peptide that was available for[32P]ADP-ribosylation.When intact cells were incubated with pertussis toxin (5

ng/ml) at 370C, a lag time of 1 hr was observed for ADP-ribosylation of the Mr 41,000 peptide. After 2 and 3 hr,significant ADP-ribosylation was observed, with 64% of thepeptide ADP-ribosylated in intact cells after 3 hr (Table 2).The entry of the toxin into the cells and the subsequentADP-ribosylation of the Mr 41,000 peptide was also tempera-ture-dependent. ADP-ribosylation was significantly reducedbut not completely eliminated when the cells were incubatedwith the toxin for 2 hr at 0°C instead of at 37°C (Table 2).When intact WBC264-9C cells were incubated for 2 hr at 37°Cwith various concentrations of pertussis toxin, significantADP-ribosylation of the Mr 41,000 peptide was observed at atoxin concentration of 5 ng/ml, and at 50 ng/ml almost all ofthe peptide in intact cells was ADP-ribosylated (Fig. 4). Theresults of these experiments show that the concentrationdependence, temperature dependence, and time lag of ap-proximately 1 hr for ADP-ribosylation of the Mr 41,000peptide in intact cells are qualitatively correlated with theinhibition of chemotaxis by pertussis toxin.

Highly purified Ni is known to have a high-affinity GTPaseactivity (15, 36). When high-affinity GTPase activity was

Table 2. ADP-ribosylation of the membrane Mr 41,000 peptideand GTPase activity after incubation of cells withpertussis toxin

Incubation Peptide available for GTPase activity,

Time, Temp., [32P]ADP-ribosylation, pmol-min-1 mg-hr 0C % control of protein

0 100 64 11 37 98 54 52 37 68 57 83 37 36 42 31 0 107 60 ± 22 0 85 59 ± 2

Cells were incubated in minimal essential medium (Eagle) contain-ing 10%6 fetal calf serum and pertussis toxin (5 ng/ml) as indicated.Membranes were prepared from the cells, and the M, 41,000 peptidewas ADP-ribosylated to completion with [32P]NAD and pertussistoxin. Incorporation of 32p into the Mr 41,000 peptide was quantifiedby densitometry of the autoradiograms of membrane proteins sepa-rated by NaDodSO4/PAGE. The [32P]ADP-ribosylation reactionswere performed in duplicate, and the data were averaged andexpressed as percent of control. GTPase assays were performed induplicate and the data are presented as mean ± SEM.

Biochemistry: Backlund et al.

2640 Biochemistry: Backlund et al.

1 2 3 4 5

94-

67-

0 43-

30-

20.1 -dye front

PT with cells, ng ml 0 0 0.5 5.0 50

PT with membranes -

FIG. 4. Concentration dependence of pertussis toxin-catalyzedADP-ribosylation of Mr 41,000 peptide in intact WBC264-9C cells.Cells were incubated for 2 hr at 370C with the indicated concentra-tions of pertussis toxin (PT). Membranes were prepared and incu-bated for 45 min with [32P]NAD in the absence (lane 1) or presence(lanes 2-5) of activated pertussis toxin (8 pyg/ml) to 32P-label any M,41,000 peptide that had not been ADP-ribosylated in the intact cellsby the toxin. Membrane proteins were then separated by NaDod-S04/PAGE and the radioactivity was detected by autoradi-ography.

measured in WBC264-9C membranes, the activity was re-duced in the membranes of cells that were treated withpertussis toxin, and the decreased GTPase activity wascorrelated with the ADP-ribosylation of Mr 41,000 peptide(Table 2). It has also been reported that fMet-Leu-Phe canstimulate high-affinity GTPase activity in broken-cell prepa-rations of human neutrophils (37), suggesting that the fMet-Leu-Phe receptor in neutrophils may interact with Ni orrelated guanine nucleotide binding proteins. However, a widerange of fMet-Leu-Phe concentrations did not stimulate thehigh-affinity GTPase activity in WBC264-9C membranes(Table 3). Therefore, even though the chemotaxis ofWBC264-9C to fMet-Lue-Phe and the high-affinity GTPase activitywere inhibited by pertussis toxin under conditions where theMr 41,000 peptide was ADP-ribosylated, it does not appearthat the binding of the tripeptide to the N-formyl peptidereceptor has an effect on the high-affinity GTPase activity.

