Development of a cyclosporin A derivative with excellent anti-hepatitis C virus potency

Article abstract of DOI:10.1016/j.bmc.2017.09.008

Synthetic modification of cyclosporin A at P3-P4 positions led to the discovery of NIM258, a next generation cyclophilin inhibitor with excellent anti-hepatitis C virus potency, with decreased transporter inhibition, and pharmacokinetics suitable for coad

Full text of DOI:10.1016/j.bmc.2017.09.008

Accepted Manuscript
Development of a cyclosporin A derivative with excellent anti-hepatitis C virus
potency
Jiping Fu, Christopher Becker, Li Cao, Michael Capparelli, Regis Denay, Roger
Fujimoto, Yu Gai, Zhaobo Gao, Christian Guenat, Subramanian Karur,
Hongyong Kim, Weikuan Li, Xiaolin Li, Wei Li, Thomas Lochmann, Amy Lu,
Peichao Lu, Alexandre Luneau, Nicole Meier, Wosenu Mergo, Simon Ng,
David Parker, Yunshan Peng, Bernard Riss, Alexey Rivkin, Silvio Roggo,
Harald Schroeder, Friedrich Schuerch, Robert L. Simmons, Feng Sun, Zachary
K. Sweeney, Meiliana Tjandra, Michael Wang, Ruidong Wang, Andrew H.
Weiss, Nicolas Wenger, Quanbing Wu, Xin Xiong, Su Xu, Wenjian Xu,
Aregahegn Yifru, Jibin Zhao, Jianguang Zhou, Christian Zürcher, Fabrice
Gallou
PII:
S0968-0896(17)31223-3
DOI:
http://dx.doi.org/10.1016/j.bmc.2017.09.008
BMC 13968
Reference:
To appear in:
Bioorganic & Medicinal Chemistry
Received Date:
Revised Date:
Accepted Date:
11 June 2017
21 August 2017
6 September 2017
Please cite this article as: Fu, J., Becker, C., Cao, L., Capparelli, M., Denay, R., Fujimoto, R., Gai, Y., Gao, Z.,
Guenat, C., Karur, S., Kim, H., Li, W., Li, X., Li, W., Lochmann, T., Lu, A., Lu, P., Luneau, A., Meier, N., Mergo,
W., Ng, S., Parker, D., Peng, Y., Riss, B., Rivkin, A., Roggo, S., Schroeder, H., Schuerch, F., Simmons, R.L., Sun,
F., Sweeney, Z.K., Tjandra, M., Wang, M., Wang, R., Weiss, A.H., Wenger, N., Wu, Q., Xiong, X., Xu, S., Xu, W.,
Yifru, A., Zhao, J., Zhou, J., Zürcher, C., Gallou, F., Development of a cyclosporin A derivative with excellent anti-
hepatitis C virus potency, Bioorganic & Medicinal Chemistry (2017), doi: http://dx.doi.org/10.1016/j.bmc.
2
017.09.008
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Development of a cyclosporin A derivative with excellent anti-
hepatitis C virus potency
o,*
o
‡
o
#
o
Jiping Fu, Christopher Becker, Li Cao, Michael Capparelli, Regis Denay, Roger Fujimoto,
‡
‡
#
o
‡
‡
Yu Gai, Zhaobo Gao, Christian Guenat, Subramanian Karur, Hongyong Kim, Weikuan Li,
o
‡
#
‡
o
#
Xiaolin Li, Wei Li, Thomas Lochmann, Amy Lu, Peichao Lu, Alexandre Luneau, Nicole
ᵻ
o
#
o
o
ᵻ
Meier, Wosenu Mergo, Simon Ng, David Parker, Yunshan Peng, Bernard Riss, Alexey
Rivkin, Silvio Roggo, Harald Schroeder, Friedrich Schuerch, Robert L. Simmons, Feng Sun,
Zachary K. Sweeney, Meiliana Tjandra, Michael Wang, Ruidong Wang, Andrew H. Weiss,
Nicolas Wenger, Quanbing Wu, Xin Xiong, Su Xu, Wenjian Xu, Aregahegn Yifru, Jibin
o
#
#
ᵻ
o
‡
o
o
‡
‡
o
ᵻ
‡
‡
‡
o
o
‡
‡
ᵻ
ᵻ,*
Zhao, Jianguang Zhou, Christian Zürcher, Fabrice Gallou
o
Novartis Institutes for BioMedical Research, 5300 Chiron way, Emeryville, California 94608,
United States
#
Novartis Institutes for BioMedical Research, Novartis Campus, CH-4056 Basel, Switzerland
ᵻ
Chemical and Analytical Development, CH-4002 Basel, Switzerland
‡
Chemical & Analytical Development, Suzhou Novartis Pharma Technology Company Limited,
Changshu, Jiangsu 215537, China
Corresponding authors: e-mail address: fabrice.gallou@novartis.com; tel.: +41 61 324 9911.
jiping.fu@novartis.com; tel.: +1 (510) 879-9593.
Abstract:
Synthetic modification of cyclosporin A at P3-P4 positions led to the discovery of
NIM258, a next generation cyclophilin inhibitor with excellent anti-hepatitis C virus potency,
with decreased transporter inhibition, and pharmacokinetics suitable for coadministration with
other drugs. Herein is disclosed the evolution of the synthetic strategy to from the original
medicinal chemistry route, designed for late diversification, to a convergent and robust

development synthesis. The chiral centers in the P4 fragment were constructed by an
asymmetric chelated Claisen rearrangement in the presence of quinidine as the chiral ligand.
Identification of advanced crystalline intermediates enabled practical supply of key
intermediates. Finally, macrocyclization was carried out at 10% weight concentration by a
general and unconventional “slow release” concept.
Keywords:
Cyclosporin A derivative, Hepatitis C Virus, Cyclophilin inhibitor
Introduction and background
Cyclosporin A (CsA) was initially discovered at Sandoz while searching for novel
1a
1b
antifungal agents.
It was later found to have immunosuppression properties and its
1
c
introduction revolutionized the field of organ transplantation.
Cyclosporin has since found
2
additional applications in the treatment of psoriasis, rheumatoid arthritis, and uveitis. The
immunosuppressive effects of cyclosporin can be attributed to the formation of a ternary
complex between the compound, cyclophilin A, and the two subunits of the phosphatase
calcineurin. Formation of this complex decreases the production of inflammatory cytokines by
3
T-lymphocytes.
Cyclosporin can be synthetically modified to decrease its immunosuppressive properties
4
while retaining cyclophilin binding. Such non-immunosuppressive cyclophilin inhibitors have
5
been explored for the treatment other diseases. In particular, alisporivir, NIM811, and SCY-635
6
(Figure 1) have demonstrated clinical efficacy in patients infected with hepatitis C virus (HCV).
Alisporivir was the first oral non-immunosuppressive cyclophilin inhibitor to enter clinical trials,
where administration alone or in combination with pegylated-interferon α resulted in a

7
significant decrease in viral RNA. As host-targeted agents, cyclophilin inhibitors maintain
activity against all major HCV genotypic variants, as well as exhibit a much higher barrier to
8
resistance development than most direct-acting antiviral agents. In a subset of patients,
however, alisporivir was associated with an increase in circulating levels of conjugated and
7
9
unconjugated bilirubin. This hyperbilirubinemia was attributed to inhibition of transporters.
We therefore undertook a program to address the transporter liability associated with alisporivir.
Figure 1. Cyclosporin related active pharmaceutical ingredients (API): cyclosporin A (CsA),
NIM811, alisporivir, and SCY-635.

