THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1994 by The American Soeiety for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 6 , Issue of February 11, pp. 4417-4423, 1994 Printed in U S A .

Receptor-independent Activation of Guanine Nucleotide-binding Regulatory Proteins by Terminal Complement Complexes*

(Received for publication, May 28, 1993, and in revised form, October 18, 1993)

Florin NiculescuS, Horea Rus, and Moon L. Shin5 From the University of Maryland, School of Medicine, Department of Pathology, Baltimore, Maryland 21201

Activation of heterotrimeric guanine nucleotide-bind- ing proteins (G proteins) by terminal complement com- plexes (TCC) was investigated on human lymphoblas- toid B-cell line JY2S and its mutant JYS deficient in glycosylphosphatidylinositol-anchored proteins. TCC assembly achieved by antibody-dependent activation of C7-deficient serum reconstituted with C7 increased spe- cific guanosine-6'-(ythio)triphosphate (GTP+) bind- ing, 4- and &fold, in JY2S and J Y S membranes, respec- tively, between 2 and 10 min, over the level without C7. TCC also increased GTPase activity 5- and 4-fold in JY2S and JYS, respectively, between 5 and 10 min. Increased GTPase activity was noted first with CSb-7 aseembly, which increased further with CSbd and CSb-9. The pres- ence of G proteins in anti-TCC immunoprecipitates of cell lysates was investigated by demonstration of Ga subunit that can be ADP-ribosylated by pertussis toxin (PTX). Immunoprecipitated TCC complexes contained a PTX-sensitive 41-kDa GidGoa subunit, as shown by SDS-PAGE and Western blotting. These complexes were functionally active as determined by GTPyS binding. We have further shown that enhanced TCC elimination from the plasma membrane induced by TCC-generated signals was inhibited by PTX. In conclusion the biologi- cal activities induced by TCC in nucleated cells may be mediated in part by activation of PTX-sensitive G pro- teins.

Complement activation in infection and inflammation plays a crucial role in host defense by generating inflammatory me- diators, such as C3a and C5a,' by opsonization of activating particles through C4b, C3b, and iC3b and lysis of target cells by forming C5b-9 channels. Assembly of the cytolytic C5b-9 com- plex is accomplished through a sequential interaction of C5-C9 plasma proteins, which results in amphipathic conformational

*This work was supported by Grants R01-AI19622 and R01- NS15662 from the National Institutes of Health. The costs of publica- tion of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

4624. $ Recipient of International Fogarty Research Fellowship F05, TWO

0 To whom correspondence should be addressed. The abbreviations used are: C3a, C5a, cleavage products of C3 and

C5, respectively; App(NH)p, adenosine-5'-(P,y-imino)triphosphate, BCA, bicinchoninic acid; BSA, bovine serum albumin; C7D, C8D, C9D, human serum deficient in C7, C8, or C9, respectively; CTX, cholera toxin; DAG, diacylglycerol; Dm, DL-dithiotreitol; G proteins, guanine nucleotide-binding regulatory proteins; Gi, guanine nucleotide-binding inhibitory protein; Gs, guanine nucleotide-binding stimulatory protein; GTPyS, guanosine-5'-(y-thio)triphosphate; NHS, normal human se- rum; PBS, phosphate-buffered saline; PMSF, phenylmethylsulfonyl fluoride; F'TX, pertussis toxin; PAGE, polyacrylamide gel electrophore-

which include C5b-7, C5b-8, and C5b-9; MHC, major histocornpatability sis; TBS, Tris-buffered saline; TCC, terminal complement complexes,

complex.

changes of C6-C9 and insertion of these peptides into the lipid bilayer. Assembly of C5b-9 channels occurs stepwise manner, by forming membrane-associated C5b-7, C5b-8, and C5b-9 com- plexes, collectively called as terminal complement complexes (TCC) (1-3). Unstable voltage-dependent pores are first de- tected when C8 interact with membrane-bound C5b-7 to form C5b-8. Binding of one C9 to a C5b-8 produces transmembrane channel of 1-3-nm pore size, which enlarges to 10 nm with increasing number of C9 and C9-C9 polymerization (3-5). When complement is activated on cells of homologous species, complement inhibitory proteins severely restrict the number and the pore size of the C5b-9 channel (6-9).

The C5b-7, C5b-8, and limited C5b-9 stimulate cells in the absence of lysis and induce a variety of biological activities, which include production of eicosanoids and their derivatives, generation of oxygen radicals, synthesis of tumor necrosis fac- tor and interleukin-1, enhanced production of collagen and col- lagenases, hydrolysis of myelin basic protein, platelets activa- tion, elimination of potentially lytic TCC, and generation of mitotic signals (1, 2, 10-13). These activities are mediated in part by increased [Ca2+li and protein kinase C activity during C5b-8/C5b-9 pore formation (14-16). Increased mass levels of sn-l,2-diacylglycerol (DAG) and ceramide, known endogenous regulators of protein kinase C activity, were achieved by C5b-7, C5b-8, and C5b-9 in intact cells (17). Production of these signal messengers by C5b-7, which does not form a channel nor causes Ca2+ influx, may explain some of cellular activities induced by C5b-7 (11, 15, 17). The TCC-induced DAG increase was inhib- ited by prior treatment of cells with pertussis toxin (m), which suggested that activation of phospholipases by TCC may involve G proteins sensitive to PTX (17).

