Cell polarity and domains

Epithelia (including endothelia and mesothelia) are organized into sheets or more complex structures (see Ch. 2) with very different environments on either side. These cells actively transfer macromolecules and ions between the two surfaces and are thus polarized in structure and function (Fig. 1.3). In polarized cells, particularly in epithelia, the cell is generally subdivided into domains that reflect the polarization of activities within it. The free surface, e.g. that facing the intestinal lumen or airway, is the apical surface, and its adjacent cytoplasm is the apical cell domain. This is where the cell interfaces with a specific body compartment (or, in the case of the epidermis, with the outside world). The apical surface is specialized to act as a barrier, restricting access of substances from this compartment to the rest of the body. Specific components are selectively absorbed from, or added to, the external compartment by the active processes, respectively, of active transport and endocytosis inwardly or exocytosis and secretion outwardly. The apical surface is often covered with small protrusions of the cell surface, microvilli, which increase the surface area, particularly for absorption.

Fig. 1.3 Absorptive epithelium from human colon, immunolabelled to show the functional division of cell surface domains. CD66 antigen expression (red) is concentrated in the apical domain facing the gut lumen (top), on striated border microvilli; diffuse labelling is present throughout the cytoplasm but is excluded from the basal zone occupied by nuclei. The intestinal cell antigen A33 (green) is a member of the immunoglobulin superfamily of transmembrane proteins, expressed here on basolateral cell surfaces.

(By courtesy of Cécile Chalouni, Ludwig Institute for Cancer Research, Yale University School of Medicine.)

The surface of the cell opposite to the apical surface is the basal surface, with its associated basal cell domain. In a single-layered epithelium, this surface is apposed to the basal lamina. The remaining surfaces are known as the lateral cell surfaces. In many instances the lateral and basal surfaces perform similar functions and the cellular domain is termed the basolateral domain. Cells actively transport substances, such as digested nutrients from the intestinal lumen or endocrine secretions, across their basal (or basolateral) surfaces into the subjacent connective tissue matrix and the blood capillaries within it. Dissolved non-polar gases (oxygen and carbon dioxide) diffuse freely between the cell and the bloodstream across the basolateral surface. Apical and basolateral surfaces are separated by a tight intercellular seal, the tight junction, which prevents the passage of even small ions through the space between adjacent cells and thus maintains the difference between environments either side of the epithelium.

The cell coat (glycocalyx)

The plasma membrane differs structurally from internal membranes in that it possesses an external, diffuse, carbohydrate-rich coat, the cell coat or glycocalyx. The cell coat forms an integral part of the plasma membrane, projecting as a diffusely filamentous layer 2–20 nm or more from the lipoprotein surface (see Fig. 1.5). The overall thickness of the plasma membrane is therefore variable, but is typically 8–10 nm. The cell coat is composed of the carbohydrate portions of glycoproteins and glycolipids embedded in the plasma membrane (Fig. 1.4).

Fig. 1.5 Intercellular junctions. The positions of the apical junctional complex and other junctional specializations of epithelial cells are shown, along with their electron microscopic appearances (see A–D; B and C are freeze fractured preparations). B shows how, in a gap junction, numerous channels (pores within connexons) are clustered to form a plaque-like junctional region between adjacent plasma membranes. C shows the anastomotic network of contacts between adjacent cell membranes forming a tight junction. D shows the attachments of basal plasma membrane to basal lamina at hemidesmosomes.

(Part B by courtesy of Professor Dieter Hülser, University of Stuttgart. Part C by courtesy of Dr Andrew Kent, King’s College London.) A,D from human tissue.

The precise composition of the glycocalyx varies with cell type: many tissue and cell type-specific antigens are located in the coat, including the major histocompatibility antigen systems and, in the case of erythrocytes, blood group antigens. It also contains adhesion molecules, which enable cells to adhere selectively to other cells or to the extracellular matrix. They have important roles in maintaining the integrity of tissues and in a wide range of dynamic cellular processes, e.g. the formation of intercommunicating neural networks in the developing nervous system and the extravasation of leukocytes. Cells tend to repel each other because of the predominance of negatively charged carbohydrates at cell surfaces. There is consequently a distance of at least 20 nm between the plasma membranes of adjacent cells, other than at specialized junctions.

Cell surface contacts

The plasma membrane is the surface which establishes contact with other cells and with structural components of extracellular matrices. These contacts may have a predominantly adhesive role, or initiate instructive signals within and between cells, or both; they frequently affect the behaviour of cells. Structurally, there are two main classes of contact, both associated with cell adhesion molecules. One class is associated with specializations at discrete regions of the cell surface that are ultrastructurally distinct. These are described on page 6. The second, general, class of adhesive contact has no obvious associated ultrastructural features.

Cell signalling

Cellular systems in the body communicate with each other to coordinate and integrate their functions. This occurs through a variety of processes known collectively as cell signalling, in which a signalling molecule produced by one cell is detected by another, almost always by means of a specific receptor protein molecule. The recipient cell transduces the signal, which it most usually detects at the plasma membrane, into intracellular chemical messages that change cell behaviour.

The signal may act over a long distance, e.g. endocrine signalling through the release of hormones into the bloodstream, or neuronal synaptic signalling via electrical impulse transmission along axons and subsequent release of chemical transmitters of the signal at synapses (p. 44) or neuromuscular junctions (p. 62). A specialized variation of endocrine signalling (neurocrine or neuroendocrine signalling) occurs when neurones or paraneurones (e.g. chromaffin cells of the suprarenal medulla) secrete a hormone into interstitial fluid and the bloodstream.

Alternatively, signalling may occur at short range through a paracrine mechanism, in which cells of one type release molecules into the interstitial fluid of the local environment, to be detected by nearby cells of a different type that express the specific receptor protein. Neurocrine cell signalling uses chemical messengers found also in the central nervous system which may act in a paracrine manner via interstitial fluid or reach more distant target tissues in the bloodstream. Cells may generate and respond to the same signal. This is autocrine signalling, a phenomenon that reinforces the coordinated activities of a group of like cells, which respond together to a high concentration of a local signalling molecule. The most extreme form of short-distance signalling is contact-dependent (juxtacrine) signalling, where one cell responds to transmembrane proteins of an adjacent cell that bind to surface receptors in the responding cell membrane. Contact-dependent signalling also includes cellular responses to integrins on its surface binding to elements of the extracellular matrix. Juxtacrine signalling is important during development and in immune responses.

These different types of intercellular signalling mechanism are illustrated in Figure 1.6. For further reading, see Alberts et al (2002) and Pollard & Earnshaw (2007).

Fig. 1.6 The different modes of cell–cell signalling.

Transport across cell membranes

Lipid bilayers are increasingly impermeable to molecules as they increase in size or hydrophilicity. Transport mechanisms are therefore required to carry essential polar molecules, including ions, nutrients, nucleotides and metabolites of various kinds, across the plasma membrane and into or out of membrane-bound intracellular compartments. Transport is facilitated by a variety of membrane transport proteins, each with specificity for a particular class of molecule, e.g. sugars. Transport proteins fall mainly into two major classes, channel proteins and carrier proteins.

Channel proteins form aqueous pores in the membrane, which open and close under the regulation of intracellular signals, e.g. G-proteins, to allow the flux of solutes (usually inorganic ions) of specific size and charge. Transport through ion channels is always passive and ion flow through an open channel depends only on the ion concentration gradient and its electronic charge, and the potential difference across the membrane. These factors combine to produce an electrochemical gradient, which governs ion flux. Channel proteins are utilized most effectively by the excitable plasma membranes of nerve cells, where the resting membrane potential can change transiently from about −80 mV (negative inside the cell) to +40 mV (positive inside the cell) when stimulated by a neurotransmitter (as a result of the opening and subsequent closure of channels selectively permeable to sodium and potassium).

Carrier proteins bind their specific solutes, such as amino-acids, and transport them across the membrane through a series of conformational changes. This latter process is slower than ion transport through membrane channels. Transport by carrier proteins can occur either passively by simple diffusion, or actively against the electrochemical gradient of the solute. Active transport must therefore be coupled to a source of energy, such as ATP generation, or energy released by the coordinate movement of an ion down its electrochemical gradient. Linked transport can be in the same direction as the solute, in which case the carrier protein is described as a symporter, or in the opposite direction, when the carrier acts as an antiporter.

Exocytosis and endocytosis

Secreted proteins, lipids, mucins, small molecules such as amines and other cellular products destined for export from the cell are transported to the plasma membrane in small vesicles released from the trans face of the Golgi apparatus. This pathway is either constitutive, in which transport and secretion occur more or less continuously, or it is regulated by external signals, as in the control of salivary secretion by autonomic neural stimulation. In regulated secretion, the secretory product is stored temporarily in membrane-bound secretory granules or vesicles. Exocytosis is achieved by fusion of the secretory vesicular membrane with the plasma membrane and release of the vesicle contents into the extracellular domain.

