Flagellar protein dynamics in Chlamydomonas

Cilia and flagella appear to be stable, terminal, microtubule-containing organelles, but they also elongate and shorten in response to a variety of signals. To understand mechanisms that regulate flagellar dynamics, Chlamydomonas cells with nongrowing flagella were labeled with (35)S, and flagella and basal body components were examined for labeled polypeptides. Maximal incorporation of label into the flagella occurred within 3 h. Twenty percent of the flagellar polypeptides were exchanged. These included tubulins, dyneins, and 80 other axonemal and membrane plus matrix polypeptides. The most stable flagellar structure is the PF-ribbon, which comprises part of the wall of each doublet microtubule and is composed of tubulin and three other polypeptides. Most (35)S was incorporated into the high molecular weight ribbon polypeptide, rib240, and little, if any, (35)S is incorporated into PF-ribbon-associated tubulin. Both wild-type (9 + 2) and 9 + 0 flagella, which lack central microtubules, exhibited nearly identical exchange patterns, so labeling is not due to turnover of relatively labile central microtubules. To determine if flagellar length is balanced by protein exchange, (35)S incorporation into disassembling flagella was examined, as was exchange in flagella in which microtubule assembly was blocked by colchicine. Incorporation of (35)S-labeled polypeptides was found to occur into flagellar axonemes during wavelength-dependent shortening in pf18 and in fla10 cells induced to shorten flagella by incubation at 33 degrees C. Colchicine blocked tubulin addition but did not affect the exchange of the other exchangeable polypeptides; nor did it induce any change in flagellar length. Basal bodies also incorporated newly synthesized proteins. These data reveal that Chlamydomonas flagella are dynamic structures that incorporate new protein both during steady state and as flagella shorten and that protein exchange does not, alone, explain length regulation.     

Calcium regulation of flagellar curvature and swimming pattern in triton X-100--extracted rat sperm

Free Ca2+ changes the curvature of epididymal rat sperm flagella in demembranated sperm models. The radius of curvature of the flagellar midpiece region was measured and found to be a continuous function of the free Ca2+ concentration. Below 10(-7) M free Ca2+, the sperm flagella assumed a pronounced curvature in the same direction as the sperm head. The curvature reversed direction at 2.5 x 10(-6) M Ca2+ to assume a tight, hook-like bend at concentrations of 10(-5) to 10(-4) M free Ca2+. Sodium vanadate at 2 x 10(-6) M blocked flagellar motility, but did not inhibit the Ca2+-mediated change in curvature. Nickel ion at 0.2 mM and cadmium ion at 1 microM interfered with the transition and induced the low Ca2+ configuration of the flagellum. The forces that maintain the Ca2+-dependent curvature are locally produced, as dissection of the flagella into segments did not significantly alter the curvature of the excised portions. Irrespective of the induced pattern of curvature, the sperm exhibited coordinated, repetitive flagellar beating in the presence of ATP and cAMP. At 0.3 mM ATP the flagellar waves propagated along the principal piece while the level of free Ca2+ controlled the overall curvature. When Ca2+-treated sperm models with hooked midpieces were subjected to higher concentrations of ATP (1-5 mM), some cells exhibited a pattern of movement similar to hyperactivated motility in capacitated live sperm. This type of motility involved repetitive reversals of the Ca2+-induced bend in the midpiece, as well as waves propagated along the principal piece. The free Ca2+ available to the flagellum therefore appeared to modify both the pattern of motility and the flagellar curvature.     

Exceptions to the prevailing pattern of tubules (9 + 9 + 2) in the sperm flagella of certain insect species

Various deviations from classical 9 + 2 flagellar structure are found in sperm of insect species. In mature spermatozoa of a psocid, Psocus, the outer flagellar tubules are not straight, but are disposed in a long-pitched helix such that they form an angle of about 8° with a single dense rod located in the position usually occupied by the central pair. In young spermatids of Psocus the outer tubules are straight; thus, spiraling of the flagellar tubules occurs during the course of spermiogenesis. Spiraling of flagella also occurs in the cat flea Ctenocephalides felis. Variations in the number and morphology of the central element or elements occur in other insect species besides Psocus. Among the observed deviations from a central pair of tubules are a 9 + 0 tubule pattern in the sperm of three species of mayflies, a 9 + 1 tubule pattern in the sperm of two species of mosquitoes, and 9 + 7 tubules in sperm of two species of caddis flies. Spermatozoa of treehoppers vary in yet another respect from the typical 9 + 9 + 2 insect flagellum. These sperm tails branch into four long tails, three of which each contain two doublet and two singlet tubules while the fourth branch contains three doublet and three singlet tubules. The wide distribution of insects with aberrant flagella suggests that the variant forms have evolved independently.     

