A chemoenzymatic approach to glycopeptide antibiotics

Article abstract of DOI:10.1021/ja045147v

Many biologically active natural products are constrained by macrocyclization and modified with carbohydrates. These two types of modifications are essential for their biological activities. Here we report a chemoenzymatic approach to make carbohydrate-mo

Full text of DOI:10.1021/ja045147v

Published on Web 10/12/2004
A Chemoenzymatic Approach to Glycopeptide Antibiotics
Hening Lin and Christopher T. Walsh*
Contribution from the Department of Biological Chemistry and Molecular Pharmacology,
Armenise 616, HarVard Medical School, Boston, Massachusetts 02115
Received August 11, 2004; E-mail: christopher_walsh@hms.harvard.edu
Abstract: Many biologically active natural products are constrained by macrocyclization and modified with
carbohydrates. These two types of modifications are essential for their biological activities. Here we report
a chemoenzymatic approach to make carbohydrate-modified cyclic peptide antibiotics. Using a thioesterase
domain from the decapeptide tyrocidine synthetase, 13 head-to-tail cyclized tyrocidine derivatives were
obtained with one to three propargylglycines incorporated at positions 3-8. These cyclic peptides were
then conjugated to 21 azido sugars via copper(I)-catalyzed cycloaddition. Antibacterial and hemolytic assays
showed that the two best glycopeptides, Tyc4PG-14 and Tyc4PG-15, have a 6-fold better therapeutic index
than the natural tyrocidine. We believe this method will also be useful for modifying other natural products
to search for new therapeutics.
Sharpless and co-workers (Figure 2B).4 This cycloaddition
reaction has been used by several groups in various bioconju-
Introduction
A variety of natural products with antibiotic activities are of
polyketide (PK) origin, such as erythromycin and daunomycin,
or of nonribosomal peptide (NRP) origin, such as novobiocin
and vancomycin, or are PK/NRP hybrids, such as bleomycin
(Figure 1). Macrocyclization and glycosylation are the two late-
stage modifications in the biosynthesis of many PK or NRP
natural products, the stage that usually confers the biological
activity per se, or an improved property such as solubility or
target affinity.1 The widespread occurrence of such tailoring
macrocyclization and glycosylation encourages us to make novel
glycosylated cyclic peptides and to study the effect of different
carbohydrates on their activity.
gation experiments and has proved to be robust.5-9 In the past
two decades, the ability to make complex glycoconjugates,
including glycopeptides and glycoproteins, has increased dra-
matically.10,11 In the field of natural product modification, both
chemical and enzymatic methods and natural and unnatural
linkages have been used to conjugate carbohydrates to
peptides.12-19 However, chemical methods to form glyco-
conjugates normally require several transformations to install
appropriate functional groups for the conjugation reaction, and
the yield could be low. On the other hand, enzymatic methods
suffer from the availability of glycosyl donor substrates
(normally TDP-sugars, which are difficult to make even via
We have designed a chemoenzymatic approach to make
carbohydrate-modified cyclic peptides, which begins, as the first
example, with the cyclic decapeptide tyrocidine (Tyc) as the
peptide scaffold onto which various carbohydrates will be
attached. The head-to-tail cyclic peptide scaffold of Tyc can be
generated enzymatically by the excised thioesterase domain (TE)
from the tyrocidine synthetase using the corresponding linear
peptide N-acetyl cysteamine (SNAC) thioester as the substrate
(Figure 2A).2 This 35 kDa TE has relaxed substrate specificity
and can tolerate substitutions at most of the substrate’s 10
residues.2,3 Cyclization is accompanied by variable amounts of
enzyme-mediated, competing thioester hydrolysis. Utilizing Tyc
TE's relaxed substrate specificity, we planned to make alkyne-
containing cyclic peptides by enzymatic macrocyclization and
then conjugated them to azido sugars to produce glycosylated
cyclic peptides using the copper(I)-catalyzed [3 + 2] cyclo-
addition, a reaction condition that was recently developed by
(4) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. Angew.
Chem., Int. Ed. 2002, 41, 2596-2599.
