ELECTRICAL CONTROL OF CELL POLARIZATION IN THE FISSION YEAST SCHIZOSACCHAROMYCES POMBE

Abstract

Electric signals surround tissues and cells and have been proposed to participate in directing cell polarity in processes such as development, wound healing and host invasion [1, 2]. The application of exogenous electric fields (EF) can direct cell polarization in cell types ranging from bacteria and fungi to neurons and neutrophils [3–7]. The mechanisms by which EFs modulate cell polarity however remain poorly understood. Here, we introduce the fission yeast Schizosaccharomyces pombe as a model organism to elucidate the mechanisms underlying this process. In these rod-shaped cells, an exogenous EF reorients cell growth in a direction orthogonal to the field, producing cells with a bent morphology. A candidate genetic screen identifies conserved factors involved in this process: an integral membrane proton ATPase pma1p that regulates intracellular pH, the small GTPase cdc42p, and the formin for3p that assembles actin cable structures. Interestingly, mutants in these genes still respond to the EF, but orient in a different direction, towards the anode. In addition, the EF also causes electrophoretic movement of membrane proteins such as the bgs cell wall synthase complex towards the anode. These data suggest molecular models for how the EF reorients cell polarization by modulating intracellular pH and steering the activity or localization of cell polarity factors in multiple directions.

RESULTS AND DISCUSSION

Endogenous electrical signals are present around the cells in the body and yet their possible contribution in globally organizing spatial aspects of cell polarization remain poorly appreciated [6, 8]. Such electrical fields have been measured across epithelial layers and have been proposed to guide cellular behavior in wound healing and development. Application of exogenous electric fields has potential clinical value, for instance in wound healing and tissue engineering. Even single cells can also organize their own local ion currents, through local asymmetries in ion influxes/effluxes, which may contribute to establishing or maintaining a polarity axis [9–13]. Fungi may encounter electric fields in soil, at the surface of plants and in host epithelium [6, 10], and thus such electrical signals could guide pathogenic processes such as host invasion [14]. It has been observed for decades that the exogenous application of an electric field (EF), on the same order of magnitude as those measured in vivo, can direct cell polarization, migration and division in diverse cell types [3, 4, 6, 7, 15]. One puzzling finding is that an EF directs cell polarization in different directions in different cell types: some cells polarize towards the cathode, the anode or even to a direction perpendicular to the field. Research on EF effects would be greatly aided by the development of genetically tractable models to study molecular mechanism. Here, we introduce the fission yeast Schizosaccharomyces pombe, whose cell polarization mechanisms have been studied extensively [16], as a model organism to study these EF effects.

Fission yeast cells reorient their growth axis perpendicular to an exogenous EF

To apply exogenous EFs to yeast cells, we placed the cells in microfluidic channels which allow for defined field lines, heat control and constant exchange of media (Figure S1). Wildtype S. pombe cells normally exhibit a straight-rod morphology and grow by tip extension. Application of a DC (direct current) EF of 50 V/cm caused about half of the cells to grow into a bent-cell morphology. No significant effects were seen on the growth rate, cell cycle periods, or cell stress pathways, suggesting that the physiology of the cells are not grossly perturbed (Figure S1). Time-lapse imaging revealed that this bent morphology was a result of new tip growth in a direction perpendicular to the field; the pre-existing part of the cell wall was not altered, and the cells do not otherwise move (Figure 1A and 1B and Movie S1). This effect was apparent 60–90 min after application of the field and showed a dose-dependent relationship (Figure S1). Although the absolute field strength is higher than those used to study much larger mammalian cells, the intensity of the EF required was similar in terms of trans-membrane potential (few mV across the cell) that takes into account the small size of these yeast cells. This reorientation to the perpendicular axis occurred regardless of the initial orientation of the cell, but was observed most often at those tips initially facing the anode (Figure 1C–E). In a subset of cells, both tips bent toward the perpendicular axis, giving rise to a S-shape cell (Figure 1B). Inverting the EF periodically every 10 min also led to a similar perpendicular reorientation (Figure S1).

Figure 1. An electric field induces a perpendicular reorientation of fission yeast cell growth.

