The movement of eukaryotic flagella and cilia is regulated by intracellular calcium. We have tested a model in which the central pair of microtubules mediate the effect of Ca2+ to modify the dynein activity. We used a novel microtubule sliding assay that allowed us to test the effect of Ca2+ in the presence or absence of the central-pair microtubules. When flagellar axonemes of sea-urchin sperm were exposed to ATP in the presence of elastase, they showed different types of sliding disintegration depending on the ATP concentration: at low concentrations of ATP (≤50μM), all the axonemes were disintegrated into individual doublets by microtubule sliding; by contrast, at high ATP concentrations (≥100 μM),a large proportion of the axonemes showed limited sliding and split lengthwise into a pair of two microtubule bundles, one of which was thicker than the other. The sliding behaviour of the axonemes was also influenced by Ca2+. Thus, at 1 mM ATP, the proportion of axonemes that split into two bundles increased from 25% at <10–9 M Ca2+to 60% at 10–4 M Ca2+, whereas the sliding velocity of doublets during the splitting did not change. Electron microscopy of split bundles showed that the thicker bundles contained five or six doublets and the central pair, whereas the thinner bundles contained three or four doublets but not the central pair. Closer examinations revealed that the thicker bundles were dominated by four patterns of doublet combinations:doublets 8-9-1-2-3-4, 8-9-1-2-3, 4-5-6-7-8 and 3-4-5-6-7-8. This indicates that the sliding occurred preferentially at one or two fixed interdoublet sites on either side of the central-pair microtubules, whereas the sliding at the remaining interdoublet sites was inhibited under these conditions. Ca2+ reduced the appearance of the 4-5-6-7-8 and 3-4-5-6-7-8 patterns and increased the 8-9-1-2-3-4 and 8-9-1-2-3 patterns. The splitting patterns are possibly related to the switching mechanism of the dynein activity underlying the cyclical flagellar bending. To investigate the role of the central pair in the regulation of the dynein activity by Ca2+,we studied the behaviour of singlet microtubules applied to the dynein arms exposed on the doublets of the split bundles that were either associated with the central pair or not. Microtubules moved along both the thicker and the thinner bundles but the frequency of microtubule sliding on the thinner (i.e. the central-pair-less) bundles was three to four times (at≤10–5 M Ca2+) and ten times (at 10–4 M Ca2+) as large as that on the thicker,central-pair-associated bundles. Furthermore, the velocity of microtubule sliding at 1 mM ATP on the thicker bundles were significantly reduced by 10–7-10–4 M Ca2+, whereas that on the thinner bundles was not changed by the concentration of Ca2+. These results indicate that Ca2+ inhibits the activity of dynein arms on the doublets through a regulatory mechanism that involves the central pair and the radial spoke complex. This mechanism might control the switching of the dynein activity within the axoneme to induce the oscillatory bending movement of the flagellum.
Introduction
Our goal is to determine the mechanism by which Ca2+ regulates flagellar motility. Changes in the intracellular concentration of Ca2+ have profound effects on the bending patterns of cilia and flagella (Naitoh and Kaneko,1972; Izumi and Miki-Noumura,1985; Holwill and McGregor,1976). In Chlamydomonas flagella, an increase in the intracellular Ca2+ changes the flagellar waveform from an asymmetrical ciliary type to a symmetrical flagellar type(Hyams and Borisy, 1978; Bessen et al., 1980; Omoto and Brokaw, 1985). In sea urchin sperm flagella, Ca2+ increases the asymmetry of waveform, ultimately inducing a stoppage (`quiescence') of flagellar beating above 10–4 M Ca2+(Brokaw, 1979; Gibbons and Gibbons, 1980). It is likely that the effect of Ca2+ on the flagellar movement is related to the regulation of microtubule sliding, but the mechanism of regulation by Ca2+ has been poorly understood.
The central pair of microtubules (central apparatus) and the radial spokes are thought to form a complex (CP/RS) that regulates the activity of dynein arms. The rotation of the central pair during ciliary and flagellar beating suggests that the central pair might act as a `distributor' to regulate the activity of dynein arms (Omoto et al.,1999). The CP/RS is also implicated in the Ca2+-dependent conversion of the waveforms. In Chlamydomonas, the flagella of mutants lacking the CP/RS do not beat,even though their axonemes are capable of sliding disintegration(Witman et al., 1978; Kamiya, 2002). At low ATP concentrations (<20 μM), however, the CP/RS-deficient axonemes are capable of beating and, in response to Ca2+, convert the waveform from an asymmetrical (at <10–6 M Ca2+) to a symmetrical (at ≥10–5 M Ca2+) pattern, although they are non-beating at high ATP concentrations(Wakabayashi et al., 1997). At physiological concentrations of ATP (∼1 mM), the central-pair microtubules are probably involved in the response to Ca2+ in the wild type(Hosokawa and Miki-Noumura,1987) as well as in the CP/RS mutants of Chlamydomonas(Smith, 2002b). The ATP-dependent role of the CP/RS in the Ca2+ regulation is also observed in sea urchin sperm flagella. Based on the analysis of flagellar waveforms of flagella `beating' under imposed head vibration(Gibbons et al., 1987), we showed that Ca2+ decreased the velocity of microtubule sliding through a trypsin-sensitive regulatory mechanism, which possibly involves the central-pair microtubules (Bannai et al.,2000). This was also supported by the artificially induced rotation of the Ca2+-induced asymmetrical bending pattern. The rotation of the beating plane under imposed head vibration occurred only at high ATP concentrations (≥100 μM), indicating that the regulation of dynein activity through the central pair occurs only at high ATP concentrations (Bannai et al.,2000). These studies suggest that, at least at ATP concentrations higher than ∼100 μM, high concentrations of Ca2+ modify the regulatory signal from the central pair, which is mediated by the radial spokes to control the activity of the dynein arms.