In polarized cells, e.g. most epithelia, exocytosis occurs at the apical plasma membrane and the cells secrete into a duct lumen or onto a free surface such as the lining of the stomach. In hepatocytes, bile is secreted across a very small area of plasma membrane forming the wall of the bile canaliculus. This region is defined as the apical plasma membrane, and is the site of exocrine secretion, whereas secretion of hepatocyte plasma proteins into the bloodstream is targeted to the basolateral surfaces facing the sinusoids. Packaging of different secretory products into appropriate vesicles takes place in the trans-Golgi network. Delivery of secretory vesicles to their correct plasma membrane domains is achieved by sorting sequences in the cytoplasmic tails of vesicular membrane proteins.

There are other mechanisms in which initial delivery of secretory products is less selective, but is followed by selective retention (or degradation) or reprocessing and redistribution by endosomes. Ultimately, secretory vesicles undergo docking, priming (to prepare the vesicle for a regulatory signal, where secretion is regulation-dependent) and fusion with the plasma membrane to release their contents. The process of exocytosis also delivers integral membrane components to the cell surface in the normal turnover and recycling of the plasma membrane. However, excess plasma membrane generated by vesicle fusion during exocytosis is rapidly removed by concurrent endocytosis.

The process of endocytosis involves the internalization of vesicles derived from the plasma membrane. The vesicles may contain: engulfed fluids and solutes from the extracellular interstitial fluid (pinocytosis); larger macromolecules, often bound to surface receptors (receptor-mediated endocytosis); particulate matter, including microorganisms or cellular debris (phagocytosis). Pinocytosis generally involves small fluid-filled vesicles and is a marked property of capillary endothelium, e.g. where vesicles containing nutrients and oxygen dissolved in blood plasma are transported from the vascular lumen to the endothelial basal plasma membrane (see Fig. 6.11). Interstitial fluid containing dissolved carbon dioxide is also taken up by pinocytosis for simultaneous transportation across the endothelial cell wall in the opposite direction, for release into the bloodstream by exocytosis. This shuttling of pinocytotic vesicles is also termed transcytosis. Larger volumes of fluid are engulfed by dendritic cells, e.g. in the process of sampling interstitial fluids by macropinocytosis in immune surveillance for antigens (p. 79). Interstitial fluid is inevitably taken up during receptor-mediated endocytosis when ligands are internalized.

Receptor-mediated endocytosis, also known as clathrin-dependent endocytosis, is initiated at specialized regions of the plasma membrane known as clathrin-coated pits. Clathrin is a protein that cross-links adjacent adaptor protein (adaptin) complexes to form a basket-like structure, bending the membrane inwards into a hemisphere. Much, but not all, fluid-phase pinocytosis also utilizes clathrin-coated pits. Ligands such as the iron-transporting protein, transferrin, and the cholesterol-transporting low-density lipoprotein bind to their receptors, which cluster in clathrin-coated pits through an interaction with adaptins. The pits then invaginate and pinch off from the plasma membrane, internalizing both receptor and ligand. The processing of endocytic vesicles and their contents is described on p. 12. For further details of the molecular mechanisms of endocytosis, see Alberts et al (2002) or Pollard & Earnshaw (2007).

Compartments and functional organization

The cytoplasm is highly concentrated, with about 200 mg/ml of proteins (about twice the concentration in blood) that must be precisely organized for correct molecular interactions to occur. It normally has extremely low levels of Ca²⁺, high K⁺ and low Na⁺ ions in comparison to extracellular fluid, differences which are important in cell signalling. Cytoplasm is also reductive, a state maintained by a high concentration of thiol-containing glutathione. The cell is able to undertake completely opposite reactions simultaneously (e.g. the synthesis and degradation of proteins; growth at one end of a cell with retraction at another) by partitioning them into different regions of the cytoplasm. The most fundamental divide is the use of oxidative reactions within the reductive cytoplasm, achieved by the compartmentalization of different environments within membranes. For example, the endoplasmic reticulum is comprised of stacks of tubules, whose lumen resembles the extracellular environment in being oxidative and Ca²⁺-rich, predominantly encrusted on the external (cytoplasmic) face with attached ribosomes (rough endoplasmic reticulum). Ribosomes are macromolecular machines for protein synthesis and those attached to RER are engaged in synthesizing proteins that will undergo post-translational modification to adapt them (in the RER lumen and within the Golgi apparatus and associated vesicles) for exposure to the oxidative extracellular environment. A key development in the evolution of the modern cell was the ability to use oxygen as an energy source. This is possible due to one organelle, the mitochondrion, whose separate genes and dual membranes suggest its origins as a symbiotic bacterium.

The key organizer of the cytoplasm, and thence of the entire cell, is the cytoskeleton (see below). This is composed of three distinct elements. Intermediate filaments are relatively stable cables of approximately 10 nm diameter that provide strength. Actin microfilaments (6–8 nm diameter) form highly branched scaffolds under the cell surface, organizing the shape of the cell surface and its specialized functions, including extracellular interactions such as signalling and adhesion, by binding to the intracellular domains of receptors and adhesive proteins, respectively. The actin scaffold under the cell surface is highly labile, forming, branching and dissolving in response to extracellular signals. Specialized myosin-family motor proteins attach to actin filaments, generating force to move membranes and to relay vesicles between the surface and the tubulin network. The tubulin network, however, is the core organizer of the cell; it is polarized, which allows motor proteins to move directionally along the tubules and convey vesicular traffic around the cell.

Endoplasmic reticulum

Endoplasmic reticulum is a system of interconnecting membrane-lined channels within the cytoplasm (Fig. 1.7). These channels take various forms, including cisternae (flattened sacs), tubules and vesicles. The membranes divide the cytoplasm into two major compartments. The intramembranous compartment includes the space where secretory products are stored or transported to the Golgi complex and cell exterior. The extramembranous cytosol is made up of the colloidal proteins such as enzymes, carbohydrates and small protein molecules, together with ribosomes and ribonucleic acids, and elements of the cytoskeleton.

Fig. 1.7 Smooth endoplasmic reticulum with associated vesicles. The dense particles are glycogen granules.

(By courtesy of Rose Watson, Cancer Research UK.)

Structurally, the channel system can be divided into rough or granular endoplasmic reticulum, which has ribosomes attached to its outer cytosolic surface, and smooth or agranular endoplasmic reticulum, which lacks ribosomes. Functionally, the endoplasmic reticulum is compartmentalized into specialized regions with unique functions. For further reading see Levine & Rabouille (2005).

Rough endoplasmic reticulum

The rough endoplasmic reticulum, studded with ribosomes, is a site of protein synthesis (Fig. 1.8). Most proteins pass through its membranes and accumulate within its cisternae, although some integral membrane proteins, e.g. plasma membrane receptors, are inserted into the rough endoplasmic reticulum membrane, where they remain. After passage from the rough endoplasmic reticulum, proteins remain in membrane-bound cytoplasmic organelles such as lysosomes, become incorporated into new plasma membrane, or are secreted by the cell. Some carbohydrates are also synthesized by enzymes within the cavities of the rough endoplasmic reticulum and may be attached to newly formed protein (glycosylation). Vesicles are budded off from the rough endoplasmic reticulum for transport to the Golgi as part of the protein-targeting mechanism of the cell.

Fig. 1.8 The Golgi apparatus and functionally related organelles. A, Golgi apparatus (G) adjacent to the nucleus (N) (V = vesicle). B, Large residual body (tertiary lysosome) in a cardiac muscle cell (M = mitochondrion). C, Functional relationships between the Golgi apparatus and associated cellular structures. D, 3-dimensional reconstruction of the Golgi apparatus in a pancreatic beta cell showing stacks of Golgi cisternae from the cis-face (pink), cis-medial cisternae (red, green), to the trans-Golgi network (blue, yellow, orange-red); immature proinsulin granules (condensing vesicles) shown in pale blue and mature (crystalline) insulin granules in dark blue. The flat colour areas represent cut faces of cisternae and vesicles; E, Rough endoplasmic reticulum (R), associated with the Golgi apparatus (G).

(Part D by courtesy of Dr Brad Marsh, Institute for Molecular Bioscience, University of Queensland, Brisbane.) A,B,E from human tissue.

Ribosomes

Ribosomes are macromolecular machines that catalyse the synthesis of proteins from amino-acids. They are granules approximately 15 nm in diameter, composed of ribosomal RNA (rRNA) molecules assembled into two unequal subunits. A large number of proteins, mostly small and basic, are applied mainly to the surfaces of the subunit cores of RNA. The subunits can be separated by their sedimentation coefficients (S) in an ultracentrifuge, into larger 60S and smaller 40S components. These are associated with 73 different proteins (40 in the large subunit and 33 in the small), which have structural and enzymatic functions. Three small, highly convoluted rRNA strands (28S, 5.8S and 5S) make up the large subunit, and one strand (18S) is in the small subunit. Their synthesis and assembly into subunits takes place in the nucleolus, and includes association with ribosomal proteins translocated from their site of synthesis in the cytoplasm. The individual subunits are then transported into the cytoplasm, where they remain separate from each other when not actively synthesizing proteins.

