The hIRE1α cLD forms a stable dimer validated by hydrogen–deuterium exchange experimental data.

(A) The PDB model: structural model of hIRE1α cLD dimer as obtained by Zhou et al., 2006 (PDB ID 2HZ6) with the addition of missing residues, in particular DR1 (cyan) and DR2 (orange). Backbone in cartoon representation, areas of interest highlighted in different colors. (B) The AF model: AlphaFold 2 Multimer prediction of hIRE1α cLD dimer (P29–P368). Representation style as in (A). (C) The deuterated fraction obtained from experimental results (dashed line, shaded area indicates the error we calculated from bootstrapping) published by Amin-Wetzel et al., 2019 and the fraction computed from molecular dynamic (MD) simulations (solid lines, blue for TIP3P water and orange for TIP4P-D water) for the PDB and AF model at incubation time point 0.5 min. This time point corresponds to experimental incubation times, not MD simulation time. Each point represents the mean value derived from three replicas and two monomers per replica. The error bars were obtained from bootstrapping. Below each absolute value plot, we report the discrepancy, which is defined as the difference between the simulated and experimental deuterated fractions, with the shaded area indicating the corresponding error. (D) Visualization of the discrepancy values from (C) onto the molecular structures of the PDB and AF model simulated in TIP4P-D water (TIP3P shown in Figure 1—figure supplement 4A). Shades of blue indicate regions where the simulated structure is more flexible and solvent-accessible than observed experimentally, while shades of red indicate regions where the simulated structure is more rigid and less solvent-accessible than expected.

The human IRE1α cLD dimer has a narrower and more occluded central groove than its yeast homolog.

(A) Top panel: structure of human IRE1α (hIRE1α) cLD dimer (AlphaFold model). Bottom panel: structure of the S. cerevisiae IRE1 (yIre1) cLD dimer (PDB ID 2BE1). (B) Close-up on hIRE1α (top panel) and yIre1 (bottom panel) structures on the central groove delimited by the αA-helices. Helices delimiting the groove are depicted in blue cartoon and side chains in stick representation, while the rest of the dimer is transparent. The arrows indicate the distance between Cα atoms used to compute the groove width. (C) Probability density distribution of the groove width of: hIRE1α cLD dimer computed between Cα of Q105 during simulations in TIP3P and TIP4P-D water (top panel); yIre1 luminal domain dimer computed between Cα of S200 during simulations in TIP3P and TIP4P-D water (bottom panel). The different shades of colors in the distributions represent the three system replicas for each water model. Dashed lines indicate the groove width of the dimer model prior to energy minimization.

Unfolded peptides can bind with specificity to the center of the cLD dimer, even if the groove is closed.

(A) AF model of the cLD dimer (gray cartoon) simulated with the unfolded peptide MPZ1-N (violet) positioned parallel to the principal axis of the central groove (t = 0 µs). During the simulation, MPZ1-N rearranges to a vertical position (t = 1 µs). The direction of the principal axis of the central groove is represented by a solid arrow, while the direction of the principal axis of the peptide is indicated by a dashed arrow. (B) AF model of the cLD dimer simulated with unfolded peptide MPZ1N-2X (orange), which stably binds to the center of the dimer. (C) AF model of the cLD dimer simulated with unfolded peptide MPZ1N-2X-RD (red), which unbinds in half of the replicas. (D) Box plot distributions of the minimum groove-peptide distance. Each box plot reports on the results from three replicas (in TIP3P or in TIP4P-D) of a system containing a dimer and one of the unfolded peptides.

Single-point mutations of cLD weaken the binding of unfolded peptide MPZ1N-2X.

