Characterisation of Modular Polyketide Synthases Designed to Make Pentaene Analogues of Amphotericin B

Abstract

Glycosylated polyene macrolides are important antifungal agents that are produced by many actinomycete species. Development of new polyenes may deliver improved antibiotics. Here, Streptomyces nodosus was genetically re-programmed to synthesise pentaene analogues of the heptaene amphotericin B. These pentaenes are of interest as surrogate substrates for enzymes catalysing unusual, late-stage biosynthetic modifications. The previous deletion of amphotericin polyketide synthase modules 5 and 6 generated S. nodosus M57, which produces an inactive pentaene. Here, the chain-terminating thioesterase was fused to module 16 to generate strain M57-16TE, in which cycles 5, 6, 17 and 18 are eliminated from the biosynthetic pathway. Another variant of M57 was obtained by replacing modules 15, 16 and 17 with a single 15-17 hybrid module. This gave strain M57-1517, in which cycles 5, 6, 15 and 16 are deleted. M57-16TE and M57-1517 gave reduced pentaene yields. Only M57-1517 delivered its predicted full-length pentaene macrolactone in low amounts. For both mutants, the major pentaenes were intermediates released from modules 10, 11 and 12. Longer pentaene chains were unstable. The novel pentaenes were not glycosylated and were not active against Candida albicans. However, random mutagenesis and screening may yet deliver new antifungal producers from the M57-16TE and M57-1517 strains.

Keywords: amphotericin B; antifungal antibiotics; glycosylated polyene macrolides; modular polyketide synthase; synthetic biology.

Figure 1

Structures of amphotericin B and pentaenes.

Figure 2

Structures of pentaene macrolactones from engineered S. nodosus strains.

Figure 3

Modular structure of the PKS and intermediates in S. nodosus M57-16TE. The tetramodular AmphC* protein replaces the native hexamodular AmphC, and the AmphJ-M16TE protein replaces the trimodular AmphJ. Inactive domains are shaded grey. Note that 5/7 is a hybrid module (KS5-AT7-DH7-KR7-ACP7) that replaces modules 5, 6 and 7 of the full-length PKS assembly line. Numbering of other modules is as for the native amphotericin PKS (Figure S1).

Figure 4

Modular structure of the PKS and intermediates in S. nodosus M57-1517. The tetramodular AmphC* protein replaces the native hexamodular AmphC and the unimodular AmphJ* protein replaces the trimodular AmphJ. Inactive domains are shaded grey. Note that 5/7 is a hybrid module (KS5-AT7-DH7-KR7-ACP7) that replaces modules 5, 6 and 7 of the full-length PKS assembly line. 15/17 is a hybrid module (KS15-AT17-DH17-KR17-ACP17) that replaces modules 15, 16 and 17. Numbering of other modules is as for the native amphotericin PKS (Figure S1).

Figure 5

HPLC analysis of pentaenes from (A) S. nodosus M57, (B) S. nodosus M57-16TE, (C) S. nodosus M57-1517.

Figure 6

HPLC analysis of highly purified pentaene from S. nodosus M57-1517. A plot of A₃₅₂ against time is shown. The diode array display of this chromatogram indicated that the pentaene was free of contaminants that absorb in the UV-visible region (Figure S12).

Figure 7

Structures of pentaene polyketide intermediates and full-length 57-1517 macrolactone. The products of extension modules 10, 11 and 12 (see Figure 3) are shown as free acids. The calculated molecular masses can be fitted to ions detected in the purified pentaene sample (see Figure S15).

Figure 8

Polyenes predicted from S. nodosus NM-16TE. (A) heptaene; (B) tetraene.