Medicinal Chemistry background
Figure 2. Synthetic modifications on CsA led to NIM258.
Since its discovery, cyclosporin A has been subjected to extensive synthetic
1
0
modifications. Our medicinal chemistry approach was therefore defined by existing structure-
1
1
activity relationships and largely focused on P3-P4 modification (Figure 2).
An extensive
survey of Novartis cyclosporin analogs containing P3 groups has revealed that small, non-
branched groups such as CH or CH OH increase cyclophilin binding affinity. At the P4
3
2
12
position, branching enhances potency as well as decreases calcineurin interaction.
13
Furthermore, polar groups could reduce MDR1 inhibition by cyclosporin, while decreasing
lipophilicity can improve pharmaceutical properties, such as solubility and plasma protein
binding.
In an effort to maintain antiviral potency comparable to alisporivir, while decreasing
transporter activities, we focused on analogs with D-alanine at the P3 position and a branched P4
with appended polar groups. Our initial survey of polar groups included hydroxy, alkoxy, cyano,
sulfone, and amines. While most analogs retained cyclophilin binding affinity and cellular

potency, only the amine analogs achieved a substantial reduction in OATP1B1 inhibition.
Further optimization on the amine moiety led to the discovery of NIM258, a next generation
cyclophilin inhibitor with excellent anti-HCV potency, decreased transporter inhibition, and
pharmacokinetics suitable for coadministration with other drugs.
Retrosynthetic analysis
1
1
NIM258 is derived from the fermentation product cyclosporin A via hemi-synthesis.
Two strategies were utilized during the development of the compound. The first one relied on
late introduction of the side-chain 3 at late-stage stage through reductive amination with
aldehyde 2, a strategy well amenable to medicinal chemistry derivatization (path A). The second
strategy emerged once the lead compound had been selected and relied on the convergent
construction of the dipeptide 5, that was directly coupled with nonapeptide 4 and then cyclized to
provide NIM258 (1) (path B). Both strategies were enabled by the practical access to advanced
1
4
crystalline intermediates identified in an earlier efforts. The dipeptide fragment 5 synthesis
also turned to be a major endeavor, as will be discussed in more details in the next section.

Scheme 1. Original and final retrosynthetic analysis
Synthesis of the dipeptide fragment
In the initial synthesis of dipeptide 5, the chiral centers in P4 were introduced in an
asymmetric organocatalytic propionaldehyde 8 syn-addition to an activated imine 7 mediated by
(2S,4R)-4-(tert-butyldiphenylsilyloxy)-pyrrolidine-2-carboxylic acid 9 to give aldehyde 10 with
modest diastereoselectivity (70-90% d.r.) and good enantioselectivy (>95% e.e. for the major
1
5
diastereoisomer).
Reduction of aldehyde resulted in alcohol 10, which was obtained in
isomerically pure form after crystallization. Subsequent silylation of alcohol and N-methylation
provided protected amine 11. Removal of Boc group, coupling with Boc-protected D-alanine
provided 13 in 90% yield over two steps. The furan was then oxidatively cleaved to result first
in the alpha-keto acid, subsequently cleaved to the desired acid, later masked as the benzyl ester

1
4. Desilylation of the TBDPS-protected precursor 14, followed successively by oxidation,
reductive amination afforded 15. Benzyl ester was then removed under hydrogenation condition
to give key dipeptide 5 later used in our final assembly.
This long linear sequence was sufficiently expedient and productive for small quantities
of dipeptide, but scale-up issues were soon encountered and especially envisioned for larger
scale. For example, formation of the activated imine was extremely slow as was the
organocatalytic propionaldehyde addition. The selectivity was at best moderate. Cleavage of the
furan was particularly undesired from a health, safety and environment standpoint. Most
intermediates were oils, which led to challenging purifications. This resulted in an overall low
yield (<1.5%).
Conditions: a) NH Boc, PhSO Na, HCOOH, MeOH, H O; b) K CO , Na SO , THF, 50 % over
2
2
2
2
3
2
4
2
steps; c) propionaldehyde, (2S,4R)-4-((tert-butyldiphenylsilyl)oxy)pyrrolidine-2-carboxylic
acid, MeCN, 70-90% dr, >95% ee of major diastereomer; d) NaBH , MeOH; e) TBDPSCl,
4
imidazole, DMF, 40% yield over 3 steps; f) NaH, MeI, THF, DMF, 80% 2 steps; g) HCl,
dioxane; h) 12, HATU, DIPEA, 90% yield over 2 steps; i) RuCl , NaIO , heptane, EtOAc, water;
3
4
j) H O , NaHCO ; k) BnOH, EtOAc, water, 87% crude yield over 2 steps; l) TBAF, THF; m)
2
2
3
Py·SO , DIPEA, DMSO; n) 3, NaBH(OAc) , AcOH, DCE, 42% yield over 3 steps; o) H ,
3
3
2
Pd(OH) , 71 % yield.
2

Scheme 2. Original route to the dipeptide 5.
We therefore envisioned a de novo synthesis based on more readily scalable chemistry,
1
6
and on high control of the dia and enantio-selectivity. The sequence started with glycine 16.
The terminal amine was masked as the trifluoroacetamide 17, and the crotyl ester 18 was
prepared. An asymmetric chelated Claisen rearrangement in the presence of chiral ligands
(quinidine) was then performed. Although a significant cost-driver, quinidine could easily be
recovered and proved to be the best and most expedient ligand for optimal selectivity, providing
in a straight manner good enantiomeric and diastereoisomeric excess (94.2% ee, 88.4% dr).
Chiral upgrade was achieved via recrystallization with (R)-(+)--methyl benzyl amine and the -
amino ester 19 was finally obtained with >98% ee, >96% de and >98% chemical purity after the
rearrangement and benzylation and N-methylation. The second amino acid 11 was introduced
using propyl phosphonic anhydride (T P) as the coupling reagent, to allow for minimal amount
3
of epimerization. A vinyl moiety was used as an alternative to furan and masked aldehyde. The
advantage was the more convenient unmasking via a simple ozonolysis, conducted in the
presence of excess propionaldehyde to prevent over-oxidation. While purification of product 5
was previously carried out by column chromatography, this could now be avoided after the
introduction of the desired piperazine and direct crystallization of the resulting intermediate.
Other methods of introduction of the piperazine via nucleophilic substitution strategies were
evaluated but all led to lower yield mostly due to -elimination, or epimerization. Overall, a
more efficient and asymmetric synthesis of the fully functionalized dipeptide in 10 steps had
been demonstrated.

