A challenging synthesis of the highly functionalized echinocandin ASP9726: A successor of micafungin

Article abstract of DOI:10.1021/op500078y

Here, we describe a practical, scalable, and challenging synthesis of the highly functionalized novel echinocandin ASP9726 (1) starting from the natural product FR901379 (3), which is a starting material of micafungin (2). The synthesis includes transformations that address significant synthetic challenges due to the need to control the chemoselectivity of the reactions during modification of the highly functionalized peptide core. In the present study, we discovered an efficient, high-yielding route to ASP9726 (1) that is suitable for large-scale production. Namely, dehydration of carboxamide (14) to nitrile (15) was accomplished by use of EDC·HCl with pyridine. Further, the transformation of nitrile (15) to primary amine (17) was conducted via hydrogenation with Sponge Nickel catalyst without decomposition, followed by one-pot debenzylation with Pd/C. Reductive amination between primary amine (17) with dihydroxyacetone (DHA) was accomplished using 2-picoline/borane complex as a reducing agent in MeOH, yielding 66.6 kg of peptide core unit (18). After the C₁₅H₃₁ chain cleavage by bioconversion, reductive amination between the core peptide unit (4) and side chain (10) was achieved in high yield by making use of tert-butyl amine/borane complex as a reducing agent. Consequently, highly pure ASP9726 (1) was obtained in a practical manner without using silica gel or ODS column chromatography purification in any step. Overall yield was drastically increased from 0.71% to 13.8% compared to that of the prior synthetic method.

Full text of DOI:10.1021/op500078y

Article
pubs.acs.org/OPRD
A Challenging Synthesis of the Highly Functionalized Echinocandin
ASP9726: A Successor of Micafungin
Shinya Yoshida,*,† Joji Hayashida,† Yasuhiro Morinaga,† Shoji Mizobata,† Akihiro Okada,‡
Kazumi Kawai,§ Shinjirou Tanoue,§ Tomohide Nakata,† Minoru Kitayama,§ Atsushi Ohigashi,†
Mitsutaka Matsuura,§ Takumi Takahashi,∥ Shigeru Ieda,† and Minoru Okada†
†Process Chemistry Labs., Astellas Pharma Inc., 160-2 Akahama, Takahagi-shi, Ibaraki 318-0001, Japan
‡Drug Discovery Research, Astellas Pharma Inc., 21, Miyukigaoka, Tsukuba-shi, Ibaraki 305-8585, Japan
§Fermentation and Biotechnology Labs., Astellas Pharma Inc., 5-2-3, Tokodai, Tsukuba-shi, Ibaraki 300-2698, Japan
∥Technology Product Management, Astellas Pharma Inc., 2-5-1, Nihonbashi-Honcho, Chuo-ku, Tokyo, 103-8411, Japan
S
* Supporting Information
ABSTRACT: Here, we describe a practical, scalable, and challenging synthesis of the highly functionalized novel echinocandin
ASP9726 (1) starting from the natural product FR901379 (3), which is a starting material of micafungin (2). The synthesis
includes transformations that address significant synthetic challenges due to the need to control the chemoselectivity of the
reactions during modification of the highly functionalized peptide core. In the present study, we discovered an efficient, high-
yielding route to ASP9726 (1) that is suitable for large-scale production. Namely, dehydration of carboxamide (14) to nitrile
(15) was accomplished by use of EDC·HCl with pyridine. Further, the transformation of nitrile (15) to primary amine (17) was
conducted via hydrogenation with Sponge Nickel catalyst without decomposition, followed by one-pot debenzylation with Pd/C.
Reductive amination between primary amine (17) with dihydroxyacetone (DHA) was accomplished using 2-picoline/borane
complex as a reducing agent in MeOH, yielding 66.6 kg of peptide core unit (18). After the C15H31 chain cleavage by
bioconversion, reductive amination between the core peptide unit (4) and side chain (10) was achieved in high yield by making
use of tert-butyl amine/borane complex as a reducing agent. Consequently, highly pure ASP9726 (1) was obtained in a practical
manner without using silica gel or ODS column chromatography purification in any step. Overall yield was drastically increased
from 0.71% to 13.8% compared to that of the prior synthetic method.
1). However, we encountered several drawbacks to this
medicinal chemistry method, with issues summarized below.
1.1. Issues of the Medicinal Chemistry Method.
1. INTRODUCTION
ASP9726 (1; Figure 1) is a novel echinocandin with potent
Aspergillus hyphal growth inhibition and significantly improved
minimum inhibitory concentrations against Candida parapsilosis
and echinocandin-resistant Candida compared to caspofungin.
Further, this compound 1 is expected to be a successor of
micafungin (2; Figure 1), which was launched by Astellas
Pharma Inc. in 2002 and received approval from the U.S. Food
and Drug Administration in 2005. ASP9726 (1) has shown
potent efficacy in treating systemic candidiasis and aspergillosis
with no concerns of side effects.1 For the preparation of 1,
some chemical modifications of the peptide core are required to
accomplish the synthesis. These transformations present
significant synthetic challenges due to the need to control the
chemoselectivity of the reactions during modification of the
highly functionalized peptide core. Chemical reactions with the
natural product FR901379 (3)1f,2 which is a starting material of
micafungin1 and its derivatives also present numerous
processing issues as a result of physical characteristics that are
inherent to this class of lipopeptides (e.g., poor solubility in
organic solvents, micellar and soap-like behavior, hygroscopicity
in the solid state, and instability to both strong acid and strong
base). Under these difficult circumstances, in the medicinal
chemistry synthetic method, final product 1 was prepared in a
lab-scale synthesis starting from the natural product 3 (Scheme
• Low yield in steps a, b, f, l, and n; further, overall yield
was only 0.71%.
• Inefficient protection and deprotection steps were
required.
• ODS column chromatography was necessary in all steps
that should be avoided in large-scale synthesis because it
requires large amounts of solvent. Actually it is
impossible to prepare 100 kg-scale of 1 in this manner.
• Use of 1,4-dioxane, which represents an environmental
hazard.
• Use of NaBH3CN as a reducing agent in steps k and n,
which is toxic and has a sluggish rate of reaction.
• Final product 1 was unstable, presenting difficulties
during operations including aging of the slurry, filtration,
and drying.
Therefore, a practical and efficient new synthetic method for
the large-scale synthesis of compound 1 was required.
Received: March 7, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Figure 1. Structure of ASP9726 (1) and micafungin (2).
a
Scheme 1. Medicinal synthetic method for ASP9726 (1)
a
Reagents and conditions: (a) benzyl chloroformate, THF, pH 6.86 standard buffer solution, ODS column chromatography, 68%; (b) Et3SiH, TFA,
CH2Cl2, ODS column chromatography, 45%; (c) H2, Pd/C, H2O, ODS column chromatography, 73%; (d) (Boc)2O, NaOH, H2O, 1,4-dioxane,
ODS column chromatography, 88%; (e) BnBr, LiOH-H2O, DMF, ODS column chromatography, 85%; (f) MsCl, NaHCO3, i-Pr2NEt, zeolite, DMF,
ODS column chromatography, 41%; (g) 10% HCl−MeOH, MeOH, ODS column chromatography; 86%; (h) MeI, LiOH·H2O, DMF, ODS column
chromatography, 76%; (i) H2, Pd/C, MeOH, ODS column chromatography, quant.; (j) NaBH4, CoCl2-6H2O, MeOH, H2O, ODS column
chromatography, 85%; (k) dihydroxyacetone (DHA), NaBH3CN, AcOH, MeOH, ODS column chromatography, 73%; (l) Fmoc-Cl, i-Pr2NEt,
DMF, ODS column chromatography, 46%; (m) TFA, Et3SiH, CH2Cl2, ODS column chromatography, 83%; (n) 10, NaBH3CN, AcOH, MeOH,
DMF, CHCl3, then piperidine, ODS column chromatography; 67%. Overall yield: 0.71%.
B
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 2. New synthetic approach of 1
between primary amine (17) with dihydroxyacetone (DHA)
was accomplished using 2-picoline/borane complex as a
reducing agent in MeOH, yielding 66.6 kg of peptide core
unit (18). After the C15H31 chain cleavage by bioconversion,
reductive amination between the core peptide unit (4) and side
chain (10) which was prepared in nine steps3 was accomplished
in high yield with tert-butyl amine/borane complex as a
reducing agent. Consequently, highly pure ASP9726 (1) was
obtained in a practical manner without using silica gel or ODS
column chromatography purification in any step. Overall yield
was drastically increased from 0.71% to 13.8% compared to that
of the prior synthesis.
2.1.1. Selective Reduction of Benzylalcohol and Aminal
with Silane Reducing Agent. The selective silane reduction of
aminal moiety and benzyl alcohol moiety was examined. As
reducing agents, NaBH4 or silane reagents4 such as Ph3SiH,
Et3SiH, TBDMSiH were tested with AcOH or TFA. We found
that the combination of TBDMSiH and TFA gave the best
reaction profile, and the desired 11 was obtained in a high yield
2. RESULTS AND DISCUSSION
2.1. New Synthetic Strategy and Brief Summary. To
resolve many of the significant issues with the medicinal
chemistry method, we considered a new synthetic strategy for
the preparation of 1 (Scheme 2).
The synthesis of ASP9726 (1) from FR901379 (3) requires
deacylation of the C15H31 chain by bioconversion and the
installation of a side chain (10). Further, we considered that
this C15H31 side chain might potentially be useful as a
protection group of the primary amine because this strategy
does not need inefficient protection and deprotection steps. In
this strategy, we discovered an efficient, high-yielding route to
ASP9726 (1) that is suitable for large-scale production. The
synthesis includes dehydration of carboxamide (14) to nitrile
(15) that was accomplished using EDC·HCl with pyridine
without decomposition. Further, the transformation of nitrile
(15) to primary amine (17) was conducted via hydrogenation
with Sponge Nickel catalyst without degradation, followed by
one-pot debenzylation with Pd/C. Reductive amination
C
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Figure 2. Structure of retro-aldol impurity 19.
(>95%). The highly reproducible and robust method of
selective reduction of benzylalcohol and aminal with a silane
reducing agent was feasible for large-scale synthesis, producing
160 kg of crude 11. After purification with DIAION HP20SS,5
11 was used in the next benzylation step.
core, and a stepwise reaction might be more appropriate for
large-scale synthesis of the highly functionalized core unit.
2.1.4. Stepwise Approach to Primary Amine (16). There-
fore, a more effective two-step procedure was developed
through dehydration to the nitrile followed by a catalytic
hydrogenation. In the medicinal chemistry method, dehydra-
tion was conducted by using MsCl, NaHCO3, and i-Pr2NEt,
resulting in 39% yield (Table 2, entry 1). This condition was
2.1.2. Benzylation, Desulfurization and Methylation.
Following telescoping benzylation with benzyl bromide and
LiOH, desulfurization under acidic conditions and further
methylation with methyl iodide and LiOH were accomplished
with good yield (86.8% in three reactions). LiOH gave the best
reaction profile for benzylation and methylation in comparison
with any other bases such as KOH or NaOH because of its
improved solubility in DMF. However, a large amount of LiOH
produced impurities derived from retro-aldol threonine moiety
(Figure 2), as identified by LC−MS. Therefore, determining
the appropriate amount of LiOH was a critical factor for this
reaction, and 1.2 equiv of LiOH was ultimately selected as the
optimum amount for the preparation of 14 with minimization
of the formation of 19.
