Signal Sequence Recognition

Soluble and membrane proteins destined for ER translocation contain a hydrophobic sequence either at their N-terminus or in transmembrane segments that serve as recognition signals for direction to the ER membrane. N-terminal leader sequences (termed signal sequences) are typically 15 to 35 amino acids long and contain a hydrophobic core of at least 6 residues, while transmembrane signal segments have a hydrophobic stretch of 16 to 25 residues. Aside from hydrophobicity, these signal sequences have no other features in common (Table 20-2). Nevertheless, when attached to proteins that are not normally targeted to the ER, these signal sequences direct the protein to the ER and not to other organelles such as to mitochondria or peroxisomes, which use unique targeting signals (see Fig. 18-1).

A protein containing a signal sequence or transmembrane signal segment is guided to the ER membrane by either a cotranslational pathway or a posttranslational pathway. In the cotranslational pathway, substrates are translocated across the ER concurrent with their synthesis by membrane-bound ribosomes. In the posttranslational pathway, the substrate is first fully synthesized in the cytoplasm and then translocated in a ribosome-independent fashion. These pathways operate in several qualitatively different ways.

Cotranslational Translocation

In the cotranslational pathway, the signal recognition particle (SRP) recognizes and binds the first hydrophobic domain of either the signal sequence or the transmembrane signal segment as it emerges from the ribosome slowing translation of the polypeptide (Fig. 20-3A). Then the complex consisting of ribosome, nascent chain, and SRP associates with the ER membrane by binding the SRP receptor (SR), a heterodimer consisting of one subunit that binds SRP and another that spans the ER membrane. Both the SRP and the SR have guanosine triphosphatase (GTPase) domains (Fig. 20-3B) similar to Ras (see Fig. 4-6). GTP binding and hydrolysis by SRP and SR provide directionality and order to the sequence of reactions that bring the nascent chain to the translocation channel. Once the ribosome–nascent chain–SRP complex is bound to SR at the ER membrane, SRP releases the signal sequence, allowing the ribosome–nascent chain complex to be transferred to the protein-conducting channel across the ER membrane. SRP and SR then dissociate after hydrolyzing their bound guanosine triphosphate (GTP), releasing SRP into the cytoplasm and allowing SR to diffuse away in the membrane. The targeting cycle delivers the ribosome–nascent chain complex to the protein-conducting channel (called the translocon) and recycles the targeting machinery (i.e., SRP and SR).

Figure 20-3 cotranslational pathway from ribosome to endoplasmic reticulum lumen. A, Signal recognition particle (SRP) and SRP-receptor use a cycle of recruitment and hydrolysis of GTP to control delivery of the ribosome-mRNA complex to the ER translocon. SRP binds a signal sequence emerging from a ribosome and arrests polypeptide translation. SRP also directs the ribosome to the SRP-receptor on the ER membrane, where the ribosome docks on the translocon and continues translation. (GDP, guanosine diphosphate.) B, Ribbon diagram of SRP and SRP-receptor-binding complex showing the close association between each protein’s GTPase domain, which permits the GTPases to undergo reciprocal activation.

The nascent chain emerging from the ribosome must then engage and open the translocon for transport across the ER membrane. The central component of the translocon is the evolutionarily conserved heterotrimeric Sec61 protein complex (Sec61 complex). The Sec61 complex provides a high-affinity docking site for the ribosome–nascent chain complex. Binding of the signal sequence to the Sec61 complex is facilitated by additional factors, including the protein TRAM (translocating chain-associating membrane protein) and the protein complex TRAP (translocon-associated protein). Once bound, the ribosome–nascent chain complex inserts the nascent chain into the Sec61 channel. The channel then opens toward the lumen, providing the nascent chain a continuous path from the peptidyl transferase center in the ribosome, through the translocation channel and into the ER lumen. Elongation of the polypeptide chain by the ribosome “pushes” the nascent chain through the channel and across the membrane. In this way, the energy used for protein synthesis is harnessed to drive translocation of the polypeptide across the membrane. Inside the ER lumen binding and release of chaperones help to “pall” the polypeptide across the membrane.

Posttranslational Translocation

Posttranslational translocation (prevalent in fungi such as yeast) differs because neither cotranslational targeting machinery (SRP and SR) nor ribosomes participate. Instead, other components fulfill the functions provided by SRP, SR, and ribosomes (with the Sec61 complex serving as channel, as it does in cotranslational translocation). Signal-containing polypeptides destined for posttranslational insertion into the ER are held in a largely unfolded state by cytoplasmic chaperones until they can be delivered to the tetrameric Sec62/63 protein complex at the ER membrane (Fig. 20-4). A similar strategy is used to import proteins into the mitochondria (see Fig. 18-4). During posttranslational translocation, the signal sequence engages and opens the Sec61 channel in a fashion similar to cotranslational translocation. Since protein synthesis is already complete, another energy source must be exploited to move the substrate through the channel into the ER. BiP, a luminal ER chaperone belonging to the Hsp70 family, binds the substrate in the ER lumen, thereby preventing it from sliding back into the cytoplasm. Repeated rounds of substrate binding and release, catalyzed by ATP hydrolysis, allow BiP to act as a molecular ratchet to drive substrate transport into the lumen. The transmembrane protein Sec63 regulates the ATPase activity of BiP and helps to recruit BiP to the translocation channel. Thus, the peptide is “pulled” across the membrane from the luminal side instead of being “pushed” from the cytoplasmic side, as during cotranslational translocation. Hsp70 family members perform a similar function during import of proteins into mitochondria and chloroplasts (see Fig. 18-4).

