Detergent-extracted Volvox model exhibits ananterior–posterior gradient in flagellar Ca2+ sensitivityNoriko Uekia and Ken-ichi Wakabayashia,1

aLaboratory for Chemistry and Life Science, Institute of Innovative Research, Tokyo Institute of Technology, Yokohama-shi, Kanagawa 226-8503, Japan

Edited by Krishna K. Niyogi, Howard Hughes Medical Institute, University of California, Berkeley, CA, and approved December 8, 2017 (received for reviewSeptember 1, 2017)

Volvox rousseletii is a multicellular spheroidal green alga contain-ing ∼5,000 cells, each equipped with two flagella (cilia). This or-ganism shows striking photobehavior without any knownintercellular communication. To help understand how the behav-ior of flagella is regulated, we developed a method to extract thewhole organism with detergent and reactivate its flagellar motil-ity. Upon addition of ATP, demembranated flagella (axonemes) inthe spheroids actively beat and the spheroids swam as if theywere alive. Under Ca2+-free conditions, the axonemes assumedplanar and asymmetrical waveforms and beat toward the poste-rior pole, as do live spheroids in the absence of light stimulation. Inthe presence of 10−6 M Ca2+, however, most axonemes beat three-dimensionally toward the anterior pole, similar to flagella inphotostimulated live spheroids. This Ca2+-dependent change inflagellar beating direction was more conspicuous near the anteriorpole of the spheroid, but was not observed near the posteriorpole. This anterior–posterior gradient of flagellar Ca2+ sensitivitymay explain the mechanism of V. rousseletii photobehavior.

flagella | calcium | phototaxis | Volvox

Motile photosynthetic organisms require photobehavioralresponses (photoresponses) to inhabit environments suit-

able for photosynthesis. Green algae of the order Volvocalesswim with flagella protruding from each cell and display appro-priate behaviors in response to surrounding light. Organisms inthe section Volvox (= Euvolvox, a genus of Volvocales) (Fig.S1A) have unique characteristics, such as large cell number, acomplex developmental process (1), high swimming velocity (2),and marked photoresponses (3). They display two kinds ofphotoresponses: phototaxis and photoshock. Phototaxis is di-rectional swimming toward or away from a light source, whereasphotoshock is a transient stop or deceleration in response tosudden changes in light intensity. Each cell of both unicellularand multicellular species of the order Volvocales commonly usesits eyespot to sense light and generates propulsive force bybeating two flagella. The eyespot, an orange spot in the cell,consists of layers of carotenoid-rich granules in the chloroplastand photoreceptor proteins (channel rhodopsins) localized in theplasma membrane right above these granules. The carotenoidlayers function as a light reflector, making the eyespot a highlydirectional photoreceptive organelle (4).Regulation of flagellar beating in photoresponses has been

extensively studied in a unicellular species of Volvocales, Chla-mydomonas reinhardtii. A C. reinhardtii cell swims by beating itstwo flagella in asymmetrical/cilia-type waveforms and in oppositedirections (Fig. S2). Because the beating planes of the two fla-gella are slightly skewed from each other, the cell rotates whileswimming. During photoshock responses, the waveform tran-siently changes to a symmetrical/flagella-type pattern and the cellswims backward for a short period (Fig. S2). In phototaxis, theforce balance between the two flagella is dependent on the di-rection of light, causing the cell to turn toward or away from thelight source (Fig. S2). Both kinds of photoresponse are likely tobe mediated by changes in intraflagellar Ca2+ concentration, as

suggested by studies using demembranated and reactivated cellsand flagella (5, 6).Multicellular spheroidal species of Volvocales, including Volvox

species, have often been regarded as colonial Chlamydomonas.However, alignment of Chlamydomonas cells on the surface of aspheroid, with each cell displaying breaststroke-like flagellarbeating, would result in spheroids unable to swim in one direction.Unlike a C. reinhardtii cell, each cell in a Volvox spheroid has twoflagella beating in the same direction. A Volvox spheroid has ananterior–posterior (A–P) axis, and its ∼10,000 flagella beat towardthe posterior pole with a slight tilt from the A–P axis (7, 8) (Fig. 1A and B, Top). Flagellar beating causes the surrounding fluid toflow posteriorly around the spheroid so that the spheroid swimsforward while rotating bodily.Volvox spheroids and C. reinhardtii cells differ in their regu-

lation of flagellar beating during photoresponses. While aspheroid swims with bodily rotation, the light signal sensed by asingle cell increases when it faces the light source and decreaseswhen it faces the opposite side because of light reflection by thecarotenoid layers of the eyespot. Phototactic turning occurswhen the flagella on Volvox carteri and Volvox aureus cells facingthe light source stop beating (9, 10) (Fig. S1A). In large-spheroidspecies, such as Volvox rousseletii and Volvox barbari, the di-rection of flagellar beating changes when light is perceived by theanterior hemisphere in a posterior-to-anterior direction, while itsciliary waveform is retained (3, 10) (Fig. 1B, Bottom and Figs.S1A and S2). This flagellar response is called ciliary reversal.Importantly, the magnitude of this response has a gradient alongthe A–P axis, in that the response is more pronounced near theanterior pole of the spheroid and is rarely observed near itsposterior pole (3). This gradient may be produced by differencesin eyespot size, which is greater in cells near the anterior pole

Significance

The multicellular green alga Volvox rousseletii displays photo-taxis by changing its flagellar beating pattern in response tophotoreception. However, the molecular mechanism underlyingflagellar regulation is unknown. This study describes a methodto demembranate whole spheroids using a nonionic detergent,with the addition of ATP reactivating flagellar motility. Thesereactivated spheroids swam like live spheroids. Flagellar beatingdirection was altered in a Ca2+-dependent manner, with agreater change in the anterior hemisphere than in the posteriorhemisphere. These findings indicate that V. rousseletii has ananterior–posterior gradient of flagellar sensitivity to Ca2+, whichlikely plays a key role in V. rousseletii phototaxis.

