专题3 细胞质膜

The lipid molecules are arranged as a continuous double layer about 5 nm thick. This lipid bilayer provides the basic fluid structure of the membrane and serves as an essentially impermeable barrier to the passage of most water-soluble molecules. Most membrane proteins span the lipid bilayer and mediate nearly all of the other functions of the membrane, including the transport of specific molecules across it, and the catalysis of membrane-associated reactions such as ATP synthesis.

The lipid bilayer provides the basic structure for all cell membranes. It is easily seen by electron microscopy, and its bilayer structure is attributable exclusively to the special properties of the lipid molecules, which assemble spontaneously into bilayers even under simple artificial conditions.

Lipid molecules constitute about 50% of the mass of most animal cell membranes, nearly all of the remainder being protein. There are approximately 5 × 106 lipid molecules in a 1 μm × 1 μm area of lipid bilayer, or about 7 × 108 lipid molecules in the plasma membrane of a red blood cell. All of the lipid molecules in cell membranes are amphiphilic; that is, they have a hydrophilic (“water-loving”) or polar end and a hydrophobic (“water-fearing”) or nonpolar end.

The most abundant membrane lipids are the phospholipids. These have a polar head group, which includes a phosphate group, and two hydrophobic hydrocarbon tails. In animal, plant, and bacterial cells, the tails are usually fatty acids, and they can differ in length (they normally contain between 14 and 24 carbon atoms). One tail typically has one or more cis-double bonds (that is, it is unsaturated), while the other tail does not (that is, it is saturated). As shown in Figure 10–2, each cis-double bond creates a kink in the tail. Differences in the length and saturation of the fatty acid tails influence how phospholipid molecules pack against one another, thereby affecting the fluidity of the membrane, as we discuss later.

The main phospholipids in most animal cell membranes are the glycerophospholipids, which have a three-carbon glycerol backbone (see Figure 10–2). Two long-chain fatty acids are linked through ester bonds to adjacent carbon atoms of the glycerol, and the third carbon atom of the glycerol is attached to a phosphate group, which in turn is linked to one of several types of head group. By combining several different fatty acids and head groups, cells make many different glycerophospholipids. Phosphatidylethanolamine, phosphatidylserine, and phosphatidylcholine are the most abundant ones in mammalian cell membranes (Figure 10–3A, B, and C).

Another important class of phospholipids is the sphingolipids, which are built from sphingosine rather than glycerol (Figure 10–3D and E). Sphingosine is a long fatty acid tail with an amino group (NH2) and two hydroxyl groups (OH) at one end. In sphingomyelin, the most common sphingolipid, a fatty acid tail

is attached to the amino group, and a phosphocholine group is attached to the terminal hydroxyl group. Together, the phospholipids phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine, and sphingomyelin constitute more than half the mass of lipid in most mammalian cell membranes (see Table 10–1, p. 610). In addition to phospholipids, the lipid bilayers in many cell membranes contain glycolipids and sterols. Glycolipids resemble sphingolipids, but, instead of a phosphate-linked head group, they have sugars attached. We discuss glycolipids later.

Sterols are rigid ring structures related to steroids but containing a single polar hydroxyl group and a short nonpolar hydrocarbon chain (Figure 10–4). Different types of sterols, distinguished primarily by the side chain attached to the ringed scaffold, are found in fungi, plants, and animal cells. Eukaryotic plasma membranes contain especially large amounts of sterols—up to one molecule for every phospholipid molecule. Cholesterol is the major sterol found in animal cells. The cholesterol molecules orient themselves in the bilayer with their hydroxyl group close to the polar head groups of adjacent phospholipid molecules (Figure 10–5).

Table 10–1 compares the lipid compositions of several biological membranes. Note that bacterial plasma membranes are often composed of one main type of phospholipid and contain no cholesterol. In archaea, lipids usually contain 20- to 25-carbon-long prenyl chains instead of fatty acids; prenyl and fatty acid chains are similarly hydrophobic and flexible (see Figure 10–18F). In thermophilic archaea, the longest lipid chains span both leaflets, making the membrane particularly stable to heat. Thus, lipid bilayers can be built from molecules with similar features but different molecular designs.

The plasma membranes of most eukaryotic cells are more varied than those of prokaryotes and archaea, not only in containing large amounts of sterols but also in containing a mixture of different phospholipids. Analysis of membrane lipids by mass spectrometry has revealed that the lipid composition of a typical eukaryotic cell membrane is much more complex than originally thought. These membranes contain a bewildering variety of perhaps 500–2000 different lipid species with even the simple plasma membrane of a red blood cell containing well over 150. Lipid heterogeneity antagonizes phase transitions and may help membrane-spanning proteins to fit better in the bilayer, avoiding leaks.

While some of this complexity reflects the combinatorial variation in head groups, hydrocarbon chain lengths, and desaturation of the major phospholipid classes, some membranes also contain many structurally distinct minor lipids, at least some of which have important functions. The inositol phospholipids, for example, are present in small quantities in animal cell membranes and have crucial functions in guiding membrane traffic and in cell signaling (discussed in Chapters 13 and 15, respectively). Their local synthesis and destruction are regulated by a large number of enzymes, which create both small intracellular signaling molecules and lipid docking sites on membranes that recruit specific proteins from the cytosol, as we discuss later.

The shape and amphiphilic nature of the phospholipid molecules cause them to form bilayers spontaneously in aqueous environments. As discussed in Chapter 2, hydrophilic molecules dissolve readily in water because they contain charged groups or uncharged polar groups that can form either favorable electrostatic

interactions or hydrogen bonds with water molecules (Figure 10–6A). Hydrophobic molecules, by contrast, are insoluble in water because all, or almost all, of their atoms are uncharged and nonpolar and therefore cannot form energetically favorable interactions with water molecules. If dispersed in water, this forces the adjacent water molecules to reorganize into icelike cages that surround the hydrophobic molecule (Figure 10–6B). Because these cage structures are more ordered than the surrounding water, their formation increases the free energy. This entropic free-energy cost is minimized, however, if the hydrophobic molecules (or the hydrophobic portions of amphiphilic molecules) cluster together so that the smallest number of water molecules is affected.

When phospholipid molecules are exposed to an aqueous environment, they behave as you would expect from the above discussion. They spontaneously pack together to minimize exposure of their hydrophobic tails to water and maximize exposure of their hydrophilic heads to water. Depending on their shape, the optimal packing arrangement is achieved in either of two ways: they can form spherical micelles, with the tails inward, or they can form double-layered sheets, or bilayers, with the hydrophobic tails sandwiched between the hydrophilic head groups (Figure 10–7).

The same forces that drive phospholipids to form bilayers also provide a self-sealing property. A small tear in the bilayer creates a free edge exposed to water; because this is energetically unfavorable, the lipids will rearrange spontaneously to eliminate the free edge. The prohibition of free edges has a profound consequence: the only way for a bilayer to avoid having edges is by closing in on

A lipid bilayer also has other characteristics that make it an ideal structure for cell membranes. One of the most important of these is its fluidity, which is crucial to many membrane functions (Movie 10.2). Around 1970, researchers first recognized that individual lipid molecules are able to diffuse freely within the plane of a lipid bilayer. The initial demonstration came from studies of synthetic (artificial) lipid bilayers, which can be made in the form of spherical vesicles, called liposomes (Figure 10–9), or planar lipid films.

