Using this book: This book is designed to be used in both introductory and advanced

cell biology courses. The primary text is generally on the left side of the vertical divider,

Cytoskeleton : and printed in black. Details that are usually left to an advanced course are printed in
blue and found on the right side of the divider. Finally, additional biomedically relevant
information can be found in red print on either side of the divider.

Structure and Movement

When a eukaryotic cell is taken out of its physiological context and placed in a plastic or glass
Petri dish, it is generally seen to flatten out to some extent. On a precipice, it would behave
like a Salvador Dali watch, oozing over the edge. The immediate assumption, particularly in
light of the fact that the cell is known to be mostly water by mass and volume, is that the cell
is simply a bag of fluid. However, the cell actually has an intricate microstructure within it,
framed internally by the components of the cytoskeleton.

As the name implies, the cytoskeleton acts much like our own skeletons in support- Although the genes are not particularly well conserved, a combi-
ing the general shape of a cell. Unlike our skeletons though, the cytoskeleton is highly nation of genetic similarity and protein structure have confirmed
dynamic and internally motile, shifting and rearranging in response to the needs of the presence of prokaryotic proteins that are related to eukary-
the cell. It also has a variety of purposes beyond simply providing the shape of the otic cytoskeletal proteins in both form and function. Compared
cell. Generally, these can be categorized as structural and transport. While all three to the eukaryotic cytoskeleton, study of prokaryotic proteins is
major components of the cytoskeleton perform each of these functions, they do not very recent, and for a long time, there was an assumption that
do so equally, as their biophysical characteristics are quite different. With respect to prokaryotes did not have or need cytoskeletal architecture. FtsZ,
structure, at some point in the life of every cell, it must change shape, whether simply the bacterial equivalent of tubulin, was discovered in 1980 but
increasing or decreasing in size, or a more drastic alteration like the super-elongated most of the work on it has occurred in the last decade. MreB is
form of neurons with axons, the cytoskeleton must be able to respond by dynamically an actin-like protein, first compared to actin in 1992, and crescen-
increasing and decreasing the size of the internal structures as needed. Structure also tin, an intermediate filament class protein, was only described in
applies to the relative position of internal cellular elements, such as organelles or pro- 2003. For comprehensive review of prokaryotic cytoskeleton pro-
teins, to one another. In many highly specialized cells, the segregation of particular teins, see Graumann, P.L., Ann. Rev. Microbiology 61:589-618, 2007.
structures within certain parts of the cell is crucial for it to function. Transport refers
to the movement of molecules and organelles within the cell as well as movement of
the cell as a whole. We just discussed intracellular movement of proteins and lipids by
way of vesicles in the last chapter. Those vesicles, as we will see in this chapter, are
not just floating from one place to another; they are moved purposefully and direction-
ally along the cytoskeleton like cargo on highways or railroad tracks. With respect
to whole cell movement, this can range from paddling or swimming by single-celled
organisms to the stereotyped and highly coordinated crawling of many cells from their
point of origin to their eventual destination during the development of a metazoan
organism or the movement of fibroblasts to heal a cut in your skin.

Chapter 12, Cytoskeleton, version 1.0 Page 175

 = Microtubules

= Intermediate Filaments

= Actin Filaments

Figure 1. Cytoskeletal element distribution in a prototypical eukaryotic cell. The purple ball is the nucleus.

The three major components of the cytoskeleton are microtubules, microfilaments, and
intermediate filaments. Each of these are polymers composed of repeating subunits in
specific arrangements. With just a quick glance (fig. 1), it is very clear that the interme-
diate filaments will likely play a significantly different role from either microtubules
or microfilaments. Because the IF’s are made of long fibrous subunits that coil around
one another to form the filament, there is clearly a great deal of contact (which facili-
tates formation of hydrogen bonds, aka molecular velcro™) between subunits provid-
ing great tensile strength. It is very difficult to break these subunits apart, and thus the
IF’s are primarily used for long-term or permanent load-bearing purposes. Looking at
the other two components of the cytoskeleton, one can see that with the globular in-
stead of fibrous shape of the subunits, the maximum area of contact between subunits
is greatly limited (think of the contact area when you push two basketballs together),
making it easier to separate the subunits or break the microfilament or microtubule.
The cell can use this characteristic to its advantage, by utilizing these kinds of cy-
toskeletal fibers in dynamic situations where formation or destruction of intermediate
filaments would take far too long. We now address these three groups of cytoskeletal
elements in more detail.

Intermediate Filaments

“Intermediate filaments” is actually a generic name for a family of proteins (grouped

into 6 classes based on sequence and biochemical structure) that serve similar func- Most intermediate filaments fall between 50-100 kDa, including
tions in protecting and shaping the cell or its components. Interestingly, they can even keratins (40-67 kDa), lamins (60-70 kDa), and neurofilaments (62-
be found inside the nucleus. The nuclear lamins, which constitute class V intermediate 110 kDa). Nestin (class VI), found mostly in neurons, is an excep-
filaments, form a strong protective mesh attached to the inside face of the nuclear tion, at approximately 240 kDa.

Chapter 12, Cytoskeleton, version 1.0 Page 176

membrane. Neurons have neurofilaments (class IV), which help to provide structure for
axons — long, thin, and delicate extensions of the cell that can potentially run meters
long in large animals. Skin cells have a high concentration of keratin (class I), which
not only runs through the cell, but connects almost directly to the keratin fibers of
neighboring cells through a type of cellular adhesion structure called a desmosome
(described in the next chapter). This allows pressure that might be able to burst a single
cell to be spread out over many cells, sharing the burden, and thus protecting each
member. In fact, malformations of either keratins or of the proteins forming the des-
mosomes can lead to conditions collectively termed epidermolysis bullosa, in which the Epidermolysis bullosa simplex is a collection of congenital diseases
skin is extraordinarily fragile, blistering and breaking down with only slight contact, caused by mutations to the keratin genes KRT5 or KRT14, or to the
compromising the patient’s first line of defense against infection. plectin gene PLEC1. These mutations either weaken the polymer-
ization of keratin into filaments, or the interaction between kera-
NH2 COOH Figure 2. Intermediate filaments
Monomer
are composed of linear subunits tin filaments. This leads to the inability of each individual cell to
that wrap around each other and maintain structural integrity under pressure. Another type of EB,
interact very tightly.
junctional epidermolysis bullosa (JEB), is caused by mutations to
Coiled-Coil Dimer integrin receptors (b4, a6) or laminins. This includes JEB gravis or
Herlitz disease, which is the most severe, often leading to early
postnatal death. JEB is also related to dystrophic epidermolysis
bullosa (DEB) diseases such as Cockayne-Touraine, each of which
Tetramer s due to a mutation in collagen type VII. The gene products in-
volved in JEB and DEB are discussed in more detail in the next
chapter. They play a role in adhering the cells to the basememnt
membrane, and without them, the disorganization of the cells
leads to incomplete connections between the epidermal cells, and
therefore impaired pressure-sharing.

