Pasteurella multocida toxin, a potent mitogen, increases inositol 1,4,5-trisphosphate and mobilizes Ca2+ in Swiss 3T3 cells.

Pasteurella multocida toxin, both native and recombinant, is an extremely potent mitogen for Swiss 3T3 cells and acts to enhance the formation of total inositol phosphates (Rozengurt, E., Higgins, T., Changer, N., Lax, A.J., and Staddon, J.M. (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 123-127). P. multocida toxin also stimulates diacylglycerol production and activates protein kinase C (Staddon, J.M., Chanter, N., Lax, A.J., Higgins, T.E., and Rozengurt, E. (1990) J. Biol. Chem. 265, 11841-11848). Here we analyze, by [3H]inositol labeling and high performance liquid chromatography, the inositol phosphates in recombinant P. multocida toxin-treated cells. Recombinant P. multocida toxin stimulated increases in [3H]inositol 1,4,5-trisphosphate ([3H]Ins(1,4,5)P3) and its metabolic products, including Ins(1,3,4,5)P4, Ins(1,3,4)P3, Ins(1,4)P2, Ins(4/5)P, and Ins(1/3)P. The profile of the increase in the cellular content of these distinct inositol phosphates was very similar to that elicited by bombesin. Furthermore, recombinant P. multocida toxin, like bombesin, mobilizes an intracellular pool of Ca2+. Recombinant P. multocida toxin pretreatment greatly reduces the Ca2(+)-mobilizing action of bombesin, consistent with Ca2+ mobilization from a common pool by the two agents. The enhancement of inositol phosphates and mobilization of Ca2+ by recombinant P. multocida toxin were blocked by the lysosomotrophic agents methylamine, ammonium chloride, and chloroquine and occurred after a dose-dependent lag period. The stimulation of inositol phosphate production by recombinant P. multocida toxin persisted after removal of extracellular toxin, in contrast to the reversibility of the action of bombesin. Recombinant P. multocida toxin, unlike bombesin and guanosine 5'-O-(gamma-thiotriphosphate), did not cause the release of inositol phosphates in permeabilized cells. These data demonstrate that recombinant P. multocida toxin, acting intracellularly, stimulates the phospholipase C-mediated hydrolysis of phosphatidylinositol 4,5-bisphosphate.

Sci. U. S. A. 87, 123-127). P. multocida toxin also stimulates diacylglycerol production and activates protein kinase C (Staddon, J. M., Chanter, N., Lax, A. J., Higgins, T. E.,  J. Biol (1,4,5)P3) and its metabolic products, including Ins(1,3,4,5)P4, Ins(1,3,4)P3, Ins(1,4)Pz, Ins(4/5)P, and Ins( 1/3)P. The profile of the increase in the cellular content of these distinct inositol phosphates was very similar to that elicited by bombesin. Furthermore, recombinant P . multocida toxin, like bombesin, mobilizes an intracellular pool of Ca2+. Recombinant P . multocida toxin pretreatment greatly reduces the Ca2+-mobilizing action of bombesin, consistent with Ca2+ mobilization from a common pool by the two agents. The enhancement of inositol phosphates and mobilization of Ca2+ by recombinant P . multocida toxin were blocked by the lysosomotrophic agents methylamine, ammonium chloride, and chloroquine and occurred after a dose-dependent lag period. The stimulation of inositol phosphate production by recombinant P . multocida toxin persisted after removal of extracellular toxin, in contrast to the reversibility of the action of bombesin. Recombinant P . multocida toxin, unlike bombesin and guanosine 5'-O-(y-thiotriphosphate), did not cause the release of inositol phosphates in permeabilized cells. These data demonstrate that recombinant P . multocida toxin, acting intracellularly, stimulates the phospholipase C-mediated hydrolysis of phos-phatidylinositol4,5-bisphosphate.
