DPAGT1 Inhibitors of Capuramycin Analogues and Their Antimigratory Activities of Solid Tumors

DPAGT1 Inhibitors of Capuramycin Analogues and Their

Antimigratory Activities of Solid Tumors

Katsuhiko Mitachi

Rita G. Kansal

Kirk E. Hevener

Cody D. Gillman

Syed M. Hussain

Hyun Gi Yun

Gustavo A. Miranda-Carboni

Evan S. Glazer

William M. Clemons Jr.

Michio Kurosu

Department of Pharmaceutical Sciences, College of Pharmacy, University of Tennessee Health

Science Center, 881 Madison Avenue, Memphis, TN 38163, USA

Department of Surgery and Center for Cancer Research, College of Medicine, University of

Tennessee Health Science Center, 910 Madison St., Suite 300, Memphis, TN 38163, USA

Division of Chemistry and Chemical Engineering, California Institute of Technology, 1200 E.

California Blvd. Pasadena, CA 91125, USA

Department of Medicine, Division of Hematology-Oncology, University of Tennessee Health

Science Center, 19 S. Manassas Avenue, Memphis, TN 38163, USA

Abstract

Capuramycin displays narrow spectrum of antibacterial activity by targeting bacterial translocase I

(MraY). In our program of development of new

-acetylglucosaminephosphotransferase1

(DPAGT1) inhibitor, we have identified that a capuramycin phenoxypiperidinylbenzylamide

analogue (CPPB) inhibits DPAGT1 enzyme with an IC

value of 200 nM. Despite a strong

DPAGT1 inhibitory activity, CPPB does not show cytotoxicity against normal cells and a series of

cancer cell lines. However, CPPB inhibits migrations of several solid cancers including pancreatic

cancers that require high DPAGT1 expression in order for tumor progression. DPAGT1 inhibition

by CPPB leads to a reduced expression level of Snail, but does not reduce E-cadherin expression

level at the IC

(DPAGT1) concentration. CPPB displays a strong synergistic effect with

paclitaxel against growth-inhibitory action of a patient-derived pancreatic adenocarcinoma,

PD002: paclitaxel (IC

1.25 μM) inhibits growth of PD002 at 0.0024–0.16 μM in combination

with 0.10–2.0 μM of CPPB (IC

35 μM).

Graphical Abstract

Corresponding Author:

Michio Kurosu - Department of Pharmaceutical Sciences, College of Pharmacy, University of Tennessee

Health Science Center, 881 Madison Avenue, Memphis, TN 38163, USA; Phone: +1-901-448-1045; Fax: 901-448-6940;

mkurosu@uthsc.edu;.

Supporting Information

This material is available free of charge via the Internet at

http://pubs.acs.org

NMR spectra of the compounds in the Experimental Section, all assay data, HPLC chromatogram of new compounds, assay

procedures, and molecular formula string (CSV)

The authors declare no competing financial interest.

DEDICATION

This article is dedicated to the memory of Dr. Isao Kitagawa, Professor Emeritus of Pharmaceutical sciences at Osaka University, an

inspirational scientist.

HHS Public Access

Author manuscript

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Published in final edited form as:

J Med Chem

. 2020 October 08; 63(19): 10855–10878. doi:10.1021/acs.jmedchem.0c00545.

Author Manuscript

Keywords

Capuramycin analogues; DPAGT1 inhibitor; Antimigratory activity; Snail zinc-finger transcription

factors; E-Cadherin; Synergistic effect; Computational chemistry

INTRODUCTION

Capuramycin (CAP (

), Figure 1) is a nucleoside antibiotic, which has a narrow-spectrum of

activity against Gram-positive bacteria including

Mycobacterium tuberculosis

(Mtb).

–

date, medicinal chemistry efforts on capuramycin have been focused on developing a new

TB drug; new capuramycin analogues have been synthesized to improve bacterial phospho-

MurNAc-pentapeptide translocase I (MraY and MurX for

Mycobacterium

spp.) enzyme as

well as antimycobacterial activities.

Somewhat recently, the anti-

Clostridioides difficile

(formerly

Clostridium difficile

) activity of a capuramycin analogue has been reported.

Capuramycin is a specific inhibitor of MraY with the IC

value of 0.13 μM,

however,

some other nucleoside antibiotics (

e.g

., muraymycin A1 and tunicamycins) display activity

against MraY, WecA (polyprenyl phosphate-GlcNAc-1-phosphate transferase), and its

human homologue, DPAGT1 (

-acetylglucosaminephosphotransferase1)-type

phosphotransferases.

–

MraY is an essential enzyme for growth of the vast majority of

bacteria that catalyzes the transformation from UDP-MurNAc-pentapeptide (Park’s

nucleotide) to prenyl-MurNAc-pentapeptide (lipid I).

WecA catalyzes the transformation

from UDP-GlcNAc to decaprenyl-P-P-GlcNAc, the first membrane-anchored

glycophospholipid that is responsible for the biosynthesis of mycolylarabinogalactan in

Mycobacterium tuberculosis

(Mtb). WecA is an essential enzyme for the growth of Mtb and

some other bacteria. Biochemical studies of WecA enzyme are hampered by lack of

selective inhibitor molecules.

Tunicamycin shows inhibitory activity against these

phosphotransferases with the IC

values of 2.9 μM (MraY/MurX), 0.15 μM (WecA), and

1.5 μM (DPAGT1).

CPZEN-45, an antimycobacterial MraY inhibitor, was reported to

exhibit WecA inhibitory activity (IC

~0.084 μM).

We showed that 2’

-methyl

capuramycin (OM-CAP (

), formerly UT-01320, Figure 1) does not exhibit MraY/MurX

inhibitory activity, but displays a strong WecA inhibitory activity (IC

0.060 μM).

vitro

cytotoxicity of tunicamycin has been documented in a number of articles.

Acute

toxicity of tunicamycin due to its narrow therapeutic window (LD

: 2.0 mg/kg, LD

100

: 3.5

mg/kg mice, IP) discourages scientists from developing tunicamycin for new antibacterial,

Mitachi et al.

Page 2

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript

antifungal, or anti-cancer agents.

A large number of scientists believe that cytotoxicity

of tunicamycin is attributable to its interaction with DPAGT1, which catalyzes the first and

rate limiting step in the dolichol-linked oligosaccharide pathway in

-linked glycoprotein

biosynthesis (Figure 2).

In contrast, OM-CAP (

) possesses the same level of DPAGT1

inhibitory activity (IC

4.5 μM) to tunicamycins (IC

1.5 μM), but it does not show

cytotoxicity against healthy cells (

e.g

., Vero and HPNE cells) at 50 μM and some cancer cell

lines (

e.g

., L1210, KB, AsPC-1, PANC-1, LoVo, SK-OV-3) at 10 μM. A sharp difference in

the cytotoxicity profiles between tunicamycins (

, Figure 1) and OM-CAP (

) raises the

question of whether selective inhibition of DPAGT1 enzyme function in certain cells/organs

by small molecules does not cause unacceptable level of toxicity against human during

chemotherapy. We have recently engineered the structure of muraymycin to yield strong

DPAGT1 inhibitors; one of these analogues, APPB (

, Figure 1), showed a promising

antiproliferative effect on a series of solid cancer cell lines that required overexpression of

the

DPAGT1

gene in their growth and cancer progression.

