Efficient chemoenzymatic synthesis of ADP-D-glycero-β-D-manno-heptose and a mechanistic study of ADP-L-glycero-D-manno-heptose 6-epimerase

Article abstract of DOI:10.1021/ol050774q

(Chemical Equation Presented) A chemoenzymatic synthesis of ADP-D-glycero-β-D-manno-heptose (ADP-D,D-Hep) is described in which D,D-Hep 7-phosphate is converted to ADP-D,D-Hep by two biosynthetic enzymes. This strategy allows access to the 6″-deuterated a

Full text of DOI:10.1021/ol050774q

ORGANIC
LETTERS
2005
Vol. 7, No. 12
2457-2460
Efficient Chemoenzymatic Synthesis of
ADP- -glycero- -manno-Heptose and a
Mechanistic Study of
ADP- -glycero- -manno-Heptose
6-Epimerase
D
â
-D
L
D
Jay A. Read, Raef A. Ahmed, and Martin E. Tanner*
Department of Chemistry, UniVersity of British Columbia, VancouVer, BC, Canada
mtanner@chem.ubc.ca
Received April 10, 2005
ABSTRACT
A chemoenzymatic synthesis of ADP-
D
-glycero-
â
-D
-manno-heptose (ADP-
D,
D-Hep) is described in which
D,
D-Hep 7-phosphate is converted to
ADP-D,D-Hep by two biosynthetic enzymes. This strategy allows access to the 6′′-deuterated analogue, which upon incubation with the epimerase
showed complete retention of the isotopic label at the 6′′-position. This provides evidence for a direct oxidation mechanism in which the
+
hydride initially transferred to the NADP cofactor is subsequently returned to the same carbon in a nonstereospecific manner.
Lipopolysaccharide (LPS) is a large glycolipid that comprises
the outer surface of the outer membrane in gram-negative
bacteria.1,2 It consists of three main parts: a Lipid A
component, which serves as a hydrophobic anchor to the
outer membrane; a core oligosaccharide, which serves as a
barrier to hydrophobic antibiotics, and; an O-antigenic
component, consisting of an immunogenic repeating oli-
gosaccharide distinct to each species. The constituents of the
core oligosaccharide are largely conserved among gram-
negative bacteria and contain the higher order sugars
3-deoxy-D-manno-octulosonic acid (KDO) and L-glycero-D-
manno-heptose (L,D-Hep).
The biosynthesis of the core component L,D-Hep begins
with the conversion of sedoheptulose 7-phosphate into D,D-
Hep 7-phosphate as catalyzed by the isomerase GmhA
(Scheme 1).3,4 The D,D-Hep 7-phosphate is then converted
into D,D-Hep 1,7-bisphosphate by the kinase activity of the
bifunctional enzyme HldE. It has recently been determined
that the unusual â-anomer is utilized in this biosynthetic
scheme. A phosphatase, GmhB, then generates D,D-Hep
1-phosphate, and the second activity of HldE catalyzes a
pyrophosphorylase reaction with ATP to give ADP-D,D-Hep.
The required stereochemistry for LPS biosynthesis is then
introduced by the action of ADP-L-glycero-D-manno-heptose
6-epimerase (HldD) that interconverts ADP-D,D-Hep and
ADP-L,D-Hep by a reversible inversion of stereochemistry
at C-6′′. Finally, a heptosyltransferase incorporates the L,D-
Hep into the LPS structure.
The mechanism employed by ADP-L-glycero-D-manno-
heptose 6-epimerase is of considerable interest in that, unlike
the majority of racemases and epimerases, it performs an
inversion of stereochemistry at an “unactivated” center.5,6
Previous biochemical and structural studies have confirmed
(3) Kneidinger, B.; Marolda, C.; Graninger, M.; Zamyatina, A.; McArthur,
F.; Kosma, P.; Valvano, M. A.; Messner, P. J. Bacteriol. 2002, 184, 363-
369.
(1) Raetz, C. R. H.; Whitfield, C. Annu. ReV. Biochem. 2002, 71, 635-
700.
(2) Raetz, C. R. H. In Escherichia coli and Salmonella; Neidhardt, F.
C., Ed.; American Society for Microbiology: Washington, DC, 1996; pp
1035-1063.
(4) Valvano, M. A.; Messner, P.; Kosma, P. Microbiol. 2002, 148, 1979-
1989.
(5) Samuel, J.; Tanner, M. E. Nat. Prod. Rep. 2002, 19, 261-277.
(6) Tanner, M. E. Acc. Chem. Res. 2002, 35, 237-246.
10.1021/ol050774q CCC: $30.25
© 2005 American Chemical Society
Published on Web 05/17/2005