Heterotrimeric G proteins transduce signals from the recep- tor, and activate effector molecules, such as adenylyl cyclases, phospholipases, and various ion channels (18-20). Ligand bind- ing causes enhanced interaction of the receptor with corre- sponding Ga and promotes rapid release of GDP from the a subunit. The activated receptor (R*) is associated with empty Ga (a,) with a high affinity together with /3 and y subunits and stabilizes R*.a;py complex. Subsequent entry of GTP to the binding site is thought to induce a transient decrease in affinity for both receptor and fir subunits (19). Hydrolysis of GTP to GDP by intrinsic GTPase activity induces the interaction of a with the effector, then the a-GDP reassociates with the Py. Interaction of a-GTP with effectors regulated by Mg2+ and Li+ is also affected by covalent modification of the a subunit, as in ADP-ribosylation induced by bacterial toxins (21, 22). G pro- teins are mostly activated by ligand-receptor interaction, and these receptors belong to proteins with seven membrane-span- ning domains or a single membrane-spanning domain with or without intrinsic tyrosine kinase activity (20, 23, 24). G pro- teins can also be activated in receptor-independent manner by complexing with metallofluorides or by mastoparan. [AlF3]- or [BeF,I- activates G proteins by interacting with the nucleotide- binding site of a subunit, mimicking the y-phosphate of GTP

4417

4418 G Protein Activation by Terminal Complement Complexes

(25, 26), whereas mastoparan binds directly to the carboxyl terminus of a subunit, similar to ligand-bound receptors (27- 29).

This study presents experimental evidence that all of the three TCC complexes are capable of activating PTX-sensitive G proteins in the membrane. Physical association of TCC with the 41-kDa Ga subunit was demonstrated by immunoprecipitation with antibodies to C7, C8, and to C5b-9 neoantigen from deter- gent-solubilized cells carrying TCC. Involvement of G proteins in biological activities of TCC was suggested by the ability of €'"X to inhibit the enhanced TCC elimination from the cell surface.

EXPERIMENTAL PROCEDURE Chemicals and Reagents-Cholera toxin (CTX) and PTX were ob-

tained from List Biological Laboratories (Campbel, CA). DL-Dithiotreitol (DIT), EDTA, EGTA, Lubrol, ATP, NAD, creatine phosphokinase, cre- atine phosphate, dextran (2 x 109, Norit A charcoal, bovine serum albumin (BSA), aprotinin, leupeptin, and phenylmethylsulfonyl fluo- ride (PMSF) were from Sigma. GTPyS, GTP, GDP, App(NH)p, and Tri- ton X-100 were from Boehringer Mannheim. [ys2P1GTP (3000 Ci/ mmol), [36SlGTPyS (1267 Ci/mmol), and [32PlNAD (30 Ci/mmol) were from DuF'ont NEN. Nitrocellulose membranes were from Millipore (Bedford, MA); BA85 nitrocellulose membrane filters (0.45 pm) were from Schleicher & Schuell; polyacrylamide reagents were from Bio-Rad; protein A-agarose and BCA reagent were from Pierce Chemical Co.

Antibodies, Complement, and Complement Components-Mouse IgG2a anti-human MHC class I1 was purified by precipitating the cul- ture supernatant of hybridoma L-243 (ATCC, HB55, Batch P-8310) with caprylic acid (30), then by purifying IgG fraction using agarose-conju- gated rabbit IgG to mouse IgG (Sigma). Normal human serum (NHS) used as source of complement components was obtained by pooling sera from 20 healthy donors and kept aliquoted at -78 "C. Human serum immunochemically depleted of C7, C8, or C9, (C7D, C8D, or C9D, re- spectively), puritied human C7, affinity-purified goat IgG to C7, C8, and C9, and monoclonal IgG to human C5b-9 neoantigen (anti-C5b-9) were purchased from Quidell (La Jolla, CA). Rabbit antiserum to common Ga subunit (GN1) and to Gi3a (EC/2) were from DuPont NEN. Affinity- purified rabbit IgG specific for Gia,,,, GiaJGoa, and to GP subunit were from Calbiochem-Novabiochem (La Jolla, CA). G protein standards were also obtained from Calbiochem-Novabiochem. Agarose-conjugated with rabbit IgG against mouse IgG or goat IgG were from Sigma.

Preparation of Cells and Cell Membranes-The JY cell line is an Epstein-Barr virus transformed human lymphoblastoid B-cell line. JY5, a mutant of the wild type JY25, has a defect in expressing glycosylphos- phatidylinositol-anchored membrane proteins which include decay-ac- celerating factor and CD59 and allow four times more C5b-9 assembly than JY25 (17, 31, 32). The antibody optimal for sublytic complement activation was determined by lysis of JY25 or JY5 sensitized by serial dilutions of a monoclonal IgG to MHC class I1 protein with excess serum complement (10% NHS or 20% C7D + 10 pg of C7), as described (17). Identical concentrations of IgG and serum complement were used in experiments performed with plasma membranes. Plasma membranes were prepared according to a procedure described (33), with several modifications. In brief, 5 x lo7 JY cells were stirred for 10 min in 100 ml of 20 m~ boric acid with 0.2 M EDTA, pH 10, and again for 15 min after addition of 10 ml, 500 m~ boric acid in PBS. Cell debris were filtered through 120-pm nylon mesh, and nuclei were removed by centrifuging for 5 min at 2000 rpm (Beckman AccuSpin FR) at 4 "C. Membranes were pelleted by centrifugation at 10,000 rpm for 30 min (Sorvall RC- 5B), resuspended in PBS, and ultracentrifuged at 45,000 rpm for 60 min at 10 "C using SW60 rotor (Beckman Instruments Inc.). Membrane pellets were suspended in PBS containing protease inhibitors (1 m~ PMSF, 10 pg/ml leupeptin, and 10 pg/ml aprotinin). Protein concentra- tion was determined by using BCA reagent and membranes were kept aliquoted at -78 "C until use.

