|Home | About | Journals | Submit | Contact Us | Français|
Signal transduction networks can display complex dynamic behavior such as oscillations in the activity of key components [1-6], but it is often unclear if such dynamic complexity is actually important for the network's regulatory functions [7, 8]. Here we found that the mitogen-activated protein kinase (MAPK) Fus3, a key regulator of the yeast mating pheromone response undergoes sustained oscillations in its phosphorylation/activation state during continuous pheromone exposure. These MAPK activity oscillations led to corresponding oscillations in mating gene expression. Oscillations in MAPK activity and gene expression required the negative regulator of G-protein signaling Sst2, and partially required the MAPK phosphatase Msg5. Peaks in Fus3 activation correlated with periodic rounds of cell morphogenesis, with each peak preceding the formation of an additional mating projection. Preventing projection formation did not eliminate MAPK oscillation, but preventing MAPK oscillation blocked the formation of additional projections. A mathematical model was developed that reproduced several features of the observed oscillatory dynamics. These observations demonstrate a role for MAPK activity oscillation in driving a periodic downstream response, and explain how the pheromone signaling pathway, previously thought to desensitize after 1-3 hours, controls morphology changes that continue for a much longer time.
In haploid yeast cells, MAPK signaling is activated in response to pheromone secreted by a cell of the opposite mating type. As a consequence, the yeast arrest their cell cycle in the G1 phase, and initiate a developmental programme characterized by alterations in gene expression, oriented growth toward the mating partner (mating projection formation) and, ultimately, fusion of the two haploid cells to form a diploid [9, 10].
At the molecular level, pheromone binding to a G-protein-coupled receptor triggers a signal transduction cascade containing the MAPKs Fus3 and Kss1. These MAPKs share over 50% sequence identity with human ERK1 and ERK2, and like ERK1/2 are activated via dual phosphorylation by a MAPK kinase (yeast Ste7) and deactivated by various MAPK phosphatases.
Previous studies of the pheromone response have indicated that it is relatively short-lived, with MAPK activity and pheromone-induced transcript levels peaking within the first hour of stimulation, and returning to their unstimulated baseline levels 2-4 hours after pheromone addition [11-13]. This apparent desensitization is a hallmark of G-protein-coupled receptor pathways .
Since the physiological response to pheromone can persist for many hours under certain circumstances, however [14-16], we wondered how physiological responses persisted if the pathway was stably desensitized. To gain some preliminary insight into this question, we performed an extended time-course in asynchronous, mid-log cultures. Surprisingly, after approximately 3-5 hours of continuous pheromone exposure, MAPK activation returned to near-peak levels (Supplementary Fig. 1).
To investigate MAPK oscillations at a higher level of resolution, cells carrying an integrated fusion of Fus3MAPK to enhanced green fluorescent protein (GFP) were synchronized by S-phase arrest/release, and pheromone was added to cells during the G1 phase, when the mating response is maximal (3 different pheromone doses were tested). The treated cells were then monitored for MAPK phosphorylation levels, Fus3-GFP fluorescence and localization, and cell morphology (Fig. 1A). The set of strains used for these experiments lacked the Bar1 protease, so as to remove the potentially confusing effects of pheromone proteolysis (in contrast, the strain used for Supplementary Fig. 1 was BAR1+). As shown in Fig. 1B, in these G1-synchronized cells, an initial peak of MAPK activation occurred within 60 minutes of pheromone addition, and then declined, consistent with the short-time course studies cited above. When observed over an 8 hour time course, however, sustained oscillations of MAPK activity were present: phospho-Fus3 levels rose again, fell again, then rose and fell yet a third time. Elevated activation of Fus3 occurred at approximately 60, 180-240 and 420-480 minutes post stimulation (Fig. 1C), with activation of Kss1MAPK roughly paralleling this trend (Fig. 1B). As assessed by microscopic examination, cells remained in G1-arrest during the entire 8 hour time course (data not shown); thus, reentry into the cell cycle did not account for the observed oscillation. In summary, MAPK activation levels oscillate, with 3 distinct peaks over 8 hours of persistent stimulation.
