Chapter 13:  Intracellular Vesicular Traffic

There Are Various Types of Coated Vesicles

Most transport vesicles form from specialized, coated regions of membranes. They bud off as coated vesicles that have a distinctive cage of proteins covering their cytosolic surface. Before the vesicle fuses with a target membrane, the coat is discarded, as is required to allow the two cytosolic membrane surfaces to interact directly and fuse.

The coat is thought to perform two principal functions. First, it concentrates specific membrane proteins in a specialized membrane patch that then gives rise to the vesicle membrane. It thus helps select the appropriate molecules for transport. Second, the assembly of the coat proteins into curved, basketlike lattices deforms the membrane patch and thereby molds the forming vesicles, which explains why vesicles with the same type of coat have a relatively uniform size.

There are three well-characterized types of coated vesicles, which differ in their coat proteins: clathrin-coated, COPI-coated, and COPII-coated vesicles (Figure 13-4). Each type is used for different transport steps in the cell. Clathrin-coated vesicles, for example, mediate transport from the Golgi apparatus and from the plasma membrane, whereas COPI- and COPII-coated vesicles most commonly mediate transport from the ER and the Golgi cisternae (Figure 13-5). There is, however, much more variety than this short list suggests. As we discuss below, there are at least three types of clathrin-coated vesicles, each specialized for a different transport step, and the COPI-coated vesicles may be similarly diverse. Moreover, still other coats have been seen in the electron microscope, whose molecular compositions and functions are not yet known.

The Assembly of a Clathrin Coat Drives Vesicle Formation

Clathrin-coated vesicles were the first coated vesicles discovered and have been the most thoroughly studied. They provide a good example of how vesicles form.

The major protein component of clathrin-coated vesicles is clathrin itself. Each clathrin subunit consists of three large and three small polypeptide chains that together form a three-legged structure called a triskelion. Clathrin triskelions assemble into a basketlike convex framework of hexagons and pentagons to form coated pits on the cytosolic surface of membranes (Figure 13-6). Under appropriate conditions, isolated triskelions spontaneously self-assemble into typical polyhedral cages in a test tube, even in the absence of the membrane vesicles that these baskets normally enclose (Figure 13-7). Thus, the geometry of the clathrin cage is determined by the clathrin triskelion alone.

A second major coat protein in clathrin-coated vesicles is a multisubunit complex called adaptin. It is required both to bind the clathrin coat to the membrane and to trap various transmembrane proteins, including transmembrane receptors that capture soluble cargo molecules inside the vesicle—so-called cargo receptors. In this way, a selected set of membrane proteins and the soluble proteins that interact with them are packaged into each newly formed clathrin-coated transport vesicle (Figure 13-8).

There are at least four types of adaptins, each specific for a different set of cargo receptors. Clathrin-coated vesicles budding from different membranes use different adaptins and thus package different receptors and cargo molecules. The formation of a clathrin-coated pit is driven by forces generated by the successive assembly of adaptins and the clathrin coat on the cytosolic surface of the membrane. The lateral interactions between adaptins and between clathrin molecules then aid in bud formation.

Both The Pinching-off and Uncoating of Coated Vesicles Are Regulated Processes

As a clathrin-coated bud grows, soluble cytoplasmic proteins, including dynamin, assemble as a ring around the neck of each bud. Dynamin is a GTPase, which regulates the rate with which vesicles pinch off from the membrane. In the pinching-off process, the two noncytosolic leaflets of the membrane are brought into close proximity and fuse, sealing off the forming vesicle (Figure 13-9). To perform this task, dynamin recruits other proteins to the neck of the budding vesicle, which together with dynamin help to bend the membrane, either by directly distorting the bilayer structure locally or by changing the lipid composition, or both. A local change in lipid composition may result from the action of lipid-modifying enzymes that are recruited into the dynamin complex.

Once the vesicle is released from the membrane, the clathrin coat is rapidly lost. A chaperone protein of the hsp70 family functions as an uncoating ATPase, using the energy of ATP hydrolysis to peel off the coat. Another protein called auxillin, which is attached to the vesicle, is believed to activate the ATPase. Because the coated bud persists much longer than the coat on the vesicle, additional control mechanisms must somehow prevent the coat from being removed before it has formed a vesicle (discussed below).

Although there are many similarities in vesicle budding at various locations in the cell, each cell membrane poses its own special challenges. The plasma membrane, for example, is comparatively flat and stiff, owing to its cholesterol-rich lipid composition and underlying cortical cytoskeleton. Thus, clathrin coats have to produce considerable force to introduce curvature, especially at the neck of the bud where dynamin and its associated proteins facilitate the sharp bends required for the pinching-off of the vesicle. In contrast, vesicle budding from many intracellular membranes occurs preferentially at regions where the membranes are already curved, such as the rims of Golgi cisternae or membrane tubules.

COPI-coated vesicles and COPII-coated vesicles transport material early in the secretory pathway: COPII-coated packages bud from the ER, and COPI-coated packages bud from pre-Golgi compartments and Golgi cisternae (see Figure 13-5). The coats of COPI and COPII vesicles consist, in part, of large protein complexes that are composed of seven individual coat-protein subunits for COPI and four for COPII coats. Some COPI coat-protein subunits show sequence similarity to adaptins, suggesting a common evolutionary origin.

Not All Transport Vesicles are Spherical

Transport vesicles occur in various sizes and shapes. When living cells that have been genetically engineered to express fluorescent membrane components are observed under the microscope, endosomes and the trans Golgi network are seen to continually send out long tubules. Coat proteins assemble onto the tubules and help recruit specific cargo. The tubules then either withdraw, or they pinch off with the help of dynamin-like proteins and thus can serve as transport vesicles. Depending on the relative efficiencies of membrane tubulation and severing, differently sized portions of a donor organelle can pinch off.

Tubules have a much higher surface-to-volume ratio than the organelles from which they form. They are therefore relatively enriched in membrane proteins compared with soluble cargo proteins. As we discuss later, this property of tubules is used for sorting proteins in endosomes. Thus, vesicular transport does not necessarily occur only through uniformly sized spherical vesicles, but can involve larger portions of a donor organelle.

Monomeric GTPases Control Coat Assembly

The vesicular transport performed by both clathrin-coated and COP-coated vesicles depends on a variety of GTP-binding proteins that control both the spatial and the temporal aspects of membrane exchange. As discussed in Chapter 3, large families of GTP-binding proteins regulate diverse processes within cells. These proteins act as molecular switches that flip between an active state with GTP bound and an inactive state with GDP bound. Two classes of proteins regulate the flipping: guanine-nucleotide-exchange factors (GEFs) activate the proteins by catalyzing the exchange of GDP for GTP, and GTPase-activating proteins (GAPs) inactivate the proteins by triggering the hydrolysis of the bound GTP to GDP (see Figure 3-72). Although both monomeric GTP-binding proteins (monomeric GTPases) and trimeric GTP-binding proteins (G proteins) have essential roles in vesicular transport, the roles of the monomeric GTPases are better understood, and we focus our discussion on them.

To ensure that membrane traffic to and from an organelle is balanced, coat proteins must assemble only when and where they are needed. Coat-recruitment GTPases, which are members of a family of monomeric GTPases, usually serve this function. They include the ARF proteins, which are responsible for both COPI coat assembly and clathrin coat assembly at Golgi membranes, and the Sar1 protein, which is responsible for COPII coat assembly at the ER membrane. Clathrin coat assembly at the plasma membrane is also thought to involve a GTPase, but its identity is unknown.

Coat-recruitment GTPases are usually found in high concentration in the cytosol in an inactive, GDP-bound state. When a COPII-coated vesicle is to bud from the ER membrane, a specific GEF embedded in the ER membrane binds to cytosolic Sar1, causing the Sar1 to release its GDP and bind GTP in its place (recall that GTP is present in much higher concentration in the cytosol than GDP and therefore will spontaneously bind after GDP is released). In its GTP-bound state, Sar1 exposes a hydrophobic tail, which inserts into the lipid bilayer of the ER membrane. The tightly bound Sar1 now recruits coat protein subunits to the ER membrane to initiate budding (Figure 13-10). Other GEFs and coat-recruitment GTPases operate in a similar way on other membranes.

Some coat protein subunits also interact, albeit more weakly, with the head groups of certain lipid molecules, in particular phosphatidic acid and phosphoinositides, as well as with the cytoplasmic tails of some of the membrane proteins they recruit into the bud. Activated coat-recruitment GTPases at sites of bud formation can locally activate phospholipase D, which converts some phospholipids to phosphatidic acid, thereby enhancing the binding of coat proteins. Together, these protein-protein and protein-lipid interactions tightly bind the coat to the membrane, causing the membrane to deform into a bud, which then pinches off as a coated vesicle.

The coat-recruitment GTPases also have a role in coat disassembly. The hydrolysis of bound GTP to GDP causes the GTPase to change its conformation so that its hydrophobic tail pops out of the membrane, causing the vesicle's coat to disassemble. Although it is not known what triggers the GTP hydrolysis process, it has been proposed that the GTPases work like timers, which hydrolyze GTP at a slow but predictable rate. COPII coats, for example, accelerate GTP hydrolysis by Sar1, thereby triggering coat disassembly at a certain time after coat assembly has begun. Thus, a fully formed vesicle will be produced only when bud formation occurs faster than the timed disassembly process; otherwise, disassembly will be triggered before a vesicle pinches off, and the process will have to start again at a more appropriate time and place. Completion of coating or contact with the target membrane may also trigger coat disassembly.

SNARE Proteins and Targeting GTPases Guide Membrane Transport

To ensure that membrane traffic proceeds in an orderly way, transport vesicles must be highly selective in recognizing the correct target membrane with which to fuse. Because of the diversity of membrane systems, a vesicle is likely to encounter many potential target membranes before it finds the correct one. Specificity in targeting is ensured because all transport vesicles display surface markers that identify them according to their origin and type of cargo, while target membranes display complementary receptors that recognize the appropriate markers. This crucial recognition step is thought to be controlled mainly by two classes of proteins: SNAREs and targeting GTPases called Rabs. SNARE proteins seem to have a central role both in providing specificity and in catalyzing the fusion of vesicles with the target membrane. Rabs seem to work together with other proteins to regulate the initial docking and tethering of the vesicle to the target membrane.

There are at least 20 different SNAREs in an animal cell, each associated with a particular membrane-enclosed organelle involved in the biosynthetic-secretory or endocytic pathway. These transmembrane proteins exist as complementary sets—vesicle membrane SNAREs, called v-SNAREs, and target membrane SNAREs, called t-SNAREs, (Figures 13-11 and 13-12). v-SNARESs and t-SNAREs have characteristic helical domains. When a v-SNARE interacts with a t-SNARE, the helical domains of one wrap around the helical domains of the other to form stable trans-SNARE complexes, which lock the two membranes together. We discuss later how the trans-SNARE complex is thought to contribute to membrane fusion. The specificity with which SNAREs interact determines the specificity of vesicle docking and fusion. In this way SNAREs specify compartment identity and govern the orderly transfer of material during vesicular transport.

SNAREs have been best characterized in nerve cells, where they mediate the docking and fusion of synaptic vesicles at the nerve terminal plasma membrane (see Figure 13-12). The SNARE complexes at neuron terminals are the targets of powerful neurotoxins that are secreted by the bacteria that cause tetanus and botulism. These toxins are highly specific proteases that enter specific neurons, cleave SNARE proteins in the nerve terminals and thereby block synaptic transmissions, often fatally.

Interacting SNAREs Need To Be Pried Apart Before They Can Function Again

Most SNARE proteins in cells have already participated in multiple rounds of membrane targeting and are sometimes present in a membrane as stable complexes with one or two partner SNAREs (see Figure 13-11). The complexes have to be disassembled before the SNAREs can mediate new rounds of transport. A crucial protein called NSF cycles between membranes and the cytosol and cata-lyzes the disassembly process. It is an ATPase that structurally resembles a minor class of cytosolic chaperone proteins that use the energy of ATP hydrolysis to solubilize and help refold denatured proteins. Similarly, NSF uses ATP to unravel the coiled-coil interaction between the helical domains of SNARE proteins, using several adaptor proteins to bind to the SNAREs (Figure 13-13).

The requirement for SNARE complex disassembly may help explain why membranes do not fuse indiscriminately in cells. If the t-SNAREs in a target membrane were always active, then any membrane containing an appropriate v-SNARE would fuse whenever the two membranes made contact. The requirement for NSF-mediated reactivation of SNAREs allows the cell to control when and where membranes fuse. In addition, t-SNAREs in target membranes are often associated with inhibitory proteins that must be released before the t-SNARE can function. This release step may be controlled by the targeting GTPases, as we discuss next.