Assays for GTPyS Binding and GTPase Actiuity-For GTPyS bind- ing, plasma membranes (50 pg of proteidsample) sensitized with 10 pg/ml anti-MHC class I1 IgG were suspended in 50 pl of Tris-buffered saline (TBS) (20 m~ Tris-HC1,lOO m~ NaCl), pH 8.0, containing 30 m~ MgCl,, 1 n" EDTA, 1 n" Dl", 0.1% Lubrol, 1 p~ GDP (34-36). Fol- lowing addition of 10% NHS or 20% C7D f 10 pg of C7 and 1 pl of [36S]GTPyS (specific activity of 1267 Ci/mmol) diluted with TBS to 0.22 n~ GTPyS, the mixtures were shaken at 37 "C. The reaction was stopped at different time points by adding 950 pl of ice-cold TBS with 25

p~ unlabeled GTPyS. Samples were vortexed and filtered through 0.45-pm BA85 filter. After washing two times with 10 ml of ice-cold TBS, filters were immersed in 10 ml of scintillation fluid, and the 3sS activity was counted. Membranes treated with NHS without IgG, or IgG and C7D without C7, were used as controls. Results, performed in duplicate, from seven experiments for NHS and three experiments with C7D were statistically analyzed by the paired Student's t test.

GTPase activity was determined as described (34, 35). Membranes (10 pg of proteidsample) sensitized with 10 @mI IgG were incubated for 5 min at 37 "C with 10% NHS, 20% C8D, C9D, or C7D C7, then centrifuged at 12,000 rpm for 1 min in a Beckman Microfuge B. Mem- branes were incubated in TBS containing 30 n" MgC12, 0.5 n" Ap- p(NH)p, 0.1 IIIM ATP, 3 n" creatine phosphate, 75 unitdml creatine phosphokinase, 0.1 n" EGTA, 1 n" Dl", and 250 n~ [y-32P]GTP (spe- cific activity, 3000 Ci/mmol), then incubated at 37 "C in a shaker. Re- action was terminated at the indicated time points by adding 1 ml of cold 5% Norit A charcoal suspension in PBS containing 0.1% dextran and 0.5% BSA. Mixtures were centrifuged at 12,000 rpm for 1 min in a Beckman Microfuge, and the radioactivity in 2 0 0 4 supernatant was counted. Membranes treated with GTPase buffer without serum were included in each assay as additional control. Results of five experiments for NHS and three for deficient serum, performed in duplicates, were statistically analyzed by the paired Student's t test.

ADP-ribosylation of JY Membranes Carrying TCC Complexes- Untreated membranes and IgG-sensitized membranes (150 pg of mem- brane protein) were incubated with 10% NHS for 15 min at 37 "C. Membranes were washed and suspended in 50 pl of ADP-ribosylation buffer (TBS with 5 rn MgCl,, 10 rn thymidine, 1 m~ ATP, 0.1 m~ GTP, 10 n" DTT, 1 m~ EDTA) (36, 37) containing 10 [32PlNAD and 10 pg/ml PTX or CTX previously activated by incubating in 20 IIIM DIT in PBS for 60 min at 37 "C. The mixtures were shaken at 37 "C for 30 min, then 1 ml of cold TBS was added. Membranes were pelleted, then solubilized in 1 ml of TBS with 1% Triton X-100 for 30 min on ice with repeated vortexing and sonication. Proteins in the membrane lysates were precipitated overnight a t 4 "C with 20% trichloroacetic acid. Iden- tical amounts of protein from each sample, determined by BCA, were boiled for 3 min in loading buffer (125 m~ Tris-HC1, pH 6.8,2% SDS, 6% mercaptoethanol, 40% sucrose, and 0.025% bromphenol blue) and ana- lyzed on 10% SDS-PAGE (38), followed by Western blotting and auto- radiography. For immunodetection, blots were washed in TBS with 5% BSA and 0.1% Triton X-100 for 60 min at 25 "C, then exposed to anti- bodies to human Ga or GP subunits. The antibody-reactive bands were visualized with horseradish peroxidase conjugated with goat IgG to rabbit IgG followed by the use of horseradish peroxidase-precipitating reagent (Harlan, Madison, WI) or by enhanced chemiluminescence de- tection reagents (ECL) (Amersham Corp.). Autoradiographic densities were quantitated by Molecular Dynamic Densitometer, and the density of each band was integrated using Imagequant Software (Molecular Dynamics, Sunnyvale, CA). The results were expressed in arbitrary scan units.

Immunoprecipitation of TCC-G Protein Complexes-Intact cells (5 x lo7) sensitized with IgG were incubated for 15 min at 37 "C with 10% NHS, 20% C9D or C8D, and also with 20% C7D f 10 pg of C7. Cells were washed with Hanks' balanced salt solution, then placed in 1 ml of cold lysis buffer (30 m~ Tris-HC1, 150 m~ NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 10 pg/ml leupeptin, 10 pg/ml aprotinin, and 1 m~ PMSF) for 30 min on ice, while vortexing and sonicating repeatedly. Cell debris were removed by centrifuging 5 min at 2000 rpm at 4 "C (Beck- man AccuSpin FR). The supernatants were precleared by incubating with protein A-agarose for 60 min at 4 "C and centrifuging at 15,000 rpm for 15 min in an Eppendorf centrifuge 5415 C. This detergent- solubilized cell lysates (100 pl) were incubated with 50 pg/ml anti- C5b-9, anti-C8, anti-C7, or anti-C3 IgG and with 100 pg/ml agarose conjugated with secondary antibodies for overnight a t 4 "C. Agarose beads were washed four times with lysis buffer, two times with TBS, then subjected to ADP-ribosylation with preactivated PTX or CTX as described for the plasma membrane. The reaction was terminated by adding 1 ml of cold TBS. After washing, beads suspended in loading buffer were bolied for 3 min, and identical amounts of protein from each sample, determined by protein concentration at A280 ,,,,,, were analyzed by 10% SDS-PAGE and autoradiography.