In human neutrophils, an increase in cAMP has beenreported when cells are incubated with fMet-Leu-Phe, sug-

Table 3. Membrane adenylate cyclase and GTPase activities

Adenylate cyclase, GTPase activity,pmol1min-'mg-1 pmol-min-lmg1

Ligand of protein of proteinNone 22 ± 1 77 ± 2fMet-Leu-Phe10-8M 23 ± 1 80 ± 310-7M 22 ± 1 73 ± 310-6M 22±1 76±410-5M 23± 1 75±4

Isoproterenol (10-5 M) 137 ± 9 ND

Adenylate cyclase and GTPase activities were determined formembranes prepared from WBC264-9C cells, and the data representmeans ± SEMs for triplicate determinations. Similar results wereobtained with two other membrane preparations from WBC264-9Ccells. ND, not done.

gesting that the N-formyl peptide receptor may be coupled toadenylate cyclase (3840). It has also been shown thatADP-ribosylation of N, by pertussis toxin attenuates theinhibition of adenylate cyclase when agonists bind to inhibi-tory receptors (28, 30-32). Since pertussis toxin also inhibitsthe chemotaxis of WBC264-9C cells to fMet-Leu-Phe underconditions where the toxin ADP-ribosylates the Mr 41,000peptide, it was of interest to determine whether adenylatecyclase was coupled to N-formyl peptide receptors. The datain Table 3 show that there was no change in adenylate cyclaseactivity when WBC264-9C membranes were exposed to10-8-10-5 M fMet-Leu-Phe, including those concentrations(10-8_10-7 M) at which optimal chemotaxis is observed (41).As a positive control for the cyclase assays, membranes wereincubated with isoproterenol, which produced a large stimu-lation of adenylate cyclase activity (Table 3). If the N-formylpeptide receptors were coupled to adenylate cyclase and Ni,then binding offMet-Leu-Phe would be expected to decreasebasal adenylate cyclase activity (24, 28). Since this inhibitionwas not observed in the WBC264-9C membranes, the datasuggest that fMet-Leu-Phe receptors in WBC264-9C cellsand chemotaxis to the tripeptide are not directly coupled toadenylate cyclase.A role for biological methylation in macrophage chemo-

taxis emerged from earlier studies in which it was shown that3-deazaadenosine, an adenosine analog capable of interactingwith S-adenosylhomocysteine hydrolase, will specificallyinhibit chemotaxis (4). The synthesis of specific proteins isinhibited when cells are treated with 3-deazaadenosine, andthis finding has led to the proposal that the continuoussynthesis ofone or more 3-deazaadenosine-sensitive proteinsis required for normal chemotaxis (42). Since both 3-deazaadenosine and pertussis toxin inhibit chemotaxis, anexperiment was performed to determine whether treatmentof cells with 3-deazaadenosine would change the amount ofthe M, 41,000 peptide available for ADP-ribosylation. Mem-branes were prepared from untreated and 3-deazaadenosine-treated RAW264 cells and in vitro ADP-ribosylation of equalamounts of the membranes by pertussis toxin was deter-mined. The amount of ADP-ribosylation was the samewhether or not the cells were treated with 3-deazaadenosine(data not shown), a finding that suggests the amount ofthe Mr41,000 peptide is not decreased in membranes of cellsincubated with 100 u&M 3-deazaadenosine for 3 hr.

DISCUSSION

Chemotaxis by the WBC264-9C cell line to two differentattractants is inhibited by pertussis toxin. Since a largefraction ofWBC264-9C cells migrate in 1 hr, preincubation ofthese cells with pertussis toxin is necessary to observecomplete inhibition. Upon incubation of the cells with toxin,a period of 15-60 min is required before chemotaxis isinhibited, and a similar lag in the ADP-ribosylation of a singleMr 41,000 membrane protein is observed. Presumably thistime is needed for the toxin to pass into or through the cellmembrane and to become activated for ADP-ribosyltrans-ferase activity. Although the temporal delay, temperaturedependence, and concentration dependence of the inhibitionof chemotaxis and of the ADP-ribosylation of the Mr 41,000membrane protein by pertussis toxin were qualitativelysimilar, inhibition of chemotaxis was quantitatively moresensitive. For example, the data in Table 2 indicate that only3040% of the membrane protein is ADP-ribosylated underconditions where chemotaxis is inhibited >80%. The dis-crepancy in the quantitative relationship between ADP-ribosylation and chemotaxis does not necessarily indicatethat the two are unrelated, since the details of the multistepchemotactic process are unknown.