Figure 20-4 posttranslational pathway from ribosome to endoplasmic reticulum lumen. As the polypeptide emerges (purple thread) from the ribosome in the cytosol, it is bound by chaperones (green) that prevent it from aggregating and guide it to the Sec62/63 protein complex at the ER membrane. After signal sequence engagement with the Sec63 protein-conducting channel, the polypeptide is pulled across the membrane from the lumenal side by repeated binding to and release from BiP (i.e., ratcheting). The signal sequence is the blue portion of the polypeptide.

Universality of Protein Translocation

In both cotranslational and posttranslational translocation, the Sec61 complex translocation channel is closed until opened by interacting with binding partners, including the translocated polypeptide. This ensures that only specific types of proteins pass through the ER membrane and that the permeability barrier of these membranes is maintained at all times. In cotranslational translocation, interactions with a ribosome and a signal sequence open the channel, and protein synthesis provides the energy for translocation. In posttranslational translocation, the Sec62/63 complex operates on both sides of the membrane. On the cytoplasmic side, it contributes to the interactions of the signal sequence that open the translocon. On the lumenal side, it recruits BiP to provide the driving force for translocation. This theme applies to protein translocation across the plasma membrane by prokaryotes. There, a homolog of the Sec61 complex (the SecYE complex) interacts with the cytoplasmic SecA adenosine triphosphatase (ATPase), a partner that drives translocation across the plasma membrane (see Fig. 18-9).

Signal Recognition Particle and Signal Recognition Particle–Receptor

Human SRP is a ribonucleoprotein composed of six proteins (named by their apparent molecular weights and a 300-nucleotide RNA (Fig. 20-5). The SRP54 protein subunit and a portion of the RNA compose the minimal hardware for targeting to the translocon and are found in both prokaryotic and eukaryotic cells. SRP54 binds signal sequences in a deep, hydrophobic groove lined by the flexible side chains of several methionines. Like bristles of a brush, the methionines accommodate the various shapes of the hydrophobic side chains of the residues of different signal sequences. Phosphates of the SRP RNA near one end of this groove may interact with basic residues that are often (but not always) adjacent to the hydrophobic core of signal sequences and transmembrane signal segments.

Figure 20-5 atomic structures of cotranslocation hardware. A–B, Comparison of bacterial and human signal recognition particles, showing the base pairing of human 7S SRP RNA and bacterial 4.5S SRP RNA and indicating protein-binding sites (blue boxes). Atomic structures of elements of the RNA and associated proteins are shown as space-filling and ribbon diagrams. Ffh GTP-binding and N-domains (PDB files: 3NG1, 2NG1), Ffh M-domain bound to RNA (PDB file: 1DUL), RNA (PDB file: 1E9S), SRP9/14 bound to RNA (PDB file: 1E80). C, Bacterial SRP-receptor subunit FtsY. (PDB file: 1FTS.) D, Bacterial signal peptidase. (PDB file: 1B12.)

SRP binding to a signal sequence slows translation, a phenomenon that is termed elongation arrest (Fig. 20-3A). The mechanism appears to involve occlusion of the elongation factor binding site on the ribosome by the SRP9 and SRP14 subunits of SRP, which structurally resembles a portion of eEF2. Slowing translation provides time to target the ribosome to the translocation channel before excessive polypeptide synthesis precludes cotranslational transport.

Interaction of SRP with SR directs the SRP–ribosome–nascent chain complex to the translocation channel (Fig. 20-3A). This interaction is regulated by GTP binding and hydrolysis by SRP and SR, whose GTPase cycles are a notable exception to the GTPase switch paradigm of Ras-like GTPases that involve GTP exchange factors (GEFs) and GTPase-activating proteins (GAPs; see Fig. 25-8). No external GEFs or GAPs are known for SRP and SR GTPases. Instead, these proteins readily exchange GDP for GTP and are in the GTP-bound state as they enter the targeting cycle. On formation of a complex, SRP and SR reciprocally activate each other’s GTPase activity, thereby obviating the need for an external GAP to drive the conversion of GTP to GDP on these proteins. Upon GTP hydrolysis, the conformations of SR and SRP change in a way that reduces their affinity. They dissociate from each other as well as from the ribosome–nascent chain and translocation channel. This frees SRP and SR for further rounds of ribosome–nascent chain targeting. Release of SRP and SR occurs only after the ribosome has become properly engaged with the translocation channel in the ER membrane, thereby ensuring that the channel is closed until the ribosome has bound.

Sec61 Complex: The Protein-Conducting Channel

The protein-conducting channel consists of three or four copies of the Sec61 complex, appearing as a donut-like structure (Fig. 20-6A). Each Sec61 complex is a heterotrimer of three transmembrane polypeptides: an a subunit with 10 transmembrane segments, and smaller b and g-subunits, each with single transmembrane sequences. The homologous bacterial proteins are called SecY, SecE, and SecG. They form the SecY complex (Fig. 20-6C), the channel for translocating proteins across the plasma membrane of bacteria.

