Proc. Nati. Acad. Sci. USA Vol. 82, pp. 869-873, February 1985 Immunology

Association of the N-formyl-Met-Leu-Phe receptor in human neutrophils with a GTP-binding protein sensitive to pertussis toxin (receptor regulation/stimulus-secretion coupling/ADP-ribosylation/inflammation)

PRAMOD M. LAD, CHARLES V. OLSON, AND PAULA A. SMILEY Kaiser Regional Research Laboratory, 4953 Sunset Boulevard, Los Angeles, CA 90027

Communicated by John D. Baldeschwieler, October 5, 1984

ABSTRACT Pertussis toxin inhibits the N-formyl-MetLeu-Phe (fMet-Leu-Phe) mediated human neutrophil functions of enzyme release, superoxide generation, aggregation, and chemotaxis. As pertussis toxin modifies the GTP binding receptor-regulatory protein "N1," the association of the fMetLeu-Phe receptor with such a protein was further examined in purified neutrophil plasma membranes. Both fMet-Leu-Phemediated guanine nucleotide exchange and nucleotide-mediated regulation of the fMet-Leu-Phe receptor are inhibited by pertussis toxin. In addition, membrane pretreatment with pertussis toxin abolishes the fMet-Leu-Phe-mediated inhibition of adenylate cyclase. Actions of pertussis toxin are due to the ADP-ribosylation of a single subunit at 41 kDa in the neutrophil plasma membrane, which comigrates on NaDodSO4 gels with the N. GTP-binding protein in the platelet plasma membrane. Our results suggest that (i) the fMet-Leu-Phe receptor is associated with a N. GTP regulatory protein, and (it) a fMetLeu-Phe-N; complex is important in the control of several neutrophil functions, probably involving multiple transduction systems, including adenylate cyclase.

MATERIALS

[32P]ATP, [3H]cAMP, and 5'-[3H]guanylyl imidodiphosphate (GuoPP[NH]P) were purchased from ICN. NCS tissue solubilizer, [3H]fMet-Leu-Phe, and Omnifluor were provided by New England Nuclear. Hydrofluor was purchased from National Diagnostics (Somerville, NJ). Special quality ATP and GuoPP[NH]P were from Boehringer Mannheim. 5'-Adenylyl imidodiphosphate, cAMP, GTP, bovine serum albumin, fMet-Leu-Phe, creatine phosphokinase, creatine phosphate, ferricytochrome c, superoxide dismutase, cytochalasin B, Micrococcus Iysodeikticus, guaiacol, and O-dianisidine were from Sigma. Dextran T-500 and Ficoll were from Pharmacia, sodium diatriazoate (Hypaque) was from Sterling Drug (New York), and Hanks' balanced salt solution was from GIBCO. Acrylamide, bisacrylamide, ammonium persulfate, and N,N,N',N'-tetramethylethylenediamine were of electrophoresis grade and were purchased from Bio-Rad. PT, purified by the method of Sekura et al. (11), was donated by J. Cowell of the National Institutes of Health (Bethesda, MD) or was from List Biologicals (Campbell, CA). Both sources of PT gave similar results. PT was pure as judged by the criteria of Sekura et al. (11). In addition, the effects of PT observed here have been highly specific for nucleotide regulatory proteins and, in experiments with isolated plasma membranes, were completely dependent on the presence of NAD.

The neutrophil is an important cell in host defense against bacterial infections and is also a key participant in adverse inflammatory reactions (1-3). Neutrophil functions such as chemotaxis, enzyme release, superoxide production, and aggregation are promoted by N-formyl-Met-Leu-Phe (fMetLeu-Phe) and C5a receptors, and are modulated by prostaglandins, f3-adrenergic, and muscarinic cholinergic receptors. TWo types of GTP regulatory proteins, termed N, and Ni, are involved in the mediation of the effects of several hormone systems (4, 5). For the human neutrophil, we have recently identified the N, protein and shown its involvement in the regulation of the f3-adrenergic and prostaglandin receptors and in mediating the effects of these receptors on adenylate cyclase (6, 7). In contrast to these receptors, the fMet-Leu-Phe receptor seemed not to be associated with N, although two recent reports have indicated that the receptor is indeed regulated by guanine nucleotides (8, 9). Involvement of a Ni GTP-binding protein would, therefore, be a logical possibility for such a regulatory process. Analysis of such a protein has been greatly facilitated by the studies of Katada and Ui (10) and of Sekura et al. (11), who have shown that an exotoxin purified from Bordetella pertussis, termed pertussis toxin (PT), specifically ADP-ribosylates Ni, using NAD as the substrate. In this study, we have used PT to probe the involvement of N1 with fMet-Leu-Phe-mediated cellular functions, in the reciprocal exchange reactions between the fMet-Leu-Phe receptor and GTP regulatory proteins (12), and in the inhibition of adenylate cyclase.