During mating, activated Fus3MAPK and Kss1MAPK phosphorylate the transcriptional regulators Ste12, Dig1 and Dig2 , resulting in the up- and down-regulation of multiple genes, including genes involved in morphogenesis and cell-cell fusion (such as FUS1), genes encoding negative regulators of the pathway (such as SST2 and MSG5), and genes encoding certain members of the signaling pathway itself (such as FUS3) . To determine whether the oscillations in MAPK activation resulted in corresponding oscillations of pheromone-induced gene transcription, we monitored the expression levels of both Fus3-GFP and Fus1-GFP by immunoblotting (Fig. 1D and 1E) and flow cytometry (data not shown). Both methods showed that Fus3-GFP and Fus1-GFP protein levels displayed strong oscillatory dynamics with frequencies similar to the oscillations of Fus3 activation. Furthermore, increasing the pheromone dose resulted in a progressive phase shift of Fus3-GFP expression peaks towards earlier time points (Fig. 1D); this trend was also observed for phospho-MAPK levels (Fig. 1C), although it was less apparent. In general, the peaks in protein expression levels roughly coincided with the peaks in phospho-Fus3, as expected given the relatively slow oscillatory period of the later (a notable exception, however, was the oscillation in Sst2 protein levels; see below). In summary, the expression levels of MAPK target genes oscillate in approximate synchrony with MAPK activation levels.
Oscillatory behavior is often a sign of the presence of at least one negative feedback in the underlying molecular interactions, although negative feedback does not necessarily entail oscillation , and oscillation does not strictly require a classical negative feedback loop . Several negative regulators have been shown to affect the mating pathway, including the MAPK phosphatase Msg5 and the regulator of G-protein signaling (RGS) protein Sst2 [20, 21]. Msg5 dephosphorylates, and thereby deactivates, Fus3, but its function is partially redundant with several other phosphatases . Thus, cells lacking Msg5 display a modest increase in pheromone sensitivity. Sst2 binds to the pheromone receptor and stimulates the GTPase activity of the pheromone receptor-coupled G protein; cells lacking Sst2 display severe hypersensitivity to pheromone . As both MSG5 and SST2 are pheromone-inducible genes, they are components of negative feedback loops that might control the oscillatory activity of the pathway [20, 22].
To gain theoretical insight regarding how negative feedback might modulate the properties of the pheromone signaling network, we developed a mathematical model of a MAPK signaling cascade with a negative feedback, and parameterized with values appropriate to the yeast pheromone response pathway (see Supplementary Materials). Model simulations (Fig. 2A) testing the role of different pheromone doses on Fus3MAPK activation rates reproduced the observed dependence of oscillation amplitude on dose strength. In addition, this analysis predicted a slight phase shift with dose variation, and increased dampening of the oscillations with decreasing dose (see a more extensive model analysis in Supplementary Material); both these trends were apparent in the experimental data. Further, the model predicted that Sst2 expression would oscillate with the same frequency as did Fus3 phosphorylation/activation levels, but should be shifted in phase (Fig. 2B); this prediction was supported by examination of a wild-type strain stimulated with 100 nM α-factor (Fig. 2C). Not surprisingly, the model also predicted that decreasing the strength of the negative feedback (by removing Sst2 or lowering its activity) would dampen MAPK activity oscillations (Fig. 2D). These theoretical results show that Sst2-mediated negative feedback is, in principle, sufficient to generate oscillations resembling those observed experimentally.
To experimentally investigate the role of Sst2 and Msg5 on MAPK oscillation, we compared a wild-type strain to an otherwise isogenic strain lacking Sst2 (i.e., an sst2Δ strain). In the strain lacking Sst2, Fus3 activation persisted for at least 8 hours following pheromone addition, without obvious oscillations (Fig. 2E). In addition, in an msg5Δ strain, the oscillation of active Fus3 was present, but substantially damped (Fig. 2F). Oscillations in the expression of pheromone-induced genes (as represented by Fus3-GFP) were also compromised in sst2Δ and msg5Δ strains (Figs. 2E and 2F, plotted in 3C and 3D). In summary, the negative regulators Sst2 and Msg5 are required for wild-type oscillatory dynamics. Further work is needed to address whether negative feedback per se is also required.