The mechanisms of action of bacterial toxins have provided novel insights into the control of cellular regulatory processes, including protein synthesis, ion channel activity, and signal * The work in the Birmingham laboratory was supported by the Medical Research Council and the Royal Society. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisenent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1) To whom correspondence should be sent. transduction (1). Indeed, cholera and pertussis toxins have been instrumental in the identification, respectively, of stimulatory and inhibitory G-proteins controlling adenylylcyclase activity (2, 3). To date, no intracellularly acting toxin has been described which leads to an activation of polyphosphoinositide-specific phospholipase C, another major transducer of transmembrane signaling (4). This phospholipase is involved in the action of many extracellular factors, including mitogenic neuropeptides and growth factors (5).
Recently, Pasteurella multocida toxin has been shown to be an extremely potent and effective mitogen for Swiss 3T3 cells, other established cell lines, and early passage cultures (6). The toxin is a monomeric 146-kDa protein and has been purified (7-13), cloned (14)(15)(16), sequenced (17,18), and expressed in Escherichia coli (14,15). The deduced amino acid sequence of P. multocida toxin did not reveal any significant homologies with other toxins or proteins (17,18). Both native and recombinant P. multocida toxin are mitogenic at picomolar concentrations (6). Several lines of evidence indicated that P. multocida toxin enters the cells and acts intracellularly to initiate and sustain DNA synthesis. Thus, a transient exposure of the cells to the toxin was sufficient to commit them to S phase and division. Furthermore, early but not lat,e addition of either the lysosomotrophic agent methylamine or P. multocida toxin antiserum selectively blocked the mitogenic action of recombinant P. multocida toxin (6).
Prior to the stimulation of DNA synthesis, recombinant P. multocida toxin enhanced the formation of total inositol phosphates (6). The toxin also increased the cellular content of diacylglycerol, caused the translocation of protein kinase C, and stimulated the phosphorylation of 80 K (19), a major substrate of protein kinase C (20-22). Furthermore, the binding of epidermal growth factor to its receptor was decreased by recombinant P. multocida toxin, an action attributable in part to protein kinase C activation (19). The stimulation of protein kinase C by recombinant P. multocida toxin, like its mitogenic action (6), required cellular entry and, via a methylamine-sensitive process, activation of the toxin. The toxin did not increase the cellular content of cyclic AMP (6).
The inositol phosphate species present in recombinant P. multocida toxin-stimulated Swiss 3T3 cells have not been analyzed in detail. It is not known if P. multocida toxin causes PtdIns(4,5)Pz' breakdown, as do many extracellular stimuli,  (23) were propagated as described previously (24). For experimental purposes, lo5 cells were subcultured in 33-mm Nunc dishes in 2.5 ml of DMEM containing 10% fetal bovine serum, 100 units/ml penicillin, and 100 Ng/ml streptomycin. The cultures were incubated in a humidified atmosphere of 10% C02, 90% air at 37 "C for 6-8 days before use, under the same conditions, in the experiments. After this time the cultures were confluent and arrested in the Cl/G, phase of the cell cycle (24). Secondary cultures of mouse embryo fibroblasts were subcultured in DMEM containing 10% fetal calf serum and then rendered quiescent by incubation for 4-6 days in DMEM containing 0.5% fetal bovine serum. HPLC Analysis of Inositol Phosphates-The cells were washed twice with 2 ml of DMEM:Waymouth medium (1:l) and then incubated for 20 h in 1 ml of this medium containing 25 pCi of [2-3H] inositol. Additions were made to the cells as required, and LiCl was added to a final concentration of 20 mM for the last 20 min of the incubation. Inositol phosphates were extracted by rapidly replacing the medium with 0.25 ml of ice-cold 