Rab Proteins Help Ensure the Specificity of Vesicle Docking

Rab proteins make an important contribution to the specificity of vesicular transport. They are monomeric GTPases, and with over 30 known members, they are the largest subfamily of these GTPases. Like the SNAREs, each Rab protein has a characteristic distribution on cell membranes and every organelle has at least one Rab protein on its cytosolic surface (Table 13-1). Rab proteins are thought to facilitate and regulate the rate of vesicle docking and the matching of v-SNAREs and t-SNAREs, as required for membrane fusion.

Like the coat-recruitment GTPases discussed earlier (see Figure 13-10), Rab proteins cycle between a membrane and the cytosol. In their GDP-bound state they are inactive and in the cytosol, and in their GTP-bound state they are active and associated with the membrane of an organelle or transport vesicle (Figure 13-14). Many transport vesicles only form if a proper complement of SNARE and Rab proteins are included in the membrane, so as to allow the vesicle to dock and fuse appropriately.

The amino acid sequences of Rab proteins are most dissimilar near their C-terminal tails. Tail-swapping experiments indicate that the tail determines the intracellular location of each family member, presumably by enabling the protein to bind to complementary proteins, including GEFs, on the surface of the appropriate organelle. Once in its GTP-bound state and membrane-bound through a lipid anchor, a Rab protein is thought to bind to other proteins (called Rab effectors) that facilitate the docking process.

In contrast to the highly conserved structure of Rab proteins, the structures of Rab effectors vary greatly from one Rab protein to the next. One Rab effector, for example, is a large protein complex that serves to direct vesicles to specific sites on the plasma membrane for exocytosis. Vesicle fusion is limited to the region where this complex resides, even though the required t-SNAREs are uniformly distributed in the membrane. Some Rab effectors are long, filamentous, tethering proteins, which may restrict the movement of vesicles between adjacent Golgi cisternae. Others bind to Rab proteins in their active GTP-bound state and prevent premature GTP hydrolysis. Yet others are motor proteins that propel vesicles along actin filaments or microtubules to their proper target.

Although the Rab proteins and their effectors use widely different molecular mechanisms to influence vesicular transport, they have a common function. They help concentrate and tether vesicles near their target site and trigger the release of SNARE control proteins. In this way Rab proteins speed up the process by which appropriate SNARE proteins in two membranes find each other. Some Rab proteins function on the vesicle, whereas others function on the target membrane. The pairing of v-SNAREs and t-SNAREs then locks the docked vesicle onto the target membrane, readying it for fusion, which we discuss next. After fusion, the Rab protein hydrolyzes its bound GTP and the inactive GDP-bound protein returns to the cytosol to participate in another cycle of transport.

SNAREs May Mediate Membrane Fusion

Once a transport vesicle has recognized its target membrane and docked there, it unloads its cargo by membrane fusion. Fusion does not always follow immediately, however. As we discuss later, in the process of regulated exocytosis, fusion is delayed until it is triggered by a specific extracellular signal.

Thus, docking and fusion are two distinct and separable processes. Docking requires only that the two membranes come close enough for proteins protruding from the lipid bilayers to interact and adhere. Fusion requires a much closer approach, bringing the lipid bilayers to within 1.5 nm of each other so that they can join. When the membranes are in such close apposition, lipids can flow from one bilayer to the other. For this close approach, water must be displaced from the hydrophilic surface of the membrane—a process that is energetically highly unfavorable. It seems likely that all membrane fusions in cells are catalyzed by specialized fusion proteins that provide a way to overcome this energy barrier. We have already discussed the role of dynamin in a related task during clathrin-coated vesicle budding (see Figure 13-9).

SNAREs are thought to have a central role in membrane fusion. The formation of the SNARE complex may work like a winch, using the energy that is freed when the interacting helices wrap around each other to pull the membrane faces together, while simultaneously squeezing out water molecules from the interface (Figure 13-15). When liposomes containing purified v-SNAREs are mixed with liposomes containing matching t-SNAREs, their membranes fuse, albeit slowly. In the cell, other proteins recruited to the fusion site presumably cooperate with SNAREs to initiate fusion. Moreover, inhibitory proteins may have to be released to allow the complete zipping-up of SNARE pairs. In some cases, such as in regulated exocytosis (discussed later), a localized influx of Ca²⁺ triggers the fusion process.

Viral Fusion Proteins and SNAREs May Use Similar Strategies

Membrane fusion is important in other processes beside vesicular transport. Examples are the fusion of the plasma membranes of sperm and egg that occurs at fertilization (discussed in Chapter 20) and the fusion of myoblasts during muscle cell development (discussed in Chapter 21). All cell membrane fusions require special proteins and are subject to tight controls, which ensure that only appropriate membranes fuse. The controls are crucial for maintaining both the identity of cells and the individuality of each type of intracellular compartment.

The membrane fusions catalyzed by viral fusion proteins are the best understood. These proteins have a crucial role in permitting the entry of enveloped viruses (which have a lipid-bilayer-based membrane coat) into the cells that they infect (discussed in Chapters 5 and 25, CD). For example, viruses—such as human immunodeficiency virus (HIV), which causes AIDS—bind to cell-surface receptors, and then the viral and plasma membranes fuse (Figure 13-16). This fusion event allows the viral nucleic acid to enter the cytosol, where it replicates. Other viruses, such as influenza virus, first enter the cell by receptor-mediated endocytosis (discussed later) and are delivered to endosomes. In this case, the low pH in endosomes activates a fusion protein in the viral envelope that catalyzes the fusion of the viral and endosomal membranes. This likewise releases the viral nucleic acid into the cytosol.

The three-dimensional structures of the fusion proteins of influenza virus and HIV provide valuable insights into the molecular mechanism of the membrane fusion catalyzed by these proteins. An exposure of the influenza fusion protein to low pH, or an exposure of HIV fusion protein to receptors on the target cell membrane, uncovers previously buried hydrophobic regions. These regions, called fusion peptides, are observed to then insert directly into the hydrophobic core of lipid bilayer of the target membrane. Thus, the viral fusion proteins are, for a moment, integral membrane proteins in two separate lipid bilayers. Structural rearrangements in the fusion proteins then bring the two lipid bilayers into very close apposition and destabilize them so that the bilayers fuse (see Figure 13-16). For viral fusion, the fusion proteins are the only components required, supporting the possibility that SNAREs are also the central players in the process of bilayer fusion in cells.

Summary

The differences between the many different membrane-enclosed compartments in a eucaryotic cell are maintained by directed, selective transport of particular membrane components from one compartment to another. Transport vesicles, which can be spherical or tubular, bud from specialized coated regions of the donor membrane. The assembly of the coat helps to collect specific membrane and soluble cargo molecules for transport and to drive the formation of the vesicle.

Of the various types of coated vesicles, the best characterized are clathrin-coated vesicles, which mediate transport from the plasma membrane and the trans Golgi network, and COPI- and COPII-coated vesicles, which mediate transport between the ER and the Golgi apparatus and between Golgi cisternae. In clathrin-coated vesicles, adaptins link the clathrin to the vesicle membrane and also trap specific cargo molecules for packaging into the vesicle. The coat is shed rapidly after budding, which is necessary for a vesicle to fuse with its appropriate target membrane.

Monomeric GTPases help regulate various steps in vesicular transport, including both vesicle budding and docking. The coat-recruitment GTPases, including Sar1 and the ARF proteins, regulate coat assembly and disassembly. A family of Rab proteins functions as vesicle targeting GTPases. Being incorporated with v-SNAREs into budding transport vesicles, the Rab proteins help ensure that the vesicles deliver their contents only to the appropriate membrane-enclosed compartment: the one that displays complementary t-SNARE proteins. Complementary v-SNARE and t-SNARE proteins form stable trans-SNARE complexes, thereby bringing their membrane bilayers into close apposition for fusion.

Proteins Leave the ER in COPII-coated Transport Vesicles

To initiate their journey along the biosynthetic-secretory pathway, proteins that have entered the ER and are destined for the Golgi apparatus or beyond are first packaged into small COPII-coated transport vesicles. These transport vesicles bud from specialized regions of the ER called ER exit sites, whose membrane lacks bound ribosomes. In most animal cells, ER exit sites seem to be randomly dispersed throughout the ER network.

Originally it was thought that all proteins that are not tethered in the ER enter transport vesicles by default. However, it is now clear that packaging into vesicles that leave the ER can also be a selective process. Some cargo proteins are actively recruited into such vesicles, where they become concentrated. It is thought that these cargo proteins display exit (transport) signals on their surface that are recognized by complementary receptor proteins that become trapped in the budding vesicle by interacting with components of the COPII coat (Figure 13-17). At a much lower rate, proteins without such exit signals can also get packaged in vesicles, so that even proteins that normally function in the ER (so-called ER resident proteins) slowly leak out of the ER. Similarly, secretory proteins that are made in high concentrations may leave the ER without the help of sorting receptors.

The exit signals that direct proteins out of the ER for transport to the Golgi and beyond are mostly not understood. There is one exception, however. The ERGIC53 protein seems to serve as a receptor for packaging some secretory proteins into COPII-coated vesicles. Its role in protein transport was identified because humans who lack it owing to an inherited mutation have lowered serum levels of two secreted blood-clotting factors (Factor V and Factor VIII) and therefore bleed excessively. The ERGIC53 protein is a lectin that binds mannose and is thought to recognize this sugar on Factor V and Factor VIII proteins, thereby packaging the proteins into transport vesicles in the ER.

Only Proteins That Are Properly Folded and Assembled Can Leave the ER

To exit from the ER, proteins must be properly folded and, if they are subunits of multimeric protein complexes, they may need to be completely assembled. Those that are misfolded or incompletely assembled are retained in the ER, where they are bound to chaperone proteins (see Chapter 6), such as BiP or calnexin. The chaperones may cover up the exit signals or somehow anchor the proteins in the ER (Figure 13-18). Such failed proteins are eventually transported back into the cytosol where they are degraded by proteasomes (discussed in Chapter 12). This quality-control step is important, as misfolded or misassembled proteins could potentially interfere with the functions of normal proteins if they were transported onward. The amount of corrective action is surprisingly large. More than 90% of the newly synthesized subunits of the T cell receptor (discussed in Chapter 24) and of the acetylcholine receptor (discussed in Chapter 11), for example, are normally degraded in the cell without ever reaching the cell surface, where they function. Thus, cells must make a large excess of many protein molecules from which to select the few that fold and assemble properly.

Sometimes, however, this quality-control mechanism is detrimental. The predominant mutations that cause cystic fibrosis, a common inherited disease, produce a plasma membrane protein important for Cl^- transport that is only slightly misfolded. Although the mutant protein would function perfectly normally if it reached the plasma membrane, it is retained in the ER. The devastating disease thus results not because the mutation inactivates the protein, but because the active protein is discarded before it reaches the plasma membrane.

Transport from the ER to the Golgi Apparatus Is Mediated by Vesicular Tubular Clusters

After transport vesicles have budded from an ER exit site and have shed their coat, they begin to fuse with one another. This fusion of membranes from the same compartment is called homotypic fusion, to distinguish it from heterotypic fusion, in which a membrane from one compartment fuses with the membrane of a different compartment. As with heterotypic fusion, homotypic fusion requires a set of matching SNAREs. In this case, however, the interaction is symmetrical, with v-SNAREs and t-SNAREs contributed by both membranes (Figure 13-19).

The structures formed when ER-derived vesicles fuse with one another are called vesicular tubular clusters, on the basis of their convoluted appearance in the electron microscope (Figure 13-20A). These clusters constitute a new compartment that is separate from the ER and lacks many of the proteins that function in the ER. They are generated continually and function as transport packages that bring material from the ER to the Golgi apparatus. The clusters are relatively short-lived because they quickly move along microtubules to the Golgi apparatus, where they fuse and deliver their contents (Figure 13-20B).

As soon as vesicular tubular clusters form, they begin budding off vesicles of their own. Unlike the COPII-coated vesicles that bud from the ER, these vesicles are COPI-coated. They carry back to the ER resident proteins that have escaped, as well as proteins that participated in the ER budding reaction and are being returned. This retrieval process demonstrates the exquisite control mechanisms that regulate coat assembly reactions. The COPI coat assembly begins only seconds after the COPII coats have been shed. It remains a mystery how this switchover in coat assembly is controlled.

The retrieval (or retrograde) transport continues as the vesicular tubular clusters move to the Golgi apparatus. Thus, the clusters continuously mature, gradually changing their composition as selected proteins are returned to the ER. A similar retrieval process continues from the Golgi apparatus, after the vesicular tubular clusters have delivered their cargo.

