Generative AI designs functional thiolation domains for reprogramming non-ribosomal peptide synthetases

preprint OA: closed
Full text JSON View at publisher

Abstract

Abstract Large language models and generative protein design promise to accelerate biotechnology, but it remains unclear whether they can engineer dynamic megasynth(et)ases whose activity depends on transient, context-specific domain interfaces. Non-ribosomal peptide synthetases (NRPSs) exemplify this challenge and produce many clinically used therapeutics. Here we integrate pretrained generative models (ESM3, ProteinMPNN and EvoDiff) with design–build–test–learn cycles and data-guided prioritization to generate 76 de novo thiolation (T) domains. We built and tested 578 recombinant NRPS variants in vivo spanning minimal, full-length and hybrid assembly lines. AI-designed T-domains supported product formation across architectures, enabled catalytically active hybrids at recombined junctions and increased yields by up to ~3-fold relative to NRPSs carrying the native T-domain. A representative design showed improved soluble expression, refolding, and a 12 °C higher melting temperature, while molecular dynamics simulations indicated preserved global stability but reshaped, state-dependent interdomain contact networks. Together, these results establish generative design as an effective route to context-conditioned optimization and reprogramming of biosynthetic assembly lines.
Full text 16,821 characters · extracted from preprint-html · click to expand
Generative AI designs functional thiolation domains for reprogramming non-ribosomal peptide synthetases | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Generative AI designs functional thiolation domains for reprogramming non-ribosomal peptide synthetases Kenan Bozhüyük, Emre Bülbül, Seounggun Bang, Kevin George, Gabriele Bianchi, and 12 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9138214/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Large language models and generative protein design promise to accelerate biotechnology, but it remains unclear whether they can engineer dynamic megasynth(et)ases whose activity depends on transient, context-specific domain interfaces. Non-ribosomal peptide synthetases (NRPSs) exemplify this challenge and produce many clinically used therapeutics. Here we integrate pretrained generative models (ESM3, ProteinMPNN and EvoDiff) with design–build–test–learn cycles and data-guided prioritization to generate 76 de novo thiolation (T) domains. We built and tested 578 recombinant NRPS variants in vivo spanning minimal, full-length and hybrid assembly lines. AI-designed T-domains supported product formation across architectures, enabled catalytically active hybrids at recombined junctions and increased yields by up to ~3-fold relative to NRPSs carrying the native T-domain. A representative design showed improved soluble expression, refolding, and a 12 °C higher melting temperature, while molecular dynamics simulations indicated preserved global stability but reshaped, state-dependent interdomain contact networks. Together, these results establish generative design as an effective route to context-conditioned optimization and reprogramming of biosynthetic assembly lines. Biological sciences/Biotechnology/Molecular engineering/Synthetic biology Biological sciences/Computational biology and bioinformatics/Protein design Full Text Additional Declarations Yes there is potential Competing Interest. K.A.J.B. is a co-founder and CSO of Myria Biosciences AG. S.S. is a co-founder and CEO of Myria Biosciences AG. H.A.M. is employed by ETH and Myria Biosciences AG. E.F.B. has a consulting agreement with Myria Biosciences AG. R.H. is employed by ETH and works on an Innosuisse collaboration project between ETH and Myria Biosciences AG. The remaining authors declare no competing interests. Supplementary Files Supplementarydata10303.xlsx Supplementaty Data 1 SupplementaryInfo1603NBT.pdf Supplementary Information Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9138214","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":608120206,"identity":"083f1f64-9170-451f-b088-1ad38700b512","order_by":0,"name":"Kenan Bozhüyük","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0002-8609-2967","institution":"Helmholtz Institute for Pharmaceutical Research Saarland (HIPS)","correspondingAuthor":true,"prefix":"","firstName":"Kenan","middleName":"","lastName":"Bozhüyük","suffix":""},{"id":608120207,"identity":"c555b64e-fec0-42b3-aa69-87175c73776c","order_by":1,"name":"Emre Bülbül","email":"","orcid":"","institution":"Helmholtz Institute for Pharmaceutical Research Saarland (HIPS)","correspondingAuthor":false,"prefix":"","firstName":"Emre","middleName":"","lastName":"Bülbül","suffix":""},{"id":608120208,"identity":"75e212ac-8970-4f8f-9df5-e23b01d98e70","order_by":2,"name":"Seounggun