Methods for identifying intermediates

Abstract

A method for identifying a target protein folding intermediate is described herein, and the intermediate is suitable for testing as a target in a drug development process. The method is implemented by computer calculation. The method provides a step of modeling the time series of events that define a protein folding pathway, and the step includes modeling and / or calculating the structure and / or energy and / or physicochemical properties of one or more protein folding intermediates along the folding pathway. Then, the method includes the following steps: based on the identification characteristics, identifying at least one candidate protein folding intermediate along the simulated folding pathway, and based on the screening characteristics, selecting one or more target protein folding intermediates in the at least one candidate protein folding intermediate. The screening characteristics are related to the pharmaceutical usability of the protein folding intermediate. The present disclosure also includes related methods for computer drug development based on targeted folding intermediates.

Description

Technical Field

[0001] This article describes a method for identifying target protein folding intermediates that are suitable for testing as targets in the drug development process. The method is implemented by computer calculation. Background Art

[0002] How proteins fold, why they fold in this way, and how the folding pathway of each protein is encoded in its sequence and structure are important questions for protein structure and design, folding and misfolding, regulation and function, clinical issues, and industrial applications.

[0003] B. Nolting 1999 (Protein Folding Dynamics: A Biophysical Approach, Springer, Berlin); WA Eaton et al., 2000 (Annu. Rev. Biophys. Biomol. Struct. 29, 327); V. Daggett and A. Fersht 2003 (Nat. Rev. Mol. Cell Biol. 4, 497) describe attempts to understand protein folding dynamics. Identifying new therapeutic targets is a long-standing need.

[0004] Here, for the first time, a new computational approach is demonstrated that enables the identification of relevant pathways in the protein folding process, showing its use in identifying folding intermediates of druggable target proteins, opening new avenues for drug discovery of previously uninvolved pharmacological targets. Summary of the invention

[0005] The objects of the invention are defined by the appended claims.

[0006] Specifically, the present invention relates to a method for identifying target protein folding intermediates, which are suitable for testing as targets in the drug development process, and to a computer drug development method based on targeting folding intermediates.

[0007] In particular, the invention allows the identification of folding intermediates that are stabilized once bound to a selected ligand, which may allow, for example, the stable folding intermediates to be removed from the cell, thereby inhibiting their activity.

[0008] In another embodiment, the invention allows for the identification of intermediates that are stabilized and potentially activated once bound to a selected ligand. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1: Frequency distribution histogram of conformations of the human PrPC folding pathway, calculated using the bias function method (BF). Trajectories were projected on two ensemble variables described in this paper: the fraction of native contacts Q and the root mean square deviation (RMSD) from the native structure (PDB code: 1QLZ).

[0010] Figure 2 : Three major clusters (called clusters C1, C2, and C3) representing the most visited structures in PrPC folding intermediates.

[0011] Figure 3 : Potentially druggable regions in the folding intermediate of prion protein PrP (FI-PrP). DETAILED DESCRIPTION

[0012] Described herein is a method for identifying target protein folding intermediates that are suitable for testing as targets in the drug discovery process.

[0013] The method comprises the following steps, which are implemented by computer calculation.

[0014] The method provides a step of modeling a temporal sequence of events defining a protein folding pathway, the step comprising modeling and / or calculating the structural and / or energetic and / or physicochemical properties of one or more protein folding intermediates along the folding pathway. The modeling step may comprise, for example, sequentially modeling the protein folding intermediates along the folding pathway.

[0015] The method then includes the steps of identifying at least one candidate protein folding intermediate along the simulated folding pathway based on identification properties (including the structure and energy and / or physicochemical properties); and selecting one or more target protein folding intermediates among the at least one candidate protein folding intermediate based on screening properties (including the structure and energy and / or physicochemical properties).

[0016] The screening properties are related to the drugability of the protein folding intermediates.

[0017] According to this specification, a "folding pathway" describes the transition of a protein from an unfolded to a natively folded state over time, i.e. how the amino acid chain reaches a thermodynamically stable state.

[0018] In one embodiment of the present invention, the application context of "protein folding" and "folding pathway" mentioned in this specification is endogenous protein synthesis, rather than denaturation or renaturation processes or conformers or "transient conformations".

