Proc. Nat!. Acad. Sci. USAVol. 86, pp. 6724-6728, September 1989Immunology

Surface immunoglobulin-mediated B-cell activation in the absenceof detectable elevations in intracellular ionized calcium: A modelfor T-cell-independent B-cell activation

(signal transducton/phosphatidyilnosdl bbphoshte hydrolysis)

MARK BRUNSWICK*, CARL H. JUNEt, FRED D. FINKELMAN*, HOWARD M. DINTZIS*, JOHN K. INMAN§,AND JAMES J. MOND*

*Department of Medicine, Uniformed Services University of the Health Sciences, Bethesda, MD 20814; tNaval Medical Research Institute, Bethesda, MD20814; tDepartment of Biophysics, Johns Hopkins University, Baltimore, MD 21205; and Laboratory of Immunology, National Institutes of Health, Bethesda,MD 20892

Communicated by Michael Heidelberger, May 30, 1989

ABSTRACT We recently showed that anti-Immunglob-ulin conjugated to high molecular weight dextran is 1000-foldmore mitogenic for B cells than unconjugated anti-immuno-globulin. This system serves as a model for T-cell-independentype 2 antigens such as haptenated Ficoll, dextran, and bac-terial polysaccharides, which can also stimulate B-cell prolif-eration and antibody production at low concentrations. Weshow here that conjugated anti-immunoglobulin, at concentra-tions that stimulate significant increases in expression of majorhistocompatibility complex class II molecules and incorpo-ration of thymidine into DNA, does not induce detectablemodulation of surface inmunoglobulin. These results indicatethat the facilitated T-cell-independent B-cell activation bypolysaccharide antigens may result from inability to modulatesurface immunoglobulin, possibly resulting in persistentand/or repetitive signaling. Early large increases in Ca2+ andbreakdown of inositol phospholipids presently thought to beinvolved in transduction of the mitogenic signal are not detect-able at low concentrations ofconjugated anti-immunoglobulin,raising the possibility that these biochemical events may not infact be central to this signaling pathway.

Crosslinking of B-cell-surface immunoglobulin by antigen oranti-Ig antibody has been demonstrated to induce phospho-lipase C-dependent phosphatidylinositol bisphosphate (PIP2)hydrolysis (1-6). The second messengers thereby generated,diacylglycerol (DAG) and inositol trisphosphate (IP3), lead toactivation of protein kinase C (PKC) and mobilization ofintracellular ionized calcium ([Ca2]jD (reviewed in refs. 7, 8),events that facilitate entry of B cells into G1 and S phases(9-14). Stimulation of B cells by other agonists does notappear to stimulate an inositol phospholipid-dependent modeof activation. Thus, stimulation of murine B cells withlipopolysaccharide or 8-mercaptoguanosine leads to DNAsynthesis in the absence of detectable increases in [Ca2+]i(ref. 15 and our unpublished observations). Likewise, stim-ulation of human B cells with antibodies to the CD20 cluster(16) or to Bgp95 (17) either alone or in combination does notstimulate increases in [Ca2+]i, indicating that many alterna-tive routes of activation can be recruited by B cells afterstimulation by agonists that do not bind to surface immuno-globulin (sIg).

Recently we showed that when anti-Ig antibody is conju-gated to high molecular weight dextran (anti-Ig-dex) it stim-ulates B-cell proliferation at concentrations 1000-fold lowerthan that stimulated by unconjugated anti-Ig antibody (18).Furthermore, at these low concentrations anti-Ig-dex induces

greater B-cell proliferation than that induced by unconju-gated anti-Ig but induces no modulation of slg and very lowlevels of binding to slg. This suggests that sIg-modulatedB-cell activation could be stimulated despite low-level re-ceptor occupancy and led us to investigate whether theseconjugates stimulate B-cell activation via stimulation of PIP2hydrolysis similar to that induced by high concentrations ofunconjugated anti-Ig antibody. The data show that anti-Ig-dex stimulates greater accumulation ofinositol phosphatesand greater increases in [Ca2+]i as compared to comparableconcentrations of anti-Ig. However, at lower concentrationsthat remain fully mitogenic we can detect no such increases.These data indicate that early large increases in [Ca2l]i maynot be essential for slg-mediated B-cell activation.

