C2′-OH of amphotericin B plays an important role in binding the primary sterol of human cells but not yeast cells

Article abstract of DOI:10.1021/ja403255s

Amphotericin B (AmB) is a clinically vital antimycotic but is limited by its severe toxicity. Binding ergosterol, independent of channel formation, is the primary mechanism by which AmB kills yeast, and binding cholesterol may primarily account for toxicity to human cells. The leading structural model predicts that the C2′ hydroxyl group on the mycosamine appendage is critical for binding both sterols. To test this, the C2′-OH was synthetically deleted, and the sterol binding capacity of the resulting derivative, C2′deOAmB, was directly characterized via isothermal titration calorimetry. Surprisingly, C2′deOAmB binds ergosterol and, within the limits of detection of this experiment, does not bind cholesterol. Moreover, C2′deOAmB is nearly equipotent to AmB against yeast but, within the limits of detection of our assays, is nontoxic to human cells in vitro. Thus, the leading structural model for AmB/sterol binding interactions is incorrect, and C2′deOAmB is an exceptionally promising new antifungal agent.

Full text of DOI:10.1021/ja403255s

Communication
pubs.acs.org/JACS
C2′-OH of Amphotericin B Plays an Important Role in Binding the
Primary Sterol of Human Cells but Not Yeast Cells
Brandon C. Wilcock, Matthew M. Endo, Brice E. Uno, and Martin D. Burke*
Howard Hughes Medical Institute, Roger Adams Laboratory, Department of Chemistry, University of Illinois at Urbana−Champaign,
Urbana, Illinois 61801, United States
S
* Supporting Information
ABSTRACT: Amphotericin B (AmB) is a clinically vital
antimycotic but is limited by its severe toxicity. Binding
ergosterol, independent of channel formation, is the
primary mechanism by which AmB kills yeast, and binding
cholesterol may primarily account for toxicity to human
cells. The leading structural model predicts that the C2′
hydroxyl group on the mycosamine appendage is critical
for binding both sterols. To test this, the C2′-OH was
synthetically deleted, and the sterol binding capacity of the
resulting derivative, C2′deOAmB, was directly character-
ized via isothermal titration calorimetry. Surprisingly,
C2′deOAmB binds ergosterol and, within the limits of
detection of this experiment, does not bind cholesterol.
Moreover, C2′deOAmB is nearly equipotent to AmB
against yeast but, within the limits of detection of our
assays, is nontoxic to human cells in vitro. Thus, the
leading structural model for AmB/sterol binding inter-
actions is incorrect, and C2′deOAmB is an exceptionally
promising new antifungal agent.
Figure 1. C2′-OH of AmB is predicted to play a critical role in binding
both ergosterol and cholesterol. Structures of the synthetic derivatives of
AmB designed to test this model.
formation is not required.8 This suggests that binding cholesterol
may account primarily for the toxicity of AmB to human cells and
that efforts to improve the therapeutic index of this clinically vital
antimycotic can focus directly on the simpler problem of
maximizing the relative binding affinity for ergosterol vs
cholesterol.
he polyene macrolide natural product amphotericin B
T(AmB) is the archetype for both small molecules that form
ion channels1 and antibiotics that are inherently refractory to
microbial resistance.2 AmB is also, unfortunately, highly toxic,3
which often limits its effective utilization as the last line of defense
against life-threatening systemic fungal infections. As a result, the
mortality rate for these infections remains near 50%.2a,3b,c
Moreover, the incidence of such fungal infections and resistance
to all other classes of antifungals are rising.2 For all of these
reasons, finding a way to improve the therapeutic index of AmB is
a critically important problem. Some progress has been made
with liposomal formulations,4,5 but these are often prohibitively
expensive4 and substantial toxicity remains.5 Despite more than
four decades of extensive efforts worldwide,6 a clinically viable
derivative of AmB with an improved therapeutic index has yet to
emerge.