A typical cell contains millions of ribosomes. They may be solitary, relatively inactive structures, or may form groups (polyribosomes or polysomes) attached to messenger RNA (mRNA), which they translate during protein synthesis. Polysomes may be attached to the membranes of the rough endoplasmic reticulum (see Fig. 1.8) or may lie free in the cytosol, where they synthesize proteins for use outside the system of membrane compartments, including enzymes of the cytosol and cytoskeletal proteins. Some of the cytosolic products include proteins that can be inserted directly into (or through) membranes of selected organelles, such as mitochondria and peroxisomes.

In a mature polysome, all the attachment sites of the mRNA are occupied as ribosomes move along it, synthesizing protein according to its nucleic acid sequence. Consequently, the number of ribosomes in a polysome indicates the length of the mRNA molecule and hence the size of the protein being made. The two subunits have separate roles in protein synthesis. The 40S subunit is the site of attachment and translation of mRNA. The 60S subunit is responsible for the release of the new protein and, where appropriate, attachment to the endoplasmic reticulum via an intermediate docking protein that directs the newly synthesized protein through the membrane into the cisternal space.

Protein synthesis on ribosomes may be suppressed by a class of RNA molecule known as small interfering RNA (siRNA) or silencing RNA. These molecules are typically 20–25 nucleotides in length and bind (as a complex with proteins) to specific mRNA molecules via their complementary sequence. This triggers the enzymatic destruction of the mRNA or prevents the movement of ribosomes along it. Synthesis of the encoded protein is thus prevented. Their normal function may have anti-viral or other protective effects; there is also potential for developing artificial siRNAs as a therapeutic tool for silencing disease-related genes.

Golgi apparatus (Golgi complex)

The Golgi apparatus is a distinct cytoplasmic region near the nucleus, and is particularly prominent in secretory cells when stained with silver or other metallic salts. The Golgi apparatus forms part of the pathway by which proteins synthesized in the rough endoplasmic reticulum undergo post-translational modification and are targeted to the cell surface for secretion or for storage in membranous vesicles. As with the endoplasmic reticulum, the Golgi apparatus is compartmentalized spatially, in a labile manner, to carry out specific functions.

Ultrastructurally, the Golgi apparatus is a membranous organelle (Fig. 1.8) consisting of a stack of several flattened membranous cisternae, together with clusters of vesicles surrounding its surfaces. Seen in vertical section, it is often cup-shaped. Small transport vesicles from the rough endoplasmic reticulum, generated by a process of budding and pinching off, are received at one face of the Golgi stack, the convex cis-face (entry or forming surface). Here, they deliver their contents to the first cisterna in the series by membrane fusion. From the edges of this cisterna, the protein is transported to the next cisterna by vesicular budding and then fusion, and this process is repeated until the final cisterna at the concave trans face (exit or condensing surface) is reached. Here, larger vesicles are formed for delivery to other parts of the cell.

In addition to these cisternae, there are other membranous structures that form an integral part of the Golgi apparatus, termed the cis-Golgi and trans-Golgi networks. The cis-Golgi network is a region of complex membranous channels interposed between the rough endoplasmic reticulum and the Golgi cis face (Golgi–rough endoplasmic reticulum complex), which receives and transmits vesicles in both directions. Its function is to select appropriate proteins synthesized on the rough endoplasmic reticulum for delivery by vesicles to the Golgi stack, while inappropriate proteins are shuttled back to the rough endoplasmic reticulum.

The trans-Golgi network, at the other side of the Golgi stack, is also a region of interconnected membrane channels engaged in protein sorting. Here, modified proteins processed in the Golgi cisternae are packaged selectively into vesicles and dispatched to different parts of the cell. The packaging depends on the detection, by the trans-Golgi network, of particular amino-acid signal sequences, leading to their enclosure in membranes of appropriate composition that will further modify their contents, e.g. by extracting water to concentrate them or by pumping in protons to acidify their contents. The membranes contain specific signal proteins, which may allocate them to microtubule-based transport pathways and allow them to dock with appropriate targets elsewhere in the cell, e.g. the plasma membrane in the case of secretory vesicles. Vesicle formation and budding at the trans-Golgi network involves the addition of clathrin on their external surface, to form coated pits.

Within the Golgi stack proper, proteins undergo a series of sequential chemical modifications that started in the rough endoplasmic reticulum. These include: changes in glycosyl groups, e.g. removal of mannose, addition of N-acetyl glucosamine and sialic acid; sulphation of attached glycosaminoglycans; protein phosphorylation. Lipids formed in the endoplasmic reticulum are also routed for incorporation into vesicles.

The role of the Golgi apparatus in the synthesis of primary lysosomes is a major activity in cells with abundant lysosomes, such as those with phagocytic roles. In glandular cells with an apical secretory zone, the Golgi apparatus lies between the secretory surface and the nucleus. In fibroblasts, there are two or more groups of Golgi stacks; up to 50 groups are found in liver cells. The Golgi apparatus is often closely associated with the centrosome (a region of the cell containing a centriole pair and related microtubules), reflecting a link with the microtubule-mediated vesicle transport system.

Endosomes, lysosomes, proteasomes and peroxisomes

The endosome system of vesicles originates in small endocytic vesicles (clathrin-coated vesicles and caveolae) or larger phagosomes and macropinocytic vesicles taken up by the cell from the exterior. Clathrin-dependent endocytosis occurs at specialized patches of plasma membrane called coated pits; this mechanism is also used to internalize ligands bound to surface receptor molecules and is also termed receptor-mediated endocytosis. Caveolae (little caves) are structurally distinct vesicles most widely used by endothelial and smooth muscle cells, where they are involved in transcytosis, signal transduction and possibly other functions. For further reading, see Pollard & Earnshaw (2007).

The endocytic system is linked functionally to a second series of membranous structures, the lysosomes. Lysosomes contain acid hydrolases, which process or degrade exogenous materials (heterophagy), and intracellular organelles that are exhausted, damaged or no longer required (autophagy). There is a continual exchange of vesicles between this system and the Golgi–rough endoplasmic reticulum complex, so that the endosomal/lysosomal system is provided with hydrolytic enzymes and the Golgi receives depleted vesicles for recharging. Once internalized, endocytic vesicles shed their coat of adaptin and clathrin, and fuse with a tubular cisterna termed an early endosome, where the receptor molecules release their bound ligands. Membrane and receptors from the early endosomes can be recycled to the cell surface as exocytic vesicles.

Lysosomes

Lysosomes are dense, spheroidal, membrane-bound bodies 80–800 nm in diameter (Fig. 1.8, Fig. 1.9), often with complex inclusions of material undergoing hydrolysis (secondary lysosomes). They contain acid hydrolases able to degrade a wide variety of substances. To date, more than 40 lysosomal enzymes have been described, including proteases, lipases, carbohydrases, esterases and nucleases. The enzymes are heavily glycosylated, and are maintained at a low pH by proton pumps in the lysosomal membranes.

Fig. 1.9 A, Mitochondria in human cardiac muscle. The folded cristae (arrows) project into the matrix from the inner mitochondrial membrane. B, The location of the elementary particles which couple oxidation and phosphorylation reactions.

Lysosomes are numerous in actively phagocytic cells, e.g. macrophages and neutrophil granulocytes, in which lysosomes are responsible for destroying phagocytosed bacteria. In these cells, the phagosome containing the bacterium may fuse with several lysosomes. Lysosomes are also frequent in cells with a high turnover of organelles, e.g. exocrine gland cells and neurones. Effete organelles are targeted for demolition by a process that is not fully understood, but which results in engulfment of areas of cytoplasm, including entire organelles, in a membranous cisterna. The structure then fuses with lysosomes and the contents are rapidly degraded.

Material that has been hydrolysed within late endosomes and lysosomes may be completely degraded to soluble products, e.g. amino-acids, which are recycled through metabolic pathways. However degradation is usually incomplete, and some debris remains. A debris-laden vesicle is called a residual body or tertiary lysosome (Fig. 1.8B), and may be passed to the cell surface, where it is ejected by exocytosis; alternatively, it may persist inside the cell as an inert residual body. Considerable numbers of residual bodies can accumulate in long-lived cells, often fusing to form larger dense vacuoles with complex lamellar inclusions. As their contents are often darkly pigmented, this may change the colour of the tissue, e.g. in neurones the end-product of lysosomal digestion, lipofuscin (neuromelanin or senility pigment), gives ageing brains a brownish-yellow colouration.

Lysosomal enzymes may also be secreted – often as part of a process to alter the extracellular matrix, as in osteoclast erosion of bone (p. 88). Abnormal release of enzymes can cause tissue damage, as in certain types of arthritis. Some drugs, e.g. cortisone, can stabilize lysosomal membranes and may therefore inhibit many lysosomal activities, including the secretion of enzymes, and their fusion with phagocytic vesicles.

