Chapter 9

Membranes and Cell Surfaces

Membranes are like thin envelopes that define the volume
of cells, like this hot-air ballon is defined in space by it's
colorful envelope. (Boar's Head Inn Balloon, Charlottesville,
VA, by Larry Swank)

Membranes serve a number of essential cellular functions. They constitute the boundaries of cells and intracellular organelles, and they provide a surface where many important biological reactions and processes occur. Membranes have proteins that mediate and regulate the transport of metabolites, macromolecules, and ions. Hormones and many other biological signal molecules and regulatory agents exert their effects via interactions with membranes. Photosynthesis, electron transport, oxidative phosphorylation , muscle contraction, and electrical activity all depend on membranes and membrane proteins. Thirty percent of the genes of at least one organism, Mycoplasma genitalium (whose entire genome has been sequenced), are thought to encode membrane proteins.
            Biological membranes are uniquely organized arrays of lipids and proteins (either of which may be decorated with carbohydrate groups). The lipids found in biological systems are often amphipathic , signifying that they possess both polar and nonpolar groups. The hydrophobic nature of lipid molecules allows membranes to act as effective barriers to polar molecules. The polar moieties of amphipathic lipids typically lie at the surface of membranes, where they interact with water. Proteins interact with the lipids of membranes in a variety of ways. Some proteins associate with membranes via electrostatic interactions with polar groups on the membrane surface, whereas other proteins are embedded to various extents in the hydrophobic core of the membrane. Other proteins are anchored to membranes via covalently bound lipid molecules that associate strongly with the hydrophobic membrane core.
            This chapter discusses the composition, structure, and dynamic processes of biological membranes.

9.1 · Membranes

Cells make use of many different types of membranes. All cells have a cytoplasmic membrane, or plasma membrane, that functions (in part) to separate the cytoplasm from the surroundings. In the early days of biochemistry, the plasma membrane was not accorded many functions other than this one of partition. We now know that the plasma membrane is also responsible for (1) the exclusion of certain toxic ions and molecules from the cell, (2) the accumulation of cell nutrients, and (3) energy transduction. It functions in (4) cell locomotion, (5) reproduction, (6) signal transduction processes, and (7) interactions with molecules or other cells in the vicinity.

Figure 9.1 · Electron micrographs of several different membrane structures: (a) Menoidium, a protozoan; (b) Gram-negative envelope of Aquaspirillum serpens; (c) Golgi apparatus; (d) pancreatic acinar cell. (a, T. T. Beveridge/Visuals Unlimited; b, © Cabisco/Visuals Unlimited; c, d, © D. W. Fawcett/Photo Researchers, Inc.)           

     Even the plasma membranes of prokaryotic cells (bacteria) are complex (Figure 9.1). With no intracellular organelles to divide and organize the work, bacteria carry out processes either at the plasma membrane or in the cytoplasm itself. Eukaryotic cells, however, contain numerous intracellular organelles that perform specialized tasks. Nucleic acid biosynthesis is handled in the nucleus; mitochondria are the site of electron transport, oxidative phosphorylation , fatty acid oxidation, and the tricarboxylic acid cycle; and secretion of proteins and other substances is handled by the endoplasmic reticulum and the Golgi apparatus. This partitioning of labor is not the only contribution of the membranes in these cells. Many of the processes occurring in these organelles (or in the prokaryotic cell) actively involve membranes. Thus, some of the enzymes involved in nucleic acid metabolism are membrane-associated. The electron transfer chain and its associated system for ATP synthesis are embedded in the mitochondrial membrane. Many enzymes responsible for aspects of lipid biosynthesis are located in the endoplasmic reticulum membrane.

Spontaneously Formed Lipid Structures

Monolayers and Micelles

Amphipathic lipids spontaneously form a variety of structures when added to aqueous solution. All these structures form in ways that minimize contact between the hydrophobic lipid chains and the aqueous milieu. For example, when small amounts of a fatty acid are added to an aqueous solution, a monolayer is formed at the air - water interface, with the polar head groups in contact with the water surface and the hydrophobic tails in contact with the air (Figure 9.2). Few lipid molecules are found as monomers in solution.

Figure 9.2 · Several spontaneously formed lipid structures.

   Further addition of fatty acid eventually results in the formation of micelles. Micelles formed from an amphipathic lipid in water position the hydrophobic tails in the center of the lipid aggregation with the polar head groups facing outward. Amphipathic molecules that form micelles are characterized by a unique critical micelle concentration, or CMC. Below the CMC, individual lipid molecules predominate. Nearly all the lipid added above the CMC, however, spontaneously forms micelles. Micelles are the preferred form of aggregation in water for detergents and soaps. Some typical CMC values are listed in Figure 9.3.

Figure 9.3 · The structures of some common detergents and their physical properties. Micelles formed by detergents can be quite large. Triton X-100, for example, typically forms micelles with a total molecular mass of 90 to 95 kD. This corresponds to approximately 150 molecules of Triton X-100 per micelle.

Lipid Bilayers

Lipid bilayers consist of back-to-back arrangements of monolayers (Figure 9.2). Phospholipids prefer to form bilayer structures in aqueous solution because their pairs of fatty acyl chains do not pack well in the interior of a micelle. Phospholipid bilayers form rapidly and spontaneously when phospholipids are added to water, and they are stable structures in aqueous solution. As opposed to micelles, which are small, self-limiting structures of a few hundred molecules, bilayers may form spontaneously over large areas (10⁸ nm² or more). Because exposure of the edges of the bilayer to solvent is highly unfavorable, extensive bilayers normally wrap around themselves and form closed vesicles (Figure 9.4).

Figure 9.4 · Drawings of (a) a bilayer, (b) a unilamellar vesicle, (c) a multilamellar vesicle, and (d) an electron micrograph of a multilamellar Golgi structure (394,000). (d, David Phillips/Visuals Unlimited)

The nature and integrity of these vesicle structures are very much dependent on the lipid composition. Phospholipids can form either unilamellar vesicles (with a single lipid bilayer ) known as liposomes , or multilamellar vesicles. These latter structures are reminiscent of the layered structure of onions. Multilamellar vesicles were discovered by Sir Alex Bangham and are sometimes referred to as “Bangosomes ” in his honor.
            Liposomes are highly stable structures that can be subjected to manipulations such as gel filtration chromatography and dialysis. With such methods, it is possible to prepare liposomes having different inside and outside solution compositions. Liposomes can be used as drug and enzyme delivery systems in therapeutic applications. For example, liposomes can be used to introduce contrast agents into the body for diagnostic imaging procedures, including computerized tomography (CT) and magnetic resonance imaging (MRI) (Figure 9.5). Liposomes can fuse with cells, mixing their contents with the intracellular medium. If methods can be developed to target liposomes to selected cell populations, it may be possible to deliver drugs, therapeutic enzymes, and contrast agents to particular kinds of cells (such as cancer cells).

