Modification of bafilomycin structure to efficiently synthesize solid-state NMR probes that selectively bind to vacuolar-type ATPase

Article abstract of DOI:10.1002/asia.201403299

Bafilomycin (Baf) is one of the most potent inhibitors of vacuolar-type ATPase, which is strongly implicated in age-related diseases. However, the binding site of the antibiotic on the protein remains unclear because of the complexity of the structure of Baf bound to the target subunit in the transmembrane region. For conducting structural studies by applying solid-state NMR, which is one of the most promising methodologies available for structural analysis in membrane system, preparing bioactive fluorinated Baf analogues is essential. In this study two Baf analogues were carefully designed and efficiently synthesized through the convergent coupling of three segments. Biological evaluation revealed that the activity of 24-F-Baf was comparable to that of Baf, indicating its utility as a potential probe for solid-state NMR analysis. By contrast, desmethylated 24-F-Baf exhibited markedly diminished activity. The absence of two methyl groups caused a critical conformational change in the macrocyclic core structure necessary for binding to the target protein.

Full text of DOI:10.1002/asia.201403299

DOI: 10.1002/asia.201403299
Full Paper
Solid-State NMR
Modification of Bafilomycin Structure to Efficiently Synthesize
Solid-State NMR Probes that Selectively Bind to Vacuolar-Type
ATPase
Hajime Shibata,[a, b] Hiroshi Tsuchikawa,[a] Tatsuru Hayashi,[a] Nobuaki Matsumori,[a]
Michio Murata,*[a, b] and Takeo Usui[c]
Abstract: Bafilomycin (Baf) is one of the most potent inhibi-
tors of vacuolar-type ATPase, which is strongly implicated in
age-related diseases. However, the binding site of the antibi-
otic on the protein remains unclear because of the complex-
ity of the structure of Baf bound to the target subunit in the
transmembrane region. For conducting structural studies by
applying solid-state NMR, which is one of the most promis-
ing methodologies available for structural analysis in mem-
brane system, preparing bioactive fluorinated Baf analogues
is essential. In this study two Baf analogues were carefully
designed and efficiently synthesized through the convergent
coupling of three segments. Biological evaluation revealed
that the activity of 24-F-Baf was comparable to that of Baf,
indicating its utility as a potential probe for solid-state NMR
analysis. By contrast, desmethylated 24-F-Baf exhibited mark-
edly diminished activity. The absence of two methyl groups
caused a critical conformational change in the macrocyclic
core structure necessary for binding to the target protein.
Introduction
but also in the plasma membranes of specialized cells such as
osteoclasts and certain tumor cells.[2–4] Moreover, because V-AT-
Pases have been implicated in numerous diseases such as os-
teoporosis, cancer, and diabetes, and have attracted considera-
ble biomedical research scrutiny as a promising drug target, V-
ATPase have been extensively investigated in terms of the
pathology of these diseases and drug development.[5–7]
Eukaryote cells contain acidic organelles such as lysosomes
and endosomes, which possess unique structures and func-
tions. The acidic nature of the organelles is considered to be
critical for various physiological processes such as protein deg-
radation and endocytosis. A membrane-bound protein, vacuo-
lar-type ATPase(V-ATPase), is the main molecule that generates
this acidic environment by actively transporting protons
against the concentration gradient by using ATP. Since the dis-
covery of V-ATPase in the vacuole of yeast,[1] it has been found
not only in the endomembrane system of all eukaryotic cells
Bafilomycin A1 (Baf, 1), which was first isolated in 1983 from
Streptomyces griseus ssp. sulphurus,[8] is a representative, highly
potent V-ATPase inhibitor.[9] Baf is composed of two main
parts, a 16-membered macrolactone and a tetrahydropyrane
(THP) structure, tethered by a 4-carbon acyclic chain, encom-
passing 9 stereogenic centers (Figure 1). Early investigation of
its mode of action implied that the Baf-binding target is the
transmembrane Vo domain of V-ATPase.[10,11] Subsequent
point-mutation[12] and photoaffinity-labeling[13] studies identi-
fied the putative binding site of Baf to be subunit c (and subu-
nit a) in the Vo domain. However, despite the intensive use of
Baf as an inhibitor in physiological studies on V-ATPase, the de-
tailed interaction of Baf with V-ATPase at the atomistic level re-
mains unclear. The most challenging problem is the structural
[a] Dr. H. Shibata, Dr. H. Tsuchikawa, T. Hayashi, Prof. Dr. N. Matsumori,+
Prof. Dr. M. Murata
Graduate School of Science
Osaka University
1-1 Machikaneyama, Toyonaka, Osaka 560-0043 (Japan)
Fax: (+81)6-6850-5774
E-mail: murata@chem.sci.osaka-u.ac.jp
[b] Dr. H. Shibata, Prof. Dr. M. Murata
ERATO, Lipid Active Structure Project
Japan Science and Technology Agency
1-1 Machikaneyama, Toyonaka, Osaka 560-0043 (Japan)
[c] Prof. Dr. T. Usui
Faculty of Life and Environmental Sciences
University of Tsukuba
Tennodai, Tsukuba, Ibaraki 305-8572 (Japan)
[+] Present address:
Graduate School of Sciences
Kyushu University
6-10-1 Hakozaki, Higashi-ku, Fukuoka, 812-8581 (Japan)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/asia.201403299.
Figure 1. Structure of bafilomycin A1.
Chem. Asian J. 2015, 10, 915 – 924
915
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
analysis of V-ATPase because it forms a macromolecular com-
plex that exhibits high heterogeneity, and this prevents the ap-
plication of X-ray crystallography. Although the entire structure
of a yeast V-ATPase has been determined using cryo-electron
microscopy and certain X-ray structures of its subunits were re-
cently reported,[14,15] detailed structural information about the
transmembrane Vo domain has remained unavailable thus far.
Moreover, subunit c, the probable binding site of Baf, forms
the two methyl groups located at C-6 and C-8 are crucial for
Baf activity. By performing NMR-based conformational analysis
of them, we have revealed for the first time that critical confor-
mational alteration of the macrolactone ring, including the di-
rection of the C-7 hydroxy group, markedly influences the bio-
logical activity of Baf.
hetero-oligomers with its isoforms (subunit c’ and subunit c’’) Results and Discussion
in yeast V-ATPases, which makes the structural analysis even
Design and synthetic plan for fluorinated bafilomycin
analogues
more challenging.
One of the useful methods for the structural analysis of non-
crystal, heterogeneous, and hardly solubilized entities is solid-
state NMR; thus, this technique has attracted considerable at-
tention over the last few decades.[16] Among the solid-state
NMR techniques, rotational echo double resonance (REDOR)[17]
has enabled us to obtain interatomic distance information.
