Docking domains from modular polyketide synthases and their use in engineeringDownload PDF Download PDF Matters ArisingOpen accessPublished: 22 July 2025Sabrina Collin ORCID: orcid.org/0000-0002-8565-82441 &Kira J. Weissman ORCID: orcid.org/0000-0002-3012-29601 Nature Communications volume 16, Article number: 6690 (2025) Cite this articleSubjectsMetabolic engineeringMultienzyme complexesNatural product synthesisSynthetic biologyarising from Y. Liu et al. Nature Communications https://doi.org/10.1038/s41467-025-55973-0 (2025)In a recent paper entitled “Improving polyketide biosynthesis by rescuing the translation of truncated mRNAs into functional polyketide synthase subunits”, Liu et al.1 discovered that >93% of the mRNAs resulting from transcription of gigantic modular polyketide synthase (PKS) genes are abridged, leading to reduced overall concentrations of complete assembly lines, and thus biosynthetic product. Although the authors’ strategy of splitting the genes into smaller fragments by introducing docking domains (DDs) did not reduce the number of truncated transcripts, it did increase the overall proportion and rate of translation of intact modules farther down the PKSs, resulting in a 5−13-fold boost in metabolite titers. However, applying this approach to additional PKS systems will depend on information that was almost entirely absent from the published manuscript. While our critique does not affect the central conclusions of the paper1, we provide a detailed analysis of the employed DDs that suggests that the engineering might have resulted in side products, and propose succinct guidelines for future docking domain-based manipulation of modular PKSs.Bacteria harbor gigantic molecular-scale assembly lines called modular polyketide synthases (PKSs)2 that build polyketide specialized metabolites. Each PKS multienzyme contains one or multiple modules, where each module is responsible for growing the polyketide chain by one building block and its chemical modification. As polyketides have found extensive use in human and veterinary medicine, there is substantial interest in enhancing their yields. In this context, Liu and colleagues’ finding that mRNAs encoding large PKS subunits are truncated is important, as it implies that smaller genes are better. Supporting this idea, inspection of all of the cis-acyltransferase (AT) PKSs in the MIBiG3 database – a curated collection of experimentally characterized systems – shows that 88% of the 462 catalogued subunits comprise 3 or fewer modules, and fully 72% have 1 or 2 (Fig. 1 and Supplementary Data 1).Fig. 1: Composition in modules of the cis-AT PKS subunits catalogued in the MIBiG database3.This analysis reveals a strong preference for polypeptides incorporating three or fewer modules. The effect on polyketide yields identified by Liu et al.1 may explain this natural bias.Full size imageTo separate the target PKSs into parts, Liu et al. replaced the short linker regions which covalently join modules together within subunits, with pairs of docking domains4— a strategy first reported in 20095, although this seminal work was not cited in the manuscript. DDs are protein-protein recognition motifs situated at the extreme C- and N-terminal ends of native PKS multienzymes. Via highly specific, non-covalent contacts, they ensure that the PKS polypeptides line up in the defined order that is critical to maintaining product fidelity. It is therefore not the case as stated by the authors that “…docking domains do not enhance the communication and activity of modular PKS”, as it has been shown by both deletion and swapping of these regions that they contribute to biosynthetic efficacy6.Based on several decades of research carried out in our and other laboratories, we now have a good understanding of both DD structure and function. Globally, this work has revealed that the cis-AT class of modular PKSs deploys at least five different types of DD, as judged by the three-dimensional structures of the docked complexes (referred to as 1a4, 1b7 and 28 (Fig. 2a−c), as well as 1-like9 and 2-like9). In each case, both the C-terminal DD (CDD) and the N-terminal DD (NDD) are composed of α-helices, but of variable number and amino acid composition, leading to overall DD assemblies of distinct topology (Fig. 2a−c). The high specificity of interaction within the DD types arises from strategically positioned electrostatic interactions that drive association4,7,8, as well as from shape complementarity, while the different DD types are intrinsically orthogonal to each other. Thus, in PKS systems, the same type of DD may operate at multiple intersubunit junctions, but equally, several types of DD (e.g. both 1a and 1b) may be present9.Fig. 2: Structures of complexes of three of the identified types of docking domains from cis-AT PKS systems.a Type 1a4. b Type 1b7. c Type 28. The C-terminal docking domains are indicated in shades of blue, and the N-terminal docking domains in shades of gold. For each docking complex, the residues comprising one of the interfaces are shown with the following color code: red: negatively charged; blue: positively charged; green: polar; white: hydrophobic. The primes denote residues