Pasteurella multocida Toxin Activates the Inositol Triphosphate Signaling Pathway in Xenopus Oocytes via Gqα-coupled Phospholipase C-β1

Pasteurella multocida toxin (PMT) has been hypothesized to cause activation of a GTP-binding protein (G-protein)-coupled phosphatidylinositol-specific phospholipase C (PLC) in intact cells. We used voltage-clamped Xenopus oocytes to test for direct PMT-mediated stimulation of PLC by monitoring the endogenous Ca2+-dependent Cl− current. Injection of PMT induced an inward, two-component Cl− current, similar to that evoked by injection of IP3 through intracellular Ca2+ mobilization and Ca2+ influx through voltage-gated Ca2+ channels. These PMT-induced currents were blocked by specific inhibitors of Ca2+ and Cl− channels, removal of extracellular Ca2+, or chelation of intracellular Ca2+. Specific antibodies directed against an N-terminal, but not a C-terminal, peptide of PMT inhibited the toxin-induced currents, implicating that the N terminus of PMT is important for toxin activity. Injection with specific antibodies against PLCβ1, PLCβ2, PLCβ3, or PLCγ1 identified PLCβ1 as the primary mediator of the PMT-induced Cl− currents. Injection with guanosine 5′-O-(2-(thio)diphosphate), antibodies to the common GTP-binding region of G-protein α subunits, or antibodies to different regions of G-protein β subunits established the involvement of a G-protein α subunit in PMT-activation of PLCβ1. Injection with specific antibodies against the α-subunits of Gq/11, Gs/olf, Gi/o/t/z, or Gi-1/i-2/i-3 isoforms confirmed the involvement of Gq/11α. Preinjection of oocytes with pertussis toxin enhanced the PMT response. Overexpression of Gqα in oocytes could enhance the PMT response by 30-fold to more than 300-fold, whereas introduction of antisense Gqα cRNA reduced the response by 7-fold. The effects of various specific antibodies on the PMT response were reproduced in oocytes overexpressing Gqα.

Infections of Pasteurella multocida are associated with such severe diseases as pasteurellosis, dermonecrosis resulting from bite wounds, and the irreversible bone atrophy of progressive atrophic rhinitis (1). Purified P. multocida toxin (PMT) 1 alone is sufficient to induce experimentally all of the major symptoms of atrophic rhinitis in animals (1)(2)(3)(4)(5). PMT appears to bind to and enter mammalian cells via receptor-mediated endocytosis (6,7) and acts intracellularly to initiate DNA synthesis (7)(8)(9). Some of the events toward the eventual stimulation of DNA synthesis that occur upon exposure to PMT in cultured fibroblasts and osteoblasts are: enhanced hydrolysis of inositolphospholipids to increase the total intracellular inositol phosphates (9 -11); mobilization of intracellular Ca 2ϩ pools (9 -12); increased production of diacylglycerol (10 -12); decreased ADPribosylation of GRP78/BiP (13); and translocation of protein kinase C and increased protein phosphorylation (12). Recently, PMT has also been shown to induce tyrosine phosphorylation of p125 Fak and paxillin, as well as actin stress fiber formation and focal adhesion assembly (14).
The reported mitogenic response caused by PMT on intact cells has been hypothesized to be the result of activation of a cellular phosphatidylinositol-specific phospholipase C (PLC) (10,11), which catalyzes the hydrolysis of phosphatidylinositol 4,5-bisphosphate (PIP 2 ) to inositol 1,4,5-triphosphate (IP 3 ) and diacylglycerol. Accordingly, PMT-induced release of these second messengers was presumed to be responsible for initiation of subsequent signaling events, including stimulation of Ca 2ϩ mobilization and activation of protein kinase C, respectively. There are a large number of different ligands and receptors that are known to activate PLC, causing release of IP 3 and diacylglycerol from PIP 2 (15)(16)(17)(18)(19)(20)(21). Receptor regulation of phosphoinositide hydrolysis is generally considered to be mediated either through protein tyrosine phosphorylation of PLC␥ or G-protein activation of PLC ␤-isoforms (15, 16, 20 -22). At least two general pathways of G-protein-regulated PIP 2 hydrolysis can be distinguished by their sensitivity to ADP-ribosylation by pertussis toxin (PT). The ␤␥ subunits of PT-sensitive G o/i -proteins preferentially stimulate PLC␤3 Ͼ PLC␤2 Ͼ PLC␤1, whereas the ␣ subunits of the PT-insensitive G q family, including ␣ q and ␣ 11 , stimulate PLC␤1 Ն PLC␤3 Ͼ Ͼ PLC␤2 (16, 17, 19 -42). It has been reported that PMT-induced phosphoinositide hydrolysis could be blocked by the addition of GDP␤S to permeabilized cells, but PMT does not increase tyrosine phosphorylation of PLC␥ (10). PMT action may thus involve a G-protein-dependent PLC activity, such as PLC␤1, PLC␤2, or PLC␤3.
