Calcineurin, the Ca2+-dependent phosphatase, regulates Rga2, a Cdc42 GTPase-activating protein, to modulate pheromone signaling.

Calcineurin, the conserved Ca2+/calmodulin-activated phosphatase, is required for viability during prolonged exposure to pheromone and acts through multiple substrates to down-regulate yeast pheromone signaling. Calcineurin regulates Dig2 and Rod1/Art4 to inhibit mating-induced gene expression and activate receptor internalization, respectively. Recent systematic approaches identified Rga2, a GTPase-activating protein (GAP) for the Cdc42 Rho-type GTPase, as a calcineurin substrate. Here we establish a physiological context for this regulation and show that calcineurin dephosphorylates and positively regulates Rga2 during pheromone signaling. Mating factor activates the Fus3/MAPK kinase, whose substrates induce gene expression, cell cycle arrest, and formation of the mating projection. Our studies demonstrate that Fus3 also phosphorylates Rga2 at inhibitory S/TP sites, which are targeted by Cdks during the cell cycle, and that calcineurin opposes Fus3 to activate Rga2 and decrease Cdc42 signaling. Yeast expressing an Rga2 mutant that is defective for regulation by calcineurin display increased gene expression in response to pheromone. This work is the first to identify cross-talk between Ca2+/calcineurin and Cdc42 signaling and to demonstrate modulation of Cdc42 activity through a GAP during mating.

INTRODUCTION

Calcineurin (CN), the Ca2+/calmodulin-activated phosphatase, translates Ca2+ signals into downstream regulatory events and is widely expressed in the animal and fungal kingdoms (Rusnak and Mertz, 2000). Although CN structure is highly conserved, its physiological functions have diverged significantly in fungi and mammals due to rapid rewiring of the CN signaling network through gain and loss of docking motifs in its substrates (Goldman ). In mammals, CN is best known as the target of immunosuppressant drugs cyclosporin A and FK506, which inhibit CN-mediated dephosphorylation of NFAT transcription factors to block T-cell activation (Crabtree and Schreiber, 2009). In fungi, CN broadly regulates stress responses and is essential under stringent growth conditions (Cyert and Philpott, 2013). These include survival in a mammalian host for several pathogenic fungi, where CN is required for virulence (Juvvadi ). Thus elucidation of fungal CN signaling pathways may identify novel targets for antifungal drug development.

In the budding yeast Saccharomyces cerevisiae, CN is dispensable for growth under standard laboratory conditions, when cytosolic Ca2+ levels are low, but essential under conditions that trigger Ca2+ signaling, that is, environmental stress, including ionic and cell wall stress, and pheromone signaling (Cyert and Philpott, 2013, Carbó ). CN regulates distinct substrates and processes under different signaling conditions (Cyert and Philpott, 2013; Goldman ; Arsenault ; Guiney ). In cells responding to pheromone, CN inhibits mating-induced gene expression and reduces levels of pheromone receptor at the cell surface by dephosphorylating Dig2 and Rod1/Art4, respectively (Cyert and Thorner, 1992; Alvaro , 2016; Goldman ). This CN-mediated down-regulation of the pheromone response is essential because CN mutants lose viability during prolonged exposure to mating pheromone (Moser ; Withee ). A systematic screen uncovered >15 proteins and their associated processes as novel targets of CN (Goldman ). One of these, Rga2, a GTPase-activating protein (GAP) for the Rho-type GTPase Cdc42, suggested previously undescribed cross-talk between CN and Cdc42 signaling and was chosen for further investigation. Specifically, we sought to identify the effect of CN-dependent dephosphorylation on Rga2 function and the physiological context in which this regulation occurs.

Members of the Rho family of small GTPases, that is, Rho, Rac, and Cdc42, are master regulators that transduce extracellular signals into complex behaviors such as wound healing and chemotaxis by coordinately regulating cytoskeletal dynamics, vesicle trafficking, gene expression, and proliferation (Hall, 2012; Abreu-Blanco ; Mócsai ). Rho GTPases control diverse effectors and act as molecular switches; the active, GTP-bound state is promoted by guanine nucleotide exchange factors (GEFs) and inactivation is mediated by GAPs, which stimulate conversion to GDP-bound Rho (Hall, 2012). These Rho GEFs and GAPs allow Rho GTPase signaling to be regulated in a tissue- and signal-specific manner and precisely control the spatiotemporal pattern of Rho GTPase cycling in vivo (Fritz and Pertz, 2016). Perturbation of Rho GAPs causes pathologies including cancer, mental retardation, and kidney disease (Cherfils, 2014), and understanding their regulation provides critical insight into the complex control of Rho GTPase signaling in vivo (Bernards and Settleman, 2004).

In S. cerevisiae, the conserved Cdc42 GTPase is essential for asymmetric cell growth and critically regulates responses to pheromone, nutrient limitation, and osmotic stress (Chen and Thorner, 2007; Bi and Park, 2012). Cdc42 activity is controlled by a single essential GEF, Cdc24, and multiple GAPs, Bem3, Rga1, and Rga2 (Zheng , 1994; Stevenson ; Chen ; Smith ), which serve both overlapping and unique functions (Smith ; Caviston ; Tong ). Bem3 and Rga2 undergo cell cycle–dependent phosphorylation (Knaus ; McCusker ; Sopko ), and cyclin-dependent kinases (Cdks) target inhibitory S/TP sites in Rga2 to facilitate activation of Cdc42 during bud emergence in G1 (Sopko ). During mating, pheromone-activated receptor recruits Cdc24 to the membrane, causing a local increase in Cdc42-GTP, which initiates a kinase cascade that culminates in activation of the Fus3 mitogen-activated protein kinase (MAPK) to induce cell cycle arrest, gene expression, and shmoo formation (Alvaro and Thorner, 2016). Although deletion of Cdc42 GAPs elevates both basal and induced levels of pheromone-dependent gene expression, whether these proteins are regulated during the mating response has not been investigated (Smith ).

In this work, we uncover a new role for Rga2 in regulating Cdc42 activity during pheromone signaling. Rga2 is phosphorylated by Fus3 on inhibitory S/TP sites and dephosphorylated by CN to dampen pheromone signaling. When exposed to mating factor, cells expressing an Rga2 mutant deficient for CN-dependent regulation show increased mating-induced gene expression. Thus we establish a novel mechanism, mediated by Ca2+ and CN, that regulates Cdc42 signaling, and identify Rga2 as an additional target through which calcineurin limits the pheromone response.

