A systematic investigation of the synthetic utility of glycopeptide glycosyltransferases

Article abstract of DOI:10.1021/ja052945s

Glycosyltransferases involved in the biosynthesis of bacterial secondary metabolites may be useful for the generation of sugar-modified analogues of bioactive natural products. Some glycosyltransferases have relaxed substrate specificity, and it has been assumed that promiscuity is a feature of the class. As part of a program to explore the synthetic utility of these enzymes, we have analyzed the substrate selectivity of glycosyltransferases that attach similar 2-deoxy-L-sugars to glycopeptide aglycons of the vancomycin-type, using purified enzymes and chemically synthesized TDP β-2-deoxy-L-sugar analogues. We show that while some of these glycopeptide glycosyltransferases are promiscuous, others tolerate only minor modifications in the substrates they will handle. For example, the glycosyltransferases GtfC and GtfD, which transfer 4-epi-L-vancosamine and L-vancosamine to C-2 of the glucose unit of vancomycin pseudoaglycon and chloroorienticin B, respectively, show moderately relaxed donor substrate specificities for the glycosylation of their natural aglycons. In contrast, GtfA, a transferase attaching 4-epi-L-vancosamine to a benzylic position, only utilizes donors that are closely related to its natural TDP sugar substrate. Our data also show that the spectrum of donors utilized by a given enzyme can depend on whether the natural acceptor or an analogue is used, and that GtfD is the most versatile enzyme for the synthesis of vancomycin analogues.

Full text of DOI:10.1021/ja052945s

Published on Web 07/08/2005
A Systematic Investigation of the Synthetic Utility of
Glycopeptide Glycosyltransferases
Markus Oberthu¨r,§,‡ Catherine Leimkuhler,‡ Ryan G. Kruger,§ Wei Lu,§
Christopher T. Walsh,§ and Daniel Kahne*,‡,§
Contribution from the Department of Chemistry and Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138, and Department of Biological Chemistry and Molecular
Pharmacology, HarVard Medical School, Boston, Massachusetts 02115
Received May 5, 2005; E-mail: kahne@chemistry.harvard.edu
Abstract: Glycosyltransferases involved in the biosynthesis of bacterial secondary metabolites may be
useful for the generation of sugar-modified analogues of bioactive natural products. Some glycosyltrans-
ferases have relaxed substrate specificity, and it has been assumed that promiscuity is a feature of the
class. As part of a program to explore the synthetic utility of these enzymes, we have analyzed the substrate
selectivity of glycosyltransferases that attach similar 2-deoxy-L-sugars to glycopeptide aglycons of the
vancomycin-type, using purified enzymes and chemically synthesized TDP â-2-deoxy-L-sugar analogues.
We show that while some of these glycopeptide glycosyltransferases are promiscuous, others tolerate
only minor modifications in the substrates they will handle. For example, the glycosyltransferases GtfC
and GtfD, which transfer 4-epi-L-vancosamine and L-vancosamine to C-2 of the glucose unit of vancomycin
pseudoaglycon and chloroorienticin B, respectively, show moderately relaxed donor substrate specificities
for the glycosylation of their natural aglycons. In contrast, GtfA, a transferase attaching 4-epi-L-vancosamine
to a benzylic position, only utilizes donors that are closely related to its natural TDP sugar substrate. Our
data also show that the spectrum of donors utilized by a given enzyme can depend on whether the natural
acceptor or an analogue is used, and that GtfD is the most versatile enzyme for the synthesis of vancomycin
analogues.
Introduction
The glycopeptide antibiotic vancomycin (Figure 1) has gained
prominence as the antibiotic of last resort for the treatment of
life-threatening infections by methicillin-resistant Gram-positive
bacteria. With the global emergence of vancomycin-resistant
enterococci (VRE)1,2 and, more recently, vancomycin-resistant
Staphylococcus aureus (VRSA),1,3 the development of novel
glycopeptides with improved activity against resistant strains
has become a major research focus. In 1988, researchers at Eli
Lilly showed that glycopeptides containing lipophilic substit-
uents on the carbohydrate moiety have significant activity
against vancomycin-resistant strains.4 As a result of these
studies, the semisynthetic glycopeptide oritavancin (Figure 1),
a derivative of the glycopeptide chloroeremomycin, was put into
human clinical trials.1,5
Figure 1. Glycopeptide antibiotics.
