The architecture of polyketide synthases

Since the discovery of penicillin over a century ago, secondary metabolites from all
kingdoms of life have proven to be of high medical value. One class of proteins
prevalent in the production of secondary metabolites are polyketide synthases
(PKSs). Their polyketide products are complex organic compounds based on carbon
chains assembled from carboxylic acid precursors. Many polyketides are produced
by their hosts with the primary purpose of gaining an advantage in their ecological
niche. To contribute to such an advantage, a significant proportion of polyketides are
active against pro- and eukaryotic microorganisms. Type I PKSs are giant
multienzyme proteins employing an assembly line logic for the synthesis of the most
complex polyketides. They are composed of one or more functional and structural
modules, each capable of carrying out one step of precursor elongation during the
formation of an extended polyketide product.
In this thesis, I address two fundamental and open questions in the biosynthesis of
polyketides: First, what is the unique architecture underlying the assembly line logic
of multimodular PKS assembly lines; and second, how is atomic accuracy achieved
in cyclization and aromatic ring formation in the final step of PKS action.
The first aim is addressed in chapter two, which provides for the first time detailed
structural insights into the organization of type I PKS multimodules. This is achieved
by cryo-electron microscopic analysis of filamentous and non-filamentous forms of
K3DAK4, a bimodular trans-acyltransferase (AT) PKS fragment from Brevibacillus
brevis. Overall reconstructions are provided at an intermediate resolution of 7 Å, with
detailed insights into individual domains at sub-3Å resolution from cryo-electron
microscopy and X-ray crystallography. The bimodule core displays a vertical
stacking of its two modules along the central dimer axis of all three enzymatic
domains involved. Additionally, K3DAK4 oligomerizes into filaments horizontally via
small scaffolding domains in a trans-AT PKS-specific manner.
In chapter three the second aim is tackled, as I visualize an intermediate of the
enigmatic targeted cyclization and aromatic ring formation in the product template
domain (PT) of the aflatoxin-producing PksA at 2.7 Å resolution using X-ray
crystallography. To this end a substrate-analogue mimicking the transient
intermediate after the first of two cyclization steps facilitated by the enzyme is
covalently crosslinked to the active site. The positioning of the ligand relative to
previously known ligands representing the pre-and post-cyclization states indicate an
outward movement of the substrate throughout the process and a substantial effect
of progressing cyclization on the meticulous positioning of the intermediates.
The work provides detailed insights into core aspects of PKS biology from the
atomistic picture of guided product modification to the giant overall assembly line
architecture. In chapter four, both of these levels are put into context with current
advances in the analysis of modular structure and dynamics of PKSs, such as recent
structural models of cis-AT PKS modules and iterative PKSs. Furthermore, it
addresses currently open questions, such as the interaction of trans-AT PKS with
their cognate trans-acting enzymes. Altogether, the current progress in mechanistic
understanding of PKS systems makes systematic and structure-guided efforts to
unleash the full potential of PKS bioengineering ever more achievable.