Proc. Natl. Acad. Sci. USA 82 (1985)

Proc. Natl. Acad. Sci. USA 82 (1985) 2641

The major membrane substrate for activated pertussistoxin in WBC264-9C cells has a subunit of Mr 41,000, asdetermined by NaDodSO4/PAGE. In other systems, a mem-brane protein with the same subunit Mr is ADP-ribosylatedby pertussis toxin, and this protein has been identified as thea subunit of Ni, the inhibitory regulatory protein of adenylatecyclase. Recently, a second membrane protein has beenfound in brain (14, 15, 36) and adipocytes (16) that isADP-ribosylated by pertussis toxin and has a subunit of Mr39,000-40,000. The function of this protein (No) and itsrelationship to Ni are unknown.

In addition to ADP-ribosylation of the Mr 41,000 peptide,treatment of cells with pertussis toxin resulted in a smallinhibition of membrane GTPase activity. Although purifiedNi does have GTPase activity (15, 36), it is not knownwhether the inhibition of GTPase activity that we observedwas due to inhibition of Ni or other enzymes. Under theconditions we used in our studies, we have not observed anyeffects of fMet-Leu-Phe on either GTPase or adenylatecyclase activity.Guanine nucleotide binding proteins are essential compo-

nents in signal transduction of a variety of different stimulisuch as light (43) and hormones (11). By analogy, thecorrelation of ADP-ribosylation of the Mr 41,000 membraneprotein by pertussis toxin with inhibition of chemotaxissuggests a pivotal role for this guanine nucleotide bindingprotein in transduction of the chemotactic signal in macro-phages.

Ni is known to couple inhibitory receptors to adenylatecyclase. However, the data reported here do not show aninteraction between the Mr 41,000 peptide and adenylatecyclase, and we would like to suggest that the effectorenzyme(s) involved in transduction of chemotactic signalsmay be different from adenylate cyclase. This hypothesis isconsistent with recent reports that a pertussis toxin substratemay couple N-formyl peptide receptors to one or morephospholipases, resulting in the release of arachidonic acidand the mobilization of calcium (44 46). Furthermore, itshould be noted that a uniform concentration of fMet-Leu-Phe elicits several activities, including cell aggregation,degranulation, and superoxide production, whereas a gradi-ent of the tripeptide is required for chemotaxis. Therefore,more than one guanine nucleotide binding protein may berequired to regulate all of these functions. Biological activi-ties measured in different cell types can also respond differ-ently to occupation of the N-formyl peptide receptor.Harvath and Aksamit (47) have reported that oxidized fMet-Leu-Phe binds to the N-formyl peptide receptor in humanneutrophils without eliciting a chemotactic response, where-as human monocytes both bind and exhibit chemotaxis to thisoxidized peptide. Based on these findings, these authors havesuggested that the N-formyl peptide receptor complex ortransduction mechanism is different in human neutrophilsand monocytes. Additional studies will be needed to deter-mine the molecular correlates ofthe various cellular activitiesand the guanine nucleotide binding proteins.

1. Aswad, D. W. & Koshland, D. E., Jr. (1975) J. Mol. Biol. 97,207-223.

2. Pike, M. C., Kredich, N. M. & Snyderman, R. (1978) Proc.Nati. Acad. Sci. USA 75, 3928-3932.

3. Pike, M. C. & Snyderman, R. (1982) Cell 28, 107-114.4. Aksamit, R. R., Falk, W. & Cantoni, G. L. (1982) J. Biol.

Chem. 257, 621-625.5. Schiffmann, E., Corcoran, B. A. & Wahl, S. M. (1975) Proc.

Nati. Acad. Sci. USA 72, 1059-1062.6. Koo, C. R., Lefkowitz, R. & Snyderman, R. (1981) Biochem.

Biophys. Res. Commun. 106, 442-449.7. Mackin, W. M., Huang, C.-K. & Becker, E. L. (1982) J.

Immunol. 129, 1608-1611.8. Seligmann, B., Fletcher, M. & Gallin, J. I. (1982) J. Biol.

Chem. 257, 6280-6286.9. Koo, C. R., Lefkowitz, J. & Snyderman, R. (1983) J. Clin.

Invest. 72, 748-753.10. Snyderman, R., Pike, M. C., Edge, S. & Lane, B. (1984) J.

Cell Biol. 98, 1441-8.11. Gilman, A. G. (1984) Cell 36, 577-579.12. Meade, B. D., Kind, P. D., Ewell, J. B., McGrath, P. P. &

Manclark, C. R. (1984) Infect. Immun. 45, 718-725.13. Van Dop, C., Yamanaka, G., Steinberg, F., Sekura, R. D.,

Manclark, C. R., Stryer, L. & Bourne, H. R. (1984) J. Biol.Chem. 259, 23-26.