Conditions: (a) TFA, Tf O, 81% yield; (b) oxalyl chloride, DMF, DCM; crotyl alcohol, 68%; (c)
2
5
.5 equiv. LiHMDS, 1.1 equiv Al(OiPr) , 2.5 equiv quinidine, THF, then recrystallization with
3
(
R)-(+)--methyl benzyl amine, 23% overall; (d) BnBr, K CO , DMF, 92%; (e) MeI, K CO ,
2
3
2
3
DMF, 95%; (f) NaBH , MeOH, 70%; (h) 11, T P,Et N, 78%; (i) ozone, DCM; (j) 3,
4
3
3
NaBH(OAc) , AcOH, 40% yield 2 steps; (k) H , Pd(OH) , 71%, >98% ee, >98% de.
3
2
2
Scheme 3. Revised route to the dipeptide 5.
This chromatography-free revised strategy resulted in a ca. 2% overall yield, high optical
purity, and enabled rapid and reliable access to the targeted dipeptide and proved suitable for
early stages of development, including the first multi-kg supply.
Preparation of key crystalline intermediate
Earlier efforts on similar scaffolds had showed the importance of crystalline
1
4
intermediates, and more particularly or tetrafluoroborate salt of complex acyclic polypeptides.
Such compounds not only facilitate the rapid access to advanced intermediates, but also enabled

efficient scale-up for supply in development. We again endeavored at finding such crystalline
compounds.
Tetrafluoroborate salt of basic amine acyclic polypeptide 22 had resulted in crystalline
products in a related synthesis, and we had capitalized on this to generate our starting point for
this project. Initial attempts to identify rapidly more advanced crystalline compounds failed.
However, we noticed some transient crystallinity on the rearranged depsipeptide 4, an unusual
potential intermediate that could nevertheless serve our purposes. Crystalline 4 would greatly
help to access a readily hydrolyzable ester, on the way to the coupling fragment. No other
intermediate up to this compound displayed any crystallinity. We therefore constructed our
synthetic strategy through this crystalline intermediate, and hoped to develop chemistry as
selective and clean as possible, to avoid generating further process-related impurities.
Scheme 4. From the advanced crystalline intermediate to the final acyclic intermediate 4

Conditions: (a) Fmoc N-hydroxysuccinimide ester, Na CO , water/toluene; (b) NaBH , glycine,
2
3
4
MeOH/toluene; (c) MeSO H, iPr-OH, then Pyridine, Ac O, 52 % over 3 steps; (d) tris(2-
3
2
aminoethyl)amine, toluene, then HBF , MeOH/water, quantitative yield.
4
The sequence starting first with Fmoc protection, followed by reduction, acidic
rearrangement, and capping as an acetamide turned to be the most selective sequence. A
scalable and reliable protocol was developed that led to crystalline material of high purity > 92%.
The critical point for the crystallization was to carefully control the water content during the
crystallization from a well-defined wet MeTHF/TBME solvent system. For the earlier steps, the
dichloromethane previously used in the Fmoc protection step could be substituted by toluene
when working under biphasic conditions and resulted in the generation of compound 23 in
quantitative yield, and purity of ca. 90%, the product being used directly as solution in toluene
for the next step. The reduction to 24 turned out to be very complex, with competing deacylation
(product 26) and elimination (product 27) (see Figure 3). We required here selective methyl
ester reduction in the presence of an acetate. Des-acetylation and elimination were observed but
could be minimized by using sodium borohydride in THF and toluene with glycine. While
numerous control experiments were done to fully understand the role of glycine, it is still not yet
fully clear how it affects positively the outcome of the transformation. It is nevertheless
unambiguous that it acts as a buffer in our case. The product was finally obtained as a solution
in iPrOH and used as such in the next step. The purity of 24 had nevertheless dropped
significantly to <80%. It is worth pointing out that activation with some methanol was critical to
favor a smooth and complete reduction. One potential rationale is that a more reactive
dimethoxy-borohydride species is formed in situ. Additional parameters which caused some

practical challenges were the hydroxide content and particle size of the sodium borohydride, that
varied with the lots. These challenges could be partly solved when dissolving the powder fully.
The best solution was found with glymes, which we however did not consider as a viable option
due the severe long-term limitations coming from undesirable reprotoxicity and SVHC
1
7
restriction. Ultimately, tight specifications on particle size and minimum sodium hydroxide
content were introduced to ensure robustness in the process.
des-acetyl impurity
elimination impurity
Figure 3. Major impurities from the reduction step.
The subsequent rearrangement step turned out to be challenging from a robustness and
scalability standpoint. This step required very careful monitoring of the pH and temperature to
promote rearrangement of the amino-ester into the N-Acetyl protected derivative 25. Proper
stirring proved crucial to generate reliably material of high purity (87-92%), in acceptable yields
(55%). We rationalized that local variation of pH should be minimized, hence the importance of
good stirring. Fmoc removal was then carried out under standard conditions using tris-
ethylamine, followed by HBF salt formation to result in the crystalline tetrafluoroborate salt 4 in
4
9
2% yield and high purity (92-95%). To our delight, we had identified a chromatography-free

process up to this key intermediate which allowed us to generate large quantities in a streamlined
manner This crystalline salt was again pivotal for practical access to advanced acyclic
derivatives of cyclosporin A as we could tolerate modest purity throughout the sequence and
purify at the end of the sequence by crystallization.
Final assembly to the API
At this stage, whether for the early multigram lots or the multi-kilogram ones, two last
key bonds remained to be formed, whether on the partial dipeptide 28 utilized in the early lots, or
on the fully functionalized dipeptide 5. In both cases, the chances to easily purify the
intermediates were deemed as very low, as it would be experienced ultimately. Indeed, no
crystalline acyclic polypeptide could be obtained, and direct crystallization of the crude
macrocycle turned out to be unfruitful despite major efforts. For path A, the decapeptide 4 was
first coupled to dipeptide 28 under standard HATU conditions to result in dodecaptide 29 in 70%
yield. We then set the stage for the macrocyclization after palladium-mediated silane cleavage
of the benzyloxy carbamate and mild ester hydrolysis of the terminal fragment ethylene glycol
methyl acetamide to 30. Cyclization was carried out under dilute conditions using BOP and
DMAP and resulted in macrocyclic polypeptide 31 in 58% yield for the three steps. The
synthesis was completed by fluoride silyl ether deprotection, oxidation to an aldehyde and
reductive amination with the piperazine moiety 3, followed by basic hydrolysis of the acetate in
2
0% overall for the 8 final steps.

