Hip to Be Square: Oxetanes as Design Elements to Alter Metabolic Pathways

Article abstract of DOI:10.1021/acs.jmedchem.9b00030

Oxetane-containing ring systems are increasingly used in medicinal chemistry programs to modulate druglike properties. We have shown previously that oxetanes are hydrolyzed to diols by human microsomal epoxide hydrolase (mEH). Mapping the enzymes that contribute to drug metabolism is important since an exaggerated dependence on one specific isoenzyme increases the risk of drug-drug interactions with co-administered drugs. Herein, we illustrate that mEH-catalyzed hydrolysis is an important metabolic pathway for a set of more structurally diverse oxetanes and the degree of hydrolysis is modulated by minor structural modifications. A homology model based on the Bombyx mori EH crystal structure was used to rationalize substrate binding. This study shows that oxetanes can be used as drug design elements for directing metabolic clearance via mEH, thus potentially decreasing the dependence on cytochromes P450. Metabolism by mEH should be assessed early in the design process to understand the complete metabolic fate of oxetane-containing compounds, and further study is required to allow accurate pharmacokinetic predictions of its substrates.

Full text of DOI:10.1021/acs.jmedchem.9b00030

Subscriber access provided by UNIV OF SOUTHERN INDIANA
Article
Hip to be square: oxetanes as design elements to alter metabolic pathways
Francesca Toselli, Marlene Fredenwall, Peder Svensson, Xue-
Qing Li, Anders Johansson, Lars Weidolf, and Martin A Hayes
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b00030 • Publication Date (Web): 16 Jul 2019
Downloaded from pubs.acs.org on July 17, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Hip to be square: oxetanes as design elements to
alter metabolic pathways
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†,§
†
‡
†
†
Francesca Toselli , Marlene Fredenwall , Peder Svensson , Xue-Qing Li , Anders Johansson ,
†
†,
Lars Weidolf and Martin A. Hayes *
†Cardiovascular, Renal and Metabolism, Innovative Medicines and Early Development Biotech
Unit, AstraZeneca, Pepparedsleden 1, Mölndal, 431 83, Sweden
‡Integrative Research Laboratories, Arvid Wallgrens Backe 20, Gothenburg, 413 46, Sweden.
KEYWORDS
Oxetane, epoxide hydrolase, metabolic clearance, drug design, spirocycles
ABSTRACT
Oxetane-containing ring systems are increasingly used in medicinal chemistry programs to
modulate drug-like properties. We have shown previously that oxetanes are hydrolyzed to diols
by human microsomal epoxide hydrolase (mEH). Mapping the enzymes which contribute to drug
metabolism is important, since an exaggerated dependence on one specific isoenzyme increases
the risk of drug-drug interactions with co-administered drugs. Herein we illustrate that mEH-
catalyzed hydrolysis is an important metabolic pathway for a set of more structurally diverse
oxetanes and the degree of hydrolysis is modulated by minor structural modification. A
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
homology model based on the Bombyx mori EH crystal structure was used to rationalize
substrate binding. This study shows that oxetanes can be used as drug design elements for
directing metabolic clearance via mEH, thus potentially decreasing the dependence on
cytochromes P450. Metabolism by mEH should be assessed early in the design process to
understand the complete metabolic fate of oxetane-containing compounds and further study is
required to allow accurate PK prediction of its substrates.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
INTRODUCTION
Early work in drug discovery programs aims at optimizing a host of compound properties like
pharmacological potency, off-target profile and pharmacokinetic properties, including metabolic
stability. All these properties are important for the prediction of dose and dosing frequency, to
reach the desired target engagement and efficacy in clinical trials. However, not only is it
essential to achieve sufficient metabolic stability of a clinical candidate, but also to understand
the metabolic pathways and the enzymes involved in the metabolism of the molecule.
The majority of marketed drugs depend on cytochrome P450 (CYP) isoenzymes for their
metabolic clearance.1,2 This is a potential concern in drug discovery programs, as there is a risk
of interaction between drugs that are metabolized to a large extent by the same CYP isoenzyme.
Such interactions can lead to elevated drug plasma levels and unwanted side effects for patients
taking several medications. This phenomenon is known as drug-drug interactions (DDI) and any
drug entering clinical trials must be carefully evaluated and monitored from a DDI perspective. It
is therefore desirable to optimize a lead series to minimize the risk of DDI. Designing away from
dependence on CYP-mediated metabolism is however not trivial and may result in switching of
the metabolism to non-CYP metabolizing enzymes e.g. aldehyde oxidase.3,4 A downside of
2
ACS Paragon Plus Environment