METHODS Neutrophil Plasma Membranes. Human neutrophils were isolated on dextran and Ficoll-Hypaque gradients and plasma membranes were prepared as described (7). The structural characteristics of the native and toxin-treated membranes were determined by NaDodSO4 gel electrophoresis, using the system of Neville (7, 14). Adenylate cyclase activity was determined by the assay of Salomon et al. (15) at an ATP concentration of 50 uM. Details of the assay are provided elsewhere (6, 7). Toxin Activation, Membrane Labeling, and Whole Cell Modification. PT (1 ,ug/ml) was activated by incubating with 0.1 M dithiothreitol for 15 min at 30'C (9). The membrane (100 ,ug) was incubated with 100 mM nicotinamide/1 mM GTP/1 mM ADP-ribose/[32P]NAD (200 Ci/mmol; 1 Ci = 37 GBq)/5 ,ul of activated PT, in a final assay vol of 100 ,l for 30 min at 30'C. Samples were centrifuged at 200,000 x g for 15 min and prepared for NaDodSO4 gels. Whole cell treatments with PT were carried out by incubating the neutrophils (1-5 x 107 cells per ml) with PT (1-10 ,ug/ml) for 1 hr at 37°C. After the treatment, cells were pelleted (400 x g for 15

The publication costs of this article were defrayed in part by page charge

Abbreviations: fMet-Leu-Phe, N-formyl-L-methionyl-L-leucyl-Lphenylalanine; PT, pertussis toxin; GuoPP[NH]P, 5'-guanylyl imidodiphosphate.

payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

869

870

Immunology: Lad et aL

min) and resuspended in Hanks' balanced salt solution at a concentration of 2 x 107 cells per ml and were evaluated for various neutrophil functions. Controls for membrane pretreatment included incubation with buffer alone (without PT) and incubation with PT in the absence of NAD. A control for the whole cell treatment consisted of treatment with buffer alone under identical incubation conditions. Cell viability was checked routinely, using trypan blue exclusion. Viability of >95% was observed in both PT-treated and control cells. Chemotaxis. Neutrophils (2 x 106 cells per ml) were suspended in Hanks' balanced salt solution and incubated for 60 min at 370C, with or without PT (3 pug/ml). The cells were then washed and resuspended in Hanks' balanced salt solution with 0.5% bovine serum albumin. The chemotaxis assay was performed in duplicate as follows: 250 ,.l of buffer (control well), fMet-Leu-Phe (1 nM to 1 uM) or zymosan-activated serum (diluted 1:10) were added to the bottom wells of the modified Boyden chambers. Millipore filters (16), 2.5-cm diameter and 3-Am pore size, were placed on top of the solution containing ligands, and 250 1Ld of the cell suspension was added to the top wells of the chamber. The chamber was placed in a moisture box and incubated in a moist 5% C02/95% air incubator at 370C for 60 min. The filters were removed and stained sequentially in 95% isopropanol (30 sec); hemotoxylin (3 min); distilled water, 2 changes; 70% isopropanol acidified with 1 M HCl (30 sec); distilled water (30 sec); 2% MgSO4/0.2% NaHCO3 (3 min); distilled water (2 min); then two immersions in each of 70%, 95%, and 100% isopropanol, and then xylene for 2 min. The filters were mounted on slides using Permount (Fisher), and the cells were counted under high magnification (x 100). Cells moving in the direction of the chemoattractant were counted at 20,um intervals. Four fields were examined at each level and an average was obtained. Incubations for random migration were determined by incubating the cells in the presence of buffer without a chemoattractant. [3HJGuoPP[NHJP Exchange. Purified neutrophil plasma membranes (0.5 mg/ml) were suspended in buffer A (10 mM MgCl2/1 mM EDTA/1 mM dithiothreitol/50 mM Tris HCl, pH 7.5) (17). The membrane suspension was pretreated with 0.2 mM adenylyl imidodiphQsphate and 1 mM Mg2' at 25°C for 2 min to reduce nonspecific binding of the labeled nucleotide. [3H]GuoPP[NH]P (10.6 Ci/mmol; ICN) was added to a final concentration of 0.3 ,uM, and the incubation was continued at 25°C for 5 min. The binding was stopped with the addition of the same volume of ice-cold buffer A. The membranes were then pelleted (50,000 x g for 15 min at 4°C), washed once, and resuspended to a final concentration of 1 mg/ml in ice-cold buffer A. Activated PT (10 IlI at 0.1 ,ug/,ul) was added to 50 ,u of the labeled membrane suspension and 40 ,ul of a buffer containing 4 mM NAD/50 mM arginine/20 mM Tris HCl, pH 7.5, and was incubated at 30°C for 10 min. This mixture was then added to 2.9 ml of pre-equilibrated (30°C) buffer A containing the ligands and unlabeled nucleotide, and was incubated for 10 min. The exchange was terminated by rapid vacuum filtration over Whatman GF/C filters, using a Millipore filtration manifold, and the filtrate activity was counted in 12 ml of Hydrofluor in a Beckman LS-7500 liquid scintillation counter. [3H fMet-Leu-Phe Binding. Neutrophil membranes (25 ,ug per assay in a total vol of 200 ,lI) were incubated with [3H]fMet-Leu-Phe (1.1 nM; 78.4 Ci/mmol) for 30 min at 25°C in Hanks' balanced salt solution (10). The incubation was terminated by the addition of 4 ml of ice-cold incubation buffer and then filtered over GF/C filters in a Millipore apparatus. The filters were washed twice with 4 ml of ice-cold incubation buffer. The filters were dried, transferred to scintillation vials, and the membrane was solubilized in 2 ml of NCS tissue solubilizer overnight at room temperature. Om-

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

nifluor scintillant (15 ml) was added to the vials, and the activity was measured in a Beckman LS-7500 counter. Other Neutrophil Function Tests. Aggregation studies were carried out as described by Craddock et al. (18). Enzyme (lysozyme) release was measured using the method of Zurier et al. (19). The percent lysozyme released was 35%-40% of the total measured in the presence of 0.5% Triton X-100. Superoxide generation was monitored as the reduction of cytochrome c (20). Controls included cells incubated without substrate and cells incubated with fMet-Leu-Phe and superoxide dismutase.

