Design of A Metabolically Stable Tritium-Tracer of the PI3Kδ-Inhibitor CDZ173 (Leniolisib) as a Tool to Study Liver Metabolites

Article abstract of DOI:10.1002/hlca.201800044

In this disclosure, we summarize the preliminary metabolic profiling of the PI3Kδ inhibitor CDZ173 (leniolisib, 1a) obtained from incubations of the unlabeled compound and the synthesis of its metabolically stable tritium isotopologue 1b used for metaboli

Full text of DOI:10.1002/hlca.201800044

(
1 of 33) e1800044
Design of A Metabolically Stable Tritium-Tracer of the PI3Kd-
Inhibitor CDZ173 (Leniolisib) as a Tool to Study Liver Metabolites
a,1
a
b
a
c
b
Carsten Bauer,* Tong Luu, Fabian Eggimann, Patrick Bross, Werner Gertsch, Cheng Hu,
c
c
a
b
b
Philippe Ramstein, Julien Bourgailh, Albrecht Gl €a nzel, Ina Dix, Christian Guenat, Nicolas
Soldermann, Karine Litherland, Sandrine Desrayaud, Jean-Claude Hengy, David Pearson,*
b
c
c
c
c
c
d
Joachim Blanz, and Christoph Burkhart
a
Isotope Laboratory, PK Sciences, Novartis Institute for Biomedical Research (NIBR), Basel,
e-mail: carsten.bauer@unibas.ch
b
Global Discovery Chemistry, NIBR, Basel
ADME Biotransformation, PK Sciences, NIBR, Basel, e-mail: david.pearson@novartis.com
c
d
Autoimmunity, Transplantation and Inflammation, NIBR, Basel
In this disclosure, we summarize the preliminary metabolic profiling of the PI3Kd inhibitor CDZ173 (leniolisib, 1a)
obtained from incubations of the unlabeled compound and the synthesis of its metabolically stable tritium
isotopologue 1b used for metabolite structure confirmation. Access to 1b was achieved when a halogenated
3
3
precursor was subject to Hal/ H-exchange. Hence, [ H]CDZ173 with specific activity 630 GBq/mmol, HPLC-RA 97%
and ee = 99.2% was obtained. Synthetic key to the precursor was using a bis-halo-pyridine in a Pd-catalyzed mono-
amination of the tetrahydropyrido-pyrimidine core. Stereochemistry of the synthetic precursors were confirmed by X-
ray analysis of the unlabeled bis-halo-pyridines and chiral HPLC of the tritiated material. The correct position of
3
tritium label in the target, was confirmed by H-NMR difference spectroscopy. Besides, we report on the validation of
the radiotracer as a tool for pre-clinical ADME in incubations with hepatocytes. Based on this data, we present a
quantitative metabolite profile of leniolisib which was confirmed by independently synthesized metabolite references.
The conformation of CDZ173 was investigated by NMR suggesting two different amide backbones each with specific
pyrrolidine puckerings.
Keywords: PI3Kd inhibitor, leniolisib, CDZ173, metabolism of pyrrolidines, conformational analysis of pyrrolidines.
dependent reduction in PI3K/Akt pathway activity
assessed ex vivo and improved immune dysregulation
Since its discovery in 1985, phosphatidylinositide 3- and it may be a new treatment option for patients with
kinases (PI3K) have become important drug targets in pathological activation of the PI3Kd pathway, as in B
medicinal chemistry. Recently we have discovered the cell malignancies or autoimmune disorders including
Introduction
[
1]
7
-phenyl-2-pyyrolidinyloxy-quinoxaline CDZ173 (le- Sj o€ gren’s syndrome, systemic lupus erythematosus or
niolisib; 1a), which is a promising PI3Kd inhibitor and rheumatoid arthritis.
[
4 – 6]
conformationally flexible side chain with (S)-
To summarize,
starting from the lead com-
[
2]
stereochemistry. Most recently, leniolisib was investi- pound BEZ522 sophisticated SAR studies yielded the
rd
gated in a 12-week, open-label and within-subject
3
generation PI3Kd inhibitor CDZ173 with improved
dose-escalation study in patients with Activated PI3K biophysical properties, promising PK/PD and encour-
[
3]
Delta Syndrome (APDS). Leniolisib led to a dose- aging clinical results. As part of the preclinical pro-
gram, the ADME properties of this compound were
investigated in animal models. For these studies it
was aimed to use a more economical and metaboli-
[
7]
1
Current address: Department of Chemistry, University
of Basel, Biopark BPR 1096, Mattenstrasse 24a, CH-4058 cally stable tritium analogue of CDZ173 as a very
ꢀ
Basel, e-mail: carsten.bauer@unibas.ch
sensitive, weak b
emitting radiotracer for the
DOI: 10.1002/hlca.201800044 Helv. Chim. Acta 2018, 101, e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
+
quantification of metabolites and elimination routes lists the observed [M + H] . Columns 7 – 11 list m/z in
2
3
with a low cost radiolabel. The development of this MS and MS together with the assigned cations.
radiotracer for preclinical studies is described in the
following sections.
O-Demethylation, amide bond hydrolysis (group A),
C-oxidation at the tetrahydropyrido moiety, at the
pyrrolidine moiety, as well as mixed oxidation at both
moieties were identified as important biotransforma-
tion steps (groups B, C, E – H). Oxidation at the pyrro-
lidine moiety also led to ring opening and subsequent
Results and Discussion
3
Defining the Ideal Position for H-Labelling of CDZ173
Based on Metabolites Found After Microsomal Incubation
3
Prior starting the synthesis work on [ H]CDZ173 (1b), oxidation of the aliphatic side chain to the corre-
metabolic profiling data of unlabeled CDZ173 was sponding acid (group D). N-Oxidation was not
acquired in order to define the most likely metaboli- detected among the metabolites that were subject of
cally stable proton/tritium position in CDZ173 (see structure elucidation.
Supplementary data for Materials And Methods). Since
members of the CYP-enzyme family account for 75% grouped in seven structural categories A – H. The
To summarize, the 26 detected metabolites can be
[
8]
of drug metabolism, liver microsomes are a practical molecular mass of several metabolites indicated a loss
system to determine metabolic stability of most com- of two, in one case four hydrogens that may be attrib-
pounds. Information on the elimination pathways of a uted to an initial oxidation and a subsequent loss of a
drug can be obtained by analyzing bile and urine in water molecule during the ionization process. For this
animal models. Hence, for metabolic profiling, unla- reason, such metabolites (e.g. M 8 or M 12) are
M M
beled CDZ173 was incubated with liver microsomes of grouped together with hydroxylation metabolites.
rat, mouse, dog, cynomolgus monkey and human and Based on its fragmentation pattern and the character-
[
9]
administered to bile-duct cannulated rat with collec- istic loss of m/z 176,
M 14 was assigned to be a
M
tion of blood, bile, urine as well as tissues at necropsy. glucuronide. Reactive intermediate trapping with glu-
Subsequently, the samples of the in vitro incubations tathione was performed but glutathione adducts were
as well as of the animal model were analyzed by LC/ not detected in any of the microsomal incubations.
[
9]
MS/MS.
However, a cysteine conjugate (M 17) was detected
M
in all in vivo samples. Hence in Table 1 these two con-
jugates are treated as a follow up products of a
hydroxylated species; for more explicit structures, see
the structure of the corresponding metabolite in
Table 12.1 in the Supplementary Material.
Most importantly, no metabolites with hydroxyla-
tion at the pyridine ring-carbon were observed. It was
shown that CDZ173 is extensively metabolized at the
tetrahydropyridopyrimidine (THP), the pyrrolidine and
O-methyl group, but not at the CF -containing pyridine
3
A total of about 50 microsomal metabolites were ring. A very similar metabolic inertness was found for
detected, but metabolic profiling and preliminary struc- the m-CF -pyridine moiety in the in vitro and in vivo
ture investigations were limited to 26 microsomal metabolism of the cannabinoid-1 receptor inverse
3
[
10]
metabolites M 1 – M 26 with larger LC/MS peak areas antagonist taranabant.
Thus, the most important
M
M
(
Table 1). LC/MS-Peak areas of the microsome incuba- scope of this analysis was achieved. This was to define
tions and body/tissue-samples of the metabolites were the synthetic target for later tritium positioning. Based
determined but will not be discussed here, because no on this reasoning and the first metabolic profiling data
reference standards were used in this preliminary analy- summarized here, the synthetic target structure for radi-
sis. Isomeric xenobiotics were distinguished by the olabeling was defined to be 1b (Figure 1).
characteristic LC-retention time (e.g. M 3 and M 15).
M
M
3
The numbering of the metabolites is given by their Synthesis of the [ H]Tracer
sequence of retention times (LC t ). Table S1 (for space
R
reasons placed in the Supplementary Material) summa- Development of the Precursor Synthesis Suitable for End-
n
rizes the data gathered from MS/MS spectrometry of Tritiation. The medicinal chemistry of CDZ173 was
[
2]
the metabolites. In this table, column 4 lists the pro- published in 2017. There, the synthesic route for
posed structure and molecular formula, and column 5 CDZ173 consists out of three principal steps, which are
www.helv.wiley.com
(2 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
Table 1. Structure categories of microsomal metabolites of CDZ173
Category Biotransformation Narrowed structures
Metabolites
M 1, M 2
of M
step
A
Demethylation and
hydrolysis
M
M
B
Pyrrolidine
mono-
M 11, M 17,
M M
M 25
M
hydroxylation
C
N-propionly-
M 6, M 7,
M M
pyrrolidine mono-
hydroxylation
M
M
9
D
Formal hydration
and oxidation
M
M
M
3, M 4,
M M
M 15, M 21
www.helv.wiley.com
(3 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
Table 1. (cont.)
Category Biotransformation Narrowed structures
Metabolites
M 5, M 19
of M
step
E
Pyrrolidine
M
M
bis-hydroxylation
F
THP mono-
M M
M 10, M 22
hydroxylation
G
THP bis-
M 8, M 12,
M
M
hydroxylation
M
M
13, M
23, M
M
14,
26
M
M
M
H
Pyrrolidine and
THP bis-
hydroxylation
M 16, M 18,
M M
M
M
19, M
24
M
20,
M
M
chemistry synthesis of CDZ173, see the Supplementary
part of reference [2].
For radiosynthesis, i.e., introduction of the tritium
in the very last step of the synthesis, this general
approach needed to be changed as outlined in
Scheme 1. The radiosynthesis started from the N-pro-
tected
5,6,7,8-tetrahydropyrido[4,3-d]pyrimidin-4-ol
which was transformed into the corresponding
chloride 3. Subsequent nucleophilic substitution with
3
Figure 1. Structure of the radiosynthesis target [ H]CDZ173.
[11]
2
,
a) the amination of an extended 4-chloropyrimidine the amine 6 in step ii gave 7. Debenzylation in step iii
with a by-functional aliphatic amine, b) the extension gave 8. In order to confirm the chiral purity of 8, an
of the pyrimidine system by Pd-catalyzed amination analytical sample was BOC-protected and analyzed by
using a X-Phos ligand and a poly-substituted pyridine, chiral HPLC. It became clear that slight racemization
and c) completion of the targets by further modification had occurred in step iii. For this reason, racemized 8
on the by-functional amine mentioned under a). For a was BOC-protected to yield 9. Slightly racemized 9
comprehensive synthesis scheme of the medicinal was purified by preparative chiral HPLC (step v) to
www.helv.wiley.com
(4 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
3
Scheme 1. Synthesis of [ H]CDZ173. Reagents and conditions: i) POCl , Et N, Tol, 100 °C, 2 h; ii) (S)-6, EtOH, AcOH, 140 °C, 16 h; iii)
3
3
Pd(OH ) on C, H (gas), EtOH, 60 °C, 16 h; iv) Boc O, Et N, CH Cl , 0 °C ? r.t., 3 h; v) preparative chiral purification; vi) 1. HCl, Et O, 0 °C
2
2
2
3
2
2
2
?
r.t., 2 h; 2. 1 M NaOH, THF/MeOH; vii) see Table 2 for conditions used for X = I, Br, and Cl; viii) 10% Pd on C, DIPEA, 19 mbar T
2
-gas,
DMF, 150 °C, 2 h; ix) Cl(CO)CH CH , Et N, THF, r.t., 2 h; x) 4 M HCl, CH Cl , r.t., 2 h.
2
3
3
2
2
yield – after the removal of the BOC group (step vi) – source for electrophilic halogen (El). Table 2 summa-
optically pure (S)-8. This precursor was used for palla- rizes the conditions under which these pyridines were
dium-catalyzed amination with the bis-halides 16, 17, obtained. In order to avoid by-products due to trans-
[
14]
and 18. The pyridines 16, 17, and 18 were prepared metalation and/or iodine-migration
strict tempera-
by metalation of the CF -pyridine core using LDA as ture control was applied with the result that stereo-
3
[
12][13]
super base
and iodine, C Br Cl , or C Cl as chemically pure 16 was obtained (Entry 1; see
2 2 4 2 6
www.helv.wiley.com
(5 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
Table 2. Experimental conditions for site-selective metalation of the halo-pyridines 16, 17, and 18 and subsequent electrophile
trapping
[
a]
X
n
T
1
[°C]
t [min]
El
equiv. of El
T
2
[°C]
Yield
60%
Remark
I (16)
1.0
< ꢀ90
180
I
2
1.5 (in portions)
ꢀ75
Cat. amount of LiBr was present and
control of T and T
1 2
I (16)
Br (17)
Cl (18)
2.5
2.0
2.5
ꢀ75
ꢀ75
ꢀ75
45
45
45
I2
4.0
2.0
2.0
0
0
0
45%
20%
48%
C
C
2
Br
Cl
2 4
Cl
2
6
[
a]
Yield and stereochemical purity were determined by GC-MS.
Supplementary Material for GC-MS of 16 prior to chro- the later content of tritium and the specific activity in
matography). However, under these conditions only 1b depends on the decreasing bond strength
2
6
0% starting material conversion was observed. between a sp -carbon and a halogen X (C–Cl > C–Br >
[
15]
Hence, since chromatography needed to be applied C–I),
achieving C(5)-coupling selectivity during the
anyway to isolate the product, an excess of LDA was amination step with the iodide and the bromide was
used in subsequent preparations (Entries 2 – 4). X-Ray prioritized first.
data of 16, 17, and 18 are shown in the Supplemen-
tary Material to ultimately prove the stereochemical amination step (vii) using the substrates 16 – 18.
Table 3 summarizes the conditions applied in the
[16 – 18]
identity of these building blocks.
