17.03.04 · mol-cell-bio / cell-organization

Endoplasmic reticulum and Golgi: protein folding, glycosylation, and vesicle budding

stub3 tiersLean: nonepending prereqs

Anchor (Master): Helenius, A. et al. — Curr. Opin. Cell Biol. 15 (2003) 431-439; Barlowe, C. — Annu. Rev. Cell Dev. Biol. 17 (2001) 1-24

Intuition Beginner

The endoplasmic reticulum (ER) is a folded membrane network that extends from the nucleus throughout the cell. It comes in two forms. The rough ER is studded with ribosomes on its outer surface and makes proteins destined for secretion, the plasma membrane, or other organelles. The smooth ER lacks ribosomes and synthesises lipids and detoxifies harmful substances.

When ribosomes on the rough ER make a protein, the growing polypeptide chain is threaded through a channel into the interior space of the ER, called the lumen. Inside the lumen, the protein folds into its correct three-dimensional shape with help from chaperone proteins. If a protein fails to fold properly, the ER retains it and eventually sends it for destruction. This quality-control system ensures that only correctly folded proteins leave the ER.

After passing quality control, proteins are packed into tiny membrane sacs called vesicles that bud off from the ER and travel to the Golgi apparatus. The Golgi is a stack of flattened membrane compartments (cisternae) that receives proteins, modifies them, sorts them, and ships them to their final destinations. One major modification is glycosylation — attaching sugar chains to proteins. These sugar tags help proteins fold correctly, protect them from degradation, and serve as address labels for sorting.

The Golgi has a receiving face (the cis-Golgi, near the ER) and a shipping face (the trans-Golgi network, or TGN). Proteins enter at the cis face, pass through the medial cisternae where they are modified step by step, and exit from the TGN. At the TGN, proteins are sorted into different vesicles depending on their destination: the plasma membrane, lysosomes, or secretory granules.

Visual Beginner

The diagram illustrates the flow of proteins through the ER and Golgi. Ribosomes (small dots) stud the rough ER membrane, feeding nascent polypeptides into the ER lumen. Inside the lumen, chaperone proteins (orange) assist folding. A vesicle buds from the ER, coated in COPII proteins, and travels to the cis-Golgi. The Golgi apparatus is drawn as a stack of four curved cisternae (cis, medial-1, medial-2, trans). Enzymes within each cisterna modify the glycan chains on cargo proteins. At the trans-Golgi network, three vesicle populations depart: one toward the plasma membrane, one toward lysosomes, and one constitutive secretory vesicle.

Worked example Beginner

Tracing a glycoprotein through the ER-Golgi pathway. Consider erythropoietin (EPO), a hormone produced by kidney cells that stimulates red blood cell production. EPO is a heavily glycosylated protein — roughly 40% of its molecular weight comes from attached sugar chains.

Step 1. Synthesis and ER entry. EPO mRNA is translated by ribosomes docked on the rough ER. A signal sequence at the N-terminus directs the nascent polypeptide into the ER lumen through the Sec61 channel. The signal sequence is cleaved off.

Step 2. Folding and N-glycosylation in the ER. As the polypeptide enters the lumen, an enzyme called oligosaccharyltransferase attaches a pre-built 14-sugar tree (two N-acetylglucosamines, nine mannoses, three glucoses) to specific asparagine residues. This is N-linked glycosylation. Chaperone proteins (BiP and calnexin) bind the glycoprotein and help it fold. If folding succeeds, the protein is released and packaged for export.

Step 3. ER-to-Golgi transport. Correctly folded EPO is captured into a COPII-coated vesicle that buds from an ER exit site and fuses with the cis-Golgi.

Step 4. Golgi modifications. As EPO passes through the Golgi cisternae, enzymes trim and remodel the sugar chains. Mannoses are removed and replaced with N-acetylglucosamine, galactose, and sialic acid. These terminal sugars are critical for EPO stability in the bloodstream — desialylated EPO is rapidly cleared from circulation.

Step 5. Sorting and secretion. At the TGN, EPO is packaged into constitutive secretory vesicles that fuse with the plasma membrane, releasing EPO into the blood.

Check your understanding Beginner

Exercise (easy, multiple choice).

What distinguishes rough ER from smooth ER?

A. Rough ER synthesises lipids; smooth ER synthesises proteins B. Rough ER has ribosomes attached; smooth ER does not C. Rough ER is found in plant cells; smooth ER is found in animal cells D. Rough ER is inside the nucleus; smooth ER is outside

Hint

Ribosomes are the small structures that synthesise proteins. Their presence or absence on the ER membrane gives each type its name and function.

Answer

Option B. The rough ER is studded with ribosomes that synthesise proteins destined for the secretory pathway (secretion, plasma membrane, endomembrane organelles). The smooth ER lacks ribosomes and instead synthesises lipids, metabolises carbohydrates, and detoxifies drugs and poisons. Both are continuous extensions of the nuclear envelope.

Formal definition Intermediate+

The endoplasmic reticulum is the largest membrane-bound organelle in most eukaryotic cells, comprising a continuous network of tubules, sheets, and cisternae that extends from the nuclear envelope throughout the cytoplasm. It is divided into functionally distinct domains:

Rough ER (RER). Studded with membrane-bound ribosomes. The site of co-translational translocation of secretory and membrane proteins via the Sec61 translocon, N-linked glycosylation, disulfide bond formation, and the initial stages of protein folding.
Smooth ER (SER). Ribosome-free. The site of lipid and steroid synthesis, drug detoxification (cytochrome P450 enzymes), and calcium storage (via the SERCA calcium pump). In specialised cells, the SER is highly developed: skeletal muscle has the sarcoplasmic reticulum (a SER derivative dedicated to calcium release for contraction), and hepatocytes have extensive SER for drug metabolism.
ER exit sites (ERES). Discrete regions where COPII-coated vesicles bud. Marked by Sec16, which nucleates COPII coat assembly. Mammalian cells typically have several hundred ERES.