Figure 20-6 Structure of the Sec61 protein-conducting translocon. A–B, Three-dimensional reconstructions from electron micrographs: native translocon (channel) isolated from the ER and purified recombinant Sec61 oligomer. C, Tetrameric assembly of SecY complexes, as seen from electron micrographs. D, Ribbon diagram of SecY oligomer, showing potential path for lateral movement of the transmembrane segment of the translocated protein out of the channel. Inset, Side view of the channel with the front half of the model cut away. The pore ring represents the narrowest constriction between the two funnels of the channel. A segment of the SecY protein (“plug”) occludes the pore until the polypeptide causes it to open. E, A ribosome bound to the Sec61 channel. F–G, Orientation of the signal sequence within the Sec61 channel or SecY channel. The signal sequence is the blue portion of the purple polypeptide that threads through each channel.

Reconstructions of Sec61 complex from electron micrographs revealed a large central hole that might have been the pore for translocation of proteins across the ER bilayer. However, reconstructions at higher resolutions showed that the hole is not an aqueous channel, but a low-density region filled with lipid. This suggested that the pore was formed somewhere else, possibly within the Sec61 complex itself.

Crystal structures of the archaeal SecY complex revealed a small, hourglass-shaped pore within a single SecY complex (Fig. 20-6D). This pore is flanked on its luminal and cytoplasmic sides by funnels. The narrow constriction between the two funnels is only about 5 to 8 Å in diameter and lined by several hydrophobic side chains that together form the pore ring. Because of their small size and flexibility, the side chains forming the pore ring could fit snugly around a translocating peptide, preventing passage of ions or other small molecules. Another segment of SecY protein (termed the plug domain) occludes the pore in its inactive state and is proposed to shift away from the pore in its active state. The arrangement of the polypeptide allows the pore to open sideways, providing a lateral gate from the aqueous interior of the channel into the hydrophobic environment of the lipid bilayer that transmembrane domains of proteins use to access the lipid bilayer.

The structures strongly suggest that a single Sec61 complex functions as the pore for proteins to translocate across the ER. Several identical Sec61p complexes compose the translocon, but owing to its large size, only one ribosome can bind a translocon. Thus, only one Sec61p complex could be active for translocation. Why then is the translocon an oligomer of several Sec61 complexes? One possibility is that this creates binding sites for components associated with the translocon, such as the ER signal peptidase, the oligosaccharide transferase, TRAP, or TRAM. Oligomerization might also stabilize ribosome binding to the ER by increasing the number of binding sites that are made between Sec61 complex and the ribosome.

Signal Sequence

Binding of a signal sequence to the Sec61p complex opens or “gates” the translocon in preparation for protein translocation. Current evidence suggests that the signal sequence forms a loop in the channel with its N-terminus exposed to the cytoplasm (Fig. 20-6F). This occurs by signal sequence binding to transmembrane helices 2 and 7 on Sec61p.

The ability to recognize signal sequences allows Sec61 to discriminate substrates for translocation from other proteins. Traditionally, this has been thought to be a constitutive process that is predetermined by the sequences on the substrate. However, various cell types differ in the efficiency with which they recognize particular signal sequences. This can be explained if additional factors at the translocation site might influence signal sequence recognition. For example, proteins at or near the site of translocation (i.e., Sec62, Sec63, p180, p34, TRAM, or TRAP complex) might stimulate or inhibit the translocation of selected substrates by recognizing diversity within the signal sequence. Selective changes in expression or modifications of these accessory components in different cell types could then affect the outcome of translocation for different substrates.

Translocation of Soluble Proteins into the Lumen of the Endoplasmic Reticulum

Soluble proteins destined for secretion or retention in the lumen of the ER are translocated fully across the Sec61 protein-conducting channel into the lumen of the ER (Fig. 20-7A). Once the polypeptide has grown to about 150 residues during this process, a signal peptidase in the ER lumen removes the signal sequence from the translocating polypeptide. The signal peptidase consists of five subunits ranging in size from 12 kD to 25 kD and associates with the translocon. After signal peptide cleavage, the new N-terminus of the growing polypeptide continues to pass through the translocon until it is released into the ER lumen. The remaining cleaved signal peptide either is degraded or has other functions elsewhere in the cell.

Figure 20-7 Targeting of Sec61-dependent proteins to the lumen and membrane of the endoplasmic reticulum.

Insertion of Membrane Proteins in the Endoplasmic Reticulum Bilayer

Most proteins destined for insertion into the bilayer of the ER membrane use the Sec61 protein-conducting channel. During translocation of transmembrane proteins, some parts of the polypeptide chain are translocated across the lipid bilayer, whereas others are not. Depending on how the transmembrane protein is translocated and oriented across the bilayer, the protein is categorized as type 1, type 2, or polytopic (Fig. 20-7B-F). All copies of a particular polypeptide have the same orientation after translocation (i.e., type 1, type 2, or polytopic), and this orientation is usually maintained as the protein is carried to different membrane destinations in the cell through membrane budding and fusion events (see Chapters 21 and 22).