RESULTS As shown in Fig. 1, PT inhibits fMet-Leu-Phe modulation of diverse neutrophil functions. For example (Fig. 1A), the aggregation profiles of control and toxin treated cells are shown at fMet-Leu-Phe concentrations of 10 and 100 nM. At 10 nM fMet-Leu-Phe, significant inhibition and reversibility of aggregation is observed in toxin-treated cells. At a higher concentration (100 nM), the degree of inhibition is lower. When fMet-Leu-Phe-mediated chemotaxis was examined (Fig. 1B), a near complete inhibition (>90%) of chemotactic response was observed at all concentrations of fMet-LeuPhe tested (1-100 nM). Analysis of fMet-Leu-Phe-mediated superoxide generation and enzyme release (Fig. 1C) showed a pattern of inhibition that was very similar to the pattern observed for inhibition of aggregation. Maximal and nearcomplete inhibition relative to control cells was observed at 10 nM fMet-Leu-Phe, while only 30%-40% of maximal inhibition is observed at 100 nM fMet-Leu-Phe. In summary, the results indicate that PT is a potent inhibitor of fMet-LeuPhe-mediated cellular response at low concentrations of fMet-Leu-Phe in aggregation, enzyme release, and superoxide production, as well as a potent inhibitor of chemotaxis at all concentrations tested. The difference in the mode of inhibition between these two sets of responses is unclear, although it is worth noting that cytochalasin B is present in those function tests in which an inverse relationship between the degree of inhibition and fMet-Leu-Phe concentration is observed. The significant inhibition of diverse functions by PT implies a target for this toxin proximal to the receptor itself. It is now well established that PT modifies the N. component of adenylate cyclase (8, 9). The role for a Ni protein in the human neutrophil was probed by examining the effects of PT on two characteristic regulatory processes that involve GTP regulatory proteins: (i) GTP-mediated regulation of the receptor and (ii) receptor-mediated release of prebound labeled guanine nucleotide. The results presented in Figs. 2 and 3 show that both guanine nucleotide regulation of the fMet-Leu-Phe receptor as well as fMet-Leu-Phe-mediated guanine nucleotide release occur in neutrophil plasma membranes. Consistent with its observed role as an "uncoupler" of receptor-Ni interactions (8), PT caused significant inhibition in both of these regulatory processes. Nucleotide release was observed in a dose-dependent manner in the presence of fMet-Leu-Phe (1 nM to 10 ,uM) for control membranes, while membranes treated with PT exhibited a 60% decrease in fMet-Leu-Phe-mediated exchange. PT modification also affects the GTP-binding protein directly, as the basal exchange values for control and PT-treated membranes were different. A clear dose-dependent inhibition of [3H]fMet-Leu-Phe binding by GuoPP[NHIP (10 nM to 10 ,uM) was observed for control membranes, while membranes treated with PT showed an 80% decrease in the nucleotidemediated inhibition. Although quantitative differences in the degree of inhibition of the two processes by PT are noted, the phenomenon of the inhibition itself is clearly evident. Another feature of the effect of PT is the altered binding of

Proc. Natl. Acad. Sci. USA 82

Immunology: Lad et aL

A.

871

(1985)

B. 400, Control 0

0 300

I-_

C58

a_-

200

PT Treated

I I I

_

0

E-

-0

c

0

7

LA

en

E

C

/

lo-8 iota 1o-6

;-9

_-J

!0-9 Ic-8 10-7

10-6

fMet-Leu-Phe. M

0

-_

0

CcmDntrol

C. a

xO

_

U-0 0)

--

(a.) tno >( 2)0 J0)

r

A/'

V/

Ct 0o

I,

9

10-8

10-7

10-6

I0-8

l0-,

10-6

fMet-Leu-Phe. M FIG. 1. PT effects on fMet-Leu-Phe-mediated neutrophil function. (A) Neutrophil aggregation was carried out as described. In the experix 106 cells were used at fMet-Leu-Phe doses of 10 (lower curves) and 100 nM (upper curves). The aggregation response is essentially complete in 3-4 min. PT treatment was carried out at a concentration of 1 Ag/ml at 370C for 1 hr. (B) Neutrophils (2 x 106 cells per ml) were preincubated with (0) and without (D) PT at 3 Ag/ml at 370C for 60 min. Before assaying for chemotaxis, the cells were washed twice to remove any residual toxin and resuspended in Hanks' balanced salt solution with 0.5% bovine serum albumin. Bars indicate total number of migrating cells over a distance of 80 Am. Assay was performed in triplicate and error bars represent SEM. (C) Neutrophils (6 x 107 cells per ml) were incubated with and without PT at 15 gg/ml (to allow for the higher cell concentration) at 370C for 60 min. Cells were washed twice and resuspended in Hanks' balanced salt solution to the concentrations described. Bars represent inhibition of lysozyme release (a) and superoxide generation (0) with respect to the non-toxin-treated cells. Percentage inhibition at each dose of fMet-Leu-Phe is measured relative to its own control without PT. Values for superoxide generation expressed as A OD per 5 min per 107 cells ranged from 0.050 at 1 nM to 0.238 at 1 ,uM, and values for enzyme release expressed as A OD per min per 107 cells ranged from 0.034 at 1 nM to 0.003 at 1 AtM. The entire set of measurements (A-C) were carried out on cells from a single donor and then repeated with two different normal donors, and the results were essentially similar in all cases.

