Synthetic Enzymology and the Fountain of Youth: Repurposing Biology for Longevity

Abstract

Caloric restriction (CR) is an intervention that can increase maximal lifespan in organisms, but its application to humans remains challenging. A more feasible approach to achieve lifespan extension is to develop CR mimetics that target biochemical pathways affected by CR. Recent studies in the engineering and structural characterization of polyketide synthases (PKSs) have facilitated their use as biocatalysts to produce novel polyketides. Here, we show that by establishing a combinatorial biosynthetic route in Escherichia coli and exploring the substrate promiscuity of a mutant PKS from alfalfa, 413 potential anti-ageing polyketides were biosynthesized. In this approach, novel acyl-coenzyme A (CoA) precursors generated by promiscuous acid-CoA ligases were utilized by PKS to generate polyketides which were then fed to Caenorhabditis elegans to study their potential efficacy in lifespan extension. It was found that CR mimetics like resveratrol can counter the age-associated decline in mitochondrial function and increase the lifespan of C. elegans. Using the mitochondrial respiration profile of C. elegans supplemented for 8 days with 50 μM resveratrol as a blueprint, we can screen our novel polyketides for potential CR mimetics with improved potency. This study highlights the utility of synthetic enzymology in the development of novel anti-ageing therapeutics.

Figure 1

E. coli constructs generated for in vivo precursor-directed combinatorial biosynthesis of polyketides. Constructs containing the CoA ligases with an empty Tom-15b vector served as controls during the subsequent high-performance liquid chromatography (HPLC) analyses. When a starter acid such as p-coumaric acid and an extender acid such as malonic acid are introduced to the E. coli constructs, the respective CoA thioesters are formed, which are utilized by 18xCHS to form polyketides like resveratrol.

Figure 2

Substrate profile of 18xCHS using 69 starter CoA thioesters and 12 extender CoA thioesters; 413 out of 828 possible combinations (49.9%) gave rise to new polyketides. Extender acyl-CoAs are abbreviated as follows: malonyl-CoA (Mal), methylmalonyl-CoA (MeMal), ethylmalonyl-CoA (EtMal), isopropylmalonyl-CoA (IsoMal), butylmalonyl-CoA (ButMal), allylmalonyl-CoA (AlMal), hydroxymalonyl-CoA (OHMal), fluoromalonyl-CoA (FMal), chloromalonyl-CoA (ClMal) and bromomalonyl-CoA (BrMal), phenylmalonyl-CoA (PhMal), and 3-thiophenemalonyl-CoA (3ThMal).

Figure 3

HPLC profile of organic extracts from E. coli constructs grown in M9 supplemented with 2-flurocinnamate and butylmalonate. Biosynthesized products in spent minimal medium containing either E. coli with CoA ligases + 18xCHS or E. coli with CoA ligases only (control construct) were extracted and subjected to HPLC analysis. An additional peak at retention time 38.7 min was observed in the extract containing the 18xCHS construct, but was not present in the extract containing the control construct, indicating that a new polyketide was biosynthesized when 2-fluorocinnamyl-CoA and butylmalonyl-CoA were supplemented to 18xCHS.

Figure 4

Substrate profile of 18xCHS based on the seven starter substrate families. Benzoyl-CoA derivatives were the most preferred starters while bicyclic aromatic CoA derivatives were least preferred by 18xCHS. A total of at least 413 novel polyketides were biosynthesized.

Figure 5

Structural mimicry of bicyclic aromatic CoA thioesters. The mimicry to phenylacetyl-CoA (for 1-naphthalenecarboxyl-CoA) and cinnamyl-CoA (for 2-naphthalenecarboxyl-CoA and 2-quinolinecarboxyl-CoA) is highlighted in blue.

Figure 6

Resonance structures depicting the stabilization of the carbanion of 3-thiophenemalonyl-CoA after decarboxylation.

Figure 7

Potential polyketide formed from one unit of p-coumaroyl-CoA and butylmalonyl-CoA without cyclization.

Figure 8

(A) Survival plots of C. elegans fed with 50 μM commercially available resveratrol (green) or 0.1% DMSO (blue). Worms exposed to 50 μM resveratrol had a significantly longer mean lifespan (27.4 days) compared to the DMSO control (17.1 days). (B) Survival plots of wild-type N2 (blue) and eat-2 (green) C. elegans. eat-2 mutants had a significantly longer mean lifespan (17.6 days) compared to the wild-type worms (14.3 days). The p-values were calculated using log rank test, and a p-value ≤ 0.05 is considered to be statistically significant.

Figure 9

Survival plots of N2 C. elegans fed with either 18xCHS E. coli strain (green) or control E. coli strain (blue) supplemented with 3-chlorocinnamic acid + malonic acid (left) or 3-(3′-chloro-4′-methoxy)phenylpropanoic acid + methylmalonic acid (right). Worms exposed to the novel polyketides produced from these two combinations of starter and extender acids had a significantly shorter mean lifespan compared to the control which is exposed to the respective carboxylic acids and CoA esters only.

Figure 10

Wild-type N2 worms exposed to 8 days of 50 μM resveratrol had a significantly higher basal OCR, maximal OCR, and spare respiratory capacity compared to N2 worms exposed to 5-fluoro-2′-deoxyuridine (FUdR) or FUdR + 0.1% DMSO. Consistent with results using the CR mimetic resveratrol, eat-2 mutants undergo chronic CR and have a higher basal respiration, maximal respiration, and spare respiratory capacity compared to wild-type C. elegans. Introducing 50 μM resveratrol to eat-2 mutants does not improve their mitochondrial function further. FUdR: worms exposed to FUdR treatment only; DMSO: worms exposed to FUdR + 0.1% DMSO; Resv: worms exposed to FUdR + 50 μM resveratrol. *p-value ≤ 0.05; **p-value ≤ 0.01; ***p-value ≤ 0.001; ****p-value ≤ 0.0001.

Figure 11

Butylmalonic acid and p-coumaric acid were incorporated into NGM agar together with either the control (Ctrl) or 18xCHS E. coli construct. C. elegans exposed to novel polyketides and CoA esters produced by the 18xCHS E. coli construct (contains CoA ligases + 18xCHS) do not have a higher basal respiration, maximal respiration, and spare respiratory capacity compared to C. elegans exposed to CoA esters only (control construct contains CoA ligases only).

Figure 12

Combinatorial biosynthesis of resveratrol or other anti-ageing polyketides in E. coli Nissle.

Figure 13

HPLC profile of organic extracts from E. coli Nissle constructs grown in M9 supplemented with 0.25 mM tyrosine. A peak corresponding to p-coumarate is present in extracts from Nissle constructs containing PAL and absent in the extract from the Nissle construct with an empty vector, indicating that PAL is active. In addition, resveratrol is produced by the Nissle construct containing PAL, 4CL, and 18xCHS.