A. Time-lapse images of a WT fission yeast cell growing under a DC EF of 50V/cm. White and blue arrows represent the initial and final direction of polarized growth, respectively. Note that the new axis has reoriented in an axis perpendicular to the EF. B. (Left) Example of two daughter cells with a bent morphology after growth in an EF. (Right) A cell that developed two bent tips after growth in an EF C. Cells at different initial angles (α₀) grown in an EF, and corresponding percentage of bent cells as a function of α₀ for cells grown 90 min under no EF and a DC EF (n > 100 cells for each condition). D. Schematic representation of the important parameters used to quantify growth reorientation in the EF: α₀ is the initial angle of tip elongation with the EF and α₁ is the final angle of tip elongation with the EF. The reorientation factor, RF, is computed as RF = sin(α₁ − α₀).sign(tan(α₁)). The RF is 0 if the cell keeps growing straight and returns a number between 0 and 1 when the cell reorients perpendicular to the EF and a number between 0 and -1 when the cell reorients toward the anode. E. Reorientation factor as a function of α₀ under a DC EF. Each point represents one cell. The ideal RF is calculated by setting α₁=90° for 0<α₀<180° and α₁=−90° for 180<α₀<360°. On the population scale, WT cells have a reproducible RF of 0.20 ± 0.03 (n>100). F. Time-lapse imaging of a tea1Δ cell recovering from starvation in a DC EF of 50 V/cm and forming a branch along the perpendicular axis. G. Percentage of straight, bent and branched tea1Δ cells recovering from starvation in no EF and a DC EF. H. Radial histogram of polarized growth direction, α₁ of tea1Δ cells recovering from starvation in no EF and a DC EF.

We also found that the EF could direct the emergence of a new site of polarization. tea1Δ mutant cells often form a new site of growth on the side of the cell, to form a branch upon recovery from starvation [17]. In the presence of an EF, almost all tea1Δ cells initiated growth along the perpendicular axis of the EF (either along the cell side or at cell tip, depending on cell orientation, Figure 1F–H). Thus, the EF affects both cell polarity establishment and maintenance.

The formin for3p and cdc42p are involved in the EF response

To begin to dissect the molecular mechanism underlying this response, we screened a set of well characterized factors that affect cell polarity. A cdc25-22 mutant, which exhibits a cell cycle delay in G2 and produces elongated cells, exhibited similar EF response as WT, showing that the response is not specific to a G1 phase and is independent of cell size (Figure S2). Microtubules were not required for this redirection, as treatment with a microtubule-inhibitory drug methyl-benzidazole-carbamate (MBC), did not alter the response. In untreated cells, EF did not cause obvious changes in microtubule dynamics (data not shown) [18, 19]. In addition, mutants in microtubule-based polarity pathways defined by tea1p, tea4p, pom1p, mal3p and moe1p all showed a normal EF response [16, 18, 20].

Our screen identified the role of the formin for3p and the small GTPase cdc42p in the EF response. Formins are conserved actin nucleating proteins that are commonly regulated by small Rho-type GTPases [21]. In fission yeast, for3p drives the assembly of actin cables from the cell tips and contributes to cell polarity regulation. for3Δ cells lack actin cables but are viable and still polarize to some degree [22]. Interestingly, for3Δ mutants still responded to the EF, but reoriented in the wrong direction: towards the anode (Figure 2A, 2B, 2C). Anodal responses were also seen in a for3 mutant allele defective in actin assembly and binding (FH2 domain, for3-I930A, [23]) and one defective in localization and actin cable organization (for3FH3Δ) [24] (Figure 2B, 2C). For3p is regulated in part by the actin-binding factor bud6p and the small GTPase cdc42p [24]. Bud6p was not required for normal EF response (Figure S2). However, a cdc42 allele (cdc42-1625), which fails to activate for3p, also reoriented to the anode (Figure 2B, 2C and Movie S2) [24].

Figure 2. The formin for3p and the small GTPase cdc42p are required for directionality of the EF response.

A. Time-lapse images of for3Δ (formin) mutant cells growing under a DC EF. Note that the EF causes a reorientation of the growth axis towards the anode. B. Images of for3 and cdc42 mutants after approximately 2h under a DC EF. C. Average reorientation factor after approximately 2–3h of growth in the absence and in the presence of an EF (n>100), for wildtype and the indicated mutants presented in B. A positive RF is representative of a perpendicular reorientation, while a negative RF stands for an anodal reorientation. Error bars represent standard deviations. D. Confocal maximal projection images of cells expressing for3p-3GFP; stained for actin with Alexa-phalloidin, and single focal plane images of cells expressing CRIB-GFP (a marker for active cdc42p). Cells grown in the absence and in the presence of an EF are shown. **P<0.01, Student’s t-test compared with the control. Scale bars, 2 µm.

