Feast or flee: bioelectrical regulation of feeding and predator evasion behaviors in the planktonic alveolate Favella sp. (Spirotrichia)

Alveolate grazers (ciliates and dinoflagellates) are key members of marine plankton communities, and as members of the microzooplankton consume ∼60% of global marine primary production in the world's oceans (Schmoker et al., 2013). They therefore mediate regeneration of nutrients in the surface oceans (Dolan, 1997; L'Helguen et al., 2005) and act as a trophic link between phytoplankton and larger metazoan consumers in marine food webs. The ecological success of alveolate grazers is underpinned by complex behaviors they have evolved to capture preferred prey efficiently and reject toxic and unpalatable prey, while avoiding predators (Taniguchi and Takeda, 1988; Buskey and Stoecker, 1989; Broglio et al., 2001). To effectively capture prey and avoid predation, alveolate grazers must be able to sense a range of hydrodynamic, mechanical, chemical, and electrostatic signals produced by prey and predators, and integrate these signals via signal transduction mechanisms to respond with effective prey capture or predator evasion behaviors. Predators and prey of alveolate grazers can generate similar mechanical stimuli so the sensory mechanisms of alveolate grazers must also be sophisticated enough to discern between the two. Despite their importance, the sensory mechanisms of alveolate grazers are poorly understood.

The marine ciliate Favella sp. (Fig. 1A) is a cosmopolitan member of the plankton community that has well-described trophic behaviors (Stoecker and Sanders, 1985; Buskey and Stoecker, 1988,, 1989; Taniguchi and Takeda, 1988; Stoecker et al., 1995, 2013; Strom et al., 2007) and has morphological and behavioral characteristics that are representative of marine planktonic ciliates; they are therefore an excellent model for studying the sensory mechanisms underlying the trophic ecology of marine alveolate grazers (Montagnes, 2013; Echevarria et al., 2014). [Note: the genus Favella has been recently redescribed and some former Favella spp. have been placed in the new genus Schmidingeralla by Agatha and Strüder-Kypke (2012). For consistency with the ecological literature, and because of uncertainties in species-level identification in previous studies, we retain the name Favella throughout this paper.] Favella sp. exhibit aggregative swimming patterns in the presence of chemical and mechanical stimuli from preferred prey that are believed to allow them to increase residence time in prey patches that typically occur in the environment (Buskey and Stoecker, 1989). Increased frequency of ciliary reversals that cause transient backward swimming and turning events in response to mechanical stimuli appear to underlie these aggregative swimming patterns (Buskey and Stoecker, 1988; Stoecker et al., 1995). However, the sensory mechanisms that underlie these ciliary reversals are unknown. Once food particles have been successfully contacted, they are processed at the peristomal cavity where they are either rejected or consumed depending on the characteristics of the prey surface (Tanguchi and Takeda, 1988; Stoecker et al., 1995; Montagnes et al., 2008). Favella sp. are also capable of performing predator avoidance behaviors. In response to predation attempts by a predatory dinoflagellate, Favella azorca contract into the lorica, although dinoflagellates are still able to extract them (Uchida et al., 1997). The lorica may be effective as protection from copepods; in feeding experiments with a copepod, crushed loricas containing intact cells were observed at the end of the experiment (Stoecker and Sanders, 1985).

Although Favella sp. exhibit complex behavioral responses to predators and prey, the cellular mechanisms that underlie these responses are unknown. In contrast to marine ciliates, the sensory mechanisms of model freshwater ciliates such as Paramecium and Tetrahymena have been well studied. Although Favella sp. may be expected to utilize similar conserved sensory mechanisms, it is likely that they implement them in different ways due to their different morphological and ecological characteristics. Nevertheless, it is known that ionotropic mechanisms – those governed by voltage-, chemo-, or mechano-sensitive ion channels – are important in regulating the behaviors of ciliates (Preston and Van Houten, 1987; Shiono and Naitoh, 1997; Grønlien et al., 2011). For example, ionotropic mechanisms are known to be important in sensory processes that relate to prey capture in ciliates. Contact of the freshwater ciliate Tetrahymena vorax with prey cells elicits ciliary reversal and backward swimming, which are regulated by transient depolarizations (Grønlien et al., 2011). Ionotropic mechanisms may also be important in regulating predator evasion behaviors. Contractile behavior of the freshwater ciliate Vorticella sp. upon mechanostimulation is regulated by rapid all-or-nothing action potentials (Shiono and Naitoh, 1997). Many ciliates contain extrusomes, membrane-bound organelles that can be ejected from the cell to deter predators and immobilize prey, whose release is coordinated by the bioelectrical activity of the cell (Iwadate et al., 1997; Harumoto et al., 1998; Rosati and Modeo, 2003). Similar ionotropic mechanisms may function in regulating predator evasion and prey capture behaviors in Favella sp.