REDOR has been successfully applied in the structure elucida-
tion of non-solubilized protein–ligand complexes[18,19] and
drug–lipid assembles.[20] Therefore, solid-state NMR measure-
ment appears to be the optimal approach for analyzing the
structure of the highly complicated Baf–ATPase complex. How-
ever, certain key challenges must be overcome. First, the prep-
aration of labeled compounds featuring an NMR-sensitive nu-
cleus is essential. Specifically, 19F is the most appropriate nu-
cleus because it exhibits a high gyromagnetic ratio and low
background signals in biological samples favorable for long
and precise distance measurements.[21] However, the introduc-
tion of 19F atoms into small compounds often causes large per-
turbations in biological activity.[22] Second, in NMR experiments
conducted with labeled ligands in complex with receptors,
most of the ligand molecules must bind to the original binding
site while a negligible amount of the ligand stays unbound;
this is because NMR signals are observed from both the bound
and the unbound ligands and occasionally these overlap sub-
stantially, which seriously hampers the measurement of intera-
tomic distances.
In designing the fluorinated Baf analogues, our first consider-
ation was that 19F labeling must be introduced into the molec-
ular framework in order to facilitate precise distance measure-
ment. Next, we attempted to design a target compound based
on considering mainly these three points: retention of biologi-
cal activity, efficient synthesis on a multi-milligram scale, and
chemical stability during several days of NMR measurement.
Initially, we aimed at substituting a hydroxy group with a flu-
orine atom because of their chemical compatibility. However,
this could not be applied to all four hydroxy groups because
replacing the 19- or 21-hydroxy group with fluorine would de-
stabilize the product and because the groups at C7 and C17
were reported to be critical for Baf activity.[24] Next, we focused
on the C-23 position in the THP ring because alteration of its
substituent to larger or smaller groups such as pentadienyl or
ethyl moieties did not compromise the activity.[25] Moreover,
we envisioned that a CF3 group could be introduced more
readily than a fluorine atom and would be compatible with
the iPr group with respect to bulkiness.[26] Consequently, we
designed 24,24-didesmethyl-24-F3-Baf (24-F-Baf, 2) as a synthet-
ic target. As described in the synthetic part, desmethyl ana-
logue 3 of the macrolide moiety was also designed as a target
to facilitate the supply of an NMR probe in tens of milligrams.
Two fluorinated Baf analogues, 2 and 3, could be synthe-
sized using a common synthetic strategy (Scheme 1). Based on
considering previous synthetic studies of 1,[27] we concluded
that CÀC bond formation at C17 and C18 of ethyl ketone 4
and macrolactone 5 through a diastereoselective aldol reac-
tion[27b] would be favorable because the CF3 group appeared
Recently, in a brief communication, we reported the synthe-
sis of a 24-fluorinated Baf analogue (24-F-Baf 2).[23] Here, we
describe the underlying concept in designing NMR probes and
synthetic schemes in greater detail. More importantly, in this
study, we synthesized a new
19F-labeled analogue, a 6,8-di-
desmethylated derivative (des-
methyl-24-F-Baf 3), which was
designed in order to reduce the
number of synthetic steps. The
first analogue 2 was found to
exhibit potent inhibitory activity
toward V-ATPase, and thus it
could serve as a useful probe
for elucidating its protein-
bound structure by using solid-
state NMR. Conversely, the
second analogue 3 unexpected-
ly showed considerably reduced
activity, which indicated that
Scheme 1. Structure of fluorinated bafilomycin analogues and their retrosynthetic analysis.
Chem. Asian J. 2015, 10, 915 – 924
www.chemasianj.org
916
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
to exert little effect on the stereoselectivity.[27a] A novel CF3-la-
beled C18ÀC24 segment 4 could be synthesized through a tri-
fluoromethylation and two-carbon elongation of aldehyde 6,
with the construction of the stereogenic center at C23. The
C1ÀC17 segment 5 could be obtained via Stille coupling be-
tween C11 and C12,[27b,c,f] followed by macrolactonization from
the known C12ÀC17 segment 7[28] and C1ÀC11 segment 8.[27c,f]
Synthesis of segments 7 and 8 could be achieved by adopting
the strategy reported for 1, with certain modifications, and the
desmethyl C1ÀC11 segment 9 could be synthesized in short
steps even more easily than 8. Notably, in order to synthesize
fluorinated analogues efficiently, difficulties caused by fluorine,
such as high electronegativity and unpredictable chemical re-
activity, would have to be controlled.
Synthesis of fluorinated bafilomycin analogues
The synthesis of the C18ÀC24 segment 4 commenced with the
conversion of 1,3-propanediol into the chiral Weinreb amide
10 in 4 steps,[29] followed by p-methoxybenzyl (PMB) protection
of the hydroxy group and diisobutylalminum hydride (DIBAL)
reduction of amide to generate aldehyde 6 (Scheme 2). Next,
stereoselective trifluoromethylation of 6 was attempted using
trifluoromethyltrimethylsilane (TMSCF3) treatment with tetrabu-
tylammonium fluoride (TBAF),[30] which unexpectedly led to
the preferential formation of the undesired 23R alcohol 11.
Thus, we sought to invert the configuration at C23 of the read-
ily separable undesired 23R-isomer 11 through an oxidation–
reduction sequence. After Dess–Martin periodinane (DMP) oxi-
dation of alcohol 11 and PMB removal, 1,3-anti-selective reduc-
tion of 13 was attempted using the b-hydroxy group. First, we
used triacetoxyborohydride[31] as a reducing agent, but the de-
sired diol 14 was obtained as a minor product. We suspected
that this unexpected result was obtained because of the unde-
sired chelation of the axial CF3 group to the reagent as shown
in TS-13. Second, we conducted the hydride transfer of benzal-
dehyde in the presence of samarium iodide;[32] however, this
resulted in simultaneous epimerization at the C22 position be-
cause of the high electron-withdrawing property of the CF3
group. Finally, after DMP oxidation and subsequent reduction
of ketone with l-Selectride, a non-chelating and bulky borohy-
dride, we found that the desired 23S alcohol 12 was preferen-
tially produced in a 4.2:1 ratio and 68% separated yield.[33]
Next, the PMB group was removed, and the resulting diol was
Scheme 2. Synthesis of the C18ÀC24 segment 4. DDQ=2,3-dichloro-5,6-di-
cyano-1,4-benzoquinone.
protected as a cyclic silyl ether to generate compound 15. The
deprotection of the primary alcohol followed by DMP oxida-
tion furnished the aldehyde, and then the addition of ethyl-
magnesium bromide and subsequent oxidation of the alcohol
generated the C18ÀC24 segment 4 in 80% yield in 4 steps.
Having obtained the common CF3-labeled segment 4, the
synthesis of 24-F-Baf (2) was next examined. First, the known
C12ÀC17 segment 7[28] was prepared by modifying Toshima’s
protocol[27b] (Scheme 3). Methyl (S)-3-hydroxyisobutyrate was
protected using the 4-monomethoxytrityl (MMTr) group, and
then DIBAL reduction and Swern oxidation were performed to
produce aldehyde 16. Next, diastereoselective methoxyallyla-
tion of aldehyde 16 by using chromium(II) chloride and acrole-
in dimethyl acetal in the presence of trimethylsilyl iodide
(TMSI)[34] generated the desired alcohol 17 as the main isomer.
The terminal olefin of purified 17 was oxidatively cleaved to an
aldehyde, and subsequent alkynation and hydrostannation
generated the desired C12ÀC17 segment 7 in 48% yield in
4 steps.
