Calcium responses to external mechanical stimuli in the multicellular stage of Dictyostelium discoideum

Calcium acts as a second messenger to regulate many cellular functions, including cell motility. In Dictyostelium discoideum, the cytosolic calcium level oscillates synchronously, and calcium waves propagate through the cell population during the early stages of development, including aggregation. In the unicellular phase, the calcium response through Piezo channels also functions in mechanosensing. However, calcium dynamics during multicellular morphogenesis are still unclear. Here, live imaging of cytosolic calcium revealed that calcium wave propagation, depending on cAMP relay, disappeared at the onset of multicellular body (slug) formation. Later, other forms of occasional calcium bursts and their propagation were observed in both anterior and posterior regions of migrating slugs. This calcium signaling also occurred in response to mechanical stimuli. Two pathways—calcium release from the endoplasmic reticulum via IP3 receptor and calcium influx from outside the cell—were involved in calcium signals induced by mechanical stimuli. These data suggest that calcium signaling is involved in mechanosensing in both the unicellular and multicellular phases of Dictyostelium development using different molecular mechanisms.

www.nature.com/scientificreports/ development 18,19 ; however, recent studies have shown that the dynamics of cAMP signaling are altered after multicellularity begins 16,20 . In addition to Ca 2+ wave propagation during aggregation 17 , transient [Ca 2+ ] i elevation has been observed in mounds and slugs 21 . These results suggest that synchronous [Ca 2+ ] i bursts and wave propagation occur not only during aggregation but also in later Dictyostelium development, including in the formation of mounds and slugs. Calcium signaling has been suggested to be involved in chemotaxis and stalk cell differentiation [22][23][24][25][26] . However, the dynamics of, and molecular mechanisms underlying [Ca 2+ ] i signaling during the morphogenesis of Dictyostelium remain unclear.
In this study, [Ca 2+ ] i signaling during Dictyostelium development was investigated using fluorescent Ca 2+ probes. This approach revealed the transition of [Ca 2+ ] I signaling dynamics during cell aggregation and the formation of slugs of Dictyostelium, both of which have robust calcium signaling mechanisms in response to mechanical stimuli.

Results
Calcium signaling dynamics change during Dictyostelium development. To investigate the relationship between the dynamics of calcium signals and multicellularity in Dictyostelium, we monitored [Ca 2+ ] i dynamics during development with genetically encoded calcium indicators (GECI). As previously reported 17 , cells expressing the Förster resonance energy transfer (FRET) sensor YC-Nano15 (K d = 15 nM) showed clear oscillations and wave propagation in aggregation streams (Fig. 1a, Supplementary Fig. 1a, and Movie 1). Moreover, [Ca 2+ ] i dynamics were investigated with the single-wavelength GECI GCaMP6s 27,28 to confirm whether the wave propagation observed using YC-Nano15 authentically reflected [Ca 2+ ] i dynamics during development using a different GECI, as well as when avoiding the phototoxicity caused by exposure to violet-blue light excitation for YC-Nano15.
In starved Dictyostelium cells, [Ca 2+ ] i transiently increases in response to external cAMP 22 and the calcium channel, IplA, the homolog of the IP3 receptor, is essential for its elevation 29 . When chemotactic-competent cells expressing GCaMP6s were stimulated by cAMP, wild-type cells showed transient rapid elevations of fluorescence with a peak, 16 s after stimulation; however, cells lacking iplA showed no such elevations after stimuli ( Supplementary Fig. 2a). Thus, GCaMP6s (K d = 144 nM) 27 is functional in Dictyostelium cells and is appropriate for visualizing [Ca 2+ ] i dynamics. Oscillations of fluorescence signals and wave propagation were observed during both early aggregation (before streaming) and mound stages of cells expressing GCaMP6s ( Fig. 1b- . These signal oscillations were not observed in iplA − cells during development ( Supplementary Fig. 2b, c, Movie 7), demonstrating that the periodic changes in GCaMP6s signals in developing Dictyostelium cells reflect [Ca 2+ ] i oscillations caused by cAMP signal relay. Oscillation periods in the early mound stage were significantly shorter than those in the early aggregation and late mound stages (p < 0.001) (Fig. 1e). The early and late mound stages described in this study correspond to the loose and tight mound stages, respectively. The periods of [Ca 2+ ] i oscillations in the early aggregation, and early and late mound stages were 5.29 ± 0.59, 2.95 ± 0.61, and 4.60 ± 0.89 min, respectively. These periods are consistent with those of [cAMP] i oscillations 16 . Wave propagation was observed until the late mound stage; however, signal oscillations and propagation in cell populations disappeared when the late mound began elongation, that is, at the onset of slug formation ( Fig. 1f-h, Movie 8). Consistent with this, periodic oscillations in the cAMP signal have also been shown to disappear after the late mound stage 16 . Signal visualization using GECIs revealed that [Ca 2+ ] i signal dynamics, as well as cAMP signal dynamics, show transitions during multicellular morphogenesis 16 .