When the pyridine-iodide 16 and the pyridine-bromide
In subsequent chemical methodology experiments, 17 were used as coupling substrates (Entries 1 and 2),
S)-8 was coupled under Buchwald–Hartwig conditions the C(5)-coupling products 12 (iodide) and 13 (bro-
(
with either of the bis-halogenated pyridines in step mide) were found in traces in the crude mixtures (see
vii. Goal of this methodology sub-project was to test Supplementary Material for analytical data). ROESY- and
to which extend a C(5)-coupling product can be HMBC-experiments of a 95:5-mixture of the bromides
favored if two halogens on C(4) and C(5) are present. 11 and 13 indicated the presence of 13 by cross peaks
Since the reaction rate of the following step viii, hence characteristic for this isomer (see Scheme 1 and
Table 3. Palladium catalyzed amination conditions tested to achieve C(5)-selectivity
Entry
Pyridine
Amine
Catalyst system
Ratio
C(4)–N:C(5)–N
[a]
[
17]
1
2
3
16 (X = I)
(S)-10
(S)-10
(S)-10
Pd
low yield
Pd (dba) , BINAP, Cs
yield isolated as fraction with C(5)-isomer enrichment
2
(dba)
3
, BINAP, KF on Al
2
O
3
, solvent free, 120 °C, MW @ 240 W, 2 h,
> 99:1
[
16]
17 (X = Br)
17 (X = Br)
2
3
2
CO
3
, dioxane, 135 °C, MW @ 240 W, 2 h,
< 5%
(higher yield in crude
BuPhos, NaO Bu, BuOH, 100 °C, MW @ 230 W, 3 h
ca. 95:5
ca. 95:5
t
t
t
[18]
Pd (dba) , BuMePhos, NaO Bu, BuOH, 100 °C, 4 h
2
3
product mixture)
[
b]
t
t
t
4
5
17 (X = Br)
18 (X = Cl)
TIQ
(S)-10
Pd
Pd
2
(dba)
(dba)
3
, Me
4
No reaction
> 1:99
t
t
t
2
3
, BuMePhos, NaO Bu, BuOH, 100 °C, 4 h, 6% isolated
[
a]
Entry 1 and 3: ratio of isomers determined by LC/MS in the crude mixture. In case of Entry 1, the structure of the isomer is given
by the substitution pattern of the pyridine 23. Entry 2: ratio of isomers was determined by LC/MS of the enriched fraction. Entry 5:
[
b]
isomer ratio in the crude product is estimated since compound 29 was the only isomer which could be isolated.
quinoline (TIQ) was used as a synthetic model for the amine 10.
Tetrahydroiso-
www.helv.wiley.com
(6 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
Supplementary Material). However, when the 95:5 mix- access to the halogenides 12 and 13 in practical
tures of 11 and 13 was tritiated, the small about of triti- yields became less important during the ongoing of
ated 1b (from 13) was lost during preparative HPLC the project. Nevertheless, a synthetic tool was pro-
purification (no RA detector could be used). Conse- vided which can fulfil the requirements of bioanalysts
quently, the major product (S)-15 was isolated and as discussed above.
characterized by correct mass peak but a different
3
HPLC-retention time as required for [ H]CDZ173.
3
Validation of the H-Tracer as ADME-Tool by Incubation
In an attempt to favor C(5)-N amination over C(4)-N
amination occurring next to the bulky, vicinal
CF -group on Py, the sterically more demanding ligand In order to test the metabolic stability of [ H]CDZ173
Me BuPhos with tetramethylation on phosphine bear- and validating it as a potential tool for ADME-studies,
with Hepatocyctes
3
3
t
4
ing phenyl was tested. This hypothesis however was a sample of 1b with a specific activity of 630 MBq/
falsified (Table 3, Entry 4). Since none of the known mmol was diluted 1:50 with the phosphate salt of
+
ꢀ
4
ligands and catalyst systems tested in step vii provided unlabeled CDZ173 ([C H F N O ] [H PO ] ). Micro-
2
1
26 3
6
2
2
the stereoelectronic effects on the transition metal grams of this diluted material were incubated with
needed to achieve C(5)-coupling product enrichment cryopreserved hepatocytes from rats and humans at
and since HPLC-isolation of 1b from a mixture of (S)-15 5 lM concentrations for up to 4 h and the incubate
and 1b was impossible, tritiating either 12 or 13 profiles were analyzed by UPLC with radioactivity
needed to be abandoned. Instead, when the relatively detection. Figure 2 shows the incubation profile after
easy accessible chloro-pyridine 18 was used in step vii, 4 h in rat hepatocytes (ratHEP) and human hepato-
the CDZ173-derivative 14 was the only coupling pro- cytes (huHEP). The metabolism of this diluted material
duct to be observed in step vii. The isolated amount of in rat and human hepatocyte incubations showed the
1
4 (< 10% yield) was sufficient to complete the tritia- formation of eleven phase I metabolites. The following
tion within this project.
baseline separated hepatic metabolites M were ana-
lyzed by LC/MS and LC/MS on a Q-ToF Synapt I mass
H
2
Radiosynthesis. In order to cope with the analytical spectrometer: M 14, M 16, M 1, M 14, M 4, M 3,
H
H
H
H
H
H
requirements of metabolic profiling from a regulatory M 10, M 21, M 5, M 6, M 7, M 25, and M 9. In this
H
H
H
H
H
H
H
[19]
perspective,
radioanalytical tools are needed, that row, the metabolites are ordered according to their
provide excellent sensitivity. Typically, the limit of retention time (see Table 4, column 2 for values of tR
detection of a tritium-labelled analyte it is 0.2 – 0.7 Bq and column 11 for relative peak areas). It was
[20]
(10 – 40 dpm).
Practically this means, that a tritium- observed that CDZ173 is heavily metabolized in
tracer is able to detect the xenobiotic analyte in a pico- human and rat hepatocytes with M 1 and M 5 being
H
H
to femtomolar range allowing accurate and sensitive the most abundant. Metabolite structural information
determination of metabolism pathways in ADME studies. was obtained to evaluate the metabolically stability of
3
Upon Cl-tritium exchange on 15 in step viii, the the isotopomer of [ H]CDZ173 shown in figure 6. In
radiotracer 1b was obtained after purification with a column 4 of Table 4, the elementary composition of
+
specific activity of 0.623 TBq/mmol after purification. the MH calculated from the observed [M + H] (col-
3
The correct position of the H-label was confirmed by umn 6) is listed. It is compared with the calculated
3
+
1
(
D- H-NMR (spectra not shown; singulet at
d
exact mass for [M + H] (column 5) to give the devia-
3
H) = 7.87 ppm) and compared with the value of the tion of the observed m/z (in ppm) relative to the cal-
fully assigned, unlabeled material (d( H) = 7.87 ppm; culated value (column 7). For possible interpretations
see Experimental Section). The radiochemical yield of and individual MS /MS +ESI high resolution spectra
8% referred to 1 9 100% tritium-labeling in one of the metabolites obtained from hepatocyte incuba-
1
1
2
5
position (max. 1.08 TBq/mmol possible) seems to be tions: see Supplementary Material.
small compared to a hypothetical iodine-tritium
exchange with ca. 90% radiochemical yield. However, Structural Conclusions from ESI+ Fragmentation Patterns
considering the requirements of analytical chemists of Hepatocyte Metabolites M and Metabolite Clustering
H
on the minimum specific activity of radiotracers in
[
21]
general,
b with a specific activity = 0.623 TBq/mmol (undi- fragmentation of CDZ173 in the MS/MS spectra
luted) has a limit of detection in the attogram range captures earlier interpretations of MS-data of 1,2,3,4-
the presumably metabolically stable tracer Reference CDZ173. Our interpretation of the ESI-
2
1
[
22][23]
of the analyte. Hence, in this context, the interesting tetrahydroisoquinolines (THQ).
In these studies,
question of C(4)/C(5)-coupling selectivity for getting the fragmentation of the piperidine moiety in THQ
www.helv.wiley.com
(7 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
3
2
Figure 2. Profiles of the hepatic metabolites M of [ H]CDZ173 after 4 h incubation with rat and human hepatocytes.
H
2
was interpreted to occur via a retro Diels–Alder
The MS of the unlabeled reference CDZ173 (Fig-
+
reaction (rDA). Yet, key to our interpretation of the ure 3) shows [M + H] at m/z 451.20 and structurally
ESI-MS data is the assumption that the tertiary, characteristic fragment ions at m/z 247.15, 191.13, and
aromatic amine function in CDZ173 (Py-N-piperidine) 174.1029. Below this, the fragment ions become noisy
should have a similar low E1/2 as the model (m/z 147.09, 124.09, and 122.07), what we interpret as
compound
E1/2 = ꢀ0.28 V)
tetramethyl-p-phenyldiamine
(TMPD; increasing tendency to form radicals and radical reac-
[
24]
and therefore easily oxidizes to tion products (e.g. m/z 149.09). One possible molecular
form a nitrogen centered radical initiating a rDA- interpretation of this fragmentation pattern is shown
process. To compare, the redox-potential of 4-amino- in Scheme 2. Upon oxidation of the piperidine nitro-
pyrimidine and pyridine was determined to be gen, a rDA in the THQ-moiety is postulated, which can
[
25]
E1/2 = ꢀ1.115 and E = 1.82 V, respectively.
The explain m/z 247.15. This fragment ion depropionlyates
1/2
0
peak potential (E ) of N,N -dimethylpropionamide was to yield m/z 191.12. Subsequent elimination of
p
[
26]
found to be E = 1.26 V.
pyrimidine in CDZ173 is unlikely.
Hence, oxidizing N(1) of
p
2
For reasons of graphic clarity, the subscript ‘H’ is
ignored in the annotations of Figure 2.
www.helv.wiley.com
(8 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
www.helv.wiley.com
(9 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
www.helv.wiley.com
(10 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
www.helv.wiley.com
(11 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
Figure 3. Q-TOF electrospray ionization MS/MS-spectra of CDZ173.
ammonia gives the tautomers of m/z 174.10. Upon additional double bond is a strong argument to narrow
+
loss of m/z 27 (probabely as [C H ] ), the distonoid down the position of the hydroxyls to the pyrrolidine
2
3
ion m/z 147.09 seems to be formed, which – due to moiety. The fundamental difference between M 6, M 7
H
H
its reactivity – gives rise to slightly smeared (noisy) and M 14, M 16 though is that the first pair contains a
H
H
fragment patterns in this spectral region. This is O-methyl-group, while the second pair does not.
expressed by losing formally C H in order to form m/ Compared to the microsomal metabolites of group B
2
4
z 122.07 and a reduced species m/z 124.09.
(Table 1), four metabolites from hepatocyte incubation
+
Similar to the metabolite observed in microsomes were clearly identified where i) a [M + H] was
M ), the hepatocyte metabolites (M ) can be chate- detected and ii) the site of hyroxylation can be
(
M
H
0
0
goterized into the groups A – E . For structures and narrowed down with sufficient confidence to the
possible fragment interpretation of the metabolites pyrrolidine ring. Few O-demethylated metabolites were
discussed in this paragraph, see Supplementary Mate- detected in microsomes. It must be noted, that in
rial 1.3.
hepatocyte incubations no M 17-like bioconjugate was
M
detected which is a further argument that M 17 is an
Metabolites: Metabolite M 1. A precise assay artefact. In this context, it is likely that
M
0
Group
A
H
+
[
(
M + H] (+ 0.655 ppm) was detected for the dehydrogenated ion peaks detected earlier for M 11
M
hydrolytic) demethylation metabolite M 1. Therefore it and M 25 in CID-MS/MS data (Table 1 and Table S1)
n
H
M
can be hypothesized that both metabolites M 1 and indeed are ionization artefacts, i.e., it is justified to keep
M
M 1 are identical. Further, that this metabolite is an M 11 and M 25 in group B of the microsomal
H
M
M
0
integral part of metabolism in microsomes and metabolites. Consequently M 6 and M 7 in group B
H
H
hepatocytes. Practically, the identity was confirmed by correspond to the precursor of M 11 and M 25 in
M
M
coinjection with independently synthesized material group B, i.e., very likely M 11 and M 25 were spectral
M
M
(
see below). Interestingly, in hepatocyte incubates, a artefacts due to in source fragmentation of H O.
2
hydrolysis product corresponding to M 2 was not
found.
M
0
Group
C
Metabolite: N-Propionyl-pyrrolidine Mono-
Metabolite M 10. Comparing the
Hydroxylation
H
0
Group B Metabolites: Pyrrolidine-mono-hydroxylation fragmentation pattern of M 10 with the fragmentation
H
0
Metabolites M 6, M 7, M 14, M 16. The common pattern of B -group metabolites reveals that
H
H
H
H
0
fragmentation motive of the B -group metabolites M 6, characteristic ion fragments of M 10 are not
H
H
M 7, M 14, and M 16 are very similar to the key dehydrogenated, i.e., m/z 191.12 and 174.10 are found.
H
H
H
+
fragments postulated for the rDA process in CDZ173 Since a [M + H] was well defined in M 10 while no
H
(
(
2
blue in Scheme 2). Compared to the parent’s fragments fragment dehydrogenation was observed, the hydroxy
Scheme 2), the key fragments of these metabolites m/z group of M 10 must be located in the propionyl side
H
0
45.15, 189.11, and 172.08 are dehydrogenated, i.e., chain. Therefore, M 10 in group C corresponds
H
contain an additional double bond compared to the structurally to one of the three metabolites in the
parents fragment. Together with the well-defined microsomal group C. Dehydrated forms of mono-
+
[
M + H] of these metabolites (see Table 4), this hydroxylated MH corresponding to M 25 and
M
www.helv.wiley.com
(12 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
n
Scheme 2. ESI-MS/MS retro-Diels–Alder fragmentation hypothesis of unlabeled CDZ173. Structural key fragments observed in
metabolites of CDZ173 are highlighted in blue.
conjugates corresponding to M 17 (Table 1) were not reduction. In both cases, the characteristic cations m/z
M
2
found in hepatocytes.
326.1194 and 326.1227 in MS were found. This
finding narrows down the site of a C–N bond
Group D Metabolites: Formally Hydrated Metabolites cleavage to the pyrrolidine moiety, however at this
M 4 and M 21. For M 4 and M 21, very precise point it is not possible to distinguish which ring
0
H
H
H
H
+
[
M + H] at m/z 469.2169 and 469.2175 were opening isomer is formed. Due to pyrroldine ring
observed (ꢁ0 ppm in the case of M 4, +1.279 ppm in opening, m/z 265.1633 (M 4) and m/z 265.1701
H
H
the case of M 21. Both metabolites have very similar (M 21) fragment differently than the corresponding
H
H
fragmentation patterns with characteristic fragments key fragment of CDZ173 (Scheme 2). Upon formal
m/z 265.1633 (M 4) and m/z 265.1701 (M 21), which elimination of propionamide, m/z 192.1 was found.
H
H
formally corresponds to a hydrate of the CDZ-key According to the nitrogen rule, this exact fragment
fragment m/z 247.15, followed by ring opening and mass only is possible, when an impair number of
www.helv.wiley.com
(13 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
nitrogen atoms is present, i.e., the cations at m/z 192.1 physiological context we are proposing for M 3 and
H
must contain oxygen. However, at this point two M 15 the structures as shown in Table 4.