A major contributor to this lack of progress has been poor
understanding of the mechanism(s) by which AmB impacts yeast
and human cells. It has for half a century been widely accepted
that AmB kills both types of cells primarily via ion channel-
mediated membrane permeabilization.2,6,7 Guided by this model,
extensive efforts have focused on the challenging problem of
selectively forming ion channels in yeast vs human cells.6,7
In contrast to this classic model, we recently discovered that
AmB kills yeast primarily by simply binding ergosterol; channel
In this vein, we have previously found that deletion of the
mycosamine appendage from AmB eliminates its capacity to bind
both ergosterol and cholesterol.8 The resulting derivative,
amphoteronolide B (AmdeB), was also found to be nontoxic
to yeast.8,9 The roles played by each heteroatom contained in the
mycosamine appendage, however, have remained unclear.
In the leading structural model, AmB interacts with both
ergosterol and cholesterol via a similar binding mode in which
the C2′ hydroxyl group of AmB forms a critical H-bond to the 3β
hydroxyl group on each sterol (Figure 1).10 Experiments
designed to probe this hypothesis, however, have yielded
conflicting results. Studies comparing the membrane permeabi-
lizing activities of conformationally restricted derivatives of AmB
concluded that such a H-bond plays a key role with both
sterols,10b whereas recent computations suggested that this H-
bond is not involved in binding cholesterol.11 The results on a
series of C41 methyl ester derivatives of AmB further modified at
C2′ were mixed: epimerization at C2′ led to retention of both
membrane permeabilizing and antifungal activities, whereas
epimerization and methyl etherification at C2′ resulted in
substantial reductions in both activities.10e Most importantly,
none of these prior studies directly measured sterol binding.
Received: April 4, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja403255s | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Scheme 1
Scheme 2
In pursuit of a definitive experiment, we aimed to delete the
C2′-OH from AmB and directly determine the impact on
binding ergosterol and cholesterol. Synthesis of the targeted
C2′deoxyAmB (C2′deOAmB), however, represented a major
challenge. This is because, in addition to all of the other problems
associated with chemically manipulating this complex and
sensitive natural product,8,9 2-deoxy sugars are more acid-
sensitive than their oxygenated counterparts.12
exploration, it was determined that the TBS-protected derivative
of this 2,3-epoxy alcohol can be regioselectively opened at C2′
using lithium triethylborohydride in THF at 60 °C to give
intermediate 4. The resulting alcohol was mesylated and
displaced by sodium azide, and subsequent removal of the
PMB group generated the deoxysugar donor 5. Importantly, 5 is
protected in such a way that the functional groups at C3′ and C4′
are inert to all of the subsequently required transformations yet
readily unmasked at the end of the synthesis using mild
conditions.16
We also prepared a similarly protected macrolide acceptor
7,9,10e,14 having suitably stable yet readily cleavable silyl ethers
protecting all of the hydroxyl groups and the carboxylic acid at
C41. Glycosylation of 7 with 5 proceeded smoothly to yield
C2′deOAmB derivative 8 as a 2:1 mixture of α and β anomers. 8
proved to be much more amenable to deprotection than 2.
Concomitant cleavage of all nine of the silyl protecting groups
was achieved with HF/pyridine, and the resulting α and β
anomers were readily separated by HPLC. Finally, deprotections
of the C3′ azide with trimethylphosphine and the C13 hemiketal
with aqueous acid completed the synthesis of C2′deOAmB (94%
pure as judged by analytical HPLC).
With several milligrams of this key probe in hand, we tested
whether deletion of the C2′-OH impacts the capacity of AmB to
bind ergosterol via an optimized isothermal titration calorimetry
(ITC)-based assay (Figure 2A). We first titrated an aqueous
solution of AmB with a suspension of large unilamellar vesicles
(LUVs) comprised of only 1-palmitoyl-2-oleoyl-sn-glycero-3-
phosphocholine (POPC), and the net exotherm was recorded.
