The four-celled Volvocales green alga Tetrabaena socialis exhibits weak photobehavior and high-photoprotection ability

Photo-induced behavioral responses (photobehaviors) are crucial to the survival of motile phototrophic organisms in changing light conditions. Volvocine green algae are excellent model organisms for studying the regulatory mechanisms of photobehavior. We recently reported that unicellular Chlamydomonas reinhardtii and multicellular Volvox rousseletii exhibit similar photobehaviors, such as phototactic and photoshock responses, via different ciliary regulations. To clarify how the regulatory systems have changed during the evolution of multicellularity, we investigated the photobehaviors of four-celled Tetrabaena socialis. Surprisingly, unlike C. reinhardtii and V. rousseletii, T. socialis did not exhibit immediate photobehaviors after light illumination. Electrophysiological analysis revealed that the T. socialis eyespot does not function as a photoreceptor. Instead, T. socialis exhibited slow accumulation toward the light source in a photosynthesis-dependent manner. Our assessment of photosynthetic activities showed that T. socialis chloroplasts possess higher photoprotection abilities against strong light than C. reinhardtii. These data suggest that C. reinhardtii and T. socialis employ different strategies to avoid high-light stress (moving away rapidly and gaining photoprotection, respectively) despite their close phylogenetic relationship.


Introduction
Light is the essential energy source for phototrophic organisms. However, excess light energy leads to the production of toxic reactive oxygen species and possibly to cell death [1]. Phototrophic organisms have several photoprotection systems to prevent this negative aspect of light, which are induced by light as a signal. Non-photochemical quenching (NPQ) is such a system in the chloroplasts of plants and algae, which mitigates photo-induced oxidative stress [2]. The major component of NPQ is the energy-dependent quenching (qE) that dissipates excess light energy absorbed by photosystems through conversion of light as heat. In the model unicellular green alga Chlamydomonas reinhardtii, qE is facilitated by LHCSR (light harvesting complex, stress related) proteins, whose expression is triggered by light illumination [3,4] or exposure to low temperature [5]. In contrast, constitutively expressed PsbS protein regulates qE in the vascular plants and thus qE is rapidly induced [6]. Unlike immotile land plants, motile phototrophic organisms can physically move in response to light [7] and locate themselves under suitable light conditions for photosynthesis. Such behaviors are called photo-induced behavioral responses or photobehaviors. Typical photobehaviors include phototaxis and the photoshock response. Phototaxis describes the behavior of an organism moving toward or away from a light source (called positive and negative phototaxis, respectively). When an organism stops moving or reverses the motion direction upon a sudden change in light intensity, the behavior is referred to as the photoshock response. Both behaviors are widely observed among algae and are thought to be critical for survival.
Volvocine green algae are excellent model organisms for investigating photobehavior mechanisms. Among them, unicellular Chlamydomonas reinhardtii has long served as a reference organism for studying photobehavior, photosynthesis, and cilia (a.k.a. flagella) [8][9][10][11]. Each C. reinhardtii cell possesses one eyespot for photoreception and two cilia that generate force for swimming. The molecular mechanisms of C. reinhardtii photobehavior have been described in detail. Typically, the cilia beat ahead of the cell in opposite directions, allowing the cell to move forward using a movement pattern akin to the human breaststroke (Fig 1A, left). When the eyespot senses light, cation influx via channelrhodopsins (ChRs) localized at the eyespot trigger intraciliary Ca 2+ concentration increases [12]. The cilia then change their beating mode in a Ca 2 + -dependent manner, and the cell exhibits the photobehavior. Phototaxis occurs when cells sense relatively weak light stimulation, raising the intraciliary Ca 2+ concentration to~10 −7 M [13]. The beating frequency and amplitude of the cilia become unbalanced, and the difference in force generated by each cilium results in a directional change in the cell's motion [14,15]. The photoshock response occurs when cells sense strong light stimulation, which induces depolarization of the plasma membrane and raises the intraciliary Ca 2+ concentration to~10 −4 M [16]. The waveform of cilia then converts from asymmetrical to symmetrical, allowing the cell to swim backward with the cell body ahead of the cilia [17,18].