Page 3 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
directing metabolism to less-studied enzymes is that our understanding of their biology is not as
mature as that for CYPs and can lead to problems in human PK prediction.5-8
A functional group increasingly used for improving drug-like properties in medicinal
chemistry is the oxetane ring.9,10 It has been used as a mimic for carbonyl groups e.g. in
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
1
enzymatically-stable peptides or as a more hydrophilic replacement of a gem dimethyl
group.12,13 Oxetane-containing spirocyclic ring systems have proven to be versatile building
blocks, by improving the physicochemical properties and providing more favorable
pharmacokinetic profiles of a lead series.14-20
Recently, the introduction of the spirocyclic oxetanylazetidinyl moiety led to discovery of the
clinical candidate AZD1979 (1), a melanin-concentrating hormone receptor 1 antagonist
21
designed for the treatment of obesity-related disorders. Careful characterization of the
enzymology of its metabolism revealed that the oxetane moiety undergoes ring-opening in liver
fractions via an unusual non-oxidative metabolic route, i.e. hydrolysis by human microsomal
22
epoxide hydrolase (mEH; EC 3.2.2.9). This was the first example of a non-epoxide substrate
for this important drug-metabolizing enzyme, which was previously known to only catalyze
hydrolytic opening of epoxide-containing drug metabolites formed via Phase I oxidation.23,24
A
subsequent study on simple oxetanes from the same chemical series along with truncated
25
analogues revealed that oxetane hydrolysis by human mEH was not limited to 1. Thus, there is
evidence that these strained rings constitute a new class of substrates for mEH. This observation
has recently been recognized by the International Union of Biochemistry and Molecular Biology
with a change in the definition of mEH, now known as microsomal oxirane/oxetane hydrolase.26
The rate of hydrolysis can be fine-tuned by structural elements in the vicinity of the oxetane.25
This implies that oxetane-containing building blocks could be used as design tools to open up an
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
alternative metabolic clearance pathway and, potentially, to decrease the dependence on CYP-
mediated metabolic clearance.
The primary aim of the current study was to investigate hydrolytic ring-opening by mEH on a
more diverse set of oxetane-containing spirocycles related to 1, broadening our understanding of
the structure-activity relationship (SAR) of spirocyclic and bicyclic motifs. Secondly, an
additional set of simple oxetanes from an entirely different chemical series were investigated, to
understand the generality of oxetane hydrolysis by human mEH. Finally, due to the lack of a
crystal structure of human mEH, a homology model was built to better understand the SAR of
oxetane hydrolysis and to rationalize substrate binding. The findings presented here further
emphasize how the hydrolysis of simple and spirocyclic oxetanes by mEH is affected by subtle
structural modifications. Our results also indicate that metabolism can be directed towards mEH,
thus decreasing dependence on oxidative metabolism by CYP enzymes though caution should be
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
exercised if non-CYP enzymes are considered as major routes of clearance since PK predictions
may be challenging.5-8
RESULTS AND DISCUSSION
The compounds studied consisted of different ring-size combinations of bicyclic O- and N-
containing heterocycles with an emphasis on spirocyclic oxetanes (compounds 1-11, Table 1),
(compounds 16-23, Table 2), (compounds 24-27, Table 3) and a set of oxetane-substituted
pyrimidines (compounds 29-34, Table 4). The bicyclic oxetanes compounds were prepared
according to three different synthetic strategies (Figure 1). The majority of the compounds were
synthesized in a linear manner starting with the preparation of the eastern region esters.
4
ACS Paragon Plus Environment

Page 5 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Western region
Linker
Eastern region
Reductive amination
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Amide formation
Ether formation
Figure 1. Synthetic routes to oxetane heterocycles.
The final step following this strategy was a reductive amination between a benzylic aldehyde
precursor and an oxetane-containing amine. This strategy allowed for preparation of several
bicyclic oxetane amines from a common building block. A second strategy was amide formation
between an azetidine precursor and an eastern region ester, which, of course, is complementary
to the reductive amination strategy and allows for variation of the eastern region on a common
western region building block. The third and most convergent synthetic strategy was to form the
central ether bond as the last step. Both amide and ether formation were less commonly used for
preparation of target compounds, but they were useful for preparing larger batches when atom
economy is more important. It should be mentioned that ether formation appears to be more
substrate dependent and in some cases quite poor yields were observed. Synthetic routes for the
preparation of building blocks and target compounds are shown in Schemes 1 and 2. Several of
2
1
the building blocks have been reported previously and in the synthetic schemes only novel
compounds are numbered.
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
Scheme 1 Preparation of aldehyde building blocks for reductive amination as the last step
ii
i
3
7: R =CHF
1 2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
iii
3
8: R =CHF , R =Me
39: R =CHF , R =R =H
1 2 3 4
1
2
2
4
4
4
4
5
3: R =MeO, R =Me, R =H
1 3 4
4: R =MeO, R =H, R =Me
1
3
4
8: R =H, R =MeO, R =H
1
3
4
9: R =H, R =H, R =MeO
1
3
4
0: R =CHF , R =MeO, R =H
1
2
3
4
51: R =CHF , R =H, R =MeO
1
2
3
4
aReagents and conditions: (i) NaCN, 3-hydroxyazetidinium chloride, TEA, MeOH, rt (55%); (ii)
TEA, MsCl or TsCl, DCM, 0°C → rt (quant.); (iii) Cs
2
CO , DMA, 110°C (48-81%).
3
a
Scheme 2 Synthesis of analogues 6, 7, 10 and 11 using ether formation as the last step
i
4
4
4
4
5: R =MeO, R =H
1 2
7: R =H, R =MeO
1
2
0: R =Cl, R =H
1
2
1: R =H, R =Cl
1
2
ii
6
7
: R =Cl, R =R =H
1 2 3
4
4
2: R =H
6: R =CHF
3
: R =H, R =Cl, R =H
1
2
3
3
2
10: R =MeO, R =H, R =CHF
1
2
3
2
2
11: R =H, R =MeO, R =CHF
1
2
3
aReagents and conditions: (i) Na(CH
DMA, 90°C (6-51%).
CO
)
3
BH, TEA, DCM, rt (43%-quant); (ii) Cs
CO , DMF or
3
2
2
3
6
ACS Paragon Plus Environment