We repeated the titration using POPC LUVs containing 10%
ergosterol. Consistent with our previous results,8 we observed a
significant increase in net exotherm when switching to
ergosterol-containing LUVs, indicating a direct AmB-sterol
binding interaction. No such binding was observed when the
same pair of titrations was repeated with AmdeB (Figure 2A).8
Surprisingly, when C2′deOAmB was subjected to the same
experiments, a significant increase in net exotherm was observed,
We pursued two different synthetic strategies toward this
probe (Schemes 1 and 2). First we targeted site-selective
deoxygenation of the decahydroxylated natural product (Scheme
1). This led to the discovery that site-selective and site-divergent
functionalizations can be achieved simply by modifying the
electronic properties of achiral reagents.13 Harnessing this
phenomenon, we achieved site-selective acylation of the C2′-
OH to generate intermediate 1, and subsequent persilylation,
deacylation, and deoxygenation of the C2′-OH generated
protected C2′deOAmB 2.13 Deprotection of this intermediate
was initiated by global desilylation with HF/pyridine. We then
employed potassium hydroxide to deprotect the methyl ester and
camphorsulfonic acid to concomitantly remove the p-methox-
ybenzylidene acetals and methyl ketal. Consistent with the
sensitivity of 2-deoxysugars, these studies revealed that
C2′deOAmB derivatives are substantially less compatible with
many chemical reagents than their mycosaminoylated counter-
parts, and this lack of compatibility manifested in low yields for
these transformations. This problem was particularly evident in
the final step. Specifically, removal of the phenylacyl group from
the C3′ amine using penicillin G amidase, a reaction that was
previously successful with both AmB and C35deOAmB,8a
resulted in a low yield of C2′deOAmB as an inseparable mixture
containing a variety of deglycosylated byproducts.
Importantly, extensive knowledge gained during these studies
bolstered an alternative semisynthetic approach10e,14 that
ultimately proved to be much more productive (Scheme 2).
We first generated a C2′-deoxygenated mycosamine (acos-
amine) donor from known intermediate 3.15 After much
B
dx.doi.org/10.1021/ja403255s | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Figure 2. ITC assay to probe the binding of AmB, AmdeB, and
C2′deOAmB to (A) ergosterol and (B) cholesterol. Values represent
the mean of at least three experiments SD. *, p < 0.05; NS, not
significant.
Figure 3. (A) Minimum inhibitory concentrations (MICs) against
ergosterol-containing fungal cells and minimum hemolytic concen-
trations (MHCs) and minimum toxic concentrations (MTCs) against
cholesterol-containing human cells. (B) Microscopy images of primary
human renal epithelial cells treated with AmB or its derivatives.
demonstrating a retained capacity for this derivative to bind
ergosterol. Thus, contrary to the leading model, the C2′-OH on
AmB is not required for ergosterol binding.
Even more surprisingly, when we repeated these same binding
studies with cholesterol, we observed a different result.
Specifically, after confirming binding and no binding of
cholesterol for AmB and AmdeB, respectively,8 we tested the
cholesterol binding capacity of C2′deOAmB (Figure 2B). In
contrast to the results with ergosterol, C2′deOAmB showed no
evidence of binding cholesterol in this experiment. Thus, the
C2′-OH of AmB plays a major role in binding cholesterol but not
ergosterol.
We next tested the minimum inhibitory concentrations
(MICs) of AmB, AmdeB, and C2′deOAmB against two
ergosterol-containing strains of yeast, S. cerevisiae and C. albicans.
The latter represents the most common cause of fungal
infections in humans. Consistent with our previous results8
and the ergosterol binding data described above, we observed
potent antifungal activity for AmB against both cell lines and no
antifungal activity for AmdeB. Importantly, and consistent with
the observation of retained ergosterol binding, when we tested
C2′deOAmB in these same assays, we observed retention of
potent antifungal activity against both strains of yeast (Figure
3A).