The design of APPB was

originated from the discovery of DPAGT1 inhibitors of capuramycin analogues. In this

article, we report structure-activity relationship (SAR) studies of capuramycin to identify

novel DPAGT1 inhibitors, and

in vitro

anti-invasion and anti-metastasis activity of a new

capuramycin analogue DPAGT1 inhibitor, CPPB (capuramycin

phenoxypiperidinylbenzylamide analogue,

) (Figure 1). A unique synergistic effect was

observed against a patient-derived pancreatic adenocarcinoma, PD002 in a combination of

CPPB with paclitaxel. We demonstrated key interactions of CPPB with DPAGT1 via

molecular docking studies. Lastly, we report a semi-synthetic method to deliver enough

CPPB for future

in vivo

studies.

RESULTS AND DISCUSSION

Chemistry and Structure activity relationship (SAR).

Sankyo (currently Daiichi-Sankyo) and Sequella have reported several capuramycin (CAP,

) analogues that have improved MraY enzyme and bacterial growth inhibitory activities.

Their SAR studies rely on a capuramycin biosynthetic intermediate, A-500359E,

which allows delivery of novel amide molecules (R

group) having an ester (R

) functional

group (Figure 3). The capuramycin analogue, SQ992, has an interesting antibacterial

characteristic with antimycobacterial activity

in vivo

We have accomplished a total

chemical synthesis of capuramycin and its analogues.

To date, we have identified an novel

analogue possessing improved MraY inhibitory activity, Cap-3-amino-1,4-

benzodiazepine-2-one analogue,

and a selective WecA inhibitor, OM-Cap (

), via our total

synthetic approach (Figure 3).

Since our first report on a total synthesis of capuramycin, a

few improvements of the synthetic scheme have been made: we introduced acid-cleavable

protecting groups for the uridine ureido nitrogen (3-postion in

) and primary alcohol (6”-

position in

) (Scheme 1).

The MDPM and MTPM groups can simultaneously be removed

with 30% TFA to form the key synthetic intermediate

for capuramycin. In our on-going

SAR of capuramycin, commercially available protecting groups (BOM and chloroacetyl

groups) for the ureido nitrogen (3-position) and C6“-alcohol were decided to apply in our

SAR tactics. Success of hydrogenolytic cleavage of the BOM group and tolerability of the

chloroacetyl group were difficult to predict in capuramycin synthesis. To establish the new

Mitachi et al.

Page 3

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript

protecting group strategy, we first demonstrated the synthesis of

iso

-capuramycin (I-Cap)

(Scheme 2A). Syntheses of all building blocks and the experimental procedures for their

synthetic steps in Scheme 2 and 3 are summarized in Supporting Information (SI).

Highlights of syntheses of new capuramycin analogues are illustrated in Scheme 2 and 3.

The (

)-cyanohydrin

was subjected to NIS-AgBF

promoted

-selective mannosylation

with the thioglycoside

to yield the

-mannosylated cyanohydrin

in 78% yield.

The

cyano group of

was hydrated using HgCl

-aldoxime in aq. EtOH, and the BOM and

chloroacetyl groups of the generated amide were deprotected stepwisely:

dechloroacetylation with thiourea followed by hydrogenation with 10% Pd-C in AcOH-

PrOH-THF provided the C6”-free alcohol

in 53% overall yield. Oxidation-elimination

reaction of

with SO

•pyridine in a solvent system (CH

/Et

N/DMSO = 10/2/1)

provided the

unsaturated aldehyde

in quantitative yield (determined by

H NMR

analysis). After all volatiles were removed, the aldehyde

was oxidized to the

corresponding carboxylic acid

by using NaClO

in the presence of NaH

and 2-

methyl-2-butene.

The resulting carboxylic acid

was coupled with (

aminocaprolactam by using a standard peptide-forming reaction condition (HOBt, EDCI,

and NMM) to yield the coupling product

in 70% overall yield from

. Saponification of

by using Et

N in MeOH provided I-Cap (

) in quantitative yield. Similarly, dimethyl-

capuramycin (DM-CAP) was synthesized in 31% overall yield from the cyanohydrin-

acetonide

. CAP (

) and its three analogues, OM-CAP (

), I-CAP (

), and DM-CAP (

)

were evaluated in enzyme inhibitory assays against bacterial phosphotransferases, MraY and

WecA, and archaeal and human dolichyl-phosphate GlcNAc-1-phosphotransferases, AglH

and DPAGT1.

CAP is a selective MraY inhibitor that does not display inhibitory

activity against WecA, AglH, and DPAGT1 (Table 1). Previously, OM-CAP (

) was

identified as a selective WecA inhibitor (IC

0.060 μM) that does not possess MraY

inhibitory activity (IC

>50 μM (entry 2 in Table 1).

In this program, it was realized that

OM-CAP (

) has inhibitory activity against AglH and DPAGT1 with the IC

values of 2.5

and 4.5 μM, respectively (entry 2). I-CAP (

) showed enzyme inhibitory activity against

MraY, WecA, AglH, and DPAGT1 with the IC

between 8.5–30 μM concentrations (entry

3). On the contrary, DM-CAP (

) showed only a weak WecA inhibitory activity (IC

35.0

μM) (entry 4). In recent reports on co-crystal structures of DAPGT1 with tunicamycin

(PDB: 5O5E and 6BW6), the fatty acid chain of tunicamycin is occupied in the hydrophobic

tunnel (the proposed dolichol-phosphate (Dol-P) binding site).

We speculated that

introduction of pharmacologically amenable hydrophobic groups that occupy the proposed

Dol-P binding site is essential to exhibit strong DPAGT1 inhibitory activity. In virtual

screening of a hydrophobic group (excluding fatty acids) using the structure of DPAGT1

with bound tunicamycin (PDB: 6BW6), an analogue (

e.g

(((trifluoromethoxy)phenoxy)piperidinyl)benzylamide) of

(((trifluoromethoxy)phenoxy)piperidin-1-yl)phenol group (

i.e

in Scheme 3) found in a

new anti-TB drug, delamanid,

was suggested to be a reasonable fatty acid surrogate,

whose capuramycin derivatives (capuramycin phenoxypiperidinylbenzylamide (CPPB,

iso

-capuramycin phenoxypiperidinylbenzylamide (I-CPPB,

-methyl capuramycin

phenoxypiperidinylbenzylamide (OM-CPPB,

), demethyl capuramycin

phenoxypiperidinylbenzylamide (DM-CPPB,

), capuramycin

phenoxypiperidinylbenzylamine (CPPA,

), and

iso

-capuramycin

Mitachi et al.

Page 4

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript

phenoxypiperidinylbenzylamine (I-CPPA,

)) could bind to DPAGT1 with high affinity.

However, the docking program used in these studies is not able to distinguish between low-

and high-binding molecules: all (((trifluoromethoxy)phenoxy)piperidin-1-

yl)phenyl)methanamino derivatives provided “good” GlideScores of between -12.9~-16.6

(see Table S1 in SI). Therefore, we decided to synthesize all analogues (CPPB, I-CPPB,

OM-CPPB, DM-CPPB, CPPA, and I-CPPA) identified in the virtual screening and evaluate

their phosphotransferase inhibitory activity.