Some forms of Charcot-Marie-Tooth disease, the most common

Structurally, as mentioned previously, all intermediate filaments start from a fibrous inherited peripheral nerve disease, are also linked to mutations
subunit (fig. 2). This then coils around another filamentous subunit to form a coiled-coil of intermediate filament genes. This disease, also known as per-
dimer, or protofilament. These protofilaments then interact to form tetramers, which oneal muscular atrophy or hereditary motor sensory neuropathy,
are considered the basic unit of intermediate filament construction. Using proteins is a non-lethal degenerative disease primarily affecting the nerves
called plectins, the intermediate filaments can be connected to one another to form of the distal arms and legs. There is a broad variety of CMT types
sheets and meshes. Plectins can also connect the intermediate filaments to other parts and causes, the most common being malformations of Schwann
of the cytoskeleton, while other proteins can help to attach the IF cytoskeleton to the cells and the myelin sheath they form. CMT type 2 is character-
cell membrane (e.g. desmoplakin). The most striking characteristic of intermediate fila- ized by malformations of the peripheral nerve axons, and is linked
ments is their relative longevity. Once made, they change and move very slowly. They to mutations of lamin A proteins and of light neurofilaments.
are very stable and do not break down easily. They are not usually completely inert, but The causal mechanism has not yet been established; however,
compared to microtubules and microfilaments, they sometimes seem to be. the neurofilaments are significant elements in maintaining the
integrity of long axons.

Chapter 12, Cytoskeleton, version 1.0 Page 177

Actin Microfilaments

Microfilaments are also known as actin filaments, filamentous actin, and f-actin, and
they are the cytoskeletal opposites of the intermediate filaments. These strands are
made up of small globular actin (g-actin) subunits that stack on one another with rela-
tively small points of contact. You might envision two tennis balls, one fuzzy and the
other covered in velcro hooks. Even if you push hard to mush them together, the area
of contact between the balls (i.e. the area available for H-bonding between subunits) is
fairly small compared to the overall surface area, or to the area of contact between IF
subunits. They will hold together, but they can also fall apart with relatively little force.
Contrast this with intermediate filaments, which might be represented as two ribbons
of velcro hooks or loops. Considerably more work is required to take them apart. Be-
cause there are fewer H-bonds to break, the microfilaments can be deconstructed very
quickly, making it suitable for highly dynamic applications.

Low G-actin Concentration Figure 3. Actin microfilaments have a (+)

and (-) end. When the free (globular) ac-
tin concentration is low, actin is primar-
_ +
ily added to the (+) end, and lost from the
(-) end. However at high levels of g-actin,
new monomers can potentially add onto
the filament from either end.

High G-actin Concentration

_ +

When the actin subunits come together to form microfilaments, they interact direc-
tionally. That is, subunits have a “top” and a “bottom”, and the top of one subunit
always interacts with the bottom of another. If we go to the “bottom”-most subunit
of a filament, the open end is called the minus (-) end, while the opposite end, which
incidentally sees more additive action, is called the plus (+) end. Microfilaments are also
said to have polarity, but again this is only in the sense of having directionality, and
has nothing to do with electrical charge. Individual microfilaments can exist, but most
microfilaments in vivo are twisted pairs. Unlike DNA; however, microfilament pairs are
not antiparallel: both strands have the same directionality.

Chapter 12, Cytoskeleton, version 1.0 Page 178

The formation of filaments from g-actin is an ATP-dependent process, although not in
the conventional sense of utilizing the energy released in hydrolysis. Instead, the globu-
lar actin subunits will only bind with another g-actin subunit if it has first bound an
ATP. If the g-actin has bound ADP, then it must first exchange the ADP for ATP before
it can be added onto a filament. This alters the conformation of the subunit to allow
for a higher-affinity interaction. A short time later, hydrolysis of the ATP to ADP (with
release of Pi) weakens the affinity but does not directly cause dissolution of the sub-
unit binding. The hydrolysis is brought about by the actin itself, which has this ATPase
enzymatic activity built in.

Although f-actin primarily exists as a pair of filaments twisted around each other, ad-
dition of new actin occurs by the addition of individual g-actin monomers to each fila-
ment (fig. 3). Accessory proteins can be used to help or hinder either the building or
breakdown of the filaments, but the primary mechanism is essentially self-regulating.
When free g-actin levels are high, elongation of actin filaments is favored, and when the
g-actin concentration falls, depolymerization of f-actin predominates. Under average
physiological conditions, though, what is often seen in actin microfilaments is an effect
called treadmilling. Since actin is mostly added onto one end but removed from the
other, the net effect is that any given actin monomer in a filament is effectively moving
from (+) end to (-) end even if the apparent length of the filament does not change.

In most cell types, the greatest concentration of actin-based cytoskeletal structures is

found in the periphery of the cell rather than towards the center. This fits well with
the tendency of the edges of the cell to be more dynamic, constantly adjusting to sense
and react to its environment. Clearly, the polymerization and depolymerization of ac-
tin filaments is much faster than for intermediate filaments. The big exception to the
actin-in-periphery rule is found in muscle cells. Actin filaments, and the myosin motor
proteins that work on them, are the basis for muscle cell contraction, and fill up most
of the muscle cells, not just the periphery. We will discuss the role of actin in both
types of cell movement later in the chapter.

Microtubules

Microtubules are made up of two equally distributed, structurally similar, globular

subunits: a and b tubulin. Like microfilaments, microtubules are also dependent on a Microtubule stability is temperature-dependent: if cooled to 4°C,
nucleotide triphosphate for polymerization, but in this case, it is GTP. Another similar- microtubules fall apart into ab-tubulin heterodimers. Warmed
ity is that microtubules have a polarity in which the (-) end is far less active than the back up to 37°C, the tubulin repolymerizes if there is GTP avail-
(+) end. However, unlike the twisted-pair microfilaments, the microtubules are mostly able.

Chapter 12, Cytoskeleton, version 1.0 Page 179

found as large 13-stranded (each strand is called a protofilament) hollow tube struc-
tures. Also, the a and b tubulin used for building the microtubules not only alternate,
but they are actually added in pairs. Both the a-tubulin and b-tubulin must bind to GTP
to associate, but once bound, the GTP bound to a-tubulin does not move. On the other
hand, GTP bound in the b-tubulin may be hydrolyzed to GDP. GDP-bound ab-dimers
will not be added to a microtubule, so similar to the situation with ATP and g-actin,
if the tubulin has GDP bound to it, it must first exchange it for a GTP before it can be
polymerized. Although the affinity of tubulin for GTP is higher than the affinity for
GDP, this process is usually facilitated by a GEF, or guanine nucleotide exchange factor.
As the signal transduction chapter will show in more detail, this type of nucleotide ex-
change is a common mechanism for activation of various biochemical pathways.
Figure 4. Microtubules ex- GTP GTP

hibit dynamic instability.