Finally, we probed the activity of these same three compounds
against human cells. Two of the most important toxic side effects
associated with AmB are anemia and nephrotoxicity caused by
damage to red blood cells and renal proximal tubule cells,
respectively.5a,6a,7h,17 AmB causes 90% hemolysis of human red
blood cells at 8.5 μM (Figure 3A).18 This is defined as the
minimum hemolytic concentration (MHC). In stark contrast, we
found the corresponding MHCs for AmdeB and C2′deOAmB,
both of which do not bind cholesterol in our ITC assay, to be
>500 μM. Similarly, AmB causes 90% loss of cell viability of
primary human renal proximal tubule epithelial cells19 at 2.4 μM
(the minimum toxic concentration). Again, in stark contrast to
AmB, both AmdeB and, most importantly, C2′deOAmB showed
no evidence of toxicity up to their limits of solubility.20 As shown
in Figure 3B, microscopy further revealed that human primary
renal cells treated with just 2 μM AmB show severe abnormalities
compared to DMSO-treated controls. In contrast, cells treated
with C2′deOAmB show no visual evidence of toxicity, even up to
80 μM.20
These findings demonstrate that the leading structural model
for AmB-sterol binding (Figure 1)10 will need to be revised. Two
alternative models are suggested by our results. In the first, AmB
interacts with these two sterols via distinct binding modes, and
the C2′-OH uniquely participates in a direct binding interaction
with cholesterol. While we cannot yet rule out this possibility, the
structural similarity of the two sterols seems to make this scenario
unlikely. We favor an alternative model in which the C2′-OH
stabilizes a conformer21 of AmB that readily binds both
ergosterol and cholesterol. Deletion of this hydroxyl group, we
propose, favors a shift to a different conformer or set of
conformers which retain the capacity to bind ergosterol but not
cholesterol. Alternatively stated, this model predicts that deletion
of the C2′-OH of AmB causes a small-molecule-based allosteric
effect that results in ligand-selective binding.22 Although further
studies are required to test this hypothesis, we note that in the X-
ray crystal structure of N-iodoacyl AmB23 there is a prominent
water-bridged H-bond between the hydroxyl groups at C2′ and
C13 that may serve to stabilize a particular conformation of the
mycosamine appendage relative to the polyene macrolide core.
To our knowledge, no AmB derivatives with the demonstrated
capacity to directly bind ergosterol but not cholesterol have
previously been reported. We are also unaware of any reported
derivatives with retained antifungal potency but no observable
toxicity to human cells. These features, combined with a
mechanistic model connecting sterol binding to cell killing,8
suggest that C2′deOAmB will have a substantially increased
therapeutic index in vivo. We further note that this derivative was
C
dx.doi.org/10.1021/ja403255s | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Matsushita, N.; Tsuchikawa, H.; Matsumori, N.; Oishi, T. Pure Appl.
Chem. 2009, 81, 1123. (h) Zietse, R.; Zoutendijk, R.; Hoorn, E. J. Nat.
Rev. Nephrol. 2009, 5, 193.
(8) (a) Gray, K. C.; Palacios, D. S.; Dailey, I.; Endo, M. M.; Uno, B. E.;
Wilcock, B. C.; Burke, M. D. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 2234.
(b) Palacios, D. S.; Dailey, I.; Siebert, D. M.; Wilcock, B. C.; Burke, M. D.
Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 6733.
generated via removal of a single atom from AmB. Thus, in
contrast to the methyl ester of AmB and many other derivatives
having modifications at the C41 and/or C3′ positions,6
C2′deOAmB retains the amphoteric nature and many other
potentially important features of the extensively clinically
validated natural product. Combining all of these considerations
with promising starting points for the development of a scalable
synthesis (Schemes 1 and 2),24 C2′deOAmB represents an
exceptional candidate for further development as a potentially
less-toxic clinical substitute for AmB. Preclinical studies to
explore this potential are currently being targeted.
(9) Palacios, D. S.; Anderson, T. M.; Burke, M. D. J. Am. Chem. Soc.
2007, 129, 13804.
(10) (a) Silberstein, A. J. Membr. Biol. 1998, 162, 117. (b) Matsumori,
N.; Sawada, Y.; Murata, M. J. Am. Chem. Soc. 2005, 127, 10667.
(c) Neumann, A.; Czub, J.; Baginski, M. J. Phys. Chem. B 2009, 113,
15875. (d) Czub, J.; Neumann, A.; Borowski, E.; Baginski, M. Biophys.
Chem. 2009, 141, 105. (e) Croatt, M. P.; Carreira, E. M. Org. Lett. 2011,
13, 1390.
ASSOCIATED CONTENT
* Supporting Information
■
S
(11) Neumann, A.; Baginski, M.; Czub, J. J. Am. Chem. Soc. 2010, 132,
18266.
Detailed synthesis, spectral data, and assay procedures. This
material is available free of charge via the Internet at http://pubs.
acs.org.
(12) (a) Borovika, A.; Nargony, P. J. Carbohydr. Chem. 2012, 31, 255.
(b) Noguiera, J. M.; Issa, J. P.; Chu, A.-H. A.; Sisel, J. A.; Schum, R. S.;
Bennett, C. S. Eur. J. Org. Chem. 2012, 2012, 4927. (c) Mergott, D. J.;
Frank, S. A.; Roush, W. R. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 11955.
