Multivalent drug design. Synthesis and in vitro analysis of an array of vancomycin dimers

Article abstract of DOI:10.1021/ja021273s

The design, synthesis, and in vitro microbiological analysis of an array of forty covalently linked vancomycin dimers are reported. This work was undertaken to systematically probe the impact of linkage orientation and linker length on biological activity

Full text of DOI:10.1021/ja021273s

Published on Web 05/03/2003
Multivalent Drug Design. Synthesis and In Vitro Analysis of
an Array of Vancomycin Dimers
John H. Griffin,* Martin S. Linsell, Matthew B. Nodwell, QiQi Chen, John L. Pace,
Kelly L. Quast, Kevin M. Krause, Lesley Farrington, Terry X. Wu,
Deborah L. Higgins, Thomas E. Jenkins, Burton G. Christensen, and
J. Kevin Judice
Contribution from TheraVance, Inc., 901 Gateway BouleVard,
South San Francisco, California 94080
Received October 15, 2002; E-mail: jgriffin@pharmix.com
Abstract: The design, synthesis, and in vitro microbiological analysis of an array of forty covalently linked
vancomycin dimers are reported. This work was undertaken to systematically probe the impact of linkage
orientation and linker length on biological activity against susceptible and drug-resistant Gram-positive
pathogens. To prepare the array, monomeric vancomycin synthons were linked through four distinct positions
of the glycopeptide (C-terminus (C), N-terminus (N), vancosamine residue (V), and resorcinol ring (R)) in
10 unique pairwise combinations. Amphiphilic, peptide-based linkers of four different lengths (11, 19, 27,
and 43 total atoms) were employed. Both linkage orientation and linker length were found to affect in vitro
antibacterial potency. The V-V series displayed the greatest potency against vancomycin-susceptible
organisms and vancomycin-resistant Enterococcus faecalis (VRE) of VanB phenotype, while the C-C,
C-V, and V-R series displayed the most promising broad-spectrum activity that included VRE of VanA
phenotype. Dimers bearing the shortest linkers were in all cases preferred for activity against VRE. The
effects of linkage orientation and linker length on in vitro potency were not uniform; for example, (1) no
single compound displayed activity that was superior against all test organisms to that of vancomycin or
the other dimers, (2) linker length effects varied with test organism, and (3) whereas one-half of the dimers
were more potent than vancomycin against methicillin-susceptible Staphylococcus aureus (MSSA), only
one dimer was more potent against methicillin-resistant S. aureus (MRSA) and glycopeptide-intermediate
susceptible S. aureus (GISA). In interpreting the results, we have considered the potential roles of
multivalency and of other phenomena.
Introduction
mature peptidoglycan that has been polymerized through the
action of transglycosylase enzymes but not yet cross-linked
The emergence and spread of antibiotic-resistant bacteria
present an ongoing challenge to the healthcare community.1 Of
particular concern has been the emergence of high-level
resistance to vancomycin and other glycopeptide antibiotics in
virulent Gram-positive pathogens, first among enterococci
(VRE)2 and, very recently, in Staphylococcus aureus that are
also resistant to methicillin (VRSA).3 The basis for the biological
effects exerted by vancomycin and other naturally occurring
glycopeptides derives from their ability to bind to and prevent
from use intermediates involved in the biosynthesis of bacterial
cell wall peptidoglycan.4 Specifically, glycopeptide antibiotics
are receptors capable of forming complexes with terminal
D-alanyl-D-alanine ligands that are displayed by lipid intermedi-
ate II (N-acetylglucosamine-â-1,4-(undecaprenyl diphospho-N-
acetylmuramyl-L-Ala-D-Glu-L-Lys-D-Ala-D-Ala)) and by im-
through the action of transpeptidases. Resistance to glycopep-
tides is most commonly associated with elaboration of altered
peptidoglycan precursors. For example, VanA VRE employ an
alternative biosynthetic pathway to generate peptidoglycan
precursors terminating in D-alanyl-D-lactate rather than D-alanyl-
D-alanine.5 This simple replacement of an amide NH group for
an ester oxygen results in the replacement of a stabilizing
hydrogen bond with a destabilizing lone pair-lone pair repulsion
within vancomycin/ligand complexes. This in turn leads to a
decrease in relative stability by a factor of 1000 (Figure 1).6
Multiple approaches have been taken to address the challenges
posed by vancomycin-resistant bacteria. First, novel antibiotics
that act against targets other than cell wall biosynthesis have
(4) (a) Reynolds, P. E. Biochim. Biophys. Acta 1961, 52, 403-405. (b) Perkins,
H. R.; Nieto, M. Ann. N.Y. Acad. Sci. 1974, 235, 348-363. (c) Barna, J.
C. J.; Williams, D. H. Annu. ReV. Microbiol. 1984, 38, 339-357. (d)
Reynolds, P. E. Eur. J. Clin. Microbiol. Infect. Dis. 1989, 8, 943-950. (e)
Nicas, T. I.; Allen, N. E. In Glycopeptide Antibiotics; Nagarajan, R., Ed.;
Marcel Dekker: New York, 1994; pp 219-241. (f) Williams, D. H.;
Bardsley, B. Angew. Chem., Int. Ed. 1999, 38, 1172-1193.
(1) Neu, H. C. Science 1992, 257, 1064-1073.
(2) (a) Leclerq, R. E.; Derlot, R. E.; Duval, J.; Courvalin, P. N. Engl. J. Med.
1988, 319, 157-161. (b) Uttley, A. H. C.; Collins, C. H.; Naidoo, J.;
George, R. C. Lancet 1988, 1, 57-58.
(3) Sievert, D. M.; Boulton, M. L.; Stoltman, G.; Johnson, D.; Stobierski, M.
G.; Downes, F. P.; Somsel, P. A.; Rudrik, J. T.; Brown, W.; Hafeez, W.;
Lundstrom, T.; Flanagan, E.; Johnson, R.; Mitchell, J.; Chang, S. Morbid.
Mortal. Weekly Rep. 2002, 51, 565-567.
(5) Walsh, C. T.; Fisher, S. L.; Park, I.-S.; Zhu, W. Chem. Biol. 1996, 3, 21-28.
(6) Popieniek, P. H.; Pratt, R. F. Anal. Biochem. 1987, 165, 108-113.
9
10.1021/ja021273s CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 6517-6531
6517

A R T I C L E S
Griffin et al.
via hydrogen bonding and hydrophobic interactions in an
antiparallel, back-to-back fashion to form specific dimers.
Noncovalent dimerization of glycopeptides can be both highly
favorable and cooperative with the binding of peptide ligands
and has been correlated with enhanced bacterial cell surface
binding and in vitro antibacterial potency.4f,15 Vancomycin self-
associates only weakly in solution (Kdim ) 700 M-1),13b which
prompted the question of whether covalently linked dimers of
this glycopeptide would display regained potency against VRE.16
This hypothesis has been confirmed through the synthesis and
study of vancomycin dimers linked through the C-terminal
carboxyl group16-18 as well as dimers of vancomycin and
chloroeremomycin linked via their aminodisaccharide moi-
eties.19,20 A small number of heterodimers comprising two
different glycopeptide subunits,19,20 the synthesis of vancomycin
dimers linked via the C-terminus of one subunit and the terminal
N-methyl group of a second subunit,21 and a single example of
a covalently linked dimer of eremomycin22 have also been
reported.
Figure 1. Structure of vancomycin (1) complexed to terminal D-Ala-D-
Ala and D-Ala-D-Lac segments of ligands displayed by vancomycin-
susceptible and -resistant bacteria, respectively. Hydrogen bonds are
indicated as dotted lines. Also shown are various chemical modifications
in synthons 2-5 (Scheme 1) used to prepare vancomycin dimers linked
through the carboxyl terminus (C), the amino terminus (N), the vancosamine
amino group (V), and the resorcinol side chain of amino acid residue 7
(R). The box presents a simplified depiction of vancomycin where the
heptapeptide core is abbreviated as a diamond shape but with the functional
groups at C-, N-, V- and R-positions explicated.
In this paper, we report the design, synthesis, and in vitro
biological evaluation of an array of forty covalently linked
vancomycin dimers. This work was undertaken in order to
systematically explore the structure-activity relationships among
these compounds in the service of antibiotic lead discovery and
as a platform on which to develop and illustrate the unique
aspects of multivalent drug design.
been developed. These include the ribosome-targeting protein
synthesis inhibitors quinupristin/dalfopristin (Synercid) and
linezolid (Zyvox), as well as the membrane-targeting antibiotic
daptomycin.7 A second approach has been to modify existing
glycopeptides in order to regain activity against resistant
bacterial strains.8 Here, the most promising results have been
obtained through addition of hydrophobic groups to the amino-
disaccharide moieties of vancomycin and (chloro)eremomycin.9
A third approach has sought to improve the affinity of
vancomycin for depsipeptide-containing glycopeptide precursors
through multivalency. Multivalency is a phenomenon wherein
multiple, simultaneous, energetically coupled receptor/ligand
interactions enhance the overall affinity and selectivity of
binding. Natural systems have evolved to make extensive use
of multivalency, for example, in the functioning of immuno-
globulins, in the adhesion of virus particles to target cells, and
in controlling cell-cell interactions.10 A diffuse yet growing
literature suggests that multivalency may also find general
applicability in the design of small molecule ligands,11 including
novel drugs for enzymes, receptors, transporters, and macro-
molecular structures representing the major classes of therapeutic
targets.12
Results
Multivalent Design. Our design process took into account
the singular considerations of multivalent design, including the
following:
(12) (a) Griffin, J. H.; Judice, J. K. WO 99/64037, 1999. (b) Marquess, D.;
Mammen, M.; Griffin, J. H. WO 99/66944, 1999. (c) Jenkins, T. E.;
Christensen, B. G.; Griffin, J. H. WO 99/64045, 1999. (d) Griffin, J. H.;
Moran, E. J.; Oare, D. A. WO 99/64036, 1999.
(13) (a) Waltho, J. P.; Williams, D. H. J. Am. Chem. Soc. 1989, 111, 2475-
2480. (b) Gerhard, U.; Mackay, J. P.; Maplestone, R. A.; Williams, D. H.
J. Am. Chem. Soc. 1993, 115, 232-237. (c) Mackay, J. P.; Gerhard, U.;
Beauregard, D. A.; Maplestone, R. A.; Williams, D. H. J. Am. Chem. Soc.
1994, 116, 4573-4580. (d) Mackay, J. P.; Gerhard, U.; Beauregard, D.
A.; Westwell, M. S.; Searle, M. S.; Williams, D. H. J. Am. Chem. Soc.
1994, 116, 4581-4590.
(14) Reviewed in: Loll, P. J.; Axelsen, P. H. Annu. ReV. Biophys. Biomol. Struct.
2000, 29, 265-289.
(15) (a) Beauregard, D. A.; Williams, D. H.; Gwynn, M. N.; Knowles, D. J. C.
Antimicrob. Agents Chemother. 1995, 39, 781-785. (b) Beauregard, D.
A.; Maguire, A. J.; Williams, D. H.; Reynolds, P. E. Antimicrob. Agents
Chemother. 1997, 41, 2418-2423. (c) Westwell, M. S.; Bardsley, B.;
Dancer, R. J.; Try, A. C.; Williams, D. H. Chem. Commun. 1996, 589-
590. (d) Allen, N. E.; LeTourneau, D. L.; Hobbs, J. N., Jr. Antimicrob.
Agents Chemother. 1997, 41, 66-71. (e) Cooper, M. A.; Williams, D. H.;
Cho, Y. R. Chem. Commun. 1997, 1625-1626. (f) Entress, R. M. H.;
Dancer, R. J.; O′Brien, D. P.; Try, A. C.; Cooper, M. A.; Williams, D. H.
Chem. Biol. 1998, 5, 329-337. (g) O′Brien, D. P.; Entress, R. M. H.;
Cooper, M. A.; O′Brien, S. W.; Hopkinson, A.; Williams, D. H. J. Am.