Biophysical studies showed that phospholipid molecules in synthetic bilayers very rarely migrate from the monolayer (also called a leaflet) on one side to that on the other. This process, known as “flip-flop,” occurs on a time scale of hours for any individual molecule. It is slow because during flip-flop, the hydrophilic head groups must transiently enter and pass through the hydrophobic core of the bilayer, which is energetically disfavored. Cholesterol is an exception to this rule and can flip-flop rapidly, having only a single hydroxyl group to accommodate transiently in the hydrophobic core.

By contrast to flip-flop, all lipid molecules rapidly exchange places with their neighbors within a monolayer (∼107 times per second). This gives rise to a rapid lateral diffusion, with a diffusion coefficient (D) of about 10–8 cm2/sec, which means that an average lipid molecule diffuses the length of a large bacterial cell (∼2 μm) in about 1 second.

These studies have also shown that individual lipid molecules rotate very rapidly about their long axis and have flexible hydrocarbon chains. Computer simulations show that lipid molecules in synthetic bilayers are very disordered, presenting an irregular, ragged surface of variously spaced and oriented head groups to the water phase on either side of the bilayer (Figure 10–10). The liquidity of membranes allows lipids to rapidly patch transient holes that may appear in the bilayer through mechanical or other stresses.

Similar mobility studies on labeled lipid molecules in isolated biological membranes and in living cells give results similar to those in synthetic bilayers. They demonstrate that the lipid component of a biological membrane is a twodimensional liquid in which the constituent molecules are free to move laterally. As in synthetic bilayers, individual phospholipid molecules are normally confined to their own monolayer. This confinement creates a problem for the growth of biological membranes. Phospholipid molecules are manufactured in only one monolayer of a membrane, mainly in the cytosolic monolayer of the endoplasmic reticulum membrane. If none of these newly made molecules could migrate reasonably promptly to the noncytosolic monolayer, the membrane would expand asymmetrically. The problem is solved by a special class of membrane proteins called phospholipid translocators, or flippases and scramblases, which catalyze the rapid flip-flop of phospholipids from one monolayer to the other, as discussed in Chapter 12.

polar lipid head groups bind water molecules and ions that need to be displaced for the bilayers of two different liposomes to come into sufficiently close contact to fuse. Biological membranes have an even larger hydration shell due to the proteins embedded or associated with them. This hydration shell insulates the many internal membranes in a eukaryotic cell and prevents their uncontrolled fusion, thereby maintaining the compartmental integrity of membrane-enclosed organelles. All cell membrane fusion events are catalyzed by tightly regulated tethers that bring appropriate membranes close together and fusion proteins that force out the water layer that keeps the bilayers apart, as we discuss in Chapter 13.

The fluidity of cell membranes has to be precisely regulated. It allows membrane proteins to interact rapidly and transiently, and certain membrane transport processes and enzyme activities, for example, cease when the bilayer viscosity is experimentally increased beyond a threshold level. The fluidity of a lipid bilayer depends on both its composition and its temperature, as is readily demonstrated in studies of synthetic lipid bilayers.

A synthetic bilayer made from a single type of phospholipid changes from a liquid state to a two-dimensional rigid crystalline (or gel) state at a characteristic temperature. This change of state is called a phase transition, and the temperature at which it occurs is lower (that is, the membrane becomes more difficult to freeze) if the hydrocarbon chains are short or have double bonds. A shorter chain length reduces the tendency of the hydrocarbon tails to interact with one another in both the same and opposite monolayer so that the membrane remains fluid at lower temperatures. Fluidity is also favored by cis-double bonds because they produce kinks in the chains that make them more difficult to pack together (Figure 10–11).

The makeup of membranes as a complex mix of many different lipid species further adjusts most membranes so that they remain liquids just above the phase-transition point. Bacteria, yeasts, and other organisms whose temperature fluctuates with that of their environment adjust the fatty acid composition of their membrane lipids to maintain a relatively constant fluidity. As the temperature falls, for instance, the cells of those organisms synthesize fatty acids with more cis-double bonds, thereby avoiding the decrease in bilayer fluidity that would otherwise result from the temperature drop.

Sterols, such as cholesterol, modulate the properties of lipid bilayers. When mixed with phospholipids, they enhance the permeability-barrier properties of the lipid bilayer. Cholesterol inserts into the bilayer with its hydroxyl group close to the polar head groups of the phospholipids, so that its rigid, platelike steroid rings interact with—and stiffen—those regions of the hydrocarbon chains closest to the polar head groups (see Figure 10–5 and Movie 10.3). By decreasing the mobility of the first few CH2 groups of the hydrocarbon chains of the phospholipid molecules, cholesterol makes the lipid bilayer less deformable in this region and thereby decreases the permeability of the bilayer to small water-soluble molecules.

Because a lipid bilayer is a two-dimensional fluid, we might expect most types of lipid molecules in it to be well mixed and randomly distributed in their own monolayer. The van der Waals attractive forces between neighboring hydrocarbon tails are not selective enough to hold groups of phospholipid molecules together. With certain lipid mixtures in artificial bilayers, however, one can observe phase transitions that lead to the lateral segregation of lipids with specific lipids coming

together in separate domains (Figure 10–12). In these cases, attractive forces between lipid molecules must outweigh the entropic cost associated with concentrating them. Phase transitions thus break the homogeneity of the bilayer into a patchwork of domains with different properties. There has been a long debate among cell biologists about whether the lipid molecules in the plasma membrane of living cells similarly segregate into specialized domains, called lipid rafts.

Although many lipids and membrane proteins are not distributed uniformly, large-scale lipid phase segregations are seen rarely in living cell membranes. Instead, specific membrane proteins and lipids are seen to concentrate in a more temporary, dynamic fashion facilitated by protein–protein interactions that allow the transient formation of specialized membrane regions (Figure 10–13). Such clusters can be tiny nanoclusters on a scale of a few molecules or larger assemblies that can be seen with electron microscopy, such as the caveolae (discussed in Chapter 13).

The tendency of mixtures of lipids to undergo phase transitions, as seen in artificial bilayers (see Figure 10–12), may help create rafts in living cell membranes—organizing and concentrating membrane proteins either for transport in membrane vesicles (discussed in Chapter 13) or for working together in protein assemblies, such as when they convert extracellular signals into intracellular ones (discussed in Chapter 15).

Most eukaryotic cells store an excess of lipids in lipid droplets, from where they can be retrieved as building blocks for membrane synthesis or as a food source fueling metabolic energy generation. Fat cells, or adipocytes, are specialized for lipid storage. They contain a giant lipid droplet that fills up most of their cytoplasm. Most other cells have many smaller lipid droplets, the number and size varying with the cell’s metabolic state. Fatty acids can be liberated from lipid droplets on demand and exported to other cells through the bloodstream.