We used voltage-clamped Xenopus oocytes to demonstrate direct PMT-mediated stimulation of PLC activity by monitor-ing the endogenous Ca 2ϩ -dependent Cl Ϫ current evoked upon microinjection with PMT. To identify the intracellular targets involved in the PMT-induced IP 3 signaling pathway, we tested the effects of specific antibodies against G pan ␣, G q/11 ␣ (C-terminal and N-terminal), G i/o/t/z ␣, G i-1/i-2/i-3 ␣ isoforms, G s/olf ␣, G pan ␤ (C-terminal, internal, and N-terminal), PLC␤1, PLC␤2, PLC␤3, PLC␥1, an N-terminal peptide of PMT (toxA 28 -42 ), or a C-terminal peptide of PMT (toxA 1239 -1253 ) on the PMT-induced Cl Ϫ currents. We also examined the effects of PT on the PMT response. Our results established the direct involvement of the free, monomeric G q ␣ protein in PMT activation of PLC␤1. The specific role of G q ␣ was further confirmed by over-and underexpression of mouse G q ␣ in Xenopus oocytes.
Synthesis of cRNA from G q ␣ cDNA-Complementary DNA coding for mouse G q protein ␣ subunit in the pcDNAI cloning vector (5.4 kilobases) was obtained as a generous gift from Dr. Petra Schnabel. The plasmid containing the cDNA insert was linearized by digestion with the restriction enzyme ApaI (Life Technologies, Inc.). Using the Ampliscribe Transcription System (Epicentre Technologies, Inc.), sense cRNA was transcribed by the T7 promoter and antisense cRNA by the SP6 promoter, according to the manufacturer's procedure. In vitro transcriptions of the cRNA were performed in the presence of methylated cap (m 7 G[5Ј]ppp[5Ј]G) and catalyzed by T7 or SP6 DNA-dependent RNA polymerase, respectively, at 37°C for 2 h. In vitro transcribed cRNA was dissolved in RNase-free water at a final concentration of 1 ng/nl.
Oocyte Preparation-Adult female Xenopus laevis frogs (Xenopus I, Michigan) were anesthetized by immersion in a 0.15% tricaine methanesulfonate (Ayerst) solution for 30 min. A small incision was made on one side of the abdomen to remove several ovarian lobes. The lobes were gently torn apart and immersed in a Ca 2ϩ -free OR-2 solution (100 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , and 5 mM Hepes-Tris, pH 7.5). Oocytes were defolliculated by incubation with 2 mg/ml collagenase (Sigma, type 1A) at room temperature (22-24°C) for 2-3 h. The oocytes were then washed five times with OR-2 solution and five times with a modified Barth's solution (88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO 3 , 0.3 mM Ca(NO 3 ) 2 , 0.4 mM CaCl 2 , 0.8 mM MgSO 4 , 15 mM Tris-HCl, pH 7.6, containing 100 g/ml penicillin and 100 g/ml streptomycin). Stage 5-6 oocytes were selected and stored at 18°C in modified Barth's solution.
Two-microelectrode Voltage Clamp-Two microelectrodes, made by a horizontal puller (PD-5, Narishige) and filled with 3 M KCl to give a resistance of 1.5-2.0 megaohms, were used for voltage clamping. Voltage clamp experiments were performed in a continuously perfused bath (10 ml/min) at room temperature (22°C). The bath was connected through an Ag-AgCl-Agar-3 M-KCl bridge to the voltage-recording amplifier (Axoclamp 2A, Axon Instruments). The data were filtered with a four-pole Bessel filter at 500 Hz. Voltage pulse protocols and data acquisition were performed on a 486 IBM computer with pCLAMP software (Axon Instruments), and graphics were obtained using Origin software (Microcal) with a pCLAMP module. Membrane currents were measured in normal Ringer's solution (96 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , 11.8 mM CaCl 2 , and 5 mM Hepes-NaOH, pH 7.4), or where indicated, in Ca 2ϩ -free Ringer's solution (96 mM NaCl, 2 mM KCl, 12 mM MgCl 2 , and 5 mM Hepes-NaOH, pH 7.4).