(5) Wang, Q.; Chan, T. R.; Hilgraf, R.; Fokin, V. V.; Sharpless, K. B.; Finn,
M. G. J. Am. Chem. Soc. 2003, 125, 3192-3193.
(6) Speers, A. E.; Adam, G. C.; Cravatt, B. F. J. Am. Chem. Soc. 4686, 125,
4686-4687.
(7) Lee, L. V.; Mitchell, M. L.; Huang, S.-J.; Fokin, V. V.; Sharpless, K. B.;
Wong, C.-H. J. Am. Chem. Soc 2003, 125, 9588-9589.
(8) Link, A. J.; Tirrell, D. A. J. Am. Chem. Soc. 2003, 125, 11164-11165.
(9) Deiters, A.; Cropp, T. A.; Mukherji, M.; Chin, J. W.; Anderson, J. T.;
Schultz, P. G. J. Am. Chem. Soc. 2003, 125, 11782-11783.
(10) Davis, B. G. Chem. ReV. 2002, 102, 579-601.
(11) Grogan, M. J.; Pratt, M. R.; Marcaurelle, L. A.; Bertozzi, C. R. Annu. ReV.
Biochem. 2002, 71, 593-634.
(12) Solenberg, P. J.; Matsushima, P.; Stack, D. R.; Wilkie, S. C.; Thompson,
R. C.; Baltz, R. H. Chem. Biol. 1997, 4, 195-202.
(13) Thompson, C.; Ge, M.; Kahne, D. J. Am. Chem. Soc. 1999, 121, 1237-
1244.
(14) Nicolaou, K. C.; Cho, S. Y.; Hughes, R.; Winssinger, N.; Smethurst, C.;
Labischinski, H.; Endermann, R. Chem. Eur. J. 2001, 7, 3798-3823.
(15) Losey, H.; Peczuh, M.; Chen, Z.; Eggert, U.; Dong, S.; Pelczer, I.; Kahne,
D.; Walsh, C. Biochemistry 2001, 40, 4745-4755.
(16) Sun, B.; Chen, Z.; Eggert, U. S.; Shaw, S. J.; LaTour, J. V.; Kahne, D. J.
Am. Chem. Soc. 2001, 123, 12722-12723.
(17) Dong, S. D.; Oberthur, 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.
(1) 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.
(2) Trauger, J. W.; Kohli, R. M.; Mootz, H. D.; Marahiel, M. A.; Walsh, C. T.
Nature 2000, 407, 215-218.
(18) Ritter, T. K.; Mong, K.-K. T.; Liu, H.; Nakatni, T.; Wong, C.-H. Angew.
Chem., Int. Ed. 2003, 42, 4657-4660.
(19) Fu, X.; Albermann, C.; Jiang, J.; Liao, J.; Zhang, C.; Thorson, J. S. Nat.
Biotechnol. 2003, 21, 1467-1469.
(3) Kohli, R. M.; Walsh, C. T.; Burkart, M. D. Nature 2002, 418, 658-661.
9
13998
J. AM. CHEM. SOC. 2004, 126, 13998-14003
10.1021/ja045147v CCC: $27.50 © 2004 American Chemical Society

Chemoenzymatic Approach to Glycopeptide Antibiotics
A R T I C L E S
Figure 1. Structures of representative natural antibiotics with carbohydrate moieties.
Figure 2. Chemoenzymatic approach (B) to make glycopeptidevariants of the antibiotic tyrocidine using Tyc TE catalyzed macrocyclization (A) and copper(I)
catalyzed [3 + 2] cylcoaddition.
enzymatic methods19) and substrate specificity of the glycosyl-
transferases. Therefore, new chemoselective ligation reactions
could potentially offer advantages.
Tyc is a NRP antibiotic, and it is believed that it targets the
lipid bilayer of bacteria to make pores.20 Because there is no
specific bacterial protein target for the cyclic peptides of the
tyrocidine class, drug resistance is rare, which is particularly
beneficial considering the increasing antibiotic resistance in
bacteria. However, Tyc has the liability that it can also insert
into eukaryotic membranes and cause lysis of human red blood
cells, which limits its systemic application.21 Here, by introduc-
ing carbohydrates into Tyc, we seek to improve its therapeutic
index which is defined by the ratio of the minimal hemolytic
concentration (MHC) to the minimal inhibitory concentration
(MIC) of bacteria growth. It should be noted that even though
carbohydrates are important for the functions of many natural
(20) Izumiya, N.; Kato, T.; Aoyagi, H.; Waki, M.; Kondo, M. Kondansha
International Ltd.: Tokyo, 1979.