Due to the complexity of the glycopeptide class of natural
products, early investigations of glycopeptide derivatives in-
volved exploring substituent changes on sites that could be
modified easily.4 Because this precluded studies on glycopeptide
derivatives containing unnatural sugars, our laboratory6 and the
Nicolaou group7 have reported chemical glycosylation strategies
to attach unnatural sugars to the vancomycin aglycon. In 1999,
we showed that a lipidated vancomycin analogue containing
‡ Harvard University.
§ Harvard Medical School.
(1) Kahne, D.; Leimkuhler, C.; Lu, W.; Walsh, C. T. Chem. ReV. 2005, 105,
425-448.
(2) (a) Cetinkaya, C.; Falk, P.; Mayhall, C. G. Clin. Microbiol. ReV. 2000, 13,
686-707. (b) Walsh, C. T.; Fisher, S. L.; Park, I. S.; Prahalad, M.; Wu, Z.
Chem. Biol. 1996, 3, 21-28.
(6) (a) Ge, M.; Thompson, C.; Kahne, D. J. Am. Chem. Soc. 1998, 120, 11014-
11015. (b) Thompson, C.; Ge, M.; Kahne, D. J. Am. Chem. Soc. 1999,
121, 1237-1244.
(7) (a) Nicolaou, K. C.; Mitchell, H. J.; Jain, N. F.; Winssinger, N.; Hughes,
R.; Bando, T. Angew. Chem., Int. Ed. 1999, 38, 240-244. (b) Nicolaou,
K. C.; Cho, S. Y.; Hughes, R.; Winssinger, N.; Smethurst, C.; Labischinski,
H.; Endermann, R. Chem.sEur. J. 2001, 7, 3798-3823.
(3) Fridkin, S. K. et al. N. Engl. J. Med. 2003, 348, 1342-1347.
(4) (a) Nagarajan, R.; Schabel, A. A.; Occolowitz, J. L.; Counter, F. T.; Ott,
J. L. J. Antibiot. 1988, 41, 1430-1438. (b) Nagarajan, R.; Schabel, A. A.;
Occolowitz, J. L.; Counter, F. T.; Ott, J. L.; Felty-Duckworth, A. M. J.
Antibiot. 1989, 42, 63-72.
(5) Allen, N. E.; Nicas, T. I. FEMS Microbiol. ReV. 2003, 26, 511-532.
9
10.1021/ja052945s CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 10747-10752
10747

A R T I C L E S
Oberthu¨r et al.
daunosamine was more active against some resistant bacterial
strains than the corresponding vancosamine derivative.8 We also
suggested that glycopeptide derivatives, such as oritavancin,
possess a second mechanism of action that is different from
vancomycin,8 which prevents maturation of the bacterial cell
wall by binding to the terminal D-alanyl-D-alanine moiety of
peptidoglycan precursors, thereby inhibiting the enzymes in-
volved in the final stages of peptidoglycan synthesis. More
recently, we could indeed show that lipidated glycopeptides
interfere with peptidoglycan synthesis by direct inhibition of
the major transglycosylases of Escherichia coli9 and S. aureus.10
These findings provided the impetus for exploring additional
carbohydrate derivatives to examine the influence of structural
changes in the lipid-disaccharide portion on antibiotic activity
and on the ability to inhibit the transglycosylase enzymes. Since
chemical glycosylation strategies do not enable rapid exploration
of significant numbers of glycopeptide derivatives because of
the number of synthetic steps necessary,11 we became interested
in exploring enzymatic approaches to generate glycopeptides.
glycopeptide analogues rapidly.15c,d,16 We wanted to know
whether the relaxed substrate selectivity of GtfE was typical of
the class because, then, the use of these glycosyltransferases in
combination with either chemically17 or enzymatically18,19
synthesized nucleotide diphosphate (NDP) sugar donors would
allow the generation of a large number of sugar-modified
analogues in a straightforward way.20 Accordingly, we decided
to compare the donor substrate selectivity of three structurally
related glycopeptide glycosyltransferases that attach similar
2-deoxy-L-sugars to glucosylated vancomycin aglycons. In the
biosynthesis of chloroeremomycin, GtfA transfers 4-epi-L-
vancosamine to the benzylic hydroxyl of amino acid 7 of the
vancomycin pseudoaglycon 1 to produce chloroorienticin B
(2).21 GtfC then transfers 4-epi-L-vancosamine to the glucose
C-2 hydroxyl of compound 2.21 GtfD, on the other hand, is part
of the vancomycin biosynthetic cluster and transfers L-van-
cosamine to the glucose C-2 hydroxyl of pseudoaglycon 1
(Figure 2).15e
We report here the first comprehensive in vitro study of
purified glycosyltransferases involved in the biosynthesis of
bacterial secondary metabolites that transfer 2-deoxy-L-sugars.