Like C. reinhardtii, multicellular volvocine algae are excellent organisms for studying photobehavior mechanisms. Volvox species are colonial algae, with thousands of cells comprising a single large spheroid. Each Volvox cell looks similar to C. reinhardtii, possessing one eyespot and two cilia. However, unlike C. reinhardtii, the two cilia on each cell beat in the same direction, i.e., to the spheroid's posterior pole (Fig 1A, right) [19]. Upon photoreception at the eyespot, cilia of the cells stop beating (e.g., Volvox carteri) or rotate their beating plane without changing the waveform [e.g., V. rousseletii) to change the spheroid direction of motion [20][21][22][23]. Within Volvox spheroids, cells are gradually differentiated from the anterior pole to the posterior pole. In the case of V. rousseletii, the rotation angle of the beating plane is~180˚at the anterior pole,~90˚degrees at the equator, and~0˚degrees at the posterior pole [20,23]. Because of this differentiation, the anterior hemisphere is specialized for steering, while the posterior hemisphere is specialized for force generation. This steering and force labor differentiation enables Volvox spheroids to rapidly exhibit photobehaviors.
Volvocine algae evolution is unique in that modern multicellular (i.e., Volvox) and unicellular (i.e., Chlamydomonas) taxa both evolved from a Chlamydomonas-like unicellular ancestor, showing morphological diversity within the clade [24][25][26]. Additionally, C. reinhardtii and Volvox spp. exhibit vastly different regulatory mechanisms for photobehaviors, as described above. Thus, raising the question of how these differences arose during the evolution of multicellularity.
Tetrabaena socialis is a volvocine alga, described as the simplest known multicellular organism, comprised of only four cells [27,28]. T. socialis looks like four C. reinhardtii cells arranged in the same orientation, forming a square (Fig 1B and 1C). In this study, we analyzed T. socialis photobehaviors to elucidate the first step in the evolution of photobehavior regulation in multicellular volvocine algae. Surprisingly, unlike C. reinhardtii or Volvox spp., T. socialis does not exhibit clear phototaxis or photoshock responses. Instead, we observed a slow accumulation of T. socialis colonies toward the light source in a photosynthesis-dependent manner.  (left) and Volvox rousseletii (right). In C. reinhardtii, two cilia beat in opposite directions to propel the cell forward. For phototaxis, the force generated by two cilia becomes unbalanced, allowing the cell to change its swimming direction. For the photoshock response, the cilia waveform changes from asymmetrical to symmetrical, allowing the cell to swim backward. In V. rousseletii, all cilia beat toward the posterior pole of the spheroid for forward swimming. For phototaxis, the beating direction of cilia in the anterior hemisphere on the lightsource side reverses. The force generated by cilia between the light-source hemisphere and the opposite hemisphere becomes imbalanced, changing the swimming direction of the spheroid. For the photoshock response, the beating direction of almost all cilia in the anterior hemisphere reverses. The forces generated by the anterior hemisphere and posterior hemisphere are balanced, stopping the spheroid motion. (B) A differential-interference-contrast (top) and a bright-field (bottom) images and (C) schematics of T. socialis NIES-571 colonies. Four cells are arranged in a square. Each T. socialis cell resembles a C. reinhardtii cell, possessing two cilia and one eyespot (arrowheads in the bottom panel). Scale bars: 10 μm. https://doi.org/10.1371/journal.pone.0259138.g001 Electrophysiological analyses revealed that the T. socialis eyespot does not function as a photoreceptor. Intriguingly, NPQ (qE) of T. socialis is constitutive like plants and higher than that observed in C. reinhardtii. These data suggest that T. socialis has a more effective photoprotective system than C. reinhardtii. Although T. socialis and C. reinhardtii belong to the same order, our findings suggest they employ diverse survival strategies.