H218O) is incorporated into product. The demonstration that
7′′-deoxy and 4′′-deoxy analogues serve as substrates for the
epimerase rules out mechanisms involving transient oxidation
at either of these positions. Finally, the demonstration that
the epimerase is capable of catalyzing the dismutation of a
truncated intermediate analogue containing an aldehyde at
C-6′′ into a 1:1 mixture of 6′′-alcohol and 6′′-acid strongly
supports the notion that oxidation/reduction takes place at
the 6′′ position during the epimerase mechanism. A necessary
consequence of the proposed mechanism is that the C-6′′
hydrogen of ADP-D,D-Hep will be transferred to the C-6′′
position of ADP-L,D-Hep. In this paper, ADP-[6′′-2H]-D,D-
Hep is used to test this proposal.
Scheme 1. Biosynthesis of
ADP-L-glycero-â-D-manno-Heptose
Progress in the study of ADP-L,D-Hep 6-epimerase has
been hampered by the difficulty in obtaining the epimeric
substrates. ADP-D,D-Hep can be isolated from mutant strains
of Escherichia coli that lack the epimerase gene; however,
only submilligram amounts can be obtained in this fashion.11
Although the synthesis of L,D-Hep and D,D-Hep has been
achieved by several groups,12 the synthesis of both epimers
of ADP-â-Hep has only recently been reported and involves
lengthy synthetic sequences with several challenging steps
(15 steps for ADP-D,D-Hep).13 Notably, it is necessary to
install the correct stereochemistry at both C-6′′ and C-1′′
(the latter gives the â-manno configuration). Furthermore,
the products are sensitive to both acidic and basic conditions,
with the latter instability attributed to formation of a cyclic
phosphodiester. In our effort to develop a more efficient
synthesis of ADP-D,D-Hep, we targeted the biosynthetic
intermediate D,D-Hep 7-phosphate that could be converted
into ADP-D,D-Hep by the action of the enzymes HldE and
GmhB (Scheme 1). This avoids the need to introduce the
â-phosphate group and to accomplish the diphosphate
coupling reaction. Our synthesis was also planned such that
ADP-[6′′-2H]-D,D-Hep could be readily obtained with the use
of NaB2H4. Our chemoenzymatic synthesis of ADP-D-
glycero-â-D-manno-heptose begins with the primary alcohol
1 (Scheme 3), which is available in four steps from
D-mannose.12b The alcohol was converted to the acid 2 by
dichromate oxidation. Acid 2 was reacted with oxalyl
chloride to produce the acid chloride, which was added to a
solution of diazomethane in ether to afford the R-diazoketone
3.14 Nucleophilic displacement of nitrogen with diben-
zylphosphoric acid gave the R-ketophosphate;15 however, this
compound was relatively unstable and decomposed on silica
gel. Instead of isolating the compound, a reductive workup
using either NaBH4 or NaBD4 was employed to give the
nondeuterated (4a and 5a) and deuterated (4b and 5b)
that it employs a tightly bound NADP+ cofactor during
catalysis.7,8 Recent studies have supported a direct oxidation-
reduction mechanism in which transient oxidation takes place
at C-6′′ to produce a 6′′-keto intermediate (Scheme 2).9,10
Scheme 2. Mechanism of the Reaction Catalyzed by
ADP-L-glycero-D-manno-Heptose 6-Epimerase
Following reorientation of the intermediate within the active
site, the hydride is then transferred to the opposite face of
the carbonyl to yield the epimeric product during catalysis.
Evidence in support of this mechanism includes the observa-
(11) Coleman, W. G., Jr. J. Biol. Chem. 1983, 258, 1985-1990.
(12) (a) Rosenfeld, D. A.; Richtmyer, N. K.; Hudson, C. S. J. Am. Chem.
Soc. 1951, 73, 4907-4910. (b) Dziewiszek, K.; Zamojski, A. Carbohydr.
Res. 1986, 150, 163-171. (c) Brimacombe, J. S.; Kabir, A. K. M. S.
Carbohydr. Res. 1986, 152, 329-334. (d) Crich, D.; Banerjee, A. Org.
Lett. 2005, 7, 1395-1398.
(13) Zamyatina, A.; Gronow, S.; Puchberger, M.; Graziani, A.; Hofinger,
A.; Kosma, P. Carbohydr. Res. 2003, 338, 2571-2589.
(14) Kende, A. S.; Fujii, Y.; Mendoza, J. S. J. Am. Chem. Soc. 1990,
112, 9645-9646.
(15) Bischofberger, N.; Waldmann, H.; Saito, T.; Simon, E. S.; Lees,
W.; Bednarski, M. D.; Whitesides, G. M. J. Org. Chem. 1988, 53, 3457-
3465.
2
tion that no solvent-derived isotope (from either H2O or
(7) Ni, Y.; McPhie, P.; Deacon, A.; Ealick, S.; Coleman, W. G., Jr. J.
Biol. Chem. 2001, 276, 27329-27334.
(8) Deacon, A. M.; Ni, Y. S.; Coleman, W. G., Jr.; Ealick, S. E. Structure
2000, 8, 453-461.
(9) Morrison, J. P.; Read, J. A.; Coleman, W. G., Jr.; Tanner, M. E.
Biochemistry 2005, 44, 5907-5915.
(10) Read, J. A.; Ahmed, R. A.; Morrison, J. P.; Coleman, W. G., Jr.;
Tanner, M. E. J. Am. Chem. Soc. 2004, 126, 8878-8879.
2458
Org. Lett., Vol. 7, No. 12, 2005