To investigate the sensory capabilities of Favella sp., electrophysiological measurements and high-speed video microscopy were performed in the presence of artificial and live prey, and predatory copepods to investigate the bioelectrical regulation of prey capture and predator evasion behaviors. Using this powerful combination of techniques, we investigated signal transduction from environmental stimuli to behavioral response in an ecologically relevant context, providing a deeper understanding of the sensory mechanisms possessed by marine alveolate grazers.

Favella sp. and the phytoplankton cultures Heterocapsa triquetra, Mantoniella squamata, and Isochrysis galbana were obtained from the Shannon Point Marine Center, Anacortes, WA. The phytoplankton Tetraselmis sp. was obtained from Michael Finiguerra at the University of Connecticut, Groton, CT. Favella sp. cultures were maintained in 200 ml batches at 14–16°C in filtered seawater (30 ppt) supplemented with a dilute trace metal solution (Strom et al., 2007). Favella sp. cultures were fed with prey and sub-cultured every 3–4 days. Phytoplankton were maintained in 40–1000 ml batches in filtered 36 ppt seawater supplemented with f/2 nutrients and Guillard's vitamins (Guillard, 1975) at 14–16°C and 50–80 μmol photons m⁻² s⁻¹ on a 12 h:12 h light:dark cycle. Phytoplankton were sub-cultured every 1–2 weeks. The copepod predator of Favella sp., Acartia tonsa, was obtained from the aquaculture supplier AlgaGen (Vero Beach, FL, USA). Acartia tonsa were maintained in 2 liter batches at 14–16°C in filtered 33 ppt seawater at 14–16°C and 50–80 μmol photons m⁻² s⁻¹ on a 12 h:12 h light:dark cycle. They were fed Tetraselmis sp. and I. galbana every 3–4 days. For physiological experiments examining interactions of Favella sp. and H. triquetra, the latter were prepared immediately prior to use by gentle filtration through 3 μm pore size Whatman polycarbonate filters (General Electric, Pittsburgh, PA, USA) followed by two rinses in artificial seawater (ASW; 450 mmol l⁻¹ NaCl, 30 mmol l⁻¹ MgCl₂, 16 mmol l⁻¹ MgSO₄, 8 mmol l⁻¹ KCl, 10 mmol l⁻¹ CaCl₂, 2 mmol l⁻¹ NaHCO₃, 10 mmol l⁻¹ HEPES, pH adjusted to 8.0 with NaOH) to remove any dissolved chemical cues present in the supernatant.

Electrophysiological measurements of the bioelectrical activity of Favella sp. were obtained using sharp electrode current clamp and voltage clamp recordings of whole-cell membrane potential (V_m) and currents, respectively (Fig. 1A). Favella sp. were prepared for electrophysiology by rinsing cultures three times in ASW by reverse filtration through a 40 μm nylon mesh cell strainer (Corning Life Sciences, ME, USA) before being placed in a recording chamber equipped with gravity-fed perfusion that was secured to the stage of an Olympus IX 71 microscope. A Peltier cooled stage (fabricated in-house) maintained the recording chamber at 14–16°C for experiments where behaviors and electrical activity of Favella sp. were recorded simultaneously. All other experiments were performed at room temperature (∼20°C). Cells were tethered by the base of the lorica (Fig. 1A) using a micro suction pipette fabricated from non-filamented borosilicate glass capillaries (1.5 mm outer diameter×0.86 mm inner diameter, Sutter Instruments, Novato, CA, USA) using a microelectrode puller (Narishge, Tokyo, Japan) and bent into a position that allowed optimal imaging and access to cells with sharp electrodes. Sharp recording electrodes (10–20 MΩ) were fabricated from filamented borosilicate glass capillaries (1.5 mm outer diameter×0.86 mm inner diameter, Sutter Instruments), coated in beeswax to minimize stray capacitance, filled with 1 mol l⁻¹ KCl and inserted into an electrode holder (Harvard Apparatus, Holliston, MA, USA) that was mounted onto the headstage of an Axon 900 A voltage clamp amplifier (Molecular Devices, Sunnyvale, CA, USA). The electrode was positioned using a Sutter MP-285 motorized micromanipulator (Sutter Instruments) until contact was made with the lateral surface of the cell, just posterior to the ring of membranelles. Electrical access to the cell was accomplished using a brief (<30 ms) capacitance overcompensation.