H
fragmentation solutions for these MS data are
In the case of M 34, the situation is different
H
possible (see Supplementary Material), and thus it is because per se the metabolite is O-demethylated. Nev-
impossible to distinguish by MS which pyrroldine ring ertheless the rDA-key fragments at m/z 279.1473
opening isomers are formed (principally, ring opening ((OH) -analogon together with a reduced species as
2
is possible either via the N–C(2) or via the N–C(5) observed for CDZ173) and at m/z 223.1182 ((OH)2-
bond, while transamidation of propionyl to OH is analgon of m/z 191.12 found for CDZ173; Scheme 2)
2
chemically very unlikely; see Table 1 for isomeric were found in MS . Based on this finding the exact
possibilities of these aminoalcohols). Based on the site of oxidation of the pyrrolidine ring is not clear.
significantly different retention times of the discussed
0
metabolites, it can be conclude that M 4 ꢂ M 4 and Group X Metabolite: N-Oxide M 9. M 9 is an previously
H
M
H
H
+
M 21 ꢂ M 21.
not detected N-oxide [M + H] at m/z 467.2018
ꢁ0 ppm). M 9 shows the same fragments m/z
H
M
(
H
0
Group E Metabolites: Formal Pyrrolidine-bis-hydroxylation 324.1045, and m/z 263.1514 as detected for the
Metabolites M 3, M 5, and M 34. M 3 shows synthetic reference 20 (see Experimental Section),
H
H
H
H
+
[
M + H] = 483.1965. Formally, for M 3, a (OH) - however the characteristic H NO-elimination fragment
analogon of the key fragment m/z 247.15 was found (radical cation [M] at m/z 435.1817) is better seen in
H 2 2
ꢃ+
2
(
m/z 279.1433). Further fragmentation yielded a vinyl + ESI MS of 20 (see Supplementary Material). This
at m/z 160.0869 as final fragment (i.e., a fragment characteristic fragment defines the oxygen to be
which contains one methylene less than the key- located at N(1) of the pyrimidine. rDA-fragmentation
fragment m/z 174.1 in Scheme 2). Precursors for m/z of M 9 yields the O-adduct of the characteristic key
H
1
3
2
60.0869 are the fragments m/z 206.0928 and fragment m/z 247.15 at m/z 263.1514. In case of M 9,
64.1476. Since the difference of m/z 160.0869 and a fragmentation competing with the rDA model is
H
06.0928 is 46.0059, it is reasonable to postulate the observed. This is seen in the fragment m/z 324.1045,
+
loss of CH O2 in this step (EM = 46.0055). This in which [M + H] losses the pyrollidine moiety and
2
indicates that i) m/z 206.0928 can be a cyclopropane formally eliminates H O.
2
with carboxyl substitution, in which the carboxyl
0
0
substituent originates from the 2 -carbon of Group Y Metabolite: THP-Dehydrogenated Metabolite
0
0
+
pyrroldine, ii) the C moiety originates from C(3 ), C(4 ) M 25. M 25 exhibits at characteristic [M + H] at m/z
3
H
H
0
and C(5 ), and iii) initial geminal bis-hydroxylation 449.1905, which proves the presence of an additional
ending in ring opening occurs on C(2 ) (for structure double bond equivalent compared to the parent.
0
proposal see Supplementary Material 1.3). In this Therefore, rDA process in the THP moiety cannot be
context, it must be noted that m/z 244.1433 was not observed. Instead M 25 loses propionyl. The pyrrolidine
H
[
27]
14
observed in other MS studies
with [ C]CDZ173, i.e., ring eliminates NH3 (– m/z 17) to form butyl and
it is very likely an artefact. On the other side, M 5 also acetylen fragments (m/z 376.1410 and 348.1068). In
H
exhibits a very intense m/z 483.1974 for its formal parallel, M 25 was prepared synthetically (compound
H
(
OH) -analogon, which either can be a geminal bis- 21, Experimental Section) to further characterize M 25.
2 H
0
hydroxy on the same carbon (e.g., on C(5 ) of M 25 may not be confused with the microsomal
pyrrolidine) or the corresponding C(5 )-acid upon ring metabolite M 25, where the double bond is located in
H
0
M
opening. m/z 483.1974 fragments into the the pyrrolidine moiety (M 25
depropionlyated rDA key fragments of its parent
6¼ M 25).
H
M
(
Scheme 2) to yield m/z 279.1431 (= m/z 247 + m/z Structure Confirmation of M 1, M 9, and M 25 by Co-
H H H
3
2). Compared to M 3, m/z 279.1431 of M 5 Injection with Synthetic References. To unambiguously
H H
fragments in a series of homologs (– m/z 29, – m/z 36, confirm the structure of M 1, M 9, and M 25, the
H
H
H
–
m/z 36). Similar homologs are observed in the ESI of synthetic references 19 (M 1), 20 (M 9), and 21
H H
[
28]
b-alanine and 1-propionic acid.
We interpret this (M 25) were synthesized (see Figure 4 for structures).
H
difference in fragmentation as an argument that M 5 The O-demethylated product 19 and the pyrimidine-
H
0
is the corresponding C(2 )-acid with pyrrolidine ring N-oxide 20 were accessible by incubation of CDZ173
opening. To conclude, M 3 and M 5 ≙ M 3 and M 15 with the soil micro-organisms S. platensis and S.
H
H
M
M
in the microsomal incubation (differences in retention griseus, respectively (see Experimental Section for
time of M 15 and M 15 are explained since different details). Facing the low one-electron oxidation-
H
M
HPLC hardware was used in these incubations). In the potential of TMPD compared to the oxidation
www.helv.wiley.com (14 of 33) e1800044 © 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
Figure 4. a) Radio-LC/MS traces of the incubation of CDZ173 with huHEP after 4 h. 1) LC/MS at 447.18 – 449.19 Da. 2) LC/MS at
437.18 Da. 3) LC/MS trace at 467.19 Da. b) LC/MS of A at the same mass points 1), 2), and 3) after adding a 1:1:1:1 mixture of 19, 20,
2
1, and 22.
potential of 4-amino-pyrimidine, the selectivity in this Section for C-atom numbering in CDZ173), the signal
aerobic biotransformation process is surprising. 20 of C(8d″) in 20 was shifted high field (d
+
was characterized by
ESI-MS/MS showing the (C) = 150.1 ppm). Compound 21 – the synthetic
ꢃ+
expected radical cation [M] at m/z 435.1817 (see reference for M 25 – was obtained as a mixture with
Supplementary Material) and by comparing the C- the two diastero-isomers of 22 by oxidation of
spectra of 20 with the C-NMR of CDZ173. There a CDZ173 with laccase from Trametes versicolor. To
high field shift for C(2″) in 20 compared to C(2″) in confirm the structures of M 1, M 9, and M 25, a
H
1
3
1
3
H
H
H
CDZ173 was observed (147.6 ppm in 20 vs. diluted mixture of 19 – 22 was prepared and co-
55.8 ppm in CDZ173). Compared to the signal for C injected with incubates from huHEP and ratHEP after
8d″) in CDZ173 (d(C) = 158.8 ppm; see Experimental 4 h incubation time. Figure 4 shows the LC/MS traces
1
(
www.helv.wiley.com
(15 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
at different mass points. It can be seen, that after co- (e.g., M1) and pyrrolidine hydroxylation (e.g., M6, M7)
injection M 1 (t = 8.31), M 9 (t = 2.09), M 25 as well as several minor oxidative pathways. Similar
H
R
H
R
H
(
t = 7.44) and two diasteroisomers of a previously metabolites were identified to those found in the
R
0
not identified hepatic metabolites M ξ and M ξ
qualitative microsomal metabolism investigation
H
H
(
t = 11.32 and t = 20.15) came up. Based on this co- (Table 1) except that the THP oxidation pathway was
R R
injection it is proven that M 1 ꢂ 19, M 9 ꢂ 20, M 25 not identified in the hepatocyte experiment.
H
H
H
0
H
0
ꢂ 21, and M ξ,M ξ ꢂ (3 S,5″-rac)-22.
Additionally, no phase II metabolites were identified in
H
hepatocytes. These differences could be assigned to
Metabolic Pathway. Based on the data presented differences in metabolic pathways between microsomes
above, we were able to outline a hepatic metabolism and hepatocytes; however, as the hepatocyte study was
pathway based on metabolites formed in vitro in performed quantitatively with a tritium radiolabel and
hepatocytes (Scheme 3). The key metabolic pathways the microsome study was qualitative, it is also feasible
were oxidative, involving predominantly O-demethylation that the additional microsomal metabolites are simply
n
Scheme 3. Metabolic pathway of CDZ173 as concluded from MS/MS data acquired in this study.
www.helv.wiley.com
(16 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
quantitatively minor and therefore not detected in
hepatocyte experiment. This hypothesis is supported by
in vivo data showing that THP oxidation metabolites are
[27]
formed but at low abundance.
The next project
steps, i.e., identifying the cytochrome complexes
responsible for the biotransformation to the discussed
metabolites were not covered by the project plan at
the time of approval, but might be addressed at a
suitable point elsewhere.
NMR of CDZ173
NMR of CDZ173 and its synthetic metabolite references
1
9 – 22 was hampered due the dynamics of the amide
bond causing pseudo-rotamers and different conforma-
tions of the pyrrolidine ring. In this paragraph, we focus
on the NMR discussion of the common parent. Our
model to understand the 2D-NMR spectra of CDZ173
takes into account the rotational barrier of the amide
Scheme 4. cis/trans - UP/DOWN conformational isomerism in
CDZ173 and its derivatives.
bond of N-trifluoroacetyl pyrroldine. The activation bar- anticipate the same conformational colascence behav-
rier for cis-trans isomerism for this compound was ior for CDZ173 and its derivatives.
‡
ꢀ1
found to be DG
= 80.4 kJmol
98
in solution and
The existence of two conformers of CDZ173 at
2
‡
ꢀ1
1
DG298 = 71.4 kJmol
in the gas phase, while the 298 K was seen in the H-NMR, e.g. in CDZ173, where
0
pseudorotational barrier in pyrrolidine was calculated H–C(3 ) of the (S)-amino functionality in the pyrrolidine
to be 0.6 kJmol ab initio taking into account that the gave two well resolved sextets at 4.70 ppm ( J = 6.6 Hz)
nitrogen inversion of the pyrrolidine ring is of little and 4.58 ppm ( J = 6.4 Hz). Complete H- and C-signal
importance in N-trifluoroacetyl pyrrolidine.
ꢀ
1
3
3
1
13
[
29][30]
On the assignment with open diastereotopicity was achieved
other side, we consider the ab initio conformational from COSY, TOCSY, HSQC, and HMBC-data. COSY-,
0
analysis of Reiher and Seebach on C(2 )-substituted pyr- TOCSY-, HSQC-, and HMBC-signal assignment using stan-
roldines, where the pyrrolidine nitrogen is bound to a dard semi-automated assignment tools has shown that
2
[31]
sp -hydridized carbon. These authors calculated that NOE peaks exist, which were interpreted as signals
UP- and DOWN ring conformers of pyrrolidine are both between the CH -group of the propionly side chain
2
significantly populated when a Boltzmann distribution and either of the two amide-vicinal CH -groups of the
2
is in equilibrium. Similar to what is known for solubi- pyrrolidine conformers (Figure 5). Key to the differentia-
[
32][33]
lized poly-Pro-peptides
one must assume cis- and tion between H110 and H111 of one conformer and
trans-isomerism along the dihedral angle C(2)–C(=O)– H53 and H54 of the second conformer were distinct
0
N–C(2 ) in combination with UP/DOWN conformers TOCSY, HSQC- and HMBC-signals for these protons in
of the pyrroldine ring (see Experimental Section for a both conformers.
C-atom-numbering in CDZ173 consistent with rules of
Although noise was found in the f2-dimension of
nomenclature). Scheme 4 shows this situation together the ROESY-spectra, the observed NOE correlations are
with the characteristic NOE signals to be expected. The consistent with the validated assignment obtained
0
dihedral angle C(2)–C(=O)–N–C(2 ) is high-lighted in red. from the data set mentioned above, i.e., for the first
For reasons of graphic clarity the flat pyrimidalization of conformer (highlight green in Figure 5) qualitative
nitrogen is neglected.
cross-peak correlations between CH -propionly (pro-
2
0
NMR Spectra of CDZ173 showed that two distinct tons 53 and 54 in Figure 5) and the CH (2 ) group of
2
structures are present in a 1:1 ratio at room tempera- the pyrrolidine moiety (protons 46 and 47 in Figure 5)
ture in (D )DMSO which were assigned to be conform- were found defining the constitution of the amide
6
ers as illustrated in Scheme 4. In case of CDZ173 and bond in CDZ173 trans to the amino group in the
its synthetic metabolites and precursors, chemical pyrrolidine ring. Based on the hypothesis outlined in
homongenity was proven by HPLC and in case of the Scheme 4, a trans-amide bond with cross peaks to
synthetic precursor (S)-8 (Scheme 1) also by high tem- only one side of the pyrrolidine requires an UP-confor-
perature NMR. At 373 K, a single structrue was mation of the pyrrolidine. In this conformation, the
0
observed for (S)-8. For this reason, it is justified to oxygen is cis to the substituent on C(3 ).
www.helv.wiley.com
(17 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
Figure 5. ROESY-spectra of CDZ173 in the cross peak region of CH
2
-propionyl and the amide-vicinal CH
2
-groups of the pyrrolidine
(
600 MHz, 25 °C, Gaussian denoised and smoothened, semi-automated assignment with MNova v11.06).
In our data, these cross-peak correlations are and trans amide bond is present. This means that the
slightly, but distinctively high field shifted. Conse- pyrrolidine ring in this conformation of CDZ173 must
quently, the intense signals which are standing out be relatively flat compared to the conformer discussed
from the fading f2-noise line (e.g. cross peak 110/111 above.
↔
105 in the upper half of the spectrum) were inter-
This qualitative NMR-analysis shows that in
preted as NOE correlations of the second conforma- (D )DMSO two conformers each with different pyrro-
6
tion (red in Figure 5). Reflected cross peaks for this lidine puckering similar to the proline puckering in
[
34]
conformer without noise were found in the lower tri- peptides
have reached the energy minimum and
angle of the spectra. Since cross peaks between are present in solution at room temperature. Further it
hydrogen 110/111 and the hydrogens 103 and 105/ means that the rotational barrier between both con-
1
06 are observed, a qualitative solution conformation formers must be at least 81 kJ/mol and slightly lower
is very likely, in which the propionly side chain stands at body temperature so that one conformer binds to
anti-periplanar to the pyrrolidine ring and the car- PI3Kd. A similar conformational homogenization upon
bonyl-oxygen must point out of plane of the pyrrol- binding was observed in conformers of small molecule
[
35]
dine, i.e. a mixed conformational form between a cis PRO-ligands upon binding to a cofactor.