Protein translocation via the Sec61 translocon. The Sec61 complex (Sec61 $α$ , $β$ , $γ$ ) forms a protein-conducting channel in the ER membrane. Co-translational translocation proceeds as follows: (1) the signal recognition particle (SRP) binds the N-terminal signal sequence as it emerges from the ribosome; (2) SRP docks the ribosome at the SRP receptor on the ER membrane; (3) the ribosome engages Sec61, and the growing polypeptide is threaded through the channel into the ER lumen; (4) signal peptidase cleaves the signal sequence. For membrane proteins, transmembrane segments partition laterally through the Sec61 gate into the lipid bilayer.

N-linked glycosylation. Oligosaccharyltransferase (OST), a multiprotein complex associated with the Sec61 translocon, transfers the preassembled oligosaccharide Glc $_{3}$ Man $_{9}$ GlcNAc $_{2}$ from the lipid carrier dolichol pyrophosphate to asparagine residues within the consensus sequon N-X-S/T (where X is any residue except proline). This transfer occurs co-translationally as the nascent chain enters the ER lumen. The dolichol-linked oligosaccharide is assembled in a stepwise pathway involving 14 enzymatic reactions on both the cytoplasmic and luminal faces of the ER membrane, with the lipid carrier dolichol serving as a scaffold.

The calnexin/calreticulin cycle. After N-glycan transfer, glucosidases I and II sequentially remove the two outer glucose residues, producing a monoglucosylated glycan (Glc $_{1}$ Man $_{9}$ GlcNAc $_{2}$ ). This structure is recognised by the ER lectin chaperones calnexin (membrane-anchored) and calreticulin (soluble). Binding to these chaperones prevents aggregation and promotes proper folding. Glucosidase II then removes the final glucose, releasing the glycoprotein. If folding is incomplete, UDP-glucose glucosyltransferase (UGGT) acts as a folding sensor: it reglucosylates incompletely folded glycoproteins, returning them to calnexin/calreticulin for another round. This cycle continues until the protein achieves its native conformation.

ER-associated degradation (ERAD). Proteins that fail to fold after prolonged cycling through the calnexin cycle are diverted to ERAD. ER mannosidases (ERManI, EDEMs) trim mannose residues, generating a timer signal. The lectins OS-9 and XTP3-B recognise trimmed glycans and deliver substrates to the Hrd1 ubiquitin ligase complex. Hrd1 forms a retrotranslocation channel, and the AAA-ATPase p97 (VCP/Cdc48) extracts ubiquitinated substrates to the cytosol for proteasomal degradation. Three ERAD branches handle substrates from different topological locations: ERAD-L (luminal), ERAD-M (membrane), and ERAD-C (cytoplasmic domains).

Golgi apparatus architecture and function. The Golgi consists of a stack of 4-8 flattened cisternae with distinct functional zones: cis-Golgi (receiving), medial-Golgi (processing), and trans-Golgi (sorting). The trans-Golgi network (TGN) is a tubulovesicular network that serves as the major sorting station. Two models explain intra-Golgi transport:

Vesicle transport model. Each cisterna is a stable compartment, and COPI-coated vesicles shuttle cargo forward between cisternae.
Cisternal maturation model. Each cisterna matures from cis to trans over time; resident enzymes are carried backward in COPI vesicles as the cisterna progresses forward. Live-cell imaging of yeast Golgi supports the cisternal maturation model.

Golgi glycosylation reactions. The Golgi harbours dozens of glycosyltransferases that remodel N-linked glycans and attach O-linked glycans. N-glycan processing in the Golgi involves: (1) $α$ -mannosidase I removes mannose residues; (2) N-acetylglucosaminyltransferase (GnT) adds GlcNAc branches; (3) $α$ -mannosidase II removes two more mannoses, creating the substrate for complex glycan formation; (4) galactosyltransferases and sialyltransferases add terminal galactose and sialic acid residues. Each enzyme is localised to a specific cisterna, creating an assembly-line processing of glycoproteins as they transit through the stack.

COPII vesicle budding from ER exit sites. The formation of COPII-coated vesicles proceeds through an ordered assembly:

Sar1 activation. Sec12 (a guanine nucleotide exchange factor) catalyses GDP-to-GTP exchange on Sar1. Sar1-GTP exposes its amphipathic N-terminal helix, which inserts into the ER membrane.
Inner coat recruitment. Sar1-GTP recruits the Sec23/Sec24 heterodimer. Sec24 is the primary cargo adaptor, recognising di-acidic (DxE), di-hydrophobic (FF, YY), and other export signals on transmembrane cargo, as well as cargo receptors for soluble proteins.
Outer coat polymerisation. The Sec13/Sec31 heterotetramer polymerises into a cuboctahedral cage that deforms the membrane into a bud. Sec31 stimulates the GAP activity of Sec23 on Sar1, creating a built-in timer for coat disassembly.
Scission. Membrane constriction at the bud neck, driven by continued cage polymerisation and Sar1-GTP hydrolysis, severs the vesicle. The uncoated vesicle (60-80 nm diameter) is then available for tethering and fusion with the cis-Golgi or ERGIC.

For large cargo such as procollagen fibrils (300-400 nm), specialised adaptors (TANGO1, cTAGE5) and the scaffold protein Sec16 generate oversized COPII carriers.

Key mechanism Intermediate+

Mechanism: The calnexin/calreticulin quality-control cycle and N-glycan-dependent folding.

