Regulation of Ste7 ubiquitination by Ste11 phosphorylation and the Skp1-Cullin-F-box complex.

Ste7 is a mitogen-activated protein kinase kinase that mediates pheromone signaling in Saccharomyces cerevisiae. We showed previously that Ste7 is ubiquitinated upon prolonged stimulation by pheromone and that accumulation of ubiquitinated Ste7 results in enhanced transcription and cell division arrest responses (Wang, Y., and Dohlman, H. G. (2002) J. Biol. Chem. 277, 15766-15772). We now report that ubiquitination of Ste7 requires Ste11 kinase and Skp1/Cullin/F-box (SCF) ubiquitin-conjugating activities. Ste7 is not ubiquitinated in Ste11-deficient cells or when the Ste11 phosphorylation sites have been mutated. Ste7 ubiquitination and degradation (but not phosphorylation) is specifically blocked in mutants defective for the E2 ubiquitin-conjugating enzyme Cdc34 or the cullin homologue Cdc53. Both are components of the SCF complex that ubiquitinates proteins during the G1-S transition of the cell cycle. Our findings suggest that SCF promotes the ubiquitination and degradation of Ste7, thereby favoring the resumption of cell division cycling after pheromone-induced growth arrest.

Ste7 is a mitogen-activated protein kinase kinase that mediates pheromone signaling in Saccharomyces cerevisiae. We showed previously that Ste7 is ubiquitinated upon prolonged stimulation by pheromone and that accumulation of ubiquitinated Ste7 results in enhanced transcription and cell division arrest responses ( . We now report that ubiquitination of Ste7 requires Ste11 kinase and Skp1/Cullin/F-box (SCF) ubiquitin-conjugating activities. Ste7 is not ubiquitinated in Ste11-deficient cells or when the Ste11 phosphorylation sites have been mutated. Ste7 ubiquitination and degradation (but not phosphorylation) is specifically blocked in mutants defective for the E2 ubiquitin-conjugating enzyme Cdc34 or the cullin homologue Cdc53. Both are components of the SCF complex that ubiquitinates proteins during the G 1 -S transition of the cell cycle. Our findings suggest that SCF promotes the ubiquitination and degradation of Ste7, thereby favoring the resumption of cell division cycling after pheromone-induced growth arrest.
Many cell regulators exert their action by binding to cell surface receptors coupled to G proteins. In mammals, receptors of this type can detect hormones, neurotransmitters, odors, taste, and light. In yeast, G protein-coupled receptors respond to mating pheromones that trigger fusion of a and ␣ haploid cells (1). Upon pheromone stimulation of its receptor (Ste2 or Ste3), the G protein ␣ subunit (Gpa1) binds to GTP and dissociates from the G protein ␤␥ subunits (Ste4/Ste18). The G␤␥ dimer can then propagate the mating signal through activation of effector proteins, including a protein kinase (Ste20), a kinase scaffolding protein (Ste5), and the Cdc42 GDP-GTP exchange factor (Cdc24). These effectors go on to activate a mitogenactivated protein kinase (MAPK) 1 cascade comprised of a MAPK kinase kinase (Ste11), which phosphorylates and activates a dual specificity MAPK kinase (Ste7), which in turn phosphorylates and activates two related MAPKs (Fus3 and Kss1). MAPK substrates include a transcription factor (Ste12) and associated regulators (Dig1 and Dig2), phosphorylation of which lead to enhanced expression of specific genes involved in mating. Another MAPK substrate is the cell cycle regulator Far1, which promotes cell division arrest in late G 1 phase where mating occurs (1).
It has long been recognized that many signaling proteins are regulated by phosphorylation. A small but growing number of signaling proteins appear to be ubiquitinated as well. Ubiquitination typically requires three distinct enzymatic activities: a ubiquitin-activating enzyme (E1), a ubiquitin-conjugating enzyme (E2), and a ubiquitin-ligase (E3). Several classes of E3 enzymes have been identified, each having a distinct catalytic mechanism. Best known are those containing a HECT or RING domain, whereas the recently described U-box protein family represents a third class of E3 enzymes (2). E3 enzymes catalyze the formation of an isopeptide bond between the C-terminal carboxyl group of ubiquitin and Lys side chains of the target protein (3). Conjugated ubiquitin is itself ubiquitinated, resulting in the formation of poly-ubiquitin chains. The polyubiquitinated protein is usually then captured and degraded by a large compartmentalized protease particle, the 26 S proteasome (4).