GTP-bound ab-tubulin dim- GDP GDP
GDP GDP
GDP
ers are added onto the mi- GTP GTP GTP
GDP
GDP
GDP
GDP
GTP GTP GDP GDP GDP GDP
crotubule. Once the GTP is GTP
GTP GTP
GTP GDP
GTP GTP
hydrolyzed, the conforma- GTP GTP
GTP
GDP
GDP
GDP GDP G
DP
GTP GTP GDP GDP
tional shift strains the mi- GTP
GTP GTP
GTP
GDP GDP
GDP
GDP
GTP
crotubule, which will tend GTP
GTP
GTP
GDP
GDP
GDP GDP
GDP GDP
GDP
GDP
GDP
P G T
to break apart unless new GT
GTP GTP
P
GTP GDP
GDP GDP GDP
GDP
GTP GTP GDP GDP
tubulin dimers are added to GTP
GTP GTP
GTP
GTP
GDP GDP GDP
GDP
GDP
GDP
stabilize the structure. GTP GTP GDP GDP GDP GDP
GDP
GDP
GDP
GD
GDPGTP TP
GTP GDP GDP P GDP
G GTP
GDP P
GDP GD GDP
GDP GDP GD
GDP GDP GDP GDP P
GDP GDP
GDPGDP GDP GDP GDP GDP GDP P
GDP GD
GDP GDP P GDP
GDP GDP GDP GDP GD
GDP GD
GDPGDP GDP GDP GDP GDP GDP GDP GDP GDPGDPGDP P
GDP P GDP GDP
GD
GDP GDP GDP
GDP GDP GDP
GDP GDP GDP GDP GDP GDP GDP GDP
GDP GDP GDP GDP GDP
GDP GDP GDP
GDP
GDP GDP
GDP
GDP GDP GDP GDP
GDP GDP GDP
GDPGDP GDP GDP GDP GDP
GDP GDP
GDP
GDP GDP GDP
GDP GDP
GDP GDP GDP GDP
GDP GDP
GDPGDP GDP GDP GDP GDP GDP GDP GDP
GDP GDP GDP GDP GDP GDP
GDP
GDP GDP GDP
GDP GDP GDP
GDP GDP GDP
GDP GDP GDP GDP
GDP GDP
GDP
GDP GDP GDP

Again like actin, the tubulin itself has enzymatic activity, and over time, the GTPase
activity hydrolyzes the GTP to GDP and phosphate. This changes the attachment be-
tween b-tubulin of one dimer and the a-tubulin of the dimer it is stacked on because
the shape of the subunit changes. Even though it isn’t directly loosening its hold on the
neighboring tubulin, the shape change causes increased stress as that part of the mi-
crotubule tries to push outward. This is the basis of a property of microtubules known
as dynamic instability. If there is nothing to stabilize the microtubule, large portions of
it will fall apart. However, as long as new tubulin (which will have GTP bound) is be-
ing added at a high enough rate to keep a section of low-stress “stable”-conformation
microtubule (called the GTP cap) on top of the older GDP-containing part, then it sta-
bilizes the overall microtubule. When new tubulin addition slows down, and there is
only a very small or nonexistent cap, then the microtubule undergoes a catastrophe

Chapter 12, Cytoskeleton, version 1.0 Page 180

in which large portions rapidly break apart. Note that this is a very different process
than breakdown by depolymerization, which is the gradual loss of only a few subunits
at a time from an end of the microtubule. Depolymerization also occurs, and like with
actin, is determined partially by the relative concentrations of free tubulin and micro-
tubules.

From a physical standpoint, the microtubule is fairly strong, but not very flexible. A
microfilament will flex and bend when a deforming force is applied (imagine the fila-
ment anchored at the bottom end standing straight up, and something pushing the tip
to one side). The microtubule in the same situation will bend only slightly, but break
apart if the deforming force is sufficient. There is, of course, a limit to the flexibility of
the microfilament and eventually, it will also break. Intermediate filaments are slightly
less flexible than the microfilaments, but can resist far more force that either microfila-
ments or microtubules.

Microtubule Organizing Centers

Microtubules, like microfilaments, are dynamic structures, changing in length and in-
teractions to react to intra- and extra-cellular changes. However, the general place-
ment of microtubules within the cell is significantly different from microfilaments,
although there is some overlap as well as interaction. Microfilaments do not have
any kind of global organization with respect to their polarity. They start and end in
many areas of the cell. On the other hand, almost all microtubules have their (-) end
in a perinuclear area known as the MTOC, or microtubule organizing center and they
radiate outward from that center. Since the microtubules all radiate outward from
the MTOC, it is not surprising that they are concentrated
more centrally in the cell than the microfilaments which, as
mentioned above, are more abundant around the periphery
of the cell. In some cell types (primarily animal), the MTOC
contains a structure known as the centrosome. This con-
sists of a centriole (two short barrel-shaped microtubule-
based structures positioned perpendicular to each other)
and a poorly defined concentration of pericentriolar mate-
rial (PCM). The centriole is composed of nine fibrils, all Inhibition of g-tubulin function by antibody blocking, RNA inter-
connected to form a cylinder, and each also connected by ference of expression, and gene knockout confirm that without
Figure 5. An electron micrograph
radial spokes to a central axis. The electron micrograph depicting the cross-section of g-tubulin function, the microtubule structures did not form. In
in figure 5 shows a cross-section of a centriole. In it, each a centriole in an embryonic addition, it appears to be play roles in coordination of late mitosis
fibril is shown to actually be a fused triplet of microtubules. mouse brain cell. L. Howard and M.
Marin-Padilla, 1985
(anaphase onwards).

Chapter 12, Cytoskeleton, version 1.0 Page 181

 GTP GTP GTP GTP
However, in each triplet, only one is a complete microtu-
GTP
GTP
GTP
GTP GTP
GTP
GTP
bule (designated the A tubule), while the B and C tubules
GTP GTP GTP
GTP
do not form complete tubes (they share a wall with the A
GTP
GTP
GTP GTP
GTP
GTP and B tubules, respectively). Interestingly, the centrioles
GTP GTP
GTP
GTP
do not appear to be connected to the cellular microtubule
GTP GTP GTP
GTP network. However, whether there is a defined centrosome
GTP GTP
GTP
GTP
GTP
γ or not, the MTOC region is the point of origin for all micro-
γ γ
γ γ γ γ tubule arrays. This is because the MTOC contains a high
γ γ γ γ
concentration of g-tubulin. Why is this important? With all
of the cytoskeletal elements, though it is most pronounced
Figure 6. g-tubulin ring com- with microtubules, the rate of nucleation, or starting a mi-
plex facilitates microtubule crotubule is significantly slower than the rate of elongating
nucleation.
an existing structure. Since it is the same biochemical in-
teraction, the assumption is that the difficulty lies in getting the initial ring of dimers
into position. The g-tubulin facilitates this process by forming a g-tubulin ring complex
that serves as a template for the nucleation of microtubules (fig. 6). This is true both
in animal and fungal cells with a single defined MTOC, as well as in plant cells, which
have multiple, dispersed sites of microtubule nucleation.

Transport on the Cytoskeleton

While it can be useful to think of these cytoskeletal structures as analogous to an ani-

mal skeleton, perhaps a better way to remember the relative placement of the micro-
tubules and microfilaments is by their function in transporting intracellular cargo from
one part of the cell to another. By that analogy, we might consider the microtubules
to be a railroad track system, while the microfilaments are more like the streets. By
the same analogy, we can suggest that the microtubule network and microfilament
network are connected at certain points so that when cargo reaches its general desti-
nation by microtubule (rail), then it can be taken to its specific address by microfila-
ment. Let’s extend this analogy a bit further. If the microtubules and microfilaments
are the tracks and streets, then what are the trains and trucks? Ah, an astute question,
Grasshopper. On the microtubules, the “trains” are one of two families of molecular
motors: the kinesins and the dyneins.