Lipid droplets store neutral lipids, such as triacylglycerols and cholesterol esters, which are synthesized from fatty acids and cholesterol by enzymes in the endoplasmic reticulum membrane. Because these lipids do not contain hydrophilic head groups, they are exclusively hydrophobic molecules, and therefore aggregate into three-dimensional droplets rather than into bilayers.

In order for these hydrophobic droplets to reside in the aqueous cytosol of the cell, their surface is covered by phospholipids oriented with their hydrophobic acyl chains facing the lipid droplet and hydrophilic head groups facing the cytosol. This is why lipid droplets are surrounded by a monolayer of phospholipids rather than the bilayer that defines all other membrane-bounded compartments of the cell. The surface of lipid droplets contains a large variety of proteins, some of which are enzymes involved in lipid metabolism.

Lipid droplets form rapidly when cells are exposed to high concentrations of fatty acids. They form from the endoplasmic reticulum membrane where many enzymes of lipid metabolism are localized. Figure 10–14 shows one model of how lipid droplets form and acquire their surrounding monolayer of phospholipids and proteins. In some specialized cells, such as liver cells and enterocytes (the absorptive cells of the gut), droplets bud into the lumen of the endoplasmic reticulum from where they are secreted as lipoprotein particles that move metabolic energy in the form of triglycerides through the body.

The lipid compositions of the two monolayers of the lipid bilayer in many membranes are strikingly different. In the human red blood cell (erythrocyte) membrane, for example, almost all of the phospholipid molecules that have choline—(CH3)3N+CH2CH2OH—in their head group (phosphatidylcholine and sphingomyelin) are in the outer monolayer, whereas almost all that contain a terminal primary amino group (phosphatidylethanolamine and phosphatidylserine) are in the inner monolayer (Figure 10–15). Because the negatively charged phosphatidylserine is located in the inner monolayer, there is a significant difference in charge between the two halves of the bilayer. We discuss in Chapter 12 how membrane-bound phospholipid translocators generate and maintain lipid asymmetry.

Lipid asymmetry is functionally important, especially in converting extracellular signals into intracellular ones (discussed in Chapter 15). Many cytosolic proteins bind to specific lipid head groups found in the cytosolic monolayer of the lipid bilayer. The enzyme protein kinase C (PKC), for example, which is

activated in response to various extracellular signals, binds to the cytosolic face of the plasma membrane, where phosphatidylserine is concentrated, and requires this negatively charged phospholipid for its activity. In other cases, specific lipid head groups must first be modified to create protein-binding sites at a particular time and place. One example is phosphatidylinositol (PI), one of the minor phospholipids that are concentrated in the cytosolic monolayer of cell membranes (see Figure 13–10A, B, and C). Various lipid kinases can add phosphate groups at distinct positions on the inositol ring, creating binding sites that recruit specific proteins from the cytosol to the membrane.

An important example of such a lipid kinase is phosphoinositide 3-kinase (PI 3-kinase), which is activated in response to extracellular signals and helps to recruit specific intracellular signaling proteins to the cytosolic face of the plasma membrane (see Figure 15–53). Similar lipid kinases phosphorylate inositol phospholipids in intracellular membranes and thereby help to recruit proteins that guide membrane transport.

Phospholipids in the plasma membrane are used in yet another way to convert extracellular signals into intracellular ones. The plasma membrane contains various phospholipases that are activated by extracellular signals to cleave specific phospholipid molecules, generating fragments of these molecules that act as short-lived intracellular messengers. Phospholipase C, for example, cleaves an inositol phospholipid in the cytosolic monolayer of the plasma membrane to generate two fragments, one of which remains in the membrane and helps activate protein kinase C, while the other is released into the cytosol and stimulates the release of Ca2+ from the endoplasmic reticulum (see Figure 15–29).

Animals exploit the phospholipid asymmetry of their plasma membranes to distinguish between live and dead cells. When animal cells undergo apoptosis (discussed in Chapter 18), phosphatidylserine, which is normally confined to the cytosolic (or inner) monolayer of the plasma membrane lipid bilayer, rapidly translocates to the extracellular (or outer) monolayer. The phosphatidylserine exposed on the cell surface signals neighboring cells, such as macrophages, to phagocytose the dead cell and digest it. The translocation of the phosphatidylserine in apoptotic cells occurs because the active mechanisms that generate and maintain lipid bilayer asymmetry are impaired.

Sugar-containing lipid molecules called glycolipids have the most extreme asymmetry in their membrane distribution: these molecules, whether in the plasma membrane or in intracellular membranes, are found exclusively in the monolayer facing away from the cytosol. In animal cells, they are made from sphingosine, just like sphingomyelin (see Figure 10–3). These intriguing molecules tend to selfassociate, partly through hydrogen bonds between their sugars and partly through van der Waals forces between their long and straight hydrocarbon chains, which causes them to partition preferentially into lipid raft phases (see Figure 10–13).

The asymmetric distribution of glycolipids in the bilayer results from the addition of sugar groups to the lipid molecules in the lumen of the Golgi apparatus. Thus, the compartment in which they are manufactured is topologically equivalent to the exterior of the cell (discussed in Chapter 12). As they are delivered to the plasma membrane, the sugar groups are exposed at the cell surface (see Figure 10–15), where they have important roles in interactions of the cell with its surroundings.

Glycolipids probably occur in all eukaryotic cell plasma membranes, where they generally constitute about 5% of the lipid molecules in the outer monolayer. They are also found in some intracellular membranes. The most complex of the glycolipids, the gangliosides, contain oligosaccharides with one or more sialic acid moieties, which give gangliosides a net negative charge (Figure 10–16). The most abundant of the more than 40 different gangliosides that have been identified are in the plasma membrane of nerve cells, where gangliosides constitute 5–10% of the total lipid mass; they are also found in much smaller quantities in other cell types.

Hints as to the functions of glycolipids come from their localization. In the plasma membrane of epithelial cells, for example, glycolipids are confined to the exposed apical surface, where they may help to protect the membrane against the harsh conditions frequently found there (such as low pH and high concentrations of degradative enzymes). Charged glycolipids, such as gangliosides, may be important because of their electrical effects: their presence alters the electrical field across the membrane and the concentrations of ions—especially Ca2+—at the membrane surface. Glycolipids also function in cell-recognition processes, in which membrane-bound carbohydrate-binding proteins (lectins) bind to the sugar groups on both glycolipids and glycoproteins in the process of cell–cell adhesion (discussed in Chapter 19).

The ubiquitous presence of glycolipids on the cell surface has been exploited by a number of bacterial toxins and viruses as a means to enter cells. For example, influenza virus interacts with sialic acid sugars on gangliosides during its entry into cells (see Figure 10–16). Polyomaviruses also enter the cell after binding initially to gangliosides. Similarly, the ganglioside GM1 acts as a cell-surface receptor for the bacterial toxin that causes the debilitating diarrhea of cholera. Cholera toxin binds to and enters only those cells that have GM1 on their surface, including intestinal epithelial cells.