[0019] According to this specification, "pharmaceutical usability" is the ability of a protein or any "conformer" of a protein to bind a drug (e.g., a small molecule, any other organic compound, a peptide, or an antibody), thereby bringing potential therapeutic benefits to patients. "Conformers" are different conformations of the same polypeptide. It reflects the statistical properties of the conformational isomers of the polypeptide and the thermodynamic state of the macromolecule.

[0020] As described above, the method provides two different steps: "identifying candidate protein folding intermediates" (based on a first set of properties of folding intermediates, referred to as "identification properties" in this specification) and "screening target protein folding intermediates" (based on a second set of properties of folding intermediates, referred to as "screening properties" in this specification).

[0021] In the following part of the description, several examples of "identification properties" and "screening properties" will be described. Obviously, "identification properties" and "screening properties" can be different, and in preferred embodiments of the present invention are actually different.

[0022] According to one embodiment of the method, the identifying characteristics comprise energetic and / or physicochemical properties of the protein folding intermediate.

[0023] According to an implementation option of this embodiment, the identification characteristic includes a free energy barrier between an intermediate and a native state of the protein, or a free energy barrier between an intermediate and a next intermediate toward the native state. In this case, the identification step includes identifying an intermediate as a "candidate protein folding intermediate" characterized in that the free energy barrier between the intermediate and the native state or the next intermediate toward the native state is greater than a free energy threshold.

[0024] In one embodiment, the free energy threshold is 7.5 kJ / mol.

[0025] According to another embodiment of said embodiment, the identifying property comprises the lifetime of the intermediate, defined as the reverse rate of transition from the intermediate to the native state or to the next intermediate along the folding pathway.

[0026] In this case, the method further comprises the step of assessing the lifetime of the intermediate, and the identifying step comprises identifying intermediates having lifetimes longer than a minimum lifetime threshold as "candidate protein folding intermediates".

[0027] In one embodiment, the minimum lifespan threshold may be at least three times the half-life of the protein under physiological conditions.

[0028] According to one embodiment of the method, the step of identifying at least one candidate protein folding intermediate comprises identifying a metastable state.

[0029] Now consider the second set of features, the “screening features”, for which the following details are provided.

[0030] According to one embodiment of the method, the screening properties include structural properties that are associated with the pharmaceutical acceptability of the candidate protein folding intermediate under consideration.

[0031] According to another embodiment of the method, the screening properties include scoring parameters for protein hot-spot identification.

[0032] In this specification, consistent with the terminology commonly used in drug research, "hot spots" are sites on the target protein that have a high ligand binding propensity and are therefore of great significance for drug development.

[0033] According to one embodiment of the method, the screening step further comprises selecting an intermediate having screening properties that does not exist in nature as the target protein folding intermediate.

[0034] According to one embodiment, the screening step further comprises selecting an intermediate having a druggable pocket that does not exist in nature as the target protein folding intermediate.

[0035] In this specification, the term "pocket" is consistent with the commonly used terminology in the field of drug research and refers to a cavity in the tertiary structure of a protein that is suitable for binding small molecules.

[0036] In particular, the concept of a druggable pocket refers to a specific binding site on a disease-associated protein target that is capable of binding drug-like molecules to thereby modulate the physiological function of the protein.

[0037] When the binding site cannot be determined from the 3D structure (e.g., ligand-protein complex) or other experimental data (e.g., drug resistance mutations), computational methods can be used to suggest possible sites.

[0038] The following part of the specification mentions several properties and / or parameters and / or global pocket descriptors (and corresponding exemplary values) used in the methods to characterize binding sites / pockets.

[0039] According to another embodiment, the screening step further includes selecting an intermediate having a druggable pocket as a target protein folding intermediate, characterized in that the root mean square deviation (RMSD) of the druggable pocket is greater than the root mean square deviation threshold of the pocket existing in the natural state.

[0040] In an embodiment, the RMS deviation threshold is equal to 2 angstroms. or greater than 2 angstroms

[0041] According to one embodiment of the method, the screening property comprises the presence of a pharmaceutically acceptable pocket in the candidate protein folding intermediate under consideration, wherein the pharmaceutically acceptable pocket is defined according to a pocket parameter.