Lysosomal storage diseases

If any of the lysosomal enzymes are defective because of gene mutations, the materials that they normally degrade will accumulate within late endosomes and lysosomes. Many such lysosomal storage diseases are known, e.g. Tay–Sachs disease, in which a faulty hexosaminidase leads to the accumulation of ganglioside in neurones, causing death during childhood. In Hurler’s syndrome, failure to metabolize certain mucopolysaccharides causes the accumulation of large amounts of matrix within connective tissue, which distorts growth of many parts of the body.

Mitochondria

Mitochondria are membrane-bound organelles (Fig. 1.9). They are the principal source of chemical energy in most cells. Mitochondria are the site of the citric acid (Krebs’, tricarboxylic acid) cycle and the electron transport (cytochrome) pathway by which complex organic molecules are finally oxidized to carbon dioxide and water. This process provides the energy to drive the production of ATP from ADP and inorganic phosphate (oxidative phosphorylation). The various enzymes of the citric acid cycle are located in the mitochondrial matrix, whereas those of the cytochrome system and oxidative phosphorylation are localized chiefly in the inner mitochondrial membrane. It is now known that in many tissues, especially smooth muscle, mitochondria also play an important role in cell signalling, especially intracellular calcium homeostasis. They are also major producers of reactive oxygen species and oxidant stress, and are involved in activation of apoptosis.

The numbers of mitochondria in a particular cell reflect its general energy requirements; e.g. in hepatocytes there may be as many as 2000, whereas in resting lymphocytes there are usually very few. Mature erythrocytes lack mitochondria altogether. Cells with few mitochondria generally rely largely on glycolysis for their energy supplies. These include some very active cells, e.g. fast twitch skeletal muscle fibres, which are able to work rapidly, but for only a limited duration. Mitochondria appear in the light microscope as long thin threads in the cytoplasm of most cells, particularly those with a high metabolic rate, e.g. secretory cells in exocrine glands. In living cells, mitochondria constantly change shape and intracellular position; they multiply by growth and fission and may undergo fusion.

In the electron microscope, mitochondria usually appear as round or elliptical bodies 0.5–2.0 μm long. Each mitochondrion is lined by an outer and an inner unit membrane, separated by a variable gap termed the intermembrane space. The lumen is surrounded by the inner membrane and contains the mitochondrial matrix. The outer membrane is smooth and sometimes attached to other organelles, particularly microtubules. The inner membrane is deeply folded to form incomplete transverse or longitudinal tubular invaginations, cristae, which create a relatively large surface area of membrane. Mitochondrial shape, and the shape and organization of the cristae, vary with the cell type. Cristae are most numerous and complex in cells with a high metabolic rate, e.g. cardiac muscle cells. The permeabilities of the two mitochondrial membranes differ considerably: the outer membrane is freely permeable to many substances because of the presence of large non-specific channels formed by proteins (porins), whereas the inner membrane is permeable to only a narrow range of molecules. The presence of cardiolipin, a phospholipid, in the inner membrane may contribute to this relative impermeability.

The mitochondrial matrix is an aqueous environment. It contains a variety of enzymes, and strands of mitochondrial DNA with the capacity for transcription and translation of a unique set of mitochondrial genes (mitochondrial mRNAs and transfer RNAs, mitochondrial ribosomes with rRNAs). The DNA forms a closed loop, about 5 μm across; several identical copies are present in each mitochondrion. The ratio between its bases differs from that of nuclear DNA, and the RNA sequences also differ in the precise genetic code used in protein synthesis. At least 13 respiratory chain enzymes of the matrix and inner membrane are encoded by the small number of genes along the mitochondrial DNA. The great majority of mitochondrial proteins are encoded by nuclear genes and made in the cytosol, then inserted through special channels in the mitochondrial membranes to reach their destinations. Their membrane lipids are synthesized in the endoplasmic reticulum.

Mitochondrial ribosomes are smaller and quite distinct from those of the rest of the cell; they (and mitochondrial nucleic acids) resemble those of bacteria. This similarity underpins the theory that mitochondrial ancestors were oxygen-utilizing bacteria that existed in a symbiotic relationship with eukaryotic cells unable to metabolize the oxygen produced by early plants. As mitochondria are formed only from previously existing ones, it follows that all mitochondria in the body are descended from those in the cytoplasm of the fertilized ovum. It has also been shown that mitochondria are of maternal origin because the mitochondria of the sperm are not generally incorporated into the ovum at fertilization. Thus mitochondria (and mitochondrial genetic variations and mutations) are passed only through the female line.

Mitochondria are distributed within a cell according to regional energy requirements, e.g. near the bases of cilia in ciliated epithelia, in the basal domain of the cells of proximal convoluted tubules in the renal cortex (where considerable active transport occurs) and around the proximal end of the flagellum in spermatozoa. They may be involved with tissue-specific metabolic reactions, e.g. various urea-forming enzymes in liver cell mitochondria. Moreover, a number of genetic diseases of mitochondria affect particular tissues exclusively, e.g. mitochondrial myopathies (skeletal muscle) and mitochondrial neuropathies (nervous tissue). For further information see Graff et al (2002).

Cytosolic organelles

The aqueous cytosol surrounds the membranous organelles described above. It also contains various non-membranous organelles, including free ribosomes, a system of filamentous proteins known as the cytoskeleton, and other inclusions, such as storage granules (e.g. glycogen) and lipid vacuoles.

Cytoskeleton

The cytoskeleton is a system of filamentous intracellular proteins of different shapes and sizes that form a complex, often interconnected, network throughout the cytoplasm. It provides mechanical support, maintains cell shape and rigidity, and enables cells to adopt highly asymmetric or irregular profiles, e.g. in neurones. The cytoskeleton plays an important part in establishing structural polarity and different functional domains within a cell. It also provides mechanical support for projections from the cell surface such as microvilli and cilia, and anchors them into the cytoplasm.

The cytoskeleton restricts specific organelles to particular cellular locations, e.g. the Golgi apparatus is near the nucleus and endoplasmic reticulum, and mitochondria are near sites of energy requirement. Most specifically, the cytoskeleton is concerned with motility, either within the cell (e.g. shuttling vesicles and macromolecules between cytoplasmic sites, or the movement of chromosomes during mitosis), or of the entire cell (e.g. in embryonic morphogenesis or the chemotactic migration of leukocytes). One of the most highly developed and specialized functions of the cytoskeleton is seen in the contractility of muscle cells.

The catalogue of cytoskeletal structural proteins is extensive and still increasing. The major filamentous structures found in non-muscle cells are microfilaments (actin), microtubules (tubulin), and intermediate filaments (assemblies of cell type-specific intermediate filament proteins). Other important components are proteins that bind to the principal filamentous types to link them together or to generate movement. These include actin-binding proteins such as myosin, which in some cells can assemble into thick filaments, and microtubule-associated proteins. Pathologies involving cytoskeletal abnormalities are reviewed in Ramaekers & Bosman (2004).

Actin filaments (microfilaments)

Actin filaments are flexible filaments with a width of 6–8 nm (Fig. 1.10), and a solid cross-section. Within most cell types, actin constitutes the most abundant protein and in some motile cells its concentration may exceed 200 μM (10 mg protein per ml cytoplasm). The filaments are formed by the ATP-dependent polymerization of actin monomer (with a molecular mass of 43 kDa) into a characteristic linear form in which the subunits are arranged in a single tight helix with a distance of 13 subunits between turns. The polymerized form is termed F-actin (fibrillar actin) and the unpolymerized form is G-actin (globular actin). Each monomer has an asymmetric structure. When monomers polymerize, they confer a defined polarity on the filament: the plus end favours monomer addition, and the minus end favours monomer dissociation. Myosins bind to filamentous actin at an angle to give the appearance of a series of arrowheads pointing towards the minus end of the filament, and the barbs point towards the plus end. There is a dynamic equilibrium between G-actin and F-actin: in most cells about 50% of the actin is estimated to be in the polymerized state.

Fig. 1.10 The cytoskeleton. A, Immunofluorescence micrograph of α actin microfilaments (green) in human airway smooth muscle cells in culture. The actin-binding protein, vinculin (red), is localized at the ends of actin filament bundles; nuclei are blue. B, Immunofluorescence micrograph of keratin intermediate filaments (green) in human keratinocytes in culture. Desmosome junctions are labelled with antibody against desmoplakin (red). Nuclei are stained blue (Hoechst). By courtesy of Prof. Dr. W.W. Franke, German Cancer Research Centre, Heidelberg. C, Electron micrograph of human nerve showing microtubules (small, hollow structures in cross-section, long arrow) in a transverse section of an axon (A), engulfed by a non-myelinating Schwann cell (S). Neuronal intermediate filaments (neurofilaments) are the solid, electron-dense profiles, also in transverse section (short arrow).

(By courtesy of Dr T Nguyen, Professor J Ward, Dr SJ Hirst, Kings College London.)

Actin-binding proteins

A wide variety of actin-binding proteins are capable of modulating the form of actin within the cell. These interactions are fundamental to the organization of cytoplasm and to cell shape. Actin-binding proteins can be grouped into bundling proteins, gel-forming proteins and filament severing proteins.