Oocyte Microinjections-Microinjections of toxins, antibodies, and other reagents were performed under voltage-clamp conditions at Ϫ80 mV holding potential using a pulse-controlled picoliter injector (Dagan model PMI200). Toxins and other reagents were prepared in 50 mM Tris-HCl, pH 7.5, or 50 mM potassium phosphate buffer, pH 7.5, prior to injection. In all oocytes tested, if a stable membrane potential could be achieved, PMT always elicited a response. However, rarely, an entire group of oocytes from a particular donor gave only a very weak response to PMT, upon which the entire group of oocytes was discarded, and the responses were not included in the data analysis; oocytes from that donor were not used again. In all experiments reported here, injection of PMT into control, untreated oocytes gave detectable responses. In a number of cases, a group of oocytes from a particular donor elicited robust responses to PMT, of which several oocytes would give extremely large responses (exceeding the detection limits of the instrument), and the responses from these oocytes were also not included in the data analysis. The number of oocytes used for data analysis is given in parentheses in the figure legends, and the total number of oocytes tested, including those giving overwhelming (but not those giving weak) responses, is indicated in brackets. All results are expressed as the mean Ϯ S.D. or the mean Ϯ S.E. (as indicated in the figure legends) of the responses assayed in oocytes from at least two different donors (except where indicated), with N denoting the number of groups tested and n denoting the number of oocytes tested in each group. To evaluate the statistical significance of the results, p values were determined for each group using the two-ended t test with unequal variance. Oocytes that were not injected with cRNA are referred to throughout this communication as normal oocytes.
G q ␣ Protein Expression-In the experiments over-or underexpressing G q ␣, the oocytes were injected with in vitro transcribed cRNA (50 ng) by positive displacement using a 10-l micropipetter 2 days prior to the electrophysiological experiments. Because of the large response observed in the oocytes overexpressing G q ␣, the amount of PMT was decreased to 0.1 ng/oocyte. Even at this dose, it was frequently observed that the peak inward current exceeded the instrument's recording range (similar to that shown in Fig. 2B, lower trace). Consequently, only those groups of oocytes showing a moderate increase (up to ϳ800 nA, similar to that shown in Fig. 5A, lower trace) in the peak current upon injecting the reduced PMT dose were arbitrarily selected for the antibody studies. The total number of oocytes tested, including those giving overwhelming responses (all oocytes gave a response), is denoted in brackets in the figure legends.
Antibody Injections-Prior to injection, the antibodies were dialyzed against 50 mM Tris-HCl, pH 7.5, or 50 mM potassium phosphate buffer, pH 7.5, and subsequently diluted to the original concentrations as supplied by the commercial source. The optimal preincubation period for oocytes with each antibody prior to toxin injection was determined to be 2-4 h (data not shown). Oocytes were injected with 50 nl of undiluted antibody 3 h prior to injection with PMT. Anti-toxA 28 -42 , diluted to a concentration that could neutralize the PMT effect in normal oocytes, as determined by titration (Fig. 1B), were used for co-injection with PMT. Anti-toxA 1239 -1253 was used at the same concentration as anti-toxA 28 -42 .
SDS-Polyacrylamide Gel Electrophoresis and Western Analysis-Each individual oocyte was solubilized in lysis buffer, containing 50 mM Hepes, pH 7.4, 1% Triton X-100, 0.5% Nonidet P-40, 100 mM NaCl, 5 mM MgCl 2 , 10 mM benzamidine, 10 g/ml aprotinin, 10 g/ml leupeptin, and 0.2 mM phenylmethylsulfonyl fluoride, followed by the addition of 2 ϫ SDS-polyacrylamide gel electrophoresis sample buffer. The mixture was heated at 95°C for 10 min, and the entire contents of each oocyte were separated by 10% SDS-polyacrylamide gel electrophoresis and transferred onto nitrocellulose membrane. The membrane was immu-noblotted first with rabbit polyclonal antibodies to G q ␣, revealing a major band due to G q ␣ (data not shown), then with rabbit polyclonal antibodies to RhoA (Fig. 5C), revealing an additional band due to RhoA and an unidentified higher molecular weight band at M r ϳ60,000. The blots were developed using secondary antibodies conjugated to alkaline phosphatase.