Single electrode current and voltage clamp experiments were recorded using Clampex software (Molecular Devices). Single electrode current clamp switching frequency was generally between 5 and 15 kHz and voltage clamp gain was generally between 0.3 and 3 nA mV⁻¹. To determine the voltage response of Favella sp. to negative current injection and the passive membrane properties of Favella sp., hyperpolarizing step protocols were used that injected 1 nA pulses of negative current for 120 ms. The passive membrane properties of Favella sp. were calculated by calculating resistance and the time constant (τ) from the voltage response to injection of negative current. Action potentials (APs) and rhythmic depolarizations (RDs) (two stereotypical patterns of bioelectrical activity in Favella sp.) were elicited by injecting 5–10 nA current for 10–20 ms. Cells were perfused with 0.8, 4, 8, and 40 mmol l⁻¹ KCl ASW while measuring free-running V_m to determine the role of K⁺ in regulating resting potential and on spontaneous electrical activity. Experiments that examined the relationship between electrical activity and behaviors were performed without perfusion to avoid mechanostimulation by the flow of media around the cell.

Data were analyzed off-line in Clampfit 11.0 (Molecular Devices). The characteristics of spontaneous electrical activity and electrical activity in the presence of prey were determined using the threshold analysis tools of the software. The characteristics of APs evoked by positive current injection were measured using manual cursors.

Simultaneous electrophysiological and video acquisition allowed us to determine how the bioelectrical phenomena described above coordinate behavior in Favella sp. High-speed (up to 250 frames s⁻¹) video recordings were accomplished using a Fastec Inline 250 frames s⁻¹ camera (Fastec Imaging, San Diego, CA, USA) with 320×228 pixel resolution. Camera acquisition was triggered with a transistor–transistor logic pulse from the electrophysiology software for simultaneous electrophysiology and behavioral recordings. To quantify spontaneous and evoked behavioral characteristics of Favella sp., videos were analyzed using MiDAS Player (Ver. 2.2.1.1., Xcitex, Cambridge, MA, USA) and Metamorph Basic (Ver. 7.7.4.0., Molecular Devices) software to define regions of interest (ROIs). ROIs were drawn around areas occupied by structures of interest, e.g. cilium, cell body, peristomal cavity, or stalk. Threshold analysis tools in Clampfit 11.0 (Molecular Devices) were used to measure changes in average pixel intensity value as the cell structure moved into and out of the ROI. Pixel intensity values were normalized to a scale of 100 and inverted for the purposes of data presentation.

To determine how the bioelectrical activity of Favella sp. coordinates feeding behaviors, electrophysiological and high-speed video recordings were obtained in the presence of natural and artificial prey (polystyrene microspheres). Favella sp. were exposed to either the preferred prey dinoflagellate H. triquetra or 15 μm polystyrene microspheres (Polysciences, Warrington, PA, USA) at concentrations of 15,000–30,000 ml⁻¹ while membrane potential and behavior were recorded simultaneously. Microspheres were coated with methylcellulose to neutralize surface charge (Stoecker et al., 1995) and were the same equivalent spherical diameter as H. triquetra, and therefore simulated the mechanical stimulus from prey while controlling for chemical prey cues. Interactions of Favella sp. with microspheres and H. triquetra were manually scored via frame-by-frame observations to determine the sequence of events from time and location of particle contact to bioelectrical response and behavioral event.

Behavioral interactions of free-swimming Favella sp. with the copepod predator A. tonsa were investigated using high-speed (125 frames s⁻¹) video microscopy. Prior to experiments, A. tonsa cultures were rinsed twice with ASW by reverse filtration through 40 μm pore size nylon cell strainers and starved for 15–20 h. Immediately prior to experiments, ciliates and copepods were rinsed twice with ASW by reverse filtration through 40 μm pore size nylon cell strainers (Corning Life Sciences). Thirty individuals of Favella sp. and six late-stage copepodite to adult female A. tonsa were combined in a 3.5 ml spectrophotometry cuvette and the volume was adjusted to 1 ml with ASW. Cuvettes were imaged through the side using an Olympus SZX12 dissecting microscope at 12.5× magnification. High-speed (125 frames s⁻¹) video recordings were accomplished using a Fastec Inline 250 frames s⁻¹ camera (Fastec Imaging) with 640×478 pixel resolution. Cuvettes were observed for 0.5 h and all events where ciliates were captured by copepods were recorded. Behavioral interactions between copepods and ciliates were manually scored as ‘captures and ingests Favella sp.’ or ‘captures and releases Favella sp.’. The amount of time it took Favella sp. to emerge from the lorica post-release was also measured.

