Numerous natural products of clinical value are biosynthesized by polyketide synthases (PKSs) and nonribosomal peptide synthetases (NRPSs), which are multienzymes comprising modules of catalytic domains. The key players in each module are carrier proteins, which serve as attachment points for the growing substrate chains. Thus, the details of carrier protein-based substrate delivery to each active site are central to understanding chain assembly in these systems. In the enterobactin NRPS, communication between a peptidyl carrier protein (PCP) and the adjacent thioesterase (TE) domain occurs through formation of a compact complex. Using NMR, we show that the corresponding interaction between a PKS acyl carrier protein (ACP) and its downstream TE is fundamentally different: chain transfer occurs in the absence of a protein-protein interface, with contact limited to the substrate acyl terminus.

Polyketide natural products such as erythromycin A and epothilone are assembled on multienzyme polyketide synthases (PKSs), which consist of modular sets of protein domains. Within these type I systems, the fidelity of biosynthesis depends on the programmed interaction among the multiple domains within each module, centered around the acyl carrier protein (ACP). A detailed understanding of interdomain communication will therefore be vital for attempts to reprogram these pathways by genetic engineering. We report here that the interaction between a representative ACP domain and its downstream thioesterase (TE) is mediated largely by covalent tethering through a short "linker" region, with only a minor energetic contribution from protein-protein molecular recognition. This finding helps explain in part the empirical observation that TE domains can function out of their normal context in engineered assembly lines, and supports the view that overall PKS architecture may dictate at least a subset of interdomain interactions.

Equilibrative nucleoside transporters (ENTs) are integral membrane proteins, which reside in plasma membranes of all eukaryotic cells and mediate thermodynamically downhill transport of nucleosides. This process is essential for nucleoside recycling, and also plays a key role in terminating adenosine-mediated cellular signaling. Furthermore, ENTs mediate the uptake of many drugs, including anticancer and antiviral nucleoside analogues. The structure and mechanism, by which ENTs catalyze trans-membrane transport of their substrates, remain unknown. To identify the core of the transporter needed for stability, activity, and for its correct trafficking to the plasma membrane, we have expressed human ENT deletion mutants in Xenopus laevis oocytes and determined their localization, transport properties and susceptibility to inhibition. We found that the carboxyl terminal trans-membrane segments are essential for correct protein folding and trafficking. In contrast, the soluble extracellular and intracellular loops appear to be dispensable, and must be involved in the fine-tuning of transport regulation.

Domains within the multienzyme polyketide synthases are linked by noncatalytic sequences of variable length and unknown function. Recently, the crystal structure was reported of a portion of the linker between the acyltransferase (AT) and ketoreductase (KR) domains from module 1 of the erythromycin synthase (6‐deoxyerythronolide B synthase), as a pseudodimer with the adjacent ketoreductase (KR). On the basis of this structure, the homologous linker region between the dehydratase (DH) and enoyl reductase (ER) domains in fully reducing modules has been proposed to occupy a position on the periphery of the polyketide synthases complex, as in porcine fatty acid synthase. We report here the expression and characterization of the same region of the 6‐deoxyerythronolide B synthase module 1 AT‐KR linker, without the adjacent KR domain (termed ΔN AT1‐KR1), as well as the corresponding section of the DH‐ER linker. The linkers fold autonomously and are well structured. However, analytical gel filtration and ultracentrifugation analysis independently show that ΔN AT1‐KR1 is homodimeric in solution; site‐directed mutagenesis further demonstrates that linker self‐association is compatible with the formation of a linker‐KR pseudodimer. Our data also strongly indicate that the DH‐ER linker associates with the upstream DH domain. Both of these findings are incompatible with the proposed model for polyketide synthase architecture, suggesting that it is premature to allocate the linker regions to a position in the multienzymes based on the solved structure of animal fatty acid synthase.

No abstract