Author contributions: N.U. and K.W. designed research, performed research, analyzeddata, and wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Published under the PNAS license.1To whom correspondence should be addressed. Email: wakaba@res.titech.ac.jp.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1715489115/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1715489115 PNAS | Published online January 8, 2018 | E1061–E1068

PLANTBIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Mar

ch 2

5, 2

020

and smaller in cells near the posterior pole (3). However, thesignal(s) that causes ciliary reversal is not known.In vitro reactivation of the motility of detergent-extracted cili-

ated organisms, such as Paramecium, C. reinhardtii, and sea urchinsperm, is a powerful method for studying the regulatory mecha-nisms of cilia and flagella (6, 11–13). However, application of thismethod to whole Volvox spheroids has been difficult, since thesespheroids are much larger in size (500–1,000 μm) than those ofother organisms studied to date. This study was designed to elu-cidate the mechanism underlying ciliary reversal and its gradientalong the A–P axis in V. rousseletii. Whole spheroids weredemembranated by detergent extraction and subsequently reac-tivated. Our results suggest that ciliary reversal in V. rousseletii isdependent on Ca2+ concentration, with the Ca2+ responsivenessof demembranated flagella having a gradient along the A–P axis.

ResultsReactivation of the Motility of V. rousseletii Axonemes.To reactivatethe flagellar motility of detergent-extracted V. rousseletii, we firsttested the method developed for C. reinhardtii. Although this

method was not successful, we found that several modificationsof the original protocol resulted in successful demembranationand reactivation (details are provided in Materials and Methods).(i) Because V. rousseletii spheroids cannot be efficiently con-centrated by centrifugation, buffer exchange entailed trapping(Fig. 2A) or a strainer-scooping method (see Fig. 5A) rather thancentrifugation. (ii) The concentration of nonionic detergent usedfor C. reinhardtii (0.1%) demembranation was too high forV. rousseletii, resulting in deflagellation during perfusion. Theoptimal concentration for V. rousseletii, as determined by thedegree of deflagellation and the efficiency of reactivation, wasmuch lower, in the range of 0.01–0.03%, and varied by de-velopmental stage (Figs. 1C and 2B). (iii) Polyethylene glycol(PEG) was omitted from the reactivation buffer. The buffer usedto reactivate C. reinhardtii cell models contained 0.5% PEG[molecular weight (Mw) = 20,000] to mimic the molecularcrowding effect in vivo (6), thereby increasing the reactivationrate. Although PEG increased the reactivation rate in V. rous-seletii, it tended to distort the spheroids (Fig. S3). PEG wastherefore added only when the trapping method was used (seeFig. 4). (iv) Although dark-field microscopy is often used tomonitor demembranated C. reinhardtii cells, the flagellar axo-nemes of V. rousseletii were not clearly visible by dark-field mi-croscopy because of strong halation from the spheroid. Thus,

Fig. 1. Life cycle and swimming style of V. rousseletii. (A) Schematic diagramsshowing the directions of flagella-generated fluid streams (blue arrows) arounda spheroid. The A–P current with a slight tilt causes the spheroid to swim whilerotating bodily in a counterclockwise direction when viewed from the posteriorpole. The dark gray circles in the spheroid represent gonidia that localize mainlyin the posterior hemisphere. (B) Schematic diagrams showing flagellar beating inthe somatic cells on the spheroid surface in an anterior region (box in A). Fla-gellar waveforms in posterior (Top) and anterior (Bottom) beating modes areshown. The blue and red arrows indicate the current generated in the twomodes. Modified from ref. 3. Note that the cell closest to the anterior pole isdepicted with a large eyespot (a red spot), with the eyespot size gradually de-creasing with distance from the pole. (C) Asexual life cycle of V. rousseletii cul-tured on a 16-h/8-h light/dark cycle. Spheroids in three stages (termed stages I–IIIin this study) were used: Stage I represents newly hatched spheroids, in whichgonidia (next-generation embryos) just start cell division (Left); stage II representsspheroids, in which gonidia are undergoing cleavage (Center); and stage IIIrepresents expanded spheroids, in which daughter spheroids have undergoneinversion and are ready to hatch (Right). A, anterior pole; P, posterior pole. (Scalebar: 200 μm; note that the bar length differs in the three photographs.)