All differences in means were tested using a one-way ANOVA followed by multiple comparisons with Tukey's HSD test. To determine the relationship between the length of rhythmic depolarizations and behavioral events, linear regressions were performed. All statistical analyses were performed with Sigmaplot Software version 11 (Systat Software). All quantitative measures are given as means±s.d. (n=number of cells, events analyzed = total number of events analyzed for all cells) unless noted otherwise.

The passive membrane properties of Favella sp. were similar to previously investigated ciliates (Table S1). All Favella sp. cells produced spontaneous RDs (n=34; Fig. 1B,C), and occasionally rapid ‘all-or-nothing’ APs were observed (n=4, events analyzed=4; Fig. 1B,C). RDs had a frequency of 0.153±0.1 Hz, a peak amplitude of 27.4±2.2 mV and an average length of 618±183 ms (n=5, events analyzed=58). When cells were voltage clamped at resting V_m (−62 mV), spontaneous inward currents were recorded that underlie the RDs observed in current clamp (Fig. 1D).

A strong dependence of V_m upon external [K⁺] was observed, with a Nernstian depolarization (50 mV change in V_m per decadal change in external [K⁺]) of V_m, over the concentration range between 4 and 40 mmol l⁻¹ KCl ASW (Fig. 2A,B). V_m deviated from the Nernstian relationship in <4 mmol l⁻¹ KCl ASW, with cells exhibiting only a slight hyperpolarization when perfused with 0.8 mmol l⁻¹ KCl ASW (Fig. 2A,B). Interestingly, changing external [K⁺] changed the characteristics of RDs. The duration and peak depolarization of RDs in all [K⁺]_ext concentrations was similar. However, decreased external [K⁺] led to decreased amplitude of the small hyperpolarizations that occurred at the end of RDs (P<0.05, ANOVA, Tukey's HSD; Fig. 2C, Table S1). In contrast, high external [K⁺] (40 mmol l⁻¹ KCl ASW) caused APs to be generated at the rising phase of the RDs (Fig. 2D). These AP-RDs were much shorter than RDs and did not exhibit hyperpolarizations following the depolarized phase of the event (n=4, events analyzed=51; Fig. 2D, Table S1). They almost always preceded bursts of APs induced by the depolarized state (up to 3 Hz; Fig. S1) that were accompanied by whole-cell contraction into the lorica (not shown).

Under normal ionic conditions, positive current injection pulses elicited depolarization and either RDs or APs (Fig. 3A,B) that were identical to spontaneous RDs and APs. Over the range of positive current injection protocols tested there did not appear to be a correlation between length (10–20 ms) or amplitude (5–10 nA) of injection and type of response elicited (AP or RD). In contrast, negative current injection elicited a typical RC voltage response (Fig. 3C). APs were elicited in seven cells where 5–10 nA current was injected for 10–20 ms (Fig. 3B). APs did not occur until V_m reached an average depolarization threshold of −3.7±17.10 mV (n=3, events analyzed=3) and V_m at peak after-hyperpolarization was −76.50±6.75 mV (n=3, events analyzed=3). The time to peak depolarization from start of the evoked AP was very rapid at 6.92±2.40 ms (n=7, events analyzed=7).

Spontaneous RDs always resulted in ciliary reversals, peristomal cavity contractions, and stalk bending events (Figs 4, 5, 6A), which all began within 70–110 ms of the start of RDs (Fig. 6A). RDs lasted approximately the same length of time as the events they regulated (between 600 and 750 ms, n=4, events analyzed=58; Fig. 6A): there was a strong 1:1 correlation (r²=0.89, P<0.001) between the length of RDs and the period of time cilia were in reverse beating. There were weaker correlations between length of RDs and duration of peristomal cavity contractions (r²=0.23, P<0.001) and stalk bending (r²=0.17, P=0.007; Fig. 5). During forward beating, the instantaneous ciliary beat frequency (CBF) was 27.7±1.5 Hz (n=5, events analyzed >10,000) and during reversed beating it was slightly higher at 33.3±7.5 Hz (n=5, events analyzed >500; Fig. S2). Forward beating transitioned to reverse beating very rapidly (within approximately 45 ms; n=5, events analyzed=48).

Contact of H. triquetra with the cilia of Favella sp. invariably resulted in depolarizations and behaviors (ciliary reversals, peristomal cavity contractions, and stalk bending; n=4, events analyzed=13) that were very similar to RDs (Figs 6B, 7A, Movie S1). Identical responses were observed when H. triquetra-sized polystyrene microspheres (15 μm) that were used as artificial prey to control for the effect of chemical cues present on the surface of natural prey particles (Figs 6C, 7B), and these events are thus termed mechanostimulated depolarizations (MSDs). There was a very short lag of 29.8±17.9 ms (n=4, events analyzed=13) between contact of H. triquetra with cilia and MSDs, with a nearly identical lag observed with beads (Figs 6B–C, 7A–B, Movie S1).