Chem. Soc. 1999, 121, 5259-5265. (h) Allen, N. E.; LeTourneau, D. L.;
Hobbs, J. N., Jr.; Thompson, R. C. Antimicrob. Agents Chemother. 2002,
46, 2344-2348.
A series of studies in solution13 and in the solid state14 have
shown that vancomycin and other glycopeptides self-associate
(7) Bush, K.; Macielag, M.; Clancy, J. Emerging Drugs 2000, 5, 347-365.
(8) Malabarba, A.; Nicas, T. I.; Thompson, R. C. Med. Res. ReV. 1997, 17,
69-137.
(9) (a) Nagarajan, R.; Schabel, A. A.; Occolowitz, J. L.; Counter, F. T.; Ott,
J. L.; Felty-Duckworth, A. M. J. Antibiot. 1989, 42, 63-72. (b) Nagarajan,
R. J. Antibiot. 1993, 46, 1181-1195. (c) Nicas, T. I.; Mullen, D. L.;
Flokowitsch, J. E.; Preston, D. A.; Snyder, N. J.; Stratford, R. E.; Cooper,
R. D. G. Antimicrob. Agents Chemother. 1995, 39, 2585-2587. (d) Cooper,
R. D. G.; Snyder, N. J.; Zweifel, M. J.; Staszak, M. A.; Wilkie, S. C.;
Nicas, T. I.; Mullen, D. L.; Butler, T. F.; Rodriguez, M. J.; Huff, B. E.;
Thompson, R. C. J. Antibiot. 1996, 49, 575-581. (e) Printsevskaya, S. S.;
Pavlov, A. Y.; Olsufyeva, E. N.; Mirchink, E. P.; Isakova, E. B.; Reznikova,
M. I.; Goldman, R. C.; Branstrom, A. A.; Baizman, E. R.; Longley, C. B.;
Sztaricskai, F.; Batta, G.; Preobrazhenskaya, M. N. J. Med. Chem. 2002,
45, 1340-1347.
(16) Sundram, U. N.; Griffin, J. H.; Nicas, T. I. J. Am. Chem. Soc. 1996, 118,
13107-13108.
(17) (a) Rao, J.; Whitesides, G. M. J. Am. Chem. Soc. 1997, 119, 10286-10287.
(b) Rao, J.; Yan, L.; Xu, B.; Whitesides, G. M. J. Am. Chem. Soc. 1999,
121, 2629-2630. (c) Rao, J.; Yan, L.; Lahiri, J.; Whitesides, G. M.; Weis,
R. M.; Warren, H. S. Chem. Biol. 1999, 6, 353-359.
(18) Adamczyk, M.; Moore, J. A.; Rege, S. D.; Yu, Z. Bioorg. Med. Chem.
Lett. 1999, 9, 2437-40.
(19) Stack, D. R.; Thompson, R. G. EP 0801075 A1, 1997.
(20) (a) Nicolaou, K. C.; Hughes, R.; Cho, S. Y.; Winssinger, N.; Smethurst,
C.; Labischinski, H.; Endermann, R. Angew. Chem., Int. Ed. 2000, 39,
3823-3828. (b) Nicolaou, K. C.; Hughes, R.; Cho, S. Y.; Winssinger, N.;
Labischinski, H.; Endermann, R. Chem.sEur. J. 2001, 7, 3824-3843.
(21) Staroske, T.; Williams, D. H. Tetrahedron Lett. 1998, 39, 4917-4920.
(22) Sztaricskai, F.; Batta, G.; Dinya, Z.; Miroshnikova, O. V.; Preobrazhenskaya,
M. N.; Hernadi, F.; Koncz, A.; Boda, Z. J. Antibiot. 2001, 54, 314-319.
(10) Mammen, M.; Choi, S.-K.; Whitesides, G. M. Angew. Chem., Int. Ed. 1998,
37, 2754-2794.
(11) (a) Portoghese, P. S. J. Med. Chem. 2001, 44, 1-11. (b) Wright, D.; Usher,
L. Curr. Org. Chem. 2001, 5, 1107-1131.
9
6518 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Synthesis and Analysis of an Array of Vancomycin Dimers
A R T I C L E S
Chart 1. Depictions of the 10 Unique Vancomycin Dimer Linkage
Orientations Produced by the Coupling of Vancomycin Synthons
2-5 (Scheme 1) through the C-, N-, V-, and R-Positions
A. Which Glycopeptide(s) to Use as Monomeric Subunits.
We chose to work with vancomycin, the best-studied of the
200+ known glycopeptides.23 To simplify the synthesis of
compounds and interpretation of data, we chose to study only
dimers of vancomycin derivatives rather than to include
heteromeric compounds comprising more than one distinct
member of the glycopeptide family.
B. Valency of the Constructs. We chose to focus on dimeric
constructs rather than higher-order structures in light of the prior
art with dimers and in order to restrict molecular weights to
below 5 kDa.
C. Linkage Orientation. The choice of linkage site dictates
the relative orientations in which individual vancomycin subunits
may be presented to their targets. We chose to link vancomycin
subunits through four distinct sites: the carboxyl- (C) and
amino- (N) termini of the dalbaheptide chain, along with the
amino group of the vancosamine moiety (V) and an amino group
that could be regioselectively installed at the 4′, resorcinol-like
(R) position on the aromatic side chain of amino acid 7. These
positions were chosen because they allow for linkage through
sites around the perimeter of the vancomycin molecule: on the
top (V), bottom (R), left (C), and right (N) of vancomycin when
the glycopeptide is considered as oriented in Figure 1. Ten
distinct linkage orientations are possible with these four distinct
sites of attachment (Chart 1): C-C, C-N, C-V, C-R, N-N,
N-V, N-R, V-V, V-R, R-R.
D. Linkage Chemistries. Reductive alkylation chemistry was
chosen for attachment of the linkers at the N- and V-positions
in order to preserve their basic character, which has been
previously shown to be optimal for the antibacterial activity of
vancomycin and chloroeremomycin.8,9 Amide bond formation
at the C-terminus of glycopeptides including vancomycin,24
eremomycin,25 and teicoplanin26 provides derivatives that typi-
cally retain their antibacterial activity in full. It has also been
shown that Mannich aminomethylation of vancomycin at the
R-position provides compounds with in vitro activity similar
to that of the parent compound.8,27 Modified vancomycin
synthons 2-5 which provide for these linkage chemistries are
depicted in Scheme 1 along with methods by which they were
synthesized.
E. Linker Characteristics (Length, Geometry, Composi-
tion, and Physical Properties). To assist in choosing linker
lengths, we used two space-filling models of vancomycin
arranged in the self-associated, back-to-back geometry observed
for vancomycin and other glycopeptides in solution and in the
solid state.13,14 Linkers 11, 19, 27, and 43 atoms in length (Chart
2) were ultimately chosen to enforce, allow, or preclude this
mode of self-association to the extent possible. For reference,
a 19-atom linker in its fully extended conformation is ap-
proximately equal to the width and height dimensions of
vancomycin when the glycopeptide is considered in the orienta-
tion presented in Figure 1. The following lists our estimates of
the minimum linker lengths needed to span the two linkage sites
in the 10 distinct linkage series without interfering with ligand
binding. In the case of some V-linked series, two lengths are
given because of the nonidentical conformations adopted by the
disaccharide moieties of the two independent vancomycin
subunits. Where informative, we also assign a preferred path
for linkage, whether it be over the top or underneath the
glycopeptide as oriented in Figure 1: for the C-C series, 27
atoms; C-N, 11 atoms; C-V, 11 or 27 atoms; C-R, 19 atoms
(underneath); N-N, 27 atoms (underneath); N-V, 19 or 27
(23) Nagarajan, R. Glycopeptide Antibiotics. Marcel Dekker: New York, 1994.
(24) (a) Shi; Z.; Griffin, J. H. J. Am. Chem. Soc. 1993, 115, 6482-6486. (b)
Sundram, U. M.; Griffin, J. H. J. Org. Chem. 1995, 60, 1102-1103.
(25) (a) Pavlov, A. Y.; Berdinikova, F.; Olsufyeva, E.; Miroshnikova, O. V.;
Filipposyanz, S.; Preobrazhenskaya, M. N. J. Antibiot. 1996, 49, 194-
198. (b) Miroshnikova, O. V.; Printsevskaya, S. S.; Olsufyeva, E. N.;
Pavlov, A. Y.; Nilius, A.; Hensey-Rudloff, D.; Preobrazhenskaya, M. N.
J. Antibiot. 2000, 53, 286-293.
(26) (a) Malabarba, A.; Trani, A.; Tarzia, G.; Ferrari, P.; Pallanza, R.; Berti,
M. J. Med. Chem. 1989, 32, 2450-2460. (b) Malabarba, A.; Ciabatti, R.;
Kettenring, J.; Scotti, R.; Candiani, G.; Pallanza, R.; Berti, M.; Goldstein,
B. J. Med. Chem. 1992, 35, 4054-4060.
(27) Pavlov, A. Y.; Lazhki, E. I.; Preobrazhenskaya, M. N. J. Antibiot. 1997,
50, 509-513.
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6519

A R T I C L E S
Griffin et al.
Scheme 1. Simplified Depictions of Synthons 2-5 and Schemes for Their Synthesisa
a Key to reagents: (a) 9-fluorenylmethyl N-(2-aminoethyl)carbamate hydrochloride/PyBOP/HOBT/ DIPEA/DMPU; (b) piperidine/DMF; (c) N-(9-
fluorenylmethoxycarbonyl)glycinal/HOBT/ DIPEA/NaCNBH3/MeOH/DMPU; (d) 3-(dimethylamino)propylamine/PyBOP/HOBT/DIPEA/ DMF; (e) quinu-
clidine/DMF; (f) N-(9-fluorenylmethoxycarbonyl)glycinal/DIPEA/TFA/BH3-pyridine/DMF/MeOH; (g) ethylenediamine/formaldehyde/water/ACN.
Chart 2. Linker Synthons A-D Used to Couple Synthons 2-5 to
Produce Dimers Linked by 11, 19, 27, or 43 Atoms
impacts antibacterial potency, for example, imparting activity
against VRE at low µg/mL levels.8,9
Synthesis. The covalently linked vancomycin dimers were
synthesized through sequential, directed amide bond forming
reactions. This was made possible by the use of the modified
vancomycin synthons 2-5 bearing uniquely reactive, sterically
unhindered primary amines in conjunction with the terminally
differentiated dicarboxylic acid, mono-ester linker synthons
A-D (Chart 2). In the case of compounds linked through the
N-, V-, and/or R-positions (synthons 3-5), the C-terminal
carboxyl was capped as a 3-dimethylaminopropylamide in order
to prevent its interference with subsequent amide bond coupling
reactions. By way of example, the N-R linked dimer joined
by a 27-atom linker (N-R(27)) was prepared as outlined in
Scheme 2.
atoms; N-R, 11 atoms (underneath); V-V, 11 atoms; V-R,
19 or 27 atoms; R-R, 19 atoms (underneath).
Coupling of R-linkage synthon 5 with linker synthon C
followed by removal of the fluorenylmethyl ester afforded an
intermediate 5-C conjugate. Conjugate 5-C was then coupled
with N-linkage synthon 3 to afford the N-R(27) dimer. Details
of the synthesis of 2-5, the solution- and solid-phase synthesis
of linker synthons A-D, and an example of the synthesis of
one member of each of the 10 distinct dimer linkage series are
described in the Experimental Section. The dimers are cationic
in nature, bearing from four (C-C series) to eight basic centers
(R-R series).
Bacterial Strains and In Vitro Susceptibility Studies.
