Investigation into the functional impact of the vancomycin C-ring aryl chloride

Abstract

A vancomycin aglycon analogue that possesses a reduced C-ring and an intact E-ring chloride was prepared and its antimicrobial activity towards S. aureus and binding affinity to model cell wall ligands were established. Comparison of the derivative with a series of vancomycin aglycon analogues that possess and lack the chloro substituents on the aryl C- and E-rings defines the impact and further refines the role the C-ring chloride plays in promoting both target binding affinity and binding selectivity for d-Ala-d-Ala and its impact on antimicrobial activity.

Vancomycin (1) and structurally related glycopeptide natural products have proven to be invaluable antibiotics over the past fifty years.¹ Most recently, vancomycin is prescribed as the drug of last resort for the treatment resistant Gram-positive bacterial infections, including methicillin-resistant Staphylococcus aureus (MRSA).² Vancomycin disrupts bacterial cell wall biosynthesis by binding to the peptide terminus d-Ala-d-Ala found in peptidoglycan precursors through a network of five hydrogen bonds, thereby inhibiting cell wall cross-linking.^2,3 Alarmingly, vancomycin-resistant bacteria have emerged which are able to overcome this mode of action. In the more prevalent resistant bacterial strains, VanA and VanB, the d-Ala-d-Ala terminus is remodeled to d-Ala-d-Lac resulting in a 1000-fold decrease in antibiotic binding affinity and an accompanying 1000-fold loss of antimicrobial activity.^2–5 Additionally, several resistant strains (VanC, VanE, VanG) have been found to possess an altered d-Ala-d-Ser peptide terminus.^6–8 Although not as severe, these bacteria display reduced sensitivity to vancomycin antimicrobial activity, resulting from a reduced binding affinity for d-Ala-d-Ser.^5,9

Our efforts to overcome vancomycin bacterial resistance have enlisted both total and semi-synthesis studies and have provided several vancomycin analogues that are effective against resistant bacteria.^10–12 As a result of these studies, we have a strong interest in understanding the intricacies of the binding properties of vancomycin.¹³ n a recent report, we disclosed synthetic efforts capable of selective Pd(0)-catalyzed conversion of the E-ring aryl chloride in vancomycin aglycon derivatives to a boronic acid, permitting the synthesis of alternatively substituted E-ring analogues to probe the impact of this characteristic feature of the glycopeptides antibiotics.¹⁴ Many E-ring aryl chloride substitution analogues were well tolerated and it was the presence, but not the nature, of a substituent on the E-ring that was important for maintaining binding affinity for a terminal d-Ala-d-Ala model peptide and antimicrobial activity. Herein, we disclose an initial study on the impact and role of the C-ring aryl chloride through the synthesis and evaluation of the reduced C-ring vancomycin aglycon possessing the E-ring aryl chloride (3) and its comparison with a series of vancomycin aglycon analogues (2–5) that posses and lack the chloro substituents on the C and E-aryl rings (Figure 1, Table 1).

Figure 1.

Vancomycin and vancomycin aglycon analogues.

Table 1. Vancomycin Analogues.

Compound	R1	R2	R3
1	Sugar	Cl	Cl
2	H	Cl	Cl
3	H	H	Cl
4	H	Cl	H
5	H	H	H

Eremomycin, a naturally occurring glycopeptide also possesses a vancomycin heptapeptide scaffold with a reduced C-ring and intact E-ring aryl chloride.^1,15 However due to the presence of alternative amino-sugars on the periphery of the molecule, comparisons to vancomycin do not provide a clear indication as to the role of the C-ring chloride in binding and antimicrobial activity. Our studies on the selective borylation of the N-Boc vancomycin aglycon (6) revealed that a major byproduct from the reaction was the reduced C-ring aryl chloride aglycon possessing an E-ring aryl boronic acid (7, Figure 2).¹⁴ With this intermediate in hand, copper(II)-mediated chlorination and N-Boc deprotection provided the desired vancomycin aglycon analogue 3, capable of direct assessment of only the impact of the C-ring aryl chloride.

Figure 2.