We next examined whether EFs affect the localization of formin for3p, cdc42p and actin in a wildtype background. For3-3GFP is normally in dots concentrated symmetrically around the growing cell tip [22, 23]. In cells bent in the EF, for3p was still in dots at the cell tip, but now slightly redistributed to the center of the bent tip (Figure 2D). Consistent with this localization of for3p, actin cables appeared to arise symmetrically from the bent cell tip (Figure 2D). GTP-bound cdc42p, visualized by a CRIB domain fused to GFP [25], followed the same behavior (Figure 2D). Importantly, these factors did not depict any obvious movement prior to cell bending. Thus, in the wildtype cell, the redistribution of these factors may not be the primary effect of the EF, but rather is secondary to the redirection of cell growth.

EF response is dependent on the plasma membrane proton ATPase pump pma1p

Because the plasma membrane may act as an insulator, it is thought that EFs do not directly affect cytoplasmic proteins, but rather exert effects on components on the outer leaflet of the membrane or on extra-cellular components [15]. Thus, we sought to identify transmembrane proteins involved in the response of the EF. One candidate class of membrane proteins is ion channels. Calcium is implicated in cell polarization, and has been proposed to be involved in the response of some cell types to EFs, including some fungi [7, 26]. However, a screen of calcium transporter mutants (yam8Δ, cch1Δ, pmc1Δ, vcx1Δ, pmr1Δ, pdt1Δ and cta4Δ [27–29]), treatments for calcium chelation by EDTA, or blocking L-type and V-type calcium channel with Verapamil or Bafilomycin A respectively, did not affect the response to the EF in S. pombe. In addition, the response to the EF was unaffected by blocking sodium transport with amiloride or in mutants of the sodium/proton anti-porter sod2 or sod22 [30] or by inhibiting the major potassium transport system dependent on trk1 and trk2 [31] (Figure S2).

We did identify a transmembrane protein necessary for EF response: the proton ATPase pump pma1p. This is a major proton transport regulator, which is conserved in plants and fungi and is thought to regulate intracellular pH by pumping protons out of the cell [32]. Although pma1 is an essential gene in S. pombe, we studied a viable mutant pma1-1 allele, which is a single point mutation that specifically reduces enzymatic ATPase activity [33, 34]. When grown in an EF, pma1-1 mutant cells oriented to the anode, in a similar manner to for3Δ and cdc42-1625 mutants (Figure 3A and 3B). pma2p, a related protein functionally redundant with pma1p [35], was dispensable for this response (Figure S2).

Figure 3. The plasma membrane proton ATPase pma1p implicates a role of pH in cell polarity and EF response.

A. Polarity reorientation of pma1-1 mutants after approximately 2–3h of growth in the absence and presence of an EF (n>100). Error bars represent standard deviations. B. Images of pma1-1 cells and pma1-1 cells in a cdc25-22 background (at 25°C) exhibiting anodal reorientation in an EF. C. DIC images of pma1-1 cells grown overnight at the indicated temperature. D. Phalloidin staining of F-actin in wildtype and pma1-1 cells. Confocal maximal projection images of wildtype and pma1-1 cells expressing for3p-3GFP; and single focal plane images of wildtype and pma1-1 cells expressing CRIB-GFP (marker for active cdc42p). E. Epifluorescence images of pma1-GFP in wildtype cells with and without EF. The graphs depict a representative plot of pma1-GFP fluorescence as a function of the position around the cell cortex of half a cell (distances are renormalized between 0 and 1). F. Computational simulation of the electric potential (Φ) landscape around a S. pombe cell created by a DC EF of 50V/cm. The cytoplasm is set at an arbitrary homogenous reference potential. The lines represent the equipotentials. G. Proposed model for the effect of EF on a putative intracellular pH gradient and reorientation of growth along the perpendicular axis: (left). In normal cells, pma1p on the cell sides extrudes protons, which leads to a trans-cellular loop of proton fluxes exiting the side and entering the tip. This loop sets a putative intracellular pH gradient, with the growing tip being more acidic than the sides. The local pH value at the cell tip may be optimal for localized activity of formin molecules there. (Right) In the presence of an exogenous EF, the trans-cellular loop of protons is modified by the local changes in transmembrane potential induced by the EF. This leads to acidification at the anode-facing side and alkalinization at the cathode –facing side. The zone of optimal growth pH is displaced to a site along the perpendicular axis, and thus guides growth in this new direction. **P<0.01, Student’s t-test compared with the control. Scale bars, 2 µm.

