An Introduction to the Cytoskeleton.

 

The cytoskeleton is a three dimensional network of filamentous protein which fills the space between organelles and gives shape and structure to cells.  The cytoskeleton also provides the cell with “motility”, that being the ability of the entire cell to move and for material to be moved within the cell..  Three main protein systems constitute the cytoskeleton, these are (in order of typical abundance):-  Microfilaments, Intermediate filaments and Microtubules.  Although the term “cytoskeleton” is well used and accepted it unfortunately gives an impression of a rather static entity whereas all three constituents are dynamic structures, they constantly change shape through cycles of polymerisation / depolymerisation and interactions with other proteins. 

 

Microfilaments

Microfilaments are linear assemblages of the 43 Kilodalton protein actin. Actin is the most abundant protein in typical eukaryotic cells, accounting for as much as 15% of total protein. It is a highly conserved protein: the amino-acid sequence of actin from Acanthamoeba, a small soil amoeba, is 95% identical to vertebrate isoforms of actin. X-ray crystallography has revealed that the actin monomer is approximately pear shaped, and when viewed conventionally with the more pointed end upper most, both the NH2 and the COOH termini are seen in the bottom right hand corner.

Figure 1  An Actin monomer

           

Actin is composed of four domains with a large cleft almost bisecting the molecule. This cleft forms both a divalent cation (most likely magnesium in cells) and nucleotide binding site. Because the actin subunit has polarity, the microfilament also has (figure 2). Traditionally, the ends of the microfilament have been referred to as "pointed" and "barbed". This nomenclature arises from the resemblance of microfilaments decorated with fragments of myosin II (see later) to arrowheads in the electron microscope. Happily, this nomenclature coincides with the pointed appearance of the actin monomer! (see figure 1 above, top is pointed end). The microfilament is a single-stranded helix with each monomer rotated 166o with respect to neighbouring subunits which means that every 36 nm, or every 13 subunits, subunits eclipse each other at what appears to be a crossover.

           

Figure 2          A Microfilament

Microtubules

Microtubules (MTs) are assemblages of 110-kDa tubulin dimers.  Each dimer is actually a heterodimer, i.e. the polymerising subunit is one 55-kDa a-tubulin associated with one 55-kDa b-tubulin.  As their name suggests MTs are small tubes.  They are 25nm in diameter with an internal diameter of 14nm.  It is not known if materials are transported within the lumen of the MT, (this is unlikely as the ends at the cell centre are most likely blocked) so MT perform a scaffold function rather that that of a pipe.  Note that the MTs are polar, i.e. they have a b-tubulin exposed at the minus end and an a-tubulin exposed at the plus end.  Each MT is typically composed of 13 tubulins arranged around the circumference, but some MTs (especially those found in protozoans) exist which break this general rule.

 

Figure 3

 

Intermediate Filaments

Intermediate Filaments (IFs), are so called because, at 10nm in diameter they are typically intermediate in size between microfilaments and microtubules.  IFs are different to microfilaments and microtubules in a number of fundamental respects.  First of all they tend to be more or less permanent structures in tissues such as skin and hair, in fact in these non-living tissues IF proteins are almost the only protein.  Thus it is true (but a little sad) to say that beauty is only IF thick!  In other cell types, IFs are modified by phosphorylation when they are required to be disassembled for example during cell division.  Unlike the highly conserved actins and tubulins more than 40 distinct IF proteins are encoded by a number of genes in mammalian cells.  All IF proteins have a similar structure with a central helical rod domain and more variable head and tail domains.  The IFs can be divided into five major classes:-

 

                        Class               Name                          Tissue

 

                        i                       Acidic Keratins             Epithelia

                        ii                      Basic Keratins              Epithelia

                        iii                     Desmin                         Muscle

                                                Glial                             Glial cells and astrocytes

                                                Peripherin                     Peripheral neurones

                                                Vimentin                       Mesenchyme

                        iv                     Neurofilaments             Neurons

                        v                      Lamins                         Nuclear envelopes

 

Figure 4.  A representation of the domain structure of the intermediate filament family.  The numbers refer to the number of amino-acids typically forming each domain.  The tail region is the most diverse, some IFs do not have any, while others (neurofilament-H), has a tail of 607 amino-acids.  IF monomers assemble in a parallel fashion:-

 

           

 

Dimerisation takes place by coiled-coil interaction of the a-helical domain.  The two helices (top left) associate with those of another molecule wrapping around each other (bottom left), so that the N terminus and C terminus lie next to each other.  The rectangles to the right give a simplified view for later comparison with higher order structures.