Conditions: (a) HATU, DIPEA, DCM, 70%; (b) Pd(OAc) , TEA, Et SiH, DCM; (c) NaOH,
2
3
THF, water; (d) BOP, DMAP, DCM, 58% over 3 steps; (e) TBAF, THF; (f) Py·SO , DIPEA,
3
DMSO, DIPEA; (g) 1-(2-methoxyethyl)piperazine, AcOH, NaBH(OAc) , DCE; (h) Me NOH,
3
4
MeOH, 48% over 3 steps.
Scheme 5. End-game for path A.
Such a lengthy sequence for an overall mediocre isolated proved particularly costly and
sub-optimal for the long-term. We therefore rapidly turned our attention to the direct
introduction of the fully decorated dipeptide 5 obtained via either the original synthesis (scheme
2
) or the revised one (scheme 3). Incorporation of the dipeptide 5 worked optimally when
o
utilizing HATU as the activating agent added in solution in DMF at -10 C in 1-2 h to a solution
of both peptides and NMM in dichloromethane. This resulted in crude material 32 that was
purified by chromatography in toluene/ethanol/ ammonium hydroxide to afford the product in
8
2% yield and 91% purity for the supply of the first few grams. Later, the crude material was
used as such for the cyclization.

Conditions: (a) HATU, NMM, CH Cl /DMF, 82%; (b) H SO , MeOH/toluene, then
2
2
2
4
BnNMe OH, MeOH, 75%; (c) Cl-HOBT, DCC, Li CO , toluene/TBME (3:1), 74 % crude yield.
3
2
3
Scheme 6. End-game to the API via path B.
The subsequent ester hydrolysis was triggered under acidic transesterification with
sulfuric acid in methanol, followed by very mild caustic hydrolysis of both methyl ester and
hydroxyl acetate functionality utilizing benzyltrimethylammonium hydroxide. The reaction
proceeded quantitatively to 33 and resulted in material of ca. 78% purity.
The final macrocyclization was originally conducted with 6-chloro-1-
hydroxybenzotriazole (Cl-HOBT) and dicyclohexylcarbodiimide (DCC), under a traditional slow
addition of the acyclic polypeptide in a very dilute mixture of dichloromethane and toluene. The
slow addition over 10 h at 30 °C minimizes accumulation of the activated intermediate and the
1
8
resulting competing oligomerization (concentration ca. 0.5% wt) while maintaining the low

concentration necessary for minimal formation of oligomer, which is controlled to a total level
below 3%. This strategy enabled the generation of the first few grams but rapidly showed
limitations upon scale-up. It indeed requires large amount of solvent and particularly of the
undesirable methylene chloride, long overall processing time, which contribute to major
1
8a
expenses. The seminal review by Yudin reports some of the recent and elegant approaches to
solve the entropic challenge associated with such a disfavored macrocyclization event. Another
elegant alternative approach relied on the pre-organization of the acyclic precursor into the
conformation favoring macrocyclization via either complexation or constraining the system into
19
a more desirable conformation. Such an approach was however not pursued as it requires
substantial experimentation with a rather uncertain success. Our preferred approach was based
on early successes on related macrocyclic systems, where we engineered the process to follow a
dual slow-release / slow-addition process. Indeed, we targeted avoidance of the accumulation of
the acyclic precursor, which would lead to oligomers at too high concentration, via the standard
slow-addition technique, but also via an uncommon slow-release concept. A prerequisite was
obviously to accelerate the reaction to minimize the required addition time. This was achieved in
a traditional manner by standard screening of reagent conditions that confirmed
dicyclohehylcarbodiimide (DCC) and 6-Chloro-Hydroxybenzotriazole (Cl-HOBt) to be the
optimal reagent combination.
The main challenge was the choice of reaction medium (solvent composition and ratio).
Out of an extensive variety of solvents (toluene, xylenes, anisole, TBME, iPr O, MeTHF,
2
EtOAc, iPrOAc, acetone, acetonitrile) and combinations thereof, toluene and TBME in a 3 to 1
ratio by volume were demonstrated to both promote the transformation and ensure a proper
dissolution profile of the acyclic polypeptide precursor. The ratio proved critical to ensure
suitable outcome (yield > 70%, total amount of oligomers < 5%), although sufficient tolerance

for robustness was observed. We indeed demonstrated that the toluene/TBME ratio could vary
from 1:2 to 1:4 with minimal variability in the outcome. A more detailed report of our
screening, and design of experiment approach to identify this set of conditions will be reported
shortly. It was besides critical that 33 displays suitable physical properties in the medium. The
suspension of 33 in toluene itself tended for example to aggregate, something that would render
the large scale dosing unmanageable. It also tended to oil out if local supersaturation was
reached in the course of the addition, as observed when we added 33 in TBME directly into
Toluene.
Base, concentration, addition time, temperature, and additive effect were then optimized
in a standard way to ultimately result in the final conditions, namely addition of compound 33 in
suspension in toluene/TBME (3/1) in 5 h, to a mixture of DCC (2.5 eq), Cl-HOBT (0.15 eq),
-
1
o
Li CO (3 eq) (overall concentration 2.2 mol.L ), at 40 C, which resulted in 74% isolated yield,
2
3
and below 3.5% total oligomers. Such an approach was later proven to be general in terms of
scope of polypeptide that undergoes macrocyclization, which will be thoroughly reported from a
process standpoint in a coming venue.
Overall, this new process allowed us to increase the concentration of the
macrocyclization from 0.5% of acyclic starting material by weight to ca. 10%, with the same
yield, and selectivity (ca. 3% oligomers, ca. 5% epimerization). The main process-related
impurities resulted from epimerization (epi-1 in Figure 4) or oligomerization.
.

Epimer (epi-1)
Figure 4. Epimeric impurity generated during the macrocyclization step.
Purification of the drug substance
With the crude Drug substance in hand, the first and most expedient option relied on
purification by reverse-phase column chromatography. The crude material was thus purified by
preparative chiral column chromatography and resulted in material of 93% purity in 45% yield,
and was further purified by RP-LC on Daiso-RP-8 column to upgrade the quality to 98.5% and a
recovery of 78%. This strategy proved useful to supply the first few grams but rapidly showed
limitations in terms of productivity and required investment of substantially large columns and
high flow preparative systems. Thus, we tried to identify (a) crystalline form(s) and develop
robust processes enabling the practical generation of high quality drug substance by
crystallization. We aimed to utilize the basic piperazine side-chain for salt formation. A rapid
solid state screening identified the fumarate salt and the free form exhibited sufficient
crystallinity to be built into our purification strategy. The best outcome was obtained when the
crude was first submitted to formation of the fumarate salt in acetonitrile/water which could
upgrade the quality from ca. 90% to ca. 97%, in 85 to 90% yield, although in a very slow
crystallization process (5 days). Subsequent free base formation and recrystallization in
TBME/heptane allowed for the isolation of a transient crystalline form, that becomes amorphous
upon drying. The process nevertheless allowed us to increase the purity to > 99% with ca. 80%
yield. This novel process, although long overall, enabled the larger scale production of the
desired drug substance in high quality.

o
Conditions: (a) fumaric acid (1.0 eq), acetonitrile/water, 55 C to rt; (b) K CO , MeTHF, then
2
3
TBME/water/heptane, 80 % overall yield.
Scheme 7. Purification to the API.
Conclusion
Overall, we have disclosed the evolution of the synthetic strategy to the target compound
NIM258 (1) from the original medicinal chemistry route, designed for late diversification, to the
development approach, built for increased convergency, expediency, and robustness. Critical to
this was the identification of advanced crystalline intermediates that enabled practical supply of
advanced intermediates and of the drug substance, and an unprecedented macrocyclization
approach to the best of our knowledge, that allowed for high throughput.
Experimentals
1
13
H/ C NMR spectra were recorded on 400’54 ascend purchased from Bruker Biospin
AG, operating at 400/100 MHz, respectively. Chemical shifts (δ) are reported in parts per
million (ppm) downfield from tetramethylsilane (TMS), which was used as internal standard.