Fig. 2. Detergent sensitivity of flagella along the A–P axis of spheroids atthe three developmental stages. (A) Schematic diagrams showing the trap-ping method for demembranation of Volvox spheroids. (i–iii) Drop of cul-ture containing several spheroids was placed on a glass slide. The mediumwas drained with a pipette, and washing solution (HMDEK; Materials andMethods) was added. (iv) After a few minutes, spheroids were withdrawnwith a pipette and (v and vi) placed in a perfusion chamber. (vii) Top view ofthe chamber. Solution containing either Igepal or ATP was perfused.(B) Degree of demembranation in the three stages of spheroids treated withdifferent concentrations of detergent. The demembranation index was cal-culated for the anterior or posterior hemisphere as follows: If visual in-spection showed that all flagella in the examined area were stopped afterdetergent perfusion, the index was 1; if any flagella were moving, the indexwas 0. Index values were counted in three to 26 spheroids, with the averagedefined as the demembranation index. The arrowhead in stage I indicatesthe conditions used for Fig. 5, and the arrows in stage II indicate the con-ditions used for Figs. 3 and 4. The filled arrow and squares are located in theanterior hemisphere, and the open arrow and circles are located in theposterior hemisphere.

E1062 | www.pnas.org/cgi/doi/10.1073/pnas.1715489115 Ueki and Wakabayashi

Dow

nloa

ded

by g

uest

on

Mar

ch 2

5, 2

020

demembranated V. rousseletii was monitored by phase-contrastmicroscopy.The optimal concentration of a nonionic detergent, Igepal

CA-630 (a substitute for Nonidet P-40; hereinafter referred to asIgepal) was determined by washing spheroids with a buffer andtrapping them between a glass slide and a coverslip. The samplewas then gently perfused with Igepal-containing buffer fordemembranation (Fig. 2A). Determinations of the demembra-nation rate (defined as 1 when no flagella moved and 0 when atleast one flagellum moved after demembranation in a givensector of a spheroid) indicated the presence of an A–P gradientin Igepal sensitivity throughout the developmental stages (Fig.S4). Cells at the anterior pole were the most sensitive to Igepal(i.e., their flagella stopped beating at lower concentrations), andthose at the posterior pole were the least sensitive (Fig. 2B andFig. S4). To reproducibly repeat experiments, we selected threeeasily distinguishable stages in development, termed stages I–III(Fig. 1C). Stage I consisted of newly hatched spheroids, withflagella most sensitive to Igepal (Fig. 2B). Stage II consisted ofmiddle-aged spheroids whose gonidia were in the midst of celldivision, with cells in the posterior hemisphere becoming lesssensitive than in stage I. Stage III consisted of the expansionstage, when gonidia finished inversion, with the Igepal concen-tration necessary to demembranate posterior flagella being thehighest among the three stages.We then determined the optimal conditions to reactivate the

motility of demembranated flagella and spheroids. Because highconcentrations of Igepal cause deflagellation, lower concentra-tions of Igepal were desirable. Stage II spheroids are best suitedfor monitoring flagella because the distance between flagella onthe spheroidal surface is longer than in stage I, which allowsmonitoring of their waveforms, and because their gonidia aresmaller than in stage III, making the A–P axis clearly identifiable.Stage I, however, is optimal for monitoring whole spheroids,because the same low concentration of Igepal (0.01%) effectivelydemembranates both anterior and posterior hemispheres (Fig.2B, arrowhead). Thus, high-speed monitoring of axonemesin trapped spheroids was performed using stage II spheroidsdemembranated with 0.01% (for anterior flagella) or 0.03%(for posterior flagella) Igepal (Fig. 2B, closed and open arrows,respectively). Reactivation of whole detergent-extracted spher-oids was assessed using stage I spheroids demembranated with0.01% Igepal.The axonemal motility of detergent-extracted V. rousseletii

spheroids (DEVs) was successfully reactivated by treatment withan ATP-containing buffer. The average beat frequencies ofreactivated axonemes in DEVs, as determined using a high-speed camera, increased with ATP concentration in a mannerconsistent with Michaelis–Menten kinetics (Fig. 3 A and C). Theaverage beat frequencies of DEVs reactivated with >1 mM ATPwere higher than those of flagella in live spheroids, which were29.2 ± 2.7 Hz at the anterior pole and 25.1 ± 1.9 Hz at theposterior pole (Fig. 3B). Under all conditions tested in this study,the average beat frequencies of flagella/axonemes were slightlyhigher near the anterior pole than near the posterior pole (Fig. 3A and B). For example, the beat frequencies in DEVs at 1 mMATP were 41.8 ± 3.1 Hz near the anterior pole and 33.8 ± 0.9 Hznear the posterior pole. Using double-reciprocal plots, themaximal beat frequencies at the anterior and posterior poleswere 43.5 Hz and 48.8 Hz, respectively, and the apparentMichaelis constants were 0.10 mM and 0.22 mM ATP, respec-tively (Fig. 3C).

Calcium Control of Flagellar Stroke Direction. Our primary interestwas the identification of the factor that causes ciliary reversal inphotostimulated V. rousseletii flagella. The optimal candidatewas Ca2+, a major factor that modulates flagella/cilia beating in

various organisms. We therefore examined the effect of Ca2+ onaxonemal motility in the DEV.We first assessed flagellar motion in live spheroids that

responded to photostimulation. Under continuous light, all fla-gella on a spheroid beat in a posterior direction in a ciliarywaveform [Fig. 4 A, Left (setup A), B, Center (setup B), and C,Right (setup C) and Movie S1]. After photostimulation (Materialsand Methods), flagella in the anterior hemisphere showed ciliaryreversal, with flagella beating in an anterior direction for a shortperiod of time, usually 2–3 s (Movie S2). The beating direction ofa ciliary-type waveform can be determined from the appearanceof a typical hook shape in a series of waveforms. The change inbeating direction from before to after photostimulation was mostclearly observed near the anterior pole, with none observed nearthe posterior pole (Fig. 4 B and C and Movies S3 and S4). Asurface view of a region close to the anterior pole, as observedfrom within the spheroid [Fig. 4 A, Center (setup B) and D andMovie S5], showed that, in response to photostimulation, thebeating direction of flagella gradually changed toward the an-terior end after one or two strokes toward the posterior end. Thedirection of beating was almost completely reversed, but did notlast long, returning to the original direction in a few seconds.Similar to anterior flagella, flagella near the spheroid’s equator(as observed from outside) [Fig. 4 A, Right (setup C) and E andMovie S6] continued beating in a posterior direction for one or