Figure 9.5 · A computerized tomography (CT) image of the upper abdomen of a dog, following administration of liposome-encapsulated iodine, a contrast agent that improves the light/dark contrast of objects in the image. The spine is the bright white object at the bottom and the other bright objects on the periphery are ribs. The liver (white) occupies most of the abdominal space. The gallbladder (bulbous object at the center top) and blood vessels appear dark in the image. The liposomal iodine contrast agent has been taken up by Kuppfer cells, which are distributed throughout the liver, except in tumors. The dark object in the lower right is a large tumor. None of these anatomical features would be visible in a CT image in the absence of the liposomal iodine contrast agent. (Courtesy of Walter Perkins, The Liposome Co., Inc., Princeton, NJ, and Brigham and Women’s Hospital, Boston, MA)

            That vesicles and liposomes form at all is a consequence of the amphipathic nature of the phospholipid molecule. Ionic interactions between the polar head groups and water are maximized, whereas hydrophobic interactions (see Chapter 2) facilitate the association of hydrocarbon chains in the interior of the bilayer . The formation of vesicles results in a favorable increase in the entropy of the solution, because the water molecules are not required to order themselves around the lipid chains. It is important to consider for a moment the physical properties of the bilayer membrane, which is the basis of vesicles and also of natural membranes. Bilayers have a polar surface and a nonpolar core. This hydrophobic core provides a substantial barrier to ions and other polar entities. The rates of movement of such species across membranes are thus quite slow. However, this same core also provides a favorable environment for nonpolar molecules and hydrophobic proteins. We will encounter numerous cases of hydrophobic molecules that interact with membranes and regulate biological functions in some way by binding to or embedding themselves in membranes.

Fluid Mosaic Model

In 1972, S. J. Singer and G. L. Nicolson proposed the fluid mosaic model for membrane structure, which suggested that membranes are dynamic structures composed of proteins and phospholipids. In this model, the phospholipid bilayer is a fluid matrix, in essence, a two-dimensional solvent for proteins. Both lipids and proteins are capable of rotational and lateral movement.

Figure 9.6 · The fluid mosaic model of membrane structure proposed by S. J. Singer and G. L. Nicolson. In this model, the lipids and proteins are assumed to be mobile, so that they can move rapidly and laterally in the plane of the membrane. Transverse motion may also occur, but it is much slower.

            Singer and Nicolson also pointed out that proteins can be associated with the surface of this bilayer or embedded in the bilayer to varying degrees (Figure 9.6).They defined two classes of membrane proteins. The first, called peripheral proteins (or extrinsic proteins), includes those that do not penetrate the bilayer to any significant degree and are associated with the membrane by virtue of ionic interactions and hydrogen bonds between the membrane surface and the surface of the protein. Peripheral proteins can be dissociated from the membrane by treatment with salt solutions or by changes in pH (treatments that disrupt hydrogen bonds and ionic interactions). Integral proteins (or intrinsic proteins), in contrast, possess hydrophobic surfaces that can readily penetrate the lipid bilayer itself, as well as surfaces that prefer contact with the aqueous medium. These proteins can either insert into the membrane or extend all the way across the membrane and expose themselves to the aqueous solvent on both sides. Singer and Nicolson also suggested that a portion of the bilayer lipid interacts in specific ways with integral membrane proteins and that these interactions might be important for the function of certain membrane proteins. Because of these intimate associations with membrane lipid, integral proteins can only be removed from the membrane by agents capable of breaking up the hydrophobic interactions within the lipid bilayer itself (such as detergents and organic solvents). The fluid mosaic model has become the paradigm for modern studies of membrane structure and function.

Membrane Bilayer Thickness

The Singer - Nicolson model suggested a value of approximately 5 nm for membrane thickness, the same thickness as a lipid bilayer itself. Low angle X-ray diffraction studies in the early 1970s showed that many natural membranes were approximately 5 nm in thickness and that the interiors of these membranes were low in electron density. This is consistent with the arrangement of bilayers having the hydrocarbon tails (low in electron density) in the interior of the membrane. The outside edges of these same membranes were shown to be of high electron density, which is consistent with the arrangement of the polar lipid head groups on the outside surfaces of the membrane.

Hydrocarbon Chain Orientation in the Bilayer

An important aspect of membrane structure is the orientation or ordering of lipid molecules in the bilayer . In the bilayers sketched in Figures 9.2 and 9.4, the long axes of the lipid molecules are portrayed as being perpendicular (or normal) to the plane of the bilayer . In fact, the hydrocarbon tails of phospholipids may tilt and bend and adopt a variety of orientations. Typically, the portions of a lipid chain near the membrane surface lie most nearly perpendicular to the membrane plane, and lipid chain ordering decreases toward the end of the chain (toward the middle of the bilayer ).

Membrane Bilayer Mobility

The idea that lipids and proteins could move rapidly in biological membranes was a relatively new one when the fluid mosaic model was proposed. Many of the experiments designed to test this hypothesis involved the use of specially designed probe molecules. The first experiment demonstrating protein lateral movement in the membrane was described by L. Frye and M. Edidin in 1970. In this experiment, human cells and mouse cells were allowed to fuse together. Frye and Edidin used fluorescent antibodies to determine whether integral membrane proteins from the two cell types could move and intermingle in the newly formed, fused cells. The antibodies specific for human cell proteins were labeled with rhodamine , a red fluorescent marker, and the antibodies specific for mouse cell proteins were labeled with fluorescein , a green fluorescent marker. When both types of antibodies were added to newly fused cells, the binding pattern indicated that integral membrane proteins from the two cell types had moved laterally and were dispersed throughout the surface of the fused cell (Figure 9.7). This clearly demonstrated that integral membrane proteins possess significant lateral mobility.

Figure 9.7 · The Frye – Edidin experiment. Human cells with membrane antigens for red fluorescent antibodies were mixed and fused with mouse cells having membrane antigens for green fluorescent antibodies. Treatment of the resulting composite cells with red- and green-fluorescent – labeled antibodies revealed a rapid mixing of the membrane antigens in the composite membrane. This experiment demonstrated the lateral mobility of membrane proteins.

            Just how fast can proteins move in a biological membrane? Many membrane proteins can move laterally across a membrane at a rate of a few microns per minute. On the other hand, some integral membrane proteins are much more restricted in their lateral movement, with diffusion rates of about 10 nm/sec or even slower. These latter proteins are often found to be anchored to the cytoskeleton (Chapter 17), a complex latticelike structure that maintains the cell’s shape and assists in the controlled movement of various substances through the cell.
            Lipids also undergo rapid lateral motion in membranes. A typical phospholipid can diffuse laterally in a membrane at a linear rate of several microns per second. At that rate, a phospholipid could travel from one end of a bacterial cell to the other in less than a second or traverse a typical animal cell in a few minutes. On the other hand, transverse movement of lipids (or proteins) from one face of the bilayer to the other is much slower (and much less likely). For example, it can take as long as several days for half the phospholipids in a bilayer vesicle to “flip” from one side of the bilayer to the other.

Membranes Are Asymmetric Structures

Biological membranes are asymmetric structures. There are several kinds of asymmetry to consider. Both the lipids and the proteins of membranes exhibit lateral and transverse asymmetries. Lateral asymmetry arises when lipids or proteins of particular types cluster in the plane of the membrane.

Lipids Exhibit Lateral Membrane Asymmetry

Lipids in model systems are often found in asymmetric clusters (see Figure 9.8). Such behavior is referred to as a phase separation, which arises either spontaneously or as the result of some extraneous influence. Phase separations can be induced in model membranes by divalent cations , which interact with negatively charged moieties on the surface of the bilayer . For example, Ca²⁺ induces phase separations in membranes formed from phosphatidylserine (PS) and phosphatidylethanolamine (PE) or from PS, PE, and phosphatidylcholine . Ca²⁺ added to these membranes forms complexes with the negatively charged serine carboxyls , causing the PS to cluster and separate from the other lipids. Such metal-induced lipid phase separations have been shown to regulate the activity of membrane-bound enzymes.