In yeast several components of the pheromone-signaling cascade are ubiquitinated, including the receptor (5-7), the G␣ subunit (8,9), the MAPK kinase Ste7 (10), and most likely the MAPK kinase kinase Ste11 (11). In the case of cell surface receptors, ubiquitination functions as a signal for ligand-induced endocytosis. This situation is unusual, however, in that the receptors are largely mono-ubiquitinated instead of polyubiquitinated and delivered to the vacuole instead of the proteasome (5-7). Ubiquitination of Ste7 is also dependent on prolonged pheromone stimulation, but in this case the protein is degraded by the proteasome (10). The mechanism of pheromone-dependent ubiquitination and the enzymes responsible for this modification have not been established. Here we show that ubiquitination and degradation of Ste7 requires Ste11 phosphorylation and Skp1-Cullin-F-box (SCF) ubiquitin-conjugating activities.

EXPERIMENTAL PROCEDURES
Strains and Plasmids-Standard methods for the growth, maintenance, and transformation of yeast and bacteria and for the manipulation of DNA were used throughout (12). The yeast Saccharomyces cerevisiae strains used in this study are listed in Table I. The UBP3::KanMX module in a BY4741-derived mutant lacking UBP3 (Research Genetics, Huntsville, AL) (RG-ubp3; Table I) was PCR-amplified and transformed into strains JTY2556 (ste7::ADE2) and ETB1 (ste5::LYS2) to generate YHD1001 (ste7⌬ ubp3⌬) and YHD1002 (ste5⌬ ubp3⌬), respectively. Replacement of the UPB3 locus by KanMX was verified by PCR. The cells were grown at 30°C unless otherwise indicated.
Ste7 Ubiquitination and Degradation Assays-Ubiquitination of Ste7 results in the formation of a high molecular weight species that is recognized by both anti-Ste7 and anti-ubiquitin antibodies following immunoprecipitation of Ste7. Detection of ubi-Ste7 requires that it be stabilized, either by using cim3-1 mutants, which are deficient in 26 S proteasome activity, a ubp3⌬ gene deletion mutant, or by expressing a catalytically inactive dominant negative form of Ubp3 (Ubp3 C469S ). This is analogous to using specific phosphatase inhibitors to preserve transient increases in protein phosphorylation. To monitor the loss of Ste7 over time, mid-log cell cultures were treated with 3 M ␣-factor for 60 min, followed by cycloheximide (final concentrations, 10 g/ml in 0.1% ethanol) for up to 120 min. Growth was stopped by the addition of 10 mM NaN 3 and transfer to an ice bath. The cells were washed and resuspended directly in boiling SDS-PAGE sample buffer (62.5 mM Tris-HCl, pH 6.8, 10% glycerol, 2% SDS, 1% 2-mercaptoethanol, 0.0005% bromphenol blue) for 10 min, subjected to glass bead homogenization, and clarified by microcentrifugation. Following SDS-PAGE and transfer to nitrocellulose, the membrane was probed with antibodies to Ste7 at 1:200 (yN-18) (Santa Cruz). Immunoreactive species were visualized by enhanced chemiluminescence detection (Pierce) of horse-radish peroxidase-conjugated anti-rabbit IgG (Bio-Rad) or anti-goat IgG (Santa Cruz). Specificity of detection was established using ste7⌬ cell extracts as negative controls (10).

RESULTS
It is well established that pheromone stimulation promotes phosphorylation of Ste7 by Ste11. We have shown recently that pheromone stimulation also promotes ubiquitination of Ste7 (10). The enzymes that ubiquitinate Ste7 have not been identified. Our goal here was to identify those enzymes and to determine the mechanism underlying pheromone regulation of ubiquitination. Our initial approach was to examine the ubiquitination and signaling properties of strains lacking each of the proteins known to bind or phosphorylate Ste7. These include Ste11, the Ste7 substrates Fus3 and Kss1, as well as the kinase scaffolding protein Ste5. We postulated that deletion of STE11 or STE5 might block ubiquitination, because both are required for activation of Ste7. Ste7 phosphorylates Fus3/Kss1, but this is not required for Ste7 activity and might therefore not be required for ubiquitination (14 -18). On the other hand Ste7 is a substrate for Fus3/Kss1, and this feedback phosphorylation event might be required (14,15,18,19).