TCC-G protein complexes isolated on agarose were also examined for the [36S]GTPyS binding. Ten pl of agarose carrying immunoprecipitated complexes, with and without 30-min exposure at 37 "C to 10 ng of preactivated PTX, were suspended in TBS (20 m~ Tris-HC1, 100 m~ NaC1, pH 8.0) containing 1 pl of [36SlGTPyS (12 pCi/sample). The mix- tures were incubated at 37 "C in shaker for 15 min. The reaction was terminated by adding 1 ml of ice-cold TBS containing 1 p~ GDP, then

G Protein Activation by

/i 7 4 /

0 2 4 6 8 10 12 +

Time, min

I lkrminal Complement Complexes 4419

O ! I 0 10 20

Time, min FIG. 1. GTP+ binding and GTPase. activity during comple-

ment activation. A, plasma membranes (50 pg of protein) of JY25 (0, M) and JY5 (0, 0) sensitized with (circles) or without (squares) anti- MHC class I1 IgG (10 pg/ml) at 37 "C for 30 min, were suspended in 50 pl of TBS buffer containing 30 m~ MgC12, 1 m~ EDTA, 1 m~ D", 0.1% Lubrol, 1 p~ GDP. Following addition of 10% NJ3S and 0.22 m [s6SlGTPyS (1267 Ci/mmol) and incubation at 37 "C, 950 pl of ice-cold TBS was added at the time indicated. The samples were filtered, and the radioactivity associated with the filter was counted. The data are mean * S.E. from seven experiments performed in duplicates. E , IgG- sensitized JY25 (0, M) and JY5 (0, 0,) membranes (10 pg of protein) were incubated with 10% NHS for 5 min at 37 "C. Pelleted membranes were incubated in 50 pl of TBS containing 30 m~ MgC12, 0.5 m~ Ap- p(NH)p, 0.1 m~ ATP, 3 m~ creatine phosphate, 75 unitdml creatine phophokinase, 0.1 m~ EGTA, 1 m~ Dm, and 250 m [-ps2P1GTP (3000 CVmmol). At the time indicated, 1 ml of cold PBS containing 5% Norit Acharcoal, 0.1% dextran, and 0.5% BSA was added, and radioactivity of 2004 supernatant was counted. Data are mean * S.E. from five ex- periments performed in duplicates.

the suspension was filtered through a 0.45-pm BA85 nitrocellulose fil- ter. Filters were washed two times with 10 ml of cold TBS, and the radioactivity was counted. Plasma membranes processed in parallel were used as a positive control.

Effects of pllx on TCC Elimination from the Plasma Membrams- The kinetics of TCC elimination from the cell surface was determined on JY25 by a functional assay (14, 15). In brief, cells treated with medium alone, F"X (500 nglml), or CTX (1000 nglml) for 3 h at 37 "C were sensitized with a lytic dose of IgG (100 pg/ml), then incubated with 20% C9D in RPMI for 15 min at 37 "C to form C5b-8. Cells were placed a t 37 "C, in triplicates, for varying time periods and the remaining C5b-8 were converted tu lytic C5b-9 by incubating with 10 pg of C9 for 60 min. Cell death was determined by lactate dehydrogenase release (14). Lactate dehydrogenase released from melittin-treated cells and that from cells carrying C5b-8 were used as maximum and background values, respectively.

RESULTS

Activation of Serum Complement and TCC Assembly Stimu- late P'SIGTPyS Binding and [Y-~~PIGTP Hydrolysis-Possible

0 : T I I 0 2 4 6 8 10 12

Time, min FIG. 2. Inhibition of GTPase activation by PTX. JY25 (M) and

JY5 (0) membranes treated with preactivated PTX (500 ng/ml) for 3 h (PTX) at 37 "C were sensitized and incubated with 10% NHS for 5 min. The kinetics of GTP hydrolysis determined in membranes treated with Fl'2C or with TBS instead of PTX (JY25 (0); JY5 (0)) were shown. The data represent mean f S.E. from three experiments performed in du- plicates.

activation of G proteins by serum complement was investigated by detecting nonhydrolyzable analogue [3SSlGTPyS binding to JY plasma membranes and also by GTPase activity. The in- crease in GTPyS binding after addition of NHS to IgG-sensi- tized membranes between 2 and 10 min was 4- and 10-fold in JY25 (p < 0.02) and JY5 (p < 0.051, respectively, over the level by NHS in the absence of IgG (Fig. IA). Hydrolysis of GTP, determined as GTPase activity, increased 13-fold between 1 and 15 min in JY25 ( p < 0.051, whereas 4-fold increase between 1 and 10 min was the maximum in JY5 ( p < 0.051, over the level with NHS alone (Fig. LB). When membranes were pretreated with PTX (500 ng/ml), the extent of GTP hydrolysis induced by IgG and NHS was reduced to 50% (Fig. 2). Inhibition was also observed when the reaction was carried out in the presence of 10-fold excess GDP (data not shown). Incomplete inhibition of GTP hydrolysis by PTX may be in part due to the presence of PTX-insensitive G proteins.

To evaluate the role of TCC, GTPyS binding and GTP hydro- lysis were determined in IgG-sensitized membranes treated with C7D f C7. As shown in Fig. 3, C7 reconstitution was required to achieve the effect of IgG and NHS. Between a 2- and 10-min period, GTPyS binding increased 4-fold in JY25 ( p < 0.02) and 8-fold in JY5 ( p < 0.05) over the level obtained by C7D without C7 (Fig. 3A) .

The increase in GTP hydrolysis in response to C7 reconsti- tution between 5 and 15 min was 5-fold in JY25 (p < 0.02) and 4-fold in JY5 ( p < 0.02) (Fig. 3B). When GTP hydrolysis was determined during assembly of C5b6, C5b-7, and C5b-8, using C7D, C8D, and C9D, the increase between 1 and 10 min was 2- and 3-fold for C5b-7 and C5b-8, respectively, over the level by C5b6 (Fig. 4). Treatment with C7D + C7 used as maximum C5b-9 effect showed 5-fold increase in GTP hydrolysis between 1 and 10 min when compared with the level by C7D, which generates C5b6. C5b-9 is a strong activator as evidenced by a 2-fold increase in GTP hydrolysis at 1 min over the C5b6 level, which was similar to the basal level obtained without serum.