Biological membranes consist of a continuous double layer of lipid molecules in which membrane proteins are embedded. This lipid bilayer is fluid, with individual lipid molecules able to diffuse rapidly within their own monolayer. The membrane lipid molecules are amphiphilic. When placed in water, they assemble spontaneously into bilayers, which form sealed compartments. Although cell membranes can contain hundreds of different lipid species, the plasma membrane in animal cells contains three major classes—phospholipids,

cholesterol, and glycolipids. Because of their different backbone structure, phospholipids fall into two subclasses—glycerophospholipids and sphingolipids. The lipid compositions of the inner and outer monolayers are different, reflecting the different functions of the two faces of a cell membrane. Different mixtures of lipids are found in the membranes of cells of different types, as well as in the various membranes of a single eukaryotic cell. Inositol phospholipids are a minor class of phospholipids, which in the cytosolic leaflet of the plasma membrane lipid bilayer play an important part in cell signaling: in response to extracellular signals, specific lipid kinases phosphorylate the head groups of these lipids to form docking sites for cytosolic signaling proteins, whereas specific phospholipases cleave certain inositol phospholipids to generate small intracellular signaling molecules.

Although the lipid bilayer provides the basic structure of biological membranes, the membrane proteins perform most of the membrane’s specific tasks and therefore give each type of cell membrane its characteristic functional properties. Accordingly, the amounts and types of proteins in a membrane are highly variable. In the myelin membrane, which serves mainly as electrical insulation for nerve-cell axons, less than 25% of the membrane mass is protein. By contrast, in the membranes involved in ATP production (such as the internal membranes of mitochondria and chloroplasts), approximately 75% is protein.

A typical plasma membrane is somewhere in between, with protein accounting for about half of its mass. Because lipid molecules are small compared with protein molecules, however, there are always many more lipid molecules than protein molecules in cell membranes—about 50 lipid molecules for each protein molecule in cell membranes that are 50% protein by mass. Membrane proteins vary widely in structure and in the way they associate with the lipid bilayer, which reflects their diverse functions.

Figure 10–17 shows the different ways in which proteins can associate with the membrane. Like their lipid neighbors, membrane proteins are amphiphilic, having hydrophobic and hydrophilic regions. Many membrane proteins extend through the lipid bilayer, and hence are called transmembrane proteins, with part of their mass extruding from the membrane on both sides (Figure 10–17, examples 1, 2, and 5).

Other transmembrane proteins are inserted with the bulk of their mass exposed almost exclusively on one or the other side of the membrane (Figure 10–17, examples 3 and 4). In all cases their hydrophobic regions pass through the membrane and interact with the hydrophobic tails of the lipid molecules in the interior of the bilayer, where they are sequestered away from water. Their hydrophilic regions are exposed to water on either side of the membrane.

Other membrane proteins are located entirely in the cytosol and are attached to the cytosolic monolayer of the lipid bilayer, either by an amphiphilic α helix exposed on the surface of the protein (Figure 10–17, example 6) or by one or more covalently attached lipid chains (Figure 10–17, example 7). The lipid-linked proteins in example 7 in Figure 10–17 are made as soluble proteins in the cytosol and are subsequently anchored to the membrane by the covalent attachment of the lipid group. Lipids can also be attached to the cytosolic facing domains of transmembrane proteins as an additional means of anchoring them to the membrane (see Figure 10–17, example 1).

Yet other membrane proteins are entirely exposed at the external cell surface, being attached to the lipid bilayer only by a covalent linkage (via a specific oligosaccharide) to a lipid anchor in the outer monolayer of the plasma membrane (Figure 10–17, example 8). These proteins are initially made and inserted into the endoplasmic reticulum (ER) by a single transmembrane segment at the C-terminus (similar to example 4 in Figure 10-17). While still in the ER, the transmembrane segment of the protein is cleaved off and a glycosylphosphatidylinositol (GPI) anchor is added, leaving the protein bound to the noncytosolic surface of the ER membrane solely by this anchor (discussed in Chapter 12); transport vesicles eventually deliver the protein to the plasma membrane (discussed in Chapter 13).

By contrast to these examples, membrane-associated proteins do not extend into the hydrophobic interior of the lipid bilayer at all; they are instead bound to either face of the membrane by noncovalent interactions with other membrane proteins (Figure 10–17, examples 9 and 10).

Many of the proteins of this type can be released from the membrane by relatively gentle extraction procedures, such as exposure to solutions of very high or low ionic strength or of extreme pH, which interfere with protein–protein interactions but leave the lipid bilayer intact; these proteins are often referred to as peripheral membrane proteins, and their association with membranes is often regulated by the cell as we discuss next. Transmembrane proteins and many proteins held in the bilayer by lipid groups or hydrophobic polypeptide regions that insert into the hydrophobic core of the lipid bilayer cannot be released in these ways.

How a membrane protein is associated with the lipid bilayer reflects the function of the protein. Only transmembrane proteins can function on both sides of the bilayer or transport molecules across it. Cell-surface receptors, for example, are usually transmembrane proteins that bind signal molecules in the extracellular space and generate different intracellular signals on the opposite side of the plasma membrane, as we discuss in Chapter 15.

To transfer small hydrophilic molecules across a membrane, a membrane transport protein must provide a path for the molecules to cross the hydrophobic permeability barrier of the lipid bilayer; the molecular architecture of multipass transmembrane proteins (Figure 10–17, examples 2 and 5) is ideally suited for this task, as we discuss in Chapter 11. Proteins that function on only one side of the lipid bilayer, by contrast, are often associated exclusively with either the lipid monolayer or a protein domain on that side.

Some intracellular signaling proteins, for example, that help relay extracellular signals into the cell interior are bound to the cytosolic half of the plasma membrane by one or more covalently attached lipid groups, which can be fatty acid chains or prenyl groups (Figure 10–18). In some cases, myristic acid is added to the N-terminal amino group of the protein during its synthesis on a ribosome. All members of the Src family of cytoplasmic protein tyrosine kinases (discussed in Chapter 15) are myristoylated in this way.

Membrane attachment through a single lipid anchor is not very strong, however, and a second lipid group is often added to anchor proteins more firmly to a membrane. For most Src kinases, the second lipid modification is the attachment of palmitic acid to a cysteine side chain of the protein.

When the signaling pathway is turned off, the palmitic acid is removed, allowing the kinase to return to the cytosol. Other intracellular signaling proteins, such as the Ras family small GTPases (discussed in Chapter 15), use a combination of prenyl group and palmitic acid attachment to recruit the proteins to the plasma membrane. Many proteins attach to membranes transiently. Some are classical peripheral membrane proteins that associate with membranes by regulated protein–protein interactions.

Others undergo a transition from soluble to membrane protein by a conformational change that exposes a hydrophobic peptide or covalently attached lipid anchor. Many of the small GTPases of the Rab protein family that regulate intracellular membrane traffic (discussed in Chapter 13), for example, switch depending on the nucleotide that is bound to the protein.

In their GDPbound state they are soluble in the cytosol, often stabilized by binding to a GDP dissociation inhibitor, or GDI, whereas in their GTP-bound state their lipid anchor is exposed and tethers them to membranes. They are membrane proteins at one moment and soluble proteins at the next. Such highly dynamic interactions greatly expand the repertoire of membrane functions.