[A functional flagella with a 6 + 0 pattern]

The male gamete of the Gregarine Lecudina tuzetae has been studied with transmission electron microscopy and microcinematography. It is characterized by a flagellar axoneme of 6 + 0 pattern, a reduction of the chondriome, and the abundance of storage polysaccharide or lipid bodies. The movements of the flagella are of the undulating type and they are performed in the three dimensions of space. They are very slow, with a cycle time of about 2s. The structure of the axoneme components are similar to those of flagella with a 9 + 2 pattern. Each doublet has overall dimensions of 350 x 220 A; the space between the adjacent doublets is about 160 A. The A subfiber bears arms like dynein arms. The diameter of the axoneme is about 1,000 A. The basal body consists of a cylinder of dense material 2,500 A long and 1,300-1,400 A in diameter; a microtubule 200 A in diameter is present in the axis. This study shows that a 6 + 0 pattern can generate a flagellar movement. The mechanism of the flagellar movement of the male gamete of L. tuzetae does not require the presence of central microtubules and it would include molecular interactions of the dynein-tubulin type between the adjacent peripheric doublets. The slowness of the movements is discussed in terms of the axoneme's structure and its energy supply. Finally, the phylogenetic significance of this flagella is examined on the basis of the morphopoietic potentialities of the centriolar structures.     

Selective inhibition of flagellar activity in Chlamydomonas by nickel.

Flagellar activity in the biflagellate chlorophyte Chlamydomonas reinhardtii is selectively inhibited by Ni2+ or by treatment with Ca2+-chelating agents. Inhibitions of swimming speed, geotaxis, phototaxis, and pattern swimming result from qualitative and quantitative losses in the activity of individual flagella and in the coordination of activity between the 2 flagella of each cell. Addition of Ca2+ (a) prevents inhibition and (b) restores normal flagellar activity in inhibited cells. Mg2+ is partially effective in reversal of inhibition. Other ions do not cause similar inhibition or reversal of nickel inhibition. The characteristics of inhibition and reversal suggest that the primary target for nickel is a component of the flagellar apparatus, and that this component uses Ca2+ to perform its normal function in the regulation of flagellar activity. A 2nd target for nickel is a Ca-requiring process specific to phototaxis (and not involved in the photophobic response).     

FLAGELLAR DEVELOPMENT AND REGENERATION IN VOLVOX CARTERI (CHLOROPHYTA)1

The biflagellate somatic cells of Volvox carteri f. nagariensis lyengar exhibit an asymmetric pattern of flagellar development. Initiallt each somatic cell has two short (4 μm) flagella but after several hours one flagellum on each cell elongates unitl it reaches a length of 12 μm. Due to the regular arrangement of somatic cells in the Volvox spheroid it is apparent that the same flagellum on each somatic is the first to elongale. The asymmetric flagellar length is maintained for about 8 h after which the second flagellum on each somatic cell elongates. When the second flagellum attains the same length (12 μm) as the first flagellum, both flagella elongale at the same rate until reaching a final length of 22 μm. Experimental removal of somatic cell flagella results in their regeneration. Somatis cells regenerate both flagella simultaneously and full length flagella are produced in about 2 h. The intial rate of flagellar regeneration is about ten times faster than the intial rate of flagllar growth in development. Cycloheximide, an inhibitor of protein synthesis, has no effect on the initial rate of flagellar regeneration but the flagella produced in the presence of the drug are half the length of flagella produced in its absence. Somatic cells are able to regenerate flagella up to the time of α and β tubulin, the major structural proteins of the flagellar axoneme, and other cellular proteins.     

A backward swimming mutant of Chlamydomonas reinhardii

A backward swimming mutant (RL-10) was isolated from Chlamydomonas reinhardii. In contrast to the wild-type flagellum which usually displays a ciliary type beating pattern, the flagella in the RL-10 cells always propagated such undulating waves as found in sperm flagella. This abnormal beating pattern was maintained after the cell was demembranated by a non-ionic detergent (Nonidet P40) and reactivated with ATP. Reactivated axonemes (demembranated flagella) of the wild-type cells changed the beating pattern from the ciliary type to the flagellar type when the Ca²⁺ concentration was increased from 10⁻⁷ to 10⁻⁶ M. However, the RL-10 axonemes did not show such a Ca-dependent change in the beating pattern. Hence the RL-10 flagella might carry defects in the controlling mechanisms of flagellar beating pattern, at sites other than the membrane.     

