Discrete Acyltransferases and Thioesterases in Iso-Migrastatin and Lactimidomycin Biosynthesis

Abstract

Iso-Migrastatin (iso-MGS) and lactimidomycin (LTM) are glutarimide-containing polyketide natural products (NPs) that are biosynthesized by homologous acyltransferase (AT)-less type I polyketide synthase (PKS) assembly lines. The biological activities of iso-MGS and LTM have inspired numerous efforts to generate analogues via genetic manipulation of their biosynthetic machinery in both native producers and model heterologous hosts. A detailed understanding of the MGS and LTM AT-less type I PKSs would serve to inspire future engineering efforts while advancing the fundamental knowledge of AT-less type I PKS enzymology. The mgs and ltm biosynthetic gene clusters (BGCs) encode for two discrete ATs of the architecture AT-enoylreductase (AT-ER) and AT-type II thioesterase (AT-TE). Herein, we report the functional characterization of the mgsB and ltmB and the mgsH and ltmH gene products, revealing that MgsB and LtmB function as type II thioesterases (TEs) and MgsH and LtmH are the dedicated trans-ATs for the MGS and LTM AT-less type I PKSs. In vivo and in vitro experiments demonstrated that MgsB was devoid of any AT activity, despite the presence of the conserved catalytic triad of canonical ATs. Cross-complementation experiments demonstrated that MgsH and LtmH are functionally interchangeable between the MGS and LTM AT-less type I PKSs. This work sets the stage for future mechanistic studies of AT-less type I PKSs and efforts to engineer the MGS and LTM AT-less type I PKS assembly lines for novel glutarimide-containing polyketides.

Figure 1.

Overview of NPs, BGCs, gene products, and enzymology relevant to this study. (A) Structures of 1, 2, and 3, the latter of which is a nonenzymatic rearrangement product of 1. The structure of 4 is included for comparison. (B) Post-PKS steps in the biosynthetic pathways of 1 and 2 and the associated mgs and ltm BGCs, highlighting the similarity of their genetic organization and associated chemistry. The genes of interest for this study from each BGC are highlighted. (C) Reactions catalyzed by ATs and type II TEs. For ATs, both half reactions of self-malonylation and ACP S-acylation are shown. The hydrolysis reaction catalyzed by type II TEs plays an editing role in the event of mispriming (path a) or aberrant decarboxylation (path b). (D) Domain architectures of the trans-ATs from characterized AT-less type I PKS-containing BGCs. The accession numbers of the protein sequences are given in Table S4. ACP, acyl carrier protein; AT, acyltransferase; DC, decarboxylase; ER, enoyl reductase; SAM, S-adenosyl methionine; TE, thioesterase; TEII, type II thioesterase.

Figure 2.

Gene inactivation, complementation, active site mutagenesis, and cross-complementation experiments involving mgsB, mgsH, ltmB, and ltmH. (A) HPLC analysis of the ΔmgsH mutant and varying complementation strains: (I) 1 standard, (II) 6 standard, (III) SB11001 (mgsH positive control), (IV) SB11028 (ΔmgsH), (V) SB11031 (ΔmgsH/mgsH, integrative), (VI) SB11029 (ΔmgsH/mgsH, replicative), (VII) SB11032 (ΔmgsH/ltmH, integrative), and (VIII) SB11030 (ΔmgsH/ltmH, replicative). (B) HPLC analysis of the ΔmgsB mutant and varying complementation strains: (I) 1 standard, (II) SB11001 (mgsB positive control), (III) SB11027 (ΔmgsB), (IV) SB11033 (ΔmgsB/mgsB), (V) SB11034 (ΔmgsB/mgsB-AT*, AT active site Ser84Ala mutation), and (VI) SB11035 (ΔmgsB/mgsB-TE*, TE active site Ser374Ala mutation). (C) HPLC analysis of the ΔltmB, ΔltmH, and varying complementation strains: (I) 2 standard, (II) S. amphibiosporus wild-type (ltmB and ltmH, positive control), (III) SB15010 (ΔltmB, y-axis is 2× magnification due to the low titer of 2), (IV) SB15011 (ΔltmB/ltmB), (V) SB15005 (ΔltmH), (VI) SB15012 (ΔltmH/ltmH), and (VII) SB15013 (ΔltmH/mgsH). (◊) 1; (∇) 2; (◆) 6.

Figure 3.

In vitro assays demonstrating that MgsH functions as an AT, while MgsB cannot catalyze self-malonylation. (A) Self-malonylation reaction catalyzed by MgsH but not MgsB. (B) Incubation of MgsB and MgsH with [2-¹⁴C]malonyl-CoA as visualized on a 100% SDS-PAGE gel with Coomassie Blue staining (I) and phosphor imaging (II). Lane 1, MgsB with [2-¹⁴C]malonyl-CoA; lane 2, MgsH with [2-¹⁴C]malonyl-CoA. (C) MgsH-catalyzed acylation of MgsG-ACP11 with [2-¹⁴C]malonyl-CoA as a substrate. (D) Incubation of eight mgs ACPs with MgsH and [2-¹⁴C]malonyl-CoA as visualized on a 4–20% SDS-PAGE gel with Coomassie Blue staining (panel I) and phosphor imaging (panel II). Lane 1, MgsC; lane 2, MgsE-ACP3; lane 3, MgsF-ACP5; lane 4, MgsF-ACP7; lane 5, MgsF-ACP8; lane 6, MgsF-ACP9; lane 7, MgsG-ACP10; lane 8, MgsG-ACP11. MgsG-ACP11: 12.4 kDa; Svp PPTase: 26.8 kDa; MgsB: 61.0 kDa; MgsH: 84.8 kDa. See Figure S4 for the full set of in vitro data.

Figure 4.

In vitro assays of MgsB-catalyzed type II TE activity using [1-¹⁴C]acetyl-S-MgsG-ACP11 (A, B, C) and [1-¹⁴C]propionyl-S-MgsG-ACP11 (D–F). MgsB was incubated with the indicated substrates as visualized on 4–20% SDS-PAGE with Coomassie blue staining (B, E) and phosphor imaging (C, F). (A) Hydrolysis reaction of [1-¹⁴C]acetyl-S-MgsG-ACP11, and visualization on SDS-PAGE (B) Lane M, molecular weight standards with the numbers to the left given in kDa; Lanes 1–5, incubation of [1-¹⁴C]acetyl-S-MgsG-ACP11 in the absence of MgsB for 0, 7.5, 15, 30, and 60 min, respectively; Lanes 6–9, incubation of [1-¹⁴C]acetyl-S-MgsB-ACP11 in the presence of MgsB for 7.5, 15, 30, and 60 min, respectively. (C) Lanes 1–9 same as in (B) except products visualized by phosphor imaging. (D) Hydrolysis reaction of [1-¹⁴C]propionyl-S-MgsG-ACP11 and visualization on SDS-PAGE (E) the same as (B) except for use of [1-¹⁴C]propionyl-S-MgsG-ACP11 instead of [1-¹⁴C]acetyl-S-MgsG-ACP11. (F) Lanes identical to (E) except products visualized by phosphor imaging. MgsG-ACP11:12.4 kDa; Svp PPTase: 26.8 kDa; MgsB: 61.0 kDa.

Figure 5.

SSN of trans-ATs from AT-less type I PKSs from characterized BGCs and the UniProt database. The sequences were curated as described in the Materials and Methods section, and an alignment score cutoff of 101 was used. The node shapes by default were depicted as ellipses, the characterized ATs shown in Figure 1 and Table S4 as triangles, and the Mgs/Ltm proteins as squares. The AT domain architectures were assigned by Interpro, and nodes were colored according to the domain architectures as shown below the SSN. Two SSNs from the same analysis with two other alignment scores are depicted in Figure S7. The yellow TE-AT-AT and TE-AT nodes were the basis of the BGC analysis shown in Figure S8.