A transmembrane protein always has a unique orientation in the membrane. This reflects both the asymmetric manner in which it is inserted into the lipid bilayer in the ER during its biosynthesis (discussed in Chapter 12) and the different functions of its cytosolic and noncytosolic domains. These domains are separated by the membrane-spanning segments of the polypeptide chain, which contact the hydrophobic environment of the lipid bilayer and are composed largely of amino acids with nonpolar side chains.

Because the peptide bonds themselves are polar and because water is absent in the bilayer, all peptide bonds in the membranespanning segments of a polypeptide are driven to form hydrogen bonds with one another (discussed in Chapter 3). There are two ways that hydrogen-bonding between peptide bonds can be maximized.

In single-pass transmembrane proteins, the polypeptide chain crosses only once (see Figure 10–17, example 1), whereas in multipass transmembrane proteins, the polypeptide chain crosses multiple times (see Figure 10–17, example 2). An alternative way for the peptide bonds in the lipid bilayer to satisfy their hydrogen-bonding requirements is for multiple transmembrane strands of a polypeptide chain to be arranged as a β sheet that is rolled up into a cylinder (a so-called b barrel; see Figure 10–17, example 5, and Figure 3–7). This protein architecture is seen in the porin proteins that we discuss later.

Progress in x-ray crystallography and single-particle cryo-electron microscopy of membrane proteins has enabled the determination of the three-dimensional structure of many of them. The structures confirm that it is often possible to predict from the protein’s amino acid sequence which parts of the polypeptide chain extend across the lipid bilayer. Segments containing about 20–30 amino acids, with a high degree of hydrophobicity, are long enough to span a lipid bilayer as an α helix, and they can often be identified in hydropathy plots (Figure 10–20). From such plots, it is estimated that about 30% of an organism’s proteins are transmembrane proteins, emphasizing their importance.

Hydropathy plots cannot identify the membrane-spanning segments of a β barrel, as 10 amino acids or fewer are sufficient to traverse a lipid bilayer as an extended β strand, and only every other amino acid side chain is hydrophobic. The strong drive to maximize hydrogen-bonding in the absence of water means that most transmembrane helices span the membrane completely.

But multipass transmembrane proteins can also contain regions that fold into the membrane from either side, squeezing into spaces between transmembrane α helices without contacting the hydrophobic core of the lipid bilayer. Because such regions interact only with other polypeptide regions, they do not need to maximize hydrogen-bonding; they can therefore have a variety of secondary structures, including helices that extend only partway across the lipid bilayer (Figure 10–21).

Such regions are important for the function of some membrane proteins, including water channel and ion channel proteins, in which the regions contribute to the walls of the pores traversing the membrane and confer substrate specificity on the channels, as we discuss in Chapter 11.

The transmembrane α helices of many single-pass membrane proteins do not contribute to the folding of the protein domains on either side of the membrane. As a consequence, it is often possible to engineer cells to produce just the cytosolic or extracellular domains of these proteins as water-soluble molecules. This approach has been invaluable for studying the structure and function of these domains, especially the domains of transmembrane receptor proteins (discussed in Chapter 15).

A transmembrane α helix, even in a single-pass membrane protein, however, often does more than just anchor the protein to the lipid bilayer. Many single-pass membrane proteins form homodimers or heterodimers that are held together by noncovalent, but strong and highly specific, interactions between the two transmembrane α helices; the sequence of the amino acids of these helices contains the information that directs the protein–protein interaction.

Similarly, the transmembrane α helices in multipass membrane proteins occupy specific positions in the folded protein structure that are determined by interactions between the neighboring helices (Figure 10–22). These interactions are crucial for the structure and function of the many receptors, channels, and transporters that communicate or move molecules across cell membranes. In these proteins, each transmembrane helix shields regions of neighboring transmembrane helices from membrane lipids.

When all the helices are packed together into the final folded structure, the outer surface of the helical bundle that is exposed to lipids is composed primarily of hydrophobic amino acids. By contrast, the interior of the helical bundle, which is not exposed directly to lipids, can contain polar and even charged amino acids that would ordinarily be disfavored in the membrane.

The ability to accommodate hydrophilic amino acids within a bundle of transmembrane helices means that multipass membrane proteins can contain binding sites and channels across the membrane for hydrophilic molecules. This property affords multipass membrane proteins considerable functional diversity and probably explains why they represent the majority of membrane proteins.

This means that β barrels all share the same overall architecture in which the amino acids facing the outside of the cylinder are hydrophobic. By contrast, the size of the barrel, and the content inside the barrel, are highly variable and suited to the function of each β-barrel membrane protein. For example, the number of β strands in a β barrel varies widely, from as few as 8 strands to as many as 22 (Figure 10–23). β-Barrel proteins are abundant in the outer membranes of bacteria, mitochondria, and chloroplasts.

Some are pore-forming proteins, which create water-filled channels that allow selected small hydrophilic molecules to cross the membrane. The porins are well-studied examples (Figure 10–23C). Many porin barrels are formed from a 16-strand, antiparallel β sheet rolled up into a cylindrical structure. Polar amino acid side chains line the aqueous channel on the inside, while nonpolar side chains project from the outside of the barrel to interact with the hydrophobic core of the lipid bilayer.

Loops of the polypeptide chain often protrude into the lumen of the channel, narrowing it so that only certain solutes can pass. Some porins are therefore highly selective: maltoporin, for example, preferentially allows maltose and maltose oligomers to cross the outer membrane of E. coli. The FepA protein is a more complex example of a β-barrel transport protein (Figure 10–23D). It transports iron ions across the bacterial outer membrane. It is constructed from 22 β strands, and a large globular domain completely fills the inside of the barrel. Iron ions bind to this domain, which changes its conformation to transfer the iron across the membrane.

Not all β-barrel proteins are transporters. Some form smaller barrels that are completely filled by amino acid side chains that project into the center of the barrel. These proteins function as receptors or enzymes (Figure 10–23A and B); the barrel serves as a rigid anchor, which holds the protein in the membrane and orients the cytosolic loops that form binding sites for specific intracellular molecules.

Most multipass membrane proteins in eukaryotic cells and in the bacterial plasma membrane are constructed from transmembrane α helices. The helices can slide against each other, allowing conformational changes in the protein that can open and shut ion channels, transport solutes, or transduce extracellular signals into intracellular ones. In β-barrel proteins, by contrast, hydrogen bonds bind each β strand rigidly to its neighbors, making conformational changes within the wall of the barrel unlikely.

The plasma membrane of a cell is exposed to the often harsh and constantly changing extracellular environment. How are the membrane and its embedded proteins protected from damage? One answer is that many cells have a layer of oligosaccharide and polysaccharide chains that are attached to the lipids and protein domains facing the outside. Most transmembrane proteins in animal cells are glycosylated. As in glycolipids, the sugar residues are added in the lumen of the endoplasmic reticulum and the Golgi apparatus (discussed in Chapters 12 and 13). For this reason, the oligosaccharide chains are always present on the noncytosolic side of the membrane.

Another important difference between proteins (or parts of proteins) on the two sides of the membrane results from the reducing environment of the cytosol. This environment decreases the likelihood that intrachain or interchain disulfide (S–S) bonds will form between cysteines on the cytosolic side of membranes. These bonds form on the noncytosolic side, where they can help stabilize either the folded structure of the polypeptide chain or its association with other polypeptide chains (Figure 10–24).