Scheme 3. Synthesis of the C12ÀC17 segment 7. DMAP=4-dimethylaminopyridine, NMO=N-methylmorpholine N-oxide.
Chem. Asian J. 2015, 10, 915 – 924 www.chemasianj.org ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
917

Full Paper
Scheme 4. Synthesis of the C1ÀC11 segment 8. TBACl=tetrabutylammonium chloride, NCS=N-chlorosuccinimide, KHMDS=potassium hexamethyldisilazane.
Next, the synthesis of the C1ÀC11 segment 8 was initiated
with the preparation of aldehyde 21 from methyl (R)-3-hydrox-
yisobutyrate in 3 steps[35] (Scheme 4). Under the conditions re-
ported by Paterson,[36] the (E)-boron enolate generated from
ketone 22 reacted with aldehyde 21 to produce the desired
anti-aldol adduct 23 as the main diastereomer (d.r.=16:1). In
this reaction, favorable selectivity was achieved through the
transition state TS-22, in which the benzoyl moiety of ketone
22 formed hydrogen bonds with the aldehyde hydrogen to
assist the reaction on the Si-face of aldehyde 21. After tert-bu-
tyldimethylsilyl (TBS) protection of the secondary alcohol, the
reduction of the ketone and removal of the benzoyl group
was performed in a one-pot reaction to generate diol 24 in
79% yield in 2 steps. The subsequent oxidative cleavage of the
diol and NaBH4 reduction furnished the primary alcohol 25,
which was further converted into the known alkyne 26[27c,f] via
two-carbon elongation with lithium acetylide and removal of
the PMB group. Lastly, following primarily a report by Mar-
C11 were constructed. On the primary alcohol of 26, Swern ox-
idation and two-carbon elongation through Wittig reaction
were performed, and then DIBAL reduction afforded an allyl al-
cohol, which was exposed to Negishi’s condition[37] to produce
the vinyl iodide 27 without any problem. Finally, the primary
alcohol of 27 was oxidized using 2,2,6,6-tetramethylpiperidine-
1-oxyl (TEMPO) and the subsequent Horner–Wadsworth–
Emmons reaction proceeded in a Z-selective manner in moder-
ate yield, and this was followed by TBS removal to generate
the desired C1ÀC11 segment 8.
Having prepared the segments 4, 7, and 8, we focused on
their coupling and the completion of the synthesis of 24-F-Baf
(2) (Scheme 5). Initially, the Stille coupling between segments 7
and 8 performed using Marshall’s condition[27f] proceeded ef-
fectively to generate diene 28 in high yield. After the hydroly-
sis of the methyl ester,[27b] the obtained seco acid was exposed
to the Yamaguchi condition[38] by closely following Carreira’s
protocol[27e] to afford macrolactone 29 in 51% yield in 2 steps.
shall,[27f] the C1ÀC4 part and the vinyl iodide moiety at C10À In this reaction, the free 7-hydroxy group was critical for suc-
Scheme 5. Synthesis of 24-F-Baf (2).
Chem. Asian J. 2015, 10, 915 – 924
www.chemasianj.org
918
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
cessful cyclization as previously reported.[27c] Next, the secon-
dary alcohol of 29 was protected using the diethylisopropylsilyl
(DEIPS) group,[39] followed by the removal of the MMTr group
by pyridinium-p-toluene sulfonate (PPTS) treatment; the result-
ing alcohol was oxidized to aldehyde 30, which was used as
the crude material in the next coupling reaction performed
with segment 4. Referring to Evans’s method,[27a,b] we exam-
ined the aldol reaction between the crude aldehyde 30 and
(E)-boron enolate formed by treating segment 4 with dichloro-
phenylborane and diisopropylethylamine. The reaction pro-
ceeded smoothly, and the desired b-hydroxyketone 31 was ob-
tained in acceptable yield over 2 steps with high diastereose-
lectivity (d.r.=20:1).[40] As we had aimed, the stereoselective
aldol reaction at C17 and C18 was not influenced by the char-
acteristic property of the CF3 group. Lastly, the cyclic silyl and
DEIPS groups were removed in a stepwise manner to afford
24-F-Baf (2) in 36% yield after HPLC purification.
As described later, 24-F3-Baf 2 was found to retain the origi-
nal biological activity of Baf. However, the synthesis of 2 was
as laborious as that of the natural product. Thus, we attempted
to enhance the synthetic efficiency by reducing the number of
asymmetric centers. We planned to remove the methyl groups
on the macrolactone, based on our hypothesis that the macro-
cyclic conformation is critical for the biological activity of Baf
and based on Macromodel calculations, which showed that re-
moving the two methyl groups at C-6 and C-8 did not appear
to substantially alter the conformation of the 16-membered
ring.[41] Other stereogenic substituents on the macrolides, such
as 7-OH and 14-OCH3, could not be removed: 7-OH is reported
to be necessary[24b] and 14-OCH3 is the common functionality
in the plecomacrolide family. These considerations led us to
design 6,8-24,24-tetradesmethyl-24-F3-Baf (desmethyl-24-F-Baf,
3) as the second synthetic target.
The facile preparation of desmethyl C1ÀC11 segment 9 was
first examined starting from 4-pentyne-1-ol. After the oxidation
of alcohol to aldehyde, the single stereogenic center at C7 was
constructed through Keck asymmetric allylation[42] to generate
the known alcohol 32 in 66% yield and 94% ee (determined
using a modified Mosher’s method[43]). Next, the terminal olefin
was converted into an aldehyde through ozonolysis, which
was followed by the same transformation as in the case of seg-
ment 8 to afford the desired desmethyl C1ÀC11 segment 9 in
5 steps. Moreover, the established procedure was applied to
the next 3 steps, that is, Stille coupling between segments 9
and 7, saponification, and Yamaguchi macrolactonization. Un-
expectedly, the cyclization reaction was not successful (the
yield of 37 was <20%), and scrutinizing its byproducts re-
vealed the isomerization of the 2,4-diene moiety and, subse-
quently, the cyclization of the 7-hydroxy group to the C1 car-
bonyl carbon was found to occur, probably triggered by
proton abstraction from the non-methylated C6 position by
a base. Therefore, we speculated that a bulky protective group
at the 7-hydroxy group could prevent this undesired side reac-
tion in contrast to the case of the dimethylated compound 29.
After the triisopropylsilyl (TIPS) ether 34 was synthesized from
segment 9, the corresponding treatment in 3 steps resulted in
the successful formation of the macrolactone 38 in 69% yield
(in 2 steps), as expected. The subsequent conversion of 38 into
aldehyde 39 and the following aldol reaction with segment 4
proceeded smoothly,[44] and the final stepwise deprotection
generated smoothly the desired desmethyl-24-F-Baf (3).
Biological evaluation of fluorinated bafilomycin analogues
We evaluated the inhibitory effects of the two fluorinated Baf
analogues on V-ATPase using two methods.[45] First, we exam-
ined the effects of the analogues on the acidification of intra-
cellular acidic organelles by V-ATPase in rat 3Y1 fibroblasts;
The synthesis of desmethyl-24-F-Baf (3) commenced with
the construction of the desmethyl intermediate 9 (Scheme 6).