RESULTS AND DISCUSSION

CN interacts with and positively regulates Rga2 in vivo

Previous work showed that CN dephosphorylates Rga2 in vitro and in addition identified a sequence that lies downstream of the N-terminal LIM domains, PQVLVS (amino acids 192–197), as matching the consensus for a known CN-docking motif termed PxIxIT (Figure 1A; Roy ; Goldman ). To examine the interaction of Rga2 with CN, we first showed that, when fused to glutathione S-transferase (GST), a 22–amino acid sequence from Rga2 containing PQVLVS copurified with recombinant hexahistidine (6xHis)-CN and that, consistent with PxIxIT-dependent binding, mutation of key residues to alanine (AQALAA) disrupted the interaction (Figure 1B). To determine whether this putative PxIxIT site is functional in the context of the full-length protein, we expressed wild-type (WT) Rga2 (Rga2WT) or PxIxIT mutants (Rga2AQALAA or a deletion of PQVLVS, Rga2PΔ) in Escherichia coli and tested them for copurification with GST-CN. Rga2WT but not Rga2AQALAA or Rga2PΔ copurified with GST-CN, and none of the Rga2 proteins copurified with GST (Figure 1C). Thus the PQVLVS sequence mediates a specific interaction with CN. To examine whether, as for other substrates (Grigoriu ; Alvaro ; Goldman ; Arsenault ; Guiney ), CN fails to regulate Rga2 PxIxIT mutants, we examined the electrophoretic mobility of Rga2 in extracts of yeast in which CN was acutely activated by addition of 200 mM CaCl2. A fast-migrating form of Rga2 was observed, whereas an additional, slower-migrating band appeared in cultures that were pretreated with the CN inhibitor FK506 (Figure 1D). Consistent with previous findings, this slower-migrating band reflects hyperphosphorylation (Sopko ; Goldman ) and confirms that Rga2 is dephosphorylated by CN in vivo. In contrast to Rga2WT, the Rga2 PxIxIT mutants Rga2AQALAA and Rga2PΔ were hyperphosphorylated under CN activating conditions, and phosphorylation levels were unaffected by FK506, demonstrating the abrogation of CN-dependent dephosphorylation. Thus Rga2 PxIxIT mutants allow us to investigate the functional consequences of CN-mediated regulation of Rga2.

FIGURE 1:

PQVLVS sequence binds CN and is required for regulation of Rga2 by CN. (A) The domain structure of Rga2 and position of the PQVLVS sequence. (B) Immunoblots showing that GST fused to amino acids 184–205 from Rga2, which contains PQVLVS, copurifies with 6xHis-CN, whereas the mutated sequence GST-AQALAA and GST do not (Pulldown). Arrowheads, positions of CN (anti-His) and GST-peptide fusion protein (anti-GST). Input, immunoblot (anti-GST) showing amounts of GST proteins added to each assay. (C) Immunoblot showing that recombinant 6xHis-Rga2WT copurifies with GST-CN but not with PxIxIT mutants Rga2AQALAA or Rga2PΔ, with residues 192–197 deleted (Pulldown). Arrowheads, positions of Rga2 (anti-His), GST-CN, and GST (anti-GST). Input, immunoblot (anti-His) showing amounts of Rga2 proteins added to each assay. (D) Calcineurin dephosphorylates Rga2 in vivo under Ca2+-activating conditions. Cells (BY263) expressing Rga2-FLAG plasmids were treated with CN inhibitor FK506 or vehicle before 200 mM CaCl2 activation of CN. Arrowheads, hyperphosphorylated (Rga2-P) and hypophosphorylated (Rga2) forms (anti-FLAG). Percentage of total Rga2 signal represented by slower-migrating band is shown below each lane. (E, F) Rga2 PxIxIT mutants are less active in vivo than Rga2WT. (E) Yeast lacking Cdc42 GAPs (bem3Δ rga1Δ rga2Δ) and containing a pheromone-responsive FUS1-LacZ reporter were incubated with pheromone before measurement of β-galactosidase activity. Strains also expressed pGAL1-FLAG vector or Rga2-FLAG plasmids as indicated. Error bars are SD from four biological replicates. (F) Growth rate (doublings/hour) of WT yeast (NLY33) treated with 20 µM estradiol and expressing pEG(KT) vector or GST-Rga2 plasmids. Error bars are SD from three biological replicates. For E and F, all values were normalized to vector, and statistical difference from Rga2WT was assessed using one-way ANOVA with Tukey’s correction for multiple comparisons, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Because Cdc42 GAPs are functionally redundant, we examined Rga2 activity in a strain lacking three Cdc42 GAPs (bem3Δ rga1Δ rga2Δ), which displays elevated levels of pheromone-induced gene expression (Smith ). In the presence of pheromone, high levels of FUS1-LacZ activity were observed in cells overexpressing vector or catalytically inactive Rga2 (K872A, Rga2GAPmut), which were reduced in cells overexpressing Rga2WT (Figure 1E). In comparison, neither of the Rga2 PxIxIT mutants was as effective as Rga2WT at reducing β-galactosidase activity (p < 0.01 and p < 0.05, Figure 1E), despite equivalent expression levels for all Rga2 variants (unpublished data). As a second measure of function, Rga2 was overexpressed in wild-type cells. As previously observed, overexpression of Rga2WT, but not Rga2GAPmut or vector, decreased the growth rate (Figure 1F; Sopko ). In contrast, neither of the Rga2 PxIxIT mutants slowed growth to the same degree as Rga2WT (p < 0.001, Figure 1F). Together these findings show that loss of regulation by CN decreases Rga2 function in vivo and suggest that CN activates Rga2, presumably through dephosphorylation.

