Role of PKD2 in Rheotaxis in Dictyostelium

The sensing of mechanical forces modulates several cellular responses as adhesion, migration and differentiation. Transient elevations of calcium concentration play a key role in the activation of cells following mechanical stress, but it is still unclear how eukaryotic cells convert a mechanical signal into an ion flux. In this study, we used the model organism Dictyostelium discoideum to assess systematically the role of individual calcium channels in mechanosensing. Our results indicate that PKD2 is the major player in the cell response to rheotaxis (i.e., shear-flow induced mechanical motility), while other putative calcium channels play at most minor roles. Mutant pkd2 KO cells lose the ability to orient relative to a shear flow, whereas their ability to move towards a chemoattractant is unaffected. PKD2 is also important for calcium-induced lysosome exocytosis: WT cells show a transient, 2-fold increase in lysosome secretion upon sudden exposure to high levels of extracellular calcium, but pkd2 KO cells do not. In Dictyostelium, PKD2 is specifically localized at the plasma membrane, where it may generate calcium influxes in response to mechanical stress or extracellular calcium changes.


Introduction
Prokaryotic and eukaryotic cells are constantly exposed to mechanical forces, both extracellular (e.g. shear force exerted by a fluid flow, gravity, contact) and intracellular (e.g. changes in osmotic pressure), to which they respond by regulating many cellular processes, such as cell adhesion and migration, mitosis, gene expression, and cell differentiation [1,2]. Mechanosensing involves the perception of mechanical forces and their conversion into intracellular biochemical signals [3,4]. The intracellular response induced by mechanical stress depends largely on ion fluxes, particularly calcium fluxes caused by the opening of calcium channels [5,6]. In bacteria, Msc (for mechanosensing) channels are directly gated by membrane deformations caused by changes in cell osmolarity, and initiate intracellular signaling [7]. However, to date no eukaryotic mechanosensing receptor and/or channel have been unambiguously identified [3,8]. Rather, mechanosensing in eukaryotic cells is thought to involve the regulatory action of protein complexes linking the extracellular matrix (ECM) or the cytoskeleton to ion channels.
Several families of ion channels, mostly non-selective calcium channels from the TRP family [9], have been implicated in mechanosensing in C. elegans, Drosophila and mammals, but it is still not clear if they are directly or indirectly gated by mechanical stress [3,10]. For example, early observations suggested that TRPC6 channel could be directly activated by changes in membrane tension, but recent findings rather indicate that this channel is indirectly activated by the angiotensin II type 1 receptor [11,12]. TRPP2 (also named PKD2 or polycystin-2) is a calcium channel that forms a complex with PKD1, and the PKD1/PKD2 complex has been implicated in intracellular calcium increases in mechanically stressed ciliated cells [13][14][15]. However some studies indicate that the PKD complex may act rather by interacting with the cytoskeleton and regulating an as yet unidentified channel [16,17]. In addition to TRP channels, metazoan candidates for mechanosensitive components include sodium channels of the ENaC family, two-pore domain potassium channels (K2P) and bacterial Msc-like channels [8,18].
The amoeba Dictyostelium discoideum is a model organism easily amenable to genetic analysis, and largely used to study cell migration and chemotaxis, as the core mechanisms involved in motility are largely conserved from amoebae to human cells [19]. Several publications have reported that migration and physiology of Dictyostelium cells are modulated by mechanical stresses induced by a fluid flow, electrical fields or compression [20][21][22][23]. Remarkably, the total number of putative ionic channels is extremely reduced in Dictyostelium compared to other organisms. The Dictyostelium genome contains only three genes encoding putative calcium channels potentially expressed at the cell surface or in endocytic compartments (mcln, pkd2, tpc) as well as one Msclike channel (mscS) [24,25]. In addition, one IP 3 receptor (iplA) is potentially present in the ER, and five P2X receptors (p2xA-E) are restricted to the contractile vacuole [26]. Since P2X receptors are thought to play a specific role in the function of the specialized osmo-regulatory contractive vacuole, they were not considered further in this study. The low number of channels and the relative ease with which specific knockout strains can be generated and analyzed makes Dictyostelium a unique system by allowing a systematic comparative analysis of the role of each channel in mechanosensing.
In this study, we generated specific knockout strains for the mcln, pkd2, tpc, mscS and iplA channels in Dictyostelium and characterized their role in rheotaxis (or shear-flow-induced cell motility). Our results reveal that PKD2 plays a key role in rheotaxis in Dictyostelium amoebae.