Fig. 3. ATP-dependent beat frequency in the axonemes of detergent-extracted Volvox (DEV). (A) Beat frequency of axonemes near the anteriorand posterior poles in DEVs at different ATP concentrations. Three to sixaxonemes were measured. (B) Beat frequency of flagella in live cells near theanterior and posterior poles. The average in six flagella is shown for each.(C) Double-reciprocal plot of the data in A. Intercepts yielded apparent max-imal beat frequencies of 43.5 Hz (anterior region) and 48.8 Hz (posterior re-gion) and apparent Michaelis constants of 0.10 mM (anterior) and 0.22 mMATP (posterior).

Ueki and Wakabayashi PNAS | Published online January 8, 2018 | E1063

PLANTBIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Mar

ch 2

5, 2

020

two strokes after photostimulation, but gradually rotated strokedirection by ∼90°. These flagella returned to their original pos-terior direction more rapidly than the anterior flagella.

We next examined the beating direction of reactivated axo-nemes in DEVs in the absence and presence of Ca2+. UnderCa2+-free conditions, axonemes in reactivated DEVs beat in a

Fig. 4. Ca2+-dependent changes in the direction of axonemal beating. (A) Experimental setups for observation of live or demembranated spheroids in achamber. Glass slides and coverslips (gray), spacers (orange), the A–P (A↔P) direction of the spheroids, and the regions observed (red boxes) are indicated.(Left) In setup A, a spheroid was sandwiched between a glass slide and a coverslip, and regions near the anterior pole (A) and posterior pole (P) were ob-served. (Center) In setup B, a spheroid, after gentle pressing, was placed between the glass slide and a coverslip separated by thick spacers, and the regionnear the anterior pole was observed. (Right) In setup C, a spheroid was attached to the glass slide with 1% polyethyleneimine, and the region near theequator was observed. (B) Frames from high-speed recordings of regions near the anterior (Top) and posterior (Bottom) poles of a live spheroid. The ob-servation using setup A was under stationary conditions in continuous light (Left) and after photostimulation (Right). The direction of the flagellar effectivestroke (arrows) was determined by the direction of the typical hook shapes (red arrowheads) in the recovery stroke. After photostimulation, the direction ofthe flagellar effective stroke reversed in the anterior region. (Scale bar: 100 μm.) (C) Typical sequential flagellar waveforms in a single beating cycle undereach condition. Waveforms recorded as in B were traced (time interval of 1/500 s). The typical hook-shaped waveforms appearing in recovery strokes aretraced in magenta. (Scale bar: 10 μm.) Photographs show the time course of directional change in the flagellar effective stroke in live spheroids, near theanterior pole in setup B (D, Top) and near the equator in setup C (E, Top). (D and E, Bottom) Photographs show the time series of the change after pho-tostimulation in the boxed cell. The direction of ciliary beating and the time after onset of illumination are shown. (Scale bars: Top, 100 μm; Bottom, 10 μm.) InD, the direction of the effective stroke was almost reversed at 0.27 s and recovered at 4.44 s. The red asterisk indicates a presumptive anterior pole. In E, thedirection of the effective stroke rotated ∼90° at 0.72 s and recovered at 1.52 s. The two-headed arrow indicates the approximate A–P direction. (F) High-speedrecording of regions near the anterior (Top) and posterior (Bottom) poles of DEVs reactivated in the presence of 0 or 10−6 M Ca2+ in setup A. The ATPconcentration was 1 mM. In the presence of Ca2+, the axonemes in the anterior region showed anteriorly directed effective strokes, opposite to those in theabsence of Ca2+. However, this change was not observed in the posterior region. Arrowheads indicate the typical hook-shaped waveforms. (Scale bar:100 μm.) (G) Typical sequential axonemal waveforms traced for a single beating cycle under each condition (time interval of 1/500 s). The hook-shapedwaveforms are shown in magenta. (Scale bar: 10 μm.)

E1064 | www.pnas.org/cgi/doi/10.1073/pnas.1715489115 Ueki and Wakabayashi

Dow

nloa

ded

by g

uest

on

Mar

ch 2

5, 2

020

posterior direction (Fig. 4 F and G and Movie S7), similar toflagella in live spheroids in response to continuous light. In thepresence of 10−6 M Ca2+, however, the axonemes beat in ananterior direction, similar to flagella in the anterior regions ofphotostimulated live spheroids (Fig. 4 F andG and Movie S8). Inaddition, the reactivation rate and the beating amplitude becamesmaller in the presence of 10−6 M Ca2+ than under Ca2+-freeconditions. Strokes in the anterior direction were most clearlyvisible around the anterior pole, with axonemes near the poste-rior pole beating in a posterior direction, similar to those underCa2+-free conditions (Fig. 4 F and G and Movies S9 and S10).Perfusion with a buffer containing 10−3 M Ca2+ resulted in ax-onemes that displayed waveforms of smaller amplitude andlower beat frequency (Fig. S5). These waveforms, however, werenot observed in photostimulated live spheroids, suggesting thatthe maximum in vivo Ca2+ concentration after photostimulationis in the range of 10−6–10−4 M.Overall, our in vitro experiments using DEVs showed that (i)

ciliary reversal is a Ca2+-dependent event and (ii) cells and ax-

onemes on spheroids show a Ca2+ sensitivity gradient along theA–P axis.

