AI藥物在近幾年發(fā)展迅速，其中，蛋白質(zhì)從頭設(shè)計無疑是AI設(shè)計藥物中的重頭戲與熱點。它不僅可以針對不可成藥靶點的藥物和自然中不存在的蛋白進行設(shè)計，而且大大縮短了藥物研發(fā)的周期。
華盛頓大學醫(yī)學院的David Baker教授是該領(lǐng)域的強者。當然，AI再牛，也需要用實驗手段去表征設(shè)計的蛋白。比如，使用Octet® 非標記分子互作系統(tǒng)檢測設(shè)計的蛋白與靶點的親和力。目前，David Baker課題組已使用Octet® 檢測蛋白親和力發(fā)表了近40篇文章，其中CNS文章就有近20篇！(文章列表見文末)
這里，陳老師就介紹一下他今年發(fā)表的幾篇CNS文章。
Nature
De novo design of protein structure and function with RF diffusion[1]
本文結(jié)合擴散模型（RF diffusion）和用深度學習算法，實現(xiàn)了從頭設(shè)計合成蛋白質(zhì)的目標，設(shè)計成功率高兩個數(shù)量級。針對已知的五個靶標，只需要不到100個候選即可達到nM級別的親和力（全部用BLI進行檢測）。同時，研究人員還設(shè)計了一種蛋白與其底物（流感血凝素）的復合物，并使用冷凍電鏡解析了其結(jié)構(gòu)。結(jié)果顯示，冷凍電鏡解析的結(jié)構(gòu)與設(shè)計的模型幾乎一模一樣，從而證明了該模型的準確性。為了進一步驗證生成的蛋白是否具有結(jié)合活性，研究人員繼續(xù)使用Octet®對這兩種蛋白的結(jié)合進行驗證。
圖1.(a-c) 從頭設(shè)計靶標蛋白的結(jié)合蛋白，針對五個靶標，從頭設(shè)計結(jié)合蛋白，BLI 響應信號值≥陽性對照1/2為候選;（b）RF diffusion成功率高出兩個數(shù)量級; (d) 親和力最高的結(jié)合物，結(jié)合KD為28nM; (e-h) 結(jié)構(gòu)學驗證
研究團隊表示，RF diffusion是對目前蛋白質(zhì)設(shè)計方法的一次綜合改進，能夠產(chǎn)生總長度達600個氨基酸殘基的結(jié)構(gòu)，其復雜性和準確度均比之前更高。研究團隊還表示，對該方法的進一步改進或能設(shè)計出復雜程度更高的全新蛋白質(zhì)。
Nature
De novo design of modular peptite-binding proteins by superhelical matching
針對內(nèi)在無序區(qū)的重復性蛋白和多肽，本文開發(fā)了一種從頭設(shè)計蛋白的通用性方法，能以大約 20% 的成功率高效設(shè)計出多肽結(jié)合蛋白。這些蛋白具備低至皮摩（pM）級別的高親和力、高特異性、高熱穩(wěn)定性以及高精度。而且還證實了該方法可以靶向更廣闊的非重復性多肽區(qū)域、以及可以延伸至人源蛋白的復雜網(wǎng)絡(luò)。
審稿人評價稱：“表征非常詳盡，親和力高達納摩爾至皮摩爾范圍，并且具有很高的特異性。設(shè)計與結(jié)構(gòu)的匹配也表明核心方法是合理的。”
這種高親和力、高特異性就是通過Octet® 來進行檢測的。
圖2. 將生物素化的目標肽加載到生物傳感器上，并和設(shè)計的binder進行結(jié)合解離，紅色矩形框表示相匹配的結(jié)合；可見，設(shè)計的binder與目標多肽的良好特異性
Science
Top-down design of protein architectures with reinforcement learning
該研究開發(fā)了一種基于強化學習的蛋白質(zhì)設(shè)計軟件，并證明了它有能力創(chuàng)造有功能的蛋白質(zhì)。這項工作用于設(shè)計高度穩(wěn)定和多功能的多聚蛋白籠組裝結(jié)構(gòu)和納米蛋白顆粒，開啟了蛋白質(zhì)設(shè)計的新時代。該技術(shù)對癌癥治療、再生醫(yī)學、強效疫苗和可生物降解日用品都有積極影響，文章同樣使用Octet® 檢測了設(shè)計的多聚蛋白顆粒與已知表位抗體的結(jié)合，以進一步驗證其結(jié)構(gòu)的正確性。
圖3. Octet® 親和力測定發(fā)現(xiàn)設(shè)計的單體蛋白聚合顆粒可以與識別不同表位抗體結(jié)合，表明單體蛋白在聚合顆粒上的構(gòu)象依舊完整。
Octet® 為什么如此受歡迎呢？
因為Octet® 用于親和力驗證的優(yōu)點在于
非標記Direct binding是趨勢，不需要標記和信號放大，可以更好的保持反應物的活性
快速測定親和力，提供結(jié)合速率常數(shù)和解離速率常數(shù)更加定量化地表征分子互作
無洗滌步驟，可測弱親和力（解離快）
寫入了美國藥典，文章多，認可度廣
萬金油技術(shù)，可以用與檢測DNA，小分子，蛋白質(zhì)等各種生物分子
操作簡便，耗材及維護成本低
盡管經(jīng)過不同的AI算法優(yōu)化，Binder設(shè)計的成功率已有所提高，但目前設(shè)計出高親和力的Binder仍然具有相當大的難度。對PPI（蛋白質(zhì)相互作用）的建模和理解仍然需要進行大量的工作。Octet® 可以真正實現(xiàn)高通量、快速的表征和檢測，這極大地加速了科研進程，實現(xiàn)預測和驗證的有機結(jié)合。
-參考文獻
[1] De novo design of protein structure and function with RFdiffusion
JL Watson, D Juergens, NR Bennett, BL Trippe, J Yim… - Nature, 2023 - nature.com
[2] De novo design of modular peptide-binding proteins by superhelical matching.nature,2023
[3] Top-down design of protein architectures with reinforcement learning.
SCIENCEVOL. 380, NO. 6642
[4] Massively parallel de novo protein design for targeted therapeutics
…, X Huang, R Jin, IA Wilson, DH Fuller, D Baker - Nature, 2017 - nature.com
[5] Optimization of affinity, specificity and function of designed influenza inhibitors using deep sequencing
…, H Kamisetty, P Blair, IA Wilson, D Baker - Nature …, 2012 - nature.com
[6] De novo design of picomolar SARS-CoV-2 miniprotein inhibitors
