J. Cell Set. 20, 639-654 (1976) 639

Printed in Great Britain

FLAGELLAR REGENERATION IN

CHLAMYDOMONAS REINHARDTII: EVIDENCE

THAT CYCLOHEXIMIDE PULSES INDUCE

A DELAY IN MORPHOGENESIS

K. W. FARRELL*

Department of Zoology, University of Edinburgh, West Mains Road,Edinburgh, EHg 3JT, U.K.

SUMMARY

The behaviour of a pool of flagellar precursors, assayed by the ability of cells to regenerateflagella in the absence of de novo protein synthesis, has been examined during organelle morpho-genesis in the biflagellate alga Chlamydomonas. The results demonstrate that flagellar elongationcan continue even when this pool is apparently empty and suggest that 2 sources of precursorsare available to the regenerating flagella: those pre-existing in the cellular pool and thosesynthesized de novo. Further evidence for this was obtained by subjecting regenerating cells topulses of cycloheximide. Cells exposed to this drug during the first 60 min post deflagellationformed only half-length (5-/im) flagella, whereas a pulse administered after this point allowedthe formation of longer flagella and suggested that some de novo protein synthesis was requiredfor the formation of full-length flagella, although it was not a prerequisite for the initiation ofregeneration. In addition, it was found that, subsequent to the removal of the cycloheximide,flagellar regeneration did not recommence immediately, but was delayed for a period ofapproximately 45 min, irrespective of length of flagella formed prior to drug inhibition.

The nature of this cycloheximide-induced delay is unclear and certain alternatives, based onthe exhaustion of structural/regulatory components are considered. Although it is not possibleto distinguish between these alternatives, tubulin is not the limiting component, since a poolof this protein is present when flagellar elongation is prevented by cycloheximide.

INTRODUCTION

The biflagellate alga Chlamydomonas reinhardtii has been widely used for flagellarmorphogenetic studies and the kinetics of regeneration of this organelle are now wellestablished. Following amputation, the flagella regenerate to their original length withthe deceleratory kinetics characteristic of many protozoans, even after repeateddeflagellations (Randall et al. 1967; Rosenbaum & Child, 1967; Rosenbaum, Moulder& Ringo, 1969). The initial rate of regeneration is dependent upon the length of theflagellar stump remaining after amputation: the greater the stump length the slowerthe initial rate of flagellar elongation (Tamm, 1967). Other studies have also shownthat if only one of the two flagella is removed only the amputated flagellum initiallyregenerates, suggesting that the intact flagellum is unable to utilize the precursor

• Present address: Imperial Cancer Research Fund Laboratories, P.O. Box 123, 44 Lincoln'sInn Fields, London, WC2A 3PX, U.K.

640 K. W. Farrell

proteins available to the regenerating flagellum (Rosenbaum et al. 1969; Coyne &Rosenbaum, 1970).

These observations suggest that the control of flagellar regeneration resides withinthe organelle itself. It has been proposed that the rate of elongation of Peranemaflagella (Tamm, 1967), Chlamydomonas and Astasia flagella (Child, 1963), and sea-urchin cilia (Burns, 1973) might be controlled by a proximo-distal concentrationgradient, the concentration of precursor at the distal (tip) end determining the rateof regeneration. The existence of a pool of flagellar precursors in Chlamydomonashas been inferred from the ability of this alga partly to regenerate its flagella in theabsence of de novo protein synthesis. However, the role of these precursors in thecontrol of flagellar regeneration is not known, although initial studies suggest thatthe size of this precursor pool does vary during elongation (Rosenbaum et al. 1969).

Since the subunit protein of microtubules, tubulin, forms approximately 30% ofthe total flagellar protein (Gibbons, 1965) and is well characterized, both chemicallyand physically (Olmsted & Borisy, 1973; Lee, Frigon & Timasheff, 1973; Shelanski,1973) many workers have monitored the intracellular levels of this protein duringorganelle morphogenesis. Oral development in the ciliate Tetrahymena (Williams &Nelsen, 1973) occurs without the de novo synthesis of large amounts of tubulin, anda cytoplasmic pool of this protein was found to remain constant throughout earlydevelopment in Arbacia embryos (Raff & Kaumeyer, 1973). Similarly, most ciliaryproteins of Strongylocentrotus, including tubulin, are made prior to ciliogenesis(Stephens, 1972), which suggests that the de novo synthesis of tubulin is unlikely to bethe limiting step in the formation of these organelles. In contrast, the differentiationof Naegleria amoebae into amoebo-flagellates is accompanied by a 35—55-fold increasein the cellular levels of tubulin, as detected by a radio-immune assay (Kowit & Fulton,1974 a), and further evidence indicates that this increase is due to the de novosynthesis of tubulin, rather than the unmasking of existing subunits (Kowit & Fulton,19746).

The availability of flagellar precursors during organelle regeneration has beenexamined in Chlamydomonas in an attempt to elucidate their role in the control ofmorphogenesis. The results suggest that a pool of flagellar precursors, assayed by theability of the cells to form flagella in the absence of de novo protein synthesis, does notinfluence the kinetics of flagellar regeneration, under normal conditions, and that thecytoplasmic pool of tubulin present in this alga greatly exceeds that estimated fromthe regeneration kinetics.

MATERIALS AND METHODS

StrainThe wild-type strain of Chlamydomonas reinhardtii, 32D-, was obtained from the Cambridge

Culture Collection of Algae.