ADP-ribosylation of G Proteins by PTX in Membranes Car- rying Inserted TCC-PTX-induced ADP-ribosylation of a pro- tein migrated as a 41-kDa band which was two times more intense in JY25 membranes than in JY5 (Figs. 5 and 7). This band was identified as GidGoa by immunoreacting with spe- cific antibodies to GidGoa subunits. When membranes previ- ously treated with I g G and NHS were ADP-ribosylated by P T X , the reaction was significantly reduced (Fig. 5). These findings

4420 G Protein Activation by Terminal Complement Complexes

71 6 A I // C7D+C7 I

C7D

I 0 2 4 6 8 10 12

Time, min

n -

P " 1

4 ; i 10 12 14

Time, min FIG. 3. GTP@ binding and GTPase activity during activation

of terminal complement proteins. A, GTPyS binding to JY25 (0, W) and JY5 ( 0 , O ) membranes was determined as described in the legend to Fig. l.4, except that sensitized membranes were incubated with 20% C7D with (circles) or without (squares) 10 pg of C7. The data are mean

S.E. from three experiments performed in duplicates. B, the time- dependent increase in GTPase activity was determined by measuring GTP hydrolysis in JY5 (0,CI) and JY25 (0, B) membranes as described in the legend to Fig. 1B, except that 20% C7D was used with (circles) or without (squares) 10 pg of C7. The data are mean = S.E. from three experiments performed in duplicates.

C5b6 C5b-7 C5b-8 C5b-9

FIG. 4. GTPase activity at various steps of TCC formation. IgG- sensitized JY25 membranes (10 pg of protein) were incubated for 5 min a t 37 "C with 20% complement deficient serum. C7D, C8D, and C9D were used to form C5b6, C5b-7, and C5b-8, respectively. C5b-9 com- plexes were formed by C7D + C7. GTP hydrolysis was determined a t 1 and 10 min, as described in the legend to Fig. 1B.

suggested that a portion of Ga in membranes carrying inserted TCC may not be available for PTX to catalyze ADP-ribosyla- tion, possibly by steric hindrance by TCC in intact membrane or due to the activation induced by inserted TCC peptides or by

PTX + + ' + - ' I + - I

41 kDa I

.@ -7- tanti-G a

0

JY25 JY5

serted TCC. A, untreated (JY25, JY5) and IgG-sensitized membranes FIG. 5. ADP-ribosylation by €'"X of membranes carrying in-

(150 pg of protein) treated with 10% NHS (JY25pI'CC, JY5pI'CC) were incubated in 50 pl of ADP-ribosylation buffer (see "Experimental Pro- cedures") with 10 p~ ["2P]NAD (30 CUmmol) and activated FTX (10 pg/ml) for 30 min a t 37 "C. Membranes were solubilyzed in 1 ml of TBS with 1% ?titon X-100, 1 m.. PMSF, 10 pg/ml leupeptin, and 10 pg/ml aprotinin for 30 min at 0 "C. Proteins were precipitated with 20% tri- chloroacetic acid and analyzed by 10% SDS-PAGE, Western blot, and autoradiography. ADP-ribosylated proteins by PTX migrated as a 41- kDa band. The same blot was then incubated with anti-Ga rabbit an- tiserum (GN1) followed by horseradish peroxidase-conjugated goat IgG to rabbit IgG and developed with horseradish peroxidase precipitating reagent. B, densitometric quantitation of the autoradiography shown in A is expressed as relative densities. B, untreated membranes; 0, mem- branes treated with IgG and NHS.

both mechanisms. With CTX, two G proteins of 43 and 21 kDa were ADP-ribosylated in JY membranes with much less inten- sity and this process was not affected by TCC (data not shown).

Association of TCC with PTX-sensitive G Proteins in the Membranes-Association of TCC peptides with G proteins was examined by immunoprecipitating the lysate of cells carrying TCC with anti-C5b-9 IgG and identifying the presence of G proteins in this immunoprecipitates by ADP-ribosylation with PTX and by Western blotting. The 41-kDa Ga subunits were ADP-ribosylated in anti-C5b-9 immunoprecipitates of cells treated with C7D + C7 or NHS (Fig. 6A). C?x was not effective in inducing ADP-ribosylation of any G protein in this complex (data not shown). As shown in Fig. 6, A and B , anti-C7 and anti-C8 IgG immunoprecipitates of cell lysates pretreated with IgG and C8D, C9D, or NHS also revealed the 41-kDa band which was ADP-ribosylated by PTX. This 41-kDa substrate for PTX was not observed when anti-C3 I&, or when heat-inacti- vated NHS (H-NHS), were used. As shown in Fig. 6C, this 41-kDa protein coprecipitated with TCC migrated in an iden- tical manner with the ADP-ribosylated band by PTX in purified membranes and also in immunoprecipitates with anti-Gia IgG. Immunoblotting with anti-Gial,2 antibody showed the presence of Gia migrated as a 41-kDa protein (Fig. 7). The TCC-G pro- tein complexes were also immunoreactive with IgG anti-Gia$ Goa as well as with anti-GP subunits (Fig. 8). These findings indicated that at least two species of G protein interacted with

G Protein Activation by Terminal Complement Complexes 4421

A C7D+C7 NHS C8D C9 D

PTX + - + - I"'1-I + - + - 41 - ADP-r ibosylat ion

43.