Transient [Ca 2+
] i bursts and their propagation in migrating slugs. During late Dictyostelium development, the late mound elongates vertically into a cylindrical structure called a finger, which subsequently falls over and starts to migrate as a slug. When monitoring [Ca 2+ ] i dynamics in migrating slugs using YC-Nano15, transient and rapid elevations of [Ca 2+ ] i , or bursts, and their propagation were observed ( Fig. 2a, b, Movie 9), although no wave propagation was observed in the finger stage ( Fig. 1). Monitoring the signal using GCaMP6s also revealed these transient signal propagations in migrating slugs, with [Ca 2+ ] i bursts observed in both the anterior and posterior of slugs, which are regarded as prestalk and prespore regions, respectively, despite the differences in expressed genes due to differentiation ( Fig. 2c-f, Movies 10,11). During slug [Ca 2+ ] i bursts, the migration velocities increased transiently, with a peak delay of approximately 2 min (Fig. 2b, Movie 9). The periodicity of [Ca 2+ ] i signals that we observed during aggregation and mound stages ( Fig. 1) was not observed in migrating slugs, which occasionally showed irregular [Ca 2+ ] i bursts ( Fig. 2 and Supplementary Fig. 3). Thus, although periodic [Ca 2+ ] i signal propagations disappeared during the process of multicellular development, the ability of Ca 2+ signaling was maintained after slug formation, and the occasional propagation of [Ca 2+ ] i bursts in migrating slugs affected the cooperative movement of cells (Movie 9).

Slug [Ca 2+ ] i bursts are induced by mechanical stimulation. On closer observation, [Ca 2+ ] i bursts in
slugs occurred when a part of the slug touched the surface of the agar (Fig. 2c, e and Movies 10, 11). This suggested that rapid [Ca 2+ ] i elevation in slugs was induced by mechanical stimuli. To confirm this, [Ca 2+ ] i dynamics were monitored using GCaMP6s after slugs were subjected to mechanical stimulation. Slugs developing on agar were excised with their supporting agar, then turned over onto glass dishes, such that they were sandwiched between glass and the agar. They were then pressed from above with a 5 mm diameter plastic rod without crushing so that the entire slug was stimulated evenly (Supplementary Fig. 4a). In all tests using wild-type expressing GCaMP6s, [Ca 2+ ] i in the anterior region of slug increased transiently with a peak at 25.0 ± 4.1 s after all stimulation (n = 9) (Fig. 3a, b and Movie 12). The posterior region also responded to the stimulus with the same peak time as the anterior region, but with attenuated changes in [Ca 2+ ] i levels (Fig. 3b). In contrast, neither an increase nor a decrease in intracellular cAMP concentration was observed when slugs expressing the cAMP fluorescent www.nature.com/scientificreports/ The IplA Ca 2+ channel is involved in calcium signaling in response to the mechanical stimulation of slugs. In the unicellular phase of Dictyostelium, the IP3 receptor, IplA, which is localized in the endoplasmic reticulum (ER), is responsible for [Ca 2+ ] i elevation in response to mechanical stimuli 30 . Expression of iplA mRNA is low during growth, peaks at about 9 h after starvation, and then decreases 29   www.nature.com/scientificreports/ slugs), and the response at the anterior region peaked at 15.7 ± 7.9 s (n = 18), earlier than in wild-type controls (Figs. 3 and 4). There was no difference in the timing of the response peak between the anterior and posterior regions (Fig. 4). Calcium responses were also observed when wild-type and iplA − slugs bumped into the agar ( Supplementary Fig. 6). These results suggest that [Ca 2+ ] i bursts in response to mechanical stimuli are partially mediated by the IplA channel.
Calcium influx allows for a rapid response to mechanical stimuli. Deletion of IplA did not completely abolish the slug calcium response to mechanical stimulation (Fig. 4), indicating that another Ca 2+ pathway contributes to mechanosensing. It has been reported that extracellular Ca 2+ influx via the Piezo channel homolog is important for mechanosensing during the unicellular stage of Dictyostelium 31 . To investigate whether   (Fig. 5a, b). However, EGTA slowed the response peak to 67.7 ± 16.8 s (p < 0.001) (Fig. 5b, e). Additionally, in the presence of 1 mM EGTA, iplA − slugs did not show any calcium response to mechanical stimulation (n = 13) (Fig. 5c). These and our iplA − mutant results indicate that Ca 2+ influx from extracellular sources allows a fast response, whereas IplA is essential for the response from intracellular sources (Figs. 4 and 5c). We constructed a pzoA null strain and confirmed that PzoA is essential for Ca 2+ influx from extracellular sources by mechanical stimulation during unicellular stages as previously reported ( Supplementary Fig. 7) 31 . However, the response of pzoA − slugs was similar to that of wild-type, suggesting that Ca 2+ influx is mediated by other pathways during the multicellular phase (Fig. 5d, e).