The carboxylic acid intermediate

for CAP (

) was subjected to peptide coupling reaction

with (((trifluoromethoxy)phenoxy)piperidin-1-yl)phenyl)methylamine (

), providing the

protected CPPB,

. Saponification of

and purification by reverse HPLC yielded CPPB

(

) in 95% overall yield from

. Similarly, I-CPPB (

), OM-CPPB (

), and DM-CPPB (

)

were synthesized from the mannuronic acid derivatives

, and

in 24–34% overall

yield (Scheme 3A). Capuramycin phenoxypiperidinylbenzylamine analogues, CPPA (

) and

I-CPPA (

), were synthesized via reductive aminations of the aldehydes

and

with the

amine

, furnishing the desired products in 63–65% overall yield from

and

after

saponification. MraY, WecA, AglH, and DPAGT1 enzyme inhibitory activity assays for the

six (((trifluoromethoxy)phenoxy)piperidin-1-yl)phenyl)methylamine analogues revealed that

CPPB (

) and I-CPPB (

) are high nM range DPAGT1 inhibitors (entries 5 and 6 in Table

1), however, the

-methylation and demethylation analogues (OM-CPPB (

) and DM-CPPB

(

)) turned out to be very low- or no-DPAGT1 inhibitor (entries 7 and 8). The secondary

amine analogues, CPPA (

) and I-CPPA (

), did not inhibit all phosphotransferases tested

in Table 1 at 50 μM concentration (entries 9 and 10). CPPB was determined to be three times

stronger DAPGT1 inhibitor (IC

0.20 μM) than I-CPPB (entry 5 vs. 6). I-CPPB did not

display MraY inhibitory activity, but showed a very weak WecA and AglH enzyme

inhibitory activity (entry 6). Interestingly, difference in these phosphotransferase inhibitory

profiles between CPPB and I-CPPB correlate with their antimycobacterial activity: CPPB

possessing MraY/WecA inhibitory activity killed

Myocobacterium tuberculosis

Rv,

Mycobacterium avium

2285,

Mycobacterium smegmatis

(ATCC607) with the MIC values

6.25–12.5 μg/mL. In contrast, I-CPPB, which does not have MraY inhibitory activity, did

not show growth inhibitory activity against these

Mycobacterium

spp. at 50 μg/mL.

Cytotoxicity of new capuramycin analogues, CPPB (5) and I-CPPB (6).

In the capuramycin analogue series, the degree of MraY inhibitory activity correlates with

their antimycobaterial activity.

–

Antimycobacterial capuramycin analogues display low

in vitro

cytotoxicity against mammalian cells, and have been recognized as safe drug leads

that have acceptable tolerability in animal models.

The toxicity of tunicamycin (

, Figure

1) has been studied extensively

in vitro

: tunicamycin inhibits growth in many cancer cell

lines without selectivity, and has a narrow therapeutic window demonstrated in

in vivo

studies using mice.

The toxicity of tunicamycin is believed to be attributable to its

ability to inhibit DPAGT1 enzyme function.

However, in our studies, tunicamycin’s

toxicity could not be explained solely by its inhibition of DPAGT1. Our DPAGT1 inhibitor,

APPB (

, Figure 1) inhibits DPAGT1 with greater than 30-times the inhibitory activity of

tunicamycin, and inhibits growth of selected solid cancer cell lines at low μM concentrations

with selective cytotoxicity (IC

normal cells / cancer cells) of >35.

The LD

value of

Mitachi et al.

Page 5

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript

APPB is >20 mg/kg (mouse), whereas it is 2.0 mg/kg (mouse) for tunicamycin.

Capuramycin-based DPAGT1 inhibitors, CPPB (

) and I-CPPB (

), identified in this

program inhibited DPAGT1 enzyme with the IC

values of 0.2 and 0.6 μM, respectively.

Unlike the MraY-antimycobacterial activity relationship observed for CAP analogues, the

DPAGT1 inhibitors, CPPB and I-CPPB, did not show antiproliferative activity against

L1210 (a leukemia cell), HPNE (a normal pancreatic ductal cell), and Vero (a normal kidney

cell) at 50 μM. They showed various levels of growth inhibitory activity against several solid

cancer cell lines such as KB (HeLa, a cervix carcinoma), SiHa (a cervical squamous cell

carcinoma), HCT-116 (a colorectal adenocarcinoma), DLP-1 (a colorectal adenocarcinoma),

Capan-1 (a pancreatic ductal adenocarcinoma), PANC-1 (a pancreatic ductal carcinoma),

AsPC-1 (a pancreatic adenocarcinoma), PD002 (a patient-derived pancreatic

adenocarcinoma) in MTT assays (IC

15–45 μM, Table 2). A lower DPAGT1 inhibitor,

tunicamycin (

), showed growth inhibition of all cell lines in Table 2 with the IC

values

of 0.78–7.5 μM concentrations (entry 5 in Table 2). Cellular behavior and morphological

changes of a patient-derived metastatic pancreatic adenocarcinoma, PD002 treated with

CPPB were monitored over time via IncuCyte® live cell analysis imaging system (Figure

4A). Interestingly, 10–13% of phase area confluent of PD002 culture (time 0h) remained the

same after 72h for the CPPB-treated cells (50 μM), whereas, ca. 70% of confluence was

reached for the control PD002 culture (PBS) (Figure 4B vs.4C). Although morphological

changes were subtle over time (0–72h), cell viability assessed by the MTT reduction assay

revealed that PD002 cells treated with CPPB (50 μM) was significantly decreased (Table 2).

Exposure of CPPB (0.2–20 μM) inhibited cell proliferation of PD002: ca. 20% of cell

proliferation was inhibited at time 72h. These results may indicate that DPAGT1 inhibitors

may have cytostatic effect again certain cancerous tumors that require DPAGT1

overexpression for their growth.

Cell migratory inhibition by CPPB (5).

DPAGT1 catalyzes the first step in

-glycan biosynthesis of mammalian cells (Figure 2).

Aberrant

-glycosylation is common in many solid cancers and important for the epithelial

to mesenchymal transition program (EMT, a mechanism of metastases).

We found high

levels of DPAGT1 protein expression in a series of pancreatic cancers (

e.g

., PANC-1,

Capan-1, and AsPC-1). Dysregulation of DPAGT1 enzyme leads to disturbances in cell-cell

adhesions and may increase epithelial to mesenchymal transition (EMT): these processes

increase migratory and invasive capabilities of malignant neoplasms that are the initiation of

metastasis in cancer progression, especially pancreatic cancer.

Interestingly, there is

significant crosstalk between DPAGT1 and the Wnt/

-catenin and Snail pathway where

DPAGT1 overexpression leads to 1) accumulation of

-catenin in the cytoplasm and then

translocation into the nucleus, and 2) reduction of the Snail expression levels, preventing

epithelial-mesenchymal transition by suppressing the E-cadherin expression (a cell-cell

adhesion glycoprotein).

As such, aberrations in these pathways occur in numerous

cancers, thus, discovery of small molecules directed towards inhibition of the Wnt and Snail

pathways represents an important area of anticancer therapeutics.