For all of our experiments we monitored Ste7 ubiquitination by immunoblotting with Ste7 antibodies. We have shown previously that ubiquitination of Ste7 results in the formation of a high molecular weight species that migrates slowly in SDS-PAGE. That species is recognized by anti-Ste7 antibodies as well as by anti-ubiquitin antibodies following Ste7 immunoprecipitation (10). In each experiment we either deleted the UBP3 gene or expressed a catalytically inactive dominant negative form of Ubp3 (Ubp3 C469S ) (20) to stabilize ubiquitinated Ste7. Cells in mid-log phase were treated with ␣-factor pheromone, harvested, lysed directly in SDS-PAGE sample buffer, and analyzed by immunoblotting. Under these conditions, ubiquitinated Ste7 accumulates slowly, becoming fully modified after 2 Q. Ge and B. Errede, manuscript in preparation. 15Dau cdc28-1 Beverly Errede uitin-modified protein is easily detected in wild-type cells, none is detected in ste5⌬, ste11⌬, or fus3⌬/kss1⌬ mutants (Fig. 2). Thus it appears that deletion of any component of the MAPK assembly is sufficient to block ubiquitination of Ste7.
We then examined whether phosphorylation per se, as opposed to expression of the phosphorylating kinase, is required for ubiquitination of Ste7. Ste11 is required for Ste7 function, because mutational replacement of the Ste11 phosphorylation sites (Ser 359 and Thr 363 ) completely abolishes the biological activity of Ste7 (21). We transformed a ste7⌬ strain with plasmids that express wild-type Ste7, a Ste7 mutant lacking the Ste11 phosphorylation sites (Ste7 A2 ) (18,21,22), or a multiply substituted mutant that does not undergo feedback phosphorylation by Fus3/Kss1 (Ste7 A7 ) (19). 2 As shown in Fig. 3, pheromone-dependent ubiquitination is blocked in the Ste7 A2 mutant but not by the Ste7 A7 mutant. Ste7 A7 does not undergo a pheromone-dependent mobility shift, consistent with loss of Fus3/Kss1 phosphorylation. 2 Thus it appears that phosphorylation by Ste11 is necessary for ubiquitination of Ste7.
We also examined whether Ste5 might play a direct role in ubiquitination of Ste7. The only known function of Ste5 is to assemble Ste11, Ste7, and Kss1/Fus3 kinases (23,24). However, Ste5 also has a RING-H2 domain found in some E3 ubiquitin ligases, and we have speculated that Ste5 might directly ubiquitinate its binding partner Ste7 (10). To test this we examined whether RING domain substitutions at Cys 177 and Cys 180 (Ste5 C177A,C180A ) would block the ubiquitination of Ste7. The RING domain is also required for binding of G␤␥ (25). Thus we fused Ste5 C177A,C180A to glutathione S-transferase (Ste5 C177A,C180A -GST). Dimerization of GST bypasses the requirement for G protein activation, resulting in a constitutive signal (26). As shown in Fig. 4, Ste7 expression is diminished in cells expressing Ste5 C177A,C180A -GST, but it is nevertheless ubiquitinated. Ubiquitination occurs even in the absence of added pheromone, consistent with a constitutive signaling phenotype exhibited by Ste5 C177A,C180A -GST (26). Thus it appears that the RING domain of Ste5 is not necessary for ubiquitination of Ste7. These data argue against the model that Ste5 is the E3 ubiquitin ligase for Ste7.
Although the Ste5 RING domain appears not to be required for Ste7 ubiquitination, Ste5 may still play a direct role in this regulatory process. It was observed previously that Ste5 binds preferentially to an inactive hypophosphorylated form of Ste7 (23), suggesting that phosphorylation and activation of Ste7 leads to dissociation from Ste5. Dissociation from Ste5 might expose Ste7 to the appropriate modifying enzymes, and this might be sufficient for ubiquitination to occur. Thus we considered whether dissociation from Ste5 could itself trigger the ubiquitination of Ste7. To this end we monitored Ste7 ubiquitination in cells that express Ste5 V763A,S861P , a mutant that is unable to bind to Ste7 yet is competent to bind Ste11 (13). As shown in Fig. 5, this mutation eliminates ubiquitination of Ste7. We conclude that dissociation of Ste7 from Ste5 is not itself a signal for ubiquitination.