Table 2. Dehydration of Amide 14 to Nitrile 15
a
results
14/15/
others
entry
1
reagents
conditions
MsCl ( × 1.1 mol), NaHCO3
( × 1.1 mol), i-Pr2NEt ( × 1.1 mol),
zeolite/DMF
cyanuric chloride ( × 2.5 mol)/DMF6
medicinal
method
5/56/39
2
−30 °C to
2/67/31
rt
3
4
cyanuric chloride ( × 2.5 mol)/DMF6
−10 °C
6/64/30
cyanuric chloride ( × 2.5 mol)/DMF,
−30 °C to
no reaction
water (250 mol %)5
rt
2.1.3. Direct Reduction of Amide to Primary Amine. The
direct reduction of amide 14 to primary amine 16 requires
chemoselective reduction of the primary amide in the presence
of other amides and hydroxyl groups. The borane complex,
ionic aluminum, and boron hydride are the most common
reducing agents for carboxamide; however, other amides are
generally reduced more readily than primary amides with these
reducing agents. Despite the potential drawbacks, we
investigated the direct reduction of an amide to a primary
amine (Table 1). As predicted, these reaction conditions did
not produce a good reaction profile. Many unknown impurities
appeared, such as over-reduced impurities during the reaction,
even though the reaction was conducted at low temperature.
The combination of BSTFA and BH3·THF complex did not
work well in this reaction.6 Given these findings, we determined
that direct reduction would not be suitable for this peptide
5
cyanuric chloride ( × 2.5 mol)/NMP6
Vilsmeier reagent ( × 2.5 mol)/DMF
Vilsmeier reagent ( × 2.5 mol)/NMP
dichlorotriazine ( × 2.5 mol)/DMF
−30 °C to
no reaction
2/65/33
rt
6
−30 °C to
rt
7
−30 °C to
5/64/31
rt
8
−30 °C to
70/25/5
rt
9
Vilsmeier reagent ( × 2.2 mol), pyridine −30 °C to
(× 4.4 mol)/DMF rt
78/10/12
75/11/14
no reaction
no reaction
10
11
12
Vilsmeier reagent ( × 2.2 mol), NaHCO3 −30 °C to
(× 4.4 mol)/DMF
rt
2-chloro-4,6-dimethoxy-1,3,5-triazine
(CDMT, × 2.5 mol)/DMF
1-chloro-N,N,2-trimethyl-1-
propenylamine (Ghosez reagent, × 2.5
mol)/DMF
−30 °C to
rt
−30 °C to
rt
13
(COCl)2( × 3.0 mol), pyridine ( × 6.0 −30 °C to
mol)/DMF rt
11/58/31
a
a
HPLC method B (see Experimental Section).
Table 1. Direct reduction of amide to primary amine
b
entry
conditions
results
problematic for large-scale synthesis, since 14 and product 15
decomposed during the reaction, resulting in a low yield. To
solve this critical issue, study of the dehydration conditions was
conducted (Table 2).
The methodology system of cyanuric chloride in DMF that
was used for the synthesis of caspofungin acetate6 gave a good
reaction profile. However, in this case, unacceptable levels of
unknown impurities were observed in the HPLC analysis
(entries 2 and 3). In contrast, unfortunately, the addition of 250
mol % of water (in situ hydroxy dichlorotriazine, 20, Figure 3)6
1
2
3
4
5
6
7
Red-Al/THF−toluene
DIBAL-H/THF
yield: 3%, complex mixture
yield: 5%, complex mixture
yield: 1%, complex mixture
yield: 1%, complex mixture
yield: 2%, complex mixture
yield: 16%, complex mixture
yield: 14%, complex mixture
LiAlH4/THF
BH3·THF complex/THF
BH3·S(CH3)2 complex/THF
CoCl2/NaBH4/THF
BSTFA then BH3·THF/THF6
a
b
Reaction was carried out at −30 to 25 °C. HPLC method B (see
Experimental Section).
D
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Figure 3. Proposed structures of chemically active species.
gave no reaction in this case (entry 4). Further, while reaction
was attempted with cyanuric chloride in NMP, the reaction
ultimately did not occur (entry 5). These results suggested that
chemically active species were generated between cyanuric
chloride and DMF. In contrast, Vilsmeier reagent in DMF or
NMP gave good conversion, but a number of unknown
impurities were observed in HPLC analysis (entries 6 and 7).
To improve the ratio of the formation of impurities, bases such
as pyridine or sodium bicarbonate were added to the reaction
mixture with Vilsmeier reagent. However, these additions
resulted in low reactivity for this reaction (entries 9 and 10). To
confirm the reactivity of this carboxamide 14, dichlorotriazine,
CDMT, and Ghosez reagent were evaluated in DMF solvent
(entries 8, 11, and 12). The use of dichlorotriazine gave
medium activity, which should have been higher than that of
hydroxy dichlorotriazine 20 (entry 4), whereas CDMT and
Ghosez reagent gave no reaction at all. Therefore, we
determined that the direct reaction of carboxamide with
cyanuric chloride was infeasible, and that the active source
would likely be either or both compound 21 and 22 (Figure 3).
Using a similar approach, in situ generation of Vilsmeier reagent
was tested, such as via oxalyl chloride and pyridine in DMF
(entry 13). However, this approach did not improve the
reaction profile, with nearly identical results in comparison with
the cyanuric chloride and Vilsmeier method. In conclusion,
these methodologies did not give a satisfactorily clean reaction,
required low temperature, and carried a risk of degradation in
large-scale synthesis.
Other coupling agents such as CIP (2-chloro-1,3-dimethylimi-
dazolidium hexafluorophosphate), TFFH (tetramethylfluoro-
formamidinium hexafluorophosphate), and DEPBT (3-(dieth-
oxyphosphoryloxy)-1,2,3-benzotriazin-4(3H)-one) were less
reactive than EDC·HCl (entries 2−4). Other conditions such
as HATU (1-[bis(dimethylamino)methylene]-1H-1,2,3-
triazolo[4,5-b]pyridinium 3-oxid hexafluorophosphate), PyBop
(benzotriazol-1-yl-oxytripyrrolidinophosphonium hexafluoro-
phosphate),7 DPPA (diphenylphosphoryl azide), CPMI (2-
chloro-1-methylpyridinium iodide), T3P (propylphosphonic
anhydride),8 iBuOC(O)Cl, and CDMT did not give the
desired nitrile 15 (entries 5−11).
To improve the yield of 15 with EDC·HCl, a base screening
test was attempted (Table 4). One of the weak and nonbulky
a
Table 4. Optimization of dehydration with EDC·HCl
b
entry
base
results 14/15 HPLC ratio
1
2
3
4
5
6
7
8
9
pyridine
2,6-lutidine
2,4,6-collidine
DMAP
61/39
75/25
82/18
97/3
Et3N
97/3
DIPEA
98/2
NMM
93/7
DBU
complex mixture
99/1
NaHCO3
a
Reaction temp: rt to 50 °C, reaction time: 4 h, base: × 28 mol,
b
solvent: NMP. HPLC area % ratio, HPLC method C (see
Experimental Section).
To improve the reaction profile for the formation of the
nitrile moiety, various kinds of dehydration agents were tested
for this reaction (Table 3). A mild coupling agent EDC·HCl
gave the best reaction profile, with a 72% yield (HPLC area) of
desired 15 and minimized formation of impurities (entry 1).
bases, pyridine, gave the best conversion among other bulky
bases such as 2,6-lutidine or 2,4,6-collidine (entries 2 and 3).
Other bases of different bulkiness did not work for this reaction
(entries 4−9).
a
Table 3. Screening of dehydration agents
To increase the yield of 15, we also attempted to use other
solvents such as THF and MeCN with EDC·HCl and pyridine.
However, no solvent improved the yield of 15. From the point
of view of large-scale synthesis, the EDC·HCl method was
considered to be better than the Vilsmeier approach. After the
isolation of crude 15, purification was conducted using
SEPABEADS SP207SS.5
2.1.5. Catalytic Hydrogenation of Nitrile to Amine. In the
next step, the condition of catalytic hydrogenation was
investigated. First, catalyst screening was attempted (Table
5). Results showed that CoCl2/NaBH4 in MeOH gave
unacceptable levels of impurities on HPLC analysis. In
addition, desired product was gradually decomposed in the
reaction mixture; therefore, an ODS column was necessary for
purification of 15 (entry 1). Further, Pt/C gave the undesired
hydrolysis of nitrile to carboxamide without reducing nitrile
(entry 2). The use of Rh on Alumina9 gave undesirable results
with the formation of unacceptable levels of impurities, and the
b
results 14/15/
entry
reagents
conditions
others
1
2
EDC·HCl ( × 10 mol)
CIP ( × 10 mol)
rt to 50 °C, 27 h
rt to 50 °C, 27 h
rt to 50 °C, 24 h
rt to 50 °C, 24 h
rt to 50 °C, 24 h
rt to 50 °C, 24 h
rt to 50 °C, 24 h
rt to 50 °C, 24 h
rt to 50 °C, 24 h
rt to 50 °C, 24 h
2/72/26
56/27/17
31/10/59
c
c
3
TFFH ( × 10 mol)
DEPBT ( × 10 mol)
HATU ( × 10 mol)
PyBop6 ( × 10 mol)
DPPA ( × 10 mol)
CMPI ( × 10 mol)
T3P7 ( × 10 mol)
c
4
73/2/25
5
no reaction
no reaction
no reaction
no reaction
no reaction
no reaction
6
7
8
9
10
iBuOC(O)Cl ( ×
10 mol)
11
CDMT ( × 10 mol)
rt to 50 °C, 24 h
no reaction
a
b
Pyridine ( × 28 mol), NMP ( × 7 vol/wt). HPLC area % ratio,
HPLC method C (see Experimental Section). Reaction stopped even
c
though there remained further reaction time.
E
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
a
reaction temperature gave the best reaction profile, with the
reaction completing within 5 h. After reaction completion,
amine compound 16 was isolated as an amorphous solid.
Unfortunately, critical problems were encountered during
filtration because of the amorphous-like character of amine
16, which resulted in a slow filtration speed and deliquescent
powder. Even after attempts to isolate amine 16 as a
hydrochloric acid salt, the condition of the filtrate was not
improved. To resolve these issues, we attempted one-pot
reduction of nitrile 15 to amine 16 following deprotection of
the benzyl group to afford 17. After the completion of the
nitrile reduction with Ni catalyst, Pd/C catalyst was added to
the reaction mixture and the mixture was aged under 30 psi
hydrogen pressure for 1 h. After completion of the
debenzylation, the desired debenzylated amine 17 was isolated
as an HCl salt. The condition of the filtrate was better than that
of 16, it gave the faster rate during the filtration. This highly
practical and reproducible method was successfully demon-
strated in large-scale synthesis and yielded 63.0 kg of
intermediate 17·HCl.