The calnexin cycle is the central quality-control system for glycoprotein folding in the ER. It operates as a gated loop: entry requires a specific glycan signal, exit requires attainment of the native fold, and persistence triggers degradation. The mechanism can be described as a series of discrete steps:

Step 1. Glycan transfer and initial trimming. Oligosaccharyltransferase (OST) transfers Glc $_{3}$ Man $_{9}$ GlcNAc $_{2}$ to an N-X-S/T sequon on the nascent polypeptide. Glucosidase I removes the outermost glucose ( $α$ 1-2 linkage), and glucosidase II removes the middle glucose ( $α$ 1-3 linkage). The product, Glc $_{1}$ Man $_{9}$ GlcNAc $_{2}$ , is the substrate for lectin chaperone binding.

Step 2. Calnexin/calreticulin binding. The single remaining glucose is recognised by the lectin domain of calnexin or calreticulin. These chaperones bind the glycoprotein with the assistance of ERp57 (a thiol oxidoreductase) that promotes disulfide bond formation. The binding holds the glycoprotein in a partially folded state, preventing aggregation and allowing structured folding of individual domains.

Step 3. Release and folding assessment. Glucosidase II removes the final glucose, releasing the glycoprotein from calnexin/calreticulin. The protein either folds to its native state or remains partially unfolded. If folding is incomplete, UGGT (UDP-glucose glucosyltransferase) scans the protein surface. UGGT is the key folding sensor: it recognises exposed hydrophobic patches that are characteristic of non-native conformations and reglucosylates the N-glycan at the same position, recreating the monoglucosylated signal.

Step 4. Re-entry or exit. Reglucosylated proteins rebind calnexin/calreticulin for another folding attempt. Successfully folded proteins are not reglucosylated by UGGT (which cannot recognise the buried hydrophobic patches of a natively folded protein) and are therefore free to exit the ER via COPII vesicles.

Step 5. Mannose trimming and ERAD commitment. Proteins that cycle unsuccessfully through the calnexin system are eventually trimmed by ER $α$ -1,2-mannosidases (ERManI, EDEM1/2/3), which remove mannose residues from the N-glycan. Mannose trimming is effectively a molecular timer: the longer a protein remains in the ER, the more mannoses are removed. Trimming to a specific Man $_{5 - 6}$ GlcNAc $_{2}$ structure is recognised by the lectins OS-9 and XTP3-B, which deliver the substrate to the Hrd1-Sel1L ubiquitin ligase complex for ERAD.

The cycle is nontrivial because UGGT must distinguish between partially folded intermediates (which should be reglucosylated and given another chance) and irreparably misfolded proteins (which should be committed to degradation). UGGT achieves this by sensing the extent of exposed hydrophobic surface area: proteins with small exposed patches receive another glucosylation cycle, while proteins with extensive exposed hydrophobicity may have their glycans trimmed by mannosidases before UGGT can act, shunting them to ERAD.

Exercises Intermediate+

Exercise 1 (easy, multiple choice).

Which enzyme attaches the pre-assembled oligosaccharide to asparagine residues during N-linked glycosylation?

A. Glucosidase I B. Oligosaccharyltransferase (OST) C. UDP-glucose glucosyltransferase (UGGT) D. $α$ -mannosidase I

Hint

The transfer of the 14-sugar oligosaccharide from dolichol pyrophosphate to the protein occurs co-translationally at the translocon. Which enzyme catalyses this transfer?

Answer

Option B. Oligosaccharyltransferase (OST) is a multiprotein complex associated with the Sec61 translocon that transfers the pre-assembled Glc $_{3}$ Man $_{9}$ GlcNAc $_{2}$ oligosaccharide from the lipid carrier dolichol pyrophosphate to asparagine residues within the N-X-S/T sequon. This occurs co-translationally as the nascent polypeptide enters the ER lumen. Glucosidase I and II then trim the glucose residues to generate the calnexin-binding substrate. UGGT reglucosylates incompletely folded proteins, and mannosidase I trims mannoses in the Golgi.

Exercise 2 (easy, multiple choice).

What is the function of UGGT in the calnexin cycle?

A. It removes the final glucose from the N-glycan B. It adds mannose residues to the glycan C. It reglucosylates incompletely folded glycoproteins, returning them to calnexin/calreticulin D. It ubiquitinates misfolded proteins for ERAD

Hint

UGGT stands for UDP-glucose glucosyltransferase. The "glucosyltransferase" part indicates its enzymatic activity. What does adding a glucose do in this context?

Answer

Option C. UGGT is the folding sensor of the calnexin cycle. After glucosidase II removes the final glucose and releases the glycoprotein from calnexin, UGGT scans the protein for exposed hydrophobic patches. If it detects incomplete folding, UGGT reglucosylates the N-glycan, recreating the monoglucosylated signal that is recognised by calnexin/calreticulin. This returns the glycoprotein for another round of chaperone-assisted folding. UGGT does not act on natively folded proteins because their hydrophobic residues are buried.

Exercise 3 (medium, short answer).

Describe the sequence of events in COPII vesicle formation at an ER exit site, naming the key proteins at each step. What is the role of Sec24 in cargo selection?

Hint

COPII assembly begins with Sar1 activation by Sec12 and proceeds through inner coat and outer coat recruitment. Sec24 has a specific role in recognising cargo signals.

Answer

Sar1 activation. Sec12 (an ER-resident GEF) exchanges GDP for GTP on Sar1. Sar1-GTP undergoes a conformational change that exposes its amphipathic N-terminal helix, inserting it into the cytoplasmic leaflet of the ER membrane.
Inner coat recruitment. Sar1-GTP recruits the Sec23/Sec24 heterodimer. Sec24 is the primary cargo adaptor: it recognises di-acidic (DxE), di-hydrophobic (FF, YY), and other export signals on the cytoplasmic tails of transmembrane cargo proteins. Sec24 also binds cargo receptors (such as ERGIC-53 for soluble glycoproteins), allowing indirect capture of soluble cargo.
Outer coat assembly. The Sec13/Sec31 heterotetramer polymerises into a polyhedral cage that deforms the membrane into a bud. Sec31 stimulates the GAP activity of Sec23 on Sar1, providing a timer for coat disassembly.
Scission and uncoating. GTP hydrolysis by Sar1 drives constriction and severing of the bud neck. The coat disassembles, and the uncoated vesicle (60-80 nm) is available for fusion with the ERGIC or cis-Golgi.