MATERIALS AND METHODSMice. Female DBA/2J mice (from The Jackson Laborato-

ries), 8 weeks old, were used within 4 weeks of arrival.Antibody and Other Reagents. Anti-T-cell antibodies for

B-cell purification were anti-Thyl.2 (clone 30-H12), anti-L3T4 (clone GK1.5), anti-Lyt-2 (clone 53-6.7), and rat anti-mouse immunoglobulin K chain (clone MAR18.5). Otherantibodies were anti-Fc-y receptor (clone 24G2), anti-IgD ofa allotypes (clone H5a/1 and clone FF1), fluoresceinatedanti-Iad (clone MKD6), fluoresceinated anti-IgD of the aallotype (clone AMS15), and IgD (clone TEPC1017). Allantibodies (except 53-6.7 and MAR18.5, which were used ascrude ascitic fluid and tissue culture fluid, respectively) werepurified by precipitation with (NH4)2SO4, DE-52 ion-exchange chromatography, and filtration with Sephadex G-200 gel as described (18). Indo-lAM was from MolecularProbes, and Dowex 1 formate was from Bio-Rad. [3H]-Thymidine (specific activity, 20 Ci/mmol; 1 Ci = 37 GBq) andmyo-[3H]inositol were from New England Nuclear.

B-Cell Purification. Resting B cells were purified as de-scribed (18): spleen cell suspensions were depleted ofT cellsby a cocktail of anti-T cells (30-H12, GK1.5, 53-6.7) followedby treatment with rabbit complement in the presence ofMAR18.5. Resting B cells were separated on a discontinuousPercoll gradient (Pharmacia). Cells banding at the 65-70%interface were then isolated and used as resting cells.

Antibody-Dextran Conjugates. Antibodies were coupled todextran (Mr 2 x 106) by the heteroligation technique recentlydescribed (18).

Abbreviations: PKC, protein kinase C; [Ca2+]i, intracellular ionizedcalcium; anti-8, antibody to the heavy chain of IgD; dex, dextran;anti-8-dex, anti-8 coupled to dextran; PIP2, phosphatidylinositolbisphosphate; DAG, diacylglycerol; IP3, inositol trisphosphate; sIg,surface immunoglobulin; MHC, major histocompatibility complex.

6724

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.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

2, 2

020

Immunology: Brunswick et al.

and-&d

1200 2 lo I A -,.

100017--i/.1 2 3 4 5

TIME (minutes)

Proc. Natl. Acad. Sci. USA 86 (1989) 6725

B

1200

6 71000

800

400

0.001

*-_-_.a7.

4 I

FIG. 1. Calcium fluxes in B cells stimulated by anti-5-dex and anti-8. Resting B cells were analyzed for 1 min to establish an appropriatebaseline after which ligand was added at the indicated concentration of anti-8-dex (A) or anti-8 (B) and analysis was continued for 6 min. Theratio of violet:blue was converted to Ca concentration (nM). Results are expressed as the mean Ca levels of the population vs. time. Controlsamples were run with medium, dex, or 24G2-dex with similar results. Cells were stimulated by various concentrations of anti-8-dex conjugatesand analyzed for the indo-1 violet:blue fluorescence ratio (directly proportional to [Ca2]i), which was digitally calculated in real time for eachindividual cell and plotted on the y axis against time on the x axis (C). Data were accumulated from -500 cells per sec and the ratio within eachcell was plotted. Samples are analyzed for 6 min 40 sec and each channel represents data for 4 sec. Results are displayed on a 100 x 100 pixelgrid on which the number of cells per pixel is represented by a 16-color progression according to an exponential function in which 0 to 1 cellis assigned a color of black, 2-10, shades of blue, 11-80, shades of green, and 81-600, shades of yellow to red. The gap in analysis occurredwhen ligand was added to the suspension.