2.1.7. Reductive Amination with Dihydroxyacetone. In the
medicinal chemistry method, reductive amination of the
primary amine with dihydroxyacetone was conducted with
NaBH3CN as a reducing agent. However, the use of NaBH3CN
should be avoided due to safety concerns. Further, ODS
column chromatography purification also should be avoided
because the point of view of large-scale operation. To resolve
these issues, the medicinal group investigated improved
conditions using NaBH4 in DMF. Unfortunately, this method-
ology encountered fatal issues in scale-up synthesis. HPLC
chromatograms of this reaction are shown in Figure 5.
Table 5. Hydrogenation of nitrile to amine
b
entry
1
catalysts
solvents
results
1f5
CoCl2/NaBH4
MeOH/water
= 5/1
85% yield, ODS column
purification was necessary
2
Pt/C
EtOH/NH3 aq hydrolysis of nitrile occurred,
= 10/1 48% yield.
3
5% Rh on
alumina8
EtOH/NH3 aq 15/16=22/48, yield 45%
= 10/1
4
Pd/C (K-type)
EtOH/NH3 aq main product: 23
= 10/1
5
Pd/C (P-type)
EtOH/NH3 aq main product: 23
= 10/1
6
Pd/C (NX-type) EtOH/NH3 aq main product: 23
= 10/1
7
Pd/C (STD-
type)
EtOH/NH3 aq main product: 23
= 10/1
8
Sponge Cobalt
OFT-55
EtOH/NH3 aq 15/16 = 82/18
= 10/1
9
Sponge Cobalt
OFT-MS
EtOH/NH3 aq 15/16 = 77/23
= 10/1
10
11
12
13
Sponge Cobalt
ODHT-60
EtOH/NH3 aq 15/16 = 85/15
= 10/1
Sponge Nickel
NDT-90
EtOH/NH3 aq 15/16 = 34/66
= 10/1
Sponge Nickel
NDHT-90
EtOH/NH3 aq 15/16 = 4/96
= 10/1
Sponge Nickel
DHT-90M
EtOH/NH3 aq 15/16 = 37/63
= 10/1
a
b
Hydrogen pressure: 43−58 psi, 40 °C, 8−9 h. HPLC area% ratio,
HPLC method D (see Experimental Section).
yield of the desired compound was poor (less than 50%; entry
3). Screening of various kinds of Pd/C catalysts10 did not give
the desired reaction, as the main product was debenzylated
undesired compound 23 without the nitrile moiety (entries 4−
7, Figure 4). To overcome these issues, Sponge Cobalt catalysts
Figure 4. Structure of debenzylated impurity 23.
and Sponge Nickel catalysts were evaluated (entries 8−13).11
Surprisingly, one of the nickel catalysts, NDHT-90, allowed the
reaction to progress markedly cleanly, and the yield was up to
90% (entry 12).
Figure 5. HPLC chromatograms of reductive amination with
NaBH(OAc)3 in DMF (HPLC method E, see Experimental Section).
2.1.6. Optimization of the Reaction Condition. To improve
the yield of amine 16, optimization of reaction conditions, such
as reaction temperature, hydrogen pressure (30 to 100 psi), and
the concentration of aqueous NH3 in EtOH were investigated.
Reaction pressure with hydrogen was not found to be an
important factor; in contrast, reaction temperature and the
concentration of aqueous NH3 were found to be critical factors.
An almost 1:1 ratio of EtOH and aqueous NH3 and a 50 °C
After 30 min at 0 °C, desired compound 18 appeared, and
the ratio of starting material 17 was decreased in HPLC
analysis. However, impurities were also detected in HPLC
analysis, and after 5 h, the target material 18 had almost entirely
decomposed, with a number of impurities observed. From these
F
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
results, we concluded that hydride reduction of imine was not
suitable for this type of poly peptide compound.
Table 7. Screening of reducing agents for reductive
amination
a
2.1.8. Reductive Amination with Catalytic Hydrogenation.
Since hydride reduction of imine was not feasible, we
investigated catalytic hydrogenation conditions (Table 6).
17/18/
b
entry
1
reducing agent
NaBH(OAc)3
others
comments
ND/80/20
excess amount of reagent was
needed for full conversion
a
2
3
4
pyridine borane
ND/90/10
ND/92/8
10/67/23
risk for the safety operation
best condition
Table 6. Reductive amination with catalytic hydrogenation
2-picoline borane
b
entry
catalyst ( × 1 wt)
17/18 (HPLC area % ratio)
5-ethyl-2-
not completed, and impurities
were observed
methylpyridine
1
Sponge Nickel NDT-90
Sponge Nickel NDHT-90
Sponge Nickel DHT-90M
Sponge Cobalt OFT-55
Sponge Cobalt OFT-MS
Sponge Cobalt ODHT-60
Pd/C PE-type
86/14
94/6
borane14
2
5
6
triethylamine
borane
23/50/27
33/41/26
not completed, and impurities
were observed
3
93/7
4
64/36
83/17
85/15
64/36
70/30
35/65
68/32
11/89
10/90
t-BuNH2 borane
not completed, and impurities
5
were observed
a
b
6
MeOH solvent; reaction temp: −5 to 25 °C. HPLC method E (see
7
Experimental Section).
8
Pd/C P-type
6907,16 in 45% yield. We then assessed the utility of the side-
chain unit installation method. First, SN2 substitution reaction
of the primary amine with side chains such as mesylate 24 and
triflate compound 25 were investigated (Scheme 3). However,
only a trace amount of desired product 1 was detected in
HPLC analysis (HPLC method G, see Experimental Section).
Given our findings, we concluded that reductive amination
with side-chain aldehyde 103 would be better for this peptide
core 4, as the reaction was able to be conducted under milder
conditions than the SN2 reaction. The medicinal chemistry
method used NaBH3CN as a reducing agent, which should be
avoided in large-scale synthesis from the point of view of the
safety operation (Table 8, entry 1). To accomplish the
complicated reductive amination between highly water-soluble
peptide core 3 and hydrophobic side chain 10, we screened
several reducing agents (Table 8).
As described in Table 8, NaBH3CN produced poor results
due to competitive reduction of the side-chain aldehyde to
alcohol (entry 1). Further, NaBH(OAc)3 also failed to reduce
the imine intermediate, and competitive reduction of side-chain
aldehyde to alcohol was also observed (entry 3). In contrast,
one of the alkyl amines, borane t-BuNH2·BH3 complex,
resulted in a clean reaction profile, and imine was reduced at
around 25 °C (entry 4). To optimize this reaction, we
compared AcOH to other acids such as TFA or formic acid.
TFA gave the best result among the examined acids. Further,
base screening was also conducted to improve the yield of
desired 1 and minimize the formation of dialkylated impurities.
A number of bases were assessed, including TEA, N,N-
diisopropylethylamine (DIPEA), pyridine, N-methyl morpho-
line (NMM), and sodium bicarbonate. The weak base, N-
methyl morpholine, gave the best results, with relatively little
formation of impurities. Using the above acid and base, yield of
1 was improved from 61% to more than 90% due to the
minimization of the formation of dimeric impurities. Since
peptide core 4 was isolated as a free base instead of
dihydrochloric acid salt in the second-generation synthesis,
we evaluated the optimal amount of acid, finding that the use of
1.65 equiv of HCl was required to minimize the formation of
dimeric impurities.
9
Pd/C K-type
10
11
12
Pd/C STD-type
Pd/C NX-type
Pt/C
a
DHA (5 equiv), hydrogen pressure: 30 psi, 40 °C, 8−9 h; solvent:
b
NMP/MeOH = 1/1, 30 °C, 7 h. HPLC method E (see Experimental
Section).
Hydrogenation with Pt/C was selected for the reductive
amination with dihydroacetone (entry 12) but failed to provide
reproducible results, even when the reaction was conducted
under the same conditions with the same reactor. Actually, in
an 80-kg scale synthesis, reductive amination stopped at almost
half conversion. In addition, further addition of Pt/C catalyst
and DHA gave decompositions of both starting material and
target compound without proceeding to the desired reaction. It
was a critical issue for scale-up synthesis. At that time, the
desired compound was recovered by temporary treatment;
however, the yield was quite low. According to the
consumption of hydrogen gas during the reaction, there was
a possibility that DHA could be reduced to the triol impurity
without reducing the imine intermediate. To resolve this critical
issue, we focused on the character of the dihydroxyacetone and
its dimer.12 Although, the reaction was conducted with
commercially available DHA monomer or dimer of DHA
under the same conditions to prevent the hydrogenation of
DHA, no differences were noted between results of these
reactions. Subsequently, we attempted the reaction with a mild
reducing agent in the presence of a protic solvent such as
MeOH, which favors imine formation (Table 7).
As we expected, excess amount of NaBH(OAc)3 was
required in MeOH solvent; however, the desired product was
obtained without decomposition. On the contrary, an alkyl
amine/borane complex such as 5-ethyl-2-methylpyridine/
borane,13 triethylamine/borane, or t-BuNH2/borane did not
give a good reaction profile (entries 4−6). Fortunately, the
borane complex with a relatively weak base such as the
pyridine/borane complex14 or the 2-picoline/borane complex15
in MeOH gave excellent results; the desired compound was
obtained without decomposition in good yield, more than 90%
as an HCl salt. From the point of view of safety operation, 2-
picoline/borane complex was selected as a reducing agent, and
the highly efficient and reproducible procedure was demon-
strated in a large-scale synthesis yielding 66.6 kg of 18·HCl.
2.1.9. Installation of Side Chain. Deacylation of C15H31
moiety was accomplished using acylase Streptomyces sp. no.
