Targeted target epitope antibody design method

Abstract

The invention discloses a target epitope targeting antibody design method, and belongs to the technical field of artificial intelligence technology and computational structure biology technology, and the method comprises the following steps: step 1, epitope information acquisition: acquiring epitope site information on a target antigen structure by using a python code; step 2, mask driving; step 3, generating a conformation; 4, matching verification: carrying out simulated docking on the designed antibody and the target antigen in a digital space; and step 5, energy efficiency evaluation: calculating the binding energy of the designed antibody and the target antigen by using the code. According to the present invention, the template antibody sequence is redesigned by using the target antigen information and the target epitope information, such that the antibody sequence capable of targeting the designated epitope and having the high affinity can be designed; a large number of targeted antibodies to be selected are designed in a short time so as to reduce the harm possibly caused by pathogens.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence technology and computational structural biology technology, and particularly relates to an antibody design method targeting target epitopes. BACKGROUND

[0002] Antibodies are immunoglobulins produced in response to antigen stimulation, have the biological activity of specific binding to specific antigen epitopes, and are the core effector molecules of the body's immune response. In recent years, therapeutic antibody drugs have played an increasingly important role in the clinical treatment of tumors, autoimmune diseases and infectious diseases due to their high targeting and controllable safety characteristics.

[0003] Current antibody development mainly relies on in vitro screening technology systems, and its standard process includes:

[0004] Construction of immune library: obtain a diverse antibody sequence group through animal model immunization or in vitro display technology;

[0005] High-throughput screening: use enzyme-linked immunosorbent assay (ELISA), surface plasmon resonance (SPR) or flow sorting technology for multiple rounds of affinity screening;

[0006] Functional verification: in vitro / in vivo verification of binding activity, neutralization potency and cytotoxicity of candidate antibodies.

[0007] The technical path has the following core defects:

[0008] Long development cycle: it usually takes 6-18 months from antigen immunization to obtain preclinical candidate molecules;

[0009] High economic cost: the average cost of developing a monoclonal antibody is 2-5 million US dollars;

[0010] Limited screening efficiency: the library capacity is limited by experimental throughput, resulting in a high screening rate of high-affinity antibodies;

[0011] Insufficient target controllability: it is difficult to accurately control the binding specificity of antibodies to specific antigen epitopes; therefore we propose an antibody design method targeting target epitopes to solve this problem. SUMMARY

[0012] The purpose of the present application is to provide an antibody design method targeting target epitopes to solve the problems raised in the background art.

[0013] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows:

[0014] An antibody design method targeting target epitopes, comprising:

[0015] Step 1, epitope information acquisition: use python code to obtain epitope site information on the target antigen structure;

[0016] Step 2, mask driving: use epitope site information and target antigen structure, based on antibody conservative amino acid alignment analysis and shadow mask mechanism, redesign the target template antibody sequence;

[0017] Step 3, conformation generation: use protein language model to predict the structure of the redesigned antibody sequence, and select and obtain the most reasonable antibody structure;

[0018] Step 4, docking verification: simulate the docking of the designed antibody and the target antigen in digital space;

[0019] Step 5, energy efficiency evaluation: use the code to calculate the binding energy of the designed antibody and the target antigen.

[0020] Preferably, the template antibody sequence in step 2 can be an antibody sequence targeting other epitopes of the antigen or an antibody sequence targeting other antigens.

[0021] Preferably, the antibody conservative amino acid alignment analysis in step 2 is as follows:

[0022] Step 201, use the characteristics of the FR region amino acid in the antibody sequence, which is more conservative, to build a database comparison of the FR region of the known antibody;

[0023] Step 202, define the amino acid and its site as the template amino acid when the occurrence frequency of the amino acid type is more than 95%;

[0024] Step 203, align the template antibody sequence with the structure template to obtain the possible structure template of the FR region of the template sequence.

[0025] Preferably, the method for redesigning the target template antibody sequence in step 2 is as follows: place a shadow paratope on the specific epitope side, combine the shadow paratope with the structure template, and use the structure template and antigen epitope information to preliminarily limit the paratope; and calculate the probability of each amino acid appearing at each site of the paratope, and select the highest probability as the sequence of the paratope.

[0026] Preferably, in step 3, the local confidence and iptm value of the predicted structure are calculated to evaluate the rationality of the predicted structure, and the plddt and iptm>75 are used as the screening standard. If there are multiple structures meeting the requirements, the highest plddt and iptm value is selected as the candidate structure.