In type 1 transmembrane proteins, the N-terminal signal sequence initiates translocation, similar to soluble proteins (Fig. 20-7B). In the crystal structure of the archaeal SecY complex, the signal sequence binds to two adjacent helices that are proposed to open laterally to let hydrophobic segments of transmembrane proteins exit from the pore into the bilayer (see Fig. 20-6D). A stop-transfer signal (usually the future transmembrane domain) stops the transfer process before the type 1 polypeptide chain is completely translocated. The signal sequence (also called a start-transfer signal) is then cleaved off by the ER signal peptidase, and the polypeptide slides out of the translocon through the lateral exit site. The protein is now oriented with its N-terminus facing the ER lumen, its C-terminus facing the cytoplasm, and its transmembrane segment spanning the ER membrane.

Some type 1 proteins, called GPI-anchored proteins, exchange their carboxyl terminal transmembrane segment for an oligosaccharide anchored to the lipid phosphatidylinositol (Fig. 20-7C; also see Fig. 7-9). An enzyme in the ER lumen cleaves off the transmembrane segment and transfers the new C-terminus to a preassembled glycosylphosphatidylinositol (GPI) membrane anchor. Many cell surface proteins are attached to the plasma membrane in this manner. This allows them to be readily released from the cell when specific phospholipases in the plasma membrane are activated. For example, during sperm capacitation, many GPI-anchored proteins are cleaved and released from the sperm plasma membrane. This reorganization of the sperm’s cell surface is essential for a sperm to fertilize an egg.

Type 2 transmembrane proteins use a transmembrane domain as an internal signal sequence (Fig. 20-7D). Once such an internal signal sequence emerges from a ribosome, it is recognized by SRP and brought to the ER membrane, where it serves as a start-transfer signal to initiate protein translocation. When the protein is fully synthesized, the start-transfer signal, which is not cleaved off, slides out of the translocation channel into the surrounding lipid bilayer, where it serves as a transmembrane anchor.

Polytopic proteins that span the membrane multiple times (such as ion channels and carriers) utilize multiple stop-transfer signals, none of which are cleaved by signal peptides (Fig. 20-7E-F). SRP is probably required to tar-get the first signal sequence to the ER membrane. Thereafter, the dynamics of the channel must accommodate sequences that specify translocation of loops in the cytoplasm or lumen alternating with the transfer of transmembrane segments to the lipid bilayer.

Association of Tail-Anchored Proteins with the Endoplasmic Reticulum Membrane

Although most integral membrane proteins that integrate into the ER bilayer do so by being targeted to and translocated through the Sec61 translocation channel, a group of integral membrane proteins known as C-tail-anchored proteins (tail-anchored proteins) do not. These proteins are held in the phospholipid bilayer by a single stretch of hydrophobic amino acids close to the C-terminus and have their entire functional N-terminus facing the cytoplasm (Fig. 20-8).

Figure 20-8 targeting of tail-anchored proteins to the endoplasmic reticulum or mitochondrial membrane. Tail-anchored proteins have a short, hydrophobic transmembrane domain flanked on one or both sides by positively charged residues. The charged residues at the C-terminus and hydrophobic domain directly cross the bilayer with the aid of cytosolic factors and ATP. After insertion, the majority of the protein faces the cytoplasm with only the C-terminal charged residues in the lumen.

Tail-anchored proteins lack an N-terminal signal sequence and their membrane-interacting region is so close to the C-terminus that it emerges from the ribosome only on termination of translation. Because of this, tail-anchored proteins do not bind to SRP, which recognizes only signal peptides or signal anchors as long as they are part of a nascent polypeptide chain (i.e., still attached to the ribosome). Tail-anchored proteins also do not posttranslationally target to the Sec61 channel. Instead, they use a still unclear mechanism for membrane insertion that involves ATP, cytosolic factors, and membrane components, and that requires that the C-terminal anchor of the protein be hydrophobic. The C-terminus of a tail-anchored protein that crosses the bilayer doubles as a transmembrane anchor after insertion into the appropriate bilayer with only two to three hydrophilic residues translocated across the membrane. By contrast, most other transmembrane and soluble proteins translocate a much larger hydrophilic region across the membrane during their biogenesis.

Tail-anchored proteins have diverse roles in membrane biogenesis and traffic, as well as in cell metabolism (Table 20-3). They include SNARE proteins, such as syntaxins and synaptobrevins, which are responsible for membrane fusion events within cells (see Chapter 21); Sec61g and b of the translocation channel; cytochrome b₅, which participates in lipid metabolism; and Bcl and Bax, which are found on the mitochondrial outer membrane and regulate apoptosis.

Table 20-3 EXAMPLES OF TAIL-ANCHORED PROTEINS

Association of Lipid-Anchored Proteins with the Cytoplasmic Surface of the Endoplasmic Reticulum

Many classes of lipid-anchored proteins, including N- and H-Ras GTPases, are targeted from the cytoplasm to the cytoplasmic leaflet of the ER bilayer by posttranslational modification of a C-terminal cysteine-alanine-alanine-x (CAAX) motif. The first step in this process is prenylation, in which a soluble prenyltransferase attaches a farnesyl or geranylgeranyl lipid to the protein via a stable thioether linkage to the cysteine residue in CAAX. A prenyl-CAAX protease localized in the ER bilayer then cleaves the AAX residues, leaving the prenylcysteine as the new C terminus. The modified cysteine is then recognized by a prenylcysteine carboxyl methyltransferase (pcCMT) also in the ER bilayer that methylesterifies the a carboxyl group. (In the case of N- and H-Ras, a further modification occurs whereby one or two other cysteine residues upstream of the CAAX motif are modified by palmitic acid via a labile thioester linkage.) These hydrophobic C-terminal modifications anchor the otherwise hydrophilic molecule to the cytoplasmic surface of the ER membrane. To reach the plasma membrane, the lipid-anchored proteins follow the secretory pathway out of the ER.