RESULTS
PMT Activation of the Ca 2ϩ -dependent Cl Ϫ Current in Xenopus Oocytes-Xenopus oocytes clamped at negative holding potentials (Ϫ80 mV) produced an inward current when microinjected with the 146.3-kDa protein toxin from P. multocida (Fig. 1A). Examination of the activation of the current showed that it was characteristic of the IP 3 -mediated Ca 2ϩ -dependent Cl Ϫ conductance, exhibiting two components: an initial faster peak current I 1 in response to mobilization of intracellular Ca 2ϩ pools; and a second, slower and larger peak current I 2 caused by Ca 2ϩ influx through voltage-gated Ca 2ϩ channels on the plasma membrane. Pre-incubation of PMT with polyclonal antibodies against a 15-amino acid synthetic peptide from the N-terminal region of PMT (anti-toxA 28 -42 ) prevented the response (Fig. 1B), as did heat inactivation of the toxin (Fig. 1A), demonstrating the specificity of the PMT-induced response and the importance of the N terminus of PMT in its action. Similar preincubation of PMT with polyclonal antibodies against a 15-amino acid peptide from the C-terminal region of PMT (anti-toxA 1239 -1253 ) had no effect on the PMT response (statistical analysis included in Fig. 3A).
To confirm that the PMT-induced Cl Ϫ currents were indeed Ca 2ϩ -dependent, the effects of removing extracellular Ca 2ϩ and applying specific blockers of Ca 2ϩ and Cl Ϫ channels on the PMT-induced currents were examined (Fig. 1C). For the IP 3 response, it has been shown that removal of extracellular Ca 2ϩ or blockage of the voltage-gated Ca 2ϩ channel by Cd 2ϩ will abolish I 2 , but not I 1 , whereas induction of both I 1 and I 2 can be blocked by intracellular chelation of Ca 2ϩ with EGTA or by anion transport inhibitors of Cl Ϫ channels, such as anthracene-9-carboxylic acid (51). In agreement with these observations, injection of PMT into oocytes bathed in Ca 2ϩ -free solution induced only I 1 , whereas I 2 was not observed. Likewise, when the Ca 2ϩ -channel blocker Cd 2ϩ was present at 1 mM concentration in the medium, I 2 was diminished. Both I 1 and I 2 were inhibited if PMT was co-injected with EGTA or if anthracene-9-carboxylic acid was perfused in the bath solution.
Dose-Response and Effect of Multiple Doses of PMT on the Ca 2ϩ -dependent Cl Ϫ Currents in Xenopus Oocytes-The overall effect of microinjecting PMT on the voltage-clamped oocytes was direct and almost immediate, occurring within 20 s from microinjection. An EC 50 value of 0.28 ng/oocyte for PMT (for the Sigma sample) in normal oocytes was determined from the dose-response curve in Fig. 2A. Contrary to what was observed for multiple IP 3 injections (Fig. 2B, upper trace), after the first injection with PMT in normal oocytes, additional injection with PMT gave little or no further response (Fig. 2B, middle trace), suggesting that the action of PMT is not readily reversible. For oocytes unable to elicit a response to a second dose of PMT, additional injection of IP 3 in the same oocyte was still able to evoke both I 1 and I 2 . In G q ␣-overexpressing oocytes, even at FIG. 1. PMT-induced Ca 2؉ -dependent Cl ؊ currents in Xenopus oocytes. A, voltage-clamped oocytes were injected with 10 ng of IP 3 (trace shown is representative of 11 oocytes) and 1 ng of PMT (representative of 74). I 1 , the first peak current from the mobilization of intracellular Ca 2ϩ ; I 2 , the second peak current from Ca 2ϩ influx through voltage-gated Ca 2ϩ channels on the plasma membrane. No response occurred if PMT was heat-denatured for 10 min at 95°C prior to injection (representative of 4). B, PMT (1.0 g) was neutralized on ice for 30 min with increasing amounts of anti-toxA 28 -42 in a total volume of 20 l prior to injection of 20 nl of the resulting mixture. The relative ratios of the PMT and antibody solutions were: 1:2 (representative of 9); 1:1 (representative of 11); 1:0.5 (representative of 6); and 1:0 (representative of 18). Injection of a neutralizing amount of anti-toxA 28 -42 alone failed to elicit a response (representative of 2), whereas co-injection of a mixture of anti-toxA 28 -42 and 10 ng of IP 3 evoked the characteristic Ca 2ϩ -dependent Cl Ϫ currents (representative of 2). C, both I 1 and I 2 were blocked if the Cl Ϫ channel blocker anthracene-9-carboxylic acid 20-fold less PMT doses, the initial injection elicited an overwhelming response (Fig. 2B, lower trace), more than 15 times that observed for normal oocytes (i.e. Ͼ300-fold more potent). After the first injection, a second dose of PMT elicited a greatly diminished response (Fig. 2B, lower trace inset). Despite numerous attempts, it was not technically possible to perform a subsequent IP 3 Fig. 3A.