Discussion
In Dictyostelium, [Ca 2+ ] i transients have been observed during both the mound and slug stages 21 ; however, the actual dynamics of Ca 2+ signaling have not been clarified, because most previous studies have focused on stages before aggregation 17,23,29,32 . In this study, [Ca 2+ ] i imaging during Dictyostelium development with highly sensitive GECIs revealed that synchronized [Ca 2+ ] i elevations and their propagation in cell populations occur continuously during the aggregation and mound stages; however, they disappear in the late stage of multicellular development. Ca 2+ wave propagation depends on cAMP relay during the early aggregation and mound stages, and cAMP wave propagation disappears during tip elongation in the late mound 16,20 . This cAMP signal has been shown to induce transient [Ca 2+ ] i elevation 23,33 . Therefore, changes in Ca 2+ dynamics of during development follow the transition of cAMP relay. We found that [Ca 2+ ] i bursts and their propagation occasionally occurred in slugs (Fig. 2). Consistent with previous reports of cAMP signaling 16 , the signal oscillation periods in the early mound stage was shorter than in the early aggregation and late mound stages (Fig. 1e). The decrease in the oscillation periods is due to increased cell density and extracellular cAMP levels 34,35 , and the increase in the oscillation periods can be explained by the expression of low-affinity cAMP receptors in the mound stage instead of the high-affinity cAMP receptors expressed in the aggregation stage 36 . Our results suggest that these occasional [Ca 2+ ] i bursts are in response to mechanical stimulus caused by bumping of the tips of migrating slugs in an environment where they are sandwiched between agar and glass ( Supplementary Fig. 4). When [Ca 2+ ] i signals were propagated in migrating slugs, migration velocity transiently increased. Ca 2+ signaling affects both cell movement at the unicellular stage and slug behavior 30,37,38 . Calcium wave propagation and its effects are also well known in animal cells, with gap junctions as essential components in cell-cell signaling 4 . Given that Dictyostelium does not have gap junction component homologs 39,40 , the mechanism for calcium signal propagation in slugs must be gap junction independent. Dictyostelium cells show [Ca 2+ ] i elevation in response to cAMP signals and mechanical stimuli in the unicellular phase 30,31,41 . Our assay showed that slug [Ca 2+ ] i bursts and their propagation were induced by mechanical stimuli. The IplA Ca 2+ channel and Ca 2+ release from ER have been shown to be involved in the Ca 2+ response to mechanical stimuli in the unicellular phase 30,31,41,42 . In an iplA − strain, there is no elevation of [Ca 2+ ] i in response to cAMP stimulation of starved single cells and cell aggregation is delayed 29,43 . In slugs lacking IpIA, the calcium response to mechanical stimulation was attenuated, suggesting that IplA is responsible for increasing the certainty of response to mechanical stimuli. The IplA channel has been shown to be essential for Ca 2+ dependent flow-directed motility, but not for chemotactic migration toward cAMP sources 44 . This suggests that both cAMP signaling and IplA-mediated Ca 2+ signaling affect downstream components independently. This hypothesis is supported by the observation that there is no clear defect in the development of iplA − slugs under laboratory conditions (Movie 7). However, mechanosensing may play important morphogenetic roles in natural environments, where soil-living Dictyostelium amoebae are exposed to more complex stimuli and physical   www.nature.com/scientificreports/ important in natural environments (Movie 2). Alternatively, the Ca 2+ response to mechanical stimulation was not completely abolished in iplA − slugs, suggesting that other pathways are involved in the elevation of [Ca 2+ ] i . In the Dictyostelium genome, other potential Ca 2+ signaling components include the mucolipin channel (mcln), two pore channels (tpc), a transient receptor potential (trp) channel, and an Msc-like channel (mscS) 42,[46][47][48][49] . In higher eukaryotes, the stretch-activated Ca 2+ permeable ion channel Piezo is involved in mechanical stimulus responses 8,9 . Recently, it has been reported that D. discoideum has a homolog of Piezo, PzoA. Disruption of the pzoA gene causes defects in the [Ca 2+ ] i response to mechanical stimuli in amoebae 31 . Additionally, cells lacking PzoA develop normally under laboratory conditions; however, a defect is present in chemotactic migration under confined conditions 31 . Slugs lacking pzoA did not show a substantial difference in calcium response from that of wild-type control slugs. Notably, the Piezo channel in higher multicellular organisms functions only in the unicellular phase in Dictyostelium. Even within Dictyostelium, there are interesting changes, with Piezo acting as the main pathway during unicellularity, and pathways from the extracellular and ER during multicellularity.
In iplA − cells, the [Ca 2+ ] i response was faster than in the wild type. This indicates that the apparent single-peak [Ca 2+ ] i burst is a mixture of a fast extracellular Ca 2+ influx with a slower, yet more efficient, response from the ER. This delayed ER response may be due to the fact that the signal from mechanoreceptors in the plasma membrane is transmitted via IP3. Moreover, in the unicellular phase of iplA − cells, no calcium response to mechanical stimuli was detected, even though IplA is not involved in blebbing, which is regulated by calcium signaling related to mechanical stimulation 31 . In a human colorectal carcinoma cell line, membrane blebbing is regulated by store-operated Ca 2+ entry, which is controlled by ER proteins 50 . Alternatively, although no homolog of stromal interaction molecule (STIM) has been found in Dictyostelium 51 , unknown store-operated calcium channels (SOCs) may be transducing mechanical stimuli. The [Ca 2+ ] i bursts at both the tip and posterior regions in slugs indicate the ability for rapid [Ca 2+ ] i elevation in both prestalk and prespore cells. Previous studies have shown that [Ca 2+ ] i in the anterior part of the slug is higher than that in the posterior 21,22 . However, in our study, this difference was not observed in slugs in a steady state, not receiving any external stimuli. On the contrary, if these previous data contain dynamic information, they are consistent with those of our study (Fig. 3). As slugs migrate with their tips protruding vertically 52 , it may be easier to generate anterior responses to obstacles or enemies, such as nematodes 53 . Thus, it has been frequently observed that [Ca 2+ ] i is higher in the anterior than in the posterior of the slug. Because Ca 2+ responses were observed when slugs collapsed (Fig. 2c-f), it is also possible that the mechanical stimulus response supports the construction of the fruiting body.
In conclusion, we have shown that [Ca 2+ ] i bursts and their propagation in Dictyostelium are dependent on cell-cell communication via diffusible chemical signals during early developmental stages. Following multicellular development, such Ca 2+ signaling is triggered by mechanical stimuli. Because mechanical stimuli could not be accurately quantified in our system, the contribution of calcium to the mechanical stimulus response remains unresolved and warrants further study. Ca 2+ signaling in response to mechanical stimuli is conserved broadly in higher eukaryotes and prokaryotes, such as Escherichia coli 4,[54][55][56] . We observed that the social amoeba, D. discoideum, belonging to Amoebozoa, uses Ca 2+ signaling as a mechanosensing signal in the multicellular phase similar to that in the unicellular phase 31 ; however, the molecular mechanism is different. Thus, this study demonstrates that mechanochemical signal transduction via Ca 2+ signaling is a universal system for response to mechanical stimuli and can be applied in any cell type or state. In this study, the pathway for extracellular Ca 2+ uptake associated with mechanical stimulation in the slugs of D. discoideum was not identified. This is because even though Ca 2+ acts as a signal across species, molecular mechanisms differ significantly. Further studies are required to clarify conserved and specific molecular mechanisms.