Figure 9.8 · An illustration of the concept of lateral phase separations in a membrane. Phase separations of phosphatidylserine (green circles) can be induced by divalent cations such as Ca²⁺.

            There are other ways in which the lateral organization (and asymmetry) of lipids in biological membranes can be altered. For example, cholesterol can intercalate between the phospholipid fatty acid chains, its polar hydroxyl group associated with the polar head groups. In this manner, patches of cholesterol and phospholipids can form in an otherwise homogeneous sea of pure phospholipid . This lateral asymmetry can in turn affect the function of membrane proteins and enzymes. The lateral distribution of lipids in a membrane can also be affected by proteins in the membrane. Certain integral membrane proteins prefer associations with specific lipids. Proteins may select unsaturated lipid chains over saturated chains or may prefer a specific head group over others.

Proteins Exhibit Lateral Membrane Asymmetry

Membrane proteins in many cases are randomly distributed through the plane of the membrane. This was one of the corollaries of the fluid mosaic model of Singer and Nicholson and has been experimentally verified using electron microscopy. Electron micrographs show that integral membrane proteins are often randomly distributed in the membrane, with no apparent long-range order.
            However, membrane proteins can also be distributed in nonrandom ways across the surface of a membrane. This can occur for several reasons. Some proteins must interact intimately with certain other proteins, forming multisubunit complexes that perform specific functions in the membrane. A few integral membrane proteins are known to self-associate in the membrane, forming large multimeric clusters. Bacteriorhodopsin , a light-driven proton pump protein, forms such clusters, known as “purple patches,” in the membranes of Halobacterium halobium (Figure 9.9). The bacteriorhodopsin protein in these purple patches forms highly ordered, two-dimensional crystals.

Figure 9.9 · The purple patches of Halobacterium halobium.

Transverse Membrane Asymmetry

Membrane asymmetries in the transverse direction (from one side of the membrane to the other) can be anticipated when one considers that many properties of a membrane depend upon its two-sided nature. Properties that are a consequence of membrane “sidedness” include membrane transport, which is driven in one direction only, the effects of hormones at the outsides of cells, and the immunological reactions that occur between cells (necessarily involving only the outside surfaces of the cells). One would surmise that the proteins involved in these and other interactions must be arranged asymmetrically in the membrane.

Protein Transverse Asymmetry

Protein transverse asymmetries have been characterized using chemical, enzymatic, and immunological labeling methods. Working with glycophorin , the major glycoprotein in the erythrocyte membrane (discussed in Section 9.2), Mark Bretscher was the first to demonstrate the asymmetric arrangement of an integral membrane protein. Treatment of whole erythrocytes with trypsin released the carbohydrate groups of glycophorin (in the form of several small glycopeptides). Because trypsin is much too large to penetrate the erythrocyte membrane, the N-terminus of glycophorin , which contains the carbohydrate moieties, must be exposed to the outside surface of the membrane. Bretscher showed that [³⁵S]-formylmethionylsulfone methyl phosphate could label the C-terminus of glycophorin with ³⁵S in erythrocyte membrane fragments but not in intact erythrocytes. This clearly demonstrated that the C-terminus of glycophorin is uniformly exposed to the interior surface of the erythrocyte membrane. Since that time, many integral membrane proteins have been shown to be oriented uniformly in their respective membranes.

Lipid Transverse Asymmetry

Phospholipids are also distributed asymmetrically across many membranes. In the erythrocyte, phosphatidylcholine (PC) comprises about 30% of the total phospholipid in the membrane. Of this amount, 76% is found in the outer monolayer and 24% is found in the inner monolayer. Since this early observation, the lipids of many membranes have been found to be asymmetrically distributed between the inner and outer monolayers . Figure 9.10 shows the asymmetric distribution of phospholipids observed in the human erythrocyte membrane. Asymmetric lipid distributions are important to cells in several ways. The carbohydrate groups of glycolipids (and of glycoproteins ) always face the outside surface of plasma membranes where they participate in cell recognition phenomena. Asymmetric lipid distributions may also be important to various integral membrane proteins, which may prefer particular lipid classes in the inner and outer monolayers . The total charge on the inner and outer surfaces of a membrane depends on the distribution of lipids. The resulting charge differences affect the membrane potential, which in turn is known to modulate the activity of certain ion channels and other membrane proteins.

Figure 9.10 · Phospholipids are arranged asymmetrically in most membranes, including the human erythrocyte membrane, as shown here. Values are mole percentages. (After Rothman and Lenard, 1977. Science 194:1744.)

            How are transverse lipid asymmetries created and maintained in cell membranes? From a thermodynamic perspective, these asymmetries could only occur by virtue of asymmetric syntheses of the bilayer itself or by energy-dependent asymmetric transport mechanisms. Without at least one of these, lipids of all kinds would eventually distribute equally between the two monolayers of a membrane. In eukaryotic cells, phospholipids, glycolipids , and cholesterol are synthesized by enzymes located in (or on the surface of) the endoplasmic reticulum (ER) and the Golgi system (discussed in Chapter 25). Most if not all of these biosynthetic processes are asymmetrically arranged across the membranes of the ER and Golgi . There is also a separate and continuous flow of phospholipids, glycolipids , and cholesterol from the ER and Golgi to other membranes in the cell, including the plasma membrane. This flow is mediated by specific lipid transfer proteins. Most cells appear to contain such proteins.

Flippases : Proteins That Flip Lipids Across the Membrane

Proteins that can “flip” phospholipids from one side of a bilayer to the other have also been identified in several tissues (Figure 9.11). Called flippases , these proteins reduce the half-time for phospholipid movement across a membrane from 10 days or more to a few minutes or less. Some of these systems may operate passively, with no required input of energy, but passive transport alone cannot establish or maintain asymmetric transverse lipid distributions. However, rapid phospholipid movement from one monolayer to the other occurs in an ATP-dependent manner in erythrocytes. Energy-dependent lipid flippase activity may be responsible for the creation and maintenance of transverse lipid asymmetries.

Figure 9.11 · Phospholipids can be “flipped” across a bilayer membrane by the action of flippase proteins.
When, by normal diffusion through the bilayer, the lipid encounters a flippase, it can be moved quickly to
the other face of the bilayer.

Membrane Phase Transitions

Lipids in bilayers undergo radical changes in physical state over characteristic narrow temperature ranges. These changes are in fact true phase transitions, and the temperatures at which these changes take place are referred to as transition temperatures or melting temperatures (T_m). These phase transitions involve substantial changes in the organization and motion of the fatty acyl chains within the bilayer . The bilayer below the phase transition exists in a closely packed gel state, with the fatty acyl chains relatively immobilized in a tightly packed array (Figure 9.12).

Figure 9.12 · An illustration of the gel-to-liquid crystalline phase transition, which occurs when a membrane is warmed through the transition temperature, T_m. Notice that the surface area must increase and the thickness must decrease as the membrane goes through a phase transition. The mobility of the lipid chains increases dramatically.