N-linked Glycosylation

The majority of proteins synthesized in association with the ER are glycoproteins with covalently attached carbohydrates. One class of protein glycosylation takes place cotranslationally in the ER by the addition of a preformed oligosaccharide complex to asparagine residues (Fig. 20-9). These asparagine or N-linked oligosaccharides form flexible hydrated branches that can extend 3 nm or more out from the polypeptide. They frequently make up a sizable portion of the mass of a glycoprotein and cover a large fraction of its surface. Because the oligosaccharides are polar, glycosylation makes the protein more hydrophilic and is less likely to aggregate. By avoiding aggregation, the protein has a higher probability of folding into its correct conformation. Hence, oligosaccharides on proteins play a key role in enabling a newly synthesized protein to fold properly. Once correctly folded, the protein can leave the ER and move through the secretory pathway, where the sugars can be modified further. The great diversity of oligosaccharides found on secreted proteins is also crucial for their functions outside the cell.

Figure 20-9 dolichol pathway. A core oligosaccharide consisting of mannose and N-acetylglucosamine is synthesized in the cytoplasm, attached through high-energy pyrophosphate bonds to dolichol in the ER membrane. Following transfer across the ER bilayer, the addition of sugars imported into the ER completes the core structure. The oligosaccharide-transferase complex transfers the completed oligosaccharide to the consensus Asn-X-Ser/Thr motif of a nascent chain as it enters the lumen of the ER.

A single, preformed oligosaccharide precursor, composed of 14 sugars (three glucoses, nine mannoses, and two N-acetyl glucosamines) (Glc₃Man₉GlcNAc₂), serves as the core for N-linked oligosaccharides. This precursor is synthesized in a stepwise fashion while attached to the ER membrane by dolichol phosphate (a long-chained, unsaturated isoprenoid alcohol with pyrophosphate at one end; Fig. 20-9). Assembly of the oligosaccharide precursor involves 14 separate transfer reactions: seven on the cytoplasmic face of the ER and seven in the lumen. The mechanism that flips the glycolipid across the bilayer is unknown. The enzyme oligosaccharyltransferase recognizes dolichol and transfers the complete oligosaccharide to asparagine side chains on the nascent polypeptide contained in the sequences Asn-X-Ser/Thr, where X is any amino acid other than proline. The sequence must have emerged a distance of 12 to 14 residues out of the translocon into the ER lumen before it can be recognized.

Calnexin/Calreticulin Cycle

Once the core oligosaccharide has been added to the protein, the glycoprotein begins a cycle of modifications that help it to achieve its fully folded state. This calnexin cycle (Fig. 20-10) starts when glucosidases I and II remove the first two glucose residues of a core glycan. The resulting monoglucosylated transmembrane or soluble proteins bind to calnexin, a type I transmembrane protein in the ER lumen (Fig. 20-10). Monoglucosylated soluble proteins also bind to calreticulin, a soluble protein in the ER lumen similar to calnexin. Both proteins are related to sugar-binding lectin proteins from legumes. Both are monomeric, calcium-binding proteins. Binding to calnexin or calreticulin sequesters the glycoprotein and prevents its aggregation. It also exposes the glycoprotein to ERp57, a thiol-disulfide oxidoreductase. A close homolog of protein disulfide isomerase, ERp57 forms a complex with calnexin and calreticulin and specifically interacts with glycoproteins. ERp57 catalyzes intramolecular disulfide bond interchange during the folding process.

Figure 20-10 calnexin cycle of protein folding in the lumen of the endoplasmic reticulum and protein degradation from the endoplasmic reticulum. A, Glucosidases I and II rapidly remove two of three glucoses from newly synthesized, unfolded glycoprotein. The calnexin-thioloxidoreductase complex binds the monoglucosylated protein. Glucosidase II removes the remaining glucose, releasing the protein. If the released protein is folded, it can exit the ER. If unfolded, it is recognized by glucosyltransferase and reglucosylated so that it reenters the folding cycle until folding is complete, or it enters the degradation pathway by interacting with EDEM and a retrograde translocation channel, which deliver the unfolded polypeptide to the proteasome for degradation. Thioloxidoreductases catalyze rearrangements of disulfide bonds during folding. (Glc, glucose; Man, mannose; UDP, uridine diphosphate; UMP, uridine monophosphate.) B, Three-dimensional structure of calnexin showing a globular, lectin-binding domain and an extended arm composed of four repeat modules that folds around a sugar residue after it binds to the globular domain.