[0042] According to one embodiment of the method, the screening step includes the steps of binding pocket identification, binding pocket characterization, and binding pocket drugability prediction.

[0043] In this case, the screening step comprises selecting one or more target protein folding intermediates among the identified candidate protein folding intermediates based on a comparison of said pocket parameter with a corresponding threshold value.

[0044] According to different implementation options of this embodiment, the above-mentioned pocket parameters include size parameters, and / or shape parameters, and / or position parameters, and / or the ratio of hydrophobicity to hydrophilicity.

[0045] In particular, the size parameters of the pocket may include the volume of the pocket and / or the depth of the pocket and / or the degree of closure and exposure of the pocket. The degree of exposure and closure provides a different way to measure the openness of the site to the solvent.

[0046] In one embodiment, the pocket volume threshold is at least

[0047] In one embodiment, the pocket exposure threshold is less than 0.49 and the pocket closed threshold is at least 0.78.

[0048] In one embodiment, the pocket depth threshold is at least

[0049] In the screening feature, the scoring parameters used for protein hotspot identification have been mentioned previously.

[0050] Regarding this feature, different embodiments of the present method provide the following scoring parameters: including "site score" (SiteScore), and / or "drugability score" (Dscore), and / or "drug score" (DrugScore) and / or pocket balance (pocketbalance).

[0051] These parameters are known per se in the art of modeling and characterization of proteins and protein intermediates and, based on the combination of values ​​associated with different properties, are suitable for assessing the "druggability" of proteins in their native state or of protein folding intermediates.

[0052] In fact, such scoring parameters are derived from known evaluation software packages.

[0053] For example, SiteMap (Halgren TA (2009) "Identifying and characterizing binding sites and assessing druggability", J. Chem. Inf. Model 49: 377-389) predicts the site score and druggability score (DScore) by a linear combination of only three single descriptors: the size of the binding pocket, its closedness, and a penalty for its hydrophilicity.

[0054] Another example is DoGSiteScorer (A. Volkamer, D. Kuhn, T. Grombacher, F. Rippmann, M. Rarey, “Combining global and local measures for structure-based druggability predictions” J. Chem. Inf. Model. 2012, 52, 360-37), which also generates a drugability score (DrugScore) ranging from 0 to 1.

[0055] Obviously, in other embodiments of the method, other known scoring parameters may be used, and / or new scoring parameters may be defined and adopted.

[0056] In this case, screening is also based on comparison of the scoring parameters with the corresponding thresholds.

[0057] In some embodiments of the method, the scoring parameter threshold is selected as follows.

[0058] In one embodiment, the pocket site score threshold is 0.8.

[0059] In one embodiment, the pocket drugability score (DScore) threshold is 0.98.

[0060] In one embodiment, the pocket drug score (DrugScore) threshold is 0.5.

[0061] In one embodiment, the pocket balance threshold is 1.0.

[0062] As described above, the present method provides basic properties for identifying candidate protein folding intermediates and screening target protein folding intermediates.

[0063] In addition, the method also provides criteria for the above identification and screening steps. These criteria are based on, for example, comparison of corresponding threshold values ​​with parameters and / or values ​​related to the selected characteristics for identification and / or screening.

[0064] Exemplary values ​​of the threshold values ​​have been provided in the above description. Nevertheless, it will be appreciated by those skilled in the art that the present method is not limited to the exemplary values ​​mentioned, as the threshold values ​​may vary specifically depending on the type of protein or other requirements.

[0065] According to one embodiment of the method, the time series of the protein folding pathway is modeled by computer simulation based on molecular mechanics (MM) or quantum mechanics-molecular mechanics (QM-MM) methods.

[0066] According to different implementation options of this embodiment, the above computer simulation is performed by computer simulation based on ratchet-and-pawl molecular dynamics, and / or bias function calculation method, and / or by self-consistent path sampling calculation method.

[0067] Those skilled in the art can easily understand that the above algorithms and calculation methods are merely exemplary and are provided for clarity of disclosure, and the present method can be implemented by applying other algorithms and calculation methods that provide similar results and are not explicitly mentioned herein.