CN opposes Fus3 to regulate Rga2 during pheromone signaling

To determine when CN dephosphorylates Rga2 in vivo, we sought physiological conditions under which the phosphorylation status of Rga2WT and Rga2AQALAA differs (as in Figure 1D). Rga2 is expressed at very low levels from its native promoter (200–300 molecules/cell; Ghaemmaghami ) and could not be reliably detected by immunoblot. Therefore we expressed FLAG-tagged Rga2 from a CEN plasmid under control of the GAL promoter in a wild-type strain (BEM3 RGA1 RGA2) carrying the GEV chimera for β-estradiol–induced expression of galactose-regulated genes (Gao and Pinkham, 2000; Veatch ). Cells were grown in the presence of low β-estradiol levels to allow detection of Rga2 without perturbing cell growth (Supplemental Figure S1, A and B). During G1 phase of the cell cycle, Cdks phosphorylate and inhibit Rga2, maximizing Cdc42 activation for bud emergence (Sopko ). The phosphatase(s) that counteract these phosphorylation events have not been identified. Because CN opposes Cdk1 regulation of the S-phase transcription factor Hcm1 (Arsenault ), we tested whether CN dephosphorylates Rga2 during the cell cycle. Yeast cells synchronized in G1 by pheromone arrest were released into the cell cycle, and the electrophoretic mobility of Rga2WT and Rga2AQALAA were compared. Rga2WT ran as a doublet, and the relative abundance of the upper band varied through the cell cycle, peaking at 30–45 min, which coincided with S phase (Supplemental Figure S1, B and C). This agrees with previous studies, which demonstrated that hyperphosphorylation of Rga2 by G1-Cdks results in its decreased electrophoretic mobility (Sopko ). Rga2AQALAA was indistinguishable from Rga2WT in these analyses, suggesting that CN does not dephosphorylate Rga2 during the mitotic cell cycle. Of interest, PxIxIT-dependent effects on growth inhibition were observed when Rga2 was overexpressed (Figure 1F), indicating possible activation of CN under these conditions.

While executing these cell cycle studies, we observed that the relative abundance of the upper band in the doublet differed strikingly between Rga2WT and Rga2AQALAA in extracts of pheromone-arrested cells. Under these conditions, ∼70% of Rga2AQALAA appeared as the slower-migrating form, whereas, consistent with CN-dependent dephosphorylation, the majority of Rga2WT was in the faster-migrating form (Figure 2A). To confirm that these electrophoretic mobility differences were due to phosphorylation, we purified Rga2WT and Rga2AQALAA from pheromone-treated cells and incubated them with nonspecific λ-phosphatase in the presence or absence of phosphatase inhibitors. Phosphatase treatment of either Rga2WT or Rga2AQALAA increased the abundance of the faster-migrating band, and abrogation of this shift by phosphatase inhibitors confirmed that it was caused by dephosphorylation (Figure 2B).

FIGURE 2:

Fus3-dependent phosphorylation of Rga2 during pheromone arrest. (A) Immunoblot showing enrichment of slow-migrating Rga2-FLAG band in extracts of pheromone-treated cells expressing Rga2AQALAA as compared with Rga2WT. (B) Shift in electrophoretic mobility is due to phosphorylation. Rga2WT or Rga2AQALAA was immunopurified from pheromone-treated cells, treated with λ-phosphatase (λP) in the presence or absence of phosphatase inhibitors (PPase I), and analyzed by immunoblot (anti-FLAG). (C) Rga2 phosphorylation in pheromone-treated cells is dependent on Fus3. Fus3-WT (NLY39 + MR4937) or Fus3-as1 (NLY39 + MR4938) cells expressing Rga2WT-FLAG were treated with or without 1Na-PP1 inhibitor. Representative blot shows Rga2 electrophoretic mobility (anti-FLAG). Graph shows quantification and SD from three biological replicates, *p < 0.05, Student’s t test. (D) Rga2 is a substrate for Fus3 in vitro. Recombinant 6xHis-Rga2 and Fus3WT (solid line) or kinase-dead Fus3K42R (dashed line) were incubated with [γ-32P]ATP-, and 32P incorporation at 0, 8, 16, and 32 min was quantified by phosphoimager analysis. The 32P signal was adjusted for Rga2 protein level, normalized to maximal signal for Rga2WT, and relative phosphorylation level is displayed. (E, F) Fus3 and CDKs regulate similar phosphosites on Rga2. (E) During pheromone signaling, Rga28A is a poor substrate in vivo. Cells (NLY33) transformed with Rga2-FLAG plasmids were induced with 5 nM estradiol and incubated with 10 µM αF. (F) Rga28A phosphomutant is a poor substrate for Fus3 in vitro. Recombinant 6xHis-Rga28A and Fus3WT were incubated with [γ-32P]ATP, and 32P incorporation at 0, 8, 16, and 32 min was quantified as described in D. For A–C and E, phosphorylation was assayed by electro­phoretic mobility, with arrowheads indicating hyperphosphorylated (Rga2-P) and hypo­phosphorylated (Rga2) forms (anti-FLAG) and Hsc82 (anti-Hsc82) loading control. Percentage of total Rga2 signal represented by slower-migrating band is shown below each lane.

Next we investigated the identity of the kinase that phosphorylates Rga2 during pheromone signaling. The Fus3 MAPK was a prime candidate because it is activated during pheromone signaling and, like Cdks, targets S/TP sites. To test for Fus3-dependent phosphorylation of Rga2 in vivo, we incubated cells expressing an analogue-sensitive allele of Fus3 (Fus3-as1; Bishop ; Matheos ) with pheromone. Ten minutes after addition of inhibitor, Rga2 phosphorylation was reduced compared with cells treated with vehicle (Figure 2C), suggesting that Fus3 phosphorylates Rga2 during pheromone signaling. In vitro, recombinant WT Fus3 (Fus3WT) but not the kinase-dead form (Fus3K42R; Parnell ) directly phosphorylated Rga2WT and Rga2AQALAA to similar degrees, confirming that differences in their phosphorylation states observed in vivo are due to differential dephosphorylation rather than phosphorylation (Figure 2D). Together these results newly identify Rga2 as a direct Fus3 substrate during the pheromone response.

To test whether Fus3 regulates similar sites as Cdks on Rga2, we used an Rga2 mutant in which eight experimentally verified Cdk-phosphorylated residues were changed to alanine (Rga28A; Sopko ). Phosphorylation of Rga28A by Fus3 was drastically reduced in vivo and in vitro, suggesting that it targets one or more of the same sites as Cdks (Figure 2, E and F). Previous work established that phosphorylation of these sites by Cdks inhibits Rga2 function, thereby increasing active Cdc42 levels and promoting bud emergence; overexpression of a nonphosphorylatable mutant decreases Cdc42-GTP levels in vivo and causes accumulation of large, unbudded cells arrested in G1 (Sopko ). Therefore phosphorylation of these sites by Fus3 strongly suggests that it similarly inhibits Rga2 during the pheromone response, when Cdk signaling is inhibited, to promote the sustained activation of Cdc42 required for downstream signaling events, that is, cell cycle arrest, expression of mating genes, and shmoo formation (Alvaro and Thorner, 2016). In contrast, we predict that, by dephosphorylating Rga2, CN down-regulates Cdc42 and pheromone signaling.