We can generalize somewhat and say that the kinesins drive towards the (+) end (to-
ward periphery of cell) while the dyneins go toward the (-) end (toward the MTOC). On
actin microfilaments, the molecular motors are proteins of the myosin family. At this
point, the analogies end, as the functioning of these molecular motors is very different

Chapter 12, Cytoskeleton, version 1.0 Page 182

from locomotion by train or truck. Finally, one might question the biological need for Although this type of transport occurs in all eukaryotic cells, a
such a transport system. Again, if we analogize to human transport, then we could say particularly well-studied case is axonal transport (also called axo-
that transport via simple diffusion is akin to people carrying packages randomly about plasmic transport) in neurons. Here, the transport of materials
the cell. That is to say, the deliveries will eventually be made, but you wouldn’t want from the cell body (soma) to the tips of the axons can some-
to count on this method for time-critical materials. Thus a directed, high-speed system times traverse very long distances up to several meters in larger
is needed to keep cells (particularly larger, eukaryotic cells) alive. animals, and must do so in a timely manner. Axonal transport
is generally classified as anterograde (from soma to axon termi-
nal) or retrograde (from terminals back). The types of material
transported in these two directions is very different: much of
the anterograde transport is protein building blocks for extend-
ing the axon or synaptic vesicles containing neurotransmitters;
retrograde transport is mostly endocytic vesicles and signaling
molecules. Axonal transport is also categorized as fast and slow.
Slow transport is primarily the movement of proteins directly
bound to the motors, and they can move from from 100 mm per
day (SCa, slow component a) up to 3mm/day (SCb). In compari-
son, fast transport is generally movement of vesicles, and can vary
from 50 to 400 mm/day. The mechanism of slow transport had
been debated for over a decade until 2000, when direct visualiza-
tion of fluorescently labeled neurofilaments in transport showed
that the actual movement of the proteins was very similar to the
movement in fast axonal transport, but there were many pauses
in the transport, a “stop and go” mechanism rather moving from
= Microtubules = Dynein
source to destination continuously.
= Actin Filaments
= Kinesin
= Myosin V
= Vesicle

Figure 7. Transport on microtubules and microfilaments.

All of the kinesins and dyneins have a few key commonalities. There is a catalytic ener-
gy-releasing “head” connected to a hinge or neck region allowing the molecule to flex
or “step”, and there is a cargo-carrying tail beyond that (fig. 8). The head of a kinesin
or dynein catalyzes the hydrolysis of ATP, releasing energy to change its conformation
relative to the neck and tail of the molecule, allowing it to temporarily release its grip
on the microtubule, swivel its “hips” around to plant itself a “step” away, and rebind to
the microtubule (fig. 9). On the actin microfilaments, the myosins, of which there are
also many types (some depicted in fig. 10) are the molecular motors. Their movement
is different from dyneins and kinesins, as will be described in the next section, but also
uses the energy of ATP hydrolysis to provide energy for the conformational changes
needed for movement. We have introduced the motors, but considering the enormous
diversity in the molecules that need to be transported around a cell, it would be impos-

Chapter 12, Cytoskeleton, version 1.0 Page 183

sible for the motors to directly bind to all of them. In fact, the motors bind to their
cargo via adapter molecules that bind the motor on one side, and a cargo molecule or
vesicle on the other. Further examination of the cargo and the routing of the cargo by
address markers (SNAREs) was discussed in the vesicular transport chapter.
Figure 8. Kinesin (A) and Dynein (B) are A

motor proteins that move along micro- ATP Heavy Chain Light Chain

tubules. Generally, kinesins move to

the (+) end while dyneins move to the
(-) end. Their motor function requires ATP
ATP hydrolysis. ATP binding sites are
marked in white. ATP Heavy Chain
B

ATP

Light Chain
Intermediate Chain

Figure 9. A cargo vesicle (yellow)

can be simultaneously bound by
dynein (green) and kinesin (blue)
via adapter proteins. This top side
also depicts the movement of the
kinesin, in which binding of ATP
causes one “foot” to release, and
hydrolysis of ATP causes the mol-
ecule to swivel the other foot in
front. +
_

Figure 10. Selected Myosins. (A) Type I myosin, pri- A B

marily for binding membranes to f-actin, including
endocytic vesicles. (B) Type II myosin, binds f-actin
on both ends to slide filaments against each other.
(C) Type V myosin, used in vesicular transport. (D) C

Type VI myosin, used in endocytosis. (E) Type XI

myosin, a fast myosin used in cytoplasmic stream-
ing in plant cells. D

Chapter 12, Cytoskeleton, version 1.0 Page 184

Actin - Myosin Structures in Muscle

The motor proteins that transport materials along the acting microfilaments are simi-
lar in some ways, such as the globular head group that binds and hydrolyzes ATP, yet
different in other ways, such as the motion catalyzed by the ATP hydrolysis. Much of
the f-actin and myosin in striated and cardiac muscle cells is found in a peculiar ar-
rangement designed to provide a robust contractile response over the entire length of
the cell. The sarcomere is an arrangement of alternating fibers of f-actin (also known
as “thin fibers” based on their appearance in electron micrographs) and myosin II (or
“thick fibers”). Although we do not normally think of the motor protein as a fiber, in
this case the tails of the myosin II molecules intertwine to form a continuous fiber of
myosin molecules. As the contractile cycle proceeds, the myosin molecules grip the
adjacent actin fibers, and move them. In fig. 11, you can see that a sarcomere is con-
structed so that the stationary myosin fibers are located centrally, with two parallel
sets of actin fibers interspersed between the myosin fibers, to the left and the right
of the center. Note that the actin fibers do not cross the center line, and that at the

M line
H zone
Z Z
line I band A band I band line

Figure 11. Sarcomere. Myosin II is depicted as in fig. 9, but here entwined with other myosins to form the Figure 12. Human skeletal muscle is organized into sarcomeres. The dark Z
thick filament. They are supported and anchored by titin (shown as long tangled orange ribbons). The myo- lines are a clear reference point in comparing this to diagram in fig.11. This elec-
sin heads act on the actin filaments (blue), pulling them towards each other in a contractile movement. tron micrograph placed in the public domain by L. Howard.

Chapter 12, Cytoskeleton, version 1.0 Page 185

center, the myosin molecules switch orientation. The physiological effect of this is that
the actin filaments are all pulled inwards toward the center of the sarcomere. The
sarcomere in turn, is merely one of many connected together to form a myofibril. The
myofibrils extend the length of the muscle cell.