The glycopeptide glycosyltransferases GtfC and GtfD show a
moderately relaxed substrate selectivity for the glycosylation
of their natural aglycons, whereas GtfA and, in addition, GtfC
with vancomycin pseudoaglycon 1 as acceptor only utilize
donors that are closely related to their natural NDP sugar
substrate. This difference in promiscuity indicates that in order
to fully exploit the synthetic utility of glycosyltransferases
involved in the biosynthesis of other bioactive natural products,
a detailed analysis of the specific enzymes will be necessary.
Recent investigations have shown that some glycosyltrans-
ferases have a relaxed substrate selectivity,12 suggesting that
these enzymes may be useful for the chemoenzymatic synthesis
of antibiotic analogues containing unnatural carbohydrates.
Some of these studies have explored the glycosylation of
unnatural aglycon substrates,13 whereas other studies focused
on variations in the sugar substrate structure.14,15 For example,
investigations of GtfE, which transfers D-glucose to the central
4-hydroxyphenylglycine of vancomycin aglycon, have shown
that this enzyme transfers a range of unnatural deoxy and amino
sugar derivatives to both the vancomycin and teicoplanin
aglycons, making it possible to prepare a number of different
Results
Synthesis of TDP Sugar Donors. In preliminary studies, we
were able to show that GtfD can transfer 4-epi-L-vancosamine
instead of its natural substrate, L-vancosamine, to vancomycin
pseudoaglycon 1, and that the glucosylated aglycon of the
(8) 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.
(9) Chen, L.; Walker, D.; Sun, B.; Hu, Y.; Walker, S.; Kahne, D. Proc. Natl.
Acad. Sci. U.S.A. 2003, 100, 5658-5663.
(10) Leimkuhler, C.; Chen, L.; Barrett, D.; Panzone, G.; Sun, B.; Falcone, B.;
Oberthu¨r, M.; Donadio, S.; Walker, S.; Kahne, D. J. Am. Chem. Soc. 2005,
127, 3250-3251.
glycopeptide teicoplanin can also serve as a substrate.15e
A
detailed study of the substrate specificity of GtfD, as well as of
GtfA and C, was hampered by the synthetically challenging
access to the required â-2-deoxy glycosyl thymidine diphosphate
(TDP) donors. We have recently developed a chemical route
to synthesize â-2-deoxy glycosyl phosphates from 2-deoxy
glycosyl chlorides using the phosphate donor tetrabutylammo-
nium dihydrogenphosphate.22 The glycosyl phosphates are then
converted to the desired TDP sugar donors with thymidine 5′-
monophosphomorpholidate (TMP morpholidate)17a and then
(11) Leimkuhler, C.; Chen, Z.; Kruger, R. G.; Oberthu¨r, M.; Lu, W.; Walsh, C.
T.; Kahne, D. Tetrahedron: Asymmetry 2005, 16, 599-603.
(12) For recent reviews, see: (a) Walsh, C. T. ChemBioChem 2002, 3, 124-
134. (b) Mendez, C.; Salas, J. A. Trends Biotech. 2001, 19, 449-456.
(13) (a) Minami, A.; Uchida, R.; Eguchi, T.; Kakinuma, K. J. Am. Chem. Soc.
2005, 127, 6148-6149. (b) Li, S.-m.; Heide, L. Curr. Med. Chem. 2005,
12, 419-427. (c) Freel Meyers, C. L.; Oberthu¨r, M.; Heide, L.; Kahne,
D.; Walsh, C. T. Biochemistry 2004, 43, 15022-15036. (d) Eusta´quio, A.
E.; Gust, B.; Li, S.-m.; Pelzer, S.; Wohlleben, W.; Chater, K. F.; Heide, L.
Chem. Biol. 2004, 11, 1561-1572. (e) Rohr, J. et al. Chem. Biol. 2004,
11, 547-555. (f) Freel Meyers, C. L.; Oberthu¨r, M.; Anderson, J. W.;
Kahne, D.; Walsh, C. T. Biochemistry 2003, 42, 4179-4189. (g) Tang,
L.; McDaniel, R. Chem. Biol. 2001, 8, 547-555.