Mating projection formation may occur either by polarization in the strongest pheromone gradient emanating from an adjacent mating partner [14, 16], or, when gradients are too shallow or variable to be detected precisely, by sequential formation of multiple projections in random directions (the so-called “default response”) [15, 23]. Cells can also form a second projection when the direction of the gradient is switched (our unpublished observations). Multiple projections may facilitate mating in environments where cells are present in dense groups, and saturating pheromone concentrations (or noisy, fluctuating gradients) may arise naturally due to rapid proliferation, flocculation, or biofilm formation.
We hypothesized that oscillations in Fus3 activation might control the timing of the formation of multiple mating projections, and that the frequency of MAPK oscillation would therefore correlate with the periodicity of projection formation. To test this prediction, we quantified projection formation by microscopic examination of the same cells that were used for determination of the dynamics of MAPK activation and gene expression.
As shown in Fig. 3, during 8 hours of pheromone exposure, wild-type cells exhibited a progressive increase in the number of projections, with more than half of the cells displaying 3 projections by the end of the time course (Figs. 3A, E, F); these results are similar to those previously reported . Importantly, the dynamics of the onset of successive projections displayed an excellent correlation with the cycling of both Fus3-GFP and Fus1-GFP protein levels (Fig. 3A). In particular, in cells exposed to 100 nM pheromone, the formation of the second projection coincided with the rise in gene expression at 3-5 hours, and the formation of the third projection coincided with the rise in gene expression at 7-8 hours. The correlation between the timing of each peak in gene expression and the formation of a new projection was preserved at the higher pheromone dose of 25 μM, with both periodic processes undergoing a noticeable phase shift towards earlier time (Fig. 3B).
A second prediction arising from the hypothesis that the MAPK oscillator drives periodic projection formation is that genetic alterations that alter the dynamics of phospho-Fus3 oscillation should also affect the periodicity of projection formation. To test this prediction, projection formation was evaluated in sst2Δ and msg5Δ strains, where oscillations are severely damped. Both mutant strains formed an initial projection at a rate similar to that of the wild-type strain. Strikingly, however, both mutants predominantly formed only this single projection, even though they were continuously exposed to pheromone for 8 hours (Figs. 3C, D, F).
Since the total level of active Fus3 is persistently increased in sst2Δ cells following incubation with 100 nM pheromone, we considered the possibility that the increase in the total amount active Fus3, rather than the loss of the oscillation, was responsible for the defect in secondary projection formation. To address this question, sst2η cells were treated with a 10-fold reduced concentration of pheromone (i.e, 10 nM). Under these conditions, the average level of active Fus3 was reduced to levels comparable to SST2η cells treated with 100 nM pheromone, yet the cells still only formed one projection (data not shown); this suggests that the lack of oscillation, rather than a change in the level of Fus3 activation, caused the defect in periodic morphogenesis in cells lacking Sst2. In summary, MAPK oscillations appear to drive the periodic formation of multiple mating projections.
To investigate whether actin dynamics and/or projection formation reciprocally regulated MAPK oscillation, we examined Fus3 phopshorylation in a strain lacking Bni1, which is required for the formation of actin cables and mating projections . As previously reported, in the bni1Δ strain, the average Fus3 activation and expression levels were considerably lower than in the wild-type, consistent with putative amplification of Fus3 signaling by actin cables . However, the oscillatory activation and expression of Fus3-GFP was nevertheless preserved, albeit being somewhat damped (Figs. 4A-C). Thus, the formation of mating projections influences, but is not essential for, MAPK oscillation.
Here we demonstrated persistent oscillations in the phosphorylation and activation of the MAPKs Fus3 and Kss1, and in the expression of MAPK-target genes, in response to continuous pheromone stimulation over an 8-hour period. In addition, we showed that these oscillations require the negative regulators SST2 and MSG5. Finally, we provided evidence linking the oscillations in MAPK signaling to periodic changes in cell shape and polarity that occur when cells form a secondary mating projection(s) if the first attempt at partner location is unsuccessful. Our results suggest that the yeast MAPK pathway functions as a timer to control periodic morphogenesis, so as to facilitate mating in shallow or noisy gradients. These findings may be relevant to understanding the regulation of other periodic morphogenetic processes, such as neuronal development, segmentation, cell locomotion, and contractility.
This work was supported by NIH-NIGMS research grants GM60366, GM75309 and GM76516 (L.B.), GM72024 and RR020839 (A.L.), GM69013 and GM84332 (L.B. and A.L.) and the NSF grant MCB-0331306 (A.L.)
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.