Scheme 6. Synthesis of desmethyl-24-F-Baf (3). BINOL=1,1’-bi-2-naphthol.
Chem. Asian J. 2015, 10, 915 – 924
www.chemasianj.org
919
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
3Y1 cells were treated with Baf (1), 24-F-Baf (2), or desmethyl-
24-F-Baf (3) for 2 h and then stained with 5 mm acridine orange
for 15 min (Figure 2). In contrast to treatment with DMSO (neg-
ative control), in which normal acidic cells were observed as
of 3 here was not as weak as what was observed in the V-ATP
assay.
Bioactivity based on conformational analysis
Based on the previous structure–activity relationship (SAR) ob-
tained for Baf,[24b] the entire characteristic macrolactone struc-
ture was considered to be a prerequisite for Baf activity; how-
ever, this conclusion might not be warranted because one-site
modification of the substituents in the macrolactone part was
not performed except in the case of the 7-hydroxy group.
Thus, our result showing that removing the two methyl groups
at C6 and C8 critically influenced the activity provides new
knowledge in terms of the SAR of synthetic derivatives. There-
fore, we were motivated to investigate in detail the effect of
the methyl groups on the conformation of the macrolactone
ring by means of NMR analysis.
Initially, the 1H chemical shifts and NOE correlation of the flu-
orinated analogues were compared with those of natural Baf
(1) (Figure 3).[47] No large difference existed between 24-F-Baf
Figure 2. Effect of fluorinated Baf analogues on acridine orange stain of 3Y1
cells.
a red fluorescence of the acridine orange stain, treatment with
100 nm 24-F-Baf (2) caused complete disappearance of fluores-
cence and was as effective as treatment with 100 nm Baf (1)
(positive control). This indicated that 24-F-Baf (2) strongly in-
hibited V-ATPase in the cells. Conversely, after treatment with
10 mm desmethyl-24-F-Baf (3), fluorescence was clearly ob-
served, indicating a loss of the inhibitory activity. Next, we
quantitatively assayed V-ATPase inhibition: we used V-ATPase
in the purified vacuole membrane of budding yeast and mea-
sured the inorganic phosphate liberated from ATP (Table 1).[46]
Figure 3. Comparison of NMR data at the C1ÀC15 position between Baf and
fluorinated Baf analogues.
(2) and Baf (1) around the C1ÀC15 position (Dd(1H)
<0.05 ppm, NOE between H5 and H8 as well as H8 and H11
was observed). By contrast, analysis of desmethyl-24-F-Baf (3)
revealed large differences in the chemical shift at H3 (1/3=
6.68/6.41, Dd=0.27) and H13 (1/3=5.16/5.24, Dd=0.08) in ad-
dition to that in the C6ÀC9 region, which was assumed to
change due to the lack of methyl groups. Moreover, distinct
NOE correlation between H7 and H9 and H9 and H11 was ob-
served instead of H5/H8 and H8/H11 correlations. These results
suggested that conformational change occurred not only at
the C6-C9 moiety but also over a wider range of the macrolac-
tone because of the absence of C6/C8 methyl groups.
Table 1. Yeast V-ATPase inhibition.
IC50 [nm]
Baf (1)
24-F-Baf (2)
Desmethyl-24-F-Baf (3)
2.3
2.5
>10000
The results clearly showed that the inhibitory activity of 24-F-
Baf (2) was similar to that of natural Baf (1) (IC50 =2.5 and
2.3 nm, respectively), and that desmethyl-24-F-Baf (3) exerted
no noticeable effect even at 10 mm. These results suggested
that 24-F-Baf (2) could be useful for structural analysis of the
Baf–V-ATPase complex by solid-state NMR, but that 24-des-
methyl-F-Baf (3), which was designed for simplifying the syn-
thetic process, was not suitable for such studies.
To perform conformational analysis in additional detail, an
energy calculation by Macromodel was conducted using the
Truncated Newton Conjugate Gradient (TNCG) algorithm con-
strained by experimentally obtained NOEs and coupling con-
stants (3JHH) (Figure 4). Although our initial calculation of the
lowest-energy conformation of natural Baf and 6,8-didesmethyl
Baf without constraints revealed that the shape of the macro-
cyclic core of desmethyl derivative was almost the same as
that of natural Baf (Supporting Information, Figure S2), the cal-
culations performed under the NMR constraints revealed an
To further examine whether 3 completely lacked biological
activity, cytotoxicity assays were performed using HL-60 and
K562 cell lines (Supporting Information, Figure S1). Unexpect-
edly, desmethyl analogue 3 inhibited cell-growth at moderate
level (ED50 was around 3 mm in the case of both cell lines); this
was 100 times weaker than the cytotoxicity of 1, but the effect
Chem. Asian J. 2015, 10, 915 – 924
www.chemasianj.org
920
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
was minimized by using appropriate synthetic route. The V-
ATPase inhibitory activity of 24-F-Baf (2) was comparable to
that of the natural product, indicating the utility of 2 as a po-
tential molecular probe for investigating the inhibitory mecha-
nism of Baf by using solid-state NMR. Moreover, although the
V-ATPase inhibitory activity of desmethyl-24-F-Baf (3) was unex-
pectedly lost, detailed conformational analysis of its macrolac-
tone moiety provided the new insight that the presence of C6
and C8 methyl groups is crucial for maintaining the proper ar-
chitecture required for exhibiting a potent biological activity.
Experimental Section
C18ÀC24 segment 4
To a solution of alcohol S15–3 (46.0 mg, 124.1 mmol) in CH2Cl2
(2.4 mL) was added NaHCO3 (156 mg, 1.86 mmol) and DMP
(158 mg, 372 mmol) at 08C. After being stirred at room tempera-
ture for 9 h, the reaction mixture was quenched with saturated
aqueous NaHCO3 and Na2S2O3. The aqueous layer was extracted
with CH2Cl2, and the organic layer was washed with brine, dried
over anhydrous Na2SO4, filtered and concentrated under reduced
pressure. Purification by silica gel column chromatography
Figure 4. The lowest-energy conformation of Baf and desmethyl-24-F-Baf
calculated using Macromodel with constraints from NOEs and JHH data.
3
unexpected conformational change of the macrolactone. The
conformation of the C6ÀC9 position, including the direction of
the critical 7-hydroxy group, was altered substantially and
moreover, the entire structure of the macrocyclic core was con-
siderably distorted and had become more compact than that
of natural Baf 1. As a result of this deformation, the relative po-
sition between the dienoate (C1ÀC5) and the diene (C10ÀC13)
was changed, which possibly led to the chemical shift change
at C3 and C13 positions.
(hexane/EtOAc=50:1 to 10:1) afforded ketone
112 mmol, 90%) as a colorless oil.