In this state, the anti conformation is adopted by all the carbon - carbon bonds in the lipid chains. This leaves the lipid chains in their fully extended conformation. As a result, the surface area per lipid is minimal and the bilayer thickness is maximal. Above the transition temperature, a liquid crystalline state exists in which the mobility of fatty acyl chains is intermediate between solid and liquid alkane . In this more fluid, liquid crystalline state, the carbon - carbon bonds of the lipid chains more readily adopt gauche conformations (Figure 9.13). As a result, the surface area per lipid increases and the bilayer thickness decreases by 10 to 15%.

Figure 9.13 · Membrane lipid phase transitions can be detected and characterized by measuring the rate of absorption of heat by a membrane sample in a calorimeter (see Chapter 3 for a detailed discussion of calorimetry). Pure, homogeneous bilayers (containing only a single lipid component) give sharp calorimetric peaks. Egg PC contains a variety of fatty acid chains and thus yields a broad calorimetric peak. Below the phase transition, lipid chains primarily adopt the anti conformation. Above the phase transition, lipid chains have absorbed a substantial amount of heat. This is reflected in the adoption of higher-energy conformations, including the gauche conformations shown.

     The sharpness of the transition in pure lipid preparations shows that the phase change is a cooperative behavior. This is to say that the behavior of one or a few molecules affects the behavior of many other molecules in the vicinity. The sharpness of the transition then reflects the number of molecules that are acting in concert. Sharp transitions involve large numbers of molecules all “melting” together.

            Phase transitions have been characterized in a number of different pure and mixed lipid systems. Table 9.1 shows a comparison of the transition temperatures observed for several different phosphatidylcholines with different fatty acyl chain compositions. General characteristics of bilayer phase transitions include the following:

1.  The transitions are always endothermic; heat is absorbed as the temperature increases through the transition (Figure 9.13).

2.  Particular phospholipids display characteristic transition temperatures (T_m). As shown in Table 9.1, T_m increases with chain length, decreases with unsaturation , and depends on the nature of the polar head group.

3.  For pure phospholipid bilayers , the transition occurs over a narrow temperature range. The phase transition for dimyristoyl lecithin has a peak width of about 0.2°C.

4.  Native biological membranes also display characteristic phase transitions, but these are broad and strongly dependent on the lipid and protein composition of the membrane.

5.  With certain lipid bilayers , a change of physical state referred to as a pretransition occurs 5° to 15°C below the phase transition itself. These pretransitions involve a tilting of the hydrocarbon chains.

6.  A volume change is usually associated with phase transitions in lipid bilayers .

7.  Bilayer phase transitions are sensitive to the presence of solutes that interact with lipids, including multivalent cations , lipid-soluble agents, peptides, and proteins.

Cells adjust the lipid composition of their membranes to maintain proper fluidity as environmental conditions change.

9.2 · Structure of Membrane Proteins

The lipid bilayer constitutes the fundamental structural unit of all biological membranes. Proteins, in contrast, carry out essentially all of the active functions of membranes, including transport activities, receptor functions, and other related processes. As suggested by Singer and Nicolson , most membrane proteins can be classified as peripheral or integral. The peripheral proteins are globular proteins that interact with the membrane mainly through electrostatic and hydrogen-bonding interactions with integral proteins. Although peripheral proteins are not discussed further here, many proteins of this class are described in the context of other discussions throughout this textbook. Integral proteins are those that are strongly associated with the lipid bilayer , with a portion of the protein embedded in, or extending all the way across, the lipid bilayer . Another class of proteins not anticipated by Singer and Nicolson , the lipid-anchored proteins, are important in a variety of functions in different cells and tissues. These proteins associate with membranes by means of a variety of covalently linked lipid anchors.

Integral Membrane Proteins

Despite the diversity of integral membrane proteins, most fall into two general classes. One of these includes proteins attached or anchored to the membrane by only a small hydrophobic segment, such that most of the protein extends out into the water solvent on one or both sides of the membrane. The other class includes those proteins that are more or less globular in shape and more totally embedded in the membrane, exposing only a small surface to the water solvent outside the membrane. In general, those structures of integral membrane protein within the nonpolar core of the lipid bilayer are dominated by a-helices or β-sheets because these secondary structures neutralize the highly polar N —H and C=O functions of the peptide backbone through H-bond formation.

A Protein with a Single Transmembrane Segment

In the case of the proteins that are anchored by a small hydrophobic polypeptide segment, that segment often takes the form of a single α-helix. One of the best examples of a membrane protein with such an α-helical structure is glycophorin . Most of glycophorin’s mass is oriented on the outside surface of the cell, exposed to the aqueous milieu (Figure 9.14).

Figure 9.14 · Glycophorin A spans the membrane of the human erythrocyte via a single α-helical transmembrane segment. The C-terminus of the peptide, whose sequence is shown here, faces the cytosol of the erythrocyte; the N-terminal domain is extracellular. Points of attachment of carbohydrate groups are indicated.

A variety of hydrophilic oligosaccharide units are attached to this extracellular domain. These oligosaccharide groups constitute the ABO and MN blood group antigenic specificities of the red cell. This extracellular portion of the protein also serves as the receptor for the influenza virus. Glycophorin has a total molecular weight of about 31,000 and is approximately 40% protein and 60% carbohydrate. The glycophorin primary structure consists of a segment of 19 hydrophobic amino acid residues with a short hydrophilic sequence on one end and a longer hydrophilic sequence on the other end. The 19-residue sequence is just the right length to span the cell membrane if it is coiled in the shape of an α-helix. The large hydrophilic sequence includes the amino terminal residue of the polypeptide chain.
            Numerous other membrane proteins are also attached to the membrane by means of a single hydrophobic α-helix, with hydrophilic segments extending into either the cytoplasm or the extracellular space. These proteins often function as receptors for extracellular molecules or as recognition sites that allow the immune system to recognize and distinguish the cells of the host organism from invading foreign cells or viruses. The proteins that represent the major transplantation antigens H2 in mice and human leukocyte associated (HLA) proteins in humans are members of this class. Other such proteins include the surface immunoglobulin receptors on B lymphocytes and the spike proteins of many membrane viruses. The function of many of these proteins depends primarily on their extracellular domain, and thus the segment facing the intracellular surface is often a shorter one.

Bacteriorhodopsin : A 7-Transmembrane Segment Protein

Membrane proteins that take on a more globular shape, instead of the rodlike structure previously described, are often involved with transport activities and other functions requiring a substantial portion of the peptide to be embedded in the membrane. These proteins may consist of numerous hydrophobic α-helical segments joined by hinge regions so that the protein winds in a zig-zag pattern back and forth across the membrane. A well-characterized example of such a protein is bacteriorhodopsin , which clusters in purple patches in the membrane of the bacterium Halobacterium halobium . The name Halobacterium refers to the fact that this bacterium thrives in solutions having high concentrations of sodium chloride, such as the salt beds of San Francisco Bay . Halobacterium carries out a light-driven proton transport by means of bacterio-rhodopsin , named in reference to its spectral similarities to rhodopsin in the rod outer segments of the mammalian retina. When this organism is deprived of oxygen for oxidative metabolism, it switches to the capture of energy from sunlight, using this energy to pump protons out of the cell. The proton gradient generated by such light-driven proton pumping represents potential energy, which is exploited elsewhere in the membrane to synthesize ATP.