T. socialis does not exhibit significant photobehavior
Before comparing the photobehaviors of T. socialis and C. reinhardtii, we measured the swimming velocity and the ciliary beating frequency (CBF) of T. socialis, which were 137 ± 27 μm/sec and 16 ± 3 Hz, respectively (S1 Fig). Wild-type C. reinhardtii cells typically swim at~150 μm/ sec with a CBF of~60 Hz [8]. If a sphere with a cross-sectional area about four times that of C. reinhardtii swims with the propulsive force of four C. reinhardtii cells, the estimated swimming velocity would be~255 μm/sec, according to Stokes' formula (S2 Fig). Thus, the actual swimming speed of T. socialis was slower than that predicted from its cell number and shape.
To investigate T. socialis photobehaviors, we tested phototaxis using two strains of T. socialis (NIES-571 and ISA2-2) and wild-type C. reinhardtii. Algae cultures were placed in Petri dishes and illuminated with green LED (λ = 525 nm, 30 μmol photons m −2 s −1 ), which prominently induces phototaxis in C. reinhardtii (Fig 2A) [29-31]. However, no significant accumulation toward the light source was observed in T. socialis NIES-571, and only a slight accumulation was observed in ISA-2-2 (Fig 2A). We also observed the motion of T. socialis colonies (S1 Movie) and C. reinhardtii cells (S2 Movie) under a microscope and measured the angle between the light beam and the swimming trajectories. Still, we did not observe phototactic motion in T. socialis colonies in contrast to C. reinhardtii cells (Fig 2B). Phototactic indices (to examine whether colonies showed positive or negative phototaxis) and parallel indices (to examine whether colonies swam in parallel to the light beam) calculated from those angles provided no evidence of phototaxis in T. socialis colonies (Fig 2C and 2D). Thus, the slight photoaccumulation (accumulation of organisms in the presence of light) observed in ISA2-2 may be attributed to a phenomenon other than phototaxis.
Next, we tested the photoshock responses of T. socialis and C. reinhardtii. Changes in the angles (θ) between swimming trajectories of cells/colonies before and after the flash illumination were measured ( Fig 3A). Immediately after the flash illumination, C. reinhardtii cells swam backward for a short period (θ = 136.0 ± 47.7 degrees) ( Fig 3B and 3C and S3 Movie) [26]. The swimming trajectories of C. reinhardtii cells formed V-shapes with the apex at the flash illumination point (Fig 3B). In contrast, T. socialis colonies did not change the swimming direction as C. reinhardtii upon the flash illumination (θ = 26.0 ± 6.7 degrees for NIES-571; 8.2 ± 2.5 degrees for ISA2-2) (Fig 3B and 3C and S4 Movie). These data provided no evidence of the photoshock response in T. socialis colonies.

Photoreceptor currents not detected in T. socialis eyespots
C. reinhardtii photobehaviors occur via changes in the cilia beating patterns after photo-sensing at the eyespot [32,35]. Thus, we proposed that the lack of photobehavior in T. socialis could be due to defects in the cilia or the eyespot. To test whether T. socialis cilia possess the waveform conversion ability that enables the C. reinhardtii photobehaviors, we demembranated T. socialis colonies using non-ionic detergent. After preparing the demembranated cell models, we reactivated their motility by adding ATP in a high-Ca 2+ buffer [13,23]. The cilia of live T. socialis showed asymmetrical waveform, whereas demembranated T. socialis cilia showed symmetrical waveform in the 10 −3 mM Ca 2+ buffer (S3 Fig). These data show that T. socialis cilia have retained the Ca 2+ -dependent waveform conversion capability. Observations of spontaneous backward swimming in live T. socialis colonies further support this finding (S5 Movie).
Next, we tested whether defects in the eyespot were responsible for the lack of photobehaviors in T. socialis. The volvocine algae eyespot comprises two parts: ChRs functioning as photoreceptors and carotenoid-rich granule layers acting as light reflectors [30,36,37]. In C. reinhardtii, the light-activated cation channel activities of ChRs can be electrophysiologically detected as photoreceptor currents (PRC) (S4 Fig) [38]. Each T. socialis cell contains one detectable red spot, and a previous study confirmed that the eyespot structure is intact ( Fig 1B) [27]. We measured the PRC of two strains of T. socialis and C. reinhardtii as a control. C. reinhardtii cells showed a prominent PRC signal after a flash illumination, which is similar to the previous studies using the same experimental setup (Fig 4) [38,39]. However, no PRC signal was detected in either T. socialis strain (Fig 4). These data suggest that the T. socialis cell eyespots do not function as photoreceptors. A ChR ortholog was identified in the T. socialis