[0068] For more detailed information on exemplary algorithms for effectively performing the time series modeling step of the above-mentioned protein folding pathway, see the scientific papers: "S. Orioli, Sa Beccara, and P. Faccioli, J. Chem. Phys. 147, 064108 (2017)"; "C. Camilloni, RA Broglia, and G. Tiana, J. Chem. Phys. 134, 045105 (2011)"; "S. a Beccara, L. Fant, and P. Faccioli, Phys. Rev. Lett. 114, 098103 (2015)".

[0069] According to other possible embodiments of the method, the step of modeling the time series of the protein folding pathway is performed by computer simulation based on any other computer method, thereby reconstructing the protein folding pathway or identifying folding intermediates.

[0070] According to other possible embodiments of the method, the step of modeling the time series of the protein folding pathway is performed by any experimental method to reconstruct the protein folding pathway or to identify folding intermediates.

[0071] The present invention describes an in silico drug discovery approach based on targeting folding intermediates.

[0072] According to any of the above embodiments, the method comprises the following steps: identifying a target protein folding intermediate, wherein the intermediate is suitable for testing as a target in a computer drug discovery method.

[0073] Based on this, the method of computer drug development has realized the study of computer drugs on the selected target protein folding intermediates.

[0074] According to one embodiment of the present method, the step of performing computer drug development for each selected target protein folding intermediate (in which a druggable pocket or hotspot has been identified) includes: identifying potential ligands based on the characteristics of one or more druggable pockets and / or hotspots identified in the target protein folding intermediate, wherein the one or more druggable pockets and / or hotspots are considered to be possible binding sites; then, modeling the interaction between each identified ligand and each identified binding site by computer simulation; and finally, selecting a ligand based on the above model.

[0075] In the different embodiments of the present method, in principle any known computer drug discovery program can be used.

[0076] Also included herein is a computer program comprising at least one instruction which, when executed by a computer, causes the computer to perform the method for identifying a target protein folding intermediate as described in any of the above embodiments.

[0077] The present invention also includes a computer program, which includes at least one program instruction. When executed by a computer, the program instruction causes the computer to perform the computer drug discovery method described in any of the above exemplary embodiments.

[0078] Obviously, in the context of this specification, the term "computer" is intended to include in the broadest sense a supercomputer and / or a computer cluster, or any other type of known electronic processor.

[0079] The present disclosure also includes a carrier and / or a medium and / or a carrier medium carrying the above-mentioned one or more computer programs.

[0080] In the following portion of the description, more details are provided on exemplary, non-limiting embodiments of the invention.

[0081] The following description outlines the actual operations performed by a technician when implementing the method.

[0082] (i) Target identification.

[0083] The target identification and screening process has been extensively described above.

[0084] (ii) Identification of druggable pockets.

[0085] The selected candidate protein folding intermediates are subjected to computer search analysis to select target protein folding intermediates that are exposed in the solvent and have a druggable pocket, wherein the pocket is unique in the selected protein folding intermediate and does not exist in the native form of the protein.

[0086] (iii a) Small molecule identification.

[0087] In one embodiment of the present invention, the screening step includes selecting the highest-level pocket, i.e., the highest-level target protein folding intermediate, which is then used for virtual drug screening to identify potential small molecule ligands. According to the research field / goal, a specific (ad-hoc) virtual chemical library is designed and constructed. The most promising candidates are selected by applying computational methods / tools in the prior art, such as docking-based virtual screening, ligand affinity evaluation, ligand efficiency (LE) and ligand lipophilicity efficiency (LLE), removal of pan-active screening interfering compounds and potential polymers, physicochemical and ADMET compound property evaluation, similarity and cluster analysis of virtual compounds.

[0088] The methods described and claimed herein allow for the selection of the most promising candidates in an efficient manner for subsequent testing in more costly and time-consuming in vitro experiments.

[0089] (iii b) Other molecular identification.

[0090] In other embodiments of the invention, compounds identified according to the methods described may include antibodies (and antibody derivatives), toxins, nucleic acids; these molecules may also be represented by endogenous metabolites.

[0091] (iv) Cell-based assays

[0092] For example, ligands predicted by virtual screening were validated in stably transfected heterologous cell systems by testing their ability to post-translationally reduce target protein expression in a dose-dependent manner.