…, L Stewart, MS Diamond, D Veesler, D Baker - Science, 2020 - science.org
[7] De novo design of bioactive protein switches
…, JE Dueber, WRP Novak, H El-Samad, D Baker - Nature, 2019 - nature.com
[8] High-throughput characterization of protein–protein interactions by reprogramming yeast mating
D Younger, S Berger, D Baker… - Proceedings of the …, 2017 - National Acad Sciences
[9] A potent anti-malarial human monoclonal antibody targets circumsporozoite protein minor repeats and neutralizes sporozoites in the liver
…, R Vistein, C Barillas-Mury, R Amino, D Baker… - Immunity, 2020 - Elsevier
[10] Quadrivalent influenza nanoparticle vaccines induce broad protection
…, MC Crank, L Stewart, KK Lee, M Guttman, D Baker… - Nature, 2021 - nature.com
[11] Transferrin receptor targeting by de novo sheet extension
…, DE Ingber, J Abraham, D Baker - Proceedings of the …, 2021 - National Acad Sciences
[12] Receptor subtype discrimination using extensive shape complementary designed interfaces
…, CJ Kuo, KC Garcia, D Baker - Nature structural & …, 2019 - nature.com
[13] Ultrapotent miniproteins targeting the SARS-CoV-2 receptor-binding domain protect against infection and disease
…, R Ravichandran, L Carter, L Stewart, D Baker… - Cell Host & Microbe, 2021 - Elsevier
[14] Targeting HIV Env immunogens to B cell follicles in nonhuman primates through immune complex or protein nanoparticle formulations
…, G Alter, WR Schief, S Crotty, NP King, D Baker… - npj Vaccines, 2020 - nature.com
[15] De novo design of potent and selective mimics of IL-2 and IL-15
…, GJL Bernardes, M Dougan, KC Garcia, D Baker - Nature, 2019 - nature.com
[16] Computational design of proteins targeting the conserved stem region of influenza hemagglutinin
…, C Dreyfus, JE Corn, EM Strauch, IA Wilson, D Baker - Science, 2011 - science.org
[17] Reconfigurable asymmetric protein assemblies through implicit negative design
…, J Decarreau, HM Morris, A Kang, AK Bera, D Baker - Science, 2022 - science.org
[18] Structural and functional evaluation of de novo-designed, two-component nanoparticle carriers for HIV Env trimer immunogens
…, JP Moore, RW Sanders, NP King, D Baker… - PLoS …, 2020 - journals.plos.org
[19] Polyclonal antibody responses to HIV Env immunogens resolved using cryoEM
…, RF Rocha, ZT Berndsen, D Baker… - Nature …, 2021 - nature.com
[20] Induction of potent neutralizing antibody responses by a designed protein nanoparticle vaccine for respiratory syncytial virus
…, KK Lee, D Veesler, CE Correnti, LJ Stewart, D Baker… - Cell, 2019 - Elsevier
[21] Designed protein logic to target cells with precise combinations of surface antigens
…, A Nguyen, S Pun, CE Correnti, SR Riddell, D Baker - Science, 2020 - science.org
[22] De novo design of tyrosine and serine kinase-driven protein switches
…, VH Wysocki, H El-Samad, D Baker - Nature structural & …, 2021 - nature.com
[23] Computational design of trimeric influenza-neutralizing proteins targeting the hemagglutinin receptor binding site
…, AB Ward, P Yager, DH Fuller, IA Wilson, D Baker - Nature …, 2017 - nature.com