[0027] Preferably, in step 4, the specific steps of simulating docking are as follows:

[0028] Step 401, simulating the docking of the antigen-antibody in space by using the rigid docking method,

[0029] Step 402, finding out whether the approximate docking site is an epitope site;

[0030] Step 403, keeping the antigen position unchanged;

[0031] Step 404, fine-tuning the antibody position in space in multiple steps to find a model in which the two are combined in space to best meet the natural conditions.

[0032] Preferably, in step 4, after each simulation docking, a relaxation is performed to reduce unreasonable structures that may be generated during the rigid docking process and further simulate the slight deformation of the protein docking process.

[0033] Preferably, after the two simulation dockings in step 4, the predicted model needs to be further screened, and the screening criteria are mainly the binding surface and the total energy score. The binding surface (dSASA) needs to be greater than >1200, and the total energy score needs to be less than <-1000.

[0034] Preferably, in step 5, the specific steps are as follows: the energy terms of the predicted antigen-antibody complex structure are weighted and summed, and the energy weighted sum of the antigen-antibody itself is subtracted to calculate the approximate value of the binding energy of the two.

[0035] The beneficial effects of the present application are:

[0036] 1、In the present application, the antibody design method targeting the target epitope, by using the target antigen information and the target epitope information to redesign the template antibody sequence, so as to design the antibody sequence that can target the specified epitope and has high affinity, and at the same time, the structure is predicted by using the space structure modeling method, and the designed antibody is simulated and detected in digital space whether it can combine with the target epitope and the combination condition under natural conditions, compared with the traditional antibody design scheme, it has the advantages of high efficiency and low cost, and a large number of target antibodies can be designed in a short time when there is no corresponding target antibody for a certain epitope, so as to reduce the harm caused by pathogenic bacteria;

[0037] 2. In the application, the antibody design method for targeting the target epitope, by analyzing the spatial position of the target epitope, and combining the highly conserved structure template of the antibody FR region for comparison and analysis, the shadow mask mechanism is used to dynamically constrain and guide the binding site design of the template antibody, so that the newly generated Paratope sequence accurately matches the target epitope, which breaks the dependence on natural antibodies of specific target points, and can efficiently generate target antibody sequences based on key epitope information of antigens in the absence of ready-made antibodies, greatly improving the targeting and feasibility of the design;

[0038] 3. In the application, the antibody design method for targeting the target epitope, by integrating the leading protein language model to predict the high-quality antibody structure, and using strict local conformation confidence and overall folding quality evaluation indicators to select the most reasonable structure model, then through the innovative two-stage simulation docking process: first, verify the epitope binding possibility through spatial coarse positioning, then fine-tune and optimize the energy for multiple rounds to simulate the real protein binding dynamics and small conformation adjustment, finally, through strict evaluation of the quality and affinity of the final complex by combining the size of the binding surface and the comprehensive binding energy strength and other multi-dimensional physical indicators, a closed loop of design-generation-verification is formed, which guarantees the structural feasibility and binding efficiency of antibody design;

[0039] 4. In the application, the antibody design method for targeting the target epitope, by using antibody sequences of any source, whether they target the same antigen or other epitopes, or even other completely different antigens, through the comparison and analysis of the conservative FR region skeleton template and the shadow mask guided reprogramming, this method can effectively utilize the structural framework information in existing antibody resources, and redesign its binding site to match the target epitope, this high flexibility greatly expands the source of design resources, effectively overcomes the obstacles caused by the lack of template antibodies corresponding to specific epitopes, and makes it valuable in the development of new pathogens and multi-epitope targeting strategies;

[0040] 5. In the application, the antibody design method for targeting the target epitope, from antigen epitope information extraction, template sequence redesign, structure prediction screening to simulation docking and binding energy evaluation, can be efficiently completed in a computing environment, compared with traditional methods that rely on a large number of physical experiments, such as time-consuming and laborious animal immunization or library screening technology, this method can generate and screen a large number of candidate antibody designs in a very short time, significantly shorten the development cycle, greatly reduce the early development cost and resource consumption, and provide efficient "digital antibody design" response capability for new infectious disease epidemics or rapid iteration of pathogens. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1A flowchart for designing a targeting antibody for an antigen epitope according to an embodiment of the present application;

[0042] Figure 2 A structure diagram of an antigen S-protein of human respiratory syncytial virus (RSV) in the embodiment, in which a target epitope site IV is marked in dark color;

[0043] Figure 3 A structure diagram of an antibody redesigned according to the method of the present application, in which a H chain structure is marked in dark color and a L chain structure is marked in light color;

[0044] Figure 4 A complex prediction structure of a target antigen S-protein of RSV and a designed antibody. DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments of the present application.