Role of pma1 in cell polarity

Although pma1p has been characterized for effects on transport, its role in cell polarity has not been well described in S. pombe. We found that even in the absence of the EF, pma1-1 mutants displayed significant defects in cell polarity regulation. At 25°C, the cells were fatter and rounder than WT, with occasional presence of ectopic bulges. Time-lapse imaging revealed abnormal growth patterns, similar to those of for3 mutants (Figure S3). At higher temperatures severe morphological defects were seen (Figure 3C). pma1-1 mutants also exhibited defective actin organization: actin cables were fainter and disorganized, and were possibly missing in a subset of cells, while actin patches were more numerous, smaller and less concentrated at cell tips (Figure 3D). For3-3GFP dots were less concentrated at cell tips (Figure 3D and Figure S3), similar to those observed in cdc42-1625 mutants [24]. However, GTP-bound cdc42p, as visualized by CRIB-GFP, was still mostly localized at cell tips in a pma1-1 mutant. Together, these data suggest that pma1p participates in cell polarization and affects for3p localization and function at a step downstream of cdc42p activation.

Next, we examined the localization of pma1p. Immunofluorescence staining of wildtype cells with anti-pma1 antibody showed that pma1p is located on the plasma membrane and concentrated along the sides of the cells. We confirmed this localization by using a functional GFP fusion that is overexpressed from a plasmid under the control of the nmt1 promoter (Figure 3E and Figure S4) [36]. Quantitative analysis of fluorescence intensities showed that pma1-GFP was about two-fold more concentrated along the sides of cells than at cell tips, and showed a peak localization along a cortical band between the center of the cell and the growing tip. Similar distributions were seen at different levels of expression (Figure 3E and Figure S4). Pma1p localization pattern was similar in for3Δ and cdc42-1625 mutants. As shown in pollen tubes [9, 11], the higher concentration of pma1p along the sides of the cells may establish a pH gradient in which the growing cell tip is more acidic than the sides of the cell. The mutant phentoype of pma1-1 suggests that this pH gradient is critical for regulation of cell polarity.

A model of EF effects based upon pH

Because of the involvement of pma1p, we suggest a model in which an EF reorients cell growth by local modulation of pH [9, 11] (see Supplementary Materials). EF treatment did not alter the localization of pma1p. Rather, EFs may directly affect local proton influx by modifying the transmembrane potential (TMP) around the cell [11]. Computational modeling predicted that the TMP is hyperpolarized at the side of the cell facing the anode and depolarized at the cathode facing side. These local differences in potential could cause local changes in cortical pH; the pH is least affected by the EF at a zone facing the perpendicular axis (Figure S4). If we assume that this pH value is at the optimal level for formin activity and cell growth, this model provides one explanation for why cells in an EF grow towards the perpendicular axis (Figure 3G).

Electrophoresis of certain membrane proteins to the anode

Another proposed effect of EFs on cells is through electrophoresis of membrane components [37]. We found that while calcofluor-staining regions of the cell wall are normally located at sites of cell growth and division, the EF caused an asymmetric accumulation of calcofluor-staining cell wall material at a site near cell tips towards the anode, distinct from the growth site (Figure 4A and 4B). This asymmetrical accumulation was specific to calcofluor staining and was not observed by staining the entire wall with a lectin (Figure S5). Increased staining was observed at both the inner and the outer curvature of the bent cell and sometimes at the septum of a dividing cell (Figure S5). Application of an EF periodically inverted every 10 min caused perpendicular bending but no accumulation of cell wall material (Figure 4A). Finally, this effect was mechanistically distinct from the perpendicular growth response: accumulation was still observed in Latrunculin A-treated cells and in for3 mutants, showing that this process was independent of actin and cell growth (Figure 4B).

Figure 4. Electrophoresis of a transmembrane cell wall synthase complex towards the anode.