The Retrieval Pathway to the ER Uses Sorting Signals

The retrieval pathway for returning escaped proteins back to the ER depends on ER retrieval signals. Resident ER membrane proteins, for example, contain signals that bind directly to COPI coats and are thus packaged into COPI-coated transport vesicles for retrograde delivery to the ER. The best-characterized signal of this type consists of two lysines, followed by any two other amino acids, at the extreme C-terminal end of the ER membrane protein. It is called a KKXX sequence, based on the single-letter amino acid code.

Soluble ER resident proteins, such as BiP, also contain a short retrieval signal at their C-terminal end, but it is different: it consists of a Lys-Asp-Glu-Leu or similar sequence. If this signal (called the KDEL sequence) is removed from BiP by genetic engineering, the protein is slowly secreted from the cell. If the signal is transferred to a protein that is normally secreted, the protein is now efficiently returned to the ER, where it accumulates.

Unlike the retrieval signals on ER membrane proteins that can interact directly with the COPI coat, soluble ER resident proteins must bind to specialized receptor proteins such as the KDEL receptor—a multipass transmembrane protein that binds to the KDEL sequence and packages any protein displaying it into COPI-coated retrograde transport vesicles. To accomplish this task, the KDEL receptor itself must cycle between the ER and the Golgi apparatus, and its affinity for the KDEL sequence must be different in these two compartments. The receptor must have a high affinity for the KDEL sequence in vesicular tubular clusters and the Golgi apparatus, so as to capture escaped ER resident proteins that are present there at low concentration. It must have a low affinity for the KDEL sequence in the ER, however, to unload its cargo in spite of the very high concentration of KDEL-containing resident proteins in the ER.

How can the affinity of the KDEL receptor change depending on the compartment in which it resides? The answer may be related to the different pH values established in the different compartments, regulated by H⁺ pumps in their membrane. The KDEL receptor could bind the KDEL sequence under the slightly acidic conditions in vesicular tubular clusters and the Golgi compartment but release it at the neutral pH in the ER. As we discuss later, such pH-sensitive protein-protein interactions form the basis for many of the sorting steps in the cell (Figure 13-21).

Most membrane proteins that function at the interface between the ER and Golgi apparatus, including v- and t-SNARES and some cargo receptors, enter the retrieval pathway to the ER. Whereas the recycling of some of these proteins is signal-mediated as just described, for others no specific signal seems to be required. Thus, while retrieval signals increase the efficiency of the retrieval process, some proteins—including some Golgi enzymes—randomly enter budding vesicles destined for the ER and are returned to the ER at a slower rate. Such Golgi enzymes cycle constantly between the ER and the Golgi, but their rate of return to the ER is slow enough for most of the protein to be found in the Golgi apparatus.

Many Proteins are Selectively Retained in the Compartments in which they Function

The KDEL retrieval pathway only partly explains how ER resident proteins are maintained in the ER. As expected, cells that express genetically modified ER resident proteins, from which the KDEL sequence has been experimentally removed, secrete these proteins. But secretion occurs at a much slower rate than for a normal secretory protein. It seems that ER resident proteins are anchored in the ER by a mechanism that is independent of their KDEL signal and that only those proteins that escape retention are captured and returned via the KDEL receptor. A suggested mechanism of retention is that ER resident proteins bind to one another, thus forming complexes that are too big to enter transport vesicles. Because ER resident proteins are present in the ER at very high concentrations (estimated to be millimolar), relatively low-affinity interactions would suffice to have most of the proteins tied up in such complexes.

Aggregation of proteins that function in the same compartment—called kin recognition—is a general mechanism that compartments use to organize and retain their resident proteins. Golgi enzymes that function together, for example, also bind to each other and are thereby restrained from entering transport vesicles.

The Length of the Transmembrane Region of Golgi Enzymes Determines their Location in The Cell

Vesicles that leave the Golgi apparatus of animal cells destined for the plasma membrane are rich in cholesterol. The cholesterol fills the space between the kinked hydrocarbon chains of the lipids in the bilayer, forcing them into tighter alignment and increasing the separation between the lipid head groups of the two leaflets of the bilayer (see Figure 10-11). Thus the lipid bilayer of the cholesterol-derived vesicles is thicker than that of the Golgi membrane itself. Transmembrane proteins must have sufficiently long transmembrane segments to span this thickness if they are to enter the cholesterol-rich transport vesicle budding from the Golgi apparatus destined for the plasma membrane. Proteins with shorter transmembrane segments are excluded.

This exclusion is thought to explain why membrane proteins that normally reside in the Golgi and the ER have shorter transmembrane segments (around 15 amino acids) than do plasma membrane proteins (around 20–25 amino acids). When the transmembrane segments of Golgi proteins are extended by recombinant DNA techniques, the proteins are no longer efficiently retained in the Golgi apparatus and are transported to the plasma membrane instead. Thus, at least some Golgi proteins seem to be retained in the Golgi apparatus mainly because they cannot enter transport vesicles heading for the plasma membrane.

The Golgi Apparatus Consists of an Ordered Series of Compartments

Because of its large and regular structure, the Golgi apparatus was one of the first organelles described by early light microscopists. It consists of a collection of flattened, membrane-enclosed cisternae, somewhat resembling a stack of pancakes. Each of these Golgi stacks usually consists of four to six cisternae (Figure 13-22), although some unicellular flagellates can have up to 60. In animal cells, many stacks are linked by tubular connections between corresponding cisternae, thus forming a single complex, which is usually located near the cell nucleus and close to the centrosome (Figure 13-23A). This localization depends on microtubules. If microtubules are experimentally depolymerized, the Golgi apparatus reorganizes into individual stacks that are found throughout the cytoplasm, adjacent to ER exit sites. In some cells, including most plant cells, hundreds of individual Golgi stacks are normally dispersed throughout the cytoplasm (Figure 13-23B).

During their passage through the Golgi apparatus, transported molecules undergo an ordered series of covalent modifications. Each Golgi stack has two distinct faces: a cis face (or entry face) and a trans face (or exit face). Both cis and trans faces are closely associated with special compartments, each composed of a network of interconnected tubular and cisternal structures: the cis Golgi network (CGN) (also called the intermediate compartment) and the trans Golgi network (TGN), respectively. Proteins and lipids enter the cis Golgi network in vesicular tubular clusters arriving from the ER and exit from the trans Golgi network bound for the cell surface or another compartment. Both networks are thought to be important for protein sorting. As we have seen, proteins entering the CGN can either move onward in the Golgi apparatus or be returned to the ER. Similarly, proteins exiting from the TGN can either move onward and be sorted according to whether they are destined for lysosomes, secretory vesicles, or the cell surface, or be returned to an earlier compartment.

The Golgi apparatus is especially prominent in cells that are specialized for secretion, such as the goblet cells of the intestinal epithelium, which secrete large amounts of polysaccharide-rich mucus into the gut (Figure 13-24). In such cells, unusually large vesicles are found on the trans side of the Golgi apparatus, which faces the plasma membrane domain where secretion occurs.

Oligosaccharide Chains Are Processed in the Golgi Apparatus

As described in Chapter 12, a single species of N -linked oligosaccharide is attached en bloc to many proteins in the ER and then trimmed while the protein is still in the ER. Further modifications and additions occur in the Golgi apparatus, depending on the protein. The outcome is that two broad classes of N-linked oligosaccharides, the complex oligosaccharides and the high-mannose oligosaccharides, are found attached to mammalian glycoproteins (Figure 13-25). Sometimes both types are attached (in different places) to the same polypeptide chain.

Complex oligosaccharides are generated by a combination of trimming the original N-linked oligosaccharide added in the ER and the addition of further sugars. By contrast, high-mannose oligosaccharides have no new sugars added to them in the Golgi apparatus. They contain just two N-acetylglucosamines and many mannose residues, often approaching the number originally present in the lipid-linked oligosaccharide precursor added in the ER. Complex oligosaccharides can contain more than the original two N-acetylglucosamines as well as a variable number of galactose and sialic acid residues and, in some cases, fucose. Sialic acid is of special importance because it is the only sugar in glycoproteins that bears a negative charge. Whether a given oligosaccharide remains high-mannose or is processed is determined largely by its position on the protein. If the oligosaccharide is accessible to the processing enzymes in the Golgi apparatus, it is likely to be converted to a complex form; if it is inaccessible because its sugars are tightly held to the protein's surface, it is likely to remain in a high-mannose form. The processing that generates complex oligosaccharide chains follows the highly ordered pathway shown in Figure 13-26.

Proteoglycans Are Assembled in the Golgi Apparatus

It is not only the N-linked oligosaccharide chains on proteins that are altered as the proteins pass through the Golgi cisternae en route from the ER to their final destinations; many proteins are also modified in other ways. Some proteins have sugars added to the OH groups of selected serine or threonine side chains. This O -linked glycosylation, like the extension of N-linked oligosaccharide chains, is catalyzed by a series of glycosyl transferase enzymes that use the sugar nucleotides in the lumen of the Golgi apparatus to add sugar residues to a protein one at a time. Usually, N-acetylgalactosamine is added first, followed by a variable number of additional sugar residues, ranging from just a few to 10 or more.

The Golgi apparatus confers the heaviest glycosylation of all on proteoglycan core proteins, which it modifies to produce proteoglycans. As discussed in Chapter 19, this process involves the polymerization of one or more glycosaminoglycan chains (long unbranched polymers composed of repeating disaccharide units) via a xylose link onto serines on the core protein. Many proteoglycans are secreted and become components of the extracellular matrix, while others remain anchored to the plasma membrane. Still others form a major component of slimy materials, such as the mucus that is secreted to form a protective coating over many epithelia.

The sugars incorporated into glycosaminoglycans are heavily sulfated in the Golgi apparatus immediately after these polymers are made, thus adding a significant portion of their characteristically large negative charge. Some tyrosine residues in proteins also become sulfated shortly before they exit from the Golgi apparatus. In both cases, the sulfation depends on the sulfate donor 3′-phosphoadenosine-5′-phosphosulfate, or PAPS, that is transported from the cytosol into the lumen of the trans Golgi network.

What Is the Purpose of Glycosylation?

There is an important difference between the construction of an oligosaccharide and the synthesis of other macromolecules such as DNA, RNA, and protein. Whereas nucleic acids and proteins are copied from a template in a repeated series of identical steps using the same enzyme or set of enzymes, complex carbohydrates require a different enzyme at each step, each product being recognized as the exclusive substrate for the next enzyme in the series. Given the complicated pathways that have evolved to synthesize them, it seems likely that the oligosaccharides on glycoproteins and glycosphingolipids have important functions, but for the most part these functions are not known.

N-linked glycosylation, for example, is prevalent in all eucaryotes, including yeasts, but is absent from procaryotes. Because one or more N-linked oligosaccharides are present on most proteins transported through the ER and Golgi apparatus—a pathway that is unique to eucaryotic cells—one might suspect that they function to aid folding and the transport process. We have already discussed a number of instances for which this is so—the use of a carbohydrate as a marker during protein folding in the ER (see Chapter 12), for example, and the use of carbohydrate-binding lectins in guiding ER-to-Golgi transport. As we discuss later, lectins also participate in protein sorting in the trans Golgi network.

Because chains of sugars have limited flexibility, even a small N-linked oligosaccharide protrudes from the surface of a glycoprotein (Figure 13-27) and can thus limit the approach of other macromolecules to the protein surface. In this way, for example, the presence of oligosaccharides tends to make a glycoprotein more resistant to digestion by proteases. It may be that the oligosaccharides on cell-surface proteins originally provided an ancestral eucaryotic cell with a protective coat that, unlike the rigid bacterial cell wall, left the cell with the freedom to change shape and move. But if so, these sugar chains have since become modified to serve other purposes as well. The oligosaccharides attached to some cell-surface proteins, for example, are recognized by transmembrane lectins called selectins, which function in cell-cell adhesion processes, as discussed in Chapter 19.

Glycosylation can also have important regulatory roles. Signaling through the cell-surface signaling receptor Notch, for example, is important for proper cell fate determination in development. Notch is a transmembrane protein that is O-glycosylated by addition of a single fucose to some serines, threonines, and hydroxylysines. Some cell types express an additional glycosyltransferase that adds an N-acetylglucosamine to each of these fucoses in the Golgi apparatus. This addition sensitizes the Notch receptor, and thus allows these cells to respond selectively to activating stimuli. In this way, glycosylation has become important to the establishment of spatial boundaries in developing tissues.