 

 

The IF dimers now associate with other dimers in an anti-parallel fashion so that there are now two N and two C termini at each end of the complex to form a tetramer (top).  The next step is the association of the N terminal head of one tetramer with the C terminal tails of another. IF assembly can then proceed in this manner infinitely.  It should be noted that the above scheme is somewhat tentative and lacks firm evidence. It is not clear for example exactly which domains are responsible for tetramers binding end to end.

Figure 5     Intermediate filament structure and assembly

 

Cellular Organisation of the Cytoskeleton

The three cytoskeletal components have distinct sub-cellular localisations.  Microfilaments are enriched in a layer known as the “cell cortex”, immediately beneath the plasma membrane, and in cell projections such as microvilli.  Microtubules extend from the perinucleus towards the cell periphery.  The plus ends of MTs point to the cell periphery.  IFs are distributed in a similar pattern to MTs except where cells are in contact where the IFs are enriched.  IFs and MTs are excluded from the actively expanding leading edge of the moving or “ruffling” cell.  In some situations a co-localisation of the cytoskeletal systems is seen.  For example, neurofilaments and MT co-localise in the axon of neurons where specific cross links are made between the two systems.  Cross linking proteins such as “filamin” also exist which bind both MTs and MFs.  MFs are also associated with a number of other structures in specific situations, such as the contractile ring in dividing cells, and in specific cell types such as the “focal contact” in fibroblasts,

 

Figure 6.  Distribution of Cytoskeletal Systems in a typical Cell

            Thin straight lines                      - Microfilaments.  Arranged in bundles in microvilli

            Wavy thicker black lines           - Intermediate  Filaments.  Connect cell to cell

            Wavy thick grey lines                - Microtubules.  Radiate from perinuclear MTOC.

 

 

Summary

Microfilaments and Microtubules are similar in many respects, they are polar, dynamic structures whose assembly state is nucleotide dependent and both interact with a host of associated proteins .  IFs are apolar and rather more static polymers which are depolymerised by phosphorylation.  The three systems differ in their mechanical properties; MFs form visco-elastic gels; MTs resist bending and compression; IFs are extremely tough fibres that resist stretching. MFs are arranged as gels or bundles in association with a large number of actin binding proteins.  MTs are usually single typically have their minus end associated with an MTOC deep within the cell, with the plus end toward the periphery.  IFs connect cell-cell junctions to give strength to tissues.  All three systems are interconnected to various extents.

 

References:-

General                   Molecular Biology of the Cell chapter 16, p787. (for third edition)

                                                Bray, D. Cell Movements Garlands press (1992).

 

Intermediate filaments         Stewart, M. Current Opinion Cell Biol. 5, 3-11 (1993).

 

The Cytoskeleton & Movement of Vesicles/Organelles.

 

It is often necessary for vesicles and organelles to be transported to other parts of the cell.  An obvious example of this is the axon of a nerve cell which may be yards long.  Vesicles full of neurotransmitters, synthesised in the cell body must travel down the length of the axon to be delivered at the synapse.  Most metabolites in cells merely diffuse throughout the cell and don’t need specific conduits, but organelles and vesicles are too large for diffusion to be useful since they would become entangled in the cytoplasm and other organelles.  Also, a simple diffusive mechanism would oblige an equal distribution of all organelles and vesicles to exist in all cell types; an obviously intolerable situation. 