Media and culture conditions

Liquid cultures of C. reinhardtii were cultured in a modified Normal Growth Medium(NGM) of Sager & Granick (1953). The composition of this medium was as follows: KtHPO4,

Flagellar regeneration in Chlamydomonas 641

01 g; KH,PO4)oig;MgSO4.7HjO,o-3 g; NH4NO3, 0-3 g; CaCl,.6HjO, oo4g; ferric citrate,001 g; citric acid, o-oi g; sodium citrate dihydrate, 005 g; distilled water, 1 litre. One millilitreof the following trace element solution was added to 1 litre of the above medium: H,BO4)i-og; ZnSCv7HaO, i-o g; MnSO4.4H.aO, 04 g; CoCl2.6H2O, 04 g; NaMoO4.2Hj,O, 02 g;distilled water 1 litre. Sulphate-free NGM was prepared as for NGM except for the omissionof theMgSO4.7Hi!O.

Asynchronous cultures were grown at 25 °C and at a light intensity of 5000 lux in 500-ml or2-1. Ehrlenmeyer flasks. All cultures were bubbled with air and continually stirred.

Stock cultures were maintained on 2 % agar (Oxoid No. 3) slopes supplemented with theabove growth medium and at a light intensity of 2500 lux. These stocks were regularly sub-cultured on to fresh slopes.

Flagellar length determination

Drops of cell suspensions were placed on glass slides and routinely fixed by exposure toOsO4 vapour for 90 s. Coverslips were then placed on the slides and the samples allowed to drydown until the cells were firmly sandwiched between slide and coverslip in a thin film of liquid.These cells were protographed using a Zeiss photomicroscope and phase-contrast optics andthe negatives projected on to a screen. The mean flagellum length was determined by measuringthat part of the flagellum external to the cell wall using a 'Curvimeter' map measuringinstrument.

In some experiments the cells were fixed by exposure to glutaraldehyde vapour, in order tocheck that the osmium treatment did not selectively shorten some (e.g. cycloheximide-treated)flagella. In these experiments quantitatively identical results were obtained, suggesting that thechanges in flagellar length are not artifacts of procedure.

Flagellar precursor pool assay

The ability of Chlamydomonas to regenerate flagella in the presence of cycloheximide hasbeen taken to indicate the existence of a pool of flagellar precursors (Rosenbaum et al. 1969;Coyne & Rosenbaum, 1970). The length of the flagellum regenerated is assumed to be pro-portional to the size of this pool and was assayed as follows.

Experimental cultures were deflagellated by the acid-pH method of Witman, Carlson,Berliner & Rosenbaum (1972) and immediately washed with fresh NGM containing 10 fig ml"1

of cycloheximide. This concentration of cycloheximide has been shown to inhibit proteinsynthesis completely and rapidly in Chlamydomonas (Rosenbaum et al. 1969). Flagellar re-generation was allowed to proceed in this medium for 75 min at 25 °C and under constant lightconditions, before fixing with OsO4 and measuring the mean flagellum length.

In order to examine the variation in the size of this flagellar pool during flagellar regeneration,experimental cultures were deflagellated and allowed to regenerate flagella in NGM. Atpredetermined intervals, cell samples were removed and further deflagellated in the presenceof cycloheximide (10 fig ml"1). Flagellar regeneration in the presence of cycloheximide wasallowed to proceed for an additional 75 min at 25 °C, before fixing the cells for flagellar lengthmeasurements.

Cycloheximide pulses

Samples to be pulsed with cycloheximide were made 10 fig ml"1 in this drug at the appropriatestages of regeneration (see Fig. 6, p. 646) and left in continuous light and at 25 °C for 45 or90 min. To terminate the pulse the samples were washed 3 times with 10 times their volume offresh NGM and finally resuspended in a volume of NGM equal to the initial volume of thesample. The entire washing procedure took no longer than 10 min.

Incorporation of [3H]lysine into TCA-insoluble protein

In vivo amino acid incorporation was carried out in 10-ml Ehrlenmeyer flasks. Cultures at acell density of 1-4 x io" cells ml"1 were incubated in continuous light with stirring and [3H]-lysine (Radiochemical Centre, Amersham; sp. act. 8-15 Ci/mmol) added to the cultures to a

642 K. W. Farreli

final concentration of 10/tCi ml"1. After predetermined intervals, 05-ml samples were removedinto 10 ml of ice-cold 5 % trichloroacetic acid (TCA) supplemented with 1 min lysine and thesuspension added to a glass-fibre filter (Whatman, GF/A) on a vacuum sampling manifold.The excess fluid was drawn through under a vacuum and the cells washed with 25 ml of theTCA solution. Finally the filters were dried before placing in vials for liquid scintillationcounting.

In the case of cultures pre-treated with cycloheximide (Fig. 1, p. 643), the [3H]lysine wasadded immediately following the termination of the pulse and then samples removed and treatedas described above.

Preparation of Chlamydomonas post-ribosomal supernatants

Two-litre cultures of Chlamydomonas were harvested using a continuous flow rotor operatingat 6000 rev/min (MSE High Speed 18) and washed twice with ice-cold 10 mM Tris-HClpH 6 8 . The cells were deflagellated in a buffer of 5% sucrose/10 mM Tris-HCl pH 6-8 usingthe acid-pH method and pelleted through 25 % sucrose/10 mM Tris-HCl pH 6 8 at 5000 rev/min (MSE ' Mistral' 2L, 4 x 50 head) for 10 min to remove the contaminating flagella (Witmanetal. 1972). The cells were then washed with ice-cold 20 mM NaH2P04/ ioo mM sodium glutamatepH 6-8 (PG) buffer, resuspended in an equal volume (v/v) of this buffer and broken either on aprecooled French press or by sonication on ice. The homogenates were then centrifuged at150000 g for 1 h at 4 °C (Beckman Spinco model L; SW 50. 1 head) to prepare the highspeed supernatants.

Labelling of the Chlamydomonas cellular tubulin

Chlamydomonas tubulin was labelled by incubating deflagellated cells for 2 h in "SOj(Radiochemical Centre, Amersham; sp. act. in culture medium 100 /iCi ml"1, or approximately80 /*Ci mraol"x sulphate). This procedure resulted in the preferential labelling of the cyto-plasmic tubulin pool, since some of the pre-existing unlabelled pool was chased into theflagella which regenerated during the incubation period.

Gel electrophoresis

Sample preparation and sodium dodecylsulphate (SDS)-gel electrophoresis were carried outaccording to the method of Laemlli (1970) using 7 5 % gels, 7 0 x 0 8 cm. The gels were stainedwith Coomassie blue R250 (Fairbanks, Steck & Wallace, 1961) and scanned using a Joyce-Loebl UV scanner. Radioactive gels were frozen on dry ice and cut into 05-mm slices usinga Mickel slicer, which were then dried down on to filter paper (Whatman No. 50, speciallyhardened) by incubation at 60 °C for 1 h, and counted using a toluene/05 % dibutyl PBD[(2-4'('e'''--butylphenyl)-5-(4"-biphenyl)-i,3,4-oxadiazole] scintillant.

The tubulin content of sample preparations was obtained by scanning the gels of the relevantpreparations and determining the weight of the tubulin peak under the densitometer tracingsrelative to the total weight of all the protein peaks. The concentration of tubulin in the samplewas calculated from this value and the total protein concentration of the sample.

This method of estimation of protein bands assumes that each band is quantitatively stainedby the Coomassie blue. However, it is known that at high protein concentrations Coomassieblue staining deviates markedly from Beer's Law (Fazekas de St Groth, Webster & Datyner,1963; Chrambach, Reisfeld, Wyckoff & Zaccari, 1967), the minor protein bands staining toa greater degree than the major bands. In order to exclude errors arising from this differentialstaining response, low amounts of protein (less than 100 fig) were loaded on to gels from whicha quantitative determination of the amount of tubulin was required.

Protein determination

Protein assays were performed by the method of Lowry, Rosebrough, Fan- & Randall (1951)using bovine serum albumin as a standard.

Flagellar regeneration in Chlamydomonas 643

RESULTS

The effects of cycloheximide on the incorporation of labelled amino acid into TCA-insoluble proteins

It has been previously shown that cycloheximide concentrations greater thani /tg ml"1 rapidly and completely inhibit the uptake of amino acids into the TCA-insoluble proteins of Chlamydomonas (Rosenbaum et al. 1969). However, the abilityof this alga to recover from cycloheximide treatment is not known. To examine thisaspect, a culture of Chlamydomonas was divided into 2 samples. The first was made

30

XEV

.: 10

C_

10 20 30 40 50 60Time, min

Fig. 1. Incorporation of pHJlysine into TCA-insoluble proteins of Chlamydovionas.05-ml samples were removed (Methods) from cultures which had previously beenexposed to a 45-min cycloheximide (10 fig ml"1) pulse (A) or left untreated ( • ) .Total washing time required to terminate the cycloheximide treatment was 10 min.

10 fig ml"1 with cycloheximide, while the control had an equivalent volume of NGMadded. Both cultures were maintained at 25 °C for either 45 min or 90 min afterwhich time the cycloheximide treatment was terminated (Methods). The controlculture was processed in parallel to the treated culture. The ability of both thesecultures to incorporate pH]lysine into the TCA-insoluble protein fraction was thenfollowed for the next 60 min (Methods). The results show quite clearly thatChlamydomonas recovers from the cycloheximide treatment immediately followingdrug removal (Fig. 1). In fact, prior exposure to the drug actually appears to stimulateamino acid incorporation. The reason for this stimulation is unclear, although possiblyit is owing to a build up of messenger RNA during the period of exposure to cyclo-heximide. However, Chlamydomonas recovers completely from exposure to cyclo-heximide and damage to the protein-synthesizing machinery of this alga does notappear to result from this treatment.

The kinetics of flagellar regeneration

Chlamydomonas cells regenerate their flagella with characteristic deceleratorykinetics and attain the original length within 150 min (Fig. 2). The apparent lagbefore elongation commences has been shown to be an artifact and in fact, flagellar

644 K. W. Farrell

regeneration starts immediately after amputation (Randall, 1969). If the flagella areagain removed after they have attained full length the kinetics of regeneration areidentical to those observed following the first deflagellation (Fig. 3). Similar observa-tions have previously been reported for Chlamydomonas (Randall, 1969), Ochromonas(Rosenbaum & Child, 1967) and Tripneutes (Burns, 1973).

1 2 r

10 -

8 -

6 -

4 -

2 -

7,

/

i i 1

——""X

A A

12 r

E 10

DO

=S 4

40 80 120 160Time, mm

Fig. 2

200 240 80 160 240 320

Time, min

Fig. 3

400 480

Fig. 2. The kinetics of flagellar regeneration in a population of Chlamydomonasdeflagellated at time zero in the presence (A) and absence (A) of cycloheximide(10 fig ml-1). Arrow indicates mean flagellum length prior to deflagellation.

Fig. 3. The kinetics of flagellar regeneration in a population of Chlamydomonassuccessively deflagellated after o and 240 min. Arrow indicates mean flagellum lengthprior to deflagellation.

In the presence of cycloheximide flagellar regeneration does not go to completionand only half-length flagella are formed. In addition, the rate of flagellar regenerationis noticeably slower than in the absence of cycloheximide (Fig. 2).

Variation of the flagellar precursor pool during regeneration

The variation in the size of the flagellar precursor pool during elongation is shownin Fig. 4. The pool quickly becomes depleted during the initial, rapid phase of regen-eration and eventually falls to a basal level 45 min post-deflagellation. This basal levelis given as zero since no flagellar stumps were observed protruding external to the cellsurface. However, this does not necessarily mean that the flagellar precursors havebecome completely exhausted, as it is known that proximal portions of the flagella arehidden within the flagellar tunnel of the cell wall (Johnson 8c Porter, 1968; Randall,1969). Very occasionally, the basal level of the pool is represented by short (ca. i-/tm)flagellar stumps. Regardless of the basal level reached, the pool begins to refill after60 min post-deflagellation and attains its initial pre-deflagellation value by 120 min,after which time the pool level remains constant at this pre-deflagellation value. The

Flagellar regeneration in Chlamydotnonas 645

size of this precursor pool, therefore, does vary during regeneration, and in additionflagellar elongation continues even when the pool is apparently empty.

By contrast, in the presence of cycloheximide the extent of flagellar regeneration isequivalent to the size of the flagellar pool: no further regeneration occurs after thepool has been exhausted (Fig. 5). In addition, the extent of flagellar regeneration, upto the time of pool exhaustion, is less in the presence of cycloheximide than in itsabsence (Fig. 2), suggesting that the amount of flagellar regeneration occurring in theabsence of cycloheximide is greater than can be accounted for on the basis of poolprecursor utilization alone.

12

1 0

M 8

12 _

E 10a.

3

40 80 120 160 200 240 280 320Time, mm

F'g- 4

40 80 120 160

Time, mm

Fig. 5

200 240

Fig. 4. Variation in the size of the flagellar precursor pool during flagellar regenerationin a population of Chlamydomonas deflagellated at time zero. The flagellar precursorpool was assayed by removing samples of regenerating cells at intervals, furtherdeflagellating in the presence of cycloheximide, and allowing the cells to regenerate inthe presence of this drug for a further 75 min. The length of the flagella regenerated isthen taken to indicate the size of the precursor pool. Pool size (A), mean flagellumlength (A). Arrows indicate size of pool (->) and mean flagellum length (-•) prior todeflagellation.

Fig. 5. Variation in the size of the flagellar precursor pool during flagellar regenerationin a population of Chlamydomonas in the presence of cycloheximide (io/igml"1).Pool size (A), mean flagellum length (A). Arrows indicate size of pool (->) and meanflagellum length (->) prior to deflagellation.

The effect of cycloheximide pulses on flagellar regeneration and pool variation

The above data suggested that flagellar regeneration is characterized by the utili-zation of both precursor proteins present in a pool and proteins synthesized de novo.Although this de novo protein synthesis was not a prerequisite for elongation to occur,it seemed likely that some protein synthesis was essential for flagellar regeneration togo to completion, since the flagella never reached full length in the presence ofcycloheximide. In order to test this, regenerating cells were pulsed with cycloheximideat intervals and the behaviour of both the elongating flagella and the precursor pool

646 K. W. Farrell

was examined (Fig. 6). The initial depletion of the fiagellar pool was unaffected bythe timing of the cycloheximide pulse and in every case fell to its basal level within45 min. In contrast, the subsequent recovery of the pool to its initial value wasnoticeably modified by the timing of this pulse: inhibition of protein synthesis for the

8-

-o4

/

40 80 120 160 200 240 280 320

E12 1

8-

-o4 - V•\

40 80 120 160 200 240 280 320

12

8 •

/ \ A A A A A

40 80 120 160 200 240 280 320

Time, mm

Fig. 6. Variation in the size of the fiagellar precursor pool during fiagellar regenerationin a population of Chlamydomonas pulsed with cycloheximide (io /tg ml"1) from0-45 min (A), 15-60 min (B), and 60-150 min (c) subsequent to deflagellation. Poolsize (A), mean flagellum length (A). Arrows indicate pool size (-(>) and meanflagellum length (-•) prior to deflagellation.

0-45 min, 15-60 min or 60-150 min post-deflagellation period prevented poolrecovery for approximately 45 min subsequent to the removal of the cycloheximide.The length of this delay period appeared to be independent of the duration of thecycloheximide pulse, since pulses of 45 or 90 min both prevented pool recovery for45 min subsequent to the removal of the drug.

Flagellar regeneration in Chlamydomonas 647

The extent of flagellar regeneration was also related to the timing of the cyclo-heximide pulse. Pulses administered over the 0-45 min and 15-60 min post-deflagellation periods permitted the formation of only half-length (5-/tm) flagellawhereas a pulse over the 60-150 min period permitted the formation of ((-/on-longflagella. Furthermore, once the precursor pool had become depleted no furtherflagellar elongation occurred until after pool recovery was apparent.

It would appear, therefore, that de novo protein synthesis is required during atleast part of the first 60 min of flagellar regeneration if the flagella are to attain fulllength. The cycloheximide-induced delay in flagellar regeneration suggests that45 minof de novo protein synthesis are required for the reinitiation of flagellar elongation.Furthermore, although under control conditions (no cycloheximide) pool recoveryduring flagellar regeneration is not apparent until after 15-30 min of reaching thebasal level, no 'break' in the flagellar regeneration curve is apparent (Fig. 4). There-fore, the 45-min delay would appear to be contingent upon pool exhaustion in thepresence of cycloheximide.

Studies on the cellular tubulin pool during flagellar regeneration

Although the above kinetic studies are useful for examining the overall availabilityof flagellar precursors, no information concerning the possible roles of individualprecursors in the morphogenetic event can be deduced from these data. Con-sequently, the behaviour of the protein tentatively identified as tubulin (K. W. Farrell,in preparation) was examined during flagellar regeneration, in an attempt to gainsome insight into the control mechanisms underlying the regeneration process.

Since the flagellar precursor pool apparently becomes depleted 45 min subsequentto deflagellation and is refilled by about 120 min (Fig. 4), high-speed supernatantswere prepared from fixed volumes of cell suspensions removed from regeneratingChlamydomonas populations o, 45 and 120 min after deflagellation. The 45- and 120-min samples were further deflagellated in order to prevent possible contamination ofsamples with flagellar matrix protein. A prominent band, corresponding to theputative tubulin, was detectable in all 3 samples by SDS-gel electrophoresis (Fig. 7),suggesting that there was little variation in the size of the tubulin pool duringflagellar regeneration and that the amount of tubulin in this pool is in excess of thatrequired to form 2 full-length flagella. In addition, the pool (measured by the abilityof flagella to regenerate in the presence of cycloheximide) does not correspond to thetotal cytoplasmic tubulin pool.

To examine this possibility further a culture of Chlamydomonas was allowed toregenerate flagella to full-length twice in sulphate-free NGM supplemented with36SO4. Following this incubation period the cells were washed thoroughly with NGMto remove the labelled sulphate from the medium, deflagellated, and transferred toNGM, the large excess of unlabelled sulphate in this medium effectively providinga 'cold chase'. The de novo synthesis of tubulin during flagellar regeneration in thischase period should result in a fall in the specific activity of the tubulin in the flagella.If the flagella regeneration studies with cycloheximide (Fig. 2) are indicative of acytoplasmic tubulin pool sufficient only to form 2 half-length flagella, then

648 K. W. Farrell

Mobility

Fig. 7. Scans of 75 % SDS gels of Chlamydomonas high-speed supernatants preparedfrom regenerating populations after omin (A), 45 min (B), and 120 min (c) post-deflagellation. Arrow indicates position of putative tubulin band. Origin to the left.

Table 2. The change in the specific activity of flagellar tubulin during flagellarregeneration

Time allowedfor regenerationin'cold' NGM,

min

0

451 2 0

Control

Specific activityof flagellar

tubulin,cpm/o.D. unit*

3 5 °2 7 0

2 7 0

o-33

The cytoplasmic and flagellar tubulin was labelled by 2 successive deflagellations in thepresence of "SO^ (100 /iCi ml"1) and a 'cold chase' was applied during a third flagellarregeneration. The control sample consisted of an untreated population of cells left in thepresence of 35SO4 for a period equivalent to the time taken for the third flagellar regeneration.

• The specific activity of the flagellar samples was determined by calculating the weightratios of the areas under the radioactivity and optical density profiles.

Flagellar regeneration in Chlamydomonas 649

a decrease of at least 50% in the specific activity of the flagellar tubulin would beanticipated. Flagellar samples were taken immediately prior to deflagellation andafter 45 min and 120 min of regeneration in the unlabelled medium and prepared forSDS-gel electrophoresis.

The results further show that the method of determination of the precursor poolby means of flagellar growth is not indicative of the size of the cytoplasmic tubulinpool: the specific activity of the flagellar tubulin decreased by only 20-25 % duringflagellar regeneration (Table 1) instead of the anticipated 50%, which suggests thatthe elongating flagellar microtubules are formed mainly from pre-existing tubulinmolecules. In addition, there is some exchange of flagellar tubulin with newlysynthesized subunits in non-regenerating flagella (Table 1), although it was notdetermined whether this turnover was with assembled tubulin or non-assembledmatrix protein, or both. Although this turnover has been noted previously (Gorovsky,Carlson & Rosenbaum, 1970), its significance remains obscure.

Calculations based on a knowledge of the total number of cells in a preparation andthe amount of protein derived from them, and assuming that each cell had two,io-,wm-long flagella, suggest that the amount of assembled flagellar tubulin is equiva-lent to only 16-17 % of the total of tubulin in the cytoplasmic pool. However, this isa minimum estimate since no allowance was made for the non-assembled tubulin inthe flagellar lumen. Also it was assumed that the Chlamydomonas protein band whichco-electrophoresed with purified gerbil brain tubulin was composed entirely of theputative algal tubulin. However, the similarity of this value with the decrease inspecific activity of flagellar tubulin during regeneration, suggests that the amount oftubulin synthesized de novo during flagellar regeneration is equivalent to the amountof tubulin utilized in forming the flagellar microtubules.

DISCUSSION

For the purpose of the following discussion the flagellar precursor pool, assayed bythe ability of the cells to regenerate flagella in the presence of cycloheximide, will betermed pool A.

During the first 45 min post-amputation (the time required to exhaust pool A,Fig. 4), Chlamydomonas cells regenerate flagella to a greater length in the absence ofcycloheximide than in the presence of this drug, even though the amount of flagellarprecursors available from pool A is the same in both cases (Figs. 4, 5). It would appear,therefore, that there are 2 sources of flagellar precursors available to regeneratingcells of Chlamydomonas: those pre-existing in the cytoplasmic pool (pool A) and thosedependent upon protein synthesis de novo, and that the total amount of precursoraccessible to the cells can influence the extent of regeneration up to final full length.In contrast, the kinetics of regeneration are normally independent of precursoravailability: the regeneration kinetics of cells recovering from cycloheximide pulsesare a function of flagellar length and are unaffected by the large amounts of flagellarprecursors made available as a result of pool A refilling (Figs. 6, 8). Similarly, undercontrol conditions (no cycloheximide), no increase in the rate of flagellar regeneration

650 K. W. Farrell

is observed as pool A refills (Fig. 4). These data are consistent with the suggestionthat flagellar regeneration may be controlled by a proximo-distal concentrationgradient (e.g. see Tamm, 1967). The only conditions under which the availability ofprecursors affect the kinetics of flagellar regeneration are when denovo protein synthesisis prevented: in the presence of cycloheximide the rate of flagellar growth is slowerthan in the absence of this drug over the first 45 min post-amputation (Fig. 2). Thiseffect might be a result of the cell's inability to replace utilized precursor, producingsuboptimal conditions for morphogenesis. Alternatively, a protein synthesized de novois concerned with the control of flagellar regeneration.

12

10

3 6

sz

40 80 120 160

Time, min

200 240

Fig. 8. Comparison of the kinetics of flagellar regeneration in populations ofChlamydomonas recovering from 0-45 min ( • ) , 15-60 min (A), and 60-150 min ( • )pulses of cycloheximide (10 fig ml"1) with those of an untreated population (A)-Those portions of the regeneration curves of treated populations subsequent to thecycloheximide-induced delay (see Fig. 6) have been shifted horizontally along the timeaxis to allow a direct comparison with the control curve.

The cessation of flagellar growth in the presence of cycloheximide (Fig. 2) pre-sumably results either from the exhaustion of a flagellar structural component (e.g.dynein, nexin, etc.) or from the depletion of a morphogenetic regulator protein, andreflects an absolute requirement for protein synthesis. Electrophoretic data suggestthat a tubulin pool in excess of that required for the formation of 2 full-lengthflagella is present in this alga (Fig. 7). In addition, flagellar regeneration inChlamydomonas can apparently occur without a heavy investment of protein synthesisin the production of tubulin (Table 1). Consequently, the ability to detect a pool oftubulin under conditions where flagellar regeneration is prevented by cycloheximide,indicates that tubulin is not a limiting structural component and possibly factorsrequiring de novo protein synthesis are concerned with the assembly of this cytoplasmictubulin into microtubules.

An analogous situation has been described for oral replacement in Tetrahymena.

Flagellar regeneration in Chlamydomonas 651

Although there is an absolute requirement for de novo protein synthesis during oralmorphogenesis in this organism (Frankel, 1970), as much as 94% of the incorporatedmicrotubule protein is derived from a pre-existing tubulin pool (Williams & Nelsen,1973). These authors therefore speculate that the required synthesis may be for aregulatory protein concerned with the morphogenesis of this organelle. It is interestingto note that, although these studies demonstrate that there is a large pool of micro-tubule protein present in this organism, totally deciliated cells of the same strain areunable to regenerate their cilia in the presence of cycloheximide (Rannestadt, 1974).In addition, in partially deciliated cells the ciliary volume regenerated is greater thanthe ciliary volume resorbed and the author suggests that a factor present inthe resorbing cilia potentiates the assembly of pre-existing ciliary precursors.Although the nature of this potentiating factor is unknown, it seems more probablethat it belongs to the regulator class of ciliary components, rather than the structuralclass, since the precise stoichiometric relationships of the latter components withinthe cilia seems to preclude a potentiating function. Finally, ciliogenesis in the embryoof the sea urchin has been reported to occur in the absence of protein synthesis(Auclair & Siegel, 1966). However, the inhibition of amino-acid incorporation intoTCA-precipitable ciliary proteins was about 93 % and consequently the synthesis ofsmall amounts of essential ciliary (possibly regulatory) components cannot be excluded.Microtubules in these organelles can therefore be formed, at least in part, from pre-existing tubulin subunits, which suggests that microtubule assembly in these casescannot be completely regulated by the de novo synthesis of tubulin. In these situations,the absolute requirement of morphogenesis for protein synthesis may reflect arequirement for a morphogenetic regulator protein, although a limiting structuralcomponent cannot be excluded.

The nature of the cycloheximide-induced delay in both flagellar regeneration andpool A recovery, subsequent to the removal of the inhibitor, is unclear. It is not dueto the cells being in a state of 'physiological shock' nor to traces of cycloheximideremaining from exposure to this drug, since the cells recover from this treatmentalmost immediately (Fig. 1). The delay is contingent upon pool A exhaustion in thepresence of cycloheximide (Fig. 6); pool A exhaustion in the absence of cycloheximidedoes not induce this delay (Fig. 4). This suggests that the delay is induced only whenthe supply of flagellar precursors from both the precursor pool and de novo synthesisis interrupted simultaneously. These data clearly indicate that the exhaustion of anessential component is required for initiation of the delay period. However, it seemsimprobable that the entire 45 min are required for the resynthesis of a limitingstructural/regulatory component, since de novo synthesis alone apparently producessufficient quantities of this essential factor to maintain flagellar regeneration (Fig. 4).Three alternatives therefore seem possible, (i) Forty-five minutes of protein synthesisare required to achieve a critical threshold concentration of a structural/regulatorycomponent for the re-initiation of flagellar regeneration. Below this threshold levelthe re-initiation of regeneration cannot occur, but once achieved de novo proteinsynthesis can maintain the concentration of this factor at or above the critical levelduring the ensuing flagellar elongation, (ii) Several flagellar components are

652 K. W. Farrell

sequentially synthesized during the delay period, the limiting component being syn-thesized last. Here, it is relevant to note that prior to ciliogenesis in the sea urchinStrongylocentrotus, several ciliary components are synthesized uniformly but that theiinitiation of ciliogenesis is characterized by the de novo synthesis of certain minorstructural components (Stephens, 1972). (iii) Flagellar morphogenesis is dependentupon an unstable regulatory factor which rapidly degenerates into a non-functionalform and which must be maintained at a critical level either from a pool (with rapidturnover) or by de novo synthesis. The net gain in the amount of this componentwould then be dependent on its relative rates of synthesis and deactivation. Con-sequently, although sufficient supplies of this regulator could be synthesized in lessthan 45 min, the decay reaction extends the delay period to the observed figure.

At present, it is not possible to distinguish between these alternatives. However, thesituation described here for Chlamydomonas is not unique and has parallels in othersystems. For example, it is apparent from the data of Kerr (1972, fig. 3) that longflagellum regeneration in Didymium is delayed for approximately 45 min followingeven a brief exposure to cycloheximide. Furthermore, this delay period is independentof the duration of exposure to the drug. However, the situation does differ fromChlamydomonas since this delay is not observed if the cells are exposed to pulses of thedrug after morphogenesis has begun. Similarly, organelle morphogenesis in Tetra-hymenacan be synchronized by a series of 'heat shocks' (Frankel, 1970). Following theend of the heat treatment a delay of 40 min is seen prior to the resumption ofmorphogenesis (Zeuthen & Williams, 1969). Since it is possible to correlate thecessation and resumption of morphogenesis with the disappearance and reappearanceof fibrils, some of which have the dimensions of microtubules, within the stomatogenicfield, the authors suggested that these structures might represent the primary site ofaction of the heat treatment. Alternatively, a site of action on a protein(s) concernedwith the polymerization of microtubules could not be excluded. (Also see Williams &Nelsen, 1973.)

Clearly, the primary effect of cycloheximide on flagellar morphogenesis inChlamydomonas cannot be at the microtubule level, since some flagellar regenerationis possible even in the presence of this drug (Fig. 2). However, the cycloheximidecould be affecting a protein or proteins (possibly unstable) concerned with thepromotion of microtubule formation. It has already been proposed that tubulin sub-units in Chlamydomonas must be modified ('activated') prior to assembly in vivo(Farrell & Burns, 1975) and it is tempting to speculate that the target protein(s) ofthe cycloheximide is a morphogenetic regulator protein concerned with the activationof tubulin subunits.

The author would like to thank Dr R. G. Burns for his help in the preparation of thismanuscript and Sir John Randall for his continued interest throughout the course of this work.This work was supported by an M.R.C. grant.

Flagellar regeneration in Chlamydomonas 653

REFERENCES

AUCLAIR, W. & SIEGEL, B. W. (1966). Cilia regeneration in the sea urchin embryo: evidence fora pool of ciliary proteins. Science, N.Y. 154, 913-915.

BURNS, R. G. (1973). Kinetics of the regeneration of sea-urchin cilia. J. Cell Set. 13, 55-67.CHILD, F. M. (1963). Flagellar regeneration: a mechanism for its initiation and regulation.

Biol. Bull. mar. biol. Lab., Woods Hole 125, 361.CHRAMBACH, A., REISFELD, R. A., WYCKOFF, M. & ZACCARI, J. (1967). A procedure for rapid

and sensitive staining of protein fractionated by polyacrylamide gel electrophoresis. Analyt.Biochem. 20, 150-154.

COYNE, B. & ROSENBAUM, J. L. (1970). Flagellar elongation and shortening in Chlamydomonas.II. Re-utilisation of flagellar proteins. J. Cell Biol. 47, 777-781.

FAIRBANKS, G., STECK, T. L. & WALLACE, D. F. H. (1971). Electrophoretic analysis of themajor polypeptides of the human erythrocyte membrane. Biochemistry, N. Y. 10, 2606-2616.