Rheotaxis in Dictyostelium
The Dictyostelium discoideum genome exhibits a reduced number of genes encoding proteins potentially involved in mechanotransduction, including some ionic channels (MscS, IplA, PKD2, TRP-ML, and TPC2) and one integrin beta-like protein (SibA) ( Table 1).
To determine the role of these different proteins in mechanotransduction, we first tested the ability of WT and specific KO cell lines for each of these six genes to respond to shear-flow induced stress. For this, Dictyostelium cells were allowed to attach to a glass coverslip and their migratory behavior was assessed before and after the initiation of a uniform fluid flow ( Figure S1 shows a schematic diagram of the flow chamber used).
As reported previously [20], WT cells respond to shear stress by moving in the same direction as the fluid flow ( Figure 1A, and Movies S1 and S2). To quantify this oriented movement, we measured the net displacement of cells on the X axis, parallel to the flow (Dx) ( Figure 1B). In the absence of flow, both WT and KO cell lines migrated randomly (Dx close to 0) and with similar speed (around 2.5 mm/min) ( Figure 1C). When exposed to a constant flow for 10 min (with an applied force equivalent to 4 Pa), WT cells moved at the same speed ( Figure 1C) and oriented in the direction of the flow (from right to left, as indicated by negative Dx values) ( Figure 1B). Of all the KO cell lines analyzed, only pkd2 KO cells showed an almost complete loss of directionality when exposed to a flux ( Figure 1B, and Movies S3 and S4). A WT phenotype was restored when pkd2 KO cells were transfected with an expression vector harboring the PKD2 coding sequence ( Figure 1D). mcln KO cells also showed a significant decrease in their response to shear stress, although not as pronounced as pkd2 KO cells, and iplA KO cells showed a weak and not statistically significant decrease in directionality ( Figure 1B). Orthologs of Msc and TPC2 channels and the beta-integrin-like SibA protein did not appear to be involved in response to mechanical stress in Dictyostelium.
Another way of analyzing the behavior of cells submitted to a shear stress is to determine their directional persistence. When submitted to a fluid flow, the directional persistence of WT cells increased two-fold, and the same happened for all the KO cells ( Figure 1E). However, pkd2 KO cells did not show any increase in persistence when submitted to fluid flow. These results indicate that the PKD2 channel plays a unique role in Dictyostelium mechanosensing, and this led us to further study its structure and localization.

Structure and localization of PKD2
Dictyostelium PKD2 belongs to the TRP (Transient Receptor Potential) family of ion channels, and phylogenetic analysis places it at the base of the Metazoan group ( Figure 2A). It presents the distinctive features of the TRP family: six transmembrane (TM) domains, a conserved pore region between TM5 and TM6 and a large extracellular loop between TM1 and TM2 [27]. In addition, the C-terminal domain contains a conserved coiled-coil region ( Figure 2B), also present in metazoans where it is responsible for interactions with other proteins, notably PKD1, ionic channels (e.g. TRPV4 and TRPC1), and cytoskeleton-related proteins (e.g. a-actinin and troponin) [28][29][30][31][32].
Two distinctive features of the human PKD2, the presence of an EF-hand domain and of a large region ensuring retention in the endoplasmic reticulum [33], are absent from the other PKD2 orthologs analyzed here (C. elegans, D. melanogaster and Dictyostelium) ( Figure 2B). As the localization of the human ortholog is still a matter of debate -PKD2 has been localized to plasma membrane, primary cilia, ER, and Golgi [34,35] -we decided to check where the Dictyostelium PKD2 ortholog was localized.
Protein localization was assessed by immunofluorescence using a Flag-tagged PKD2 construct ( Figure 3). The majority of the protein was present at the plasma membrane, as shown by the extensive co-localization with a plasma membrane marker (H36). No significant co-localization was seen with a marker of late endosomal compartments (p80) or contractile vacuole (Rhesus). The internal structures in which PKD2 can also be detected colocalized partially with recycling endosomes (p25 marker) and with newly formed endosomes (actin-positive).
These observations suggest that in Dictyostelium, PKD2 is mostly localized at the cell surface and in early endocytic compartments. Given the surface localization of Dictyostelium PKD2, it seems reasonable to hypothesize that its major role in the response to mechanical stress is to mediate transient entry of extracellular calcium in response to mechanical signals.