Diffusion cannot account for vesicle transport  (see Bloom & Goldstein, 1998)

 

The equation for three dimensional diffusion is:-                                  and for one dimensions:-

 

                                                                                             

 

Where t is the time taken to travel the length L and D is the diffusion constant  2.5x10-10 cm2/sec for a 200nm diameter neutrophil granule ( Felder  & Kam, 1994).  The diffusion constant was arrived at by measurement in neutrophils devoid of microfilaments or microtubules.  In an average cell this means that it would take about 10 minutes to diffuse from the Golgi to the PM  in the absence of a cytoskeletal networks that would otherwise impede it.  This one imagines may be sufficient but if one does the same calculation for a 1m long axon this figure (1 D diffusion) some 630,000 years would be required to transport a vesicle from one end to the other (this is slower that the west coast Virgin service through Carlisle!)

 

There are two mechanisms by which the cell might produce localised distribution of vesicles, one is to allow the vesicles unhindered access to all parts of the cytosol, but to trap them when they arrived at that particular region, alternatively the cell may move the vesicles to specific regions.  Cells seems to adopt both mechanisms although the difference between the two methods is not always apparent.

            Vesicular and organelle motility is most obviously associated with the microtubule system (see second lecture handout  MPT-10).  Microtubules are ideal candidates for the provision of tracks along which cargoes may be transported since they are rigid, long, straight and polar.  Actin is not straight (unless it is bundled), and often arranged in orthogonal gels (not polar).  However microfilaments do provide tracks and anchors for organelles in a number of important situations.

Microtubules are ideally suited as tracks to deliver vesicles and organelles to distant cellular regions.  They are rigid and so comparatively straight, they are polar with the plus end almost invariably pointing to the periphery.  Two main groups propel vesicles and organelles along MT in the cytoplasm; cytoplasmic dynein and a more diverse group, the kinesins

 

Cytoplasmic Dynein

 

Cytoplasmic dynein is also known as MAP1C, it is a huge, two headed ATP-ase which is minus end directed.  Some dyneins from the flagella of Chlamydomonas and Tetrahymena have three heads, but most other flagellar dyneins have just the two.  Cytoplasmic dynein moves vesicles at around 0.3 µm/sec in vitro, is unaffected by NEM and Vanadate and can utilise ATP, GTP or ITP for force production.  The enormous size of the dynein molecule has meant that our understanding of its structure has not progressed as fast as kinesin.  The heavy chain is 4644 amino acids in rat and smaller in both Dictyostelium and Saccharomyces, about 3300 AA of this form the “head” while the remainder form the stalk (compared to 340 AAs for the kinesin head and 850 AAs for myosin).

The dynein heavy chain contains 4 conserved nucleotide-binding “P-loop” motifs.  Only the first of these P-loops is actually involved in nucleotide hydrolysis, but the others may also bind nucleotide, perhaps to regulate the molecule (Mocz & Gibbons, 1996).  It is known that vanadate is capable of cleaving the dynein (and other ATPases) at the ATP binding site, this is, as expected in the first P-loop.  The microtubule binding site is exposed where one might imagine it would be on the three dimensional structure of the dynein molecule, at the top, on an extended structure (see part a) (Gee et al, 1997).  However, the microtubule site is encoded by a stretch of sequence toward the C-terminus, next to the stalk domain.  Solving the 3D structure of the dynein molecule will be a major task!  Not much is presently known about the regulation of dynein at the molecular level.  What is known is largely phenomenological, for example it is known that okadaic acid (an inhibitor of phosphatases PP1 and PP2A) causes a 27 fold increase in the number of ER tubules moving on microtubules in an in vitro reconstituted experiment (Allan, 1995), but it is not known which of the many proteins in the complex that gets phosphorylated is the regulating component.  

Figure 7

 

The Kinesins

The kinesins are a large group of microtubule based motor proteins that are formed from multi-component complexes.  The first kinesin was identified in the axon in 1985 but several other kinesin related proteins (KRP) have come to light.  Three kinesin subfamilies have been implicated in the plus end directed movement of vesicles and organelles.  All three share a similar kinesin “head” domain but are otherwise quite different.  The kinesin head domain has a fold with surprising similarity to the myosin head and some switch type G-proteins (Vale, 1996).  Although the kinesins carry vesicles it is suspected that each type carries a specific sub-set of vesicle.  Kinesin moves vesicles at around 1 µm/sec, is inhibited by NEM and Vanadate and can only utilise ATP for force production

            Many other kinesin related motor proteins have been discovered which are involved in the spindle pole formation and karyokinesis.  These have largely been identified at the gene level by genetic screen and so not much is known about their biochemical properties.  However, one of these, NCD is a kinesin related protein which drives minus end movement (see below).

 

______________________________________________________________________________________________________

            Protein                         direction           Mol.Wt (kDa)                                      Speed              Cargo

                                                                        of heavy chain                                       (µm/min)

______________________________________________________________________________________________________

            Kinesin                         plus                  110 (heterotetrameric)                          30-54               Vesicles

 

            Unc-104 (KIF1A)        plus                  192 (monomeric)                                  72                    precursor                                                                                                                                                                                                                      vesicles

                         (KIF1B)          plus                  130 (monomeric)                                  40                    Mitochondria

 

            Kinesin II (KRP85/95) plus                  79 (heterotrimer)                                  24                    Vesicles

            Rabkinesin-6                ?                      ~100 (?)                                               ?                      Golgi & TGN

            ncd                               minus                ~100                                                                            chromosomes

 

Figure 8

 

Control of Dynein and Kinesin motor activity.

In most cell types, MTs have their plus ends facing the plasma membrane and the minus end associated with a MTOC deep in side the cell.  Kinesins will therefore generally move organelles toward the plasma-membrane and dyneins take them towards the nucleus.  A spectacular example of the is the movement of pigment granules along the MT in the chromatophore from  the skin of a variety of fish and amphibians.  Chromophores are stellate flattened cell types which contain a multitude of dense light absorbing granules.  When the granules are packed in the centre of the cell, the skin appears light, but darkens when the pigment granules are dispersed throughout the cell.  This affords the animal protective camouflage.  Granule movement requires microtubules.  Dispersion depends on cAMP or IP3/diacylglycerol, the activity of PKA or PKC, and kinesin.  Aggregation depends on phosphatase activity and it thought to be dynein dependent, although this step is very rapid at 5µm/sec and dynein driven motility is only 0.3 µm/sec in vitro.  Cellular cofactors absent from the in vitro studies may account for this difference.

Figure 9

Figure 10

 

Other membranous organelles such as the Gogli, RER, lysosomes and mitochondria are moved around the cell on microtubules.  Many of these share the same motors so it is difficult to see how the cell switches on motility of one particular organelle and not others.  Perhaps motor receptors are the key (see later), localised phosphorylation of motor protein complexes is perhaps another controlling mechanism.

 

 

Figure 11

Transport of material along the nerve axon.  Materials such as neurotransmitter peptides are synthesised in the cell body and sequestered in vesicles at the golgi.  These vesicles are then transported down the axon towards the synapse by kinesin motor proteins.  This distance may be yards in the case of a giraffe sciatic nerve!  Other materials are transported from the synapse to the cell body by dynein motors (see Nixon 1998 for a recent review of transport of cytoskeletal components)

.

Figure 12
 

Are there specific Kinesin and Dynein receptors on cargo membranes?

Recently a putative receptor “kinectin” has been identified by passing detergent-solubilized microsomal membranes over a kinesin affinity column (see Burlhardt, 1996).  Kinectin has a molecular weight of 160 kDa and the cloned cDNA reveals that the protein has a hydrophobic N-terminus and a high probability of forming coiled-coil   Kinectin has numerous putative phosphorylation sites and is known to be phosphorylated on serine, however there has been no correlation between the state of phosphorylation and kinesin binding activity to date.  Kinectin is most abundant on the ER membranes, and is not thought to bind dynein.  It is very likely that receptors other than kinectin exist in or on other membranes.  A possible candidate for a dynein receptor is “dynactin”, a 150-170 kDa protein complex (see left).  At least two mRNAs and therefore two proteins are produced from the kinectin gene in humans and this may allow a measure of specificity.  Early evidence suggests that kinectin is present of E.R. but not Golgi complex membranes.