5% (w/v) CCI3COOH containing 1 pl of phytate hydrolysate (2.3 pg of phosphorus/pl, prepared from phytic acid by the method of Wreggett et al. (25)). After 15 min at 4 "C, the cells were scraped into the medium with a rubber policeman, and the extracts from five dishes were combined. The dishes were then washed sequentially with 0.25 ml of 1% CC1,COOH which was then added to the extract. After a 5-min centrifugation at 10,000 X g at 4 "C, the CC13COOH in the supernatant was removed by extraction with 6 equal volumes of water-saturated diethylether. EDTA/Na, 0.1 M, pH 7.0, was then added to the extract to a final concentration of 5 mM. The extracts were stored at -20 "C until chromatography. HPLC analysis was performed essentially as described by Michell et al. (26) but using a Whatman 25-cm Partisphere SAX column. Inositol phosphate standards, I4C and 32P labeled as described below, were added to an aliquot of extract, and the volume was made up to 1 ml with deionized H20. The sample was injected onto the column (equilibrated with deionized H20) at 1 ml/min. The gradient was as follows, with pump A, containing deionized H20, and pump B, 0.1 M trilo)]tetraacetic acid; GTPyS, guanosine 5'-O-(y-thiotriphosphate); PBt2, phorbol 12,13-dibutyrate; Ins, InsP, InsPz, InsP3, InsP,, InsP5, and InsPG (and equivalent abbreviations with added locants, e.g. Ins(1,4,5)P3), myo-inositol and the appropriate myo-inositol phosphates (numbered by reference to D-myo-inositol 1-phosphate as InslP (see the recommendations of the Nomenclature Committee of the International Union of Biochemistry (1989) Biochem. J. 258, 1-2, in which when a single peak in the HPLC analysis is likely to contain more than one species, a hybrid abbreviation is used, e.g. Ins(l/3)P for a mixture of the enantiomers InslP and Ins3P)); GroPIns4P and GroPIns(4,5)P2, glycero-3-phospho-1-inositol 4phosphate and 4,5-bisphosphate, respectively; InsP,, total inositol phosphates present in a cell extract, analyzed as an unresolved pool. [32P]InsP6 was produced as described by Graf (33). ["C] Ins(1,3,4,5,6)P6 was obtained by HPLC from extracts of HL-60 cells labeled with [14C]inositol (Amersham Corp.).
FPLC Analysis of Inositol Phosphates-The cells were labeled, and an extract was prepared as for the HPLC analysis except that the cells were extracted into 3% (w/v) HCIO, instead of CC13COOH. After centrifugation, the supernatants were neutralized on ice with 0.5 M KOH containing 25 mM HEPES, 5 mM EDTA, and 0.01% phenol red. Precipitated KClO, was removed by centrifugation. FPLC analysis was essentially by the method of Meek (34) using a Pharmacia LKB Biotechnology Inc. 5/5 Mono Q column. The entire extract was diluted to 10 ml in 10 mM HEPES/Na, 0.1 mM EDTA, pH 7.4, and then loaded onto the column at 1 ml/min. After washing the column with the dilution buffer for a further 10 min to remove [3H]inositol the inositol phosphates were eluted by linearly increasing the concentration of Na2S0, (in 10 mM HEPES/Na, 0.1 mM EDTA, pH 7.4) from 0 to 80 mM in 25 min. The gradient was then increased to 110 mM Na2SOa in 30 min. Fractions (1 ml) were collected, mixed with 4 ml of Pic0 Fluor 15 (Packard), and then counted by liquid scintillation. The peak of radioactivity corresponding to Ins(1,4,5)P3 in the sample was assigned on the basis of coelution with standard [3H]Ins(1,4,5)P3.
Analysis of Total Inositol Phosphates-The cultures were labeled in 2 ml of medium as above but containing 10 pCi of [2-3H]inositol. Using one dish/data point, inositol phosphates were extracted by replacing the medium with 1 ml of ice-cold 3% HCIO,. After 15 min at 4 "C the extract was neutralized with KOH as described above. Analysis of total inositol phosphates was by anion exchange column chromatography (35). Samples were diluted to 10 ml with 5 mM Na2B40, and then loaded onto 1 ml of Dowex AG 1-X8 (200-400 mesh, HCOO-form) in Bio-Rad Econo-Columns. After washes with 4 X 10 ml of H20 and 2 X 8 ml of 60 mM NH4COOH, 5 mM Na2B407, inositol phosphates were eluted with 5 ml of 1 M NH,COOH, 0.1 M HCOOH. An aliquot (1 ml) of the eluate was counted in 10 ml of Pic0 Fluor 15.