Under maximal signaling conditions, CN positively regulates Rga2 to suppress the pheromone response

Next we examined the timing of Rga2 dephosphorylation by CN during the pheromone response. The electrophoretic mobility of Rga2WT and Rga2AQALAA (expressed at low levels as in Supplemental Figure S1) were compared in extracts of WT cells (BEM3 RGA1 RGA2) that were synchronized by release from hydroxyurea (HU)-triggered S-phase arrest and stimulated with mating factor before G1 (t = 0; Figure 3A). Dephosphorylation of Rga2WT was detected 30 min after pheromone addition, which coincided with entry into G1, the onset of pheromone signaling as detected by the appearance of active Fus3 (Fus3-P) and initial appearance of cells exhibiting early shmoo morphology (Figure 3, A–D). Compared to Rga2WT, Rga2AQALAA remained hyperphosphorylated throughout the time course, suggesting that CN was activated early and throughout the pheromone response to regulate Rga2 (Figure 3D). Consistent with this idea, studies of single yeast cells exposed to pheromone revealed that intracellular Ca2+ spikes initiate early during pheromone signaling and persist throughout the response, with the frequency and total number of Ca2+ signals increasing as a function of pheromone concentration (Carbó ). However, these Ca2+ signals are asynchronous and heterogeneous in amplitude and duration, suggesting that each time point of our experiment included a variable subset of cells that were undergoing Ca2+/CN-activated dephosphorylation of Rga2. Thus, in four independent experiments, we consistently observed that Rga2AQALAA was hyperphosphorylated compared with Rga2WT throughout the pheromone response, although the extent of hyperphosphorylation observed at each time point varied (Figure 3D and Supplemental Figure S1D, p < 0.01). Together these observations indicate that high levels of pheromone trigger Ca2+ signals that activate CN and suggest that CN and Fus3 oppose each other in regulating Rga2. By positively regulating Rga2, CN may inhibit the ability of Cdc42 to regulate cell morphology, cell cycle arrest, and/or gene transcription in response to pheromone (Alvaro and Thorner, 2016).

FIGURE 3:

CN regulates Rga2 throughout the pheromone response. Analysis of yeast (NLY33) synchronized at S-phase with HU and then released; addition of 5 µM αF occurred at time 0, and samples were removed every 15 min thereafter. Strains expressed either Rga2WT (pBA2059) or Rga2AQALAA (pNL111) induced with 5 nM estradiol. (A) FACS analysis shows progressive arrest in G1. (B) Immunoblots show accumulation of active Fus3 (Fus3-P, anti-p44/42 MAPK) and total Fus3 (anti-Fus3). (C) Representative images of Rga2WT cells stained with FITC-concanavalin A at indicated times show progressive accumulation of shmoo morphology. Scale bar, 5 µm. (D) Representative blot of Rga2WT and Rga2AQALAA phosphorylation state. Arrowheads, hyperphosphorylated (Rga2-P) and hypophosphorylated (Rga2) forms (anti-FLAG) and Hsc82 loading control (anti-Hsc82). Percentage of total Rga2 signal represented by slower-migrating band is shown below each lane, and results from four biological replicates are quantified in the graph to the right, with error bars represented as SD. Rga2AQALAA differs from Rga2WT, p < 0.01, Wilcoxon paired nonparametric test. Refer to Supplemental Figure S1D for quantification of each biological replicate.

To assess the phenotypic consequences of Rga2 dephosphorylation, we constructed isogenic strains that contained all Cdc42 GAPs and differed only in the susceptibility of Rga2 to CN-mediated dephosphorylation. Replacement of genomic RGA2 through integration of isogenic, tagged RGA2 or RGA2 alleles at the endogenous locus allowed for their expression at physiological levels. Characterization of these integrated Rga2WT or Rga2AQALAA strains through the pheromone response showed no detectable differences in the timing of cell cycle arrest and shmoo formation (Supplemental Figure S2, A and B) or in shmoo morphology (unpublished data). To test whether CN regulates pheromone-induced gene expression through Rga2, we assayed FUS1-LacZ activity in cells treated with a range of mating factor concentrations. Although Rga2WT and Rga2AQALAA responded similarly to low pheromone concentrations, there was a reproducible and statistically significant difference at higher pheromone concentrations (>5 µM), with Rga2AQALAA cells exhibiting higher levels of β-galactosidase activity than Rga2WT (Figure 4A, p < 0.01 and p < 0.05). This difference between Rga2WT and Rga2AQALAA was consistent, as additional assays with a larger sample size (n ≥ 15) showed increased statistical significance (Figure 4B, p < 0.0001 at 5 and 10 µM mating factor). Thus, under conditions of sustained pheromone signaling, which trigger Ca2+ bursts, CN activates Rga2 to decrease pheromone-induced gene expression.

FIGURE 4:

CN regulates Rga2 to oppose pheromone signaling. (A) Rga2AQALAA yeast cells, compared with Rga2WT, are more sensitive to high pheromone concentrations. Integrated Rga2 alleles (WT, NLY52; AQALAA, NLY54) were assayed for FUS1-lacZ activity after treatment with pheromone at indicated concentrations. FUS1-lacZ activity was normalized to maximum Rga2WT value occurring at 5 μM to represent relative levels. Error bars represent SD from four biological replicates. *p < 0.05, **p < 0.01. Values statistically significant from Rga2WT were assessed using Student’s t test. (B) Yeast expressing Rga2AQALAA display higher levels of pheromone-induced FUS1-lacZ activity. Cells were treated with 5 or 10 µM αF as indicated. Statistical significance assessed by Student’s t test with n = 16 biological samples for Rga2WT and n = 15 for Rga2AQALAA, performed over 2 d (blue triangles, day 1; red circles, day 2). Middle bar represents mean with errors bars as SD, ****p < 0.0001.