In this study, we tested the idea that Ca2+ alters the dynein activity, using a novel microtubule sliding assay. As the protease necessary to initiate sliding between the doublets, we used elastase(Brokaw, 1980) instead of trypsin, although trypsin is more widely used, to digest axonemal structures that restrict free sliding of the doublets(Summers and Gibbons, 1971). Unlike the trypsin-treated axonemes, which do not show oscillatory bending movements in response to local, repetitive application of ATP, the elastase-treated axonemes were capable of oscillatory bending movements,indicating that they retained certain regulatory mechanisms for producing local cyclical bending (Shingyoji and Takahashi, 1995).
When the elastase-treated axonemes were exposed to physiological concentrations (e.g. 1 mM) of ATP, they split lengthwise into two unequal microtubule bundles: a thicker bundle that contained the central-pair microtubules and a thinner one that did not. We found that 10–4 M Ca2+ did not affect the velocity at which the two bundles slid along each other during the splitting, although it affected the patterns into which the axonemes split. Electron microscopic analysis of the split bundles showed that the axonemes were split into a pair of bundles at some preferred interdoublet sites that were close to either of the central pair.
To study the effect of the central pair on the dynein activity at high Ca2+, we used the microtubule sliding assay developed by Yoshimura and Shingyoji (Yoshimura and Shingyoji,1999), which was a modification of the method of Shingyoji et al.(Shingyoji et al., 1998). In the present study, we analysed the behaviour of singlet microtubules that were made to interact with either the thicker or the thinner bundles obtained from the splitting of the axonemes. We found that the velocity of microtubule sliding on the thicker bundles, which contained the central-pair microtubules,was significantly decreased by 10–7-10–4 M Ca2+. The frequency of microtubule sliding on the bundles was significantly reduced by the presence of the central pair but was not affected by Ca2+ except at 10–4 M. These results indicate that Ca2+ alters the sliding activity of dynein arms in flagella through the CP/RS complex.
Materials and Methods
Axonemes and microtubules
Sperm of the sea urchins Pseudocentrotus depressus and Clypeaster japonicus were suspended in the same volume of Ca2+-free artificial sea water (465 mM NaCl, 10 mM KCl, 25 mM MgSO4, 25 mM MgCl2 and 2 mM Tris-HCl, pH 8.0) and demembranated with 15 volumes of demembranating solution [0.04% (w/v) Triton X-100, 0.15 M potassium acetate, 2 mM MgSO4, 2 mM glycoletherdiamine-N,N,N′,N′-tetra-acetic acid(EGTA), 1 mM dithiothreitol (DTT), 10 mM Tris-HCl, pH 8.0] for 45 seconds at 20-23°C (Pseudocentrotus) or 24-27°C (Clypeaster). The demembranated sperm suspension was diluted with 16 volumes of reactivating solution [0.15 M potassium acetate, 2 mM MgSO4, 2 mM EGTA, 1 mM DTT, 2% (w/v) polyethyleneglycol (MW 20,000) and 10 mM Tris-HCl, pH 8.0]without ATP. For Clypeaster (in the new sliding assay experiment), 20 mM HEPES (pH 7.8) was used instead of Tris (pH 8.0) in both demembranating and reactivating solutions. Preparations that showed less than 90% reactivation at 1 mM ATP were discarded.
The demembranated sperm were fragmented by passing a sperm suspension through a 22-gauge hypodermic needle (for sliding disintegration experiments and electron microscopy) or by homogenizing the sperm suspension (for new sliding assay experiments). Sperm heads were removed by centrifugation at 2,000-4000 g; the supernatant containing axonemal fragments was centrifuged at 17,000-20,000 g and resuspended in reactivating solution (without ATP). For observation of the sliding at 1 mM ATP (0.6-0.9 mM MgATP) and 10–4 M Ca2+, ATP and CaCl2 were added to the reactivating solution following calculations according to Goldstein(Goldstein, 1979).