Bang","email":"","orcid":"https://orcid.org/0009-0001-2964-3259","institution":"Helmholtz Institute for Pharmaceutical Research Saarland (HIPS)","correspondingAuthor":false,"prefix":"","firstName":"Seounggun","middleName":"","lastName":"Bang","suffix":""},{"id":608120209,"identity":"fb3c4b8f-bfe4-4fc8-b025-0375166649e9","order_by":3,"name":"Kevin George","email":"","orcid":"","institution":"Center for Bioinformatics","correspondingAuthor":false,"prefix":"","firstName":"Kevin","middleName":"","lastName":"George","suffix":""},{"id":608120210,"identity":"039d6c36-3d8a-4215-b1d2-ad0e1f6aa923","order_by":4,"name":"Gabriele Bianchi","email":"","orcid":"","institution":"Helmholtz Institute for Pharmaceutical Research Saarland (HIPS)","correspondingAuthor":false,"prefix":"","firstName":"Gabriele","middleName":"","lastName":"Bianchi","suffix":""},{"id":608120211,"identity":"f01578b7-6f68-456b-85df-62dbc6a25430","order_by":5,"name":"Prateek Raj","email":"","orcid":"","institution":"Helmholtz-Zentrum für Infektionsforschung GmbH","correspondingAuthor":false,"prefix":"","firstName":"Prateek","middleName":"","lastName":"Raj","suffix":""},{"id":608120212,"identity":"62316667-5134-4574-8031-7bfc88c45ace","order_by":6,"name":"Seonyong Chung","email":"","orcid":"","institution":"Helmholtz Institute for Pharmaceutical Research Saarland (HIPS)","correspondingAuthor":false,"prefix":"","firstName":"Seonyong","middleName":"","lastName":"Chung","suffix":""},{"id":608120213,"identity":"1990314c-4630-4765-bac4-9bf063d0bd88","order_by":7,"name":"Vincent Pauline","email":"","orcid":"","institution":"Technical University of Munich","correspondingAuthor":false,"prefix":"","firstName":"Vincent","middleName":"","lastName":"Pauline","suffix":""},{"id":608120214,"identity":"5fd3e723-d862-445d-afd7-54721c820c9b","order_by":8,"name":"Ramon Hochstrasser","email":"","orcid":"","institution":"Eidgenössische Technische Hochschule (ETH) Zürich","correspondingAuthor":false,"prefix":"","firstName":"Ramon","middleName":"","lastName":"Hochstrasser","suffix":""},{"id":608120215,"identity":"24e1182d-b9f5-4d65-a59c-b57c9283623f","order_by":9,"name":"Hannah Minas","email":"","orcid":"https://orcid.org/0000-0003-2997-6915","institution":"Myria Biosciences AG","correspondingAuthor":false,"prefix":"","firstName":"Hannah","middleName":"","lastName":"Minas","suffix":""},{"id":608120216,"identity":"dbbde010-8231-41aa-8a56-58c4c5a1fedb","order_by":10,"name":"Walid Elgaher","email":"","orcid":"https://orcid.org/0000-0002-8766-4568","institution":"HIPS","correspondingAuthor":false,"prefix":"","firstName":"Walid","middleName":"","lastName":"Elgaher","suffix":""},{"id":608120217,"identity":"2f438e39-bd97-416b-b813-55d41154ddf6","order_by":11,"name":"Andreas Kany","email":"","orcid":"","institution":"HIPS","correspondingAuthor":false,"prefix":"","firstName":"Andreas","middleName":"","lastName":"Kany","suffix":""},{"id":608120218,"identity":"7979172c-f443-48a1-8c2b-334bd09c18c5","order_by":12,"name":"Anna Hirsch","email":"","orcid":"https://orcid.org/0000-0001-8734-4663","institution":"HIPS","correspondingAuthor":false,"prefix":"","firstName":"Anna","middleName":"","lastName":"Hirsch","suffix":""},{"id":608120219,"identity":"c41e8e9c-3900-493e-94cf-40044be6229c","order_by":13,"name":"Steven Schmitt","email":"","orcid":"","institution":"Myria Biosciences AG","correspondingAuthor":false,"prefix":"","firstName":"Steven","middleName":"","lastName":"Schmitt","suffix":""},{"id":608120220,"identity":"f82ac759-8564-45a2-8fca-3c47fb6239d7","order_by":14,"name":"Dirk Heinz","email":"","orcid":"","institution":"HZI","correspondingAuthor":false,"prefix":"","firstName":"Dirk","middleName":"","lastName":"Heinz","suffix":""},{"id":608120221,"identity":"bb12032c-10fc-40d7-985f-6e7d05ce5c86","order_by":15,"name":"Olga Kalinina","email":"","orcid":"https://orcid.org/0000-0002-9445-477X","institution":"Helmholtz Institute for Pharmaсeutical Research Saarland (HIPS) / Helmholtz Centre for Infection Research (HZI)","correspondingAuthor":false,"prefix":"","firstName":"Olga","middleName":"","lastName":"Kalinina","suffix":""},{"id":608120222,"identity":"131bf01f-615d-4585-8cdb-0ee0b4d90d17","order_by":16,"name":"Dietrich Klakow","email":"","orcid":"","institution":"Saarland University","correspondingAuthor":false,"prefix":"","firstName":"Dietrich","middleName":"","lastName":"Klakow","suffix":""}],"badges":[],"createdAt":"2026-03-16 13:00:51","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9138214/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9138214/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105565057,"identity":"926b5742-e66a-47c7-b6b1-48d88d15c2b1","added_by":"auto","created_at":"2026-03-27 12:51:44","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3192801,"visible":true,"origin":"","legend":"Article File","description":"","filename":"MANUSCNBTv2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9138214/v1_covered_2692f0d4-5c61-453d-be96-bc08e52807f3.pdf"},{"id":105353469,"identity":"d80b9384-c196-4c45-a87e-f7238adfbb0b","added_by":"auto","created_at":"2026-03-25 06:21:06","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":161721,"visible":true,"origin":"","legend":"Supplementaty Data 1","description":"","filename":"Supplementarydata10303.