14. Sternweis, P. C. & Robishaw, J. D. (1984) J. Biol. Chem. 259,13806-13813.

15. Neer, E. J., Lok, J. M. & Wolf, L. G. (1984) J. Biol. Chem.259, 14222-14229.

16. Malbon, C. C., Rupiejko, P. J. & Garcia-Sainz, J. A. (1984)FEBS Lett. 176, 301-306.

17. Raschke, W. C., Baird, S., Ralph, P. & Nakoinz, I. (1978) Cell15, 261-267.

18. Aksamit, R. R., Falk, W. & Leonard, E. J. (1981) J. Immunol.126, 2194-2199.

19. Sekura, R. D., Fish, F., Manclark, C. R., Meade, B. & Zhang,Y. (1983) J. Biol. Chem. 258, 14647-14651.

20. Meade, B. D., Kind, P. D. & Manclark, C. R. (1984) Infect.Immun. 46, 733-739.

21. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254.22. Laemmli, U. K. (1970) Nature (London) 227, 680-685.23. O'Farrell, P. H. (1975) J. Biol. Chem. 250, 4007-4021.24. Koski, G., Streaty, R. A. & Klee, W. A. (1982) J. Biol. Chem.

257, 14035-14040.25. Sharma, S. K., Nirenberg, M. & Klee, W. A. (1975) Proc.

Natl. Acad. Sci. USA 72, 590-594.26. Katada, T. & Ui, M. (1980) J. Biol. Chem. 255, 9580-9588.27. Olansky, L., Myers, G. A., Pohl, S. L. & Hewlett, E. L.

(1983) Proc. Natl. Acad. Sci. USA 80, 6547-6551.28. Kurose, H., Katada, T., Amano, T. & Ui, M. (1983) J. Biol.

Chem. 258, 4870-4875.29. Katada, T., Bokoch, G. M., Northup, J. K., Ui, M. & Gilman,

A. G. (1984) J. Biol. Chem. 259, 3568-3577.30. Katada, T. & Ui, M. (1982) Proc. Natl. Acad. Sci. USA 79,

3129-3133.31. Katada, T. & Ui, M. (1982) J. Biol. Chem. 257, 7210-7216.32. Murayama, T. & Ui, M. (1983) J. Biol. Chem. 258, 3319-3326.33. Bokoch, G. M, Katada, T., Northup, J. K., Ui, M. & Gilman,

A. G. (1984) J. Biol. Chem. 259, 3560-3567.34. Katada, T., Bokoch, G. M., Smigel, M. D., Ui, M. & Gilman,

A. G. (1984) J. Biol. Chem. 259, 3586-3595.35. Codina, J., Hildebradt, J., Iyengar, R., Birnbaumer, L.,

Sekura, R. D. & Manclark, C. R. (1983) Proc. Natl. Acad. Sci.USA 80, 4276-4280.

36. Milligan, G. & Klee, W. A. (1985) J. Biol. Chem. 260,2057-2063.

37. Hyslop, P. A., Oades, Z. G., Jesaitis, A. J., Painter, R. G.,Cochrane, C. G. & Sklar, L. A. (1984) FEBS Lett. 166,165-169.

38. Simchowitz, L., Fischbein, L. C., Spilberg, I. & Atkinson,J. P. (1980) J. Immunol. 124, 1482-1491.

39. Marx, R. S., McCall, C. E. & Bass, D. A. (1980) Infect.Immun. 29, 284-288.

40. Smolen, J. E., Korchak, H. M. & Weissmann, G. (1980) J.Clin. Invest. 65, 1077-1085.

41. Aksamit, R. R. (1985) Fed. Proc. Fed. Am. Soc. Exp. Biol. 44,1459 (abstr.).

42. Aksamit, R. R., Backlund, P. S., Jr., & Cantoni, G. L. (1983)J. Biol. Chem. 258, 20-23.

43. Fung, B. K.-K., Hurley, J. B. & Stryer, L. (1981) Proc. Natl.Acad. Sci. USA 78, 152-156.

44. Okajima, F. & Ui, M. (1984) J. Biol. Chem. 259, 13863-13871.45. Bokoch, G. M. & Gilman, A. G. (1984) Cell 39, 301-308.46. Molski, T. F. P., Naccache, P. H., Marsh, M. L., Kermode,

J., Becker, E. L. & Sha'afi, R. I. (1984) Biochem. Biophys.Res. Commun. 124, 644-650.

47. Harvath, L. & Aksamit, R. R. (1984) J. Immunol. 133,1471-1476.

Biochemistry: Backlund et al.