This chapter describes the popular idea that the cytoskeleton of a cell is being continuously assembled at the front of the cell and disassembled at the back. This is believed to lead to a flow of cytoskeleton just beneath the cell membrane which, so it is said, leads to capping and motility.
6.1 Cytoskeletal Flow Model
6.2 Why this Name?
6.3 Criticisms of the Cytoskeletal Flow Model
6.4 The Cytoskeletal Flow Model and Motility
6.5 My View of the Cytoskeletal Flow Model
6.6 The Cytoskeletal Paradigm is Vacuous
6.7 Why is the Cytoskeletal Flow Model so Popular?
I once took my son to the Pleasurewood Hills Theme Park, full of rides for children. He went on all of them, while his terrified father, looked on from a distance. One of the few rides I dared, accompanied by my scornful 9 year old, was a small boat ride. Here the boats, with passengers on board, gently floated round a circular course before returning to the station to unload one set of passengers and load another. The boats had flat bottoms and did not float through the station, they were guided through by a conveyor belt located just beneath the surface of the water. This way the boat could be stopped and started by an operator, while passengers such as myself, still nervous after their experiences elsewhere in the park, were not subject to any unseemly rocking motions as they embarked or disembarked.
The cytoskeletal flow model proposes this conveyor belt mechanism for driving patches and particles on the cell surface (see Fig. 6.1). Parts of the cytoskeleton tend to occur just beneath the membrane. The model proposes that these structures are the conveyor belt driving the motion of patches or particles. To move this way the cytoskeleton must be assembled at the front of the cell and broken down at the rear, the assembly and disassembly process giving rise to, and powering, continuous motion. As pieces of cytoskeleton are disassembled at the rear of the cell, they must be rapidly transported through the cytoplasm to the front for reassembly into cytoskeleton. The whole filament, indeed the whole cytoskeleton, is bodily moved along by the impetus given by this turnover.
Advocates of this mechanism point out the properties of actin, the protein forming the cytoskeletal backbone. Actin has the property of existing in two forms - one form in solution with separate molecules the filamentous. The joining into fibres and separation of molecules are known as polymerisation and depolymerisation. The model suggests actin fibres are continuously polymerising at the front of the cell and depolymerising at the rear. The dissolved actin is transported through the cell. The cytoskeleton is proposed to circulate beneath the membrane (see for example Kucik, Elson & Sheetz (1990)).
There is much of the membrane flow model in the idea. Like that model, a conveyor belt mechanism is being proposed but the conveyor belt lies just beneath the membrane rather than being the membrane itself. The link with motility implies that all amoeboid cells move this way. The circulating cytoskeleton is said to carry trans-membrane proteins with it so they too continuously circulate and sieve the surface of the cell. Their sieving action catches large particles or patches while small particles pass through the mesh and remain unmoved. At present, most major researchers seem to believe this concept.
6.2 Why this Name?
Conceptually, the idea is closer to the membrane flow model than to the original cytoskeletal model. The interior of the cell does not recognise or respond to external objects, instead they are moved by an ongoing cellular process. The original cytoskeletal model was a recognition-response model but this is an entrainment model. The two cytoskeletal models are distinct from one another and mutually incompatible. The idea of actin fibres flowing is quite at odds with the implications of planned location of fibres contained within the original cytoskeletal model.
In the literature the two ideas go by the same name, the cytoskeletal model. Their name is virtually the only thing they have in common. The shared name is odd and the source of much confusion. Workers can jump from one model to the other, without realising the incompatibility of the ideas, workers supporting the cytoskeletal flow model do so without realizing that they are rejecting the original ideas. This book gives different names to the different ideas but the name "cytoskeletal flow model" is not used in the scientific literature only here. It aptly describes what is being suggested and how the ideas arose.
Much of the underlying rationale for the cytoskeletal flow model comes from the membrane flow model with the ideas transferred to the cytoskeleton. In substance what is proposed is a flow model, that the cell's muscle, its cytoskeleton, can flow. The cell's muscle proteins are said to generate force not by contraction, as they do in muscle, but by forming a conveyor belt. In two different situations, the same proteins are said to generate force by two entirely separate mechanisms; in muscle they are contractile, in the cell they are flow generating.
The conversion of the cytoskeletal model to an entrainment model moves it closer to reality but in other respects the idea is not sensible.
- It suggests transmembrane proteins will cap without first being patched. This is not generally the case.
- Just as the membrane flow model needed a mechanism for transporting vesicles through the cell cytoplasm, so the cytoskeletal flow model requires some mechanism for the transport of the solubilised actin. How this could happen is unclear.
- The model predicts that cell surface lipids and non- transmembrane proteins, those components not passing right across the membrane, can cap only by being dragged along by other membrane components. They should cap only by co-capping with the transmembrane proteins that do contact the flowing cytoskeleton. This seems a testable prediction but no report supporting it is known to the author.
- The observation widely seen as a "proof" of the original cytoskeletal model was that actin and myosin accumulate beneath patches and caps. Such accumulations play no obvious part in cytoskeletal flow. This is ironic, since most workers seem to think both that the cytoskeletal flow model is the cytoskeletal model and that the accumulations prove the cytoskeletal model.
