Upgrading the ferribactin pre-chromophore – synthesis, modification and polymerization

Abstract

The global rise in antibiotic resistance underscores the urgent need for alternative antimicrobial strategies. One approach involves the conjugation of iron-chelating moieties to macromolecular scaffolds to disrupt bacterial iron homeostasis and inhibit cellular uptake mechanisms. In this work, the pre-chromophoric unit of the siderophore ferribactin served as the structural template for the development of antimicrobial polymer precursors. A series of L-tyrosine and L-DOPA-derived pre-chromophore analogues were synthesized and chemically modified to introduce polymerizable functionalities. These monomers were copolymerized with N-vinylpyrrolidone via reversible addition–fragmentation chain-transfer (RAFT) polymerization to afford well-defined, bifunctional copolymers. Antimicrobial testing of the monomers and polymers showed varying levels of activity, depending on the bacterial species.

---

1. Introduction

The increase in antimicrobial resistance is a serious threat to human health causing an estimated 1.27 million annual deaths worldwide.^1–3 Therefore, the issue is under ongoing surveillance and discussion in the scientific and health communities.^1–3 Different approaches to combat these resistances are being pursued, including new technologies like artificial intelligence (AI),^3,4 or targeting enzymes in cell wall synthesis or in respiratory chain processes.^5,6 Other strategies focus on reducing the pathogenicity and virulence of bacteria rather than killing them to avoid selective evolutionary pressures, thus reducing the spread of antibiotic resistance by “diluting” the bacterial gene pool.^7–9 Targeting the nutrient supply of bacteria, such as the iron uptake, is another promising strategy to inhibit bacterial growth, especially in combination with antibiotics.¹⁰

Iron as a stoichiometric building block is an essential element in the bacterial growth phase. Nature developed various chelators for iron uptake, among which enterobactin 1 is a prominent example (Fig. 1). Enterobactin 1 belongs to a class of small iron chelating molecules named siderophores, produced by bacteria to maintain iron uptake, especially in iron deficient conditions.^10–12 Conjugates of siderophores and antibiotics show promising results as “Trojan horses”, increasing the activity of antibiotics by using the bacteria's own pathways to bring antibiotics into the cell thus circumventing some bacterial resistance mechanisms.^13,14 Pyoverdin D 2 is an example for a class of siderophores used by the Pseudomonas species that has been thoroughly investigated.^15–21 The first total synthesis of pyoverdin D 2 was reported in 2013 by Mashiach and Meijler.²² Synthetic approaches towards the chromophore unit 3-chr of pyoverdin D 2 (marked blue in Fig. 1) have been known since 1990, but challenges remain due to the multistep syntheses and toxic reagents required.^23–25

Fig. 1 Examples of siderophores, the (pre)chromophoric subunits of pyoverdin 3-chr and ferribactin 4-chr as well as deferiprone 5.

Ferribactin 4-ferri, i.e. the biosynthetic precursor of pyoverdin 3-pyo, is a weaker siderophore,²⁶ which is also found in Pseudomonas supernatants.²¹ Ferribactin 4-ferri differs from pyoverdin 3-pyo only in the chromophore unit. Although some literature refers to the ferribactin substructure 4-chr as a chromophore,^27–29 it should be noted that 4-chr is a colorless compound. Therefore, in this work, unit 4-chr is referred to as a pre-chromophore. Ferribactin 4-ferri has been less studied than pyoverdin 3-pyo, presumably due to the fact, that its binding affinity towards Fe(III) is 4 orders of magnitude lower than that of the corresponding Fe(III)-pyoverdin complexes. This vast difference is caused by the varying amounts of binding sites. Both siderophores have two binding sites in the form of N-formyl-N-hydroxy-L-ornithine moieties (marked pink in Fig. 1). But while the catechol unit of the pyoverdin chromophore 3-chr provides a third binding site of pyoverdin, the phenol unit of the ferribactin pre-chromophore 4-chr does not participate in complexation. This was shown by Budzikiewicz with ¹³C NMR studies of Ga(III)-complex analogues of ferribactin compared with data for pyoverdin 3-pyo from other literature.²⁶ Ga(III) serves as a substitute for Fe(III) with equal charge and comparable ion radius and does not have the unfavorable paramagnetic property afflicting NMR studies.²⁶ Most publications deal with the structure of ferribactin 4-ferri and its role as precursor in the biosynthesis of pyoverdin 3-pyo.^26–29 The synthesis of the ferribactin pre-chromophore 4-chr was reported in 1993 by Jones, who achieved the formation of a tetrahydropyrimidine unit in 4-chrvia a chiral iminothioester.³⁰ However, epimerization occurred during the synthetic reaction and only a diastereomeric mixture was obtained.³⁰ Later, Abdallah performed the coupling via an amide and Meerwein's reagent, introducing the first stereoselective synthesis step.^31,32 Begley³³ and Jones³⁴ utilized truncated ferribactin derivatives for biomimetic oxidative cyclizations to the pyoverdin derivative.

Regarding their use as antimicrobials, siderophores and their iron-binding subunits seem to be highly attractive lead compounds for drug development.^10,35 However, combining a low molecular weight iron chelator with a macromolecular backbone is an effective strategy to limit the growth of pathogenic bacteria, because small iron binding molecules are likely to be ingested and digested by bacteria. Kizhakkedathu connected a hyperbranched polyglycerol with the hexadentate Fe(III) chelator N,N-bis(2-hydroxybenzyl)ethylenediamine-N,N-diacetic acid (HBED) units (subunit 6, shown in Fig. 2).³⁶ These macromolecular Fe(III) chelators successfully slowed down the growth of Staphylococcus aureus and showed bactericidal activity when administered as adjuvants together with antibiotics.³⁶ An alternative approach was reported by Berkland who developed a biomimetic iron-sequestering polymer PAI-DHBA 7via crosslinking of polyallylamine (PAI) with 2,3-dihydroxybenzoic acid (DHBA).³⁷ Pretreatment of culture media with this polymer effectively inhibited the growth of Pseudomonas aeruginosa.³⁷ Other catechol-based polymers with antibiotic activity against a variety of bacteria are known as well.³⁸ An attractive approach towards polymer-based drug delivery systems is the reversible addition–fragmentation chain transfer (RAFT) polymerization.^39–41 RAFT enabled, for example, the synthesis of amphiphilic copolymers with sequence-controlled alternating hydrophilic–hydrophobic pendant side chains.⁴² Solvent-dependent self-assembly of these copolymers led to micelles, vesicles or reverse micelles, which were used for drug encapsulation.⁴² Amino acid-based alternating copolymers displayed a pH-dependent reversal of amphiphilicity.^43,44 Previously, Ang et al. disclosed a deferiprone-functionalized acrylamide monomer 8 that was co-polymerized with N-vinylpyrrolidone 9 using RAFT polymerization to yield an iron-binding co-polymer, DIBI 10, with enhanced antimicrobial activity.⁴⁵ The DIBI copolymer inhibited the growth of the antibiotic-resistant bacteria Acinetobacter baumannii, the methicillin-resistant Staphylococcus aureus as well as the yeasts Candida albicans and Candida vini.^46–48