–

In order to obtain

insights into anti-metastatic ability of DPAGT1 inhibitors, we explored the degree of cell

migration in several commercially available cell lines (Capan-1, PANC-1, and AsPC-1), a

patient-derived pancreatic ductal adenocarcinoma cell line (PD002), a cervical carcinoma

Mitachi et al.

Page 6

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript

(SiHa), and a colorectal adenocarcinoma (HCT-116) to determine the effects on cellular

motility. After 24h of three treatment doses (0.05, 0.1, and 0.2 μM) with CPPB (

inhibition of migration (closing the gap) was measured in a scratch assay (wound healing

assay).

Only the images for PD002 are shown in Figure 5 and all other images obtained

through these experiments are illustrated in Supporting Information (SI). The pancreatic

cancer cell lines treated with CPPB migrated far less than the PBS treated (control) cells

(Figure 5A–D). In these assays, the wound-healing rate of the untreated PD002 cells was

63% in 24h. In sharp contrast, CPPB treated cells inhibited the wound-healing effectively at

its IC

level against DPAGT1 (0.2 μM): the wound-healing rate was approximately 20%

(Figure 5B). We thoroughly evaluated migration inhibition ability of CPPB compared to

gemcitabine, one of the main chemotherapy drugs used to treat pancreatic cancer, and

tunicamycin using PD002 cells. Gemcitabine shows a wound-healing rate of 43% at 0.2 μM,

and tunicamycin shows 35% at 0.2 μM (SI). Thus, it was concluded that CPPB is more

effective in inhibiting cancer cell migration than gemcitabine and tunicamycin. These trends

were further confirmed by an endpoint migration assays via Boyden chambers for

PD002.

In these assays, the cell migrations of PD002 treated with CPPB (0.1 μM) were

inhibited on a higher level compared to those with tunicamycin and gemcitabine at the same

concentration (0.1 μM) (Figure 6).

Inhibition of a zinc-finger transcription factor, Snail1 (Snail) in the selected cancer cell

lines by CPPB.

Snail protein is one of the most important transcription factor that induces epithelial to

mesenchymal transition (EMT), which converts epithelial cells into migratory mesenchymal

cells that are more efficient at metastasizing.

EMT induced by overexpression of Snail

produces cancer stem-like properties in a number of solid organ cancers. Aberrant

expression of Snail leads to loss of expression of E-cadherin.

Thus, suppression of Snail

expression or inhibition of Snail functions represents a potent targeted therapeutic strategy

for many cancers.

Immunofluorescence assays using an anti-Snail antibody revealed that the fluorescence

intensity of Snail was strong in a series of pancreatic cancer cells (PANC-1, AsPC-1,

Capan-1, and PD002), and the expression of Snail was decreased by the treatment with

CPPB in a concentration dependent manner. Among pancreatic cancer cell lines, only the

data for PD002 and PANC-1 are shown in Figure 7A and 7B (see SI for AsPC-1 and

Capan-1). The Snail expression level of a non-metastatic pancreatic cancer, PANC-1, was

much lower than metastatic pancreatic cancers (

e.g

., PD002 and Capan-1). A few other types

of cancer cells such as a colorectal cancer (HCT-116) and a cervical cancer (SiHa) were

examined by similarly designed immunofluorescence assays or Western blot assays (Figures

7C and 7D). The Snail expression in SiHa was inhibited by treatment of CPPB in a

concentration dependent manner; at the IC

concentration (0.2 μM against DPAGT1),

CPPB effectively inhibited the Snail expression (Figure 7D). In contrast, the Snail

expression level in HCT-116 was not noticeably changed by the treatment of CPPB between

0.05 and 2.0 μM concentrations. Interestingly, cell migration of HCT-116 was not inhibited

by CPPB demonstrated in the wound healing (scratch) assays (Figure 5F). At the

concentrations tested in the scratch assays (0.05–2.0 μM), the E-Cadherin expression levels

Mitachi et al.

Page 7

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript

of PD002, PANC-1, and HCT-116 were not changed significantly (Figure 8). To support the

above discussion based on the immunefluorescent staining (Figure 7 and 8), the relative

expression levels of Snail and E-cadherin in PD002, PANC-1, and HCT-116 treated with

CPPB were measured by Western blot analyses (Figure 9). The relative expression levels

were obtained by using Image Studio™ Lite quantification software, and these quantified

data were summarized in SI (see Figure S5). PD002 and PANC-1 treated with CPPB (0 to

2.0 μM) lead to a dose dependent decrease in Snail. The E-Cadherin expression level of a

metastatic pancreatic cancer, PD002 was increased in a CPPB concentration depended

manner (0 to 2.0 μM), whereas, a non-metastatic pancreatic cancer, PANC-1 exhibited the

same expression level of E-Cadherin at 0–2.0 μM concentrations of CPPB (Figure 9A and

9B). A low DPAGT1 expressed-colon cancer, HCT-116 did not display noticeable difference

in the expression levels of Snail and E-Cadherin at 0–2.0 μM CPPB concentrations (Figure

9C).

Inhibition of DPAGT1 by CPPB.

CPPB decreased the DPAGT1 expression level in all pancreatic cancer cell lines examined

in Figure 5: the DPAGT1 expression was apparently inhibited by the IC

concentration of

CPPB (0.2 μM) (only the data for PD002 and PANC-1 shown Figure 10 and see Figure S4 in

SI) for AsPC-1 and Capan-1). An important observation is that the DPAGT1 expression of

the pancreatic cancer cell lines could not completely be inhibited at a high concentration of

CPPB (2–20 μM) (Figures 9 and 10). In MTT assays, we realized that all pancreatic cancers

tested remained viable at 20 μM concentration of CPPB (Table 2). We confirmed that the

DPAGT1 expression level in a colorectal adenocarcinoma, HCT-116 is significantly lower

than that in the pancreatic cancer cell lines (PD002 and PANC-1); DPAGT1 was not

detectable in Western blot assays for the lysate obtained by a standard protocol (50 μg/30 μL

of total protein sample). A 10 times concentrated HCT-116 cell lysate (prepared by

ultracentrifugation, 130,000xg for 1h) enabled us to detect DPAGT1 in Western blotting. By

treatment of HCT-116 with CPPB at 0.1–20 μM concentration, the DPAGT1 expression

levels of HCT-116 remained higher fluorescence intensity in immunefluorescent assay

(Figure 10C) and 20–90% in Western blot assays (Figure 9C). These data imply that the

inhibitory effect of CPPB on cell migration varies depending on degree of inhibition of the

DPAGT1 expression: immunefluorescent assays at 0.2–2.0 μM concentrations of CPPB, the

degree of DPAGT1 expression was decreased by the following order: PANC-1 > PD002 >>

HCT-116. Migration inhibition observed in the scratched assays (Figure 5) is well-correlated

with the degree of the DPAGT1 expression inhibition. Although CPPB decreased the protein

expression of DPAGT1 without significantly decreasing its gene (

DPAGT1

) expression,

tunicamycin decreased both the gene expression of

DPAGT1

and DPAGT1 protein

expression (Figure 11). These down-regulation in

DPAGT1

gene expression by tunicamycin

may be attributable to its high cytotoxicity against mammalian cells without selectivity.

