Tuning mechanism-based inactivators of neuraminidases: Mechanistic and structural insights

Article abstract of DOI:10.1002/anie.201309675

3-Fluorosialosyl fluorides are inhibitors of sialidases that function by the formation of a long-lived covalent active-site adduct and have potential as therapeutics if made specific for the pathogen sialidase. Surprisingly, human Neu2 and the Trypanosoma cruzi trans-sialidase are inactivated more rapidly by the reagent with an equatorial fluorine at C3 than by its axial epimer, with reactivation following the same pattern. To explore a possible stereoelectronic basis for this, rate constants for spontaneous hydrolysis of the full series of four 3-fluorosialosyl fluorides were measured, and ground-state energies for each computed. The alpha (equatorial) anomeric fluorides hydrolyze more rapidly than their beta anomers, consistent with their higher ground-state energies. However ground-state energies do not explain the relative spontaneous reactivities of the 3-fluoro-epimers. The three-dimensional structures of the two 3-fluoro-sialosyl enzyme intermediates of human Neu2 were solved, revealing key stabilizing interactions between Arg21 and the equatorial, but not the axial, fluorine. Because of changes in geometry these interactions will increase at the transition state, likely explaining the difference in reaction rates. Understanding reactivity and selectivity: The mechanistic basis for the surprisingly different rates of inactivation and reactivation of human and Trypanosoma cruzi sialidases by the two 3-fluoro epimers of 2,3-difluorosialic acid was probed through spontaneous hydrolysis kinetics, computational analysis, and X-ray crystallography.

Full text of DOI:10.1002/anie.201309675

.
Angewandte
Communications
DOI: 10.1002/anie.201309675
Biochemistry
Tuning Mechanism-Based Inactivators of Neuraminidases: Mechanistic
and Structural Insights**
Sabrina Buchini, FranÅois-Xavier Gallat, Ian R. Greig, Jin-Hyo Kim, Soichi Wakatsuki,
Leonard M. G. Chavas,* and Stephen G. Withers*
Abstract: 3-Fluorosialosyl fluorides are inhibitors of sialidases
that function by the formation of a long-lived covalent active-
site adduct and have potential as therapeutics if made specific
for the pathogen sialidase. Surprisingly, human Neu2 and the
Trypanosoma cruzi trans-sialidase are inactivated more rap-
idly by the reagent with an equatorial fluorine at C3 than by its
axial epimer, with reactivation following the same pattern. To
explore a possible stereoelectronic basis for this, rate constants
for spontaneous hydrolysis of the full series of four 3-
fluorosialosyl fluorides were measured, and ground-state
energies for each computed. The alpha (equatorial) anomeric
fluorides hydrolyze more rapidly than their beta anomers,
consistent with their higher ground-state energies. However
ground-state energies do not explain the relative spontaneous
reactivities of the 3-fluoro-epimers. The three-dimensional
structures of the two 3-fluoro-sialosyl enzyme intermediates of
human Neu2 were solved, revealing key stabilizing interactions
between Arg21 and the equatorial, but not the axial, fluorine.
Because of changes in geometry these interactions will increase
at the transition state, likely explaining the difference in
reaction rates.
acid from epithelial cell surfaces, exposing bacterial receptors,
though the roles of these neuraminidases in the pathogenesis
of infection are not well understood. The influenza virus
likewise carries a surface neuraminidase to liberate itself from
cell surface confinement after budding. Neuraminidase inhib-
itors have therefore attracted attention, not only as antivirals
(Relenza and Tamiflu for treatment of flu) but also as
potential antimicrobials.[3] Humans also produce neuramini-
dases: hNeu1 (intra lysosomal), hNeu2 (cytosolic), hNeu3
(plasma membrane), and hNeu4 (lysosomal or mitochondrial
membrane). The four enzymes all belong to the same
sequence-related family (GH33 in the CAZy classification)
as their bacterial and trypanosomal counterparts, though
distinct from that (GH34) of the viral enzymes. Consequently
any inhibitor design must incorporate specificity-determining
elements to avoid unwanted effects on the host.