Effects of Specific Antibodies to Different G-protein ␣ and ␤ Subunits on the PMT-induced Response in Normal Oocytes-We investigated the ability of specific antibodies against different G-protein subunits, G pan ␣, G s/olf ␣, G i/o/t/z ␣, G i-1 ␣, G i-1/i-2␣, G i-3 ␣, G q/11 ␣, or G pan ␤ to block the PMT-induced currents. Specific antibodies against the common GTP-binding region of most G-protein ␣ subunits (G pan ␣) greatly diminished the PMT-mediated response (p Ͻ 0.07). Specific antibodies to the C-terminal regions of G s/olf ␣, G i/o/t/z ␣, or to the N-terminal region of G q ␣ had no significant effect (p Ͼ 0.4) on the PMTinduced Cl Ϫ currents. Antibodies to the C-terminal regions of G i-1 ␣, G i-1/i-2 ␣, or G i-3 ␣ only slightly increased the PMT response (p Ͻ 0.3). On the other hand, antibodies directed against the unique C-terminal region of G q/11 ␣ caused a pronounced reduction (p Ͻ 0.006) in the PMT response, strongly supporting a direct role for G q ␣-coupled PLC␤1 in PMT action.
To determine whether release of the ␤␥ subunits from the ␣ subunits of the G-proteins might account for activation of the PLC activity, antibodies to the common N-terminal, internal, and C-terminal regions of the 1-4 isoforms of G␤ (anti-G pan ␤) were tested. Rather than blocking the PMT-induced response, results revealed a marked 4-fold increase in the PMT response (p Ͻ 0.0003) for the C-terminally directed antibodies and a 2-fold increase (p Ͻ 0.0001) for the internally directed antibodies. Antibodies against the N terminus of G pan ␤ had little effect on the PMT response (p Ͼ 0.1). The results are summarized in Fig. 3B.
Effects of Over-and Underexpression of G q ␣ on the PMTinduced Response-It was determined that injection of 1.0 ng of PMT into normal oocytes elicited a response comparable to that evoked by 10 ng of IP 3 . Oocytes preinjected with sense mouse G q ␣ cRNA two days prior to injection of PMT resulted in a marked increase in the PMT-induced Cl Ϫ current (Fig. 4, A and  B). Among those sense G q ␣ cRNA-treated oocytes showing moderate responses (see "Experimental Procedures"), there was a ϳ3-fold increase in the PMT-induced response with 10-fold less toxin (p Ͻ 0.001). The PMT response could be blocked by preinjection of the nonhydrolyzable GDP analog, GDP␤S (final intracellular concentration was estimated to be 500 M), even at a dose of 1.0 ng/oocyte. Preinjection with antisense G q ␣ cRNA into Xenopus oocytes reduced the response mediated by PMT (p Ͻ 0.01) by ϳ7-fold, compared to that observed for normal oocytes (Figs 4, A and B). Western blot analysis of total cell lysates confirmed that the G q ␣ protein was indeed being overexpressed in the oocytes showing overwhelming response to PMT (Fig. 4C) but was not as evident in oocytes showing only a moderate response (data not shown). In the G q ␣-underexpressed oocytes, only a slight reduction in G q ␣ protein level was noticed. The effects of various specific antibodies on the PMT-induced response were also investigated in sense G q ␣ cRNA-treated oocytes. The results, summarized in Fig. 5, were consistent with those found for normal oocytes

FIG. 2. The effect of multiple injections of PMT on the Ca 2؉dependent Clcurrents in Xenopus oocytes.
A, dose-dependence of the PMT-induced response in normal oocytes, using PMT concentrations of 0.07 ng/oocyte (n total ϭ 12), 0.14 ng/oocyte (n total ϭ 6), 0.28 ng/oocyte (n total ϭ 15), 0.42 ng/oocyte (n total ϭ 9), and 0.56 ng/oocyte (n total ϭ 9). The mean of the peak inward Cl Ϫ current (nA) induced by injection with PMT (from Sigma) is shown for each data point; bars, S.E. An EC 50 value of 0.28 ng/oocyte was determined from the data using Sigma Plot. B, voltage-clamped oocytes were treated with multiple injections of 10 ng of IP 3 (upper, representative of 10). Voltageclamped oocytes were injected with 0.5 ng of PMT, followed by a second injection with 0.5 ng of PMT and then an injection with 10 ng of IP 3 (middle, representative of 5); the gap between the two traces indicates a time lapse of ϳ10 min. In oocytes overexpressing G q ␣, an additional injection of PMT (0.05 ng/injection) elicited a greatly diminished response after the initial injection of PMT (lower, representative of 16). This oocyte is representative of oocytes eliciting an overwhelming response to PMT (Ͼ3000 nA); the inset shows an amplification of the second response (ϳ30 nA). (Fig. 3).