Figure 9.15 · An electron density profile illustrating the three centers of threefold symmetry in arrays of bacterio-rhodopsin in the purple membrane of Halobacterium halobium, together with a computer-generated model showing the seven α-helical transmembrane segments in bacteriorhodopsin. ( Electron density map from Stoecknius, W., 1980. Purple membrane of halobacteria: A new light-energy converter. Accounts of Chemical Research 13:337–344. Model on right from Henderson , R., 1990. Model for the structure of bacteriorhodopsin based on high-resolution electron cryo-microscopy. Journal of Molecular Biology 213:899–929.)

            Bacteriorhodopsin clusters in hexagonal arrays (Figure 9.15) in the purple membrane patches of Halobacterium , and it was this orderly, repeating arrangement of proteins in the membrane that enabled Nigel Unwin and Richard Henderson in 1975 to determine the bacteriorhodopsin structure. The polypeptide chain crosses the membrane seven times, in seven α-helical segments, with very little of the protein exposed to the aqueous milieu. The bacteriorhodopsin structure has become a model of globular membrane protein structure. Many other integral membrane proteins contain numerous hydro-phobic sequences that, like those of bacteriorhodopsin , could form α-helical transmembrane segments. For example, the amino acid sequence of the sodium - potassium transport ATPase contains ten hydrophobic segments of length sufficient to span the plasma membrane. By analogy with bacterio-rhodopsin , one would expect that these segments form a globular hydrophobic core that anchors the ATPase in the membrane. The helical segments may also account for the transport properties of the enzyme itself.

Porins — A β-Sheet Motif for Membrane Proteins

The β-sheet is another structural motif that provides extensive hydrogen bonding for transmembrane peptide segments. Porin proteins found in the outer membranes ( OM ) of Gram-negative bacteria such as Escherichia coli, and also in the outer mitochondrial membranes of eukaryotic cells, span their respective membranes with large β-sheets. A good example is maltoporin , also known as LamB protein or lambda receptor, which participates in the entry of maltose and maltodextrins into E. coli. Maltoporin is active as a trimer . The 421-residue monomer is an aesthetically pleasing 18-strand β-barrel (Figure 9.16).

Figure 9.16 · The three-dimensional structure of maltoporin from E. coli.

The β-strands are connected to their nearest neighbors either by long loops or by β-turns (Figure 9.17). The long loops are found at the end of the barrel that is exposed to the cell exterior, whereas the turns are located on the intracellular face of the barrel. Three of the loops fold into the center of the barrel.

Figure 9.17 · The arrangement of the peptide chain in maltoporin from E. coli.

            The amino acid compositions and sequences of the β-strands in porin proteins are novel. Polar and nonpolar residues alternate along the β-strands, with polar residues facing the central pore or cavity of the barrel and non-polar residues facing out from the barrel where they can interact with the hydrophobic lipid milieu of the membrane. The smallest diameter of the porin channel is about 5 Å. Thus, a maltodextrin polymer (composed of two or more glucose units) must pass through the porin in an extended conformation (like a spaghetti strand).

Lipid-Anchored Membrane Proteins

Certain proteins are found to be covalently linked to lipid molecules. For many of these proteins, covalent attachment of lipid is required for association with  a membrane. The lipid moieties can insert into the membrane bilayer , effectively anchoring their linked proteins to the membrane. Some proteins with covalently linked lipid normally behave as soluble proteins; others are integral membrane proteins and remain membrane-associated even when the lipid is removed. Covalently bound lipid in these latter proteins can play a role distinct from membrane anchoring. In many cases, attachment to the membrane via the lipid anchor serves to modulate the activity of the protein.

            Another interesting facet of lipid anchors is that they are transient. Lipid anchors can be reversibly attached to and detached from proteins. This provides a “switching device” for altering the affinity of a protein for the membrane. Reversible lipid anchoring is one factor in the control of signal transduction pathways in eukaryotic cells (Chapter 34).
            Four different types of lipid-anchoring motifs have been found to date. These are amide-linked myristoyl anchors, thioester -linked fatty acyl anchors, thioether -linked prenyl anchors, and amide-linked glycosyl phosphatidylinositol anchors. Each of these anchoring motifs is used by a variety of membrane proteins, but each nonetheless exhibits a characteristic pattern of structural requirements.

Amide-Linked Myristoyl Anchors

Myristic acid may be linked via an amide bond to the α-amino group of the N-terminal glycine residue of selected proteins (Figure 9.18). The reaction is referred to as N-myristoylation and is catalyzed by myristoyl - CoA:protein N-myristoyltransferase , known simply as NMT. N-Myristoyl -anchored proteins include the catalytic subunit of cAMP -dependent protein kinase , the pp60^src tyrosine kinase , the phosphatase known as calcineurin B, the α-subunit of G proteins (involved in GTP-dependent transmembrane signaling events), and the gag proteins of certain retroviruses, including the HIV-1 virus that causes AIDS.

Figure 9.18 · Certain proteins are anchored to biological membranes by lipid anchors. Particularly common are the N-myristoyl – and S-palmitoyl – anchoring motifs shown here. N-Myristoylation always occurs at an N-terminal glycine residue, whereas thioester linkages occur at cysteine residues within the polypeptide chain. G-protein – coupled receptors, with seven transmembrane segments, may contain one (and sometimes two) palmitoyl anchors in thioester linkage to cysteine residues in the C-terminal segment of the protein.

Thioester -Linked Fatty Acyl Anchors

A variety of cellular and viral proteins contain fatty acids covalently bound via ester linkages to the side chains of cysteine and sometimes to serine or threonine residues within a polypeptide chain (Figure 9.18). This type of fatty acyl chain linkage has a broader fatty acid specificity than N-myristoylation . Myristate , palmitate , stearate , and oleate can all be esterified in this way, with the C₁₆ and C₁₈ chain lengths being most commonly found. Proteins anchored to membranes via fatty acyl thioesters include G-protein-coupled receptors, the surface glycoproteins of several viruses, and the transferrin receptor protein.

Thioether -Linked Prenyl Anchors

As noted in Chapter 7, polyprenyl (or simply prenyl ) groups are long-chain polyisoprenoid groups derived from isoprene units. Prenylation of proteins destined for membrane anchoring can involve either farnesyl or geranylgeranyl groups (Figure 9.19). The addition of a prenyl group typically occurs at the cysteine residue of a carboxy -terminal CAAX sequence of the target protein, where C is cysteine , A is any aliphatic residue, and X can be any amino acid. As shown in Figure 9.19, the result is a thioether -linked farnesyl or geranylgeranyl group. Once the prenylation reaction has occurred, a specific protease cleaves the three carboxy -terminal residues, and the carboxyl group of the now terminal Cys is methylated to produce an ester. All of these modifications appear to be important for subsequent activity of the prenyl -anchored protein. Proteins anchored to membranes via prenyl groups include yeast mating factors, the p21^ras protein (the protein product of the ras oncogene ; see Chapter 34), and the nuclear lamins , structural components of the lamina of the inner nuclear membrane.

Figure 9.19 · Proteins containing the C-terminal sequence CAAX can undergo prenylation reactions that place thioether-linked farnesyl or geranylgeranyl groups at the cysteine side chain. Prenylation is accompanied by removal of the AAX peptide and methylation of the carboxyl group of the cysteine residue, which has become the C-terminal residue.