4 (41.4 mg,
Rf =0.70 (hexane/EtOAc=3:1); ½aŠ24 = +50.2 (c=1.91, CHCl3); IR
D
(film): n˜ =2971, 2895, 1721, 1469, 1383, 1365, 1218, 1096, 1011,
1
961, 916, 826, 651 cmÀ1; H NMR (500 MHz, CDCl3): d=4.73 (1H, dt,
J=10.0, 4.0 Hz), 4.10 (1H, quin, J=7.0 Hz), 2.72 (1H, dd, J=15.0,
10.0 Hz), 2.53 (1H, q, J=7.5 Hz), 2.40–2.31 (2H, m), 1.07 (3H, t, J=
7.5 Hz), 1.03–0.99 ppm (21H, m); 13C NMR (125 MHz, CDCl3): d=
209.1, 124.8 (q, J=282 Hz), 74.2 (q, J=31.2 Hz), 71.1, 45.5, 36.8,
36.1, 27,2, 27.1, 21.4, 21.2, 13.4, 7.5 ppm; 19F NMR (470 MHz, CDCl3):
d=À78.15 ppm (d, J=7.5 Hz); HRMS (ESI-TOF) m/z calcd for
C17H31F3O3SiNa [M+Na+] 391.1887, found: 391.1890.
Based on these results, we suggest that removal of the two
methyl groups in Baf caused a substantial conformational
change in the macrocyclic core and this led to the drastic drop
in the V-ATPase inhibitory activity. Although it remains unclear
to what extent the directional change of the 7-hydroxy group
and the conformational alteration of the entire macrocyclic
structure contribute to the biological function of Baf, we di-
rectly confirmed that these perturbations critically affect it. By
contrast, desmethyl-24-F-Baf 3, which completely lacked V-
ATPase inhibitory activity, exhibited moderate cytotoxicity
toward leukemia cell lines, which raises the possibility that 3
exerted its activity through target proteins other than V-
ATPase; for example, Shacka et al.[48] have reported that 1 inhib-
its autophagy through a mechanism other than V-ATPase in-
hibition. Thus, the synthesis of slightly modified analogues can
help elucidate hidden mechanisms of action of plecomacrolide
family members, which might otherwise be masked by their
powerful inhibition of V-ATPases.
24-F-Baf (2)
To a solution of alcohol 31 in THF (0.30 mL) was added 18% HF·Py
(10.3 mL, 102 mmol) at 08C. After being stirred at room temperature
for 1 h, the reaction mixture was poured into saturated aqueous
NaHCO3. The aqueous layer was extracted with ether, the organic
layer was washed with brine, dried over anhydrous Na2SO4, filtered
and concentrated under reduced pressure. Rough purification by
silica gel column chromatography (hexane/EtOAc=7:1 to 2:1) af-
forded 7-protected 24-F-Baf (5.3 mg). To a solution of 7-protected
24-F-Baf in CH2Cl2 (38 mL), MeCN (380 mL), H2O (95 mL) was added
TsOH·H2O (14.6 mg, 76.8 mmol) at 08C. After being stirred at room
temperature for 3 h, TsOH·H2O (7.3 mg, 38.4 mmol) was added. The
resulting mixture was stirred for 12 h and quenched with saturated
aqueous NaHCO3. The aqueous layer was extracted with CH2Cl2,
the organic layer was dried over anhydrous Na2SO4, filtered and
concentrated under reduced pressure. Purification by silica gel
column chromatography (hexane/EtOAc=5:1 to 1:1), then by
HPLC (COSMOSIL 5SL-II 10”250 mm, eluent: hexane/EtOAc=1:1,
Conclusions
Two fluorine-labeled Baf analogues, 24-F-Baf (2) and desmeth-
yl-24-F-Baf (3), were designed and convergently synthesized
from two common segments, C18ÀC24 segment 4 and C12À
C17 segment 7, and one respective segment, C1ÀC11 segment
8 or desmethyl C1ÀC11 segment 9, through Stille coupling,
macrolactonization, and diastereoselective aldol reaction. In
this synthesis, the adverse effect of trifluoromethyl substitution
flow rate: 2.0 mLminÀ1
, tR=15.7 min) to afford 24-F-Baf (2)
(1.2 mg, 1.87 mmol, 36% over 2 steps) as a white solid: Rf =0.38
(hexane/EtOAc=1:1); ½aŠ22 = +10.3 (c=0.0318, CHCl3); IR (film):
D
n˜ =3404, 2959, 2924, 2850, 1685, 1461, 1259, 1099, 799 cmÀ1
;
1H NMR (500 MHz, CDCl3): d=6.63 (1H, s), 6.51 (1H, d, J=15.0,
10.5 Hz), 6.17 (1H, d, J=2.0 Hz), 5.81 (1H, d, J=10.5 Hz), 5.75 (1H,
d, J=9.0 Hz), 5.15 (1H, dd, J=15.5, 9.0 Hz), 4.94 (1H, dd, J=9.0,
Chem. Asian J. 2015, 10, 915 – 924
www.chemasianj.org
921
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1.0 Hz), 4.87 (1H, d, J=4.0 Hz), 4.09–4.02 (1H, m), 4.04 (1H, qd, J=
11.0, 6.5 Hz), 3.90 (1H, d, J=9.5 Hz), 3.75 (1H, ddd, J=11.0, 11.0,
5.0 Hz), 3.63 (3H, s), 3.33–3.26 (1H, m), 3.25 (3H, s), 2.56–2.49 (1H,
m), 2.34 (1H, dd, J=12.0, 5.0 Hz), 2.18–2.10 (1H, m), 2.15 (1H, ddq,
J=10.5, 7.0, 1.0 Hz), 2.00–1.92 (1H, m), 1.95–1.86 (1H, m), 1.87–
1.79 (1H, m), 1.98 (3H, s), 1.94 (3H, s), 1.65–1.55 (1H, m), 1.31–1.20
(1H, m), 1.12 (3H, d, J=6.5 Hz), 1.07 (3H, d, J=7.0 Hz), 1.06 (3H, d,
J=7.0 Hz), 0.93 (3H, d, J=6.0 Hz), 0.83 ppm (3H, d, J=7.0 Hz);
13C NMR (151 MHz, CDCl3): d=167.4 143.3 142.6 141.4, 133.2, 133.2
133.1, 127.1 125.3 100.6, 81.9, 81.2, 76.5, 71.9 (q, J=30.0 Hz), 70.4
69.7 59.6, 55.5, 42.9 41.3, 41.0, 40.2 38.9 37.1 36.7 21.7 20.2 17.2
14.0, 11.9, 9.6 7.0 ppm; 19F NMR (470 MHz, CDCl3): d=À73.36 ppm
(d, J=7.0 Hz); HRMS (ESI-TOF) m/z calcd for C33H51F3O9Na [M+Na+
] 671.3377, found: 671.3385.