Effect of Pertussis Toxin on the PMT-induced Response-Normal oocytes were injected with PT (1 ng/oocyte) at Ϫ80 mV to give a PT-induced conductance (Fig. 6A). When a stable base line was recovered (after ϳ2 h), this was followed by injection of PMT (0.5 ng/oocyte) at various time intervals. After PT injection, a progressively increased PMT response was observed (Fig. 6B), which was enhanced more than 20-fold after 3.5 h (Fig. 6A), compared to that of control oocytes not preinjected with PT (p Ͻ 0.0001). DISCUSSION Transient elevation of free Ca 2ϩ concentration in the cytosol is one of the first events observed in various cell types on stimulation with hormones and growth factors (13-19, 44 -49). This elevation is due both to Ca 2ϩ release from intracellular stores and to Ca 2ϩ influx through Ca 2ϩ channels on the plasma membrane (15, 18, 19, 44 -49). It has been reported that treatment of cultured cells with PMT causes an increase in intracellular inositol phosphate and Ca 2ϩ levels within 3-4 h after exposure to toxin (9 -12). The lag period, before detectable intracellular responses are observed, has been attributed to the requirement for PMT to first bind to cell surface receptors and be internalized and presumably processed through endocytic vesicles (6,7). With the objective of deciphering early events in the action of the toxin, we bypassed this lengthy internalization process by directly injecting toxin into Xenopus oocytes. Upon microinjection of PMT into oocytes, we observed an immediate IP 3 -like response (Fig. 1A) within 20 s. Control experiments confirmed that this IP 3 -like response was indeed a Ca 2ϩ -dependent Cl Ϫ current (Fig. 1C).
The observed response was PMT-dependent, as demonstrated by the lack of response from heat-denatured toxin (Fig.  1A) and by the ability of specific antibodies against an Nterminal peptide of PMT, anti-toxA 28 -42 , to block the activity (Fig. 1B). Unlike anti-toxA 28 -42 , specific antibodies to a Cterminal peptide of PMT, anti-toxA 1239 -1253 , did not block the activity (Fig. 3A), strongly implicating that the N terminus of PMT is crucial for its activity. This PMT-induced response was dose-dependent ( Fig. 2A). The action of PMT was not readily reversible, for although repeated IP 3 injections produced repeated responses, a second injection of PMT did not (Fig. 2B). PMT did not impair the IP 3 response, because oocytes first injected with PMT and showing no response to a second dose of PMT were still able to show an IP 3 -dependent response (Fig.  2B, middle trace); therefore, PMT action must occur upstream to IP 3 release.
oocytes from the same donor. For IP 3 , N ϭ 4, n ϭ 2-3 (n total ϭ 11 [20]); normal oocytes, N ϭ 14, n ϭ 3-8 (n total ϭ 74 [104]); oocytes injected with antisense G q ␣ cRNA, N ϭ 4, n ϭ 3-4 (n total ϭ 14 [14]) (p Ͻ 0.01); oocytes injected with sense G q ␣ cRNA, N ϭ 11, n ϭ 2-6 (n total ϭ 37 [69]) (p Ͻ 0.001); oocytes injected with sense G q ␣ cRNA and GDP␤S, N ϭ 3, n ϭ 3 (n total ϭ 9 [9]) (p Ͻ 0.01). C, Western analysis of over-and underexpression of mouse G q ␣ protein in Xenopus oocytes. The entire content of each oocyte lysate was analyzed by 10% SDS-polyacrylamide gel electrophoresis and immunoblotted with anti-G q ␣ and anti-RhoA antibodies, as described under "Experimental Procedures." The results shown are representative of five independent experiments. Lane 1, a single oocyte 48 h after injection with sense G q ␣ cRNA and showing overwhelming response to PMT similar to that shown in Fig. 2B, lower trace; Lane 2, a single oocyte 48 h after injection with antisense G q ␣ cRNA and having diminished response similar to that shown in Fig. 4A, middle trace; Lane 3, a single normal oocyte without cRNA treatment and having a response similar to that in Fig. 4A, upper trace. cating a G␣-dependent pathway. Anti-G pan ␤ did not block, but instead enhanced, the response by as much as 4-fold (see discussion below).