(14) For recent examples of syntheses of carbohydrate analogues of bacterial
secondary metabolites using an in vivo approach, see: (a) Pe´rez, M,;
Lombo´, F.; Zhu, L.; Gibson, M.; Bran˜a, A. F.; Rohr, J.; Salas, J. A.; Me´ndez,
C. Chem. Commun. 2005, 1604-1606. (b) Melanc¸on, C. E., III; Takahashi,
H.; Liu, H.-w. J. Am. Chem. Soc. 2004, 126, 16726-16727. (c) Lombo´,
F.; Gibson, M.; Greenwell, L.; Bran˜a, A. F.; Rohr, J.; Salas, J. A.; Me´ndez,
C. Chem. Biol. 2004, 11, 1709-1718. (d) Hoffmeister, D.; Dra¨ger, G.;
Ichinose, K.; Rohr, J.; Bechthold, A. J. Am. Chem. Soc. 2003, 125, 4678-
4679. (e) Rodriguez, L.; Aguirrezabalaga, I.; Allende, N.; Bran˜a, A. F.;
Me´ndez, C.; Salas, J. A. Chem. Biol. 2002, 9, 721-729. (f) Borisova, S.
A.; Zhao, L.; Sherman, D. H.; Liu, H.-w. Org. Lett. 1999, 1, 133-136.
(15) Examples for the chemoenzymatic synthesis of analogues using purified
glycosyltransferases: (a) Lu, W.; Leimkuhler, C.; Oberthu¨r, M.; Kahne,
D.; Walsh, C. T. Biochemistry 2004, 43, 4548-4558. (b) Albermann, C.;
Soriano, A.; Jiang, J.; Vollmer, H.; Biggins, J. B.; Barton, W. A.; Lesniak,
J.; Nikolov, D. B.; Thorson, J. S. Org. Lett. 2003, 5, 933-936. (c) Fu, X.;
Albermann, C.; Jiang, J.; Liao, J.; Zhang, C.; Thorson, J. S. Nat. Biotechnol.
2003, 21, 1467-1469. (d) Losey, H. C.; Jiang, J.; Biggins, J. B.; Oberthu¨r,
M.; Ye, X.-Y.; Dong, S. D.; Kahne, D.; Thorson, J. S.; Walsh, C. T. Chem.
Biol. 2002, 9, 1305-1314. (e) Losey, H. C.; Peczuh, M. W.; Chen, Z.;
Eggert, U. S.; Dong, S. D.; Pelczer, I.; Kahne, D.; Walsh, C. T. Biochemistry
2001, 40, 4745-4755.
(16) 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.
(17) (a) Wittmann, V.; Wong, C.-H. J. Org. Chem. 1997, 62, 2144-2147. (b)
Fang, X.; Gibbs, B. S.; Coward, J. K. Bioorg. Med. Chem. Lett. 1995, 5,
2701-2706. (c) Arlt, M.; Hindsgaul, O. J. Org. Chem. 1995, 60, 14-15.
(18) Syntheses of NDP sugars using multienzyme systems: (a) Chen, H.;
Thomas, M. G.; Hubbard, B. K.; Losey, H. C.; Walsh, C. T.; Burkart, M.
D. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 11942-11947. (b) Amann, S.;
Dra¨ger, G.; Rupprath, C.; Kirschning, A.; Elling, L. Carbohydr. Res. 2001,
335, 23-32.
(19) Enzymatic synthesis of NDP sugars using nucleotidyltransferases: (a)
Wong, C.-H.; Halcomb, R. L.; Ichikawa, Y.; Kajimoto, T. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 521-546. (b) Jiang, J.; Albermann, C.; Thorson,
J. S. ChemBioChem 2003, 4, 443-446. (c) Mizanur, R. M.; Zea, C. J.;
Pohl, N. L. J. Am. Chem. Soc. 2004, 126, 15993-15998.
(20) Yang, J.; Hoffmeister, D.; Liu, L.; Fu, X.; Thorson, J. S. Bioorg. Med.
Chem. 2004, 12, 1577-1584.
(21) Lu, W.; Oberthu¨r, M.; Leimkuhler, C.; Tao, J.; Kahne, D.; Walsh, C. T.
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 4390-4395.
(22) Oberthu¨r, M.; Leimkuhler, C.; Kahne, D. Org. Lett. 2004, 6, 2873-2876.
9
10748 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Synthetic Utility of Glycopeptide Glycosyltransferases
A R T I C L E S
Figure 2. Action of the glycosyltransferases GtfA, GtfC, and GtfD involved in the biosynthesis of glycopeptide antibiotics.