(a) (HO)₄B₂ (3 eq), X-Phos-Pd-G2 (0.3 eq), KOAc (3 eq), X-Phos (0.6 eq), EtOH, 60 °C, 6 h, 19% 7. (b) CuCl₂, MeOH/H₂O (60 °C, 2 h) then TFA/CH2Cl2, (25 °C, 20 min), 52% (two steps) R¹ = R² = H, R³ = Cl, R⁴ = H.

For comparison, a complete series of vancomycin aglycon analogues that possess and lack the aryl chloride substituents was assembled. Compounds 4 and 5 which lacked the E-ring chloride and both C- and E-ring aryl chlorides, respectively, were prepared as previously described by Harris through hydrogenation of vancomycin aglycon.¹⁶

The antimicrobial activity of vancomycin and the vancomycin aglycon analogues 2–5 against sensitive S. aureus were obtained according to a standard microtiter plate-based antimicrobial assay (Table 2).^11,17 Ligand binding studies were performed according to the procedure of Nieto and Perkins with the ligands N,N′-Ac₂-l-Lys-d-Ala-X (X = d-Ala, d-Ser, Gly, d-Phe) and the results are summarized in Table 2.¹⁸ These ligands represent known (X = Gly) and additional possible vancomycin binding sites in the bacterial cell wall of resistant bacteria (X = d-Ser).

Table 2.

Binding and Antimicrobial Properties.

Compd	MIC (μg/mL)	Association constant for N,N′-Ac2-L-Lys-D-Ala-X
	S. aureusa	Ka, M−1 d-Ala (×105)	ΔΔGbobs,b	Ka, M−1 d-Ser (× 103)	Ka, M−1 Gly (× 104)	Ka, M−1 d-Phe (× 102)
1	1.25	1.8
2	1.25	1.4	-	3.3	1.0	9.1
3	10	0.38	0.77	4	0.49	9
4	5	0.68	0.42
5	20	0.18	1.21

^aVancomycin-sensitive Staphylococcus aureus (ATCC 25923).

^bThe free energy of binding (ΔΔG_b) was tabulated from the experimentally determined K_a using the equation ΔΔG_b = RTlnK, where K=¹K_a/^xK_a; ¹K_a is the association constant for the complex of 2 with N,N′-Ac₂-l-Lys-d-Ala-d-Ala; ^xK_a is the association constant for the complex of compound X with N,N′-Ac₂-l-Lys-d-Ala-d-Ala.

Removal of just the C-ring aryl chloride with 3 led to a >8-fold drop in activity against sensitive S. aureus compared to the vancomycin aglycon. This loss in activity was more substantial than the decrease in activity seen upon the selective removal of the E-ring aryl chloride in 4. The difference in antimicrobial activity can be explained through ligand binding studies with N,N′-Ac₂-l-Lys-d-Ala-d-Ala, which revealed that selective removal of the C-ring chloride with 3 led to an approximate 4-fold loss in binding, whereas the selective E-ring chloride removal with 4 led to a smaller 2-fold decrease in binding affinity. Interestingly, the loss of binding for 3 and 4 for d-Ala-d-Ala are additive and explain the poorer binding observed for the analogue 5, lacking both the C- and E-ring chlorides.

For the model d-Ala-d-Ser ligand, analogue 3 was slightly more effective at binding than the aglycon. Nevertheless, the binding affinity was reduced ca. 30-fold for both analogues. The lower binding with d-Ala-d-Ser accounts for the decreased antimicrobial activity of vancomycin on VanC, VanE, and VanG bacterial strains. For d-Ala-Gly, analogue 3 was 2-fold less effective than that of the vancomycin aglycon which is a slightly smaller decrease compared to that seen with d-Ala-d-Ala (ca. 3-fold). For the larger d-Ala-d-Phe ligand, it was possible that the removal of the C-ring aryl chloride would enable better binding due to decreased steric constraints. However, both 2 and 3 showed equally reduced binding towards the model d-Ala-d-Phe ligand.

These observations further refine the role of the C-ring aryl chloride in target binding selectivity and binding affinity. Its presence helps define the vancomycin binding pocket and increase binding affinity for terminal d-Ala-d-Ala ligands over alternative bacterial cell wall binding sites and its impact is larger than that of the E-ring chloride. Additionally, the comparative binding studies of 2 and 3 with model cell wall ligands provide a valuable insight into and starting point for evaluating C-ring analogues replacing the chloride with alternative substituents.