29

D

FIG. 6. Coprecipitation of Ga with TCC by anti-C6b-9 I@, anti- c?, or anti43 I&. IgG-sensitized JY25 cells (5 x 10') were incubated with serum complement for 15 min at 37 "C. Cells were solubilyzed in lysis buffer (described in the legend to Fig. 5) containing additional 0.5% sodium deoxycholate. The lysates were immunoprecipitated with anti- bodies to TCC, and the complexes were recovered on agarose conjugated with secondary antibodies. Agarose beads were washed with lysis buffer and treated with activated YlX in ADP-ribosylation buffer. Proteins were then released from the beads by boiling and analyzed by 10% SDS-PAGE, Western blot, and autoradiography. A, cells treated with C7D + C7 or NHS were immunoprecipitated with anti-C5b-9 IgG, and cells treated with C8D or C9D were immunoprecipitated with anti-C7 or anti-C8, respectively, as described. The ADP-ribosylated proteins migrated as a 41-kDa band. B, controls used for the specificity of im- munoprecipitation are shown. ADP-ribosylated proteins by PIX were not detected in C9D-treated cell lysates immunoprecipitated with iso- type control IgG (CSDlagarose), NHS-treated cell lysates immunopre- cipitated with anti-C3 IgG (NHS/anti-C3), or heated NHS-treated cell lysates immunoprecipitated with a n t i 4 8 IgG (H-NHS/ant i -C8) . C , im- munoprecipitates with anti-C5b-9 (NHSlanti-TCC) or anti-Gial.z con- tained similar ADP-ribosylated proteins induced by PTX. Untreated membranes and immunoprecipitation with anti-C3 or isotype IgG served as controls.

TCC, and these complexes were also associated with Py sub- units. As shown in Fig. 9, complexes isolated by anti-C5b-9 anti-

body displayed constant increase in [35S]GTPyS binding be- tween 5 and 15 min which was inhibited by the pretreatment with PTX. Therefore, G proteins precipitated with TCC were functionally active, and the GTP binding sites on Ga were available for GTPyS, even after they were eluted from the membrane.

Involvement of G Protein in Biological Activities of TCC- Functional implication of TCC-activated G proteins was evalu- ated by examining FTX-mediated inhibition of cellular activi- ties induced by TCC. Cells have the capacity to survive limited

41- -- t i r an t i -Ga i

FIG. 7. ADP-ribylation and immunoreaction with anti-GiaIz antibodies. Lysates of IgG-sensitized JY25 cells (5 x lo7) treated with C8D to form C5b-7 were immunoprecipited with anti-C7. The complexes were immunoadsorbed on agarose and subjected to ADP-ribosylation by PIX as described in the legend to Fig. 5, then analyzed by 10% SDS- PAGE, Western blot, and autoradiography. The 41-kDa ADP-ribo- sylated bands are shown in the upper panel. The same blot was reacted with rabbit anti-Gial.z IgG, then developed by ECL (lower panel). The immunoreactive band is a 41-kDa protein, as identified by ADP-ribo- sylation.

43 - " 011)

29 -

anti-Gia 1.2 anti-Gia 3lGoa anti-Gp

FIG. 8. Identification of G protein subunits in TCC-G protein complexes. Lysates form JY25 cells treated with NHS were immuno- precipitated with anti-C5b-9 or anti-C3 IgG and analyzed by 10% SDS- PAGE, Western blot, and immunoreacted with affinity-purified anti- Gial,z IgG, anti-GiaJGoa IgG, and also with anti-GB IgG followed by peroxidase-conjugated secondary antibodies, then the reactivity was developed by ECL. Immunoreactive bands with anti-Ga subunits mi- grated as 41 kDa and that with anti-GB migrated as 36 kDa, identical to respective Ga or GP subunit standards (Gp-st).

I PTX

Y

0 4 6 8 10 1 2 1 4

Time, min

6

TCC. TCC-G protein complexes isolated from JY25 cells as in Fig. 6 FIG. 9. Functional activity of G protein coprecipitated with

were incubated with [3sSlGTPyS, as described in the legend to Fig. 1A, with (D) and without (0) PTX pretreatment. At the indicated time points, the binding of [3sS1GTPyS was determined as BA85 filter-asso- ciated radioactivity.

C5b-9 attack by actively eliminating potentially lytic TCC from the cell surface by vesiculation of membrane fragments and by endocytosis (2, 14). This process is mediated by signals gener- ated through TCC-membrane interaction (15, 17). As shown in

4422 G Protein Activation by Terminal Complement Complexes

100

? n

rr 5 40-

.S 60-

- 80

- CD

._ m

8 $ 20 -

" PTX

Control P

CTX t

L 1 I 1 1 1 10 20 30 40

Time, rnin

FIG. 10. Effects of PTX on CSb-8 elimination from the surface of living cell. JY25 cells treated with 500 ng /d €'"x (0) or CTX (0) for 3 h or with medium (0) were incubated with 100 pg/ml IgG and 20% C9D for 15 min to form C5b-8. The remaining C5b-8 was determined by adding 10 pg of C9 at the time indicated to convert the C5b-8 to lytic C5b-9. Cells were further incubated for 60 min at 37 "C, then cell death was determined by lactate dehydrogenase release. Data are expressed as percent C5b-8 remaining on the plasma membrane. The lactate dehydrogenase released from C5b-8 carrying cells treated with C9 at zero time was used as loo%, and the lactate dehydrogenase released from cells with C5b-8 at zero time was used as background control.

Fig. 10 pretreatment of cells with PTX completely abolished the C5b-8 removal from the cell surface as demonstrated in this kinetic assay which determines the rate of C5b-8 elimination by converting remaining C5b-8 to lytic C5b-9 on cell surface. The C5b-8 elimination was slightly increased by CTX, which was not effective in blocking C5b-8 removal (Fig. 10).