To determine which of the G␣ families is involved in activating the PLC␤ isoforms leading to IP 3 release, the effects of specific antibodies against G s/olf ␣, G i/o/t/z ␣, G i-1 ␣, G i-1/i-2 ␣, G i-3 ␣, or G q/11 ␣ on PMT action were examined. Anti-G q/11 ␣ completely blocked the PMT-induced response, similar to that observed for anti-G pan ␣, anti-PLC␤1, or anti-toxA 28 -42 . Anti-G s/olf ␣, anti-G i-1 ␣, anti-G i-1/i-2 ␣, anti-G i-3 ␣, and anti-G i/o/t/z ␣ did not block the PMT-induced response, consistent with the known non-PLC effector targets of the respective G␣ proteins (52). Based on these results using specific antibodies to identify the key mediators of PMT action, the observed PMT-induced Cl Ϫ current involves a signal transduction pathway composed of G q/11 ␣-dependent activation of PLC␤1, subsequent hydrolysis of PIP 2 to release IP 3 , Ca 2ϩ mobilization, and eventual activation of the Ca 2ϩ -dependent Cl Ϫ channels.
The N-terminal half of the cytotoxic necrotizing factors type 1 and 2 (CNF1 and CNF2) from enteropathogenic Escherichia coli show 24 -27% homology to the first ϳ600 amino acids of PMT (53,54). The Rho family of Ras-related, small GTP-binding proteins, involved in regulating the assembly of focal adhesion and stress fibers in eukaryotic cells, have recently been implicated as possible intracellular targets of the cytotoxic necrotizing factors (54). Ca 2ϩ and Rho signaling pathways cooperate to regulate reorganization of actin filaments (55,56), and Rho regulation of cytoskeletal function appears to be activated by PLC via a PKC/diacylglycerol/phorbol ester-sensitive factor (56). We have observed no inhibitory effect on the PMTinduced Cl Ϫ currents using antibodies specific for the GTPbinding region unique to RhoA or RhoB, as well as antibodies to the conserved GTP-binding region of Ras-related proteins. 2 Our results do not preclude Rho-or Ras-related proteins as additional PMT targets or as potential signaling proteins important in the mitogenic effect stimulated by PMT (14). However, these proteins do not appear to be required for the PMTtriggered IP 3 release in Xenopus oocytes. The cellular mechanisms that lead to PMT-induced mitogenesis remain unclear, and there is as yet no direct evidence linking the effect of PMT on the G q -coupled PLC pathway to cell proliferation.
The cloned amino acid sequence of the Xenopus G q ␣ shares 96% identity to that from mouse (43,57) and retains all of the characteristics that distinguish G q ␣ from other G-protein ␣ subunits (43,57). To confirm that G q ␣ is involved in PMT action, we overexpressed the mouse G q ␣ subunit in Xenopus oocytes. Overexpressing G q ␣ in oocytes could increase the PMTdependent Cl Ϫ current 30-to more than 300-fold (Fig. 2B), whereas oocytes treated with antisense G q ␣ cRNA showed a 7-fold decreased response (Fig. 4, A and B). Western blot analysis confirmed an enhancement of G q ␣ production in the G q ␣ cRNA-treated oocytes (Fig. 4C). GDP␤S, a known inhibitor of G q ␣and other G␣-mediated signaling pathways (10,17,34,58), blocked the PMT-induced response in G q ␣-overexpressing oocytes, even at much higher PMT doses (Fig. 4B). As observed for normal oocytes, a second dose of PMT elicited a dramatically decreased response in oocytes overexpressing G q ␣ (Fig.  2B, lower trace). These findings further support the direct involvement of G q ␣ in PMT-mediated signaling pathways.