Motility Reactivation of Whole Multicellular Spheroids of V. rousseletii.Finally, we examined whether the Ca2+ effect on axonemalbeating direction contributes to the changes in spheroid swim-ming velocities that occur during the photoshock response. Afterphotostimulation, live V. rousseletii spheroids show photoshockresponse, with their swimming velocity decreasing to 50–60%(3) (Fig. 5D). To test if this response can be reproduced invitro by changing the Ca2+ concentration, stage I spheroids weredemembranated with 0.01% Igepal by the strainer-scoopingmethod, which demembranates spheroids while preserving theiroverall morphology. No spheroids displayed movement in thisdemembranation solution (Fig. 5A and Movies S11 and S12).Subsequently, the spheroids in the demembranation solutionwere transferred to a buffer solution containing 0 M or 10−6 MCa2+ for reaction. Upon the addition of 1 mM ATP, most of theDEVs started to swim like live spheroids. The average swimming ve-locities, as calculated from the swimming tracks, were 588 ± 99 μm·s−1

Fig. 5. DEV swimming. (A) Schematic diagrams showing the strainer-scooping method for demembranation and reactivation. Spheroids were successivelytransferred from the culture medium to the washing, demembranation, and reactivation solutions, with ATP added during the final step to reactivatemotility. (B) Swimming trajectories of spheroids in live and demembranated/reactivated spheroids. Movements of the spheroids were video recorded andtracked for 5 s. (Left) Live spheroids under continuous light before demembranation. (Center and Right) DEVs reactivated with 1 mM ATP in the absence andpresence of 10−6 M Ca2+. (Scale bars: 5 mm.) (C) Swimming velocities (n = 20 for live spheroids, n = 30 for DEVs under each condition) calculated from theswimming trajectories for 5 s. DEVs that did not move at all were not counted. (D) Swimming velocities of live spheroids before and right after photo-stimulation calculated from the swimming trajectories for 3 s.

Ueki and Wakabayashi PNAS | Published online January 8, 2018 | E1065

PLANTBIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Mar

ch 2

5, 2

020

for live spheroids, 499 ± 97 μm·s−1 for DEVs reactivated inCa2+-free solution (∼85% of live spheroids), and 285 ±83 μm·s−1 for DEVs reactivated in buffer containing 10−6 M Ca2+

(∼49% of the velocity of live spheroids before photostimulationand ∼94% of the velocity after photostimulation) (Fig. 5 B and Cand Movies S11 and S12). These findings suggest that the pho-toshock response is caused by Ca2+-mediated ciliary reversal onthe anterior hemisphere.

DiscussionTo understand how changes in flagellar waveform are regulatedin V. rousseletii upon photostimulation, we developed a methodto demembranate whole spheroids with a detergent and reac-tivate their motility by addition of ATP. This method enabled usto examine the motility of “dead” multicellular spheroids in vitroand to assess the effects of Ca2+. We found that the flagellaraxonemes show ciliary reversal in a Ca2+-dependent manner andthat these axonemes are sensitive to Ca2+ with a gradient alongthe A–P axis, such that the cells near the anterior pole show thegreatest changes in axonemal motility.

Ca2+ Mediates Light-Induced Ciliary Reversal in V. rousseletii. Thebeating patterns of Volvox flagella change upon photostimulation.In V. aureus and V. carteri (Eudorina group), the most sensitivecells are localized at the anterior pole, at which flagella stopbeating in response to photostimulation (9, 10, 14–16). InV. rousseletii, cells near the anterior pole show ciliary reversal (3).However, the intracellular signals that trigger these flagellar re-sponses have remained unclear. Our results using DEVs stronglysuggest that ciliary reversal in V. rousseletii is induced by Ca2+

influx into the flagella in response to photoreception. At intra-flagellar Ca2+ concentrations <10−6 M, flagella beat in a posteriordirection, whereas at intraflagellar Ca2+ concentrations ≥10−6 M,flagella near the anterior pole beat in an anterior direction. Sim-ilar to C. reinhardtii, photoreception by channelrhodopsin at theeyespot may cause membrane depolarization, followed by Ca2+

influx through voltage-dependent Ca2+ channels localized at theflagellar tips (17).Ca2+ has been shown to modulate flagellar/ciliary beating in

various organisms. The modulation includes (i) waveform con-version, (ii) reversal of bend propagation, (iii) reversal of beatingdirection, (iv) rotation of beating direction, (v) increase in beatfrequency, and (vi) arrest of beating (18) (Fig. S1B). TheseCa2+-dependent modulations are responsible for various im-portant biological events, such as (i) waveform conversion insperm flagella for sperm chemotaxis toward eggs during fertiliza-tion and (ii) increase in beat frequency in mammalian trachealepithelial cilia for airway clearance (19, 20). Ciliary reversal hasbeen found in several organisms, including Ctenophora, Parame-cium, and sea urchin, and is regarded as (iii) reversal of beatingdirection in these organisms (11, 21, 22) (Fig. S1B). Our present