Bundling proteins tie actin filaments together in longitudinal arrays to form cables or core structures. The bundles may be closely spaced, e.g. in microvilli, microspikes and filopodia, where parallel filaments are tied tightly together to form stiff bundles orientated in the same direction. Proteins with this function include fimbrin and villin (also classified as a severing protein). Other actin-bundling proteins form rather looser bundles of filaments that run anti-parallel to each other with respect to their plus and minus ends. They include myosin II, which can form cross-links with ATP-dependent motor activity, and cause adjacent actin filaments to slide on each other, and either change the shape of cells or (if the actin bundles are anchored into the cell membrane at both ends), maintain a degree of active rigidity.

Gel-forming proteins, such as filamin, interconnect adjacent actin filaments to produce loose filamentous meshworks (gels) composed of randomly orientated F-actin. These networks are frequently found in the outer cortical regions of cells, e.g. fibroblasts. They form a semi-rigid zone from which most other organelles are excluded. Severing proteins, such as gelsolin and severin, bind to F-actin filaments and sever them, which produces profound changes within the actin cytoskeleton and in its coupling to the cell surface.

Other classes of actin-binding protein link the actin cytoskeleton to the plasma membrane either directly or indirectly through a variety of membrane-associated proteins. The latter may also create links via transmembrane proteins to the extracellular matrix. Best known of these is the family of spectrin-like molecules, which can bind to actin and also to each other and various membrane-associated proteins to create supportive networks beneath the plasma membrane. Defects in such molecules are linked to a number of inherited diseases (reviewed in Bennett & Healy 2008). Spectrin is found in erythrocytes, and closely related molecules are present in many other cells; for instance, fodrin is found in nerve cells, and dystrophin occurs in muscle cells, linking the contractile apparatus with the extracellular matrix via integral membrane proteins. Proteins such as ankyrin (which also binds actin directly), vinculin, talin, zyxin and paxillin connect actin-binding proteins to integral plasma membrane proteins such as integrins (directly or indirectly), and thence to focal adhesions. Myosin I and other unconventional myosins connect actin filaments to membranous structures, including the plasma membrane and transport vesicle membranes. Tropomyosin, an important regulatory protein of muscle fibres, is also present in non-muscle cells, where its function may be primarily to stabilize actin filaments against depolymerization. For further reading see Pollard & Earnshaw (2007).

Myosins – the motor proteins

The myosin family of microfilaments is often classified within a distinct category of motor proteins. Myosin proteins have a globular head region consisting of a heavy and a light chain. The heavy chain bears an α-helical tail of varying length. The head has an ATPase activity and can bind to and move along actin filaments – the basis for myosin function as a motor protein. The best-known class is myosin II, which occurs in muscle and in many non-muscle cells. Its molecules have two heads and two tails, intertwined to form a long rod. The rods can bind to each other to form long, thick filaments, as seen in striated and smooth muscle fibres, myoepithelial cells and myofibroblasts. Myosin II molecules can also assemble into smaller groups, especially dimers, which can cross-link individual actin microfilaments in stress fibres and other F-actin arrays. The ATP-dependent sliding of myosin on actin forms the basis for muscle contraction and the extension of microfilament bundles, as seen in cellular motility or in the contraction of the ring of actin and myosin around the cleavage furrow of dividing cells. There are a number of known subtypes of myosin II: they assemble in different ways and have different dynamic properties. In skeletal muscle the myosin molecules form bipolar filaments 15 nm thick. Because these filaments have a symmetric anti-parallel arrangement of subunits, the midpoint is bare of head regions. In smooth muscle the molecules form thicker, flattened ribbons and are orientated in different directions on either face of the ribbon. These arrangements have important consequences for the contractile force characteristics of the different types of muscle cell.

Related molecules include the myosin I subfamily of single-headed molecules with tails of varying length. Functions of myosin I include the movements of membranes in endocytosis, microspike formation in neuronal growth cones, actin–actin sliding and attachment of actin to membranes, e.g. of microvilli. Myosin V is implicated in the movements of membranous organelles on actin filaments. So, for example, vesicles track along F-actin in a similar manner to kinesin and dynein-related movements along microtubules. Other myosins have been isolated; the significance of their diversity is not fully understood.

Other thin filaments

A heterogeneous group of filamentous structures with diameters of 2–4 nm occur in various cells. The two most widely studied forms, titin and nebulin, constitute around 13% of the total protein of skeletal muscle. They are amongst the largest known molecules, and have subunit weights of around 10⁶; native molecules are about 1 μm in length. Their repetitive bead-like structure gives them elastic properties that are important for the effective functioning of muscle, and possibly for other cells.

Intermediate filaments

Intermediate filaments are about 10 nm thick and formed by a heterogeneous group of filamentous proteins. They are found in different cell types and are often present in large numbers, either where structural strength is needed (Fig. 1.10B,C), or to provide scaffolding for the attachment of other structures. It is likely that more complex, non-mechanical functions of intermediate filaments, with implications for human disease, remain to be discovered (see Toivola et al 2005). Intermediate filaments of different molecular classes are characteristic of particular tissues or states of maturity. They are therefore important indicators of the origins of cells or levels of differentiation, and are of considerable value in histopathology.

Of the different classes of intermediate filaments, keratin (cytokeratin) proteins are found in epithelia, where keratin filaments are always composed of equal ratios of types I (acidic) and II (basic to neutral) keratins. About 20 types of each of the acidic and basic/neutral keratin proteins are known. Within the epidermis, expression of keratin heterodimer combinations changes as keratinocytes mature during their transition from basal to superficial layers. Genetic abnormalities of keratins are known to affect the mechanical stability of epithelia. For example, the disease epidermolysis bullosa simplex causes lysis of epidermal basal cells and blistering of the skin after mechanical trauma. It is caused by defects in genes encoding keratins 5 and 14, which produce cytoskeletal instability and thus cellular fragility in the basal cells. When keratins 1 and 10 are affected, cells in the prickle cell layer of the epidermis lyse, and this produces the intraepidermal blistering of epidermolytic hyperkeratosis. For a review, see Porter & Lane (2003).

Vimentins occur in mesenchyme-derived cells of connective tissue, desmins in muscle cells, glial fibrillary acidic protein in glial cells, and peripherin in peripheral axons. Neurofilaments are a major cytoskeletal element in neurones, particularly in axons (Fig. 1.10C), where they are the dominant protein. They are heteropolymers of low, medium and high molecular weight neurofilament proteins; the low molecular weight form is always present in combination with either the medium- or the high-molecular weight neurofilament. Abnormal accumulations of neurofilaments (neurofibrillary tangles) are characteristic features of a number of neuropathological conditions.

Other intermediate filament proteins include nestin, a molecule resembling neurofilament protein which forms intermediate filaments in neurectodermal stem cells in particular. Nuclear lamins are intermediate filaments that line the inner surface of the nuclear envelope of all nucleated cells. They provide a mechanical framework for the nucleus and act as attachment sites for a number of proteins that organize the chromatin at the periphery of the nucleus. They are unusual in that they form an irregular anastomosing network of filaments rather than linear bundles.

The exact manner in which intermediate filament proteins polymerize to form linear filaments is much more complex than that of tubulin or actin, and has not been fully determined. The individual intermediate filament proteins are chains with a middle α-helical region flanked on either side by non-helical domains. The proteins associate as coiled coil dimers that form short rods about 48 nm long. These assemble in pairs in a staggered antiparallel formation to form soluble tetramers, eight of which pack together laterally and twist into the rope-like 10 nm intermediate filament. The 32 α-helices in parallel give the filaments their tensile strength. However, unlike actin and myosin, the antiparallel arrangement of the dimers produces a filamentous protein with no intrinsic polarity. The non-coiled regions of the subunits project outwards as side arms that can link intermediate filaments into bundles or attach them to other structures. The existence of different combinations of subunit proteins within one filament is the basis of their functional diversity. In the living cell they have been shown to be quite dynamic structures, possibly as a result of reversible phosphorylation.

Microtubules

Microtubules are polymers of tubulin with the form of hollow, relatively rigid cylinders, approximately 25 nm in diameter and of varying length (up to 70 μm in spermatozoan flagella). They are present in most cell types, and are particularly abundant in neurones (Fig. 1.10C), leukocytes and blood platelets. They are the predominant constituent of the mitotic spindles of dividing cells. They also form part of the structure of cilia, flagella and centrioles.