For the
www.helv.wiley.com (18 of 33) e1800044 © 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
synthetic metabolites 19 – 22, the same 1:1 ration choice to provide a safe precursor, which is ready for
1
was observed in the H-NMR spectra, however a quali- tritiation in a metabolically stable position at a time
tative conformational analysis was not undertaken for when the pharmaceutical compound still is in phase 0.
these metabolites, i.e. analytical data of these con- Nevertheless, following a standard operating protocol
formers are reported as ‘conformer A’ and ‘conformer (SOP) of the isotope lab unit, every tritiation candidate
B’ in the Experimental Section In the biologically rele- was tested for hydrogen isotope exchange using the
[
2][36]
vant X-ray structure of CDZ173 bound to PI3Kd,
Crabtree catalyst (typical procedure is first max.
the backbone of CDZ173 is cis, the oxygen is neither 5 mol-% Crabtree catalyst loading in CH Cl with
2
2
0
cis nor trans relative to the axial C(3 ) substituent and p(D ) = 1 atm using the balloon technique for gas
2
the pyrroldine puckering corresponds to the con- transfer, then up to equimolar amounts of catalyst
former IV in Scheme 4, while the carbonyl side chain is under the same conditions). The synergetic purpose of
in plane with the pyrrolidine. To conclude, during the this SOP is to test the general chemical reactivity
binding of solubilized CDZ173 to PI3Kd, a conforma- towards transition metal insertion, hence the possibility
tional change of CDZ173 occurs via homogenisation of metabolic loss of the label as it occurs in cytochrome
towards a planar cis backbone and a DOWN-puckering c enzyme by the iron cofactors. When this SOP condi-
of the ring. In (D )DMSO, this situtation corresponds tions were applied to CDZ173, no significant mass
6
0
to the second conformer with d(H) H–C(3 ) = change was detected by LC/MS. Under the SOP condi-
4
.58 ppm. From such a conformer the energy min- tions no exchange was observed in the methyl of the
imun seen in the biological X-ray structure must be esterified 2-pyrollidon moiety of CDZ173 as it was it
[37]
reached most easily.
was mentioned in the scientific literature
and repro-
duced with high specific activity by Crabtree-exchange
3
for the structurally related orexin 2 receptor EMBA by
Summary, Conclusions and Outlook for H-
Labelling Strategies
[38]
the first author.
In case of CDZ173, we assign this
finding to the presence of an electronic withdrawing
During this project, a first metabolic landscape based CF -group in CDZ173 compared to EMBA, where this
3
on ESI was created. Extensive metabolism was particular Ir-catalyzed exchange must follow an
observed; however the pyridine moiety of CDZ173 unorthodox exchange mechanism. However, even if this
remained chemically inert. Based on these data, a syn- Ir-catalyzed exchange reaction was successful in the
thetic target of a metabolically stable tritium tracer of OCH -group of CDZ173, the label would have been
3
3
CDZ173 ([ H]CDZ173, 1b) was designed and devel- lost during tracer application (see Table 1, group A
oped yielding several batches of enantiomerically pure metabolites).
3
[
H]CDZ173 with a specific activity of 600 GBq/lmol
After the development of the tritium label for this
and radiochemical purity > 98% HPLC. In a diluted for- project was completed, a pincer-type, oxygen sensitive
3
mulation [ H]CDZ173 was incubated with human and iron-catalyst was applied by the Chirik group in 2016
rat hepatocytes and the incubates were analyzed by for the labelling of structurally relevant model com-
high resolution ESI tandem MS. During the validation pounds (e.g. compounds 12 and 17 in figure 2 in the
3
of [ H]CDZ173, no tritated water was observed qualify- discussed publication) and the application of this cata-
ing this tracer as a tool to study liver metabolites. lyst for the hydrogen isotope exchange of e.g. the
3
Consequently, this H labelled tracer is suitable for use Merck compounds MK-7246 and MK-8228 as well as to
in preclinical and potentially clinical ADME studies the Amgen compound Cinacalcet (figure 3 in the dis-
[
39]
once the cytochrome C enzymes are identified cussed publication).
In case of MK-7246, exchange
0
responsible for the biotransformations. Hence, the was observed in an aromatic ortho, ortho position of
hypothesis postulated by Krauser that H-labelling is a fluoro-benzene. In case of Cinaclacet and MK-8228,
[
7]
3
a sophisticated and cost effective tool to i) replace exchange occurred in the meta-position of CF -phenyl
3
1
4
expensive C-multi step labelling and ii) to study and the meta-position of MeO–phenyl. Unfortunately,
ADME properties of modern pharmaceuticals was pro- the Chirik catalyst is chemically stable in C D for
6
6
[
40]
ven experimentally. Compared to tritium labels com- 1 day only.
Generally seen, the authors applied this
monly prepared by catalytic hydrogen-isotope exchange technology to aromatic compounds with a
exchange methods yielding more or less a random lower degree of substitution in the exchanging ben-
and difficult to predict distribution of tritium in the zene ring. To which extend the mainly di-substituted
most reactive positions of the molecule, non-radioactive substituents of the application molecules can be com-
organic synthesis combined with microsome screening pared with the substitution pattern of the pyridine in
and structure elucidation appears to be the tool of CDZ173 is unclear. In any case, relevant pyridine
www.helv.wiley.com
(19 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
positions in CDZ173 are occupied or sterically hin- human liver microsomes at a protein concentration of
dered. Consequently it remains to be tested if the car- 0.3 mg/mL in 0.1 M KH PO buffer (pH 7.4), CDZ173
2
4
bene-iron complex indeed can exchange in the (5 lM) and MgCl2 (5 mM). The mixture was pre-
electronically and sterically challenging positions H– incubated for 3 min at 37 °C before addition of GSH
C(6) of the pyridine in CDZ173.
(1.5 mM), alamethicin (1.25 lM), UDPGA (2.4 mM), and
In this context, it must be emphasized that our an NADPH regenerating system containing isocitrate-
microsome and hepatocyte incubations, which cer- dehydrogenase (1 U/mL), NADP (1 mM), isocitrate
tainly hydroxylate CDZ173 based on cytochrome (5 mM). Reactions were stopped after 1 h by addition
P450, non-heme diiron hydroxylases and a-keto-gluta- of an equal volume of ice-cold MeCN. The samples
rate dependent non-heme iron hydroxylases following were then vortex-mixed and centrifuged for 5 min at
[
41]
an oxygen rebound mechanism
are incompatible 10000 g. 0.1 mL of supernatant was diluted with
with the mechanism proposed for a Crabtree-type cat- 0.4 mL water and filtered (Ultrafree MC filters,
alyst. If a-keto-glutarate dependent non-heme iron Durapore PVDF, 0.45 lm, Millipore). The resulting
halogenases following a non-oxygen rebound mecha- supernatants were transferred to a clean autosampler
n
nism are present in our assays is not addressed. To vials for LC/MS analysis.
the best of our knowledge the last-named enzyme
class plays a minor role in liver metabolism. On the In vivo Experiments Using Bile Duct Cannulated
other side cytochrome c3 and cytochrome c6, which Sprague–Dawley Rats. White coat Sprague Dawley male
both have hydrogenase activity under certain condi- rats (Crl:CD (SD), 320 g) were used and were obtained
[
42]
tions
are indeed relevant e.g. in the liver metabo- from Charles River Laboratories (Iffa-Credo, France). The
[
43]
lism of taxol.
To which extend [FeFe]-hydrogenase in vivo experiment was performed according to the
cluster exchange in H–C(6) of the pyridine and there- regulations effective in the Canton Basel-City,
fore are a mechanistic model for the Chirik catalyst is Switzerland, specifically according to experimental
unclear. This argumentation further supports the license No. 2241. The study was performed as
[
45]
experiment suggested with this catalyst.
described previously.
Briefly, the Sprague–Dawley
To conclude, our synthetic approach at the time male rat as specified in the general experimental part
when the tritium-synthesis was developed is defined by was used for this study. The day before drug
nature and bio-analytics rather than ab initio manual administration, the animal was anaesthetized and
screening of catalysts in order to achieve the goal. Of then, under aseptic conditions, three catheters were
course, in this context it must be noted, that analytical successively implanted and fixed into its femoral
assays are known to accelerate catalyst screening (e.g. artery, femoral vein, and bile duct for blood collection,
[
44]
ref. ). Seen from an economic point of view, it must be drug administration, and continuous bile collection,
evaluated at which point in the pharmaceutical develop- respectively. The animals were allowed to recover for
ment process bigger resources shall be spent on auto- 24 h following surgery. The duration of the study was
mated exchange catalyst screening. Only future basic 8 h, throughout which, the animal was housed in a
research can show to which extend the hydrogenase metabolic cage. CDZ173 (3 mg/kg, 0.5 mL/kg
a
activity in a hepatocyte incubation assay of a pharma- solution in NMP/PEG, 30:70, v/v) was administered
cophore run under 95% air and 5% CO in aqueous buf- intravenously via the femoral vein catheter. Bile was
2
fer (see Experimental Section) can predict if a synthetic collected for up to 8 h post-dose at pre-defined time
catalyst, which – ideally – should be identified at least for intervals (0 – 2 h, 2 – 4 h, 4 – 6 h, 6 – 8 h) while
every lead structure of a med-chem library, can be sufficient urine was only obtained during 4 – 6 h.
applied successfully. Achieving this goal would certainly Blood samples were collected at 1, 4, and 8 h post-
raise the practicability of the hydrogen-isotope technol- dose. All samples were collected on ice, then frozen
ogy in case of some clinical candidates.
as soon as possible at ꢀ80 °C until analysis.
n
Preparation of in vivo Samples for LC/MS . Bile and
urine samples were thawed, centrifuged, and pooled
Experimental Section
Incubation with Liver Microsomes and Sample Preparation
from Bile Cannulated rat and LC/MS Analytical Method
from each time intervals and diluted with H O/MeCN
(9:1). The diluted samples were then filtered to
2
n
remove any precipitated material. Blood samples
In vitro Incubations. Incubations (0.5 mL each) were (0.1 mL) were treated with 0.3 mL ice-cold MeCN,
performed at 37 °C in a shaking incubator. The thoroughly mixed and then centrifuged (10 000 g,
incubations contained rat, mouse, dog, monkey, and 5 min). A volume of 0.4 mL of the supernatant was
www.helv.wiley.com
(20 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
transferred to a new vial and concentrated using a CH Cl : dichloromethane, DEA: diethanolamine, Et O:
2
2
2
Cyclone high speed evaporator. The residue was then diethyl ether, DMSO: dimethyl sulfoxide, AcOEt: ethyl
reconstituted with 0.25 mL H O/MeCN (9:1) and acetate, ETPH: ethyl-benzene, ESI: electrospray ionization,
2
filtered. The resulting supernatants were transferred to FA: formic acid, GSH: glutathione, HEP: heptane, HEX:
n
clean autosampler vials for LC/MS analysis.
hexane, AcOH: acetic acid, IPA: isopropylalcohol, i.v.:
intravenous, LC/MS : liquid chromatography/multi-stage
n
n
LC/MS
Analytical Method. Capillary HPLC was mass spectrometry, MeOH: methanol, MeONa: sodium
performed using a Chorus-220 HPLC pump (CTC methoxide, Me - BuPhos: di-tert-butyl(2 ,4 ,6 -triisopropyl-
Analytics, Switzerland), a HotDog-5090 column oven 3,4,5,6-tetramethyl-[1,1 -biphenyl]-2-yl)phosphine,
t
0
0
0
4
0
(
Prolab, Switzerland) and a HTS-PAL or an LC-PAL MOPS: 4-morpholinepropanesulfonic acid, NADP:
autosampler with cooled sample stacks (CTC Analytics, nicotinamide-adenine-dinucleotide-phosphate, NMP: N-
Switzerland). Chromatographic separations were Methylpyrrolidone, Pd dba : Tris(dibenzylideneacetone)-
2
3
performed using a Reprosil Basic HD C18 capillary HPLC dipalladium(0), PEG: polyethylenglycol, r.t.: room
column (3.0 lm, 0.3 mm 9 150 mm, Morvay Analytik, temperature, t : retention time, SD: Sprague Dawley, Et N:
R
3
Basel, Switzerland) with a flow rate was 4.5 lL/min. The triethylamine, TFA: trifluoroacetic acid, TEMPO: (2,2,6,6-
injection volume was 1 lL. Mobile phase A consisted of tetramethylpiperidin-1-yl)oxidantyl,
aqueous ammonium formate (10 mM, with 0.02% TFA, 5 -diphosphoglucuronic acid.
UDPGA:
uridine
0
pH 6)/MeCN (95:5, v/v). Mobile phase B consisted of
aqueous ammonium formate (10 mM, with 0.02% TFA, Biochemicals, Chemicals, and Reagents: Analytical
â
pH 6/MeCN/MeOH (5:90:5, v/v/v).Chromatographic Reagents. CHROMASOLV Plus high-performance liquid
separations used a linear solvent gradient: 0 min (5% chromatography (HPLC) grade (Sigma, Switzerland) or
B), 2 min (5% B), 25 min (95% B), and 32 min (95% B, HiPerSolv ChromaNorm HPLC/MS grade (VWR
6
.5 lL/min). Re-equilibration of the column was Chemicals, France) water was further purified by
performed at 5% B for 5 min. The column temperature distillation using a Water Still Distinction D4000 (GMB
was maintained at 40 °C. Samples were stored at 10 °C Glasmechanik, Switzerland) and stored in plastic
in the autosampler until injection. The Capillary HPLC labware for usage on the capillary HPLC system.
â
system was coupled to a LTQ XL Ion Trap/Orbitrap CHROMASOLV HPLC grade acetonitrile (MeCN), HPLC
(
Thermo Scientific, CA, USA) mass spectrometer grade methanol (MeOH), phosphate buffer (1.0 M pH
equipped with an ESI source (Thermo Scientific, CA, 7.4), TFA was purchased from Sigma (Switzerland).
USA) operating in positive mode with a source voltage Ammonium formate was purchased from Fluka
of approximately 4 kV and a capillary temperature of (Switzerland). Nitrogen (purity 99.5%) and Argon
2
75 °C. Auxiliary and sheath gas flow rates (both 99% (purity 99.995%) were purchased from Carbagas
purity N ) were 1 and 15 (arbitrary units), respectively, (Switzerland). Pooled rat, mouse, dog, cynomolgus
2
whereas sweep gas was not used. Mass spectral data monkey, and human liver microsomes were
were acquired in data-dependent mode with full-scan purchased from BD Biosciences (Woburn, MA). Culture
spectra obtained from m/z 150 – 1000 with a mass media NL148S: 0.2% soluble starch (Sigma), 0.8% soya
resolution of 30 000. CID data-dependent MS/MS scan peptone (Sigma), 0.4% Oxoid Lab-Lemco, 0.05% yeast
events were carried out with normalized CE of 25 extract (Sigma), 0.15% NaCl, 0.21% MOPS, adjusted to
2
arbitrary units. MS scan events were triggered for the pH 6.50 (1 N NaOH/1 N HCl), and trace element
3
two most intense ions and MS scan events were solution (1 mL/L). The trace element consisted out of
2
triggered for the two most intense ions from each MS
Na MoO ꢃ 2 H O (30 mg/L), FeSO ꢃ 7 H O (5.5 g/L),
2
4
2
4
2
5
4
scan. AGC target settings were 5 9 10 and 1 9 10
CuSO ꢃ 5 H O (80 mg/L), MnCl ꢃ 4 H O (180 mg/L),
4
2
2
2
for full MS and MS/MS, respectively. The polysiloxane ZnSO ꢃ 7 H O (4.4 g/L), and H SO 97% (2 mL/L).