contributed by the second monomer. (Note: the dashed lines indicate regions whose structures have not been characterized). The structures in this figure were produced using PYMOL15.Full size imageBased on this information, it is evident that the docking domains used to split PKS multienzymes into shorter proteins should not interact with those native to the system. In this way, the engineering will preserve the normal ordering of the subunits/modules. Conversely, if aberrant cross-talk occurs between the inserted and natural DDs, the subunits may assume non-native configurations. The result may be alternative products or stalled assembly lines, both of which would reduce yields of the target product. The manuscript by Li et al. lacks any discussion of the compatibility of the transplanted DDs with the recipient PKSs, and in particular, their potential to engage in mis-associations. We therefore carried out a comprehensive analysis of docking in the three PKSs investigated by the authors (butenyl-spinosyn (Bus), avermectin (AveA) and epothilone (Epo)), as well as of the three pairs of inserted DDs. For this, the sequences of the putative docking regions were compared to representatives of the 1a, 1b and 2 DD types9 (Supplementary Figs. 1−3).This analysis revealed that the butenyl-spinosyn PKS incorporates 3 type 1a DD pairs (at the BusB/BusC, BusC/BusD and BusD/BusE interfaces), and 1 set of type 1b DDs (BusA/BusB junction) (Supplementary Figs. 1, 2). The salinomycin (Sln) DDs (SlnA1 CDD/SlnA2 NDD and SlnA7 CDD/SlnA8 NDD) that were used to split the trimodular BusA into two or three fragments are both type 1a (Supplementary Fig. 1). This choice, although motivated by the absence of terminators in the respective sln genes, set the bar high in terms of specificity. Indeed, the basic mode of docking in the engineered PKSs is the same at four or five intersubunit junctions, meaning that partner choice is entirely dependent on a limited set of interface residues. More detailed inspection of the twelve charged and hydrophobic residues that make up the interface in each pair4 suggests that both SlnA1 CDD and SlnA7 CDD could be compatible with BusD NDD, and reciprocally, that the BusC CDD could interact productively with either SlnA1 NDD or SlnA8 NDD (Table 1). No attempt was made by the authors to search for side products potentially arising from these incorrect associations, and thus it cannot be excluded that the yield increases would have been even greater if alternative DD pairs were employed.Table 1 Comparison of the ensemble of interface residues in the systems that were engineered, relative to the introduced pairs of docking domainsaFull size tableIn the case of the avermectin system, two type 1a and 1 type 1b junctions are present (Supplementary Figs. 1, 2). The differences between the native AveA and the introduced type 1a Sln DD pairs are more substantial across the key residues (Table 1), but again, no direct evidence was provided that orthogonality, and thus the native sequence of modules, was maintained. Concerning the Epo system, it is a hybrid between four cis-AT PKS subunits (EpoA, C−F) and a single subunit of nonribosomal peptide synthetase (NRPS) (EpoB). The PKS/NRPS and NRPS/PKS docking elements are distinct from the five types typical of cis-AT PKSs (e.g. a SLIM/β-hairpin pair operates at the EpoA/EpoB interface10,11), while those at the remaining PKS/PKS interfaces are type 2 (Supplementary Fig. 3). In this case, the authors elected to use an additional pair of type 2 domains from the stigmatellin PKS (StiB CDD/StiC NDD, Supplementary Fig. 3) to split the tetramodular EpoD into two bimodular proteins. Again, the result of this choice was to increase to four the number of junctions in the resulting PKS dependent on the same type of docking domain, necessitating direct verification that improper interactions do not occur.The failure to address specificity issues is surprising, as it is already possible based on existing literature to define guidelines for docking domain-based engineering that ensure that orthogonality is preserved. These include: (i) inserting a type of DD which is not already present in the system; (ii) using a non-natural (e.g. synthetic) pair of DDs, such as the SYNZIPs that have been shown to function in engineered PKSs12; and (iii) replacing all sets of DDs within the engineered PKS with the native pairs of DDs from a system incorporating at least as many interfaces (e.g. the stambomycin PKS, which comprises 6 pairs of type 1a DDs and 2 pairs of type 1b DDs13). We do not recommend attempting to modify partner choice within DD types by site-directed mutagenesis, as it is not yet possible to reliably predict docking specificity. Ultimately, direct confirmation of the intrinsic orthogonality of all DD sets can be obtained via biophysical analysis of recombinant or synthetic DD peptides, as previously described9,13.MethodsAnalysis of subunit composition and docking domainsThe subunits used for the analysis of modular composition (Fig. 1 and Supplementary Data 1) were sourced from the MIBiG database (https://mibig.secondarymetabolites.org/)3. The sequences of representative type 1a, type 1b and type 2 CDDs and