Page 7 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Synthesis and medicinal chemistry of all compounds from the pyrimidine series (Table 4) have
been described earlier.27-29
To properly assess the extent of mEH-mediated metabolism, it was crucial to discriminate
between oxido-reductive and mEH-mediated hydrolytic ring-opening of the spiro-oxetane. Both
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
these metabolic steps can lead to the formation of hydrated (+18.0106 Da, i.e. +H
2
O) metabolites
(a mechanism for CYP-mediated oxetane hydration is proposed below). Experiments were
carried out in human liver microsomes (HLM) in the presence or absence of NADPH, a cofactor
for CYP enzymes but not for EH.23
First, a set of analogues containing the spiro-oxetanylazetidinyl building block with different
substitution patterns were investigated (entries 1-11, Table 1). These compounds formed two
distinct hydrated (+18 Da) metabolites from either oxido-reductive ring-opening of the azetidine
or via hydrolytic and/or oxido-reductive ring opening of the oxetane. Structures of both +18 Da
metabolites were confirmed by UPLC-MS analysis of HLM incubations of the analogue 1
alongside synthetic standards of both hydrolyzed oxetane (12) and ring-opened azetidine (13)
metabolites (See Supporting Information, Figure S1). For all compounds in this set (1-11),
formation of diol could be detected in the presence of NADPH and could account for as much as
1
5% of the total parent-related material (cf 1 and 6; Table 1). In the absence of NADPH, diol
formation could be also detected for compounds 1-11, suggesting mEH-mediated diol formation
Table 1). For the majority of compounds in the spiro-oxetanylazetidinyl series (Table 1), an
(
increase in diol formation was observed in the absence of NADPH, likely reflecting the higher
substrate availability to mEH-mediated hydrolysis in absence of a functional CYP system.
However, it is clear that subtle structural changes have a large impact on the metabolic fate of
compounds from this chemical series. For instance, meta-chloro substitution of the western
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
phenyl ring (entry 7) followed the pattern observed for most compounds in the series, with an
increase of diol formation in the absence of NADPH. In contrast, ortho-chloro substitution of the
western phenyl ring (entry 6) led to a higher proportion of diol formation in the presence of
NADPH. Diol formation could also be detected in the absence of NADPH for 6 but CYP-
mediated oxetane ring-opening to diol is the major pathway for this compound, similar to what
has previously been reported for other oxetane-containing compounds.14,30 A possible reaction
mechanism for oxetane ring-opening to diol by CYP-mediated oxidation, proceeds via α-
hydroxylation, spontaneous ring-opening to the aldehyde and finally reduction to diol by
aldehyde reductase.14,30 The eastern phenyl chloro-substituted pair 6 and 7 exhibit similar
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
physicochemical properties, with slightly higher logD (3.0 and 2.7, respectively) and lower pK
7.5 and 7.7, respectively), as compared to the non-chlorinated analogue 2 (logD 1.8, pK 8.0).
The western phenyl methyl-substituted matched pair 8 and 9 both show an increase of diol
a
(
a
formation in the absence of NADPH. For 8, the diol metabolite increased fourfold in the absence
of NADPH (22% vs 5.2%, Table 1) and for 9, a two-fold increase was observed. This pair can be
compared to 1, which in addition, is also 4-methoxy-substituted in the eastern phenyl ring and
for which a decrease in diol formation was observed without NADPH. The introduction of a
methyl group in the western phenyl ring of 1 to yield 8 and 9 gave an increase in lipophilicity of
around 0.5 log units, an increase in overall hepatic metabolism and a higher proportion of non-
NADPH mediated diol formation.
8
ACS Paragon Plus Environment