4
2
2
4
background ion [C H SiO] with m/z 445.12003 was Streptomyces platensis: strain type 40041 (DSM).
2
6
6
used as lock mass.
Preculture of streptomyces griseus (strain ATCC No.
5185) was grown in 100 mL NL148s medium in
5
Synthesis and Radiosynthesis
500 mL shake flasks at 26 °C and 220 rpm for 3 days.
Nutrient solution: 40 g/L glucose, 40 g/L lactose, 60 g/
General Abbreviations. AGC: Auto gain control, MeCN: L sodium citrate dihydrate. The reference 1a for the
0
0
0
acetonitrile, BINAP: (2,2 -bis(diphenylphosphino)-1,1 - release analysis of 1b was a registered sample from
t
t
binaphthyl),
methylbiphenyl, BuOH: tert-butanol, CE: collision energy, 4.5: prepared according to.
CID: collision induced dissociation, CYH: cyclohexane, molecule organic synthesis: All starting materials and
BuMePhos:
2-(di- butylphosphino)-2 - the Novartis compound archive. McIllvaine buffer pH
t
[46]
Reagents for small
www.helv.wiley.com
(21 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
solvents were of common commercial origin (Sigma) separation on Venusil chiral OD-H, 4.6 9 250 mm,
and were used without further purification.
5 lm, 25 °C, k = 254 nm, flow 1 mL/min, P = 4.0 –
.4 MPa, isocratic separation with 0.1% Et N in Hex/
5
3
Special Equipment for Synthesis and Glassware for Non- EtOH 90:10. HPLC-Method II (chiral preparative): Gilson.
Radioactive Synthesis. Only were explicitly stated, Enantiomer baseline separation on Venusil Chiral OD-H
reactions were carried out in an inert atmosphere of 211 9 250 mm, 5 lm, 25 °C, k = 220 nm, 254 nm;
nitrogen (99.5% purity). For the synthesis of 17, mobile phases for gradient elution A = 0.2% DEA in Hex
Schlenck technique was applied for reaction and and B = 0.2% DEA in EtOH. HPLC-Method III: Agilent 1100
[
47]
reagent transfer at low temperature.
Cooling bath Series. Column: Waters Sunfire RP C18, 19 9 250 mm,
5 lm, 25 °C, k = 230 nm, flow 25 mL/min, mobile
temperature < ꢀ80 °C was achieved by adding liq N
2
to a MeOH bath. Microwave assisted synthesis was phases for gradient elution phase A = water and phase
performed using CEM Microwave Discover apparatus B = MeCN, gradient: 20 – 55% B (16 min), 55 – 95% B
ramping with 200 W to 110 °C. Automated flash (1 min), 95% (2 min), 95 – 5% B (1 min), 20% B (2 min),
chromatography: ISCO flash chromatography system injection volume: 1 mL, sample size: 2 9 3 mg in 1 mL
or Analogix Intelli Flash 280 with Supel flash cartridge, H2O/MeCN (1:1). HPLC-Method IV: Agilent. Column:
4
0 – 63 lm, active silica bed.
Machery Nagel Nucleodur Sphinx, 46 9 150 mm, 5 lm,
0 °C, k = 254 nm, flow 1.4 mL/min, mobile phases for
2
Tritiation. Halogen–tritium exchange reactions were gradient elution phase A = 0.05% aq. TFA and phase
performed on a 316L stainless steel, helium-leak B = MeCN, gradient: 55% B (5 min), 55 – 95% B
tested vacuum manifold custom made by RC Tritec (4.5 min), 95%. HPLC-Method V: Agilent. Column:
AG, CH-9053 Teufen. The getter technology used on Machery Nagel Nucleodur Sphinx, 8 9 150 mm, 5 lm,
this manifold is based on the thermal, reversible 20 °C, k = 254 nm, 280 nm; flow 4.0 mL/min, mobile
equilibrium of uranium(III) tritide, uranium and tritium phases for gradient elution phase A = 0.05% aq. TFA
238
238
[48][49]
according to 2
UT ⇋ 2
U + 3 T2.
The and phase B = MeCN, gradient: 30% B (0 min), 36.5% B
3
manifold contains a depleted uranium reservoir with (6.5 min), 95% B (7 – 9.5 min), 30% B (10 min). HPLC-
2
38
1
85TBq T per 15 g U (forward reaction), a depleted Method VI: Agilent 1200. Column Waters Sunfire RP C18,
2
uranium recycle flow reservoir (reverse reaction) and a 4.6 9 150 mm, 5 lm, 40 °C, k = 258 nm, flow: 1.0 mL/
waste ampule to recover contaminated solvents and min, mobile phases for gradient elution phase A = H O/
2
organic volatiles. Tritium gas generated in the forward MeCN/TFA 95:5:0.1 (v/v) and phase B = H O/MeCN/TFA
2
reaction is 99% pure with Aspec = 95.9 GBq/mL. 5:95:0.1, gradient: 10% B (2 min), 40% B (15 min), 95% B
Reactions on this manifold are performed at r.t. in (16 min), 95% B (20 min), 95% B(20.1 min), 10% B
2
mm Pyrex glassware (2 mL two-neck-round-bottom (25 min), sample: 0.2 – 0.4 mg/mL in MeCN/H O 1:1.
2
flasks and 25 mL lyophilisation ampule each with 3/8″ HPLC-Method VII: Agilent 1200. Column Diacel OD-H,
cylindrical fitting). Tight connection of glassware is 2.1 9 150 mm, 5 lm, 30 °C, k = 240 nm, flow: 1.0 mL/
assured by using a Keel-F ferrule system. The manifold min, mobile phase A = HEX/IPA/EtOH 95:2.5:2.5 (v/v/v)
system contains integrated reservoir heating system for and mobile phase B = IPA/EtOH 1:1, isocratic for
the reservoirs and pressure measurement systems in 45 min, A/B 95:5. For HPLC-methods VI and VII a
each compartment.
Berthold LB509 radio detector with a 100 lL Z-cell was
used in all cases. Scintillation cocktails either were Perkin
Hardware for Instrumental Analytics of Synthetic Elmer’s Flow-Scint A (3.0 mL/min) or Zinsser’s Quickzint
Compounds. For flash chromatography, in process TLC Flow 302 (2.8 mL/min). UPLC/MS Method. Electrospray
3
control, radio-detection of [ H]CDZ173, determination ionization positive mode. Column: Acquity HSS T3
of the specific activity by MS, open access LC/MS, and 1.8 lm, 2.1 9 50 mm at 60 °C; eluent A: water + 0.05%
service-MS of intermediates and final products, the formic acid + 3.75 mM NH OAc; eluent B: MeCN + 0.04%
4
same hard ware equipment was used as described in formic acid; gradient from 5 to 98% B in 1.4 min, flow
detail in the experimental part of a publication from 1.0 mL/min. In case of method VII, the limit of detection
[
50]
the isotope lab unit.
For GC-MS analytics of the bis- (LOD) was set to a signal to noise ratio (SN) of 3. All
halogenated pyridines 16 – 18, a Waters GTC GC/MS peaks ≥ LOD were detected.
accurate mass instrument was used.
NMR. Chemical shifts (d) are given in ppm referenced
HPLC Methods for Analytics, Preparative Chromatography, to the residual solvent peak. Multiplicities are abbreviated
and Release of Synthetic Compounds. HPLC-Method as follow: s = singulet, br. = broad, d = doublet,
I
(chiral analytical): Gilson. Enantiomer baseline dd = doublet of a doublet, dq = doublet of a quadruplet,
www.helv.wiley.com (22 of 33) e1800044 © 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
1
Table 5. H-NMR data for compound CDZ173 (1a)
1
[a]
1
Conformer I, d( H) [ppm]
Mixed form of conformer , d( H) [ppm]
4
4
H–C(2‴)
H–C(2″)
H–C(4‴)
NH
8.28 (d, J = 2.81)
8.35 (s)
H–C(2‴)
H–C(2″)
H–C(4‴)
NH
8.28 (d, J = 2.81)
8.34 (s) 4
4
7.87 (d, J = 3.00)
7.87 (d, J = 3.00)
3
3
6.76 (d, J = 6.43)
6.73 (d, J = 6.10)
0
2
3
0
2
3
H–C(3 )
4.70 (pseudo-sext., J = 6.57, J = 6.60)
H–C(3 )
4.58 (dt, J = 6.35, J = 6.38)
CH
2
(5″)
3.97 (s)
3.92 (s)
CH
2
(5″)
3.97 (s)
3.92 (s)
MeO–C(6‴)
MeO–C(6‴)
0
0
0
2
3
0
0
0
2
3
H –C(2 )
3.84 (dd, J = 10.42, J = 6.90)
3.55 (dd, J = 7.67, J = 5.54)
3.52 (dd, J = 6.78, J = 6.78)
3.37 (m)
3.33 (dd, J = 10.75, J = 4.81)
2.80 (m)
2.20 (m)
2.15 (dq, J = 12.64, J = 6.14)
1.92 (dq, J = 12.74, J = 7.45)
0.98 (m)
H –C(2 )
3.71 (dd, J = 11.99, J = 6.93)
3.60 (ddd, J = 10.25, J = 7.64, J = 5.89)
3.52 (dd, J = 6.78, J = 6.78)
3.50 (d, J = 7.96)
0
2
3
0
2
3
3
H –C(5 )
H –C(5 )
2
3
2
3
CH
H″–C(5 )
H″–C(2 )
CH
2
(7″)
CH
H″–C(5 )
H″–C(2 )
CH
2
(7″)
0
0
0
2
0
2
3
2
3
3.29 (dd, J = 12.06, J = 5.62)
2.80 (m)3
2
(8″)
2
(8″)
CH
2
(2)
CH
2
(2)
2.26 (q, J = 7.40)
0
0
2
3
0
0
2
3
H –C(4 )
H″–C(4)
Me(3)
H –C(4 )
H″–C(4)
Me(3)
2.22 (dt, J = 7.38, J = 3.90)
2.01 (dq, J = 13.88, J = 6.99)
0.98 (m)
2
3
2
3
[
b]
[b]
[
a]
Refers to a conformation where the propionly side chain is periplanar to the pyrroldine ring (see cross peaks, Figure 5 main
[b]
part).
q overlaid with dd.
dquint. = doublet of a quintuplet, exch. = exchangeable,
1
t = triplet, sext. = sextuplet. 400 MHz H-NMR spectra
were recorded on VNMRS-400 at 296 K. 600 MHz Spectra
on Bruker DRX600 with 1.7 mm-cryoprobe TCI-C/N probe
3
head. H-NMR spectra of compound 1b were acquired
on Bruker DPX400 with SXI-3H probe head using a
registered sample of the phosphate salt of 1a
+
21 26 3 6 2 2 4
ꢀ
(
[C H F N O ] [H PO ] , 548.85 g/mol) as external
13
reference for NMR-difference spectroscopy. C-NMR
Shifts were extracted from the HMBC and HSQC data set
as much as possible. Processing and assignment software
in all cases: XWin NMR and MNova v11.0.6
Figure 6. Atom numbering of compound 1a (anti-conformer is
shown).
(
MestreReNova). Qualitative conformational analysis of 1a
1
(CDZ173) was done based on the ROESY datasets. H-
1
3
n
and C-coupling constants ( J) are given in Hz indicating MeOH (300 mL) was added acetic acid methanimid-
the bond order n. ROESY spectra of 22 were smoothed amide (4.04 g, 38.8 mmol) in MeOH (10 mL) and
with the Savitzky–Golay method and denoised with the 0.5 M MeONa in MeOH (207 mL, 104 mmol). Subse-
Gaussian method. Numbering of protons and carbons in quently, the solution was stirred for 3 h at 90 °C (oil
the experimental part is in line with the IUPAC name of bath) while being monitored by UPLC/MS. Then
the compounds.
CH Cl (ca. 300 mL) and AcOH (5.9 mL, 6.22 g,
2 2
1
1
NMR of CDZ173 (1a). H-NMR ( H, COSY, TOCSY, 104 mmol) were added and the mixture was stirred
HSQC, HMBC, ROESY, (D )DMSO, 22 °C): See Table 5. over night at r.t. The reaction was quenched with
6
See also Scheme 4 and Figure 5 in main text to align water and extracted with CH Cl (1 9 500 mL) and
2
2
1
3
with given MNova atom numbering (Figure 6). C-NMR CH Cl /IPA (3:1, 2 9 500 mL). The organic layers
2
2
13
(
C, HSQC, HMBC, (D )DMSO, 25 °C): See Table 6.
were combined, dried over Na SO , and concen-
2 4
6
trated under vacuum. The residue was purified by
automated flash chromatography eluting with
CH Cl /1 – 10% MeOH gradient. This resulted in
Synthetic Procedures and Analytical Data
2
2
6
4
-Benzyl-5,6,7,8-tetrahydropyrido[4,3-d]pyrimidin- 5.2 g (83%) of 2 as acetic acid salt (> 95% UV on
[
11][51]
+
-ol (2).
To a solution of methyl 1-benzyl-4- UPLC/MS) UPLC/MS (ESI+): 242.1 (100, [M + H] ; calc.
+
oxopiperidine-3-carboxylate (6.4 g, 25.9 mmol) in 242.13), 243.3 (15, [M + H] ; calc. 243.13).
www.helv.wiley.com (23 of 33) e1800044 © 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
6
)DMSO, 25 °C)
13
13
Table 6. C-NMR ( C, HSQC, HMBC, (D
13
[a]
13
Conformer I, d( C) [ppm]
Mixed form of conformer , d( C) [ppm]
C(1)
171.21
141.20
158.04 or 158.00
158.51 or 158.41
155.43
153.55
137.90
125.55
123.20 (q, J(C,F) = 271.97)
111.06 (q, J(C,F) = 32.23)
109.64 or 109.62
53.65
50.91
C(1)
171.11
141.20
158.00 or 158.04
158.41 or 158.51
155.43
153.55
137.90
125.55
123.20 (q, J(C,F) = 271.97)
111.06 (q, J(C,F) = 32.23)
109.62 or 109.64
53.65
50.26
C(3‴)
C(4″)
C(8d″)
C(2″)
C(6‴)
C(2‴)
C(4‴)
C(3‴)
C(4″)
C(8d″)
C(2″)
C(6‴)
C(2″)
C(4‴)
[
[
b]
a]
1
1
CF
3
CF
3
2
2
C(5‴)
C(4a″)
C(5‴)
C(4a″)
[
a]
MeO–C(6‴)
MeO–C(6‴)
0
0
C(2 )
C(2 )
0
0
C(3 )
50.72
C(3 )
49.33
[
[
c]
C(5″)
C(7″)
45.81 or 45.79
45.79 or 45.81
43.65
C(5″)
C(7″)
45.81 or 45.79
45.79 or 45.81
44.24
b]
0
0
C(5 )
C(5 )
C(8″)
C(4 )
31.18
29.63
C(8″)
C(4 )
31.18
31.34
0
0
C(2)
C(3)
26.89
8.90
C(2)
C(2)
26.53
8.90
[
a]
Refers to a conformation where the propionly side chain is periplanar to the pyrroldine ring (see cross peaks, Figure 5 main
[b]
[c]
part).