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9138214/v1/0762b9babfa0930a4ffb08e9.xlsx"},{"id":105353470,"identity":"379a72d9-853d-4aa7-babf-6d3ad403acec","added_by":"auto","created_at":"2026-03-25 06:21:06","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":11214808,"visible":true,"origin":"","legend":"Supplementary Information","description":"","filename":"SupplementaryInfo1603NBT.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9138214/v1/872ac636d872e867d13d6b56.pdf"}],"financialInterests":"\u003cb\u003eYes\u003c/b\u003e there is potential Competing Interest.\nK.A.J.B. is a co-founder and CSO of Myria Biosciences AG. S.S. is a co-founder and CEO of Myria Biosciences AG. H.A.M. is employed by ETH and Myria Biosciences AG. E.F.B. has a consulting agreement with Myria Biosciences AG. R.H. is employed by ETH and works on an Innosuisse collaboration project between ETH and Myria Biosciences AG. The remaining authors declare no competing interests.","formattedTitle":"Generative AI designs functional thiolation domains for reprogramming non-ribosomal peptide synthetases","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-9138214/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9138214/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Large language models and generative protein design promise to accelerate biotechnology, but it remains unclear whether they can engineer dynamic megasynth(et)ases whose activity depends on transient, context-specific domain interfaces. Non-ribosomal peptide synthetases (NRPSs) exemplify this challenge and produce many clinically used therapeutics. Here we integrate pretrained generative models (ESM3, ProteinMPNN and EvoDiff) with design–build–test–learn cycles and data-guided prioritization to generate 76 de novo thiolation (T) domains. We built and tested 578 recombinant NRPS variants in vivo spanning minimal, full-length and hybrid assembly lines. AI-designed T-domains supported product formation across architectures, enabled catalytically active hybrids at recombined junctions and increased yields by up to ~3-fold relative to NRPSs carrying the native T-domain. A representative design showed improved soluble expression, refolding, and a 12 °C higher melting temperature, while molecular dynamics simulations indicated preserved global stability but reshaped, state-dependent interdomain contact networks. Together, these results establish generative design as an effective route to context-conditioned optimization and reprogramming of biosynthetic assembly lines.","manuscriptTitle":"Generative AI designs functional thiolation domains for reprogramming non-ribosomal peptide synthetases","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-25 06:21:02","doi":"10.21203/rs.3.rs-9138214/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"d70c50a6-faa4-43d4-82d9-49be273d78e4","owner":[],"postedDate":"March 25th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":64708836,"name":"Biological sciences/Biotechnology/Molecular engineering/Synthetic biology"},{"id":64708837,"name":"Biological sciences/Computational biology and bioinformatics/Protein design"}],"tags":[],"updatedAt":"2026-03-25T06:21:02+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-25 06:21:02","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9138214","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9138214","identity":"rs-9138214","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}

Text is read by the "Ask this paper" AI Q&A widget below. Extraction quality varies by source — PMC NXML preserves structure cleanly, OA-HTML may include some navigation residue, and OA-PDF can have broken hyphenation. The publisher copy (via DOI) is the canonical version.

My notes (saved in your browser only)

Ask this paper AI returns verbatim quotes from the full text · source: preprint-html

Answers must be backed by verbatim quotes from this paper's full text. Hallucinated quotes are dropped automatically; if no verbatim passage answers the question, we say so. How this works

Citation neighborhood (no data yet)

We don't have any in-corpus citations linked to this paper yet. This is a recent paper (2026) — citers typically take a year or two to land, and the OpenAlex reference graph may still be filling in.

Source provenance

europepmc
last seen: 2026-05-20T01:45:00.602351+00:00