Cdc42 signaling is regulated through its GAP, Rga2, during pheromone signaling

The yeast response to pheromone has been intensively studied and yet continues to yield new insights into the fundamental regulatory features of signaling pathways. Our findings provide the first evidence that a Rho GTPase-activating protein, Rga2, modulates the output of this response and is antagonistically regulated by Fus3, the pheromone-activated MAPK, and CN, the Ca2+-activated phosphatase (Figure 5A).

FIGURE 5:

Model. (A) In response to high pheromone concentration, Ca2+ activates CN, which dephosphorylates and activates Rga2 to reduce Cdc42-GTP levels and MAPK-dependent pheromone signaling. See the text for details. (B) In response to high pheromone concentration, CN inhibits the pheromone response by dephosphorylating an α-arrestin, Rod1/Art4, to promote endocytosis; by dephosphorylating a Cdc42 GAP, Rga2, to down-regulate Cdc42 signaling; and by dephosphorylating a Ste12 inhibitor, Dig2, to inhibit mating-induced gene expression (Alvaro , 2016; Goldman ). See the text for details.

During mating, Fus3 substrates induce cell cycle arrest, gene expression, actin polarization and cell fusion (Alvaro and Thorner, 2016). Fus3 promotes polarization by phosphorylating the Bni1 formin, which is also an effector of Cdc42 (Elion ; Tedford ; Matheos ). Therefore our finding that Fus3 regulates Rga2 suggests that this MAPK elevates Cdc42 signaling to create positive feedback that reinforces activation of Bni1 and of Fus3 itself, which is stimulated in a Cdc42-dependent manner.

The CN signaling network is essential for survival of yeast cells during prolonged exposure to high concentrations of pheromone and acts through multiple substrates to down-regulate signaling. Previous studies established that CN dephosphorylates and activates Dig2, an inhibitor of Ste12, to reduce mating-induced gene expression (Figure 5B). CN reverses Fus3-mediated phosphorylation of Dig2, and directly competes with Fus3 at overlapping CN- and MAPK-docking sites for Dig2 binding (Goldman ). We speculate that CN signaling substantially decreases mating-induced gene expression through its combined regulation of Rga2 and Dig2. Furthermore, CN reduces signaling by activating Rod1/Art4-mediated receptor endocytosis (Alvaro , 2016) and may dampen other aspects of pheromone signaling through additional substrates. Candidates include the Elm1 kinase, which phosphorylates Gα (GPA1; Clement ), proteins associated with polarized growth (Kic1, Boi2, Mlf3; Goldman ), and the Crz1 transcription factor, which is activated by CN in response to pheromone (Stathopoulos ). Given these multiple points of regulation, it is remarkable that disrupting CN-dependent regulation of a single target, Rga2, in cells that contain multiple Cdc42 GAPs and functional CN substrates noticeably increased pheromone-dependent gene expression.

Although pheromone-induced Ca2+ influx is well documented (Iida ; Moser ; Withee ), new studies of Ca2+ dynamics in individual yeast cells revealed that pheromone induces asynchronous, heterogeneous Ca2+ bursts whose frequency increases with pheromone concentration (Carbó ). Consistent with these findings, we observed partial dephosphorylation of Rga2 throughout the pheromone response. Furthermore, we identified increased expression of a mating gene in Rga2AQALAA -expressing cells specifically under maximal pheromone signaling conditions, when Ca2+ signaling is most intense. The Ca2+/CN-dependent regulation of pheromone signaling is known only to promote cell survival under conditions of prolonged cell cycle arrest (Withee ). However, in many cell types, such as fungal hyphae, pollen tubes, neuronal axons, and mast cells, localized Ca2+ gradients regulate Rho GTPase signaling to direct polarized growth and migration (Brand ; Sutherland ; Chen ; Holowka ). Thus Ca2+/calcineurin-dependent regulation of Cdc42 signaling may play undiscovered roles in directional aspects of pheromone signaling, such as chemotrophic growth (McClure ; Ismael ).

Are other Cdc42 GAPs, Rga1 and Bem3, modulated during pheromone signaling? Single deletion or mutation of these GAPs has little or no effect on the pheromone response (Stevenson ; Bidlingmaier and Snyder, 2004). Furthermore, Rga2 is the only Cdc42 GAP that harbors a CN-docking site (Goldman ) and forms a complex with Cdc24, Bem1, Boi1, and Boi2 (another CN substrate). Proposed to regulate Cdk1-dependent expansion of the yeast cell surface during the cell cycle (McCusker ), this complex might also function during pheromone signaling, as most of its components contribute to shmoo formation or localize to the shmoo tip (Chenevert , 1994; Nern and Arkowitz, 1999; Narayanaswamy ). Further analyses are required to determine whether Fus3 phosphorylates additional GAPs and examine possible regulation of Rga2 by other MAPKs, including Hog1, whose activation is regulated by Cdc42 during the yeast response to hyperosmotic stress, a condition under which CN is also active (Chen and Thorner, 2007; Guiney ).

Regulation of Rho GAP activity is a fundamental mechanism to spatially and temporally regulate the dynamics of Rho GTPase signaling. In a fungal human pathogen, Candida albicans, phosphorylation and inhibition of Rga2 by a hyphal-specific Cdk are required for the morphogenetic switch from yeast to hyphal growth, which promotes virulence (Zheng ). Unidentified phosphatases may therefore prevent pathogenesis by maintaining the fungi in the yeast growth phase. In humans, epidermal growth factor promotes both phosphorylation and subsequent dephosphorylation of a Rho GAP, Deleted in Liver Cancer (DLC1), and disruption of this regulation increases cell spreading, which is linked to cell invasion and metastasis (Ravi ). Thus elucidating how and when Rho GAPs are controlled in vivo is a critical step toward unraveling the mechanisms that perturb Rho GTPase signaling to cause disease.

MATERIALS AND METHODS

Growth media and general methods

Yeast media and culturing methods were followed, as described, in synthetic complete medium (Sherman, 1991), except that twice the amount of amino acids and nucleotides was used. Yeast transformations were performed by the lithium acetate method (Ausubel ). Plasmids and yeast strains used in this study are listed in Supplemental Tables S1 and S2, respectively.