A

800 _

IIs

ansi-

100

600o

400 1_

200 1

0 1 2 3TIME (mI"nu)

2TIME (ins)

3 4

Dow

nloa

ded

by g

uest

on

Mar

ch 2

2, 2

020

6726 Immunology: Brunswick et al.

Analysis of [Ca2J]i. B lymphocytes were labeled at 5 x 107cells per ml with 1.5 gM indo-1 for 30 min, washed, andanalyzed for [Ca2+]J as described in detail (19). Five hundredcells per second were analyzed for [Ca2+]i; responses wereindependent of the cell concentration used. The violet-blueratio obtained from the cytofluorograph was converted tocalcium concentration (nM) and digitally calculated in realtime for each individual cell. In some experiments the cu-mulative calcium influx was determined by finding the areaunder the curve of "mean [Ca2+]i vs. time" after stimulationand is expressed as tM-sec of calcium. Net cytoplasmiccalcium elevation in the 6 min after stimulation with mediumis subtracted from the net cytoplasmic calcium elevationinduced by ligands.

Proliferative Responses. B-cell proliferation was deter-mined by culturing 105 cells per well in 96-well flat-bottommicrotiter plates for 48 hr in Mishell-Dutton medium. Onemicrocurie of [3H]thymidine was added for the last 18 hr ofculture and proliferative responses were determined by har-vesting onto glass fiber strips (PHD, Cambridge, MA). Spe-cific incorporation was analyzed by liquid scintillation spec-troscopy and results are expressed as mean CPM of thymi-dine incorporation in triplicate wells; standard deviation fromthe mean was always <10% and has been omitted from thefigures for the sake of clarity (18).

Determination of Expression of Class H Maijor Histocom-patibility Complex (MHC) Antigens. Cells cultured overnightwere stained for expression of class II MHC antigens withfluoresceinated MKD6 (anti-Iad) or for sIgD with fluores-cence labeled monoclonal antibody to the heavy chain of IgD(anti-8) (AMS15), which binds to an epitope of IgD notblocked by the mitogenic anti-8 (H18/1). Results are ex-pressed as channel scale area, which is the sum of theproducts of cell number and their staining intensity.Dete tion ofPIP2 Hydrolysis. For measurement ofPIP2

breakdown B cells were cultured overnight at 2 x 107 cellsper ml in the presence of myo-[3H]inositol. (20 ,uCi/ml) (5),washed three times, placed in polypropylene tubes at 5 x 106per ml at 0.2 ml per tube in the presence of 10 mM LiCl and24G2 (10 ,ug/ml), and kept at 37°C for 30 min. Ligands wereadded and incubation was continued for 40 min. Water-soluble inositol phosphates were isolated by addition of 0.75ml of chloroform/methanol (1:1) and then addition of 0.25 mlof chloroform followed by 0.25 ml of H20. Cultures weremixed in a Vortex and spun at 1000 rpm and the water phasewas separated and applied to a Dowex column from which theinositol phosphates were eluted with 1 M sodium formate/0.1M formic acid.

RESULTS AND DISCUSSIONThese studies were undertaken to determine whether lowconcentrations of anti-IgD-dex (anti-8-dex), which stimu-lated high levels of B-cell proliferation (18), also stimulatedcomparable increases in early events ofB-cell activation suchas PIP2 breakdown and increases in [Ca2+]i (1-5) induced byhigh concentrations of unconjugated anti-Ig. Although 1 and0.1 ,ug of anti-8-dex per ml stimulated increases in [Ca2+]i in>95% ofB cells from a resting level of 130 nM to 864 nM and690 nM, respectively, 100-fold higher concentrations of un-conjugated anti-8, (100 and 10 ,g/ml) stimulated increasesonly to 439 nM and 291 nM, respectively (Fig. 1), and theincreases in [Ca2+]i were more sustained than that by uncon-jugated anti-8 (Fig. 1). Low concentrations of anti-8-dex,which were highly mitogenic, stimulated barely detectableincreases in levels of [Ca2+]i measured early (6 min) (Fig. 1Inset) or later (3.5 and 4.5 hr) (Table 1). In the experimentshown in Fig. 2, 10-3 ,ug of anti-8-dex per ml stimulatedthymidine incorporation, which was =35% as great as themaximal response, but resulted in an increase in [Ca2+], that

Table 1. Analysis of the mean [Ca2+]i in individual B cellsstimulated for 3.5 and 4.5 hr with anti-B-dex antibodies