The LC-HRMS analyses were performed by using electrospray ionization in positive ion
modus after separation by liquid chromatography (Nexera from shimadzu). The elemental
composition was derived from the mass spectra acquired at the high resolution of about 30’000
on an LTQ Orbitrap XL mass spectrometer (Thermo Scientific). The high mass accuracy below
1
ppm was obtained by using a lock mass.
Reaction solvents were obtained commercially, and used without further purification.
Commercial reagents were used as received. Reaction were monitored by thin-layer
chromatography (TLC) on 0.25mm precoated Merck Silica Gel 60 F254, visualizing with
ultraviolet light.
N-(2,2,2-Trifluoroacetyl)glycine (17)
To a 30 L reactor equipped with a mechanical stirrer, a thermometer, a reflux condenser, an
addition funnel, and nitrogen inlet-outlet was charged TFA (2 L), glycine 16 (2.0 kg, 26.6 mol)
o
with stirring while keeping the inner temperature below 0 C resulting in a suspension. Then to
the suspension was added dropwise TFAA (6.6 kg, 31.4 mol) with stirring while keeping the
o
o
inner temperature below 70 C. After addition, the reaction mixture was kept at 70 C and stirred
for another 2 h until all solid was dissolved. Completion was demonstrated by disappearance of
the starting material by HPLC after 2 h. The mixture was concentrated to dryness under vacuum
o
at 50 C to obtain the crude 17 as a light yellow solid. The crude material was purified by re-
crystallization from EtOAc (6 L) and heptane (12 L) to afford pure 17 (3.7 kg, 81 %) as a white
1
13
solid: mp 120-121 °C. H NMR (400 MHz, D O) δ 4.01 (s, 2 H). C NMR (101 MHz, D O) δ
2
2
−
4
0.86, 117.08, 159.37, 171.66; HRMS (ESI) [C H F NO −H] ) calcd 170.0065, found 170.0044.
4 4 3 3
(2E)-But-2-en-1-yl N-(2,2,2-trifluoroacetyl)glycinate (18)

To a 30 L reactor equipped with a mechanical stirrer, a thermometer, a reflux condenser, an
addition funnel, and nitrogen inlet-outlet was charged CH Cl (20 L), 17 (3.4 kg, 19.6 mol),
2
2
DMF (0.01 L) with stirring to afford a clear solution. And then to this solution was added
dropwise oxalyl chloride (2.8 kg, 21.6 mol) for 30 min at 15 °C under nitrogen protection. After
o
addition, the reaction mixture was stirred at 40 C for 4 h. Completion was demonstrated by
disappearance of the starting material by HPLC after 4 h. The solvent was removed under
vacuum at 40 °C to afford 2-(2, 2, 2-trifluoroacetamido) acetyl chloride (3.7 kg, 100%) as a
brown oil which was directly used in the next step without purification.
To a stirred solution of crotyl alcohol (1.4 kg, 18.7 mol) in CH Cl (20 L) was added dropwise a
2
2
solution of 2-(2, 2, 2-trifluoroacetamido) acetyl chloride (3.7 kg, 19.7 mol) in CH Cl (1 L) at
2
2
2
5 °C over 30 min. After addition, the reaction mixture was stirred at 30 °C overnight. At
completion, the mixture was washed with H O (3 x 5 L) and 5% NaHCO aqueous solution (3 x
2
3
5
L). The separated organic layer was dried over anhydrous Na SO and evaporated to dryness to
2 4
obtain the crude product 18 which was purified by re-crystallization from CH Cl : THF: heptane
2
2
1
(
0.5 L: 0.5 L: 10 L) to obtain purified 18 (3.0 kg, 68%) as a white solid: mp 48-49 °C; H NMR
(
400 MHz, CDCl ) 1.70 (br d, J = 6.36, 3 H), 4.08 (br d, J = 5.26, 2 H), 4.58 (br d, J = 6.48, 2
3
1
3
H), 5.45 - 5.65 (m, 1 H), 5.72 -5.90 (m, 1 H), 7.32 (br s, 1 H); C NMR (101 MHz, CDCl )
3
−
1
7.53, 41.21, 66.62, 117.01, 123.91, 132.78, 157.57, 168.07; HRMS (ESI) [C H F NO −H]
8 10 3 3
calcd 224.0535, found 224.0545.
(2S,3R)-3-Methyl-2-(2,2,2-trifluoroacetamido)pent-4-enoic acid
To a solution of 18 (4.9 kg, 21.7 mol), Al(OiPr) (4.9 kg, 23.0 mol) and quinidine (16.2 kg, 54.4
3
mol) in THF (192.0 kg) was added dropwise a solution of LiHMDS in THF (1 M, 93.0 kg) at -78

°
C for 3 h. After addition, the reaction mixture was then warmed to 15-25 °C with stirring for 12
h. Completion was demonstrated by disappearance of the starting material by HPLC after 12 h.
20.0 kg of 15% KHSO aqueous solution was added dropwise to the reaction mixture at 0-15
4
4
°
C (pH <2). The aqueous layer was extracted with EtOAc (2 x 75 kg). The combined organic
phase was washed with 64 kg of 15% KHSO aqueous solution, and then concentrated under
4
vacuum to dryness to obtain the crude product (1.7 kg, 34.6%) which was directly used in the
next step without purification.
A solution of crude product above (1.7 kg, 7.5 mol) in EtOAc (10.0 kg) was washed with 7%
NaHCO aqueous solution (3 x 6 kg). The combined aqueous layer was extracted with EtOAc (2
3
kg). The aqueous layer was acidified with KHSO (2.5 kg) until pH <2 and extracted with IPAc
4
(3 x 7 kg). The organic phase was washed with 3 kg of 25% NaCl aqueous solution and
concentrated to about 3-5 kg. IPAc (5 kg) was charged to the vessel and concentrated to 3-5 kg
to make sure the water content is below 0.1%. Then the solution was heated to 60 °C and a
solution of R-(+)--methyl benzyl amine (0.9 kg, 7.0 mol) in IPAc (1 kg) was added dropwise
over 1 h. The solution was stirred for 1 h and n-heptane (14.0 kg) was added dropwise at 60-65
°
C. The mixture was stirred at 60-65 °C for 2 h and cooled to 0 °C and stirred at 0 °C for 6 h.
The mixture was filtered and washed with a mixture solution (4 kg) of n-heptane/IPAc=6/1. The
chiral salt was obtained with 99.9 % ee and 96.8 % dr. The solid was then dissolved in IPAc (8
kg) and 15 kg of 10 % KHSO aqueous solution was added. The two layers were separated and
4
the aqueous layer was extracted with IPAc (3 kg). The combined organic phase was washed with
3
kg of 25% NaCl aqueous solution and concentrated to dryness to obtain the purified product
1
(
1.1 kg, 23.1 %) as a yellow liquid: ee 99.1 %; de 96.9 %; H NMR (400 MHz, CDCl ) 1.15 (d,
3
J = 6.97, 3 H), 2.76 - 2.88 (m, 1 H), 4.68 (dd, J = 8.56, 4.89, 1 H), 5.09 - 5.19 (m, 2 H), 5.62 -