The data so far indicate that phosphorylation by Ste11 is required for ubiquitination of Ste7. Ste5 is also required, but it does not appear to directly ubiquitinate Ste7 (Ste5 could still act as an E3 ligase for other substrates). Thus we considered whether the SCF complex could instead serve this function, because it is required for progression through the G 1 -S transition where pheromone arrest occurs. Moreover, most if not all SCF substrates are phosphorylated prior to being ubiquitinated (27). The SCF complex is comprised of four proteins: Cdc53 (cullin), the E2 ubiquitin conjugating enzyme Cdc34 (also known as Ubc3), the RING-type E3 ubiquitin ligase Hrt1,

FIG. 3. Phosphorylation by Ste11 but not phosphorylation by
Fus3/Kss1 is necessary for pheromone-dependent ubiquitination of Ste7. Strain YHD1001 (ste7⌬ ubp3⌬) was transformed with plasmid pRS316 expressing wild-type STE7, a STE7 mutant that blocks phosphorylation by Ste11 (ste7 A2 ), or a multiply substituted STE7 mutant that blocks feedback phosphorylation by MAPKs (ste7 A7 ). The cells were grown to mid-log phase and treated with 3 M ␣-factor for 60 min as indicated. The cell extracts were analyzed by immunoblotting (IB) as described in the legend to Fig. 1. and the substrate adapter protein Skp1 (27). For these experiments we compared ubiquitination of Ste7 using temperature sensitive cdc34-2 and cdc53-1 mutants (27). Growth at 37°C inactivates each of these mutants, resulting in cell cycle arrest at the late G 1 phase. As shown in Fig. 6A, pheromone-dependent ubiquitination is blocked in both the cdc34-2 and cdc53-1 mutant strains. Ubiquitination is unaffected by cdc28-1, another temperature-sensitive mutant that also blocks cell division at late G 1 but that is not part of the SCF complex (Fig. 6B). We conclude from these experiments that SCF function is needed for ubiquitination of Ste7.
As an additional control we tested Ste7 ubiquitination in a panel of mutants affecting other E2 family members in yeast, including ubc1, ubc2, ubc4, ubc5, ubc6, ubc7, ubc8, ubc10,  ubc11, ubc12, and ubc13. In some cases double mutants (ubc1/ ubc4, ubc2/ubc4, ubc4/ubc5, and ubc6/ubc7) were used, because some Ubc isoforms are partially redundant in function. None of these other enzymes altered the ubiquitination of Ste7 (Table I and data not shown). Thus it appears that Ste7 ubiquitination requires SCF activity specifically and is not blocked simply as a result of arrest in G 1 or of growth at 37°C.
The data presented above indicated that SCF is required for ubiquitination of Ste7. The defect in Ste7 ubiquitination is not the result of cell cycle arrest, because ubiquitination is still observed in the cdc28 mutant (Fig. 6B). We also considered whether cell cycle arrest is necessary for Ste7 ubiquitination. This seemed likely given that prolonged pheromone stimulation is required before ubiquitination is evident (Fig. 1). To this end we tested fus3⌬ and far1⌬ mutants, which diminish and eliminate pheromone-dependent growth arrest response, respectively. As shown in Fig. 7, Ste7 ubiquitination is diminished in the fus3⌬ mutant and eliminated in the far1⌬ mutant. These data suggest that G 1 arrest is indeed necessary for Ste7 ubiquitination. It is clearly not sufficient, however, because ubiquitination in the cdc28 mutant still requires pheromone addition (Fig.  6B). Moreover, the far1 mutant does not block Ste7 phosphorylation. This result indicates that although phosphorylation is necessary for Ste7 ubiquitination, it too is not sufficient. FIG. 5. Dissociation from Ste5 is not sufficient to trigger ubiquitination of Ste7. Strain YHD1002 (ste5⌬ ubp3⌬) was transformed with plasmid YCplac33 expressing wild-type STE5 or a mutant that is unable to bind to Ste7 but is competent to bind Ste11 (ste5 V763A,S861P ). The cells were grown to mid-log phase and treated with 3 M ␣-factor for 60 min, as indicated. The cell extracts were analyzed by immunoblotting (IB) as described in the legend to Fig. 1.   FIG. 6. Pheromone-dependent ubiquitination of Ste7 is blocked in cdc34 and cdc53 mutants. A, strains MTY235 (WT), MTY670 (cdc34-2), and MTY740 (cdc53-1) were transformed with plasmid pYES2.1 expressing a dominant negative UBP3 C469S allele. The cells were grown at 24°C to an A 600 nm of ϳ0.2-0.4, shifted to 37°C for 3 h and treated for 60 min with 3 M ␣-factor, as indicated. The cell extracts were analyzed by immunoblotting (IB) as described in the legend to Fig. 1. B, strains 15Dau (WT) and D13au (cdc28-1) were analyzed as described for A.