Photosynthesis-dependent photoaccumulation of T. socialis
As described above, T. socialis did not display typical photobehaviors. However, the ISA2-2 strain colonies accumulated slowly toward the green light source (photoaccumulation) without showing phototaxis (Fig 2A). The absorption peak of C. reinhardtii ChR1 is~470 nm, which is close to a chlorophyll b absorption peak (~460 nm) [12,41]. As photosynthetic activity modulates phototaxis signals, green light (~525 nm) is preferred for phototaxis analyses to minimize the effects of photosynthesis [42,43].
We surmised that ISA2-2 photoaccumulation may be related to photosynthesis and conducted a phototaxis assay using a red LED (λ = 640 nm, 30 μmol photons m −2 s −1 ) to test our hypothesis. When we used red light for illumination, ISA2-2 colonies accumulated toward the light source in~15 min but C. reinhardtii cells and NIES-571 colonies did not ( Fig 5A). Our cell/colony-level observations suggest that the accumulation could not be attributed to phototaxis: neither the phototactic index nor the parallel index in ISA2-2 showed significant difference from the random swimming (Fig 5B-5D). Interestingly, the photoaccumulation of ISA2-2 was inhibited by treatment with photosynthesis inhibitor DCMU ( Fig 5E) but not ribosome inhibitor cycloheximide (CHX) (Fig 5F). Together, these findings indicate that photoaccumulation of T. socialis ISA2-2 toward red light is photosynthesis-dependent and does not require de novo protein expression.
So, how do T. socialis ISA2-2 colonies photoaccumulate without showing phototaxis or photoshock responses? We hypothesized that colonies might alter their swimming trajectory a 30-s illumination. (C) Phototactic indices (average of cosθ) calculated from (B). If all cells/colonies show positive or negative phototaxis, the value will be 1 or −1, respectively. If all cells/colonies swim in random directions, the value will be 0. Asterisk represents a significant difference from random swimming (p < 0.01; Student's t-test). (D) Parallel indices (average of |cosθ|) calculated from (B). If all cells/colonies exhibit phototaxis (positive or negative) or swim in a completely random direction, the values would be 1 or 0.622, respectively. Asterisk represents a significant difference from random swimming (p < 0.01; Student's t-test). https://doi.org/10.1371/journal.pone.0259138.g002

High non-photochemical quenching (qE) in T. socialis
Photobehaviors are believed to be crucial for survival under changing light environments. In particular, strong light conditions can lead to photo-oxidative damage in the phototrophic organisms, resulting in excessive reactive oxygen species generation [1]. We predicted that T. socialis might have a more effective photoprotection system than volvocine relatives, allowing the species to survive in environments with strong and varied light without showing immediate photobehaviors.
NPQ is a typical chloroplast photoprotection system that dissipates excess absorbed energy from the photosystems [2,45]. The major component of NPQ is energy-dependent qE [46], facilitated by the LHCSR protein expression in green algae [47,48]. In C. reinhardtii, the expression of LHCSR3 and LHCSR1, the major factors for qE, is induced by blue or UV illumination, respectively [3,48]. To evaluate the qE capability of the algae as well as the photosynthetic capability, we illuminated cells with UV, blue, green, or red light. The low-temperature exposure was also tested to evaluate the effects of growth temperature difference in C. reinhardtii (23˚C) and T. socialis (15˚C). Following these treatments, we measured qE as well as other photosynthetic parameters Fv/Fm (maximal photochemical efficiency of PSII representing photosynthetic efficiency) and φII (effective quantum yield of PSII) by a chlorophyll fluorometer (Fig 6). Under visible high-light conditions, Fv/Fm values in T. socialis were significantly higher than in C. reinhardtii, especially at 470 nm and 660 nm (Fig 6A). Moreover, qE values in T. socialis were relatively stable among the conditions and significantly   (Fig 6B). In contrast, φII values were not different between species under all conditions tested (Fig 6C), suggesting that the higher Fv/Fm values in T. socialis than C. reinhardtii were not due to the differences in photochemical quenching. Altogether, these data suggest that T. socialis possesses a higher photoprotection ability than C. reinhardtii.
These qE phenomena imply that T. socialis might constitutively accumulate LHCSR proteins, where C. reinhardtii does not. To test this idea, we assessed the expression level of LHCSR proteins in the algae under the same light/temperature conditions as used in Fig 6 by  western blotting (Fig 7). LHCSR3-and LHCSR1-orthologs are present in T. socialis transcriptome database constructed by the data from NIES-571 [40], though the LHCSR3-like protein in T. socialis is slightly smaller than C. reinhardtii LHCSR3 (S7 Fig). These proteins possess a peptide sequence similar to the epitope of the antibody that detects both C. reinhardtii LHCSR3 and LHCSR1 (S7 Fig) [49]. As reported previously, the expression of LHCSR3 and LHCSR1 in C. reinhardtii was prominently induced by blue or UV illumination, respectively [3,48], and slightly by low light illumination at 15˚C (Fig 7) [5]. In contrast, in T. socialis NIES-571, the band reacted with the anti-LHCSR protein antibody (probably two bands of LHCSR3-like and LHCSR-1 like proteins overlap) were detected in all the conditions tested, even without blue-or UV-light illumination (Fig 7). Moreover, in T. socialis ISA2-2, the bands were not detected in all the conditions (Fig 7). Together with the qE data in Fig 6B, these data suggest that molecular mechanisms to facilitate qE are different between C. reinhardtii and T. socialis and even between the strains in T. socialis.