1
3
5
4
2
.80 (m, 1 H), 7.00 (br d, J = 8.31, 1 H), 8.84 (br s, 1 H); C NMR (101 MHz, CDCl ) 14.99,
3
−
0.13, 56.11, 117.02, 117.54, 136.99, 157.33, 173.88; HRMS (ESI) [C H F NO −H] ) calcd
8
10
3
3
24.0535, found 224.0510.
Benzyl (2S,3R)-3-methyl-2-(methylamino)pent-4-enoate (19)
To a solution of the above product (1.9 kg, 8.2 mol) in DMF (8 L) was added K CO (1.1 kg, 8.2
2
3
mol) and BnBr (1.4 kg, 8.2 mol) at -10 °C (approximately 2 h) and stirred for 18 h at room
temperature under nitrogen atmosphere. Completion was demonstrated by disappearance of the
starting material by HPLC after 18 h. To the reaction mixture was added K CO (3.4 kg, 24.6
2
3
mol) and CH I (2.3 kg, 16.4 mol) at -10 °C (approximately 2 h) and stirred for 18 h at room
3
temperature under nitrogen atmosphere. At completion, water (2 L) and methyl tert-butyl ether
(
5 L) were added and the mixture was stirred for 1 h. The organic layer was washed with water
3 x 5 L), then brine (3 x 5 L). The aqueous layer successively extracted with methyl tert-butyl
(
ether (3 x 5 L) and the combined organic layer was dried over anhydrous Na SO , filtered, and
2
4
concentrated to afford the crude product (2.6 kg, 95.0%) as a light-brown oil which was directly
used in the next step without purification.
To the above solution (2.0 kg, 6.1 mol) in MeOH (9 L) and CH Cl (9 L) was added NaBH (0.8
2
2
4
o
o
kg, 21.2 mol) slowly at 0 C for 2 h. After addition, the mixture was stirred for 30 min at 0 C.
Completion was demonstrated by disappearance of the starting material by HPLC after 1 h. To
o
the reaction mixture was added dropwise H O (1 L) over 20 min at 20 C with stirring, and then
2
concentrated under vacuum to 20 kg. To the residue was added dropwise an aqueous HCl
o
solution (2.0 N, 7 L) over 20 min at 20 C with stirring. After addition, the mixture was
extracted with EtOAc (2 x 10 L). The pH of the aqueous layer was adjusted to 10 by adding 10%

K CO solution, and then extracted with EtOAc (3 x 10 L). The combined organic layer was
2
3
washed with brine, dried over anhydrous Na SO , and concentrated to afford crude 19 (1.5 kg,
2
4
1
crude, quantitative) as a yellow oil. H NMR (400 MHz, CDCl ) 1.04 (d, J = 6.97, 3 H), 2.26 -
3
2
.40 (m, 3 H), 2.41 - 2.59 (m, 1 H), 3.12 (d, J = 5.87, 1 H), 4.90 - 5.11 (m, 2H), 5.13 - 5.22 (m, 2
1
3
H), 5.58 - 5.84 (m, 1 H), 7.30 - 7.43 (m, 5 H); C NMR (101 MHz, CDCl ) 16.02, 34.79, 41.18,
3
6
6.03, 67.65, 115.38, 128.06, 128.21, 128.30, 135.60, 139.56, 173.80; HRMS (ESI)
+
[
C H NO +H] calcd 234.1494, found 234.1491.
1
4
19
2
Benzyl N-(tert-butoxycarbonyl)-N-methyl-ᴅ-alanyl-(2S,3R)-3-methyl-2-(methylamino)pent-4-
enoate (20)
To a mixture of 19 (1.5 kg, crude from earlier step) and 11 (1.6 kg, 7.7 mol) in EtOAc (15 L)
was added Et N (2.6 kg, 25.7 mol). T3P as a 50% solution in EtOAc (3.1 kg, 9.7 mol) was then
3
o
o
added dropwise over 20 min at 0 C. The reaction mixture was stirred at 10 C overnight.
Completion was demonstrated by disappearance of the starting material by HPLC after 16 h. To
the mixture was added EtOAc (5 L) and then quenched by adding water (2 L) dropwise over 20
o
min at 20 C. The organic layer was washed with H O (2 x 10 L), brine, dried over anhydrous
2
Na SO , and concentrated to afford 20 (2.0 kg, 78.0 %) as a yellow oil which was directly used
2
4
in the next step without purification.
Benzyl
N-(tert-butoxycarbonyl)-N-methyl-ᴅ-alanyl-(2S,3S)-3-methyl-2-(methylamino)-4-
oxobutanoate (21)
To a stirred solution of 20 (1.28 kg, 3.1 mol) in CH Cl (13 L) was added propionaldehyde (355
2
2
o
o
g, 6.0 mol) at 25 C. The reaction mixture was cooled to -78 C, and then to the reaction mixture
o
o
was bubbled O for 40 min at -78 C. The reaction mixture was slowly warmed to -10 C. To the
3