(21) Lambert, H. P.; O’Grady, F. W. Antibiotic and Chemotherapy, 6th ed.;
Churchill Livingstone: Edinburgh, U.K., 1992; pp 232-233.
9
J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004 13999

A R T I C L E S
Lin and Walsh
Figure 3. 21 azido sugars used in this study.
Table 1. Cyclization Efficiency of PG-Substituted Peptide-SNACs
products, modification of a natural product’s sugar portion may
not always confer beneficial effects, as indicated by related work
on vancomycin.14,18,19 Therefore, to maximize the chance of our
success with the modification of tyrocidine, we decided to take
a combinatorial approach, and as will be shown, the choice of
Sharpless’s click chemistry greatly facilitates its implementation.
peptide
−
SNAC
complete consumption
of substrate?a
cyclization-to-hydrolysis
ratioa
cyclic peptide
prepared?
substrate
2PG
3PG
4PG
5PG
6PG
7PG
8PG
56PG
57PG
58PG
67PG
68PG
78PG
567PG
568PG
578PG
678PG
-
+
+
+
+
+
+
-
+
+
-
+
+
-
-
+
+
NDb
3.5
6.2
3.7
11.1
3.2
3.3
0.04
1.2
2.3
0.1
3.8
2.4
0.9
NDb
0.5
3.4
-
+
+
+
+
+
+
-
+
+
-
+
+
+
-
+
+
Results
Macrocyclization of Propargylglycine Substituted Tyc
Linear Peptide-Thioesters by TE. Tyc linear peptides
containing propargylglycine (PG) at different positions were
synthesized on solid support. For monosubstituted peptides,
position 2-8 were used. For di- and tri-substituted peptides,
only positions 5-8 were used because these positions are more
tolerant to substitutions than others for subsequent TE-mediated
head-to-tail macrolactamization. After cleavage from solid
support, these peptides were then coupled to N-acetylcysteamine
and deprotected using published procedures.2 Seven mono-
substituted, six di-substituted, and four trisubstituted propar-
gylglycine variant peptide-SNACs were synthesized. These 17
peptide-SNACs were then subjected to TE-catalyzed cyclization.
Most of them (13 out of 17) can be cyclized with reasonable
efficiency (Table 1). The competing reaction is TE-mediated
hydrolysis of the peptidyl-SNAC as noted by cyclization/
hydrolysis ratio. For these 13 peptide-SNACs, preparative scale
reactions were carried out to give enough quantities of the cyclic
peptides to be used in the subsequent cycloaddition reaction.
Choice and Synthesis of Azido Monosacchrides. For azido
sugars, we have focused initially on monohexoses (Figure 3),
most of which (1-12) can be easily accessed following
published procedures.22-28 The acyl chain and the biphenyl
group in azido sugars 14-18 were incorporated to increase
a Determined with 100 µM substrate and 100 nM TE, room temperature,
1 h. b ND: not determined because almost no reaction occurred.
membrane affinity as inspired by teicoplanin and other vanco-
mycin-type lipoglycopeptide antibiotics.29 The detailed synthesis
of 13-21 can be found in the Supporting Information.
Preparation of the Glycopeptide Library by Click Chem-
istry. We carried out the cycloaddition reaction in 96-well plates
to facilitate subsequent MIC and MHC assays. Sharpless’s
(23) Wang, P.; Shen, G. J.; Wang, Y. F.; Ichikawa, Y.; Wong, C. H. J. Org.
Chem. 1993, 58, 3985-3990.
(24) May, J. A.; Sartorelli, A. C. J. Med. Chem. 1979, 22, 971-976.
(25) Maier, M. A.; Yannopoulos, C. G.; Mohamed, N.; Roland, A.; Fritz, H.;
Mohan, V.; Just, G.; Manoharan, M. Bioconjugate Chem. 2003, 14, 18-
29.