Inositol Phosphate Release from Permeabilized Cells-Confluent and quiescent cultures of Swiss 3T3 cells were labeled with [2-3H] inositol as described for the HPLC analysis. The following procedure was then performed at 37 "C. The cultures were washed with 2 ml of a solution comprising 154 mM NaCl, 1 mM MgC12, 2 mM EGTA, 0.5 mM CaC12, 1 mM KH2P04, 10 mM HEPES, KOH to give pH 7.3 at 37 "C, and hence pCa 7 (from the fluorescence of fura-2). The cells were permeabilized for 4 min with 1 ml of "K-medium" containing 140 mM KCI, 20 mM NaC1, 2 mM MgC12, 1 mM ATP, 2 mM EGTA, 0.5 mM CaCI2, 25 mM HEPES, 1 mM KHzPO,, KOH (to give pH 7.3 and hence pCa 7 at 37 "C) and 60 pg/ml saponin. The permeabilized cells were then washed with 1 ml of K-medium lacking saponin but containing 5 mM LiCl and other agents as indicated. The reaction was terminated by adding 1 ml of ice-cold 6% HCIO, (w/v). After 30 min at 4 "C the acid extract was removed from the dish and neutralized with 1 M KOH containing 20 mM EDTA and 0.01% phenol red.   Total inositol phosphates were analyzed as described above.
Cell Ca'+-''Ca'+ was added to the medium (2-5 pCi/ml) in which the cells were cultured. After 16 h, additions were made directly to the labeling medium, and cellular 4sCa2+ was then determined by rapidly washing the cultures seven times with 2 ml of DMEM (pH 7.4) containing 3 mM EGTA (36). The cultures were solubilized with 1 ml of 0.1 M NaOH containing 1% sodium dodecyl sulfate which was then counted by liquid scintillation.
To determine the intracellular concentration of Ca2+, [Ca'+],, 5 X lo5 cells were subcultured with 10 ml of DMEM containing 10% fetal bovine serum in 90-mm Nunc dishes. After 6-8 days, when the cells were confluent and quiescent, the medium was replaced with 5 ml of DMEM:Waymouth medium (1:l). The cells were pretreated as indicated, and then 5 pl of 1 mM fura-2/AM (37-39) was added directly to the medium. After a further 10 min the cells were washed twice at 37 "C with 3 ml of electrolyte solution, comprising 140 mM NaCl, 5 mM KCl, 1.8 mM CaC12, 0.9 mM MgCI,, 25 mM ghCOSe, 16 mM HEPES, 6 mM Tris, pH 7.2, and the amino acid content of DMEM. The cells were then gently removed from the dish by scraping into 2 ml of the electrolyte solution. The cell suspension was transferred to a 1-cm' quartz cuvette and stirred at 37 "C for 4 min prior to any additions. Fluorescence was measured in a Perkin-Elmer LS-5 fluorometer with an excitation wavelength of 336 nm and an emission wavelength of 510 nm. [Ca"], was calculated from the maximum and minimum fluorescence of the fura-2, as described (37, 40).