The Golgi Cisternae Are Organized as a Series of Processing Compartments

Proteins exported from the ER enter the first of the Golgi processing compartments (the cis Golgi compartment), after having passed through the cis Golgi network. They then move to the next compartment (the medial compartment, consisting of the central cisternae of the stack) and finally to the trans compartment, where glycosylation is completed. The lumen of the trans compartment is thought to be continuous with the trans Golgi network, where proteins are segregated into different transport packages and dispatched to their final destinations—the plasma membrane, lysosomes, or secretory vesicles.

The oligosaccharide processing steps occur in a correspondingly organized sequence in the Golgi stack, with each cisterna containing a characteristic abundance of processing enzymes. Proteins are modified in successive stages as they move from cisterna to cisterna across the stack, so that the stack forms a multistage processing unit. This compartmentalization might seem unnecessary, since each oligosaccharide processing enzyme can accept a glycoprotein as a substrate only after it has been properly processed by the preceding enzyme. Nonetheless, it is clear that processing occurs in a spatial as well as a biochemical sequence: enzymes catalyzing early processing steps are concentrated in the cisternae toward the cis face of the Golgi stack, whereas enzymes catalyzing later processing steps are concentrated in the cisternae toward the trans face.

The functional differences between the cis, medial, and trans subdivisions of the Golgi apparatus were discovered by localizing the enzymes involved in processing N-linked oligosaccharides in distinct regions of the organelle, both by physical fractionation of the organelle and by labeling the enzymes in electron microscope sections with antibodies. The removal of mannose residues and the addition of N-acetylglucosamine, for example, were shown to occur in the medial compartment, while the addition of galactose and sialic acid was found to occur in the trans compartment and the trans Golgi network (Figure 13-28). The functional compartmentalization of the Golgi apparatus is summarized in diagrammatic form in Figure 13-29.

The functional and structural divisions of the Golgi stack pose two important questions. How are molecules transported from one Golgi cisterna to the next, and how are Golgi resident proteins retained in their appropriate places?

Transport Through the Golgi Apparatus May Occur by Vesicular Transport or Cisternal Maturation

It is still uncertain how the Golgi apparatus achieves and maintains its polarized structure and how molecules move from one cisterna to another. Functional evidence from in vitro transport assays and the finding of abundant transport vesicles in the vicinity of Golgi cisternae initially led to the view that these vesicles transport proteins between the cisternae, budding from one cisterna and fusing with the next. According to this vesicular transport model, the Golgi apparatus is a relatively static structure, with its enzymes held in place, while the molecules in transit are moved through the cisternae in sequence, carried by transport vesicles (Figure 13-30A). Retrograde flow retrieves escaped ER and Golgi proteins and returns them to preceding compartments. Directional flow is achieved as forward-moving cargo molecules are selectively packaged into forward-moving vesicles, whereas proteins to be retrieved are selectively packaged into retrograde vesicles. Although both types of vesicles are likely to be COPI-coated, the coats may contain different adaptor proteins to confer selectivity on the packaging of cargo molecules. Alternatively, transport vesicles that shuttle between Golgi cisternae may not be directional at all, transporting cargo material randomly back and forth; directional flow would then occur because of the continual input at the cis cisterna and output at the trans cisterna. In either case, the movement of vesicles from each cisterna to an adjacent one is helped by a neat trick: the budding vesicles remain tethered by filamentous proteins that restrict their movement, so that their fusion with the correct target membrane is facilitated.

According to an alternative hypothesis, called the cisternal maturation model, the Golgi is viewed as a dynamic structure in which the cisternae themselves move through the Golgi stack. The vesicular tubular clusters that arrive from the ER fuse with one another to become a cis Golgi network, and this network then progressively matures to become a cis cisterna, then a medial cisterna, and so on. Thus, at the cis face of a Golgi stack, new cis cisternae would continually form and then migrate through the stack as they mature (Figure 13-30B). This model is supported by microscopic observations demonstrating that large structures such as collagen rods in fibroblasts and scales in certain algae—which are much too large to fit into classical transport vesicles—move progressively through the Golgi stack.

In the maturation model, the characteristic distribution of Golgi enzymes is explained by retrograde flow. Everything moves continuously forward with the maturing cisterna, including the processing enzymes that belong in the early Golgi apparatus. But budding COPI-coated vesicles continually collect the appropriate enzymes, almost all of which are membrane proteins, and carry them back to the earlier cisterna where they function. A newly formed cis cisterna would therefore receive its normal complement of resident enzymes primarily from the cisterna just ahead of it and would later pass them back to the next cis cisterna that forms.

As we discuss later, when a cisterna finally moves up to become part of the trans Golgi network, various types of coated vesicles bud off of it until this network disappears, to be replaced by a maturing cisterna just behind. At the same time, other transport vesicles are continually retrieving membrane from post-Golgi compartments and returning this membrane to the trans Golgi network.

The vesicular transport and the cisternal maturation model are not mutually exclusive. Indeed, evidence suggests that transport may occur by a combination of the two mechanisms, in which some cargo is moved forward rapidly in transport vesicles, whereas other cargo is moved forward more slowly as the Golgi apparatus constantly renews itself through cisternal maturation.

Matrix Proteins Form a Dynamic Scaffold That Helps Organize the Apparatus

The unique architecture of the Golgi apparatus depends on both the microtubule cytoskeleton, as already discussed, and cytoplasmic Golgi matrix proteins, which form a scaffold between adjacent cisternae and give the Golgi stack its structural integrity. Some of the matrix proteins form long, filamentous tethers that are thought to help retain Golgi transport vesicles close to the organelle. When the cell prepares to divide, mitotic protein kinases phosphorylate the Golgi matrix proteins, causing the Golgi apparatus to fragment and disperse throughout the cytosol. During disassembly, Golgi enzymes are returned in vesicles to the ER, while other Golgi fragments are distributed to the two daughter cells. There, the matrix proteins are dephosphorylated, leading to the reassembly of the Golgi apparatus.

Remarkably, the Golgi matrix proteins can assemble into appropriately localized stacks near the centrosome even when Golgi membrane proteins are experimentally prevented from leaving the ER. This observation suggests that the matrix proteins are largely responsible for both the structure and location of the Golgi apparatus.

Summary

Correctly folded and assembled proteins in the ER are packaged into COPII-coated transport vesicles that pinch off from the ER membrane. Shortly thereafter the coat is shed and the vesicles fuse with one another to form vesicular tubular clusters, which move on microtubule tracks to the Golgi apparatus. Many resident ER proteins slowly escape, but they are returned to the ER from the vesicular tubular clusters and the Golgi apparatus by retrograde transport in COPI-coated vesicles.

The Golgi apparatus, unlike the ER, contains many sugar nucleotides, which are used by a variety of glycosyl transferase enzymes to perform glycosylation reactions on lipid and protein molecules as they pass through the Golgi apparatus. The N-linked oligosaccharides that are added to proteins in the ER are often initially trimmed by the removal of mannoses, and further sugars are added. Moreover, the Golgi is the site where O-linked glycosylation occurs and where glycosaminoglycan chains are added to core proteins to form proteoglycans. Sulfation of the sugars in proteoglycans and of selected tyrosines on proteins also occurs in a late Golgi compartment.

The Golgi apparatus distributes the many proteins and lipids that it receives from the ER and then modifies the plasma membrane, lysosomes, and secretory vesicles. It is a polarized structure consisting of one or more stacks of disc-shaped cisternae, each stack organized as a series of at least three functionally distinct compartments, termed cis, medial, and trans cisternae. The cis and trans cisternae are both connected to special sorting stations, called the cis Golgi network and the trans Golgi network, respectively. Proteins and lipids move through the Golgi stack in the cis-to-trans direction. This movement may occur by vesicular transport, by progressive maturation of the cis cisternae that migrate continuously through the stack, or by a combination of these two mechanisms. The enzymes that function in each particular region of the stack are thought to be kept there by continual retrograde vesicular transport from more distal cisternae. The finished new proteins end up in the trans Golgi network, which packages them in transport vesicles and dispatches them to their specific destinations in the cell.

Lysosomes Are the Principal Sites of Intracellular Digestion

Lysosomes are membrane-enclosed compartments filled with hydrolytic enzymes that are used for the controlled intracellular digestion of macromolecules. They contain about 40 types of hydrolytic enzymes, including proteases, nucleases, glycosidases, lipases, phospholipases, phosphatases, and sulfatases. All are acid hydrolases. For optimal activity they require an acid environment, and the lysosome provides this by maintaining a pH of about 5.0 in its interior. In this way, the contents of the cytosol are doubly protected against attack by the cell's own digestive system. The membrane of the lysosome normally keeps the digestive enzymes out of the cytosol, but even if they should leak out, they can do little damage at the cytosolic pH of about 7.2.

Like all other intracellular organelles, the lysosome not only contains a unique collection of enzymes, but also has a unique surrounding membrane. Transport proteins in this membrane allow the final products of the digestion of macromolecules—such as amino acids, sugars, and nucleotides—to be transported to the cytosol, from where they can be either excreted or reutilized by the cell. An H⁺ pump in the lysosomal membrane uses the energy of ATP hydrolysis to pump H⁺ into the lysosome, thereby maintaining the lumen at its acidic pH (Figure 13-31). A similar or identical vacuolar H ⁺ ATPase is thought to acidify all endocytic and exocytic organelles, including lysosomes, endosomes, selected compartments of the Golgi apparatus, and many transport and secretory vesicles. Most of the lysosomal membrane proteins are unusually highly glycosylated, which helps to protect them from the lysosomal proteases in the lumen.

Lysosomes Are Heterogeneous

Lysosomes were initially discovered by the biochemical fractionation of cell extracts; only later were they seen clearly in the electron microscope. Although extraordinarily diverse in shape and size, they can be identified as members of a single family of organelles by staining them with specific antibodies. They can also be identified by histochemistry, using the precipitate produced by the action of an acid hydrolase on its substrate to indicate which organelles contain the hydrolase (Figure 13-32). By this criterion, lysosomes are found in all eucaryotic cells.

The heterogeneity of lysosomal morphology contrasts with the relatively uniform structures of most other cell organelles. The diversity reflects the wide variety of digestive functions mediated by acid hydrolases, including the breakdown of intra- and extracellular debris, the destruction of phagocytosed microorganisms, and the production of nutrients for the cell. For this reason, lysosomes are sometimes viewed as a heterogeneous collection of distinct organelles whose common feature is a high content of hydrolytic enzymes. It is especially hard to apply a narrower definition than this in plant cells, as we see next.

Plant and Fungal Vacuoles Are Remarkably Versatile Lysosomes

Most plant and fungal cells (including yeasts) contain one or several very large, fluid-filled vesicles called vacuoles. They typically occupy more than 30% of the cell volume, and as much as 90% in some cell types (Figure 13-33). Vacuoles are related to the lysosomes of animal cells, containing a variety of hydrolytic enzymes, but their functions are remarkably diverse. The plant vacuole can act as a storage organelle for both nutrients and waste products, as a degradative compartment, as an economical way of increasing cell size (Figure 13-34), and as a controller of turgor pressure (the osmotic pressure that pushes outward on the cell wall and keeps the plant from wilting). Different vacuoles with distinct functions (e.g., digestion and storage) are often present in the same cell.

The vacuole is important as a homeostatic device, enabling plant cells to withstand wide variations in their environment. When the pH in the environment drops, for example, the flux of H⁺ into the cytosol is balanced, at least in part, by an increased transport of H⁺ into the vacuole to keep the pH in the cytosol constant. Similarly, many plant cells maintain an almost constant turgor pressure in the face of large changes in the tonicity of the fluid in their immediate environment. They do so by changing the osmotic pressure of the cytosol and vacuole—in part by the controlled breakdown and resynthesis of polymers, such as polyphosphate, in the vacuole, and in part by altering the transport rates of sugars, amino acids, and other metabolites across the plasma membrane and the vacuolar membrane. The turgor pressure controls these fluxes by regulating the activities of the distinct sets of transporters in each membrane.

Substances stored in plant vacuoles are often harvested for human use: in different species, these range from rubber to opium to the flavoring of garlic. Many stored products have a metabolic function. Proteins, for example, can be preserved for years in the vacuoles of the storage cells of many seeds, such as those of peas and beans. When the seeds germinate, these proteins are hydrolyzed and the resulting amino acids provide a food supply for the developing embryo. Anthocyanin pigments stored in vacuoles color the petals of many flowers so as to attract pollinating insects, while noxious molecules released from vacuoles when a plant is eaten or damaged provide a defense against predators.

Multiple Pathways Deliver Materials to Lysosomes

Lysosomes are usually meeting-places where several streams of intracellular traffic converge. Digestive enzymes are delivered to them by a route that leads outward from the ER via the Golgi apparatus, while substances to be digested are fed in by at least three paths, depending on their source.