- The model suggests the membrane itself should flow. If trans-membrane proteins are continuously dragged through the membrane, they might be expected to act like oars in water dragging the membrane along and providing a motor for bulk membrane flow. However, membrane flow does not occur.
- The idea induces the same sense of disbelief as did the lipid flow hypothesis. It seems unworldly, unrelated to previous knowledge or experiment. Nowhere have motile mechanisms remotely analogous to tank tracks or conveyor belts evolved. The idea, like its parents, breaches the evolvability criterion. (By contrast, more creatures move around on waves than on legs.)
Cytoskeletal flow needs direct positive supportive evidence for its metaphysical elements. There seems to be no direct evidence for the existence of sub-membrane cycling at the rates required. Where measurements of the rate of cytoskeletal movements have been made, they are too slow to account for capping and particle movement. For example, Theriot and Mitchison (1991) measured the rate of movement of the cytoskeleton as the cell moved. They found the cytoskeleton moved backward with respect to the cell but remained stationary with respect to the substrate the cell moved over.
However, particles move backwards much faster than this, typically twice as fast, but in some systems particles can move at 50-100 times the rate of the cell. Later the same authors (Theriot and Mitchison (1992)) directly measured the rate of particle migration and actin turnover on the same cells (fibroblasts) and found actin turnover to be three times slower than particle movement. They observe in their conclusion that "the rearward transport .... is not, at least in fibroblasts, a direct reflection of the movement of the bulk actin meshwork inside the lamellipodium." They remark that these transports must occur either by independent mechanisms or be only weakly coupled to one another. These results indicate that cytoskeletal turnover is not nearly fast enough to drive capping and particle movement.
The cytoskeletal flow model also fails as an attempt to relate capping and particle movement to motility. If motility is the function related to particle movement, then the model suggests the cell moves by riding on its transmembrane proteins. However, the cell moves more slowly than particles on its surface.
In addition the model suggests that actin and myosin generate the force needed to move the cell by assembly and disassembly, quite differently from the way these proteins generate force in muscle where they are contractile proteins. That is, as muscle filaments slide over one another they make the muscle shorten, generating a tension force in doing so. This is a breach of the evolvability criterion which requires that actin and myosin should function in similar ways in the two situations.
This is a good point to rebut an argument often found in the scientific literature. Some readers may prefer to pass on to the next section. There have been many studies of actin filament assembly, both in living cells and, more often, in vitro. These studies show clearly that when actin filaments assemble from soluble actin they do so unidirectionally. This much is fact but that fact is not grounds for believing any physiologically relevant force is generated by this unidirectional assembly. After all, how a car engine is built tells us little about how it works once assembled. The forces assembling a car engine are not the forces that drive the car forward, it is the operation of the assembled engine does that. In muscle, actin filaments assemble unidirectionally but having assembled, they work by sliding filament contraction with myosin. In other words assembly and operation once assembled are quite separate, just as they are in a car engine.
The cytoskeletal flow model suggests the actin/myosin molecules can work in two quite different ways to produce force. In muscle these filaments are contractile but in the cell they form a conveyor belt. The same proteins could no more function in two ways like this than could an internal combustion engine from a car function as a jet if installed in an aeroplane! Once it is assembled, a machine functions in one way. If a different function is desired, a different machine must be built. Because the actin filaments of the cell are very similar to those in muscle they would be expected to function in similar ways, by contraction. If a conveyor belt is desired, other proteins should be found to make it.
The cloth they wove (so they said) wasn't only exceptionally beautiful but had magical properties; even when made into clothes it was invisible to anyone who was either unfit for his job or particularly stupid. (Hans Christian Anderson)
The cytoskeletal flow model is ad hoc. It evolved from the original cytoskeletal model, by axiomatic drift, as the advocates of this notion attempted to avoid its problems. There is no direct evidence to support any of its major contentions and attempts to obtain evidence fail. The model is also cyclic in its logic; the only evidence for cytoskeletal flow is capping and particle movement which is the only thing cytoskeletal flow explains. In other words, cytoskeletal flow is not founded upon or derived from existing knowledge or experiment, it is a "green cheese" hypothesis.
Neither is it really sensible to assert that the cytoskeleton, inside the cell is moving while its membrane, on the outside, is stationary. A car designer would find no advantage in arranging for the drive wheels to rotate while the tyres remained stationary. Even so, that is the essence of the cytoskeletal flow model.
This section examines further the relationship between the original cytoskeletal model and the cytoskeletal flow model. The aim is to examine the informative value of the statement "capping is caused by the cytoskeleton," a claim summarising what will be called the cytoskeletal paradigm. The paradigm makes no further statement about what the cytoskeleton does to cause capping.
This section will careful about terminology. The model described in Chapter 4 is the "original cytoskeletal model". The term "cytoskeletal paradigm" describes an overarching model incorporating both the original cytoskeletal model and the cytoskeletal flow model. These two ideas are submodels to the cytoskeletal paradigm but it will also be shown that the membrane flow model and the wave model (of Chapter 7) are also submodels, as is virtually any other possible mechanism.