ments 1

nucleotide that accompanies the abolition of the effect of fMet-Leu-Phe on nucleotide exchange. Alterations in the levels of the hormone and of nucleotide binding have also been noted in the actions of cholera toxin on the N, component (12), and similar considerations may apply to the actions of PT on Ni. The target enzyme for a fMet-Leu-Phe-Nj complex was then considered. An obvious choice involves the N1-mediated inhibition of adenylate cyclase (Fig. 4), which was tested in the presence of forskolin, in both whole homogenates and purified plasma membranes, to permit easier detection of inhibition through enhanced basal activity. Inhibition of cyclase was also tested in the presence of prostaglandin E1/GTP (data not shown). fMet-Leu-Phe-mediated inhibition of cyclase was observed with both preparations in the

presence of forskolin and prostaglandin E1/GTP. However, fMet-Leu-Phe-mediated inhibition was observed at higher concentrations of fMet-Leu-Phe (ED50, 1 ,4M) than required for maximal cellular effects of this peptide (e.g., Fig. 1). PT modification of the membrane resulted in a clear abolition of the inhibitory phase, as observed in other N8 systems. The molecular basis for the effects of PT were examined using [a-32P]NAD as the substrate. NaDodSO4 gel electrophoresis and autoradiography of the labeled membranes revealed a single band at 41 kDa (Fig. 5, lane 2). This band comigrated with a band in human platelet membranes, isolated according to Barber and Jamieson (21), labeled under identical conditions (lane 1). No labeling was observed in control incubations with buffer and [a-32P]NAD in the absence of PT.

872

Immunology: Lad et aL

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

N;0.9 _

S0.80

3.2 0

0.6

00-6

iO-5

10-9

10-8

10-7

10-6

10-5

fMet-Leu-Phe, M

fMet-Leu-Phe, M

FIG. 2. fMet-Leu-Phe-mediated nucleotide exchange. Purified neutrophil plasma membranes were loaded with [3H]GuoPP[NH]P. Control membranes, *; membranes treated with PT. . Basal exchange values (no fMet-Leu-Phe) were 4.26 and 6.33 pmol/mg for control and toxin-treated membrane, respectively. Data were averaged from two separate experiments. Total [3H]GuoPP[NH]P binding was -6517 ± 2 pmol/mg, maximal displacement with 10 ,uM unlabeled GuoPP[NH]P was 11 ± 1 pmol/mg, and ED50 was ~10 nM. The experiment was repeated twice with two different batches of neutrophil plasma membrane, with similar results.

FIG. 4. fMet-Leu-Phe-mediated adenylate cyclase inhibition. Neutrophil plasma membranes were treated with PT (m), and in buffer without added toxin (e). Adenylate cyclase was then measured in the presence of forskolin (10 .M) to enhance basal activity and thus permit easier detection of inhibition. Basal values were 0.6 pmol/min per mg and were essentially unaltered by the treatment with PT. Bars denote SEM of triplicate measurements. Experiments were repeated twice with purified plasma membranes, twice with whole homogenates, and in the presence of forskolin as well as prostaglandin E1/GTP.

DISCUSSION The results described here suggest that the fMet-Leu-Phe .c receptor is associated with a GTP-binding regulatory protein. Two reactions characteristic of such regulation have been.0described for several receptors (4, 5). The first is guanine nucleotide-dependent 0 10 108 decrease 1o6I1o7in hormone binding and the second, representing the reciprocal process, is hormoneinduced guanine nucleotide exchange or release. Both types

of experiments are presented in this study and demonstrate a clear role for GTP-binding proteins in the regulation of the fMet-Leu-Phe receptor. Our studies also indicate the nature of the nucleotide-binding protein involved in receptor regulation. Distinct stimulatory (Ns) and inhibitory (Ni) binding proteins have been described. PT is known to ADP-ribosylate Ni, while cholera toxin ADP-ribosylates Ns, and both toxins use NAD as substrate (11). In the experiments reported here, we show that as observed in the human platelet, PT modifies the N, component in the human neutrophil. The selective labeling of a single protein supports the specificity of the toxin's effects. As a result of the modification of Ni, PT uncouples the reciprocal regulatory interactions between various receptors and the Ni protein (8, 9). As similar effects are observed with the fMet-Leu-Phe receptor, the involve-

°' 100,