For the new sliding assay experiments, demembranated sperm were labelled with tetramethylrhodamine before fragmentation, and singlet microtubules were assembled from tetramethylrhodamine-labelled bovine tubulin according to Yoshimura and Shingyoji (Yoshimura and Shingyoji, 1999). Strongly labelled singlet microtubules were distinguished from weakly labelled fragmented axonemes by adjusting the illumination.
Observation of sliding disintegration and microtubule sliding
Sliding disintegration of doublet microtubules was observed by the method described by Takahashi et al. (Takahashi et al., 1982). The axonemal fragments in ATP-free reactivating solutions containing or not containing CaCl2 were first placed between two coverslips separated by two strips of plastic adhesive tape used as spacers, thus forming a chamber that was open on two sides. They were then perfused with the reactivating solution containing the same concentration of CaCl2 and 20 μM or 1 mM ATP. Finally, they were perfused with a reactivating solution containing elastase (5 μg ml–1elastase, Sigma type III and 5 μg ml–1 trypsin inhibitor,Sigma type I-S) as well as the same concentrations of CaCl2 and ATP.
Observation was made using an inverted microscope (Nikon, TMD) fitted with a 40× objective lens (Nikon), a dark-field condenser (Nikon,NA=0.95-0.80), heat-absorbing filters and a halogen lamp. Sliding disintegration was recorded on videotape using a CCD camera (Hamamatsu Photonics, C2400-77) and a U-matic videocassette recorder (Sony, VO-5800),enhancing the contrast of the image with a control unit (Hamamatsu Photonics,ARGUS-10) and a CCD camera unit (Hamamatsu Photonics, C2400). To analyse the sliding velocity, the recorded images were imported into a personal computer by using LF-3 Scientific Frame Grabber (Scion Corporation) and the displacement of the doublet microtubule was measured by using the NIH Image software. The sliding velocity was determined by plotting the displacement against time and applying the least-squares method to the plots.
For the observation of microtubule sliding on the doublet bundles in the new sliding assay experiment, we first induced sliding disintegration of elastase-treated rhodamine-labelled axonemes by the same procedure as used in the previous study (Yoshimura and Shingyoji, 1999), except for the following: the two sides of the slide of the 5 μl chamber were sealed with adhesive plastic tape instead of enamel in order to obtain a constant depth of the perfusion solution. A suspension of axonemal fragments was first introduced into the chamber and treated with elastase (5 μg ml–1 elastase, Sigma type III and 5 μg ml–1 trypsin inhibitor, Sigma type I-S, in the reactivating solution containing 10–4 M Ca2+without ATP) for 1-1.5 minutes. The elastase treatment was stopped with ovoinhibitor (50 μg ml–1 ovoinhibitor, Sigma type IV-O, in reactivating solution containing 10–4 M Ca2+without ATP), and followed by perfusion with casein (∼1 mg ml–1 in reactivating solution containing 10–4 M Ca2+ without ATP). The perfusion of reactivating solution containing 10–4 M Ca2+ and 1 mM ATP (0.6 mM MgATP) then induced sliding disintegration of the axonemal fragments into two parts. As the next step, to reduce the Ca2+concentration, the assay buffer (see below) without EGTA was perfused followed by perfusion of the assay buffer containing 1 mM ATP (0.9 mM MgATP) with or without Ca2+. The rhodamine-labelled microtubules were suspended in the assay buffer to which 2% (v/v) β-mercaptoethanol, 40 mM glucose, 430μg ml–1 glucose oxidase, 70 μg ml–1catalase, 10 μM Taxol and the same concentration of MgATP and Ca2+ had been added. The assay buffer contained 70 mM potassium acetate, 5 mM magnesium acetate, 20 mM HEPES, 2 mM EGTA, 0.1 mM ethylenediaminetetraacetic acid (EDTA) and 1 mM DTT, pH 7.8. The concentrations of Ca2+ in the solution used for the observation of microtubule sliding were <10–9 M, 10–7 M,10–5 M and 10–4 M.
Microtubule sliding on doublet bundles was observed under a fluorescent microscope (Olympus, BX 60) with a 100× oil-immersion objective lens(Olympus PlanApo, NA=1.4) and recorded on videotape using a high-sensitivity silicon-intensified target camera (SIT camera, Hamamatsu Photonics, C2400-08)and a VHS videocassette recorder. To analyse the sliding velocity, the video images were traced by hand from the screen of a video monitor onto a sheet of transparent film. The sliding velocity was determined from the time measured by counting the number of video fields and the distance of microtubule movement measured on the traced images.