[0093] In principle, any compound that binds to a protein folding intermediate with sufficiently high affinity can reduce its energy state, stabilize its structure, and thus extend its half-life. In a cellular environment, this stabilization may produce abnormally long-lived folding intermediates that are recognized by the cell's folding quality monitoring mechanism, thereby hindering the correct addition of modifications after translation, and / or causing its degradation (e.g., degradation through the proteasome pathway and / or autophagy). Following this principle, in a standard heterologous cell system (e.g., HEK293, CHO, SH-SY5Y or HeLa cells), in a dose-dependent manner, the computer-predicted candidate ligands of the identified folding intermediates are tested by standard biochemical techniques (e.g., Western blot) to select compounds that can reduce or completely inhibit the overall expression of the target protein (e.g., proteins for which suitable target folding intermediates have been identified). If the target protein is not endogenously expressed in the available cell system, an expression vector is designed and the cells are stably transfected to express the target protein.

[0094] Example

[0095] Prion diseases are associated with the conformational conversion of PrPC, an endogenous glycosylphosphatidylinositol (GPI)-anchored cell surface glycoprotein, into a misfolded isoform termed the “itchy form of PrP” (or PrPSc), which accumulates in the central nervous system of affected individuals. PrPSc is an infectious protein (prion) that lacks any detectable encoding nucleic acid information and replicates by directly binding to PrPSc and triggering its conformational rearrangement into new PrPSc molecules. Considerable evidence suggests that the necessary information to specify prion biological properties is encoded solely in the structure of PrPSc and that different PrPSc conformers can give rise to distinct strain properties, including the neuropathological and clinical features of various forms of prion disease. Disease-associated mutations in the PrP gene are thought to favor the misfolding of PrPC into aggregated and pathogenic PrPSc-like forms. Despite these unique features, increasing evidence from genetic, biophysical, and biochemical studies suggests that the pathogenic mechanisms of prion diseases may be at the root of neurodegenerative pathways occurring in several other diseases. Increasing evidence suggests that compounds that can modulate PrPC expression and / or activity could offer novel therapeutic perspectives for a variety of neurodegenerative diseases.

[0096] The methods described herein have been used to identify small molecules that target PrPC folding intermediates and thus have the potential to post-translationally inhibit PrPC expression.

[0097] The folding pathway of PrP was calculated using the BF procedure in explicit solvent, with the Amberff99SB-ILDN force field and the TIP3P solvent model of Gromacs 4.6.5, where the BF method is integrated in the Plumed 2.0.2 plugin. Twelve independent unfolded conformations were obtained through MD simulations of thermal unfolding starting from the energy-minimized native structure of PrP. For each of the 12 initial unfolding conditions, all 20 rMD trial trajectories generated by rMD were scored by evaluating their BF function. Three sets of trajectories were discarded because none of these three sets had folding pathways converging to the native state. Applying this protocol, nine independent folding trajectories were collected by selecting the least-biased trajectories (LBTs) according to the BF procedure, see S. a Beccara, L. Fant, and P. Faccioli, Phys. Rev. Lett. 114, 098103 (2015).

[0098] To search for candidate folding intermediates, the folding trajectories calculated using the ratchet-and-pawl molecular dynamics algorithm (see S. a Beccara, L. Fant, and P. Faccioli, Phys. Rev. Lett. 114, 098103 (2015)) were used to calculate the two-dimensional frequency distribution histogram of the collective variable Q of the native contact fraction and the ratio of RMSD. The variable Q of the protein conformation is obtained by dividing the number of atom pairs with a relative distance less than 7.5 Å in the conformation by the number of atom pairs with a relative distance less than 7.5 Å in the native conformation.

[0099] Folding intermediates have been observed on the pathway of 0.5 nm < RMSD < 0.9 nm and 0.65 < Q < 0.85 (see Figure 1 , circled part). To extract the conformations describing the obtained folding pathways, a two-step filtering of the LBTs was applied:

[0100] (i) For each set of conformations explored by the LBTs, only the conformations located in Figure 1 the high-density region were retained. To perform this task, the negative logarithm of the probability of observing a given conformation (defined according to Q and RMSD) was calculated, and all points with a stability deviation of the global minimum higher than 3.5 kBT (long-lived state region) were excluded.