Rapid and slow mechanisms for loss of cell adhesiveness during fertilization in Chlamydomonas

Although vegetative cells, gametes, and zygotes of the biflagellated alga Chlamydomonas bear flagella, only the flagella of mt+ and mt- gametes are adhesive. The molecules responsible for adhesiveness, mt+ and mt- agglutinins, are long rod-shaped glycoproteins displayed on the flagellar membrane. These flagellar agglutinins, which gametes use both as adhesion and signaling molecules during the early events of fertilization, are lost from the flagella during adhesion. Flagellar adhesiveness can be maintained, however, by recruitment and activation of preexisting, inactive agglutinins from the plasma membrane of the cell body (Hunnicutt et al, 1990, J. Cell Biol. 111, 1605-1616) unless the gametes of opposite mating types fuse to form zygotes. Upon cell fusion, flagellar adhesiveness is lost. In the studies presented here, we have employed an in vitro bioassay to measure agglutinins in both cell bodies and flagella at various times during gametogenesis, during fertilization, and after zygote-formation. By use of the bioassay, which can detect agglutinins that are functionally inactive in vivo, we found that vegetative cells are devoid of agglutinins. These adhesion molecules appear only after gametogenesis is underway with the cell body agglutinins appearing first and then the flagellar agglutinins. Surprisingly, 30 min after zygote formation, when the zygotes' flagella are no longer adhesive, the flagellar agglutinin activity detectable with the bioassay remains high. One interpretation of these results is that zygotes continue to recruit agglutinins from the cell body to the flagella, but cell fusion abrogates activation of the agglutinins. Within 45-90 min after fusion both the cell body and flagellar agglutinins are lost and can be detected in the medium. These mechanisms, which render the zygotes nonadhesive to other zygotes and unmated gametes, contribute to the Chlamydomonas equivalent of a block to polyspermy.     

Anomalies in the motion dynamics of long-flagella mutants of Chlamydomonas reinhardtii

Chlamydomonas reinhardtii has long been used as a model organism in studies of cell motility and flagellar dynamics. The motility of the well-conserved ‘9+2’ axoneme in its flagella remains a subject of immense curiosity. Using high-speed videography and morphological analyses, we have characterized long-flagella mutants (lf1, lf2-1, lf2-5, lf3-2, and lf4) of C. reinhardtii for biophysical parameters such as swimming velocities, waveforms, beat frequencies, and swimming trajectories. These mutants are aberrant in proteins involved in the regulation of flagellar length and bring about a phenotypic increase in this length. Our results reveal that the flagellar beat frequency and swimming velocity are negatively correlated with the length of the flagella. When compared to the wild-type, any increase in the flagellar length reduces both the swimming velocities (by 26–57%) and beat frequencies (by 8–16%). We demonstrate that with no apparent aberrations/ultrastructural deformities in the mutant axonemes, it is this increased length that has a critical role to play in the motion dynamics of C. reinhardtii cells, and, provided there are no significant changes in their flagellar proteome, any increase in this length compromises the swimming velocity either by reduction of the beat frequency or by an alteration in the waveform of the flagella.     

Effects of mechanical and chemical disruption on the ATP-phosphohydrolase activity and ultrastructure of sperm flagella.

The Ca²⁺- and Mg²⁺-ATP-phosphohydrolase (ATPase) activities measurable in suspensions of and extracts from bull epididymal spermatozoa flagella, sonicated for 2 min at 8 W, were compared with those of control flagella after 0, 3, and 18 h dialysis against an alkaline 0.5 M KCl extracting solution. Activity was measured in the absence of and in the presence of oligomycin. The effects of sonication and extraction on structural components within the flagella were visualized electron microscopically. Sonication caused fragmentation of flagella, extensive disruption of mitochondria and an immediate (0 h) increase in both Ca²⁺- and Mg²⁺-ATP-ase activity. Prolonged dialysis resulted in solubilization of specific flagellar structures, partial disruption of mitochondrial integrity, and increases in ATPase activity. Mg²⁺-ATPases of flagellar suspensions and extracts were greater than Ca²⁺-ATPases, and a part of this Mg²⁺-ATPase activity was inhibited by oligomycin. Therefore, Mg²⁺-ATPases from disrupted mitochondria contribute to the Mg²⁺-ATPase activities measurable in suspensions of and extracts from bull sperm flagella. This study emphasizes the necessity of evaluating the effects of the mechanical and chemical treatments used in fractionating cells before interpreting the biochemical information derived from their isolated components.     