The data presented above indicated that phosphorylation and cell cycle arrest are necessary (but not sufficient) for Ste7 ubiquitination. We then investigated whether ubiquitination leads to Ste7 degradation. This is not self-evident, because some proteins undergo ubiquitination without being degraded or targeted to the proteasome. For instance in yeast the ubiquitinated form of the transcription factor Met4 is stabile and remains associated with target promoters, although it fails to form a functional transcription complex (28). Thus we tested whether pheromone stimulation promotes Ste7 degradation. Wild-type cells were treated with ␣-factor to induce phosphorylation and ubiquitination and then treated with cycloheximide to block new protein synthesis. Expression of remaining Ste7 was monitored by immunoblotting. As shown in Fig. 8, Ste7 abundance declined more quickly after pheromone treatment. The treated sample was initially more abundant, but by the end of the time course it was less abundant than the untreated sample. The more quickly migrating unphosphorylated form of Ste7 was even more stabile, with little or no decrease in expression even after 90 min of cycloheximide treatment. The apparent stability of unphosphorylated Ste7 is consistent with our finding that phosphorylation is necessary for ubiquitination. We then investigated whether mutants that diminish Ste7 ubiquitination also diminish Ste7 degradation. The cdc34-2 and cdc53-1 mutants were shifted to 37°C to inactivate SCF, treated with ␣-factor to induce phosphorylation and ubiquitination, and then treated with cycloheximide. As shown in Fig. 9, the loss of Ste7 is slowed at least 3-fold in the ubiquitination-deficient cdc34-2 or cdc53-1 mutant cells, as compared with wild-type cells grown under identical conditions. Taken together, these results indicate that pheromone stimulation accelerates phosphorylation, ubiquitination, and ultimately the degradation of Ste7.

DISCUSSION
Many biological processes are regulated through the ubiquitination and degradation of cellular proteins. Ubiquitination has been shown to play a key role in regulation of the cell cycle (via destruction of mitotic cyclins), transcription (p53, NF〉), differentiation, and development (growth factor receptors). There is also a growing number of G protein signaling proteins shown to be ubiquitinated. Examples can be found in mammals and yeasts and include G protein-coupled receptors (5-7, 29, 30), G protein ␣ and ␥ subunits (8,9,31,32), RGS proteins (33,34), MAPK kinase kinases (11,35,36), and MAPK kinases (10,37). In yeast, ubiquitination of cell surface receptors serves as a signal for ligand-induced endocytosis and delivery to the vacuole/lysosome (5,7,38). Ubiquitination of the G␣ subunit Gpa1 is needed for sustained pheromone signaling in vivo. Gpa1 is also notable because it is the only substrate for which the in vivo ubiquitination site has been mapped directly by mass spectrometry (8).
