Dedicated to Professor Anton Rieker on the occasion of his 70th birthday
The glycopeptide antibiotic vancomycin (Scheme 1) has been used for more than 30 years against enterococcal and staphylococcal infections and meanwhile has the status of an antibiotic of last resort. [1] The significance of vancomycin together with its unique properties have attracted many researchers to perform work on chemical, biological, and medicinal aspects. [2] Over the past decade the increasing number of vancomycin-resistant bacterial strains has led scientists to investigate resistance mechanisms and to seek new approaches to overcome vancomycin-resistance problems. [2,3] As a consequence, a vast number of natural and semisynthetic glycopeptide derivatives have been characterized and tested. Considerable success was obtained with the semisynthetic derivative LY333328, a potent glycopeptide, which displays antibiotic activity against both methicillinresistant Staphylococcus aureus strains (MRSA) and vancomycin-resistant enterococci (VRE) and which is meanwhile being tested in clinical studies. [4] However, these variations relate primarily to peripheral features of vancomycin and the characteristic structure of the aglycon has been left mainly unchanged because of the highly demanding chemical and biochemical accessibility of the tricyclic aglycon.
Herein we present the first in vivo modified glycopeptides with non-natural fluorine substituents. In previous contributions, [5,6] we used the balhimycin-producing [7] (Scheme 1) strain Amycolatopsis mediterranei to obtain insights into the biosynthesis through gene disruption mutants. Based on these results we focused on the two diastereomeric 3-chloro-bhydroxytyrosine moieties (Scheme 1) as an attractive target for the introduction of structural variations, since they represent a basic element of the tricyclic aglycon structure. The presence of a chlorine substituent in each of the two bhydroxytyrosine moieties in natural balhimycin plays a significant role for enhancing antibiotic activity. [8] From the A. mediterranei wild-type strain (Figure 1, 1 a) we generated a deletion null-mutant in the bhp gene (Figure 1, 1 b), OP696, [6] which is deficient in b-hydroxytyrosine biosynthesis and thus is unable to produce balhimycin (Figure 1, 3 a). In mutasynthesis experiments, [9] we supplemented OP696 with several analogues of this amino acid. b-Hydroxytyrosine and derivatives were easily obtained as a racemic mixture of four stereoisomers in good yields by using a three-step synthesis. [10] Supplementing a culture of OP696 with 3-fluoro-b-hydroxytyrosine (3-Fht) resulted in antibiotic activity of the culture filtrate against the indicator strain Bacillus subtilis (Figure 1, 3 b). The amino acid 3-Fht did not show any antibiotic activity. Isolation from culture filtrates revealed an antibiotically active compound, and subsequent analysis with ESI-FTICR mass spectrometry revealed a mass of [Mþ2 H] 2þ ¼ 707.7439 Da (Dm ¼ 0.3 ppm; Figure 2) that corresponds to the molecular mass of 1413.4878 Da and to the elemental composition C 66 H 73 F 2 N 9 O 24 of a twofold fluorinated balhimycin, which we named fluorobalhimycin (Figure 1).
One-and two-dimensional NMR experiments (TOCSY, NOESY, HSQC, HMBC; Tables 1 and2) allowed the structure to be assigned as shown in Figure 1. The presence of two fluorine substituents in fluorobalhimycin was verified independently by 19 F NMR spectroscopy with d( 19 F) ¼ À131 and À134 ppm compared to d( 19 F) ¼ À139 ppm for the racemic amino acid 3-Fht. The positioning of the fluorine substituents in the biaryl ether ring systems is analogous to that of the chlorine atoms in balhimycin, [8,11] as confirmed by the observed NOE contacts.
In further experiments we evaluated the substrate specificity of the mutant OP696. To this end, we synthesized three sets of b-hydroxytyrosines (Scheme 2). The first set comprised bhydroxytyrosines with the phenolic hydroxy group in 2-, 3-, COMMUNICATIONS