(26) Chernyak, A. Y.; Sharma, G. V. M.; Kononov, L. O.; Krishna, P. R.;
Levinsky, A. B.; Kochetkov, N. K. Carbohydr. Res. 1992, 223, 303-309.
(27) Ni, J.; Singh, S.; Wang, L.-X. Bioconjugate Chem. 2003, 14, 232-238.
(28) Roy, R.; Kim, J. M. Tetrahedron 2003, 59, 3881-3893.
(29) Kerns, R.; Dong, S. D.; Fukuzawa, S.; Carbeck, J.; Kohler, J.; Silver, L.;
Kahne, D. J. Am. Chem. Soc. 2000, 122, 12608-12609.
(22) Fazio, F.; Bryan, M. C.; Blixt, O.; Paulson, J. C.; Wong, C. H. J. Am.
Chem. Soc. 2002, 124, 14397-14402.
9
14000 J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004

Chemoenzymatic Approach to Glycopeptide Antibiotics
A R T I C L E S
Table 2. MIC and MHC Values for Reagents Used in the [3 + 2]
tyrocidine’s MIC and MHC values measured in the presence
of all three reagents in 2.5- or 5-fold excess. The MIC and MHC
values of the 21 azido sugars were also tested and are all greater
than 200 µM. Therefore the use of more than 1 equiv of azido
sugars should cause no problems for the assays.
Cycloaddition Reaction
reagent
MIC/
µM
MHC/µM
Cu2+
ascorbate
TCEP
1250
>10000
5000
160
>10000
1250
The tested concentration range of glycopeptides is from 200
to 1.5 µM for MHC and 25 to 0.1 µM for MIC. The MIC and
MHC data for the mono-glycosylated peptides are shown in
Table S2 in the Supporting Information. These data reveal two
trends: first, most simple sugars cannot increase the therapeutic
index (MHC/MIC) of the peptides, and the best sugars (14 and
15) are those that have either a nonoyl group or a biphenyl
group; second, the best position to put the sugars is position 4,
while positions 5 and 6 are also acceptable.
Among the di- and tri-glycosylated peptides, none has better
therapeutic index (data not shown) than tyrocidine. In fact, most
of them have MIC's higher than 25 µM. It seems that modifying
more than one position with the 20 or so sugars is detrimental
to the antibiotic activity, at least for positions 5-8.
Biological Activity Assays of Selected Glycopeptides Using
Purifed Products. To verify the library MIC and MHC assay
results, the cycloaddition was repeated to synthesize the two
glycopeptides with the best therapeutic index, Tyc4PG-14 and
Tyc4PG-15 (Figure 4). MIC and MHC assays were then carried
out using HPLC purified compounds. The result indicates that
Tyc4PG-14 and Tyc4PG-15 have at least a 6-fold better
therapeutic index compared with wild-type tyrocidine (entries
1-3 in Table 3). Two other glycopeptides Tyc4PG-18 and
Tyc7PG-1 were also purified and assayed (entries 4 and 5 in
Table 3) to further examine the results from library screening.
The assay results using the cycloaddition mixtures are similar
to those obtained using purified compounds.
Divalent Display of Azido Sugars at Postion 4. Since the
best position to bear the sugar is position 4, it was of interest
to see whether more than one sugar at position 4 would further
affect MHC and MIC values. To this end, two branched glycan
chain peptides, Tyc4PG-142 and Tyc4PG-152 (Figure 4), were
synthesized. However, the MIC and MHC assays showed these
two glycopeptides have no advantage compared with Tyc4PG-
14 and Tyc4PG-15, even though Tyc4PG-152 is still slightly
better than tyrocidine itself (Table 3, entries 6 and 7).
tris(triazolyl)amine ligand
tyrocidine
5000
>10000
1.5
1.5
25
tyrocidine with 2.5 equiv of Cu2+
5 equiv of ascorbate, and
5 equiv of ligand
,
12.5
original condition for azido-alkyne [3 + 2] cycloaddition is to
use sodium ascorbate to reduce CuII to CuI.4 However, in later
publications, slightly different conditions have been used; for
example, the use of Cu wire or tris(carboxyethyl)phosphine
(TCEP) as reducing reagent and the use of a tris(triazolyl)amine
ligand.5-8 For our experiments, we chose to run the reaction at
about millimolar concentrations and then carry out MIC and
MHC assays initially without any purification process. There-
fore, we utilized reagents that give fast reaction rates and do
not interfere with the MIC and MHC assays. Sodium ascorbate
was the reducing reagent of choice since it is less toxic than
TCEP to both bacteria and human red blood cells (Table 2).