A. Calcofluor staining of the cell wall in wildtype and for3Δ cells grown under a DC EF, a periodically inverted EF, and no EF for approximately 2h. The staining is applied at the end of these 2h. Yellow arrows highlight the anodal accumulation of calcofluor-staining material. B. Anodal accumulation of the glucan synthase bgs4-RFP and corresponding calcofluor staining in wildtype and for3Δ cells, with and without Latrunculin A treatment. C. Time-lapse sequence illustrating the gradual anodal accumulation of bgs4-RFP in a wildtype cell bending perpendicular to the EF. D. Computational simulation of the electric field landscape around a S. pombe cell created by an average DC EF of 50V/cm. The lines represent the EF lines (along which the EF has the same intensity). The field in the cytoplasm is set to 0 by construction. E. Analytical calculation of the EF at the plasma membrane for different initial orientations of the cell. The orientation and the average EF direction are illustrated in the left inset. Dotted arrows represent the direction that follows the x-axis of the plot. F. Velocity of bgs4p-RFP movements as a function of the initial angle of the cell with the EF axis. The correlation coefficient between the velocity and the initial angle data sets is 0.54, with a t-test P-value of 1.4.10⁻⁸. The dotted line is depicted to guide the eyes. Each point corresponds to one cell, and is computed by tracking bgs4p movement away from the initial growth axis. G. Experimental percentage of bent for3Δ cells as a function of their initial orientation (represented by the corresponding colors) with the EF, the bars are binned on 30° (n>50 for each bar). H. Electrophoresis of glucan-synthase and consequent accumulation of cell wall building material to the anode of the EF may drive anodal polarization when the main mode of EF-driven polarization is turned off (in a pma1-1 or for3Δ cell, for instance).

Calcofluor preferentially stains (1,3)β-D-glucan in the cell wall, which is assembled by a large membrane-associated glucan synthase complex. Two components of this complex, bgs1p (also known as cps1p) and bgs4p, showed that these proteins also accumulated at the calcofluor-staining regions at the anodal side of the cell tips in both wildtype and for3Δ cells under a DC EF (Figure 4B–4C, Figure S5 and Movie S3) [38, 39]. This behavior was not seen with several other plasma membrane proteins, including pma1p, pmd1p and a GFP-CAAX construct that is anchored on the inner plasma membrane [36, 40] (data not shown).

We tested whether these membrane proteins move by direct electrophoresis. FRAP studies showed that bgs proteins are normally stable in the membrane in the absence of EF (our unpublished data). These glucan synthase proteins have large extracellular domains that are negatively charged, and thus could be directly influenced by electrophoretic forces. For instance, bgs4p has a very large predicted N-terminal extra-cellular domain that contains 84 negatively charged and 52 positively charge residues, with a pI of 4.65. An electrophoresis mechanism predicts that the rate of bgs4p movements should correlate with EF intensity, which varies with the initial orientation of the cell. The predicted EF values around the cell membrane are maximal at the tip, and this maximum increases with the initial angle of the cell with the EF axis (Figure 4D–4E). The velocity of bgs4-RFP movements around the tip was found to increase with initial orientation angles (Figure 4F and Figure S5), supporting the view of an electrophoretically driven movement.

These movements were significantly faster in for3Δ cells over wildtype cells (Figure S5), suggesting that for3p-dependent processes may normally impede the movement of bgs4p. The proportion of for3Δcells bent by the EF also increased with the initial angle with the EF (Figure 4G), showing that faster bgs movements correlate with a higher probability of cell bending in these mutant cells.

The accumulation of the bgs complex itself may provide an explanation why for3, cdc42 or pma1 cells polarize towards the anode (Figure 4H). Consistent with this view, bgs1p is critical for polarized cell growth in a for3Δ background (M. Ramos and J. Ribas, personal communication). Thus, although the bgs proteins and for3p– dependent processes are normally spatially coupled through actin-cable dependent membrane transport, the application of the EF produces a curious scenario in which these sets of proteins are spatially uncoupled and appear to compete with each other to establish a growth axis. This observation challenges a prevalent model of fungal cell growth in which cdc42p and formin-dependent actin cables guide the remodeling of the cell wall by targeted delivery of glucan synthases to the cell tip. Rather, these findings suggest that growth is not solely guided by the accumulation of these bgs proteins. These studies illustrate how the exogenous control of protein behavior with EFs can bring new insights into polarity mechanisms.

In summary, these are initial studies in fission yeast to define molecular mechanisms of electric field effects on cell polarization. Identification and characterization of a set of conserved polarity factors in this process highlight the functions of actin, pH regulation and electrophoresis and lead to models postulating how EFs cause cells to polarize in different directions. These results are likely to be generally relevant to other cell types, as they are consistent with recent findings implicating the Rho/Rac pathways and pH regulators in EF response in animal cells [6, 41, 42].