[Ca2+]i, nMStimulus, ,ug/ml 3.5 hr* 4.5 hr*

Control 130.8 130.8anti-8-dex

10-2 228.5 235.0lo-, 161.7 167.6lo-4 130.1 134.9

Resting B lymphocytes were cultured at 370C with medium only orwith anti-6-dex at 10-2, 1O-3, or 1o-4 1Lg/ml for 3.5 or 4.5 hr. Thirtyminutes before the end of the culture period indo-1 was added to thecells and they were analyzed for the indo-1 ratio of violet:bluefluorescence. Histograms were generated to show the distribution offluorescence within the population and were then further analyzed todetermine the mean [C2 +i of the population. Results are expressedas mean [Ca2+]i in nM.*Stimulation time.

was <4% of that observed when cells were maximallystimulated. At this low concentration (10-3ig/ml) of anti-8-dex '-25% of the B cells were in the G2, S, and M stages ofthe cell cycle (as determined by propidium iodide staining)compared to 43% of B cells that were stimulated with 10-1,ug/ml (data not shown). This suggests that the high level ofthymidine incorporation that we observe does not simplyreflect a small number of B cells that synthesize largeamounts ofDNA but rather reflects a large fraction of the Bcells that are induced to enter the cell cycle. Concentrationsof anti-6-dex that stimulated slight or no detectable earlyincreases in B-cell [Ca2e]i were much more mitogenic thanconcentrations of unconjugated anti-8 that induced substan-tial [Ca2e]i increases (Fig. 2). The data in Fig. 2 were derivedby integrating the area under the curve of "mean calcium vs.time" (from an experiment similar to that described in Fig. 1).Since this computation is a summation of the mean of thecalcium transients from analysis of each individual cellanalyzed over the duration of the experiment (6 min), itprovides a very accurate and sensitive way of detecting smallbut potentially significant changes in a population of B cells.The assay system that we employ analyzes events in single

anti-8-dex anti-8

0V-

x

E

a.

1o0- 101 i0n3Concentration, jsg/ml

C,0:.

E

C)co0

FIG. 2. Comparison of anti-Ig-mediated stimulation of increasesin [Ca2]1i and thymidine incorporation in B lymphocytes. To com-pare directly changes in [Ca2+]i and anti-Ig-mediated proliferativeresponses, resting B cells were analyzed for cumulative flux bydetermining the area under the curve, plotted for mean [Ca2+]i vs.time after stimulation, and expressed as ,uM-sec of calcium. Netcytoplasmic Ca elevation that occurred in the 6 min after stimulationwith medium is subtracted from that induced by various anti-8 andanti-5-dex concentrations. Cells from the same preparation werecultured in triplicate in 96-well microculture plates at 105 cells perwell and indicated concentrations of anti-S-dex or anti-8 were added,specific incorporation was determined, and results were plotted asthe mean cpm of thymidine incorporation of triplicate cultures.

Proc. Natl. Acad Sci. USA 86 (1989)

Dow

nloa

ded

by g

uest

on

Mar

ch 2

2, 2

020

Proc. Natl. Acad. Sci. USA 86 (1989) 6727

cells and is capable of detecting responses in <1% of apopulation of cells (19). The data shown were derived fromindividually determined responses of -500 cells analyzed persecond. Although we do not measure calcium transients inindividual cells over real time as was done by Wilson et al.(20), we do analyze the response of -180,000 cells over thecourse of a 6-min experiment. The dot plot (Fig. 1C) displaysthe response of individual cells and demonstrates that thereduction of data to the mean is justified. When 24G2 wasadded to the cultures containing anti-6 or anti-&.dex tominimize Fc-y receptor-mediated inhibition (21), there wasno qualitative difference in the results. These data indicatethat mitogenicity mediated by crosslinking of sIg with anti-8-dex appears to be independent of the levels of stimulatedincreases in [Ca2+]1.