Release of bound glycoprotein chains from calnexin/ERp57 occurs once glucosidase II removes the remaining glucose residue on the core glycan. The glycoprotein is now free to leave the ER unless recognized by a soluble enzyme, UDP-Glc:glycoprotein glucosyltransferase (GT). GT reglucosylates only incompletely folded glycoproteins, so it serves as a folding sensor in the cycle. When reglucosylated by GT, the glycoprotein rebinds to calnexin or calreticulin. The glycoprotein stays in the cycle until it is properly folded and oligomerized, in which case it enters the secretory pathway. If the protein cannot fold or oligomerize properly, it is removed from the cycle by being translocated out of the ER into the cytoplasm, where it is degraded. When association with the cycle is inhibited, for example, by blocking the action of the glucosidases, the folding efficiency decreases. In this case, the glycoprotein may associate with BiP, which cooperates with the calnexin cycle in helping the protein to fold correctly.

Protein Degradation in Endoplasmic Reticulum

The ER has a highly regulated mechanism to prevent the export of dysfunctional proteins into the secretory pathway. Improperly folded polypeptides, excess subunits of oligomeric assemblies, or incorrectly assembled oligomers are degraded rather than being exported from the ER (Fig. 20-10). The degradation process, termed ER-associated degradation (ERAD), prevents accumulation of unsalvageable, misfolded proteins in the ER. ERAD occurs in four sequential steps: a misfolded glycoprotein is first recognized, retrotranslocated across the ER membrane to the cytoplasm, ubiquitinated (see Fig. 23-8), and then degraded by a proteasome in the cytoplasm.

For a protein to be targeted for ERAD, it must be misfolded or unassembled. The cell distinguishes a misfolded or unassembled protein subunit from a bona fide folding intermediate by linking degradation of glycoproteins to the trimming of mannoses (Fig. 20-10). The longer an unfolded glycoprotein remains in the ER, the more likely it will be exposed to mannosidases, which trim terminal mannose residues from the core oligosaccharide. When such trimming occurs, the glycoprotein becomes a substrate for ERAD and is recognized by a membrane-bound ER protein called EDEM (for “ER degradation-enhancing α-mannosidase-like protein”), which helps to direct the glycoprotein to the retrotranslocation channel. (Elimination of the EDEM homolog in yeast retards ERAD of glycoproteins but not of nonglycosylated proteins, which use a different pathway for degradation.) Because the concentration of mannosidases in the ER is low, newly synthesized proteins and nascent chains have time to fold correctly and thereby avoid having their terminal mannoses trimmed. Once mannose trimming on these molecules has occurred, EDEM begins to compete for these substrates with the calnexin/calreticulin cycle. The channel that exports ERAD substrates back into the cytoplasm for degradation (termed the retrotranslocon) remains to be identified.

Endoplasmic Reticulum Stress Responses and Endoplasmic Reticulum Folding Diseases

The folding pathway in the ER is tightly linked to the physiological state of the cell. Conditions that flood the ER with excess protein or result in accumulation of misfolded proteins trigger the unfolded protein response (UPR; Fig. 20-11). Essentially, any condition that exceeds the protein-folding capacity of the ER triggers the unfolded protein response: misfolding of mutant proteins, inhibition of ER glycosylation (by the drug tunicamycin), inhibition of disulfide formation (by reducing agents), or even overproduction of normal proteins. To compensate for these events, this stress-induced signaling pathway upregulates genes that are required to synthesize the entire ER, including folding machinery. In yeast, the unfolded protein response activates more than 300 genes involved with all aspects of ER function, including lipid synthesis, protein translocation, protein folding, glycosylation, and degradation, as well as export to and retrieval from the Golgi apparatus. Developmental programs might work through the same genetic controls to determine the abundance of ER in differentiated cells, producing, for example, extensive ER in secretory cells such as plasma, liver, and pancreatic acinar cells.

Figure 20-11 unfolded protein response pathways to stress in the lumen of the endoplasmic reticulum. A, ATF6 pathway. B, IRE1 pathway. C, PERK pathway.

The Unfolded Protein Response

The response to ER stress (i.e., UPR) is regulated by three key ER transmembrane proteins: IRE1 (inositol requiring 1), PERK/PEK (PKR-like endoplasmic reticulum kinase/pancreatic eIF2a kinase), and ATF6 (activating transcription factor 6). These proteins serve as stress sensors to regulate the production of bZIP (basic leucine zipper) domain–containing transcription factors (see Fig. 15-17) that upregulate genes involved in ER function. This allows cells to adjust the capacity of the ER to promote ER folding depending on the demand. Whereas only IRE1 is present in yeast, all three transmembrane proteins are present and function in metazoan cells.

Accumulation of unfolded proteins in the ER lumen of metazoan cells results in activation of ATF6, IRE1, and PERK, each by a different mechanism (Fig. 20-11). Under normal conditions, free BiP is thought to inhibit the UPR pathway. BiP binds to the lumenal domains of IRE1 and PERK, preventing their dimerization. BiP also associates with ATF6, retaining it in the ER. When unfolded proteins accumulate during ER stress, BiP binds to them rather than to IRE1, ATF6, and PERK.

When released from BiP, ATF6 is transported to the Golgi, where it is cleaved by S1P and S2P proteases to produce a cytoplasmic fragment. The fragment moves to the nucleus, where it activates the transcription of responsive genes, including XBP1.