Undistinguishable carbons of conformers. Undistinguisbale carbons.
6
-benzyl-4-chloro-5,6,7,8-tetrahydropyrido[4,3-d]- organic phases were dried over Na SO , filtered, and
2 4
pyrimidine (3). In an inert atmosphere, compound 2 concentrated under vacuum to yield 5 (2.4 g, > 95%
+
(
acetic acid salt, 4 g, 1.63 mmol) was dissolved in UV). UPLC/MS (ESI+): 265.2 (15, [M + Na] ; calc.
t
+
toluene (50 mL) and Et N (4.62 mL, 33.2 mmol) was 265.15), 187.1 (100, [(M – Bu) + H] ; calc. 187.11).
3
added. POCl (2.32 mL, 24.9 mmol) was added dropwise. 1-[(3S)-3-Aminopyrrolidin-1-yl]propan-1-one
(6).
3
The mixture was stirred at 100 °C (oil bath), while the Compound 5 (2.4 g, 9.90 mmol) was dissolved in
reaction was monitored by UPLC/MS. After 2 h, the mix- CH Cl (15 mL). Then, 4 N HCl in dioxane (15 mL,
2
2
ture was cooled to r.t., the reaction was quenched with 60.0 mmol) were added, and the solution was stirred
NaHCO (20 g), the resulting mixture was stirred for 1 h, for 2 h when UPLC/MS indicated completion. The mix-
3
filtered, and evaporated. The residue was dissolved in ture was concentrated, dissolved in MeCN (20 mL),
water (200 mL), pH was adjusted with sat. NaHCO to and solid K CO (10 g) was added. The mixture was
3
2
3
pH 8 and extracted with CH Cl (3 9 200 mL). The com- stirred overnight, then filtered, and the filtrate was
2
2
bined organic layers where dried over Na SO , filtered, evaporated to yield 6 (1.41 g, quant.) UPLC/MS (ESI+):
2
4
1
+
and evaporated to dryness to yield 3 (2.5 g, 58%) H- 165.2 ([M + Na] ; calc. 165.10).
NMR ((D )DMSO): 8.76 (s, 1 H); 7.34 – 7.25 (m, 5 H); 3.73 1-{(3S)-3-[(6-Benzyl-5,6,7,8-tetrahydropyrido[4,3-d]-
6
(
s, 2 H); 3.54 (s, 2 H); 2.89 (m, 2 H); 2.77 (m, 2 H). UPLC/ pyrimidin-4-yl)amino]pyrrolidin-1-yl}propan-1-one (7).
+
MS (method 1): (100, [M + H] ; calc. 260.10), 262.1 (31, In a microwave vial, compound 3 (480 mg, 185 lmol),
+
+
[M + H] ; calc. 262.10), 263.2 (5, [M + H] ; calc. 263.10).
(S)-6 (263 mg, 185 lmol), and cat. AcOH (31.7 lL,
(
S)-tert-Butyl (1-Propanoylpyrrolidin-3-yl)carbamate 0.55 lmol) were dissolved in EtOH (10 mL). The mixture
[
52]
(
5). (S)-tert-butyl pyrrolidin-3-ylcarbamate
(4; 2 g, was ramped for 20 min at 200 W to 110 °C, then
1
0.7 mmol) was dissolved in THF (20 mL). Et N slightly concentrated in the vacuum, and heating was
3
(
1
1.96 mL, 14.0 mmol) and propanoyl chloride (1.1 g, continued in an oil bath at 140 °C for 18 h. When
1.8 mmol) were added. The resulting solution was UPLC/MS control indicated end of reaction, the solvent
stirred for 2 h at r.t. when UPLC/MS indicated comple- was evaporated, the resulting mixture dissolved in
tion of the reaction, and the mixture was quenched CH Cl (10 mL) and injected into the automated flash
2
2
with sat. aq. NaHCO3 (20 mL). The mixture was chromatography system (CH Cl /0 – 10% MeOH gradi-
2
2
extracted with AcOEt (3 9 50 mL), the combined ent). 260 mg of 7 were obtained (38%, ca. 80% UV on
www.helv.wiley.com (24 of 33) e1800044 © 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
UPLC/MS). In a scale up batch, 3 (14.4 g, 55.44 mmol) enantiomer undetectable with HPLC method 1)). For
and (S)-6 (10.2 g, 71.7 mmol, 1.30 equiv.) were placed assignment of the absolute configuration of the (S)-iso-
into a 250-mL round-bottom flask and dissolved in EtOH mer, (R)-9 was synthesized independently from (R)-4 as
(
200 mL). Then, AcOH (0.953 lL, 16.6 mmol, 0.30 equiv.) described for (S)-6 from (S)-4. With HPLC-method I, a
was added, and the solution was stirred overnight at 1:1 mixture of (S)-9 and (R)-9 gave baseline separation
40 °C (oil bath). The solution was cooled to 25 °C, of the isomers with t ((R)-9) = 9.263 and t ((S)-
1
R
R
diluted with CH Cl (100 mL) and concentrated under 9) = 13.908 (for the mixture (R)/(S) = 49.57:50.43). MS
2
2
+
+
vacuum. The residue was applied on a silica gel column (ESI ): 376 ([M + H] ; calc. 376.23).
and eluted with a CH Cl /MeOH gradient (1 – 15% 1-[(3S)-3-(5,6,7,8-Tetrahydropyrido[4,3-d]pyrimidin-
2
2
MeOH). 17 g of 7 (84%) were obtained. UPLC/MS (ESI+): 4-ylamino)pyrrolidin-1-yl]propan-1-one ((S)-8, ee
+
+
3
66.2 (100, [M + H] ; calc. 366.23), 367.2 (10, [M + H] ; > 99%). Compound (S)-9 (6.0 g, ee = 100%,
calc. 267.23). 15.98 mmol) was placed into a 500 mL 3-necked
1
4
-[(3S)-3-(5,6,7,8-Tetrahydropyrido[4,3-d]pyrimidin- round-bottom flask and dissolved in Et O (500 mL).
-ylamino)pyrrolidin-1-yl]propan-1-one (8). Com- Into this solution, dry HCl gas was introduced, and
2
pound 7 (260 mg, 0.71 mmol) was dissolved in EtOH the mixture was stirred for 2 h at r.t. The solids were
10 mL), and Pd(OH ) on carbon (110 mg) was added. collected by filtration, and the filter cake was washed
(
2
The system was stirred in a mixed atmosphere of with Et O (3 9 50 mL). The crude product was re-
2
hydrogen and air for 18 h at 60 °C. The catalyst was crystallized from 10% MeOH in Et O (100 mL). The
2
removed by filtration, the filtrate was evaporated, and collected material was dissolved in water (50 mL), the
the residue was purified by automated flash chro- solution was neutralized with 1 M NaOH (2 equiv.)
matography (CH Cl /0 – 10% MeOH gradient). 78 mg and lyophilized. Subsequently, the product was dis-
2
2
of compound 8 (40%, ca. 90% UV on UPLC/MS) were solved in THF/1% MeOH (500 mL), and the precipitate
obtained. In a scale up batch, 7 (17 g, 46.5 mmol, of NaCl was removed by filtration. The filtrate was
1
.00 equiv.) was placed into a 250-mL round-bottom concentrated under vacuum and the product dried in
flask and dissolved in EtOH (150 mL). Pd(OH) on car- the high vacuum to yield 2.97 g (64%) of (S)-8. Since
2
bon (6 g) was added, the flask was closed with a rub- non-racemizing conditions were used for deprotec-
ber stopper, the suspension was frozen in liquid N2, tion, the optical purity of basic (S)-8 (ee > 99%) was
and the flask was evacuated in the high vacuum. deduced from the optical purity of the less basic (S)-9
1
Hydrogen gas was expanded into the flask using the (ee > 99%).
H-NMR, COSY, HSQC ((D )DMSO,
6
balloon technique for gas transfer, the suspension T = 296 K): For conformer A, see Table 7, for con-
was thawed up and stirred overnight at 60 °C (oil former B, see Table 8. Atom numbering is depicted in
1
bath). Subsequently, the solids were filtered and the Figure 7. H-NMR ((D )DMSO, T = 373 K): For a single
6
+
resulting mixture was concentrated under vacuum conformer, see Table 9. HighRes-MS (ESI ): 276.18175
giving 11 g of compound 7. UPLC/MS (ESI+, m/z): cal- (100, [M + H] ; calc. 276.1824), 277.18487 (16,
culated for C H N O 276.2 (100, [M + H] ; calc. [M + H] ; calc. 277.1858).
+
+
+
1
4 21 5
+
2
76.18), 277.1 (13, [M + H] ; calc. 277.19).
5-Bromo-4-iodo-2-methoxy-3-(trifluoromethyl)pyri-
tert-Butyl 4-{[(3S)-1-Propanoylpyrrolidin-3-yl]amino}- dine (16). Under N , LDA (4.75 mL, 9.5 mmol,
2
7
,8-dihydropyrido[4,3-d]pyrimidine-6(5H)-carboxy- c = 2.0 M in THF/HEP/EtPh) was added to THF
late ((S)-9, ee = 100%). Compound 8 (11 g, 40 mmol) (4.75 mL) kept at –85 to ꢀ95 °C. Then, a solution of
was placed in a 500 mL-round-bottom flask and dis- 5-bromo-2-methoxy-3-(trifluoromethyl)pyridine (9.5 mmol,
solved in CH Cl (100 mL). Et N (12 g, 118.6 mmol) 2.43 g) in THF (5.5 mL) was added dropwise. Finally, a
2
2
3
was added. Then di-tert-butyl dicarbonate (12.4 g, catalytic amount of anhydrous LiBr (0.95 mmol,
6.8 mmol) was added at 0 °C and the solution was 82.5 mg) was added, and the mixture was kept at
5
stirred for 3 h at r.t. The reaction was then quenched ꢀ85 to ꢀ95 °C for 180 min. Then, a solution of I
2
with water (100 mL) and the suspension was extracted (1.35 mmol, 3.43 g) pre-cooled to ꢀ85 to ꢀ95 °C was
with CH Cl2 (3 9 200 mL). The organic layers were added in small portions to the lithiated species using
2
combined, dried over Na SO , and concentrated under a Ar-filled graduated pipette. The mixture was kept for
2
4
vacuum. The residue was applied onto a silica gel col- further 60 min at ꢀ85 to ꢀ95 °C and then warmed
umn and eluted with CH Cl /MeOH gradient (1 ? up to r.t. over night. After this time, the reaction was
2
2
1
5%) to get 10 g of 9 (ee = 95.1% with HPLC-method quenched with 1% aq. Na S O (20 mL) and extracted
2 2 3
I). This product (10 g) was purified by chiral semi-prep. with CH Cl (3 9 50 mL). The combined organic
2
2
HPLC using HPLC-method II to get 7.9 g of (S)-9 (ee phases were washed with 5% aq. Na S O until they
2
2 3
<
99% with HPLC-method I; (S)/(R) = 100:0 ((R)- were colorless, then dried over MgSO4 and
www.helv.wiley.com (25 of 33) e1800044 © 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
1
13
1
Table 7. H- and C-NMR data for conformer A of (S)-8, at
Table 9. H-NMR data of one single conformer ((D )DMSO) at
6
296 K
T = 373 K
1
13
1
d( H) [ppm]
d( C) [ppm]
d( H) [ppm]
H–C(2″)
8.28 (s)
6.57 (d, J = 12.0)
4.65 (pseudo-sext., J = 6.3)
3.77 (dd, J = 9.9, J = 7.0)
3.60 (br. s)
3.45 (m)
3.35 (m)
3.32 (m)
3.01 (m)
2.58 (m)
2.24 (q, J = 7.5)
2.07 (m)
1.90 (m)
C(2″)
156.2
–
H–C(2″)
8.25
6.21 (s, large)
4.61 (m)
3.8 – 3.1 (m)
3.65 (s)
3.8 – 3.1 (m)
3.8 – 3.1 (m)
3.8 – 3.1 (m)
3.03 (t, J = 6.0)
2.60 (t, J = 6.0)
2.22 (m)
2.1 – 1.8 (m)
2.1 – 1.8 (m)
3
NH
–
C(3 )
C(2 )
NH
0
0
3
0
0
H
a
0
–C(3 )
51.2
51.4
42.1
44.0
–
51.4
42.5
31.2
27.0
29.9
–
0
0
H –C(3 )
a
2
3
H –C(2 )
CH (5″)
0
H –C(2 )
2
C(5″)
CH
H –C(5 )
H″–C(5 )
H″–C(2 )
(5″)
2
0
0
0
0
0
H –C(5 )
C(5 )
0
0
H″–C(5 )
–
C(2 )
0
0
0
H″–C(2 )
CH (7″)
CH (8″)
2
2
C(7″)
C(8″)
C(2)
CH (7″)
2
2
2
CH (8″)
2
3
CH
2
(2)
CH
2
(2)
0
0
0
H –C(4 )
H″–C(4)
Me(3)
C(4 )
0
0
0
H –C(4 )
–
C(3)
0
H –C(4 )
3
3
0.99 (t, J = 6.7)
9.3
Me(3)
1.01 (t, J = 6.9)
1
13
Table 8. H- and C-NMR data for conformer B of (S)-8, at
an aliquot of this material was purified by chromatogra-
phy. In a second experiment, a solution of 5-bromo-2-
methoxy-3-(trifluoromethyl)pyridine (1.5 g, 5.85 mmol,
296 K
1
13
d( H) [ppm]
d( C) [ppm]
H–C(2″)
8.28 (s)
6.54 (d, J = 11.5)
4.50 (pseudo-sext., J = 6.0)
3.64 (m)
3.60 (br. s)
3.55 (m)
C(2″)
156.2 1 equiv.) in THF (11 mL) was cooled to ꢀ75 °C in a N -
2
3
NH
–
C(3 )
C(2 )
–
atmosphere. Then LDA (7.3 mL, 14.6 mmol, 2.5 equiv.)