Discussion
Our findings revealed that T. socialis does not display immediate photobehaviors after light illumination like C. reinhardtii and Volvox species. However, the photoprotection ability of T. socialis, represented by NPQ (qE), is higher than that of C. reinhardtii. These data indicate that T. socialis and C. reinhardtii employ different survival strategies, despite being closely related.

Weak photobehaviors of T. socialis
T. socialis did not show immediate photobehaviors, which occur within~1 s after photostimulation in C. reinhardtii and Volvox species (Figs 2 and 3) [20,35,50]. In C. reinhardtii and Volvox, the photobehavior mechanism involves Ca 2+ -dependent cilia movement pattern alterations after photostimulation [15,17,23,32,51]. The results of our demembranated model experiment suggest that T. socialis cilia have retained the ability for Ca 2+ -dependent waveform conversion (S3 Fig). Thus, the lack of immediate photobehaviors in T. socialis might be attributed to the loss of PRC (Fig 4). Comparing the amino-acid sequences of T. socialis, C. reinhardtii, V. carteri, and Gonium pectrale (S5 Fig), we found an amino-acid substitution in the ChR1 gene [12,52]. The mutated amino acid contributes to retinal binding in ChR [53], which may cause ChR dysfunction in T. socialis.
Although T. socialis did not show an immediate photoresponse, the ISA2-2 strain exhibited slow photoaccumulation in a photosynthesis-dependent manner (Fig 5). However, no evidence of phototaxis or photoshock responses was observed, and we could not clarify the underlying mechanism of the observed photoaccumulation. Time-lapse colony observations in a wide field of view will be necessary to elucidate how the colonies respond to light and change their swimming directions. Among the two T. socialis strains tested, NIES-571 did not show slow photoaccumulation (Fig 5). This ability may have been lost during the domestication of long-term cultures in the laboratory [9], as NIES-571 was isolated in 1982, while ISA2-2 was isolated in 2014.
Some photosynthesis-dependent cellular motions have been previously reported, such as movements of the nucleus and mitochondria in Arabidopsis thaliana cells [54,55]; however, the direct signal regulating the motion is unclear. In this study, the photoaccumulation of ISA2-2 was not inhibited by cycloheximide, suggesting that it was not induced by "retrograde signaling," wherein nucleus-encoded gene expression is regulated by signals from the chloroplast [56]. Reactive oxygen species have been reported to regulate ciliary beatings and phototactic signals [29,57], representing a candidate for further study on the photoaccumulation mechanism in ISA2-2.
To interpret the results of LHCSR expression in T. socialis, we considered two possibilities. First, qE is completely independent from LHCSR proteins in both strains of T. socialis, and LHCSR proteins in NIES-571 serve a different function from qE facilitation. Second, LHCSR proteins contribute to qE facilitation in NIES-571 strain, but their expression is constitutive experiments are shown. In (A) and (B), different letters indicate significant differences (P < 0.01, one-way ANOVA and Tukey honest significance difference [HSD]). In (C), no significant difference between any two groups was found in each parameter (P > 0.01, one-way ANOVA and Tukey honest significance difference [HSD]).
https://doi.org/10.1371/journal.pone.0259138.g006 regardless of light conditions. In this case, it is possible that ISA2-2 is a mutant with defects in LHCSR expression and some alternative pathway that facilitates qE is enhanced. In any case, further analyses will necessary to elucidate the difference of molecular mechanism of the qE facilitation in T. socialis and C. reinhardtii.
It is interesting to consider the weak-photobehavioral and high-photoprotective phenotypes in T. socialis in relation to its habitat. T. socialis was isolated from shallow freshwater areas or pools in Antarctic ice [58,59]. In such an environment, algae may be unable to escape or avoid strong sunlight by changing the swimming manner. We propose two hypotheses for the evolution of high-photoprotection function in T. socialis. First, T. socialis may have gained its high and constitutive photoprotection ability after multicellularity. The four-celled arrangement of T. socialis is not ideal for controlling swimming direction; thus, only those acquired mutations for high-photoprotection ability survived. Second, a Chlamydomonas-like unicellular ancestor may have gained high-photoprotection ability. Only the ancestor with this adaptation was able to evolve multicellularity, producing multicellular offspring that all possess high-photoprotection abilities, regardless of their photobehavior capabilities. In either case, T. socialis no longer required immediate photoresponses, and the ChR mutation may have been fixed.
The T. socialis phenotype paradoxically emphasizes the importance of immediate photobehaviors for photoprotection in microalgae. Photobehavior has been reported in various microalgae, though its importance in nature has not been clearly demonstrated. The negative correlation between photobehavior and photoprotection abilities in C. reinhardtii and T. socialis suggests that the photobehavior is essential for protecting photosystems in environments where light conditions fluctuate. Unlike C. reinhardtii, NPQ (qE) in land plants can always be induced because PsbS, an essential NPQ component, is constitutive [6]. It is interesting to note that phototrophic organisms with low-motility properties (land plants) and with weak-photobehavior (T. socialis) share a similar photoprotection capability (constitutive NPQ (qE)). However, a recent study suggested that NPQ is constitutive even in C. reinhardtii under conditions that mimic the wild environment [60]. The negative correlation between C. reinhardtii and T. socialis in terms of photobehavior and photoprotection abilities, as well as contribution of photobehavior to survival in the wild environment, will need to be tested by further analyses.
In conclusion, our results showed that T. socialis does not show immediate photobehaviors after light illumination, which may be due to the loss of function of the eyespot. Instead, it has a high and constitutive photoprotection capability. These properties of T. socialis contrast with relatively lower and induced photoprotection capability in C. reinhardtii, which exhibits agile photobehaviors. This study suggests the diversity of survival strategies of phototrophic organisms that differ even among closely related species.