DISCUSSION In this study we demonstrated that PTX-sensitive G proteins

were activated during the assembly of TCC, which include C5b-7, C5b-8, and C5b-9 complexes. As shown in Figs. 1 and 2, antibody-dependent activation of serum complement induced GTPyS binding and GTP hydrolysis. To assess the role of TCC, serum depleted of terminal complement proteins, C7, C8, or C9, was used to assemble various TCC complexes. The use of C7D on IgG-sensitized membranes allows complement activation up to C6 and generates inflammatory peptides, such as C3a and C5a, that could activate G proteins through specific receptors (36). Since activation of C7D produced minimal effect, G pro- tein activation by serum complement was primarily due to the TCC assembly. As in DAG and ceramide generation (171, G protein activation was first noted with C5b-7 assembly, which increased further with C5b-8 and C5b-9 (Fig. 4). Although PTX- induced ADP-ribosylation of G proteins was much less in JY5 than JY25 (Figs. 5 and 7), increase in G protein activity by complement was similar between JY25 and JY5. This finding, together with the higher GTPyS binding activity in JY5, may be explained by the fact that TCC formation is four times more efficient in JY5 cells deficient in glycosylphosphatidylinositol- anchored proteins (17, 32).

Direct association of inserted TCC peptides with G proteins was demonstrated by the presence of PTX-sensitive Ga in im- munoprecipitates obtained with antiX5b-9 neoantigen, anti- C7, or anti-C8, but not with ant i43 antibody from TCC-carry- ing cell or membrane lysates. When cells treated with heat- inactivated serum were immunoprecipiated with anti-C5b-9, PTX failed to induce Ga ADP-ribosylation. The Ga precipitated with TCC was a 41-kDa protein, migrating in an identical manner as the ADP-ribosylated Gicu by PTX in untreated plasma membranes. The capacity for PTX to ADP-ribosylate Ga was reduced in membranes carrying preformed TCC. It is possible that certain amino acid residues of Ga, that are targets for PTX after TCC interaction, may not be available for ADP-

ribosylation in intact membranes due to steric hindrance, in contrast to easy accessibility of such amino acids in TCC-G protein complexes isolated from detergent lysates. This 41-kDa protein identified by ADP-ribosylation was reactive with anti- bodies specific for GialVz and also with antibodies specific for Gia3/Goa, indicating that TCC may interact with more than one G protein. The presence of Gp subunit in anti-C5b-9 im- munoprecipitates suggested that G proteins complexed with TCC may exist as a trimeric form, that may also explain the ability of PTX to ADP-ribosylate Ga after elution from the membrane. In addition, the isolated TCC-G protein complexes contained functionally active Ga, as shown by the capacity to bind GTPyS, which was blocked by PTX. The nature of molecu- lar interaction between TCC peptides and G proteins is un- known. It is possible that TCC-G protein interaction may be similar to that of mastoparan, a tetradecapeptide toxin from wasp venom which activates heterotrimeric Gi, Gs, and Go in a receptor-independent manner by binding to the amphiphilic helix of Ga (27-291, as well as small G proteins from ras-related rholrac family (39). Although, diverse G proteins sensitive to PTX or CTX were present in JY membranes, TCC peptides predominantly interacted with GidGoa species with a molecu- lar size of 41 kDa in this cell line. Whether TCC would interact with G proteins with a similar restriction in other cell types remains to be determined.

Physiological significance of G protein activation during sub- lytic complement attack was evaluated by testing the TCC elimination from the cell surface by endocytosis and membrane shedding, processes known to depend on signals generated by TCC (2, 15, 17). Elimination of C5b-8 was completely inhibited by P T X , as shown by the functional assay. On the other hand, CTX, which failed to block the elimination, showed a slightly enhanced TCC effect. Interestingly, endocytosis or exocytosis are influenced by GTPyS, [AlF&, or mastoparan, also through a process involving activation of GidGoa (3M2). In addition, other biologically important inflammatory mediators, such as interleukin-8, tumor necrosis factor, fMLP (formylmethionyl- leucylphenylalanine), and C5a act on target cells through a family of receptors that are linked to Gi proteins (43-46). Therefore, elimination of TCC from cell surface may be medi- ated through activation of PTX-sensitive GVGo proteins by a receptor-independent mechanism.

Formation of sublytic TCC occurs during complement acti- vation under diverse pathophysiological conditions in vivo, as disparate as infection, ischemia, tissue necrosis, and neoplasia (47-50). Activation of target cells by TCC through involvement of G protein, in parallel with the increase in [Ca2+li, may serve as bioregulatory mechanisms important in host defense by pro- moting cellular repair and tissue healing.

REFERENCES 1. Shin, M. L., and Carney, D. F. (1988) Prog. Allergy 40 ,4441

3. Bhakdi, S., and TranumJensen, J. (1983) Biochim. Biophys. Acta 737, 343- 2. Morgan, B. P. (1989) Biochem. J. 264,l-14

372 4. "mhopp, J., and Podack, E. R. (1988) in Cytolytic Lymphocytes and Comple-

ment:El)*ectors ofthe Immune System (Podack. E. R.. ed) Vol. 1, pp. 185-206,

5. Ram, L. E., Michaels, D. W., Whitlow, M. B., and Mayer, M. M. (1983) in CRC Press Inc., Boca Raton, FL

BioloaicalResDonse Mediators and Modulators (August, T., ed) PP. 117-139, Academic Press, San Diego, CA