2.1.10. Purification Method. In the next stage, we
investigated purification methods that did not use ODS column
chromatography. Many kinds of synthetic adsorbents that could
be used repeatedly after regeneration, such as SEPABEADS,5
SP207SS,5 DIAION HP20SS,5 and DIAION HP20,16 were
tested, with DIAION HP20SS found to be the most efficient
G
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Scheme 3. Installation of side chain
a
Table 8. Reductive amination for side-chain installation
b
entry
reducing agents
HPLC yield of 1 (%)
1
2
NaBH3CN × 2.5 mol, AcOH × 3.2 mol, Et3N × 2.0 mol
61%; competitive reduction of side-chain aldehyde observed
82%
2-picoline/borane × 2.5 mol, AcOH × 3.2 mol, Et3N ×
2.0 mol
3
4
NaBH(OAc)3 × 2.5 mol, AcOH × 3.2 mol, Et3N ×
1%; imine was not reduced and competitive reduction, aldehyde to alcohol of side-chain aldehyde
observed
2.0 mol
t-BuNH2·BH3 × 2.5 mol, AcOH × 3.2 mol, Et3N ×
92%
2.0 mol
a
b
Reaction was carried out at −5 to 50 °C, solvent: DMF/CHCl3/MeOH, HPLC yield was determined in comparison with standard sample (HPLC
method E, see Experimental Section).
resin synthetic adsorbent. In addition, we found that the most
efficient method of regenerating the DIAION HP20SS5 was:
wash with (1) acetone, (2) water, (3) 1:1 acetone:1 M aqueous
NaOH, (4) water, (5) 0.1 M aqueous H2SO4, (6) water, (7)
acetone, (8) 0.1 M aqueous NaOH, (9) water, and finally (10)
0.5 M aqueous NaHCO3.
We also conducted a stability test at 40 °C/75% RH for the
various kinds of samples that have a different molar ration of
HCl (Figure 7). Stability of 1 was dramatically improved with a
decrease in additional amount of HCl.
2.1.11. Improvement of Stability of Final Product. We
encountered a significant issue concerning the stability of the
final product 1, which gradually decomposed during stirring
slurry, drying, and storage. In the early stage of development, 1
was isolated as a HCl salt. However, the element analysis of the
final product indicated that these samples had 2.5−2.7 equiv of
HCl even if 2 equiv of HCl was added against the result of assay
with HPLC analysis. We suspected that the additional HCl
might be the cause of the instability of 1. To resolve this critical
issue, the adequate pH range to prepare 2.0 equiv of HCl salt
was investigated. The relationship between pH and HCl
concentration of 1 is shown in Figure 6. To obtain the desired
2.0 equiv of HCl salt, the pH should be prepared between 5.0
to 6.0, with a target point of 5.5.
Figure 7. Results of the stability test at 40 °C/75% RH.
Further, for the preparation of preclinical samples, the final
product was isolated using large amounts of acetone (vol/wt)
and EtOH aq (vol/wt). However, this method required long
filtration times (15 min/2 g-scale), and the wet cake was
difficult to transfer to a dryer because of the muddy character.
To resolve these issues, we evaluated methods of precipitation.
EtOAc was added to a solution of ASP9726 in EtOH−water,
which resulted in drastically improved filtration speed (from 15
to 0.5 min in 2-g scale synthesis). SEM images of these
obtained solids are shown in Figure 8. Agglutinated powder was
observed on the right SEM picture that should be suitable for
the filtration.
3. CONCLUSION
A challenging synthesis of the highly functionalized echino-
candin ASP9726 was accomplished. A novel echinocandin,
ASP9726 (1), was prepared in an efficient and scalable method
from fermentation product 3, in an overall yield that was
Figure 6. Relationship between pH and amount of HCl in ASP9726
(1).
H
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Figure 8. SEM images of ASP9726. (Left) A solution of ASP9726/EtOH−water was added to acetone (3000×). (Right) EtOAc was added to a
solution of ASP9726 in EtOH−water (3000×).
dramatically improved in comparison with the medicinal
chemistry method (13.8% vs 0.71%). This synthesis features
complex and unstable substrates reacting with appropriate
reagents and conditions that can provide chemoselective
reactions. The formation of primary amine was achieved by
the dehydration of nitrile with EDC·HCl in pyridine followed
by hydrogenation with Sponge Ni (NDHT-90). The reductive
amination of primary amine with dihydroxyacetone was
accomplished by using 2-picoline/borane in MeOH. Further,
reductive amination of core peptide 4 with side chain 10 was
achieved using of t-BuNH2/borane complex. In this study,
practical purification and precipitation methods were devel-
oped, and pure final product 1 was obtained.
with UV detection at 210 nm; 3: 5.1 min; mono reduced
intermediate: 6.6 min, 11, 7.6 min.
HPLC Method B. Column: YMC-Pack ODS-AM, 5 μm, 4.6
mm × 150 mm I.D. (YMC), elution with 20 mM phosphate
buffer (KH2PO4: 1.36 g + Na2HPO4·12H2O 3.58g + Sodium 1-
decanesulfonate: 1.22 g/water 1000 mL, adjusted pH 6.86 with
aq H3PO4)/CH3CN (4/6), over 30 min, 1.0 mL/min, at 35 °C,
with UV detection at 210 nm; DMF: 1.8 min, 11: 3.4 min; 12,
4.2 min, 13: 9.9 min, 14: 11.9 min.
HPLC Method C. Column: YMC-Pack ODS-AM, 5 μm, 4.6
mm × 150 mm I.D. (YMC), elution with 10 mM K2HPO4
buffer (adjusted 6.86 with aq H3PO4)/CH3CN (2/8), over 30
min, 1.0 mL/min, at 35 °C, with UV detection at 210 nm;
DMF: 1.8 min, 14: 3.5 min; 15: 5.1 min.
HPLC Method D. Column: YMC-Pack ODS-AM, 5 μm, 4.6
mm × 150 mm I.D. (YMC), elution with phosphate buffer
(KH2PO4: 1.36 g + Na2HPO4·12H2O 3.58 g + sodium 1-
decane sulfonate: 1.22 g/water 1000 mL, adjusted pH 6.86 with
aq H3PO4)/CH3CN (4/6), over 40 min, 1.0 mL/min, at 35 °C,
with UV detection at 210 nm; DMF: 1.8 min, 15: 18.0 min; 16:
6.4 min, 17: 3.5 min.
HPLC Method E. Column: Unison UK-C18, 3 μm, 4.6 mm
I.D. × 250 mm (Imtakt), elution with 10 mM K2HPO4 buffer
(adjusted pH 7.5 with aq H3PO4)/CH3CN (1/1), over 20 min,
1.0 mL/min, at 35 °C, with UV detection at 210 nm; 17: 11.0
min, 18: 8.0 min.
HPLC Method F. Column: Unison UK-C18, 3 μm, 4.6 mm
I.D. × 150 mm (Imtakt), elution with 20 mmol/L aq
NH4H2PO4 buffer (pH 4.0 with H3PO4)/CH3CN (11/9),
over 25 min, 1.0 mL/min, at 15 °C, with UV detection at 210
nm; 18: 13.0 min.
HPLC Method G. Column: YMC-Pack ProC18, AS-302, 5
μm, 4.6 mm I.D. × 150 mm (YMC), elution with phosphate
buffer (pH 7.5)/CH3CN (55/45), over 40 min, 1.0 mL/min, at
40 °C, with UV detection at 210 nm; 4: 1.4 min, 1: 11.8 min,
imine intermediate: 13.5 min.
HPLC Method H. Column: Unison UK-Phenyl, AS-302, 3
μm, 4.6 mm I.D. × 250 mm (Imtakt), elution A: 200 mM
NaClO4 aq (adjusted pH 2.0 with HClO4)/CH3CN (7/3),
elution B: 200 mM NaClO4 aq (adjusted pH 2.0 with HClO4)/
CH3CN (2/8), at 50 °C, with UV detection at 210 nm, 1: 34
min. Gradient program; see Table 9.
4. EXPERIMENTAL SECTION
General Information. Starting materials, reagents, and
solvents were obtained from commercial suppliers and used
without further purification. 1H NMR spectra were recorded on
a Varian Unity Inova 600 (600 MHz) spectrometer. Chemical
shifts are reported in ppm from the solvent resonance DMSO-
d6 (2.50 ppm) or the tetramethylsilane (0.0 ppm) resonance as
the internal standard [DMSO-d6]. Data are reported as follows:
chemical shift, integration, multiplicity (s = singlet, d = doublet,
t = triplet, m = multiplet, and brs = broad singlet) and coupling
constants (Hz). 13C NMR spectra were recorded on a Unity
Inova 600 (150 MHz) spectrometer with complete proton
decoupling. Chemical shifts are reported in ppm from the
solvent resonance as the internal standard [tetramethylsilane
1
(0.0 ppm) in DMSO-d6]. Structures are assigned by H, 13C,
DEPT, COSY, TOCSY, HSQC, and HMBC spectra. HR-mass
spectra were measured using LC/MS-IT-TOF (Shimadzu) or
LTQ Orbitrap Discovery (Thermo Fisher Scientific). HPLC
was performed using a Hitachi D-2500 or D-7500 system and
Agilent 1100. HPLC methods are described below; see Table 9
for the gradient program.
HPLC Method A. Column: YMC-Pack ODS-AM, 5 μm, 4.6
mm I.D. × 150 mm (YMC), elution with 20 mM phosphate
buffer (KH2PO4: 1.36 g + Na2HPO4·12H2O 3.58g + Sodium 1-
decanesulfonate: 1.22 g/water 1000 mL, adjusted pH 6.86 with
aq H3PO4)/CH3CN (1/1), over 30 min, 1.0 mL/min, at 35 °C,
Table 9. Gradient program of HPLC Method H
Sodium 5-[(2R)-2-{(2R,6S,9S,11R,14aS,15S,16S,20S,23-
S,25aS)-20-[(1R)-3-amino-1-hydroxy-3-oxopropyl]-9-hexadeca-
namido-2,11,15-trihydroxy-6-[(1R)-1-hydroxyethyl]-16-methyl-
5,8,14,19,22,25-hexaoxotetracosahydro-1H-dipyrrolo[2,1-
0 min
40 min
70 min
80 min
elution A
elution B
100%
0%
100%
0%
0%
0%
100%
100%
I
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
c:2′,1′-l][1,4,7,10,13,16]hexaazacyclohenicosin-23-yl}-2-hy-
droxyethyl]-2-hydroxyphenyl sulfate (11).
water 160 L) at 19 to 23 °C. The reaction mixture was stirred at
23 °C for 1 h; HPLC analysis subsequently indicated that <1%
11 remained (HPLC method B). To the reaction mixture, 7 M
NH3 in MeOH (35 L) was added, and the mixture was aged at
24 °C for 4 h; HPLC analysis subsequently indicated that <1%
benzyl bromide remained and formation of 12 had occurred
(HPLC method B). The mixture was cooled to 7 °C, 2 M HCl/
MeOH (594 L, 1370 mol) was added, and the mixture was
warmed to 20 °C and aged for 17 h; HPLC analysis
subsequently indicated that <1% 12 remained (HPLC method
B). The batch was concentrated in vacuo to 640 L, and the
residue was divided into three equal batches for isolation. To a
solution of MeOH (27L)/water (2136 L), one of the divided
batches was added at 25 °C, and the resulting slurry was
agitated for 1 h, filtered, and washed with water (800 L). Loss
to the filtrate was less than 0.1%. The wet cake was dried in
vacuo at 40 °C to afford the desired crude 13 (46.7 kg, 88%
yield). This precipitation procedure was repeated three times,
and 140.5 kg of crude 13 was prepared in total. Crude 13 was
divided into two batches and washed with water as follows:
crude 13 (70.0 kg) was added to water (1400 L) at 25 °C and
aged for 3 h, filtered, and washed with water (700 L). Loss to
the filtrate was less than 0.1%. The wet cake was dried in vacuo
at 40 °C to afford the desired 13 (66.9 kg, 95.6% yield). This
precipitation procedure was repeated twice, and 132.5 kg of 13
was prepared in total (83.5% yield). The analytically pure
sample of 13 was purified via ODS column chromatography.