If the relevant stimulus for an excitatory event in B cellswere the establishment of a domain (22) of crosslinked B-cellslg, multivalent ligands, such as anti-8-dex, would be espe-cially effective in establishing such domains. Thus, a singleinteraction between a molecule of anti-8-dex and a B cellmight be more likely to trigger excitation than an interactionbetween the same cell and unconjugated, divalent anti-S. Atthe very low concentrations of anti-8-dex observed as stillmitogenic, ligand-receptor interactions might also occur atlow levels. Such infrequent and low-level B-cell-ligand in-teractions, though providing effective signal transduction, asreflected by increased expression ofMHC class II molecules,do not modulate much IgD from the B-cell surface at all timestested, including 2, 6, 11 (data not shown), and 24 hr (Table2) after addition of antibody. Absence of sIg modulationincreases the likelihood of subsequent interactions of anti-5-dex with sIgD, which would allow repetitive B-cell signal-ing to occur over longer periods. Lack of a direct relationshipbetween modulation of sIg, PIP2 hydrolysis (Table 2), andB-cell activation is supported by the observation that con-centrations of anti-6-dex too low to induce detectable in-

Table 2. B-cell stimulation induced by anti-B-dex does notcorrelate with its ability to induce receptor modulation,increases in [Ca2+1], or increases in PIP2 breakdown

Water-solubleinositol CSA

Stimulus, Mean max phosphates,,ug/ml [Ca2+]i, nM cpm IgD+ cells Ia+ cells

None 150 98 24,200 12,000anti-B-dex

1 848 1110 1,600 31,40010-1 668 606 10,300 32,00010-2 395 134 12,700 31,30010-3 196 78 19,200 28,70010-4 157 83 24,700 26,300lo-, 146 ND 25,000 21,000

anti-8100 393 346 1,000 31,70010 265 308 1,000 29,4001 194 286 1,200 28,800

10-1 158 132 3,800 23,70010-2 150 103 14,400 18,300l0o- 143 ND 22,400 12,600

Resting B lymphocytes were placed in culture at 106 per ml in thepresence of indicated concentrations of anti-8-dex or anti-8 + 10 ,ugof 24G2 per ml. Results are expressed as channel scale area (CSA),the sum of the products of cell number and their staining intensity.Results are compared to mean max. [Ca2+]i levels were obtainedfrom cells treated as described in the legend to Fig. 1. Mean max isthe mean of four consecutive channels of data that attained highestlevels of [Ca]2+, in a 16-sec period during which i500 cells per secwere analyzed. Parallel cultures were analyzed for PIP2 hydrolysis.Results presented are representative of five different experimentsthat gave similar results. ND, not determined.

creases in mean B-cell [Ca2+]1 or in breakdown of PIP2 (asmeasured over 40 min) are also too low to modulate B cellsIgD yet induce considerable increases in B-cell expressionof class II MHC and B-cell proliferation (Table 2 and Fig. 2).This is not the case when B cells are stimulated withunconjugated bivalent anti-6, which would theoretically haveto be present at considerably greater concentration thananti-&.dex to crosslink sIgD sufficiently to activate B cells.These high concentrations required for B-cell activation arealso very effective at modulating B-cell sIg (Table 2) and thusmay limit their own capacity to activate B cells repetitively.To examine this hypothesis more carefully we precultured

B cells for 2 hr with a nonstimulatory anti-6 antibody (FF1)(23) demonstrated to modulate surface IgD and then added tothe cultures anti-S-dex antibody over a broad range of mito-genic concentrations. We reasoned that the mitogenicity ofthe lower concentrations of anti-8-dex antibody was depen-dent on their relative inability to modulate sIgD and hencewould only stimulate proliferation in B cells whose sIg hadnot been modulated during the 2-hr preculture period. Incontrast, higher concentrations of anti-8-dex antibody that

50,000 1 A

40,000

30,000

20,000-

10,000

Cez0a. 0USwr

40,0000IL

00r. 'f

20,000

10,000

E2Normal B cellsB sig-modulated B cells

0.01 0.

and-8-dex (Wg/ml)