Exercise 4 (medium, short answer).

Compare the cisternal maturation model and the vesicle transport model for intra-Golgi trafficking. Which model is currently favoured, and what evidence supports it?

Hint

Consider how large cargo (such as procollagen fibrils, which are much larger than typical COPI vesicles) moves through the Golgi. Also consider what happens to resident Golgi enzymes over time.

Answer

The vesicle transport model proposes that each Golgi cisterna is a stable compartment and that COPI-coated vesicles shuttle cargo forward between adjacent cisternae. The cisternal maturation model proposes that each cisterna matures from cis to trans over time: new cis-cisternae form from incoming ERGIC/COPII material, and resident enzymes are transported backward in COPI vesicles as the cisterna itself progresses forward.

The cisternal maturation model is currently favoured. Key evidence includes: (1) Large cargo such as procollagen fibrils (300-400 nm) are too large to fit into 60-80 nm COPI vesicles, yet they traverse the Golgi efficiently — they remain within a single cisterna that matures around them. (2) Live-cell imaging in yeast (Losev et al. 2006 Nature 441, 1002-1006; Matsuura-Tokita et al. 2006 Nature 441, 1007-1010) showed that cis-Golgi resident enzymes are replaced by medial-Golgi enzymes within the same cisterna over time, directly visualising maturation. (3) COPI vesicles carry resident enzymes backward but rarely carry forward-moving cargo, consistent with a maturation mechanism.

Exercise 5 (medium, short answer).

Explain the role of mannose trimming in the ER quality-control system. How does it function as a molecular timer for ERAD commitment?

Hint

The initial N-glycan has nine mannose residues. ER mannosidases remove these slowly. What recognises the trimmed glycan, and what is the consequence?

Answer

The initial N-linked glycan transferred by OST contains nine mannose residues (Man $_{9}$ GlcNAc $_{2}$ after glucose removal). ER $α$ -1,2-mannosidases (ERManI and EDEM proteins) progressively remove mannose residues from the glycan. This trimming is slow — each mannose removal takes time — so a protein that rapidly achieves its native fold exits the ER before significant trimming occurs. A protein that remains in the ER for an extended period, cycling repeatedly through the calnexin system, accumulates mannose loss.

When the glycan is trimmed to approximately Man $_{5 - 6}$ GlcNAc $_{2}$ , it is recognised by the lectins OS-9 and XTP3-B, which are components of the Hrd1 ubiquitin ligase complex. These lectins deliver the substrate to Hrd1 for retrotranslocation and ubiquitination. Thus, mannose trimming acts as a molecular timer: the number of remaining mannoses correlates with the time spent in the ER, and below a critical threshold, the protein is committed to ERAD regardless of its folding state. This timer prevents indefinitely repeated folding attempts on proteins that are unlikely to ever achieve their native conformation.

Exercise 6 (hard, short answer).

The most common cystic fibrosis mutation, $Δ$ F508 in the CFTR protein, causes the protein to be retained in the ER and degraded by ERAD, even though the mutant protein retains partial chloride channel activity if it reaches the plasma membrane. Explain why $Δ$ F508-CFTR is recognised as misfolded by the ER quality-control system, and discuss the implications for therapeutic strategies.

Hint

$Δ$ F508 is a deletion of phenylalanine at position 508 in the first nucleotide-binding domain (NBD1). This domain still folds but with altered thermodynamic stability. Consider what the ER quality-control system detects and whether a partially functional protein can be "rescued."

Answer

The $Δ$ F508 mutation deletes phenylalanine 508 from the first nucleotide-binding domain (NBD1) of CFTR. This residue sits at the interface between NBD1 and the intracellular loop 4 (ICL4) of the transmembrane domain. The deletion does not prevent NBD1 from folding entirely, but it destabilises the domain by approximately 2-3 kcal/mol and disrupts the NBD1-ICL4 interdomain interaction. This exposes hydrophobic surfaces that are normally buried, which are detected by the ER quality-control system. Calnexin and Hsp70 chaperones bind the exposed hydrophobic patches, retaining $Δ$ F508-CFTR in the ER. The protein is eventually diverted to ERAD (via the Hrd1 complex) and degraded by the proteasome.

Crucially, $Δ$ F508-CFTR retains partial chloride channel function — if experimentally forced to the plasma membrane (by low-temperature incubation at 26-30 $^{\circ}$ C, or by pharmacological chaperones), the protein reaches the surface and conducts chloride at roughly 25-50% of wild-type activity. This means the ER quality-control system is making a binary decision (degrade or export) on a protein that falls below the folding threshold but is not non-functional.

Therapeutic strategies target this gap. Corrector compounds (lumacaftor, tezacaftor, elexacaftor) bind $Δ$ F508-CFTR and stabilise the NBD1-ICL4 interface, reducing the exposure of hydrophobic surfaces and allowing some fraction of the protein to escape ER quality control and reach the plasma membrane. Potentiator compounds (ivacaftor) then enhance the gating of the rescued channels at the surface. The triple-combination therapy elexacaftor/tezacaftor/ivacaftor (Trikafta) restores approximately 50-60% of CFTR function in $Δ$ F508 patients and is the current standard of care.

Exercise 7 (hard, short answer).