Our laboratory has introduced 3-fluorosialosyl fluorides
(difluorosialic acids, DFSAs) as a class of mechanism-based
neuraminidase inhibitors that function by forming transient
covalent adducts, and has had some success in developing
analogs with improved specificity for pathogen neuramini-
dases over their human counterparts, with the recent work on
influenza neuraminidase inhibitors being particularly rele-
vant.[4] Three-dimensional structures of the trapped 3-fluo-
rosialyl enzyme intermediates of the influenza virus enzyme
and of the T. cruzi trans-sialidase (TcTS) have been pub-
lished, but not that of a complex with a human neuramini-
dase.[4,5] To optimize specificity, such structural insights are
needed, as they will provide a better understanding of the
active-site environment of the human enzyme and of the
consequences on reaction rates of changing stereochemistry
at the site of fluorine substitution (C3). Here we report the
first three-dimensional structures of covalent sialyl-enzyme
intermediates formed with a human neuraminidase using
both 3-fluoro (equatorial) and 3-fluoro (axial) sialic acid
reagents. We also explore the consequences of modifying the
stereochemistry at C3 on enzyme inactivation and reactiva-
tion behaviors with two different GH33 sialidases, as well as
underlying effects on the nonenzymatic solvolysis, coupled to
computational analysis.
S
ialic acid is found as the terminal sugar on many
mammalian cell surface glycans, where it plays an important
role in mediating interactions of cells with their environ-
ment.[1] These interactions may be involved in normal cell
functions or may occur when a cell is attacked by pathogens
bearing sialic acid-binding proteins on their surfaces. Some
parasites such as Trypanosoma cruzi are incapable of
biosynthesizing sialic acid, and require the expression of
a cell surface trans-sialidase to scavenge sialic acids from host
glycans and thereby decorate their own surfaces as part of the
infection process.[2] Bacteria such as Hemophilus influenzae,
Streptococcus pneumoniae, and Pseudomonas aeruginosa also
produce neuraminidases that are thought to remove sialic
[*] Dr. S. Buchini, Dr. I. R. Greig, Dr. J.-H. Kim, Prof. Dr. S. G. Withers
Department of Chemistry, University of British Columbia
Vancouver, B. C., V6T 1Z1 (Canada)
E-mail: withers@chem.ubc.ca
F.-X. Gallat, Prof. S. Wakatsuki, Prof. Dr. L. M. G. Chavas
Structural Biology Research Center
Photon Factory Institute of Materials Structure Science
High Energy Accelerator Research Organization (KEK)
Tsukuba, Ibaraki 305-0801 (Japan)
Sialidases and trans-sialidases from the GH33 family
employ a two-step double displacement mechanism in which
an active-site tyrosine functions as a nucleophile and reaction
proceeds through oxocarbenium ion-like transition states.[6]
The 3-fluorosialyl fluorides function as mechanism-based
inactivators by forming a long-lived sialyl-enzyme intermedi-
ate (Figure 1). The fluorine at C3 destabilizes the oxocarbe-
nium ion-like transition states, slowing both steps in the
mechanism, while the equatorial anomeric fluoride at C2
E-mail: leonard.chavas@kek.jp
[**] We thank the Canadian Institutes of Health Research for financial
support of this work.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201309675.
3382
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 3382 –3386

Angewandte
Chemie
also obtained by fluori-
nation of protected
DANA (2-deoxy-2,3-
didehydro-N-
acetyl-
neuraminic acid) using
xenon difluoride. This
yielded
the
axial
anomeric
fluoride
derivative directly and
the equatorial deriva-
tive by aqueous acid
hydrolysis of the axial
anomeric fluoride fol-
lowed by re-protection
and the treatment de-
scribed above (see
Scheme S2).
Figure 1. Mechanism of neuraminidase inactivation by 2,3-difluorosialosyl fluorides and structures of difluorosialic
acids studied. Slowing of the de-sialylation step (kr) by inductive destabilization of the oxocarbenium-ion-like
transition state leads to accumulation of the covalent intermediate, thus inactivation.
Kinetic analysis of
serves as a good leaving group, making the intermediate
accessible.[7] All inactivation studies of GH33 enzymes to date
with this DFSA class of inhibitor have employed the
derivative with an axial fluorine at C3. However, our recent
studies with the GH34 influenza
the interaction of 2e3e with TcTS revealed rapid inactivation
(Table 1). Indeed measurements at concentrations above
0.5 mm could not be made because inactivation was too fast,
thwarting attempts to observe saturation behavior. Nonethe-
neuraminidase revealed significant
Table 1: Kinetic parameters for inactivation and reactivation of TcTS and human Neu2 by 3ax- and 3eq-
differences in rates of inactivation
and reactivation for sialic acids with
axial versus equatorial fluorines at
C3. Likewise, a single earlier report
had shown that GT80 sialyltransfer-
ases process the corresponding
equatorial CMP-3-fluorosialic acid
derivative faster than the axial
analog,[8] piquing our interest in
how the two C3 epimers are pro-
cessed by human sialidases and by
TcTS, and what underlies this
behavior. We chose hNeu2 as the
representative human sialidase
since it is the only such mammalian
enzyme for which structural infor-
difluorosialic acids.
