Modular type I polyketide synthase acyl carrier protein domains share a common N-terminally extended fold

Luisa Moretto,Rachel Heylen,Steven Vance,R William Broadhurst,Natalie Holroyd

doi:10.1038/s41598-019-38747-9

Abstract

Acyl carrier protein (ACP) domains act as interaction hubs within modular polyketide synthase (PKS) systems, employing specific protein-protein interactions to present acyl substrates to a series of enzyme active sites. Many domains from the multimodular PKS that generates the toxin mycolactone display an unusually high degree of sequence similarity, implying that the few sites which vary may do so for functional reasons. When domain boundaries based on prior studies were used to prepare two isolated ACP segments from this system for studies of their interaction properties, one fragment adopted the expected tertiary structure, but the other failed to fold, despite sharing a sequence identity of 49%. Secondary structure prediction uncovered a previously undetected helical region (H0) that precedes the canonical helix-bundle ACP topology in both cases. This article reports the NMR solution structures of two N-terminally extended mycolactone mACP constructs, mH0ACPa and mH0ACPb, both of which possess an additional α-helix that behaves like a rigid component of the domain. The interactions of these species with a phosphopantetheinyl transferase and a ketoreductase domain are unaffected by the presence of H0, but a shorter construct that lacks the H0 region is shown to be substantially less thermostable than mH0ACPb. Bioinformatics analysis suggests that the extended H0-ACP motif is present in 98% of type I cis-acyltransferase PKS chain-extension modules. The polypeptide linker that connects an H0-ACP motif to the preceding domain must therefore be ~12 residues shorter than previously thought, imposing strict limits on ACP-mediated substrate delivery within and between PKS modules.

Highlights

The polyketide family of natural products is remarkable for both its structural diversity and biological activity, containing clinically important molecules such as macrolide, polyene and polyether antibiotics[1]
In Mycobacterium ulcerans, the toxin mycolactone is produced by a modular type I cis-AT polyketide synthase (PKS) that comprises 16 chain-extension modules spread across three polypeptide chains, MLSA1, MLSA2 and MLSB (Fig. 1)[7]
We have used nuclear magnetic resonance (NMR) spectroscopy to show that the Ppant arm swings freely around its attachment point when loaded with short, polar substrate mimics, whereas longer, more saturated chains can adhere to the surface of the Acyl carrier protein (ACP) domain without significantly altering its structure[10]

Summary

Introduction

The polyketide family of natural products is remarkable for both its structural diversity and biological activity, containing clinically important molecules such as macrolide, polyene and polyether antibiotics[1]. Polyketides are constructed from simple carboxylic acid-derived components, often by filamentous bacteria using modular type I polyketide synthases (PKSs), gigantic protein complexes that house multiple modules of covalently-linked catalytic domains[2] Inside these systems, acyl carrier protein (ACP) domains act as key interaction hubs, shuttling substrate chains between the active sites in one module before handing their cargo over to the next. In Mycobacterium ulcerans, the toxin mycolactone is produced by a modular type I cis-AT PKS that comprises 16 chain-extension modules spread across three polypeptide chains, MLSA1, MLSA2 and MLSB (Fig. 1)[7] This system employs two varieties of KR domain with characteristic amino acid sequence signatures, termed A1- and B1-type, which use an NADPH cofactor to generate substrate β-hydroxyl groups with 3 S and 3 R stereochemistry, respectively, but which leave the 2 R orientations of α-methyl substituents unchanged[8]. Sequence analysis indicates that the extended H0-ACP unit is a common feature in type I cis-AT PKS systems

Methods

Results

Conclusion