The best studied of these paths to degradation in lysosomes is that followed by the macromolecules taken up from extracellular fluid by endocytosis. As discussed in detail later, endocytosed molecules are initially delivered in vesicles to small, irregularly shaped intracellular organelles called early endosomes. Some of these ingested molecules are selectively retrieved and recycled to the plasma membrane, while others pass on into late endosomes. It is here that endocytosed materials first meet the lysosomal hydrolases, which are delivered to the endosome from the Golgi apparatus. The interior of the late endosomes is mildly acidic (pH ~6), and it is the site where the hydrolytic digestion of the endocytosed molecules begins. Mature lysosomes form from the late endosomes, accompanied by a further decrease in internal pH. Lysosomes are thought to be produced by a gradual maturation process, during which endosomal membrane proteins are selectively retrieved from the developing lysosome by transport vesicles that deliver these proteins back to endosomes or the trans Golgi network.

A second pathway to degradation in lysosomes is used in all cell types for the disposal of obsolete parts of the cell itself—a process called autophagy. In a liver cell, for example, an average mitochondrion has a lifetime of about 10 days, and electron microscopic images of normal cells reveal lysosomes containing (and presumably digesting) mitochondria, as well as other organelles. The process seems to begin with the enclosure of an organelle by membranes of unknown origin, creating an autophagosome, which then fuses with a lysosome (or a late endosome). The process is highly regulated, and selected cell components can somehow be marked for lysosomal destruction during cell remodeling. The smooth ER that proliferates in a liver cell in response to the drug phenobarbital (discussed in Chapter 12), for example, is selectively removed by autophagy after the drug is withdrawn.

As we discuss later, the third pathway that brings materials to lysosomes for degradation is found mainly in cells specialized for the phagocytosis of large particles and microorganisms. Such professional phagocytes (macrophages and neutrophils in vertebrates) engulf objects to form a phagosome, which is then converted to a lysosome in the manner described for the autophagosome. The three pathways are summarized in Figure 13-35.

A Mannose 6-Phosphate Receptor Recognizes Lysosomal Proteins in the Trans Golgi Network

We now consider the pathway that delivers lysosomal hydrolases and membrane proteins to lysosomes. Both classes of proteins are synthesized in the rough ER and transported through the Golgi apparatus to the trans Golgi network. The transport vesicles that deliver these proteins to late endosomes (which later form lysosomes) bud from the trans Golgi network. The vesicles incorporate the lysosomal proteins and exclude the many other proteins being packaged into different transport vesicles for delivery elsewhere.

How are lysosomal proteins recognized and selected in the trans Golgi network with the required accuracy? For the lysosomal hydrolases the answer is known. They carry a unique marker in the form of mannose 6-phosphate (M6P) groups, which are added exclusively to the N-linked oligosaccharides of these soluble lysosomal enzymes as they pass through the lumen of the cis Golgi network (Figure 13-36). The M6P groups are recognized by transmembrane M6P receptor proteins, which are present in the trans Golgi network. The receptor proteins bind to lysosomal hydrolases on the lumenal side of the membrane and to adaptins in assembling clathrin coats on the cytosolic side. In this way, they help package the hydrolases into clathrin-coated vesicles that bud from the trans Golgi network. The vesicles subsequently deliver their contents to a late endosome.

The M6P Receptor Shuttles Between Specific Membranes

The M6P receptor protein binds its specific oligosaccharide at pH 6.5–6.7 in the trans Golgi network and releases it at pH 6, which is the pH in the interior of late endosomes. Thus, in the late endosomes, the lysosomal hydrolases dissociate from the M6P receptor. As the pH drops further during endosomal maturation, the hydrolases begin to digest the endocytosed material delivered from early endosomes. Having released their bound enzymes, the M6P receptors are retrieved into transport vesicles that bud from late endosomes; the receptors are then returned to the membrane of the trans Golgi network for reuse (Figure 13-37). Transport in either direction requires signal peptides in the cytoplasmic tail of the M6P receptor that specify transport of this protein to the late endosome or back to the Golgi apparatus. Thus, the recycling of the M6P receptor closely resembles the recycling of the KDEL receptor, discussed earlier.

Not all of the hydrolase molecules that are tagged with M6P for delivery to lysosomes get to their proper destination. Some escape the normal packaging process in the trans Golgi network and are transported “by default” to the cell surface, where they are secreted into the extracellular fluid. Some M6P receptors, however, also take a detour to the plasma membrane, where they recapture the escaped lysosomal hydrolases and return them by receptor-mediated endocytosis to lysosomes via early and late endosomes. As lysosomal hydrolases require an acidic milieu to work, they can do little harm in the extracellular fluids, which usually has a neutral pH.

A Signal Patch in the Hydrolyase Polypeptide Chain Provides the Cue for M6P Addition

The sorting system that segregates lysosomal hydrolases and dispatches them to late endosomes works because M6P groups are added only to the appropriate glycoproteins in the Golgi apparatus. This requires specific recognition of the hydrolases by the Golgi enzymes responsible for adding M6P. Since all glycoproteins leave the ER with identical N-linked oligosaccharide chains, the signal for adding the M6P units to oligosaccharides must reside somewhere in the polypeptide chain of each hydrolase. Genetic engineering experiments have revealed that the recognition signal is a cluster of neighboring amino acids on each protein's surface, known as a signal patch.

Two enzymes act sequentially to catalyze the addition of M6P groups to lysosomal hydrolases. The first is a GlcNAc phosphotransferase that specifically binds the hydrolase and adds GlcNAc-phosphate to one or two of the mannose residues on each oligosaccharide chain (Figure 13-38). A second enzyme then cleaves off the GlcNAc residue, leaving behind a newly created M6P marker. Since most lysosomal hydrolases contain multiple oligosaccharides, they acquire many M6P residues, providing a high affinity signal for the M6P receptor.

Defects in the GlcNAc Phosphotransferase Cause a Lysosomal Storage Disease in Humans

Lysosomal storage diseases are caused by genetic defects that affect one or more of the lysosomal hydrolases. The defect results in the accumulation of the undigested substrates in lysosomes, with severe pathological consequences, most often in the nervous system. In most cases there is a mutation in a structural gene that codes for an individual lysosomal hydrolase. This occurs in Hurler's disease, for example, in which the enzyme required for the breakdown of glycosaminoglycans is defective or missing. The most severe form of lysosomal storage disease, however, is a very rare disorder called inclusion-cell disease (I-cell disease). In this disease almost all of the hydrolytic enzymes are missing from the lysosomes of fibroblasts, and their undigested substrates accumulate in lysosomes, which consequently form large “inclusions” in the patients' cells.

I-cell disease is due to a single gene defect and, like most genetic enzyme deficiencies, it is recessive—that is, it is seen only in individuals in whom both copies of the gene are defective. In I-cell disease patients, all the hydrolases missing from lysosomes are found in the blood. Because they fail to be sorted properly in the Golgi apparatus, the hydrolases are secreted rather than transported to lysosomes. The missorting has been traced to a defective or missing GlcNAc-phosphotransferase. Because lysosomal enzymes are not phosphorylated in the cis Golgi network, they are not segregated by M6P receptors into the appropriate transport vesicles in the trans Golgi network. Instead, they are carried to the cell surface and secreted by a default pathway.

In I-cell disease the lysosomes in some cell types, such as hepatocytes, contain a normal complement of lysosomal enzymes, implying that there is another pathway for directing hydrolases to lysosomes that is used by some cell types but not others. The nature of this M6P-independent pathway is unknown. Similarly, the membrane proteins of lysosomes are sorted from the trans Golgi network to late endosomes by an M6P-independent pathway in all cells, and they are therefore normal in I-cell disease. These membrane proteins exit from the trans Golgi network in clathrin-coated vesicles distinct from those that transport the M6P-tagged hydrolases.

It is unclear why cells need more than one sorting pathway to construct a lysosome, although it is perhaps not surprising that different mechanisms operate for soluble and membrane-bound lysosomal proteins, especially since—unlike the M6P receptor—those membrane proteins are lysosomal residents and hence need not be returned to the trans Golgi network.

Some Lysosomes May Undergo Exocytosis

Targeting of material to lysosomes is not necessarily the end of the pathway. Lysosomal secretion (also called defecation) of their undigested content enables all cells to eliminate indigestible debris. For most cells, this seems to be a minor pathway, used only when cells are stressed. Some cell types, however, contain specialized lysosomes that have acquired the necessary machinery for fusion with the plasma membrane. Melanocytes in the skin, for example, produce and store pigments in their lysosomes. These pigment-containing melanosomes release their pigment into the extracellular space by exocytosis. The pigment is then taken up by keratinocytes, leading to normal skin pigmentation. In some genetic disorders, this transfer process is blocked owing to defects in melanosome exocytosis, leading to forms of hypopigmentation (albinism).

Summary

Lysosomes are specialized for the intracellular digestion of macromolecules. They contain unique membrane proteins and a wide variety of hydrolytic enzymes that operate best at pH 5, which is the internal pH of lysosomes. This low pH is maintained by an ATP-driven H⁺ pump in the lysosomal membrane. Newly synthesized lysosomal proteins are transferred into the lumen of the ER, transported through the Golgi apparatus, and then carried from the trans Golgi network to late endosomes by means of clathrin-coated transport vesicles.

The lysosomal hydrolases contain N-linked oligosaccharides that are covalently modified in a unique way in the cis Golgi network so that their mannose residues are phosphorylated. These mannose 6-phosphate (M6P) groups are recognized by an M6P receptor protein in the trans Golgi network that segregates the hydrolases and helps package them into budding transport vesicles that deliver their contents to late endosomes (the organelle that matures into lysosomes). The M6P receptors shuttle back and forth between the trans Golgi network and these endosomes. The low pH in the late endosome dissociates the lysosomal hydrolases from these receptors, making the transport of the hydrolases unidirectional. A separate transport system uses clathrin-coated vesicles to deliver resident lysosomal membrane proteins from the trans Golgi network.

Specialized Phagocytic Cells Can Ingest Large Particles

Phagocytosis is a special form of endocytosis in which large particles such as microorganisms and dead cells are ingested via large endocytic vesicles called phagosomes. In protozoa, phagocytosis is a form of feeding: large particles taken up into phagosomes end up in lysosomes, and the products of the subsequent digestive processes pass into the cytosol to be utilized as food. However, few cells in multicellular organisms are able to ingest such large particles efficiently. In the gut of animals, for example, the particles of food are broken down extracellularly and their hydrolysis products are imported into cells.

Phagocytosis is important in most animals for purposes other than nutrition, and it is mainly carried out by specialized cells—so-called professional phagocytes. In mammals, three classes of white blood cells act as professional phagocytes—macrophages, neutrophils, and dendritic cells. These cells all develop from hemopoietic stem cells (discussed in Chapter 22), and they defend us against infection by ingesting invading microorganisms. Macrophages also have an important role in scavenging senescent cells and cells that have died by apoptosis (discussed in Chapter 17). In quantitative terms, the latter function is by far the most important: our macrophages phagocytose more than 10¹¹ senescent red blood cells in each of us every day, for example.

Whereas the endocytic vesicles involved in pinocytosis are small and uniform, phagosomes have diameters that are determined by the size of the ingested particle, and they can be almost as large as the phagocytic cell itself (Figure 13-39). The phagosomes fuse with lysosomes inside the cell, and the ingested material is then degraded. Any indigestible substances will remain in lysosomes, forming residual bodies. Some of the internalized plasma membrane components never reach the lysosome, because they are retrieved from the phagosome in transport vesicles and returned to the plasma membrane.

To be phagocytosed, particles must first bind to the surface of the phagocyte. However, not all particles that bind are ingested. Phagocytes have a variety of specialized surface receptors that are functionally linked to the phagocytic machinery of the cell. Unlike pinocytosis, which is a constitutive process that occurs continuously, phagocytosis is a triggered process, requiring that receptors be activated that transmit signals to the cell interior and initiate the response. The best-characterized triggers are antibodies, which protect us by binding to the surface of infectious microorganisms to form a coat in which the tail region of each antibody molecule, called the Fc region, is exposed on the exterior (discussed in Chapter 24). This antibody coat is recognized by specific Fc receptors on the surface of macrophages and neutrophils, whose binding induces the phagocytic cell to extend pseudopods that engulf the particle and fuse at their tips to form a phagosome (Figure 13-40).