Plasmid construction and genetic manipulation. Plasmids are listed in Supplementary
Imaging. In all experiments, cells were observed at 22 °C. Confocal images were taken using an A1 confocal laser microscope (Nikon, Japan) with an oil immersion lens (Plan Fluor 40 ×/1.30 NA, Nikon), or using an inverted microscope (Eclipse Ti, Nikon) equipped with a CSU-W1 confocal scanner unit (Yokogawa), two sCMOS cameras (ORCA-Flash4.0v3, Hamamatsu Photonics, Japan), and oil immersion lenses (Plan Apo 60 ×/1.40 NA or CFI Apo TIRF 60 ×/1.49, Nikon). GCaMP6s and YC-Nano15 were excited using 488 and 440 nm solid-state CW lasers, respectively. Epifluorescence micrographs were acquired using an inverted epifluorescence microscope (IX83, Olympus, Japan) equipped with a 130 W mercury lamp system (U-HGLGPS, Olympus), sCMOS cameras (Zyla4.2, Andor Technology or Prime 95B, Photometrics, USA), and objective lenses (UPLSAPO 4 ×/0. 16  All images were processed and analyzed using the Fiji 59 and R software. GCaMP6s oscillations were calculated by averaging differences between peaks. Data with at least three peaks in the oscillation were used for the analysis. In general, fluorescence intensities of GCaMP6s, Flamindo2, and the ratio of YFP/CFP channels of YCNano15 were normalized to their values at t = 0.