Polystyrene microspheres were never observed to be ingested on contact with the peristomal cavity (n=5, events analyzed=6) (although one cell contained a microsphere that was ingested before video recording started; Fig. 6B,D). Instead, a significant lag occurred before a depolarization was elicited that resulted in ciliary reversal and ejection of microspheres from the peristomal cavity (Figs 6E, 7D). In some cases when H. triquetra contacted the peristomal cavity, they also elicited this response (Fig. 6D, Fig. 7C). Although the lag between particle contact at the peristomal cavity and subsequent MSD was longer for H. triquetra (750±552 ms) than for microspheres (596±777 ms), lag times were highly variable and therefore not statistically significant. However, in other instances, contact of H. triquetra with the peristomal cavity resulted in ingestion after a processing period of 1.01±0.84 s (n=7, events analyzed=7) that invariably resulted in hyperpolarization that began 212.1±384.7 ms (n=3, events analyzed=3) after ingestions started, lasted 1700±800 ms (n=8, events analyzed=8), and had peak amplitudes of 3.5±0.8 mV (n=8, events analyzed=8) (Figs 6F, 7E, Movie S2). No change in CBF was observed during hyperpolarizations associated with prey ingestion (n=4, data not shown), suggesting that swimming was unaffected. The striking discrimination against ingesting microspheres suggests that ingestion and associated hyperpolarization are mediated by chemical cues present on the cell surface, but not on the microspheres.

Capture of Favella sp. by A. tonsa resulted in contraction of Favella sp. into the lorica (Fig. 8). Favella sp. were released 73±25% of the time and ingested the rest of the time (six experiments, n=16 events). Released ciliates were invariably contracted into the lorica, but re-emerged and resumed swimming after 10.6±3.8 s (six experiments, n=5 events; Fig. 8).

The bioelectrical basis of this predator evasion behavior was investigated using sharp electrode current clamp recordings and simultaneous high-speed video microscopy. Spontaneous and evoked APs caused cessation of cilia beating and rapid contraction into the lorica (Figs 6F, 9, Movie S3). The coupling between electrical event and behavioral response was extremely rapid, with contraction starting 12.2±5.9 ms after the initiation of AP (n=7, events analyzed=7), with cells reaching 95% of their full contraction at 121±77 ms (n=5, events analyzed=5). Cells gradually recovered to their pre-AP state after 8.8±2.0 s (n=5, events analyzed=5; Figs 6F, 9, Movie S3).

Favella sp. exhibited a resting membrane potential that was more negative than previously studied marine and freshwater ciliates reported in the literature (Table S2). This indicates that their resting permeability and the electromotive force across the plasma membrane are determined by somewhat different conductances. The passive membrane properties of Favella sp. show a high input capacitance in comparison with other ciliates (Table S2) that likely reflects their large cell size and large membrane surface area. Additionally, their input resistance was rather low compared with that of other ciliates (Table S2), suggesting a high resting ionic permeability. Similar to other eukaryotes, resting permeability to K⁺ was important in determining resting V_m in Favella sp. The near-Nernstian relationship of V_m and [K⁺]_ext at concentrations above 8 mmol l⁻¹ KCl indicated that V_m was primarily determined by permeability to K⁺ ions. However, below 8 mmol l⁻¹ KCl the non-linear curve suggests that permeability to ions other than K⁺ contributes to resting V_m.

The bioelectrical activity of Favella sp. is crucial in regulating endogenous swimming patterns (Fig. 10). Favella sp. exhibited frequent spontaneous RDs that regulated ciliary reversals, stalk bending, and peristomal cavity contractions. Ciliary reversals result in periods of backward swimming in free-swimming cells, and stalk bending changes the orientation of the cell and would therefore affect swimming angle. Peristomal cavity contractions may function in clearing the surface of the peristomal cavity of particles during ciliary reversals. The strong correlation between length of RDs and length of ciliary reversals indicates that Favella sp. can modify the length of RDs and ciliary reversals based on environmental and endogenous signals.