Because the extracellular parts of most plasma membrane proteins are glycosylated, carbohydrates extensively coat the surface of all eukaryotic cells. These carbohydrates occur as oligosaccharide chains covalently bound to membrane proteins (glycoproteins) and lipids (glycolipids). They also occur as the polysaccharide chains of integral membrane proteoglycan molecules. Proteoglycans, which consist of long polysaccharide chains linked covalently to a protein core, are found mainly outside the cell, as part of the extracellular matrix (discussed in Chapter 19). But, for some proteoglycans, the protein core either extends across the lipid bilayer or is attached to the bilayer by a glycosylphosphatidylinositol (GPI) anchor.

The terms cell coat or glycocalyx are sometimes used to describe the carbohydrate-rich zone on the cell surface. This carbohydrate layer can be visualized by various stains, such as ruthenium red (Figure 10–25A), as well as by its affinity for carbohydrate-binding proteins called lectins, which can be labeled with a fluorescent dye or some other visible marker.

Although most of the sugar groups are attached to intrinsic plasma membrane molecules, the carbohydrate layer also contains both glycoproteins and proteoglycans that have been secreted into the extracellular space and then adsorbed onto the cell surface (Figure 10–25B). Many of these adsorbed macromolecules are components of the extracellular matrix, so that the boundary between the plasma membrane and the extracellular matrix is often not sharply defined. One of the many functions of a slippery carbohydrate layer is to protect cells against mechanical and chemical damage; it also keeps various other cells at a distance, preventing unwanted cell–cell interactions.

The oligosaccharide side chains of glycoproteins and glycolipids are enormously diverse in their arrangement of sugars. Although they usually contain fewer than 15 sugars, the chains are often branched, and the sugars can be bonded together by various kinds of covalent linkages—unlike the amino acids in a polypeptide chain, which are all linked by identical peptide bonds. Even three sugars can be put together to form hundreds of different trisaccharides. How sugars can form such a vast variety of different structures is discussed in Chapter 2.

Both the diversity and the exposed position of the oligosaccharides on the cell surface make them especially well suited to function in specific cell-recognition processes. Plasma-membrane-bound lectins that recognize specific oligosaccharides on cell-surface glycolipids and glycoproteins mediate a variety of transient cell–cell adhesion processes, including those occurring in lymphocyte recirculation and inflammatory responses (see Figure 19–28).

In general, only agents that disrupt hydrophobic associations and disassemble the lipid bilayer can liberate membrane proteins in a soluble form. The most useful of these for the membrane biochemist are detergents, which are small amphiphilic molecules of variable structure (Movie 10.4). Detergents are much more soluble in water than in lipids. Their polar (hydrophilic) ends can be either charged (ionic), as in sodium dodecyl sulfate (SDS), or uncharged (nonionic), as in b-octylglucoside and Triton X-100 (Figure 10–26A).

At low concentration, detergents are monomeric in solution, but when their concentration is increased above a threshold, called the critical micelle concentration (CMC), they aggregate to form micelles (Figure 10–26B, C, and D). Above the CMC, detergent molecules rapidly diffuse in and out of micelles, keeping the concentration of monomer in the solution constant, no matter how many micelles are present.

Because of this dynamic behavior, the structure of a micelle changes constantly: at any moment most, but not all, of the hydrophilic ends of the detergent molecules will be external facing the water phase and most, but not all, of the hydrophobic ends will be internal to the micelle. Both the CMC and the average number of detergent molecules in a micelle are characteristic properties of each detergent, but they also depend on the temperature, pH, and salt concentration. Detergent solutions are therefore complex systems and are difficult to study.

When mixed with membranes, the hydrophobic ends of detergents bind to the hydrophobic regions of the membrane proteins, where they displace lipid molecules with a collar of detergent molecules. Because the other end of the detergent molecule is polar, this binding tends to bring the membrane proteins into solution as detergent–protein complexes (Figure 10–27). Usually, some lipid molecules also remain attached to the protein.

Strong ionic detergents, such as SDS, can solubilize even the most hydrophobic membrane proteins. This allows the proteins to be analyzed by SDS polyacrylamidegel electrophoresis (discussed in Chapter 8). Such strong detergents, however, unfold (denature) proteins by disrupting their internal hydrophobic cores, thereby rendering the proteins inactive and unusable for functional studies. Nonetheless, proteins can be readily separated and purified in their SDS-denatured form.

Many membrane proteins can be solubilized and then purified in an active form by the use of mild detergents. These detergents cover the hydrophobic regions on membrane-spanning segments that become exposed after lipid removal but do not unfold the protein. Especially when working with multipass membrane proteins, it is often important to maintain a thin layer of lipids upon detergent extraction to retain the protein’s activity.

In the presence of an excess of phospholipid molecules in such a solution, however, membrane proteins incorporate into small liposomes that form spontaneously. In this way, functionally active membrane protein systems can be reconstituted from purified components, providing a powerful means of analyzing the activities of membrane transporters, ion channels, signaling receptors, and so on (Figure 10–28). Such functional reconstitution can determine which proteins are both necessary and sufficient for a particular cell function. For example, the approach provided proof for the hypothesis that the enzymes that make ATP (ATP synthases) use H+ gradients in mitochondrial, chloroplast, and bacterial membranes to produce ATP.

Membrane proteins can also be reconstituted from detergent solution into nanodiscs, which are small, uniformly sized patches of membrane that are surrounded by a belt of a specially designed protein, which covers the exposed edge of the bilayer to keep the patch in solution (Figure 10–29). The belt protein is derived from high-density lipoproteins (HDLs), whose normal function is to keep lipids soluble for transport in the blood.

In nanodiscs the membrane protein of interest can be studied in its native lipid environment and is experimentally accessible from both sides of the bilayer, which is useful, for example, for ligand-binding experiments. Proteins contained in nanodiscs can also be analyzed by single-particle electron microscopy techniques to determine their structure.

By this technique (discussed in Chapter 9), the structure of a membrane protein can be determined to high resolution without a requirement of the protein of interest to crystallize into a regular lattice, which is often hard to achieve for membrane proteins. These developments have led to a rapid increase in the number of three-dimensional structures of membrane proteins and protein complexes that are known, although they are still few compared to the known structures of water-soluble proteins and protein complexes.

In Chapter 11, we consider how multipass transmembrane proteins mediate the selective transport of small hydrophilic molecules across cell membranes. But a detailed understanding of how such a membrane transport protein works requires precise information about its three-dimensional structure in the bilayer. Bacteriorhodopsin was the first membrane transport protein whose structure was determined, and it emerged as the prototype of many multipass membrane proteins that have a similar structure.

The “purple membrane” of the archaeon Halobacterium salinarum is a specialized patch in the plasma membrane that contains a single species of protein molecule, bacteriorhodopsin (Figure 10–30A). The protein functions as a light-activated H+ pump that transfers H+ out of the archaeal cell.