Bacterial strains were selected to represent the major Gram-
positive human pathogens (streptococci, staphylococci, and
enterococci) and various forms of drug resistance, including
resistance to vancomycin: (1) a strain of penicillin-resistant
Streptococcus pneumoniae (PRSP SU-10), (2) a strain of
methicillin-susceptible Staphylococcus aureus (MSSA 13709),
(3) a strain of methicillin-resistant Staphylococcus aureus
(MRSA 33591), (4) a strain of methicillin-resistant Staphylo-
coccus aureus that displays an intermediate level of resistance
We also estimated the minimum linker lengths needed to
bridge attachment points without disrupting ligand binding when
two vancomycin subunits were juxtaposed face-to-face as
observed for vancomycin/ligand complexes in the solid
state:14,28 for the C-C series, 43 atoms (underneath); C-N, 11
atoms; C-V, 11 or 19 atoms; C-R, 27 atoms (underneath);
N-N, 19 atoms (underneath); N-V, 27 atoms; N-R, 11 atoms;
V-V, 19 atoms; V-R, 19 or 27 atoms; R-R, 11 atoms
(underneath).
We chose neutral yet polar, peptide-based linkers in order to
avoid substantial changes in the physical properties of the dimers
with changes in linker length (Chart 2). These linkers comprise
multiple rotatable bonds and, in conjunction with linker length,
allow the dimeric constructs to adopt different conformations.
We specifically avoided the use of hydrophobic linkers given
that addition of hydrophobic appendages to the vancosamine
residue or the C- and N-termini of the glycopeptide significantly
(28) (a) Loll, P. J.; Bevivino, A. E.; Korty, B. D.; Axelsen, P. H. J. Am. Chem.
Soc. 1997, 119, 1516-1522. (b) Loll, P. J.; Miller, R.; Weeks, C. M.;
Axelsen, P. H. Chem. Biol. 1998, 5, 293-298.
9
6520 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Synthesis and Analysis of an Array of Vancomycin Dimers
A R T I C L E S
Scheme 2. Directed Synthesis of the N-R(27) Dimera
tions allow for convenient visual analysis of the impact of
variations in key structural parameters of vancomycin dimers
on biological activity. In cases where only limiting experimental
data were obtained (e.g., MIC < 0.1 µg/mL or > 200 µg/mL),
the limiting values were used to make the graphs and in linear
regression analyses. For reference, the data for vancomycin is
also included in each figure panel.
Activity as a Function of Linkage Orientation and Linker
Length. A. C-C Series. Relative to vancomycin, compounds
in this series were as much as 30-fold more potent against PRSP,
ESFVS, and the three VRE strains. C-C dimers were either
slightly more or slightly less potent than vancomycin against
MSSA. All compounds in this series were less potent than
vancomycin against MRSA and GISA, by 3- to 8-fold. Greatest
absolute potency was observed for all members of the series
against PRSP. Lowest absolute potency was observed for
C-C(27) and C-C(43) against VRE. Against the vancomycin-
susceptible organisms, potency either was independent of linker
length or, in the case of MSSA, was greatest at an intermediate
linker length of 27 atoms. In contrast, potency against the three
VRE strains decreased substantially with increasing linker
length.
a Key to reagents: (a) R-linkage synthon 5/linker synthon C/PyBOP/
HOAT/DIPEA/DMF; (b) quinuclidine/DMF; (c) N-linkage synthon 3/Py-
BOP/HOAT/DIPEA/DMF.
B. C-N Series. Relative to vancomycin, compounds in this
series were as much as 20-fold more potent against PRSP and
ESFVS, and some members of the series displayed measurable
activity against EFSVB and EFMVA but not EFSVA. C-N
dimers were also up to 4-fold more potent against MSSA. All
compounds in this series were less potent than vancomycin
against MRSA and GISA, by up to 12-fold. Greatest absolute
potency was observed for C-N(11) and C-N(19) against PRSP.
Lowest absolute potency was observed against EFSVA. Potency
was greatest at shorter linker lengths for PRSP, EFSVS, and
EFSVB and was either independent of linker length or maximal
at intermediate length against the staphylococci.
to glycopeptides (GISA CDC-5836), (5) a strain of vancomycin-
and beta-lactam-susceptible Enterococcus faecalis (EFSVS
29212), (6) a strain of vancomycin-resistant Enterococcus
faecalis of VanB phenotype (EFSVB 51575), (7) a strain of
vancomycin-resistant Enterococcus faecium of VanA phenotype
that is also highly resistant to beta-lactams (EFMVA KPB-01),
and (8) a strain of vancomycin-resistant Enterococcus faecalis
of VanA phenotype (EFSVA MGH-01).
The susceptibilities of these strains to vancomycin, the
monomer synthons 2-5, and the forty vancomycin dimers were
determined by standard broth microdilution assays. The collected
minimum inhibitory concentrations (MICs) are presented in
Table 1. Here, it may be seen that against six of the eight test
strains (MSSA, MRSA, GISA, EFSVS, EFMVA, EFSVA), no
significant difference in potency is displayed by synthons 2-5
relative to vancomycin. Synthons 2-5 are 2.5- to 11-fold less
potent than vancomycin against PRSP. In contrast, against the
VanB strain of VRE tested, EFSVB, synthons 2 and 4 exceed
the potency of vancomycin by factors in excess of 50 and 6,
respectively.
The data for the covalently linked vancomycin dimers
presented in Table 1 is coded such that MIC values for dimers
that are at least 2-fold lower than those for vancomycin against
the same organism are shown in bold type, while values that
differ by less than a factor of 2 or are at least 2-fold higher
than those for vancomycin are shown in normal or italic type,
respectively. Of the 320 MIC values for dimers, 144 are at least
2-fold lower than the corresponding values for vancomycin.
Visual inspection of the data coded in this way reveals that the
effects of dimerization on in vitro potency are not uniform across
all test organisms. For example, whereas many to most of the
dimers are more potent than vancomycin against PRSP, MSSA,
EFSVS, EFSVB, and EFMVA, only the V-V(11) dimer is even
modestly more potent against MRSA and GISA.
C. C-V Series. Relative to vancomycin, compounds in this
series were as much as 30-fold more potent against PRSP,
MSSA, ESFVS, and the three VRE strains. The C-V(27) dimer
was equipotent against MRSA, and the C-V(19) dimer was
equipotent against GISA; otherwise, C-V dimers were less
potent than vancomycin against these organisms. Greatest
absolute potency was observed for C-V(11), C-V(19), and
C-V(27) against PRSP and for C-V(11) against EFSVS.
Lowest absolute potency was observed for C-V(19) and
C-V(43) against VanA VRE. Potency was greatest at an
intermediate linker length (19 atoms) against GISA but inde-
pendent of linker length against all other vancomycin-susceptible
organisms. Potency against the three VRE strains decreased
substantially between the shortest and longest linker lengths.
D. C-R Series. Relative to vancomycin, compounds in this
series were as much as 15-fold more potent against PRSP,
MSSA, and ESFVS, and some members of the series were
potent against EFSVB and EFMVA but not EFSVA. The
C-R(27) dimer was equipotent against MRSA, and the
C-R(19) dimer was equipotent against GISA; otherwise, C-R
dimers were less potent than vancomycin against these organ-
isms. Greatest absolute potency was observed against PRSP and
EFSVS. Lowest absolute potency was observed against VanA
VRE. Potency was greatest at intermediate linker lengths against
MRSA and GISA, independent of linker length against PRSP,
The data for the covalently linked dimers are presented in
graphical form as a function of linkage orientation in Figure 2
and as a function of test organism in Figure 3. These presenta-
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6521

A R T I C L E S
Griffin et al.
Table 1. Analytical and In Vitro Susceptibility Data for Vancomycin (1), Monomer Synthons 2-5, and Vancomycin Dimers
e
MIC (µg/mL)
d
linkage
orientation
linker
lengtha
m/z
(calcd)b
m/z
(obsd)c
HPLC
rt
PRSP
MSSA
MRSA
GISA
EFSVS
EFSVB
EFMVA
EFSVA
1
2
3
4
NA
NA
NA
NA
NA
11
19
27
43
11
19
27
43
11
19
27
43
11
19
27
43
11
19
27
43
11
19
27
43
11
19
27
43
11
19
27
43
11
19
27
43
11
19
27
43
0.58
1.6
6.3
2.3
4.7
<0.1
<0.1
0.1
<0.15
<0.1
<0.1
0.2
0.39
<0.15
<0.15
<0.15
0.2
0.39
0.2
0.2
0.39
0.2
0.78
0.78
2.3
0.78
0.78
2.3
0.78
2.3
1.2
1.2
0.3
0.59
1.2
1.6
1.6
3.1
1.6
3.1
9.4
9.4
7.8
4.7
4.7
6.3
2.3
4.7
3.1
9.4
1.2
2.3
4.7
2.3
1.2
4.7
7.8
4.7
4.7
19
1.6
1.6
3.6
4.7
1.6
4.7
2.3
3.1
0.4
0.9
1.2
3.1
3.1
2.3
1.6
3.1
9.4
3.1
3.1
9.4
6.3
2.3
6.3
3.1
6.3
50
50
50
50
75
25
25
50
19
6.3
13
25
50
6.3
13
25
38
38
75
200
19
19
50
50
25
25
25
75
3.1
4.7
6.3
13
19
13
15
25
50
38
19
25
3.1
3.1
6.3
3.1
3.1
0.3
0.3
0.2
0.4
0.4
0.3
0.4
1.2
0.2
<0.1
0.2
0.4
0.4
0.2
0.2
0.3
1.2
3.1
3.1
6.3
0.3
0.4
0.8
0.8
1.2
1.2
0.8
1.6
0.2
<0.1
0.2
0.3
0.8
0.8
0.8
0.6
1.6
0.8
1.2
0.6
>200
4.7
200
38
100
6.3
6.3
>200
>200
>200
>200
>200
9.4
15
75
>150
120
>200
>200
>200
>200
>200
38
1492.3
1577.5
1577.5
1606.5
1539.0
1610.5
1681.5
1823.5
1582.0
1653.0
1724.1
1866.3
1581.9
1653.0
1724.1
1866.2
1596.8
1667.6
1738.6
1880.9
1624.5
1695.0
1766.7
1908.8
1624.5
1695.0
1766.7
1908.8
1639.0
1710.1
1781.2
1923.4
1624.5
1695.0
1766.7
1908.8
1639.0
1710.1
1781.2
1923.4
1653.6
1724.6
1795.7
1937.9
1492.2
1577.7
1577.7
1606.9
1540.2
1611.3
1682.2
1824.5
1582.8
1654.0
1724.8
1867.7
1582.5
1653.0
1724.5
1867.4
1597.2
1668.4
1739.5
1881.8
1625.4
1696.6
1767.7
1909.7
1625.4
1696.3
1767.1
1909.7
1639.8
1711.3
1782.1
1923.5
1625.1
1696.6
1767.7
1909.1
1639.8
1711.3
1781.8
1924.7
1654.9
1725.7
1796.5
1939.1
5
C-C
C-C
C-C
C-C
C-N
C-N
C-N
C-N
C-V
C-V
C-V
C-V
C-R
C-R
C-R
C-R
N-N
N-N
N-N
N-N
N-V
N-V
N-V
N-V
N-R
N-R
N-R
N-R
V-V
V-V
V-V
V-V
V-R
V-R
V-R
V-R
R-R
R-R
R-R
R-R
4.30
4.27
4.22
4.02
3.90
4.01
3.98
3.39
3.99
3.15
3.92
3.92
3.95
3.70
3.70
3.83
4.15
4.15
4.04
4.03
3.78
3.68
3.89
3.94
3.95
3.65
3.75
3.84
3.37
2.78
3.63
3.79
3.84
3.76
3.75
3.78
3.63
3.49
3.47
3.67
50
13
>200
>200
200
>200
>200
>200
38
>150
19
50
0.2
75
0.39
0.78
0.39
0.15
0.2
0.39
0.59
0.2
200
>200
4.7
>200
>200
19
>150
200
>200
13
150
75
38
>200
75
100
7.3
150
>200
200
50
100
150
>200
6.3
>200
>150
>150
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
150
>200
>150
>200
>200
>200
>200
>200
75
0.2
200
0.39
0.59
0.78
1.2
3.1
0.3
>200
58
150
>200
>200
19
0.2
0.2
0.3
38
100
0.39
0.59
0.59
0.78
0.59
1.6
<0.1
<0.1
<0.15
0.30
0.78
0.78
0.39
0.59
2.3
0.78
0.78
1.6
0.39
0.39
1.2
0.15
0.59
0.2
>150
>200
13
> 150
>200
38
75
150
>200
150
170
>200
>200
9.4
38
150
19
75
200
<0.2
100
150
150
4.7
6.3
25
100
13
25
0.2
0.59
0.59
1.2
0.39
1.6
0.39
0.39
0.78
200
>200
>200
150
>200
>200
>150
>200
9.4
13
1.2
1.2
1.2
>200
>150
150
>150
a Total linker length in atoms. NA ) Not applicable. b Calculated mass-to-charge ratios for singly charged parent ions for linkage synthons 2-5 ([M +
H+]+) and for doubly charged parent ions for the dimers ([M + 2H+]2+). c Observed mass-to-charge ratios for singly charged parent ions for linkage
synthons 2-5 ([M + H+]+) and for doubly charged parent ions for the dimers ([M + 2H+]2+). d Observed retention time in minutes in reversed-phase
high-performance liquid chromatography carried out as described in the Experimental Section. e Minimum inhibitory concentration in micrograms per milliliter
as determined by microdilution broth assays in vitro. Values for dimers indicated in bold type are lower than the corresponding values for vancomycin.