Fig. 2 Previous work on siderophore-polymer hybrids and the ferribactin-derived polymers described in this work.

Our aim was to synthesize a library of polymerizable ferribactins with different protecting groups, to integrate them into polymers in order to obtain antibiotic polymers 11, 12, and to gauge the roles of the chromophores and pre-chromophores in the native compounds (Fig. 2). Although the synthesis of L-tyrosine-based ferribactin pre-chromophores 14-T is known,^30,31 no synthetic route for the analogous L-DOPA-based derivatives 14-D has been published yet. This should provide the basis for a biomimetic approach towards the pyoverdin chromophore 3-chrvia the ferribactin pre-chromophore 14-T or 14-D instead of truncated systems without stereogenic centers.^33,34 Since the synthesis route is complex, an incorporation of small amounts of the ferribactin-based monomers 13-T and 13-D with other easily available iron chelating monomers such as MAHMP 8 is favorable, to enhance the effects of already existing polymers. This was tested under the RAFT polymerization method that afforded the iron-chelating co-polymer DIBI.⁴⁵

The ability to bind iron was evaluated by UV-vis measurements, a simpler and more sensitive method in comparison with NMR studies with Ga(III).²⁶ This should allow for better comparison between phenol- and catechol-derived ferribactins and different protecting groups, as well as a first indication of potential antimicrobial activity.

Finally, biological tests with synthesized ferribactins 13-T and 13-D and the obtained polymers 11 and 12 were performed with different bacteria to determine their antimicrobial activity. This data indicated the influence of structural motives in the ferribactins 13-T, 13-D and the incorporation into the polymers 11 and 12, thus pointing to the first structure activity relationships (SAR).

2. Results and discussion

2.1. Retrosynthetic approach to the ferribactin pre-chromophore

In order to synthesize the ferribactin pre-chromophore derivatives 14-T and 14-D, the synthetic route via an iminoether, following a method by Jones³⁰ and Abdallah,³¹ appeared to be the most promising (Fig. 2). The protection of the amino acid precursors was necessary for the synthesis of target polymers 11 and 12 in order to avoid undesired side reactions during the different synthetic steps. The set of O-methyl and N-Cbz protecting groups (PGs) was chosen to obtain chemically stable derivatives and to minimize accidental deprotection during acidic or other deprotecting reaction conditions throughout the planned synthetic route. But the deprotection of the methoxy ethers may not be viable for the target molecules 14-T or 14-D due to their stability. Therefore, another set, containing O-silyl and N-Boc protecting groups was chosen to synthesize derivatives that can be easily deprotected after synthesis of the target molecules 14-T and 14-D. Compounds 14-T, 14-D should be obtained via coupling of amides 15-T and 15-D with 2,4-diaminobutyric acid (DABA) 16 (prepared from L-glutamine 17) and subsequent cyclization. Amides 15-T and 15-D should be synthesized from L-tyrosine 18a and L-DOPA 18b, respectively.

2.2. Synthesis of ferribactins derived from tyrosine and DOPA amides

First, the amide derivatives of L-tyrosine 18a and L-DOPA 18b had to be prepared. As shown in Scheme 1, the amino groups of L-tyrosine 18a and L-DOPA 18b were protected first by modification of known procedures.^49–51 Tyrosine 18a was treated with Boc₂O and NaOH (method A) and gave Boc-tyrosine 19a-T in quantitative yield, while the Boc protection of DOPA 18b was performed with NaHCO₃ and Boc₂O and gave Boc-DOPA 19b-D in quantitative yield (entries 1, 2). The Cbz protection of tyrosine 18a was performed with CbzCl (method B) and yielded 53% of Cbz-tyrosine 19c-T. The Cbz-DOPA 19d-D was obtained in a similar reaction with CbzCl and K₂CO₃ in 48% (entries 3, 4). Then the vicinal hydroxy groups of the phenol moiety were protected with different protecting groups according to literature to avoid oxidation reactions towards quinoid systems.^50,52 Boc-DOPA 19b-D was treated with TBDMSCl (method C) to give the TBDMS and Boc-protected 20b-D in 43% yield (entry 5). The Cbz-protected tyrosine 19c-T and DOPA 19d-D were alkylated with an excess of MeI and the obtained methyl esters were saponified with 2 M NaOH to yield the O-methyl-N-Cbz-protected tyrosine 20c-T in 80% and the O-methyl-N-Cbz-protected DOPA 20d-D in 78%, respectively (method D, entries 6, 7).

Scheme 1 Synthesis and results of tyrosine and DOPA amides. (A) Boc₂O, NaOH or NaHCO₃, H₂O, THF, rt, 22–25 h, quant.; (B) CbzCl, NaOH or K₂CO₃, H₂O, THF, rt, 23–26 h, 48–53%; (C) TBDMSCl, imidazole, MeCN, rt, 42 h, 43%; (D) MeI, K₂CO₃, DMF, rt, 18–26 h, then 2 M NaOH, rt, 6–19 h, 78–80%; (E) N-methylmorpholine, isobutyl chloroformate, concentrated NH₃, CH₂Cl₂, 0 °C to rt, 21–24 h, 49–58%, (F) N-methylmorpholine, isobutyl chloroformate, concentrated NH₃, CH₂Cl₂, 0 °C to rt, 18 h, then 1 M NaOH, THF, rt, 3 h, 54%; (G) S3-T, TBDPSCl, imidazole, CH₂Cl₂, rt, 25 h, 38%. For further details see Chapter S2, SI.

The amidations of the protected derivatives 20b-D, 20c-T and 20d-D were performed according to a modified procedure by Moreno-Cinos with isobutyl chloroformate and concentrated NH₃ solution, yielding the O-TBDMS-N-Boc-DOPA amide 15b-D in 58%, the O-methyl-N-Cbz-tyrosine amide 15c-T in 55% and the O-methyl-N-Cbz-DOPA amide 15d-D in 49% respectively (method E, entries 9, 10, 11).⁵³ To obtain O-TBDPS-N-Boc-tyrosine amide 15a-T, the N-Boc-tyrosine 19a-T was first treated according to Moreno-Cinos, then with TBDPSCl, yielding 21% of 15a-T over two steps (method F, then method G, entry 8, see SI, Chapter S3).⁵³

2.3. Synthesis of coupling agent 2,4-diaminobutyric acid (DABA)

The second building block for the coupling of the tetrahydropyrimidine ring in 14-T and 14-D is 2,4-diaminobutyric acid (DABA) 16. The synthesis of DABA 16 started with glutamine 17 (Scheme 2), which was treated with Boc₂O (analogue to tyrosine 18a)⁴⁹ and Boc-glutamine 21 was isolated in 67% yield (method A). A Hofmann rearrangement with N-Boc-glutamine 21 was performed with Br₂ and NaOH, then the resulting intermediate was protected with Boc₂O, so N,N-Boc₂-DABA 22 could be obtained in 44% yield (method B).⁵⁴ Boc₂-DABA 22 was treated with an excess of concentrated HCl to yield DABA 16 as dihydrochloride in 84% (method C).⁵⁵

Scheme 2 Synthesis of DABA 16. (A) Boc₂O, NaOH, H₂O, THF, rt, 22 h, 67%; (B) Br₂, NaOH, H₂O, 0 °C to 80 °C, 2 h, then Boc₂O, NaOH, H₂O, 1,4-dioxane, rt, 24 h, 44%; (C) concentrated HCl, rt, 2 h, 84%.

2.4. Formation of ferribactin pre-chromophore monomers

After optimization (see SI, Chapter S3), pre-chromophore synthesis was performed according to Jones.³⁴ Therefore, the silyl-protected amines 15a-T, 15b-D and the methyl-protected amides 15c-T, 15d-D were stirred with MeOTf, then coupled with DABA 16 to obtain the corresponding ferribactin pre-chromophores (14a-T, 14b-D, 14c-T, 14d-D) (Scheme 3). The silyl-protected derivatives yielded 94% for 14a-T and 59% for 14b-D (method A, entries 1, 2), while the methyl-protected derivatives yielded 73% for 14c-T and 40% for 14d-D (entry 3, 4).

Scheme 3 Coupling reaction to ferribactin pre-chromophores and modification to monomers. (A) MeOTf, CH₂Cl₂ or CHCl₃, reflux to rt, 22–27 h, then DABA 16, iPr₂EtN, EtOH, rt to reflux, 20–23 h, 40–93%; (B) TFA, CH₂Cl₂, rt, 4 h or 6 M HCl, reflux, 6 h then Boc-β-alanine S2 EDC·HCl, HOBt·H₂O, N-methylmorpholine, CH₂Cl₂ or DMF rt, 16–22 h, 5–83%; (C) TFA, CH₂Cl₂, rt, 2–4 h then methacrylic acid (MAA), EDC·HCl, HOBt·H₂O, N-methylmorpholine, CH₂Cl₂, rt, 16–23 h, 47–80%. For further details see Chapter S2, SI.

For further functionalization to the ferribactin monomers 13a-T, 13b-D, 13c-T and 13d-D deprotection was necessary. For the silyl-protected, Boc-containing derivatives 14a-T and 14b-D, the acid-catalyzed reaction with trifluoroacetic acid (TFA) according to a method by Imramovský was successfully applied.⁵⁶ The derivatives 14a-T and 14b-D were deprotected with TFA and used without isolation in the subsequent coupling reaction.⁵⁷ Activation of N-Boc-protected β-alanine S2 with EDC·HCl and HOBt, followed by addition of the deprotected derivatives of 14a-T and 14b-D, yielded the silyl-protected spacer derivatives 23a-T in 83% and 23b-D in 42% (method B, entries 5, 6). After optimization (see SI, Chapter S3), the method of Ghorai was used to deprotect the methylated Cbz-derivatives 14c-T and 14d-D.⁵⁸ Sequential treatment with 6 M HCl, then coupling with N-Boc-protected β-alanine S2 yielded the methyl-protected compounds 23c-T in 5% and 23d-D in 35% respectively (method B, entries 7, 8).⁵⁸ To introduce the polymerizable side group, Boc deprotection with TFA was carried out on the silyl-protected (23a-T, 23b-D) and methyl-protected (23c-T, 23d-D) spacer derivatives. The subsequent coupling reaction (optimized as described in the SI, Chapter S3) was performed with methacrylic acid, EDC·HCl and HOBt. The silyl-protected monomers 13a-T and 13b-D were obtained in 80% and 49% yield. The methyl-protected monomers 13c-T and 13d-D were obtained in 68% and 47% yield (method C, entries 9, 10, 11, 12).

2.5. Investigation of iron binding properties of ferribactin pre-chromophore monomers

The silyl-protected (13a-T, 13b-D) and methyl-protected (13c-T, 13d-D) monomers were tested regarding their Fe(III)-binding properties by adding increasing amounts of FeCl₃ (4 mM solution in MeOH) to a solution of monomers in MeOH and monitoring the absorbance via UV-vis spectroscopy (Fig. 3). To study how the protecting groups alter the Fe(III)-binding properties, samples of the silyl-protected monomers 13a-T and 13b-D were deprotected with TBAF to yield deprotected 24a-T and 24b-D. The TBAF was not removed before the measurements. In Fig. 3 the absorbance of all variants with an added amount of 80 nmol of Fe(III) in approx. 1 mL solvent is shown exemplarily (all spectra are shown in the SI, Fig. S2–S7). The enlarged section of the absorption spectrum at λ = 400–800 nm (Fig. 3b) shows that for all protected derivatives no additional band was detected as was expected for the Fe(III) binding according to the literature.^59–61 None of the protected monomers showed Fe(III) binding properties, even the sterically less demanding methyl ethers in 13d-D were suppressing coordination. The phenol/catechol-unit of the molecule was proven to be the only binding site by detecting a shift in absorbance, although several peptide bonds and the tetrahydropyrimidine ring were present. The peptide bonds and the tetrahydropyrimidine ring remained unchanged and free during the complexation experiments. The deprotected 24b-D (catechol-type) showed a wide band with a maximum absorbance around 570 nm, which indicates an Fe(III) binding at the catechol unit. This aligns perfectly with the ligand system containing two catechol units at neutral pH as reported by Bijlsma.⁵⁹ The deprotected 24a-T (phenol-type) showed a narrow and much weaker band with a maximum absorbance at 460 nm, which would fit in the range of phenol-containing Fe(III) complexes.^60,61 Due to its small intensity, it is assumed that a very weak complexation occurred.

Fig. 3 (a) UV-vis spectra of 1 mL of stock solutions in MeOH with molar concentrations of 13a-T: 64.9 µM, 13b-D: 92.7 µM, 13c-T: 122.5 µM, 13d-D: 112.0 µM, 24a-T: 64.9 µM (TBAF not removed), 24b-D: 92.7 µM (TBAF not removed), with added 20 µL of 4 mM FeCl₃ solution (80 nmol FeCl₃); (b) Enlarged section of (a) at λ = 400–800 nm; (c) structures of 13a-T, 13b-D, 13c-T, 13d-D and 24a-T, 24b-D.

Since UV-vis measurements showed that the free catechol moiety is needed for Fe(III) binding, we surmised that the silyl-protected monomers (13a-T, 13b-D) and the methyl-protected monomers (13c-T, 13d-D) probably cannot induce a state of iron deficiency and show antimicrobial effects from iron depletion. Only the enzymatic catechol deprotection of the DOPA-derived derivatives 13b-D (silyl-protected) or 13d-D (methyl-protected) might lead to this effect. The Fe(III) binding of the tyrosine-based derivatives 13a-T (silyl-protected), 13c-T (methyl-protected) in their free phenolic form is probably too weak to show an effect. It can be concluded that the complexation behavior of ferribactin and polymer derivatives occurs through the catechol moiety, as there is an absence of metal complexation towards peptide bonds and the tetrahydropyrimidine.

2.6. Polymerization of ferribactin pre-chromophore monomers

Copolymerization of the silyl-protected monomers (13a-T, 13b-D) and the methyl-protected monomers (13c-T, 13d-D) with N-vinylpyrrolidinone (NVP) 9 and N-[2-(3-hydroxy-2-methyl-4-oxopyridin-1(4H)-yl)ethyl]methacrylamide (MAHMP) 8 was attempted. DIBI 10 was synthesized as a control reaction according to the original conditions by Ang⁴⁵ (Scheme 4), which yielded DIBI 10 in 83%, to show successful polymerization conditions and to obtain DIBI as non-(pre-)-chromophore containing co-polymer for control experiments.

Scheme 4 Polymerization conditions towards DIBI 10 and methyl-protected 13c-T with NVP (9), MAHMP (8), RAFT-agent (25), tetramethylethylenediamine (TMEDA) and tert-butyl hydroperoxide (TBHP).

The reaction conditions were adjusted for the silyl-protected (13a-T, 13b-D) and methyl-protected (13c-T, 13d-D) monomers as shown exemplarily for the polymerizations of methyl-protected 13c-T in Scheme 4. The reactions were performed in µM scale in gas chromatography (GC) vials. The monomers, RAFT-agent 25, NVP 9 and tetramethylethylenediamine (TMEDA) were dissolved in the solvent or solvent mixture (Table 1). Due to solubility issues we, unfortunately, could not use the reported RAFT conditions.^42–44Tert-butyl hydroperoxide (TBHP) was added and stirred at 40 °C.

Table 1 Reaction conditions for the polymerization attempts of silyl-protected (13a-T, 13b-D) and methyl-protected (13c-T, 13d-D) monomers with NVP 9 and MAHMP 8

Entry	Monomer	Comonomer	Solvent (v : v)	Time [h]	Yield [%]	Observation
1	13c-T	NVP 9	H2O	19	65	copolymer 11c-T formed
2	13c-T	NVP 9, MAHMP 8	H2O	20	71	copolymer 12c-T formed
3	13d-D	NVP 9	H2O	18	—	no reaction
4	13a-T	NVP 9	H2O	—	—	13a-T does not dissolve
5	13a-T	NVP 9	H2O/DMSO 1 : 1	18	—	oligomers formed
6	13b-D	NVP 9	H2O/DMSO 1 : 1	—	—	gelation of reaction mixture
7	13b-D	NVP 9	H2O/DMF 1 : 1	—	—	gelation of reaction mixture
8	13b-D	NVP 9	Dioxane/H2O 5 : 3	18	—	no reaction

---

The methyl-protected monomers 13c-T and 13d-D were soluble in water. As shown in Table 1 (entry 1) and Scheme 4 (middle), the methyl-protected monomer 13c-T reacted with NVP 9 without further change to the reaction conditions of the DIBI 10 control and yielded polymer 11c-T in 65%. Copolymerization of NVP 9 and MAHMP 8 (Scheme 4, bottom and Table 1, entry 2) was performed and yielded DIBI polymer 12c-T in 71%, confirming reagent and reaction conditions. For the methyl-protected monomer 13d-D, a similar reactivity was expected as compared to methyl-protected 13c-T, but no reaction occurred under aqueous conditions (entry 3). Due to their hydrophobic protection groups, the silyl-protected monomers 13a-T and 13b-D could not be dissolved in H₂O, therefore addition of a co-solvent was necessary (entry 4). For silyl-protected 13a-T, a mixture of H₂O and DMSO was tested and yielded a solid. The NMR spectrum of the solid showed an overlapping mixture of broadened singlets and sharp doublets of the tyrosine phenyl ring protons. The broadened singlets indicate polymerization and the remaining sharp doublets indicate that silyl-protected monomer 13a-T is still present (see SI, page S158). Additionally, the matrix-assisted laser desorption (MALDI) mass spectra data shows no polymer distribution and only one relevant signal at 1156.98, since the mass of silyl-protected 13a-T is 640.31 and the mass of NVP 9 is 111.07 (see SI, Chapter S13). We deduced from this data that the formation of short chains of oligomers had likely occurred (entry 5). The silyl-protected DOPA derivative 13b-D did not show any reaction under the RAFT conditions. Silyl-protected 13b-D gelled the (1 : 1) mixture of H₂O/DMSO as well as the (1 : 1) mixture of H₂O/DMF with the intended concentration of 26 mM while dissolving, before the start of the reaction (entries 6, 7). The gelation of solvents by DOPA derivatives and short peptides containing DOPA is known.^62,63 Moreover Kar⁶⁴ observed the gelation of dioxane, THF, DMSO and DMF by TBDMS protected (catechol unit) DOPA, that was solely caused by the silyl units and showed liquefaction after cleavage of the TBDMS ether with fluoride. The silyl-protected monomer 13b-D was soluble in a (5 : 3) mixture of 1,4-dioxane/H₂O, but the monomer showed no conversion under RAFT conditions (entry 8). Incorporation of silyl-protected monomers 13b-D and 13a-T as well as additional silyl deprotection was desirable. For silyl-protected 13b-D, this would lead to a free catechol, that could bind iron (as shown for the deprotected 24b-D in chapter 2.5) or enhance the iron binding properties of existing polymers. Deprotected derivative 24a-T could mimic ferribactin and deprotected derivative 24d-D pyoverdin and their (pre-)chromophoric units, and, if recognized, interfere with the siderophore uptake of bacteria.

The co-polymerization with NVP 9 was performed to increase water solubility with the aim of increasing the iron binding and bioactivity of the silyl-protected monomers 13a-T and 13b-D and the methyl-protected monomers 13c-T and 13d-D. Poly-NVP polymers and co-polymers are known for their great water solubility and have been used as blood plasma extenders.⁶⁵ A polymer has additional advantages. It cannot be as easily taken up and digested by bacteria³⁶ and can improve the antibacterial activity with an effect of cooperation through several active units in the vicinity, as shown by Ang⁴⁵ in DIBI 10. Another effect mentioned by Ang is the induction of steric hinderance by the polymer chain, later described as the polymer molecule wrapping around the Fe(III), that increases the antimicrobial effect.^45,47 All of these effects might combine in our targeted polymers. After polymerization, the silyl groups of the incorporated 13a-T and 13b-D could be deprotected to benefit from the iron binding capacity due to the free phenol or catechol moiety. The β-alanine as a short spacer was chosen to mirror the ethyl amide linker of MAHMP 8 and methacrylate as polymerizable group was used to get the same reactivity as MAHMP 8 by Ang, since the same polymerization method and conditions were chosen.⁴⁵ Increasing solubility by polymerization worked, because the obtained polymers 11c-T and 12c-T of the monomer 13c-T were easily water soluble. Even the oligomers of silyl-protected 13a-T with 9 showed an increased water solubility, compared with the silyl-protected monomer 13a-T itself.

For the successfully synthesized polymers 10, 11c-T and 12c-T, diffusion-ordered spectroscopy (DOSY) NMR spectra and MALDI mass spectra were measured (see SI, Chapter S7). For 10, the MALDI distribution showed a maximum in the same order of magnitude as the molecular weight for 10 synthesized by Ang.⁴⁵ For 11c-T and the mixed polymer 12c-T, the maxima were lower. DOSY spectra are 2D spectra, showing the cross signals of the ¹H NMR and the diffusion coefficient (see SI, Table S2). No homopolymer formation for 10, 11c-T or 12c-T was observed and the extracted diffusion coefficients lie roughly in between the reported ones for NVP 9 and PVP.⁶⁶

2.7. Biological tests

The biological tests were performed in aqueous medium with small amounts of DMSO. Due to their solubility in H₂O and H₂O/DMSO mixtures, the silyl-protected monomers (13a-T, 13b-D), the methyl-protected monomers (13c-T, 13d-D) and the polymers (11c-T, 12c-T) were chosen for biological testing. For comparison, the known DIBI polymer 10 was included in the assays.⁴⁵

The cytotoxicity tests were performed with L929 mouse cells and the Alamar Blue assay was used to evaluate the viability of the cells.⁶⁷ As an example of the obtained results, the data for 13a-T are shown in Fig. 4.

Fig. 4 Cell viability assay of the monomer 13a-T.

The silyl-protected monomers 13a-T and 13b-D and the methyl-protected monomers 13c-T and 13d-D (see SI, Fig. S8) displayed a similar influence on cell viability regardless of the protecting groups. A variation of around 90% to 110% was obtained for concentrations from 0.4 µM to 25 µM, showing no cytotoxicity in this concentration range and lying within the expected experimental measurement error range. For concentrations higher than 50 µM, the cell viability decreased with increasing concentration (70% cell viability at 100 µM), indicating cytotoxic properties at higher concentrations. For polymers 11c-T, 12c-T and DIBI 10 the cell viability increased slightly with increasing concentrations (see SI, Fig. S9). The results indicated that the studied polymers 12c-T, 11c-T and DIBI 10 were not cytotoxic.

Antibacterial tests were performed with Escherichia coli wild type (WT), Escherichia coli ΔTolC, Staphylococcus aureus, Klebsiella pneumoniae and Pseudomonas aeruginosa. Four of which (S. aureus, K. pneumoniae, P. aeruginosa, E. coli) belong to the ESKAPE pathogens, which are representative for nosocomial multidrug resistant bacteria.⁶⁸ Examples for the results obtained are shown in Fig. 5 (For all results see SI, Fig. S10–S19).

Fig. 5 Bacterial growth after incubation with 12c-T. (a) E. coli ΔTolC, (b) K. pneumoniae.

For the E. coli WT no antibacterial activity of the silyl-protected monomers (13a-T, 13b-D) and the methyl-protected monomers (13c-T, 13d-D) was observed (Fig. S10). The polymers 10 and 12c-T slightly inhibited growth of E. coli WT at the high concentrations of 50 µg mL⁻¹ and 100 µg mL⁻¹, whereas the polymer 11c-T showed no activity. This indicates that the activity is probably caused by the hydroxypyridone unit in the polymer 12c-T (derived from monomer 8), which agrees with the published results for DIBI 10.⁴⁵ Presumably, at high concentrations both polymers created an iron poorer medium that led to decreased bacterial growth.

Next, tests were performed for E. coli ΔTolC. Since the ΔTolC mutant is defective in the TolC protein, which is in charge of the transport of molecules through the outer membrane, the results were expected to be more distinct.⁶⁹

The methyl-protected monomers 13c-T and 13d-D did not show any significant activity (Fig. S12). In contrast, the silyl-protected monomers 13a-T and 13b-D led to growth of about 80% at concentrations of 100 µM. Both monomers seem to be active against E. coli ΔTolC at high concentrations. As expected, the polymers showed a more distinct activity against the ΔTolC variant than against the wild type (Fig. S13). While 11c-T had no activity, the polymers 10 and 12c-T (Fig. 5a) reduced the growth of E. coli ΔTolC to 70–75%. The growth rate already decreased at concentrations of 10 µg mL⁻¹.

The tests with S. aureus showed no activity for the silyl-protected monomers (13a-T, 13b-D) and the methyl-protected monomers (13c-T, 13d-D) at low concentrations. Above 25 µg mL⁻¹, an increased growth of S. aureus occurred for all four monomers 13a-T, 13b-D, 13c-T and 13d-D (Fig. S14). This might be explained by metabolization of the monomers by the bacterium. However, this interpretation has to be approached with great care, because experimental edge effects might also be involved. The polymers 10 and 12c-T had only little effect on the growth of S. aureus (Fig. S15), whereas 11c-T showed a slight increase in growth at 100 µg mL⁻¹.

Tests with K. pneumoniae revealed increased bacterial growth with increasing concentration for the silyl-protected monomers (13a-T, 13b-D) and for the methyl-protected monomer 13c-T, while the methyl-protected monomer 13d-D did not show any activity (Fig. S16). In comparison, the polymers 10, 11c-T and 12c-T (Fig. 5b) slightly decreased growth (Fig. S17).

For P. aeruginosa no significant activity for the silyl-protected monomers (13a-T, 13b-D) and the methyl-protected monomers (13c-T, 13d-D) could be observed. The bacterial growth rate changed little irrespective of the concentrations (Fig. S18). Although the silyl-protected monomers (13a-T, 13b-D) and the methyl-protected monomers (13c-T, 13d-D) have a similar structure as the ferribactin pre-chromophore 4-chr, no interference with bacterial biosynthesis seemed to occur. Presumably, the methoxy or silyloxy protecting groups are not cleaved by these bacteria. Neither polymer 12c-T nor 11c-T changed the growth of P. aeruginosa significantly, while DIBI 10 led to a slight decrease.

3. Conclusion

In conclusion, several protected ferribactin pre-chromophores were synthesized that, under the right conditions, can successfully be incorporated into polymers. Less polar monomers have lower solubility in water resulting in poor conversion during RAFT polymerization. UV-vis experiments showed that protected monomers cannot bind to iron. The unprotected catechol unit clearly shows iron binding behavior. We thus speculate that in naturally occurring iron-binding compounds, peptide polymer bonding and stabilization of the Fe(III) center is of little to no effect.

Biological tests showed no cytotoxicity at low concentrations but weak cytotoxicity at high concentrations, while the polymers showed no cytotoxic behavior. Non-cytotoxic properties are of interest for potential future antibacterial drug developments to selectively affect Fe(III) availability to pathogens and the human host.

Our results provided new ferribactin derivatives in monomeric form, which were embedded in a polymer. Future work is required to tackle the synthesis of polymers with deprotected ferribactin pre-chromophores and pyoverdin chromophore subunits, possibly through incorporation of protected precursors in the polymer in order to study their antibiotic potential and obtain detailed structure–activity relationships of these polymers.

Author contributions

A. P. G. synthesized the compounds and wrote the first draft of the manuscript. M. T. C. A. supervised the polymerization and wrote the manuscript. S. H. helped with the pre-chromophore synthesis, B. N. G. and I. M. helped with the synthesis of precursors of the aza-Wittig route. D. C. M. C. performed the antibacterial tests, B. P. performed the cytotoxicity tests. U. B. supervised the biological testing and wrote the manuscript. M. B. supervised the polymerization and wrote the manuscript. A. Z. checked and managed the data. S. L. supervised the project and wrote the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting the findings of this study are available within the article and its supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d5nj03251b.

Additional raw data files are available from the corresponding author upon reasonable request.

Acknowledgements

The travel grant for A. Greulich by the Global Glimpse Fellowship Program of the University of Stuttgart as well as the Ministerium für Wissenschaft, Forschung und Kunst des Landes Baden-Württemberg, the European Regional Development Fund (EFRE, FEIH_778511), and the German Research Foundation DFG: Project-ID 358283783 – SFB 1333/2 2022 (for providing the infrastructure), HBFG shared instrumentation grants no. 513030456-INST 41/1175-1 FUGG for MALDI-MS, INST 41/897-1 FUGG for the 700 MHz NMR, INST 41/1136-1 FUGG for LC-Orbitrap-MS: Exactive Plus Orbitrap MS System and INST 41/1135-1 FUGG for GC-Orbitrap-MS: Exactive GC Orbitrap MS System are gratefully acknowledged. We acknowledge the support of Birgit Claasen and her department at the University of Stuttgart for all help surrounding NMR measurements and mass spectrometry. Thanks to Franziska Welsch and Ruben Pereira Rebelo from the University of Stuttgart for the MALDI measurements. The support of Friederike Adams with ideas for polymer analytics is acknowledged.

References

1. World Health Organization, Global Antimicrobial Resistance and Use Surveillance System (GLASS), 2022, vol. 9.
2. I. N. Okeke, M. E. A. de Kraker, T. P. Van Boeckel, C. K. Kumar, H. Schmitt, A. C. Gales, S. Bertagnolio, M. Sharland and R. Laxminarayan, Lancet, 2024, 403, 2426–2438.
3. S. K. Ahmed, S. Hussein, K. Qurbani, R. H. Ibrahim, A. Fareeq, K. A. Mahmood and M. G. Mohamed, J. Med. Surg. Public Health, 2024, 2, 100081.
4. M. F. Chellat, L. Raguž and R. Riedl, Angew. Chem., Int. Ed., 2016, 55, 6600–6626.
5. G. Niro, S. C. Weck and C. Ducho, Chem. – Eur. J., 2020, 26, 16875–16887.
6. S. Kruth and M. Nett, Antibiotics, 2023, 12, 1067.
7. M. Duplantier, E. Lohou and P. Sonnet, Pharmaceuticals, 2021, 14, 1262.
8. J. Konstantinović, A. M. Kany, A. Alhayek, A. S. Abdelsamie, A. Sikandar, K. Voos, Y. Yao, A. Andreas, R. Shafiei, B. Loretz, E. Schönauer, R. Bals, H. Brandstetter, R. W. Hartmann, C. Ducho, C. M. Lehr, C. Beisswenger, R. Müller, K. Rox, J. Haupenthal and A. K. H. Hirsch, ACS Cent. Sci., 2023, 9, 2205–2215.
9. C. Jonkergouw, N. K. Beyeh, E. Osmekhina, K. Leskinen, S. M. Taimoory, D. Fedorov, E. Anaya-Plaza, M. A. Kostiainen, J. F. Trant, R. H. A. Ras, P. Saavalainen and M. B. Linder, Nat. Commun., 2023, 14, 2141.
10. U. Bilitewski, J. A. V. Blodgett, A. K. Duhme-Klair, S. Dallavalle, S. Laschat, A. Routledge and R. Schobert, Angew. Chem., Int. Ed., 2017, 56, 14360–14382.
11. M. Miethke and M. A. Marahiel, Microbiol. Mol. Biol. Rev., 2007, 71, 413–451.
12. R. Saha, N. Saha, R. S. Donofrio and L. L. Bestervelt, J. Basic Microbiol., 2013, 53, 303–317.
13. C. Rohrbacher, R. Zscherp, S. C. Weck, P. Klahn and C. Ducho, Chem. – Eur. J., 2023, 29, e202202408.
14. T. N. Pham, P. Loupias, A. Dassonville-Klimpt and P. Sonnet, Med. Res. Rev., 2019, 39, 2343–2396.
15. G. Briskot, K. Taraz and H. Budzikiewicz, Z. Naturforsch., C: J. Biosci., 1986, 41, 497–506.
16. H. Beiderbeck, D. Risse, H. Budzikiewicz and K. Taraz, Z. Naturforsch., C: J. Biosci., 1999, 54, 1–5.
17. C. Lenz, C. Amann, G. Briskot, K. Taraz and H. Budzikiewicz, Z. Naturforsch., C: J. Biosci., 2000, 55, 146–152.
18. A. M. Albrecht-Gary, S. Blanc, N. Rochel, A. Z. Ocaktan and M. A. Abdallah, Inorg. Chem., 1994, 33, 6391–6402.
19. C. Cezard, N. Farvacques and P. Sonnet, Curr. Med. Chem., 2014, 22, 165–186.
20. C. Cézard, P. Sonnet and B. Bouvier, J. Biol. Inorg. Chem., 2019, 24, 659–673.
21. P. Visca, F. Imperi and I. L. Lamont, Trends Microbiol., 2007, 15, 22–30.
22. R. Mashiach and M. M. Meijler, Org. Lett., 2013, 15, 1702–1705.
23. T. Kolasa and M. J. Miller, J. Org. Chem., 1990, 55, 4246–4255.
24. P. Seubert, M. Freund, R. Rudolf, Y. Lin, L. Altevogt, U. Bilitewski, A. Baro and S. Laschat, Synlett, 2020, 1177–1181.
25. T. Zhang, A. Bolhuis and I. M. Eggleston, Eur. J. Org. Chem., 2024, e202400745.
26. K. Taraz, L. Seipold, C. Amann and H. Budzikiewicz, Z. Naturforsch., C: J. Biosci., 2000, 55, 836–839.
27. H. Budzikiewicz, M. Schäfer, D. U. Fernández, S. Matthijs and P. Cornelis, Biometals, 2007, 20, 135–144.
28. U. Hohlneicher, M. Schäfer, R. Fuchs and H. Budzikiewicz, Z. Naturforsch., 2001, 56c, 308–310.
29. U. Hohlneicher, R. Hartmann, K. Taraz and H. Budzikiewicz, Z. Naturforsch., C: J. Biosci., 1995, 50, 337–344.
30. R. C. F. Jones and A. K. Crockett, Tetrahedron Lett., 1993, 34, 7459–7462.
31. A. Zamri, F. Sirockin and M. A. Abdallah, Tetrahedron, 1999, 55, 5157–5170.
32. C. Hennard, Q. Chi Truong, J. F. Desnottes, J. M. Paris, N. J. Moreau and M. A. Abdallah, J. Med. Chem., 2001, 44, 2139–2151.
33. P. C. Dorrestein, K. Poole and T. P. Begley, Org. Lett., 2003, 5, 2215–2217.
34. R. C. F. Jones, S. C. Yau, J. N. Iley, J. E. Smith, J. Dickson, M. R. J. Elsegood, V. McKee and S. J. Coles, Org. Lett., 2009, 11, 1519–1522.
35. C. Kurth, H. Kage and M. Nett, Org. Biomol. Chem., 2016, 14, 8212–8227.
36. S. Abbina, A. Gill, S. Mathew, U. Abbasi and J. N. Kizhakkedathu, ACS Appl. Mater. Interfaces, 2020, 12, 37834–37844.
37. N. El-Gendy, J. Qian, K. Eshelman, M. Rivera and C. Berkland, Biomacromolecules, 2015, 16, 1480–1488.
38. S. Razaviamri, K. Wang, B. Liu and B. P. Lee, Molecules, 2021, 26, 559.
39. D. J. Keddie, Chem. Soc. Rev., 2014, 43, 496–505.
40. M. D. Nothling, Q. Fu, A. Reyhani, S. Allison-Logan, K. Jung, J. Zhu, M. Kamigaito, C. Boyer and G. G. Qiao, Adv. Sci., 2020, 7, 2001656.
41. G. Moad, E. Rizzardo and S. H. Thang, Chem. – Asian J., 2013, 8, 1634–1644.
42. K. G. Goswami, S. Mete, S. S. Chaudhury, P. Sar, E. Ksendzov, C. Das Mukhopadhyay, S. V. Kostjuk and P. De, ACS Appl. Polym. Mater., 2020, 2, 2035–2045.
43. P. Sar, S. Ghosh, Y. D. Gordievskaya, K. G. Goswami, E. Y. Kramarenko and P. De, Macromolecules, 2019, 52, 8346–8358.
44. P. Sar, S. G. Roy, P. De and S. Ghosh, Macromol. Mater. Eng., 2020, 305, 1900809.
45. M. T. C. Ang, R. Gumbau-Brisa, D. S. Allan, R. McDonald, M. J. Ferguson, B. E. Holbein and M. Bierenstiel, Med. Chem. Commun., 2018, 9, 1206–1212.
46. M. del C. Parquet, K. A. Savage, D. S. Allan, M. T. C. Ang, W. Chen, S. M. Logan and B. E. Holbein, Antimicrob. Agents Chemother., 2019, 63, 1–15.
47. M. del C. Parquet, K. A. Savage, D. S. Allan, R. J. Davidson and B. E. Holbein, Front. Microbiol., 2018, 9, 1–11.
48. B. E. Holbein and R. Mira De Orduña, FEMS Microbiol. Lett., 2010, 307, 19–24.
49. N. A. Bindman, S. C. Bobeica, W. R. Liu and W. A. Van Der Donk, J. Am. Chem. Soc., 2015, 137, 6975–6978.
50. M. J. Sever and J. J. Wilker, Tetrahedron, 2001, 57, 6139–6146.
51. M. Li, B. Luo, Q. Liu, Y. Hu, A. Ganesan, P. Huang and S. Wen, Org. Lett., 2014, 16, 10–13.
52. J. P. T. A. Guerra, A. Lindner, C. R. Nicoleti, V. G. Marini, M. Silva and V. G. Machado, Tetrahedron Lett., 2015, 56, 4733–4736.
53. C. Moreno-Cinos, E. Sassetti, I. G. Salado, G. Witt, S. Benramdane, L. Reinhardt, C. D. Cruz, J. Joossens, P. Van Der Veken, H. Brötz-Oesterhelt, P. Tammela, M. Winterhalter, P. Gribbon, B. Windshügel and K. Augustyns, J. Med. Chem., 2019, 62, 774–797.
54. B. C. Askew, R. A. Bednar, B. Bednar, D. A. Claremon, J. J. Cook, C. J. McIntyre, C. A. Hunt, R. J. Gould, R. J. Lynch, J. J. Lynch, S. L. Gaul, M. T. Stranieri, G. R. Sitko, M. A. Holahan, J. D. Glass, T. Hamill, L. M. Gorham, T. Prueksaritanont, J. J. Baldwin and G. D. Hartman, J. Med. Chem., 1997, 40, 1779–1788.
55. M. Đud, Z. Glasovac and D. Margetić, Tetrahedron, 2019, 75, 109–115.
56. R. Jorda, J. Dušek, E. Řezníčková, K. Pauk, P. P. Magar, A. Imramovský and V. Kryštof, Eur. J. Med. Chem., 2017, 135, 142–158.
57. W. J. Wu, Y. Wu and B. Liu, Tetrahedron, 2017, 73, 1265–1274.
58. B. G. Das and P. Ghorai, Chem. Commun., 2012, 48, 8276–8278.
59. J. Bijlsma, W. J. C. de Bruijn, J. A. Hageman, P. Goos, K. P. Velikov and J. P. Vincken, Sci. Rep., 2020, 10, 1–11.
60. M. Emami, R. Bikas, N. Noshiranzadeh, J. Sanchiz, K. Ślepokura and T. Lis, J. Mol. Struct., 2019, 1180, 392–398.
61. B. K. Kanungo, M. Baral, S. K. Sahoo and S. E. Muthu, Spectrochim. Acta, Part A, 2009, 74, 544–552.
62. D. Giuri, P. Ravarino and C. Tomasini, Org. Biomol. Chem., 2021, 19, 4622–4636.
63. D. Giuri, N. Zanna and C. Tomasini, Gels, 2019, 5, 27.
64. S. K. Mandal, P. Seth and T. Kar, New J. Chem., 2024, 48, 5429–5438.
65. M. Teodorescu and M. Bercea, Polym. Plast. Technol. Eng., 2015, 54, 923–943.
66. N. T. M. Hai, J. Furrer, F. Stricker, T. M. T. Huynh, I. Gjuroski, N. Luedi, T. Brunner, F. Weiss, A. Fluegel, M. Arnold, I. Chang, D. Mayer and P. Broekmann, J. Electrochem. Soc., 2013, 160, D3116–D3125.
67. J. O’Brien, I. Wilson, T. Orton and F. Pognan, Eur. J. Biochem., 2000, 267, 5421–5426.
68. W. R. Miller and C. A. Arias, Nat. Rev. Microbiol., 2024, 22, 598–616.
69. G. Dhamdhere and H. I. Zgurskaya, Mol. Microbiol., 2010, 77, 743–754.

---