Scheme 1. Synthetic Strategy toward TDP â-2-Deoxy Sugars
we changed the location of the amino group to position C-4
(TDP 4-amino-2-deoxy-L-fucose (7) and TDP 4-amino-2-deoxy-
L-rhamnose (8)), exchanged the amino for a hydroxy group
(TDP 2-deoxy L-fucose (9) and TDP 2-deoxy L-rhamnose (10)),
added a 6-hydroxy group (TDP 2-deoxy L-glucose (11)), or
changed the stereochemistry of the C-3 amino group (TDP
L-ristosamine (12)).
Characterization of the Substrate Specificity of GtfA,
GtfC, and GtfD. All three enzymes were tested for their ability
to transfer the various TDP sugars to the vancomycin pseudoa-
glycon 1 (Scheme 2) in order to evaluate their possible use as
tools for the chemoenzymatic synthesis of novel vancomycin
derivatives carrying unnatural 2-deoxy sugars. The results
obtained for these transfers are summarized in Table 1. These
data show that both epi-vancosaminyl transferases GtfA and C
were able to transfer L-acosamine in addition to their natural
substrate, that is, they tolerated the removal of the 3-C-methyl
group (entry 4). Additional modifications, for example, a change
of stereochemistry at C-4, led to a dramatic decrease of transfer
efficiency (entries 1 and 3). Of all other substrates used, only
2-deoxy L-glucose (GtfA and C) and L-vancosamine, L-
daunosamine, and 2-deoxy L-rhamnose (GtfC) were transferred,
albeit in trace quantities.
For the glycosylation of vancomycin pseudoaglycon 1, GtfD
showed the most relaxed substrate specificity. Besides its natural
substrate, L-vancosamine, it also transferred 4-epi-L-van-
cosamine, L-daunosamine, and L-acosamine with good turnover,
that is, both the removal of the C-3-methyl group and the
inversion of stereochemistry at C-4 were tolerated (entries 2-4).
In addition, 2-deoxy L-glucose, 2-deoxy L-rhamnose, and
L-ristosamine were transferred in trace amounts, whereas for
substrates containing an amino group at C-4, no product was
observed.
deprotected (Scheme 1).23 In a shorter approach, the 2-deoxy
glycosyl chlorides can be coupled directly with the tetrabuty-
lammonium salt of TDP.17c Although the stereoselectivities
obtained are lower for the shorter route, it is still possible to
obtain sufficient amounts of the desired â-isomer after HPLC
separation to characterize the enzymes.
To test the substrate specificity of GtfA, C, and D, we
prepared a range of potential TDP 2-deoxy sugar substrates
(Figure 3), following the approaches outlined above (see ref
22 and Supporting Information). The TDP sugars were stable
under all reaction conditions and could be stored at -20 °C
without decomposition for a prolonged period of time (weeks
to months). The set of TDP 2-deoxy sugars included the natural
substrates (TDP L-vancosamine (3) and TDP 4-epi-L-van-
cosamine (4)) and the corresponding 3-desmethyl derivatives
(TDP L-daunosamine (5) and TDP L-acosamine (6)). In addition,
The natural aglycon substrate for GtfC is chloroorienticin B
(2) (see Figure 2),21 which contains the epi-vancosamine
sugar already attached to the benzylic position. To determine
if GtfC shows a broader donor substrate selectivity when its
(23) So far, there have been no reports regarding nucleotidyltransferases that
accept â-configured 2-deoxy-L-glycosyl phosphates as substrates.
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10749

A R T I C L E S
Oberthu¨r et al.
Figure 3. TDP sugar substrates used to study the specificity of GtfA, C, and D.
L-vancosamine (2.3 min-1, entry 2). The overall catalytic
efficiency, however, decreases only about 2-fold for the un-
natural substrate because of substantially lower Km for L-
acosamine (38 versus 218 µM). The only other TDP sugar
accepted by GtfA, 2-deoxy-L-glucose, was transferred in trace
quantities with a kcat < 0.001 min-1 (entry 6).