Live imaging of [Ca 2+
] i dynamics during Dictyostelium development. Imaging was performed as described previously 16 . To induce development upon starvation, exponentially growing cells (1.5-3 × 10 6 cells ml −1 ) were harvested and washed three times in KK2 phosphate buffer (20 mM KH 2 PO 4 /K 2 HPO 4 , pH 6.0). To monitor development, cells were plated on the entire surface of 2% water agar (2% w/v Difco Bacto-agar in ultrapure water) in 35 mm plastic dishes at a density of 5-7 × 10 5 cells cm −2 (Iwaki, Japan) and incubated at 21 °C. Thereafter, plates were filled with liquid paraffin (Nacalai Tesque, Japan) to attenuate light scattering and for microscopy. Additionally, the "2D slug" method 60,61 was applied for observing slug migration without threedimensional scroll movement ( Fig. 4a and Supplementary Fig. 2). One microliter of cell suspension (4 × 10 7 cells mL −1 ) was dropped on 2% water agar plates with 2 μL liquid paraffin. A coverslip was placed over the suspension and incubated at 21 °C for a minimum of 15 h.

GCaMP6s as an indicator of [Ca 2+
] i . Dictyostelium cells expressing GCaMP6s were suspended in 1 mL developmental buffer (DB: 5 mM Na/KPO 4 , 2 mM MgSO 4 , 0.2 mM CaCl 2 , pH 6.5) at a density of 5 × 10 5 cells mL −1 and incubated for 1 h. Thereafter, cells were exposed to 100 nM cAMP pulses at 6 min intervals for the next 5 h. Following starvation with cAMP pulses, cells were washed three times with 1 mL DB and resuspended in DB at a density of 10 6 per mL. Cell suspension (40 µL) was deposited onto a glass bottom dish. Cells were stimulated by adding 160 μL of 12.5 μM cAMP (Sigma Aldrich, USA) to the cell droplet (final concentration 10 μM). During stimulation, confocal fluorescent micrographs of starved cells were acquired at 5 s intervals. Averaged fluorescence intensities of GCaMP6s in 5 μm 2 regions positioned within the cytosol were estimated for each time point.

Slug [Ca 2+
] i response to mechanical stimulation. Five microliter of cell suspension, at a density of 4 × 10 7 cells mL −1 , was deposited on 2% water agar with or without 1 mM EGTA and incubated at 21 °C for 12-15 h. Following slug formation, a piece of agar containing slugs was excised and placed upside down on a spacer attached to a 35 mm glass bottom dish (12 mm diameter glass, Iwaki). The spacer was filled with liquid paraffin to prevent desiccation during observation and to avoid light scattering. Slugs covered with agar were pushed with a 5 mm diameter plastic rod using a micromanipulator system (MM-94 and MMO-4, Narishige, Japan) ( Supplementary Fig. 4a). In the micropipette assay, a piece of agar with slugs was excised and placed directly on a 35 mm glass bottom dish (12 mm diameter glass, Iwaki). A wet paper was placed in the dish and the agar piece was covered with liquid paraffin. A Femtotip microcapillary (1 µm tip diameter, Eppendorf, Germany) was mounted onto a Femtojet pump and micromanipulator (Eppendorf). The slug was pricked with the pipette using manual operation with the manipulator (Supplementary Fig. 4b). During mechanical stimulation, images of slugs expressing GCaMP6s were acquired at 5 s intervals using epifluorescence microscopy.

Data availability
The data supporting the findings of this study are available from the corresponding author upon request.