Glycosyl Phosphatidylinositol Anchors

Glycosyl phosphatidylinositol , or GPI, groups are structurally more elaborate membrane anchors than fatty acyl or prenyl groups. GPI groups modify the carboxy -terminal amino acid of a target protein via an ethanolamine residue linked to an oligosaccharide, which is linked in turn to the inositol moiety of a phosphatidylinositol (Figure 9.20). The oligosaccharide typically consists of a conserved tetrasaccharide core of three mannose residues and a glucosamine , which can be altered by modifications of the mannose residues or addition of galactosyl side chains of various sizes, extra phosphoethanolamines , or additional N-acetylgalactose or mannosyl residues (Figure 9.20).

Figure 9.20 · The glycosyl phosphatidylinositol (GPI) moiety is an elaborate lipid-anchoring group. Note the core of three mannose residues and a glucosamine. Additional modifications may include fatty acids at the inositol and glycerol OOH groups.

The inositol moiety can also be modified by an additional fatty acid, and a variety of fatty acyl groups are found linked to the glycerol group. GPI groups anchor a wide variety of surface antigens, adhesion molecules, and cell surface hydrolases to plasma membranes in various eukaryotic organisms. GPI anchors have not yet been observed in prokaryotic organisms or plants.

9.3 · Membranes and Cell-Surface Polysaccharides

Bacterial Cell Walls

Some of nature’s most interesting polysaccharide structures are found in bacterial cell walls. Given the strength and rigidity provided by polysaccharide structures, it is not surprising that bacteria use such structures to provide protection for their cellular contents. Bacteria normally exhibit high internal osmotic pressures and frequently encounter variable, often hypotonic exterior conditions. The rigid cell walls synthesized by bacteria maintain cell shape and size and prevent swelling or shrinkage that would inevitably accompany variations in solution osmotic strength.

Peptidoglycan

Bacteria are conveniently classified as either Gram-positive or Gram-negative depending on their response to the so-called Gram stain. Despite substantial differences in the various structures surrounding these two types of cells, nearly all bacterial cell walls have a strong, protective peptide - polysaccharide layer called peptidoglycan . Gram-positive bacteria have a thick (approximately 25 nm) cell wall consisting of multiple layers of peptidoglycan . This thick cell wall surrounds the bacterial plasma membrane. Gram-negative bacteria, in contrast, have a much thinner (2 to 3 nm) cell wall consisting of a single layer of peptidoglycan sandwiched between the inner and outer lipid bilayer membranes. In either case, peptidoglycan , sometimes called murein (from the Latin murus for “wall”), is a continuous cross-linked structure — in essence, a single molecule — built around the cell. The structure is shown in Figure 9.21.

Figure 9.21 · The structure of peptidoglycan. The tetrapeptides linking adjacent backbone chains contain an
unusual γ-carboxyl linkage.

The backbone is a β(1→4) linked polymer of alternating N-acetylglucosamine and N-acetylmuramic acid units. This part of the structure is similar to chitin, but it is joined to a tetrapeptide , usually L-Ala·D-Glu·L-Lys·D -Ala , in which the L-lysine is linked to the γ-COOH of D-glutamate. The peptide is linked to the N-acetylmuramic acid units via its D-lactate moiety. The e-amino group of lysine in this peptide is linked to the -COOH of D-alanine of an adjacent tetrapeptide . In Gram-negative cell walls, the lysine e-amino group forms a direct amide bond with this D-alanine carboxyl (Figure 9.22). In Gram-positive cell walls, a pentaglycine chain bridges the lysine ε-amino group and the D-Ala carboxyl group.

Figure 9.22 · (a) The cross-link in Gram-positive cell walls is a pentaglycine bridge. (b) In Gram-negative cell walls, the linkage between the tetrapeptides of adjacent carbohydrate chains in peptidoglycan involves a direct amide bond between the lysine side chain of one tetrapeptide and d -alanine of the other.

Cell Walls of Gram-Negative Bacteria

In Gram-negative bacteria, the peptidoglycan wall is the rigid framework around which is built an elaborate membrane structure (Figure 9.23). The peptidoglycan layer encloses the periplasmic space and is attached to the outer membrane via a group of hydrophobic proteins. These proteins, each having 57 amino acid residues, are attached through amide linkages from the side chains of C-terminal lysines of the proteins to diaminopimelic acid groups on the peptidoglycan . Diaminopimelic acid replaces one of the d -alanine residues in about 10% of the peptides of the peptidoglycan . On the other end of the hydrophobic protein, the N-terminal residue, a serine, makes a covalent bond to a lipid that is part of the outer membrane.

Figure 9.23 · The structures of the cell wall and membrane(s) in Gram-positive and Gram-negative bacteria. The Gram-positive cell wall is thicker than that in Gram-negative bacteria, compensating for the absence of a second (outer) bilayer membrane.

            As shown in Figure 9.24, the outer membrane of Gram-negative bacteria is coated with a highly complex lipopolysaccharide , which consists of a lipid group (anchored in the outer membrane) joined to a polysaccharide made up of long chains with many different and characteristic repeating structures (Figure 9.24). These many different unique units determine the antigenicity of the bacteria; that is, animal immune systems recognize them as foreign substances and raise antibodies against them. As a group, these antigenic determinants are called the O antigens, and there are thousands of different ones. The Salmonella bacteria alone have well over a thousand known O antigens that have been organized into 17 different groups. The great variation in these O antigen structures apparently plays a role in the recognition of one type of cell by another and in evasion of the host immune system.

Figure 9.24 · Lipopolysaccharide (LPS) coats the outer membrane of Gram-negative bacteria. The lipid portion of the LPS is embedded in the outer membrane and is linked to a complex polysaccharide.

Cell Walls of Gram-Positive Bacteria

In Gram-positive bacteria, the cell exterior is less complex than for Gram-nega-tive cells. Having no outer membrane, Gram-positive cells compensate with a thicker wall. Covalently attached to the peptidoglycan layer are teichoic acids, which often account for 50% of the dry weight of the cell wall (Figure 9.25). The teichoic acids are polymers of ribitol phosphate or glycerol phosphate linked by phosphodiester bonds. In these heteropolysaccharides , the free hydroxyl groups of the ribitol or glycerol are often substituted by glycosidically linked monosaccharides (often glucose or N-acetylglucosamine ) or disaccharides. D-Alanine is sometimes found in ester linkage to the saccharides. Teichoic acids are not confined to the cell wall itself, and they may be present in the inner membranes of these bacteria. Many teichoic acids are antigenic, and they also serve as the receptors for bacteriophages in some cases.

Figure 9.25 · Teichoic acids are covalently linked to the peptidoglycan of Gram-positive bacteria. These polymers of (a, b) glycerol phosphate or (c) ribitol phosphate are linked by phosphodiester bonds.