Describe the unfolded protein response (UPR), naming the three ER transmembrane sensors and their downstream signalling pathways. How does the UPR shift from a pro-survival to a pro-apoptotic output under prolonged ER stress?

Hint

The three sensors are IRE1, PERK, and ATF6. All three are normally kept inactive by binding of the chaperone BiP. When misfolded proteins accumulate, BiP is titrated away. Consider the timescale of each pathway and what happens if stress persists.

Answer

The UPR activates when misfolded proteins accumulate in the ER, titrating BiP away from three ER transmembrane sensors:

IRE1 (inositol-requiring enzyme 1) is a kinase-endoribonuclease. Upon activation by BiP release and direct binding of unfolded proteins to its lumenal domain, IRE1 dimerises, autophosphorylates, and activates its RNase domain. The RNase splices XBP1 mRNA, removing a 26-nucleotide intron that creates a frameshift, producing the active transcription factor XBP1s. XBP1s upregulates chaperones (BiP, PDI), ERAD components (Hrd1, EDEM), and lipid biosynthesis genes to expand ER capacity. Under prolonged stress, IRE1's RNase activity becomes less specific (RIDD — regulated IRE1-dependent decay), degrading ER-localised mRNAs to reduce the protein-folding load.
PERK (PKR-like ER kinase) phosphorylates eIF2 $α$ (eukaryotic initiation factor 2 alpha subunit) on Ser51. This attenuates general translation initiation, reducing the influx of new proteins into the ER. However, ATF4 mRNA is preferentially translated under eIF2 $α$ phosphorylation due to upstream open reading frames in its 5' UTR. ATF4 induces CHOP (C/EBP homologous protein), which upregulates pro-apoptotic genes and downregulates Bcl-2 family anti-apoptotic members. Under sustained stress, CHOP also induces GADD34, a phosphatase subunit that dephosphorylates eIF2 $α$ , creating a negative feedback loop that restores translation and further loads the stressed ER.
ATF6 (activating transcription factor 6) transits to the Golgi upon BiP release. In the Golgi, the proteases S1P and S2P cleave ATF6 in its transmembrane domain, releasing its cytosolic domain (ATF6f) as a transcription factor. ATF6f upregulates chaperones, ERAD components, and XBP1 mRNA (feeding into the IRE1 pathway).

The shift from pro-survival to pro-apoptotic output occurs through several mechanisms: (1) CHOP, induced by the PERK-ATF4 pathway, transcriptionally activates pro-apoptotic BIM and suppresses anti-apoptotic Bcl-2. (2) Prolonged IRE1 activation recruits TRAF2, activating ASK1 and the JNK pathway, which promotes apoptosis. (3) Sustained calcium release from the stressed ER activates calpain proteases and mitochondrial apoptosis pathways. (4) The GADD34 negative feedback loop restores translation under conditions where the ER cannot handle the load, further aggravating stress. The UPR is a timed decision: if homeostasis is not restored within a certain window (hours in most cell types), apoptosis is triggered to eliminate the damaged cell.

Protein folding diseases, the unfolded protein response, and glycosylation disorders Master

Protein folding in the ER is inherently error-prone. An estimated 30% of nascent polypeptides fail to achieve their native conformation on the first attempt and require chaperone-assisted refolding; a subset is ultimately committed to ERAD. In specialised secretory cells — plasma cells secreting immunoglobulins, pancreatic $β$ -cells producing insulin, hepatocytes synthesising albumin and clotting factors — the ER folding burden is enormous. These cells have expanded ER networks and elevated chaperone expression, and they are particularly susceptible to diseases caused by protein misfolding.

Cystic fibrosis and $Δ$ F508-CFTR. The $Δ$ F508 mutation in the cystic fibrosis transmembrane conductance regulator (CFTR) is carried by approximately 1 in 25 individuals of Northern European descent and accounts for roughly 70% of cystic fibrosis alleles worldwide. The mutation deletes phenylalanine at position 508 in the nucleotide-binding domain 1 (NBD1). The deleted residue sits at a critical interdomain interface between NBD1 and the intracellular loop 4 (ICL4) of the second membrane-spanning domain (MSD2). Biophysical studies show that $Δ$ F508 destabilises NBD1 by approximately 2-3 kcal/mol and reduces the efficiency of domain-domain assembly. The mutant protein is recognised by the ER quality-control system through exposure of hydrophobic surfaces at the disrupted interface and is targeted to ERAD via the Hrd1 ubiquitin ligase complex. Less than 1% of $Δ$ F508-CFTR reaches the plasma membrane under physiological conditions.

The therapeutic rescue of $Δ$ F508-CFTR represents a landmark in precision medicine. The small molecule lumacaftor (VX-809) was identified in a high-throughput screen for compounds that increase $Δ$ F508-CFTR maturation and surface expression. Lumacaftor binds to the first membrane-spanning domain of CFTR, stabilising the MSD1-NBD1 interface and improving folding efficiency. Elexacaftor (VX-445) binds to a different site on NBD1, directly addressing the $Δ$ F508 structural defect. The triple combination elexacaftor/tezacaftor/ivacaftor restores approximately 50-60% of CFTR function, transforming the clinical trajectory for most cystic fibrosis patients. This success validates the principle that pharmacological chaperones can override the ER quality-control system for therapeutic benefit, but it also raises questions about the long-term consequences of allowing partially misfolded proteins to escape ER surveillance.

Alpha-1-antitrypsin deficiency. The Z mutation (Glu342Lys) in the SERPINA1 gene encoding alpha-1-antitrypsin (A1AT) is the most common cause of genetic emphysema and liver disease. The Z mutation disrupts a salt bridge in the serpin fold, causing the mutant protein to polymerise within the ER of hepatocytes. These polymers accumulate as Periodic Acid-Schiff-positive inclusion bodies, leading to hepatocyte injury, cirrhosis, and hepatocellular carcinoma. Simultaneously, the lack of secreted A1AT removes a critical anti-protease shield from the lung, allowing neutrophil elastase to destroy alveolar walls and causing panacinar emphysema.