Inhibitor
TcTS ki/Kd
TcTS kr
Neu2 ki/Kd
Neu2 Ki (app)
[mm][b]
[ꢀ10À3 minÀ1 mmÀ1
]
[ꢀ10À3 minÀ1
]
[ꢀ10À3 minÀ1 mmÀ1
]
8.0Æ0.4
471Æ76
ꢀ0[a]
1440Æ72
ND
28
17.4Æ1.6
ND
ND=Not Detected. [a] Reactivation occurs only in the presence of sugar acceptor (e.g. Lactose).
[b] Determined from analysis of the Dixon plot.
mation is available.[9] Further, in order to better understand
the effects of the two fluorine substitutions on the inherent
reactivity of sialosides we synthesized the full set of DFSAs
(listed in Figure 1), measured rates of nonenzymatic hydrol-
ysis for each, and performed quantum calculations of ground-
state energies.
Synthesis of 3-fluorosialic acid for conversion into its two
anomeric fluorides was achieved by incubation of 3-fluoro-
pyruvic acid with N-acetylmannosamine in the presence of
the E. coli Neu5Ac aldolase, yielding a mixture of the two 3-
fluoro epimers, as described previously.[8,10] After protection
of the acid and alcohols as esters the two compounds were
separated by silica gel chromatography and subjected to
anomeric deprotection with hydrazine acetate. Reaction of
these hemiketals, individually, with DAST (diethylaminosul-
fur trifluoride) yielded mixtures of the two anomeric fluo-
rides, which were separated by column chromatography and
then deprotected (see Scheme S1 in the Supporting Informa-
tion). The 3-fluoro (equatorial) sialic acid derivatives were
less a second-order rate constant of 471 ꢀ 10À3 minÀ1 mmÀ1 was
measured from the slope of the plot of the inactivation rate
versus the 2e3e concentration. This is approximately 60-fold
faster than the inactivation caused by the axial epimer
2e3a.[10a] To assess turnover (reactivation) rates, a sample of
TcTS inactivated by 2e3e was freed of excess inactivator by
gel filtration, incubated in buffer at 258C, and samples
removed and assayed at time intervals. Fitting of the data to
a simple first-order expression revealed a reactivation rate
constant of 1.7 ꢀ 10À2 minÀ1, corresponding to a half-life of
approximately 40 minutes (see Figures S2 and S3). The
formation of 3-fluorosialic acid as shown by mass spectrom-
etry confirmed the hydrolytic turnover nature of the reac-
tivation (Figure S6). By comparison, no significant hydrolytic
reactivation of the enzyme trapped by the axial epimer was
observed.[10] Taken together, these data show that an equa-
torial fluorine at C3 clearly has a less rate-retarding effect on
both steps of the enzymatic reaction than does an axial
fluorine.
Angew. Chem. Int. Ed. 2014, 53, 3382 –3386
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3383

.
Angewandte
Communications
Table 2: pKa values, rate constants for spontaneous hydrolysis and
computed relative ground state energies for difluorosialic acids.
A similar trend, though manifested somewhat differently,
was seen for hNeu2. In this case the axial anomer 2e3a
inactivated the enzyme according to a second-order rate
constant of ki/Kd = 1.44 minÀ1 mmÀ1 while reactivation oc-
curred with khyd = 6 ꢀ 10À4 minÀ1, corresponding to a half-life
of approximately 20 h (see Figure S4). By contrast, no time-
dependent inactivation was observed with the equatorial
epimer 2e3e, which instead functioned as a slow substrate, but
one for which the deglycosylation step (khyd) is slower than the
inactivation step (ki). This kinetic behavior is revealed in the
surprisingly tight apparent binding of 2e3e when tested as
a competitive inhibitor (Kiapp = 28 mm), which arises because
most of the enzyme accumulates as its fluorosialyl-enzyme
species (Figure S5). This is a much higher “affinity” than
would be anticipated based upon its simple structure, thus
suggesting that, as with TcTS, both formation and hydrolysis
of the 3-fluorosialyl-hNeu2 intermediate are much faster with
the equatorial 3-epimer than the axial and that the inter-
mediate is accumulating under assay conditions.