6. Hansch, M. G., Hammer, C. H., Van@, P., and Shin, M. L. (1981) P m . Natl.

7. Nicholson-Weller, A., Burge, J., Fearon, D. T., Weller, P. F., and Austen, K F.

8. Schonermark, S., Rauterberg, E . W., Shin, M. L., Lake, S., Roelcke, D., and

9. Sugita, Y., Nakano, Y., and Tomita, M. (1988) J. Biochem. (Tokyo) 104,633-635

Acad. Sci. U. S. A. 78,5112-5121

(1982) J. Imnrmnol. 129,184-189

Hansch, G. M. (1986) J. Zmmunol. 138, 1772-1776

10. Hansch, G. M. (1992) Zmmunopharmacology '24,107-117 11. Vanguri, P., and Shin, M. L. (1988) J. Biol. Chem. 263,7228-7234 12. Seeger, W., Suttorp, N., Hellwig, A, and Bhakdi. S . (1986) J. Zmmunol. 157,

13. Halperin, J. A,, 'hratuska, A,, and Nicholson-Weller, A. J. (1993) J. Clin. 1286-1293

G Protein Activation by Terminal Complement Complexes 4423

14. Carney, D. F., Hammer, C., and Shin, M. L. (1986) J. Immunol. 137.263-270 15. Carney, D. F., Lang, T. J., and Shin, M. L. (1990) J. Immunol. 146, 623429 16. Wiedmer. T., Ando, B., and Sims, P. J. (1987) J. Biol. Chem. 262,13674-13681 17. Niculescu, F., Rus, H., Shin, S . , Lang, T., and Shin, M. L. (1993) J. Immunol.

18. Birnbaumer, L., Abramowitz, J., and Brown, A. M. (1990) Biochim. Biophys.

20. Cockcroft, S . (1992) Biochim. Biophys. Acta 1113, 135-160 19. Conklin, B. R., and Bourne, H. R. (1993) Cell 73,631-641

21. Takei, Y.. Kurosu, H., Takahashi, K, and Katada, T. (1992) J. Biol. Chem. 267,

22. Bommakanti, R. K., Bokoch, G. M., 'Iblley, J. O., Schreiber, R. E., Siemsen, D.

23. Yang, L., Camoratto, A. M., Bat& G., hj, S. , Manning, D. R., and Williamson,

24. Wu, D., Lee, C. H., R h e e , S . G., and Simon, M. I. (1992) J. Biol. Chem. 267,

25. Yatani, A,, and Brown, A. M. (1991) J. Biol. Chem. 266,22872-22877

27. Higashijima, T., Burnier, J., and Ross, E. M. (1990) J. Biol. Chem. 265,14176- 26. Antonny, B., and Chabre, M. (1992) J. Biol. Chem. 267,6710-6718

28. Weingarten, R., Ransnas, L., Mueller, H., Sklar, L. A,, and Bokoch, G. M.

29. Higashijima, T., and Ross, E. M. (1991) J. Biol. Chem. 266,12655-12661

31. Hollander, N., Selvaraj, P., and Springer, T. A. (1988) J. Immunol. 141,4283- 30. Steinbuch, M., and Audran, R. (1969) Arch. Biochem. Biophys. 134,279-284

32. Hollander, N., Shin, M. L., Rosse, W. F., and Springer, T. A. (1989)J. Immunol.

33. Thom, D., Powell, A. J., Lloyd, C. W., and Rees, D. A. (1977) Biochem. J. 168,

huest. 91, 1974-1978

160,214-224

Acta 1031, 163-224

5085-5089

W., Klotz, K N., and Jesaitis, A. J. (1992) J. Biol. Chem. 267, 7576-7581

J. R. (1993) J. Biol. Chem. 286,37393746

1811-1817

14186

(1990) J. Biol. Chem. 266,11044-11049

4290

142,3913-3916

187-194 34. Imamura, K, and Kufe, D. (1988) J. Biol. Chem. 263,14093-14098 35. O'Neill, L. A. J., Bird, T. A., Gearing, A. J. H., and Saklatvala, J. (1990) J. Biol.

36. Siciliano, S . J., Rollins, T. E., and Springer, M. S . (1990) J. Biol. Chem. 286,

37. Luttrell, L., Kilgour, E., Lamer, J., and Romero, G. (1990) J. Bwl. Chem. 265,

38. Laemmli, U. K. (1970) Nature 227,680-685 39. Koch, G., Haberman, B., Mohr, C., Just, I., and Aktories, K. (1991) FEBS Lett.

40. Lillie, T. H. W., and Gomperts, B. D. (1992) Biochem. J. 288,181-187 41. Leyte, A,, Barr, F. A., Kehlenbach, R. H., and Huttner, W. B. EMBO J. 11,

4795-4804 42. Carter, L. L., Redelmeier, T. E., Woollenweber, L. A,. and Schmid, S . L. (1993)

J. Cell Biol. 120, 3745 43. Rollins, T. E., Siciliano, S . , Kobayashi, S., Cianciarulo, D. N., Bonilla-Argudo,

V., Collier, K, and Springer, M. S . (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 971-975

Chem. 265,314&3152

19568-19574

16873-16879

291,336-340

44. Schepers, T. M., and McLeish, K. R. (1993) Bioehem. J. 289,469473 45. Kupper, R. W., Dewald, B., Jakobs, K H., Baggiolini, M., and Gierschik, P.

46. Yanaga, F., Abe, M., Koga, T., and Hirata, M. (1992) J. Biol. Chem. 287,

47. Sanders, M. E., Kopicky, F., Wigley, F., Shin, M. L., Frank, M. M., and Joiner,

48. Rus, H. G., Niculescu, F., and Vlaicu, R. (1987) Immunol. Lett. 16, 15-20 49. Koski, C. L., Saunders, M. E., Swoveland, P. T., Lawley, T. I., Shin, M. L.,

50. Niculescu, F., Rus, H. G., Retegan, M., and Vlaicu, R. (1992) Am. J. Pathol.

(1992) Biochem. J. 282,429434

5114-5121

K A. (1986) J. Rheumatol. 13, 102S1034

Frank, M. M., and Joiner, K. A. (1987) J. Clin. Inuest. 80,149Z1502

140, 1039-1043