o
reaction mixture was added triethylamine (400 g, 4 mol) over 10 min at -10 C. After addition,
o
the mixture was stirred at -10 C for 1 h, and then washed with NH Cl (2 x 0.3 L). The separated
4
organic layer was dried over anhydrous Na SO and concentrated under vacuum to obtain 21
2
4
(1.29 kg) as a yellow oil which was directly used in the next step without purification.
Benzyl N-(tert-butoxycarbonyl)-N-methyl-ᴅ-alanyl-(2S,3R)-4-[4-(2-methoxyethyl)piperazin-1-
yl]-3-methyl-2-(methylamino)butanoate
To a stirred solution of 21 (882 g, 6.1 mol) in CH Cl (18 L) was added AcOH (294 g, 4.9 mol)
2
2
o
o
and NaBH(OAc) (3.12 kg, 15 mol) at 25 C. The reaction mixture was cooled to -10 C and
3
o
compound 3 (1.28 kg, 3.0 mol) was then added. The reaction mixture was stirred at -10 C for 16
h. Completion was demonstrated by disappearance of the starting material by HPLC after 16 h.
The reaction mixture was quenched with NH Cl (3 L), and washed with NH Cl (2 x 3 L),
4
4
NaHCO (2 x 3 L) and brine (2 x 3 L). The separated organic layer was dried over anhydrous
3
Na SO and concentrated under vacuum to obtain the crude product, which was filtered through
2
4
a plug of silica gel (ca. 1 kg, elution with CH Cl ) to afford the product (651 g, 40 % yield) as a
2
2
1
light yellow oil. H NMR (400 MHz, DMSO-d6) 0.46 - 0.83 (m, 3 H); 0.99 - 1.19 (m, 3 H); 1.23
-
1.52 (m, 9 H); 2.07 - 2.46 (m, 13 H); 2.54 - 3.04 (m, 5 H); 3.21 (s, 3 H); 3.37 - 3.50 (m, 3 H);
1
3
4
.73 - 5.36 (m, 4 H); 7.16 - 7.48 (m, 5 H); C NMR (101MHz, DMSO-d6) 14.28 - 15.72, 21.83
23.62, 28.18 - 28.37, 28.41, 28.55, 29.15, 29.26 - 29.99, 31.26, 47.92 - 52.51, 53.43, 53.25 -
3.47, 53.49 - 53.57, 53.58, 53.69, 55.37, 57.53, 58.44, 60.12, 62.31 - 64.38, 65.41 - 67.13,
0.35, 79.06 - 81.24, 127.90, 128.34 - 128.46, 128.50, 128.85, 128.92, 135.50 - 137.21, 153.98 -
-
5
7
+
1
56.15, 168.79 - 170.36, 170.89 - 173.14; HRMS (ESI) [C29H N O +H] calcd 549.3652,
4
8
4
6
found 549.3621.

N-(tert-Butoxycarbonyl)-N-methyl-ᴅ-alanyl-(2S,3R)-4-[1-(2-methoxyethyl)piperazin-1-yl]-3-
methyl-2-(methylamino)butanoic acid (5)
To a solution of the above product (220.0 g, 401.0 mmol) in EtOH (0.2 L00 mL) was added
Pd(OH) /C 10 wt% (41.2 g). The reaction mixture was stirred at 25 °C under 1 atm of H for 16
2
2
h. Completion was demonstrated by disappearance of the starting material by HPLC after 16 h.
The reaction mixture was filtered and concentrated to dryness. To this residue was added EtOAc
(0.3 L) followed by heptane dropwise (2 L) over 20 min at 10 °C. After addition, the mixture
was stirred at 10 °C for 40 min. The slurry was filtered and concentrated to dryness to obtain 5
1
(
130 g, 71 %) as a white solid, mp 94-96 °C. H NMR (400 MHz, DMSO-d6) 0.73 (br d, J =
5
-
.75, 3 H), 1.05 - 1.21 (m, 3 H), 1.33 - 1.47 (m, 9 H), 2.39 (br s, 2 H), 2.55 - 2.97 (m, 15 H), 3.16
3.28 (m, 3 H), 3.43 - 3.58 (m, 2 H), 3.97 (br d, J = 8.68, 1 H), 4.59 (br s, 1 H), 4.70 - 5.26 (m, 1
13
H), 4.77 - 5.22 (m, 1 H); C NMR (101MHz, DMSO-d6) 14.60 - 16.69, 28.39, 28.44, 28.45 -
2
6
8.49, 28.55, 28.77, 29.21, 29.45, 51.52, 51.62, 51.81, 51.85 - 52.07, 55.86 - 57.01, 8.42, 58.44,
2.30, 65.32, 68.29 - 71.40, 79.25 - 80.26, 154.45 - 155.22, 170.84 - 171.42, 171.67 - 172.56;
+
HRMS (ESI) [C H N O +H] calcd 459.3183, found 459.3159.
2
2
42
4
6
Methyl N-{[(9H-fluoren-9-yl)methoxy]carbonyl}-ʟ-valyl-N-methyl-ʟ-leucyl-ʟ-alanyl-ᴅ-alanyl-N-
methyl-ʟ-leucyl-N-methyl-ʟ-leucyl-N-methyl-ʟ-valyl-(2S,3R,4R,6E)-3-(acetyloxy)-4-methyl-2-
(methylamino)oct-6-enoyl-(2S)-2-aminobutanoyl-N-methylglycinate (23)
To a 3 L double-jacketed reactor at rt charged with sodium carbonate (101.7 g), in water (0.63
kg) and toluene (1.1 kg) was added portionwise solid Methyl ʟ-valyl-N-methyl-ʟ-leucyl-ʟ-
alanyl-ᴅ-alanyl-N-methyl-ʟ-leucyl-N-methyl-ʟ-leucyl-N-methyl-ʟ-valyl-(2S,3R,4R,6E)-3-
(acetyloxy)-4-methyl-2-(methylamino)oct-6-enoyl-(2S)-2-aminobutanoyl-N-methylglycinate—
trifluoroborane—hydrogen fluoride (1/1/1) 22 (0.25 kg, 0.20 mol, ca. 90% purity). The lumps

went into an emulsion within a few minutes of vigorous stirring. (9H-fluoren-9-yl)methyl (2,5-
dioxopyrrolidin-1-yl) carbonate (61.5 g, 0.18 mol) was added portionwise in ca. 30 min, and the
transfer line was rinsed with toluene (0.05 L). The resulting mixture was stirred for 1 h, until
completion of the transformation. Methanol (0.05 L) was added, the emulsion was warmed to 50
°
C and stirred for 1 h. Stirring was discontinued and the layers were separated. The organic
phase was cooled to rt, and a 1 M NaHSO aqueous solution (1 L) was added slowly under
4
gentle stirring (Caution CO -evolution!). Stirring was discontinued and the layers were
2
separated. The organic extract was washed with water (1 L), and the yellowish turbid organic
layer was concentrated at ca. 50 °C and reduced pressure 250 to 500 mbar to a concentrated
solution of ca. 500 g. An aliquot was further dried under high vacuum to provide an off-white
solid and calculate the corresponding yield (278 g by HPLC quantitative assay compared to pure
1
reference, ca. 93.5% purity). C H N O : H NMR (600 MHz, DMSO-d6): one proton is
7
3
114 10 15
missing (broad line ) due to conformer mixture and very broad line proton spectra: one proton
couldn’t be located precisely. 2.29 3H 7.17 2H,7.24 2H, 7.14 1H, Carbon 21.47 1C 129.40
2
1
-
C,128.64 2C and 125.75 1C) 0.50 - 0.97 (m, 36H), 1.09 - 1.20 (m, 6H), 1.30 - 1.46 (m, 4H),
.45 - 1.78 (m,7H), 1.59 (br d, 3H), 1.93 (m, 3H), 2.02 (m, 1H), 2.14 (m, 1H), 2.20 (m, 1H), 2.64
3.14 conformer mixture of singulets (s, 18H), 3.63 and 3.68 (s, 3H), 3.87 (d, J= 17.10, 1H),
.14 - 4.44 (m, 6H), 4.60 - 4.72 (m, 2H), 5.04 (d, J =10.52, 1H), 5.09 (m, 1H), 5.21 - 5.31 (m,
H), 5.31 - 5.48 (m, 6H), 7.31 (m, 2H), 7.40 (t, J = 7.4, 2H), 7.58 (d, J = 7.46, 1H),7.65 (d, J =
4
2
8
1
3
.76, 1H), 7.71 (br, 2H), 7.88 (d, J = 7.62, 2H), 8.04 - 8.28 (m, 2H); C NMR (151 MHz,
DMSO): due to conformers mixture at room temperature, carbon signal could be splitted into 2
different signal, 10.14, 10.25, 16.13, 16.32, 17.51, 18.24, 18.32, 18.72, 19.06, 19.38, 19.56,
2
1.07, 21.49, 21.90, 22.33, 22.78, 23.00, 23.33, 23.57, 24.60 to 24.76, 27.32, 29.95, 30.18,