Tipping and mating-structure activation induced in Chlamydomonas gametes by flagellar membrane antisera

Antisera raised against vegetative and gametic flagella of Chlamydomonas reinhardi have been used to probe dynamic properties of the flagellar membranes. The antisera, which agglutinate cells via their flagella, associate with antigens that are present on both vegetative and gametic membranes and on membranes of both mating types (mt+ and mt-). Gametic cells respond to antibody presentation very differently from vegetative cells, mobilizing even high concentrations of antibody towards the flagellar tips; the possibility is discussed that such "tipping" ability reflects a differentiated gametic property relevant to sexual agglutinability. Gametic cells also respond to antibody agglutination by activating their mating structures, the mt+ reaction involving a rapid polymerization of microfilaments. Several impotent mt+ mutant strains that fail to agglutinate sexually are also activated by the antisera and procede to form zygotes with normal mt- gametes. Fusion does not occur between activated cells of like mating type. Monovalent (Fab) preparations of the antibody fail to activate mt+ gametes, suggesting that the cross-linking properties of the antisera are essential for their ability to mimic, or bypass, sexual agglutination.     

Electron microscope study of spermatogenesis in two cestodes acanthobothrium filicolle benedenii Loennberg, 1889 and Onchobothrium uncinatum (Rud., 1819) (Tetraphyllidea, Onchobothriidae) (author's transl)]

The spermatogenesis, spermatozoon differentiation and fine structure of Acanthobothrium filicolle benedenii Loennberg, 1889 and Onchobothrium uncinatum (Rud., 1819) were studied by means of light and electron microscopy. Spermatogenesis in both species is of the rosette type. During sperm differentiation, the five following stages have been distinguished: (1) formation of arching membranes and differentiation zone; (2) nuclear elongation; (3) formation of two flagella with a median cytoplasmic process; (4) flagellar rotation; (5) fusion of two flagella with the median cytoplasmic process induced by the migration of nucleus into the latter. The mature spermatozoa of both species are threadlike structures about 250 mum long. They consist of two axonemes of the platyhelminth type (9+1 pattern), elongated nucleus and cortical microtubules embedded in a cytoplasmic matrix containing numerus beta-glycogen particles. The body which appears on cross-sections as crest, lateral with respect to the axoneme, is present in both species. Such a body is reported for the first time in Cestode spermatozoa. There is no mitochondrion in the two Onchobothriidae sperms studied.     

Selective Inhibition of Flagellar Activity in Chlamydomonas by Nickel

Flagellar activity in the biflagellate chlorophyte Chlamydomonas reinhardtii is selectively inhibited by Ni²⁺ or by treatment with Ca²⁺-chelating agents. Inhibitions of swimming speed, geotaxis, phototaxis, and pattern swimming result from qualitative and quantitative losses in the activity of individual flagella and in the coordination of activity beween the 2 flagella of each cell. Addition of Ca²⁺ (a) prevents inhibition and (b) restores normal flagellar activity in inhibited cells. Mg²⁺ is partially effective in reversal of inhibition. Other ions do not cause similar inhibition or reversal of nickel inhibition. The characteristics of inhibition and reversal suggest that the prmary target for nickel is a component of the flagellar apparatus, and that this component uses Ca²⁺ to perform its normal function in the regulation of flagellar activity. A 2nd target for nickel is a Carequiring process specific to phototaxis (and not involved in the photophobic response).     