Page 9 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 1. Physicochemical parameters, metabolic clearance and fraction of diol metabolites
formed in HLM from 2-oxa-6-azaspiro[3.3]heptanes (1-11).
c
% Amino
Hhep
% Diol in
HLM
Solubil
% Diol in
HLM
CLint
(µl/min/1
a,b
a
alcohol
w
NADPH
Compound
Structure
ity
µM)
LogD
pKa
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
w/o
a
d
(
w NADPH
6
d
0
cells)
NADPH
1
2
3
4
5
9
8
8
1
9
6
2.0
1.8
2.4
2.5
2.0
8.2
8.0
8.5
NA
8.2
10.5
15.6±0.3
5.9±0.1
2.7±0.3
16±1
0.7
6.1
5.3
1.7±0.1
7.3±0.1
<1%
<1
3.0±1.4
2.4±0.6
1.9±0.5
51
27.2
1.9
4.2±0.1
3.5±0.2
7
35
1.2±0.0
6
9
0
3.0
7.7
32.7
15±0
6.3±0.4
1.7±1.7
7
8
9
7
6
3
2.7
2.6
2.5
7.5
8.2
8.4
12.4
17.0
18.6
2.9±2.9
5.2±1.1
10±0
8.6±0.2
22±1
0
7
<1
74
19±1
3.1±0.2
1
1
0
1
8
7
2.1
2.2
8.2
8.9
19.4
19.5
6.2±0.7
2.1±0.6
17±1
5.1±0.3
2.7±0.9
82
3.2±0.0
a
31 b
Solubility, logD, pK
a
and CLint were measured as described previously ; LogD7.4 HPLC;
c
d
Hhep, human hepatocytes; HLM, human liver microsomes; average diol levels ± range were
determined from duplicate incubations (60 min) of 10 µM substrate either with HLM or without
NADPH. Diol metabolite levels are expressed as a percentage of the sum of the total parent-
related peak areas. All data are from reactions in absence of inhibitors.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
For the 1, 8 and 9 trio, the correlation to logD can be made, but comparing 1-5, all unsubstituted
in the western phenyl region and with different 4-substituents of the eastern phenyl region, there
is no clear correlation to lipophilicity and hepatic stability. For example, the most lipophilic
compound of these (4, logD 2.5), which as expected shows the highest hepatic turnover, is
metabolized to diol to a higher extent without NADPH. This is the case for all the examined
unsubstituted western phenyl analogues, with the exception of 1. This compound and the ortho-
chloro substituted western phenyl 6 are the only analogues of those examined to show a higher
degree of diol formation in the presence of NADPH.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In conclusion, for the spiro-oxetanylazetidinyl series there is no clear correlation between the
measured physicochemical properties, logD and pK , and mEH-mediated diol formation. Neither
a
does a higher metabolic turnover in complementary studies using human hepatocytes (Table 1)
correlate with a higher fraction of diol formed. However, it should be pointed out that the logD
range in the series is not very wide. The difference in the degree of mEH-mediated metabolism is
on the other hand quite significant, suggesting that introduction of an oxetane moiety can greatly
affect metabolic clearance pathways even within a rather narrow physicochemical property
space.
In an effort to better understand the observed SAR, we constructed a protein structure model of
human mEH and performed docking studies of the substrates to the model structure. The human
mEH model was constructed by comparative homology modelling using the crystal structure of
3
2
juvenile hormone epoxide hydrolase from the silkworm Bombyx mori (PDB ID: 4QLA) as
template and using the aligned inhibitor valpromide from the crystal complex with the
3
3
Aspergillus niger EH (PDB ID: 3G0I) in the model construction. Compared to earlier published
models,34,35 the structure model presented here is based on templates with a significantly higher
1
0
ACS Paragon Plus Environment

Page 11 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
degree of sequence identity overall and particularly so in the active site (see Figures S2 and S3
for more details of the homology model).
2
5
Initially, the N-methylated 3-methyloxetane analogue 14 was used to guide docking and energy
minimizations in the active site (Figure 2a). Analysis of the low-energy binding mode of 14
suggested that ring closure to a 3-oxetane piperidine spirocycle should constitute an excellent fit
in the active site, presenting the oxetane oxygen in an optimal orientation for catalysis (Figure
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
b). A methyl-substituted analogue of the spiro-piperidine was available in the AstraZeneca
compound collection, 15. When docking 15, the western methyl substituent could adopt two
rotamers that both fitted into the enzyme binding site. The rotamer where the methyl is pointing
inwards has a slightly more optimal orientation of the oxetane moiety in the catalytic site. This
rotamer is shown in Figure 2b. The spiro-piperidine 15 indeed showed almost complete
conversion to diol in the absence of NADPH (Figure 2b).
1
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 2. Docking poses in the active site of the mEH protein structure model. The accessible
surface of the protein, represented by how close an oxygen atom in a water molecule could be
without any steric repulsion, is shown for the parts of the protein that are in close proximity to
the ligand. Green color: hydrophobic areas, and purple: hydrophilic areas. The two tyrosines in
the active site coordinating with the ring oxygen in the ligand (Y299 & Y374, OH distances in
Å) and the catalytic aspartic acid (D226) are shown. a) A docked pose and the 2D representation
of 14; b) A docked pose and the 2D representation of 15. Data for the observed mEH-catalyzed
oxetane hydrolysis of 14 and 15 is shown beneath each structure.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In contrast, the spiro compound 2 (Table 1), which shows lower degree of diol formation,
cannot bind in a pose optimal for catalysis without minor steric clashes in the vicinity to the
oxetane (Figure S4). The docking pose for 2 suggests a steric clash, with the oxygen atom of the
oxetane penetrating through the interaction surface, shortened hydrogen bond distances and a
slightly bent oxetane ring. This steric hindrance for compound 2 compared to the corresponding
spiro-oxetane piperidine 15 suggests a marked difference in substrate turnover.
3
6
Attempts to crystallize the mEH protein have been unsuccessful to date. In the absence of X-
ray data, we applied the homology model and guided dockings to the bicyclic oxetane series. For
this series, the structural insights have been useful for understanding where the ligand vs active
site shape and electrostatic complementarity set limits for the ligand to bind in a pose that is
optimal for hydrolysis. The model was used to increase our qualitative understanding of mEH-
mediated oxetane hydrolysis for the bicyclic oxetane series.
We then wanted to expand the investigation of mEH-catalyzed hydrolysis to other ring
systems. The metabolism of a set of 2-spiro-oxetane pyrrolidines were investigated under the
same conditions as before (Table 2, compounds 16-20). NADPH-independent ring-opening of
1
2
ACS Paragon Plus Environment