Cell Surface Polysaccharides

Compared to bacterial cells, which are identical within a given cell type (except for O antigen variations), animal cells display a wondrous diversity of structure, constitution, and function. Although each animal cell contains, in its genetic material, the instructions to replicate the entire organism, each differentiated animal cell carefully controls its composition and behavior within the organism. A great part of each cell’s uniqueness begins at the cell surface. This surface uniqueness is critical to each animal cell because cells spend their entire life span in intimate contact with other cells and must therefore communicate with one another. That cells are able to pass information among themselves is evidenced by numerous experiments. For example, heart myocytes , when grown in culture (in glass dishes) establish synchrony when they make contact, so that they “beat” or contract in unison. If they are removed from the culture and separated, they lose their synchronous behavior, but if allowed to reestablish cell-to-cell contact, they spontaneously restore their synchronous contractions. Kidney cells grown in culture with liver cells seek out and make contact with other kidney cells and avoid contact with liver cells. Cells grown in culture grow freely until they make contact with one another, at which point growth stops, a phenomenon well known as contact inhibition. One important characteristic of cancerous cells is the loss of contact inhibition.
            As these and many other related phenomena show, it is clear that molecular structures on one cell are recognizing and responding to molecules on the adjacent cell or to molecules in the extracellular matrix, the complex “soup” of connective proteins and other molecules that exists outside of and among cells. Many of these interactions involve glycoproteins on the cell surface and proteoglycans in the extracellular matrix. The “information” held in these special carbohydrate-containing molecules is not encoded directly in the genes (as with proteins), but is determined instead by expression of the appropriate enzymes that assemble carbohydrate units in a characteristic way on these mole-cules . Also, by virtue of the several hydroxyl linkages that can be formed with each carbohydrate monomer, these structures can be more information-rich than proteins and nucleic acids, which can form only linear polymers. A few of these glycoproteins and their unique properties are described in the following sections.

9.4 · Glycoproteins

Many proteins found in nature are glycoproteins because they contain covalently linked oligo - and polysaccharide groups. The list of known glycoproteins includes structural proteins, enzymes, membrane receptors, transport proteins, and immunoglobulins, among others. In most cases, the precise function of the bound carbohydrate moiety is not understood.

Figure 9.26 · The carbohydrate moieties of glycoproteins may be linked to the protein via (a) serine or threonine residues (in the O-linked saccharides) or (b) asparagine residues (in the N-linked saccharides). (c) N-Linked glycoproteins are of three types: high mannose, complex, and hybrid, the latter of which combines structures found in the high mannose and complex saccharides.

            Carbohydrate groups may be linked to polypeptide chains via the hydroxyl groups of serine, threonine , or hydroxylysine residues (in O-linked saccharides ) (Figure 9.26a) or via the amide nitrogen of an asparagine residue (in N-linked saccharides ) (Figure 9.26b). The carbohydrate residue linked to the protein in O-linked saccharides is usually an N-acetylgalactosamine , but mannose, galactose , and xylose residues linked to protein hydroxyls are also found (Figure 9.26a). Oligosaccharides O-linked to glycophorin (see Figure 9.14) involve N-acetylgalactosamine linkages and are rich in sialic acid residues (Figure 9.14). N-linked saccharides always have a unique core structure composed of two N-acetylglucosamine residues linked to a branched mannose triad (Figure 9.26b, c). Many other sugar units may be linked to each of the mannose residues of this branched core.
            O-Linked saccharides are often found in cell surface glycoproteins and in mucins , the large glycoproteins that coat and protect mucous membranes in the respiratory and gastrointestinal tracts in the body. Certain viral glycoproteins also contain O-linked sugars. O-Linked saccharides in glycoproteins are often found clustered in richly glycosylated domains of the polypeptide chain. Physical studies on mucins show that they adopt rigid, extended structures so that an individual mucin molecule (M_r =10⁷) may extend over a distance of 150 to 200 nm in solution. Inherent steric interactions between the sugar residues and the protein residues in these cluster regions cause the peptide core to fold into an extended and relatively rigid conformation. This interesting effect may be related to the function of O-linked saccharides in glycoproteins . It allows aggregates of mucin molecules to form extensive, intertwined networks, even at low concentrations. These viscous networks protect the mucosal surface of the respiratory and gastrointestinal tracts from harmful environmental agents.
            There appear to be two structural motifs for membrane glycoproteins containing O-linked saccharides . Certain glycoproteins , such as leukosialin , are O-glycosylated throughout much or most of their extracellular domain (Figure 9.27).

Figure 9.27 · The O-linked saccharides of glycoproteins appear in many cases to adopt extended conformations that serve to extend the functional domains of these proteins above the membrane surface. ( Adapted from Jentoft, N., 1990, Trends in Biochemical Sciences 15:291–294.)

Leukosialin , like mucin , adopts a highly extended conformation, allowing it to project great distances above the membrane surface, perhaps protecting the cell from unwanted interactions with macromolecules or other cells. The second structural motif is exemplified by the low density lipoprotein (LDL) receptor and by decay accelerating factor (DAF). These proteins contain a highly O-glycosylated stem region that separates the transmembrane domain from the globular, functional extracellular domain. The O-glycosylated stem serves to raise the functional domain of the protein far enough above the membrane surface to make it accessible to the extracellular macromolecules with which it interacts.

Antifreeze Glycoproteins

A unique family of O-linked glycoproteins permits fish to live in the icy seawater of the Arctic and Antarctic regions where water temperature may reach as low as -1.9°C. Antifreeze glycoproteins (AFGPs ) are found in the blood of nearly all Antarctic fish and at least five Arctic fish. These glycoproteins have the peptide structure

                                                    [Ala-Ala-Thr]_n-Ala-Ala

where n can be 4, 5, 6, 12, 17, 28, 35, 45, or 50. Each of the threonine residues is glycosylated with the disaccharide β-galactosyl -(1→3)-α-N-acetylgalactosamine (Figure 9.28). This glycoprotein adopts a flexible rod conformation with regions of threefold left-handed helix. The evidence suggests that antifreeze glycoproteins may inhibit the formation of ice in the fish by binding specifically to the growth sites of ice crystals, inhibiting further growth of the crystals.

Figure 9.28 · The structure of the repeating unit of antifreeze glycoproteins, a disaccharide consisting of β-galactosyl-(1ℯ3)-α-N-acetylgalactosamine in glycosidic linkage to a threonine residue.

N-Linked Oligosaccharides

N-Linked oligosaccharides are found in many different proteins, including immunoglobulins G and M, ribonuclease B, ovalbumin , and peptide hormones (Figure 9.29). Many different functions are known or suspected for N-glycosylation of proteins. Glycosylation can affect the physical and chemical properties of proteins, altering solubility, mass, and electrical charge. Carbohydrate moieties have been shown to stabilize protein conformations and protect proteins against proteolysis. Eukaryotic organisms use posttranslational additions of N-linked oligosaccharides to direct selected proteins to various intracellular organelles.

Figure 9.29 · Some of the oligosaccharides found in N-linked glycoproteins.

Oligosaccharide Cleavage as a Timing Device for Protein Degradation

The slow cleavage of monosaccharide residues from N-linked glycoproteins circulating in the blood targets these proteins for degradation by the organism. The liver contains specific receptor proteins that recognize and bind glycoproteins that are ready to be degraded and recycled. Newly synthesized serum glycoproteins contain N-linked triantennary (three-chain) oligosaccharides having structures similar to those in Figure 9.30, in which sialic acid residues cap galactose residues.

Figure 9.30 · Progressive cleavage of sialic acid residues exposes galactose residues. Binding to the asialoglycoprotein receptor in the liver becomes progressively more likely as more Gal residues are exposed.