RDs have been observed in several ciliates (Machemer, 1970; Lueken et al., 1996), including the marine benthic spirotrich Euplotes vannus, in which RDs mediate alternating periods of slow and fast walking and backward walking behavior (Machemer, 1970; Lueken et al., 1996). Increases in inward conductance of Ca²⁺ and decreases in outward conductance of K⁺ appear to be important in mediating RDs in E. vannus (Lueken et al., 1996). The specific ionic conductances responsible for RDs in Favella sp. are unknown, but the importance of voltage-gated channels in mediating RDs in Favella sp. was suggested by the fact that transient depolarizations and ciliary reversals were elicited by the injection of positive current (Fig. 3A). We observed increased after-hyperpolarization amplitude with decreased [K⁺]_ext, indicating that K⁺ channels were important in returning the cell to resting V_m following RDs (decreased [K⁺]_ext resulted in an increased electromotive force of K⁺ from Favella sp.). The cellular mechanisms that regulate the rhythmicity of RDs in Favella sp. and other ciliates have not been identified. The high frequency of RDs and backward swimming events in Favella sp. and presence of these mechanisms in multiple taxa indicates that they most likely play an important role in structuring ciliate search patterns for food particles and in determining encounter rates with prey and predators.

The rate of ciliary beating in Favella sp. was also regulated by bioelectrical activity of the cell. The increased reverse CBF of Favella sp. during RDs will correspond to increases in swimming speed in free-swimming cells. Intracellular Ca²⁺ dynamics are most likely involved in regulating ciliary beating in Favella sp. In Paramecium sp., Ca²⁺ dynamics determine the frequency and direction of ciliary beating, with Ca²⁺ concentrations greater than 1 μmol l⁻¹ causing a reversal of ciliary beating in Paramecium sp. (Nakaoka et al., 1984; Plattner et al., 2006). Powerful Ca²⁺ buffering systems must also be present in the cilia that return cilia to resting Ca²⁺ levels following depolarization-induced ciliary reversals (Husser et al., 2004). The cyclic nucleotides cAMP and cGMP are also important in regulating CBF and direction, and have interactive effects with Ca²⁺ (Bonini and Nelson, 1988; Noguchi et al., 2004). Similar mechanisms may regulate ciliary beating in Favella sp. Higher CBF will lead to increased swimming velocities and increased encounter rates with prey, but will also increase the susceptibility of Favella sp. to detection by predators.

Although earlier studies linked RDs and ciliary movements, a novel observation in the present study is the coordination of RDs with other behaviors such as peristomal contractions and stalk bending (Fig. 4). Moreover, the differences in bioelectrical signaling at the cilia and peristome during prey handling indicate that Favella sp. are able to discriminate the type and location of stimuli from prey. Contact of H. triquetra and 15 μm microspheres at the cilia elicited almost instantaneous MSD-mediated ciliary reversals, whereas contact of H. triquetra with the peristomal cavity elicited variable responses, sometimes resulting in hyperpolarization-mediated ingestions and sometimes resulting in MSD-mediated ciliary reversals preceded by a lag period. In contrast, contact of 15 μm microspheres at the peristomal cavity invariably elicited ciliary reversals preceded by a lag period (Fig. 10B,C). We hypothesize that spatially distinct classes of ion channels and receptors, allowing for sophisticated prey-handling behaviors, regulate interactions of planktonic alveolates, such as Favella sp., with prey.

MSD-mediated ciliary reversals upon particle contact at the cilia may allow the cell to briefly re-orientate to a particle on the periphery of the cell that could not otherwise be directly ingested. MSD-mediated reversals could also result in increased encounter rates of Favella sp. with prey because of the patchy small-scale distributional patterns of phytoplankton in nature (Mitchell et al., 2008). We propose that the ciliary reversals produced by MSDs increase residence time of Favella sp. within food patches by decreasing their diffusivity, and thus may contribute to mechanostimulus-induced aggregative swimming behavior previously reported for Favella sp. (Buskey and Stoecker, 1988, 1989) (Fig. 10B). Therefore, this is the first demonstration of the cellular basis of population-level responses in Favella sp. Mechanostimulated aggregative swimming behaviors may be an important planktonic alveolate foraging strategy.

The production of MSDs is presumably due to the activation of mechanosensitive ion channels (Fig. 10B). In E. vannus, the bioelectrical regulation of depolarizations and backward movement events induced by mechanical stimulus has been well described (albeit in a non-ecological context). In these cells, mechanical stimulus at the anterior results in an influx of Ca²⁺ and Mg²⁺ through stretch-activated ion channels and Na⁺ through Ca²⁺-dependent Na⁺ channels, resulting in a reversal of the cirri (compound cilia responsible movement) beating direction that results in backward movement (Krüppel and Lueken, 1990; Krüppel et al., 1995). The V_m of E. vannus is returned to resting level by the efflux of K⁺ from slower voltage-gated K⁺ channels (Krüppel et al., 1995). As with RDs, the involvement of voltage-gated channels in mediating MSDs in Favella sp. is suggested by the fact that transient depolarizations and reversals may be elicited by the injection of positive current (Echevarria et al., 2014).