Page 13 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the oxetane was again observed, but for these compounds diol formation was generally higher in
the presence of NADPH, suggesting CYP-mediated oxidative diol formation is a more
pronounced competing pathway for these substrates. It should be noted that these pyrrolidines
are more lipophilic than the azetidines. However, this does not account for the increase in CYP-
mediated oxetane ring openings observed per se. As shown, more lipophilic substrates, e.g. spiro
piperidine 15, are completely turned over by mEH (Figure 2). For the spiro pyrrolidines, western
ortho-methoxy phenyl substitution attenuates CYP-independent diol formation. For 17, no diol
was formed in the absence of NADPH, suggesting that this compound is not a substrate of mEH.
The analogue 19 was turned over by mEH, which is somewhat surprising as the eastern phenyl
group according to our homology model would be partly outside the enzyme and substitutions of
this region should have a minor impact on the turnover.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 2. Physicochemical parameters, metabolic clearance and fraction of diol metabolites
formed in HLM from spiro oxetanes and tetrahydrofurans (16-22) and an embedded oxetane
(23).
c
Hhep
%
Diol in
HLM
% Diol in
HLM
%
Amino
CLint
Compound
Structure
Solubility
a,b
a
LogD
pKa
(µl/min
alcohol w
NADPH
a
(µM)
w
w/o
6
/
10
d
NADPHd NADPH
cells)
1
1
1
6
7
8
3
3.9
3.8
3.5
6.0
8.0
6.0
48.7
12±1
22±2
23±2
4.5±0.2
0
0
0
60
27.5
25.0
<1
2
21
15±3
1
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
9
4
3
8
3
9
2
4.2
3.8
2.2
7.0
6.1
8.6
42.7
38.5
10.8
5.1±0.4
22±1
0
2.9±0.1
28±0
<1
0
0
20
21
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
<1
2
2
2
3
495
3
2.7
2.9
ND
7.3
10.4
38.0
0
<1
<1
4.4±0.4
21±1
2.7±0.2
aSolubility, logD, pK
and CLint were measured as described previously ; LogD7.4 HPLC;
Hhep, human hepatocytes; HLM, human liver microsomes; average diol levels ± range were
determined from duplicate incubations (60 min) of 10 µM substrate either with HLM or without
NADPH. Diol metabolite levels are expressed as a percentage of the sum of the total parent-
related peak areas. All data are from reactions in absence of inhibitors.
31
b
a
c
d
The meta-methoxy substituted western phenyl spiro-pyrrolidines 18 and 20 were hydrolyzed
by mEH to a much greater extent and for 20 even an increase in diol could be observed. Again,
this finding was unexpected and warrants further study of the SAR of CYP- vs mEH-mediated
oxetane ring-cleavage.
For spiro-tetrahydrofuran azetidines (Table 2, 21 and 22), no diol formation of the less strained
THF ring was detected in the absence of NADPH. The tetrahydrofuran ring of these compounds
was also stable to oxidative ring opening. Nor did an extended set of compounds containing only
tetrahydrofurans or tetrahydropyrans form any metabolites corresponding to the addition of
water (data not shown). However, the bridged and rather strained oxetanyl piperidine ring (Table
2, 23) was hydrolyzed to diol in the absence of NADPH. The observed increase in diol formation
1
4
ACS Paragon Plus Environment