When the myosin head is in its resting state, it is tightly attached to the actin filament.
In fact, rigor mortis occurs in dead animals because there is no more ATP being made,
and thus the sarcomeres are locked into place. Rigor begins approximately 2-3 hours OK, maybe you’ve watched CSI or Bones, etc enough to already know
after death in humans, after reserves of ATP are depleted. When the body relaxes again this, but pretty neat nonetheless, right?
in about 3 days, it is due to the decomposition and breakdown of the actin and myosin

Ca2+

Ca2+ Ca2+

ATP

Ca2+ Ca2+
Ca2+ Ca2+

ATP

ADP

Ca2+ Ca2+
Ca2+ Ca2+
Pi

ADP
ADP Pi

Figure 12. The myosin power stroke. Myosin can only attach to f-actin if there Ca++ available to bind troponin
(green) and move tropomyosin (yellow) out of the binding groove. When ATP binds to the myosin head, it
releases the f-actin. Hydrolysis of the ATP leads to cocking of the myosin head (moving it relative to the
f-actin). As Pi leaves the myosin head, it reattaches to the f-actin, but slightly displaced from its original
binding site. ADPis then released and the myosin undergoes a power stroke in which it springs back to its
original position, moving the f-actin along with it.

Chapter 12, Cytoskeleton, version 1.0 Page 186

proteins. However, while they are still living animals, ATP is generally available, and it
can bind to the myosin head, causing it to lose affinity for the f-actin, and let go (fig.
12). At this point, no significant movement has occurred. Once the ATP is hydrolyzed
though, the myosin head can reattach to the f-actin a little further down the filament
than it had originally. The energy released is stored in the neck region. The ADP and Pi
are still attached to the myosin head as well. The next step is for the Pi to drop off the
myosin, leading to the power stroke. The neck of the myosin swivels around, leading
to a translocation of the head by approximately 10 nm for myosin II. The distance of
translocation varies depending on the type of myosin, but it is not yet clear whether
the length of the neck is proportional to the displacement of the head. Finally, the ADP
drops off the myosin head, increasing the affinity of the head for the f-actin.

The sarcomere structure described in the first paragraph was incomplete in order to
place the major players clearly in their roles. There are other proteins in the sarcomere
with important structural and regulatory functions. One of the key regulatory com-
ponents is tropomyosin. This is a fibrous protein that lies in the groove of an actin
microfilament and blocks access to the myosin binding site. Tropomyosin attaches to
the microfilament in conjunction with a multi-subunit troponin complex. When Ca++ is
available, it can bind to troponin-C, leading to a conformational change that shifts the
position of tropomyosin to reveal the myosin binding site. This is the primary point of
control for muscle contraction: recall that intracellular Ca++ levels are kept extremely
low because its primary function is in intracellular signaling. One way that the Ca++ lev-
els are kept that low is to pump it into a reservoir, such as the endoplasmic reticulum.

Low Ca++ in cytoplasm

Ca++ released from SR

binds to troponin-C

Fig. 13. In low Ca++ conditions, tropomyosin (yellow line) is held in the
myosin-binding groove of f-actin (blue) by a tripartite troponin complex
(light green). Once Ca++ levels increase, it can bind to troponin-C, causing
a conformational shift that moves the tropomyosin out of the way so that
myosin (orange) can bind the actin microfilaments.

Chapter 12, Cytoskeleton, version 1.0 Page 187

In muscle cells, there is a specialization of the ER called the sarcoplasmic reticulum (SR) The SR is a specialization of part of the endoplasmic reticulum,
that is rich in Ca++ pumps and Ca++. When a signal is sent from a controlling nerve cell and contains a high concentration of Ca++ ions because the SR
to the muscle cell, it causes a depolarization of the muscle cell membrane. This conse- membrane is embedded with Ca++ pumps (ATPases) to keep the
quently depolarizes a set of membranes called the transverse tubules (T-tubules) that cytoplasmic concentration low and sequester the Ca++ ions in-
lie directly on parts of the sarcoplasmic reticulum. There are proteins on the t-tubule side the SR. This is regulated by phosphorylation and [Ca++] via
surface that directly interact with a set of Ca++ channel proteins, holding the channel a regulatory protein such as phospholamban (in cardiac muscle).
closed normally. When the t-tubule is depolarized, the proteins change shape, which Phospholamban is an integral membrane protein of the SR that
changes the interaction with the Ca++ channels on the SR, and allows them to open. normally associates with and inhibits the Ca++ pump. However
Ca++ rushes out of the SR where it is available to troponin-c. Troponin-C bound to Ca++ when it is phosphorylated, or as cytoplasmic Ca++ levels rise, the
shifts the tropomyosin away from the actin filament, and the myosin head can bind to phospholamban releases from the Ca++ pump and allows it to
it. ATP can bind the myosin head to start the power stroke cycle, and voila, we have function.
controlled muscle cell contraction.

In addition to the “moving parts”, there are also more static, structural, proteins in
the sarcomere (fig. 11). Titin is a gigantic protein (the largest known, at nearly 3 MDa),
and can be thought of as something of a bungee cord tether to the myosin fiber. Its Disturbances to the proper formation of the titin-based support
essential purpose is to prevent the forces generated by the myosin from pulling the structure can be a cause of dilated cardiomyopathy, and from
fiber apart. Titin wraps around the myosin fiber and attaches at multiple points, with that, congestive heart failure. Some 20-30% of cases of dilated
the most medial just near the edge of the H zone. At the Z-line, titin attaches to a cardiomyopathy are familial, and mutations have been mapped to
telethonin complex, which attach to the Z-disk proteins (antiparallel a-actinin). Titin the N-terminal region of titin, where the protein interacts with
also interacts with obscurin in the I-band region, where it may link myofibrils to the telethonin. Defects in titin are also being investigated with re-
SR, and in the M-band region it can interact with the Ca++-binding protein calmodulin-1 spect to chronic obstructive pulmonary disease, and some types
and TRIM63, thought to acts as a link between titin and the microtubule cytoskeleton. of muscular dystrophy.
There are multiple isoforms of titin from alternative splicing, with most of the variation
coming in the I-band region.

Of course in an actual muscle (fig. 14), what happens is that nerves grow into the
muscle and make synaptic connections with them. At these synaptic connections, the
nerve cell releases neurotransmitters such as acetylcholine (ACh), which bind to recep-
tors (AChR) on the muscle cell. This then opens ion channels in the muscle cell mem-
brane, triggering a voltage change across that membrane, which also happens to affect
the nearby membrane of the transverse tubules subsequently opening Ca++ channels in
the SR. The contraction of sarcomeres can then proceed as already described above.

Chapter 12, Cytoskeleton, version 1.0 Page 188

 Muscle
Neuron

Muscle Fiber
T-tubule Sarcoplasmic reticulum

Myofibril
Sarcomere
Figure 14. The sarcomere in context. The sarcomere of figures 11 and 12 is one tiny contractile unit within
an array that forms a myofibril. The myofibril is one of many within a muscle cell, surrounded by the sar-
coplasmic reticulum, a specialized extension of the ER that sequesters Ca++ until T-tubule excitation causes
its release.

Cytoskeletal Dynamics

In the early development of animals, there is a huge amount of cellular rearrange-

ment and migration as the roughly spherical blob of cells called the blastula starts to
differentiate and form cells and tissues with specialized functions. These cells need
to move from their point of birth to their eventual positions in the fully developed
animal. Some cells, like neurons, have an additional type of cell motility - they extend
long processes (axons) out from the cell body to their target of innervation. In both
neurite extension and whole cell motility, the cell needs to move first its attachment
points and then the bulk of the cell from one point to another. This is done gradually,
and uses the cytoskeleton to make the process more efficient. The major elements in
cell motility are changing the point of forward adhesion, clearing of internal space by

Chapter 12, Cytoskeleton, version 1.0 Page 189

myosin-powered rearrangement of actin microfilaments and the subsequent filling of
that space with microtubules.