RESULTS AND DISCUSSION
Chromatographic Analysis of Inositol Phosphates-To understand further the basis of P. multocida toxin action we have chromatographically analyzed the inositol phosphates present in recombinant P. multocida toxin-stimulated cells. Quiescent cultures of Swiss 3T3 cells were prelabeled with [2-3H]inositol for 16 h. The cells were then treated for a further 4 h with 20 ng/ml of recombinant P. multocida toxin, conditions known to result in a commitment to DNA synthesis (6) and protein kinase C activation (19). The inositol phosphate response was amplified by adding LiCl for 20 min prior to the termination of the incubation (4,41,42). Analysis by HPLC demonstrated that recombinant P. multocida toxin caused an increase in the cellular content of [3H]Ins(1,4,5)P3 (Fig. l), a product of PtdIns(4,5)P2 hydrolysis. The toxin also caused a dramatic increase in the cellular content of [3H]Ins(1,3,4)P3 and a lesser accumulation of [3H]Ins(1,3,4,5)P4. Recombinant P. multocida toxin also caused the accumulation of distinct InsP, and InsPl species, Ins(1,4)P2, Ins(4/5)P, and Ins(l/S)P. The accumulation of these products is clearly in accordance with known pathways of Ins(1,4,5)P3 metabolism (4,41,42). The small accumulation in the acid extracts of Ins(2)P probably arises from the acid hydrolysis of Ins(l:2)cyclic P formed as a side product of phospholipase C action, either directly or via Ins(l:2cyc,4)P2 and/or Ins(l:2cyc,4,5)P3.
It is well established that in Swiss 3T3 cells the mitogenic neuropeptide bombesin (43), acting via a specific plasma membrane receptor (44) and an unidentified G-protein (45)(46)(47)(48), activates phospholipase C, thereby leading to increased formation of inositol phosphates (39,49,50). The accumulation of inositol phosphates in recombinant P. multocida toxintreated cells was compared with that occurring upon activation of the bombesin receptor. The cells were treated with toxin for 4 h and with bombesin for 30 min. Since the toxin only causes an accumulation of inosito! phosphates after a lag period of several hours (6; and see below) and bombesin acts within seconds (39), the incubation times were chosen such that the cells would have been activated for approximately the same period of time. Under these conditions, the profile of inositol phosphates in the recombinant P. multocida toxinstimulated cells was very similar to that in cells treated in parallel with bombesin except for the presence of a small additional Inspa peak (coeluting with standard Ins(4,5)P2) in the recombinant P. multocida toxin-treated cells and unidentified material eluting at 16 min (Fig. 1). It is clear that recombinant P. multocida toxin causes an activation of inositol phospholipid breakdown which is very similar to the receptor-mediated breakdown induced by bombesin.
For the HPLC analysis the cells were incubated with LiCl to facilitate the accumulation of inositol phosphate metabolites in the cells. The recombinant P. multocida toxin-induced the elution of the following internal standards:   (Table I). Chromatographic analysis by FPLC of the individual inositol phosphates revealed that the enhancement by Li+ of the recombinant P. multocida toxin response occurred in the InsP, and InsPn fractions (Fig. 2) (51). Indeed, the HPLC analysis (Fig. 1)   in response to growth factors acting via Ins( 1,4,5)P3, including bombesin, is subsequently released into the extracellular medium, leading to a decrease in the cellular Ca2+ content (38,39,52). Recombinant P. multocida toxin mobilized Ca2+ in Swiss 3T3 cells. Thus, the 45Ca2+ content of 45Ca2+-equilibrated cells was decreased by recombinant P. multocida toxin (20 ng/ml, 4-h incubation) to 62 +-8% of the control value (mean k S.D., 14 independent cell preparations). This decrease occurred after a lag period and paralleled, in a timeand dose-dependent manner, the increases in inositol phosphates (Fig. 3). The kinetics of Can+ mobilization, InsP, accumulation, protein kinase C activation (19), and commitment to DNA synthesis (6) are very similar.
Recombinant P. multocida toxin also increased the accumulation of [3H]inositol phosphates and decreased the content of 45Ca2+ in tertiary passage mouse embryo fibroblasts (Fig. 3). In these cells, the toxin also activated protein kinase C (19) and stimulated DNA synthesis (6).