Scheme 3. Synthesis of ADP-D,D-Hep and ADP-[6′′-2H]-D,D-Hep
dibenzyl phosphates, respectively, each as an equimolar
labeled compound was incubated with ADP-L,D-Hep 6-epi-
merase, and the reaction was monitored by 1H NMR
spectroscopy (Figure 1). A convenient signal that can be used
to monitor the epimerization reaction is that from H5′′, which
appears at 3.32 ppm in ADP-D,D-Hep and 3.23 ppm in ADP-
L,D-Hep. In ADP-[6′′-2H]-D,D-Hep, the H5′′ signal appears
mixture of diastereomers. An attempt to control the diaste-
reoselectivity of the reduction employed Zn(BH4)2 as the
reducing agent.16 This gave predominantly one diastereomer;
however, further studies confirmed that it was the L,D-isomer
that would not be carried forward by the biosynthetic
enzymes. A global deprotection of the mixture of 4 and 5
using Pearlman’s catalyst under 50 psi of H2 produced a
mixture of D,D-Hep 7-phosphate and L,D-Hep 7-phosphate,
which were carried forth without further purification. To
accomplish the final enzymatic step of the synthesis, the gene
products HldE and GmhB were overexpressed and purified.
The enzymes were generated with N-terminal hexahistidine
tags to facilitate purification by immobilized metal ion
affinity chromatography. The mixture of D,D-Hep 7-phos-
phate and L,D-Hep 7-phosphate was incubated with HldE,
GmhB, and 2 equiv of ATP. Inorganic pyrophosphatase was
also added in order to hydrolyze the pyrophosphate produced
and drive the reaction to completion. The reaction was
monitored by 31P NMR spectroscopy, and it was clear that
only one of the isomers was converted into a sugar nucleotide
upon overnight incubation. The resulting sugar nucleotide
was readily separated from the other reaction components
by anion-exchange chromatography and found to be identical
with ADP-D,D-Hep that had been generated via chemical
synthesis.10 Thus, the enzymes would readily accept the
natural D,D-Hep 7-phosphate as a substrate and would not
accept the unnatural L,D-isomer, effectively separating the
mixture that had been generated by the borohydride reduc-
tion.
3
as a doublet with a J value of 9.5 Hz due to its coupling
With ADP-[6′′-2H]-D,D-Hep in hand, it was possible to
unequivocally show that the hydride that is transferred to
the NADP+ during the initial oxidation is subsequently
transferred back to the same carbon of the product. The
Figure 1. 1H NMR spectra illustrating the enzyme-catalyzed
epimerization of ADP-[6′′-2H]-D-glycero-D-manno-heptose.
(16) Jarosz, S.; Skora, S.; Stefanowicz, A.; Mach, M.; Frelek, J. J.
Carbohydr. Chem. 1999, 18, 961-974.
Org. Lett., Vol. 7, No. 12, 2005
2459