study has shown that ciliary reversal in V. rousseletii is (iv) rotationof beating direction with various rotation angles. In the presenceof Ca2+, V. rousseletii flagella near the anterior pole show almostfull reversal (rotation of ∼180°) of beating direction. The degreeof rotation of beating direction decreases with distance from theanterior pole, and becomes ∼90° at the equator and 0° near theposterior pole. The mechanism of rotation of beating direction hasbeen suggested to be different from that of reversal of beatingdirection in sea urchin larvae, and is not well understood to date(21). We may expect that identification of molecular/structuraldifferences between anterior and posterior flagella in V. rousseletiiwill provide clues to understand the molecular mechanism.

The A–P Gradient in Spheroids. The amplitude of Ca2+-dependentdirectional change in axonemal beating showed a gradient alongthe A–P axis, being high in cells close to the anterior pole andalmost zero near the posterior pole. The A–P gradient in fla-gellar photoresponse has been attributed to the gradient ineyespot size, with eyespots larger in cells near the anterior pole(3, 8, 15, 23). In V. rousseletii, our results clearly showed that theCa2+ sensitivity of the flagellar axonemes also displays an A–Pgradient. To date, several Ca2+-binding proteins have beenidentified in C. reinhardtii axonemes, although none has beenattributed as responsible for Ca2+ sensitivity in axonemal beat-ing. However, a promising candidate for the flagellar regulator isLC4, a Ca2+-binding subunit in outer arm dynein (18, 24). Thisprotein may show a gradient in concentration in properties alongthe A–P axis of V. rousseletii axonemes.The light sensitivity of an eyespot may increase with size, such

that a larger eyespot more readily promotes light-induced in-crease in cytoplasmic Ca2+ concentration. If so, V. rousseletiispheroids may have a strong A–P gradient in photosensitivity,resulting from the combined effects of two kinds of gradients:one in the mechanism that regulates light-induced Ca2+ influxand the other in the flagellar mechanism to change the waveformin a Ca2+-dependent manner. In addition, our study showed anA–P gradient in detergent sensitivity (i.e., the Igepal concen-tration required for complete cessation of flagellar beating), withthe detergent sensitivity in the posterior region decreasing withage. These findings indicate that flagellar membrane properties,such as the contents of particular lipids or membrane proteins,also differ along the A–P axis and with age. All of these featuresof V. rousseletii spheroids may jointly constitute a robust systemthat enables efficient photobehavior of this organism.

Volvox Evolution for Phototactic Steering. Results obtained in thisand previous studies indicate how flagellar regulatory mecha-nisms for phototactic steering have changed during evolutionfrom Chlamydomonas-like unicellular organisms to multicellularVolvox. At the cellular level, the perception of a sudden increasein light intensity by the eyespot and the increase in intraflagellar

Table 1. Ca2+-buffered reactivation solutions

Components, M Calculated concentrations, M

Approximate free [Ca2+] [CaCl2] [EGTA] [EDTA] [MgATP]

10−3 M 2.992 × 10−3 0 2.00 × 10−3 8.86 × 10−4

10−4 M 1.507 × 10−3 0 2.00 × 10−3 9.56 × 10−4

10−5 M 0.490 × 10−3 0 2.00 × 10−3 9.54 × 10−4

10−6 M 4.742 × 10−3 5.00 × 10−3 0 9.70 × 10−4

10−7 M 3.277 × 10−3 5.00 × 10−3 0 9.67 × 10−4

10−8 M 0.837 × 10−3 5.00 × 10−3 0 9.61 × 10−4

Ca-free 0 1.00 × 10−3 0 9.69 × 10−4

Final concentrations in reactivated samples. Components other than CaCl2, EGTA, and EDTA: 30 mM Hepes,5 mM MgSO4, 1 mM DTT, 50 mM potassium acetate, 1% PEG (average Mw = 20,000) (for experiments in Fig. 4)and 1 mM ATP. The pH of all reactivation solutions was adjusted to pH 7.4.

E1066 | www.pnas.org/cgi/doi/10.1073/pnas.1715489115 Ueki and Wakabayashi

Dow

nloa

ded

by g

uest

on

Mar

ch 2

5, 2

020

Ca2+ seem to be common to C. reinhardtii and V. rousseletii.However, the process that follows the intracellular Ca2+ increaseseems to differ significantly. In C. reinhardtii, flagellar beat fre-quency and amplitude of bending increase with increasing Ca2+

concentrations up to 10−6 M, causing phototactic turningthrough a change in the beating balance between the two flagella(6, 25, 26), and at 10−4 M Ca2+, when the photophobic responsetakes place, flagellar waveform is converted from a ciliary type toa flagellar type causing backward swimming (5). In V. rousseletii,the direction of ciliary-type beating is reversed at ≥10−6 M Ca2+