Compound 12. HR-MS (HPLC/MS ESI-TOF) calcd for
C58H88N8O19S 1231.5814 [M − H]−, found 1231.5810 [M −
H]−.
To a solution of tert-butyldimethylsilane (29.8 kg, 256 mol)
and trifluoroacetic acid (381.4 kg, 3345 mol) was added 3 (51.2
kg, 42.8 mol) at 0−10 °C, and the mixture was stirred at 9 °C
for 8 h; HPLC analysis subsequently indicated that <1% 3
remained (HPLC method A). The reaction mixture was then
added to a solution of water (512 L), anhydrous sodium acetate
(51.2 kg, 624 mol), and NaCl (102.4 kg) at 5−20 °C, and,
during addition, the pH was adjusted to between 6 and 8 with
an aqueous solution of sodium hydroxide prepared in another
vessel. After the addition, the pH was adjusted to 6.4, and the
resulting slurry was stored at 18 °C for 12 h without agitation.
The batch was then stirred, filter aid (Radiolite, 102.4 kg) was
added, and the mixture was filtered and washed with an
aqueous solution of NaCl (NaCl 20.5 kg/water 102 L). The
resulting filtrate was added to water (1024 L), and the batch
was agitated at 21 °C for 5 h, filtered, and washed with water
(512 L). To the resulting filtrate was added NaCl (384 kg), and
the mixture was agitated at 22 °C for 14 h, filtered, and washed
with an aqueous solution of NaCl (NaCl 30.7 kg/water 154 L).
Loss to filtrate was less than <1%. The wet cake was dried in
vacuo at 40 °C to afford the desired crude 11 (52.5 kg, 105.4%
yield). This procedure was repeated three times, and 160.3 kg
of crude 11 was prepared in total (108% yield). 160.3 kg of
crude 11 was purified using DIAION HP20SS5 (eluent: water/
MeOH = 6/4) in seven batches. Regeneration method of the
DIAION HP20SS5 was as follows: wash with (1) DMF, (2)
MeOH, (3) water/MeOH = 1/1. Desired fractions were
combined and concentrated in vacuo to 640 L. To the residue
was added DMF (640 L), and the solution was concentrated in
vacuo to 640 L. The resulting solution was used in the next step
without isolation. The analytically pure sample of 11 was
purified with ODS column chromatography.
Compound 13. HR-MS (HPLC/MS ESI-TOF) calcd for
C58H88N8O16 1151.6246 [M − H]−, found 1151.6221 [M −
H]−.
1H NMR (600 MHz, DMSO-d6) δ 8.89 (1H, s), 8.06−8.15
(2H, m), 7.69 (1H, brs), 7.51 (1H, brs), 7.46 (2H, d, J = 7.2
Hz), 7.28−7.40 (4H, m), 7.20 (1H, brs), 6.84 (1H, d, J = 8.4
Hz), 6.78 (1H, brs), 6.64−6.67 (1H, m), 6.48−6.51 (1H, m),
5.27 (1H, brs), 5.14−5.21 (2H, m), 5.06 (2H, s), 5.01−5.03
(1H, m), 4.94−5.00 (1H, brs), 4.83−4.90 (1H, m), 4.79−4.81
(1H, m), 4.74−4.77 (1H, m), 4.14−4.44 (8H, m), 3.98−4.03
(1H, m), 3.94−3.97 (1H, m), 3.78−3.85 (2H, m), 3.68−3.76
(1H, m), 3.38−3.44 (1H, brs), 3.17 (1H, t, J = 9.6 Hz), 2.92
(1H, d, J = 13.8 Hz), 2.37−2.47 (3H, m), 2.29−2.36 (1H, m),
2.21−2.28 (1H, m), 2.16−2.21 (1H, m), 2.01−2.15 (2H, m),
1.81−1.92 (2H, m), 1.63−1.72 (1H, m), 1.40−1.48 (2H, m),
1.15−1.32 (24H, m), 1.00−1.06 (3H, m), 0.95 (3H, d, J = 7.2
Hz), 0.85 (3H, t, J = 6.9 Hz).
HR-MS (HPLC/MS ESI-TOF) calcd for C51H82N8O19S
1141.5344 [M − H]−, found 1141.5332 [M − H]−.
1H NMR (600 MHz, DMSO-d6) δ 8.71 (1H, s), 8.08 (2H, d,
J = 7.8 Hz), 7.71 (1H, brs), 7.55 (1H, brs), 7.34 (1H, d, J = 8.4
Hz), 7.19 (1H, s), 6.97 (1H, s), 6.76−6.80 (2H, m), 6.71 (1H,
d, J = 8.4 Hz), 5.36 (1H, brs), 5.14−5.19 (2H, m), 5.10 (1H, d,
J = 6.6 Hz), 4.86−4.97 (1H, brs), 4.83−4.85 (1H, m), 4.81
(1H, d, J = 5.4 Hz), 4.75 (1H, d, J = 6.0 Hz), 4.14−4.47 (8H,
m), 3.98−4.02 (1H, m), 3.92−3.96 (1H, m), 3.77−3.83 (2H,
m), 3.71 (1H, d, J = 10.8 Hz), 3.38−3.43 (1H, m), 3.17−3.22
(1H, m), 2.92 (1H, d, J = 13.2 Hz), 2.38−2.47 (3H, m), 2.30−
2.36 (1H, m), 2.21−2.27 (1H, m), 2.14−2.20 (1H, m), 2.02−
2.13 (2H, m), 1.84−1.93 (2H, m), 1.63−1.72 (1H, m), 1.40−
1.47 (2H, m), 1.17−1.32 (24H, m), 1.03 (3H, d, J = 6.0 Hz),
0.96 (3H, d, J = 6.6 Hz), 0.86 (3H, t, J = 6.9 Hz).
13C NMR (150 MHz, DMSO-d6) δ 172.7, 172.2, 171.7,
171.0, 170.3, 169.8, 169.5, 167.7, 146.5, 144.8, 137.5, 131.3,
128.2, 127.5 (2 carbons), 119.8, 116.9, 114.0, 73.8, 72.0, 69.8,
69.0, 68.7, 68.6, 66.1, 65.8, 60.7, 56.6, 55.7, 53.5, 51.0, 49.9,
42.2, 38.3, 37.5, 36.9, 34.9, 33.7, 31.2, 29.0, 28.9, 28.7, 28.6,
28.4, 25.2, 22.0, 18.9, 13.9, 10.8.
N-{(2R,6S,9S,11R,14aS,15S,16S,20S,23S,25aS)-20-[(1R)-3-
Amino-1-hydroxy-3-oxopropyl]-23-{(1R)-2-[4-(benzyloxy)-3-
methoxyphenyl]-1-hydroxyethyl}-2,11,15-trihydroxy-6-[(1R)-
1-hydroxyethyl]-16-methyl-5,8,14,19,22,25-hexaoxotetracosahy-
dro-1H-dipyrrolo[2,1-c:2′,1′-l][1,4,7,10,13,16]-
hexaazacyclohenicosin-9-yl}hexadecanamide (14).
13C NMR (150 MHz, DMSO-d6) δ 172.7, 172.2, 171.6,
171.0, 170.4, 169.7, 169.5, 167.8, 147.4, 140.3, 129.7, 125.8,
124.1, 116.8, 73.9, 71.9, 69.0, 68.7, 68.6, 66.1, 65.8, 60.7, 56.6,
56.2, 55.7, 53.4, 51.0, 49.8, 42.2, 38.3, 37.5, 37.0, 34.9, 33.7,
32.0, 29.0, 28.9, 28.7, 28.6, 28.4, 25.2, 22.0, 18.9, 13.9, 10.8.
N-{(2R,6S,9S,11R,14aS,15S,16S,20S,23S,25aS)-20-[(1R)-3-
Amino-1-hydroxy-3-oxopropyl]-23-{(1R)-2-[4-(benzyloxy)-3-
hydroxyphenyl]-1-hydroxyethyl}-2,11,15-trihydroxy-6-[(1R)-1-
hydroxyethyl]-16-methyl-5,8,14,19,22,25-hexaoxotetracosahy-
dro-1H-dipyrrolo[2,1-c:2′,1′-l][1,4,7,10,13,16]-
hexaazacyclohenicosin-9-yl}hexadecanamide (13).
Methylation. To a solution of 13 (132.5 kg, 115 mol) in
DMF (663 L) was added methyl iodide (22.8 kg, 161 mol) at
25 °C. To this solution was added an aqueous solution of
lithium hydroxide (LiOH: 3.7 kg, 154 mol/water 106 L) at 1 to
Benzylation and Desulfurization. To a solution of 11 in
DMF was added benzyl bromide (37.6 kg, 220 mol) and an
aqueous solution of lithium hydroxide (LiOH 3.9 kg, 163 mol/
J
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
15 °C and agitated at 24 °C for 6 h; HPLC analysis
subsequently indicated that <3% 13 remained (HPLC method
B). To the batch was added 7 M NH3 in MeOH (25 L), and
the solution was aged at 25 °C for 4 h. The mixture was divided
into three equal batches for isolation. To a solution of NaCl
(199 kg), conc. HCl (10.6 kg), and water (2077 L) was added
one of the divided batches at 25 °C. The solution was then
washed with MeOH (33 L) and agitated for 1 h, filtered, and
washed with water (663 L). Loss to the filtrate was less than
0.1%. The wet cake was dried in vacuo at 50 °C to afford the
desired 14 (47.1 kg, 105% yield). This precipitation procedure
was repeated three times, and 139.1 kg of 14 was prepared in
total (104% yield). The analytically pure sample of 14 was
purified with ODS column chromatography.
performed as follows: wash with (1) DMF, (2) acetone, and
(3) water/MeOH = 1/1. Desired fractions were combined, and
the solution was divided into two batches, which were
concentrated in vacuo to 277 L. MeOH (347 L) was then
added to the residue, and the mixture was again concentrated in
vacuo to 277 L. The residue was then added to a solution of
NaCl (277.2 kg) in water (2772 L) at 20 °C, and the mixture
was washed with MeOH (35 L) and agitated for 2 h, filtered,
and washed with water (1040 L). Loss to the filtrate was less
than 0.1%. The wet cake was dried in vacuo at 40 °C to afford
the desired 15 (36.9 kg). This precipitation procedure was
repeated twice, and 74.4 kg of 15 was prepared in total (54.5%
yield). The analytically pure sample of 15 was purified with
ODS column chromatography.
HR-MS (HPLC/MS ESI-TOF) calcd for C59H90N8O16
HR-MS (HPLC/MS ESI-TOF) calcd for C59H88N8O15
1165.6402 [M − H]−, found 1165.6383 [M − H]−.