Our data suggest that recombinant P. multocida toxin and bombesin, by different routes, both cause a similar activation of PtdIns(4,5)P2 hydrolysis in Swiss 3T3 cells. Therefore, the effects of the toxin and the neuropeptide on Ca2+ mobilization were compared. The magnitude of the decrease in the 45Ca2+ content of cells pretreated with recombinant P. rnultocida toxin or bombesin is very similar (Table 11). Furthermore, the addition of 100 nM bombesin to recombinant P. multocida toxin-pretreated cultures only caused a small, but statistically insignificant, further decrease in the cellular 45Ca2+ content (Table 11). These data are consistent with an identity of the intracellular pools of Ca2+ mobilized by the two agents. Further evidence for a common source of mobilized Ca'+ comes from a study of the rapid and transient increase in [Ca2+]; elicited by bombesin (38,39,52). Pretreatment of the cells with recombinant P. multocida toxin under conditions The permeabilized cells were incubated for 10 min in the absence of other factors or in the presence of 10 nM bombesin, 10 p M GTPyS, 10 nM bombesin plus 10 p~ GTP-yS, 20 ng/ml recombinant P. multocida toxin (rPMT) or 100 ng/ml recombinant P. multocida toxin. The values shown are the means t S.D. of four incubations. It was established that permeabilization of the cells during the 4-min incubation with saponin caused the release of 75% of the cellular lactate dehydrogenase activity, as assayed by the standard spectrophotometric procedure of pyruvate-dependent oxidation of NADH.  (Fig. 4). Protein kinase C activation is known to attenuate Ca'+ mobilization by low concentrations of bombesin (38), and recombinant P. multocida toxin activates protein kinase C (19). However, under identical conditions, it was confirmed that treatment of the cells with the protein kinase C activator PBt2 did not decrease the Ca2+ mobilizing action of the higher concentrations of bombesin used in the experiments presented in Fig. 4 and Table 11. Attenuation by recombinant P. multocida toxin pretreatment was observed in the presence or absence of extracellular Ca2+, indicating that the source of the Ca2+ mobilized by bombesin was intracellular. Furthermore, the decrease in the Ca2+mobilizing action of bombesin caused by recombinant P. multocida toxin was dependent upon the time of exposure of the cells to the toxin, exhibiting a lag period (Fig. 4) similar to that observed for the enhancement of InsP, formation and 45Ca2+ mobilization (Fig. 3). The decrease of Ca2+ mobilization by bombesin in recombinant P. multocida toxin-treated cells is consistent with a common mechanism of Ca2+ mobilization by the two agents, i.e. both act via Ins (1,4,5)P3.
Evidence for the Cellular Entry and Processing of Recombi- nant P. multocida Toxin-Many toxins require entry into cells and processing for biological activity (53, 54). The lag period in the action of recombinant P. multocida toxin may reflect its cellular entry and possible processing and activation. Indeed, the lysosomotrophic agent methylamine selectively blocked the increase in InsP, and Ca2+ mobilization caused by recombinant P. multocida toxin (Fig. 5 ) . In contrast, this agent did not block the same responses to bombesin. An antiserum raised against native P. multocida toxin (Fig. 5 ) , which recognizes recombinant P. multocida toxin by immunoblotting (6), blocked the actions of recombinant P. multocida toxin but did not prevent the responses to bombesin (Fig.  5 ) . Addition of the antiserum 3 h after the toxin did not block the InsP, response (results not shown), suggesting that the toxin had entered the cells and was therefore unavailable to the antiserum.
In addition to methylamine, two other lysosomotrophic agents, ammonium chloride and chloroquine, also inhibited the stimulation of InsP, production by recombinant P. multocida toxin (Fig. 6). The selectivity of the inhibitory action of these agents is again demonstrated by the ability of bombesin to stimulate InsP, production in their presence (although there was a slight inhibitory effect of chloroquine).
Inhibition of recombinant P. multocida toxin-stimulated InsP, production by methylamine was time dependent (Fig.  6). Addition of the lysosomotroph with the toxin, or 30 min after it, completely inhibited the InsP, response. Addition a t later times, up to 2 h, caused progressively less inhibition after which time there was essentially no effect (Fig. 6). Similarly, in 4-h incubations with 20 ng/ml recombinant P. multocida toxin, the addition of methylamine 3 h after the addition of recombinant P. multocida toxin did not inhibit the decrease in 4sCaz+ (results not shown). The time dependence of the inhibitory action of methylamine is consistent with a requirement for the cellular entry of recombinant P. multocida toxin via endocytosis and involving endosomal/ lysosomal trafficking.