(Fig. 4 and Figs. S1A and S2). An important feature ofV. rousseletii behavior is that both photoshock response and pho-totaxis involve a Ca2+-induced ciliary reversal (3) (Fig. S2).When photostimulation is sufficiently strong for all cells in aspheroid, all flagella in the anterior hemisphere show ciliary re-versal with a gradient of amplitude that halts or greatly slowsspheroid swimming. By contrast, when a spheroid is stimulated bycontinuous light coming from the side, only the flagella on theilluminated side in the anterior hemisphere should undergo ciliaryreversal, causing the spheroid to turn to the light source (3) (Fig.S2). Thus, the gradient along the A–P axis is very important: Ifflagella on the illuminated side in the posterior hemisphere alsoreverse their beating direction, then the entire spheroid wouldrotate rather than gradually turn to the light source.The A–P gradient in V. rousseletii may have other advantages

in photobehavior. For example, the gradient may enable aspheroid to continue rotation around the A–P axis even duringphotoshock response; such continuous bodily rotation should beimportant for a spheroid to promptly resume phototactic steer-ing. In addition, the gradient may contribute to fine-tuning of theamplitude of photoresponse. Because of the A–P gradient, thenumber of cells (or the surface area of a spheroid) responding tophotostimulation changes depending on the light intensity. Thismight help the spheroid to respond to a wide range of light in-tensities with varying amplitude of photoshock response orphototaxis.In conclusion, our DEV experiments showed that, in addition

to eyespot size, the properties of flagella in V. rousseletii, includingtheir Ca2+ sensitivity, have a gradient along the A–P axis. Duringevolution from Chlamydomonas-like unicellular organisms toVolvox-like multicellular organisms, Volvocales have acquired aspheroidal shape with a functional gradient along the A–P axis. Asthe spheroid increased in size, it may have differentiated into ananterior hemisphere for steering or braking after photoreceptionand a posterior hemisphere for constant propulsion of thespheroid. This division of roles in the two hemispheres may beimportant for the effective photobehavior of spheroids.

Materials and MethodsStrain and Culture Conditions. The wild-type V. rousseletii strain MI01 (NIES-4029) was grown in Volvox thiamin acetate (VTAC) medium (27, 28) at 28 °Con a 16-h/8-h light/dark cycle under white fluorescent light at 120 μmol ofphotons m−2·s−1.

Reactivation of Detergent-Extracted Volvox Spheroids. DEVs were preparedsimilar to the method used to produce C. reinhardtii cell models (6). Spheroidswere obtained 12–24 h after transfer to a new medium. For experiments inCa2+-free conditions, spheroids were washed in HMDEK solution [30 mMHepes, 5 mM MgSO4, 1 mM DTT, 1 mM EGTA, and 50 mM potassium acetateat pH 7.4], demembranated in HMDEK solution containing an appropriateconcentration (as described in main text) of Igepal (no. 18896; Sigma–Aldrich),and reactivated in HMDEK with ATP. For experiments in the presence of Ca2+,spheroids were washed in HMDEK or HMDKP solution [HMDEK without EGTAcontaining 1% PEG (average Mw = 20,000) at pH 7.4], demembranated inHMDEK or HMDKP solution containing Igepal, and reactivated in one of theCa2+-buffered reactivation solutions (Table 1) containing ATP. The free Ca2+

concentration of each sample was calculated using CALCON software,written by Shinji Kamimura of Chuo University (www.bio.chuo-u.ac.jp/nano/calcon.html).

Twomethods were used for buffer exchange. First, in the trappingmethodfor monitoring of flagellar motion, washed spheroids were placed in aperfusion chamber made of a glass slide, spacers (0.2-mm-thick vinyl tape),and a coverslip. The thickness of the spacer was adjusted to match the size(developmental stage) of the spheroid (single layer for stage I, double layerfor stage II, and triple layer for stage III). The sample was perfused first withthe demembranation solution and then with the reactivation solution.Second, in the strainer-scooping method for monitoring of swimmingspheroids, spheroids swimming in medium in a Petri dish (35 mm in diameter,10 mm thick) were scooped with a 40-μm nylon cell strainer (BD Falcon352340; BD Biosciences), sequentially dipped in washing solution anddemembranation solution, and soaked in reactivation solution.

Beat Frequency Analysis and Waveform Monitoring of Flagella/Axonemes.Flagellar and axonemal beating were video recorded at 500 frames persecond with a phase-contrast microscope (BX-53; Olympus) equipped with ahigh-speed CCD camera (HAS-L2; DITECT Corporation). Beat frequency wascalculated from the average time interval required for 20 strokes. Flagellar/axonemal waveforms were traced frame by frame on transparency films onthe monitor screen and digitized using Adobe Illustrator software.

Tracking of Reactivated DEVs. Reactivated DEVs were monitored under astereomicroscope (SMZ1000; Nikon) and recorded using a Macromax ScopeMVC-DU CCD camera (GOKO Camera). To analyze swimming velocity,swimming paths of live spheroids and DEVs were tracked for 3 s and 5 s,respectively, using Image Hyper software (Science Eye). The swimming ve-locities were measured from the trajectories.

ACKNOWLEDGMENTS. We thank Dr. Toru Hisabori (Tokyo Institute ofTechnology) for fruitful discussions, Akinori Koitabashi and Asuka Tanno(Tokyo Institute of Technology) for maintenance of Volvox cultures, andDr. Ritsu Kamiya (Gakushuin University) for critical reading of this manu-script. This work was supported by Japan Society for the Promotion of Sci-ence KAKENHI Grants 15H01206, 15H01314, and 16K14752 (to K.W.).

