Zwittermicin A: Synthesis of Analogs and Structure-Activity Studies

Abstract

Analogs and diastereomers of the natural product zwittermicin A were prepared. SAR studies of these compounds reveal the antifungal activity to be dependent singularly upon the natural constitution and configuration.

(+)-Zwittermicin A (1) (ZwA) is a highly polar aminopolyol antibiotic isolated from the soil-borne bacterium Bacillus cereus.¹ Compound 1 was first reported in 1994 and shows significant activity against human pathogenic yeast and phytopathogenic fungi.^1,2 More importantly 1 has been shown to work synergistically with endo-toxin produced by Bacillus thuringensis for control of gypsy moth.³ Research has shown strains of B. cereus producing 1 to be ubiquitous in the soil^1c and may be more benign to the environment than some synthetic pesticides.⁴ Biosynthesis of 1 arises from a non-ribosomal peptide synthetase/polyketide synthase pathway (NRPS/PKS) involving two new type 1 PKS extender units: hydroxymalonyl-acyl carrier protein (ACP) and aminomalonyl-ACP.⁵

Preliminary studies show that 1 appears to exhibit a unique mechanism of action.⁶ Investigations of ZwA-resistant mutants found that the resistance mapped to rpoB and rpoC, genes that encode subunits of RNA polymerase.⁶ However, 1 showed no effect on total RNA or DNA synthesis implying a mechanism of action that differs from that of other antibiotics that target RNA polymerase.⁶

Previously, we synthesized (−)-1, the enantiomer of ZwA, its diastereomer 2, and analogs 5 through 10 (Figure 1).⁷ This report describes the synthesis of the additional diastereomers 3 and 4 as well as analogs ent-6, 11 and 12 together with a comprehensive SAR study of all compounds.

Figure 1.

ZwA diastereomers and analogs for SAR studies.

During the course of our synthesis of (−)-1 a number of models were made to evaluate synthetic strategies and some of these were later used to make analogs 11 and 12 as shown in Scheme 1. Analog ent-6 was synthesized by reduction of diazide 13 (Scheme 1).⁸ The known alcohol 14⁹ was converted to aldehyde 15 in three high-yielding steps; MOM protection of the primary OH group,¹⁰ desilylation of the primary OTBS group,¹¹ followed by Swern¹² oxidation. Evan’s aldol addition of the boron enolate of chiral glycolate equivalent 16¹³ to 15 (dr 47:1) followed by separation and removal of the chiral auxiliary under standard conditions afforded carboxylic acid 17 (57% over two steps).¹⁴ Coupling of 17 to known N-ureido- L-1,3-diaminopropionamide (−)-18⁷ gave 19 in 83% yield; global deprotection of the latter provided the truncated ZwA analog 11 in 76% yield. Similarly, analog 12 was prepared from (+)-18.

Scheme 1.

Synthesis of ent-6, 11, and 12. Reagents and conditions: (a) H₂O, H₂ (1 atm), Pd/C, 2 h, 100%; (b) MeOCH₂Cl, EtN(i-Pr)₂, CH₂Cl₂, 0 °C-rt, 14 h, 91%; (c) TBAF, THF, −10 °C, 16 h, 93%; (d) (i) (COCl)₂, DMSO, CH₂Cl₂, −78 °C, (ii) Et₃N, 94%; (e) (i) 16, n-Bu₂BOTf, Et₃N, CH₂Cl₂, −78 to 0 °C, 3 h, (ii) 15, −78 to 0 °C, 2.5 h, 85%, dr 47:1; (f) H₂O₂, LiOH, 0 °C, 30 min, 67%; (g) (i) 17, EDCI, HOBt, DMF, 0 °C, 10 min, (ii) (−)-18, Et₃N, 0 °C-rt, 2.5 h, 83%; (h) (i) HCl, MeOH, H₂ (5 atm), Pd/C, 1 h, (ii) HCl, H₂O, H₂ (5 atm), Pd/C, 1 h, 76%; (i) (i) 17, EDCI, HOBt, DMF, 0 °C, 10 min, (ii) (+)-18, Et₃N, 0 °C-rt, 1.5 h, 67%; (j) (i) HCl, MeOH, H₂ (5 atm), Pd/C, 1 h, (ii) HCl, H₂O, H₂ (5 atm), Pd/C, 1 h, 73%. EDCI= 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloide, HOBt= 1-hydroxybenzotriazole.

Azidodiol 20, prepared from L-serine as described earlier,⁷ was refunctionalized by TBDPS protection¹⁵ of the terminal alcohol, MOM protection of the secondary alcohol and removal of the TBDPS group to give 21 in high yield (86% three steps, Scheme 2). Transformation of the azido group in 21 to an N,N-dibenzyl group by hydrogenolysis (Lindlar’s catalyst¹⁶) followed by N-benzylation¹⁷ gave a primary alcohol that was easily oxidized to the stable aldehyde 22 (74% three steps).

Scheme 2.