Inhibitor
pKa
khyd
Relative ground-state
energy [kJmolÀ1
[ꢀ10À6 minÀ1
]
]
2.3
100
0
2.4
2.3
2.3
820
1.3
2.8
À1.4
À6.6
À9.0
To gain some insight into the source of this fascinating
difference in reactivity of the 3-fluoro equatorial and axial
derivatives, we investigated the spontaneous hydrolysis of the
full series of the four DFSAs (2e3e, 2e3a, 2a3a, 2a3e;
Figure 1) to establish whether this difference arose from
inherent stereoelectronic effects, or was primarily a conse-
quence of enzymatic interactions. Hydrolysis rates were
determined by incubating samples in 50 mm buffer containing
1m NaClO4 in sealed vials, removing aliquots at time intervals
and freezing them immediately. At the end of the time-course,
samples were thawed and their fluoride concentrations
measured using a fluoride-selective electrode. The pH de-
pendence of hydrolysis was first probed for 2e3a and 2e3e at
a series of pH values, and reaction rates shown to be pH-
independent between pH 5 and 8 (Figure S7), as previously
reported for the hydrolysis of aryl sialosides.[11] To ensure
which ionic forms are present under these conditions, pKa
values for the carboxylic acids of each DFSA were measured
by back-titration of their acid forms using sodium hydroxide
and found to lie between 2.3 and 2.4 (Table 2; Table S1 and
Figure S9). Such values are lower than those of 2.8 and 2.6
measured for the parent sialic acid and its methyl glycoside,
respectively, consistent with the presence of a second fluorine
in the DFSAs. Further, the near identity of these four DFSA
pKa values shows that the different stereochemistries of the
fluorines play no role. The pH independence of the hydrolysis
rates and the near identical pKa values therefore confirm that
the reaction of the deprotonated (monoanionic) substrate
under these conditions occurs by spontaneous heterolysis of
gain deeper insight into these relative ground-state stabilities
we computed the energy differences between the ground
states of the four DFSAs. Indeed, the sialosyl fluorides
possessing axial anomeric fluorides were found to be more
stable than those possessing equatorial anomeric fluorides as
predicted by the anomeric effect, with 2a3a being 2.4 kJmolÀ1
more stable than 2a3e, and 2e3e 1.4 kJmolÀ1 more stable than
2e3a (Table 2; Figure S10). This pattern of stabilities of the 3-
epimeric pairs of compounds is consistent with expected
stabilities arising from the interaction of carbon–heteroatom
bond dipoles between C2 and C3 centers. For example, while
2a3a possesses a stabilizing antiparallel orientation between
À
À
C3 F and C2 F bond dipoles, the antiparallel orientation of
À
À
2a3e only exists between C3 F and C2 O6 bond dipoles, with
À
the C2 O7 bond dipole expected to be smaller. Similarly,
À
À
2e3e shows an antiparallel orientation between C3 Fand C2
O6 bond dipoles, but 2e3a has no antiparallel orientations
between C2-heteroatom and C3-heteroatom dipoles. These
relative ground state stabilities are therefore consistent with
the relative rates observed for compounds with axial and
equatorial leaving groups. The present conclusions also
concur with previously reported results in which mono-
anions of aryl a-sialosides were found to hydrolyse 100-fold
faster than their b counterparts.[11b] In this earlier report, the
difference was largely attributed to the release of steric strain
arising from the axial carboxy group in the ground state of the
a-sialosides, though undoubtedly the effects seen in that case
were moderated by the counteracting release of steric strain
arising from departure of the bulky aryl leaving group. That
counteracting steric effect would not be significant in the
present case, with sterically modest fluoride leaving groups.
The computed relative ground-state stabilities do not,
however, reflect the measured rate differences between C-3
epimers possessing leaving groups of the same geometry
since, for example, 2a3a is found to have a lower energy than
2a3e, yet is hydrolyzed two-times faster (corresponding to
a transition state theory (TST)-predicted activation energy
difference of 2.1 kJmolÀ1; Table 2). Likewise, 2e3e is found to
À
the C F bond. With this established, first-order rate constants
for hydrolysis of the four DFSA substrates (2e3e, 2e3a, 2a3e,
2a3a) were measured at pH 7 (Table 2; Figures S7 and S8).