1147.6296 [M − H]−, found 1147.6290 [M − H]−.
1H NMR (600 MHz, DMSO-d6) δ 8.16 (1H, brs), 8.09 (1H,
d, J = 7.2 Hz), 7.70 (1H, brs), 7.53 (1H, brs), 7.42−7.46 (2H,
m), 7.37−7.41 (2H, m), 7.29−7.36 (2H, m), 7.21 (1H, brs),
6.92 (1H, d, J = 8.4 Hz), 6.79 (1H, brs), 6.75 (1H, brs), 6.67
(1H, d, J = 7.8 Hz), 5.25 (1H, brs), 5.19−5.20 (1H, m), 5.15−
5.18 (1H, m), 5.03 (2H, s), 4.96−5.01 (2H, m), 4.84−4.89
(1H, m), 4.80 (1H, d, J = 5.4 Hz), 4.76−4.78 (1H, m), 4.38−
4.44 (2H, m), 4.35 (1H, brs), 4.17−4.29 (5H, m), 4.00 (1H, t, J
= 8.1 Hz), 3.95 (1H, brs), 3.79−3.86 (2H, m), 3.68−3.78 (4H,
m), 3.38−3.46 (1H, m), 3.17 (1H, t, J = 9.6 Hz), 2.91 (1H, d, J
= 13.2 Hz), 2.52−2.58 (1H, m), 2.45−2.49 (1H, m), 2.41−2.44
(1H, m), 2.29−2.36 (1H, m), 2.23−2.28 (1H, m), 2.17−2.22
(1H, m), 2.01−2.16 (2H, m), 1.84−1.97 (2H, m), 1.64−1.72
(1H, m), 1.40−1.48 (2H, m), 1.14−1.33 (24H, m),1.03 (3H, d,
J = 6.0 Hz), 0.95 (3H, d, J = 6.6 Hz), 0.85 (3H, t, J = 6.9 Hz).
13C NMR (150 MHz, DMSO-d6) δ 172.8, 172.2, 171.7,
171.1, 170.4, 169.8, 169.5, 167.6, 148.6, 146.0, 137.2, 131.1,
128.2, 127.7, 127.6, 121.1, 113.5, 73.9, 72.3, 69.8, 69.0, 68.7,
66.1, 65.8, 60.8, 56.6, 55.7, 55.3, 53.5, 51.0, 49.9, 42.1, 38.3,
37.5, 36.9, 34.9, 33.6, 31.2, 29.0, 28.9, 28.8, 28.7, 28.6, 28.4,
25.2, 22.0, 18.9, 13.9, 10.8.
1H NMR (600 MHz, DMSO-d6) δ 8.23 (1H, d, J = 7.2 Hz),
7.94 (1H, d, J = 7.8 Hz), 7.68−7.75 (1H, m), 7.28−7.48 (7H,
m), 6.93 (1H, d, J = 8.4 Hz), 6.72 (1H, s), 6.64−6.69 (1H, m),
5.85 (1H, d, J = 6.0 Hz), 5.18 (2H, d, J = 5.4 Hz), 5.00−5.08
(3H, m), 4.94−4.98 (1H, m), 4.83−4.88 (1H, m), 4.74−4.82
(2H, m), 4.35−4.45 (3H, m), 4.10−4.25 (5H, m), 4.00 (1H, t, J
= 8.1 Hz), 3.93−3.97 (1H, m), 3.80−3.88 (2H, m), 3.73−3.78
(3H, m), 3.72 (1H, brs), 3.44−3.54 (1H, m), 3.20 (1H, t, J =
9.3 Hz), 2.89 (1H, d, J = 13.8 Hz), 2.78 (1H, d, J = 15.0 Hz),
2.48−2.65 (2H, m), 2.44−2.47 (1H, m), 2.30−2.38 (1H, m),
2.21−2.29 (1H, m), 2.01−2.12 (2H, m), 1.84−1.98 (2H, m),
1.62−1.72 (1H, m), 1.40−1.49 (2H, m), 1.14−1.34 (24H,
m),1.04 (3H, d, J = 6.0 Hz), 0.95 (3H, d, J = 7.2 Hz), 0.85 (3H,
t, J = 6.9 Hz).
13C NMR (150 MHz, DMSO-d6) δ 171.9 (2 carbons), 171.1,
170.3, 170.1, 169.3, 166.5, 148.6, 146.1, 137.2, 130.9, 128.3,
127.6, 121.0, 119.4, 113.4, 74.0, 72.5, 69.8, 69.1, 68.4, 68.0,
66.3, 65.7, 60.8, 56.3, 55.6, 55.2, 55.0, 53.6, 51.2, 49.7, 41.8,
37.4, 36.9, 35.0, 33.5, 31.2, 29.0, 28.9, 28.8, 28.6, 28.4, 25.3,
22.0, 21.6, 18.9, 13.9, 10.9.
N-{(2R,6S,9S,11R,14aS,15S,16S,20S,23S,25aS)-20-[(1R)-3-
Amino-1-hydroxypropyl]-2,11,15-trihydroxy-6-[(1R)-1-hydrox-
yethyl]-23-[(1R)-1-hydroxy-2-(4-hydroxy-3-methoxyphenyl)-
ethyl]-16-methyl-5,8,14,19,22,25-hexaoxotetracosahydro-1H-
dipyrrolo[2,1-c:2′,1′-l][1,4,7,10,13,16]hexaazacyclohenicosin-9-
yl}hexadecanamide monohydrochloride (17·HCl).
N-{(2R,6S,9S,11R,14aS,15S,16S,20S,23S,25aS)-23-{(1R)-2-
[4-(Benzyloxy)-3-methoxyphenyl]-1-hydroxyethyl}-20-[(1R)-
2-cyano-1-hydroxyethyl]-2,11,15-trihydroxy-6-[(1R)-1-hydrox-
yethyl]-16-methyl-5,8,14,19,22,25-hexaoxotetracosahydro-1H-
dipyrrolo[2,1-c:2′,1′-l][1,4,7,10,13,16]hexaazacyclohenicosin-9-
yl}hexadecanamide (15).
Sponge Nickel NDHT-9011 (36.5 kg wet) was washed with
EtOH twice. Compound 15 (36.5 kg, 31.8 mol) in a solution of
EtOH/28% NH3 aq = 1/1 (365 L) was added to a 1500-L
autoclave, and then prewashed sponge Nickel NDHT-90 was
added. The autoclave was flushed with nitrogen gas three times
and then with hydrogen gas three times. The hydrogenation
was conducted under 33 psi pressure of hydrogen at 50 °C for
5 h; HPLC analysis subsequently indicated that <1% 15
remained and formation of 16 (HPLC method D). Then, 7.5%
Pd/C10 (3.7 kg, 50% wet) and EtOH (20 L) were added to the
batch, and the reaction was conducted under 35 psi pressure of
hydrogen at 35 °C for 1 h; HPLC analysis subsequently
indicated that <1% 16 remained (HPLC method D). After the
completion of the reaction, MeOH (146 L) was added to the
mixture and filtered and then was washed with MeOH (220 L)
and concentrated in vacuo to 292 L. EtOH (183 L) was added
to the residue, and the mixture was concentrated in vacuo to
292 L. EtOH (183 L) was then added again to the residue, and
the mixture was concentrated in vacuo to 292 L. EtOH (365 L)
and 2 M HCl/EtOH were then added to the batch until pH
reached 2.3. The batch was concentrated in vacuo to 146 L,
Dehydration. To a solution of 14 (69.3 kg, 59.4 mol) in
NMP (485 L) was added pyridine (131.5 kg, 1662 mol) and
EDC·HCl (136.6 kg, 713 mol). The mixture was warmed to 75
°C and aged for 6 h; HPLC analysis subsequently indicated that
<5% 14 remained (HPLC method C). The batch was cooled to
40 °C, and EDC·HCl (34.1 kg, 178 mol) was added. The
mixture was warmed to 75 °C and aged for 15 h; HPLC
analysis subsequently indicated that <2% 14 remained (HPLC
method C). The reaction mixture was then added to a mixture
of EtOAc (1386 L) and water (693 L) at 27 °C. The resulting
aqueous layer was extracted with EtOAc (693 L). These
organic layers were combined and washed with an aqueous
solution of NaCl (NaCl 139 kg/water 693 L) 5 times, and the
resulting organic layer was concentrated in vacuo to 208 L.
MeOH (347 L) was added to the residue, and the mixture was
concentrated in vacuo to 208 L three times. This dehydration
procedure was conducted in two batches. The combined
resulting residue was purified with SEPABEADS SP207SS5 in
eight batches (elution: 82% aqueous MeOH then MeOH). The
regeneration method of the SEPABEADS SP207SS5 was
K
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
poured into CH3CN (1825 L) at 3 °C, and then washed with
MeOH (36 L). The resulting slurry was aged at 3 °C for 1 h,
filtered, and washed with CH3CN (111 L). Loss to the filtrate
was 2%. The wet cake was dried in vacuo at 50 °C to afford the
desired 17·HCl (30.8 kg). This hydrogenation procedure was
repeated twice, and 63.0 kg of 17·HCl was prepared in total
(89.6% yield). The analytically pure sample of 17·HCl was
purified with ODS column chromatography.
Compound 16. HR-MS (HPLC/MS ESI-TOF) calcd for
C59H92N8O15 1151.6609 [M − H]−, found 1151.6585 [M −
H]−
Compound 17·HCl. HR-MS (HPLC/MS ESI-TOF) calcd
for C52H86N8O15 1097.5907 [M + Cl]−, found 1097.5884 [M +
Cl]−.
1H NMR (600 MHz, DMSO-d6) δ 8.72 (1H, brs), 8.52 (1H,
d, J = 6.6 Hz), 8.23−8.35 (2H, m), 8.20 (1H, brs), 7.86 (1H, d,
J = 6.6 Hz), 7.55 (1H, brs), 7.34 (1H, d, J = 9.0 Hz), 6.62−6.70
(2H, m), 6.53−6.58 (1H, m), 5.41 (1H, brs), 5.05−5.35 (4H,
m), 4.83 (1H, brs), 4.72−4.81 (2H, m), 4.48 (1H, brs), 4.42
(1H, br), 4.32−4.36 (1H, m), 4.23−4.24 (1H, m), 4.16−4.22
(2H, m), 4.08−4.14 (1H, m), 4.02−4.07 (1H, m), 3.97−3.99
(1H, m), 3.90−3.96 (2H, m), 3.84−3.89 (1H, m), 3.72−3.76
(3H, m), 3.61−3.71 (3H, m), 3.54−3.60 (2H, m), 3.48−3.52
(1H, m), 3.23−3.28 (1H, m), 3.18−3.24 (1H, m), 3.14 (1H,
brs), 3.07 (1H, brs), 2.78−2.84 (1H, m), 2.58−2.64 (1H, m),
2.48−2.54 (1H, m), 2.30−2.40 (1H, m), 2.18−2.24 (1H, m),
2.06−2.15 (2H, m), 1.91−2.05 (3H, s), 1.68−1.80 (2H, m),
1.41−1.49 (2H, m), 1.13−1.30 (24H, m), 1.00 (3H, d, J = 5.4
Hz), 0.95 (3H, d, J = 7.2 Hz), 0.86 (3H, t, J = 6.9 Hz).