The A1AT polymerisation mechanism is a domain-swapping event: the reactive centre loop (RCL) of one A1AT molecule inserts into the $β$ -sheet A of an adjacent molecule, forming an extended chain of head-to-tail polymers. This polymerisation is temperature-dependent and concentration-dependent, explaining why the ER — a confined space with high protein concentration — is the site of polymer accumulation. The ER quality-control system recognises the polymer as abnormal but cannot efficiently extract and degrade it, leading to accumulation. Therapeutic strategies include: (1) small molecules that block polymerisation by stabilising the native serpin fold (the GSK256066 derivative programme); (2) RNA interference to reduce Z-A1AT synthesis and hepatic accumulation; and (3) augmentation therapy with intravenous wild-type A1AT purified from human plasma to protect the lungs.

The unfolded protein response in detail. The UPR is a homeostatic signalling network that attempts to restore ER proteostasis. Its three sensor pathways have overlapping but distinct outputs:

IRE1/XBP1 pathway. IRE1 is the most evolutionarily conserved UPR sensor (present in all eukaryotes). Its lumenal domain detects unfolded proteins through direct binding to hydrophobic peptide segments and through release of BiP. Upon activation, IRE1 oligomerises into dynamic clusters on the ER membrane. The cytoplasmic kinase domain trans-autophosphorylates, activating the RNase domain. The RNase has two activities: (1) XBP1 splicing: cleavage at two sites in XBP1 mRNA removes a 26-nucleotide intron, and RtcB ligase rejoins the exons, producing XBP1s (a potent transcription factor that upregulates chaperones, ERAD components, and lipid synthesis genes); (2) RIDD (regulated IRE1-dependent decay): under prolonged stress, IRE1 degrades a subset of ER-localised mRNAs and microRNAs, reducing the protein-folding load. The switch from specific XBP1 splicing to promiscuous RIDD activity correlates with the transition from pro-survival to pro-apoptotic signalling. Small molecules that stabilise IRE1 oligomers (enhancing XBP1 splicing without activating RIDD) are being explored as therapeutic UPR modulators.

PERK/ATF4 pathway. PERK is an ER kinase that phosphorylates eIF2 $α$ on Ser51, converting eIF2 from a translation initiator into an inhibitor of eIF2B. This reduces general protein synthesis by approximately 50-70% within 30 minutes of ER stress, rapidly decreasing the influx of new substrates into the ER. However, ATF4 mRNA contains upstream open reading frames (uORFs) that normally suppress its translation; eIF2 $α$ phosphorylation paradoxically enhances ATF4 translation by allowing ribosomes to bypass the inhibitory uORFs. ATF4 induces a transcriptional programme that includes CHOP (pro-apoptotic), GADD34 (a regulatory subunit of PP1 that dephosphorylates eIF2 $α$ , creating negative feedback), and genes involved in amino acid metabolism and redox homeostasis. The PERK pathway is integrated with the broader integrated stress response (ISR), which also activates via eIF2 $α$ phosphorylation in response to amino acid starvation (GCN2), viral infection (PKR), and heme deficiency (HRI).

ATF6 pathway. ATF6 is a type II ER transmembrane protein whose lumenal domain binds BiP under unstressed conditions, retaining ATF6 in the ER. When BiP is titrated to unfolded proteins, ATF6 is packaged into COPII vesicles and transported to the Golgi. In the Golgi, the Site-1 Protease (S1P) and Site-2 Protease (S2P) — the same proteases that process SREBP in cholesterol regulation — cleave ATF6 within its transmembrane domain, releasing ATF6f (the cytosolic fragment) to the nucleus. ATF6f is a bZIP transcription factor that induces chaperones (BiP, GRP94, calreticulin), ERAD components, and XBP1 mRNA. The ATF6 pathway is the first to activate (within 1-2 hours) and is also the first to attenuate, as ATF6f induces expression of BiP that re-sequesters full-length ATF6 in the ER.

Golgi glycan diversity and congenital disorders of glycosylation. The Golgi apparatus synthesises an astonishing diversity of glycan structures. N-linked glycans alone can be classified as high-mannose, hybrid, or complex type, with complex-type glycans exhibiting variable branching (bi-, tri-, and tetra-antennary), variable terminal modifications (sialylation, fucosylation, sulfation), and variable core modifications (core fucose, bisecting GlcNAc). This diversity is generated by the sequential action of glycosyltransferases, each localised to a specific Golgi cisterna. The human genome encodes approximately 200 glycosyltransferases, and the combinatorial use of these enzymes produces an estimated $1 0^{5}$ - $1 0^{6}$ distinct glycan structures on cell-surface glycoproteins.

Congenital disorders of glycosylation (CDG) are a rapidly growing family of genetic diseases caused by defects in glycan assembly or processing. Over 150 CDG types have been identified. PMM2-CDG (phosphomannomutase 2 deficiency) is the most common, with an estimated incidence of 1 in 20,000-50,000. PMM2 converts mannose-6-phosphate to mannose-1-phosphate, a precursor for GDP-mannose and dolichol-linked oligosaccharide assembly. PMM2 deficiency reduces the supply of dolichol-linked oligosaccharides, causing incomplete N-glycan assembly and hypoglycosylation of numerous serum and cellular glycoproteins. Clinical features include developmental delay, ataxia, stroke-like episodes, coagulation abnormalities, and failure to thrive — reflecting the ubiquitous role of N-glycans in protein folding, stability, and function.

