Trojan horse strategy: synthesis of piperazine-based siderophores

Resistance to antibiotics is an emerging phenomenon and a major medical problem. The resistance of Gram-negative bacteria such as Pseudomonas aeruginosa and the Burkholderia group to conventional antibiotics leads to therapeutic failure and requires new antibiotic therapies. The use of iron transport systems is a promising strategy to overcome this phenomenon. These TonB-dependent receptors, essential for the survival of microorganisms, allow specific recognition of ferric siderophore complexes in order to transport iron within bacteria1. Bacteria, according to their kind, express different types of receptors that allow them to recognize their endogenous siderophores but also xenosiderophores. Pseudomonas aeruginosa and Burkholderia pseudomallei in particular possess FptA receptors allowing the recognition of pyochelin.2 These specific systems may allow the introduction of antibacterial agents by forming antibiotic-siderophore conjugates or toxic complexes such as gallium complexes, in the bacteria to kill it. Siderophores have three types of chelating function: catechols, hydroxamates and hydroxy-carboxylates. Previous work in the laboratory has shown that piperazine 1,4-dicatechol structures (MPPS0225) could be recognized by Pseudomonas aeruginosa strains. In order to further investigate this piperazine platform, we have synthesized iron chelators bearing 3-hydroxypyridin-4-ones and 1,3-dihydroxypyridin-4-one ligands. At the same time, we were interested in the synthesis of a more complex 2,5-dioxopiperazine platform, part of the rhodotorulic acid (RA), a natural siderophore produced by Rhodotorula pilimanae showing an interesting iron affinity (pFe = 21,8). Two RA synthesis strategies will be developed as well as the corresponding 3,6-disubstituted analogs. Through the synthesis of these chelators, we would like to study the influence, on the iron complexation, of: i) the nitrogen platform (piperazine or dioxopiperazine), ii) the presence of stereogenic centers (3,6-disubstituted dioxopiperazine vs 1,4 -disubstituted piperazines) and iii) the nature of the iron ligands (hydroxypyridinone vs catechol). An evaluation of the siderophore-like potential and a measurement of the complexing force will be carried out. 
We would like to thank the DGA and the Haut de France region for their financial support.
References

Miethke M.; Marahiel MA. Microbiology and Molecular Biology Reviews. 2007, 71, 413-451.
 Butt AT.; Thomas MS. Frontiers in Cellular and Infection Microbiology. 2017, 7.

Under iron limited conditions, many bacteria then synthesize molecules of low molecular weight called siderophores able to chelate the surrounding iron. These siderophore-Fe (III) complexes are then recognized by specific receptors responsible for bringing the essential iron element to the bacteria. Interestingly, the bacteria is able to recognize its endogenous siderophores but also siderophores synthesized by other bacteria or synthetic siderophores 1 . In particular, P. aeruginosa and B. pseudomallei both possess FptA receptors for the recognition of pyocheline, the endogenous siderophore of P. aeruginosa. These two types of bacteria are also capable of recognizing catecholate siderophores such as cepaciachelin for B. pseudomallei 2 , and enterobactin for P. aeruginosa 3 (Fig. 3). These specific systems may allow the introduction of antibacterial agents by forming antibiotic-siderophore conjugates or toxic complexes such as gallium complexes, in the bacteria. Previous work in the laboratory has shown that piperazine 1,4-dicatechol structures (MPPS0225) can be recognized by strains of Pseudomonas aeruginosa. Bacterial growth has been observed as a function of the ratio MPPS0225/Fe(III) in Medium Minimum Succinate. With a ratio MPPS0225/Fe(III) equal to 1,5, bacterial growth has been observed for DSM1117 strains, producing pyoverdine and pyocheline (Fig. 4). The same observation was found for PAD07 strains, which don't produce pyoverdine and pyocheline. We can assume that our compound is recognized by the bacteria (Fig. 5) and there is a competition between MPPS0225 and the endogenous siderophores of P. aeruginosa (Fig. 4).  Figure 6 illustrates the determination of the stoichiometry of the complex Fe(III)-MPPS0225 at pH = 5.70 using the JOB plot method. We can see the evolution of the ligand-to-metal chargetransfer (LMCT) absorption at 550 nm up to a 0.58 molar fraction of ligand, corresponding to an iron(III)/MPPS0225 2:3 stoichiometry. At physiological pH this stoichiometry is kept.

Results and discussion
To complete these results, MPPS0225 has been synthesized for physicochemicals studies like the pFe measurement. In order to further investigate this piperazine platform, we have synthesized iron chelators bearing 3-hydroxypyridin-4-ones ligands, bioisosteres of catechol groups. The bidentate ligands precursors 3 and 4 were synthesized with para-methoxybenzyl (PMB) as a protective group in order to be coupled with the 1,4-bis(3-aminopropyl)piperazine 1.  The bidentate ligands precursors 3 and 4 were synthesized with para-methoxybenzyl (PMB) as a protective group with a 70% and 65% yield.

Results and discussion
As mentioned before, 3 and 4 were synthesized to be coupled with the 1,4-bis(3aminopropyl)piperazine 1. The common hydrogenation step was optimized and carried out using a H-cube system generating hydrogen by electrolysis of water. MPPS0225 and 7 were, respectively, obtained with a 35% and 30% yield.
12 Figure 9: Synthesis of MPPS0225 and the hydroxypyridinone analog

Results and discussion
At the same time, we were interested in the synthesis of a more complex 2,5-dioxopiperazine platform, with analogs of the rhodotorulic acid (RA), a siderophore recognized by Gram-negative bacteria, showing an interesting iron affinity (pFe = 21,8). Different ways to obtain RA have been described 4 but the one we propose should be more efficient. Indeed, it is a convergent strategy which could lead to the synthesis of analogs thanks to asymmetric alkylations of a key intermediate carrying two cleavable chiral inductors (13). RA should be obtained in ten steps from the commercially available (S)-phenylglycinol (Fig. 10).  At this time, we synthesized the dioxopiperazine 13 in 6 steps from the (S)-phenylglycinol. This one was protected with TBDMSCl to afford 8 which was alkylated in presence of methyl 2bromoacetate to afford the secondary amine 9. This product was engaged in an amidification reaction to give 10 which was cyclized in presence of 8 to dioxopiperazine 11. Then, the TBDMS protective group was cleaved with TBAF to afford 12. Thanks to the (S)-phenylglycinol chiral inductors, the cis-dialkylation product 13 was obtained in a 4% global yield.
We managed to synthesize two 1,4-disubstituted piperazines, MPPS0225 and 7 in, respectively 3 and 4 steps with 35% and 30% yields. The RA synthesis has been carried out until the formation of a key intermediate 13 in 6 steps with a 4% yield.
As for the dicatechol siderophore analogs, MPPS0225, we will first study the bacterial recognition of synthesized chelators 7 and RA by measuring the siderophore-like effect and evaluate their ability to complex iron by physicochemical methods. The siderophores having shown the best capacity for recognition and complexation of iron will then be linked to an antibiotic, via a cleavable spacer or not depending on whether the target of the antibiotic is periplasmic or cytoplasmic. The antibacterial activity of these conjugates (or complexes) will then be evaluated and compared to the antibiotic alone to estimate the vectorization capacity of the siderophore. In parallel, cytotoxicity studies will also be conducted on these compounds to determinate their therapeutic index.