 

The Golgi complex membrane skeleton and motor proteins.

It has very recently come to light that the Golgi complex has its own, Golgi-specific isoforms of a number of proteins involved with the actin (and microtubule to a lesser extent) cytoskeleton:-

                                   

                        ____________________________________________________________

                        Protein             Function                                  

                        ____________________________________________________________

                        Spectrin                        a linker between other membranous proteins.

                        Ankyrin                        Binds 4.1 type proteins and spectrin

                        ARP1               a component of the dynactin complex

                        Comitin                        an “annexin” actin and phospholipid binding protein

 

 

Figure 13
 

Other Golgi-specific protein isoforms are also expected to exist (Beck & Nelson, 1998), permitting the postulation of a scheme whereby dynein may be targeted to Golgi membranes (see left).  The dynactin complex is composed of a short “filament” of 9 ARP1 molecules. ARP1 is an actin related protein that is known to bind Golgi spectrin and so it is tempting to speculate that this may form the basis for a dynein receptor.  Additionally, S-100 proteins bind CapZ which it turn binds the ARP1 mini-filament, CapZ also binds phosphatidyl- inositol 4,5-bisphosphate which may also be relevant.  The binding partner for golgi specific ankyrin has not yet been identified, but there are many splice variants of Band4.1 proteins known to bind ankyrin in other membranes and there are many Band4.1 related proteins.

 

Regulation and specificity mediated by G-proteins

Many events such as fusion, and recognition of membranes is brought about by an enormous family of G-proteins, particularly the G proteins of the rab family (see Martinez & Goud 1998).  Rab6 is just one of these numerous proteins, it is restricted to Golgi and the TGN, recently, a kinesin related protein which binds to Rab6 has been identified (Rabkinesin-6; Echard et al, 1998) so it seems possible that the plus end directed kinesin-like activity already detected in the golgi, is localised and controlled by Rab6.  As Rabkinesin-6 has very limited homology to the other kinesins, it is possible that there are many of these in the genome that are also too little conserved to be recognised in EST libraries and other sequencing databases.

 

Why are organelles arranged and moved along MTs and MFs?

Organelles are required to be unequally distributed in certain situations and so motor proteins and tracks are needed to set up and maintain this distribution against chaotic influences.  Secretary vesicles, for example are required only at the synapse and so are taken there by axonal transport (by kinesin).  The golgi apparatus performs an assembly line function where proteins are processed in a linear fashion one modification taking place only after another is completed.  This process is made more efficient by having specific vesicles arranged on microtubules in the cis - trans configuration with intermediate vesicles shuttling products between them.

 

Co-operation between Microtubules and Microfilament systems in vesicular transport

In many cases there appears to be considerable overlap in the vesicular motility driven by the MT and MF systems.  Mitochondria are moved along MTs by KIF1B, and along MFs at 1.4µm/min by a myosin I-like activity (at least in yeast, see bottom of page 7 on previous handout).  Also yeast cell with disabled kinesin are rescued by overexpression of myosin I!  The sea urchin coelomocyte has been mentioned (page 6) in connection with transport of vesicles towards the cell periphery upon stimulation.  In addition to using microfilaments as track, these same cells use microtubules to transport mitochondria, in fact the motility of the mitochondria is increased 1.5 fold in the absence of actin filaments suggesting that the presence of the filaments otherwise impedes there transport possibly as a result of the increased cytosolic viscosity (Krendel et al, 1998). The two motor protein tracks, microfilaments and microtubules support seemingly exactly the same type of vesicular traffic in different cell type.  Melanin containing vesicles for example, are transported by Myosin V along microfilaments in mammalian melanoma cells while they are moved along microtubules in fish skin.  Two very similar organisms, Reticulamyxa and Laberinthula also demonstrate this duplicity.  Both organisms are large syncytical “amoeba” living in fresh and salt water respectively.  Reticulamyxa moves its nuclei and other organelles around on a vast arrays of microtubule bundles, while Laberinthula has bundles of actin to perform the same function.

 

Why have no Motor Proteins associated with Intermediate Filaments been discovered?

In order for a motor protein to do useful work, some directing influence must exist. In the case of the other two systems MT and MF, the motors are directed my the polarity of the polymers, however IFs are not polar and so it is difficult (but not impossible) to imagine how a motor could proceed along an IF in one direction  Motor proteins evolved very early in the Eukaryotes and today many protozoans express a great variety of MT and MF motor proteins, however, the IF system is much more recent (probably) being absent from the protozoan and possibly from the arthropods.  Consequently, it may be that IFs have not been around long enough for specific IF motor proteins to evolve, also this function would seem to be adequately fulfilled by MTs and MFs.

 

References:-

 

Allan, V. Membrane traffic motors. FEBS letters 369, 101 1995.

Barton, N.R. & Goldstein, L.S.B. Going mobile: Microtubule motors and chromosome segregation. Proc.Nat.Acad.Sci. USA.  March 1996.  93, 1735-1742.

Brady, S.T. A kinesin medley: biochemical and functional heterogeneity. Trends in Cell Biology 5, 159 April 1995.

Beck, K.A., Buchanan, J.A., and Nelson, W.J. (1997) Golgi membrane skeleton: identifiaction, localization and oligomerization of a 195 kDa ankyrin isoform associated with the golgi complex. J.Cell Sci. 110; 1239-1249.

Beck, K.A., and Nelson, W.J. (1998) A spectrin membrane skeleton of the golgi complex. Biochim.Biophys.Acta. 1404; 153-1160.

Burkhardt, J.K. (1996) In search of membrane receptors for microtubule-based motors - Is kinectin a kinesin receptor? Trend Cell Biol. 6; 127-131.

Echard, A., Jollivet, F., Martinez, O., Lacapere, J.-J., Rousselet, A., Janoueix-Lerosey, I., Goud, B. (1998) Interaction of a golgi-associsted kinesin-like protein with Rab6.  Science 279; 580-585.

Gee, M.A., Heuser, J.E. and Vallee, R.B. (1997) An extended microtubule-binding structure within the dynein motor domain. Nature 390; 636-639.

Goodson HV, Valetti, C & Kreis, T.E. Current Opinion in Cell Biol. Feb. 1997. 9, 18-28.

Hirokawa, N. (1998) Kinesin and dynein superfamily proteins and the mechanism of organelle transport. Science 279; 519-526.

Krendel, M., Sgourdas, G., Bonder, E.M. (1998) Disassembly of actin filaments leads to increased rate and frequency of mitochonrial movement along microfilaments. Cell Mot.Cytoskeleton. 40; 368-378.

Mocz, G. and Gibbons, I.R. (1996) Phase partition analysis of nucleotide binding to axonal dynein.  Biochemistry 35; 9204-9211.

Moore, J.D. & Endow, S.A “Kinesin proteins: a phylum of motors for microtubule- based motility” BioEssays 18 207, April 1996.

Nixon, R.A. (1998) Dynamic behaviour and organization of cytosketal proteins in neurons: reconciling old and new findings BioEssays, 20; 798-807.

Scholey, J.M. “Kinesin-II, a membrane traffic motor in axons, axonemes, and spindles.” J.Cell Biol. 1996. 133, 1-4. 

Stow, J.L. and Heimann, K. (1998) Vesicle budding on Golgi membranes: regulation by G proteins and myosin motors Biochim.Biophys.Acta. 1404; 161-171.

Weiner, O.H., Murphy, J., Griffiths, G., Schleicher, M. and Noegel, A.A. (1993). The actin-binding protein comitin (p24) is a

component of the Golgi apparatus. J.Cell Biol. 123; 23-34.