NDDs used in Fig. 2, Table 1, and Supplementary Figs. 1−3 were obtained from previously published analyses9,10. To identify the CDDs in the Bus, AveA and Epo systems which had not been analyzed earlier, the C-terminal boundaries of the upstream domains (ACP or peptidyl carrier protein (PCP)) were determined by multiple sequence alignment of the subunit ends using ClustalOmega (https://www.ebi.ac.uk/jdispatcher/msa/clustalo)14, with the remaining region considered as CDD. Similarly, the extent of the NDDs was determined by alignment of the first portion of the corresponding subunits, which allowed identification of the start sites of the KS domains downstream of the NDDs. Supplementary Figs. 1−3 were generated using ClustalOmega14.Reporting summaryFurther information on research design is available in the Nature Portfolio Reporting Summary linked to this article.Data availabilityData supporting the findings of this work are available within the paper and its Supplementary Information files. A reporting summary for this Article is available as a Supplementary Information file.ReferencesLiu, Y. et al. Improving polyketide biosynthesis by rescuing the translation of truncated mRNAs into functional polyketide synthase subunits. Nat. Commun. 16, 774 (2025).CAS PubMed PubMed Central Google Scholar Weissman, K. J. The structural biology of biosynthetic megaenzymes. Nat. Chem. Biol. 11, 660–670 (2015).CAS PubMed Google Scholar Zdouc, M. M. et al. MIBiG 4.0: advancing biosynthetic gene cluster curation through global collaboration. Nucleic Acids Res. 53, D678–D690 (2025).PubMed Google Scholar Broadhurst, R. W., Nietlispach, D., Wheatcroft, M. P., Leadlay, P. F. & Weissman, K. J. The structure of docking domains in modular polyketide synthases. Chem. Biol. 10, 723–731 (2003).CAS PubMed Google Scholar Yan, J., Gupta, S., Sherman, D. H. & Reynolds, K. A. Functional dissection of a multimodular polypeptide of the pikromycin polyketide synthase into monomodules by using a matched pair of heterologous docking domains. ChemBioChem 10, 1537–1543 (2009).CAS PubMed PubMed Central Google Scholar Tsuji, S. Y., Cane, D. E. & Khosla, C. Selective protein-protein interactions direct channeling of intermediates between polyketide synthase modules. Biochemistry 40, 2326–2331 (2001).CAS PubMed Google Scholar Buchholz, T. J. et al. Structural basis for binding specificity between subclasses of modular polyketide synthase docking domains. ACS Chem. Biol. 4, 41–52 (2009).CAS PubMed PubMed Central Google Scholar Whicher, J. R. et al. Cyanobacterial polyketide synthase docking domains: a tool for engineering natural product biosynthesis. Chem. Biol. 20, 1340–1351 (2013).CAS PubMed Google Scholar Risser, F. et al. Towards improved understanding of intersubunit interactions in modular polyketide biosynthesis: Docking in the enacyloxin IIa polyketide synthase. J. Struct. Biol. 212, 107581 (2020).CAS PubMed Google Scholar Richter, C. D., Nietlispach, D., Broadhurst, R. W. & Weissman, K. J. Multienzyme docking in hybrid megasynthetases. Nat. Chem. Biol. 4, 75–81 (2008).CAS PubMed Google Scholar Dowling, D. P. et al. Structural elements of an NRPS cyclization domain and its intermodule docking domain. Proc. Natl. Acad. Sci. USA 113, 12432–12437 (2016).ADS CAS PubMed PubMed Central Google Scholar Klaus, M., D’Souza, A. D., Nivina, A., Khosla, C. & Grininger, M. Engineering of chimeric polyketide synthases using SYNZIP docking domains. ACS Chem. Biol. 14, 426–433 (2019).CAS PubMed PubMed Central Google Scholar Su, L. et al. Engineering the stambomycin modular polyketide synthase yields 37-membered mini-stambomycins. Nat. Commun. 13, 515 (2022).ADS CAS PubMed PubMed Central Google Scholar Madeira, F. et al. The EMBL-EBI Job Dispatcher sequence analysis tools framework in 2024. Nucleic Acids Res. 52, W521–W525 (2024).PubMed PubMed Central Google Scholar PyMOL: The PyMOL molecular graphics system version 2.0. (Schrödinger, LLC., 2015).Download referencesAcknowledgementsThe authors acknowledge financial support from the Agence Nationale de la Recherche (BioSynKADH/ANR-24-CE92-0032-02 and PKSEng/ANR-24-CE93-0010-01 to K.J.W.) and the Lorraine Université d’Excellence (LUE).Author informationAuthors and AffiliationsUniversité de Lorraine, CNRS, IMoPA, F-54000, Nancy, FranceSabrina Collin & Kira J. WeissmanAuthorsSabrina CollinView author publicationsSearch author on:PubMed Google ScholarKira J. WeissmanView author publicationsSearch author on:PubMed Google ScholarContributionsS.C. created Fig. 2. K.J.W. conceived the study, wrote the manuscript and produced the remaining figures.Corresponding authorCorrespondence to Kira J. Weissman.Ethics declarationsCompeting interestsThe authors declare no competing interests.Peer reviewPeer review informationNature Communications thanks the anonymous reviewers for their contribution to the peer review of this work.Additional informationPublisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary informationSupplementary InformationDescription of Additional Supplementary FilesSupplementary Data 1Reporting SummaryRights and permissionsOpen Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.Reprints and permissionsAbout this articleDownload PDF