Page 15 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
was quite pronounced, demonstrating that mEH-catalyzed ring opening of oxetanes is not only
confined to spirocyclic ring systems or simple oxetanes.
These data suggested that bicyclic oxetane rings were opened via hydrolysis and that this
metabolic route may be enhanced in the absence of NADPH, i.e. in the absence of a functional
CYP system. We further investigated whether spiro-oxetane metabolism by mEH would increase
also in the presence of a potent CYP inhibitor. A subset of compounds was selected to
investigate this hypothesis: the 3-oxetane azetidines 3, 4 and 5, the 3-oxetane piperidine 15, the
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
-oxetane pyrrolidine 20, the 3-tetrahydrofuran azetidine 21 and the bridged oxetane 23. With
the exception of 21, which was included as a negative control, the selected compounds all had
shown an increase in diol formation in the absence of NADPH (Tables 1 and 2). The compounds
were incubated with HLM in the presence of a high concentration of the potent CYP inhibitor
ketoconazole and NADPH. The observed diol formation pattern matched our previous
observations in the +/- NADPH experiments. Compounds 3, 4, 5, 15, 20 and 23 showed an
increase in diol formation in the presence of ketoconazole (see Figure 3 for a representative plot
for 23), whereas this, as expected, was not the case for the 3-tetrahydrofuranyl-azetidine 21 (data
not shown). As seen in incubations devoid of NADPH (Figure 3 and Tables 1-2), an increase in
diol formation likely results from higher availability of the substrate to enzymes other than
CYPs, when these are inhibited by ketoconazole. To confirm the involvement of mEH, the same
set of compounds was also incubated with a high concentration of the selective mEH inhibitor
valpromide, this time abolishing diol formation (see Figure 3, for a representative plot for 23).
1
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Hydration of 23 in selective conditions. No HLM, incubations without human liver
microsomes; CTRL, incubations without inhibitors ([NADPH], 1 mM); No NADPH, incubations
in absence of NADPH; KTZ, incubations in presence of the CYP inhibitor ketoconazole (200
µM; [NADPH], 1 mM); VPD, incubations in presence of the mEH inhibitor valpromide (1 mM;
[
NADPH], 1 mM). Results are expressed as a % of CTRL reactions; bars are averages of
duplicate measurements + range.
These results suggest that introduction of an oxetane can direct metabolism away from CYP
enzymes, which can have an impact on the risk of DDI in a chemical series. Of course, this will
vary considerably between different chemical lead series and, as we have shown, between
similar compounds from the same series. Nonetheless, introduction of an oxetane in a lead series
can be a way to avoid a high degree of metabolism by CYP enzymes.
In an effort to understand the minimum structural requirements for mEH-catalyzed hydrolysis,
we prepared a representative set of truncated spirocyclic analogues and analyzed their
biotransformation as described earlier (compounds 24-27, Table 3).
1
6
ACS Paragon Plus Environment

Page 17 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. Physicochemical parameters, metabolic clearance and fraction of diol metabolites
formed in HLM from truncated spiro-oxetanes
c
Solubility
HLM CLint
(µl/min/mg)
% Diol in HLM
w NADPH
% Diol in HLM
w/o NADPH
LogDb
Compound
Structure
pK
a
a
d
d
(µM)
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
2
4
5
6
364
0.40
8.25
8.08
8.56
<3
0
0
744
750
1.50
28.3
20.6
8.3±0.4
<1
10.9±0.2
1±0
0.80
2
7
610
1.80
ND
8.10
9.2±0.6
0
aSolubility, logD, pK
and CLint were measured as described earlier ; LogD7.4 HPLC; HLM,
31
b
c
a
d
human liver microsomes; Average diol levels ± range were determined from duplicate incubations
(60 min) of 10 µM substrate with HLM. Diol metabolite levels are expressed as a percentage of
the sum of the total parent-related peak areas. All data are from reactions in absence of inhibitors.
The shortened 3-oxetanyl-azetidine 24 was metabolically stable to human liver microsomes,
presumably due to low lipophilicity. Both oxetane piperidines 25 and 26 were hydrolyzed to a
greater extent by mEH in the absence of NADPH, but the turnover increase of 25 is much less
pronounced than for the “full-length” analogue 15. For 27, formation of the diol metabolite was
NADPH-dependent and completely inhibited by ketoconazole, rather than valpromide (Figure
4
a). This is consistent with a CYP-mediated oxidative biotransformation and matches the
findings for the spiro-oxetanyl-pyrrolidines of the long-chain series, particularly 17. To confirm
that addition of water occurred on the oxetane and not on the pyrrolidine, an authentic oxetane
ring-opened diol standard of 27 was synthesized (compound 28). When the reaction mixture
from incubations of 27 with HLM and NADPH was co-chromatographed with 28 they were
found to co-elute by UPLC-MS, thus confirming the spiro-oxetane of this truncated compound
was ring-opened via an oxidation/reduction sequence catalyzed by CYPs and a reductase (e.g.
1
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
30
aldehyde reductase) as mentioned previously , forming the diol metabolite 28 (Figures 4b and
4
c).
We were then interested to understand whether mEH-catalyzed oxetane hydrolysis was
significant in a structurally-unrelated chemical series. A search in the AstraZeneca compound
collection was made and a series of oxetane-substituted pyrimidines were selected.27-29
For the set of compounds selected (Table 4), matched pairs were available where the oxetane
moiety was attached to the pyrimidine core via either amine or ether linkers. Again, the overall
metabolism in HLM and diol formation in the presence and absence of NAPDH was investigated
as described earlier. For the amine-linked pyrimidines 29, 31 and 33, oxetane ring-opening
increased in the absence of NADPH (Table 4). Additional experiments showed that oxetane
cleavage was completely inhibited by valpromide, was unaffected by ketoconazole and was
NADPH-independent in HLM (see Figure 4d for a representative plot for 29). Similar results
were obtained in human hepatocytes (data not shown). Hydrolysis was also unaffected by the
soluble EH (sEH) inhibitor t-AUCB and was absent in human liver cytosol (HLC), ruling out the
contribution of sEH (data not shown). Interestingly, the oxetane ether-linked pyrimidine matched
pairs 30, 32 and 34 were completely stable to oxetane hydrolysis (Table 4).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
8
ACS Paragon Plus Environment