Synergistic effect of CPPB with paclitaxel.

The FOLFIRINOX (a combination of folinic acid, 5-fluorouracil, irinotecan, and

oxaliplatin) and nab-paclitaxel (albumin-bound paclitaxel)-gemcitabine regimens have been

adopted into clinical practice for patients with metastatic pancreatic cancers.

Median

progression-free survival was reported in one study of patients with metastatic pancreatic

Mitachi et al.

Page 8

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript

cancer to be 6.4 months in the FOLFIRINOX group and 3.3 months in the gemcitabine

group.

Over the past years, the clinical data have not supported that FOLFIRINOX is

associated with any better (or worse) survival rates compared to the nab-paclitaxel-

gemcitabine regimen as there have been no head-to-head trials.

However, the inclusion of

paclitaxel and its derivatives in combination regimens remains an important therapeutic

strategy in pancreatic cancer chemotherapy since nab-paclitaxel-gemcitabine is associated

with less adverse effects (toxicity in patients) than FOLFIRINOX.

In this regard, we were

very interested in synergistic or additive effects of DPAGT1 inhibitor in combination with

paclitaxel. The synergistic or antagonistic activities of CPPB were assessed

in vitro

via

checkerboard technique.

In these experiments, CPPB displayed strong synergistic

effects with paclitaxel in a wide range of concentrations against PD002. Table 3 summarizes

the results of FIC index analyses for selected combinations of CPPB (IC

35.0 μM) plus

paclitaxel (IC

1.25 μM) that showed synergistic combination (

FIC<0.5). The FIC index

below 0.50 was observed for 20 combinations of two molecules out of 96 different

concentrations (see Figure S6 in SI). The IC

value of paclitaxel against PD002 was

lowered (0.024–0.61 μM) in combination with CPPB (0.1–2.0 μM).

Interaction of CPPB with DPAGT1 (Modeling Studies).

DPAGT1 encompasses ten transmembrane segments, three loops on the ER (endoplasmic

reticulum) side, and five loops on the cytoplasmic side. Four loops (CL-1, CL-5, CL-7, and

CL-9) on the cytoplasmic side form the UDP-GlcNAc-binding domain.

The

hydrophobic tunnel created by the transmembrane segments (TM-4, TM-5 and CL-9) within

the lipid bilayer is predicted to interact with dolychol-phosphate (Dol-P). The weak

DPAGT1 inhibitors,

-methyl capuramycin (OM-CAP) and

iso

-capuramycin (I-CAP)

(Table 1), yielded low docking scores (Glide Scores) using Schrödinger’s Glide program: the

score for OM-CAP was -12.4 and for I-CAP was -10.9. These scores predicted that OM-

CAP and I-CAP possess significantly lower affinity for DPAGT1 than CPPB, which showed

a docking score of -16.6 (see Table S1 in SI).

–

The docked CPPB-DPAGT1 structure

illustrated in Figure 12 shows several key interactions. The C2’-OH acts as a donor in a

hydrogen bond to the Glu56 carboxylate. This interaction is likely be lost when the C2’-OH

is methylated. Pi-stacking interactions are observed between 1) Phe249 and the uracil ring,

and 2) Trp122 and the trifluoromethoxybenzene in the hydrophobic moiety. Asn185 and

Ash252 (protonated Asp) form hydrogen-bond(s) to the primary amide and the

dihydropyran-hydroxy (C3”-OH) group, respectively. Additional hydrogen-bonds between

the uridine ureido group and the backbone amides (Gln44 and Leu46) strengthen the ligand

interaction. In this program, capuramycin (CAP), a strong MraY inhibitor with no DPAGT1

inhibitory activity, was successfully engineered to be a relatively strong DPAGT1 inhibitor

by introducing a hydrophobic functional group, (((trifluoromethoxy)phenoxy) piperidin-1-

yl)phenyl)methylamine at the C6“-position. This hydrophobic group allows the preferable

conformation of the uridine-enopyranosiduronic moiety in the DPAGT1 biding domain,

resulting in increased interactions with Asn185, Ash252, Glu56, and Phe249.

A semi-synthesis of CPPB from A500359F.

Pharmacological studies of CPPB (

) and its related analogues using appropriate animal

models will be a focus of our future research efforts. Previously, a natural product

Mitachi et al.

Page 9

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript

A-5003659E was used to develop novel capuramycin analogues with strong MraY inhibitory

activity (Hotoda et. al. 2003) (Figure 3).

Its free-carboxylic acid analogue, A-500359F was

also isolated from the capuramycin-producing strain,

Streptomyces griseus

Sank 60196.

Saponification of A-5003659E to A-500359F was established by the Sankyo group.

Although, the currently available synthetic schemes for capuramycin analogues (

e.g

Scheme 3) include a relatively short number of chemical steps,

a semi-synthetic

approach is more feasible to deliver a large quantities of CPPB for pharmacological studies.

To establish semi-synthesis of CPPB, A-500359F was first synthesized from the CAP-

synthetic intermediate

(Scheme 3A) in a single step. Amide-forming reaction of synthetic

A-500359F with (((trifluoromethoxy)phenoxy) piperidin-1-yl)phenyl)methylamine (

) was

performed under an optimized condition using EDCI, glyceroacetonide-Oxyma, and NMM

in DMF.

All coupling reagents could be removed by partitions between CHCl

and water

and evaporation. The crude product was passed through DOWEX 50Wx4 column (MeOH :

OH = 4 : 1) to provide CPPB with >95% purity, which was further purified by C

reverse HPLC (MeOH : H

O = 65 : 35) to yield pure CPPB (Scheme 4).

CONCLUSION

This paper describes the identification of a new DPAGT1 inhibitor of capuramycin analogue,

capuramycin phenoxypiperidinylbenzylamide (CPPB,

) and its isostere I-CPPB (

Previously, tunicamycin is the only DPAGT1 inhibitor that has been widely applied to the

studies associated with protein misfolding

in vitro

Tunicamycin displays cytotoxicity

against cancer and healthy cells with low selectivity ratio.

One of cytotoxicity mechanisms

of tunicamycin is believed to be its ability to inhibit DPAGT1 enzyme functions. The

DPAGT1 expression levels vary depending on the cell types; renal cancers and lymphomas

express low-levels of DPAGT1, whereas, a majority of solid cancers express high-levels.

Thus, the observed antiproliferative activity of tunicamycin against all types of cancer cells

are difficult to understand solely by its DPAGT1 inhibitory activity. In our studies, CPPB

showed ~7.5 times stronger DPAGT1 inhibitory activity than tunicamycin. However, unlike

tunicamycin, CPPB did not inhibit growth of cancer cell lines at the IC

values observed for

tunicamycin (0.45 –7.5 μM). CPPB is a cell-permeable molecule which was demonstrated

by IncuCyte® live cell analyses and immunofluorescence assays. In this article, we have

studied effectiveness of CPPB on the expression levels of Snail, E-cadherin, and DPAGT1

primary in pancreatic cancers (Figures 7–10). CPPB decreased the Snail expression in

commercially available pancreatic cancer cell lines (PANC-1, AsPC-1, and Capan-1) and a

patient-derived pancreatic ductal adenocarcinoma cell line (PD002) in a dose dependent

manner. On the other hand, the E-cadherin expression level was increased in PD002 (a

metastatic pancreatic cancer cell) or was not noticeably changed in PANC-1 at between

0.05–0.2 μM of CPPB. These biochemical data may support that a selective DPAGT1

inhibitor, CPPB is effective in inhibiting metastasis spread of the pancreatic cancer cells

observed in scratch and transwell assays (Figures 5 and 6). Other than pancreatic cancers, a

lower DPAGT1 expression cell, a colorectal adenocarcinoma (HCT-116) and a higher

DPAGT1 expression cell, a cervical carcinoma (SiHa) were examined. CPPB did not inhibit

migration of HCT-116, but strongly inhibited migration of SiHa in scratch assays at 0.2 μM

(IC

concentration against DPAGT1). These observations were supported by the

Mitachi et al.