Two peaks for OH protons were not confirmed.
1H NMR (600 MHz, DMSO-d6) δ 8.71 (1H, s), 8.48 (1H,
m), 8.33 (1H, d, J = 7.8 Hz), 7.76−7.81 (1H, m), 7.64−7.75
(3H, m), 7.62 (1H, brs), 7.36 (1H, d, J = 9.0 Hz), 6.62−6.69
(2H, m), 6.54−6.59 (1H, m), 5.33 (1H, brs), 5.14−5.24 (3H,
m), 4.80−4.85 (1H, brs), 4.70−4.79 (3H, m), 4.46 (1H, brs),
4.42 (1H, brs), 4.32−4.38 (1H, m), 4.20−4.28 (2H, m), 4.16−
4.19 (1H, m), 4.10−4.15 (1H, m), 4.02−4.08 (1H, m), 3.90−
3.98 (3H, m), 3.82−3.88 (1H, m), 3.73 (3H, s), 3.69 (1H, d, J
= 10.8 Hz), 3.42−3.51 (1H, m), 3.22 (1H, t, J = 9.3 Hz), 2.88−
2.94 (1H, m), 2.78−2.87 (2H, m), 2.56−2.64 (1H, m), 2.50−
2.54 (1H, m), 2.28−2.37 (1H, m), 2.18−2.24 (1H, m), 2.05−
2.16 (2H, m), 1.83−1.98 (3H, m), 1.72−1.78 (1H, m), 1.65−
1.71 (1H, m), 1.38−1.47 (2H, m), 1.13−1.34 (24H, m), 1.02
(3H, d, J = 6.0 Hz), 0.95 (3H, d, J = 7.2 Hz), 0.86 (3H, t, J =
6.9 Hz).
13C NMR (150 MHz, DMSO-d6) δ 173.1, 171.7, 171.1,
170.1 (2 carbons), 169.6, 167.3, 147.0, 144.6, 128.8, 121.3,
115.1, 113.5, 73.9, 73.2, 70.2, 69.1, 67.8, 65.4, 61.2, 59.1, 57.4,
57.1, 56.5, 55.5, 55.2, 54.5, 51.4, 49.9, 42.8, 41.6, 39.0, 37.2,
36.9, 35.0, 33.1, 31.2, 29.0, 28.9 (2 carbons), 28.7, 28.6, 28.3,
27.4, 25.4, 22.0, 19.1, 13.9, 11.0.
(2R,6S,9S,11R,14aS,15S,16S,20S,23S,25aS)-9-Amino-20-
{(1R)-3-[(1,3-dihydroxypropan-2-yl)amino]-1-hydroxypropyl}-
2,11,15-trihydroxy-6-[(1R)-1-hydroxyethyl]-23-[(1R)-1-hy-
droxy-2-(4-hydroxy-3-methoxyphenyl)ethyl]-16-methyloctade-
cahydro-19H-dipyrrolo[2,1-c:2′,1′-l][1,4,7,10,13,16]-
hexaazacyclohenicosine-5,8,14,19,22,25(9H)-hexone (4).
To a solution of 18·HCl (13.0 g, 11.1 mmol) in MeOH (1.0
L) was added water (2.0 L) and a solution of acylase
Streptomyces sp. no. 6907 (0.46 L), and then the pH was
adjusted to 8.0 with aq NaOH. The reaction mixture was then
agitated at 37 °C for 6 h; HPLC analysis subsequently indicated
that 0.1 mg/mL of 18 remained (HPLC method F). After the
completion of the reaction, pH was adjusted from 8 to 3.5 with
78% aqueous solution of H2SO4, and the mixture was
concentrated in vacuo, filtered through the filter aid, and
washed with water. The pH of the resulting filtrate was adjusted
to 2.5 with 78% aqueous solution of H2SO4, and then the
mixture was concentrated in vacuo followed by polish filtration.
MeOH (30 mL) was added to the resulting filtrate, and the pH
was adjusted to 9.0 with aq NaOH. The resulting solution was
purified with SP207SS5 followed by SEPABEADS SP700.5 The
collected desired fractions were concentrated in vacuo to 32
mL. EtOAc (600 mL) was added to the residue, and the
resulting slurry was agitated for 1 h, filtered, and washed with
EtOAc (32 mL). The wet cake was dried in vacuo at 40 °C to
afford the desired 4 (4.44 g, 44.6% yield).
13C NMR (150 MHz, DMSO-d6) δ 172.8, 171.6, 171.2,
170.2, 170.1, 169.4, 167.2, 147.0, 144.6, 128.9, 121.3, 115.1,
113.4, 73.9, 73.0, 69.5, 69.1, 68.1, 65.5, 65.4, 61.2, 56.6, 55.7,
55.6, 55.2, 54.4, 51.3, 49.9, 42.0, 37.2, 36.9 (2 carbons), 34.9,
33.3, 31.2, 29.1, 29.0, 28.9, 28.7, 28.6, 28.3, 25.3, 22.0, 19.2,
13.9, 10.9.
N-{(2R,6S,9S,11R,14aS,15S,16S,20S,23S,25aS)-20-{(1R)-3-
[(1,3-Dihydroxypropan-2-yl)amino]-1-hydroxypropyl}-2,11,15-
trihydroxy-6-[(1R)-1-hydroxyethyl]-23-[(1R)-1-hydroxy-2-(4-
hydroxy-3-methoxyphenyl)ethyl]-16-methyl-5,8,14,19,22,25-
hexaoxotetracosahydro-1H-dipyrrolo[2,1-c:2′,1′-l]-
[1,4,7,10,13,16]hexaazacyclohenicosin-9-yl}hexadecanamide
monohydrochloride (18·HCl).
To a solution of 17·HCl (30.7 kg, 27.9 mol) in MeOH (307
L) was added dihydroxyacetone (7.54 kg, 83.7 mol). The batch
was cooled to 2 °C, and 2-picoline/borane complex (4.48 kg,
41.9 mol) was added. The mixture was then warmed to 23 °C
and agitated for 5 h. Dihydroxyacetone (7.54 kg, 83.7 mol) was
added at 23 °C, and the mixture was agitated for 12 h; HPLC
analysis subsequently indicated that <1% 17 remained (HPLC
method E). To the batch was added 2 M HCl/MeOH (49 L),
and then the mixture was concentrated in vacuo to 92 L.
MeOH (307 L) was added to the residue, and the mixture was
concentrated in vacuo to 92 L, poured into CH3CN (1535 L)
at 3 °C, and washed with MeOH (30 L). The resulting slurry
was agitated at 3 °C for 1 h, filtered, and washed with CH3CN
(93 L). Loss to the filtrate was 2%. The wet cake was dried in
vacuo at 50 °C to afford the desired 18·HCl (32.8 kg). This
reductive amination procedure was repeated twice, and 66.6 kg
of 18·HCl was prepared in total (99.4% yield). The analytically
pure sample of 18 was purified with ODS column.
HR-MS (HPLC/MS ESI-TOF) calcd for C39H62N8O16
921.4176 [M + Na]+, found 921.4188 [M + Na]+.
1H NMR (600 MHz, DMSO-d6) δ 7.68−7.75 (1H, m), 7.64
(1H, brs), 7.52 (1H, brs), 7.42−7.48 (1H, m), 6.62−6.68 (2H,
m), 6.52−6.58 (1H, m), 5.20−5.24 (1H, m), 5.16−5.19 (1H,
m), 4.76−4.81 (1H, m), 4.73 (1H, brs), 4.36−4.48 (2H, m),
4.32 (1H, brs), 4.10−4.20 (3H, m), 4.00−4.06 (1H, m), 3.96−
3.99 (1H, m), 3.80−3.89 (3H, m), 3.70−3.78 (4H, m), 3.48−
3.56 (1H, m), 3.21−3.47 (7H, m), 3.10−3.20 (1H, m), 2.72−
2.78 (1H, m), 2.58−2.67 (2H, m), 2.52−2.57 (1H, m), 2.45−
2.48 (1H, m), 2.39−2.44 (1H, m), 2.32−2.38 (1H, m), 2.20−
2.28 (1H, m), 1.85−1.98 (2H, m), 1.51−1.60 (1H, m), 1.40−
1.50 (1H, m), 1.13−1.22 (1H, m), 1.09 (3H, d, J = 6.6 Hz),
0.95−0.99 (3H, m). Eight peaks for OH, NH and NH2 protons
were not confirmed.
HR-MS (HPLC/MS ESI-TOF) calcd for C55H92N8O17
1171.6274 [M + Cl]−, found 1171.6252 [M + Cl]−.
L
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
13C NMR (150 MHz, DMSO-d6) δ 174.6, 171.1, 170.4,
169.7, 169.3, 168.3, 147.0, 144.6, 128.8, 121.3, 115.1, 113.4,
73.4, 72.4, 69.8, 69.0, 68.5, 67.8, 65.9, 60.9 (2 carbons), 60.8,
60.7, 59.7, 56.7, 55.7, 55.4, 55.2, 53.6, 52.6, 50.9, 43.8, 43.3,
37.5, 37.1 (2 carbons), 32.8, 19.1, 10.8.
(2R,6S,9S,11R,14aS,15S,16S,20S,23S,25aS)-9-{[(trans-4-{5-
[4-(4-cyclohexyl-4-methoxypiperidin-1-yl)phenyl]-1,3,4-thia-
diazol-2-yl}cyclohexyl)methyl]amino}-20-{(1R)-3-[(1,3-dihy-
droxypropan-2-yl)amino]-1-hydroxypropyl}-2,11,15-trihy-
droxy-6-[(1R)-1-hydroxyethyl]-23-[(1R)-1-hydroxy-2-(4-hy-
droxy-3-methoxyphenyl)ethyl]-16-methyloctadecahydro-19H-
dipyrrolo[2,1-c:2′,1′-l][1,4,7,10,13,16]hexaazacyclohenicosine-
5,8,14,19,22,25(9H)-hexone dihydrochloride (1).