Clearly, improved understanding of how ubiquitination is regulated is important to our understanding of cell regulation. A particular challenge in carrying out these studies has been the detection of ubiquitination substrates in vivo. This has much to do with the inherent instability of most ubiquitinmodified proteins. For reasons that are unclear, this has been especially difficult in yeast. To address this problem we previously used mutant strains that lack UBP ubiquitin-specific processing proteases, which cleave the amide bond between ubiquitin A wild-type strain (BY4741) was treated with 3 M ␣-factor for 60 min, as indicated. The cells were then treated with the protein synthesis inhibitor cycloheximide for the indicated times. The cell extracts were analyzed by immunoblotting as described in the legend to Fig. 1.   FIG. 9. Ubiquitination promotes the degradation of Ste7. Strains MTY235 (WT), MTY670 (cdc34-2) and MTY740 (cdc53-1) were grown at 24°C to an A 600 nm of ϳ0.2-0.4, shifted to 37°C for 3 h, and treated with 3 M ␣-factor for 60 min. The cells were treated with the protein synthesis inhibitor cycloheximide for the indicated times. The cell extracts were analyzed by immunoblotting, as described in the legend to Fig. 1, and densitometry. and the substrate protein (39). Deletion of the UBP3 gene results in stable expression of the ubiquitinated form of Ste7, and this allowed us to demonstrate that the ubiquitination is elevated following pheromone stimulation (10). Ste7 is notable because it was the first example of a MAPK kinase that is ubiquitinated. Moreover, this modification is likely to be functionally important, because ubp3⌬ mutants exhibit elevated MAPK-regulated gene transcription and growth arrest responses (10).
Another challenge has been determining the mechanisms by which specific substrates are ubiquitinated. One way specificity is achieved is by the large number of E2 and E3 isoforms, which bind protein substrates with a high degree of selectivity (40). Here we have shown that the Cdc34 and Cdc53 components of SCF are required for ubiquitination and efficient degradation of Ste7. Cdc34 and Cdc53 were first identified genetically, as mutants defective in G 1 /S transition in yeast (27). Cdc53 was independently identified through its association with the G 1 cyclin Cln2 (41,42). The cell cycle defect exhibited by each mutant was traced to an inability to eliminate the G 1 cyclin Cln2 (42) and the cyclin-dependent kinase inhibitor Sic1 (43), which are normally destroyed as the cell enters S phase. Cdc34 and Cdc53 were later shown to assemble with Skp1 and Hrt1 to form the SCF complex (44,45). Skp1 binds one of several dozen F box proteins, which together recruit specific substrates to the core complex for ubiquitination.
In yeast the criteria most commonly used to identify SCF substrates is accumulation or stabilization in SCF mutants. Substrates identified in this manner vary widely in function and include G 1 cyclins (Cln1, Cln2, and Cln3), Cdc28 cyclin-dependent kinase inhibitors (Far1 and Sic1), signaling proteins (Gic2), transcriptional regulators (Gcn4), and F box proteins associated with SCF itself (Cdc4 and Grr1) (reviewed in Ref. 27). Here we have shown that mutations in Cdc34 and Cdc53 inhibit Ste7 ubiquitination and degradation. No other E2 enzyme affects Ste7 in this manner.
Another way to achieve specificity is through phosphorylation of the intended substrate. Here we have shown that Ste7 must be phosphorylated by Ste11 for it to be ubiquitinated. This fits the pattern of known SCF substrates, all of which must undergo phosphorylation prior to being ubiquitinated (27). One example is Gic1, an effector for the small GTPase protein Cdc42, which is stabilized by mutations that block phosphorylation (46). Another example is Far1, an inhibitor of cyclin dependent kinase Cdc28. Far1 is itself phosphorylated by Cdc28, and this event is required for binding to an SCF adapter protein Cdc4 (47). Thus it appears that the phosphate group constitutes part of the SCF substrate recognition structure.
We have shown previously that pheromone promotes ubiquitination of Ste7, and this leads to an attenuation of signaling (10). Here we have demonstrated that pheromone-dependent ubiquitination requires phosphorylation by Ste11 (ste7 A2 and ste11⌬), cell division arrest (far1⌬), and a functional SCF (cdc34-2 and cdc53-1). Ubiquitination of Ste7 also promotes its degradation. We conclude that pheromone promotes signal inactivation through accelerated ubiquitination and degradation of Ste7. Pheromone also promotes signal activation through increased Ste7 kinase function. Thus our current efforts are aimed at understanding how such obviously opposing events are coordinated. Given that no Ste7 phosphatase has yet been identified, our present model is that Ste7 degradation serves to limit the activity of the kinase, in lieu of Ste7 dephosphorylation. A cycle of pheromone-dependent activation and degradation of Ste7 would prevent sustained activation of the MAPK and MAPK-dependent responses, allowing the cell to eventually desensitize and resume normal cell division. The requirement of SCF for ubiquitination and degradation of Ste7, as well as for ubiquitination and degradation of cell cycle regulators, would in this way facilitate the resumption of cell division cycling after prolonged pheromone stimulation.