(d) Oberhur, M.; Leimkuhler, C.; Kahne, D. Org. Lett. 2004, 6, 2873.
(e) Nicolaou, K. C.; Ladduwahetty, T.; Randall, J. L.; Chucholowski, A.
J. Am. Chem. Soc. 1986, 108, 2466.
AUTHOR INFORMATION
Corresponding Author
burke@scs.uiuc.edu
■
(13) Wilcock, B. C.; Uno, B. E.; Bromann, G. L.; Clark, M. J.;
Anderson, T. M.; Burke, M. D. Nat. Chem. 2012, 4, 996.
(14) Nicolaou, K. C.; Daines, R. A.; Ogawa, Y.; Chakraborty, T. K. J.
Am. Chem. Soc. 1988, 110, 4696.
(15) Guo, H.; O’ Doherty, G. A. Angew. Chem., Int. Ed. 2007, 46, 5206.
(16) (a) Nicolaou, K. C.; Daines, R. A.; Chakraborty, T. K. J. Am. Chem.
Soc. 1987, 109, 2208. (b) Nicolaou, K. C.; Chakraborty, T. K.; Ogawa,
Y.; Daines, R. A.; Simpkins, N. S.; Furst, G. T. J. Am. Chem. Soc. 1988,
110, 4660. (c) Nicolaou, K. C.; Daines, R. A.; Chakraborty, T. K.;
Ogawa, Y. J. Am. Chem. Soc. 1988, 110, 4685.
(17) (a) Sundar, S.; Jha, T. K.; Thakur, C. P.; Sinha, P. K.; Bhattacharya,
S. K. N. Engl. J. Med. 2007, 356, 2571. (b) Zager, R. A. Am. J. Kidney Dis.
2000, 36, 238. (c) Sawaya, B. P.; Briggs, J. P.; Schnermann, J. J. Am. Soc.
Nephrol. 1995, 6, 154.
(18) Kinsky, S. C.; Avruch, J.; Permutt, M.; Rogers, H. B.; Schonder, A.
A. Biochim. Biophys. Res. Commun. 1962, 9, 503.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We acknowledge Bristol-Myers Squibb for a gift of AmB and the
NIH (GM080436) for financial support. M.D.B. is an HHMI
Early Career Scientist.
REFERENCES
■
(1) Ermishkin, L. N.; Kasumov, K. M.; Potzeluyev, V. M. Nature 1976,
262, 698.
(2) (a) Monk, B. C.; Goffeau, A. Science 2008, 321, 367. (b) Cannon, R.
D.; Lamping, E.; Holmes, A. R.; Niimi, K.; Tanabe, K.; Niimi, M.; Monk,
B. C. Microbiology 2007, 153, 3211.
(3) (a) Mora-Duarte, J.; Betts, R.; Rotstein, C.; Colombo, A. L.;
Thompson-Moya, L.; Smietana, J.; Lupinacci, R.; Sable, C.; Kartsonis,
N.; Perfect, J. N. Engl. J. Med. 2002, 347, 2020. (b) Eggimann, P.;
Garbino, J.; Pittet, D. Lancet Infect. Dis. 2003, 3, 722. (c) Martin, C. A. J.
Pharm. Pract. 2005, 18, 9.
(4) (a) Johnson, R. H.; Einstein, H. E. Ann. N.Y. Acad. Sci. 2007, 1111,
434. (b) Schneemann, M.; Imhof, A. N. Engl. J. Med. 2005, 352, 410.
(5) (a) Walsh, T. J.; Teppler, H.; Donowitz, G. R.; Maertens, J. A.;
Baden, L. R.; Dmoszynska, A.; Cornely, O. A.; Bourque, M. R.;
Lupinacci, R. J.; Sable, C. A.; dePauw, B. E. N. Engl. J. Med. 2004, 351,
1391. (b) Walsh, T. J.; Pappas, P.; Winston, D. J.; Lazarus, H. M.;
Petersen, F.; Raffalli, J.; Yanovich, S.; Stiff, P.; Greenberg, R.; Donowitz,
G.; Lee, J. N. Engl. J. Med. 2002, 346, 225.