E. coli extract

For GST-peptide binding assays, plasmids for GST-Rga2 PxIxIT peptide (pNL118 and 120) or GST alone (pGEX4T-3) were transformed into BL21 DE3 cells (Invitrogen). Single colonies were grown in Luria broth supplemented with 50 µg/ml ampicillin overnight with shaking at 37°C. Overnight cultures were diluted to OD600 = 0.01 and grown with shaking at 37°C until OD600 = 0.5. Protein expression was induced with 1 mM isopropyl-β-d-thiogalactoside (IPTG) for 2 h with shaking at 37°C. Cultures were harvested by centrifugation (5 min, 5000 rpm at 4°C in a Sorvall GS-3 rotor). Pellets were washed in cold water and stored at −80°C. Pellets were resuspended in lysis buffer (50 mM Tris-HCl, pH 8, 100 mM NaCl, 2 mM ethylene glycol tetraacetic acid [EGTA], 2 mM EDTA, 5 mM dithiothreitol [DTT], protease inhibitors) and lysed by sonication 5 × 30 s at 30% output with 1-min breaks. Cold NaCl was immediately added to lysate to a final concentration of 1.5 M. Cell debris was pelleted (5 min, 10,000 rpm at 4°C in a Sorvall SS-34 rotor) and further clarified twice (20 min, 17,000 rpm at 4°C in a Sorvall SS-34 rotor). Tween-20 was added to clarified extracts to 0.1% final. Total protein concentration of extracts was determined by Bradford assay, and extracts were aliquoted and flash-frozen in liquid nitrogen and stored at −80°C.

For Rga2 full-length binding assays, His-Rga2 constructs (pNL143, 145,146) were transformed in BL21 DE3 cells (Invitrogen), and GST-CNA1 trunc + CNB1 was transformed in BL21 DE3 pLys cells (Invitrogen). Single colonies were grown in Luria broth supplemented with 50 µg/ml ampicillin overnight with shaking at 37°C. Overnight cultures were backdiluted to OD600 = 0.01 and grown with shaking at 37°C until OD600 = 0.5. Cultures were grown for an additional 1 h at 25°C and then induced with 1 mM IPTG at 18°C for 20 h. Cultures were harvested by centrifugation (5 min, 5000 rpm at 4°C using a Sorvall GS-3 rotor), flash-frozen in liquid nitrogen, and stored at −80°C. His-Rga2 pellets were lysed in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10% glycerol, 1 mM β-mercaptoethanol [β-ME], protease inhibitors) and sonicated 5 × 30 s at 30% output with 1-min breaks. GST-CNA1 and CNB1 pellets were lysed in lysis buffer (50 mM Tris-HCl, pH 8, 100 mM NaCl, 5 mM MgCl2, 1 mM β-ME, and protease inhibitors) and sonicated 5 × 30 s at 30% output with 1-min breaks. Cell debris was pelleted (5 min, 10,000 rpm at 4°C in a Sorvall SS-34 rotor) and then clarified twice (20 min, 17,000 rpm at 4°C in a Sorvall SS-34 rotor). Total protein concentration of extracts was determined by Bradford assay, and extracts were aliquoted and flash-frozen in liquid nitrogen and stored at −80°C.

Purification of constitutively active truncated yeast CN

Plasmids for 6xHis-CN (CNA1trunc-p11) and Cnb1 (pET15b-Cnb1) were transformed into BL21 DE3 cells (Invitrogen). Single colonies were grown overnight in Luria broth supplemented with 50 µg/ml ampicillin with shaking at 37°C. Overnight cultures were diluted to OD600 = 0.01 in Luria broth supplemented with 100 µg/ml carbenicillin and 100 µg/ml kanamycin and grown with shaking at 37°C until OD600 = 0.5. Culture was grown for an additional 1 h with shaking at 25°C and then protein expression induced with 1 mM IPTG with shaking at 18°C for 20 h. Cultures were harvested by centrifugation (5 min, 5000 rpm at 4°C with a Sorvall GS-3 rotor), and pellets were washed in cold water and frozen at −80°C. Pellets were resuspended in lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM β-ME, and protease inhibitors) and lysed by sonication: 5 × 30 s at 30% output with 1-min breaks. Cell debris was pelleted (5 min, 10,000 rpm at 4°C with a Sorvall SS-34 rotor) and further clarified twice (20 min, 17,000 rpm at 4°C with a Sorvall SS-34 rotor). CN complex was purified over a nickel–nitriloacetic acid (Ni-NTA) agarose column (FP2400000; 5 Prime) according to manufacturer’s specifications. Purified protein was eluted in fractions with 150 mM imidazole and assayed by SDS–PAGE, pooled, and dialyzed into 50 mM Tris-HCl, pH 7.5, plus 100 mM NaCl buffer and stored at −80°C with 10% glycerol.

GST-peptide and CN trunc binding assay

To test the putative CN-docking site on Rga2, 50 µg of E. coli extract containing GST or GST-tagged peptide was mixed with 10 µg of purified 6xHis-CN trunc in binding buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Triton X-100, 1 mM β-ME, 5 mM imidazole) for 1 h at 4°C with end-over-end mixing. A 25-µl amount of washed Ni-NTA agarose beads (FP2400000; 5 Prime) was added to the binding mixture and incubated for another 2 h. Beads were washed three times in binding buffer containing 15 mM imidazole, and bound proteins were eluted by boiling in Laemmli buffer for 5 min. From 5 to 20 µl of sample was resolved by SDS–PAGE and immunoblotted with primary antibodies to GST (MMS112R500, mouse monoclonal; Covance/Berkeley Antibody Co.) and histidine (34660, mouse monoclonal; Qiagen). Primary antibodies were detected on film by chemiluminescence with secondary antibody α-mouse conjugated to horseradish peroxidase (NA931-1ML; GE Healthcare).

His-Rga2 and CN trunc binding assay

To test whether the CN-docking site is functional within the context of the full-length protein, 2.5 mg of E. coli extract containing 6xHis-tagged Rga2 was mixed with 30 µg of GST extract or 450 µg of GST-CNA1 trunc extract in binding buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Triton X-100, 1 mM β-ME, and protease inhibitors) for 1 h at 4°C with end-over-end mixing. A 25-µl amount of washed glutathione–Sepharose 4B beads (17075601; GE Healthcare) was added to the binding mixture and incubated for another 2 h. Beads were washed three times in binding buffer, and bound proteins were eluted by boiling in Laemmli buffer for 5 min. A 20-µl amount of sample was resolved by 10% acrylamide SDS–PAGE gel and immunoblotted with primary antibodies to GST (MMS112R500, mouse monoclonal; Covance) and histidine (34660, mouse monoclonal; Qiagen). Primary antibodies were detected by infrared (IR) fluorescence with secondary antibody anti-mouse conjugated to Alexa 680 (A-21058; Invitrogen) and visualized with an Odyssey scanner (Li-Cor Biosciences, Lincoln, NE).