0
0
3
0
0
H
a
0
–C(3 )
49.6
50.7
42.1
44.5
–
was added and the solution was stirred at ꢀ75 °C for
H –C(2 )
4
5 min. Subsequently, the mixture was poured into a
CH (5″)
C(5″)
2
solution of I2 (5.95 g, 23.4 mmol, 4 equiv.) in THF
(11 mL) cooled to 0 °C. Stirring was kept 5 min at 0 °C.
0
0
0
0
0
H –C(5 )
H″–C(5 )
H″–C(2 )
C(5 )
3.49 (m)
3.27 (m)
–
C(2 )
0
50.7 Then 1% aq. Na S O was added, and the mixture was
2 2 3
CH
CH
2
(7″)
(8″)
3.01 (m)
2.58 (m)
2.20 (m)
2.16 (m)
C(7″)
C(8″)
C(2)
C(4 )
–
42.5 extracted with Et O (2 9 100 mL). The combined
2
2
31.2
organic layers were dried over Na SO and evaporated
2
4
CH (2)
27.0
31.6
–
2
to dryness. The residue was purified on silica gel (elu-
ent: CYH) and the product was further purified by prep.
HPLC on Sunfire C18 (eluent: H O/0.1% HCO H in MeCN
0
0
0
H –C(4 )
H″–C(4)
Me(3)
2.00 (m)
3
2
2
0.97 (t, J = 6.8)
C(3)
9.3
6
:4 ? 4:6). The fractions containing compound 16 were
extracted with Et O to give after drying (Na SO ) and
2
2
4
evaporation 16 (1 g, 45%, UV-UPLC > 99%). The
evaporated. GC/MS of the crude material confirmed
the presence of one single isomer in 60% yield (see
Supplementary Material 1.2.1). For analytical purposes,
3
configuration of 16 was confirmed by X-ray analysis
1
(
see Supplementary Material 1.2.2). H-NMR ((D )DMSO):
6
13
8
1
3
.58 (s, 1 H); 3.94 (s, 3 H). C-NMR ((D )DMSO): 159.4;
49.3; 123.1; 120.7; 118.9; 116.81 (m); 54.0. GC/MS (ESI ):
80.8 (100, M ; calc. 380.85), 381.8 (40, M ; calc.
6
+
+
+
3
Crystallographic data (excluding structure factors) for
1
6 – 18 have been deposited with the Cambridge Crys-
tallographic Data Centre as supplementary publication
number CCDC 1844455, CCDC 1844456 and CCDC
1
844457. These data can be obtained free of charge via
http://www.ccdc.cam.ac.uk/data_request/cif, by emailing
data_request@ccdc.cam.ac.uk, or by contacting The Cam-
bridge Crystallographic Data Centre, 12 Union Road, Cam-
bridge CB2 1EZ, UK; fax: 441223 336033.
Figure 7. Atom numbering of compound (S)-8.
www.helv.wiley.com
(26 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
+
+
3
3
81.85), 382.8 (90, M ; calc. 382.85), 383.8 (3, M ; calc. by HPLC. When in process control indicated that the
83.85).
maximum amount of 14 was reached, the mixture
4
,5-Dibromo-2-methoxy-3-(trifluoromethyl)pyridine was diluted with brine (10 mL) and extracted with
(
(
17). Under N , a solution of 5-bromo-2-methoxy-3- AcOEt (2 9 20 mL). The combined organic layers were
2
trifluoromethyl)pyridine (1.0 g, 3.9 mmol, 1 equiv.) in dried over Na SO , filtered, and evaporated to dryness.
2
4
THF (10 mL) was cooled to ꢀ75 °C and LDA (2 M in The residue was adsorbed on silica and added on top
Hex/ETPH, 3.9 mL, 7.8 mmol, 2 equiv.) was added, of a 10 g SiOH chromabond column. The product was
while controlling the temperature carefully during eluted with a gradient of CH Cl (100%) ? CH Cl /
2
2
2
2
addition of the base. The solution was stirred at MeOH/NH OH 98:2:1. The product fractions were fur-
4
ꢀ
75 °C for 1 h, then added to a solution of C Br Cl
2.54 g, 7.8 mmol, 2 equiv.) in THF (10 mL) cooled to lyophilization of the product fractions, 6 mg of 15
°C. The solution was stirred for 3 min at 0 °C. Then, (UV-HPLC > 90%) were obtained. H-NMR, COSY,
ther purified by HPLC using HPLC-method III. After
2
2
4
(
1
0
sat. aq. NH Cl (50 mL) was added. The mixture was ROESY, HSQC, HMBC ((D )DMSO, T = 296 K): see
4
6
1
3
extracted with AcOEt (3 9 100 mL), the combined Table 10. For atom numbering, see Figure 8. C-NMR:
organic layers were washed with brine, dried over see Table 11. MS (ESI+): 485.0 (100, [M + H] ; calc.
Na SO , and evaporated to dryness. The residue was 485.17), 487.0 (34, [M + H] ; calc. 487.17), 486.0 (26,
purified on silica gel (eluent: CYH) to give 0.8 g of 17 [M + H] ; calc. 486.17), 488.0 (10, [M + H] ; calc.
60%, UV-HPLC 95%). The configuration of pure 17 488.17).
+
+
2
4
+
+
(
4
was confirmed by X-ray analysis (see Supplementary 1-[(3S)-3-({5,6,7,8-Tetrahydro-6-[6-methoxy-5-(tri-
1
3
Material 1.2.2). H-NMR ((D )DMSO): 8.70 (s, 1 H); 3.97 fluoromethyl)(4- H)pyridin-3-yl]pyrido[4,3-d]pyrim-
6
1
3
(
s, 3 H). C-NMR ((D )DMSO): 160.1; 151.7; 135.4; idin-4-yl}amino)pyrrolidin-1-yl]propan-1-one (1b).
6
+
+
1
3
3
17.2; 55.2. GC/MS (ESI ): 334.8 (100, M ; calc. 334.86), In a two-neck tritiation flask, Pd on C (10%, BASF/
+
+
32.8 (49, M ; calc. 332.86), 336.8 (35, M ; calc. Engelhard 4505) and 22 (2.23 mg, 4.6 lmol) were sus-
36.86). pended in DMF (800 lL), the flask was attached
-Bromo-4-chloro-2-methoxy-3-(trifluoromethyl)pyr- tightly to the tritium manifold, DIPEA (4.4 lL, 3.3 mg,
idine (18). Under N , a solution of 5-bromo-2-meth- 25.9 lmol) was added via the septum neck, and the
5
2
oxy-3-(trifluoromethyl)pyridine (1.0 g, 3.9 mmol,
equiv.) in THF (10 mL) was cooled to ꢀ75 °C and LDA flask was degassed 39 in a freeze-thaw cycle. In the
2 M in HEX/ETPH, 3.9 mL, 7.8 mmol, 2 equiv.) was meanwhile, tritium gas was released at 775 K from
1
flask was frozen in liquid nitrogen. Subsequently, the
(
added, while controlling the temperature carefully the depleted uranium reservoir into the gas delivery
during addition of the base. The solution was stirred arm (V = 4.5 mL) to give p = 1415 mbar (263 lmol)
at ꢀ75 °C for 1 h, then added to a solution of C Cl
of tritium gas. The tritium gas was expanded at 291 K
1.84 g, 7.8 mmol, 2 equiv.) in THF (10 mL) cooled to into the evacuated, frozen tritiation flask. After closing
2
6
(
0
°C. The solution was stirred for 3 min at 0 °C. Then, the reaction system at r.t., p(291 K) = 353 mbar was
sat. aq. NH Cl (50 mL) was added. The mixture was observed in the volume above the frozen mixture cor-
4
extracted with AcOEt (3 9 100 mL), the combined responding to 202 lmol T . Once the mixture was
2
organic layers were washed with brine, dried over thawed up and reached 291K, p(t = 0) = 603 mbar was
Na SO , and evaporated to dryness. The residue was observed in the volume above the mixture. After
2
4
purified on silica gel (eluent: CYH) to give 0.55 g of 18 12 min, the exothermic system equilibrated at
48%, UV-HPLC ca. 95%). The configuration of pure 18 T = 296 K and p(t = 12 min) = 597 mbar. After 150 min
(
4
was confirmed by X-ray analysis (see Supplementary at 296 K, gas up take stopped at p(t = 150 min)
Material 1.2.2). H-NMR (CDCl ): 8.47 (s, 1 H); 4.03 (s, 3 591 mbar. The mixture was frozen until constant
H). GC/MS (ESI ): 291.0 (100, M ; calc. 290.91), 289.0 p = 334 mbar was reached in the volume above the
=
1
3
+
+
+
+
(
1
28, M ; calc. 288.91), 291.9 (5, M ; calc. 291.91).
-[(3S)-3-({6-[4-Chloro-6-methoxy-5-(trifluoromethyl)-
frozen mixture corresponding to 10.9 lmol T con-
sumption (Dp291 K = 19 mbar). The residual amount
2
pyridin-3-yl]-5,6,7,8-tetrahydropyrido[4,3-d]pyrimidin- of tritium gas was pumped off the system into the
4
-yl}amino)pyrrolidin-1-yl]propan-1-one (14, ee recycling flow reservoir kept between 480 – 570 K,
>
99%). In an inert atmosphere, 18 (116 mg, 0.40 the reaction mixture was frozen in liquid N , the sys-
2
t
mmol), (S)-8 (ee > 99%, 121 mg, 0.44 mmol), NaO Bu tem was evacuated and the organic volatiles were
(
57.6 mg, 0.6 mmol), Pd dba (18.3 mg, 20 lmol), and lyophilized off. Volatile radioactivity in the vacuum
2 3
t
BuMePhos (7.48 mg, 20 lmol) are stirred in anhydr. dried crude product was expelled by suspending in
BuOH (4.0 mL) at r.t. Then, the mixture was stirred at MeOH (3 9 750 lL) and lyophilizing the carrier sol-
t
9
5 °C. In between 3 – 16 h, the reaction was followed vent into the waste ampule. The reaction flask was
www.helv.wiley.com (27 of 33) e1800044 © 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
1
Table 10. H-NMR, COSY, ROESY, HSQC, HMBC ((D
6
)DMSO, T = 296 K) of conformers A and B of compound 14
1
1
Conformer A, d( H) [ppm]
Conformer B, d( H) [ppm]
H–C(2‴)
H–C(2″)
8.48 (s)
8.37 (s)
6.66 (d, J = 6.2)
H–C(2‴)
H–C(2″)
8.47 (s)
8.36 (s) 3
3
NH
NH
6.62 (d, J = 5.8)
0
3
0
3
H
a
–C(3 )
4.68 (pseudo-sext., J = 6.4)
H
a
–C(3 )
4.55 (pseudo-sext., J = 6.0)
MeO–C(6‴)
3.96 (s)
3.91 (s)
3.80 (dd, J = 10.5, J = 6.8)
3.42 (m)
3.41 (m)
3.31 (m)
MeO–C(6‴)
3.96 (s)
3.91 (s)
3.65 (dd, J = 12.0, J = 6.8)
3.56 (m)
3.50 (m)
3.30 (m)
CH
2
(5″)
CH
2
(5″)
0
0
2
3
0
0
2
3
H –C(2 )
H –C(2 )
0
0
0
0
0
0
0
0
H –C(5 )
H –C(5 )
H″–C(5 )
H″–C(5 )
0
H″–C(2 )
H –C(2 )
CH (7″)
3.32 (m)
CH (7″)
3.32 (m)
2
2
CH
2
(8″)
2.82 (t, not resolved)
2.25 (m)
CH
2
(8″)
2.82 (t, not resolved)
2.18 (m)
CH
2
(2)
CH
2
(2)
0
0
0
0
H –C(4 )
H″–C(4)
Me(3)
2.11 (m)
H –C(4 )
H″–C(4)
Me(3)
2.21 (m)
3
2
2
3
1.88 (dq, J = 13.6, J = 6.9)
1.98 (dquint., J = 13.1, J = 6.6)
3
3
0.99 (t, J = 7.4)
0.97 (t, J = 7.9)
released from the manifold, the crude product was
suspended in EtOH (ca. 5 mL in several portions), the
suspension was filtered over a 0.2 lm Whatman–Ano-
top filter membrane into a graduated volumetric flask
(200 mL) and toped up with MeOH to yield a stock
solution of 1b (1.295 MBq, 52% rad. chem. purity and
product/educt ratio 1:2 (UVk = 254 nm) on HPLC-
method IV). From this stock solution, a first aliquot
(200 lL) was purified by HPLC-method V to obtain
4
48 MBq tritiated 1b in 10.84 mL eluent. This solution
Figure 8. Atom numbering of compound 14.
of 1b was speed-vaced down to ca. 6 mL and neutral-
ized with sat. aq. NaHCO . The compound was
)DMSO, T = 296 K) of conformers A and desalted on a 3 mL-Strata X column (100 mg sorbent)
3
1
Table 11. C-NMR ((D
B of compound 14
6
by washing with H O (2 9 2 mL) and eluting with
2
13
13
EtOH to yield 306 lg of 1b (426 MBq, 58% rad. chem.
yield) in ethanol solution (c = 35.6 lg/mL, SA =
171.51 626.6 GBq/mmol, SA = 49.52 MBq/mL, HPLC-UV 100%
Conformer A, d( C) [ppm]
Conformer B, d( C) [ppm]
C(1)
171.51
140.97
158.20
157.60
155.50
144.70
141.00
n.d.
C(1)
C(3‴)
C(4″)
C(6‴)
C(2″)
C(2‴)
C(4‴)
C(5‴)
C(4a″)
CF3
C(3‴)
C(4″)
C(6‴)
C(2″)
C(2‴)
C(4‴)
C(5‴)
C(4a″)
CF3
140.97
(HPLC-method VI), HPLC-RA 96.8% (HPLC-method VI),
158.20
157.60
155.50
144.70
chiral purity: 99.6%, ee = 99.2%, HPLC-method VII).
The compound identity was confirmed by HPLC-con-
injection with a registered reference sample of the
+
ꢀ
4
141.00 phosphate salt of 1a ([C H F N O ] [H PO ] ,
21
26 3
6
2
2
n.d.
110.47
n.d.
548.45 g/mol) having chemo-physical characteristics as
[
2] 3
110.47
n.d.
described in.
(
H-NMR ((D )DMSO): 7.87 (s, 1 T). MS
6
+
+
1a; APCI ): 450.4 (0.90), 451.2 (100, [M + H] , calc.