the organic layer was dried over anhydrous Na2SO4, filtered and
concentrated under reduced pressure. Purification by silica gel
column chromatography (hexane/EtOAc=3:1 to 1:2), then by
HPLC (COSMOSIL 5SL-II 10”250 mm, eluent: hexane/EtOAc=1:1,
flow rate: 2.0 mLminÀ1, tR =19.6 min) to afford desmethyl-24-F-Baf
(3) (15.8 mg, 25.4 mmol, 60% over 2 steps) as a white solid: Rf =
0.27 (hexane/EtOAc=1:1); ½aŠ23 = +92.6 (c=0.0343, CHCl3); IR
D
(film): n˜ =3413, 2918, 2850, 2363, 2339, 1641, 1631, 1219, 1102,
;
772 cmÀ1 1H NMR (500 MHz, CDCl3): d=6.53 (1H, dd, J=15.0,
11.0 Hz), 6.41 (1H, s), 6.13 (1H, d, J=2.0 Hz), 5.82 (1H, d, J=
10.5 Hz), 5.57 (1H, dd, J=10.0, 5.0 Hz), 5.24 (1H, dd, J=15.0,
9.5 Hz), 4.91 (1H, dd, J=9.0, 1.5 Hz), 4.83 (1H, dd, J=4.0, 1.0 Hz),
4.10–4.04 (1H, m), 4.09–3.99 (1H, m), 3.90 (1H, t, J=9.0 Hz), 3.74
(1H, ddd, J=10.0, 10.0, 5.0 Hz), 3.60 (3H, s), 3.55–3.48 (1H, m), 3.27
(3H, s), 2.65 (1H, ddd, J=13.0, 10.0, 4.0 Hz), 2.52–2.43 (1H, m), 2.34
(1H, dd, J=12.0, 5.0 Hz), 2.20 (1H, ddq, J=10.0, 7.0, 1.5 Hz), 2.15–
2.08 (1H, m), 2.13–2.07 (1H, m), 1.96 (3H, s), 1.96–1.91 (1H, m),
1.84 (3H, s), 1.84–1.82 (1H, m), 1.64–1.55 (1H, m), 1.26–1.23 (1H,
m), 1.12 (3H, d, J=5.5 Hz), 1.06 (3H, d, J=7.0 Hz), 0.84 ppm (3H,
d, J=7.0 Hz); 13C NMR (151 MHz, CDCl3): d=167.2, 142.2, 141.0,
135.4, 133.4, 133.0, 131.6, 129.0, 125.5, 125.0 (q, J=279 Hz), 100.7,
81.9, 76.1, 74.5, 72.0 (q, J=30 Hz,), 70.4, 69.6, 59.3, 56.6, 42.8, 41.1,
38.8, 38.2, 37.4, 36.6, 36.5, 16.3, 14.0, 11.9, 9.5, 7.0 ppm; 19F NMR
(470 MHz, CDCl3): d=À73.36 ppm (d, J=7.0 Hz); HRMS (ESI-TOF)
m/z calcd for C31H47F3O9Na [M+Na+] 643.3064, found: 643.3070.
Desmethyl C1-C11 segment 9
To a solution of iodoolefin 33 (974 g, 3.14 mmol) in CH2Cl2 (31 mL)
and pH8.6 buffer (31 mL) was added TEMPO (73 mg, 0.47 mmol),
TBACl (86 mg, 0.31 mmol) and NCS (629 mg, 4.71 mmol) at 08C.
After being stirred at room temperature for 2 h, the reaction mix-
ture was quenched with saturated aqueous NaHCO3 and Na2S2O3.
The aqueous layer was extracted with CH2Cl2, and the organic layer
was washed with brine, dried over anhydrous Na2SO4, filtered and
concentrated under reduced pressure. The residue was passed
through a silica gel (hexane/EtOAc=10:1 to 1:1) to afford aldehyde
S33, which was used in the next step without further purification.
To a solution of (iPrO)2P(O)CH(OMe)CO2Me (2.34 g, 8.74 mmol) in
THF (87 mL) was added 18-crown-6 ether (2.31 g, 8.74 mmol) and
KHMDS (0.5m in THF, 15.9 mL, 7.95 mmol) at 08C. After being
stirred at 08C for 30 min, a solution of aldehyde S33 in THF
(5.3 mL) was added at À158C. The resulting mixture was stirred at
À158C for 12 h and quenched with saturated aqueous NH4Cl. The
aqueous layer was extracted with ether, and the organic layer was
washed with brine, dried over anhydrous Na2SO4, filtered and con-
centrated under reduced pressure. Purification by silica gel column
chromatography (hexane/EtOAc=7:1 to 2:1) afforded segment 9
(754 mg, 1.91 mmol, 61% over 2 steps) as a colorless oil: Rf =0.58
Staining of intracellular acidic organellae
Cells were stained with acridine orange as described previous-
ly.[49,50] 3Y1 cells were seeded onto coverslips and incubated at
378C for 2 h with 100 nm Baf(1) or 100 nm 24-F-Baf (2) or 10 mm
desmethyl-24-F-Baf (3). They were then incubated for 15 min with
5 mm acridine orange. After three washes with phosphate buffered
saline, the coverslips were photographed using a cooled charge-
coupled device camera.
Yeast V-ATPase inhibition assay
(hexane/EtOAc=1/1); ½aŠ24 = +1.8 (c=1.06, CHCl3); IR (film): n˜ =
V-ATPase inhibition activities of Baf(1) or 24-F-Baf (2) or desmethyl-
24-F-Baf (3) were assayed against V-ATPases obtained from purified
vacuole membrane of budding yeast.[46] The reaction mixture
(135 mL) containing 5 mg of vacuolar membrane vesicles, 5 mm
MgCl2, 10 mm NH4Cl, 5 mm NaN3, 0.1 mm Na3VO4 and 25 mm Mes-
Tris (pH 6.9) was incubated with or without inhibitors as DMSO so-
lution at the indicated concentration for 10 min on ice. Then, the
reaction was started by adding 15 mL of 50 mm of Na2ATP. Activity
assays were run for 60 min at 378C. 150 mL of 0.6m perchloric acid
was added to stop the reaction and the liberated inorganic phos-
phate was measured by malachite green method.[45]
D
3473, 2949, 1721, 1436, 1267, 1021, 776, 668 cmÀ1
;
1H NMR
(400 MHz, CDCl3): d=6.61 (1H, s), 5.93 (1H, q, J=1.0 Hz), 5.84 (1H,
t, J=7.5 Hz), 3.79 (3H, s), 3.72–3.62 (1H, m), 3.66 (3H, s), 2.44–2.25
(4H, m), 1.99 (3H, d, J=1.0 Hz), 1.84 (3H, d, J=1.0 Hz), 1.69–
1.50 ppm (2H, m); 13C NMR (125 MHz, CDCl3): d=165.4, 147.7,
143.4, 134.3, 133.8, 129.0, 75.3, 70.8, 60.4, 52.2, 36.7, 35.9,
35.1, 24.0, 15.1 ppm; HRMS (ESI-TOF) m/z calcd for C15H23IO4Na
[M+Na+] 417.0539, found: 417.0538.