Ciliary reversals are rarely elicited upon contact of Favella sp. with smaller non-preferred prey; therefore, smaller particles (<4 µm) do not elicit aggregative changes in swimming in swimming behavior (Buskey and Stoecker, 1988; Stoecker et al., 1995). This suggests that smaller particles do not produce enough force on contact with membranelles to activate mechanosensitive ion channels. The mechanosensitivity of Favella sp. and other alveolates may therefore be tuned to sense optimally sized prey items, supporting the role of MSD-mediated ciliary reversals in capturing and encountering prey.

In contrast with ciliary contact by particles, direct peristomal contact resulted in a lag period before initiation of MSD (Fig. 10B). This may allow particles in the peristomal cavity to undergo a processing period during which they are sampled for their surface properties and are subsequently ingested, or rejected via ciliary reversal and backward swimming. The longer response time to particles that contact the peristomal cavity indicates that Favella sp. possess different mechanisms that regulate mechanosensitivity at the cilia and peristomal cavity; for example, Favella sp. may have more sensitive or higher densities of mechanosensors located on cilia than the peristomal cavity. The spatial distribution of ion channels is known to be important in regulating complex behaviors in Euplotes sp. and Paramecium (Machemer and Ogura, 1979; Krüppel et al., 1993). Particle processing at the peristome most likely includes crosstalk between mechanosensory and chemosensory signaling pathways. The lag between particle contact and depolarization-mediated ciliary reversals may give ciliates sufficient time to assess particles with chemosensors on the peristomal cavity and determine whether they are suitable for ingestion. Further work is now required to understand the spatial distribution of sensory mechanisms and how this allows alveolate grazers to perceive mechanical and chemical prey cues during processing and ingestion.

Unlike cilia, detection of particles at the peristomal cavity likely involves chemical information, as 15 μm microspheres were rejected, unlike similarly sized H. triquetra (Fig. 10B). Chemical recognition of prey particles is known to be important in allowing Favella sp. to identify and ingest favorable prey. For example, Favella sp. are able to recognize and reject toxic Heterosigma akashiwo upon contact, indicating that unfavorable surface compounds may be present on H. akashiwo (Taniguchi and Takeda, 1988), while ingesting similarly sized H. triquetra. The mechanism by which chemical recognition occurs in Favella sp. is unknown, but is likely similar to those known for other ciliates. Sugar–lectin interactions at the interface of prey and predator cell surfaces seem to be a common and important method of prey particle identification by ciliates and other planktonic alveolates (Scott and Hufnagel, 1983; Esteve, 1984; Casci and Hufnagel, 1988; Sakaguchi et al., 2001; Wilks and Sleigh, 2004; Roberts et al., 2006; Wootton et al., 2007). Surface proteins that are attached to the extracellular surface by glycosylphosphatidylinositol are also important in predator–prey recognition processes in freshwater ciliates (Simon and Kusch, 2013).

Recognition of favorable prey particles most likely involves cross-talk between metabotropic (enzyme-linked or second messenger-linked) signaling and ionotropic signaling pathways. G-protein coupled receptor (GPCR) pathways may be important in linking these signal transduction mechanisms (Echevarria et al., 2014). GPCR signaling pathways are crucial for chemosensation of prey in ciliates, as demonstrated in experiments with the marine ciliate Uronema sp., where chemosensory responses to prey were decreased by application of pharmacological compounds that inhibited the GPCR signal transduction pathway (Hartz et al., 2008). These GPCR signaling pathways provide a feedback link to modulate activity of ion channels through the activity of cyclic AMP, and thus the behaviors described above that these channels regulate. For example, genes for adenylyl cyclase were cloned from P. tetraurelia and found to localize to cilia, where they potentially regulate K⁺ channels that regulate ciliary movements (Weber et al., 2004). One potential consequence of such a pathway may be the ingestion-mediated hyperpolarization observed in Favella sp., which to our knowledge has not been described in the literature. Hyperpolarization may make the cell less excitable, decreasing the likelihood that either RDs or MSDs will be activated that would cause particle ejection from the peristomal cavity during ingestion, as observed by Taniguchi and Takeda (1988), where toxic H. akashiwo cells were rejected by Favella sp. (Fig. 10C). Thus, membrane hyperpolarization during ingestion may be a general mechanism for allowing planktonic alveolates to process food particles.