COPII vesicle reconstitution and the discovery of minimal budding machinery. The COPII budding machinery was reconstituted in vitro by Salama, Schekman, and colleagues using defined components and synthetic liposomes. This landmark experiment demonstrated that only five purified proteins — Sar1, Sec23/Sec24, and Sec13/Sec31 — are sufficient to generate coated vesicles from artificial membranes, establishing COPII as a minimal self-assembly system. Subsequent structural work by Stagg, Noble, and colleagues solved the COPII cage structure by cryo-EM, revealing that Sec13/Sec31 heterotetramers form a cuboctahedral lattice with octahedral symmetry, analogous to clathrin cages but with distinct geometry. The edge length of the COPII cage determines vesicle size at approximately 60-70 nm for the basic lattice.

The discovery that procollagen and other supramolecular cargoes (chylomicrons, lipoprotein particles) are too large to fit into standard COPII vesicles prompted the identification of the TANGO1/cTAGE5 adaptor system. TANGO1 (Transport and Golgi organisation 1) is a transmembrane protein at ERES that binds procollagen in the ER lumen via its SH3-like domain and recruits the COPII inner coat via its cytoplasmic domain. TANGO1 also interacts with Sec16 (a large scaffold at ERES) and with the ER-Golgi tethering factor CTAGE5. The current model is that TANGO1 acts as a linker that holds procollagen at the ERES while the COPII coat assembles an oversized carrier, potentially through a different lattice geometry or through a COPII-coated tubular extension rather than a spherical vesicle. This remains an active area of investigation.

Connections Master

Cellular organisation: organelles 17.03.01. The ER and Golgi were introduced as organelles of the endomembrane system in the organelles overview unit. This unit deepens that introduction into the mechanistic details of protein folding, quality control, glycosylation, and vesicle budding that occur within these compartments.
Vesicle trafficking 17.02.04 pending. COPII vesicle budding from the ER, COPI retrograde transport, and SNARE-mediated vesicle fusion are covered in the vesicle trafficking unit. This unit provides the biological context — ER exit sites, the calnexin cycle, glycan processing — in which those vesicle budding and fusion events occur. The two units together give a complete picture of the secretory pathway from protein synthesis to final destination.
Membrane transport: passive and active 17.02.02. The SERCA calcium pump on the ER membrane (which maintains the low cytosolic calcium concentration and high ER calcium store) and the V-ATPase in Golgi-derived vesicles both use active transport mechanisms introduced in the membrane transport unit. The pH gradient between the Golgi (pH 6.0-6.5) and the ER (pH 7.2) that drives KDEL receptor-ligand dissociation is maintained by these pumps.
Protein structure 17.01.02 pending. The principles of protein folding (hydrophobic collapse, disulfide bond formation, chaperone-assisted folding) introduced in the protein structure unit are directly applied here to the ER folding environment. The calnexin cycle and ERAD pathway are the cellular machinery that enforces the thermodynamic requirement for native conformation before export.
Cell signalling 17.07.01. The unfolded protein response sensors (IRE1, PERK, ATF6) are themselves transmembrane signalling molecules that transmit information from the ER lumen to the nucleus and cytoplasm, analogous to receptor tyrosine kinases and GPCRs at the plasma membrane. The IRE1 RNase activity and PERK kinase activity represent non-canonical signalling mechanisms.
Cellular neuroscience 17.09.01. The ER is the primary intracellular calcium store, and ER calcium release through IP3 receptors and ryanodine receptors triggers downstream signalling cascades including muscle contraction and synaptic plasticity. ER stress and dysregulated calcium release are implicated in excitotoxic neuronal death.

Historical notes Master

The endoplasmic reticulum was first observed by Keith Porter in 1945 using electron microscopy of cultured fibroblasts, where it appeared as a "lace-like" network in the cytoplasm. The term "endoplasmic reticulum" was coined by Porter and Thompson in 1945-1948. George Palade's systematic electron microscopy studies in the 1950s and 1960s established the distinction between rough and smooth ER and demonstrated that the ER is the entry point for the secretory pathway, work recognised by the 1974 Nobel Prize.

The signal hypothesis, proposed by Blobel and Sabatini in 1971 and experimentally confirmed by Blobel and Dobberstein in 1975 (J. Cell Biol. 67, 835-851), explained how proteins are targeted to the ER. Blobel's in vitro translation system with added microsomal membranes (isolated ER vesicles) demonstrated that the N-terminal signal sequence is both necessary and sufficient for ER translocation. This work received the 1999 Nobel Prize.

N-linked glycosylation was discovered through the study of viral glycoproteins. Robbins and colleagues demonstrated in 1964 that infection with vesicular stomatitis virus (VSV) produced a glycoprotein (VSV-G) whose glycan was assembled from dolichol-linked intermediates. The dolichol pathway was mapped by Leloir, Parodi, and Hirschberg in the 1970s-1980s, establishing the stepwise assembly of the 14-sugar oligosaccharide on the lipid carrier. The calnexin cycle was elucidated in the 1990s by Helenius, Aebi, and colleagues, who showed that monoglucosylated glycoproteins are specifically recognised by ER lectin chaperones.

The Golgi apparatus was discovered by Camillo Golgi in 1898 using a silver staining method on Purkinje cells. For decades, some researchers doubted its existence as a distinct organelle, viewing it as an artefact of the staining procedure. The development of electron microscopy in the 1950s definitively established the Golgi as a real organelle. The cisternal maturation model was first proposed by Bainton in 1966 and refined by Glick, Lupashin, and others in the 1990s-2000s using live-cell imaging of fluorescently tagged Golgi enzymes in yeast.

COPII vesicle budding was reconstituted in vitro by Barlowe, Schekman, and colleagues in 1994 (Cell 77, 895-907), demonstrating that five purified cytosolic components (Sar1, Sec23, Sec24, Sec13, Sec31) are sufficient to generate coated vesicles from ER membranes. This minimalist reconstitution was a landmark in membrane biology, showing that coat assembly is a self-organising process driven by a small GTPase cycle and protein-protein interactions. The structural basis of the COPII cage was determined by cryo-EM (Stagg et al. 2006 J. Mol. Biol. 363, 479-491; Noble et al. 2013 Cell 154, 1130-1141).