Values for dimers in normal type are equal to the corresponding values for vancomycin. Values in italic are higher than the corresponding values for
vancomycin. Strains are abbreviated as follows: PRSP, penicillin-resistant Streptococcus pneumoniae strain SU-10; MSSA, methicillin-susceptible S. aureus
strain ATCC 13709; MRSA, methicillin-resistant S. aureus strain ATCC 33591; GISA, glycopeptide-intermediate resistant S. aureus strain CDC-5836;
EFSVS, vancomycin-susceptible E. faecalis strain ATCC 29212; EFSVB, vancomycin-resistant E. faecalis strain ATCC 51575 (VanB phenotype); EFMVA,
vancomycin-resistant Enterococcus faecium strain KPB-01 (VanA phenotype); EFSVA, vancomycin-resistant Enterococcus faecalis strain MGH-01 (VanA
phenotype).
MSSA, and EFSVS, and greatest at the shortest linker length
against EFSVB and EFMVA.
MRSA, where potency appeared to be greatest at intermediate
linker lengths.
E. N-N Series. The N-N(11) dimer was the only member
of this series that displayed even modestly enhanced potency
against PRSP and EFSVS relative to vancomycin. Some
compounds displayed measurable activity against EFSVB and
EFMVA but not EFSVA. All compounds in this series were
less potent than vancomycin against MRSA and GISA. Greatest
absolute potency was observed for N-N(11) against PRSP.
Lowest absolute potency was observed against EFSVA. Potency
was greatest at shorter linker lengths against all organisms except
F. N-V Series. Relative to vancomycin, compounds in this
series were as much as 10-fold more potent against PRSP,
MSSA, and ESFVS. N-V linked dimers were equipotent to as
much as 8-fold less potent against MRSA and GISA. Some
compounds displayed measurable activity against EFSVB and
EFMVA but not EFSVA. Greatest absolute potency was
observed for N-V(11) and N-V(19) against PRSP, MSSA,
and EFSVS. Lowest absolute potency was observed against
EFSVA. Potency was greatest at short linker lengths for PRSP,
9
6522 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Synthesis and Analysis of an Array of Vancomycin Dimers
A R T I C L E S
Figure 2. In vitro susceptibility data presented as a function of linkage orientation. In each plot, minimum inhibitory concentrations (MICs) are depicted
as cones along the y-axis, the eight bacterial test strains are arranged along the x-axis, and linker length increases going back along the z-axis.
MRSA, EFSVB, and EFMVA but was independent of linker
length for MSSA, GISA, and EFSVS.
in one case, EFSVA. Greatest absolute potency was observed
against PRSP, MSSA, and EFSVS. Lowest absolute potency
was observed against EFSVA. Potency was greatest at short
linker lengths for GISA and the three VRE strains, at intermedi-
ate linker lengths against MSSA, but was independent of linker
length for PRSP, MRSA, and EFSVS.
G. N-R Series. Relative to vancomycin, compounds in this
series were equipotent to slightly less potent against PRSP,
slightly more to less potent against MSSA, and more potent
against EFSVS. N-R linked dimers were equipotent to as much
as 12-fold less potent against MRSA and GISA. Some com-
pounds displayed potent activity against EFSVB, EFMVA, and,
H. V-V Series. V-V linked dimers were generally more
potent than vancomycin against PRSP, MSSA, and EFSVS with
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6523

A R T I C L E S
Griffin et al.
Figure 3. In vitro susceptibility data presented as a function of bacterial test strain. In each plot, minimum inhibitory concentrations (MICs) are depicted
as cones along the x-axis, the 10 dimer linkage series are arranged along the x-axis, and linker length increases going back along the z-axis.
MIC values lower by up to a factor of more than 30. The
V-V(11) compound was the only one of the forty dimers that
displayed even modestly enhanced potency against MRSA and
GISA. Other compounds in this series were equipotent to 2-fold
less potent against these organisms. Against EFSVB, the
compound bearing the shortest linker, V-V(11) was more than
400-fold more potent than vancomycin. In contrast, the V-V
series was the least impressive overall against VanA VRE.
Greatest absolute potency was observed with V-V(11) and
V-V(19) against PRSP, EFSVS, and EFSVB. Lowest absolute
potency was observed against EFSVA. Potency was greatest at
short linker lengths against PRSP, MRSA, GISA, and EFSVB
but was independent of linker length for MSSA and EFSVS.
Activity against EFSVB decreased dramatically with increases
in linker length.
ranged from equipotent to 4-fold less potent than vancomycin
against MRSA and GISA. Some compounds displayed potent
activity against EFSVB, EFMVA, and, in one case, EFSVA.
Greatest absolute potency was observed against PRSP, MSSA,
and EFSVS. Lowest absolute potency was observed for
V-R(27) and V-R(43) against EFSVA. Potency was indepen-
dent of linker length for all vancomycin-susceptible strains and
greatest at the shortest linker length for the VRE.
J. R-R Series. Relative to vancomycin, compounds in this
series were less potent against PRSP, slightly more to slightly
less potent against MSSA, and as much as 5-fold more potent
against EFSVS. R-R linked dimers were 2-8-fold less potent
than vancomycin against MRSA and GISA. Some compounds
displayed potent activity against EFSVB and EFMVA and, in
one case, measurable activity against EFSVA. Greatest absolute
potency was observed against MSSA and EFSVS. Lowest
absolute potency was observed against EFSVA. Potency was
greatest at the shortest linker length against VRE, at intermediate
I. V-R Series. Relative to vancomycin, compounds in this
series were similarly potent against PRSP and MSSA and up
to 5-fold more potent against EFSVS. V-R linked dimers
9
6524 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Synthesis and Analysis of an Array of Vancomycin Dimers
A R T I C L E S
Table 2. Linker Length Effects on the In Vitro Potency of
B. MSSA. One-half of the dimers were more potent than
vancomycin against this strain: 20 dimers were more potent,
17 dimers were equipotent, and 3 dimers were less potent. The
MIC values for 8 dimers were lower than that of vancomycin
by a factor of at least 4. The N-N(43) dimer was 4-fold less
potent. The C-V, C-R, and V-V linkage series were the most
potent overall, and the N-N series was the least potent overall
against MSSA. In 4 of the 10 series (C-C, C-N, N-R, R-R),
dimers with intermediate-length linkers (19 or 27 atoms)
displayed maximum relative potency.
C. MRSA. The majority of dimers were less potent than
vancomycin against this strain: 1 dimer was more potent, 13
dimers were equipotent, and 26 dimers were less potent. The
V-V(11) dimer was a factor of 4 more potent than vancomycin,
and the N-N(43) dimer was 12-fold less potent than vanco-
mycin against this strain. The V-V linkage series was the most
potent overall, and the C-C, N-N, and R-R series were least
potent overall against MRSA. In 3 of the 10 series (C-R, N-N,
R-R), dimers with intermediate-length linkers displayed maxi-
mum relative potency.
Vancomycin Dimersa
observed effect
orientation PRSP MSSA MRSA GISA EFSVS EFSVB EFMVA EFSVA
C-C
C-N
C-V
C-R
N-N
N-V
N-R
V-V
V-R
R-R
U
S
U
U
S
S
U
S
I
I
U
U
U
I
U
U
I
U
S
U
U
S
U
U
U
U
U
S
S
S
S
S
S
S
S
S
S
S
U
S
S
S
S
S
X
S
S
S
X
S
X
X
X
X
X
S
U
U
S
U
I
U
U
I
I
I
S
U
S
S
U
U
S
U
S
U
I
U
U
X
a Legend: (U) in vitro antibacterial potency is relatively unaffected by
changes in linker length in this linkage orientation series against this
organism; (I) maximal potency is observed at an intermediate linker length
in this linkage orientation series against this organism; (S) potency is
maximal at short linker lengths and decreases with increasing linker length
in this linkage orientation series against this organism; (X) data are
insufficient to assign a linker length effect. Bacterial strains are abbreviated
as described in the text and in the footnotes to Table 1. In making the
assignments, we considered any difference in MICs of more than a factor
of 3 to be outside the limits of experimental variation. In the series where
the “I” linker effect was designated, the MIC value for one or both of the
dimers joined by 19- or 27-atom linkers was at least 3 times lower than the
average of the MICs for corresponding dimers linked with the shortest and
longest linkers.
D. GISA. Nearly all of the dimers were less potent than
vancomycin against this strain: 1 dimer was more potent, 4
dimers were equipotent, and 35 dimers were less potent. The
N-N(43) dimer was more than 30-fold less potent. In 2 of the
10 linkage series (C-V, C-R), dimers with intermediate-length
linkers displayed maximum relative potency.
linker lengths against MSSA and MRSA, and was independent
of linker length for PRSP, GISA, and EFSVS.
E. EFSVS. All but 3 dimers were more potent than
vancomycin against this strain, and the MIC values for 15
compounds were lower by a factor of more than 10. Along with
the V-V linkage series, each of the C-linked series (C-C,
C-N, C-V, C-R) displayed impressive potency overall, while
the N-N series was the least potent overall against EFSVS.
F. EFSVB. The majority of dimers were more potent than
vancomycin against this strain: 32 dimers were more potent in
that they displayed measurable activity in vitro, while 8 dimers,
like vancomycin, did not display measurable activity at con-
centrations of 150 µg/mL or greater. Each dimer series has
members that are at least a factor of 10 more potent than
vancomycin against EFSVB. The V-V(11) dimer is remarkably
potent, with an MIC value that is lower than those for other
members of the array by at least 20-fold and lower than that
for vancomycin by more than 400-fold. It must be considered
that synthons 2 and 4 are significantly potent against EFSVB.
In this respect, only the V-V(11) dimer is more potent than
synthon 2, by at least 24-fold. In each series, potency decreases
substantially between members with the shortest linkers and
those with the longest linkers, from 4-fold in the N-N series
to 750-fold in the V-V series. The N-N series is least potent
overall.
Effects of Linker Length. As described within each linkage
orientation series, the effects of linker length on the in vitro
antibacterial potency of dimers fell within one of three discern-
ible categories as summarized in Table 2. Potency was relatively
unaffected by changes in linker length within the range studied
(U), was maximal at the shorter linker lengths and decreased
with increasing linker length (S), or was greatest at an
intermediate linker length (I). In the case of the V-V series
against EFMVA and in seven of the 10 series against EFSVA,
the data gathered for these essentially inactive compounds were
insufficient to assign a linker length effect with confidence (X).