To a solution of 4 (3.5 g, 3.9 mmol) in DMF (21 mL) was
added 2 M HCl/MeOH (3.22 mL, 6.44 mmol), MeOH (17.5
mL), and CH2Cl2 (17.5 mL). The batch was agitated for 0.5 h,
and then 10 (2.18 g, 4.7 mmol) was added, followed by the
addition of TFA (1.42 g, 12.5 mmol). To the resulting solution
was added t-BuNH2·BH3 (0.51 g, 5.9 mmol) at −12 °C; the
solution was warmed to 25 °C (warming rate: 10 °C/h) and
agitated for 12 h; HPLC analysis subsequently indicated that
<1% of 4 remained (HPLC method G). After quenching with 4
M HCl/EtOAc (14 mL, 56 mmol), CH3CN (350 mL) was
added to the mixture at 0 °C. The resulting slurry was then
agitated at 0 °C for 1 h, filtered, and washed with CH3CN (35
mL). The wet cake was dried in vacuo at 40 °C to afford the
desired crude 1 (5.71 g, 103% yield). The crude product (1.3 g,
0.91 mmol) was purified with HP20SS5 elution in a solution of
MeOH/water (7/3). pH of the collected desired fractions was
adjusted from 2.57 to 6.10 with aq NaOH. Half of the volume
of the solution was then purified with HP20SS5 elution in
MeOH/EtOH (1/1). The desired fraction was concentrated in
vacuo to 8.8 mL, and the residue was added to EtOH (30 mL).
The resulting mixture was concentrated in vacuo to 8.8 mL, and
then EtOH (2.2 mL) and water (0.44 mL) was added. The pH
was then adjusted to 5.68 with 2 M HCl/EtOH at 0 °C. EtOAc
(55 mL) was added to the resulting solution at 0 °C, and the
slurry was agitated for 1 h, filtered, and washed with EtOAc (5
mL). The wet cake was dried in vacuo at 40 °C to afford the
desired 1 (0.443 g, 68.2% yield from crude 1; HPLC purity:
96.6%, HPLC method H).
carbons), 113.7, 74.5, 73.2, 72.4, 68.9, 68.5, 68.2, 67.6, 67.4,
66.9, 64.6, 61.9, 60.2, 60.1, 57.5, 57.3, 57.2, 56.3, 55.9, 55.6,
55.5, 53.8, 52.2, 51.0, 47.1, 44.3, 43.0 (2 carbons), 42.2, 41.5,
38.5, 37.6, 37.2, 34.0, 33.5, 32.3 (2 carbons), 29.3 (2 carbons),
28.6 (2 carbons), 26.5 (2 carbons), 26.3 (2 carbons), 26.2, 19.6,
10.9.
ASSOCIATED CONTENT
* Supporting Information
NMR data for the products. This material is available free of
charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*shinya.yoshida@astellas.com
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We gratefully and sincerely thank Dr. Masaki Tomishima, Dr.
Natsuko Kayakiri, and Mr. Hiroshi Morikawa for their helpful
discussions about the medicinal synthetic method for 1.
Further, we also thank Dr. Tatsuhiro Tokunaga for helpful
discussion of NMR study.
REFERENCES
■
(1) (a) Tawara, S.; Ikeda, F.; Maki, K.; Morishita, Y.; Otomo, K.;
Teratani, N.; Goto, T.; Tomishima, M.; Ohki, H.; Yamada, A.;
Kawabata, K.; Takasugi, H.; Sakane, K.; Tanaka, H.; Matsumoto, F.;
Kuwahara, S. Antimicrob. Agents Chemother. 2000, 44, 57. (b) Ikeda, F.;
Wakai, Y.; Matsumoto, S.; Maki, K.; Watabe, E.; Tawara, S.; Goto, T.;
Watanabe, Y.; Matsumoto, F.; Kuwahara, S. Antimicrob. Agents
Chemother. 2000, 44, 614. (c) Matsumoto, S.; Wakai, Y.; Nakai, T.;
Hatano, K.; Ushitani, T.; Ikeda, F.; Tawara, S.; Goto, T.; Matsumoto,
F.; Kuwahara, S. Antimicrob. Agents Chemother. 2000, 44, 619.
(d) Hatano, K.; Morishita, Y.; Nakai, T.; Ikeda, F. J. Antibiot. 2002,
55, 219. (e) Tomishima, M.; Ohki, H.; Yamada, A.; Takasugi, H.;
Maki, K.; Tawara, S.; Tanaka, H. J. Antibiot. 1999, 52, 674.
(f) Morikawa, H.; Tomishima, M.; Kayakiri, N.; Araki, T.; Barrett,
D.; Akamatsu, S.; Matsumoto, S.; Uchida, S.; Naka, T.; Takeda, S.;
Maki, K. Bioorg. Med. Chem. Lett. 2014, 24, 1172.
(2) FR901379: (a) Boeck, L. D.; Fukuda, D. S.; Abbott, B. J.;
Debono, M. J. Antibiot. 1989, 42, 382. (b) Debono, M.; Abbott, B. J.;
Fukuda, D. S.; Barnhart, M.; Willard, K. E.; Molloy, R. M.; Michel, K.
H.; Turner, J. R.; Butler, T. F.; Hunt, A. H. J. Antibiot. 1989, 42, 389.
(3) Yoshida, S.; Ohigashi, A.; Morinaga, Y.; Hashimoto, H.;
Takahashi, Y.; Ieda, S.; Okada, M. Org. Process Res. Dev. 2013, 17,
1252.
(4) (a) Mayr, H.; Dogan, B. Tetrahedron Lett. 1997, 38, 1013.
(b) Auerbach, J.; Zamore, M. F.; Weinreb, S. M. J. Org. Chem. 1976,
41, 725. (c) Chung, J. Y. L.; Wasicak, J. T.; Arnold, W. A.; May, C. S.;
Nadzan, A. M.; Holladay, M. W. J. Org. Chem. 1990, 55, 270.
(d) Yamamoto, K.; Takemae, M. Bull. Chem. Soc. Jpn. 1989, 62, 2111.
(e) Hart, D. J.; Tsai, Y. M. Tetrahedron Lett. 1981, 22, 1567.
(5) Web site of Mitsubishi Chemical Corporation http://www.
diaion.com/en/products/index.html and http://www.diaion.com/en/
products/synthesis_0201.html.
HR-MS (ESI) calcd for C66H99N11O17S 1350.70139 [M +
H]+, found 1350.70215 [M + H]+.
1H NMR (600 MHz, DMSO-d6, 45 °C) δ 8.60 (1H, s), 8.48
(1H, brs), 8.39 (1H, brs), 8.16 (1H, brs), 7.72 (2H, d, J = 9.0
Hz), 7.62−7.71 (3H, m), 7.01 (2H, d, J = 9.6 Hz), 6.69−6.71
(1H, m), 6.64−6.68 (1H, m), 6.55−6.60 (1H, m), 5.44−5.50
(1H, m), 5.18−5.38 (6H, m), 4.74−4.82 (2H, m), 4.49 (1H,
brs), 4.45 (1H, brs), 4.38−4.44 (1H, m), 4.18−4.25 (3H, m),
4.12−4.17 (2H, m), 4.04−4.11 (2H, m), 4.02 (1H, brs), 3.85−
3.92 (1H, m), 3.72−3.78 (4H, m), 3.66−3.71 (4H, m), 3.60−
3.65 (2H, m), 3.41 (1H, brs), 3.10−3.22 (5H, m), 3.02−3.10
(3H, m), 2.93−3.02 (2H, m), 2.85−2.90 (1H, brs), 2.80−2.84
(1H, m), 2.69−2.74 (1H, m), 2.64−2.68 (1H, m), 2.52−2.58
(1H, m), 2.35−2.44 (1H, m), 2.20−2.30 (2H, m), 2.10−2.18
(2H, m), 1.97−2.04 (1H, m), 1.92−1.96 (1H, m), 1.84−1.90
(1H, m), 1.70−1.82 (5H, m), 1.61−1.68 (7H, m), 1.50−1.60
(4H, m), 1.12−1.27 (7H, m), 1.04−1.10 (2H, m), 0.50−1.02
(5H, m). Two peaks for NH and HCl protons were not
confirmed.
(6) (a) Leonard, W. R., Jr.; Belyk, K. M.; Conlon, D. A.; Bender, D.
R.; DiMichele, L. M.; Liu, J.; Hughes, D. L. J, Org. Chem. 2007, 72,
2335. (b) Journet, M.; Cai, D.; DiMichele, L. M.; Hughes, D. L.;
Larsen, R. D.; Verhoeven, T. R.; Reider, P. J. J, Org. Chem. 1999, 64,
2411.
(7) Bose, D. S.; Narsaiah, A. V. Synthesis 2001, 3, 373.
13C NMR (150 MHz, DMSO-d6, 45 °C) δ 172.7, 171.1,
170.0 (2 carbons), 169.3, 168.1, 167.2, 152.5, 147.2, 144.7,
129.1, 128.5 (2 carbons), 121.4, 118.5, 115.2, 114.2 (2
(8) Meudt, A.; Scherer, S.; Nerdinger, S. WO/2005/070879, 2005.
(9) Kwang-Chung, L.; Yen-Shih, T.; Hao-Ling, F. U.S. Pat.
Appl.2010/0168415 A1.
M
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
(10) Pd catalysts (Kawaken Fine Chemicals Co., Ltd.) http://www.
kawakenfc.co.jp/e/pdf/e_precios_metal_01.pdf.
(11) Cobalt and Nickel catalysts (Kawaken Fine Chemicals Co., Ltd.)
http://www.kawakenfc.co.jp/e/specialty.html and http://www.
kawakenfc.co.jp/e/pdf/e_sponge_metal_01.pdf.
(12) Shia, Y. J.; Freya, L. F.; Tschaena, D. M.; Verhoevena, T. R.
Synth. Commun. 1993, 23, 2623.
(13) Matos, K.; Burkhardt, E. R. Direct Reductive Amination with
Amine Boranes. In Pharmaceutical Process Chemistry; Wiley Online
Library: New York, 2011; Section 6, p 127.
(14) (a) Pelter, A.; Rosser, R. M. J. Chem. Soc. Perkin Trans. I 1984,
717. (b) Bomann, M. D.; Guch, I. C.; DiMare, M. J. Org. Chem. 1995,
60, 5995. (c) Moormanna, A. E. Synth. Commun. 1993, 23, 789.
(15) (a) Cosenza, V. A.; Navarro, D. A.; Stortz, C. A. ARKIVOC
2011, 182. (b) Ruhaak, L. R.; Steenvoorden, E.; Koeleman, C. A. M.;
Deelder, A. M.; Wuhrer, M. Proteomics. 2010, 10, 2330. (c) Sato, S.;
Sakamoto, T.; Miyazawa, E.; Kikugawa, Y. Tetrahedron 2004, 60, 7899.
(d) Kawase, Y.; Yamagishi, T.; Kutsuma, T.; Ueda, K.; Iwakuma, T.;
Nakata, T.; Yokomatsu, T. Heterocycles 2009, 78, 463.
(16) Ueda, S.; Kinoshita, M.; Tanaka, F.; Tsuboi, M.; Shimizu, S.;
Oohata, N.; Hino, M.; Yamada, M.; Isogai, Y.; Hashimoto, S. J. Biosci.
Bioeng. 2011, 112, 409.
N
dx.doi.org/10.1021/op500078y | Org. Process Res. Dev. XXXX, XXX, XXX−XXX