51.21), 452.2 (22.17, [M + H] , calc. 452.21), 453.2 (3.19,
[M + H] , calc. 453.21). MS (1b; APCI ): 450.4 (0.77),
51.3 (68.29), 452.4 (16.70), 453.2 (100, [[ H]M + H] ,
MeO–C(6‴)
C(5″)
54.98
48.73
51.47
50.95
44.16
49.03
32.19
30.05
27.05
MeO–C(6‴)
C(5″)
54.98
48.73
50.92
49.64
+
4
+
+
0
0
C(2 )
C(2 )
3
+
0
0
4
C(3 )
C(3 )
0
0
3
+
C(5 )
C(5 )
44.68 calc. 453.22), 454.2 (22.81, [[ H]M + H] , calc. 454.22),
3
+
C(7″)
C(8″)
C(7″)
C(8″)
C(4 )
49.03
32.19
31.87
26.97
4
55.2 (4.63, [[ H]M + H] , calc. 455.22), 456.2 (0.59).
0
0
C(4 )
Dilution and Phosphate Salt Formation of 1b. For
biotransformation studies, aliquots of stock
C(2)
C(2)
www.helv.wiley.com
(28 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
solutions of pure 1b were diluted with a stock (266 lmol, 120 g, free base) in MeCN/H O (22.2 mM,
2
solution of CDZ173-phosphate (2.474 lmol, 1.356 mg, 12 mL, v/v 50:50) was prepared and aliquots of 1 mL
c(EtOH) = 1.50 mg/mL) in a ratio 1:50. After co- (26.6 lmol/flask, 12.0 mg/flask) were added to each
evaporation and solubilization in MeCN/water 7:3, a flask. Shaking was continued for additional 2 days at
3
5
mM solution of [ H]CDZ173 phosphate salt was 220 rpm/28 °C when NaCl (42 g/flask) was added.
obtained (C H F N O •H PO (unlabeled), molecular Shaking was continued for additional 0.5 – 1 h. The
21
25 3
6
2
3
4
weight (unlabeled) 548.5 g/mol, specific radioactivity culture solutions were combined and extracted with
8.10 MBq/mg (free base) and 23.07 MBq/mg (as AcOEt (3 9 650 mL). The combined organic phases
2
phosphate salt), 12.65 GBq/mmol (i.e. ca. 1% T- were dried over MgSO , filtered, and evaporated. The
4
labelled); HPLC-RA: 99.3%, HPLC-UV: 99.1% both on crude material was pre-purified on a custom made
HPLC-method VI). Remarks: A 4.3 lg sample of this RP18 stainless steel column (200 9 50 mm, stationary
material can be handelled in an uncontrolled lab. Due phase: LiChroprep RP18 (Merck KGaA 1.13900)) using a
to the high dilution factor, the increase of the flat gradient of A (H O/0.05% TFA) and B (MeCN/
2
+
characteristic m/z 453.2 in 1b was 18%. LC/MS (APCI ) 0.05% TFA); flow: 100 mL/min; 25 °C; k = 264 nm;
found for 1a with [(C H F N O )H] m/z 451.2 fraction size 60 mL. The product fractions were com-
+
2
1 25 3 6 2
(
100%), 452. (22.1%), 453.1 (3.10%), found for 1b bined and concentrated under reduced pressure to
with [(C H F N O )H]+ m/z 451.2 (100%), 452.2 V ꢄ 250 mL. A second prep. HPLC purification was
21 25 3 6 2
(
1
22.9%), 453.2 (3.79%).
-[(3S)-3-({5,6,7,8-Tetrahydro-6-[6-hydroxy-5-(trifluoro- gradient and detection conditions as described above
methyl)pyridin-3-yl]pyrido[4,3-d]pyrimidin-4-yl}- (flow: 50 mL/min). The product fractions were com-
amino)pyrrolidin-1-yl]propan-1-one (19).
1a bined, concentrated to V ꢄ 150 mL and lyophilized to
111 lmol, 50 mg, free base) was dissolved in media yield 20 (169 lmol, 79 mg, 37% absolute, > 96%
performed on Sunfire RP18 (30 9 100 mm) using the
(
1
NL148S (100 mL). Then, a culture of streptomyces HPLC/DAD). H-NMR ((D )DMSO, 303 K, 1:1 mixture of
6
platensis (1 mL) was added, and the mixture was sha- conformers of A and B): 8.92 (s); 8.91 (s); 8.30 (s); 7.95
ken at 220 rpm/28 °C for 3 days. At this point, sub- (s); 6.08 (m); 4.73 (m, conformer A); 4.63 (m, conformer
strate conversion was 32% (LC/MS). Subsequently, B); 4.07 (s); 3.73 (m); 3.56 (m); 3.54 (m); 3.29 (m); 2.98
0
NaCl (2 g) were added, and the mixture was extracted (m); 2.26 (m); 2.19 (m, conformer A); 2.08 (m, H , con-
with Et O (3 9 100 mL). The organic phases were former B); 1.99 (m, H″, conformer B); 1.00 (m). UPLC/
2
+
combined, dried over MgSO , filtered, evaporated MS (ESI+): 451.3 (18, [M – O + H] ; calc. 451.21), 467.2
under reduced pressure, and the crude material was (100, [M + H] ; calc. 467.20), 468.3 (40, [M + H] ; calc.
dissolved in MeCN (10 mL). This solution was diluted 468.21), 933.4 (5, [2M + H] ; calc. 933.40).
4
+
+
+
with HPLC-eluent (H O/MeCN/formic acid 95:5:0.05) 1-[(3S)-3-({5,6-Dihydro-6-[6-methoxy-5-(trifluoro-
2
and applied onto a Nucleodur RP18 (250 9 25 mm). methyl)pyridin-3-yl]pyrido[4,3-d]pyrimidin-4-yl}amino)-
Compound 19 eluted using a gradient H O/MeCN/ pyrrolidin-1-yl]propan-1-one (21) and 1-[(3S)-3-
2
formic acid 95:5:0.05
? H2O/MeCN/formic acid ({5,6,7,8-Tetrahydro-5-(5R,S)-hydroxy-6-[6-methoxy-
7
5:25:0.05. The fractions of 19 were combined, con- 5-(trifluoromethyl)pyridin-3-yl]pyrido[4,3-d]pyrimidin-
centrated under reduced pressure, and lyophilized to 4-yl}amino)pyrrolidin-1-yl]propan-1-one (22). A solu-
yield pure 19 (12 mg, 27 lmol, 24% absolute; tion of 1a (110 lmol, 50 mg) in MeCN (10 mL) was
1
7
6% rel. to conversion, 94% UPLC/DAD). H-NMR added to a solution of TEMPO (224 lmol, 35 mg) in
(
(D )DMSO, 303 K, 1:1 mixture of conformers A and B): McIlvaine buffer (500 mL, pH 4.5). Subsequently, laccase
6
8
.34 (s); 8.33 (s); 8.04 (br. s); 7.35 (s, exch.); 6.82 – 6.62 from Trametes versicolor (800 mg, > 10 U/mg, SIGMA
(
(
0
[
4
m, exch.); 4.83 – 4.50 (m, conformers A and B); 3.73 51639) was added and the suspension was incubated at
s); 3.63 – 3.34 (m); 2.81 – 2.75 (m); 2.34 – 2.11 (m); 30 °C/160 rpm for 15 min. Afterwards NaCl (50 g) and
3
.99 (q, J = 7.2). UPLC/MS (ESI+): 437.2+ (100, NaOH (10 M, 5 mL) were added, the suspension was fil-
+
M + H] , calc. 437.19), 438.2 (24, [M + H] , calc. tered and the filtrate was extracted with AcOEt
38.19). (2 9 1 L). The combined organic phases were com-
1
-[(3S)-3-({5,6,7,8-Tetrahydro-6-[6-methoxy-5-(trifluoro- bined, dried over MgSO , filtered, and evaporated. The
4
methyl)pyridin-3-yl]-1-oxidopyrido[4,3-d]pyrimidin-
crude material was dissolved in MeCN/H O (60 mL, 5:1,
2
4
-yl}amino)pyrrolidin-1-yl]propan-1-one (20). Twelve v/v), diluted with eluent A and applied to Sunfire RP18
shake flasks were filled with NL148S (200 mL each). To (30 9 100 mm). Subsequent chromatographic separa-
each flask, preculture of Streptomyces griseus (5 mL) tion of 21 and diastereomeric 22 using a gradient of A
was added and the flasks were shaken at 220 rom/ (H O/0.05% FA) and B (100% MeCN); flow: 100 mL/min;
2
28 °C for 2 days. After this time, a solution of CDZ173 25 °C; k = 254 nm gave 21 (4.4 lmol, 4.5 mg, 4%) and
www.helv.wiley.com (29 of 33) e1800044 © 2018 Wiley-VHCA AG, Zurich, Switzerland

Helv. Chim. Acta 2018, 101, e1800044
22 (33 lmol, 20 mg, 30%). The purity of dissolved 21 female, supplier = Celsis, product # X008000, lot # PQP,
and 22 after chromatography was > 99% HPLC-UV. After number of pooled individuals = 20. The cryopreserved
fraction concentration and lyophilization, purity dropped hepatocytes were stored in liq. N . On the day of the
2
to 50% in case of 21 and to 76% in case of 22. Analyti- incubations, the hepatocytes were thawed according to
1
cal Data of 21: H-NMR ((D )DMSO, 303 K, 1:1 mixture the instructions provided by the supplier.
6
of conformers): 8.38 (m); 8.23 (m); 7.96 (s); 7.29 (m); 6.76
(
m); 6.71 (m); 5.33 (s); 4.79 (br. s); 4.66 (m); 6.71 (m); 5.33 Viability Measurements, Hepatocycte Incubations and
m); 4.79 (br. s); 4.66 (m, conformer A); 4.53 (m, con- Analytical Sample Preparation. In two parallel
former B); 3.97 (s); 3.82 – 3.79 (m, conformer A); experiments, the cryopreserved hepatocytes (either rat
.70 – 3.68 (m, conformer B); 3.62 – 3.60 (m, conformer or human origin as described above) were thawed,
(
3
A); 3.55 – 3.50 (m, conformer B); 3.49 – 3.47 (m, con- and the cells were suspended in serum-free
former A); 3.34 – 3.25 (m); 2.27 – 2.19 (m); 2.17 – 2.13 InVitroGRO HT medium (Celsis IVT) using 25 mL tissue
(
m); 2.12 – 2.04 (m, conformer A); 2.03 – 1.93 (m, con- culture flasks (Becton Dickinson Franklin Lakes). At
13
former B); 0.97 (m). C-NMR ((D )DMSO, 303 K, 1:1 mix- t = 0 min, the viability of the suspended hepatocytes
6
ture of conformers): 156.98; 138.57; 125.45; 138.57; was determined by
38.33; 137.43; 135.70; 125.45; 101.10; 54.26; 51.01; 50.71; (Chemometec) and was found to be 79.5% for rat
0.66; 49.61; 45.09; 44.54; 44.53; 43.97; 43.96; 31.62; 31.55; hepatocyctes and 80.0% for human hepatocyctes.
9.94; 27.15; 26.90; 9.16. UPLC/MS (ESI+): 449.2 (100, After cell counting, the cell density was adjusted to
a
Nucleocounter NC-200
1
5
2
+
+
6
[M + H] , calc. 449.19), 450.2 (23, [M + H] , calc. 450.19), approx. 1 9 10 viable cells per mL by adding
+
+
451.2 (2.5, [M + H] , calc. 451.19), 897.4 (5, [2M + H] , InVitroGRO KHB medium (Celsis IVT) to reach a total
calc. 897.37). HighRes MS: 449.1913 (+ 0.9 ppm) fitting volume of each suspension of 4 mL. The incubations
þ
3
for C H N O F with 99.59% confidence.
were started by adding 10 lL [ H]CDZ173 phosphate
21
24
6
2
3
salt as a solution in MeCN/DMSO/H O to each
Analytical Data of 22. H-NMR ((D )DMSO, 303 K, 1:1 hepatocyte suspension. Incubates were shaken at
2
1
6
mixture of conformers and diastereoisomers): 8.40 (s); 25 rpm and 37 °C under a humidified atmosphere of
8
.30 (s); 7.89 (s); 6.63 (m, conformer A); 6.56 (m, 95% air and 5% CO in a Heraeus incubator/Cytoperm.
2
3
conformer B); 6.10 (m); 5.90 (m); 3.94 (s); 3.84 – 3.69 The initial molar concentration of [ H]CDZ173 in the
m); 3.54 – 3.45 (m); 3.31 (m); 2.91 (m, conformer A); incubates was 5 lM (corresponding to 413 kBq/mL)
.67 (m, conformer B); 2.26 – 2.10 (m); 2.03 (m, and the final concentration of DMSO in each incubate
(
2
conformer A); 1.93 (m, conformer B); 0.99 (m). UPLC/ was below 0.5% (v/v). Aliquots of 250 lL were taken
+
MS (ESI+): 449.2 (100, [(M – H O) + H] ), 450.2 (22, [(M from either of the suspensions at t = 0 and
2
+
+
–
4
9
H O) + H] ), 452.2 ([M – (H O)] ), HighRes MS: t = 240 min (thus two samples of rat hepatocyctes
2 2
65.1866 (+ 0.9 ppm) fitting for C H N O F with and two samples of human hepatocyctes). At t = 0
2
1 24 6 2 3
9.73% confidence.
and t = 240 min, the metabolic reaction in each
sample was stopped by the addition of 500 lL
Incubation with Rat and Human Hepatocytes, Sample volumes of ice-cold MeCN, each sample was vortexed
n
Preparation, and LC/MS Analytical Method
and was kept at ꢀ20 °C overnight to assure the
protein precipitation to be completed. Afterwards
3
Material. [ H]CDZ173 phosphate salt ([C H F N O ] each sample was centrifuged for 20 min at 35 000 g
2
1 25 3 6 2
[
H PO ], 548.5 g/mol (unlabeled)) used for this (Sigma 3K30 centrifuge, Rotor# 12154), and the pellet
3 4
hepatocyte incubation was the material described was separated from the supernatant. The supernatants
earlier in the experimental part. No further dilution of these four samples were decanted. In a negative
was performed, however the purity of the stock control experiment the chemical stability of
3
solution was re-measured prior incubation by injection [ H]CDZ173 under the incubation conditions was
3
onto UPLC-RA system and was found to be 98.4%. A tested, i.e. [ H]CDZ173 was incubated in 4 mL
5
lM stock solution of this material in H O/MeCN (3:7, InVitroGRO KHB medium in the absence of
2
v:v) was used. The cryopreserved rat hepatocytes used hepatocytes and two aliquots were taken and
in this study had the following specifications: analyzed as described above.
strain = Sprague Dawley, gender = male, supplier Celsis,
product # M00005, lot # LNZ, number of pooled ies 20 lL weighted aliquots of each supernatant were
individuals = 17. The cryopreserved human transferred in a 6 mL LSC vial and a volume of 5 mL of
To determine the percentage radioactivity recover-
hepatocytes used in this study had the following scintillation cocktail (Rialuma, Lumac, The Netherlands)
specifications: strain = mixed, gender = male and was added. The radioactivity was measured (up to
www.helv.wiley.com
(30 of 33) e1800044
© 2018 Wiley-VHCA AG, Zurich, Switzerland