Role of PKD2 in calcium-stimulated lysosome exocytosis
Another cellular function directly linked to transient increases in cytosolic calcium is the secretion of lysosomes. In mammalian cells, lysosome exocytosis may be triggered by several different stimuli that promote rises in cytoplasmic calcium, including a sudden increase in extracellular calcium levels [36][37][38]. In Dictyostelium, secretory lysosomes are highly enriched in the endosomal p80 protein, and their fusion with the plasma   Figure 4A) [39]. In nutrient medium (containing approximately 30 mM calcium [24]), secretory lysosomes fuse constitutively with the cell surface. Consequently, 4.160.2% of WT cells exhibit an exocytic patch, and pkd2 KO cells present a similar phenotype (5.160.5%). When cells were exposed suddenly to a higher extracellular calcium concentration (1 mM), a burst of lysosome fusion was observed in WT cells, as shown by a rapid and transient 2-fold increase in the number of exocytic patches ( Figure 4B). On the contrary, in the same conditions no increase in fusion of lysosomes with the cell surface was observed in pkd2 KO cells ( Figure 4B). Indeed for pkd2 KO cells, the variations over time were not significantly different from the control values at time 0. This result suggests that PKD2 plays a role in calcium-induced lysosome secretion, probably by mediating a rapid influx of extracellular calcium.

PKD2 is not involved in folate chemotaxis
To evaluate if PKD2 was implicated in cell orientation and taxis in a more general manner, we analyzed the ability of vegetative cells to migrate towards folate. Chemotaxis assays were conducted either on an agar surface or in submerged conditions. Chemotaxis on buffered agar was assessed by spotting cells in close proximity to a folate source, and observing the ability of cells to move towards the chemoattractant after 5 hours. As can be seen in Figure 5A, both WT and pkd2 KO cells were able to orientate towards folate: the front of cells moved to the left, where the folate source was located. The distance travelled by both cell types was also the same ( Figure 5B).
Similarly, cells on phosphate buffer subjected to a folate gradient generated by a micropipette were able to move towards the source of folate, as denoted by the increased number of cells near the tip of the micropipette ( Figure 5C). The directional persistence (measured as the sustained cell movement in the n = 4. E) Persistence was measured as the net distance between initial and final cell positions divided by the total distance. Here it is shown the ratio between the persistence when cells migrate randomly and when exposed to a shear flow. Only pkd2 KO cells did not show an increased persistence when submitted to a shear stress. ** p,0.01, compared to WT values; n = 5. doi:10.1371/journal.pone.0088682.g001  Figure 5D). Similarly, the oriented displacement towards the pipette tip was the same in WT and pkd2 KO cells ( Figure 5E). Altogether, these results indicate that the PKD2 channel is not necessary for chemotaxis towards folate in Dictyostelium.