The first and most striking observation is that the a-sialyl
fluorides (equatorial leaving group) hydrolyze much faster
than the b anomers: 36-fold faster in the 3-fluoro (ax) pair and
631-fold faster in the 3-fluoro (eq) pair. This reactivity
difference is reminiscent of the 40-fold rate difference
reported between a- and b-glucosyl fluoride solvolyses and
attributed to differences in ground-state stabilities that are
minimized at the (highly dissociative) transition state.[12] To
3384
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 3382 –3386

Angewandte
Chemie
have a lower energy than 2e3a, yet is hydrolyzed
eight-times faster (TST activation energy difference
of 5.7 kJmolÀ1). Clearly other factors are more
important in determining transition-state stabilities.
One possibility, consistent with earlier proposals, is
that interactions between electronegative pyranose
ring substituents and the water nucleophile may be
important in determining rates of pyranosyl fluoride
hydrolysis.[12b] In this instance, 3-fluoro substituents
disposed in an anti-fashion to the leaving group
anomeric fluoride may enhance the nucleophilicity of
the incoming water molecules through a hydrogen-
bonding interaction. Similar effects were seen in the
hydrolysis of 2,4-dinitrophenyl b-glycosides and their
2-deoxy-2-fluoro analogs, where substrates with an
axial C-2 substituent (manno- and 2-fluoromanno)
hydrolyzed five times slower than their 2-epimers.[13]
Whatever the sources of these differences, the
effects are relatively small compared to those seen
Figure 2. X-ray structures of the covalent intermediates trapped on Neu2 with
3F-equatorial and 3F-axial substituents. Left panels: structure with 3F-equatorial
sialic acid. Right panels: structure with 3F-axial sialic acid. Insets show
alternative views.
with the enzyme, though indeed in the same sense. To gain Experimental Section
Details of all synthetic steps and product characterization are
deeper insight into the nature of the differences in the
interactions of the compounds with active-site residues, we
determined the three-dimensional structures of hNeu2 trap-
ped in the form of the two 3-fluorosialyl-enzyme intermedi-
ates, both with the axial and equatorial fluorine (Table S2).
The two epimers sit inside the active site of hNeu2 in an
essentially identical manner, covalently bound to the human
enzyme by Tyr334 and stabilized by an extended hydrogen-
bond network involving nine residues and two structured
water molecules (Figure 2). While the axial epimer orients its
fluorine to the solvent, the equatorial fluorine makes contacts
(2.6 and 3.1 Angstroms) with the two terminal nitrogen atoms
(N1 and N2 respectively) of Arg21. This arginine also
interacts closely with the substrate C1 carboxylate. Interac-
tions between fluorine and protonated nitrogen are well-
documented and have been computed at up to 13.5 kcal
molÀ1.[14] Therefore, especially if optimized at the transition
state as they are likely to be due to the conformational
changes occurring, these interactions could readily explain
the difference in rates between the equatorial and axial
derivatives. Interestingly similar interactions are seen in the
influenza neuraminidase between the equatorial fluorine of
the covalently bound DFSA and Arg118 in that case.[4]
While the large differences in rates of spontaneous
hydrolysis of DFSAs of opposite anomeric stereochemistry
are due largely to differences in ground-state stability, the
more subtle differences in rate constants for the 3-epimeric
DFSA pairs arise instead largely from differences in the
transition-state structure. Likewise the very different rate
constants for inactivation and reactivation of TcTS and Neu2
by these 3-epimeric DFSAs arise primarily from differential
transition-state stabilization. Three-dimensional structures of
the two 3-fluorosialyl-enzyme intermediates of human Neu2
suggest that interactions between the equatorial fluorine and
the highly conserved Arg21 could play particularly important
roles in binding of the transition states for the two steps and
are responsible for the large rate differences observed. These
factors should be considered in any future design of inhibitors
of this class for other neuraminidases.
provided in the Supporting Information, as are full details of
enzymatic and spontaneous kinetic analyses, computational methods,
and X-ray structure determination. Optimized geometries and
frequencies were determined at the IEF-PCM/M06-2X/6-31 + G-
(d,p) level of theory. Energies for these optimized geometries were
recalculated at the SMD/M06-2X/6-31 + G(d,p) level of theory. These
methods have previously proved reasonably accurate in the inves-
tigation of the reaction energetics of a series of fluoroxylosides.[12b]
Received: November 6, 2013
Published online: March 3, 2014
Keywords: biochemistry · enzymes · hydrolysis ·
.
neuraminidase inhibition · stereochemistry
[1] A. Varki, Trends Mol. Med. 2008, 14, 351.
[2] A. C. C. Frasch, Parasitol. Today 2000, 16, 282.