Strains and culture
T. socialis NIES-571 and ISA2-2 were grown in standard Volvox medium (SVM) at 15˚C under a 12 h/12 h light/dark cycle using 20 μmol photons m −2 s −1 white light [61]. NIES-571 was isolated from Kanagawa prefecture, Japan, in 1982 [62]. ISA2-2 was isolated from Saitama prefecture, Japan, in 2014. C. reinhardtii strain CC-125 was grown in tris-acetate-phosphate (TAP) medium at 25˚C under a 12 h/12 h light/dark cycle using 20 μmol photons m −2 s −1 white light [63]. Before the photosynthetic parameter measurements, C. reinhardtii cells in the mid-log phase grown in TAP medium were collected a few hours after the beginning of the light period, resuspended in high-salt minimal medium (HSM) [64], and cultured by shaking at 23˚C under a 12 h/12 h light/dark cycle for 24 h.
To observe the swimming trajectories of T. socialis, we observed the colony suspensions under a bright-field microscope (BX-53, Olympus) with white light (10 μmol photons m -2 s -1 ). Half of the view field was covered with an ND10 filter. Videos of the swimming colonies were auto-tracked using Image Hyper software (Science Eye) for 6 s (3 s before and 3 s after swimming across the light-dark border or 6 s in the same light area), and angles were measured from the trajectories.

Photoshock response assay and measurement of ciliary beating frequency
T. socialis colonies and C. reinhardtii cells were washed using the method described for the phototaxis assay. The suspensions were kept under dim red light for 15 min, then cells/colonies were observed under a dark-field microscope (BX-53, Olympus) with dim red light (λ > 630 nm) and recorded using a high-speed camera (HAS-L2M, DITECT) at 400 fps. Videos of swimming cells/colonies were auto-tracked using Image Hyper software (Science Eye) and angles were measured from the trajectories before and after the flash illumination (white light; TT560 Speedlite, Neewer). Ciliary beating frequency was calculated from the time it takes for a cilium to beat once.

Photocurrent measurements
PRCs were assessed in populations of T. socialis colonies and C. reinhardtii cells according to the method of Sineshchekov et al. (1992) with modifications [38,66]. In brief, 1 ml of the T. socialis colony suspension (2.5 × 10 6 colonies/mL) or C. reinhardtii cell suspension (1.0 × 10 7 cells/mL) in a measuring solution (0.5 mM Hepes, pH 8.0, 0.1 mM CaCl 2 ) was added to a cuvette (10 × 10 × 15 mm) with one electrode on each side of its rectangular bottom. A flash light (λ = 500 nm, 1560 μmol photons m −2 s −1 ) was applied using an LED source (NSPE510S, Nichia Chemical) from one side of the electrode for 1 msec. The current was measured using a patch-clamp amplifier (Axoclamp 200B, Axon).

Photosynthetic parameter measurements
One hour after the beginning of the light period, chlorophyll contents in each alga were measured by the method of [67]. T. socialis colonies in a late-log phase and C. reinhardtii cells after replacement to HSM (see "Strains and culture") were harvested by centrifugation 720 × g for 3 min, resuspended in SVM (T. socialis) or HSM (C. reinhardtii) to the colony/ cell density in that the chlorophyll concentration was 2.5 μgChl/mL and cultured again for 4 h for recovery from the centrifugation stress. Algal cells were then irradiated with low light (

Western blotting of LHCSR proteins
Protein samples of whole-cell extracts (corresponding to~2.0 × 10 6 cells) were loaded onto 11% (w/v) acrylamide SDS-PAGE gels containing 7 M urea and blotted onto PVDF membranes. Antiserum against the beta subunit of ATP synthase (ATPB) control protein was obtained from Agrisera (AS05 085, rabbit polyclonal); antiserum against LHCSRs (detecting both LHCSR1 and LHCSR3) was raised and affinity-purified against the peptide LGLKPTDPEELK as reported previously [49]. An anti-rabbit horseradish peroxidase-conjugated antiserum (#7074, Cell Signaling Technology) was used as the secondary antibody. Blots were developed using EzWestLumi plus ECL detection reagent (ATTO), and images of the blots were obtained using a ChemiDocTouch System CCD imager (Bio-Rad Laboratories). The upper LHCSR3 band represents the phosphorylated form of LHCSR3 [3].

Reactivation of demembranated cell models and ciliary waveform trace
For C. reinhardtii, cell models were prepared using a previously described method [13]. For T. socialis, colonies were harvested by centrifugation, washed with SVM without Ca(NO 3 ) 2 and HES (10 mM Hepes-NaOH pH7.4, 5 mM MgSO 4 , 1 mM EGTA-NaOH pH 7.0, 4% sucrose) sequentially, and suspended in a demembranation buffer (30 mM Hepes-NaOH pH 7.4, 5 mM MgSO 4 , 1 mM EGTA-NaOH pH 7.0, 50 mM K-acetate, 0.015% Igepal CA-630 (I3021; Sigma-Aldrich), and 0.25 M DTT). Demembranated cells/colonies were suspended in a reactivation buffer containing 10 −3 M free Ca 2+ [23] and reactivated with final 1 mM ATP. Most ciliary axonemes were detached, and freely swimming axonemes were observed under a dark-field microscope (BX-53, Olympus) and recorded using a highspeed camera (HAS-L2M, DITECT) at 400 fps. Live cells/colonies were trapped on a surface of a glass slide coated with 0.1% polyethylenimine, and the ciliary beatings were observed and recorded using the same observation setup as above. The video of ciliary/axonemal beatings were played back frame-by-frame and the waveforms were manually traced by the mouse using PowerPoint (Microsoft). Microalgae inhabit low Reynolds number environments. In this situation, the force acting on the moving cell body is proportional to the product of the viscosity (η), cell size, and moving velocity (v); if we approximate the cell body as a sphere of radius r, then the viscous force acting on it is 6πηrv (Stokes' formula). Here, we approximate the cross-section of a C. reinhardtii cell body as a circle with its radius r c (left). We approximate the cross-section of the cell bodies of a T. socialis colony as a circle with its radius r t inscribed with four C. reinhardtii, and the colony is a sphere of radius r t , r t = r c / (