1. Höhn S, Hallmann A (2011) There is more than one way to turn a spherical cellular

monolayer inside out: Type B embryo inversion in Volvox globator. BMC Biol 9:89.2. Solari CA, Michod RE, Goldstein RE (2008) Volvox barberi, the fastest swimmer of the

Volvocales (Chlorophyceae). J Phycol 44:1395–1398.3. Ueki N, Matsunaga S, Inouye I, Hallmann A (2010) How 5000 independent rowers

coordinate their strokes in order to row into the sunlight: Phototaxis in the multi-

cellular green alga Volvox. BMC Biol 8:103.4. Foster KW, Smyth RD (1980) Light antennas in phototactic algae. Microbiol Rev 44:

572–630.5. Bessen M, Fay RB, Witman GB (1980) Calcium control of waveform in isolated flagellar

axonemes of Chlamydomonas. J Cell Biol 86:446–455.6. Kamiya R, Witman GB (1984) Submicromolar levels of calcium control the balance of

beating between the two flagella in demembranated models of Chlamydomonas.

J Cell Biol 98:97–107.7. Mast SO (1926) Reactions to light in Volvox, with special reference to the process of

orientation. J Comp Physiol 4:637–658.8. Hoops H (1997) Motility in the colonial and multicellular Volvocales: Structure,

function, and evolution. Protoplasma 199:99–112.9. Sakaguchi H, Tawada K (1977) Temperature effect on the photo‐accumulation and

phobic response of Volvox aureus. J Protozool 24:284–288.

10. Solari CA, Drescher K, Goldstein RE (2011) The flagellar photoresponse in Volvox

species (Volvocaceae, Chlorophyceae). J Phycol 47:580–583.11. Naito Y, Kaneko H (1972) Reactivated triton-extracted models o paramecium: Mod-

ification of ciliary movement by calcium ions. Science 176:523–524.12. Witman GB, Plummer J, Sander G (1978) Chlamydomonas flagellar mutants lacking

radial spokes and central tubules. Structure, composition, and function of specific

axonemal components. J Cell Biol 76:729–747.13. Gibbons BH, Gibbons IR (1972) Flagellar movement and adenosine triphosphatase

activity in sea urchin sperm extracted with triton X-100. J Cell Biol 54:75–97.14. Hand WG, Haupt W (1971) Flagellar activity of the colony members of Volvox aureus

Ehrbg. during light stimulation. J Eukaryot Microbiol 18:361–364.15. Sakaguchi H, Iwasa K (1979) Two photophobic responses in Volvox carteri. Plant Cell

Physiol 20:909–916.16. Schletz K (1976) [Phototaxis in Volvox-pigments involved in the perception of light

direction]. Z Pflanzenphysiol 77:189–211. German.17. Fujiu K, Nakayama Y, Yanagisawa A, Sokabe M, Yoshimura K (2009) Chlamydomonas

CAV2 encodes a voltage- dependent calcium channel required for the flagellar

waveform conversion. Curr Biol 19:133–139.18. Inaba K (2015) Calcium sensors of ciliary outer arm dynein: Functions and phyloge-

netic considerations for eukaryotic evolution. Cilia 4:6.

Ueki and Wakabayashi PNAS | Published online January 8, 2018 | E1067

PLANTBIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Mar

ch 2

5, 2

020

19. Shiba K, Baba SA, Inoue T, Yoshida M (2008) Ca2+ bursts occur around a local minimal

concentration of attractant and trigger sperm chemotactic response. Proc Natl Acad

Sci USA 105:19312–19317.20. Girard PR, Kennedy JR (1986) Calcium regulation of ciliary activity in rabbit tracheal

epithelial explants and outgrowth. Eur J Cell Biol 40:203–209.21. Wada Y, Mogami Y, Baba S (1997) Modification of ciliary beating in sea urchin larvae

induced by neurotransmitters: Beat-plane rotation and control of frequency fluctu-

ation. J Exp Biol 200:9–18.22. Nakamura S, Tamm SL (1985) Calcium control of ciliary reversal in ionophore-treated

and ATP-reactivated comb plates of ctenophores. J Cell Biol 100:1447–1454.23. Fritsch FE (1935) Introduction. The Structure and Reproduction of the Algae (Cam-

bridge Univ Press, Cambridge, UK), Vol 1, pp 1–59.

24. King SM, Patel-King RS (1995) Identification of a Ca(2+)-binding light chain withinChlamydomonas outer arm dynein. J Cell Sci 108:3757–3764.

25. Omoto CK, Brokaw CJ (1985) Bending patterns of Chlamydomonas flagella: II. Cal-cium effects on reactivated Chlamydomonas flagella. Cell Motil 5:53–60.

26. Wakabayashi K, Ide T, Kamiya R (2009) Calcium-dependent flagellar motility activa-tion in Chlamydomonas reinhardtii in response to mechanical agitation. Cell MotilCytoskeleton 66:736–742.

27. Nozaki H, Kuroiwa H, Mita T, Kuroiwa T (1989) Pleodorina japonica sp. nov.(Volvocales, Chlorophyta) with bacteria-like endosymbionts. Phycologia 28:252–267.

28. Kasai F, Kawachi MME, Watanabe MM (2004) NIES-Collection List of Strains: Micro-algae and Protozoa (National Institute for Environmental Studies, Tsukuba, Japan),7th Ed.

E1068 | www.pnas.org/cgi/doi/10.1073/pnas.1715489115 Ueki and Wakabayashi

Dow

nloa

ded

by g

uest

on

Mar

ch 2

5, 2

020