Receptor-Mediated Calcium Influx in Chlamydomonas reinhardtii

ABSTRACT. Alcian blue acts as a secretagogue and chemorepellent in a variety of unicellular eukaryotes. We report that alcian blue stimulates flagellar excision and induction of RNA encoding flagellar proteins in Chlamydomonas reinhardtii . Flagellar excision by alcian blue is dependent on extracellular Ca²⁺ and is blocked by La³⁺, ruthenium red, and neomycin, and so is similar to flagellar excision by acid shock. However, the adf-l mutant excises its flagella following alcian blue treatment, but not following acid shock, thus genetically distinguishing alcian-blue-induced excision from acid-shock-induced excision. Wild-type, but not adf-1, cells regrow their flagella in the continued presence of alcian blue. Wild-type cells that regrow flagella in the presence of alcian blue fail to excise their flagella in response to either increased concentrations of alcian blue or to acid shock. Alcian blue treatment of cells also induces RNA encoding flagellar components, but in a manner distinct from other means of stimulation. These results suggest that treating Chlamydomonas with the secretagogue alcian blue initiates a Ca²⁺ influx pathway and that prolonged treatment with alcian blue desensitizes the acid-shock-activated Ca²⁺ influx pathway to acid treatment. Alcian blue will thus be a useful excitatory ligand in future studies of receptor-mediated Ca²⁺ signaling in the Chlamydomonas flagellar regeneration system.     

Protection against Pseudomonas aeruginosa infection by passive transfer of anti-flagellar serum

The specificity of adsorbed flagellar antisera for H-antigen was demonstrated in vitro by cross-agglutination assays, motility inhibition, and an ELISA. The specific flagellar antibody was determined to be an IgG. Complete protection against burn wound sepsis was achieved with flagellar antisera. Cross-protection experiments revealed that protection was not only H-antigen dependent, but specific for the flagella antigen type. Antiserum raised against b-type flagella would only protect against homologous bacterial challenge and not against a-type flagellated strains. Results using a-type antisera were consistent, giving protection only against the homologous strain. In contrast, protective capacity was selectively removed from antisera by adsorbing with Fla+ cells. Bacteria colonized the burn wounds of passively protected mice to similar levels as seen in nonprotected animals, but the colonization remained localized and did not result in systemic infection, a pattern similar to infections with motility mutants observed in other studies. Animals rendered neutropenic prior to burning were not protected with flagellar antisera. These data suggested a role for phagocytic cells in protection. Immobilization by flagellar antiserum was observed both by microscopic studies and by inhibition of colony spreading. Antiflagellar antibody is hypothesized as exerting its protective capacity possibly in two ways; first by inhibiting the motility of invading bacteria by binding to the flagellum and immobilizing the bacteria, and secondly by acting as an opsonin, targeting either immobilized or mobile cells for phagocytosis.     

Flagellar microtubule dynamics in Chlamydomonas: cytochalasin D induces periods of microtubule shortening and elongation; and colchicine induces disassembly of the distal, but not proximal, half of the flagellum

To study the mechanisms responsible for the regulation of flagellar length, we examined the effects of colchicine and Cytochalasin D (CD) on the growth and maintenance of Chlamydomonas flagella on motile wild type cells as well as on pf 18 cells, whose flagella lack the central microtubules and are immobile. CD had no effect on the regeneration of flagella after deflagellation but it induced fully assembled flagella to shorten at an average rate of 0.03 microns-min. Cells remained fully motile in CD and even stubby flagella continued to move, indicating that flagellar shortening did not selectively disrupt machinery necessary for motility. To observe the effects of the drug on individual cells, pf 18 cells were treated with CD and flagella on cells were monitored by direct observation over a 5-hour period. Flagella on control pf 18 cells maintained their initial lengths throughout the experiment but flagella on CD-treated cells exhibited periods of elongation, shortening, and regrowth suggestive of the dynamic behavior of cytoplasmic microtubules observed in vitro and in vitro. Cells behaved individually, with no two cells exhibiting the same flagellar behavior at any given time although both flagella on any single cell behaved identically. The rate of drug-induced flagellar shortening and elongation in pf 18 cells varied from 0.08 to 0.17 microns-min-1, with each event occurring over 10-60-min periods. Addition of colchicine to wild type and pf 18 cells induced flagella to shorten at an average rate of 0.06 microns-min-1 until the flagella reached an average of 73% of their initial length, after which they exhibited no further shortening or elongation. Cells treated with colchicine and CD exhibited nearly complete flagellar resorption, with little variation in flagellar length among cells. The effects of these drugs were reversible and flagella grew to normal stable lengths after drug removal. Taken together, these results show that the distal half to one-third of the Chlamydomonas flagellum is relatively unstable in the presence of colchicine but that the proximal half to two-thirds of the flagellum is stable to this drug. In contrast to colchicine, CD can induce nearly complete flagellar microtubule disassembly as well as flagellar assembly. Flagellar microtubules must, therefore, be inherently unstable, and flagellar length is stabilized by factors that are sensitive, either directly or indirectly, to the effects of CD.     