Synthesis of ZwA diastereomers 3 and 4. Reagents and conditions: (a) TBDPSCl, imidazole, DMF, 0 °C-rt, 3.5 h, 91%; (b) MeOCH₂Cl, Hünig’s base, CH₂Cl₂, 0 °C-rt, 48 h, 96%; (c) TBAF, THF, −10 °C, 4 h, 98%; (d) Lindlar’s cat., H₂, (1 atm), EtOH, 15 h, 89%; (e) BnBr, K₂CO₃, CH₃CN, 83 h, 92%; (f) (i) (COCl)₂, DMSO, CH₂Cl₂, −78 °C, (ii) Et₃N, 90%; (g) (i) 23, n-Bu₂BOTf, Hünig’s base, Et₂O, −78 °C, 1.5 h, (ii) 22, −78 to 0 °C, 2 h, 49%; (h) LiOH, H₂O:MeOH:THF, 0 °C, 4 h, 84%; (i) (i) EDCI, HOBt, DMF, 0 °C, 10 min, (ii) (−)-18, Et₃N, 0 °C-rt, 2.5 h, 86%; (j) (i) HCl, MeOH, H₂ (5 atm), Pd/C, 1 h, (ii) HCl, H₂O, H₂ (5 atm), Pd/C, 1.3 h, 57%; (k) (i) 24, EDCI, HOBt, DMF, 0 °C, 10 min, (ii) (+)-18, Et₃N, 0 °C-rt, 1.5 h, 86%; (l) (i) HCl, MeOH, H₂ (5 atm), Pd/C, 1 h, (ii) HCl, H₂O, H₂ (5 atm), Pd/C, 1 h, 73%. TBDPSCl= tert-butyldiphenylsilyl chloride.

Aldol addition of the glycolate equivalent 23¹⁸ to aldehyde 22 (9:1 dr) followed by hydrolysis with lithium hydroxide under standard conditions¹⁹ afforded carboxylic acid 24 in (41% over two steps).

EDCI coupling of 24 to (−)-18 gave 25 (86% yield) and global deprotection provided zwittermicin A diastereomer 3 in 57% yield. In a similar manner, diastereomer 4 was synthesized from (+)-18 and 24.

Antifungal assay of natural (+)-1 and the 13 synthetic compounds was conducted against the fungal strains Candida albicans 96-489, C. glabrata, C. albicans UCDFR1, C. albicans ATCC 144503, C. krusei, and the phytopathogenic bacteria Erwinia carotovora, and E. amylovora and oomycete Phytophthora infestans (Table 1). Natural zwittermicin A [(+)-1] showed activity against C. albicans, C. glabrata, E. carotovora, and E. amylovora. Despite structural similarities of the synthetic compounds to natural ZwA, particularly the enantiomer (−)-1 and stereoisomeric 3 and 4, only the natural product showed detectable activity.

Table 1.

Biological testing of zwittermicin A and synthetic compounds.

Pathogenic microbes	(+)-1 MICa,b (μg/mL)	(−)-1, 2–10, ent-6, 11 and 12 MICa,b (μg/mL)
Candida albicans 96-489	55.7	>128
C. glabrata	59.5	>128
C. albicans UCDFR1	>128	>128
C. albicans ATCC 144503	>128	>128
C. krusei	>128	>128
Erwinia carotovora	22.2	>128c
E. amylovora	18.8	>128c
Phytophthora infestansd	>32	>32c

^aThe MIC endpoint is defined as the lowest concentration (μg/mL) with 90% growth inhibition.

^b(+)-1, (−)-1, 2–4, 11 and 12 were converted to the free amine before testing. Compounds 5-10 were used as their HCl salts.

^c5-10 not tested.

^dThe maximum concentration used was 32 μg/mL due to limitations of the nutrient agar well diffusion assay.

The biological data indicate that the mechanism of action is highly stereospecific. The enantiomer (−)-1 and all of the diastereomers were inactive.⁷ Changing the configuration of 1 at only two stereocenters (C-13 and C-14 in 3) resulted in complete loss of activity, even in the most susceptible strain, E. amylovora. None of the truncated analogs (Figure 1) showed activity. In fact, the activity of (+)-1 is quite singular; it not mimicked by simple synthetic stereoisomers or analogs. Antifungal activity in simple vicinal aminoalkanols has been observed for other natural products, such as oceanapiside,²⁰ oceanalin,²¹ even sphingosine and its synthetic short-chain analogs (C₆), irrespective of the relative configurations.²² Consequently, the stringent stereochemical requirements observed for antifungal activity properties in 1 were unexpected and surprising. The mechanism of action of 1 is presently unknown, but clearly differs from that of the former compounds which appear to interdict sphingolipid metabolism.²³

In summary, two zwittermicin A diastereomers and three analogs were synthesized and compared with natural zwittermicin A and eight other previously synthesized analogs and stereoisomers in antifungal assays. The microbial susceptibility profile of 1–12 indicates that the mechanism of action is highly stereospecific, and the complete complement of functionality found in the natural enantiomer zwittermicin A [(+)-1] is required for efficacy. Further investigations to identify the mechanism of action may benefit from selective affinity tagging of 1 and deployment in cell-free pull-down experiments to identify the cellular target.²⁴