The unfolded protein response was discovered independently by Peter Walter (IRE1) and David Ron (PERK) in the late 1990s and early 2000s. Walter's group identified IRE1 as an ER transmembrane kinase/RNase in yeast (Cox et al. 1993 Cell 73, 1197-1206) and showed that it splices HAC1 mRNA (the yeast orthologue of XBP1). Ron's group identified PERK as the kinase that phosphorylates eIF2 $α$ during ER stress (Harding et al. 1999 Nature 397, 271-274). Kazutoshi Mori and Peter Walter shared the 2014 Albert Lasker Basic Medical Research Award for the discovery of the UPR.

Congenital disorders of glycosylation were first described by Jaak Jaeken in 1980, who identified a group of children with unexplained neurological symptoms and abnormal serum transferrin glycosylation patterns. The molecular basis of PMM2-CDG (the most common CDG) was identified in 1997 by Matthijs and colleagues (Nat. Genet. 16, 88-92). The field has since expanded to include over 150 distinct CDG types, each caused by a defect in a different step of glycan assembly, transfer, or processing. CDG research has been instrumental in demonstrating the in vivo functional importance of glycosylation, which had previously been underappreciated because glycosylation defects were often lethal or severely disabling.

Bibliography Master

Alberts, B. et al. Molecular Biology of the Cell, 7th ed. Garland Science, 2022. Ch. 13 (Intracellular Membrane Traffic) and Ch. 14 (Intracellular Compartments).
Lodish, H. et al. Molecular Cell Biology, 9th ed. W. H. Freeman, 2021. Ch. 13 (The Endoplasmic Reticulum) and Ch. 14 (Vesicular Traffic).
Helenius, A. and Aebi, M. "Intracellular protein processing." Curr. Opin. Cell Biol. 15 (2003): 431-439.
Barlowe, C. "COPII and the regulation of protein trafficking." Annu. Rev. Cell Dev. Biol. 17 (2001): 1-24.
Blobel, G. and Dobberstein, B. "Transfer of proteins across membranes. I. Presence of proteolytically processed and unprocessed nascent immunoglobulin light chains on membrane-bound ribosomes of murine myeloma." J. Cell Biol. 67 (1975): 835-851.
Barlowe, C. et al. "COPII: a membrane coat formed by Sec proteins that drive vesicle budding from the endoplasmic reticulum." Cell 77 (1994): 895-907.
Cox, J. S., Shamu, C. E., and Walter, P. "Transcriptional induction of genes encoding endoplasmic reticulum resident proteins requires a transmembrane protein kinase." Cell 73 (1993): 1197-1206.
Harding, H. P., Zhang, Y., and Ron, D. "Protein translation and folding are coupled by an endoplasmic-reticulum-resident kinase." Nature 397 (1999): 271-274.
Losev, E. et al. "Golgi maturation visualized in living yeast." Nature 441 (2006): 1002-1006.
Matsuura-Tokita, K. et al. "Live imaging of yeast Golgi cisternal maturation." Nature 441 (2006): 1007-1010.
Matthijs, G. et al. "Mutations in PMM2, a phosphomannomutase gene on chromosome 16p13, in carbohydrate-deficient glycoprotein syndrome type I (Jaeken syndrome)." Nat. Genet. 16 (1997): 88-92.
Noble, A. J. et al. "A pseudoatomic model of the COPII cage obtained from cryo-electron microscopy and mass spectrometry." Cell 154 (2013): 1130-1141.
Ward, C. L. and Kopito, R. R. "Intracellular processing of mutant deltaF508 CFTR results in partial degradation." J. Biol. Chem. 269 (1994): 25710-25718.
Loo, T. W. and Clarke, D. M. "Correctors promote folding of the CFTR mutant deltaF508." Biochemistry 45 (2006): 8140-8148.
Hebert, D. N. and Molinari, M. "In and out of the ER: protein folding, quality control, degradation, and related human diseases." Physiol. Rev. 87 (2007): 1377-1408.
Schekman, R. "Discovery of the cellular vesicular trafficking system: a historical perspective." Proc. Am. Philos. Soc. 157 (2013): 117-123.
Walter, P. and Ron, D. "The unfolded protein response: from stress pathway to homeostatic regulation." Science 334 (2011): 1081-1086.

Prerequisites

17.03.01 pending
17.02.04 pending

Tier anchors

beginner: Alberts et al., Molecular Biology of the Cell, 7th ed. (2022), Ch. 13-14
intermediate: Lodish et al., Molecular Cell Biology, 9th ed. (2021), Ch. 13-14
master: Helenius, A. et al. — Curr. Opin. Cell Biol. 15 (2003) 431-439; Barlowe, C. — Annu. Rev. Cell Dev. Biol. 17 (2001) 1-24

References

Alberts et al. — Molecular Biology of the Cell, 7th ed. (Garland Science, 2022) · Ch. 13 Intracellular Membrane Traffic; Ch. 14 Intracellular Compartments
Lodish et al. — Molecular Cell Biology, 9th ed. (W. H. Freeman, 2021) · Ch. 13 The Endoplasmic Reticulum; Ch. 14 Vesicular Traffic
Helenius, A. & Aebi, M. — Intracellular protein processing · Curr. Opin. Cell Biol. 15 (2003) 431-439
Barlowe, C. — COPII and the regulation of protein trafficking · Annu. Rev. Cell Dev. Biol. 17 (2001) 1-24

Estimated time

beginner: 15m
intermediate: 30m
master: 45m