(A) Analysis of contacts between unfolded peptide and cLD dimer for all molecular dynamic (MD) simulations performed in TIP4P-D water. The contact score reported on the AlphaFold model of the cLD dimer was obtained by computing the contacts from all the simulations, summing them, and normalizing them by the number of frames. Lower values indicate fewer contacts observed. (B) Important areas for peptide binding on the surface of the cLD dimer: the red area identifies a hydrophobic–aromatic binding pocket characterized by residues Y117, F98, and L103; the yellow area identifies a deeper polar-charged binding pocket and is characterized by residues Y161, D181, and D79. Residue E102 on the αA-helices is a hub for contacts on the surface of the dimer. (C) Side view snapshot after 1 µs of simulation of wild-type (WT) hIRE1α cLD dimer (gray) in complex with MPZ1N-2X (orange). The amino acids E102 and Y161 on both monomers are represented in green sticks. (D) Side view snapshot after 1 µs of simulation of Y161R hIRE1α cLD dimer (gray) in complex with MPZ1N-2X (orange). The amino acid R161 on both monomers is represented in blue stick. (E) Side view snapshot after 1 µs of simulation of E102R hIRE1α cLD dimer (gray) in complex with MPZ1N-2X (orange). The amino acid R102 on both monomers is represented in magenta sticks. (F) Time series of the minimum groove-peptide distance for MPZ1N-2X simulated in complex with WT, E102R, and Y161R hIRE1α cLD dimer in TIP3P (three replicas) and TIP4P-D (three replicas) water. The darker lines show the rolling average over 25 frames, while the shaded lines represent the raw data. (G) Fluorescence anisotropy measurements of labeled MPZ1N-2X binding to hIRE1α LD WT and mutants E102R and Y161R.

hIRE1α cLD intermolecular interactions.

(A) Model of two cLD dimers interacting via an unfolded peptide (MPZ1N-2X): the system setup is shown (t = 0 ns) and a snapshot of the simulation (t = 200 ns). (B) BiP–cLD monomer complex as predicted by AlphaFold (BiP in shades of purple, cLD in orange) before the simulation (t = 0 μs) and at the end of the simulation (t = 1 μs). The substrate-binding domain (SBD) (residues E19–D408) is colored in light purple, and the NDB (residues C420–E650) in dark purple, and the interdomain linker (residues D409–V419) and KDEL motif (residues K651–L654) in light purple.

Schematic illustrating the proposed sequence of events occurring in the early stages of hIRE1α activation within the ER lumen.

Before the onset of ER stress conditions, hIRE1α is prevented from forming assemblies by the chaperone BiP, which occupies the cLD’s oligomerization interface. As ER stress begins, unfolded proteins start to accumulate in the ER lumen, and pre-formed hIRE1α cLD dimers are released from BiP. Now, hIRE1α can interact with unfolded proteins, which brings multiple copies of the protein together to allow the formation of larger assemblies.

Detailed comparison of the structural models used in this study.

(A) Superposition of AF model (light violet) and PDB model (orange) of the hIRE1α cLD dimer before the simulation (RMSD = 3.34 Å). (B) Superposition of cLD dimer models predicted by AlphaFold 2 Multimer (light violet) and AlphaFold 3 (dark violet). (C) Superposition of the five structures predicted by AlphaFold 2 Multimer for the cLD dimer and colored by confidence prediction score (pLDDT). (D) Superposition of the five structures predicted by AlphaFold 3 for the cLD dimer and colored by confidence prediction score (pLDDT). (E) PDB model of the hIRE1α cLD dimer after 2 µs of production run from one of the TIP3P replicas. (F) AF model of the hIRE1α cLD dimer after 2 µs of production run from one of the TIP3P replicas.

Characterizing the interactions that stabilize the dimers.

(A) Details of the dimeric interface of the hIREα cLD dimer, consisting of an antiparallel β-sheet with four hydrogen bonds between residues G119–D123. The renders show the dimer interface in the AF model (upper row) and of the PDB model (lower row) prior to energy minimization and production run (t = 0 µs) and after 2 µs of simulation. The backbone atoms of residues G119–D123 are in sticks (C cyan, H white, O red, N blue), and the rest of the structure is rendered as a transparent ribbon. The backbone of the protein is shown as a transparent cartoon. (B) Hydrogen bond analysis of all the hIREα cLD dimer simulation replicas. This analysis shows the distance between donor and acceptor atoms (left y-axis, blue line) and the angle between donor–hydrogen–acceptor (right y-axis, orange line) for all replicas of PDB and AF models in TIP3P and TIP4P-D water. The hydrogen bonds analyzed here are those formed at the dimeric interface by residues G119, K121, and N123. The horizontal blue and orange lines indicate, respectively, the ideal distance (3.2 Å) and angle (180°) for a hydrogen bond. Whenever the distance is greater than 3.2 Å or the angle is lower than 120°, the frame is highlighted by a gray or red vertical bar, respectively. Therefore, the vertical bars indicate when the hydrogen bond is lost. The data are plotted every 50th frame for the lines and every frame for the vertical bars.