[0101] (ii) To focus only on the region of interest, only the conformations showing 0.5 nm < RMSD < 0.9 nm and 0.65 < Q < 0.85 were retained.

[0102] Cluster analysis of intermediate conformations was performed using the k-means clustering method (RStudio: Integrated Development for R. RStudio, Inc., Boston, MA). The number of clusters was selected using the "elbow method". The contact map distance was used as the clustering metric. The clustering process generated three different clusters, which are referred to as clusters C1, C2, and C3 (see Figure 2 ). Conformations sampled from the densest such clusters (C2 and C3) showed a displaced helix-1, with a potential binding site. To identify a single representative conformation for each of the three clusters (C1, C2, and C3), first, we calculated the average contact map in the group, and then we identified the structure in the group such that the distance between its contact map and the average contact map in the cluster was minimal.

[0103] In silico and virtual drug screening have been used to identify potential ligands for FI-PrP, focusing on unique binding sites present in representative elements of cluster C3 that are absent in the native state of PrP ( Figure 3 Computerized drug screening was performed at this site according to the following procedures:

[0104] Starting from a representative conformation of cluster C3, we performed a 50 ns molecular dynamics (MD) study in an explicit solvent model at 300 K. In such a model, the relative positions of the backbone atoms are kept fixed in order to sample exclusively the arrangements of the side chains.

[0105] The conformations visited by this MD trajectory are concentrated in two groups because the MD trajectory indicates the existence of two major pockets. We randomly sampled 10 conformations from each group.

[0106] We analyzed the resulting 20 conformations using Sitemap to identify the one containing the most druggable site.

[0107] The conformations identified in the previous step were targeted for drug screening using the Asinex commercial library containing approximately 250,000 small molecules.

[0108] The results of this virtual screening were screened according to expected pharmacodynamics and pharmacokinetics, resulting in a final library of 275 virtually active compounds. For illustrative purposes, we report here the chemical structures of the group of drug candidates predicted to have the highest binding affinity:

[0109] Formula 1

[0110] In principle, compounds that bind to protein folding intermediates with sufficient affinity could lower their energy state, stabilize their structure, and thus extend their half-life. In the cellular context, for proteins synthesized directly within the lumen of the endoplasmic reticulum (ER), such as PrPC, such stabilization could generate abnormally long-lived folding intermediates that could be recognized by the ER quality control (ERQC) machinery, hinder the correct addition of post-translational modifications, and potentially lead to degradation (e.g., ER-associated degradation and / or autophagy). Following this principle, putative ligands for FI-PrP identified by the proposed screening protocol could be tested for their ability to induce degradation and / or post-translational modifications of wild-type (WT) PrPC expressed in stably transfected HEK293 cells. First, cells should be incubated in the presence of increasing concentrations of each molecule (indicatively between 0.01 and 50 μM). Then, PrPC expression levels should be obtained by Western blot analysis. High-affinity small molecules are expected to cause a dose-dependent reduction in PrP cellular expression.

[0111] As described above, the described method allows implementing the scope of the present invention based on the above-mentioned features.

[0112] In particular, the present method allows for the identification of binding pockets in folding intermediates in a unique manner (that cannot be identified by known solutions), which brings clear advantages related to the identification of potential drug targets.

Claims

1. A method for identifying a target protein folding intermediate, wherein the intermediate is suitable for testing as a target in a drug development process, the method comprising the following steps, implemented by computer calculation: - Modeling the time sequence of events that define a protein folding pathway, including modeling and / or calculating the structural and / or energetic and / or physicochemical properties of one or more protein folding intermediates along said folding pathway; wherein said modeling step is performed by computer simulation based on molecular mechanics (MM) or quantum mechanics-molecular mechanics (QM-MM) methods; - identifying at least one candidate protein folding intermediate along the simulated folding pathway based on identification characteristics including said structural and / or energetic and / or physicochemical properties; wherein the identifying step comprises identifying intermediates having lifetimes longer than a minimum lifetime threshold as candidate protein folding intermediates; - selecting one or more target protein folding intermediates from among the at least one candidate protein folding intermediate based on screening properties including the structural and / or energetic and / or physicochemical properties, wherein the screening properties are related to the drugability of the protein folding intermediate; The screening step further comprises: -Selecting intermediates with screening properties that do not exist in the natural state as target protein folding intermediates; and / or - Selecting an intermediate with a druggable pocket that does not exist in the natural state as a target protein folding intermediate, or selecting an intermediate with a druggable pocket, characterized in that the root mean square deviation of the druggable pocket is greater than the root mean square deviation threshold of the pocket existing in the natural state.

2. The method according to claim 1, wherein the identifying characteristics include energetic and / or physicochemical properties of protein folding intermediates, and / or The screening characteristics include: - structural properties relevant to the drugability of the candidate protein folding intermediate under consideration, and / or - Scoring parameters for protein hotspot identification.

3. The method according to claim 2, wherein: The identifying characteristics include a free energy barrier between an intermediate and a native state of the protein, or a free energy barrier between an intermediate and a next intermediate toward the native state, and The step of identifying the intermediate as a candidate protein folding intermediate is characterized in that the free energy barrier between the intermediate and the native state or the next intermediate toward the native state is greater than a free energy threshold.

4. The method according to any one of the preceding claims, further comprising: -Evaluation of the lifetime of intermediates, defined as the reverse speed of the transition from an intermediate to the native state or to the next intermediate along the folding pathway.

5. The method of claim 4, wherein the step of identifying at least one candidate protein folding intermediate comprises identifying a metastable state.

6. The method according to claim 5, wherein: The screening properties include the presence of a druggable pocket in the candidate protein folding intermediate under consideration, wherein the druggable pocket is defined according to pocket parameters, and The screening step comprises selecting one or more target protein folding intermediates from the identified candidate protein folding intermediates based on a comparison of the pocket parameter with a corresponding threshold value, and / or The pocket parameters include size parameters, and / or shape parameters, and / or position parameters, and / or the ratio of hydrophobicity to hydrophilicity, wherein the pocket size parameters include the volume of the pocket and / or the depth of the pocket, and / or the degree of closure and / or exposure of the pocket.

7. The method according to claim 2, wherein the scoring parameters include "site score", and / or "drugability score", and / or "drug score", and / or pocket balance.

8. The method according to claim 3, wherein: - the free energy threshold is 7.5 kJ / mol; and / or - the intermediate has a lifetime threshold of at least three times the half-life of the protein under physiological conditions; and / or - The RMS deviation is equal to or greater than 9. The method according to claim 6, wherein: - The pocket volume is at least and / or - the pocket depth is at least and / or - the pocket exposure value is ≤ 0.49; and / or - The pocket closure value is ≥ 0.

78.

10. The method according to claim 7, wherein: - the site score value is ≥ 0.8; and / or - the pharmaceutical acceptability score value is ≥ 0.98; and / or - The drug score value is ≥ 0.5; -The pocket balance value is ≥ 1.

11. The method according to claim 1, wherein the step of time series modeling of the protein folding pathway is performed by computer simulation based on ratchet mechanism molecular dynamics and / or bias function calculation method and / or by self-consistent pathway sampling calculation method.

12. The method according to claim 1, wherein the step of modeling the time series of the protein folding pathway is performed by any computer method or experimental method to reconstruct the protein folding pathway or identify folding intermediates.

13. A computer drug development method based on targeting folding intermediates, comprising: - implementing the method for identifying a target protein folding intermediate according to any one of claims 1 to 12, wherein the intermediate is suitable for testing as a target in a computer drug development method; -In silico drug discovery of selected target protein folding intermediates.

14. The method of claim 13, wherein once druggable pockets or hotspots have been identified, the step of performing in silico drug discovery on each selected target protein folding intermediate comprises: - identifying potential ligands based on the properties of one or more druggable pockets and / or hotspots identified in the target protein folding intermediate, which are considered to be possible binding sites; - Modeling the interaction of each identified ligand with each identified binding site by computer simulation; - Selecting ligands based on said modelling.

15. A storage medium carrying a computer program, wherein the computer program comprises at least one program instruction, and when executed by a computer, the program instruction causes the computer to execute the method for identifying a target protein folding intermediate according to any one of claims 1 to 12, Or causing a computer to execute the computer drug development method according to any one of claims 13 to 14.

Images