Ca2+ influx activated by low pH in Chlamydomonas

Cytosolic acidification stimulates an influx of Ca2+ which results in shedding of the two flagella of Chlamydomonas. Ca2+ influxes are also involved in the photoresponses of this alga, but it is not understood how the acidification-activated Ca2+ influx is distinguished from the Ca2+ influxes which mediate phototaxis and the photophobic response. The present study focuses on the deflagellation-inducing Ca2+ influx pathway. Influx occurs through an ion channel or transporter with low abundance or low permeability to Ca2+ (approximately 500 fmol/s/10(6) cells in 50 microM Ca2+). Ca2+ influx was potently blocked by Cd3+ (EC50 approximately 5 microM), but was insensitive to Cd2+ (Quarmby, L.M., and H.C. Hartzell. 1994. J. Cell Biol. 124:807) and organic blockers of Ca2+ channels including SKF-96365 (up to 100 microM) and flufenamic acid (up to 1 mM). Experiments with a flagella-less mutant (bald-2), isolated flagella, and a blocker of flagellar assembly (colchicine) indicated that the acidification-stimulated Ca2+ influx pathway is not localized to the flagellar membrane. The acid-stimulated influx pathway was transiently inactivated after cells shed their flagella. Inactivation did not occur in the deflagellation mutant, fa- 1, although acidification-stimulated Ca2+ influx was normal. This suggests that inactivation of this pathway in wild-type cells is probably not a direct consequence of acidification nor of Ca2+ influx, but may be related to deflagellation. Recovery of deflagellation- inducing Ca2+ influx occurred within 30 min after a 30 s exposure to acid and did not require flagellar assembly. The regulation, drug sensitivity, and subcellular localization identify acidification- stimulated Ca2+ influx as a specific Ca2+ entry pathway distinct from established Ca2+ channels.     

Flagellar elongation and shortening in Chlamydomonas. IV. Effects of flagellar detachment, regeneration, and resorption on the induction of flagellar protein synthesis

Synthesis of new proteins is required to regenerate full length Chlamydomonas flagella after deflagellation. Using gametes, which have a low basal level of protein synthesis, it has been possible to label and detect the synthesis of many flagellar proteins in whole cells. The deflagellation-induced synthesis of the tubulins, dyneins, the flagellar membrane protein, and at least 20 other proteins which co- migrate with proteins in isolated axonemes, can be detected in gamete cytoplasm, and the times of initiation and termination of synthesis for each of the proteins can be studied. The nature of the signal that stimulates the cell to initiate flagellar protein synthesis is unknown. Flagellar regeneration and accompanying pool depletion are not necessary for either the onset or termination of flagellar protein synthesis, because colchicine, which blocks flagellar regeneration, does not change the pattern of proteins synthesized in the cytoplasm after deflagellation or the timing of their synthesis. Moreover, flagellar protein synthesis is stimulated after cells are chemically induced to resorb their flagella, indicating that the act of deflagellation itself is not necessary to stimulate synthesis. Methods were defined for inducing the cells to resorb their flagella by removing Ca++ from the medium and raising the concentration of K+ or Na+. The resorption was reversible and the flagellar components that were resorbed could be re-utilized to assemble flagella in the absence of protein synthesis. This new technique is used in this report to study the control of synthesis and assembly of flagella.     

Formation of flagella during interphase in secondary spermatocytes from Xenopus laevis in vitro

In cell culture, single motile flagella, 1 micron in length, were observed to grow from secondary spermatocytes of Xenopus laevis within 2-3 hours after telophase I, at 22 degrees C. About 90% of the secondary spermatocytes formed flagella as observed by phase-contrast microscopy. The flagella grew up to 2-6 microns in length during interphase II, which lasted about 18 hours. The presence of the "9 + 2" microtubular structure of the flagellar axonemes of secondary spermatocytes was confirmed by electron microscopy. When chromosomal condensation began (prophase II), the flagella were resorbed into the cells and, after the second meiotic division, a flagellum was formed again by each of the round spermatids. Thus, there appears to be a close relationship between the meiotic division cycle and the formation of flagella. The possible contribution of Sertoli cells to the formation of flagella in secondary spermatocytes was examined by reducing the number of Sertoli cells to less than ten per culture. Under these conditions, flagella formed in secondary spermatocytes with very high efficiency. It is very likely that secondary spermatocytes form flagella in vivo, since the secondary spermatocytes were observed to have flagella immediately after dissociation of the testes.