Several other classes of receptors that promote phagocytosis have been characterized. Some recognize complement components, which collaborate with antibodies in targeting microbes for destruction (discussed in Chapter 25). Others directly recognize oligosaccharides on the surface of certain microorganisms. Still others recognize cells that have died by apoptosis. Apoptotic cells lose the asymmetric distribution of phospholipids in their plasma membrane. As a consequence, negatively charged phosphatidylserine, which is normally confined to the cytosolic leaflet of the lipid bilayer, is now exposed on the outside of the cell, where it triggers the phagocytosis of the dead cell.

Remarkably, macrophages will also phagocytose a variety of inanimate particles—such as glass, latex beads, or asbestos fibers—yet they do not phagocytose live animal cells. It seems that living animal cells display “don't-eat-me” signals in the form of cell-surface proteins that bind to inhibiting receptors on the surface of macrophages. The inhibitory receptors recruit tyrosine phosphatases that antagonize the intracellular signaling events required to initiate phagocytosis, thereby locally inhibiting the phagocytic process. Thus phagocytosis, like many other cell processes, depends on a balance between positive signals that activate the process and negative signals that inhibit it.

Pinocytic Vesicles Form from Coated Pits in the Plasma Membrane

Virtually all eucaryotic cells continually ingest bits of their plasma membrane in the form of small pinocytic (endocytic) vesicles, which are later returned to the cell surface. The rate at which plasma membrane is internalized in this process of pinocytosis varies between cell types, but it is usually surprisingly large. A macrophage, for example, ingests 25% of its own volume of fluid each hour. This means that it must ingest 3% of its plasma membrane each minute, or 100% in about half an hour. Fibroblasts endocytose at a somewhat lower rate (1% per minute), whereas some amoebae ingest their plasma membrane even more rapidly. Since a cell's surface area and volume remain unchanged during this process, it is clear that the same amount of membrane that is being removed by endocytosis is being added to the cell surface by exocytosis, the converse process, as we discuss later. In this sense, endocytosis and exocytosis are linked processes that can be considered to constitute an endocytic-exocytic cycle.

The endocytic part of the cycle often begins at clathrin-coated pits. These specialized regions typically occupy about 2% of the total plasma membrane area. The lifetime of a clathrin-coated pit is short: within a minute or so of being formed, it invaginates into the cell and pinches off to form a clathrin-coated vesicle (Figure 13-41). It has been estimated that about 2500 clathrin-coated vesicles leave the plasma membrane of a cultured fibroblast every minute. The coated vesicles are even more transient than the coated pits: within seconds of being formed, they shed their coat and are able to fuse with early endosomes. Since extracellular fluid is trapped in clathrin-coated pits as they invaginate to form coated vesicles, any substance dissolved in the extracellular fluid is internalized—a process called fluid-phase endocytosis.

Not All Pinocytic Vesicles Are Clathrin-coated

In addition to clathrin-coated pits and vesicles, there are other, less well-understood mechanisms by which cells can form pinocytic vesicles. One of these pathways initiates at caveolae (from the Latin for “little cavities”), originally recognized by their ability to transport molecules across endothelial cells, which form the inner lining of blood vessels. Caveolae are present in the plasma membrane of most cell types, and in some of these they are seen as deeply invaginated flasks in the electron microscope (Figure 13-42). They are thought to form from lipid rafts, which are patches of the plasma membrane that are especially rich in cholesterol, glycosphingolipids, and GPI-anchored membrane proteins (see Figure 12-57). The major structural protein in caveolae is caveolin, a multipass integral membrane protein that is a member of a heterogeneous protein family.

In contrast to clathrin-coated and COPI- or COPII-coated vesicles, caveolae are thought to invaginate and collect cargo proteins by virtue of the lipid composition of the calveolar membrane, rather than by the assembly of a cytosolic protein coat. Caveolae pinch off from the plasma membrane and can deliver their contents either to endosome-like compartments or (in a process called transcytosis, which is discussed later) to the plasma membrane on the opposite side of a polarized cell. Some animal viruses also enter cells in vesicles derived from caveolae. The viruses are first delivered to an endosome-like compartment, from where they are moved to the ER. In the ER, they extrude their genome into the cytosol to start their infectious cycle. It remains a mystery how material endocytosed in caveolae-derived vesicles can end up in so many different locations in the cell.

Cells Import Selected Extracellular Macromolecules by Receptor-mediated Endocytosis

In most animal cells, clathrin-coated pits and vesicles provide an efficient pathway for taking up specific macromolecules from the extracellular fluid. In this process, called receptor-mediated endocytosis, the macromolecules bind to complementary transmembrane receptor proteins, accumulate in coated pits, and then enter the cell as receptor-macromolecule complexes in clathrin-coated vesicles (see Figure 13-41). Receptor-mediated endocytosis provides a selective concentrating mechanism that increases the efficiency of internalization of particular ligands more than a hundredfold, so that even minor components of the extracellular fluid can be specifically taken up in large amounts without taking in a correspondingly large volume of extracellular fluid. A particularly well-understood and physiologically important example is the process whereby mammalian cells take up cholesterol.

Many animals cells take up cholesterol through receptor-mediated endocytosis and, in this way, acquire most of the cholesterol they require to make new membrane. If the uptake is blocked, cholesterol accumulates in the blood and can contribute to the formation in blood vessel walls of atherosclerotic plaques, deposits of lipid and fibrous tissue that can cause strokes and heart attacks by blocking blood flow. In fact, it was through a study of humans with a strong genetic predisposition for atherosclerosis that the mechanism of receptor-mediated endocytosis was first clearly revealed.

Most cholesterol is transported in the blood as cholesteryl esters in the form of lipid-protein particles known as low-density lipoproteins (LDL) (Figure 13-43). When a cell needs cholesterol for membrane synthesis, it makes transmembrane receptor proteins for LDL and inserts them into its plasma membrane. Once in the plasma membrane, the LDL receptors diffuse until they associate with clathrin-coated pits that are in the process of forming (Figure 13-44A). Since coated pits constantly pinch off to form coated vesicles, any LDL particles bound to LDL receptors in the coated pits are rapidly internalized in coated vesicles. After shedding their clathrin coats, the vesicles deliver their contents to early endosomes, which are located near the cell periphery. Once the LDL and LDL receptors encounter the low pH in the endosomes, LDL is released from its receptor and is delivered via late endosomes to lysosomes. There the cholesteryl esters in the LDL particles are hydrolyzed to free cholesterol, which is now available to the cell for new membrane synthesis. If too much free cholesterol accumulates in a cell, the cell shuts off both its own cholesterol synthesis and the synthesis of LDL receptor proteins, so that it ceases either to make or to take up cholesterol.

This regulated pathway for the uptake of cholesterol is disrupted in individuals who inherit defective genes encoding LDL receptor proteins. The resulting high levels of blood cholesterol predispose these individuals to develop atherosclerosis prematurely, and many die at an early age of heart attacks resulting from coronary artery disease. In some cases, the receptor is lacking altogether. In others, the receptors are defective—in either the extracellular binding site for LDL or the intracellular binding site that attaches the receptor to the coat of a clathrin-coated pit (see Figure 13-44B). In the latter case, normal numbers of LDL-binding receptor proteins are present, but they fail to become localized in the clathrin-coated regions of the plasma membrane. Although LDL binds to the surface of these mutant cells, it is not internalized, directly demonstrating the importance of clathrin-coated pits in the receptor-mediated endocytosis of cholesterol.

More than 25 different receptors are known to participate in receptor-mediated endocytosis of different types of molecules, and they all apparently use the same clathrin-coated-pit pathway. Many of these receptors, like the LDL receptor, enter coated pits irrespective of whether they have bound their specific ligands. Others enter preferentially when bound to a specific ligand, suggesting that a ligand-induced conformational change is required for them to activate the signal sequence that guides them into the pits. Since most plasma membrane proteins fail to become concentrated in clathrin-coated pits, the pits must function as molecular filters, preferentially collecting certain plasma membrane proteins (receptors) over others.

Signal peptides guide transmembrane proteins into clathrin-coated pits by binding to the adaptins. Despite a common function, their amino acid sequences vary. A common endocytosis signal consists of only four amino acids Y-X-X-Ψ, where Y is tyrosine, X any polar amino acid, and Ψ a hydrophobic amino acid. This short peptide, which is shared by many receptors, binds directly to one of the adaptins in clathrin-coated pits. By contrast, the cytosolic tail of the LDL receptor contains a unique signal (Asn-Pro-Val-Tyr) that apparently binds to the same adaptin protein.

Electron-microscope studies of cultured cells exposed simultaneously to different labeled ligands demonstrate that many kinds of receptors can cluster in the same coated pit. The plasma membrane of one clathrin-coated pit can probably accommodate up to 1000 receptors of assorted varieties. Although all of the receptor-ligand complexes that use this endocytic pathway are apparently delivered to the same endosomal compartment, the subsequent fates of the endocytosed molecules vary, as we discuss next.

Endocytosed Materials That Are Not Retrieved From Endosomes End Up in Lysosomes

The endosomal compartments of a cell can be complex. They can be made visible in the electron microscope by adding a readily detectable tracer molecule, such as the enzyme peroxidase, to the extracellular medium and leaving the cells for various lengths of time to take it up by endocytosis. The distribution of the molecule after its uptake reveals the endosomal compartments as a set of heterogeneous, membrane-enclosed tubes extending from the periphery of the cell to the perinuclear region, where it is often close to the Golgi apparatus. Two sequential sets of endosomes can be readily distinguished in such labeling experiments. The tracer molecule appears within a minute or so in early endosomes, just beneath the plasma membrane. After 5–15 minutes, it moves to late endosomes, close to the Golgi apparatus and near the nucleus. Early and late endosomes differ in their protein compositions; they are associated with different Rab proteins, for example.

As mentioned earlier, the interior of the endosomal compartment is kept acidic (pH ~6) by a vacuolar H⁺ ATPase in the endosomal membrane that pumps H⁺ into the lumen from the cytosol. In general, later endosomes are more acidic than early endosomes. This acidic environment has a crucial role in the function of these organelles.

We have already seen how endocytosed materials that reach the late endosomes become mixed with newly synthesized acid hydrolases and end up being degraded in lysosomes. Many molecules, however, are specifically diverted from this journey to destruction. They are recycled instead from the early endosomes back to the plasma membrane via transport vesicles. Only molecules that are not retrieved from endosomes in this way are delivered to lysosomes for degradation.

Specific Proteins Are Removed from Early Endosomes and Returned to the Plasma Membrane

The early endosomes form a compartment that acts as the main sorting station in the endocytic pathway, just as the cis and trans Golgi networks serve this function in the biosynthetic-secretory pathway. In the acidic environment of the early endosome, many internalized receptor proteins change their conformation and release their ligand, just as the M6P receptors unload their cargo of acid hydrolases in the even more acidic late endosomes. Those endocytosed ligands that dissociate from their receptors in the early endosome are usually doomed to destruction in lysosomes, along with the other soluble contents of the endosome. Some other endocytosed ligands, however, remain bound to their receptors, and thereby share the fate of the receptors.

The fates of the receptor proteins—and of any ligands remaining bound to them—vary according to the specific type of receptor. (1) Most receptors are recycled and return to the same plasma membrane domain from which they came; (2) some proceed to a different domain of the plasma membrane, thereby mediating a process called transcytosis; and (3) some progress to lysosomes, where they are degraded (Figure 13-45).

The LDL receptor follows the first pathway. It dissociates from its ligand LDL in the early endosome and is recycled to the plasma membrane for reuse, leaving the discharged LDL to be carried to lysosomes (Figure 13-46). The recycling vesicles bud from long, narrow tubules that extend from the early endosomes. It is likely that the geometry of these tubules helps the sorting process. Because tubules have a large membrane area enclosing a small volume, membrane proteins tend to accumulate there. Transport vesicles that return material to the plasma membrane begin budding from the tubules, but tubular portions of the early endosome also pinch off and fuse with one another to form recycling endosomes, a way-station for the traffic between the early endosomes and the plasma membrane. During this process, the tubules and then the recycling endosome continuously shed vesicles that return to the plasma membrane.

The transferrin receptor follows a similar recycling pathway, but it also recycles its ligand. Transferrin is a soluble protein that carries iron in the blood. Cell-surface transferrin receptors deliver transferrin with its bound iron to early endosomes by receptor-mediated endocytosis. The low pH in the endosome induces transferrin to release its bound iron, but the iron-free transferrin itself (called apotransferrin) remains bound to its receptor. The receptor-apotransferrin complex enters the tubular extensions of the early endosome and from there is recycled back to the plasma membrane (Figure 13-47). When the apotransferrin returns to the neutral pH of the extracellular fluid, it dissociates from the receptor and is thereby freed to pick up more iron and begin the cycle again. Thus, transferrin shuttles back and forth between the extracellular fluid and the endosomal compartment, avoiding lysosomes and delivering the iron that cells need to grow to the cell interior.