[3] a) G.-Y. Chen, X. Chen, S. King, K. A. Cavassani, J. Cheng, X.
Zheng, H. Cao, H. Yu, J. Qu, D. Fang, W. Wu, X.-F. Bai, J.-Q. Liu,
S. A. Woodiga, C. Chen, L. Sun, C. M. Hogaboam, S. L. Kunkel,
P. Zheng, Y. Liu, Nat. Biotechnol. 2011, 29, 428; b) M.
von Itzstein, Nat. Rev. Drug Discovery 2007, 6, 967; c) M.
von Itzstein, W. Y. Wu, G. B. Kok, M. S. Pegg, J. C. Dyason, B.
Jin, T. Van Phan, M. L. Smythe, H. F. White, S. W. Oliver, P. M.
Colman, J. N. Varghese, D. M. Ryan, J. M. Woods, R. C. Bethell,
V. J. Hotham, J. M. Cameron, C. R. Penn, Nature 1993, 363, 418.
[4] J. H. Kim, R. Resende, T. Wennekes, H. M. Chen, N. Bance, S.
Buchini, A. G. Watts, P. Pilling, V. A. Streltsov, M. Petric, R.
Liggins, S. Barrett, J. L. McKimm-Breschkin, M. Niikura, S. G.
Withers, Science 2013, 340, 71; C. J. Vavricka, Y. Liu, H. Kiyota,
N. Sriwilaijaroen, J. Qi, K. Tanaka, Y. Wu, Q. Li, Y. Li, J. Yan, Y.
Suzuki, G. F. Gao, Nat. Commun. 2013, 4, 1491.
[5] M. F. Amaya, A. G. Watts, I. Damager, A. Wehenkel, T. Nguyen,
A. Buschiazzo, G. Paris, A. C. Frasch, S. G. Withers, P. M. Alzari,
Structure 2004, 12, 775.
[6] D. L. Zechel, S. G. Withers, Acc. Chem. Res. 2000, 33, 11.
[7] a) I. Damager, S. Buchini, M. F. Amaya, A. Buschiazzo, P.
Alzari, A. C. Frasch, A. Watts, S. G. Withers, Biochemistry 2008,
47, 3507; b) A. G. Watts, I. Damager, M. L. Amaya, A.
Buschiazzo, P. Alzari, A. C. Frasch, S. G. Withers, J. Am.
Chem. Soc. 2003, 125, 7532.
Angew. Chem. Int. Ed. 2014, 53, 3382 –3386
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3385

.
Angewandte
Communications
[8] H. A. Chokhawala, H. Z. Cao, H. Yu, X. Chen, J. Am. Chem.
Soc. 2007, 129, 10630.
[11] a) M. Ashwell, X. M. Guo, M. L. Sinnott, J. Am. Chem. Soc.
1992, 114, 10158; b) V. Dookhun, A. J. Bennet, J. Am. Chem.
Soc. 2005, 127, 7458.
[12] a) A. Konstantinidis, M. L. Sinnott, Biochem. J. 1991, 279, 587;
b) S. S. Lee, I. R. Greig, D. J. Vocadlo, J. D. McCarter, B. O.
Patrick, S. G. Withers, J. Am. Chem. Soc. 2011, 133, 15826.
[13] M. N. Namchuk, J. D. McCarter, A. Becalski, T. Andrews, S. G.
Withers, J. Am. Chem. Soc. 2000, 122, 1270.
[9] a) L. M. Chavas, R. Kato, N. Suzuki, M. von Itzstein, M. C.
Mann, R. J. Thomson, J. C. Dyason, J. McKimm-Breschkin, P.
Fusi, C. Tringali, B. Venerando, G. Tettamanti, E. Monti, S.
Wakatsuki, J. Med. Chem. 2010, 53, 2998; b) L. M. G. Chavas, C.
Tringali, P. Fusi, B. Venerando, G. Tettamanti, R. Kato, E. Monti,
S. Wakatsuki, J. Biol. Chem. 2005, 280, 469.
[10] a) S. Buchini, A. Buschiazzo, S. G. Withers, Angew. Chem. 2008,
120, 2740; Angew. Chem. Int. Ed. 2008, 47, 2700; b) A. G. Watts,
S. G. Withers, Can. J. Chem. 2004, 82, 1581.
[14] L. Shimoni, J. P. Glusker, C. W. Bock, J. Phys. Chem. 1995, 99,
1194.
3386
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 3382 –3386