Freed of BiP, IRE1 dimerizes. This activates IRE1′s cytoplasmic endoribonuclease activity allowing it to remove a small intron from the XBP1 mRNA. This alters the translational reading frame of XBP1 to make a protein that is a potent transcriptional activator.

When dissociated from BiP, PERK phosphorylates the eukaryotic translation initiation factor 2 (eIF2). This reduces the frequency of AUG codon recognition and thereby slows the rate of translation initiation on many mRNAs. The mRNAs that are preferentially translated under these conditions are all involved in cell survival and ER functions.

Aspects of the UPR pathway involving IRE1 and PERK are also important for promoting differentiation in higher eukaryotic cells. For example, IRE1 is activated during β-lymphocyte differentiation into a plasma cell (see Fig. 28-9), in which the ER expands 5-fold to accommodate immunoglobulin synthesis. Activation of the innate immune response (i.e., the inflammatory response), for example by lipopolysaccharide (LPS) treatment, also activates IRE1. Furthermore, PERK activity is required for β-cell differentiation and/or survival. These findings have led to the view that the UPR allows cells to respond to ER stress, to provide a way to sense nutrients and to promote differentiation.

Endoplasmic Reticulum Folding Diseases

Given the central role of the ER in the synthesis of proteins for the entire exocytic and endocytic pathways, it is not surprising that many inherited diseases are a direct consequence of proteins failing to pass ER quality control. Many metabolic disorders, including some lysosomal storage diseases (see Appendix 23-1), are a direct consequence of key enzymes failing to be exported from the ER. The most common form of cystic fibrosis is due to the inability of cells to export a mutant form of the cystic fibrosis transmembrane regulator (CFTR) to the cell surface, where it would normally function as a chloride channel for the respiratory system and pancreas (see Fig. 11-4). Similarly, the inability of the liver to secrete mutated forms of a1-antitrypsin predisposes to the lung disease emphysema. Normally, a1-antitrypsin protects tissues by inhibiting extracellular proteases such as elastase, which is produced by neutrophils. Mutations prevent a1-antitrypsin folding in the ER, resulting in its degradation. The resulting deficiency in a1-antitrypsin circulating in the blood allows elastase to destroy lung tissue, leading to emphysema. In severe cases, mutant forms of the protein not only fail to be exported from the ER but also elude degradation pathways, accumulating as insoluble aggregates that induce stress responses and liver failure.

In some conditions, the ER uses the unfolded protein response to compensate, in part, for mutations in cargo proteins. In congenital hypothyroidism, mutant thyroglobulin (the precursor of thyroid hormone) is not exported efficiently from the ER. Excess protein accumulates as insoluble aggregates in the ER. Feedback pathways trigger massive proliferation of ER in an attempt to produce normal levels of circulating hormone. Similarly, in mild forms of osteogenesis imperfecta (see Chapter 32), osteoblasts assemble and secrete defective procollagen chains for bone synthesis, even though the resulting bone tissue is weak. The alternative, complete loss of procollagen by retention and degradation of the mutant procollagen in the ER would be lethal. Faulty ER quality control may also contribute to diseases of the central and peripheral nervous systems, including Alzheimer’s disease.

Cholesterol Synthesis and Metabolism

Cholesterol is maintained in animal cells by a combination of de novo synthesis (Fig. 20-13) in the ER and receptor-mediated endocytosis of lipoprotein particles containing esterified cholesterol (see Chapter 22). Coordinated regulation of these two processes precisely maintains the physiological level of cholesterol in cellular membranes. Peroxisomes also may participate in aspects of cholesterol synthesis and metabolism.

Figure 20-13 Key steps in the biosynthesis of cholesterol from acetyl CoA. Synthesis of mevalonate from 3-hydroxy-3-methylglutaryl CoA is closely regulated by controlling the concentration of the enzyme, HMG-CoA reductase, as shown in Figure 20-14. Five-carbon isoprene groups are the building blocks (yellow) for making a 10-carbon geranyl-pyrophosphate (PP), a 15-carbon farnesyl-pyrophosphate, and 30-carbon squalene intermediates. Several reactions add a hydroxyl group and join squalene into four rings to make cholesterol. ADP, adenosine diphosphate; NADP, nicotinamide adenine dinucleotide phosphate.

Enzymes in the cytoplasm and ER use 22 sequential steps to synthesize cholesterol from acetate (Fig. 20-13). Cytoplasmic enzymes catalyze the initial steps, using water-soluble molecules to produce farnesyl-pyrophosphate (farnesyl-PP) from acetyl coenzyme A (acetyl CoA). An important exception is the step going from 3-hydroxy-3methylglutaryl CoA (HMG-CoA) to mevalonate. A carefully regulated, integral membrane protein of the ER (Fig. 20-13), HMG-CoA reductase, catalyzes this key step. Enzymes in the ER bilayer catalyze the subsequent condensation of farnesyl-PP to make squalene, and cyclization to cholesterol, progressively less-polar molecules. Although the final steps leading to cholesterol take place in the ER, cholesterol is not a resident ER lipid and is rapidly exported to post-ER membranes, including the plasma membrane, where it constitutes up to 50% of the bilayer. Chapter 21 discusses the distribution of cholesterol in different organelles and current ideas regarding its transport.