To explore further the role of endocytosis in the activation of recombinant P. multocida toxin, we investigated the effect of the toxin on inositol phosphate release in permeabilized cells. The direct addition of recombinant P. multocida toxin at 20 or 100 ng/ml to saponin-permeabilized 3T3 cells failed to stimulate the release of inositol phosphates (Table 111). Under the same conditions we verified that either bombesin or GTPyS alone stimulated inositol phosphate release. Addition of both agents together resulted in synergistic stimulation. Thus, the permeabilized cell preparation was operationally defined to contain the molecular components necessary to couple the bombesin receptor via a G-protein to phospholipase C activation. The lack of effect of recombinant P. multocida toxin in this system lends further support to the notion that recombinant P. mukocida toxin has to enter the cells via the physiological process of endosomal trafficking, resulting in release of active toxin into the cytosol to stimulate polyphosphoinositide breakdown.
In theory, the intracellular action of recombinant P. multocida toxin could be the result of the direct activation of intracellular processes controlling polyphosphoinositide breakdown. Alternatively, recombinant P. multocida toxin could act indirectly causing the synthesis and release of a factor into the culture medium which then stimulates inositol phosphate formation. The latter possibility is untenable since ( a ) cycloheximide (25 p~) did not prevent the recombinant P. multocida toxin-induced increases in [3H]In~Pt accumulation and 45Ca2+ mobilization (results not shown), indicating that do novo protein synthesis was not required for its actions; and ( 6 ) the incubation medium from cells treated for 4 h with 20 ng/ml recombinant P. multocida toxin failed to provoke a ['HH]InsP, increase in untreated cells (Table IV). This result also argues against the possibility of extracellular processing of the toxin into an active form.
If recombinant P. multocida toxin acts intracellularly, the stimulation of InsP, formation should persist after removal of extracellular toxin. As shown in Fig. 7, the InsP, increase induced by recombinant P. multocida toxin persisted after removal of extracellular toxin. In marked contrast, the InsP, response induced by bombesin was reversed by removal of extracellular ligand (Fig. 7). Collectively, the dose-dependent lag period, the methylamine block, the inactivity of the toxin in saponin-permeabilized cells, the persistence of toxin action after removal from activated cells, the lack of effect of toxin-conditioned medium argue that recombinant P. multocida toxin, in order to stimulate polyphosphoinositide breakdown, enters the cells by the action of endosomal/lysosomal traffic and thereby gains access to the cytosol in an activated state.

CONCLUSIONS
We have identified a toxin, P. multocida toxin, that enters cells both to enhance the formation of Ins(1,4,5)P3 and to mobilize intracellular Ca2+. The apparent requirement for cellular entry of the toxin to elicit InsP, increases, and Ca2+ mobilization is also a feature of both its mitogenic action (6) and its ability to activate protein kinase C (19). Furthermore, the similarities in the inositol phosphate species formed in response to recombinant P. multocida toxin and bombesin suggest that P. multocida toxin modifies cellular regulatory processes physiologically involved in polyphosphoinositide hydrolysis. Thus, we propose that P. multocida toxin stimulates the phospholipase C-mediated hydrolysis of PtdIns (4,5)P2,hence causing the release of Ins(1,4,5)Pa, Ca2+ mobilization, increase in diacylglycerol, translocation of protein kinase C, and the phosphorylation of the protein kinase C substrate 80 K. The molecular basis of the action of P. multocida toxin may provide new insights into the regulation of receptor-phospholipase C coupling. P. multocida toxin could also provide a novel tool with which to study cellular Ca2+ signaling under conditions free from constraints such as ligand-induced cellular desensitization.