[0046] Reference Figures 1-4 A method for designing an antibody targeting a target epitope, comprising:

[0047] Step 1, epitope information acquisition: obtaining epitope site information on a target antigen structure by using a python code.

[0048] Step 2, mask driving: based on antibody conservative amino acid alignment analysis and shadow mask mechanism, the target template antibody sequence is redesigned by using the epitope site information and the target antigen structure.

[0049] Step 3, conformation generation: the antibody sequence redesigned is subjected to structure prediction by using a protein language model, and the most reasonable antibody structure is screened and obtained.

[0050] Step 4, fitting verification: the designed antibody and the target antigen are simulated and docked in digital space.

[0051] Step 5, energy efficiency evaluation: the binding energy of the designed antibody and the target antigen is calculated by using a code.

[0052] In the embodiment, the template antibody sequence in step 2 can be an antibody sequence targeting other epitopes of the antigen or an antibody sequence targeting other antigens.

[0053] In the embodiment, the specific steps of the antibody conservative amino acid alignment analysis in step 2 are as follows:

[0054] Step 201, the FR region of the known antibody is subjected to library comparison and alignment by using the characteristics that the amino acid occurrence in the FR region of the antibody sequence is relatively conservative.

[0055] Step 202, define the amino acid and its site as template amino acid when the occurrence frequency of the site and the amino acid type is more than 95% in the sequence;

[0056] Step 203, align the template antibody sequence with the sequence to obtain the possible structure template of the FR region of the template sequence.

[0057] In this embodiment, the method for redesigning the target template antibody sequence in step 2 is as follows: placing a shadow paratope on the specific epitope side, combining the shadow paratope with the structure template, and using the structure template and antigen epitope information to preliminarily limit the paratope; and calculating the probability of each amino acid appearing at each site of the paratope, and selecting the highest probability as the sequence of the paratope.

[0058] In this embodiment, in step 3, when the most reasonable antibody structure is screened and obtained, the local confidence and iptm value of the predicted structure are calculated to evaluate the rationality of the predicted structure, and the screening standard is plddt and iptm>75, if there are multiple structures meeting the requirements, the highest value of plddt and iptm is selected as the candidate structure.

[0059] In this embodiment, in step 4, the specific steps of the simulation docking are as follows:

[0060] Step 401, using the rigid docking method to simulate the antigen-antibody docking in space,

[0061] Step 402, finding out whether the approximate docking site is an epitope site;

[0062] Step 403, keeping the antigen position unchanged;

[0063] Step 404, adjusting the antibody position in space in multiple steps to find a model that best meets the natural conditions of the combination of the two in space.

[0064] In this embodiment, in step 4, after each simulation docking, a relax is performed to reduce the unreasonable structures that may be generated in the rigid docking process and further simulate the slight deformation generated in the protein docking process.

[0065] In this embodiment, after the two simulation dockings in step 4, the predicted model needs to be further screened, and the main screening criteria are the binding surface and the overall energy score. The binding surface (dSASA) needs to be greater than >1200, and the overall energy score needs to be lower than <-1000.

[0066] In this embodiment, in step 5, the specific steps are as follows: the energy terms of the predicted antigen-antibody complex structure are weighted and summed, and the weighted sum is subtracted from the energy of the antigen-antibody itself, so as to calculate the approximate value of the binding energy of the combination of the two.

[0067] In this embodiment, the antibody targeting the site IV epitope on the S-protein of RSV is designed.

[0068] First, search and download the structure of the S-protein of RSV from the public database of proteins such as PDB, and according to the current reports, mark the site of site IV on the S-protein (422-471), and the epitope sequence is: CTASNKNRGIIKTFSNGCDYVSNKGMDTVSVGNTLYYVNKQEGKSLYVKG, and the antigen structure diagram is shown in Figure 2 As shown, the epitope site information is saved as a json file using a python script for subsequent use.

