The undulating beat of eukaryotic flagella and cilia produces forces that propel cells (or the constituents of their environment) in creatures ranging from algae to humans. The beats arise from bends that travel from a flagellum's base to its tip — by repetitively generating these bends, flagella assume waveforms appropriate for their specialized functions. The timing mechanisms that coordinate these beats remain mysterious, but models indicate that the underlying metronome requires a complicated structure called the axoneme1,2,3, at the core of the flagellum. On page 711of this issue, however, Shingyoji et al.4 report that oscillatory movements can be generated by surprisingly simple sub-structures of the axoneme.
The axoneme is built from about 250 different polypeptide components, including nine long, thin doublet microtubules, arranged parallel to one another in a cylindrical array. Each doublet consists of an A-microtubule that is bound lengthwise to an incomplete B-microtubule (Fig. 1). Molecules of dynein, an ATP-fuelled motor protein, are arranged along the A-microtubule in two parallel rows that project towards an adjacent doublet. The doublets are locked together at the base of the axoneme so, when dynein generates shearing forces between them, they bend rather than slide apart.
Shingyoji et al.4 have used a focused laser beam (optical tweezers) as a micromanipulator and force transducer. They find that the direction of the movement generated by a small number of dynein molecules (perhaps even one) on an isolated doublet microtubule prepared from sea-urchin flagella oscillates when the dynein molecules work against an elastic load. The frequency of the oscillations is similar to the beat of an intact flagellum, in contrast to the faster, smaller oscillations observed in intact non-beating axonemes5.
Understanding the geometry of the assay is crucial for interpreting these results. Shingyoji et al. used optical tweezers to manipulate a bead bound to a microtubule, so that this microtubule was positioned across a doublet microtubule at an obtuse angle (see Fig. 1a on page 711). Despite this unusual non-parallel orientation, the dyneins on the doublet were compliant enough to move the microtubule, displacing the bead within the optical tweezers. The effect of the optical tweezers was similar to that of attaching one end of the microtubule to a spring — the restoring force increased with the displacement.
How might these oscillations arise? If more than one molecule of dynein is involved, oscillations could result from variations in the number of motors that interact with the microtubule, as suggested in Fig. 1 . As the motors displace the bead-bound microtubule, the restoring force from the optical tweezers will impart a torque on the doublet microtubule about its attachment to the glass slide. Eventually, rotation of the doublet could cause a subset of the dynein motors to be pulled away from the microtubule. The remaining motors, unable to sustain their increased load, would reverse direction and ‘walk’ backwards along the microtubule until enough strain is released to allow the lost motors to re-engage — then, forward movement would resume.
This is plausible. Shingyoji et al. have found a decrease of speed in both directions with diminishing concentrations of ATP, indicating that dynein molecules can indeed walk backwards. This differs from the saw-tooth movement traces that are seen when an elastic load is placed on single molecules of kinesin, a motor that detaches from microtubules when overloaded6,7. Chlamydomonas mutants that lack subgroups of dynein arms8 would be well suited for testing this model and for sorting out the biophysical contributions of different dyneins within an axoneme.
The most intriguing suggestion made by Shingyoji et al. is that the oscillations might be generated by a single molecule of dynein. If correct, this implies that a doublet-bound dynein molecule can retain a ‘memory’ of previous events, because it varies in a time-dependent manner in its response to an opposing force. Considering the structure of dynein, this is not a trivial task. Sea-urchin flagellar dyneins are large, multi-subunit molecules that consist of a stalk and two globular heads. The heads contain sites for microtubule binding and ATP hydrolysis. Because the heads are only about 13 nm in diameter9,10, dynein probably takes more than one ATP-hydrolysing ‘step’ to span 32 nm, which is the average amplitude of the oscillations. So it is difficult to imagine how the oscillations might be simply associated with the different chemical states of dynein's chemo-mechanical ATP hydrolysis cycle.
Backward movements of a single dynein might persist until some unspecified strain is released from the dynein molecule. However, a simple elastic relaxation cannot adequately describe the data. Any strain that is large enough to be distinguished from random thermal fluctuations would be dissipated in microseconds at most, yet the period of the oscillations was many milliseconds. But perhaps there is a way to slow dynein's relaxation. For example, at peak strain, rotation of the doublet might pull a single dynein head away from the microtubule, leaving the remaining head to walk backwards under its increased load. The weakness of this theory is that it requires a single dynein head to maintain its association with a microtubule that is under tension, while simultaneously shifting between binding sites on the microtubule to allow movement. Even kinesin, single molecules of which have been shown to be functional, cannot do this without using both of its heads11.
Are there alternatives to these purely mechanical explanations? Although the biochemical milieu of the cell has been removed, these oscillations may nonetheless depend on biochemical modifications. A dynein head has a total of four ATP binding sites12, but only one of these shows an appreciable rate of hydrolysis. So perhaps strain-dependent changes in the ATP affinity or rates of hydrolysis at the other sites could modulate force generation by dynein.
These new results are an exciting twist in the hunt for the origin of the flagellar beat. Even more exciting is the possibility that the biophysical techniques used will be yet more illuminating when combined with established methods in biochemistry and genetics.
Brokaw, C. Math. Biosci. 90, 247–263 (1987).
Murase, M., Hines, M. & Blum, J. J. Theor. Biol. 139, 413–430 (1989).
Lindemann, C. Cell Motil. Cytoskel. 29, 141–154 (1994).
Shingyoji, C., Higuchi, H., Yoshimura, M., Katayama, E. & Yanagida, T. Nature 393, 711–714 (1998).
Kamimura, S. & Kamiya, R. Nature 340, 476–478 (1989).
Svoboda, K., Schmidt, C., Schnapp, B. & Block, S. Nature 365, 721–727 (1993).
Meyhofer, E. & Howard, J. Proc. Natl Acad. Sci. USA 92, 574–578 (1995).
Dutcher, S. Trends Genet. 11, 398–404 (1995).
Goodenough, U. & Heuser, J. J. Mol. Biol. 180, 1083–1118 (1984).
Gee, M., Heuser, J. & Vallee, R. Nature 390, 636–639 (1997).
Hancock, W. & Howard, J. J. Cell Biol. 140, 1395–1405 (1998).
Mocz, G. & Gibbons, I. Biochemistry 35, 9204–9211 (1996).
About this article
Physical Review E (2016)
Geophysical Journal International (2012)
Global and Planetary Change (2009)
Annals of Glaciology (2008)
Biophysical Chemistry (2008)