Yeast extracts

For in vivo dephosphorylation under Ca2+-activating conditions, cells at OD600 = 0.6 were treated with FK506 (1 μg/ml final concentration) or vehicle (90% ethanol and 10% Tween-20) for 30 min before protein induction with 2% galactose for 4 h. CN was then activated with 200 mM CaCl2 for 20 min and then harvested with pellets stored at −80°C. Yeast pellets were resuspended in RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 1 mM DTT, and protease inhibitors) and homogenized by glass-bead lysis on a vortexer, 5 × 2 min with 1-min breaks. Lysate was separated from beads by needle-puncture spin collection and further clarified at 14,000 rpm for 20 min. Total protein concentration of extracts was determined by Bradford assay. A 45-µg amount of total protein was analyzed by SDS–PAGE gel and immunoblotted with primary antibodies: anti-FLAG (F3165, mouse monoclonal; Sigma-Aldrich).

For experiments probing protein expression or phosphorylation state from yeast extracts, cultures were pelleted and washed with cold water before flash freezing in liquid nitrogen. Pellets were resuspended in urea lysis buffer (20 mM Tris-HCl, pH 7.5, 7 M urea, 2 M thiourea, 65 mM 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate, 65 mM DTT, 50 mM NaF, 100 mM β-glycerophosphate, 1 mM NaVO3, 1 mM phenylmethylsulfonyl fluoride) and homogenized by glass-bead lysis on a vortexer, 5 × 2 min, with 1-min break in between. Lysate was separated from beads by needle-puncture spin collection, and lysate was further clarified at 14,000 rpm for 10 min. Total protein concentration of extracts was determined by Bradford assay, and 30 µg of total protein was loaded to detect plasmid-­expressed tagged-Rga2. For λ-phosphatase assay, yeast pellets were resuspended in RIPA buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 1 mM DTT, and protease inhibitors) and homogenized by glass-bead lysis on a vortexer, 5 × 2 min with 1-min breaks. Lysate was separated from beads by needle-puncture spin collection and further clarified at 14,000 rpm for 20 min. Total protein concentration of extracts was determined by Bradford assay. Samples were analyzed by 5.5% acrylamide SDS–PAGE gel and immunoblotted with primary antibodies: anti-FLAG (sc-807, rabbit polyclonal, Santa Cruz Biotechnology; or F3165-1MG, mouse monoclonal, Sigma-Aldrich), anti-p44/42 MAPK (9101S, rabbit polyclonal; Cell Signaling Technologies), anti-Fus3 (sc-28548, rabbit polyclonal; Santa Cruz Biotechnology), and loading control, anti-Hsc82 (ab30920, rabbit polyclonal; Abcam). Primary antibodies were detected by IR fluorescence with secondary antibody anti-rabbit conjugated to Alexa 790 (A11369; Invitrogen) or anti-mouse conjugated to Alexa 680 (A-21058; Invitrogen) and visualized with an Odyssey scanner.

β-Galactosidase assay

YGS57 transformed with plasmids (pBA2059, pNL110-112) was grown in SC-LEU plus 2% raffinose to mid log phase, OD600 = 0.4. Rga2 expression was induced by adding 2% final galactose for 2 h. Cells were then backdiluted to OD600 = 0.5 and treated for 90 min with 5 µM α-factor (αF; RP01002; GenScript). A 100-µl amount of pheromone-treated cells was transferred to 96-well plates for FUS1-lacZ analysis, and 5-ml cultures were saved to assay Rga2 protein expression in yeast extracts as described. FUS1-lacZ activity was determined using substrate FDG (F-1179; Invitrogen) as previously described (Hoffman ), and fluorescence emission at 530 nm was detected using a fluorescence plate reader (BioTEK, Winooski, VT). For FUS1-lacZ in the integrated strain, cells (NLY52 and 54) were grown in SCD to mid log phase, OD600 = 0.3, and treated for 90 min at indicated αF conditions. For Figure 4B, 15 or 16 biological samples over the course of 2 d for each Rga2WT and Rga2AQALAA were assayed for FUS1-lacZ activity. FUS1-lacZ activity was determined using as substrate fluorescein di-β-d-galactopyranoside and normalized to OD600.

Yeast liquid growth assay

Yeast strain (NLY33) transformed with plasmids (pNL126-128, pNL135, or pEGKT) was grown in SCD-URA to mid log phase. Cells were then backdiluted to a starting OD600 = 0.01 in a 96-well plate and protein expression induced with 20 µM estradiol (E1024-1G; Sigma-Aldrich). Absorbance at 600 nm was read every 5 min for 24 h, and a growth rate of doubling/hour was calculated from at least three technical replicates. Data were analyzed by one-way analysis of variance (ANOVA) with Tukey’s correction for multiple comparisons.

λ-Phosphatase treatment of Rga2 WT and PxIxIT mutant

Cells (NLY33) transformed with plasmids (pBA2059 or pNL111) were grown in SCD-LEU plus 5 nM estradiol to mid log phase. At OD600 = 0.35, cells were arrested in G1 with 10 µM αF for 2 h and then harvested. Yeast extracts were made in RIPA buffer as detailed earlier with 500 µg of total protein used to purify Rga2-FLAG by anti-FLAG immunoprecipitation. Protein-bound beads were washed with RIPA buffer and then washed into PMP buffer (50 mM 4-(2-­hydroxyethyl)-1-piperazineethanesulfonic acid [HEPES], 10 mM NaCl, 2 mM DTT, and 0.01% Brij 25). Beads were then incubated with 100–400 U of λ-protein phosphatase (P0753S; New England BioLabs) in PMP buffer supplemented with 1 mM MgCl2 for 45 min at 30°C. For λ-phosphatase treatment in the presence of phosphatase inhibitors, a cocktail of sodium pyrophosphate (10 mM), sodium fluoride (10 mM), sodium metavanadate (0.4 mM), sodium orthovanadate (0.4 mM), and 3-glycerol-phosphate (0.1 mM) was used. Proteins were eluted from beads by boiling in Laemmli buffer for 5 min. Samples were resolved by electrophoretic mobility in 5.5% acrylamide SDS–PAGE gel, and phosphorylation was assayed by immunoblot (anti-FLAG, rabbit) and Hsc82 loading control (anti-Hsc82). Primary antibody was detected by IR fluorescence with secondary antibody anti-rabbit conjugated to Alexa 790 (A11369; Invitrogen) and visualized with an Odyssey scanner. Percentage of total Rga2 signal represented by slower-migrating band was quantified by Image Studio (Li-Cor Biosciences).