Each bacteriorhodopsin molecule is folded into seven closely packed transmembrane α helices and contains a single light-absorbing group, or chromophore (in this case, retinal), which gives the protein its purple color. Retinal is vitamin A in its aldehyde form and is identical to the chromophore found in rhodopsin of the photoreceptor cells of the vertebrate eye (discussed in Chapter 15). Retinal is covalently linked to a lysine side chain of the bacteriorhodopsin protein.

When activated by a single photon of light, the excited chromophore changes its shape and causes a series of small conformational changes in the protein, resulting in the transfer of one H+ from the inside to the outside of the cell (Figure 10–31A).

In bright light, each bacteriorhodopsin molecule can pump several hundred protons per second. The light-driven proton transfer establishes an H+ gradient across the plasma membrane, which in turn drives the production of ATP by a second protein in the cell’s plasma membrane.

Thus, bacteriorhodopsin converts solar energy into an H+ gradient, which provides energy to the archaeal cell. The high-resolution crystal structure of bacteriorhodopsin reveals many lipid molecules bound in specific places on the protein surface (Figure 10–31B). Interactions with specific lipids are thought to help stabilize many membrane proteins, which work best and sometimes crystallize more readily if some of the lipids remain bound during detergent extraction or if specific lipids are added back to the proteins in detergent solutions.

The specificity of these lipid–protein interactions helps explain why eukaryotic membranes contain such a variety of lipids, with head groups that differ in size, shape, and charge. This layer of lipids also helps to fill gaps between the jagged hydrophobic surface of the membrane protein and the hydrophobic core of the lipid bilayer and to maintain the permeability barrier of the membrane for ions and other solutes.

These lipids are in dynamic association with the protein surface, binding and dissociating at a millisecond time scale. We can think of the membrane lipids as constituting a two-dimensional solvent for the proteins in the membrane, just as water constitutes a three-dimensional solvent for proteins in an aqueous solution: some membrane proteins can function only in the presence of specific lipid head groups, just as many enzymes in aqueous solution require a particular ion for activity.

Bacteriorhodopsin is a member of a large superfamily of membrane proteins with similar structures but different functions. For example, rhodopsin in rod cells of the vertebrate retina and many cell-surface receptor proteins that bind extracellular signal molecules are also built from seven transmembrane α helices. These proteins function as signal transducers rather than as transporters: each responds to an extracellular signal by activating a GTP-binding protein (G protein) inside the cell, and they are therefore called G-protein-coupled receptors (GPCRs), as we discuss in Chapter 15 (see Figure 15–6B).

Although the structures of bacteriorhodopsins and GPCRs are strikingly similar, they show no sequence similarity and thus probably belong to two evolutionarily distant branches of an ancient protein family. A related class of membrane proteins, the channelrhodopsins that green algae use to detect light, form ion channels when they absorb a photon. When engineered so that they are expressed in animal brains, these proteins have become invaluable tools in neurobiology because they allow specific neurons to be stimulated experimentally by shining light on them, as we discuss in Chapter 11 (Figure 11–47).

Many membrane proteins function as part of multicomponent complexes. One is a bacterial photosynthetic reaction center, which was the first membrane protein complex to be crystallized and analyzed by x-ray diffraction. In Chapter 14, we discuss how such photosynthetic complexes function to capture light energy and use it to pump H+ across the membrane.

Many of the membrane protein complexes involved in photosynthesis, proton pumping, and electron transport are even larger than the photosynthetic reaction center. The enormous photosystem II complex from cyanobacteria, for example, contains 19 protein subunits and well over 60 transmembrane helices (see Figure 14–49). Membrane proteins are often arranged in large complexes, not only for harvesting various forms of energy but also for transducing extracellular signals into intracellular ones (discussed in Chapter 15).

Like most membrane lipids, membrane proteins do not tumble (flip-flop) across the lipid bilayer. Tumbling would require large hydrophilic domains to pass through the membrane’s hydrophobic core, which is energetically prohibitive. But just like membrane lipids, proteins can rotate rapidly about an axis perpendicular to the plane of the bilayer (rotational diffusion) and move laterally within the membrane (lateral diffusion).

artificially fused with human cells to produce hybrid cells (heterokaryons) provided the first direct evidence that some plasma membrane proteins are mobile in the plane of the membrane. Two differently labeled antibodies were used to distinguish selected mouse and human plasma membrane proteins.

Although at first the mouse and human proteins were confined to their own halves of the newly formed heterokaryon, the two sets of proteins diffused and mixed over the entire cell surface in about half an hour (Figure 10–32). Lateral membrane protein mobility is important as it allows many cell signaling proteins to assemble and disassemble into protein complexes in response to extracellular ligands, turning their signaling functions on and off.

The lateral diffusion rates of membrane proteins and lipids can be measured by using the technique of fluorescence recovery after photobleaching (FRAP). The method usually involves marking the membrane protein of interest with a specific fluorescent group. This can be done either with a fluorescent ligand such as a fluorophore-labeled antibody that binds to the protein or with recombinant DNA technology to express the protein fused to a fluorescent protein such as green fluorescent protein (GFP; discussed in Chapter 9).

The fluorescent group is then bleached in a small area of membrane by a laser beam, and the time taken for adjacent membrane proteins carrying unbleached ligand or GFP to diffuse into the bleached area is measured (Figure 10–33). From FRAP measurements, we can estimate the diffusion coefficient for the marked cell-surface protein.

The values of the diffusion coefficients for different membrane proteins in different cells are highly variable, because interactions with other proteins impede the diffusion of the proteins to varying degrees. Measurements of proteins that are minimally impeded in this way indicate that cell membranes have a viscosity comparable to that of olive oil.

One drawback to the FRAP technique is that it monitors the movement of large populations of molecules in a relatively large area of membrane; one cannot follow individual protein molecules. If a protein fails to migrate into a bleached area, for example, one cannot tell whether the molecule is truly immobile or just restricted in its movement to a very small region of membrane—perhaps by cytoskeletal proteins.

Single-particle tracking techniques overcome this problem by labeling individual membrane molecules with antibodies coupled to fluorescent dyes or tiny gold particles and tracking their movement by video microscopy. Using single-particle tracking, one can record the diffusion path of a single membrane protein molecule over time. Results from all of these techniques indicate that plasma membrane proteins differ widely in their diffusion characteristics, as we now discuss.

The recognition that biological membranes are two-dimensional fluids was a major advance in understanding membrane structure and function. It has become clear, however, that the picture of a membrane as a lipid sea in which all proteins float freely is greatly oversimplified. Most cells confine membrane proteins to specific regions in a continuous lipid bilayer.

ATP synthase complexes in the inner mitochondrial membrane associate into long double rows, as we discuss in Chapter 14 (see Figure 14–33). Large aggregates of this kind diffuse very slowly. In epithelial cells, such as those that line the gut or the tubules of the kidney, certain plasma membrane enzymes and transport proteins are confined to the apical surface of the cells, whereas others are confined to the basal and lateral surfaces (Figure 10–34).

This asymmetric distribution of membrane proteins is often essential for the function of the epithelium, as we discuss in Chapter 11 (see Figure 11–11). The lipid compositions of these two membrane domains are also different, demonstrating that epithelial cells can prevent the diffusion of lipid as well as protein molecules between the domains. The barriers set up by a specific type of intercellular junction (called a tight junction, discussed in Chapter 19; see Figure 19–18) maintain the separation of both protein and lipid molecules.