For force to be transmitted, the membrane

must be attached to the cytoskeleton. In Ankyrin
Band 3
fact, signaling (chap. 14) from receptors in
the membrane can sometimes directly in- Actin Spectrin

duce rearrangements or movements of the

cytoskeleton via adapter proteins that con- Filamin
nect actin (or other cytoskeletal elements)
Glycophorin
to transmembrane proteins such as inte- Band 3
grin receptors. One of the earliest experi-
mental systems for studies of cytoskeleton- Band 4.1 Adducin

membrane interaction was the erythrocyte

(red blood cell). The illustrations at right Actin Spectrin

(fig. 15) show some of the interactions of

Figure
α-integrin 15. Membrane to microfilament β-integrin
linkage
an extensive actin microfilament network complexes in erythrocytes involve spectrin.
with transmembrane proteins. Ankyrin and Talin

spectrin are important linkage proteins between the transmembrane proteins and
Actin Vinculinthe

microfilaments. This idea of building a protein complex around the cytoplasmic side
of a transmembrane protein is ubiquitous, andα-actinin
scaffolding (linking) proteins are used
not only in connecting the extracellular substrate (via transmembrane protein) to the
cytoskeleton, but also to physically connect signaling molecules and thus increase the
speed and efficiency of signal transduction.

Accessory proteins to actin filaments and microtubules were briefly mentioned earlier.
Among other functions, they can control polymerization and depolymerization, form
bundles, arrange networks, and bridge between the different cytoskeletal networks.
For actin, the primary polymerization control proteins are profilin, which promotes
polymerization and thymosin b4, which sequesters g-actin. The minus end capping
proteins Arp 2/3 complex and tropomodulin, and the plus end capping proteins CapZ,
severin, and gelsolin can stabilize the ends of f-actin. Finally, cofilin can increase depo-
lymerization from the (-) end.

Profilin has two activities that promote polymerization. First, it is a nucleotide ex-
change factor that removes ATP bound to g-actin, and replaces it with ADP. This
sounds counterintuitive, but keep reading through to the next paragraph. Second,
when bound to a g-actin, it increases the rate of addition to actin microfilaments. It
does so by binding to the end opposite the ATP-binding site, leaving that site and that
side open to binding both ATP and the (+) end of a microfilament. Profilin can be found

Chapter 12, Cytoskeleton, version 1.0 Page 190

both in the cytoplasm at large, and associated with phospholipids (PIP2) and membrane
proteins, to control such processes as leading edge remodeling of f-actin cytoskeletal
structures.

Thymosin b4 regulates microfilament assembly by controlling the available pool of g-

actin. We already stated that greater concentrations of g-actin can increase polymer-
ization rates. However, because of the highly dynamic nature of the actin cytoskeleton,
the time constraints of degrading and producing new actin would prevent the fast-
response control necessary. Therefore, the optimal mechanism is to maintain a large
pool of g-actin monomers, but regulate its availability by tying it up with a sequester-
ing protein - thymosin b4. Thymosin b4 has a 50x higher affinity for g-actin-ATP than for
g-actin-ADP, so here is where profilin comes back into the picture. Profilin exchanges
the ATP of a Tb4-g-actin-ATP complex for an ADP. The result is that the Tb4 releases the
g-actin-ADP, allowing it to enter the general pool for building up filaments.

Increased depolymerization and slowing or cessation of polymerization can gradually Gelsolin is inhibited by the phospholipid PIP2. Phospholipase C,
break down f-actin structures, but what if there is a need for rapid breakdown? Two which breaks down PIP2 can also increase cytosolic Ca++, which is
of the capping proteins previously mentioned, gelsolin and severin, have an alternate an activator of gelsolin. Thus it is possible to rapidly upregulate
mode of action that can sever actin microfilaments at any point by binding alongside gelsolin activity by PLC signaling.
an actin filament and altering the conformation of the subunit to which it is bound.
The conformational change forces the actin-actin interaction to break, and the gelsolin
or severin then remains in place as a (+) end capping protein.

On the microtubule side of things, due to dynamic instability, one might think that Mutations in spastin are linked to 40% of those spastic paraple-
a severing enzyme is not needed, but in fact, spastin and katanin are microtubule- gias distinguished by degeneration of very long axons. The sever-
severing proteins found in a variety of cell types, particularly neurons. There is also ing ability of spastin appears to be required for remodeling of the
a Tb4-like protein for tubulin: Op18, or stathmin, which binds to tubulin dimers (not cytoskeleton in response to neuronal damage.
monomers), acting to sequester them and lower the working concentration. It is regu-
lated by phosphorylation (which turns off its tubulin binding). Microtubule-associated
proteins MAP1, MAP2, and tau (t) each work to promote assembly of microtubules, as Tau has a complicated biomedical history. Its normal function is
well as other functions. MAP1 is the most generally distributed of the three, with tau clear - assembling, stabilizing, and linking microtubules. Howev-
being found mostly in neurons, and MAP2 even more restricted to neuronal dendrites. er, it is also found in hyperphosphorylated neurofibrillary tangles
These and some other MAPs also act to stabilize microtubules against catastrophe by that are associated with Alzheimer’s disease. A cause for Alzheim-
binding alongside the microtubule and reinforcing the tubulin-tubulin interactions. er’s is not yet known, so it is still unclear whether the tau protein
tangles are play a major role in any of the symptoms.
Finally, with respect to microfilament and microtubule accessory proteins, there are
the linkers. Some of the aforementioned MAPs can crosslink microtubules either into
parallel or mesh arrays, as can some kinesins and dyneins, although they are conven-
tionally considered to be motor proteins. On the microfilament side, there are many
known proteins that crosslink f-actin, many of which are in the calponin homology

Chapter 12, Cytoskeleton, version 1.0 Page 191

domain superfamily, including fimbrin, a-actinin, b-spectrin, dystrophin, and filamin. FG Syndrome is a genetically linked disease characterized by men-
Although they all can bind to actin, the shape of the protein dictates different types of tal retardation, enlarged head, congenital hypotonia, imperforate
interaction: for example, fimbrin primarily bundles f-actin in parallel to form bundles, anus, and partial agenesis of the corpus callosum. It has been
while filamin brings actin filaments together perpendicularly to form mesh networks. linked to mutations in several X chromosome genes, including
filamin A (FLNA, FLN1, located Xq28).