Desmethyl-24-F-Baf (3)
To a solution of alcohol S39 (38.9 mg, 42.5 mmol) in THF (4.3 mL)
was added 18% HF·Py (42.9 mL, 425 mmol) at 08C. After being
stirred at room temperature for 1 h, the reaction mixture was
poured into saturated aqueous NaHCO3. The aqueous layer was ex-
tracted with ether, the organic layer was washed with brine, dried
over anhydrous Na2SO4, filtered and concentrated under reduced
pressure. Rough purification by silica gel column chromatography
(hexane/EtOAc=7:1 to 2:1) afforded 7-protected desmethyl-24-F-
Baf (33.0 mg). To a solution of 7-protected desmethyl-24-F-Baf in
CH2Cl2 (0.63 mL), MeCN (3.1 mL), H2O (0.79 mL) was added
TsOH·H2O (123 mg, 650 mmol) at 08C. After being stirred at room
temperature for 3 h, TsOH·H2O (123 mg, 650 mmol) was added. The
resulting mixture was stirred for 3 h and quenched with saturated
aqueous NaHCO3. The aqueous layer was extracted with CH2Cl2,
Ab initio calculation of the lowest-energy conformation of
Baf (1) and desmethyl-24-F-Baf (3) with constraints in the
macrolactone ring
Conformations were calculated using the MacroModel software
version 9.9. Initial atomic coordinates and structure files were gen-
erated from the crystal data of Baf.[51] The lowest-energy structure
was obtained by the molecular mechanics simulation using
a 7000 steps of Monte Carlo conformation search and TNCG
energy minimization with MMFFs in vacuum. In order to emphati-
cally observe the difference of conformation of the macrolactone
ring, the conformation of the other part was preliminarily fixed by
using two distance constraints for both compounds (2.5Æ1 Š be-
tween C1-carbonyl oxygen and C17-OH hydrogen atoms and C17-
Chem. Asian J. 2015, 10, 915 – 924
www.chemasianj.org
922
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
gelski, D. Steuber, A. K. Mehta, D. W. Kulp, P. H. Axelsen, J. Schaefer, J.
Mol. Biol. 2006, 357, 1253–1262.
OH oxygen and C19-OH hydrogen atoms), which were derived
from the hydrogen bonds data in Baf crystal structure.[51] Three di-
hedral angle constraints from NOE correlations and JH,H coupling
3
[20] a) S. Matsuoka, H. Ikeuchi, N. Matsumori, M. Murata, Biochemistry 2005,
44, 704–710; b) Y. Kasai, N. Matsumori, Y. Umegawa, S. Matsuoka, H.
Ueno, H. Ikeuchi, T. Oishi, M. Murata, Chem. Eur. J. 2008, 14, 1178–1185;
c) Y. Umegawa, N. Matsumori, T. Oishi, M. Murata, Biochemistry 2008, 47,
13463–13469.
[21] a) A. S. Ulrich, Prog. Nucl. Magn. Reson. Spectrosc. 2005, 46, 1–21; b) J.-x.
Yu, V. D. Kodibagkar, W. Cui, R. P. Mason, Curr. Med. Chem. 2005, 12,
819–848.
constants (1808Æ358 for O1-C1-C2-C3, C2-C3-C4-C5, and H14-C14-
C15-H15) were used in the case of Baf (1). Two distance constraints
from NOE correlations (2.5Æ1 Š between C7-hydrogen and C9-a-
hydrogen atoms and C9-b-hydrogen and C11-hydrogen atoms)
3
and four dihedral angle constraints from NOE correlations and JH,H
coupling constants (1808Æ358 for O1-C1-C2-C3, C2-C3-C4-C5, H6a-
C6-C7-H7, H14-C14-C15-H15) were used in the case of desmethyl-
24-F-Baf (3).
[22] B. E. Smart, J. Fluorine Chem. 2001, 109, 3–11.
[23] Preliminary communication: H. Shibata, H. Tsuchikawa, N. Matsumori,
M. Murata, T. Usui, Chem. Lett. 2014, 43, 474–476.
[24] a) S. Drçse, K. U. Bindseil, E. J. Bowman, A. Siebers, A. Zeeck, K. Alten-
dorf, Biochemistry 1993, 32, 3902–3906; b) S. Gagliardi, P. A. Gatti, P. Bel-
fiore, A. Zocchetti, G. D. Clarke, C. Farina, J. Med. Chem. 1998, 41, 1883–
1893.
Acknowledgements
We thank Dr. N. Inazumi (Osaka University) for his help in NMR
measurements. This work was supported in part by JSPS KA-
KENHI 25242073, and Grants for Excellent Graduate Schools,
MEXT, Japan. H.S. expresses his special thanks for the supports
from Global COE Programs of Osaka University.
[25] a) C. Lu, Y. Shen, J. Antibiot. 2003, 56, 415–418; b) E. Ohta, N. K. Kubota,
S. Ohta, M. Suzuki, T. Ogawa, A. Yamasaki, S. Ikegami, Tetrahedron 2001,
57, 8463–8467.
[26] a) C. Wolf, W. A. Kçnig, C. Roussel, Liebigs Ann. 1995, 781–786; b) F.
Leroux, ChemBioChem 2004, 5, 644–649.
[27] a) D. A. Evans, M. A. Calter, Tetrahedron Lett. 1993, 34, 6871–6874; b) K.
Toshima, T. Jyojima, H. Yamaguchi, Y. Noguchi, T. Yoshida, H. Murase, M.
Nakata, S. Matsumura, J. Org. Chem. 1997, 62, 3271–3284; c) K. A.
Scheidt, T. D. Bannister, A. Tasaka, M. D. Wendt, B. M. Savall, G. J. Fegley,
W. R. Roush, J. Am. Chem. Soc. 2002, 124, 6981–6990; d) S. Hanessian, J.
Ma, W. Wang, J. Am. Chem. Soc. 2001, 123, 10200–10206; e) F. Klein-
beck, G. J. Fettes, L. D. Fader, E. M. Carreira, Chem. Eur. J. 2012, 18,
3598–3610; f) J. A. Marshall, N. D. Adams, J. Org. Chem. 2002, 67, 733–
740; g) E. QuØron, R. Lett, Tetrahedron Lett. 2004, 45, 4539–4543; h) J. S.
Yadav, K. B. Reddy, G. Sabitha, Tetrahedron 2008, 64, 1971–1982.
[28] K. Toshima, T. Jyojima, N. Miyamoto, M. Katohno, M. Nakata, S. Matsu-
mura, J. Org. Chem. 2001, 66, 1708–1715.
[29] V. Rauhala, M. Nevalainen, A. M. P. Koskinen, Tetrahedron 2004, 60,
9199–9204.
[30] G. K. S. Prakash, R. Krishnamurti, G. A. Olah, J. Am. Chem. Soc. 1989, 111,
393–395.
[31] D. A. Evans, K. T. Chapman, E. M. Carreira, J. Am. Chem. Soc. 1988, 110,
3560–3578.
[32] D. A. Evans, A. H. Hoveyda, J. Am. Chem. Soc. 1990, 112, 6447–6449.
[33] The C23 configuration of 12 was determined by NOE experiments of its
p-methoxybenzylideneacetal derivative 12’.
Keywords: bafilomycin · chemical synthesis · fluorine label ·
solid-state NMR · V-ATPase
[1] Y. Ohsumi, Y. Anraku, J. Biol. Chem. 1981, 256, 2079–2082.
[2] M. Forgac, Nat. Rev. Mol. Cell Biol. 2007, 8, 917–929.
[3] B. Ma, Y. Xiang, L. An, Cell. Signalling 2011, 23, 1244–1256.