[0069] Based on the antigen information and the json file carrying the site of the epitope, the antibody targeting another epitope of RSV-site III (PDB number: 4JHW) is designed, and the original antibody sequence is: H-chain: QVQLVQSGAEVKKPGSSVMVSCQASGGPLRNYIINWLRQAPGQGPEWMGGIIPVLGTVHYAPKFQGRVTITADESTDTAYIHLISLRSEDTAMYYCATETALVVSTTYLPHYFDNWGQGTLVTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHKPSNTKVDKKVEPKSC;

[0070] L-chain: DIQMTQSPSSLSAAVGDRVTITCQASQDIVNYLNWYQQKPGKAPKLLIYVASNLET GVPSRFSGSGSGTDFSLTISSLQPEDVATYYCQQYDNLPLTFGGGTKVEIKRTVAAPSVFIFPPSDEQ LKSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC; the corresponding conserved template structure of the template antibody sequence is obtained according to the analysis of the conserved amino acids of the antibody; and the 6 CDR sequences of the template antibody sequence are all shaded masked, CDR-H3 is the main shaded paratope, the amino acid occurrence probability of all positions is calculated, the selected antibody sequence is obtained, and the protein language model is further completed to ensure the sequence integrity of the designed antibody and the targeting and high affinity of the epitope. The redesigned antibody sequence is: H-chain: QVQLVQSGASEVKKPGSSVMVSCQASGGTFTSTFSTYAINWLRQAPGQGPEWMGGIIPIIPFGTAHYA PKFQQGRVTITADESTDTAYIHLISLRSEDTAMYYCARDRGYGYYYFDYWGQGTLVTVS; L-chian: DIQMTQSPSSLSAAVGDRVTITCQASQDILNSGNQVNYLNWYQQKPGKAPKLLIYVASNAPGVVSNLE TGVPPSRFSGSGSGSGTDFSLTISSLQPEDVATYYCQQYDNYSYTLPLTFGGGTKVEIK.

[0071] The structure prediction tool is run in the server, and AlphaFold3 or RosettaFold2 can be selected for structure prediction. Here, the spatial structure of the redesigned antibody sequence is predicted, and the plddt and iptm and ptm of each predicted structure are calculated and integrated as a ranking-score to evaluate the structure rationality. The ranking-score is sequentially ordered from high to low, and the ranking-score of the top 10 is shown in Table 1. The structure with the highest ranking-score is selected as the designed antibody structure, and is saved in a computer-recognizable file format, such as a pdb file.

[0072] The designed antibody structure is shown in Figure 3 )

[0073] seed sample ranking_score 1 0 0.846464 1 1 0.849866 1 2 0.840173 1 3 0.854403 23 0 0.846972 23 1 0.847134 23 2 0.839615 23 3 0.851606 23 4 0.844674

[0074] Table 1

[0075] With the predicted antibody structure, first, the rigid docking method is used to calculate the possible docking mode and docking situation of the antibody and antigen in the natural state. Here, we set the antibody as the receptor and the antigen as the ligand, that is, in the posture of keeping the antibody unchanged, the position and orientation of the antigen in space are changed to view its binding model with the antibody, to determine whether the designed antibody can dock to the corresponding antigen epitope. After confirming that the antibody can dock to the corresponding antigen epitope, the corresponding model is subjected to relax processing to reduce the potential unreasonable structure caused by rigid docking. After the relax is finished, in order to further find the docking state of the antigen and antibody that is most consistent with the natural situation, the antibody is rotated, slightly translated and fine-tuned around the binding site under the condition of keeping the position of the antigen unchanged, to view the differences in the binding interface and energy caused thereby. At the same time, according to the binding surface (dSASA) and the overall energy score, these models are screened, and at the same time, the rationality of the binding interface and other conditions are combined to screen multiple models. The screened results are subjected to relax to simulate the slight deformation of the protein itself for binding during the docking process, and to further reduce the unreasonable conformation caused by rigid docking. (The final designed antibody and antigen binding diagram is shown in FIG. 6) Figure 4 )

[0076] Finally, for the results after relax, the binding energy is calculated, and the binding interface is finally analyzed. At the same time, in order to compare the advantages and disadvantages of the designed antibody with the natural antibody, the binding energy of the natural antibody (PDB number: 8WSQ) targeting site IV of the S-protein of RSV is also calculated. The binding energy calculation of the designed antibody and the natural antibody 8WSQ is shown in Table 2:

[0077] antibody binding energy calculation designing antibodies 51.239 8WSQ 48.878

[0078] Table 2

[0079] In summary, the antibody design method of the present application can be based on a specific antigen epitope, and only rely on a computer to efficiently generate an antibody sequence and structure with high affinity and capable of targeting the antigen epitope in digital space, providing important support for subsequent antibody drug development. Not only can it greatly save experimental costs, but it is also not limited by wet experimental conditions, and can explore all possible sequences and conformations during generation, and use set parameters and other means to artificially control the generation of antibodies and design goals. It improves the efficiency and quality of antibody engineering, reduces the trial-and-error cost in the wet experiment process, improves the efficiency of antibody generation and expression, and provides more drug possibilities for antibody molecules.