The second pathway that endocytosed receptors can follow from endosomes is taken both by opioid receptors (see Figure 13-47) and by the receptor that binds epidermal growth factor (EGF). EGF is a small, extracellular signal protein that stimulates epidermal and various other cells to divide. Unlike LDL receptors, EGF receptors accumulate in clathrin-coated pits only after binding EGF, and most of them do not recycle but are degraded in lysosomes, along with the ingested EGF. EGF binding therefore first activates intracellular signaling pathways and then leads to a decrease in the concentration of EGF receptors on the cell surface, a process called receptor down-regulation that reduces the cell's subsequent sensitivity to EGF (discussed in Chapter 15).

Multivesicular Bodies Form on the Pathway to Late Endosomes

It is still uncertain how endocytosed molecules move from the early to the late endosomal compartment so as to end up in lysosomes. A current view is that portions of the early endosomes migrate slowly along microtubules toward the cell interior, shedding tubules of material to be recycled to the plasma membrane. The migrating endosomes enclose large amounts of invaginated membrane and internally pinched-off vesicles and are therefore called multivesicular bodies (Figure 13-48). It is unknown whether multivesicular bodies eventually fuse with a late endosomal compartment or if they fuse instead with each other to become late endosomes. At the end of this pathway, the late endosomes are converted to lysosomes as a result of their fusion with hydrolase-bearing transport vesicles from the trans Golgi network and their increased acidification (Figure 13-49).

The multivesicular bodies carry specific endocytosed membrane proteins that are to be degraded but exclude others that are to be recycled. As part of the protein-sorting process, specific proteins—for example, the occupied EGF receptor described previously—selectively partition to the invaginating membrane of the multivesicular bodies (Figure 13-50). In this way, the receptors, as well as any signaling proteins strongly bound to them, are rendered fully accessible to the digestive enzymes that will degrade them (see Figure 13-50).

Membrane proteins that are sorted into the internal membrane vesicles of a multivesicular body are first covalently modified with the small protein ubiquitin. Unlike multi-ubiquitylation which typically targets substrate proteins for degradation in proteasomes (discussed in Chapter 6), ubiquitin tagging for sorting into the internal membrane vesicles of a multivesicular body requires the addition of only a single ubiquitin molecule that is added to activated receptors while still at the plasma membrane. The ubiquitin tag facilitates the uptake of the receptors into endocytic vesicles and is then recognized again by proteins that mediate the sorting process into the internal membrane vesicles of multivesicular bodies. In addition, membrane invagination in multivesicular bodies is regulated by a lipid kinase that phosphorylates phosphatidylinositol. The phosphorylated head groups of these lipids are thought to serve as docking sites for the proteins that mediate the invagination process. Local modification of lipid molecules is thus another way in which specific membrane patches can be induced to change shape and destiny.

In addition to endocytosed membrane proteins, multivesicular bodies also contain most of the soluble content of early endosomes destined for digestion in lysosomes.

Macromolecules Can Be Transferred Across Epithelial Cell Sheets by Transcytosis

Some receptors on the surface of polarized epithelial cells transfer specific macromolecules from one extracellular space to another by transcytosis (Figure 13-51). These receptors are endocytosed and then follow a pathway from endosomes to a different plasma membrane domain (see Figure 13-46). A newborn rat, for example, obtains antibodies from its mother's milk (which help protect it against infection) by transporting them across the epithelium of its gut. The lumen of the gut is acidic, and, at this low pH, the antibodies in the milk bind to specific receptors on the apical (absorptive) surface of the gut epithelial cells. The receptor-antibody complexes are internalized via clathrin-coated pits and vesicles and are delivered to early endosomes. The complexes remain intact and are retrieved in transport vesicles that bud from the early endosome and subsequently fuse with the basolateral domain of the plasma membrane. On exposure to the neutral pH of the extracellular fluid that bathes the basolateral surface of the cells, the antibodies dissociate from their receptors and eventually enter the newborn's bloodstream.

The transcytotic pathway from the early endosome to the plasma membrane is not direct. The receptors first move from the early endosome to an intermediate endosomal compartment, the recycling endosome described previously (see Figure 13-51). The variety of pathways that different receptors follow from endosomes implies that, in addition to binding sites for their ligands and binding sites for coated pits, many receptors also possess sorting signals that guide them into the appropriate type of transport vesicle leaving the endosome and thereby to the appropriate target membrane in the cell.

A unique property of a recycling endosomes is that the exit of membrane proteins from the compartment can be regulated. Thus, cells can adjust the flux of proteins through the transcytotic pathway according to need. Although the mechanism of regulation is uncertain, it allows recycling endosomes an important role in adjusting the concentration of specific plasma membrane proteins. Fat cells and muscle cells, for example, contain large intracellular pools of the glucose transporters that are responsible for the uptake of glucose across the plasma membrane. These proteins are stored in specialized recycling endosomes until the cell is stimulated by the hormone insulin to increase its rate of glucose uptake. Then transport vesicles bud from the recycling endosome and deliver large numbers of glucose transporters to the plasma membrane, thereby greatly increasing the rate of glucose uptake into the cell (Figure 13-52).

Epithelial Cells Have Two Distinct Early Endosomal Compartments But a Common Late Endosomal Compartment

In polarized epithelial cells, endocytosis occurs from both the basolateral domain and the apical domain of the plasma membrane. Material endocytosed from either domain first enters an early endosomal compartment that is unique to that domain. This arrangement allows endocytosed receptors to be recycled back to their original membrane domain, unless they contain signals that mark them for transcytosis to the other domain. Molecules endocytosed from either plasma membrane domain that are not retrieved from the early endosomes end up in a common late endosomal compartment near the cell center and are eventually degraded in lysosomes (Figure 13-53).

Whether cells contain a few connected or many unconnected endosomal compartments seems to depend on the cell type and the physiological state of the cell. Like many other membrane-enclosed organelles, endosomes of the same type can readily fuse with one another (an example of homotypic fusion, discussed earlier) to create large continuous endosomes.

Summary

Cells ingest fluid, molecules, and particles by endocytosis, in which localized regions of the plasma membrane invaginate and pinch off to form endocytic vesicles. Many of the endocytosed molecules and particles end up in lysosomes, where they are degraded. Endocytosis occurs both constitutively and as a triggered response to extracellular signals. Endocytosis is so extensive in many cells that a large fraction of the plasma membrane is internalized every hour. To make this possible, most of the plasma membrane components (proteins and lipid) that are endocytosed are continually returned to the cell surface by exocytosis. This large-scale endocytic-exocytic cycle is mediated largely by clathrin-coated pits and vesicles.

Many cell-surface receptors that bind specific extracellular macromolecules become localized in clathrin-coated pits. As a result, they and their ligands are efficiently internalized in clathrin-coated vesicles, a process called receptor-mediated endocytosis. The coated endocytic vesicles rapidly shed their clathrin coats and fuse with early endosomes.

Most of the ligands dissociate from their receptors in the acidic environment of the endosome and eventually end up in lysosomes, while most of the receptors are recycled via transport vesicles back to the cell surface for reuse. But receptor-ligand complexes can follow other pathways from the endosomal compartment. In some cases, both the receptor and the ligand end up being degraded in lysosomes, resulting in receptor down-regulation. In other cases, both are transferred to a different plasma membrane domain, and the ligand is thereby released by exocytosis at a surface of the cell different from that where it originated, a process called transcytosis. The transcytosis pathway includes recycling endosomes, where endocytosed plasma membrane proteins can be stored until they are needed.

Many Proteins and Lipids Seem to Be Carried Automatically from the Golgi Apparatus to the Cell Surface

In a cell capable of regulated secretion, at least three classes of proteins must be separated before they leave the trans Golgi network—those destined for lysosomes (via late endosomes), those destined for secretory vesicles, and those destined for immediate delivery to the cell surface. We have already noted that proteins destined for lysosomes are tagged for packaging into specific departing vesicles (by mannose-6-phosphate for lysosomal hydrolases), and analogous signals are thought to direct secretory proteins into secretory vesicles. Most other proteins are transported directly to the cell surface by the nonselective constitutive secretory pathway. Because entry into this pathway does not require a particular signal, it is also called the default pathway (Figure 13-55). Thus, in an unpolarized cell such as a white blood cell or a fibroblast, it seems that any protein in the lumen of the Golgi apparatus is automatically carried by the constitutive pathway to the cell surface unless it is either specifically returned to the ER, retained as a resident protein in the Golgi apparatus itself, or selected for the pathways that lead to regulated secretion or to lysosomes. In polarized cells, where different products have to be delivered to different domains of the cell surface, we shall see that the options are more complex.

Secretory Vesicles Bud from the Trans Golgi Network

Cells that are specialized for secreting some of their products rapidly on demand concentrate and store these products in secretory vesicles (often called secretory granules or dense-core vesicles because they have dense cores when viewed in the electron microscope). Secretory vesicles form from the trans Golgi network, and they release their contents to the cell exterior by exocytosis in response to extracellular signals. The secreted product can be either a small molecule (such as histamine) or a protein (such as a hormone or digestive enzyme).

Proteins destined for secretory vesicles (called secretory proteins) are packaged into appropriate vesicles in the trans Golgi network by a mechanism that is believed to involve the selective aggregation of the secretory proteins. Clumps of aggregated, electron-dense material can be detected by electron microscopy in the lumen of the trans Golgi network. The signal that directs secretory proteins into such aggregates is not known, but it is thought to be composed of signal patches that are common to proteins of this class. When a gene encoding a secretory protein is transferred to a secretory cell that normally does not make the protein, the foreign protein is appropriately packaged into secretory vesicles. This observation shows that although the proteins that an individual cell expresses and packages in secretory vesicles differ, they all contain common sorting signals, which function properly even when the proteins are expressed in cells that do not normally make them.

It is unclear how the aggregates of secretory proteins are segregated into secretory vesicles. Secretory vesicles have unique proteins in their membrane, some of which might serve as receptors for aggregated protein in the trans Golgi network. The aggregates are much too big, however, for each molecule of the secreted protein to be bound by its own cargo receptor, as proposed for transport of the lysosomal enzymes. The uptake of the aggregates into secretory vesicles may therefore more closely resemble the uptake of particles by phagocytosis at the cell surface, where the plasma membrane zippers up around large structures.

Initially, most of the membrane of the secretory vesicles that leave the trans Golgi network is only loosely wrapped around the clusters of aggregated secretory proteins. Morphologically, these immature secretory vesicles resemble dilated trans Golgi cisternae that have pinched off from the Golgi stack. As the vesicles mature, their contents become concentrated (Figure 13-56), probably as the result of both the continuous retrieval of membrane that is recycled back to late endosomes and the progressive acidification of the vesicle lumen that results from the progressive concentration of ATP-driven H⁺ pumps in the vesicle membrane. The degree of concentration of proteins during the formation and maturation of secretory vesicles is small, however, compared with the total 200–400-fold concentration that occurs after they leave the ER. Secretory and membrane proteins become concentrated as they move from the ER through the Golgi apparatus because of an extensive retrograde retrieval process mediated by COPI-coated transport vesicles that exclude them (see Figure 13-21).

Membrane recycling is important for returning Golgi components to the Golgi apparatus, as well as for concentrating the contents of secretory vesicles. The vesicles that mediate this retrieval originate as clathrin-coated buds on the surface of immature secretory vesicles, often being seen even on budding secretory vesicles that have not yet severed from the Golgi stack (see Figure 13-56B).

Because the final mature secretory vesicles are so densely filled with contents, the secretory cell can disgorge large amounts of material promptly by exocytosis when triggered to do so (Figure 13-57).

Proteins Are Often Proteolytically Processed During the Formation of Secretory Vesicles

Condensation is not the only process to which secretory proteins are subject as the secretory vesicles mature. Many polypeptide hormones and neuropeptides, as well as many secreted hydrolytic enzymes, are synthesized as inactive protein precursors from which the active molecules have to be liberated by proteolysis. These cleavages begin in the trans Golgi network, and they continue in the secretory vesicles and sometimes in the extracellular fluid after secretion has occurred. Many secreted polypeptides have, for example, an N-terminal pro-peptide that is cleaved off to yield the mature protein. These proteins are thus synthesized as pre-pro-proteins, the pre-peptide consisting of the ER signal peptide that is cleaved off earlier in the rough ER (see Figure 12-40). In other cases, peptide-signaling molecules are made as polyproteins that contain multiple copies of the same amino acid sequence. In still more complex cases, a variety of peptide-signaling molecules are synthesized as parts of a single polyprotein that acts as a precursor for multiple end-products, which are individually cleaved from the initial polypeptide chain. The same polyprotein may be processed in various ways to produce different peptides in different cell types (Figure 13-58).