In vivo kinase inhibition

Fus3-WT (NLY39+MR4937) or Fus3-as1 (NLY39+MR4938) was transformed with Rga2WT-FLAG (pBA2059). Strains were grown in SCD-URA-LEU plus 5 nM estradiol at 30°C to OD600 = 0.3 before addition of 10 µM αF pheromone for 2 h, followed by addition of 10 µM 1Na-PP1 (10954; Cayman Chemical Company) or vehicle (dimethyl sulfoxide) for 10 min. Cells were harvested and processed as described in Hughes Hallett , and 30 µg of total protein was loaded to detect electrophoretic mobility by immunoblot. Samples were resolved by electrophoretic mobility in 5.5% acrylamide SDS–PAGE gel, and phosphorylation was assayed by immunoblot (anti-FLAG, mouse) and Hsc82 loading control (anti-Hsc82). Primary antibody was detected by IR fluorescence with secondary antibody anti-rabbit conjugated to Alexa 790 (A11369; Invitrogen) and visualized with an Odyssey. Percentage of total Rga2 signal represented by the slower-migrating band was quantified by Image Studio.

In vitro kinase assay

GST-Fus3 kinase was purified and assayed for in vitro activity as described in Parnell . Briefly, BL21 DE3 cells expressing pGEX-2T6-FUS3 or pGEX-2T6-FUS3 K42R were purified by glutathione–Sepharose (17075601, GE Healthcare). Rga2 constructs were purified by Ni-NTA in buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10% glycerol, 1 mM β-ME, protease inhibitors) from E. coli extracts prepared as described earlier. Approximately 100 nM GST-Fus3WT or GST-Fus3K42R was activated with 300 µM ATP in kinase buffer (25 mM HEPES, pH 7.2, 15 mM MgCl2, 5 mM EGTA, 1 mM MnCl2, 3.5 mM DTT, and protease inhibitors) before being incubated with 450 nM 6xHis-Rga2, 6xHis-Rga2AQALAA, or 6xHis-Rga28A supplemented with 5 µCi of [γ-32P]ATP (PerkinElmer Cetus). Reaction was carried out at room temperature. Samples were collected at indicated time points and reaction terminated with addition of Laemmli buffer and boiling for 5 min. Samples were resolved in 7.5% SDS–PAGE, stained with Instant Blue Stain (Expedeon Protein Solutions), dried, and exposed to phosphoimager screen. The 32P incorporation was detected with Typhoon scanner, and storage phosphor count signal was quantified using ImageJ (Schneider ) and normalized to Coomassie signal.

Pheromone time course

Cells (NLY33) transformed with plasmids (pBA2059 or pNL111) were grown in SCD-LEU plus 5 nM estradiol to mid log phase. At OD600 = 0.35, cells were treated with 0.2 M HU for 3 h and then washed and released into SCD-LEU plus 5 nM estradiol at OD600 ≈ 0.5 and allowed to recover for 20 min before 10 µM αF was added at time 0. The 10-ml time points were taken every 15 min for fluorescence-activated cell sorting (FACS) analysis, microscopy, and electrophoretic mobility from yeast extracts prepared as described. A 35-µg amount of total protein was analyzed for electrophoretic mobility in 5.5% acrylamide SDS–PAGE gel, and phosphorylation was assayed by immunoblot (anti-FLAG, rabbit) and Hsc82 loading control (anti-Hsc82). Primary antibody was detected by IR fluorescence with secondary antibody anti-rabbit conjugated to Alexa 790 (A11369; Invitrogen) and visualized with an Odyssey scanner. Percentage of total Rga2 signal represented by the slower-migrating band was quantified by Image Studio.

Samples for Fus3 activation (phospho-Fus3) blot were resolved on 12% acrylamide SDS–PAGE gel and immunoblotted with primary antibodies anti-p44/42 MAPK and anti-Fus3. Primary antibodies were detected by IR fluorescence with secondary antibody anti-rabbit conjugated to Alexa 790 (A11369; Invitrogen) and visualized with an Odyssey scanner. Percentage of total Rga2 signal represented by the slower-migrating band was quantified by Image Studio.

FACS

A 1-ml amount of harvested yeast cells was fixed in 70% ethanol overnight at 4°C. Cells were washed in 50 mM sodium citrate, pH 7.4, sonicated briefly, and treated with 0.25 mg/ml RNase A for 2 h at 50°C and then 0.8 mg/ml Proteinase K for another 2 h at 50°C. Cells were washed in 50 mM sodium citrate before DNA was stained with 16 µg/ml propidium iodide in sodium citrate and analyzed on a BD FACS Calibur (BD Biosciences, San Jose, CA).

Microscopy

A 1-ml amount of harvested yeast cells was fixed in 70% ethanol overnight at 4°C. Cells were washed in 50 mM sodium citrate, pH 7.4, sonicated briefly, and stained with 24 mg/ml fluorescein isothiocyanate (FITC)–concanavalin A (C7642-2MG; Sigma-Aldrich) in P-buffer (10 mM sodium phosphate, 150 mM NaCl, pH 7.2) at room temperature for 5 min. Cells were washed in P-buffer and sonicated briefly before imaging with Zeiss Axio Imager M1 microscope (Carl Zeiss, Jena, Germany) with an LED lamp with a 100×/1.30 numerical aperture oil immersion objective. A 474/40-nm excitation and a 530/50-nm emission filter were used for FITC–concanavalin A imaging. Images were captured on a Hamamatsu Orca-ER digital camera (Bridgewater, NJ) coupled to Openlab Software 4.0.4 (PerkinElmer-Cetus).