Furthermore, the effects of specific antibodies, anti-PLC␤1, anti-G pan ␣, anti-G q/11 ␣, or anti-toxA 28 -42 on the PMT-induced Cl Ϫ currents in normal oocytes were reproduced in G q ␣-overexpressing oocytes (Fig. 5). In the G q ␣-overexpressing oocytes, anti-PLC␤2 did not enhance the PMT response, as observed in normal oocytes (compare Fig. 5 with Fig. 3A). On the other hand, the 25% decrease in PMT response observed in normal oocytes for anti-PLC␤3 was reproduced in the G q ␣-overexpressing oocytes (Figs. 3A and 5). This effect of anti-PLC␤3 suggested the possibility of a minor role for PLC␤3 in a partial G q ␣or G␤␥-mediated pathway (40,42). However, anti-G pan ␤ (C-terminal) was not able to block PMT action, and in fact, a 2.5-fold enhancement (4-fold for normal oocytes) of the response was observed (Figs. 3B and 5). Although it is conceivable that anti-G pan ␤ may not be effective in blocking the activation of PLC␤1 or PLC␤3 mediated by the G␤␥ subunits, the anti-G pan ␤ used in this study was directed against the Cterminal 20 amino acids of G␤1-4. This region has been shown to be important for G␣ association with G␤␥ (52, 59 -62), as well as dimerization of G␤␥, which is important for G␤␥-dependent activation of PLC␤2 and PLC␤3 (23-27, 30, 42). On the other hand, the enhancement of the PMT response due to anti-G pan ␤ (C-terminal) could be accounted for by antibody sequestration of the G␤␥ subunits, causing dissociation of G␣ from the G␣␤␥ heterotrimer. This is consistent with the finding in normal oocytes that antibodies to an internal region of G pan ␤ also enhanced the PMT response, whereas antibodies to the N terminus of G pan ␤ had little effect on the response (Fig. 3B). The preferred substrate for PT is the heterotrimeric G i/o/t ␣␤␥ complex (63), and ADP-ribosylation of the GDP-bound ␣ subunit 2 B. A. Wilson and X. Zhu, unpublished observations. FIG. 6. The effect of PT on the PMT-induced Cl ؊ currents in normal Xenopus oocytes. A, normal oocytes were first injected with PT (1 ng) at Ϫ80 mV to give a PT-induced conductance, which upon recovery of a stable membrane potential after ϳ2 h, was followed by injection of PMT (0.5 ng) at 3.5 h to evoke an enhanced PMT response (representative of 4); the time shown indicates the time lapsed between the end of the first recording and the beginning of the second. B, time course of the effect of PT on the PMT response. PMT was injected into normal oocytes at various time intervals after PT injection, as in A. The data are shown as the mean for each experiment using control oocytes not preinjected with PT as reference (N ϭ 1); bars, S.E. For control oocytes, n total ϭ 6; 2 h, n total ϭ 11 (p Ͼ 0.6); 2.5 h, n total ϭ 8 (p Ͻ 0.01); 3 h, n total ϭ 14 (p Ͻ 0.0001); 3.5 h, n total ϭ 10 (p Ͻ 0.0001). locks the complex in its inactive heterotrimeric form (52,64). To test if sequestration of the G␤␥ subunits to release more G q ␣ might enhance the PMT response, we examined the time-dependent effect of PT on the PMT response. Preinjection of PT enhanced the subsequent PMT response in a time-dependent manner, with a more than 20-fold increased PMT response at 3.5 h after PT injection (Fig. 6). These combined results suggest that the direct target of PMT action is the free, monomeric form of G q ␣, which agrees with the enhancement of the response observed in oocytes overexpressing G q ␣ (Fig. 4, A and B).
Although there was a greater than 300-fold enhancement in the response to a 20-fold lower dose of PMT due to overexpression of G q ␣ (Fig. 2B, lower trace), the Western blot of such oocyte lysates showed at most a 2-3-fold increase in total G q ␣ protein expression (Fig. 4C). For those oocytes showing only moderate increase in response (3-fold enhancement with 10fold lower PMT dose), little difference in total G q ␣ expression was observed (data not shown). Likewise, the antisense G q ␣ cRNA-treated oocytes showed a 7-fold decrease in response (Fig. 4A, middle trace), whereas the Western blot indicated only a slight reduction in total G q ␣ protein levels (Fig. 4C). These findings are consistent with the hypothesis that the magnitude of the PMT-induced response is dependent on the level of free, monomeric G q ␣ protein, instead of the total amount of G q ␣ protein present in the oocytes.
In light of our findings and the above discussion, we propose a possible mechanism for PMT action in Xenopus oocytes, in which PMT acts on free G q ␣, possibly the GDP-bound form, and converts G q ␣ into an active form, which stimulates PLC␤1. The activated PLC␤1 causes PIP 2 hydrolysis, leading to IP 3 release and eventual Ca 2ϩ mobilization that results in the Ca 2ϩ -dependent Cl Ϫ current. The PMT-induced, G q ␣-mediated PLC␤1 activation appears to be transient, and the presumably modified G q ␣ involved in this activation is not readily available for further PMT action.