with H4′′. Upon incubation with the epimerase, this signal
was diminished, and a new doublet appeared corresponding
to the H5′′ signal of the other epimer. The two doublets
reached equilibrium after 1 h, and their signals remained
unchanged after incubation for an additional 16 h. The fact
that the ADP-[6′′-2H]-L,D-Hep produced also displays an H5′′
signal that is a doublet confirms that the label was retained
during catalysis. Its location can be assigned to C-6′′ on the
carbonyl to interconvert ketone and aldehyde intermediates,
and a final reduction at a different carbon (C-7′′ or C-6′′
respectively). Such mechanisms could be invoked in order
to avoid proposing a nonstereospecific hydride transfer that
requires major reorientation of the substrate and cofactor;
however, they are somewhat more complex in terms of the
number of chemical steps. Instead, ADP-L,D-Hep 6-epimerase
appears to employ the direct two-step oxidation/reduction
mechanism that is reminiscent of the approaches taken by
the enzymes UDP-galactose 4-epimerase and CDP-tyvelose
2-epimerase.17,18 It is intriguing to see that this strategy was
adopted by three epimerases that are all evolutionarily related
yet operate at positions of a sugar nucleotide that are spacially
quite distinct (C-2′′, C-4′′, and C-6′′). In each case, the
chemical strategy is the same; however, the required
reorientation of the oxidized intermediate is dramatically
different.
3
basis of the observed large J4,5 value of 9.5 Hz and the
knowledge that the 3J5,6 value is small (see the spectrum of
the nondeuterated sample). As expected, the mass spectrum
of the epimeric mixture was identical to that of the starting
material.
The synthesis outlined in this letter provides a convenient
source of deuterated and nondeuterated ADP-D,D-Hep and
will greatly aid further mechanistic studies, including the
measurement of kinetic isotope effects and isotope crossover
experiments. In the latter studies, ADP-[6′′-18O]-D,D-Hep
could readily be prepared using a minor modification of the
chemoenzymatic synthesis and epimerized in the presence
of ADP-[6′′-2H]-D,D-Hep to test for the formation of com-
pounds containing both isotopic labels. This would test
whether the 6′′-keto intermediate is transferred between
active sites during catalysis.
The observation that the C-6′′ hydrogen retains its position
during catalysis is completely consistent with the proposed
nonstereospecific hydride transfer mechanism (Scheme 2).
It rules out mechanisms involving an initial oxidation at one
carbon (C-6′′ or C-7′′), followed by an isomerization of the
Acknowledgment. This work was supported by the
Canadian Institutes of Health Research (CIHR)
Supporting Information Available: Experimental pro-
cedures and NMR spectra for compounds 2, 3, and ADP-
[6′′-2H]-D,D-Hep. This material is available free of charge
via the Internet at http://pubs.acs.org.
OL050774Q
(17) Frey, P. A. FASEB J. 1996, 10, 461-470.
(18) Hallis, T. M.; Zhao, Z.; Liu, H.-W. J. Am. Chem. Soc. 2000, 122,
10493-10503.
2460
Org. Lett., Vol. 7, No. 12, 2005