Feedback loops sense the cholesterol content of the ER membrane and regulate both the synthesis and degradation of enzymes that synthesize cholesterol (Fig. 20-14). High cholesterol levels inhibit the synthesis and stimulate the destruction of key synthetic enzymes. A novel transcription factor precursor, steroid regulator element-binding proteins (SREBP), controls the expression of these genes. When cholesterol is abundant, SREBP cleavage-activating protein (SCAP), a partner protein with cholesterol-sensing transmembrane segments, retains SREBP in the ER owing to interactions with the protein Insig. When the membrane cholesterol level is low, SCAP and Insig proteins do not interact. The SREBP-SCAP complex is then free to move to the Golgi apparatus, where two successive proteolytic cleavages release the N-terminal domain of SREBP, a basic helix-loop-helix leucine zipper transcription factor, into the cytoplasm. (The Golgi proteases that are responsible for cleaving SREBP are SP1 and SP2, the same ones used to process ATF6 during UPR.) In the nucleus, the transcription factor binds steroid regulatory elements, enhancers for a wide range of genes encoding enzymes that synthesize cholesterol and other lipids, as well as low-density lipoprotein receptors that take up cholesterol from the medium (see Fig. 23-9). Cholesterol also regulates degradation of HMG-CoA reductase, which has cholesterol-sensing transmembrane domains similar to SCAP. Abundant cholesterol targets HMG-CoA reductase for degradation by the proteolytic pathway that disposes of unfolded proteins through the proteasome (Fig. 20-10 and Chapter 23). The enzyme acyl-CoA-cholesterol transferase (ACAT) helps to lower cholesterol levels in the ER bilayer by catalyzing the formation of cholesterol esters, a storage form of cholesterol.

Figure 20-14 control of cholesterol biosynthesis by proteolysis. A, High-cholesterol conditions. Cholesterol-sensing transmembrane segments (pink) of SCAP (SREBP cleavage-activation proteins) retain intact SREBP in the ER through an interaction with Insig. Similar cholesterol-sensing transmembrane segments of HMG-CoA reductase stimulate its destruction by proteolysis. B, Low-cholesterol conditions. Insig dissociates from SCAP and SREBP, allowing these molecules to move to the Golgi apparatus where the membrane-anchored site 1 serine protease cleaves the loop of SREBP in the lumen. Subsequently, an unusual transmembrane zinc-protease cleaves SREBP at a second site within the bilayer, releasing the basic helix-loop-helix-zip (bHLH-zip) transcription factor. This factor enters the nucleus, where it activates genes for cholesterol biosynthetic enzymes by steroid-response elements (SREs).

Ceramide Synthesis

Ceramide, the backbone of all sphingolipids (see Fig. 7-3), also begins its synthesis on the cytoplasmic face of ER membranes. It is made through sequential condensation of the amino acid serine with two fatty acids. Ceramide is transported to the Golgi apparatus by a nonvesicular pathway (see later), where enzymes on the lumenal leaflet either add oligosaccharide chains to it to form glycosphingolipids or add a choline head group to form sphingomyelin (see Chapter 21).

Lipid Movement between Organelles

While lipids can rapidly diffuse along one side of the bilayer as well as flip-flop across the bilayer, the hydrophobic nature of lipids makes free diffusion between adjacent bilayers thermodynamically unfavorable. Thus, lipids use two other mechanisms to move between organelles (Fig. 20-15). The major pathway is movement through the use of vesicle carriers (see Chapter 21). In addition, small, soluble phospholipid exchange proteins promote movements between membranes by a nonvesicular mechanism. These proteins—called lipid-transfer proteins (LTPs)—have a common overall structure and are specific for particular phospholipids. One example of an LTP is ceramide transport protein (CERT), which mediates ceramide transport to the Golgi apparatus by extracting newly synthesized ceramide from the ER and carrying it to the Golgi apparatus. To accomplish this, CERT has a domain that recog-nizes ceramide and mediates its intermembrane transfer. CERT also has a motif for targeting to the ER and a PH domain (see Fig. 25-11) for binding polyphosphoinositides in the Golgi apparatus.

Figure 20-15 mechanisms of lipid movement within, across, and between membrane bilayers. Membrane lipids can (1) laterally diffuse within the bilayer, (2) translocate between two leaflets of the bilayer, (3) move through the cytosol from one bilayer to another by attaching to a cytosolic lipid-transfer protein, or (4) be transported from one membrane compartment to another by incorporating into membrane-bound transport carriers.

LTPs catalyze lipid exchange but not net transfer. When the protein delivers a lipid to a target membrane, the protein exchanges that lipid for another one and returns with the second lipid (Fig. 20-15). The lipid-binding pocket of LTPs is largely lined with hydrophobic residues that stabilize the internal lipid. In the open conformation, a hinged cover projects outward away from the protein, allowing the LTP to embed itself partially into the cytosolic leaflet of the bilayer. Interactions with other membrane components (either lipids or proteins) dictate what compartments the LTP targets to. Zones of close apposition between the ER and other organelles—such as mitochondria, chloroplasts, lipid droplets, TGN, endosomes, and lysosomes—are enriched in certain types of LTPs, suggesting that lipid trafficking occurs across these sites. Some proteins with lipid transfer domains might function as lipid sensors rather than as lipid carriers.