Using cyclic peptide Tyc3PG and azido sugar 1, we found
conditions (0.7 mM peptide, 1.4 mM azido sugar, 1.7 mM
CuSO4, 3.4 mM ascorbate in a 60 µL reaction) that give
complete conversion of starting material for mono-PG-
substituted peptides in 2 h. The reaction products, however,
contain both the desired triazole product and a minor product
whose mass is the desired triazole product plus 16. We assigned
this M + 16 product as the 5-hydroxytriazole product, which
has been observed before.4 Under similar conditions, the
reactions involving di- and trisubstituted peptides could not go
to completion. However, this problem was solved by the addition
of tris(triazolyl)amine ligand (0.7 mM peptide, 2.7 or 3.6 mM
azido sugar, 1.8 mM CuSO4, 3.6 mM ligand, and 22 mM
ascorbate in a 55 µL reaction). Under this condition, the
formation of 5-hydroxytriazole is almost completely suppressed.
Using these two sets of conditions, a library of 13 × 21
cycloaddtion reactions were conducted. The quality of the
product library was determined by LCMS. For most monosub-
stituted peptides the quality is good, with >90% conversion to
products and minimal 5-hydroxytriazole products. But for
disubstituted peptides, incomplete reactions were observed when
using azido sugars 17 and 18. For trisubstituted peptides,
incomplete reactions were observed when using azido sugars
13, 15, 17, 18, 20, and 21. Therefore, the total number of
glycopeptides made in this two-step enzymatic macrocyclization,
triazole-forming glycosylation is 247.
Discussion
We have generated a library of 247 glycopeptides by
combining enzymatic macrocyclization by TE and the ligation
reaction via click chemistry. The relaxed substrate specificity
of TE from the tyrocidine synthetase enabled the preparation
of unprotected Tyc cyclic peptide derivatives with propargyl-
glycine substituted at various positions, further demonstrating
the utility of this enzyme. The unique features of the copper(I)-
catalyzed [3 + 2] cycloaddition, such as the mild reaction
condition, the orthogonality of azide and alykne with other
functional groups, the ease of preparing azido glycosides, and
the ability of the reaction to proceed at less than 1 mM
concentration on a 40 nmol scale, greatly facilitated the
construction of the library in a relatively short period of time.
It also facilitated the subsequent MIC and MHC assays because
the reaction is clean and the reagents used do not interfere with
the assays at the concentrations used. For a combinatorial library
approach, these features are essential and are advantageous
Biological Activity Assays of the Glycopeptide Library.
We then carried out biological activity assays of these glyco-
peptides. After the cycloaddition reaction, the reaction mixtures
were used directly for MIC and MHC assays.3 As a control it
was important to make sure that the reagents (Cu2+, ascorbate,
and the ligand) used in the cycloaddition do not interfere with
the assays at the concentrations used. Data in Table 2 showed
that all the reagents used have MIC and MHC values greater
than 1 mM, except that copper(II) has a MHC of 160 µM. Since
tyrocidine’s MIC is only 1.5 µM and MHC is 25 µM, these
reagents, when used in severalfold excess, should not interfere
with the MIC and MHC assays. This is further supported by
9
J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004 14001

A R T I C L E S
Lin and Walsh
Figure 4. Structures of glycopeptide Tyc4PG-14, Tyc4PG-15, Tyc4PG-142, and Tyc4PG-152.
Table 3. MIC and MHC Values Obtained Using Purified
approach. For example, it is not easy to simultaneously ligate
different azido sugars to the Tyc scaffold with multiple
propargylglycine residues.