[0080] The above describes in detail the antibody design method provided by the present application for targeting an epitope. The principles and implementation methods of the present application are described using specific examples. The above examples are only used to help understand the method of the present application and its core idea. It should be noted that for those skilled in the art, without departing from the principles of the present application, the present application can be improved and modified in several ways, and these improvements and modifications also fall within the scope of protection of the claims of the present application.

Claims

1. A method of antibody design targeting a target epitope, characterized in that, The application relates to a method for designing an antibody sequence, comprising the following steps: step 1, epitope information acquisition: obtaining epitope site information on a target antigen structure by using a python code; step 2, mask driving: based on antibody conservative amino acid alignment analysis and a shadow mask mechanism, a target template antibody sequence is redesigned by using the epitope site information and the target antigen structure; step 3, conformation generation: a protein language model is used to predict the structure of the redesigned antibody sequence, and the most reasonable antibody structure is screened and obtained; step 4, docking verification: the designed antibody and the target antigen are simulated to dock in a digital space; and step 5, energy efficiency evaluation: the binding energy of the designed antibody and the target antigen is calculated by using a code. The template antibody sequence in step 2 can be an antibody sequence targeting other epitopes of the antigen or an antibody sequence targeting other antigens. The specific steps of the antibody conservative amino acid alignment analysis in step 2 are as follows: step 201, the FR region of the known antibody is subjected to library alignment by using the characteristics that the amino acid occurrence in the FR region of the antibody sequence is relatively conservative; step 202, the amino acid with a higher occurrence frequency and the amino acid type with more than 95% of the amino acid at the site are defined as the template amino acid; and step 203, the FR region of the template antibody sequence is aligned with the structure template to obtain the possible structure template of the FR region of the template sequence. In step 2, the method for redesigning the target template antibody sequence is as follows: a shadow paratope is placed on the specific epitope side, the shadow paratope is combined with the structure template, the paratope is preliminarily limited by using the structure template and the antigen epitope information, the probability of the possible occurrence of the amino acid at each site of the paratope is calculated, and the amino acid with the highest probability is selected as the sequence of the paratope. In step 3, the local confidence and the iptm value of the predicted structure are calculated to evaluate the rationality of the predicted structure, the plddt and iptm>75 are used as the screening standard, if multiple structures meet the requirement, the structure with the highest plddt and iptm value is selected as the candidate structure. In step 4, the specific steps of the simulated docking are as follows: step 401, the rigid docking method is used to simulate the antigen-antibody docking in space; step 402, whether the approximate docking site is an epitope site is ascertained; step 403, the antigen position is kept unchanged; and step 404, the antibody position is adjusted in space in multiple steps to find a model in which the combination of the two in space is most consistent with the natural condition.

2. The antibody design method targeting an epitope of interest according to claim 1, wherein, In step 4, after each simulated docking, relaxation is carried out once to reduce unreasonable structures possibly generated in the rigid docking process and further simulate the slight deformation generated in the protein docking process.

3. The antibody design method targeting an epitope of interest according to claim 1, wherein, ​ ​ ​ ​ 4. The antibody design method targeting an epitope of interest according to claim 1, wherein, ​ 5. The antibody design method targeting an epitope of interest according to claim 1, wherein, ​ 6. The antibody design method targeting an epitope of interest according to Claim 1, wherein, ​ ​ ​ ​ ​ 7. The antibody design method targeting an epitope of interest according to Claim 1, wherein, ​ 8. The antibody design method targeting an epitope of interest of claim 1, wherein, After the two rounds of docking, the predicted models need to be further filtered, the main criteria are the buried surface area (dSASA) and the overall energy score, the dSASA needs to be greater than >1200 and the overall energy score needs to be lower than <-1000.

9. The antibody design method targeting an epitope of interest of claim 1, wherein, In step 5, the specific steps are as follows: the energy terms of the predicted antigen-antibody complex structure are weighted and summed, and the weighted sum of the energy of the antigen-antibody itself is subtracted, so as to calculate the approximate value of the binding energy of the two.

Images