[24] Scaffolding protein functional sites using deep learning
…, F DiMaio, B Correia, S Ovchinnikov, D Baker - Science, 2022 - science.org
[25] M*lent designed proteins neutralize SARS-CoV-2 variants of concern and confer protection against infection in mice
…, BS Freedman, JD Bloom, H Ruohola-Baker… - Science translational …, 2022 - science.org
[26] A computationally designed hemagglutinin stem-binding protein provides in vivo protection from influenza independent of a host immune response
…, IA Wilson, A Dagley, DF Smee, D Baker… - PLoS …, 2016 - journals.plos.org
[27] Computational design of a synthetic PD-1 agonist
…, F DiMaio, KV Tarbell, D Baker - Proceedings of the …, 2021 - National Acad Sciences
[28] First critical repressive H3K27me3 marks in embryonic stem cells identified using designed protein inhibitor
…, SH Orkin, D Baker, H Ruohola-Baker - Proceedings of the …, 2017 - National Acad Sciences
[29] Designed proteins assemble antibodies into modular nanocages
…, H Ruohola-Baker, J Mathieu, D Veesler, D Baker - Science, 2021 - science.org
[30] Designed protein logic to target cells with precise combinations of surface antigens
…, A Nguyen, S Pun, CE Correnti, SR Riddell, D Baker - Science, 2020 - science.org
[31] De novo design of tyrosine and serine kinase-driven protein switches
…, VH Wysocki, H El-Samad, D Baker - Nature structural & …, 2021 - nature.com
[32] Computational design of trimeric influenza-neutralizing proteins targeting the hemagglutinin receptor binding site
…, AB Ward, P Yager, DH Fuller, IA Wilson, D Baker - Nature …, 2017 - nature.com
[33] Scaffolding protein functional sites using deep learning
…, F DiMaio, B Correia, S Ovchinnikov, D Baker - Science, 2022 - science.org
[34] M*lent designed proteins neutralize SARS-CoV-2 variants of concern and confer protection against infection in mice
…, BS Freedman, JD Bloom, H Ruohola-Baker… - Science translational …, 2022 - science.org
[35] A computationally designed hemagglutinin stem-binding protein provides in vivo protection from influenza independent of a host immune response
…, IA Wilson, A Dagley, DF Smee, D Baker… - PLoS …, 2016 - journals.plos.org
[36] Computational design of a synthetic PD-1 agonist
…, F DiMaio, KV Tarbell, D Baker - Proceedings of the …, 2021 - National Acad Sciences
[37] First critical repressive H3K27me3 marks in embryonic stem cells identified using designed protein inhibitor
…, SH Orkin, D Baker, H Ruohola-Baker - Proceedings of the …, 2017 - National Acad Sciences
[38] Designed proteins assemble antibodies into modular nanocages
…, H Ruohola-Baker, J Mathieu, D Veesler, D Baker - Science, 2021 - science.org
[39] Computational design of a pH-sensitive IgG binding protein
…, SJ Fleishman, D Baker - Proceedings of the …, 2014 - National Acad Sciences
[40] Design of protein-binding proteins from the target structure alone
…, S Bernard, L Stewart, IA Wilson, H Ruohola-Baker… - Nature, 2022 - nature.com
[41] De novo design of modular and tunable protein biosensors
…, J Wi, HJ Hong, L Stewart, BH Oh, D Baker - Nature, 2021 - nature.com
21
立即詢價
您提交后，專屬客服將第一時間為您服務