Scheme 2. Glycosylation of Vancomycin Pseudoaglycon 1 with
GtfA, C, and D
Using the unnatural substrate vancomycin aglycon 1 as the
acceptor substrate, GtfC transferred its natural substrate, 4-epi-
L-vancosamine, with a kcat of 9.7 min-1 (entry 2). Removal of
the 3-C-methyl group in the sugar donor (TDP L-acosamine (6))
led to a 1000-fold decrease of the kcat value (<0.01 min-1, entry
4), and the kcat values for other substrates for which transfer
was observed (TDP sugars 3, 5, 10, and 11) were even lower
(<0.001 min-1, entries 1, 3, 5-7).GtfC shows a dramatically
higher catalytic efficiency for its natural aglycon substrate,
chloroorienticin B (2). This is reflected in the Km value (4 µM)
for this aglycon, which is more than 200-fold lower than that
for vancomycin pseudoaglycon 1 (Km ) 923 µM). GtfC transfers
its natural substrate, 4-epi-L-vancosamine, to aglycon 2 with a
k
cat of 41 min-1 and a Km of 199 µM (entry 2). The removal of
the 3-C-methyl group in the sugar substrate (TDP L-acosamine
(6)) is well tolerated and actually leads to an increase of catalytic
efficiency because the 3-fold reduction of kcat is compensated
by the lower Km value for this NDP sugar (entry 4).
Table 1. Results of the Transfers Using TDP Sugars 3-12a
GtfA
GtfC
GtfD
A change of stereochemistry at C-4 (TDP L-vancosamine (3)
and TDP L-daunosamine (5)), on the other hand, leads to a larger
decrease of kcat (ca. 100-fold) and overall catalytic efficiency
(entries 1 and 3). In addition, other TDP substrates, such as
TDP 2-deoxy-L-rhamnose (10), TDP 2-deoxy-L-glucose (11),
and TDP L-ristosamine (12), were also transferred with a
comparable decrease of the turnover rate (entries 5-7).
In contrast to GtfC, the vancosaminyl transferase GtfD
transfers the TDP sugars with differing stereochemistry at C-4,
TDP L-vancosamine (3) and TDP 4-epi-L-vancosamine (4), with
almost the same kcat (128 and 135 min-1, respectively, entries
1 and 2) and with a 6-fold increase of Km for the unnatural
substrate 4 (38 versus 232 µM). For this enzyme, however, the
removal of the C-3 methyl group has a more pronounced effect
on kcat (500- and 50-fold decrease for L-daunosamine and
L-acosamine, respectively, entries 3 and 4), whereas almost the
same Km values are observed. The transfers that were detected
by HPLC for the TDP sugars 10-12 occurred with a kcat <0.001
min-1 (entries 5-7).
entry
TDP sugar
(aglycon 1)
(aglycon 1)
(aglycon 2)
(aglycon 1)
1
2
3
4
-
+
trace
+
+
+
+
+
-
-
-
+
+
+
+
+
3
5
-
trace
+
+
4
6
+
+
5
7
-
-
-
6
8
-
-
-
7
9
-
-
-
8
10
11
12
-
trace
trace
-
trace
trace
trace
9
trace
-
10
a
(+): transfer observed by HPLC; (trace): transfer observed by HPLC,
but with kcat < 0.01 min-1; (-): no transfer observed by HPLC.
natural aglycon is used, we prepared quantities of chloro-
orienticin B (2) by a large-scale transfer of 4-epi-L-vancosamine
to aglycon 1 using GtfA. The transfers to chloroorienticin B
(2) revealed that, in this case, GtfC shows a substrate specificity
similar to that of GtfD (Table 1). The enzyme accepted a range
of structural changes to its natural substrate, 4-epi-L-van-
cosamine, for example, removal of the C-3-Me, the presence
of an axial 3-amino group, and inversion at C-4 and hydroxyl
groups at C-3 and C-6.
Discussion
Kinetic Characterization of the Successful Transfers. To
compare the efficiency of the different enzymes to transfer both
their natural and non-natural TDP substrates, we undertook a
kinetic characterization of the robust transfers.