Key residues controlling dimer stability in the dimeric interface.

(A) Initial PDB model of the hIRE1α cLD dimer after 2 µs of production run. Replica showing a partial dissociation of the dimer. (B) Detail of the side of the dimeric interface of the initial PDB model that dissociates during simulations. The side chains of residues L116, Y166, and W125 are highlighted in three different models: the initial PDB model (gray), the final PDB model (orange), and the AlphaFold model (light violet). The backbone of the protein is shown as transparent cartoon. (C) The DR1 is folded in the AF model (Figure 1B). The modified AF model with unfolded DR1 (left, t = 0 µs). The AlphaFold model with unfolded DR1 is stable after 500 ns of simulation (right, t = 500 ns).

Detailed validation of different structure and water models by hydrogen–deuterium exchange experimental data.

(A) Visualization of the discrepancy values from Figure 1C for TIP3P and TIP4P-D water at incubation times of 0.5 and 5 min, mapped onto the molecular structures of the PDB and AF models. Time points correspond to experimental incubation times, not molecular dynamic (MD) simulation time. Shades of blue indicate regions where the simulated structure is more flexible and solvent-accessible than observed experimentally, while shades of red indicate regions where the simulated structure is more rigid and less solvent-accessible than expected. (B) Comparison of the deuterated fraction obtained from experimental results (dashed line) published by Amin-Wetzel et al., 2019 and the fraction computed from MD simulations (solid lines, blue for TIP3P water and orange for TIP4P-D water, error bars indicate the standard deviation obtained from bootstrapping) for the PDB and AF model at incubation time point 5 min. Each point represents the mean value derived from three replicas and two monomers per replica. The error bars were obtained from bootstrapping. Below each absolute value plot, we report the discrepancy, which is defined as the difference between the simulated and experimental deuterated fractions.

Only a combination of different models fits the hydrogen–deuterium exchange experimental data accurately.

(A) Comparison of the deuterated fraction from Amin-Wetzel et al., 2019 with the deuterated fraction computed from molecular dynamic (MD) simulations (see Methods section ‘HDX data comparison’). For each residue, the plot identifies the model that best agrees with experimental data (‘best structure’) at each time examined point (t = 0.5, 5 min, or both). The histogram displays the cumulative number of residues for which each model most closely matches experimental data. Each of the four model variants is evaluated residue-wise at both time points, earning a point if it better reproduces the experimental deuterated fraction for a given residue compared to the other three models. The total score, obtained by summing points for each model at both time points, is represented in the histogram. (B) Best structural model of hIRE1α cLD dimer: colors indicate the best structural representation of each residue at either time point, as shown in (A).

Dynamics of the central groove opening.

(A) Probability density distribution of the groove width of: hIRE1α cLD dimer (AlphaFold model) computed between Cα of Q105 during simulations in TIP3P and TIP4P-D water (top panel); hIRE1α cLD dimer (PDB model) computed between Cα of Q105 during simulations in TIP3P and TIP4P-D water (middle panel); yIre1 luminal domain dimer computed between Cα of S200 during simulations in TIP3P and TIP4P-D water (bottom panel). Dashed lines indicate the groove width of the dimer model prior to energy minimization. (B) Representative Q105 side chains arrangements during simulations of the human cLD dimer AlphaFold model (top row) and of the PDB model (bottom row), snapshots are taken from three independent replicas in TIP3P and TIP4P-D.

Additional examples of peptide binding.

(A) Here, we report snapshots from simulations at the initial time step (t = 0 µs) and final time step (t = 1 µs) of the unfolded peptides not shown in Figure 3A-C, in order: Valine8, MPZ1-N (perpendicular placement), MPZ1-C, 8ab1, OR-1, V1rb2-1, V1rb2-2. (B) Heterogeneous binding. Peptides that are unstable in the central groove placement tend to rearrange to different parts of IRE1, either to the side of the groove as for V1rb2-1 and MPZ1-C or to the outer part of the dimer in the proximity of the αB-helices, as in the case of MPZ1N-2X-RD.

Peptide orientation with respect to the central groove principal axis.