B O No additionmIgD, 0 hrM IgD, 6 hr0 IgD, 24 hr

ADDITION

FIG. 3. (A) Anti-8-dex-stimulated B-cell proliferation in sIgD-modulated cells. B cells were cultured at 8 x 106 wells per ml inmedium only or with the nonmitogenic anti-6 antibody FF1 at 400,ug/ml for 2 hr to modulate sIg and then cultured with variousconcentrations of anti-3-dex in the absence or continued presence ofFF1 at 100 ±g/ml. Specific incorporation was determined, andresults are expressed as the mean cpm of thymidine incorporation oftriplicate wells. (B) Interruption of anti-8-dex-mediated B-lymphocyte activation by addition of purified IgD at various timesafter culture initiation. Resting B lymphocytes were placed in cultureat 105 cells per well with either medium or anti-5-dex (10-3i ,g/ml).At the indicated times, purified IgD (from the myeloma TEPC1017)was added at a final concentration of 50 ,ug/ml. Specific incorpo-ration was determined, and results are expressed as the mean cpm oftriplicate cultures.

Immunology: Brunswick et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

2, 2

020

6728 Immunology: Brunswick et al.

were both stimulatory and effective modulators of sIg wouldstimulate proliferation of B cells even when sIg had alreadybeen modulated. This proved to be the case, in that B cellswhose sIg had been modulated by prior exposure to thenonmitogenic anti-6 antibody responded without any signif-icant diminution to the higher concentrations (10-1 and 10-2,ug/ml) but not to the lower concentrations (10-3 or 1i-0Ag/ml) of anti-6-dex antibody (Fig. 3A). This supports theview that the low nonmodulatory concentrations of anti-6-dex antibody are mitogenic only when encountering B cellson which sIg molecules are not reduced in number and so maytransmit repetitive and prolonged signal transduction. Toaddress the issue of repetitive signal transduction directly weemployed purified IgD to inhibit competitively the binding ofanti-8-dex to sIgD and thereby serve to interrupt continuoussignal transduction. IgD was added at a 50,000-fold molarexcess at the start of culture as well as at 6 and 24 hr after theinitiation of a 48-hr culture (Fig. 3B). Addition of IgD as lateas 24 hr still had a significant suppressive effect (62%suppression) on anti-8-dex-stimulated B-cell proliferation.Taken together, these two experiments suggest that optimalstimulation of B-cellDNA synthesis by low concentrations ofanti-8-dex requires prolonged and repetitive signaling. Suchcells that undergo persistent signal transduction may notrequire the rapid and high-level increases in [Ca2+]i to initiateor maintain their activational cycle as is required by cellsstimulated by higher concentrations of unconjugated anti-Ig,which rapidly modulates slg. Rather, in such cells compara-ble biochemical events may occur at later time points, and atvery low levels, after the initiation ofcell activation and couldpersist over longer periods. Alternatively, low levels ofcalcium transients may be induced at the plasma membranethat may not be reflected in measurements of cytoplasmic[Ca2+],. Such a mode of B-cell activation may therefore notrequire the additional stimulation provided by T cells orT-cell-derived lymphokines and therefore can be consideredto be relatively T-cell independent.Other interpretations of these results that cannot be ex-

cluded by these data include these possibilities: (i) Very lowconcentrations of multimeric antigens or anti-Ig-dex mayrecruit inositol phospholipid-independent pathways of acti-vation as is stimulated by the polyclonal B-cell-activatorlipopolysaccharide (15) and recently also shown to be stim-ulated by ligation of the T-cell-antigen receptor (24, 25). Thispossibility is also consistent with our previous observationthat anti-Ig can deliver stimulatory signals to B cells depletedof PKC (26). (ii) The mitogenic effect of even low concen-trations ofanti-5-dex is dependent upon its capacity to induceindividual B cells to experience recurrent, transient increasesin [Ca2+], (20, 27). In view of the finding of Wilson et al. (20)that anti-8 induces oscillations in [Ca2+], that persist for15-60 sec and the observation that our system is sensitive tochanges in the level of [Ca2+], in <1% of cells that last for <1sec, we view this possibility as unlikely.

We thank C. Sarma for his expert technical assistance and Ms.Mary Wise for expert secretarial help. We are also very grateful to

D. Smoot for expert assistance with flow cytometry. This work wassupported in part by Uniformed Services University of the HealthSciences Protocol RO83BQ, National Institutes of Health Grant5ROlAI24273, and U.S. Naval Medical Research and DevelopmentCommand Grant G183CV.