FARRELL, K. W. & BURNS, R. G. (1975). Inability to detect Chlamydomonas microtubuleassembly in vitro: possible implications to the in vivo regulation of microtubule assembly.J. Cell Sci. 17, 669-681.

FAZEKAS DE ST GROTH, S., WEBSTER, R. G. & DATYNER, A. (1963). Two new staining pro-

cedures for quantitative estimation of proteins on electrophoretic strips. Biochim. biophys.Acta 71, 337-391-

FRANKEL, J. (1970). The synchronization of oral development without cell division inTetrahymena pyriformis GL-C. J. exp. Zool. 173, 79-100.

GIBBONS, I. R. (1965). Chemical dissection of cilia. Archs Biol., Liege 76, 317-352.GOROVSKY, M. A., CARLSON, K. & ROSENBAUM, J. L. (1970). Simple method for quantitative

densitometry of polyacrylamide gels using fast green. Analyt. Biochem. 35, 359—370.JOHNSON, U. G. & PORTER, K. (1968). Fine structure of cell division in Chlamydomonas

reinhardtii. J. Cell Biol. 38, 403-425.KERR, S. J. (1972). Flagellum growth and regeneration in the true slime mould Didymium

nigripes. J. gen. Microbiol. 72, 429-437.KOWIT, J. D. & FULTON, C. (1974a). Purification and properties of flagellar outer doublet

tubulin from Naegleria gruberi and a radioimmune assay for tubulin. J. biol. Chem. 249,3638-3646.

KOWIT, J. D. & FULTON, C. (19746). Programmed synthesis of tubulin for the flagella thatdevelop during cell differentiation in Naegleria gruberi. Proc. natn. Acad. Sci. U.S.A. 71,2877-2881.

LAEMLLI, U. K. (1970). Cleavage of structural proteins during the assembly of the head ofbacteriophage T4. Nature, Lond. 277, 680-685.

LEE, J. C , FRIGON, R. P. & TIMASHEFF, S. N. (1973). The chemical characterisation of calfbrain microtubule protein subunits. J. biol. Chem. 248, 7253-7262.

LOWRY, O. H., ROSEBROUGH, N. J., FARR, A. L. & RANDALL, R. J. (1951). Protein measure-ment with the Folin phenol reagent. J. biol. Chem. 193, 265-275.

OLMSTED, J. B. & BORISY, G. G. (1973). Microtubules. A. Rev. Biochem. 42. 507-539.RAFF, R. A. & KAUMEYER, J. F. (1973). Soluble microtubule proteins of the sea urchin embryo:

partial characterisation of the proteins and behaviour of the pool in early development.Devi Biol. 32, 309-320.

RANDALL, Sir J. T. (1969). The flagellar apparatus as a model organelle for the study of growthand morphopoiesis. Proc. R. Soc. B 173, 31-62.

RANDALL, Sir J. T., CAVALIER-SMITH, T., MCVITTIE, A., WARR, J. & HOPKINS, J. M. (1967).Developmental and control processes in basal bodies and flagella of Chlamydomonasreinhardtii. Symp. Soc. dev. Biol., Suppl. 1, 43-83.

RANNESTADT, J. (1974). The regeneration of cilia in partially deciliated Tetrahymena. J. CellBiol. 63, 1009-1017.

ROSENBAUM, J. L. & CHILD, F. M. (1967). Flagellar regeneration in protozoan flagellates.J. Cell Biol. 34, 345-364-

ROSENBAUM, J. L., MOULDER, J. E. & RINGO, D. L. (1969). Flagellar shortening and elongationin Chlamydomonas. J. Cell Biol. 41, 600—619.

41-2

654 K. W. Farrell

SAGER, R. & GRANICK, S. (1953). Nutritional studies with Chlamydomonas. In Biochemistry andPhysiology of Protozoa, vol. 3 (ed. S. H. Hutner), pp. 297-318. New York and London:Academic Press.

SHELANSKI, M. (1973). Chemistry of the filaments and tubules of brain. J. Histochem. Cytocliem.2i, 529-539-

STEPHENS, R. E. (1972). Studies on the development of the sea urchin Strongylocentrotusdrobachiensis. III. Embryonic synthesis of ciliary proteins. Biol. Bull. mar. biol. Lab., WoodsHole 142, 489-504.

TAMM, S. L. (1967). Flagellar development in the protozoan Peranema trichophorum. J. exp.Zool. 164, 163-186.

WILLIAMS, N. E. & NELSEN, E. M. (1973). Regulation of microtubules in Tetrahymena.II. Relation between turnover of microtubule proteins and microtubule dissociation andassembly during oral replacement. J. Cell Biol. 56, 458—465.

WITMAN, G. B., CARLSON, K., BERLINER, J. & ROSENBAUM, J. L. (1972). Chlamydomonasflagella. I. Isolation and electrophoretic analysis of microtubules, membranes andmastigonemes. J. Cell Biol. 54, 507-539.

ZEUTHEN, E. & WILLIAMS, N. E. (1969). Division-limiting morphogenetic processes inTetrahymena. In Nucleic Acid Metabolism, Cell Differentiation and Cancer Growth (ed. E. V.Cowdry & S. Seno), pp. 203-216. Oxford: Pergamon.

(Received 12 May 1975)