Why is proteolytic processing so common in the secretory pathway? Some of the peptides produced in this way, such as the enkephalins (five-amino-acid neuropeptides with morphine-like activity), are undoubtedly too short in their mature forms to be co-translationally transported into the ER lumen or to include the necessary signal for packaging into secretory vesicles. In addition, for secreted hydrolytic enzymes—or any other protein whose activity could be harmful inside the cell that makes it—delaying activation of the protein until it reaches a secretory vesicle or until after it has been secreted has a clear advantage: it prevents it from acting prematurely inside the cell in which it is synthesized.

Secretory Vesicles Wait Near the Plasma Membrane Until Signaled to Release Their Contents

Once loaded, a secretory vesicle has to get to the site of secretion, which in some cells is far away from the Golgi apparatus. Nerve cells are the most extreme example. Secretory proteins, such as peptide neurotransmitters (neuropeptides) that are to be released from nerve terminals at the end of the axon, are made and packaged into vesicles in the cell body, where the ribosomes, ER, and Golgi apparatus are located. They must then travel along the axon to the nerve terminals, which can be a meter or more away. As discussed in Chapter 16, motor proteins propel the vesicles along axonal microtubules, whose uniform orientation guides the vesicles in the proper direction. Microtubules also guide vesicles to the cell surface for constitutive exocytosis.

Whereas vesicles containing materials for constitutive release fuse with the plasma membrane once they arrive there, secretory vesicles in the regulated pathway wait at the membrane until the cell receives a signal to secrete and then fuse. The signal is often a chemical messenger, such as a hormone, that binds to receptors on the cell surface. The resulting activation of the receptors generates intracellular signals, often including a transient increase in the concentration of free Ca²⁺ in the cytosol. In nerve terminals, the initial signal for exocytosis is usually an electrical excitation (an action potential) triggered by a chemical transmitter binding to receptors elsewhere on the same cell surface. When the action potential reaches the nerve terminals, it causes an influx of Ca²⁺ through voltage-gated Ca²⁺ channels. The binding of Ca²⁺ ions to specific sensors then triggers the secretory vesicles (called synaptic vesicles) to fuse with the plasma membrane and release their contents to the extracellular space.

The speed of transmitter release indicates that the proteins mediating the fusion reaction do not undergo complex, multistep rearrangements. After vesicles have been docked to the presynaptic plasma membrane, they undergo a priming step, which prepares them for rapid fusion. The SNAREs may be partly paired, but their helices are not fully wound into the final four-helix bundle required for fusion (see Figure 13-12). Other proteins are thought to keep the SNAREs from completing the fusion reaction until the Ca²⁺ influx releases this brake. At a typical synapse, only few of the docked vesicles seem to be primed and ready for exocytosis. The use of only a few vesicles at a time allows each synapse to fire over and over again in quick succession. With each firing, new synaptic vesicles become primed to replace those that have fused and released their contents.

Regulated Exocytosis Can Be a Localized Response of the Plasma Membrane and Its Underlying Cytoplasm

Histamine is a small molecule secreted by mast cells. It is released by the regulated pathway in response to specific ligands that bind to receptors on the mast cell surface. Histamine is responsible for many of the unpleasant symptoms that accompany allergic reactions, such as itching and sneezing. When mast cells are incubated in fluid containing a soluble stimulant, massive exocytosis occurs all over the cell surface (Figure 13-59). But if the stimulating ligand is artificially attached to a solid bead so that it can interact only with a localized region of the mast cell surface, exocytosis is now restricted to the region where the cell contacts the bead (Figure 13-60).

This experiment shows that individual segments of the plasma membrane can function independently in regulated exocytosis. As a result, the mast cell, unlike a nerve cell, does not respond as a whole when it is triggered; the activation of receptors, the resulting intracellular signals, and the subsequent exocytosis are all localized in the particular region of the cell that has been excited. Such localized exocytosis enables a killer lymphocyte, for example, to deliver the proteins that induce the death of a single infected target cell precisely without endangering normal neighboring cells (see Figure 16-97).

Secretory Vesicle Membrane Components Are Quickly Removed from the Plasma Membrane

When a secretory vesicle fuses with the plasma membrane, its contents are discharged from the cell by exocytosis, and its membrane becomes part of the plasma membrane. Although this should greatly increase the surface area of the plasma membrane, it does so only transiently, because membrane components are removed from the surface by endocytosis almost as fast as they are added by exocytosis, reminiscent of the exocytosis-endocytosis cycle discussed earlier. After their removal from the plasma membrane, the proteins of the secretory vesicle membrane are thought to be shuttled to lysosomes for degradation. The amount of secretory vesicle membrane that is temporarily added to the plasma membrane can be enormous: in a pancreatic acinar cell discharging digestive enzymes for delivery to the gut lumen, about 900 μm² of vesicle membrane is inserted into the apical plasma membrane (whose area is only 30 μm²) when the cell is stimulated to secrete.

Control of membrane traffic thus has a major role in maintaining the composition of the various membranes of the cell. To maintain each membrane-enclosed compartment in the secretory and endocytotic pathways at a constant size, the balance between the forward and retrograde flows of membrane needs to be precisely regulated. For cells to grow, the forward flow needs to be greater than the retrograde flow, so that the membrane can increase in area. For cells to maintain a constant size, the forward and retrograde flows must be equal. We still know very little about the mechanisms that coordinate these flows.

Polarized Cells Direct Proteins from the Trans Golgi Network to the Appropriate Domain of the Plasma Membrane

Most cells in tissues are polarized and have two (and sometimes more) distinct plasma membrane domains to which different types of vesicles must be directed. This raises the general problem of how the delivery of membrane from the Golgi apparatus is organized so as to maintain the differences between one cell-surface domain and another. A typical epithelial cell has an apical domain, which faces the lumen and often has specialized features such as cilia or a brush border of microvilli; it also has a basolateral domain, which covers the rest of the cell. The two domains are separated by a ring of tight junctions (see Figure 19-3), which prevent proteins and lipids (in the outer leaflet of the lipid bilayer) from diffusing between the two domains, so that the compositions of the two domains are different.

A nerve cell is another example of a polarized cell. The plasma membrane of its axon and nerve terminals is specialized for signaling to other cells, whereas the plasma membrane of its cell body and dendrites is specialized to receive signals from other nerve cells. The two domains have distinct protein compositions. Studies of protein traffic in nerve cells in culture suggest that, with regard to vesicular transport from the trans Golgi network to the cell surface, the plasma membrane of the nerve cell body and dendrites resembles the basolateral membrane of a polarized epithelial cell, while the plasma membrane of the axon and its nerve terminals resembles the apical membrane of such a cell (Figure 13-61). Thus, some proteins that are targeted to a specific domain in the epithelial cell are also found to be targeted to the corresponding domain in the nerve cell.

Cytoplasmic Sorting Signals Guide Membrane Proteins Selectively to the Basolateral Plasma Membrane

In principle, differences between plasma membrane domains need not depend on the targeted delivery of the appropriate membrane components. Instead, membrane components could be delivered to all regions of the cell surface indiscriminately, but then be selectively stabilized in some locations and selectively eliminated in others. Although this strategy of random delivery followed by selective retention or removal seems to be used in certain cases, deliveries are often specifically directed to the appropriate membrane domain. Epithelial cells, for example, frequently secrete one set of products—such as digestive enzymes or mucus in cells lining the gut—at their apical surface, and another set of products—such as components of the basal lamina—at their basolateral surface. Thus, cells must have ways of directing vesicles carrying different cargoes to different plasma membrane domains.

By examining polarized epithelial cells in culture, it has been found that proteins from the ER destined for different domains travel together until they reach the trans Golgi network. Here they are separated and dispatched in secretory or transport vesicles to the appropriate plasma membrane domain (Figure 13-62).

Membrane proteins destined for delivery to the basolateral membrane contain sorting signals in their cytoplasmic tail. Two such signals are known, one containing a characteristic conserved tyrosine and the other two adjacent leucines. When present in an appropriate structural context, these amino acids are recognized by coat proteins that package them into appropriate transport vesicles in the trans Golgi network. The same basolateral signals that are recognized in the trans Golgi network also function in endosomes to redirect the proteins back to the basolateral plasma membrane after they have been endocytosed.

Lipid Rafts May Mediate Sorting of Glycosphingolipids and GPI-anchored Proteins to the Apical Plasma Membrane

The apical plasma membrane of most cells is greatly enriched in glycosphingolipids, which help protect this exposed surface from damage—by the digestive enzymes and low pH in such sites as the stomach or the lumen of the gut, for example. Plasma membrane proteins that are linked to the lipid bilayer by a glycosylphosphatidylinositol (GPI) anchor are also found exclusively in the apical plasma membrane. If recombinant DNA techniques are used to attach a GPI anchor to a protein that would normally be delivered to the basolateral surface, the protein is now delivered to the apical surface instead.

GPI-anchored proteins are thought to be directed to the apical membrane because they associate with the glycosphingolipids in lipid rafts that form in the membrane of the trans Golgi network. As discussed in Chapter 10, lipid rafts form in the trans Golgi network and plasma membrane when glycosphingolipids and cholesterol self-associate into microaggregates (see Figure 10-13). Membrane proteins with unusually long transmembrane domains also accumulate in the rafts. In addition, the rafts preferentially contain GPI-anchored proteins and some carbohydrate-binding proteins (lectins) that may help stabilize the assemblies (Figure 13-63).

Having selected a unique set of cargo molecules, the rafts then bud from the trans Golgi network into transport vesicles destined for the apical plasma membrane.

Synaptic Vesicles Can Form Directly from Endocytic Vesicles

Nerve cells (and some endocrine cells) contain two types of secretory vesicles. As for all secretory cells, these cells package proteins and peptides in dense-cored secretory vesicles in the standard way for release by the regulated secretory pathway. In addition, however, they make use of another specialized class of tiny (~50-nm diameter) secretory vesicles, which are called synaptic vesicles and are generated in a different way. In nerve cells, these vesicles store small neurotransmitter molecules, such as acetylcholine, glutamate, glycine, and γ-aminobutyric acid (GABA), that mediate rapid signaling from cell to cell at chemical synapses. As discussed earlier, the vesicles are triggered to release their contents within a fraction of a millisecond when an action potential arrives at a nerve terminal. Some neurons fire more than 1000 times per second, releasing neurotransmitters each time. This rapid release is possible because some of the vesicles are docked and primed for fusion, which will occur only when an action potential causes an influx of Ca²⁺ into the terminal.

Only a small proportion of the synaptic vesicles in the nerve terminal fuse with the plasma membrane in response to each action potential. But for the nerve terminal to respond rapidly and repeatedly, the vesicles need to be replenished very quickly after they discharge. Thus, most synaptic vesicles are generated not from the Golgi membrane in the nerve cell body but by local recycling from the plasma membrane in the nerve terminals. It is thought that the membrane components of the synaptic vesicles are initially delivered to the plasma membrane by the constitutive secretory pathway and then retrieved by endocytosis. But instead of fusing with endosomes, most of the endocytic vesicles immediately fill with transmitter to become synaptic vesicles.

The membrane components of a synaptic vesicle include carrier proteins specialized for the uptake of neurotransmitter from the cytosol, where the small-molecule neurotransmitters that mediate fast synaptic signaling are synthesized. Once filled with neurotransmitter, the vesicles return to the plasma membrane, where they wait until the cell is stimulated. After they have released their contents, their membrane components are retrieved in the same way and used again (Figure 13-64).

Summary

Proteins can be secreted from cells by exocytosis in either a constitutive or a regulated fashion. In the regulated pathways, molecules are stored either in secretory vesicles or synaptic vesicles, which do not fuse with the plasma membrane to release their contents until an appropriate signal is received. Secretory vesicles bud from the trans Golgi network. The secretory proteins they contain condense during the formation and maturation of secretory vesicles. Synaptic vesicles, which are confined to nerve cells and some endocrine cells, form from endocytic vesicles and from endosomes, and they are responsible for the regulated secretion of small-molecule neurotransmitters. Whereas the regulated pathways operate only in specialized secretory cells, a constitutive secretory pathway operates in all eucaryotic cells, mediated by continual vesicular transport from the trans Golgi network to the plasma membrane.

Proteins are delivered from the trans Golgi network to the plasma membrane by the constitutive pathway unless they are diverted into other pathways or retained in the Golgi apparatus. In polarized cells, the transport pathways from the trans Golgi network to the plasma membrane operate selectively to ensure that different sets of membrane proteins, secreted proteins, and lipids are delivered to the different domains of the plasma membrane.