1. Maino, V. C., Hayman, M. J. & Crumpton, M. J. (1975) Bio-chemistry 146, 247-252.

2. Coggershall, K. M. & Cambier, J. C. (1984) J. Immunol. 133,3382-3386.

3. Bijsterbosch, M. K. & Klaus, G. G. B. (1985) J. Exp. Med.162, 1825-1836.

4. Fahey, K. A. & DeFranco, A. L. (1987) J. Immunol. 138,3935-3942.

5. Mizuguchi, J., Yong-Yong, J., Nakabayaschi, H., Huang,K.-P., Beaven, M. A., Chused, T. & Paul, W. E. (1987) J.Immunol. 139, 1054-1059.

6. Grupp, S. A., Snow, E. C. & Harmony, J. A. K. (1987) Cell.Immunol. 109, 181-191.

7. Berridge, M. J. (1984) Biochem. J. 220, 345-349.8. Nishizuka, Y. Nature (London) 308, 693-697.9. Monroe, J. G., Niedel, J. E. & Cambier, J. C. (1984) J. Im-

munol. 132, 1472-1475.10. Monroe, J. G. & Kass, M. T. (1985) J. Immunol. 135, 1674-

1679.11. Rothstein, T. L., Barker, T. R., Miller, R. A. & Kolber, D. L.

(1986) Cell. Immunol. 102, 364-369.12. Klaus, G. G. B., O'Garra, A., Bijsterbosch, M. K. & Hart-

man, M. (1986) Eur. J. Immunol. 137, 2215-2219.13. Mizuguchi, J., Beaven, M. A., O'Hara, J. & Paul, W. E. (1986)

J. Immunol. 137, 2215-2219.14. Dennis, G., June, C., Mizguchi, J., Ohara, J., Witherspoon, K.,

Finkelman, F. D., McMillan, V. & Mond, J. J. (1987) J.Immunol. 139, 2516-2523.

15. Grupp, S. A. & Harmony, J. A. K. (1985) J. Immunol. 134,4087-4094.

16. Smeland, E. B., Beiske, K., Ohlsson, R., Holte, H., Ruud, E.,Blomhoff, H. K., Jeffries, R. & Godal, T. (1987) J. Immunol.138, 3179-3184.

17. Valentine, M. A., Clark, E. A., Shu, G. L., Norris, N. A. &Ledbetter, J. A. (1988) J. Immunol. 140, 4071-4078.

18. Brunswick, M., Finkelman, F. D., Highet, P. F., Inman, J. K.,Dintzis, H. M. & Mond, J. J. (1988) J. Immunol. 140, 3364-3372.

19. Rabinovitch, P. S., June, C. H., Grossman, A. & Ledbetter,J. A. (1986) J. Immunol. 137, 952-961.

20. Wilson, H. A., Greenblatt, D., Peonie, M., Finkelman, F. D. &Tsien, R. Y. (1987) J. Exp. Med. 166, 601-606.

21. Phillips, N. E. & Parker, D. C. (1984) J. Immunol. 132, 627-632.

22. Dintzis, H. M., Dintzis, R. L. & Vogelstein, B. (1976) Proc.Natl. Acad. Sci. USA 73, 3771-3775.

23. Goroff, D. K., Stall, A., Mond, J. J. & Finkelman, F. D. (1986)J. Immunol. 136, 2382-2392.

24. Sussman, J. J., Mercap, M., Saito, T., Germain, R. N., Bon-vini, E. & Ashwell, J. D. (1988) Nature (London) 334, 625-628.

25. O'Rourke, A. M. & Mesch, M. F. (1988) J. Biol. Chem. 263,18596-18597.

26. Mond, J. J., Feuerstein, N., Finkelman, F. D., Huang, F.,Huang, K.-P. & Dennis, G. (1987) Proc. Natl. Acad. Sci. USA84, 8588-8592.

27. Cuthbertson, K. S. R. & Cobbolt, P. H. (1985) Nature (Lon-don) 316, 541-542.

Proc. Natl. Acad. Sci. USA 86 (1989)

Dow

nloa

ded

by g

uest

on

Mar

ch 2

2, 2

020