Cell Motility Mutations in dystrophin, which is a major muscle protein of the

CD-domain superfamily, can result in Duchenne Muscular Dys-
There are a number of ways in which a cell can move from one point in space to an- trophy or the related but less severe Becker Muscular Dystrophy.
other. In a liquid medium, that method may be some sort of swimming, utilizing ciliary The most distinctive feature is a progressive proximal muscular
or flagellar movement to propel the cell. On solid surfaces, those mechanisms clearly degeneration and pseudohypertrophy of the calf muscles. Onset
will not work efficiently, and the cell undergoes a crawling process. In this section, of DMD is usually recognized before age 3 and is lethal by age 20.
we begin with a discussion of ciliary/flagellar movement, and then consider the more However, symptoms of BMD may not present until the 20s, with
complicated requirements of cellular crawling. good probability of long-term survival. Although it is primarily a
muscle-wasting disease, dystrophin is present in other cell types,
Cilia and flagella, which differ primarily in length rather than construction, are microtu- including neurons, which may explain a link to mild mental re-
bule-based organelles that move with a back-and-forth motion. This translates to “row- tardation in some DMD patients. Like FLNA, the dystrophin gene
ing” by the relatively short cilia, but in the longer flagella, the flexibility of the structure is also located on the X chromosome (Xp21.2).
causes the back-and-forth motion to be propagated as a wave, so the flagellar move-
ment is more undulating or whiplike (consider what happens as you waggle a garden
hose quickly from side to side compared to a short piece of the same hose). The core
of either structure is called the axoneme, which is composed of 9 microtubule doublets
connected to each other by ciliary dynein motor proteins, and surrounding a central
core of two separate microtubules. Central Microtubule Nexin Bridge PDG PDG

This is known as the “9+2” formation, (13 protofilaments) PDG PDG GDP GDP

although the nine doublets are not the PDG PDG

Microtubule Doublet
GDP GDP GDP GDP GDP GDP GDP
PDG GDP GDP GDP
PDG

same as the two central microtubules. (13 + 10 protofilaments)

The A tubule is a full 13-protofilaments,
PDG PDG

GDP GDP GDP GDP GDP GDP GDP

PDG GDP GDP GDP
PDG

but the B tubule fused to it contains GDP GDP

GDP GDP
PDG PDG

GDP GDP GDP GDP

GDP

only 10 protofilaments. Each of the

PDG PDG GDP

Dynein Arm

central microtubules is a full 13 proto- GDP GDP

PDG PDG
GDP GDP
PDG PDG

GDP GDP
GDP GDP
GDP GDP

filaments. The 9+2 axoneme extends GDP

PDG PDG

GDP
GDP GDP GDP

the length of the cilium or flagellum

GDP GDP GDP
PDG GDP GDP
PDG

from the tip until it reaches the base, GDP GDP

PDG PDG
GDP GDP
PDG PDG

GDP GDP
GDP GDP
GDP GDP

and connects to the cell body through

GDP GDP

a basal body, which is composed of 9

GDP GDP GDP
GDP GDP GDP

microtubule triplets arrange in a short

barrel, much like the centrioles from Central Sheath Radial Spoke
which they are derived. Figure 16. Partial (cutaway) diagram of an axoneme, the
central bundle of cilia and flagella.

Chapter 12, Cytoskeleton, version 1.0 Page 192

The ciliary dyneins provide the motor capability, but there are two other linkage pro- This section refers only to eukaryotes. Some prokaryotes also
teins in the axoneme as well. There are nexins that join the A-tubule of one doublet have motile appendages called flagella, but they are completely
to the B-tubule of its adjacent doublet, thus connecting the outer ring. And, there are different in both structure and mechanism. The flagella them-
radial spokes that extend from the A tubule of each doublet to the central pair of micro- selves are long helical polymers of the protein flagellin, and the
tubules at the core of the axoneme. Neither of these has any motor activity. However, base of the flagellin fibers is connected to a rotational motor
they are crucial to the movement of cilia and flagella because they help to transform a protein, not a translational motor. This motor (fig. 18) utilizes ion
sliding motion into a bending motion. When ciliary dynein (very similar to cytoplasmic (H+ or Na+ depending on species) down an electrochemical gradi-
dyneins but has three heads instead of two) is engaged, it binds an A microtubule on ent to provide the energy to rotate as many as 100000 revolutions
one side, a B microtubule from the adjacent doublet, and moves one relative to the oth- per minute. It is thought that the rotation is driven by conforma-
er. A line of these dyneins moving in concert would thus slide one doublet relative to tional changes in the stator ring, nestled in the cell membrane.
the other, if (and it’s a big “if”) the two doublets had complete freedom of movement.
However, since the doublets are interconnected by the nexin proteins, what happens
as one doublet attempts to slide is that it bends the connected structure instead (fig.
17). This bend accounts for the rowing motion of the cilia, which are relatively short, as
well as the whipping motion of the long flagella, which propagate the bending motion
down the axoneme.
PDG
PDG
PDG
PDG

PDG
PDG

PDG

PDG
PDG
PDG
PDG

PDG
PDG

PDG

PDG
PDG
PDG
PDG

PDG
PDG

PDG

PDG
PDG
PDG
PDG

PDG
PDG

PDG

Unattached
PDG
PDG
PDG
PDG

PDG
PDG

PDG

PDG
PDG
PDG
PDG

PDG
PDG

PDG

Force Generation
PDG
PDG
PDG
PDG

PDG
PDG

PDG

PDG
PDG
PDG
PDG

PDG
PDG

PDG

PDG PDG
PDG PDG
PDG PDG
PDG PDG

PDG PDG
PDG PDG

PDG PDG

PDG PDG
PDG PDG
PDG PDG
PDG PDG

PDG PDG
PDG PDG

PDG PDG

Attached
via nexin

Figure 18. The bacteria flagellum is completely different from

eukaryotic flagella. It is moved by a rotary motor driven by
Force Generation proton or Na+ ion flow down the electrochemical gradient. Il-
lustration released to public domain by M.R. Villareal.

Figure 17. The nexin bridges connecting adjacent microtubule doublets transform
the sliding motion generated by the ciliary dynein into a bending motion.

Although we think of ciliary and flagellar movement as methods for the propulsion of
a cell, such as the flagellar swimming of sperm towards an egg, there are also a num-
ber of important places in which the cell is stationary, and the cilia are used to move
fluid past the cell. In fact, there are cells with cilia in most major organs of the body.
Several ciliary dyskinesias have been reported, of which the most prominent, primary
ciliary dyskinesia (PCD), which includes Kartagener syndrome (KS), is due to mutation

Chapter 12, Cytoskeleton, version 1.0 Page 193

of the DNAI1 gene, which encodes a subunit (intermediate chain 1) of axonemal (ciliary)
dynein. PCD is characterized by respiratory distress due to recurrent infection, and the
diagnosis of KS is made if there is also situs inversus, a condition in which the normal
left-right asymmetry of the body (e.g. stomach on left, liver on right) is reversed. The
first symptom is due to inactivity of the numerous cilia of epithelial cells in the lungs.
Their normal function is to keep mucus in the respiratory track constantly in motion.
Normally the mucus helps to keep the lungs moist to facilitate function, but if the mu-
cus becomes stationary, it becomes a breeding ground for bacteria, as well as becoming
an irritant and obstacle to proper gas exchange.

Situs inversus is an interesting malformation because it arises in embryonic develop-

ment, and affects only 50% of PCD patients because the impaired ciliary function causes
randomization of left-right asymmetry, not reversal. In very simple terms, during early
embryonic development, left-right asymmetry is due in part to the movement of mo-
lecular signals in a leftward flow through the embryonic node. This flow is caused by
the coordinated beating of cilia, so when they do not work, the flow is disrupted and
randomization occurs.