Electron microscopy
The axonemal fragments were resuspended in the reactivating solution containing 1 mM ATP (0.6 mM MgATP) and Ca2+ (10–4M, 10–3 M or <10–9 M) and incubated at 20°C for 5 minutes. After reactivation, the axonemes were digested with elastase (5 μg ml–1 elastase, Sigma type III and 5 μg ml–1 trypsin inhibitor, Sigma type I-S) for 7 minutes. To stop the digestion, glutaraldehyde was added to a final concentration of 2% on ice. After 15 minutes fixation at 10°C, the axonemes were pelleted by centrifugation at 28,000 g for 10 minutes at 4°C. The pelleted samples were fixed with 2% glutaraldehyde in the rinse buffer (0.15 M potassium acetate, 25 mM MgSO4 and 10 mM phosphate buffer, pH 8.0)for 1 hour at 0°C. Then the samples were washed with the rinse buffer,postfixed in 1% OsO4 in the rinse buffer for 45 minutes at 0°C,dehydrated in a graded ethanol series and embedded in Epon 812 (TAAB). Silver-gold sections were cut with a glass knife on Sorvall, MT2-B ultramicrotome, collected on copper grids, stained with uranyl acetate and lead citrate, and observed at 80 kV with a JEOL JEM100CX electron microscope.
When the sliding pattern of disintegrated axonemes was studied, we examined thicker bundles consisting of doublets and the central pair. We used the conventional numbering system for the doublets originally proposed by Afzelius(Afzelius, 1959). The doublet microtubules in the thicker bundles were identified from the position of the 5-6 bridge and the orientation of the central pair, but we could not identify all the doublets of the thinner bundles. The identification of microtubules in the thicker bundles based on the position of the 5-6 bridge and the orientation of the central pair would be appropriate, because the central pair of P. depressus flagella does not rotate with respect to the peripheral doublet microtubules (Takahashi et al., 1991). In order to identify the sliding pattern of thicker bundles, we photographed randomly selected cross sections of thicker bundles at a magnification of 50,000×. The photographs were labelled with code numbers and shuffled before the identification to avoid subjectivity in the judgement.
Results
Sliding disintegration of elastase-treated axonemes
ATP induced different patterns of sliding disintegration in the elastase-treated axonemes depending on its concentration. At a low concentration of ATP, such as 20 μM, the elastase-treated axonemes separated into a few groups of doublets or even into individual doublets as a result of more than two times (≥3) of sliding events, in which all sliding events were continuously induced (Fig. 1A and the left graph of Fig. 1C). (A sliding event is an ATP-induced, long-distance sliding that appears to occur between two adjacent doublets.) Similar sliding is observed in trypsin-treated axonemes(Summers and Gibbons, 1971),regardless of the ATP concentration. By contrast, at ATP concentrations higher than 0.1 mM, about half of the elastase-treated axonemes showed only one sliding events or two, so that the axonemes separated lengthwise into two or three bundles of doublets (Fig. 1B and the middle graph of Fig. 1C). The remaining half of the axonemes went through more than two sliding events, the first of which separated the axoneme into two bundles and the second and later sliding events occurred a few seconds after the first sliding. When the concentration of Ca2+ was increased to 10–4 M, the axonemes showing one or two sliding events at 1 mM ATP increased to more than 80%, dominated by those that were split into a pair of bundles as the result of only a single sliding event (right graph of Fig. 1C). These observations indicate a possibility that the activity of dynein arms is regulated by ATP and Ca2+.
The effect of Ca2+ on the sliding velocity observed in the elastase-treated axonemes that slid into paired bundles at 1 mM ATP (0.6 mM MgATP) is summarized in Fig. 2. The average sliding velocities were 9.8±2.8 μm second–1 (mean±s.d., n=45) and 9.2±2.5μm second–1 (n=53) at low(<10–9 M) (Fig. 2A) and high (10–4 M)(Fig. 2B) concentrations of Ca2+, respectively. The sliding velocities at low and high Ca2+ concentrations were not significantly different(P>0.2, Mann–Whitney U test). This seems to accord with previous reports that Ca2+ does not affect the sliding velocity in trypsin-treated axonemes(Walter and Satir, 1979; Mogami and Takahashi, 1983; Okagaki and Kamiya, 1986; Vale and Toyoshima, 1989). The following experiments, however, show that microtubule sliding in flagellar axonemes is regulated by Ca2+. They also give us a clue to understanding why the velocity of sliding disintegration in the elastase-treated axonemes is apparently not affected by Ca2+.
Splitting patterns of doublet bundles
When elastase-treated axonemes split into thinner and thicker bundles, the thickness of these bundles were clearly distinguished one from the other by dark-field microscopy, suggesting that the number of doublets in each of the two kinds of bundles are somewhat fixed. Electron microscopy of thin-sectioned bundles revealed that the thicker bundles consisted of five or six doublets and the central pair, whereas the thinner bundles consisted of four or fewer doublets and were without the central pair. Because the doublet microtubules do not rotate around the central pair in the species we used(Takahashi et al., 1991 for Pseudocentrotus; unpublished data for Clypeaster), we were able to identify each of the doublets in the thicker bundles from its position relative to the 5-6 bridge and the orientation of the central pair(Fig. 3A,B).