In recent years, there has been an increasing interest in the
development of novel glycopeptide antibiotics due to the
emergence of resistance to vancomycin. Research showed that
changes in the carbohydrate portion of lipid-containing vanco-
mycin derivatives can have a significant effect on activity against
both glycopeptide sensitive and resistant bacteria,4,8 which makes
GtfA transfers L-acosamine (entry 4) with a kcat of 0.18 min-1
,
about 10-fold slower than that of the natural substrate, 4-epi-
9
10750 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005

Synthetic Utility of Glycopeptide Glycosyltransferases
A R T I C L E S
Scheme 3. Synthesis of Chloroorienticin B (2) and Glycosylation with GtfC
Table 2. Kinetic Data for the Successful Transfersa
GtfA
GtfC
GtfC
GtfD
entry
TDP sugar
(aglycon 1)
(aglycon 1)
(aglycon 2)
(aglycon 1)
1
3
-
kcat < 0.001
kcat ) 0.6
Km ) 31
cat/Km ) 0.02
kcat ) 41
kcat ) 128
Km ) 38
kcat/Km ) 3.4
kcat ) 135
Km ) 232
kcat/Km ) 0.58
kcat ) 0.28
Km ) 43
kcat/Km ) 0.007
kcat ) 2.4
Km ) 36
k
2
3
4
5
6
7
4
5
kcat ) 2.3
Km) 218
kcat ) 9.7
kcat < 0.001
kcat < 0.01
kcat < 0.001
kcat < 0.001
-
Km ) 199
k
-
cat/Km ) 0.01
k
cat/Km ) 0.21
kcat ) 0.24
Km ) 88
k
cat/Km ) 3 × 10-3
6
kcat ) 0.18
Km ) 38
kcat ) 14
Km ) 15
k
-
cat/Km ) 5 × 10-3
k
cat/Km ) 0.9
kcat/Km ) 0.07
10
11
12
kcat ) 0.5
kcat < 0.001
kcat < 0.001
kcat < 0.001
Km ) 34
k
cat/ Km ) 0.01
kcat < 0.001
kcat ) 0.04
Km ) 90
k
cat/Km ) 4 × 10-4
-
kcat ) 0.25
Km ) 100
k
cat/Km ) 2 × 10-3
a
The kinetic data for the transfers of the natural substrate pairs are shown in bold. kcat [min-1], Km [µM], kcat/Km [µM-1 min-1]. (-): no transfer (see
Table 1).
an efficient synthetic access to such derivatives crucial for the
development of analogues with even better activity.
cin (vancomycin carrying 4-epi-L-vancosamine instead of L-
vancosamine), with catalytic turnovers (kcat < 0.05 min-1) that
are too slow to be synthetically useful.21 Our results from the
present study demonstrate that GtfC also shows a significant
decrease in its ability to transfer 4-epi-L-vancosamine to the
unnatural acceptor, vancomycin pseudoaglycon 1, compared to
the natural substrate, aglycon 2. More importantly, GtfC was
also considerably less promiscuous toward unnatural sugar
substrates in the transfers to aglycon 1, which limits the use of
GtfC as a tool for the chemoenzymatic synthesis of glycopeptide
analogues.
GtfD, on the other hand, is more promising in this context
because of its ability to transfer 2-deoxy sugars, such as
L-vancosamine, 4-epi-L-vancosamine, and L-daunosamine, to
vancomycin pseudoaglycons with modifications in the glucose
portion15d and to different teicoplanin aglycons.15e,16,24 The latter
observation is especially significant because the chemical
glycosylation approach is so far limited to the vancomycin
system.
A chemoenzymatic approach could enable the rapid genera-
tion of a large number of novel glycopeptides that contain
changes in the carbohydrate portion, and the glycosyltransferase
GtfE has shown promise in this regard.15c-e To determine
whether promiscuity is a general quality of glycosyltransferases
involved in the biosynthesis of glycopeptide antibiotics, we
undertook the first systematic evaluation of the synthetic utility
of three glycopeptide glycosyltransferases, GtfA, GtfC, and
GtfD, which transfer similar 2-deoxy-L-sugars (4-epi-L-van-
cosamine or L-vancosamine, respectively, see Figure 2).
Our results for the transfers of a variety of TDP substrate
analogues22 to the natural aglycon acceptors show that the three
enzymes differ markedly with respect to their promiscuity
toward these substrates. Whereas GtfA accepted only one other
TDP sugar besides its natural substrate, both GtfC and GtfD
are comparably promiscuous and were able to transfer a number
of 2-deoxy sugar analogues of the natural substrates L-epi-
vancosamine and L-vancosamine, respectively.
Finally, it is interesting to note that the differences in substrate
specificities described above for the three enzymes correlate
with their turnover rates for their natural substrate pairs, that
is, the faster enzymes, GtfC and GtfD, also proved to be the
most promiscuous ones. This result is in agreement with our
For GtfA and GtfC, the substrate specificity can change
substantially when unnatural acceptor substrates were glyco-
sylated, which reduces the synthetic utility of these enzymes.
We have reported earlier that GtfA transfers even its natural
sugar substrate 4-epi-L-vancosamine to two alternate substrates,
the fully deglycosylated vancomycin aglycon and epivancomy-
(24) Kruger, R.; Lu, W.; Oberthu¨r, M.; Tao, J.; Kahne, D.; Walsh, C. T. Chem.