Discussion
In this work, we showed by systematic comparative analysis of KO strains that in Dictyostelium, PKD2 is the most important protein for rheotaxis. Of all mutants analyzed, only pkd2 KO cells were unable to respond to a flow-induced shear stress, and a WT phenotype was restored by complementation with a full-length PKD2. This is the first time that PKD2 has been implicated as a molecular player in mechanotaxis in Dictyostelium. Other potential candidates were also assayed for their role in shear-flow-induced cell motility, notably other calcium channels and orthologs of a bacterial mechanosensing channel (MscS) and of a metazoan integrin-beta (SibA). Of all these, only TRP-ML (or mucolipin) deficiency led to a significant, though limited, reduction in mechanosensing.
Previous studies have assessed the response of Dictyostelium cells after mechanical stresses caused by electric fields [23], compression [22], stretching [40] or a fluid flow [20,21]. In all these studies, depletion of extracellular calcium completely abolished the response to stimuli, suggesting a role for calcium transporters in the process. In addition, gadolinium (Gd 3+ ), a known blocker of plasma membrane calcium channels and stretch-activated channels, also impaired the response to mechanical stress [21,23,40]. Moreover, one of the hallmarks of the response to mechanical stress is an increase in cytosolic calcium, both in mammalian and Dictyostelium cells [6,13,23,40]. However, it is a matter of debate if the calcium originates from the extracellular medium or from the intracellular stores (and, in consequence, if the major contribution for such response comes from intracellular or plasma calcium channels).
In the aforementioned studies, the potential role of the Dictyostelium IP 3 receptor ortholog (iplA) in mechanosensing was assessed. Mammalian IP 3 receptors are implicated in cellular calcium homeostasis by controlling calcium release from ER stores. In Dictyostelium, depletion of the iplA gene did not impair chemotaxis (either towards cAMP or folic acid) [41,42] or the mechanotactic response to electric fields [23] or to flow-induced shear stress [21]. Most of these experiments were performed in the presence of an excess of extracellular calcium, a condition similar to that used in our study. It remains possible that in different conditions, notably when the extracellular calcium concentration is lower, release by IplA of intracellular stores of calcium may play a more critical role in mechanosensing, as suggested previously [21].
In summary, our observations are in agreement with previous results suggesting that mechanotaxis involves primarily a direct transfer of calcium from the extracellular medium to the cytosol. They further suggest that PKD2 may be the main effector of this calcium transport across the plasma membrane by showing that PKD2 is localized primarily at the cell surface in Dictyostelium and is  a key element in mechanosensing. This hypothesis is reinforced by our observation that PKD2 is essential for calcium-induced exocytosis of secretory lysosomes (or post-lysosomes). Indeed, since we observe that calcium-induced lysosome secretion is PKD2-dependent and is maximal two minutes after raising the extracellular calcium concentration, it seems probable that lysosome secretion is caused by a direct transfer of calcium from the extracellular medium to the cytosol through PKD2. Unfortunately, we have been unable to measure cytosolic calcium levels in pkd2 KO cells, either by using fluorimetric and ratiometric probes or with an aequorin genetic system (we were never capable of measuring a signal above background values). So, it remains to be seen if depletion of PKD2 channel really impairs entry of extracellular calcium, after a mechanical stimulus or after addition of extra calcium on the medium.
How does PKD2 open in response to mechanical stress? In mammalian cells, a number of proteins associated to PKD2 have been proposed to play a key role in its activation. In ciliated cells from the kidney and vascular endothelium, the PKD1/PKD2 complex has been implicated in mechanosensing [43,44]. Other results have suggested that this complex does not act as a calcium channel, but rather regulates the function of other potential channels (as TRPV4 and TRPC1), potentially via interactions with cytoskeleton components such as filamin [17,28,32]. Remarkably, in Dictyostelium, PKD1 as well as TRP channels from the C and V families are absent, suggesting that PKD2 can act as a mechanosensor in the absence of other associated membrane proteins, or making use of an entirely different set of interacting partners. PKD2 may even act as a bona fide stretch-activated channel of Dictyostelium, ensuring both detection of the mechanical stress and calcium entry following activation. If new candidates implicated in mechanosensing are identified in various systems, the validity and the generality of these observations may be checked in Dictyostelium by generating the corresponding knockout strains and analyzing their phenotype.

Cells and reagents
The Dictyostelium strains employed here were all derived from the subclone DH1-10 [45] of the DH1 strain, referred to as wildtype (WT) for simplicity. Cells were grown in HL5 medium at 21uC and subcultured twice a week to maintain the cell density below 10 6 cells/ml. Migration experiments were conducted using either phosphate buffer (2 mM Na 2 HPO 4 , 14.7 mM KH 2 PO 4 , pH 6.0), or MES buffer (50 mM MES, pH 6.0) when calcium was added to the medium.
KO vectors for pkd2, mscS, iplA and tpc disruption were constructed using a blasticidin-resistance cassette flanked by two gene segments (Table 1 and Figure S2). The PvuI-digested plasmid was introduced into WT cells by electroporation, transfected cells were selected in the presence of 10 mg/ml blasticidin and individual clones were screened by PCR (Figures S2B and S2C). Three independent KO clones for each gene were used in parallel in this study, with identical phenotypes. The sibA and mcln KO cell lines were described previously [24,46]. iplA KO cell lines using Ax2 [42] and JH10 [47] as parental backgrounds have also been described previously, but were not employed during this study.
A PKD2-Flag expression vector was constructed by introducing a C-terminal Flag epitope (DYKDDDDK) in frame with the PKD2 coding sequence into pDXA-3C [48]. This plasmid was transfected into pkd2 KO cells by electroporation, and transfected cells were selected in the presence of 10 mg/ml G418.