Glycopeptides
peptides
MIC/
µM
MHC/
µ
M
MHC/MIC
1
2
3
4
5
6
7
tyrocidine
1.5
3
3
6
6
25
>300
>300
100
25
16
>100
>100
16
From this 247-member library, glycopeptides with more than
6-fold improved therapeutic indexes over the native Tyc were
obtained while maintaining Tyc’s antibacterial potency. It is
worth noting that carbohydrate randomization of natural products
does not always improve activity.14,18,19 This is probably because
the functional mechanisms of the carbohydrate portion and/or
the rest of the natural product in each case could be different.
Such complication is also clearly evident in our study. For
example, we found that position 4 is the best position to attach
the sugars, and the best sugars are those with lipophilic chains
(14 and 15). While it is easy to explain the function of the
Tyc4PG-14
Tyc4PG-15
Tyc4PG-18
Tyc7PG-1
Tyc4PG-142
Tyc4PG-152
4
2
32
12
1.5
25
50
compared with the in Vitro glycorandomization method reported
in a recent publication,19 in which only a portion of the library
generated was assayed for biological activity. Of course, as with
any other method, limitations do exist for the click chemistry
9
14002 J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004

Chemoenzymatic Approach to Glycopeptide Antibiotics
A R T I C L E S
purified (20-100% acetonitrile/0.1% TFA over 50 min). Cyclic peptides
were prepared using 100 µM peptide-SNAC thioesters and 100 nM
TE in 25 mM MOPS buffer (pH 7) with 0.1% Brij58 at room
temperature for 3 h. The cyclization-to-hydrolysis ratio was calculated
on the basis of the areas of absorption at 220 nm from LCMS. The
cyclic peptides were purified by HPLC (30-100% acetonitrile/0.1%
TFA over 40 min). The lyophilized peptides were dissolved in methanol,
and the concentrations were determined by analytical HPLC using a
known concentration of Tyc peptide-SNAC thioester.
[3 + 2] Cycloaddition for Mono-PG-Substituted Tyc Peptides.
The following solutions were made: 1.5 mM cyclic peptides in
methanol (solution A), 50 mM azido sugars in methanol (solution B),
10 mM CuSO4 in water (solution C), and 10 mM sodium ascorbate
(solution D, make before use). The reactions were carried out in 96-
well plates. Reagents were added in the following sequence: 27 µL of
solution A, 1.6 µL of solution B, 10 µL of solution C, and 20 µL of
solution D. The wells were then immediately sealed with Thermowell
sealers (Corning Incorporated, Catalog No. 6570), and the plates were
incubated with shaking at room temperature for 3 h. The reactions were
checked by LCMS to ensure the quality of the glycopeptide library.
lipophilic chains’ since they are known to increase membrane
affinity, it is difficult to explain why similar azido sugars with
only slight changes in the sugar portion, such as 16-18, gave
different results from those of 14 and 15. Similarly, without
molecular details of the interactions of these glycopeptides with
bacterial and human membranes, it is difficult to explain the
selectivity observed. This emphasizes the importance of a
combinatorial library approach for this kind of study.
It is also interesting that displaying multiple sugars on the
Tyc scaffold did not further improve the therapeutic index of
tyrocidine, even with azido sugars 14 and 15. For Tyc cyclic
peptides with multiple propargylglycine residues at positions
5-8, one could argue that the sugars are not displayed at the
right positions. However, for Tyc4PG-142 and Tyc4PG-152, the
result is still negative. For Tyc4PG-152, both MIC and MHC
are lower than those of Tyc4PG-15, indicating that displaying
two copies of 15 does increase membrane affinity. However,
the change in MHC is more significant than the change in MIC,
resulting in a lower therapeutic index compared with Tyc4PG-
15. Multivalent effect can exist when a targeted partner has
multiple binding sites.30 A priori, it is not clear that this condition
would apply to tyrocidine, and it is possible that the conjugated
sugars could somehow block the function of tyrocidine as
implied by the MIC and MHC of Tyc4PG-142. Therefore, it
may not be surprising that displaying multiple sugars on the
Tyc scaffold did not further improve the therapeutic indexes in
this study.