3
4
7
1
1
0.26, 30.30, 31.22, 32.67, 34.66, 34.84, 35.00, 36.47, 37.01, 38.00, 38.03, 45.18, 17.13, 48.48,
9.69, 48.85, 50.25, 51.21, 51.51, 52.11, 52.51, 54.13, 56.27 (brs), 56.72, 57.85, 66.20, 72.63,
3.47, 74.64, 120.47 to 102.55 (br.s), 121.83, 125.68, 125,86, 126.66, 127.44, 127.49, 127.74,
29.72, 137.80, 141.14, 141.16, 144.35, 156.64, 167.74, 167.91, 169.70, 169.79, 170.01, 170.34,
70.40, 170.98, 171.09, 171.58, 171.79, 172.05, 172.05, 172.47, 173.13; HRMS for
+
[
C H N O +2H] calcd 1372.86106, obsd 1372.86096.
7
3
114 10 15
(2S)-2-[{[(9H-Fluoren-9-yl)methoxy]carbonyl}-ʟ-valyl-N-methyl-ʟ-leucyl-ʟ-alanyl-ᴅ-alanyl-N-
methyl-ʟ-leucyl-N-methyl-ʟ-leucyl-N-methyl-ʟ-valyl-(2S,3R,4R,6E)-3-(acetyloxy)-4-methyl-2-
methylamino)oct-6-enamido]-N-hydroxyethyl-N-methylbutanamide (24)
(
A suspension of sodium borohydride (14.5 g, 0.38 mol) in THF (0.4 L) was added at rt, to a 2 L
double-jacketed reactor equipped with an impeller stirrer, a temperature sensor, an addition
o
funnel and Argon inertization bubbler and the mixture was heated to ca. 45 C. To the slightly
turbid thin suspension cooled to 15 °C was added glycine (14.4 g, 0.19 mol), then the freshly
prepared 50% wt solution of 23 into toluene (276 g, ca. 138 g, 0.10 mol 23) via the addition
funnel in ca. 30 min. At the end of the addition, the transfer line was rinsed with toluene (10
mL). Within 1 h, the reduction started. A mixture of toluene and methanol (0.05 L each) were
added continuously within 3 h, to allow for continuous hydrogen release and controlled gas
release, at which point the reaction reached > 99% conversion, as monitored by HPLC. The
o
reaction mixture was then cooled to 0 C, and added to a second 3 L double-jacketed reactor
equipped with an impeller stirrer, a temperature sensor, an addition funnel and Argon
o
inertization bubbler, charged with water (0.83 L), and acetic acid (69 g) at 0 C. Care should be
taken as gas evolution ensued. The reaction vessel and the transfer line was rinsed with toluene

(0.2 L), and the resulting white emulsion was stirred overnight at rt. Stirring was discontinued,
and the phases were separated. The organic layer was washed twice with water (0.25 L), and
concentrated at ca. 50 °C and reduced pressure to a viscous colorless solution. Isopropyl
alcohol (200 mL) was added and the concentration process was repeated one more time until the
water content was below 1% and toluene < 0.1 % as measured by Karl Fischer titration and GC.
The solution was then diluted further to provide a 67% wt solution of 24 (333 g, ca. 225 g, and
1
8
1
1
3
4
7
0% purity). H NMR (600 MHz, DMSO-d6): 0.50 - 1.00 (m, 36H), 1.10 - 1.21 (d, J = 6.9, 6H),
.31 - 1.43 (m, 5H), 1.44 - 1.73 (m, 6H), 1.60 (br s, 3H), 1.93 (m, 3H), 2.03 (m, 1H), 2.13 (m,
H), 2.21 (m, 1H), 2.68 to 3.10 conformer mixture of singlets (s, 18H), 3.18 and 3.23 (m, 1H),
.45 and 3.59 (m, 1H), 3.47 and 3.54 (m,2H), 4.14 - 4.21 (m, 2H), 4.21 - 4.33 (m, 3H), 4.52 -
.72 (m, 3H), 4.80 (br s, 1H), 5.01 - 5.16 (m, 2H), 5.21 - 5.53 (m, 7H), 7.31 (m, 2H), 7.41 (t, J =
13
.3, 2H), 7.51 - 7.68 (m, 2H), 7.71 (m, 2H), 7.87 (d, J = 7.5, 2H), 8.09 (m, 2H); C NMR (151
MHz, DMSO) due to conformer mixtures at room temperature, carbon signal could be splitted
into 2 different signal. 9.76, 9.93, 15.73, 15.85, 17.07, 17.81, 17.88, 18.27, 18.62, 18.94, 19.10,
1
2
9.13, 20.62 (br s), 21.46, 21.90, 22.33, 22.41, 22.56, 22.86 (br s), 23.13, 24.18, 24.24, 24.32,
4.53, 24.93, 26.87, 29.51, 29.73, 29.82, 29.86, 30.78, 32.24 (br s), 32.44 (br s), 33.62, 34.26 (br
s), 34.40, 34.46, 34.56, 35.97, 37.58 (br s), 41.86, 44.74, 46.69, 48.05 (br s), 49.52, 49.99, 50.13,
5
1
1
1
1
1
0.77, 51.07, 53.70, 55.88 (br s), 56.28, 57.34, 57.41, 58.49, 58.67, 65.77, 74.26, 74.41, 119.85,
19.94, 120.02, 120.09 (br s ), 121.38, 124.16, 125.24, 125.42, 126.29, 126.78, 126.97, 127.00,
27.04, 127.15, 127.29, 127.63, 128.92, 129.18 (br s), 139.91, 140.72 (br s), 143.71, 143.90,
48.64, 156.20, 167.21, 167.28, 169.23, 169.34, 169.91, 169.96, 170.35, 170.57, 170.63, 170.68,
+
70.79, 171.61, 172.03, 172.69; HRMS for [C H N O +NH ] calcd 1360.88543, obsd
7
2
114 10 14
4
360.88538.