Page 19 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 4. Physicochemical parameters, metabolic clearance and fraction of diol metabolites
formed in HLM from N- and O-linked oxetane substituted pyrimidines (29-34) and N and O-
linked methyl oxetanes 35 and 36.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
% Diol in
HLM w
NADPH
% Diol in
HLM w/o
NADPH
c
Solubility
HLM CLint
Compound
Structure
LogDb pK
a
a
(µM)
(µl/min/mg)
d
d
2
3
9
0
580
1.1
2.0
2.8
<2
21±0
0
27±0
0
867
ND
2.4
ND
2.6
2.6
<2
<3.5
6.1
<2
31
>2520
818
1.6
1.6
2.1
1.9
8.5±0.2
12±0
3
3
2
3
0
9.4±0.6
0
0
>1820
476
14±0
3
3
4
5
2.8
0
61
3
2
8.2
103
32±0
31±0
98±3
36
ND
ND
>300
79±21
aSolubility, logD, pK
31
b
c
a
and CLint were measured as described earlier ; LogD7.4 HPLC; HLM,
d
human liver microsomes; Average diol levels ± range were determined from duplicate incubations
60 min) of 10 µM substrate with HLM. Diol metabolite levels are expressed as a percentage of
(
the sum of the total parent-related peak areas. All data are from reactions in absence of inhibitors.
ND, not determined.
1
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. A) Hydrolysis of 27; B) and C) UPLC-MS extracted ion chromatograms for:
compound 27 and its hydrated metabolite formed in an incubation with HLMs and NADPH (B);
synthetic diol 28 (C). Peak labels are m/z values and correspond to parent (m/z 234.1484) and
hydrated metabolite (m/z 252.1596). D) Hydrolysis of 29. Abbreviations in panels A and D are as
defined in the legend for Figure 3. Results are expressed as a % of CTRL reactions; bars are
averages of duplicate measurements + range.
The matched pair 29 and 30 could both be docked into the active site of the homology model
in a pose suitable for catalysis, as shown in Figure 5. Hence, the lack of diol formation of 30 in
the absence of NADPH could not be explained by how it fits into the active site of the model.
The proximity of the amine/ether link to the site of hydrolysis suggests that this may be due to a
reactivity difference between the anilinic and phenolic oxetane substituent in 29 and 30,
2
0
ACS Paragon Plus Environment

Page 21 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
respectively. Furthermore, dockings with the amine-linked pyrimidines 29, 31 and 33 (Figure
S5) revealed a trend towards more steric hindrance with increasing size of cyclic alkyls resulting
in a docking pose for the cyclopentyl-substituted pyrimidine 33 clearly less optimal for catalysis.
However, the difference in diol formation in absence of NADPH between the cyclopropyl-
substituted pyrimidine 29 and the cyclobutyl-substituted pyrimidine 31 could not be resolved by
the model.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
This prompted us to incorporate an ether linker between the proximal phenyl ring and the
terminal methyl oxetane. For this purpose, compound 36 was synthesized to form a matched pair
with the simple oxetane 35 and subjected to the same battery of assays (Table 4).
To our surprise, the opposite relationship was observed, with oxetane hydrolysis in the absence
of NADPH increasing from 31% with the amine-linked compound 35 to 98% with the ether
linked 36 (Table 4). In an attempt to rationalize this finding, the homology model was revisited.
A closer look at the binding site suggested that the substrates with spirocyclic amines or their
corresponding ring-opened analogues would bind more strongly in their neutral states.
This can be deduced from the lipophilic nature of the binding site region. Introduction of a
positive charge in this lipophilic cavity would decrease affinity for the ligand and reduce diol
formation. This may be an explanation as to why the ether analogue 36 shows such a high degree
of diol formation, as it is neutral. However, it should be pointed out that there is no general trend
towards higher degree of diol formation for amine substrates with a lower pK
a
within the pK
a
range examined. Again, this is an example of how structural changes in other regions of the
molecule aside from the oxetane ring system can have a marked effect on the metabolic fate and
these effects appear to vary between chemical series.
2
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Docking poses of 29 (left) and 30 (right) in the active site of the mEH protein
structure model. The accessible surface of the protein, represented by how close an oxygen atom
in a water molecule could be without any steric repulsion, is shown for the parts of the protein
that are in close proximity to the ligand. Green color: hydrophobic areas and purple: hydrophilic
areas. The two tyrosines in the active site coordinating with the ring oxygen in the ligand (Y299
& Y374, OH distances in Å) and the catalytic aspartic acid (D226) are shown.
A set of electronic properties were also calculated using both semi-empirical QM (am1 and pm3)
3
7
and density functional theory B3LYP/6-31G* calculations. Both LUMO energies and partial
charges for the oxygen and surrounding carbon atoms in the oxetane/tetrahydrofuran rings were
calculated using the methods described above. Calculations were performed for the oxetane/
tetrahydrofuran containing fragments which included substitutions that affect reactivity (11
fragments in total were identified in the dataset). LUMO energy calculations using semi-
2
2
ACS Paragon Plus Environment