A cell can also create membrane domains without using intercellular junctions. As we already discussed, regulated protein–protein interactions in membranes can create nanometer-scale raft domains that are thought to function in signaling and membrane trafficking. A more extreme example is seen in the mammalian spermatozoon, a single cell that consists of several structurally and functionally distinct parts covered by a continuous plasma membrane.

by immunofluorescence microscopy with a variety of antibodies, each of which reacts with a specific cell-surface molecule, the plasma membrane is found to consist of at least three distinct domains (Figure 10–35). Some of the membrane molecules can diffuse freely within the confines of their own domain.

While the molecular nature of most “fences” that prevent the molecules from leaving their domain is not known, some molecular mechanisms for restricting membrane protein movements are understood. The plasma membrane of nerve cells, for example, contains a domain enclosing the cell body and dendrites, and another enclosing the axon; in this case a belt of actin filaments tightly associates with the plasma membrane at the cell-body–axon junction and forms part of the barrier.

As shown in Figure 10–36B and C, a common way in which a cell restricts the lateral mobility of specific membrane proteins is to tether them to macromolecular assemblies on either side of the membrane. The characteristic biconcave shape of a red blood cell (Figure 10–37), for example, results from interactions of its plasma membrane proteins with an underlying cytoskeleton, which consists mainly of a meshwork of the filamentous protein spectrin.

Spectrin is a long, thin, flexible rod about 100 nm in length. As the principal component of the red blood cell cytoskeleton, it maintains the structural integrity and shape of the plasma membrane, which is the cell’s only membrane, as the cell has no nucleus or other organelles. The spectrin cytoskeleton is attached to the membrane through various membrane proteins.

The final result is a deformable, netlike meshwork that covers the entire cytosolic surface of the cell membrane (Figure 10–38). This spectrin-based cytoskeleton enables the red blood cell to withstand the stress on its membrane as it is forced through narrow capillaries.

spherical (instead of concave) and fragile; the severity of the anemia increases with the degree of spectrin deficiency. An analogous but much more elaborate and highly dynamic cytoskeletal network exists beneath the plasma membrane of most other cells in our body. This network, which constitutes the cortex of the cell, is rich in actin filaments, which are attached to the plasma membrane in numerous ways.

remodeling of the cortical actin network provides a driving force for many essential cell functions, including cell movement, endocytosis, and the formation of transient, mobile plasma membrane structures such as filopodia and lamellipodia discussed in Chapter 16. The cortex of nucleated cells also contains proteins that are structurally homologous to spectrin and the other components of the red cell cytoskeleton. We discuss the cortical cytoskeleton in nucleated cells and its interactions with the plasma membrane in Chapter 16.

The cortical cytoskeletal network restricts diffusion of plasma membrane proteins beyond those that are directly anchored to it. Because the cytoskeletal filaments are often closely apposed to the cytosolic surface of the plasma membrane, they can form mechanical barriers that obstruct the free diffusion of proteins in the membrane. These barriers partition the membrane into small domains, or corrals (Figure 10–39A), which can be either permanent, as in the sperm (see Figure 10–35), or transient.

The barriers can be detected when the diffusion of individual membrane proteins is followed by high-speed, single-particle tracking. The proteins diffuse rapidly but are confined within an individual corral (Figure 10–39B); occasionally, however, thermal motions cause a few cortical filaments to detach transiently from the membrane, allowing the protein to escape into an adjacent corral.

The extent to which a transmembrane protein is confined within a corral depends on its association with other proteins and the size of its cytoplasmic domain; proteins with a large cytosolic domain will have a harder time passing through cytoskeletal barriers.

When a cell-surface receptor binds its extracellular signal molecules, for example, large protein complexes build up on the cytosolic domain of the receptor, making it more difficult for the receptor to escape from its corral. It is thought that corralling helps concentrate such signaling complexes, increasing the speed and efficiency of the signaling process (discussed in Chapter 15).

Cell membranes assume many different shapes, as illustrated by the elaborate and varied structures of cell-surface protrusions and membrane-enclosed organelles in eukaryotic cells. Flat sheets, narrow tubules, round vesicles, fenestrated sheets, and pita-bread–shaped cisternae are all part of the repertoire. Often, a variety of shapes will be present in different regions of the same continuous bilayer.

Membrane shape is controlled dynamically, as many essential cell processes—including vesicle budding, cell movement, and cell division—require elaborate transient membrane deformations. In many cases, membrane shape is influenced by dynamic pushing and pulling forces exerted by cytoskeletal or extracellular structures, as we discuss in Chapters 13 and 16. A crucial part in producing these deformations is played by membrane-bending proteins, which control local membrane curvature.

1. Some insert hydrophobic protein domains or attached lipid anchors into one of the leaflets of a lipid bilayer. Increasing the area of only one leaflet causes the membrane to bend (Figure 10–40B). The proteins that shape the convoluted network of narrow endoplasmic reticulum tubules work in this way (discussed in Chapter 12).

3. Asymmetric distribution of cone-shaped and inverted cone-shaped lipids in the inner or outer leaflets can cause membrane bending. Some membrane-bending proteins cause particular membrane lipids to cluster together, thereby inducing membrane curvature. The ability of a lipid to induce positive or negative membrane curvature is determined by the relative cross-sectional areas of its head group and its hydrocarbon tails. For example, the large head group of phosphoinositides make these lipid molecules wedge-shaped, and their accumulation in a domain of one leaflet of a bilayer therefore induces positive curvature (Figure 10–40D). By contrast, phospholipases that remove lipid head groups produce inversely shaped lipid molecules that induce negative curvature.

Whereas the lipid bilayer determines the basic structure of biological membranes, proteins are responsible for most membrane functions, serving as specific receptors, enzymes, transporters, and so on. Transmembrane proteins extend across the lipid bilayer. Some of these membrane proteins are single-pass proteins, in which the polypeptide chain crosses the bilayer as a single a helix. Others are multipass proteins, in which the polypeptide chain crosses the bilayer multiple times—either as a series of a helices or as a b sheet rolled up into the shape of a barrel.

All proteins responsible for the transport of ions and other small water-soluble molecules through the membrane are multipass proteins. Some membrane proteins do not span the bilayer but instead are attached to either side of the membrane: some are attached to the cytosolic side by an amphipathic a helix on the protein surface or by the covalent attachment of one or more lipid chains, others are attached to the noncytosolic side by a GPI anchor.

Some membrane-associated proteins are bound by noncovalent interactions with transmembrane proteins. In the plasma membrane of all eukaryotic cells, most of the proteins exposed on the cell surface and some of the lipid molecules in the outer lipid monolayer have oligosaccharide chains covalently attached to them. Like the lipid molecules in the bilayer, many membrane proteins are able to diffuse rapidly in the plane of the membrane. However, cells have ways of immobilizing specific membrane proteins, as well as ways of confining both membrane protein and lipid molecules to particular domains in a continuous lipid bilayer. The dynamic association of membrane-bending proteins confers on membranes their characteristic three-dimensional shapes.