There are two major classes of tubulin: α- and β-tubulins. Before microtubule assembly, tubulins are associated as dimers with a combined molecular mass of 100 kDa (50 kDa each). Each protein subunit is approximately 5 nm across and arranged along the long axis in straight rows of alternating α- and β-tubulins, forming protofilaments. Typically, 13 protofilaments (the number can vary between 11 and 16), associate in a ring to form the wall of a hollow cylindrical microtubule. Each longitudinal row is slightly out of alignment with its neighbour, so that a spiral pattern of alternating α and β subunits appears when the microtubule is viewed from the side. There is a dynamic equilibrium between the dimers and assembled microtubules: dimeric asymmetry creates polarity (α-tubulins are all orientated towards the minus end, β-tubulins towards the plus end). Tubulin is added preferentially to the plus end; the minus end is relatively slow growing. Microtubules exhibit a dramatic behaviour, known as dynamic instability, in which growing tubules can undergo a ‘catastrophe’, abruptly shifting from net growth to rapid shrinkage. This can result in disappearance of the microtubule, or the catastrophe can be rescued and growth resumed. Tubulins are guanosine triphosphate (GTP)-binding proteins, and growth is accompanied by hydrolysis of GTP. This may regulate the dynamic behaviour of the tubules. Microtubule growth is initiated at specific sites known as microtubule-organizing centres, the best known of which are centrosomes, from which most cellular microtubules polymerize and basal bodies, from which cilia grow.

Various drugs (e.g. colcemid, vinblastine, griseofulvin, nocodazole) cause microtubule depolymerization by binding the soluble tubulin dimers and so shifting the equilibrium towards the unpolymerized state. Microtubule disassembly causes a wide variety of effects, including the inhibition of cell division by disruption of the mitotic spindle. Conversely, the drug taxol stabilizes microtubules and promotes abnormal microtubule assembly. This can cause a peripheral neuropathy, but taxol is widely used as an effective chemotherapeutic agent in the treatment of cancer.

Different microtubules possess varying degrees of stability, e.g. microtubules in cilia are generally unaffected by many drugs that cause microtubular demolition. There are also differences between tissues, e.g. neurones have a special tubulin subclass. Microtubule organizing centres include a specialized tubulin isoform known as γ-tubulin, that is essential for the nucleation of microtubule growth.

Microtubule-associated proteins

Various proteins that can bind to assembled tubulins may be concerned with structural properties or associated with motility. One important class of microtubule-associated proteins (MAPs) consists of proteins that associate with the plus ends of microtubules. They regulate the dynamic instability of microtubules as well as interactions with other cellular substructures. Structural MAPs form cross-bridges between adjacent microtubules or between microtubules and other structures such as intermediate filaments, mitochondria and the plasma membrane. Microtubule-associated proteins found in neurones include: MAPs 1A and 1B, which are present in neuronal dendrites and axons; MAPs 2A and 2B, found chiefly in dendrites; and tau, found only in axons. MAP 4 is the major microtubule-associated protein in many other cell types. Structural microtubule-associated proteins are implicated in microtubule formation, maintenance and demolition, and are therefore of considerable significance in cell morphogenesis, mitotic division, and the maintenance and modulation of cell shape. Motility-associated microtubule-associated proteins are found in situations in which movement occurs over the surfaces of microtubules, e.g. the transport of cytoplasmic vesicles, bending of cilia and flagella, and some movements of mitotic spindles. They include a large family of motor proteins, the best known of which are the dyneins and kinesins. Another protein, dynamin, is involved in endocytosis. The kinetochore proteins assemble at the chromosomal centromere during mitosis and meiosis. They attach (and thus attach chromosomes) to spindle microtubules; some of the kinetochore proteins are responsible for chromosomal movements in mitotic and meiotic anaphase.

All of these microtubule-associated proteins bind to microtubules and either actively slide along their surfaces or promote microtubule assembly or disassembly. Kinesins and dyneins can simultaneously attach to membranes such as transport vesicles and convey them along microtubules for considerable distances, thus enabling selective targeting of materials within the cell. Such movements occur in both directions along microtubules. Kinesin-dependent motion is usually towards the plus ends of microtubules, e.g. from the cell body towards the axon terminals in neurones, and away from the centrosome in other cells. Conversely, dynein-related movements are in the opposite direction, i.e. to the minus ends of microtubules. Dyneins also form the arms of peripheral microtubules in cilia and flagella, where they make dynamic cross-bridges to adjacent microtubule pairs. When these tethered dyneins try to move, the resulting shearing forces cause the axonemal array of microtubules to bend, generating ciliary and flagellar beating movements. Kinesins form a large and diverse family of related microtubule-stimulated ATPases. Some kinesins are motors that move cargo, others cause microtubule disassembly, whilst still others cross-link mitotic spindle microtubules to push the two centriolar poles apart during mitotic prophase.

Centrioles, centrosomes and basal bodies

Centrioles are microtubular cylinders 0.2 μm in diameter and 0.4 μm long (Fig. 1.11). They are formed by a ring of nine microtubule triplets linked by a number of other proteins. At least two centrioles occur in all animal cells that are capable of mitotic division (eggs, which undergo meiosis instead of mitosis, lack centrioles). They usually lie close together, at right angles or, most usually, at an oblique angle to each other (an arrangement often termed a diplosome), within the centrosome, a densely filamentous region of cytoplasm at the centre of the cell. The centrosome is the major microtubule-organizing centre of most cells; it is the site at which new microtubules are formed and the mitotic spindle is generated during cell division. Centriole biogenesis is a complex process that takes more than a single cell cycle to complete. At the beginning of the S phase (DNA replication phase) of the cell cycle, a new daughter centriole forms at right angles to each separated maternal centriole. Each mother–daughter pair forms one pole of the next mitotic spindle, and the daughter centriole becomes fully mature only as the progeny cells are about to enter the next mitosis. Because centrosomes are microtubule organizing centres, they lie at the centre of a network of microtubules all of which have their minus ends proximal to the centrosome. The association of membrane vesicles with dynein motors means that certain membranes (including the Golgi apparatus) concentrate near the centrosome. This is convenient, as the microtubules provide a means of targeting Golgi vesicular products to different parts of the cell.

Fig. 1.11 A duplicated pair of centrioles in a human carcinoma specimen. Each centriole pair consists of a mother and daughter, oriented approximately at right angles to each other so that one is sectioned transversely (T) and the other longitudinally (L). The transversely sectioned centrioles are seen as rings of microtubule triplets (arrowed).

The microtubule-organizing centre contains complexes of γ-tubulin that nucleate microtubule polymerization at the minus ends of microtubules. Basal bodies are microtubule-organizing centres that are closely related to centrioles, and are believed to be derived from them. They are located at the bases of cilia and flagella, which they anchor to the cell surface. The outer microtubule doublets of cilia and flagella originate from two of the microtubules in each triplet of the basal body.

Cell surface projections

The surfaces of many different types of cell are specialized to form structures that project from the surface. These projections may permit movement of the cell itself (flagella), or of fluids across the apical cell surface (cilia), or increase the surface area available for absorption (microvilli). Infoldings of the basolateral plasma membrane also increase the area for transport across this surface of the cell. In most non-dividing epithelial cells, the centriole gives rise to a non-motile primary cilium, which has an important sensory role.

Cilia and flagella

Cilia and flagella are motile, hair-like projections of the cell surface which create currents in the surrounding fluid, movements of the cell to which they are attached, or both. Cilia occur on many internal surfaces of the body, in particular: the epithelia of most of the respiratory tract; parts of the male and female reproductive tracts; the ependyma that line the central canal of the spinal cord and ventricles of the brain. They also occur at the endings of olfactory receptors and vestibular hair cells, and, in modified form, as portions of the rods and cones of the retina. A single cell may bear many cilia, e.g. in bronchial epithelium, or only one or two. Each male gamete possesses a single flagellum 70 μm long.

A cilium or flagellum consists of a shaft (0.25 μm diameter) constituting most of its length, a tapering tip and a basal body at its base, which lies within the surface cytoplasm of the cell (Fig. 1.12). Other than at its base, the entire structure is covered by plasma membrane. The core of the cilium is the axoneme, a cylinder of nine microtubule doublets that surrounds a central pair of single microtubules (Fig. 1.12). However, there is a class of cilia (primary cilia and nodal cilia) that are composed of nine microtubule doublets and no central microtubules.

Fig. 1.12 A, Structure of a cilium shown in longitudinal (left) and transverse (right) section. A and B are subfibres of the peripheral microtubule doublets (see text); the basal body is structurally similar to a centriole, but with microtubule triplets. B, Apical region of respiratory epithelial cells, showing the proximal parts of three cilia sectioned longitudinally, anchored into the cytoplasm by basal bodies (BB). Other cilia project out of the plane of section and are cut transversely, showing the ‘9 + 2′ arrangement of microtubules.

(Part B by permission from Young B, Heath JW 2000 Wheater’s Functional Histology. Edinburgh: Churchill Livingstone.)

Several filamentous structures are associated with the microtubules in the shaft, e.g. radial spokes extend inwards from the outer microtubules towards the central pair. The outer doublet microtubules bear two rows of tangential dynein arms attached to the A subfibre of the doublet, which point towards the B subfibre of the adjacent doublet. Adjacent doublets are also linked by thin filaments. Other filaments partially encircle the central pair of microtubules, which are also united by ladder-like spokes.

Movements of cilia and flagella are broadly similar. Flagella move by rapid undulation, which passes from the attached to the free end. In human spermatozoa there is an additional helical component to this motion. In cilia, the beating is planar, but asymmetric. In the effective stroke, the cilium remains stiff except at the base, where it bends to produce an oar-like stroke. The recovery stroke follows, during which the bend passes from base to tip, returning the cilium to its initial position for the next cycle. The activity of groups of cilia is usually coordinated so that the bending of one is rapidly followed by the bending of the next and so on, resulting in long travelling waves of metachronal synchrony. These pass over the tissue surface in the same direction as the effective stroke.

When a cilium bends, the microtubules do not change in length, but slide on one another. The dynein arms of peripheral doublets slant towards the base of the cilium from their attached ends. Dynein has an ATPase activity, which causes mutual sliding of adjacent doublets by initially attaching sideways to the next pair, then swinging upwards towards the tip of the cilium. There is a group of genetic diseases (reviewed in Afzelius 2004) in which cilia beat either ineffectively or not at all, e.g. Kartagener’s immotile cilia syndrome. Affected cilia exhibit various ultrastructural defects, such as a lack of dynein arms or missing spokes. Patients with this syndrome suffer various respiratory problems caused by the accumulation of particles in the lungs; males are typically sterile because of the loss of sperm motility, and 50% have an alimentary tract that is a mirror image of the usual pattern (situs inversus) – i.e. it rotates in the opposite direction during early development.

Microvilli

Microvilli are finger-like cell surface extensions usually 0.1 μm in diameter and up to 2 μm long (Fig. 1.13). When arranged in a regular parallel series, they constitute a striated border, as typified by the absorptive surfaces of the epithelial enterocytes of the small intestine. When they are less regular, as in the gallbladder epithelium and proximal kidney tubules, the term brush border is used.

Fig. 1.13 Microvilli sectioned longitudinally in the striated border of an intestinal absorptive cell in a human duodenal biopsy specimen. Actin filaments fill the cores of the microvilli and insert into the apical cytoplasm. There is a prominent glycocalyx (formed by the extracellular domains of plasma membrane glycoproteins), seen as a fuzzy coat at the tips of and between microvilli; it includes enzymes concerned with the final stages of digestion.

Microvilli are covered by plasma membrane and supported internally by closely packed bundles of actin filaments linked by cross-bridges of the actin-bundling proteins, fascin and fimbrin. Other bridges composed of myosin I and calmodulin connect the filament bundles to the plasma membrane. At the tip of each microvillus, the free ends of microfilaments are inserted into a dense mass that includes the protein, villin. The actin filament bundles of microvilli are embedded in the apical cytoplasm amongst a meshwork of transversely running actin filaments stabilized by spectrin to form the terminal web, which is underlain by keratin intermediate filaments. The web is anchored laterally to the zonula adherens of the apical epithelial junctional complex. Myosin II and tropomyosin are also found in the terminal web, which may explain its contractile activity.

Microvilli greatly increase the area of cell surface (up to 40 times), particularly at sites of active absorption. In the small intestine, they have a very thick cell coat or glycocalyx, which reflects the presence of integral membrane glycoproteins, including enzymes concerned with digestion and absorption. Irregular microvilli, filopodia, are also found on the surfaces of many types of cell, particularly free macrophages and fibroblasts, where they may be associated with phagocytosis and cell motility.

Long, regular microvilli are called stereocilia, an early misnomer, as they are not motile and lack microtubules. They are found on cochlear and vestibular receptor cells, where they act as sensory transducers, and also in the absorptive epithelium of the epididymis.

Nuclear membrane

The nucleus is surrounded by two concentric lipid bilayers which together form the nuclear membrane or envelope. The outer membrane bilayer and the lumen between the two bilayers are continuous with the rough endoplasmic reticulum. Like the rough endoplasmic reticulum, the outer membrane of the nuclear envelope is studded with ribosomes that are active in protein synthesis; the newly synthesized proteins pass into the perinuclear space between the two membrane layers.

A special class of intermediate filaments known as lamins is associated with the inner surface of the nuclear membrane. The lamins form a dense meshwork beneath the membrane, the nuclear lamina. The lamin filaments cross each other at right angles to create an irregular anastomosing network that covers the interior surface of the nuclear membrane. In so doing, they reinforce the nuclear membrane mechanically, determine the shape of the nucleus and provide a binding site for a range of proteins that anchor chromatin. Nuclear lamin A, with over 350 mutations, is the most mutated protein linked to human disease. Lamin A mutations cause a surprisingly wide range of diseases, from progeria to various dystrophies (reviewed in Mattout et al 2006 and Pollard and Earnshaw 2007).

Condensed chromatin (heterochromatin) also tends to aggregate near the nuclear membrane during interphase. At the end of mitotic and meiotic prophase (see below), the lamin filaments disassemble, causing the nuclear membranes to vesiculate and disperse into the endoplasmic reticulum. During the final stages of mitosis (telophase), proteins of the nuclear periphery, including lamins, associate with the surface of the chromosomes, providing docking sites for membrane vesicles. Fusion of these vesicles reconstitutes the nuclear compartment.

The transport of molecules between the nucleus and the cytoplasm occurs via specialized nuclear pore structures that perforate the nuclear membrane (Fig. 1.14A). They act as highly selective directional molecular filters, permitting proteins such as histones and gene regulatory proteins (which are synthesized in the cytoplasm but function in the nucleus) to enter the nucleus, and molecules that are synthesized in the nucleus but destined for the cytoplasm (e.g. ribosomal subunits, transfer RNAs and messenger RNAs), to leave the nucleus.

Fig. 1.14 A, Nuclear envelope with nuclear pores (arrowed) in transverse section, showing the continuity between the inner and outer phospholipid layers of the envelope on either side of the pore. The fine ‘membrane’ appearing to span the pore is formed by proteins of the pore complex. Note that the chromatin is less condensed in the region of nuclear pores. Nucleus (N); cytoplasm (C). B, Nuclear pores seen ‘en face’ as spherical structures (arrows) in a tangential section through the nuclear envelope. The appearance of the envelope varies in electron density as the plane of section passes through different regions of the curved double membrane, which is interrupted at intervals by pores through the envelope (see also Fig. 1.1). The surrounding cytoplasm with ribosomes is less electron-dense. Human tissues.

Ultrastructurally, nuclear pores appear as disc-like structures with an outer diameter of 130 nm and an inner pore with an effective diameter for free diffusion of 9 nm (Fig. 1.14B). The nuclear membrane of an active cell is bridged by up to 4000 such pores. The nuclear pore complex has an octagonal symmetry and is formed by an assembly of more than 50 proteins, the nucleoporins. The inner and outer nuclear membranes fuse around the pore complex (Fig. 1.14A). Nuclear pores are freely permeable to small molecules, ions and proteins up to about 17 kDa. Most proteins that enter the nucleus do so as complexes with specific transport receptor proteins known as importins. Importins shuttle back and forth between the nucleus and cytoplasm. Binding of the cargo to the importin requires a short sequence of amino acids known as a nuclear localization sequence (NLS), and can be either direct or via an adapter protein. Interactions of the importin with components of the nuclear pore move it together with its cargo through the pore by an energy-independent process still not understood. A complementary cycle functions in export of proteins and RNA molecules from the nucleus to the cytoplasm using transport receptors known as exportins. For further explanation, see Pollard & Earnshaw (2007).

Chromatin

DNA is organized within the nucleus in a DNA–protein complex known as chromatin. The protein constituents of chromatin are the histones and the non-histone proteins. Non-histone proteins are an extremely heterogeneous group that includes structural proteins, DNA and RNA polymerases and gene regulatory proteins. Histones are the most abundant group of proteins in chromatin, primarily responsible for the packaging of chromosomal DNA into its primary level of organization, the nucleosome. There are four core histone proteins: H2A, H2B, H3 and H4 that combine in equal ratios to form a compact octameric nucleosome core. A fifth histone, H1, is involved in further compaction of the chromatin. The DNA molecule (one per chromosome) winds twice around each nucleosome core, taking up 165 nucleotide pairs. This packaging organizes the DNA into a chromatin fibre 11 nm in diameter, and imparts to this form of chromatin the electron microscopic appearance of beads on a string, in which each bead is separated by a variable length of DNA, typically about 35 nucleotide pairs long. The nucleosome core region and one of the linker regions constitute the nucleosome proper, which is typically about 200 nucleotide pairs in length. However, chromatin rarely exists in this simple form and is usually packaged further into a 30 nm thick fibre, involving a single H1 histone per nucleosome, which interacts with both DNA and protein to impose a higher order of nucleosome packing. Usually, 30 nm fibres are further coiled or folded into larger domains. Individual domains are believed to decondense and extend during active transcription. In a typical interphase nucleus, euchromatin (nuclear regions that appear pale in appropriately stained tissue sections, or relatively electron-lucent in electron micrographs; Fig. 1.2) is likely to consist mainly of 30 nm fibres and loops, and contains the transcriptionally active genes. Transcriptionally active cells, such as most neurones, have nuclei that are predominantly euchromatic and often described as ‘open face’ nuclei.

Heterochromatin (nuclear regions that appear dark in appropriately stained tissue sections or electron-dense in electron micrographs) is characteristically located mainly around the periphery of the nucleus, except over the nuclear pores (Fig. 1.14A), and around the nucleolus (Fig. 1.2). It is a relatively compacted form of chromatin in which the histone proteins carry a specific set of post-translational modifications, including methylation at characteristic residues. This facilitates the binding of specific heterochromatin-associated proteins. Heterochromatin includes non-coding regions of DNA, such as centromeric regions, which are known as constitutive heterochromatin. DNA that is inactivated (becoming resistant to transcription) in some cells as they differentiate during development or cell maturation contributes to heterochromatin, and is known as facultative heterochromatin. The inactive X chromosome in females is an example of facultative heterochromatin and can be identified in the light microscope as the deeply staining Barr body (drumstick chromosome), often located near the nuclear periphery.

In transcriptionally inactive cells, chromatin is predominantly in the condensed, heterochromatic state, and may comprise as much as 90% of the total. Examples of such cells are mature neutrophil leukocytes (in which the condensed chromatin is present in a multilobed, densely staining nucleus), and the highly condensed nuclei of orthochromatic erythroblasts (late-stage erythrocyte precursors). In most mature cells, a mixture of the two occurs, indicating that only a proportion of the DNA is being transcribed. A particular instance of this is seen in the mature B lymphocyte (plasma cell), in which much of the chromatin is in the condensed condition and is arranged in regular masses around the perimeter of the nucleus, producing the so-called ‘clock-face’ nucleus (Figs 4.6, 4.12). Although this cell is actively transcribing, much of its protein synthesis is of a single immunoglobulin type, and consequently much of its genome is in an inactive state.

During mitosis, the chromatin is further reorganized and condensed to form the much shortened chromosomes characteristic of metaphase. This shortening is achieved through further levels of close packing of the chromatin. The mechanism of condensation is unknown, but the condensed chromosomes are stabilized by protein complexes known as condensins. Progressive folding of the chromosomal DNA by interactions with specific proteins can reduce 5 cm of chromosomal DNA by 10,000 fold, to a length of 5 μm in the mitotic chromosome.

Chromosomes and karyotypes

The nuclear DNA of eukaryotic cells is organized into linear units called chromosomes. The DNA in a normal human diploid cell contains 6 × 10⁹ nucleotide pairs organized in the form of 46 chromosomes (44 autosomes and two sex chromosomes). The largest human chromosome (number 1) contains 2.5 × 10⁸ nucleotide pairs, and the smallest (the Y chromosome) 5 × 10⁷ nucleotide pairs.

Each chromosomal DNA molecule contains a number of specialized nucleotide sequences that are associated with its maintenance. One is the centromere. During mitosis, a disc-shaped structure composed of a complex array of proteins, the kinetochore, forms as a substructure of the centromeric region of DNA in order to attach it to the microtubular spindle. Another sequence, the telomere, defines the end of each chromosomal DNA molecule. Telomeres consist of hundreds of repeats of the nucleotide sequence (TTAGGG)_n. The very ends of the chromosomes cannot be replicated by the same DNA polymerase as the rest of the chromosome, and are maintained by a specific enzyme called telomerase which contains an RNA subunit which acts as the template for lengthening the TTAGGG repeats. Thus telomerase is a specialized type of polymerase known as a reverse transcriptase that turns sequences in RNA back into DNA. The number of tandem repeats of the telomeric DNA sequence varies. It appears to shorten with successive cell divisions, because telomerase activity reduces or is absent in differentiated cells with a finite lifespan. It is believed that this mechanism contributes to regulation of cell senescence and may protect against proliferative disorders, including cancer (reviewed in Flores et al 2006).

Classification of human chromosomes

A number of genetic abnormalities can be directly related to the chromosomal pattern. The characterization or karyotyping of chromosome number and structure is therefore of considerable diagnostic importance. The identifying features of individual chromosomes are most easily seen during metaphase, although prophase chromosomes can be used for more detailed analyses.

Lymphocytes separated from blood samples, or cells taken from other tissues, are used as a source of chromosomes. Diagnosis of fetal chromosome patterns is generally carried out on samples of amniotic fluid containing fetal cells aspirated from the uterus by amniocentesis, or on a small piece of chorionic villus tissue removed from the placenta. Whatever their origin, the cells are cultured in vitro and stimulated to divide by treatment with agents that stimulate cell division. Mitosis is interrupted at metaphase with spindle inhibitors. The chromosomes are dispersed by first causing the cells to swell in a hypotonic solution, then the cells are gently fixed and mechanically ruptured on a slide to spread the chromosomes. They are subsequently stained in various ways to allow the identification of individual chromosomes by size, shape and distribution of stain (Fig. 1.15). General techniques show the obvious landmarks, e.g. lengths of arms and positions of constrictions. Banding techniques demonstrate differential staining patterns, characteristic for each chromosome type. Fluorescence staining with quinacrine mustard and related compounds produces Q bands, and Giemsa staining (after treatment that partially denatures the chromatin) gives G bands (Fig. 1.15A). Other less widely used methods include: reverse-Giemsa staining, in which the light and dark areas are reversed (R bands); the staining of constitutive heterochromatin with silver salts (C-banding); T-banding to stain the ends (telomeres) of chromosomes. Collectively, these methods permit the classification of chromosomes into numbered autosomal pairs in order of decreasing size, from 1 to 22, plus the sex chromosomes.

Fig. 1.15 Chromosomes from normal males, arranged as karyotypes. A, G-banded preparation; B, preparation stained by multiplex fluorescence in situ hybridization to identify each chromosome.

(By courtesy of Dr Denise Sheer, Cancer Research UK.)

A summary of the major classes of chromosomes is given below:

Group	Features
1–3 (A)	Large metacentric chromosomes
4–5 (B)	Large submetacentric chromosomes
6–12 + X (C)	Metacentrics of medium size
13–15 (D)	Medium-sized acrocentrics with satellites
16–18 (E)	Shorter metacentrics (16) or submetacentrics (17,18)
19–20 (F)	Shortest metacentrics
21–22 + Y (G)	Short acrocentrics; 21, 22 with satellites, Y without

Methodological advances in banding techniques improved the recognition of abnormal chromosome patterns. The use of in situ hybridization with fluorescent DNA probes specific for each chromosome (Fig. 1.15B) permits the identification of even very small abnormalities.

Nucleolus

Nucleoli are a prominent feature of an interphase nucleus (Fig. 1.2). They are the site of most of the synthesis of rRNA and assembly of ribosome subunits. Ultrastructurally, the nucleolus appears as a pale fibrillar region (non-transcribed DNA), containing dense fibrillar cores (sites of rRNA gene transcription) and granular regions (sites of ribosome subunit assembly) within a diffuse nucleolar matrix. Five pairs of chromosomes carry rRNA genes organized in clusters of tandemly repeated units on each chromosome. Each rRNA unit is transcribed individually and encodes a large precursor RNA that is processed to yield the 28S, 18S and 5.8S rRNA molecules. This processing takes place in the nucleolus, as does the processing of a number of other stable RNAs, including the RNA component of the signal recognition particle (SRP), which is essential for protein secretion. During mitosis the nucleolus breaks down. It reforms after nuclear envelope reformation in telophase, in a process associated with the onset of transcription in nucleolar organizing centres on each chromosome. The 28S, 18S and 5.8S rRNA molecules are assembled into their ribosomal subunits in the granular region of the nucleolus together with the 5S rRNA, which is not synthesized in the nucleolus. The newly formed ribosomal subunits are then translocated to the cytoplasm through the nuclear pores.

Mitosis

New DNA is synthesized during the S phase of the cell cycle interphase. This means that the amount of DNA in diploid cells has doubled to the tetraploid value by the onset of mitosis, although the chromosome number is still diploid. During mitosis, this amount is halved between the two daughter cells, so that DNA quantity and chromosome number are diploid in both cells. The cellular changes that achieve this distribution are conventionally divided into four phases called prophase, metaphase, anaphase and telophase (Fig. 1.16, Fig. 1.17).

Fig. 1.16 The stages in mitosis, including the appearance and distribution of the chromosomes.

Fig. 1.17 Immunofluorescence images of stages in mitosis in human carcinoma cells in culture. A, Metaphase, with spindle microtubules (green), the microtubule-stabilizing protein (HURP; red) and chromosomal DNA (blue). B, Anaphase, with spindle microtubules (green), the central spindle (Aurora-B kinase, red) and segregated chromosomes (blue). C, Late anaphase, with spindle microtubules (green), the central spindle (Plk1 kinase, red, appearing yellow where co-localized with microtubule protein) and segregated chromosomes (blue).

(By courtesy of Dr Herman Silljé, Max-Planck-Institute fur Biochemie, Martinsried, Germany.)

Meiosis

There are two cell divisions during meiosis (Fig. 1.18). Details of this process differ at a cellular level for male and female lineages.

Fig. 1.18 The stages in meiosis, depicted by two pairs of maternal and paternal homologues (dark and pale colours). DNA and chromosome complement changes and exchange of genetic information between homologues are indicated.