Our research represents the first demonstration of behavioral interactions of Favella sp. with copepod predators. Previous research suggests that copepods may extract Favella sp. from the lorica during predation (Stoecker and Sanders, 1985), indicating that the lorica is an awkward size or shape for ingestion or is non-nutritive (the lorica is highly refractory in nature and does not degrade in the presence of strong acids and bases; Agatha et al., 2013). We showed that upon capture, Favella sp. contracted into the lorica and were often subsequently released. Therefore, contraction into the lorica may prevent copepods from detecting the nutritive cell inside and impede extraction. Upon release by copepods, Favella sp. remained contracted for several seconds, not beating their cilia, instead sinking through the water column. This might decrease their hydrodynamic signature and decrease their encounter rate with predators that still occupy the area. These results suggest that mechanostimulated contraction into the lorica is a predator evasion behavior that could be common in the plankton given the prevalence of loricated organisms.

Contraction into the lorica was directly regulated by APs (Fig. 10D). Evoked AP-mediated contractions in tethered cells and contractions caused by predation attempts in free-swimming cells were approximately the same length of time, indicating that similar molecular mechanisms mediate both processes. The ionic mechanisms underlying changes in V_m during AP production are unknown, although voltage-activated channels are most likely involved because current injection resulted in rapid (ms) excitation–contraction coupling. In addition, spontaneous depolarizations in low [K⁺]_ext changed into AP-RDs in higher [K⁺]_ext, a change that is likely due to a shift in V_m sufficient to reach the threshold for voltage-gated cation channels responsible for APs. How predation attempts trigger the escape response is unclear, although similar classes of mechanosensitive ion channels that are involved in production of MSDs during interactions with prey are most likely activated on contact of Favella sp. with predators. However, stronger mechanostimulus by predators may activate more ion channels, or a different class of mechanoreceptors, resulting in a greater depolarization that reaches the threshold necessary for all-or-nothing APs (Fig. 10D). Such threshold-dependent sensory mechanisms likely underpin distinct prey capture and predator evasion behaviors exhibited by alveolate grazers.

Although there is a clear relationship between fast APs and whole-cell contractions in Favella sp., the mechanism is yet to be determined. A similar excitation–contraction mechanism has been well studied in the benthic ciliate Vorticella sp., albeit in a non-ecological context. Stimulation of mechanoreceptors in the cell body results in an all-or-nothing rise in Ca²⁺, immediately followed by an all-or-nothing rise in V_m and contraction of Vorticella sp. along its stalk (Katoh and Kikuyama, 1997; Shiono and Naitoh, 1997). This rapid rise in Ca²⁺ is due to an influx of Ca²⁺ from membranous tubules that surround the stalk via a calcium-induced calcium release mechanism. Calcium interacts with contractile proteins in the stalk (spasmins), resulting in contraction (Routledge, 1978; Katoh and Naitoh, 1994; France, 2007). Whether a similar mechanism functions in Favella sp. remains to be tested (Echevarria et al., 2014). Efflux of K⁺ ions may be important in returning cells to resting V_m following APs. This is indicated by the abolishment of after-hyperpolarizations in AP-RDs observed in 40 mmol l⁻¹ KCl ASW that would result from a decrease in the driving force of K⁺ due to an increase in [K⁺]_ext.

Favella sp. exhibited complex and diverse bioelectrical and behavioral properties, with responses to stimuli dependent on their type and location. Additionally, bioelectrical responses and resulting behaviors were intensity-dependent: from single-particle-triggered MSD and ciliary reversal, to high-frequency APs that resulted in extended contraction into the lorica. This may allow fine-tuned behavioral responses based on the strength of stimuli. Overall, these findings show that Favella sp. have sophisticated cellular mechanisms that enable them to perceive predators and prey, discriminate between them and respond with appropriate behaviors (Echevarria et al., 2014).

With the techniques developed in this study, it is now possible to examine the behavioral and bioelectrical responses of Favella sp. to a range of chemical, mechanical (contact) and hydromechanical (non-contact) cues from both predators and prey, and ultimately the ionic mechanisms and downstream signaling processes such as GPCR signal transduction pathways and calcium-induced calcium release mechanisms that link bioelectrical responses to behavior. These mechanisms underpin the integration of multiple environmental cues to maximize fitness in these single-celled organisms that are crucial consumers of phytoplankton and play important roles in the marine microbial food web.

We thank Suzanne Strom for helpful discussion and for Favella sp. strains. We thank Chris Finelli for loan of the Fastec camera for high-speed video microscopy and Michael Finiguerra for providing Tetraselmis sp. cultures.