Other symptoms of PCD patients also point out the work of cilia and flagella in the
body. Male infertility is common due to immotile sperm. Female infertility, though
less common, can also occur, due to dysfunction of the cilia of the oviduct and fallopian
tube that normally move the egg along from ovary to uterus. Interestingly there is also
a low association of hydrocephalus internus (overfilling of the ventricles of the brain
with cerebrospinal fluid, causing their enlargement which compresses the brain tissue
around them) with PCD. This is likely due to dysfunction of cilia in the ependymal cells
lining the ventricles, and which help circulate the CSF, but are apparently not com-
pletely necessary. Since CSF bulk flow is thought to be driven primarily by the systole/
diastole change in blood pressure in the brain, some hypothesize that the cilia may be
involved primarily in flow through some of the tighter channels in the brain.

Chapter 12, Cytoskeleton, version 1.0 Page 194

Cell crawling (fig. 19) requires A
Leading edge advancing
the coordinated rearrangement
of the leading edge microfila-
Actin polymerization
ment network, extending (by
both polymerization and slid- Integrins Collagen substrate

ing filaments) and then forming B

adhesions at the new forward-
most point. This can take the Release of Adhesion of
leading edge
trailing edge
form of filopodia or lamellipodia,
and often both simultaneously.
Filopodia are long and very thin
C
projections with core bundles of Cell body motion

parallel microfilaments and high Internalization

concentrations of cell surface of receptor

receptors. Their purpose is pri-

marily to sense the environment.
Lamellipodia often extend be- Figure 19. Cells crawl by (a) extending the leading edge primarily
through remodeling of the actin cytoskeleton, (b) forming new
tween two filopodia and is more adhesive contacts at that leading edge while releasing adhesions
of a broad ruffle than a finger. to the rear, and (c) bulk internal movement forward to “catch
up” with the leading edge.
Internally the actin forms more
into meshes than bundles, and the broader edge allows for more adhesions to be made
to the substrate. The microfilament network then rearranges again, this time opening
a space in the cytoplasm that acts as a channel for the movement of the microtubules
towards the front of the cell. This puts the transport network in place to help move
intracellular bulk material forward. As this occurs, the old adhesions on the tail end
of the cell are released. This release can happen through two primary mechanisms:
endocytosis of the receptor or deactivation of the receptor by signaling/conformational
change. Of course, this oversimplification belies the complexities in coordinating and
controlling all of these actions to accomplish directed movement of a cell.
One model of microfilament force generation, the Elastic Brown-
Once a cell receives a signal to move, the initial cytoskeletal response is to polymerize ian Ratchet Model (Mogilner and Oster, 1996), proposes that due
actin, building more microfilaments to incorporate into the leading edge. Depending to Brownian motion of the cell membrane resulting from con-
on the signal (attractive or repulsive), the polymerization may occur on the same or tinuous minute thermal fluctuation, the actin filaments that push
opposite side of the cell from the point of signal-receptor activation. Significantly, the out towards the edges of the membrane are flexed to varying
polymerization of new f-actin alone can generate sufficient force to move the mem- degrees. If the flex is large enough, a new actin monomer can fit
brane forward, even without involvement of myosin motors! Models of force genera- in between the membrane and the tip of the filament, and when
tion are being debated, but generally start with the incorporation of new g-actin into the now longer filament flexes back, it can exert a greater push
a filament at its tip; that is, at the filament-membrane interface. Even if that might on the membrane. Obviously a single filament does not geneate
technically be enough, in a live cell, myosins are involved, and help to push and arrange much force, but the coordinated extension of many filaments can
filaments directionally in order to set up the new leading edge. In addition, some push the membrane forward.

Chapter 12, Cytoskeleton, version 1.0 Page 195

filaments and networks must be quickly severed, and new connections made, both be-
tween filaments and between filaments and other proteins such as adhesion molecules
or microtubules.

How is the polymerization and actin rearrangement controlled? The receptors that
signal cell locomotion may initiate somewhat different pathways, but many share some
commonalities in activating one or more members of the Ras-family of small GTPases.
These signaling molecules, such as Rac, Rho, and cdc42 can be activated by receptor ty-
rosine kinases (see RTK-Ras activation pathways, Chap. 14). Each of these has a slightly
different role in cell motility: cdc42 activation leads to filopodia formation, Rac activates
a pathway that includes Arp2/3 and cofilin to lamellipodia formation, and Rho activates
myosin II to control focal adhesion and stress fiber formation. A different type of recep-
tor cascade, the G-protein signaling cascade (also Chapter 14), can lead to activation of
PLC and subsequent cleavage of PIP2 and increase in cytosolic Ca++. These changes, as
noted earlier, can also activate myosin II, as well as the remodeling enzymes gelsolin,
cofilin, and profilin. This breaks down existing actin structures to make the cell more
fluid, while also contributing more g-actin to form the new leading edge cytoskeleton.

In vitro experiments show that as the membrane pushes forward, new adhesive con-
tacts are made through adhesion molecules or receptors that bind the substrate (often
cell culture slides or dishes are coated with collagen, laminin, or other extracellular
matrix proteins). The contacts then recruit cytoskeletal elements for greater stability
to form a focal adhesion (fig. 20). However, the formation of focal adhesions appears
to be an artifact of cell culture, and it is unclear if the types of adhesions that form in
vivo recruit the same types of cytoskeletal components.

α-integrin β-integrin

Talin

Actin Vinculin

α-actinin

Figure 20. Focal adhesions form when integrins bind to an arti-

ficial ECM surface in cell culture dishes.

The third step to cell locomotion is the bulk movement of the cellular contents for-
ward. The mechanisms for this phase are unclear, but there is some evidence that
using linkages between the actin cytoskeleton at the leading edge and forward parts
of the microtubule cytoskeleton, the microtubules are rearranged to form an efficient

Chapter 12, Cytoskeleton, version 1.0 Page 196

transport path for bulk movement. Another aspect to this may be a “corralling” effect Much of the work on microtubule-actin interactions in cell motil-
by the actin networks, which directionally open up space towards the leading edge. ity has been done through research on the neuronal growth cone,
The microtubules then enter that space more easily than working through a tight actin which is sometimes referred to as a cell on a leash, because it acts
mesh, forcing flow in the proper direction. almost independently like a crawling cell, searching for the proper
pathway to lead its axon from the cell body to its proper synaptic
Finally, the cell must undo its old adhesions on the trailing edge. This can happen in a connection (A.W. Schaefer et al, Dev. Cell 15: 146-62, 2008).
number of different ways. In vitro, crawling cells have been observed to rip themselves
off of the substrate, leaving behind tiny bits of membrane and associated adhesion
proteins in the process. The force generated is presumed to come from actin-myosin
stress fibers leading from the more forward focal adhesions. However, there are less
destructive mechanisms available to the cells. In some cases, the adhesivity of the cel-
lular receptor for the extracellular substrate can be regulated internally, perhaps by
phosphorylation or dephosphorylation of a receptor. Another possibility is endocytosis
of the receptor, taking it off the cell surface. It could simply recycle up to the leading
edge where it is needed (i.e. transcytosis), or if it is no longer needed or damaged, it
may be broken down in a lysosome.

Chapter 12, Cytoskeleton, version 1.0 Page 197