Page 10

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript

biochemical analyses of the Snail and E-cadherin expression levels. Snail plays an important

role in cancer progression. The accumulated evidences on Snail indicate that over-expression

of Snail promotes drug resistance, tumor recurrence and metastasis.

Although, only

limited data have been generated in this article, CPPB’s Snail inhibitory activity observed in

pancreatic and a several high DPAGT1-expressed cell lines suggests that selective DPAGT1

inhibitors have the potential to develop into less toxic anticancer therapeutics than anticancer

drugs that are cytotoxic to all dividing cells in the body. CPPB is not cytotoxic against a

series of cancer and healthy cell lines at 10 μM or higher concentrations. However, it showed

a cytostatic activity against pancreatic cancers and a strong synergistic effect with paclitaxel;

cytotoxic activity of paclitaxel was improved over 250-times against PD002 in combination

with CPPB (0.2–2.0 μM). Docking studies of CPPB with the available DPAGT1 crystal

structures provided insight into unique interactions, and thus, structure-based molecule

design may be a fruitful approach to improve CPPB’s DPAGT1 affinity. A collaboration

with Daiichi-Sankyo is essential to efficiently produce CPPB for

in vivo

studies using large-

animal models such as dogs, pigs, and monkeys. We have demonstrated a semi-synthesis of

CPPB from a capuramycin biosynthetic intermediate, A-500359F, that will secure a

production of large amount of CPPB (Scheme 4). Our total synthetic scheme is amenable to

produce gram-quantity of CPPB (Scheme 3). Extensive synergistic, toxicity, and

pharmacokinetic studies of CPPB will be performed using preclinical animal models, and

these data including detailed evaluation on

in vivo

efficacy against pancreatic and cervical

cancers will be reported elsewhere.

EXPERIMENTAL SECTION

Chemistry. General Information.

All chemicals were purchased from commercial sources and used without further

purification unless otherwise noted. THF, CH

, and DMF were purified via Innovative

Technology’s Pure-Solve System. All reactions were performed under an Argon atmosphere.

All stirring was performed with an internal magnetic stirrer. Reactions were monitored by

TLC using 0.25 mm coated commercial silica gel plates (EMD, Silica Gel 60F

254

). TLC

spots were visualized by UV light at 254 nm, or developed with ceric ammonium molybdate

or anisaldehyde or copper sulfate or ninhydrin solutions by heating on a hot plate. Reactions

were also monitored by using SHIMADZU LCMS-2020 with solvents: A: 0.1% formic acid

in water, B: acetonitrile. Flash chromatography was performed with SiliCycle silica gel

(Purasil 60 Å, 230–400 Mesh). Proton magnetic resonance (

H-NMR) spectral data were

recorded on 400, and 500 MHz instruments. Carbon magnetic resonance (

C-NMR)

spectral data were recorded on 100 and 125 MHz instruments. For all NMR spectra,

chemical shifts (

C) were quoted in parts per million (ppm), and

values were quoted in

Hz.

H and

C NMR spectra were calibrated with residual undeuterated solvent (CDCl

H = 7.26 ppm,

C = 77.16 ppm; CD

CN:

H = 1.94 ppm,

C = 1.32ppm; CD

OD:

=3.31 ppm,

C =49.00 ppm; DMSO-d

H = 2.50 ppm,

C = 39.52 ppm; D

H = 4.79

ppm) as an internal reference. The following abbreviations were used to designate the

multiplicities: s = singlet, d = doublet, dd = double doublets, t = triplet, q = quartet, quin =

quintet, hept = heptet, m = multiplet, br = broad. Infrared (IR) spectra were recorded on a

Perkin-Elmer FT1600 spectrometer. HPLC analyses were performed with a Shimadzu

Mitachi et al.

Page 11

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript

LC-20AD HPLC system. HR-MS data were obtained from a Waters Synapt G2-Si (ion

mobility mass spectrometer with nanoelectrospray ionization). All assayed compounds were

purified by reverse HPLC to be ≥95% purity.

)-2-((1

)-((2

)-3-Acetoxy-5-(3-((benzyloxy)methyl)-2,4-dioxo-3,4-

dihydropyrimidin-1(2

)-yl)-4-methoxytetrahydrofuran-2-yl)(cyano)methoxy)-6-((2-

chloroacetoxy)methyl)tetrahydro-2

-pyran-3,4,5-triyl triacetate (15).

To a stirred suspension of

(0.17 g, 0.38 mmol),

(0.37 g, 0.75 mmol), MS3Å (0.50 g)

and SrCO

(0.28 g, 1.88 mmol) in CH

(9.4 mL) were added AgBF

(0.037 g, 0.19

mmol) and NIS (0.25 g, 1.13 mmol) at 0 °C. After being stirred for 19h, the reaction mixture

was added Et

N (1.0 mL) and passed through a silica gel pad (hexanes/EtOAc = 1/4). The

filtrate was concentrated

in vacuo

. The crude mixture was purified by silica gel column

chromatography (hexanes/EtOAc = 2/1 to 1/2) to afford

(0.24 g, 78%):

H NMR (400

MHz, Chloroform-

)

7.40 – 7.27 (m, 6H), 6.07 (d,

= 4.7 Hz, 1H), 6.00 (d,

= 8.2 Hz,

1H), 5.51 (d,

= 9.8 Hz, 1H), 5.47 (d,

= 9.6 Hz, 1H), 5.37 (dd,

= 3.4, 2.0 Hz, 1H), 5.28

(d,

= 10.2 Hz, 1H), 5.22 – 5.15 (m, 2H), 4.84 (d,

= 3.2 Hz, 1H), 4.71 (s, 2H), 4.41 (dd,

= 5.5, 3.2 Hz, 1H), 4.39 – 4.35 (m, 1H), 4.33 (d,

= 6.0 Hz, 1H), 4.27 (d,

= 2.3 Hz, 1H),

4.14 (d,

= 1.7 Hz, 2H), 4.03 – 4.01 (m, 1H), 3.93 (ddd,

= 8.9, 5.9, 2.4 Hz, 1H), 3.47 (s,

3H), 2.20 (s, 3H), 2.19 (s, 3H), 2.07 (s, 3H), 2.02 (s, 3H);

C NMR (101 MHz, CDCl

)

170.20, 169.84, 169.50, 166.92, 162.19, 150.97, 137.75, 137.46, 128.32 (2C), 127.72,

127.69 (2C), 113.96, 103.58, 96.29, 88.68, 80.93, 80.09, 72.30, 70.42, 69.60, 68.46, 67.94,

65.16, 64.35, 63.58, 59.26, 40.61, 31.58, 20.70, 20.66, 20.62, 20.57; HRMS (ESI+)

calcd for C

ClN

[M + H] 810.2125, found: 810.2151.

)-2-((1

)-1-((2

)-3-Acetoxy-5-(2,4-dioxo-3,4-dihydropyrimidin-1(2

yl)-4-methoxytetrahydrofuran-2-yl)-2-amino-2-oxoethoxy)-6-(hydroxymethyl)tetrahydro-2

pyran-3,4,5-triyl triacetate (16).

To a stirred solution of

(0.24 g, 0.29 mmol) in a 9:1 mixture of EtOH and H

O (2.9 mL)

were added HgCl

(0.16 g, 0.59 mmol) and acetaldoxime (0.18 mL, 2.9 mmol). After being

stirred for 13h at r.t., the reaction mixture was concentrated

in vacuo

. The residue was

quenched with aq. NaHCO

, and extracted with CHCl

. The combined organic extracts were

dried over Na

and concentrated

in vacuo

. The crude product was purified by silica gel

column chromatography (hexanes/EtOAc = 1/2 to CHCl

/MeOH = 96/4) to afford the amide

(0.21 g, 87%). To a solution of the amide (0.21 g, 0.26 mmol) in a 1:1 mixture of THF and

MeOH (2.6 mL) was added thiourea (0.059 g, 0.77 mmol). After being stirred for 11h at 50

°C, the reaction mixture was concentrated

in vacuo

. The residue was diluted with H

O and

extracted with CHCl

. The combined organic extracts were dried over Na

and

concentrated

in vacuo

. The crude product was purified by silica gel column chromatography

(CHCl

/MeOH = 98/2 to 97/3 to 96/4) to afford the primary alcohol (0.15 g, 75%). To a

stirred solution of the primary alcohol (0.15 g, 0.19 mmol) and AcOH (0.040 mL) in a 1:1

mixture of THF and

PrOH (2.0 mL) was added 10% Pd/C (0.12 g) under N

. H

gas was

introduced and the reaction mixture was stirred under H

atmosphere for 3h, the solution

was filtered through Celite and concentrated

in vacuo

. The crude product was purified by

silica gel column chromatography (CHCl

/MeOH = 97/3 to 92/8) to afford

(0.10 g,

81%):

H NMR (400 MHz, Chloroform-

)

9.08 (brs, 1H), 7.60 (d,

= 8.2 Hz, 1H), 6.87

Mitachi et al.

Page 12

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript

(brs, 1H), 6.12 (brs, 1H), 6.00 (d,

= 8.1 Hz, 1H), 5.87 (d,

= 3.6 Hz, 1H), 5.55 – 5.52 (m,

1H), 5.26 – 5.22 (m, 2H), 5.20 (t,

= 6.0 Hz, 1H), 5.00 (d,

= 1.7 Hz, 1H), 4.53 (dd,

= 6.5,

2.3 Hz, 1H), 4.46 (d,

= 2.3 Hz, 1H), 4.06 (dd,

= 5.5, 3.7 Hz, 1H), 3.76 – 3.71 (m, 1H),

3.69 – 3.57 (m, 2H), 3.44 (s, 3H), 2.17 (s, 3H), 2.15 (s, 3H), 2.09 (s, 3H), 2.04 (s, 3H);

NMR (101 MHz, CDCl

)

171.42, 170.46, 170.46, 170.40, 163.10, 150.03, 139.50, 103.61,

97.15, 96.95, 88.50, 81.26, 80.96, 75.33, 72.91, 69.67, 68.89, 65.54, 61.27, 59.05, 50.86,

20.82, 20.75, 20.67, 20.59; HRMS (ESI+)

calcd for C

[M + H] 632.1939,

found: 632.1963.

)-3,4-Diacetoxy-2-((1

)-1-((2

)-3-acetoxy-5-(2,4-dioxo-3,4-

dihydropyrimidin-1(2

)-yl)-4-methoxytetrahydrofuran-2-yl)-2-amino-2-oxoethoxy)-3,4-

dihydro-2

-pyran-6-carboxylic acid (18).

To a stirred solution of

(0.10 g, 0.16 mmol) and DMSO (0.11 mL, 1.57 mmol) in a 5:1

mixture of CH

and Et

N (0.80 mL) was added SO

•pyridine (0.25 g, 1.57 mmol). After

being stirred for 3h at r.t., the reaction mixture was added H

O (0.16 mL) and passed

through a silica gel pad (CHCl

/MeOH = 92/8) to provide the crude

. To a stirred solution

of the crude mixture in

BuOH (1.0 mL) and 2-methyl-2-butene (0.5 mL) was added a

solution of NaClO

(0.071 g, 0.78 mmol) and NaH

•2H

O (0.12 g, 0.78 mmol) in H

(1.0 mL). After being stirred for 4h at r.t., the reaction was quenched with H

O and extracted

with CHCl

/MeOH (9/1). The combined organic extracts were dried over Na

and

concentrated

in vacuo

. The crude product was purified by silica gel column chromatography

(CHCl

/MeOH = 9/1) to afford

(0.078 g, 85%):

H NMR (400 MHz, Methanol-

)

7.77 (d,

= 8.1 Hz, 1H), 5.94 (d,

= 11.7 Hz, 1H), 5.94 (s, 1H), 5.78 (t,

= 2.1 Hz, 1H),

5.66 (dd,

= 4.5, 2.5 Hz, 1H), 5.56 (ddd,

= 4.7, 3.2, 1.6 Hz, 1H), 5.32 (d,

= 3.3 Hz, 1H),

4.99 (t,

= 5.2 Hz, 1H), 4.85 (d,

= 2.1 Hz, 1H), 4.60 (dd,

= 5.2, 2.0 Hz, 1H), 3.95 (t,

5.0 Hz, 1H), 3.39 (s, 3H), 2.13 (s, 3H), 2.08 (s, 3H), 2.05 (s, 3H);

C NMR (101 MHz,

MeOD)

172.79, 172.06, 171.60, 167.79, 166.01, 152.05, 148.16, 140.97, 104.23, 103.98,

97.91, 88.13, 83.38, 82.91, 76.56, 71.95, 65.17, 65.06, 59.41, 20.75, 20.63, 20.57; HRMS

(ESI+)

calcd for C

[M + H] 586.1520, found: 586.1549.

)-2-((1

)-1-((2

)-3-Acetoxy-5-(2,4-dioxo-3,4-dihydropyrimidin-1(2

)-yl)-4-

methoxytetrahydrofuran-2-yl)-2-amino-2-oxoethoxy)-6-(((S)-2-oxoazepan-3-

yl)carbamoyl)-3,4-dihydro-2

-pyran-3,4-diyl diacetate (19).