The cytoskeletal paradigm is quite vacuous as can be demonstrated by presenting a hierarchical series of "theories" of capping.
- Capping is caused by the cytoskeleton. (The cytoskeletal paradigm).
- Capping is caused by the cell.
- Capping is caused by evolution.
- Capping is caused by God.
Statements 2 and 3 are, by conventional science, manifestly true, while, depending upon religious standpoints, statement 4 may well be. However, from a scientific perspective, all the statements 2, 3 & 4 have no value. Their very generality means they incorporate all possible experimental outcomes and so cannot predict, or be falsified by, the outcome of any experiment. This was the criterion Popper used to distinguish a scientific from a non-scientific hypothesis. Statements 2, 3 and 4 are not scientific theories with respect to capping.
Statement 1 is hardly more scientific. It has some predictive merit in excluding the idea that other organelles, for example the mitochondria or the nucleus are moving particles and patches. However, nobody has advanced the idea that other organelles are involved and the cytoskeleton is simply an organelle, not a mechanism or a dynamic process. The debate is essentially about how a cell produces force. By far the most natural supposition is that the cytoskeleton, or cytosinew, made from muscle proteins, will be the cell's force generating organelle. This being so, statement 1 above hardly restricts the range of statement 2 at all.
Statement 1 becomes predictively valuable only if it is narrowed down further e.g.
- 1A. Capping is caused by the cytoskeleton taking hold of patches and producing forces that reel them in to the rear of the cell. (The original cytoskeletal model).
- 1B. Capping is caused by the cytoskeleton producing forces that cause the membrane to flow. (The membrane flow model).
- 1C. Capping is caused by the cytoskeleton continuously assembling at the front of the cell and disassembling at the rear of the cell. (The cytoskeletal flow model).
- 1D. Capping is caused by the cytoskeleton producing forces that cause travelling wave patterns on the cell surface. (The wave model, Chapter 7).
These statements all narrow down statement 1 further and so are consistent with a narrower range of experimental observations. To some extent, they compete with one another and predict differing results, hence they can be falsified and, by Popper's criterion, are scientific hypotheses. They are the four models discussed in "A Habit of Lies". One could add to the list by suggesting other behaviour for the cytoskeleton.
In the scientific literature it is often hard to know which version of the cytoskeletal model is advocated. Often interpretation seems limited to the cytoskeletal paradigm itself. This is a big part of the problem. The cytoskeletal paradigm (theory 1) simply does not exclude any of theories 1A, 1B etc. Section 2.4 noted that a theory having the property of not excluding other theories, is referred to by scientific philosophers as "vacuous". The cytoskeletal paradigm, theory 1, is vacuous until it is narrowed down by a statement of how the cytoskeleton is supposedly acting. It must be converted to a theory 1A, 1B etc.
Though it has few scientific merits, the cytoskeletal (flow) model has displayed real survival abilities. There are hundreds of papers interpreted in these terms. Be careful here, the papers do not contain evidence to support the model as the data presented can be interpreted in other ways. The authors merely choose to construe their observations in this fashion.
Why this should be so is a question more to be addressed by a sociological analysis of science than a review of the evidence. It raises issues about both the culture and administration of modern science. The cytoskeletal flow model fuses the cytoskeletal model with the flow model to form a hybrid. It seems, instead of a rationalist rejection of the failed cytoskeletal and flow models, the cytoskeletal flow model reflects the search for compromise and consensus, keeping the spirit of the early ideas and maintaining the honour of the parties. Much of the willingness to accept the cytoskeletal flow model seems to come, not from the merits of the model itself, but from its origins in the cytoskeletal and flow models, enabling workers to claim they were right after all. In fact, it is quite different to the original cytoskeletal ideas and no more correct.
Honour alone is not enough to explain the survival of the cytoskeletal flow model. A flowing cytoskeleton is basically absurd and, compromise or not, the idea would likely have died except that those making the decisions were also those whose honour was preserved. There are indications of an inability to think outside the paradigms they themselves laid down. Moreover, the confusion between the original cytoskeletal model and cytoskeletal flow model, enables workers to respond to criticism by reversion to vacuity and holding up the cytoskeletal paradigm as the theory they advocate. A much greater degree of clarity is essential.
The debate in this field has raged backwards and forwards now for many years. All parties permitted to be involved in the argument have effectively disproved the other's ideas and all have largely ignored a far more viable alternative, namely the wave model proposed, amongst others, by this author.
This Chapter has :-
- Reviewed the cytoskeletal flow model as a current version of the cytoskeletal model.
- Asserted that this hypothesis is advanced in the scientific literature as a virtual fact.
- Summarised the criticisms that can be levelled at the cytoskeletal flow model.
- Pointed out that the unmodified cytoskeletal paradigm is vacuous.
- Stated that an alternative, the wave model, does exist but it has been largely ignored.
© Copyright John A Hewitt.
For contact information, see copyright page.
Last Modified 21 October 2005