We recognized eight patterns of cross section among the bundles that contained the central pair. Of these, four patterns composed more than 90% of the bundles that were obtained at low concentrations of Ca2+(Fig. 3C). For simplicity, we will hereafter call, for example, the cross-sectional pattern of a thicker bundle that contains the doublets 8, 9, 1, 2, 3 and the central pair a `8-3 pattern'. According to this convention, the four main patterns of thicker bundles obtained at <10–9 M Ca2+ were 8-4, 8-3,4-8 and 3-8 (Fig. 3A,B). The thinner bundles corresponding to the 8-4 and 8-3 patterns were assumed to consist of doublets 5-7 and 4-7, respectively(Fig. 3B). Lacking the central pair, they could be identified by the presence of the 5-6 bridge. The thinner bundles corresponding to the 4-8 and 3-8 patterns were assumed to be doublets 9-3 and 9-2, respectively. There were also bundles of three or four doublets that contained neither the 5-6 bridge nor the central pair. In these bundles,we could not identify the individual doublets.
The splitting pattern was affected by Ca2+. As the concentration of Ca2+ increased, the frequency of 4-8 patterns gradually decreased so that at 10–6-10–5 M Ca2+, other three patterns (8-4, 8-3 and 3-8) were mainly observed. Similarly, at Ca2+ concentrations higher than 10–4M, the frequency of 3-8 decreased so that at 10–3 M Ca2+, we found mainly two patterns, 8-4 and 8-3. At Ca2+concentrations lower than 10–7 M, the frequencies of 8-4 and 8-3 combined were about the same as those of 4-8 and 3-8 put together but, at 10–3 M Ca2+, 8-4 plus 8-3 occurred more than seven times as frequently as did 4-8 plus 3-8(Fig. 3C).
To induce the four patterns of thicker bundles, the dynein arms on the doublets 7, 3 and 2 of the thinner bundles and/or those on the doublets 3, 4 and 8 of the thicker bundles should be active during the sliding separation of the axonemes. To determine the doublets whose dynein arms were active, we examined the sliding direction of the thinner bundles with respect to the thicker bundles along which they slid, using elastase-treated axonemes that still had the sperm heads attached to them(Fig. 4). Demembranated sperm with the heads were placed on the glass surface of the perfusion chamber. Under a dark-field microscope, we cut each flagellum carefully into two or three short lengths using a glass microneedle without changing the orientation of the fragmented flagellum with respect to the head(Fig. 4, upper panel). When reactivating solution containing elastase, 1 mM ATP, and either 10–3 M or <10–9 M Ca2+ was perfused, sliding disintegration of the fragments was induced as a bundle of doublets slid over the remaining part of the axoneme.
We found that, in most cases, the thinner bundles moved toward the head (or the thicker bundles moved away from the head)(Fig. 4, lower panel). Movements in this direction were observed in 84% of the axonemal fragments examined at <10–9 M Ca2+ and this rate increased to 92% at 10–3 M Ca2+(Table 1). Because the basal end of the axoneme is the `minus', or more stable, end of the microtubule, and dynein is a minus-end-directed motor, the above observation indicates that the dynein arms on the thinner bundles were mainly active to induce the separation of the axonemes, regardless of the concentration of Ca2+. We suggest that, at low concentrations of Ca2+, the dynein arms on doublets 7, 3 and 2 of the thinner bundles are mainly active to induce the 8-4, 8-3, 4-8 and 3-8 patterns, and that, at high concentration of Ca2+ (10–3-10–4 M), the dynein arms on doublet 7 of the thinner bundles are mainly active to induce the 8-4 and 8-3 patterns (Fig. 3). This implies that the activity of the dynein arms of the thicker bundles is inhibited.
Microtubule sliding on thinner and thicker bundles in an improved assay system
To understand the role of the central pair in the regulation of the dynein activity, we examined the effect of Ca2+ on microtubule sliding on bundles with or without the central pair. The thinner and thicker bundles were obtained from elastase-treated axonemes by application of 1 mM ATP in the presence of 10–4 M Ca2+. When singlet microtubules were added to the split bundles, we observed sliding of the singlet microtubules not only on the thinner bundles but also on the thicker bundles at <10–9 M-10–4 M Ca2+(Fig. 5). This is apparently inconsistent with the above suggestion that the dynein arms on the thinner bundles are mainly active when the axonemes slid into two bundles(Table 1). To see what exactly happens, we analysed the behaviour of microtubules on the bundles.
Before describing the behaviour of microtubules, some improvements in our assay method should be realized. In the previous study(Yoshimura and Shingyoji,1999), in which we developed the new sliding assay system,microtubule sliding was not always smooth. In the present study, we reduced the flow rate and kept a constant depth of the perfusion in order to maintain the activity of dynein arms that were exposed on the doublets. After these improvements, the behaviour of microtubules on the bundles changed. First, the number of bundles supporting active microtubule sliding increased (from 30% in the original method to >50% in the new method). Second, the back-and-forth movements of microtubules observed in the previous study were not observed in the present study, whereas a different type of back-and-forth movements was observed. Third, microtubules slid on thinner and thicker bundles at significantly different speeds in the previous study, but no such difference was found in the present study.
Frequency of microtubule sliding on thinner and thicker bundles
Fig. 6A summarizes the behaviour of microtubules attached on the thinner and thicker bundles. Because only one of the doublets of the bundles has dynein arms exposed(Fig. 3), orientations of doublet bundles on the glass surface are important to determine whether singlet microtubules are able to interact with the bundles. Thinner bundles showed straight and curved configuration, while thicker bundles were almost straight. The number of bundles to which microtubules attached and the number of microtubules attached to the bundles are shown in the right-hand column of Fig. 6A. The difference in configuration of bundles is probably related to the different numbers of thinner and thicker bundles to which microtubules attached. Comparison of the number of bundles with the number of microtubules, however, indicates that the thicker bundles had lower affinity to microtubules than the thinner bundles. The frequency of microtubule sliding was about 30-40% on the thinner bundles but was less than 20% (at <10–9-10–5 M Ca2+) and 9% (at 10–4 M Ca2+) on the thicker bundles. Fig. 6B shows ratio of the number of occurrences of microtubule sliding on the thinner bundles to that on the thicker bundles. At Ca2+ concentration≤10–5 M, the ratio was about 4 but, at 10–4 M Ca2+, it increased to 9.6, indicating that the sliding on the thicker bundles occurs one tenth as frequently as that on the thinner bundles in the presence of 10–4 M Ca2+.
In the present improved sliding assay system, the microtubule sliding was smooth on both the thinner and the thicker bundles. But a back-and-forth movement with a distance of about 0.5-1 μm was observed only on curved regions along the thinner bundles at fewer than 10% of the microtubules attached to the bundles (hatched boxes in Fig. 6A). We need more detail analysis of the back-and-forth movement to know whether it is related to the regulation of dynein activity.
Velocity of microtubule sliding on thinner and thicker bundles
Most of the microtubules moved smoothly throughout the sliding, and this smooth movement was not changed by the presence of Ca2+(Fig. 7). The time course of microtubule sliding on the thinner bundles and that on the thicker bundles was similar in the present assay system (Fig. 7A). An increase of Ca2+ concentration, however,affected the velocity of microtubule sliding on the thicker bundles(Fig. 7B).
Fig. 8 summarizes the velocity of microtubule sliding at 1 mM ATP (0.9 mM MgATP) on the thinner(left panels) and the thicker (right panels) bundles at lower (A) and higher(B) concentrations of Ca2+. The sliding velocities did not show a Gaussian distribution but showed a broad distribution with a peak at around the middle and a longer tail towards higher sliding velocities(Fig. 8). At<10–9 M Ca2+, the distribution of sliding velocity was similar on the thinner and on the thicker bundles(Fig. 8A). At 10–7-10–4 M Ca2+, the distribution of the sliding velocities on the thinner bundles was also similar to that at low concentration of Ca2+, whereas that on the thicker bundles shifted towards the lower velocities. Statistical differences between the sliding velocities on the thinner and the thicker bundles at<10–9 and at 10–7-10–4 M Ca2+ were examined by using the Mann–Whitney U test. We found that the sliding velocity on the thinner bundles and that on the thicker bundles were not different at low concentrations of Ca2+but were significantly different in the presence of Ca2+(P<0.01). At any Ca2+ concentration of 10–7-10–4 M, the sliding velocity on the thicker bundles was lower than that on the thinner bundles. This showed clearly that the sliding velocity decreases with Ca2+ and the regulation of dynein activity by Ca2+ requires the presence of the central pair microtubules.
Discussion
The present study has shown that Ca2+ alters the dynein activity in the flagellar axoneme through a regulatory mechanism that involves the central-pair microtubules. We used a novel assay system that enabled us to analyse the sliding movements of a microtubule on a bundle of doublet microtubules with or without the associated central pair. The doublet bundles were obtained by ATP-induced splitting of elastase-treated axonemes. We found that Ca2+ significantly decreased the velocity and the frequency of microtubule sliding on the thicker doublet bundles, which contained the central pair. Because the microtubule sliding on the thinner bundles without the central pair did not change by Ca2+, we conclude that the regulation of dynein activity by Ca2+ depends directly on the central-pair microtubules. Thus, the elastase-treated axonemes allowed us to understand the activity of dynein arms under the regulatory mechanism involving the central pair (Shingyoji and Takahashi, 1995; Yoshimura and Shingyoji, 1999).
The Ca2+-induced changes in the microtubule sliding velocity had not been demonstrated until our previous work using beating flagella(Bannai et al., 2000). In earlier studies, Ca2+ was shown to affect neither the sliding disintegration of axonemes treated with trypsin (∼1 μg ml–1) (Walter and Satir,1979; Mogami and Takahashi,1983; Okagaki and Kamiya,1986) nor the gliding of microtubules on isolated axonemal dynein(Vale and Toyoshima, 1989). More recently, we found that trypsin digested some axonemal structures responsible for cyclical bending whereas elastase did not(Shingyoji and Takahashi,1995). The present study using elastase-treated axonemes has demonstrated that the regulatory mechanism, which is necessary for cyclical bending and is retained in the elastase-treated axonemes, is also essential for the Ca2+ regulation of the microtubule-sliding activity of dynein. This regulatory mechanism requires the presence of the central-pair microtubules and the functional dynein arms that are attached to their natural location on the doublet. These results are consistent with the previous studies suggesting involvement of the central-pair microtubules in the Ca2+ regulation of flagellar bending movement(Sale, 1986; Bannai et al., 2000).
The present study has shown that Ca2+ causes three major changes in the dynein activity: (1) a decrease in the velocity of microtubule sliding on thicker bundles; (2) a decrease in the frequency of microtubule sliding on thicker bundles; and (3) changes in the splitting patterns of axonemes, which seem to correspond to the sites of active sliding among the axonemal microtubules. These changes were all closely associated with the presence of the central-pair microtubules but were induced at different concentrations of Ca2+. This suggests that the three kinds of response to Ca2+ might be regulated by different mechanisms.
Our previous study has shown that trifluoperazine inhibits the `quiescence'of sperm flagella, implying that Ca2+-calmodulin is involved in the regulation of sliding that leads to the quiescence at 10–4 M Ca2+ (Bannai et al.,2000). The present study suggests that the quiescence is caused not only by the decrease in sliding frequency but also by the decrease in sliding velocity of dynein arms at one side (probably on doublet 3) of the central-pair microtubules. Smith (Smith,2002b) has recently shown in Chlamydomonas flagella that the microtubule sliding velocity of axonemes lacking the central pair is reduced at <10–8 M Ca2+ compared with that of the wild-type axonemes, but is restored by 10–4 M Ca2+. She has also shown that calmodulin is involved in this Ca2+ regulation of microtubule sliding. This suggests a similar regulation of flagellar motility in both the sea urchin sperm and the Chlamydomonas flagella.
The patterns of splitting of the elastase-treated axonemes into two doublet bundles with or without the central pair are probably related to the mechanism regulating the dynein activity. We have found four main patterns of splitting,and the frequency of their appearance depended on the Ca2+concentration. The four patterns were the thicker bundles consisting of the 8-4, 8-3, 4-8 and 3-8 doublet groups, with their corresponding thinner bundles of the remaining four or three doublets. Spontaneous axonemal fracture into a thinner and a thicker bundle has previously been reported in the quiescent sea urchin sperm flagella at high Ca2+ concentrations, in which the 8-3 pattern of the thicker bundles was dominant(Sale, 1986). Holwill and Satir (Holwill and Satir,1994) have developed a physical model of axonemal splitting based on the observation of splitting of the axonemes, including six-and-three doublet groups and five-and-four doublet groups. This model provides a physical basis for the axonemal splitting that was observed in the present study.
Our analysis of the splitting patterns provided some interesting findings. The dynein arms on the thinner bundles, but not those on the thicker bundles,were active to produce microtubule sliding and to split the axonemes into paired bundles. Furthermore, we found that the occurrence of the 8-4 and 8-3 patterns and of the 4-8 and 3-8 patterns were similar at lower concentrations of Ca2+, whereas the 8-4 and 8-3 patterns increased and the 4-8 and 3-8 patterns decreased with an increase of the Ca2+ concentration. These changes in the frequency of occurrence might be related to the waveforms of reactivated flagella: the flagellar beating is nearly symmetrical at lower concentrations of Ca2+, becoming increasingly asymmetrical with an increase of Ca2+ and finally becomes quiescent with a large principal bend at the base of the flagella at ≥10–4 M Ca2+. The increase in asymmetry and the quiescence are induced by the inhibition of reverse-bend formation(Brokaw, 1979; Gibbons and Gibbons, 1980; Bannai et al., 2000). Thus, it is likely that the occurrence of the 4-8 and 3-8 patterns, which decreases with Ca2+ concentration, is related to the reverse-bend formation. If this is the case, switching between the group of the 8-4 and 8-3 patterns and the group of the 4-8 and 3-8 patterns – in other words, switching of the activity of dynein arms between doublet 7 (of thinner bundles) and doublet 3 or 2 (of the thinner bundles) – would be the basis for the alternate formation of the principal and the reverse bends.
The mechanism regulating the dynein activity to split the axoneme into a thinner and a thicker bundle has been unclear. From the present result, we can speculate about the roles of the two microtubules of the central pair. Structural and biochemical differences between the C1 and C2 microtubules with their associated projections have been described in Chlamydomonasflagella, suggesting possible different roles of C1 and C2 in the regulatory mechanism (Smith and Lefebvre,1997; Mitchell and Sale,1999; Porter and Sale,2000; Smith,2002a). In sea urchin sperm flagella, two microtubules of the central pair and their associated projections look similar in electron micrographs, and their biochemical and physiological differences have not been well documented. If we assume that the microtubules that are near doublets 7 and 3 are C1 and C2, respectively, similar to those of Chlamydomonas,the alternate activation of dynein might be regulated through the radial spokes by some signal from C1 and C2. The signal might be associated with a protein phosphorylation (Yang et al.,2000; Roush-Gaillard et al.,2001), with Ca2+-calmodulin(Bannai et al., 2000; Smith, 2002b) or with changes in the mechanical states (Omoto et al.,1999; Bannai et al.,2000).
The elastase-treated axonemes showed different sliding features with varying ATP concentration. Low concentrations of ATP induced sliding disintegration of the axonemes into individual doublets, indicating that the dynein arms on any doublet of the axonemes are active. By contrast, high concentrations of ATP induced splitting of the axonemes into two or three bundles, indicating that the dynein arms attached to the doublets (3 or 7)near the central pair are active but those on the remaining doublets are inactive. These observations show that the dynein activity is inhibited at high concentrations of ATP but that the inhibition is overridden by the CP/RS. The inhibitory effect of a physiological concentration of ATP on the number of doublets that slide has been demonstrated in Tetrahymena cilia by Kinoshita et al. (Kinoshita et al.,1995), although, in their study, the sliding was also inhibited at low concentration of ATP, which is different from the present finding. Similar inhibition of the outer arms by a high level of ATP has also been reported in CP/RS-deficient Chlamydomonas mutants, in which lower concentrations of ATP induce beating of paralysed flagella(Omoto et al., 1996). The CP/RS complex is thought to override the ATP inhibition by activating the inner arms in a coordinated fashion in Chlamydomonas(Porter and Sale, 2000). The roles of ATP concentrations in the mechanism regulating the dynein activity might be similar in cilia and flagella of different species and the CP/RS might release the inhibitory effect of high ATP concentrations(Bannai et al., 2000).
The role of the CP/RS in the Ca2+ regulation of dynein activity also depends on the ATP concentration. At low concentrations of ATP, the axonemes of the CP/RS-deficient Chlamydomonas mutants are able to beat and show Ca2+-dependent waveform conversion(Wakabayashi et al., 1997). At a high (physiological) level of ATP, however, the CP/RS is suggested to play a key role in the Ca2+ regulation of waveform and microtubule sliding(Hosokawa and Miki-Noumura,1987; Smith,2002b). The present result in sea urchin sperm flagella show that the CP/RS is essential for the Ca2+ regulation of dynein activity at the physiological level of ATP.
We thus conclude that the outer and probably also the inner dynein arms are inhibited by physiological levels of ATP and that the CP/RS complex might override the inhibition by activating both the outer and inner dynein arms. When the CP/RS complex is functional, both the C1 and C2 microtubules might alternately activate the dynein arms on the near microtubules. We also postulate that mechanism regulating the activity of dynein by C2 is sensitive to Ca2+.
Finally, we discuss the velocity of sliding of the elastase-treated axonemes to induce thicker and thinner bundles. The velocity of sliding disintegration was not affected by Ca2+ even in the elastase-treated axonemes. This is apparently inconsistent with the Ca2+-induced inhibition of dynein activity on the thicker bundles that was shown in the novel sliding assay. This inconsistency could be explained, however, if we take all results of the present study into consideration. When the elastase-treated axonemes split lengthwise into two bundles, the dynein arms on the thinner bundles are predominantly active at low as well as at high concentrations of Ca2+. This means that the velocity of sliding induced by dynein arms on the thicker bundles could not be measured in the sliding disintegration. Our study also showed that the velocity and the frequency of sliding of singlet microtubules that interacted with the dynein arms exposed on the thicker bundles, but not with those on the thinner bundles, were significantly inhibited by Ca2+. Ca2+ inhibited the activity of the dynein arms on the thicker bundles but not that on the thinner bundles, and the activity of dynein arms on the thinner bundles was required for the sliding disintegration. Thus, it is very plausible that the inhibitory effect of Ca2+ could not be detected in the sliding disintegration of the axonemes.
Acknowledgements
We would like to express our gratitude to Professor Keiichi Takahashi for his kind guidance, encouragement throughout the study and reading the manuscript. We thank the director and staff of Misaki Marine Biological Station (University of Tokyo) for providing sea urchins. This work was supported by grants from Takeda Science Foundation, Hayashi Memorial Foundation for Female Natural Scientists and Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (to C.S.).