[4] A. Qin, T. S. Cheng, N. J. Pavlos, Z. Lin, K. R. Dai, M. H. Zheng, Int. J. Bio-
chem. Cell Biol. 2012, 44, 1422–1435.
[5] C. Farina, S. Gagliardi, Drug Discovery Today 1999, 4, 163–172.
[6] C. Farina, S. Gagliardi, Curr. Pharm..Des. 2002, 8, 2033–2048.
[7] H. Izumi, T. Torigoe, H. Ishiguchi, H. Uramoto, Y. Yoshida, M. Tanabe, T.
Ise, T. Murakami, T. Yoshida, M. Nomoto, K. Kohno, Cancer Treat. Rev.
2003, 29, 541–549.
[8] a) G. Werner, H. Hagenmaier, K. Albert, H. Kohlshorn, Tetrahedron Lett.
1983, 24, 5193–5196; b) G. Werner, H. Hagenmaier, H. Drautz, A. Baum-
gartner, H. Zähner, J. Antibiot. 1984, 37, 110–117.
[9] E. J. Bowman, A. Siebers, K. Altendorf, Proc. Natl. Acad. Sci. USA 1988,
85, 7972–7976.
[34] K. Takai, K. Nitta, K. Utimoto, Tetrahedron Lett. 1988, 29, 5263–5266.
[10] H. Hanada, Y. Moriyama, M. Maeda, M. Futai, Biochem. Biophys. Res.
Commun. 1990, 170, 873–878.
[11] J. P. Mattsson, D. J. Keeling, Biochim. Biophys. Acta Biomembr. 1996,
1280, 98–106.
[12] a) B. J. Bowman, E. J. Bowman, J. Biol. Chem. 2002, 277, 3965–3972;
b) E. J. Bowman, L. A. Graham, T. H. Stevens, B. J. Bowman, J. Biol. Chem.
2004, 279, 33131–33138; c) B. J. Bowman, M. E. McCall, R. Baertsch, E. J.
Bowman, J. Biol. Chem. 2006, 281, 31885–31893.
[13] a) M. Huss, G. Ingenhorst, S. Kçnig, M. Gaßel, S. Drçse, A. Zeeck, K. Al-
tendorf, H. Wieczorek, J. Biol. Chem. 2002, 277, 40544–40548; b) C. Os-
teresch, T. Bender, S. Grond, P. V. Zezschwitz, B. Kunze, R. Jansen, M.
Huss, H. Wieczorek, J. Biol. Chem. 2012, 287, 31866–31876.
[14] S. Benlekbir, S. A. Bueler, J. L. Rubinstein, Nat. Struct. Mol. Biol. 2012, 19,
1356–1363.
[15] a) M. Sagermann, T. H. Stevens, B. W. Matthews, Proc. Natl. Acad. Sci.
USA 2001, 98, 7134–7139; b) O. Drory, F. Frolow, N. Nelson, EMBO Rep.
2004, 5, 1148–1152; c) R. A. Oot, L. S. Huang, E. A. Berry, S. Wilkens,
Structure 2012, 20, 1881–1892; d) S. Basak, J. Lim, M. S. Manimekalai,
A. M. Balakrishna, G. Grüber, J. Biol. Chem. 2013, 288, 11930–11939.
[16] M. J. Duer, Introduction to Solid-State NMR, Blackwell Science Oxford,
2004, pp. 349.
[17] T. Gullion, J. Schaefer, J. Magn. Reson. 1989, 81, 196–200.
[18] S. D. Cady, K. S-Rohr, J. Wang, C. S. Soto, W. F. DeGrado, M. Hong, Nature
2010, 463, 689–693.
[19] a) S. J. Kim, L. Cegelski, D. R. Studelska, R. D. O’Connor, A. K. Mehta, J.
Schaefer, Biochemistry 2002, 41, 6967–6977; b) A. K. Mehta, L. Cegelski,
R. D. O’Connor, J. Schaefer, J. Magn. Reson. 2003, 163, 182–187; c) L. Ce-
[35] A. B. Smith, Y. Qiu, D. R. Jones, K. Kobayashi, J. Am. Chem. Soc. 1995,
117, 12011–12012.
[36] I. Paterson, S. B. Blakey, C. J. Cowden, Tetrahedron Lett. 2002, 43, 6005–
6008.
[37] E. Negishi, D. E. V. Horn, T. Yoshida, J. Am. Chem. Soc. 1985, 107, 6639–
6647.
[38] J. Inanaga, K. Hirata, H. Saeki, T. Katsuki, M. Yamaguchi, Bull. Chem. Soc.
Jpn. 1979, 52, 1989–1993.
[39] This protective group is critical for both the selective removal of MMTr
group and the final deprotection. See ref. 27b and references therein.
[40] The configuration at C17 and C18 of the major product 31 was deter-
mined by analogy with the known derivatives in references 24b and
27b, See Supporting Information for details.
[41] See Supporting Information for details. Synthesis of bioactive demethy-
lated aplysiatoxins was extensively examined by Irie’s group; Y. Nakaga-
wa, R. C. Yanagita, N. Hamada, A. Murakami, H. Takahashi, N. Saito, H.
Nagai, K. Irie, J. Am. Chem. Soc. 2009, 131, 7573–7579.
Chem. Asian J. 2015, 10, 915 – 924
www.chemasianj.org
923
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[42] G. E. Keck, K. H. Tarbet, L. S. Geraci, J. Am. Chem. Soc. 1993, 115, 8467–
8468.
[48] J. J. Shacka, B. J. Klocke, M. Shibata, Y. Uchiyama, G. Datta, R. E. Schmidt,
K. A. Roth, Mol. Pharmacol. 2006, 69, 1125–1136.
[43] I. Ohtani, T. Kusumi, Y. Kashman, H. Kakisawa, J. Am. Chem. Soc. 1991,
113, 4092–4096.
[49] M. J. Geisow, G. H. Beaven, P. D. Hart, M. R. Young, Exp. Cell Res. 1980,
126, 159–165.
[44] The configuration at C17 and C18 of the major product desmethyl-31
was determined to be the same as reported in ref. 40, see the Support-
ing Information for details.
[45] S. Kazami, M. Muroi, M. Kawatani, T. Kubota, T. Usui, J. Kobayashi, H.
Osada, Biosci. Biotechnol. Biochem. 2006, 70, 1364–1370.
[46] E. Uchida, Y. Ohsumi, Y. Anraku, J. Biol. Chem. 1985, 260, 1090–1095.
[47] See the Supporting Information for 1H chemical shifts data of Baf and
fluorinated analogues.
[50] T. Yoshimori, A. Yamamoto, Y. Moriyama, M. Futai, Y. Tashiro, J. Biol.
Chem. 1991, 266, 17707–17712.
[51] G. H. Baker, P. J. Brown, R. J. J. Dorgan, J. R. Everett, S. V. Ley, A. M. Z.
Slawin, D. J. Williams, Tetrahedron Lett. 1987, 28, 5565–5568.
Received: November 13, 2014
Published online on January 21, 2015
Chem. Asian J. 2015, 10, 915 – 924
www.chemasianj.org
924
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim