The typical structure of the eukaryotic flagellum consists of a central pair of singlet microtubules surrounded by nine doublet microtubules, called the axoneme. Much has been discovered regarding the mechanism by which axonemes produce motion: ATP is used by dynein arms found on the A tubules of the doublet microtubules to produce shear force against the B tubules. These shear forces are then converted to bending. However, if all the dynein arms along the length of the axoneme and on all doublets attempted to produce shear simultaneously, no effective movement would result. Thus, regulation of active shear force is required. Evidence suggests that the central pair–radial spoke complex is involved in this regulation. The first evidence came from an electron micrograph study in which the central pair microtubules of Paramecium, “instantaneously fixed” and serially sectioned, appeared to be oriented in systematically changing angles. This was interpreted as rotation of the central pair with respect to the nine outer doublets per beat cycle (Omoto and Kung, 1979 , 1980 ). It was suggested that the central pair may act as a “distributor” to regulate the activity of dyneins.
Rotation of the Central Pair Microtubules of Micromonas pusilla
To test the hypothesis that the central pair microtubules rotate, an organism with the central pair extending well beyond the 9 + 2 region was used to try to directly visualize movement of the central pair. Such an organism is the small marine alga, M. pusilla. The central pair of the single flagellum of this alga extends 4–5 μm beyond an extremely short (<1 μm) 9 + 2 region. The central pair of this flagellum is similar to that of other cilia and flagella in that it is helical (Omoto and Witman, 1981 ). Movie 1 shows M. pusilla swimming with what appears to be the helical central pair rotating and pushing the cell (Figure (Figure1A).1A). The basal 9 + 2 region is also beating; however, it is obscured by the glare from the cell body in these dark-field images. Note that the videos were made from negatives of original dark-field cinemicrographs, so the cell body and flagella appear dark against a light background. An appearance of rotation can result from propagation of a helical wave along a nonrotating central pair. To distinguish between this and true central pair rotation, an experiment analogous to that used to demonstrate rotation of the bacterial flagellum (Silverman and Simon, 1974 ) was used. If the movement is true rotation rather than helical wave propagation, then a cell attached to the slide by its flagellum should rotate. Movie 2 shows such an experiment (Figure (Figure1B);1B); the flagellum is clearly visible and unmoving, and the cell body rotates. These images clearly and directly demonstrate that the central pair of microtubules of the 9 + 2 flagellum of M. pusilla rotate. The direction of central pair rotation in M. pusilla is clockwise as viewed from base to tip. This is the same as that for central pair rotation in Paramecium inferred from the electron microscopic observations.
Extrusion and Rotation of the Central Pair of Chlamydomonas reinhardtii
When Chlamydomonas cell models are kept in the presence of 1 mM ATP, they can beat for >30 min. With time, however, the axonemes tend to disintegrate, frequently accompanied by extrusion of the central pair of microtubules (Kamiya, 1982 ). Movie 3 shows a demembranated cell model with one of the two flagella extruding the central pair and rotation of that central pair (Figure (Figure2).2). This phenomenon is facilitated by a mild elastase treatment of axonemes to the point that >90% of the axonemes extrude the central pair (Hosokawa and Miki-Noumura, 1987 ). This central pair extrusion and rotation can be seen in isolated axonemes (Movies 4 and 5 [Figure 3]). The helical twist of the central pair and the direction of rotation correspond to propagation of helical waves distally. Although these videos show the movement after much of the central pair has extruded, upon initial extrusion of the central pair, the bend amplitude greatly decreases. This suggests that the mechanism that causes the central pair extrusion and rotation may be coupled to the mechanism of bend formation and propagation. The clockwise rotation, as viewed from the base to the tip, and the left-handed helix of the central pair are the same as those in M. pusilla.
Although it remains to be determined whether the central pair rotates in intact Chlamydomonas axonemes, electron microscopic observations of central pair orientation in the basal portion of the two axonemes suggest that the plane of the central pair is not fixed within the cylinder of nine outer doublets (Kamiya et al., 1982 ). Therefore, the central pair may rotate in beating Chlamydomonas flagella in vivo. Such a rotation may facilitate propagation of bending waves in the axoneme.
Mechanism and Force for Central Pair Rotation
What drives the central pair rotation? Central pair rotation does not require the basal body, because central tubules can rotate and extrude out from the distal end of detached flagella, as seen above. By the same argument, it seems that the force for rotation cannot be localized at the base. There are then two general possibilities left for what drives the rotation. One is that some enzymes, possibly kinesins, which have recently been found to be associated with the central pair (Bernstein et al., 1994 ; Fox et al., 1994 ), are actively rotating the central microtubules. Alternatively, the helical central pair is passively rotated by the bending of the axoneme. The central pair free in solution takes on a helical conformation (Kamiya et al., 1982 ). When such a helical shape is confined within the cylinder of the nine doublets, the helical shape may orient itself to conform to the bend. When the bend propagates distally, the helical shape will rotate to place the helix in the lowest energy position to conform to the bend. This type of mechanism is consistent with the left-handed helix of the central pair and the clockwise rotation of the central pair as viewed from the base. It is also consistent with the following observation in experiments manipulating the beat plane of sea urchin sperm.
Rotation of Plane of Bend of Sea Urchin Sperm
Sea urchin sperm flagella normally beat in a plane (Gray, 1955 ). However, this beat plane can be manipulated by holding the sperm head in a micropipette and vibrating it in a plane (Gibbons et al., 1987 ; Shingyoji et al., 1991 ; Takahashi et al., 1991 ). When the plane of imposed vibration was gradually rotated, the flagellar bend plane rotated along with it, up to 11 complete revolutions. The surprising observation shown in Movie 6 is that when that imposed vibration was stopped, the sperm flagellum spontaneously “unwound” its bend plane back to the original orientation (Figure (Figure4).4). By imposing rotation on an axoneme that had a marker bead stuck to the outer doublet microtubules on one side, it was possible to show that the imposed rotation of the bend plane involves a rotation in the coordinated pattern of sliding between the microtubules, rather than a twisting of the whole flagellar structure (Figure (Figure5).5). It is hypothesized that the rotating pattern of sliding or the resultant bending forces the rotation of the central pair microtubules. This rotation would cause a twisting of the central pair if the basal end is firmly anchored. When the imposed vibration is stopped, the central pair presumably untwists elastically back to its original orientation, in the process regulating the pattern of sliding of the outer doublets which is manifest as rotation of the bending plane. In organisms with flagella that beat only in one plane, such as these sea urchin sperm, the central pair may be firmly anchored at its basal end. Only with such a firm anchoring can we explain the twisting and untwisting observed here.
Function of the Central Pair
What might be the function of central pair? It is clear that the central pair is not needed for bending per se, because there are naturally occurring motile flagella that lack the central microtubules (Schrevel and Besse, 1975 ; Prensier et al., 1980 ; Goldstein and Schrevel, 1982 ; Gibbons et al., 1985 ; Ishijima et al., 1988 ). Although the central pair–deficient mutants of Chlamydomonas are paralyzed, they can move in the presence of extragenic suppressor mutations (Huang et al., 1982 ; Brokaw and Luck, 1985 ), under altered nucleotide conditions in reactivation (Omoto et al., 1996 ; Frey et al., 1997 ), or under mechanical stimulation (Hayashibe et al., 1997 ). Yet a great majority of axonemal structures possess the central pair, and mutants defective in them are paralyzed under physiological conditions (Witman et al., 1978 ; Afzelius, 1985 ). Thus we propose that the nine outer doublets exhibit a default movement in the absence of central pair–radial spoke complex. The presence and activity of the central pair and radial spokes imposes a higher-order regulation on this default movement to enable a more complex three-dimensional waveform or ciliary-type waveform.