[3 + 2] Cycloaddition for Di- and Tri-PG-Substituted Tyc
Peptides. The condition was similar to that for the mono-PG-substituted
Tyc peptides unless otherwise stated. The following solutions were
made in addition to solutions A and B: 20 mM CuSO4 in water
(solution E), 120 mM sodium ascorbate (solution F, make before use),
and 20 mM tris(triazolyl)amine ligand (solution G). Reagents were
added in the following sequence: 27 µL of solution A, 3 µL (for di-
PG-substituted peptides) or 4 µL (for tri-PG-substituted peptides) of
solution B, 5 µL of solution E, 10 µL of solution G, and 10 µL of
solution F.
There is no reason to think this approach will be limited to
the tyrocidine skeleton. We believe this combination of enzyme-
mediated linear precursor macrocyclization followed by azide
glycoside cycloaddition will be generalizable to other TE-
mediated macrocyclizations in both polyketide and hybrid
polyketide/nonribosomal peptide scaffolds.31-33 By combining
chemical modifications to introduce alkyne groups, this strategy
should also be useful for modification of other natural products
that are not accessible through TE-mediated cyclization to search
for new therapeutics.
MHC Assay of the Glycopeptide Library. To the reaction mixtures
were added appropriate amounts of methanol to make the final volume
100 µL. Then 50 µL was transferred to new 96-well plates to make
serial 2× dilutions (200 µM to 1.6 µM) of the cyclic peptides with
methanol. Methanol was then removed with a speedvac. Human red
blood cells (Research Blood Components, Boston, MA) were diluted
100× with PBS buffer (pH 7.4), and 50 µL was added to each well.
The plates were incubated at room temperature with rocking, and the
concentrations required for complete lysis were determined visually
after overnight incubation.
Experimental Section
MIC Assay of the Glycopeptide Library. A 10 µL aliquot of the
diluted reaction mixture was transferred to 96-well plates to make serial
2× dilutions (25 µM to 0.2 µM) with methanol. Methanol was then
removed with a speedvac. An overnight B. subtilis PY79 culture was
diluted 10000× with LB media, and 80 µL of the diluted culture was
added to each well. After incubation in a shaker at 30 °C overnight,
the concentrations required for complete inhibition of bacterial cell
growth were determined visually.
Preparation of PG-Substituted Tyc Cyclic Peptides. Tyc TE was
expressed and purified as previously reported.2 Linear peptides were
synthesized on a Symphony peptide synthesizer using standard Fmoc
solid-phase peptide synthesis (the last residue D-Phe10 was Boc
protected). The peptides were cleaved from the 2-Cl Trt resin (3:1:1
dichloromethane:trifluoroethanol:acetic acid), and solvents were re-
moved by azeotroping with hexanes. The peptides (∼50 µmol) were
then coupled to N-acetylcysteamine (10 equiv) using dicyclohexylcar-
bodiimide (2 equiv)/1-hydroxybenzotriazole hydrate (2 equiv)/N,N-
diisopropylethylamine (4 equiv) in tetrahydrofuran (THF; 1 mL). After
removal of THF, the peptide-SNAC thioesters were deproteced using
95:2.5:2.5 trifluoroacetic acid (TFA):water:triisopropylsilane and HPLC
Acknowledgment. We gratefully acknowledge the National
Institutes of Health Grant NIH GM20011 (to C.T.W.), and The
Jane Coffin Childs Memorial Fund for Medical Research for a
fellowship award (to H.L.).
(30) Mammen, M.; Choi, S.-K.; Whitesides, G. M. Angew. Chem., Int. Ed. 1998,
37, 2754-2794.
Supporting Information Available: The synthesis of azido
sugars 13-21, 142, and 152 and MS, MIC, and MHC data for
the monosubstituted glycopeptide. This material is available free
of charge via the Internet at http://pubs.acs.org.
(31) Tseng, C. C.; Bruner, S. D.; Kohli, R. M.; Marahiel, M. A.; Walsh, C. T.;
Sieber, S. A. Biochemistry 2002, 41, 13350-13359.
(32) Sieber, S. A.; Tao, J.; Walsh, C. T.; Marahiel, M. A. Angew. Chem., Int.
Ed. 2004, 43, 493-498.
(33) Grunewald, J.; Sieber, S. A.; Marahiel, M. Biochemistry 2004, 43, 2915-
2925.
JA045147V
9
J. AM. CHEM. SOC. VOL. 126, NO. 43, 2004 14003