As these glycoproteins circulate, enzymes on the blood vessel walls cleave off the sialic acid groups, exposing the galactose residues. In the liver, the asialoglycoprotein receptor binds the exposed galactose residues of these glycoproteins with very high affinity (K_D = 10^-9 to 10^-8 M). The complex of receptor and glycoprotein is then taken into the cell by endocytosis , and the glycoprotein is degraded in cellular lysosomes . Highest affinity binding of glycoprotein to the asialoglycoprotein receptor requires three free galactose residues. Oligosaccharides with only one or two exposed galactose residues bind less tightly. This is an elegant way for the body to keep track of how long glycoproteins have been in circulation. Over a period of time, anywhere from a few hours to weeks, the sialic acid groups are cleaved one by one. The longer the glycoprotein circulates and the more sialic acid residues are removed, the more galactose residues become exposed so that the glycoprotein is eventually bound to the liver receptor.

9.5 · Proteoglycans

Proteoglycans are a family of glycoproteins whose carbohydrate moieties are predominantly glycosaminoglycans . The structures of only a few proteoglycans are known, and even these few display considerable diversity (Figure 9.31).

Figure 9.31 · The known proteoglycans include a variety of structures. The carbohydrate groups of proteoglycans are predominantly glycosaminoglycans O-linked to serine residues. Proteoglycans include both soluble proteins and integral transmembrane proteins.

They range in size from serglycin , having 104 amino acid residues (10.2 kD ) to versican , having 2409 residues (265 kD ). Each of these proteoglycans contains one or two types of covalently linked glycosaminoglycans (Table 9.2). In the known proteoglycans, the glycosaminoglycan units are O-linked to serine residues of Ser-Gly dipeptide sequences. Serglycin is named for a unique central domain of 49 amino acids composed of alternating serine and glycine residues. The cartilage matrix proteoglycan contains 117 Ser-Gly pairs to which chondroitin sulfates attach. Decorin , a small proteoglycan secreted by fibro-blasts and found in the extracellular matrix of connective tissues, contains only three Ser-Gly pairs, only one of which is normally glycosylated . In addition to glycosaminoglycan units, proteoglycans may also contain other N-linked and O-linked oligosaccharide groups.

Functions of Proteoglycans

Proteoglycans may be soluble and located in the extracellular matrix, as is the case for serglycin , versican , and the cartilage matrix proteoglycan , or they may be integral transmembrane proteins, such as syndecan . Both types of proteoglycan appear to function by interacting with a variety of other molecules through their glycosaminoglycan components and through specific receptor domains in the polypeptide itself. For example, syndecan (from the Greek syndein meaning “to bind together”) is a transmembrane proteoglycan that associates intracellularly with the actin cytoskeleton (Chapter 17). Outside the cell, it interacts with fibronectin , an extracellular protein that binds to several cell surface proteins and to components of the extracellular matrix. The ability of syndecan to participate in multiple interactions with these target molecules allows them to act as a sort of “glue” in the extracellular space, linking components of the extracellular matrix, facilitating the binding of cells to the matrix, and mediating the binding of growth factors and other soluble molecules to the matrix and to cell surfaces (Figure 9.32).

Figure 9.32 · Proteoglycans serve a variety of functions on the cytoplasmic and extracellular surfaces of the plasma membrane. Many of these functions appear to involve the binding of specific proteins to the glycosaminoglycan groups.

            Many of the functions of proteoglycans involve the binding of specific proteins to the glycosaminoglycan groups of the proteoglycan . The glycosaminoglycan binding sites on these specific proteins contain multiple basic amino acid residues. The amino acid sequences BBXB and BBBXXB (where B is a basic amino acid and X is any amino acid) recur repeatedly in these binding domains. Basic amino acids such as lysine and arginine provide charge neutralization for the negative charges of glycosaminoglycan residues, and in many cases, the binding of extracellular matrix proteins to glycosaminoglycans is primarily charge-dependent. For example, more highly sulfated glycosaminoglycans bind more tightly to fibronectin . Certain protein- glycosaminoglycan interactions, however, require a specific carbohydrate sequence. A particular pentasaccharide sequence in heparin, for example, binds tightly to antithrombin III (Figure 9.33), accounting for the anticoagulant properties of heparin. Other glycosaminoglycans interact much more weakly.

Figure 9.33 · A portion of the structure of heparin, a carbohydrate having anticoagulant properties. It is used by blood banks to prevent the clotting of blood during donation and storage and also by physicians to prevent the formation of life-threatening blood clots in patients recovering from serious injury or surgery. This sulfated pentasaccharide sequence in heparin binds with high affinity to antithrombin III, accounting for this anticoagulant activity. The 3-O-sulfate marked by an asterisk is essential for high-affinity binding of heparin to antithrombin III.

Proteoglycans May Modulate Cell Growth Processes

Several lines of evidence raise the possibility of modulation or regulation of cell growth processes by proteoglycans . First, heparin and heparan sulfate are known to inhibit cell proliferation in a process involving internalization of the glycosaminoglycan moiety and its migration to the cell nucleus. Second, fibro-blast growth factor binds tightly to heparin and other glycosaminoglycans , and the heparin - growth factor complex protects the growth factor from degradative enzymes, thus enhancing its activity. There is evidence that binding of fibro-blast growth factors by proteoglycans and glycosaminoglycans in the extracellular matrix creates a reservoir of growth factors for cells to use. Third, transforming growth factor β has been shown to stimulate the synthesis and secretion of proteoglycans in certain cells. Fourth, several proteoglycan core proteins, including versican and lymphocyte homing receptor, have domains similar in sequence to epidermal growth factor and complement regulatory factor. These growth factor domains may interact specifically with growth factor receptors in the cell membrane in processes that are not yet understood.

Proteoglycans Make Cartilage Flexible and Resilient

Cartilage matrix proteoglycan is responsible for the flexibility and resilience of cartilage tissue in the body. In cartilage, long filaments of hyaluronic acid are studded or coated with proteoglycan molecules, as shown in Figure 9.34. The hyaluronate chains can be as long as 4 μm and can coordinate 100 or more proteoglycan units. Cartilage proteoglycan possesses a hyaluronic acid binding domain on the NH₂-terminal portion of the polypeptide, which binds to hyaluronate with the assistance of a link protein. The proteoglycan - hyaluronate aggregates can have molecular weights of 2 million or more.

Figure 9.34 · Hyaluronate (see Figure 7.33) forms the backbone of proteoglycan structures, such as those found in cartilage. The proteoglycan subunits consist of a core protein containing numerous O-linked and N-linked glycosaminoglycans. In cartilage, these highly hydrated proteoglycan structures are enmeshed in a network of collagen fibers. Release (and subsequent reabsorption) of water by these structures during compression accounts for the shock-absorbing qualities of cartilaginous tissue.

            The proteoglycan - hyaluronate aggregates are highly hydrated by virtue of strong interactions between water molecules and the polyanionic complex. When cartilage is compressed (such as when joints absorb the impact of walking or running), water is briefly squeezed out of the cartilage tissue and then reabsorbed when the stress is diminished. This reversible hydration gives cartilage its flexible, shock-absorbing qualities and cushions the joints during physical activities that might otherwise injure the involved tissues.