Biol. 2005, 12, 121-130.
9
J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005 10751

A R T I C L E S
Oberthu¨r et al.
previously described.22 The synthesis of TDP sugars 7, 8, and 10-12
is described in the Supporting Information.
earlier results regarding the promiscuity of the two glucosyl-
transferases involved in the biosynthesis of chloroeremomycin
and vancomycin, in which case the faster GtfE showed also
the more relaxed substrate specificity compared to that of the
slower GtfB.15d,e If this holds true for other glycosyltransferases
involved in the biosynthesis of bacterial secondary metabolites,
this result might aid the search for other synthetically useful
enzymes because, then, the turnover rate for the natural donor-
acceptor pair could serve as an indicator for enzyme promiscuity.
Chemoenzymatic Synthesis of Chloroorienticin B (2). Vancomycin
pseudoaglycon 1, obtained from vancomycin by acid degradation
(trifluoroacetic acid/H2O 9:1, room temperature, 8 h) and HPLC
purification, was enzymatically glycosylated with TDP 4-epi-L-van-
cosamine (4) using the chloroeremomycin biosynthetic glycosyltrans-
ferase GtfA to yield the doubly glycosylated chloroorienticin B (2).21
The reaction conditions were as follows: 75 mM Tris (pH ) 7.0), 8
mM MgCl2, 2.5 mM TCEP, 1 mg/mL BSA, 10% (v/v) DMSO, 500
µM TDP 4-epi-L-vancosamine (4), 500 µM vancomycin pseudoaglycon
1, 5 µM GtfA. The reaction was monitored by HPLC, and additional
aliquots of GtfA were added twice daily until the reaction proceeded
to 50% completion (up to 3 days). Glycopeptide 2 was purified by
preparative HPLC (Vydac C18 column, 0-30% acetonitrile, 0.1% TFA
in 25 min) and was verified by LCMS.
Kinetic Analysis of GtfA, GtfC, and GtfD. The glycosyltrans-
ferases GtfA, GtfC, and GtfD were expressed and purified as previously
described.15e,21 The enzymes were tested for activity against TDP sugars
3-12 using a previously reported HPLC-based assay.15e In GtfA and
GtfD assays, vancomycin pseudoaglycon 1 was used as the glycopeptide
acceptor scaffold, while chloroorienticin B (2) was used in GtfC assays
as we have previously shown that chloroorienticin B is the preferred
substrate for GtfC.21 Kinetic parameters for GtfA using sugar TDP
4-epi-L-vancosamine (4) have been previously reported,21 and the same
methodology was used to generate kinetic parameters for sugars that
demonstrated suitable activity.
Conclusion
This first systematic study of a subgroup of glycosyltrans-
ferases that transfer 2-deoxy-L-sugars to natural product aglycons
shows that these enzymes can differ rather dramatically in terms
of their substrate specificity. The rather high promiscuity of
enzymes studied earlier, for example, GtfE, does not seem to
be a general quality of glycosyltransferases per se. It follows
that to further expand the repertoire of synthetically useful
glycosyltransferases, a detailed study of other enzymes will be
necessary. The set of TDP 2-deoxy-L-sugar analogues obtained
through chemical synthesis, however, will facilitate the screening
of such glycosyltransferases.
For the generation of vancomycin analogues carrying un-
natural sugars, GtfD is the most promising enzyme and can be
used to attach 2-deoxy sugars that are structurally related to
the natural sugar substrate L-vancosamine, for example, 4-epi-
L-vancosamine, L-daunosamine, and L-acosamine. We are cur-
rently using both GtfD and GtfE as reagents in the synthesis of
lipidated glycopeptide analogues to study the influence of
modifications in the carbohydrate portion on biological activity.
Acknowledgment. This work was supported by NIH Grant
66174 (to D.K.) and NIH Grant GM 49338 (to C.T.W.).
Supporting Information Available: Full author list for refs
3a and 13e; experimental procedures and compound character-
ization for the synthesis TDP sugars 7, 8, and 10-12. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Experimental Section
Synthesis of TDP Sugars 3-12. TDP â-L-vancosamine (3), TDP
4-epi-â-L-vancosamine (4), TDP â-L-daunosamine (5), TDP â-L-
acosamine (6), and TDP 2-deoxy-â-L-fucose (9) were synthesized as
JA052945S
9
10752 J. AM. CHEM. SOC. VOL. 127, NO. 30, 2005