(6) (a) Keim, G. R.; Poutsiaka, J. W.; Kirpan, J.; Keysser, C. H. Science
1973, 179, 584. (b) Ellis, W. G.; Sorbel, R. A.; Nielsen, S. L. J. Infect. Dis.
1982, 146, 125. (c) Cheron, M.; Cybulska, B.; Mazerski, J.; Grzybowska,
J.; Czerwinski, A.; Borowski, E. Biochem. Pharmacol. 1988, 37, 827.
(d) Driver, M. J.; Greenless, A. R.; Macpherson, D. T. J. Chem. Soc.,
Perkin Trans. 1 1992, 3155. (e) Slisz, M.; Cybulska, B.; Mazerski, J.;
Grzybowska, J.; Borowski, E. J. Antibiot. 2004, 57, 669. (f) Paquet, V.;
Volmer, A. A.; Carreira, E. M. Chem.Eur. J. 2008, 14, 2465.
(g) Volmer, A. A.; Szpilman, A. M.; Carreira, E. M. Nat. Prod. Rep.
2010, 27, 1329.
(19) Tominaga, H.; Ishiyama, M.; Ohseto, F.; Sasamoto, K.;
Hamamoto, T.; Suzuki, K.; Watanabe, M. Anal. Commun. 1999, 36, 47.
(20) Due to the limited solubility of AmB and its derivatives in the
media required for growing primary human renal cells (ATCC, PCS-
400-030), 80 μM is the highest concentration achievable in this
experiment.
(21) Fowler, B. S.; Laemmerhold, K. M.; Miller, S. J. J. Am. Chem. Soc.
2012, 134, 9755.
(22) Examples of ligand-selective allosteric effects in proteins:
(a) Duggan, K. C.; Hermanson, D. J.; Musee, J.; Prusakiewicz, J. J.;
Scheib, J. L.; Carter, B. D.; Banerjee, S.; Oates, J. A.; Marnett, L. J. Nat.
Chem. Biol. 2011, 7, 803. (b) Neant-Fery, M.; Garcia-Ordonez, R. D.;
Logan, T. P.; Selkoe, D. J.; Li, L.; Reinstatler, L.; Leissring, M. A. Proc.
Natl. Acad. Sci. U.S.A. 2008, 105, 9582. (c) Knight, Z. A.; Shokat, K. M.
Chem. Biol. 2005, 12, 621. (d) Koike, K.; Oleschuk, C. J.; Haimeur, A.;
Olsen, S. L.; Deeley, R. G.; Cole, S. P. C. J. Biol. Chem. 2002, 277, 49495.
(e) Changeux, J. P.; Edelstein, S. J. Neuron 1998, 21, 959.
(23) (a) Ganis, P.; Avitabile, G.; Mechlinski, W.; Schaffner, C. P. J. Am.
Chem. Soc. 1971, 93, 4560. (b) Jarzembska, K. N.; Kaminski, D.; Hoser,
A. A.; Malinska, M.; Senczyna, B.; Wozniak, K.; Gagos, M. Cryst. Growth
Des. 2012, 12, 2336.
(24) (a) Gantt, R. W.; Peltier-Pain, P.; Cournoyer, W. J.; Thorson, J. S.
Nat. Chem. Biol. 2011, 7, 685. (b) Hutchinson, E.; Murphy, B.; Dunne,
T.; Breen, C.; Rawlings, B.; Caffrey, P. Chem. Biol. 2010, 17, 174.
(7) (a) Finkelstein, A.; Holz, R. In Lipid Bilayers and Antibiotics;
Eisenman, G., Ed.; Dekker: New York, 1973; p 377. (b) Andreoli, T. E.
Ann. N.Y. Acad. Sci. 1974, 235, 448. (c) de Kruijff, B.; Demel, R. A.
Biochim. Biophys. Acta 1974, 339, 57. (d) Bolard, J. Biochim. Biophys.
Acta 1986, 864, 257. (e) Baginski, M.; Resat, H.; Borowski, E. Biochim.
Biophys. Acta 2002, 1567, 63. (f) Cereghetti, D. M.; Carreira, E. M.
Synthesis 2006, 6, 914. (g) Murata, M.; Kasai, Y.; Umegawa, Y.;
D
dx.doi.org/10.1021/ja403255s | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX