Direct modifications of the cyclic peptide Polymyxin B leading to analogues with enhanced in vitro antibacterial activity

A recent report from the World Health Organisation highlights the lack of treatment options available for multidrug-resistant Gram-negative pathogens, including carbapenem-resistant Enterobacteriaceae (such as Klebsiella and Escherichia coli), Pseudomonas and Acinetobacter, which can cause severe and often life-threatening infections.¹ Polymyxins remain one of the last classes of antibiotics with activity against these pathogens, and this, despite concerns over nephrotoxicity leading to poor therapeutic index, has led to a resurgence in their clinical use as a therapy of last resort.² Strains with reduced susceptibility to polymyxins are however, beginning to emerge in the clinic, although this is currently at a low level.³ An increased understanding of the structure-activity relationships (SAR) of this important class of compounds is required to further the development of the next generation of polymyxin-related antibiotics.

Polymyxin B (PMB, Figure 1) is a naturally-occurring lipopeptide antibiotic comprising a cyclic heptapeptide core, with a linear exocyclic tripeptide chain capped at the N-terminus with a fatty acyl group. PMB demonstrates bactericidal activity via a mechanism involving specific electrostatic and hydrophobic interactions with the lipid A component of lipopolysaccharide (LPS) in the outer membrane of Gram-negative bacteria. Nuclear magnetic resonance (NMR) studies of PMB with LPS indicate that in addition to the electrostatic interactions of the positively charged Dab (diaminobutyric acid) residues with the negatively charged phosphate and carboxylate groups of lipid A, the hydrophobic residues of PMB – phenylalanine, leucine, and fatty acyl group – align with the lipid chains of the LPS.⁴

Figure 1:

Polymyxin B major components

Commercially available sources of PMB comprise two major components polymyxin B1 and polymyxin B2 (Figure 1a and b), which have been shown to have similar activity to each other and to the commercial natural product mixture.^5,6 Over the last ten years, our knowledge of the SAR of polymyxins has steadily developed and has been the subject of recent reviews.^7–10

We have recently reported a series of novel polymyxins with improved antibacterial potency and reduced renal cell cytotoxicity, achieved by replacing the amino acid at position 1 and the fatty acyl chain with branched amine-containing acyl moieties of specific regio- and stereochemistry.¹¹ These compounds are based on a nonapeptide scaffold with the amino acid Dap (diaminopropionic acid) at the position adjacent to the cyclic core. Examples include the N-terminal heterocycles, and beta branched amino butyrates shown in Figure 2. As part of our SAR programme we sought to modify the lipophilicity of the cyclic core of polymyxins by modification of position 6 (D-Phe). To date, modified core Polymyxins have been prepared via solid phase peptide synthesis on resin, often requiring an off-resin cyclisation step.^12–14 We sought to develop methodology which would enable modification of the D-Phe position on the polymyxin molecule itself, utilising the readily available and low-cost starting material, PMB.

Figure 2:

Amino acyl nonapeptides¹¹

Our initial approach was to attempt to introduce a bromine substituent on the phenylalanine ring, to afford an intermediate amenable to palladium-catalysed cross coupling reactions. While the phenylalanine system was resistant to bromination with N-bromosuccinimide, harsher acidic conditions (60% sulphuric acid, bromoisocyanuric acid)¹⁵ led to decomposition of the peptide. However, we were encouraged by the reports of the Olah group on the mild electrophilic bromination of deactivated aromatic rings with Nbromo succinimide in the presence of boron trifluoride monohydrate.¹⁶ Indeed we found that boron trifluoride, used as the commercially-available dihydrate served as both a solvent for the cyclic peptide and a Lewis-acid catalyst for the bromination reaction (Scheme 1). Bromination proceeded at room temperature to afford mono-brominated material, from which the major component was isolated and purified for antibacterial testing. ¹H nmr confirmed this to be the (4-bromo)phenylalanine derivative 3(a).

Scheme 1:

Reagents and conditions: (i) N-bromosuccinimide, BF₃.2H₂O; (ii) BOC₂O, Et₃N, CH₃CN, H₂O, 27% over 2 steps; (iii) PhB(OH)₂ (or (4-pyridyl)B(OH)₂), PdCl₂(Ph₃P)₂, Na₂CO₃, DMF, H₂O; (iv) TFA, DCM, followed by prep HPLC, 10% over 2 steps

For further derivatisation, protection of the amino groups was required, thus the crude brominated material was protected as the penta-BOC derivative (4), before being used as a substrate in palladium catalysed cross-coupling. Initial attempts at Suzuki coupling with phenyl boronic acid in the presence of bis(triphenylphosphine)palladium(II)dichloride gave the desired cross-coupled product. After deprotection with TFA and purification by preparative HPLC the 4-phenyl phenylalanine derivative 5a was obtained. Suzuki coupling of 4 with 4-pyridyl-boronic acid under the same conditions led to 5b.

A key intermediate in the semi-synthetic route to polymyxin derivatives with a modified exocyclic chain is the corresponding cyclic heptapeptide, produced by enzymatic cleavage of the natural product.^{11, 17} We were pleased to find that the crude BOC-protected bromo derivative 4 was a substrate for the commercially-available protease Savinase®, enabling the synthesis of the brominated heptapeptide core (Scheme 2). Purification at this stage enabled the isolation of the brominated isomers 6(a) and (b) in a 6:1 ratio. The major component, 6(a) was assigned the (4-bromo)phenylalanine regiochemistry based on the ¹H nmr in the aromatic region, while the minor isomer was consistent with the (2-bromo)phenylalanine.¹⁵ (Supplementary). The 4-bromo regioisomer 6a was then coupled to an appropriate dipeptide fragment as previously described¹¹ to form the nonapeptide 7 containing the protected 1,3-diaminopropionic acid at the position adjacent to the core. An improved Suzuki reaction with phenyl boronic acid, utilising X-Phos (2-Dicyclohexylphosphino-2′,4′,6′-triisopropylbiphenyl) in the presence of palladium acetate proceeded in 50% yield to the biphenyl nonapeptide 8a. N-terminal deprotection of the CBZ group was effected by catalytic transfer hydrogenolysis in the presence of ammonium formate to give 8b. Coupling to the corresponding BOC protected N-terminal acid followed by deprotection and purification gave the biphenyl nonapeptide analogues 9a and b.

Scheme 2:

Reagents and conditions: (i) Savinase, 1,4-butanediol, pH 8, 27%; (ii) CBZThr (O-tBu)-Dap(BOC)OH, HATU, NEtⁱPr₂, DCM, quant.; (iii) PhB(OH)₂, Pd(OAc)₂, X-Phos, K₃PO₄,toluene, 100°C, 50%; (iv) NH₃⁺HCOO^-, 10% Pd/C, MeOH; 93%; (v) RCO₂H (N-BOC protected), HATU, NEtⁱPr₂, DCM, DMF; (vi) TFA, DCM, followed by prep HPLC, 23% over 2 steps.

A second approach to the modification of the C-6 position of polymyxins was the conversion of the D-phenylalanine residue to cyclohexyl alanine (Cha) by reduction (Scheme 3). Hydrogenation at atmospheric pressure over platinum oxide catalyst afforded clean conversion to the cyclohexyl derivative 10. For the synthesis of nonapeptide derivatives, reduction of the previously described key nonapeptide intermediates 11a and 11b¹¹ under similar conditions gave excellent conversion to the cyclohexyl derivatives 12a and 12b (Scheme 4). These could then be elaborated further to the desired aminoacyl nonapeptides, for example 13a-c as previously described.¹¹

Scheme 3:

Reagents and conditions: (i) H₂/PtO₂, CH₃CO₂H, followed by prep HPLC, 29%

Scheme 4.

Reagents and conditions with example yields (i) H₂/PtO₂, CH₃CO₂H, 97%; (ii) RCO₂H (N-BOC protected), HATU, NEtⁱPr₂, DCM; (iii) TFA, DCM, followed by prep HPLC, 29% over 2 steps.

Compounds 3a, 5a-b, 9a-b, 10, and 13a-c were tested against a panel of Gram-negative bacteria including clinical strains with reduced susceptibility to PMB (Table 1). Cytotoxicity was measured against the human kidney proximal tubule epithelial cell line (HK-2) and is presented as the ratio of the IC₅₀ of test compound to that of PMB¹⁹. PMB and previously reported nonapeptides 2a-d¹¹ were included for comparison. Introduction of a bromine or additional phenyl ring to the phenylalanine (position 6) of polymyxin (3, 5a) served to enhance the activity against resistant strains. This is consistent with previous reports on analogues of PMB where increased lipophilicity of the core of polymyxins at positions 6 or 7 has been indicated to increase activity against resistant strains.^12,13 The 4-pyridyl phenylalanine compound 5b, however, was less active against all strains. This may be due to unfavourable interactions with the lipid chain of LPS. In the nonapeptide series with a heterocyclic N-terminus, the biphenyl compounds 9a and 9b also showed enhanced activity against resistant strains compared to the corresponding phenylalanine analogues 2a and 2b. While the conversion of phenylalanine to the more bulky cyclohexyl alanine (Cha) at position 6 was tolerated, the effect on activity was less dramatic than the bromo or biphenyl modifications. In the DCha-6 polymyxin analogue 10, clear improvement in activity was noted against the less-sensitive Klebsiella strain. The cytotoxicity of the novel derivatives was, however, increased compared to their counterparts with the polymyxin scaffold, likely due to the overall increased lipophilicity. Nevertheless, the synthetic approach described herein, allowing separate modification of the core and N-terminus from a common starting material, could enable independent modification of the lipophilicity of the cyclic core and the N-terminus without the need for total synthesis. The systematic evaluation of the effect of lipophilicity on activity and cytotoxicity of polymyxin derivatives will be the subject of a further communication.

Table 1.

MIC (μg/ml) and in vitro cytotoxicity of test and reference compounds

	E.coli		K.pneumoniae		P.aeruginosa		A.baumannii		Cytotox (HK-2 IC50 relative to PMB)
Cpd	ATCC 25922	IHMA 940398	ATCC 4352	IHMA 652780	ATCC 27853	IHMA 517175	NCTC 13424	IHMA 863866
1	0.25	16	0.125	4.0	0.5	8.0	0.25	32	1
2a	0.25	32	0.25	16	0.5	32	0.5	>64	16.6
2b	0.03	2.0	0.03	2.0	0.25	2.0	0.125	2.0	8.8
2c	0.06	16	0.125	4.0	0.06	8.0	0.25	>64	6.9
2d	0.25	32	0.125	ND	0.125	8.0	0.125	>64	12.0
3a	0.5	4.0	0.5	0.5	0.5	2.0	0.5	4.0	ND
5a	1.0	8.0	0.5	0.5	1.0	2.0	1.0	16	0.14
5b	1.0	32	1.0	ND	1.0	32	2.0	ND	0.7
9a	0.125	8.0	0.25	0.5	0.5	4.0	0.125	4.0	2.4
9b	0.25	2.0	0.25	0.5	0.5	1.0	0.25	0.5	1.2
10	0.5	8.0	0.5	1.0	1.0	4.0	0.5	32	ND
13a	0.5	8.0	ND	2.0	0.5	2.0	0.125	32	ND
13b	0.25	16	0.06	4.0	0.25	4.0	0.125	64	5.2
13c	0.125	32	ND	4.0	0.25	8.0	0.06	64	2.4

MICs determined by broth microdilution using cation-adjusted Mueller-Hinton broth (Oxoid, CM0405) according to CLSI guidelines.¹⁸ In vitro cytotoxicity was measured against the human kidney proximal tubule epithelial cell line (HK-2) and is presented as the ratio of the IC₅₀ of test compound to that of polymyxin B¹⁹.

In conclusion, we have described synthetic methodology enabling the exploration of modifications at the C-6 position of polymyxins and leading to increased knowledge of SAR in this region.

Supplementary Information provides details of the synthesis and characterisation of final compounds 3a, 5a-b, 9a-b, 10, 13a-c, and intermediates 6a and b.