To a stirred solution of

(31 mg, 0.053 mmol), 2-(

)-aminocaprolactam (26 mg, 0.16

mmol), HOBt (21 mg, 0.16 mmol) and NMM (58 μL, 0.53 mmol) in DMF (0.26 mL) was

added EDCI (51 mg, 0.26 mmol). After being stirred for 6h at r.t., the reaction was quenched

with aq. NaHCO

, and extracted with CHCl

/MeOH (9/1). The combined organic extracts

were dried over Na

and concentrated

in vacuo

. The crude product was purified by silica

gel column chromatography (CHCl

/MeOH = 95/5) to afford

(30 mg, 82%):

H NMR

(400 MHz, Methanol-

)

7.68 (d,

= 8.2 Hz, 1H), 6.02 (t,

= 2.1 Hz, 1H), 5.91 (d,

= 4.2

Hz, 1H), 5.89 (d,

= 6.3 Hz, 1H), 5.73 (dd,

= 4.4, 2.5 Hz, 1H), 5.68 – 5.65 (m, 1H), 5.51

(d,

= 3.0 Hz, 1H), 5.01 (d,

= 4.6 Hz, 1H), 4.76 (d,

= 2.0 Hz, 1H), 4.62 (d,

= 12.5 Hz,

1H), 4.60 – 4.57 (m, 1H), 3.98 (d,

= 6.1 Hz, 1H), 3.35 (s, 3H), 2.11 (s, 3H), 2.08 (s, 3H),

2.05 (s, 3H), 2.03 – 1.98 (m, 2H), 1.91 – 1.83 (m, 2H), 1.65 – 1.56 (m, 2H), 1.44 – 1.37 (m,

2H);

C NMR (101 MHz, MeOD)

176.12, 172.32, 171.66, 171.29, 171.21, 165.96,

Mitachi et al.

Page 13

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript

160.96, 151.90, 144.61, 141.42, 106.25, 103.54, 98.81, 88.77, 83.41, 79.28, 77.39, 74.79,

64.91, 64.49, 59.26, 57.36, 53.25, 42.36, 32.19, 29.75, 28.94, 20.54, 20.41, 20.32; HRMS

(ESI+)

calcd for C

[M + H] 696.2364, found: 696.2391.

)-2-((1

)-2-Amino-1-((2

)-5-(2,4-dioxo-3,4-dihydropyrimidin-1(2

)-yl)-3-

hydroxy-4-methoxytetrahydrofuran-2-yl)-2-oxoethoxy)-3,4-dihydroxy-

-((

)-2-oxoazepan-3-

yl)-3,4-dihydro-2

-pyran-6-carboxamide (3).

A solution of

(30 mg, 0.043 mmol) in a 5:1 mixture of MeOH and Et

N (0.60 mL) was

stirred for 5h at r.t., and concentrated

in vacuo

. The crude mixture was purified by reverse-

phase HPLC [column: Luna® (C18, 10 μm, 100 Å, 250 × 10 mm), solvents: 15:85

MeOH:H

O, flow rate: 3.0 mL/min, UV: 254 nm, retention time: 20 min] to afford I-CAP

(

, 24 mg, 98%):

H NMR (400 MHz, Methanol-

)

7.92 (d,

= 8.1 Hz, 1H), 6.02 (d,

3.8 Hz, 1H), 5.88 (d,

= 5.1 Hz, 1H), 5.74 (d,

= 8.1 Hz, 1H), 5.23 (d,

= 5.6 Hz, 1H), 4.67

(d,

= 2.0 Hz, 1H), 4.59 – 4.54 (m, 2H), 4.38 (t,

= 4.2 Hz, 1H), 4.29 (t,

= 5.1 Hz, 1H),

3.98 (t,

= 5.0 Hz, 1H), 3.84 (t,

= 4.6 Hz, 1H), 3.43 (s, 3H), 2.06 – 1.99 (m, 2H), 1.89 –

1.81 (m, 2H), 1.62 – 1.45 (m, 2H), 1.44 – 1.33 (m, 2H);

C NMR (101 MHz, MeOD)

176.27, 173.46, 166.15, 161.85, 152.32, 144.23, 141.91, 109.37, 102.82, 101.22, 90.27,

83.49, 81.02, 78.93, 74.54, 68.51, 63.53, 58.67, 53.35, 42.48, 32.36, 29.91, 29.06; HRMS

(ESI+)

calcd for C

[M + H] 570.2048, found: 570.2071.

)-2-((

)-((3a

,6R,6a

)-6-(3-((Benzyloxy)methyl)-2,4-dioxo-3,4-

dihydropyrimidin-1(2

)-yl)-2,2-dimethyltetrahydrofuro[3,4-

][1,3]dioxol-4-yl)

(cyano)methoxy)-6-((2-chloroacetoxy)methyl)tetrahydro-2

-pyran-3,4,5-triyl triacetate (21).

To a stirred suspension of

(0.24 g, 0.56 mmol),

(0.55 g, 1.12 mmol), MS3Å (0.72 g),

and SrCO

(0.41 g, 2.80 mmol) in CH

(14.0 mL) were added AgBF

(0.055 g, 0.28

mmol) and NIS (0.25 g, 1.12 mmol) at 0 °C. After being stirred for 12h, the reaction mixture

was added Et

N (1.0 mL), and passed through a silica gel pad (hexanes/EtOAc = 1/4). The

filtrate was concentrated

in vacuo

. The crude mixture was purified by silica gel column

chromatography (hexanes/EtOAc = 6/4 – 4/6) to afford

(0.39 g, 88%):

H NMR (400

MHz, Chloroform-

)

7.37 – 7.28 (m, 5H), 7.20 (d,

= 8.1 Hz, 1H), 5.82 (d,

= 8.1 Hz,

1H), 5.61 (d,

= 1.6 Hz, 1H), 5.50 (d,

= 9.8 Hz, 1H), 5.42 (d,

= 9.8 Hz, 1H), 5.35 – 5.27

(m, 1H), 5.23 (dd,

= 10.0, 3.3 Hz, 1H), 5.07 (d,

= 1.9 Hz, 1H), 5.00 (d,

= 3.5 Hz, 1H),

4.98 (dd,

= 6.4, 1.6 Hz, 1H), 4.84 (d,

= 7.1 Hz, 1H), 4.69 (s, 2H), 4.43 (dd,

= 7.1, 3.6

Hz, 1H), 4.20 (dd,

= 12.2, 4.3 Hz, 1H), 4.17 – 4.07 (m, 2H), 4.05 (d,

= 1.6 Hz, 2H), 3.97

– 3.92 (m, 1H), 2.16 (s, 3H), 2.02 (s, 3H), 1.98 (s, 3H), 1.59 (s, 3H), 1.39 (s, 3H);

C NMR

(101 MHz, CDCl

)

169.80, 169.57, 169.47, 166.80, 162.08, 150.84, 140.89, 137.67,

128.33 (2C), 127.91, 127.77, 127.63 (2C), 115.23, 114.72, 102.84, 96.56, 96.30, 86.55,

84.15, 80.95, 72.42, 70.41, 69.41, 68.39 (2C), 66.04, 65.35, 62.90, 40.52, 27.03, 25.20,

20.73, 20.64, 20.59; HRMS (ESI+)

calcd for C

ClN

[M + H] 794.2175, found:

794.2198.

Mitachi et al.

Page 14

J Med Chem

. Author manuscript; available in PMC 2020 October 14.

Author Manuscript