Linker effects varied with both linkage orientation and test
organism. Maximum potency at intermediate lengths was only
observed for staphylococci. Compounds bearing the shortest
linkers were clearly superior against VRE. In no case did
compounds joined by the longest linkers display the greatest
potencies in their series. Each of the four C-linked series
displayed maximum activity at an intermediate linker length
against at least one strain. Three of the four R-linked series,
except for V-R, also displayed this effect against at least one
strain. In contrast, maximum activity at intermediate linker
lengths was not preferred with V-linkage; only one V-linked
series, C-V, displayed this behavior.
G. EFMVA. Out of 40 dimers, 25 displayed measurable
activity (i.e., MIC less than or equal to 200 µg/mL) against this
VanA strain, and 3 compounds (C-C(11), V-R(11), R-R(11))
displayed MIC values less than 10 µg/mL. In each series,
potency decreases substantially as linker length increases. The
least potent series overall was V-V.
H. EFSVA. This strain was the most resistant to the array;
only 9 out of 40 dimers displayed measurable activity against
this highly vancomycin-resistant organism. Compounds from
the C-C and C-V series were the most potent, but no
compound displayed an MIC value less than 38 µg/mL. Four
Activity as a Function of Test Strain. A. PRSP. Nearly
one-half of the dimers were more potent than vancomycin
against this strain: 19 dimers were more potent, 15 dimers were
equipotent, and 6 dimers were less potent. The MIC values for
12 dimers were lower than that of vancomycin by a factor of
more than 4. These results are especially impressive in light of
the finding that all of the modified linkage synthons 2-5 were
less potent than vancomycin against PRSP. The C-C, C-N,
C-V, and V-V linkage series were the most potent overall,
and the N-N, N-R, and R-R series were least potent overall
against PRSP.
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6525

A R T I C L E S
Griffin et al.
series (C-R, N-N, N-V, V-V) displayed no detectable
activity at 200 µg/mL.
nonuniformities in the observed effects. First, no one dimer is
more potent than the other dimers or vancomycin against all
target organisms. While the V-V(11) dimer is exceptionally
potent against both EFSVS and EFSVB, this compound is
among the least active against VanA VRE. Second, linker length
effects for each linkage series vary depending upon test
organism. Third, whereas the majority of dimers are more potent
than vancomycin against MSSA, only the V-V(11) dimer is
more potent than vancomycin against the other staphylococcal
strains tested, MRSA and GISA.
Statistical Correlations. Linear regression was used to make
pairwise comparisons of the MIC data sets arranged by linkage
series and by test organism (see Supporting Information). All
45 pairwise correlations between the 10 linkage series data sets
were found to be highly statistically significant (P < 0.0001).
The highest correlation coefficients were observed with the
N-V and N-R linkage series, while the lowest correlation
coefficients were observed with the C-C and C-V linkage
series. With respect to comparisons between the test organism
data sets, 9 out of 10 correlations involving the 5 vancomycin-
susceptible organisms achieved statistical significance (P <
0.05), but none of the correlations between the data for these
organisms and the data for vancomycin-resistant E. faecalis
(VanB) reached significance.
That the V-V(11) dimer is exceptionally potent against both
EFSVS and the VanB VRE strain but is at best only weakly
active against VanA VRE suggests that this compound does
not trigger the inducible resistance mechanism of VanB organ-
isms which leads to the production of peptidoglycan precursors
terminating in D-Ala-D-Lac rather than D-Ala-D-Ala.30 This
explanation may also account for the enhanced potency of
synthons 2 and 4 toward EFSVB. More generally, the stark
contrast in activity observed for V-V dimers joined by peptide
linkers against vancomycin-susceptible organisms on one hand
and against VanA VRE on the other may be considered in
evolutionary terms.31 The noncovalent self-association of gly-
copeptides into back-to-back dimers has been quantitatively
correlated with in vitro antibacterial potency against vancomy-
cin-susceptible bacteria that utilize peptidoglycan precursors
terminating in D-Ala-D-Ala.15 Accordingly, back-to-back self-
association appears to be a naturally selected property of
glycopeptides, and models indicate that back-to-back self-
association is directly enforced by V-V linkage. It has also
been established that glycopeptide-producing organisms, such
as the vancomycin producer Amycolatopsis orientalis, possess
homologues of the VanH, VanA, and VanX genes needed to
elaborate peptidoglycan precursors terminating in D-Ala-D-Lac
and to confer high-level resistance to glycopeptides upon
enterococci and staphylococci.32 If this mechanism is sufficient
to confer upon producing organisms resistance to their own
spontaneously self-associating secondary metabolites, it seems
reasonable that such resistance would not be overcome by
merely enforcing the self-associated state. Accordingly, in the
absence of other effects, one would not expect significant
activity for V-V linked dimers against VanA VRE. Consistent
with this argument, against our VanA VRE test strains, we
observed measurable activity with only one V-V dimer joined
by an amphiphilic peptide linker. However, this result stands
in contrast to the reports from the Lilly group and from Nicolaou
Discussion
We have designed and synthesized an array of 40 covalently
linked vancomycin dimers. In this array, the effects of varying
the relative orientations of the individual glycopeptide subunits
were systematically explored by varying two key parameters
of multivalent design, the sites of linkage and the lengths of
the linkers employed. We substantially fixed the parameters of
linker chemical structure and physical properties by employing
a nested set of amphiphilic, peptide-based linkers. The in vitro
susceptibility data obtained with these compounds using a panel
of representative Gram-positive bacterial strains are rich in
content. Overall, the data indicate that dimerization with peptide-
based linkers represents a useful but not fully general approach
to enhancing the potency of vancomycin toward both susceptible
and drug-resistant Gram-positive pathogens.
Both linkage orientation and linker length influence the in
vitro activity of vancomycin dimers. With respect to linkage
orientation, the V-V series displayed the greatest potency
against vancomycin-susceptible organisms and VanB VRE, the
C-C, C-V, and V-R series displayed the most promising
broad spectrum of activity that included VanA VRE, while the
N-N linked series was the least potent overall against the test
organisms. Many to most of the dimers, including one or more
compounds from each linkage series, were more potent than
vancomycin against PRSP, MSSA, EFSVS, EFSVB, and
EFMVA. At least 1 compound in 6 of the 10 linkage series
also displayed measurable levels of activity against the highly
resistant EFSVA strain. In addition, statistically significant
correlations were observed for all pairwise comparisons of the
data for each linkage series. It is possible to interpret this large
subset of the results with simple models for the effects of
dimerization. In one such model, many combinations of linkage
orientation and linker length allow for multivalent interactions
that improve affinity for intermediates involved in the biosyn-
thesis of bacterial cell wall peptidoglycan, whether they
terminate in either D-Ala-D-Ala or D-Ala-D-Lac, and this in turn
provides for enhanced in vitro antibacterial potency. Alterna-
tively, many combinations of linkage orientation and linker
length may provide vancomycin dimers with the ability to exert
a novel, potent mechanism of action against PRSP, MSSA,
EFSVS, and VRE, such as direct inhibition of transglycosylase
enzymes.9e,29
(29) (a) Kerns, R.; Dong, S. D.; Fukuzawa, S.; Carbeck, J.; Dorso, K.; Kohler,
J.; Onishi, R.; Silver, L.; Kahne, D. J. Am. Chem. Soc. 1998, 122, 12608-
12609. (b) Ge, M.; Chen, Z.; Onishi, H. R.; Kohler, J.; Silver, L. L.; Kerns,
R.; Fukuzawa, S.; Thompson, C.; Kahne, D. Science 1999, 284, 507-511.
(c) Goldman, R. C.; Baizman, E. R.; Longley, C. B.; Branstrom, A. A.
FEMS Microbiol. Lett. 2000, 183, 209-214. (d) Eggert, U. S.; Ruiz, N.;
Falcone, B. V.; Branstrom, A. A.; Goldman, R. C.; Silhavy, T. J.; Kahne,
D. Science 2001, 294, 361-364. (e) Roy, R. S.; Yang, P.; Kodali, S.; Xiong,
Y.; Kim, R. M.; Griffin, P. R.; Onishi, H. R.; Kohler, J.; Silver, L. L.;
Chapman, K. Chem. Biol. 2001, 139, 1-12. (f) Pavlov, A. Y.; Miroshi-
nokova, O. V.; Printsevaskaya, S. S.; Olsufyeva, E. N.; Preobrazhenskaya,
M. N.; Goldman, R. C.; Branstrom, A. A.; Baizman, E. R.; Longley, C. B.
J. Antibiot. 2001, 54, 455-459.
(30) Dong, S. D.; Oberthu¨r, M.; Losey, H. C.; Anderson, J. W.; Eggert, U. S.;
Peczuh, M. W.; Walsh, C. T.; Kahne, D. J. Am. Chem. Soc. 2002, 124,
9064-9065.
(31) (a) Williams, D. H.; Stone, M. J.; Hauck, P. R.; Rahman, S. K. J. Nat.
Prod. 1989, 52, 1189-1208. (b) Stone, M. J.; Williams, D. H. Mol.
Microbiol. 1992, 6, 29-34. (c) Maplestone, R. A.; Williams, D. H.; Stone,
M. J. Gene 1992, 115, 151-157.
However, no single model in its simplest form appears to be
sufficient to account for all of the results, including several
(32) Marshall, C. G.; Lessard, I. A. D.; Park, I.-S.; Wright, G. D. Antimicrob.
Agents Chemother. 1998, 42, 2215-2220.
9
6526 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Synthesis and Analysis of an Array of Vancomycin Dimers
A R T I C L E S
et al. that V-V dimers joined by hydrophobic, benzyl-based
linkers display potent activity against VanA VRE.19,20 Linkers
12-34 atoms in length were employed in those studies, with
maximum activity observed for compounds having linkers at
least 18 atoms in length. Appending hydrophobic groups to the
V-position of monomeric glycopeptides enhances their self-
association, their association with bacterial cell surfaces and
models thereof4f,15 as well as their activity against VanA VRE.9
In light of this and the findings with our V-V linkage series,
we propose that the anti-VRE activity of V-V linked dimers
joined by hydrophobic linkers is dominated not by effects that
derive from enforced back-to-back self-association but rather
by effects that derive from having added hydrophobic substit-
uents to the amino disaccharide moiety.
N-N linkage series) are dimers joined by the shortest linkers
the most potent against EFSVS. We favor an alternative
interpretation of the linker length effects observed with VRE,
which involves the conformational entropy of the linker.33 In
the case of VRE, where individual vancomycin/D-Ala-D-Lac
interactions are relatively weak, the absolute energetic benefits
of multivalency can be offset by energy losses due to restricting
the conformational entropy of longer linkers. Overall, our data
suggest that one rational next step in seeking to optimize the
activity of vancomycin dimers against VRE would be to develop
focused structure-activity relationships around dimers joined
with an array of shorter and more hydrophobic linkers. Whereas
there is no evidence which indicates there is a difference in
cell wall structure in vancomycin-susceptible enterococci versus
vancomycin-resistant enterococci, it has been established that
GISA strains elaborate an exceptionally thick, poorly cross-
linked cell wall that presents many unprocessed D-Ala-D-Ala
termini.34 This structure presents a barrier to penetration by and
a sequestering sink for vancomycin, and we expect that these
effects would be exacerbated for the larger, multivalent van-
comycin dimers.
Covalently linked glycopeptide dimers may self-assemble into
extended, polyvalent networks of receptors for precursors and
intermediates involved in peptidoglycan biosynthesis. This
phenomenon, as envisioned by Staroske and Williams21 and by
Adamczyk et al.,18 may be particularly critical in achieving
activity against highly resistant organisms where individual
vancomycin/D-Ala-D-Lac interactions are relatively weak (Ka
≈ 1000 M-1).6 Models indicate that, relative to dimers linked
in other orientations, V-V linked dimers are hindered from
engaging in the back-to-back interdimer interactions needed to
form extended networks. Whitesides and co-workers observed
enhanced binding of C-C linked dimers to surface displays of
(Nꢀ-acetyl)-L-Lys-D-Ala-D-Ala, relative to vancomycin, as mea-
sured by surface plasmon resonance.17b A similar result was
found by Adamczyk et al. who furthermore demonstrated that
two V-V linked dimers did not display enhanced surface
binding relative to vancomycin.18
Experimental Section
Chemistry. A. General. Reagents and solvents were used as received
from commercial suppliers, and all reactions were carried out at room
temperature and without rigorous exclusion of ambient atmosphere
unless otherwise noted. Ion-spray mass spectra (IS-MS) were obtained
on a PE Sciex API 150EX mass spectrometer. Nuclear magnetic
resonance (NMR) spectra were recorded at 300 MHz. Chemical shifts
(δ) are reported in parts per million downfield of tetramethylsilane.
Analytical reversed-phase HPLC (RP-HPLC) was performed on an
HP1100 instrument using a 2.1 mm × 50 mm, 3.5 µm C18 Zorbax
Plus Bonus-RP column. For the analytical separations, a 0.5 min
isocratic period was followed by a 4.5 min 2-30% gradient of 0.1%
trifluoroacetic acid/acetonitrile (ACN) in 0.1% water at a flow rate of
0.5 mL/min. Preparative RP-HPLC was performed using trifluoroacetic
acid (TFA) buffered ACN/water gradients on a Varian ProStar system
using 2.5 or 10 cm × 25 cm, 8 µm C18 Rainin Dynamax columns and
flow rates of 10 or 50 mL/min, respectively.
The effects of linker length on in vitro potency vary with
linkage series as well as with organism. We discern no
correlation between the observed effects and the minimal linker
lengths that models predicted would enforce the back-to-back
or face-to-face associated geometries of vancomycin. That the
effects of linker length would differ from one linkage series to
another is not surprising. However, the finding that linker length
effects for a given series vary from organism to organism
suggests that the factors determining in vitro potency also vary
among different organisms. Only one series (V-R) displays
the same linker length effect across all of the non-VRE strains.
Maximal potency at intermediate linker lengths is observed only
with the three staphylococcal strains used (MSSA, MRSA,
GISA), even though PRSP and EFSVS also utilize peptidogly-
can precursors terminating in D-Ala-D-Ala. The linker length
behavior observed with the staphylococci is independent of
whether the compounds are generally more potent (MSSA) or
less potent (MRSA, GISA) than vancomycin against the test
organism.
B. C-Linkage Synthon 2. This compound was prepared by the
coupling of vancomycin with 9-fluorenylmethyl N-(2-aminoethyl)-
carbamate, followed by deprotection.
Vancomycin hydrochloride (7.3 g, 4.7 mmol) was dissolved in 75
mL of dimethyl sulfoxide (DMSO). To this solution was added in
sequence N,N-diisopropylethylamine (DIPEA, 4.1 mL, 24 mmol),
9-fluorenylmethyl N-(2-aminoethyl)carbamate hydrochloride (1.8 g, 5.6
mmol), and a solution of benzotriazol-1-yl-oxy-tris(pyrrolidino)phos-
phonium hexafluorophosphate (PyBOP, 2.7 g, 5.2 mmol) and 1-hy-
droxybenzotriazole hydrate (HOBT, 0.63 g, 4.7 mmol) in 75 mL of
1,3-dimethyl-3,4,5,6-tetrahydro-2(1H)-pyrimidinone (DMPU). The re-
sulting solution was stirred for 2 h at room temperature (rt) and then
poured into 800 mL of diethyl ether (Et2O), giving a gum. The liquid
phase was decanted, and the gum containing crude vancomycin
[9-fluorenylmethyl N-(2-amidoethyl)carbamate] (6) was washed with
additional Et2O. This compound was used without further purification.
Compound 6 was dissolved in 40 mL of N,N-dimethylformamide
(DMF) and then treated with 10 mL of piperidine. The solution was
left to stand at room temperature for 20 min and then added dropwise
to 450 mL of ACN. The resulting precipitate was isolated by
Dimers bearing the shorter linkers are in all cases the most
potent against the VRE strains (EFSVB, EFMVA, EFSVA),
but dimers joined by short linkers are preferred in only 12 of
the 50 cases involving EFSVS or nonenterococci. This could
indicate that penetration of the cell wall to the sites at which
glycopeptides exert antibacterial effects is hindered in VRE
relative to vancomycin-susceptible organisms. However, sig-
nificant changes in cell wall structure in vancomycin-resistant
versus vancomycin-susceptible enterococci have not been
reported, and in only 2 out of 10 cases (i.e., with the C-N and
(33) Mammen, M.; Shakhnovich, E. I.; Whitesides, G. M. J. Org. Chem. 1998,
63, 3168-3175.
(34) (a) Sieradzki, K.; Tomasz, A. J. Bacteriol. 1997, 179, 2557-2566. (b)
Hanaki, H.; Labischinski, H.; Inaba, Y.; Kondo, N.; Murakami, H.;
Hiramatsu, K. J. Antimicrob. Chemother. 1998, 42, 315-320.
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6527

A R T I C L E S
Griffin et al.
centrifugation/decantation, washed twice with 450 mL of ACN and
once with 450 mL of Et2O, and then air-dried. C-Linkage synthon 2
was purified by RP-HPLC and isolated by lyopholization as the TFA
salt. IS-MS, m/z for [M + H+]+: calcd 1492.3; obsd 1492.2.
C. N-Linkage Synthon 3. This compound was prepared by
regioselective reductive alkylation at the amino terminus of vancomycin
with N-(9-fluorenylmethoxycarbonyl)glycinal, followed by coupling
with 3-(dimethylamino)propylamine and deprotection.
vacuum. Compound 9 was purified by RP-HPLC and isolated by
lyopholization as the TFA salt. IS-MS, m/z for [M + H+]+: calcd
1715.6; obsd 1715.2.
Compound 9 (10 g, 5.1 mmol) was dissolved in 50 mL of DMF
and treated sequentially with DIPEA (3.6 mL, 21 mmol) and 3-(dim-
ethylamino)propylamine (0.97 mL, 7.7 mmol), followed by a solution
of PyBOP (2.9 g, 5.6 mmol) and HOBT (0.87 g, 5.6 mmol) in 6 mL
of DMF. The solution was stirred for 1 h at rt and then poured into
500 mL of Et2O. The resulting precipitate was isolated by filtration,
washed with Et2O, and dried under vacuum to yield crude compound
10. IS-MS, m/z for [M + H+]+: calcd 1798.7; obsd 1799.5. This
compound was used without further purification
Compound 10 (5.14 mmol) was dissolved in 50 mL of DMF and
treated with solid quinuclidine (2.9 g, 26 mmol). The solution was
stirred at rt for 1.5 h and then added to 300 mL of Et2O. The resulting
precipitate was isolated by filtration, washed with Et2O, and dried under
vacuum. V-Linkage synthon 4 was purified by RP-HPLC and isolated
by lyopholization as the TFA salt. IS-MS, m/z for [M + H+]+: calcd
1577.5; obsd 1577.7.
2-N-(9-Fluorenylmethoxycarbonyl)ethanol (20 g, 71 mmol) was
suspended in 200 mL of dichloromethane (DCM), cooled under a
nitrogen atmosphere to -40 °C (ACN/CO2(s)), and treated with DIPEA
(49.2 mL, 283 mmol). A freshly made solution of pyridine-SO3
complex (45.0 g, 283 mmol) in 200 mL of DMSO was dropped into
the reaction mixture over 10 min. The reaction was stirred for an
additional 10 min and then poured into a mixture of 700 mL of saturated
sodium chloride solution and 500 cm3 of crushed ice. After the mixture
was vigorously stirred, the DCM was removed under vacuum. The
resulting white precipitate was isolated by filteration, washed with water,
and dried under vacuum. The crude N-(9-fluorenylmethoxycarbonyl)-
1
glycinal was used without further purification. H NMR (DMSO-d6)
E. R-Linkage Synthon 5. This compound was prepared by Mannich
reaction of vancomycin 3-N,N-dimethylaminopropanamide with eth-
ylenediamine/formaldehyde.
δ3.82 (d, 2H, J ) 7 Hz), 4.21 (t, 2H, J ) 7 Hz), 4.35 (d, 2H, J ) 7
Hz), 7.31 (m, 4H), 7.70 (d, 2H, J ) 7 Hz), 7.91 (d, 2H, J ) 7 Hz),
9.21 (s, 1H).
Vancomycin hydrochloride (10 g, 6.7 mmol) was slurried in 75 mL
of anhydrous DMF. DIPEA (4.7 mL, 30 mmol) was added, followed
by 3-(dimethylamino)propylamine (1.3 mL, 10 mmol). A solution of
PyBOP (3.9 g, 7.4 mmol) and HOBT (1.1 g, 7.4 mmol) in 7 mL of
DMF was added, and the reaction was stirred at rt for 1 h. The reaction
mixture was poured into 500 mL of Et2O. The resulting precipitate
was filtered, washed with Et2O, and dried under vacuum. Vancomycin
3-N,N-dimethylaminopropanamide (11) was purified by RP-HPLC and
isolated by lyopholization as the TFA salt. IS-MS, m/z for [M + H+]+:
calcd 1534.4; obsd 1534.5.
Compound 11 (12 g, 7.5 mmol) was dissolved in 200 mL of 1:1
water/ACN. Ethylenediamine (5.0 mL, 75 mmol) was added, followed
by aqueous formaldehyde (37%, 0.53 mL, 7.1 mmol). The reaction
was stirred at rt for 18 h. The ACN was removed under vacuum, and
the solution was acidified with TFA and subjected to RP-HPLC.
R-Linkage synthon 5 was isolated by lyopholization as the TFA salt.
IS-MS, m/z for [M + H+]+: calcd 1606.5; obsd 1606.9.
Vancomycin hydrochloride (10 g, 6.8 mmol) was slurried in 100
mL of 1:1 methanol (MeOH)/DMPU. DIPEA (1.3 mL, 7.5 mmol) was
added, and the mixture was stirred at rt until the solids dissolved. HOBT
(2.1 g, 14 mmol) was added, followed by a solution of N-(9-
fluorenylmethoxycarbonyl)glycinal in 20 mL of 1:1 MeOH/DMPU. The
reaction was allowed to stir at room temperature for 4 h. Sodium
cyanoborohydride (NaCNBH3, 1.3 g, 21 mmol) was added as a solid.
The reaction was stirred for another 2 h, at which point the MeOH
was removed under vacuum, and the remaining solution was poured
into 700 mL water. The resulting white precipitate was isolated by
filtration, washed with water, and dried under vacuum to yield
compound 7 as a white solid. IS-MS, m/z for [M + H+]+: calcd 1715.6;
obsd 1715.5. This compound was used without further purification.
Compound 7 (9.6 g, 5.6 mmol) was dissolved in 60 mL of DMF
and treated sequentially with DIPEA (2.9 mL, 17 mmol) and 3-(dim-
ethylamino)propylamine (1.1 mL, 8.34 mmol). A solution of PyBOP
and HOBT (1 M in DMF, 6.1 mL, 6.14 mmol each) was added, and
the reaction was stirred for 1 h at rt. The reaction mixture was poured
into 300 mL Et2O. The resulting white precipitate was isolated by
filtration, washed with Et2O, and dried under vacuum to yield compound
8 as a white solid. IS-MS, m/z for [M + H+]+: calcd 1798.7; obsd
1799.5. This compound was used without further purification.
Compound 8 (12 g, 5.6 mmol) was dissolved in 50 mL of DMF
and treated with solid quinuclidine (3.0 g, 28 mmol). The reaction was
stirred at rt for 1.5 h and then poured into 300 mL of Et2O. The resulting
white precipitate was isolated by filtration, washed with Et2O, and dried
under vacuum. N-Linkage synthon 3 was purified by RP-HPLC and
isolated by lyopholization as the TFA salt. IS-MS, m/z for [M + H+]+:
calcd 1577.5; obsd 1577.7.
D. V-Linkage Synthon 4. This compound was prepared by
regioselective reductive alkylation at the vancosamine residue of
vancomycin with N-(9-fluorenylmethoxycarbonyl)glycinal, followed by
coupling with 3-(dimethylamino)propylamine and deprotection.
Vancomycin hydrochloride (82 g, 55 mmol) was slurried in 600 mL
of anhydrous DMF and treated sequentially with DIPEA (29 mL, 170
mmol) and a solution of N-(9-fluorenylmethoxycarbonyl)glycinal in
100 mL of DMF. This solution was stirred at rt for 1.5 h, then treated
with TFA (26 mL, 330 mmol) and 200 mL of MeOH, and stirred for
another 40 min. A borane/pyridine complex (8 M, 7.6 mL, 61 mmol)
was added, and the reaction was stirred at rt for 18 h. The mixture was
concentrated under vacuum to a damp paste. Water (1.5 L) was added,
and the pH was adjusted to 4 using 1 M sodium hydroxide. The resulting
precipitate was filtered, washed extensively with water, and dried under
F. Linker Synthons A-D. The terminally differentiated linker
synthons were prepared by a combination of solution- and solid-phase
peptide synthesis approaches.
1. Linker Synthon A. Glutaric anhydride (10 g, 88 mmol) was
dissolved in 100 mL of anhydrous DCM and treated with 9-fluorenyl-
methanol (19 g, 96 mmol), followed by DIPEA (17 mL, 96 mmol).
The reaction was stirred at rt for 18 h, at which time thin-layer
chromatography (TLC, 5% MeOH in DCM) indicated that the reaction
was complete. The reaction mixture was washed twice with 100 mL
of 1 M hydrochloric acid (HCl), dried over anhydrous magnesium
sulfate (MgSO4), and concentrated to a yellow oil. Linker synthon A
was purified by chromatography on silica gel using 5% MeOH in DCM
as eluent to afford a pale oil which crystallized upon standing. 1H NMR
(CDCl3) δ 1.95 (qn, 2H, J ) 7 Hz), 2.39 (t, 2H, J ) 7 Hz), 2.49 (t,
2H, J ) 7 Hz), 4.22 (t, 1H, J ) 7 Hz), 4.44 (d, 2H, J ) 7 Hz), 7.35
(m, 4H), 7.60 (d, 2H, J ) 7 Hz), 7.78 (d, 2H, J ) 7 Hz).
2. Linker Synthon B. This compound was prepared by condensation
of glutaric acid with the fluorenylmethyl ester of â-alanine, followed
by deprotection.
Boc â-Ala-OH (50 g, 0.26 mol) and 9-fluorenylmethanol (54.3 g,
0.28 mol) were dissolved in 800 mL of DCM and treated sequentially
with 4-N,N-(dimethylamino)pyridine (DMAP, 1.6 g, 13.2 mmol) and
a solution of 1,3-diisopropylcarbodiimide (DIPC, 43.4 mL, 0.28 mol)
in 100 mL of DCM. The slightly exothermic reaction was stirred at rt
for 18 h. The reaction volume was reduced to approximately 200 mL,
and the white precipitate was filtered off, washed with 100 mL of DCM,
and discarded. TFA (200 mL) was added to the filtrate, which was
9
6528 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Synthesis and Analysis of an Array of Vancomycin Dimers
A R T I C L E S
stirred at rt for 1 h, concentrated under vacuum to a thick oil, and
treated with 1 L of Et2O. The resulting precipitate was filtered, washed
with Et2O, and dried under vacuum to yield the TFA salt of â-alanine,
fluorenylmethyl ester (H-â-Ala-OFm TFA). IS-MS, m/z for [M + H+]+:
calcd 268.3; obsd 268.8.
by the coupling of the N-terminus of the resin-bound chain with linker
synthon A followed by deprotection.
The linkers were subsequently elongated in the N-C direction, again
using standard Fmoc protocols. H-â-Ala-OFm TFA/PyBOP/HOAT/
DIPEA (4 equiv) was used for each coupling step. At the end of each
synthesis, the terminal fluorenylmethyl esters were not removed.
3.1. Linker Synthon C. Resin-bound linker synthon C was stirred
in 100 mL of 1:1 TFA/DCM for 1 h at rt. The mixture was filtered,
and the resin was washed with 50 mL of DCM. The filtrate was
concentrated to a pale yellow oil which was treated with 100 mL of
Et2O. The resulting precipitate was recovered by filtration, washed with
Et2O, and dried under vacuum. Linker synthon C was purified by RP-
HPLC and isolated by lyopholization. IS-MS, m/z for [M + H+]+: calcd
595.6; obsd 595.2.
3.2. Linker Synthon D. Resin-bound linker synthon D was stirred
in 100 mL of 1:1 TFA/DCM for 1 h at rt. The mixture was filtered,
and the resin was washed with 50 mL of DCM. The filtrate was
concentrated to a pale yellow oil which was treated with 100 mL of
Et2O. The resulting precipitate was recovered by filtration, washed with
Et2O, and dried under vacuum. Linker synthon D was purified by RP-
HPLC and isolated by lyopholization. IS-MS, m/z for [M + H+]+: calcd
879.9; obsd 879.5 (M)+.
Glutaric anhydride (5.0 g, 44 mmol) was dissolved in 70 mL of
anhydrous DCM. H-â-Ala-OFm TFA (18 g, 48 mmol) was added,
followed by DIPEA (8.4 mL, 48 mmol). The reaction mixture was
stirred at rt for 24 h, diluted with 100 mL of DCM, and washed with
1 M HCl. The resulting precipitate was isolated by filtration, washed
with a minimum of DCM, and dried to yield glutaryl-â-alanine,
fluorenylmethyl ester (HO-Glut-â-Ala-OFm). IS-MS, m/z for [M +
1
H+]+: calcd 382.4; obsd 383.0. H NMR (CDCl3) δ 1.21 (t, 2H, J )
7 Hz), 1.67 (qn, 2H, J ) 7 Hz), 2.06 (t, 2H, J ) 7 Hz), 2.16 (t, 2H, J
) 7 Hz), 2.47 (t, 2H, J ) 7 Hz), 3.21 (q, 1H, J ) 6.7 Hz) 4.26 (t, 1H,
J ) 7 Hz), 4.37 (d, 2H, J ) 7 Hz), 7.37 (m, 4H), 7.60 (d, 2H, J ) 7
1
Hz), 7.85 (d, 2H, J ) 7 Hz); H NMR δ 2.64 (t, 2H, J ) 7 Hz), 2.95
(t, 2H, J ) 7 Hz), 4.21 (t, 1H, J ) 7 Hz), 4.34 (d, 2H, J ) 7 Hz), 7.30
(m, 4H), 7.58 (d, 2H, J ) 7 Hz), 7.82 (d, 2H, J ) 7 Hz), 7.89 (s, 3H).
HO-Glut-â-Ala-OFm (3.0 g, 7.9 mmol) and H-â-Ala-OtBu HCl (1.6
g, 8.6 mmol) were slurried together in 75 mL of DCM and treated
sequentially with DIPEA (1.5 mL, 8.6 mmol), DMAP (∼50 mg), and
solid 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride
(EDC, 1.7 g, 8.6 mmol). The reaction was stirred at rt for 45 min,
treated with 10 mL of DMF, and then stirred for an additional 18 h, at
which time TLC (10% MeOH in DCM) indicated that the reaction was
complete. The reaction mixture was washed twice with 75 mL of 1 M
HCl and twice with 75 mL of water. The organic phase was dried over
MgSO4, filtered, and concentrated. The solid was purified by chroma-
tography on silica gel using 10% MeOH in DCM eluent to afford the
mono-t-butyl, mono-fluorenylmethyl ester of bis(â-alanyl)glutaramide
(tBuO-â-Ala-Glut-â-Ala-OFm) as a pale yellow oil.
tBuO-â-Ala-Glut-â-Ala-OFm was dissolved in 100 mL of DCM and
treated with 100 mL of TFA. The reaction was stirred for 5 h at rt and
then concentrated. The residue was subjected to cycles of redissolution
in DCM, concentration, and drying under vacuum until all TFA had
been removed. Linker synthon B was purified by chromatography on
silica gel using 10% MeOH in DCM as eluent to afford a colorless oil
which crystallized upon standing. IS-MS, m/z for [M + H+]+: calcd
453.5; obsd 454.5.
G. C-C Series. The synthesis of compounds in this series is
exemplified by the synthesis of the dimer joined by an 11-atom linker,
C-C(11).
C-Linkage synthon 2 (2.0 g, 1.1 mmol) and glutaric acid (65 mg,
0.50 mmol) were dissolved in 10 mL of DMF and treated sequentially
with DIPEA (0.86 mL, 5.0 mmol), and a solution of PyBOP (0.57 g,
1.1 mmol) and HOAT (150 mg, 1.1 mmol) in 1.5 mL of DMF. The
reaction was stirred at rt for 2 h and then treated with 100 mL of Et2O.
The resulting precipitate was isolated by filtration, washed with Et2O,
and dried under vacuum. The C-C(11) homodimer was purified by
RP-HPLC and isolated by lyopholization as the TFA salt. IS-MS, m/z
for [M + 2H+]2+: calcd 1539.0; obsd 1540.2.
H. C-N Series. The synthesis of compounds in this series is
exemplified by the synthesis of the dimer joined by a 43-atom linker,
C-N(43).
Linker synthon D (0.10 g, 0.11 mol) was dissolved in anhydrous
DMSO with heating. The solution was cooled to rt and treated
sequentially with N-linkage synthon 3 (0.30 g, 0.15 mmol), DIPEA
(150 µL, 0.89 mmol), and a solution of PyBOP/HOAT (1 M each in
DMF, 160 µL, 0.16 mmol). The solution was stirred at rt for 1 h and
then treated with solid quinuclidine (0.1 g). The solution was stirred
for an additional 1.5 h and then treated with 75 mL of Et2O. The
resulting precipitate was isolated by filtration, washed with Et2O, and
dried under vacuum. The 3-D conjugate was purified by RP-HPLC
and isolated by lyopholization as the TFA salt. IS-MS, m/z for [M +
H+]+: calcd 1130.6; obsd 1130.4.
The 3-D conjugate (45 mg, 0.017 mmol) and C-linkage synthon 2
(41 mg, 0.022 mmol) were dissolved in 1 mL of anhydrous DMF and
treated sequentially with DIPEA (21 µL, 0.12 mmol) and a solution of
PyBOP/HOAT (1 M each in DMF, 28 µL, 0.027 mmol). The reaction
was stirred at room temperature for 1 h and then treated with 20 mL
of Et2O. The resulting precipitate was collected by centrifugation,
washed with Et2O, and dried under vacuum. The C-N(43) heterodimer
was purified by RP-HPLC and isolated by lyopholization as the TFA
salt. IS-MS, m/z for [M + 2H+]2+: calcd 1866.3; obsd 1867.7.
I. C-V Series. The synthesis of compounds in this series is
exemplified by the synthesis of the dimer joined by a 19-atom linker,
C-V(19).
3. Linker Synthons C and D. These compounds were prepared by
application of solid-phase synthesis methodology using both C-N and
N-C directional couplings.
Wang resin (6.1 g, 1.2 mequiv/g loading, 7.3 mmol) was swelled
for 1 h in 150 mL of DCM. The solvent was drained, and 100 mL of
fresh DCM was added. Fmoc-â-Ala-OH (11 g, 35 mmol) was added
as a solid, followed by ∼100 mg DMAP. Neat DIPC (5.5 mL, 35 mmol)
was added, and the mixture was sparged with nitrogen for 5 h with
occasional additions of DCM. The resin was then washed with 4 ×
200 mL of DMF and 4 × 200 mL of DCM and then dried under
vacuum to a constant weight which indicated 94% loading of the first
â-Ala residue.
Linkers were initially elongated in the C-N direction using standard
protocols for Fmoc-protected amino acids. Coupling and deprotection
steps were carried out for 1 h, each followed by washing with 4 × 200
mL of DMF and 4 × 200 mL of DCM. Fmoc-â-Ala-OH/PyBOP/1-
hydroxy-7-azabenzotriazole (HOAT)/DIPEA (3 equiv) was used for
each coupling step, and deprotections were carried out with 20%
piperidine in DMF. Couplings and deprotections were monitored on
the resin by use of the qualitative Kaiser test.35
V-Linkage synthon 4 (0.30 g, 0.19 mmol) and linker synthon B (50
mg, 0.15 mmol) were dissolved in 2 mL of DMF and treated
sequentially with DIPEA (33 µL, 0.19 mmol) and a solution of PyBOP
(90 mg, 0.17 mmol) and HOAT (30 mg, 0.17 mmol) in 0.5 mL of
DMF. The solution was stirred at rt for 1 h and then treated with solid
quinuclidine (0.11 g). The reaction was stirred for an additional 1.5 h
Once the addition of â-Ala residues in the C-N direction was
complete (two residues for linker synthon C and four residues for linker
synthon D), the direction (polarity) of chain elongation was reversed
(35) Kaiser, E.; Colescott, R. L.; Bossinger, C. D.; Cook, P. I. Anal. Biochem.
1970, 34, 595-598.
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6529

A R T I C L E S
Griffin et al.
and then treated with 75 mL of Et2O. The resulting precipitate was
isolated by filtration, washed with Et2O, and dried under vacuum. The
4-B conjugate was purified by RP-HPLC and isolated by lyopholization
as the TFA salt. IS-MS, m/z for [M + H+]+: calcd 1833.7; obsd 1833.8.
The 4-B conjugate (16 mg, 0.0074 mmol) and C-linkage synthon 2
(18 mg, 0.0096 mmol) were dissolved in 1 mL of anhydrous DMF
and treated sequentially with DIPEA (9 µL, 0.052 mmol) and a solution
of PyBOP/HOAT (1 M each in DMF, 12 µL, 0.012 mmol). The reaction
was stirred at room temperature for 1 h and then treated with 20 mL
of Et2O. The resulting precipitate was collected by centrifugation,
washed with Et2O, and dried under vacuum. The C-V(19) heterodimer
was purified by RP-HPLC and isolated by lyopholization as the TFA
salt. IS-MS, m/z for [M + 2H+]2+: calcd 1653.0; obsd 1653.0.
J. C-R Series. The synthesis of compounds in this series is
exemplified by the synthesis of the dimer joined by a 27-atom linker,
C-R(27).
5-C conjugate was purified by RP-HPLC and isolated by lyopholization
as the TFA salt. IS-MS, m/z for [M + H+]+: calcd 2004.9; obsd 2004.1.
The 5-C conjugate (33 mg, 0.013 mmol) and N-linkage synthon 3
(34 mg, 0.017 mmol) were dissolved in 1 mL of anhydrous DMF and
treated sequentially with DIPEA (20 µL, 0.12 mmol) and a solution of
PyBOP/HOAT (1 M each in DMF, 21 µL, 0.021 mmol). The reaction
was stirred at room temperature for 1 h and then treated with 20 mL
of Et2O. The resulting precipitate was collected by centrifugation,
washed with Et2O, and dried under vacuum. The N-R(27) heterodimer
was purified by RP-HPLC and isolated by lyopholization as the TFA
salt. IS-MS, m/z for [M + 2H+]2+: calcd 1781.2; obsd 1782.1.
N. V-V Series. The synthesis of compounds in this series is
exemplified by the synthesis of the dimer joined by a 19-atom linker,
V-V(19).
The 4-B conjugate was prepared as described previously for the
synthesis of the C-V(19) heterodimer.
The 4-B conjugate (11 mg, 0.0049 mmol) and V-linkage synthon 4
(13 mg, 0.0064 mmol) were dissolved in 1 mL of anhydrous DMF
and treated sequentially with DIPEA (6.9 µL, 0.039 mmol) and a
solution of PyBOP/HOAT (1 M each in DMF, 7.9 µL, 0.0079 mmol).
The reaction was stirred at room temperature for 1 h and then treated
with 20 mL of Et2O. The resulting precipitate was collected by
centrifugation, washed with Et2O, and dried under vacuum. The
V-V(19) homodimer was purified by RP-HPLC and isolated by
lyopholization as the TFA salt. IS-MS, m/z for [M + 2H+]2+: calcd
1695.0; obsd 1696.6.
O. V-R Series. The synthesis of compounds in this series is
exemplified by the synthesis of the dimer joined by a 27-atom linker,
V-R(27).
The 5-C conjugate was prepared as described previously for the
synthesis of the N-R(27) heterodimer.
The 5-C conjugate was prepared as described in the following for
the synthesis of the N-R(27) heterodimer.
The 5-C conjugate (98 mg, 0.049 mmol) and C-linkage synthon 2
(110 mg, 0.059 mmol) were dissolved in 1 mL of anhydrous DMF
and treated sequentially with DIPEA (16 µL, 0.10 mmol) and a solution
of PyBOP/HOAT (1 M each in DMF, 79 µL, 0.079 mmol). The reaction
was stirred at room temperature for 1 h and then treated with 20 mL
of Et2O. The resulting precipitate was collected by centrifugation,
washed with Et2O, and dried under vacuum. The C-R(27) heterodimer
was purified by RP-HPLC and isolated by lyopholization as the TFA
salt. IS-MS, m/z for [M + 2H+]2+: calcd 1738.6; obsd 1739.5.
K. N-N Series. The synthesis of compounds in this series is
exemplified by the synthesis of the dimer joined by a 43-atom linker,
N-N(43).
The 3-D conjugate was prepared as described previously for the
The 5-C conjugate (31 mg, 0.013 mmol) and V-linkage synthon 4
(34 mg, 0.017 mmol) were dissolved in 1 mL of anhydrous DMF and
treated sequentially with DIPEA (20 µL, 0.12 mmol) and a solution of
PyBOP/HOAT (1 M each in DMF, 21 µL, 0.021 mmol). The reaction
was stirred at room temperature for 1 h and then treated with 20 mL
of Et2O. The resulting precipitate was collected by centrifugation,
washed with Et2O, and dried under vacuum. The V-R(27) heterodimer
was purified by RP-HPLC and isolated by lyopholization as the TFA
salt. IS-MS, m/z for [M + 2H+]2+: calcd 1781.2; obsd 1781.8.
P. R-R Series. The synthesis of compounds in this series is
exemplified by the synthesis of the dimer joined by a 27-atom linker,
R-R(27).
synthesis of the C-N(43) heterodimer.
The 3-D conjugate (41 mg, 0.049 mmol) and N-linkage synthon 3
(42 mg, 0.021 mmol) were dissolved in 1 mL of anhydrous DMF and
treated sequentially with DIPEA (22 µL, 0.10 mmol) and a solution of
PyBOP/HOAT (1 M each in DMF, 25 µL, 0.025 mmol). The reaction
was stirred at room temperature for 1 h and then treated with 20 mL
of Et2O. The resulting precipitate was collected by centrifugation,
washed with Et2O, and dried under vacuum. The N-N(43) homodimer
was purified by RP-HPLC and isolated by lyopholization as the TFA
salt. IS-MS, m/z for [M + 2H+]2+: calcd 1908.8; obsd 1909.7.
L. N-V Series. The synthesis of compounds in this series is
exemplified by the synthesis of the dimer joined by a 43-atom linker,
N-V(43).
The 5-C conjugate was prepared as described previously for the
synthesis of the N-R(27) heterodimer.
The 3-D conjugate was prepared as described previously for the
The 5-C conjugate (98 mg, 0.049 mmol) and R-linkage synthon 5
(130 mg, 0.059 mmol) were dissolved in 1 mL of anhydrous DMF
and treated sequentially with DIPEA (51 µL, 0.29 mmol) and a solution
of PyBOP/HOAT (1 M each in DMF, 79 µL, 0.079 mmol). The reaction
was stirred at room temperature for 1 h and then treated with 20 mL
of Et2O. The resulting precipitate was collected by centrifugation,
washed with Et2O, and dried under vacuum. The R-R(27) homodimer
was purified by RP-HPLC and isolated by lyopholization as the TFA
salt. IS-MS, m/z for [M + 2H+]2+: calcd 1795.7; obsd 1796.5.
Microbiology. Bacterial strains were obtained from the American
Type Culture Collection (ATCC), Stanford University Hospital (SU),
Kaiser Permanente Regional Laboratory in Berkeley (KPB), Mas-
sachusetts General Hospital (MGH), or the Centers for Disease Control
and Prevention (CDC). Vancomycin-resistant enterococci were phe-
notyped as VanA or VanB based on their sensitivity to teicoplanin.
Minimum inhibitory concentrations (MICs) were measured in a
microdilution broth procedure under NCCLS guidelines.36 Compounds
were 2-fold serially diluted into Mueller-Hinton broth in 96-well
synthesis of the C-N(43) heterodimer.
The 3-D conjugate (34 mg, 0.049 mmol) and N-linkage synthon 3
(42 mg, 0.021 mmol) were dissolved in 1 mL of anhydrous DMF and
treated sequentially with DIPEA (22 µL, 0.10 mmol) and a solution of
PyBOP/HOAT (1 M each in DMF, 25 µL, 0.025 mmol). The reaction
was stirred at room temperature for 1 h and then treated with 20 mL
of Et2O. The resulting precipitate was collected by centrifugation,
washed with Et2O, and dried under vacuum. The N-V(43) heterodimer
was purified by RP-HPLC and isolated by lyopholization as the TFA
salt. IS-MS, m/z for [M + 2H+]2+: calcd 1908.8; obsd 1909.7.
M. N-R Series. The synthesis of compounds in this series is
exemplified by the synthesis of the dimer joined by a 27-atom linker,
N-R(27).
R-Linkage synthon 5 (210 mg, 0.094 mmol) and linker synthon C
(43 mg, 0.072 mmol) were dissolved in 1.5 mL of anhydrous DMF.
DIPEA (110 µL, 0.66 mmol) was added, followed by a solution of
PyBOP/HOAT (1 M each in DMF, 120 µL, 0.11 mmol). The solution
was stirred at rt for 1 h and then treated with solid quinuclidine (40
mg, 0.50 mmol). The reaction was stirred for an additional 1.5 h and
then treated with 20 mL of Et2O. The resulting precipitate was collected
by centrifugation, washed with Et2O, and dried under vacuum. The
(36) Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria that
Grow Aerobically; ApproVed Standard 5th Ed. M7-A5. 20: 1-32. NCCLS.;
2000.
9
6530 J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003

Synthesis and Analysis of an Array of Vancomycin Dimers
A R T I C L E S
microtiter plates. Overnight cultures of bacterial strains were diluted
based on absorbance at 600 nm so that the final concentration in each
well was 5 × 105 cfu/mL. Plates were returned to a 35 °C incubator.
The following day (or 24 h in the case of enterococci), MICs were
determined by visual inspection of the plates. Because of the inability
of penicillin-resistant Streptococcus pneumoniae (PSRP) to grow well
in Mueller-Hinton broth, MICs with this strain were determined using
TSA broth supplemented with defibrinated blood.
Acknowledgment. We thank Professor George M. Whitesides
for many helpful discussions of this work.
Supporting Information Available: Linear regression analy-
sis comparison of MIC data sets. This material is available free
of charge via the Internet at http://pubs.acs.org.
JA021273S
9
J. AM. CHEM. SOC. VOL. 125, NO. 21, 2003 6531