Fluorescence microscopy
Immunofluorescence for PKD2 localization and for quantification of the number of exocytic p80 patches was performed as described previously [24,49]. For measurement of calciuminduced lysosome exocytosis, 10 6 cells were allowed to attach to glass coverslips in HL5-MES medium for 3 hs, then transferred to HL5-MES containing 1 mM CaCl 2 , incubated between 0 and 8 minutes as indicated, fixed with paraformaldehyde 4%, permeabilized with Triton X-100 (0.08%) and labeled with mouse monoclonal antibody anti-p80.

Cell migration under shear-flow stress
For measuring cell motility under flow conditions, the experimental setup was adapted from Decave et al. and Mennesson et al [20,54]. 10 6 Dictyostelium cells were allowed to attach on glass coverslips (24650 mm) for 30 min in MES buffer containing 1 mM CaCl 2 . Coverslips were assembled in a parallel plate laminar flow chamber (Immunetics, Cambridge, MA), and the chamber connected to input and output tanks ( Figure S1). Flow rates were controlled by the differential height between both tanks, and shear stress values were deduced by using the formula s = 6Dg/wh 2 , where D is the flow rate (14 ml/min), g the fluid viscosity (0.001 Pa?s), h the chamber height (250 mm), and w the chamber width (5.5 mm). Cells were subjected to a 4 Pa shear stress and imaged every 15 seconds during 10 min in a phasecontrast, wide-field inverted Zeiss Axiovert 100M, with a Plan-Neofluar 106 objective. The images were acquired with a Hamamatsu CCD cooled camera and assembled into a movie using Metamorph (Molecular Devices, Sunnyvale, CA). Particle tracking application for Metamorph was used to track the individual trajectories and the total distance travelled by at least 15 cells per experiment. Speed was calculated as total distance divided by total time. Persistence was estimated by the ratio of the net distance (from initial to final point) to the total distance. Net displacement on the X axis (Dx) is given by the sum of all displacements on the X axis.

Folate chemotaxis
To evaluate chemotaxis towards folate, two different assays were employed. The first assay was done by depositing 1 ml of 5610 7 cells/ml on a phosphate agar plate, 4 mm away from a folate source (5 mM) and analyzing cell orientation after 5 h [55]. A black mark on the bottom of the petri dish allowed us to align pictures taken at different time points. The travelled distance was calculated by measuring the displacement of the cell front.
For the second assay, cells were incubated overnight in HL5 in the presence of 1 mM folate, washed in phosphate buffer, and allowed to adhere for 15 min in 43 mm petri dishes. A folate gradient was created with a micropipette (pressure of 20 hPa) filled with 250 mM folate (Fiedler et al, submitted), and cells were imaged every 20 seconds for 90 minutes. Cell tracking was done as described above. The distance to the micropipette (d) was measured as the final distance of the cell to the micropipette minus the initial distance to the micropipette (cells moving towards the micropipette show negative Dd values, while cells moving randomly have Dd>0).

Sequence and phylogenetic analysis
Sequence similarity analyses were performed using BlastP program against the protein databases deposited at NCBI server. For phylogenetic analysis, protein sequences were aligned with CLUSTALX 2.0 [56] and maximum likelihood trees were done with MEGA 5.0 (WAG+F model, and parameters for invariable sites and gamma-distributed rate heterogeneity) [57]. One hundred bootstrap replicates were executed and bootstrap values drawn up on the consensus tree.

Statistical analysis
Unless otherwise specified, for quantified data, the values represent the arithmetical mean and s.e.m. (the number of independent experiments is indicated by n). Statistical comparisons were done with student t-tests (two-tailed, unpaired). Movie S1 WT cells moving randomly, without any flow passing through the system. Phase-contrast images were taken every 15 sec, during 10 min. Size: 160695 mm.

(AVI)
Movie S2 WT cells under shear-flow stress (4 Pa, from right to left). Phase-contrast images were taken every 15 sec, during 10 min. Size: 160695 mm.

(AVI)
Movie S3 pkd2 KO cells moving randomly, without any flow passing through the system. Phase-contrast images were taken every 15 sec, during 10 min. Size: 160695 mm.