(A) The angle was computed as the dihedral angle described by the Cα atoms of Y161 residues (groove principal axis) and the Cα atoms of residues L1 and A12 of the MPZ1N peptide. The dark lines indicate the rolling average of the fraction of native contacts over 10 frames, while the shaded lines indicate the value per frame. (B) Number of contacts between hIRE1α cLD dimer and MPZ1N peptide. The dark lines indicate the rolling average of the fraction of native contacts over 50 frames, while the shaded lines indicate the value per frame. The analysis was performed on three sets of simulations: ‘90°’ orientation, the peptide is initially placed perpendicular to the central groove principal axis; ‘270°’ orientation, the peptide is initially placed perpendicular to the central groove principal axis but flipped 180° with respect to the 0°; ‘0°’ orientation, the peptide is placed parallel to the groove principal axis.

Prediction of AlphaFold 3 for hIRE1α cLD dimer in complex with peptides.

(A) Distributions of the groove width of peptide-bound cLD dimers throughout all simulations performed. The left column shows the values for the three replicas in TIP3P water, while the right column displays those for the three replicas in TIP4P-D water. (B) Minimum groove-peptide distance over time for all simulations of cLD dimer in complex with a peptide. The left column shows the values for the three replicas in TIP3P water, while the right column displays those for the three replicas in TIP4P-D water.

Prediction of AlphaFold 3 for hIRE1α cLD dimer in complex with peptides.

(A) Colors represent the confidence of the prediction (plDDT). (B) Difference in enthalpy (enthalpy of binding, ΔH) as an estimate of the binding free energies of unfolded polypeptides to hIRE1α cLD dimer derived from molecular mechanics/Poisson–Boltzmann surface area (MM/PBSA) calculations of our peptide simulations.

Analysis of contacts between unfolded peptide and cLD dimer for all molecular dynamic (MD) simulations performed in TIP3P water.

(A) The contact score reported on the AlphaFold model of the cLD dimer was obtained by computing the contacts from all the simulations, summing them, and normalizing them by the number of frames. Lower values indicate fewer contacts observed. (B) Contact analysis of MPZ1N-2X in complex with wild-type hIRE1α cLD dimer. (C) Contact analysis of MPZ1N-2X in complex with E102R hIRE1α cLD dimer. (D) Contact analysis of MPZ1N-2X in complex with Y161R hIRE1α cLD dimer. The analysis of panels B–D took into consideration all six 1-μs-long replicas in TIP3P and TIP4P-D water for each system.

Peptide contacts and purification of wild-type hIRE1α luminal domain and its E102R and Y161R mutants.

(A) Number of contacts between residues 102 on both monomers and the MPZ1-N-2X peptide during simulations of wild-type (WT) hIREα LD and mutants E102R and Y161R. The dark lines indicate the rolling average of the fraction of native contacts over 25 frames, while the shaded lines indicate the value per frame. (B) Number of contacts between residues 161 on both monomers and the MPZ1-N-2X peptide during simulations of WT hIREα LD and mutants E102R and Y161R. The dark lines indicate the rolling average of the fraction of native contacts over 25 frames, while the shaded lines indicate the value per frame. (C) Protein purification of WT hIREα LD and mutants E102R and Y161R.

Prediction of AlphaFold 3 for hIRE1α cLD monomer in complex with ATP-bound BiP.

(A) The colors are as in Figure 5B. (B) Prediction of AlphaFold 3 for hIRE1α cLD monomer in complex with ADP-bound BiP. (C) Prediction of AlphaFold 3 for hIRE1α cLD monomer in complex with BiP not bound to any nucleotide. (D) Structure of hIRE1α cLD-BiP-ATP after 2 µs of simulation. (E) Structure of hIRE1α cLD-BiP-ADP after 2 µs of simulation. (F) Structure of hIRE1α cLD-BiP after 2 µs of simulation.

BiP–cLD contacts in apo, ADP-, and ATP-bound states. Then A, followed by current text.

Fraction of native contacts between BiP and cLD monomer in simulations of the structures predicted by AlphaFold 3 without ligands or in complex with ADP or ATP. The dark lines indicate the rolling average of the fraction of native contacts over 100 frames, while the shaded lines indicate the value per frame. The fraction of native contacts (Q) was calculated according to the definition of Best et al., 2013: . For N pairs of native contacts, (i,j) where is the distance of the pair in the initial configuration (here the AlphaFold 3 prediction), is the distance at frame X, β is a smoothing parameter (β = 50 nm−1), λ is the tolerance of the reference distance (λ = 1.8) and the cutoff used to define a contact between heavy atoms was 0.45 nm.