Endoplasmic Reticulum

Featured Lecture — Tom Rapoport (1/3)

The first Rapoport iBiology lecture: Organelle Biosynthesis and Protein Sorting. This is the canonical treatment of how proteins get to the ER in the first place — signal peptides, SRP recognition, the Sec61 translocon, and how membrane-protein topology is set at the moment of insertion. Essential viewing for this module and the entire course.

1. Cotranslational Targeting & the Sec61 Channel

Newly synthesised secretory or membrane proteins emerge from the ribosome bearing an N-terminal signal peptide: a short hydrophobic stretch recognised by the signal recognition particle (SRP). SRP binds the signal, halts elongation, and delivers the ribosome–nascent-chain complex to the ER membrane via the SRP receptor. Elongation resumes, threading the nascent chain through the Sec61 translocon, a heterotrimeric channel with a conserved “plug” that seals the lumen until opened by an incoming peptide (Günter Blobel, 1999 Nobel; Rapoport structural biology 2000s).

Membrane proteins are inserted laterally through a lateral gate in Sec61: when a hydrophobic transmembrane segment enters the translocon, it partitions sideways into the lipid bilayer rather than continuing into the lumen. Topology (N-in / N-out) is set by the charge distribution flanking the first TM segment — the positive-inside rule(von Heijne). An entire membrane-protein topology can be predicted from sequence alone with high accuracy.

Post-translational insertion — used for tail-anchored proteins and some small proteins — proceeds via distinct machinery: the GET pathway (yeast) / TRC40 (mammalian), and the more recently characterised EMC (ER membrane protein complex, Ng / Hegde 2018) for polytopic substrates with moderately hydrophobic TM segments.

2. Folding as Search in a Funnelled Landscape

Proteins enter the ER unfolded. The cost of searching a generic conformational space would violate Levinthal’s bound — the famous observation that an unbiased random search of a 100-residue chain would take longer than the age of the universe. Folding in the ER is tractable because the landscape is a funnel: most stochastic trajectories are downhill in free energy, and the chaperone network functions as an iterative error-correcting code.

The first-passage time distribution is well-approximated by

\[ P_{\mathrm{fold}}(t) \approx 1 - \exp(-t/\tau_{\mathrm{fold}}), \quad \tau_{\mathrm{fold}} \sim \tau_0 \exp(\beta \Delta G^{\ddagger}) \]

with ΔG^‡ the barrier to the native state and τ₀ ~ μs the prefactor. For most secretory proteins the folded-state yield on first try is 60–80%; everything else becomes substrate for ERAD, the ER-associated degradation pathway, in which misfolded species are recognised, retrotranslocated through Hrd1/Sel1L, ubiquitinated, and digested by the cytoplasmic proteasome.

Simulation: The Folding Funnel

3. The ER Chaperone Network

The ER lumen runs three partially-redundant chaperone circuits:

• BiP / Hsp70 cycle: ATP-dependent binding and release of exposed hydrophobic patches. BiP also serves as the luminal load sensor that inhibits the UPR transducers IRE1, PERK, ATF6 at low demand.
• Calnexin / calreticulin cycle: lectin-based, recognising the single-glucose intermediate Glc₁Man₉GlcNAc₂on N-glycans. A substrate is repeatedly trimmed and re-glucosylated (by UGGT) until it either folds or is routed to ERAD. This is effectively a timer: how many cycles a glycoprotein has attempted is encoded in its glycan state.
• Protein disulphide isomerases (PDI family): oxidative folding of disulphide bonds, using Ero1 as the terminal electron acceptor and molecular oxygen as the ultimate sink. Each disulphide formed releases one H₂O₂ — a small but non-zero ROS tax on secretion.

4. The Unfolded Protein Response as a Control Loop

When the misfolded load exceeds chaperone capacity, the ER activates the unfolded protein response (UPR): three transmembrane sensors — IRE1, PERK, ATF6 — each couple luminal protein load to transcriptional and translational output.

• IRE1: oligomerises and splices the XBP1 mRNA, producing an active transcription factor that upregulates chaperones and ERAD machinery.
• PERK: phosphorylates eIF2α, globally attenuating translation while selectively promoting ATF4. Sustained PERK activity triggers CHOP-mediated apoptosis.
• ATF6: upon stress, translocates to the Golgi where it is cleaved by S1P/S2P proteases, releasing its cytosolic transcription factor domain.

The system is a canonical proportional controller with threshold, described by a minimal two-variable ODE:

\[ \dot U = k_{\mathrm{syn}} - k_{\mathrm{fold}} U \dfrac{C}{C+K} - k_{\mathrm{deg}} U, \quad \dot C = g(U) - \gamma C \]

where U is unfolded load, C is chaperone concentration,g(U) is the sigmoidal UPR response function, and γ is chaperone turnover. Under chronic stress the system loses stability and transitions — via sustained PERK signalling and CHOP induction — to programmed cell death. Much of neurodegenerative biology can be framed as a failure mode of this controller.

Simulation: UPR Dynamics Under Stress

5. ER Morphology: Sheets, Tubules, Three-Way Junctions

The ER occupies distinct morphological sub-domains: flat cisternae (ribosome-studded, 50–100 nm thick, rich in calnexin and translocons) and narrow tubules (~30–50 nm diameter, sparse in ribosomes, enriched in reticulons and DP1/REEP/Yop1) connected at three-way junctions. The high curvature of tubules requires specialised proteins: reticulons and DP1 form hairpin insertions in the outer leaflet, stabilising the tubular geometry against the bending penalty of Module 1.

Atlastin (a membrane-anchored GTPase) catalyses homotypic fusion of tubules to form three-way junctions, producing the characteristic ER polygonal network. Loss-of-function of atlastin causes hereditary spastic paraplegia — one of many diseases traceable to defects in ER architecture (Blackstone 2018). Rapoport’s lecture in Module 1 is the field-defining treatment of this geometry.

6. The ER as a Calcium Store

The ER lumen stores Ca²⁺ at ~0.5–1 mM, a 10⁴-fold gradient over cytosolic (~100 nM). Release via IP₃R and RyR channels initiates second-messenger signalling; refilling is via SERCA pumps consuming one ATP per two Ca²⁺. Luminal Ca²⁺ is buffered by calreticulin and GRP94; both chaperones are calcium-sensitive, so loss of luminal Ca²⁺ directly compromises folding capacity. This is why prolonged ER Ca²⁺ depletion (thapsigargin) is a potent UPR inducer in experimental systems.