Page 23 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
empirical methods were also performed on intact molecules. No correlations between these
electronic properties and the degree of diol formation w/o NADPH were found in the dataset
(data not shown). The most likely explanation for this lack of correlation is that the degree of
hydrolytic ring-opening is related to both the fit of the substrate to the active site and the
reactivity of the substrate. These factors would not necessarily be individually correlated with the
calculated physicochemical properties and the combination most certainly not.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CONCLUSIONS
Building on our previous work22,25 we have considerably expanded our knowledge on the
chemical space of non-epoxide substrates of mEH. Our findings demonstrate that short- and
long-chain bicyclic oxetanes as well as a second structurally distinct compound series containing
simple oxetanes are turned over by mEH, confirming that hydrolysis by mEH is not confined to
oxiranyl substrates only.
A homology model of human mEH was developed, that is improved with respect to sequence
similarity as compared to earlier models described in the literature.34,35 The structural insights
from the homology model and guided dockings were useful for understanding the limits for the
ligand to bind in a pose which was optimal for hydrolysis by mEH. With this model we can
partially understand and explain the SAR in terms of degree of diol formation.
There were no clear correlations between the degree of oxetane hydrolysis and
physicochemical properties such as pK and logD or from calculated electronic properties i.e.
a
37
LUMO energies or partial charges. This study shows that hydrolytic ring-opening is an
important metabolic pathway for oxetanes over different ring systems and that substitution and
small structural changes within a chemical series can greatly affect non-oxidative pathways. This
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
study also demonstrates that it is possible to direct a compound’s metabolic clearance via mEH
by careful selection of substituted bicyclic or simple oxetanes. Thus, by decreasing the
dependence on CYP-mediated oxidations, metabolic stability may be optimized and drug-drug
interaction risk decreased.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Attempts at modulating the clearance of small molecules to non-CYP mediated pathways is
however not without risk. Human PK may be poorly predicted, since knowledge around mEH
biology is relatively limited compared to the CYP system. Problems have been observed with
AO-driven clearance recently and have led to clinical failure,5-8 though improvements to AO
3
8
scaling using e.g. multispecies allometry have recently been reported. mEH is known to be
widely expressed in human tissues, where it may also take part in physiological function, and
polymorphic variants have been reported23, 24, 39, 40 indicating possible risks for exposure
underprediction and inter-individual variability for the metabolism of its substrates.
Further fundamental research into this important non-CYP drug-metabolizing enzyme is thus
warranted and we strongly recommend that screening for mEH-mediated metabolism is included
in drug discovery projects with oxetane-containing lead series.
EXPERIMENTAL SECTION
General Chemistry. Purchased chemicals were reagent-grade and used without purification.
1H NMR spectra were recorded at 400, 500 or 600 MHz Chemical shifts (ppm) and were
determined relative to internal solvent (δ 7.26 ppm; CDCl
3
, δ 2.50 ppm; DMSO-d , δ 3.31 ppm;
6
MeOD). Microwave-assisted synthesis was carried out in an Initiator synthesizer single mode
cavity instrument producing controlled irradiation with a power range 0–400 W at 2450 MHz
(Biotage AB, Uppsala, Sweden). Analytical HPLC/MS was conducted on a QTOF mass
2
4
ACS Paragon Plus Environment

Page 25 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
spectrometer using a UV detector monitoring either at (a) 210 nm with a BEH C18 column (2.1
mm × 100 mm, 1.7 μm, 0.7 mL/min flow rate), using a gradient of 2% v/v ACN in H
(ammonium carbonate buffer, pH 10) to 98% v/v ACN in H O, or at (b) 230 nm with an HSS
C18 column (2.1 mm × 100 mm, 1.8 μm, 0.7 mL/ min flow rate), using a gradient of 2% v/v
ACN in H O (ammonium formate buffer, pH 3) to 98% v/v ACN in H O. Preparative HPLC was
performed by Waters Fraction Lynx with ZQ MS detector on either a Waters Xbridge C18 OBD
µm column (19×150 mm, flow rate 30 mL/min or 30×150 mm, flow rate 60 mL/min) using a
gradient of 5–95% ACN with 0.2% NH at pH 10 or a Waters SunFire C18 OBD 5 µm column
19×150 mm, flow rate 30 mL/min or 30×150 mm, flow rate 60 mL/min) using a gradient of 5–
95% ACN with 0.1 M formic acid. Automated flash chromatography was performed on a
Biotage system using a SiO column and UV detection. High-resolution mass spectrometry
HRMS) was carried out using high-resolution electrospray ionization mass spectrometry where
2
O
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
5
3
(
2
(
the spectrometer was linked together with an Acquity UPLC system. All tested compounds were
determined to be ≥95% pure using the analytical method a or b described above based on the
peak area percentage.
Compounds 1, 2, 16, 21 and 22 as well as intermediates 6-(4-(azetidin-3-yloxy)benzyl)-2-oxa-6-
azaspiro[3.3]heptane, 4-(1-(5-(4-methoxyphenyl)-1,3,4-oxadiazole-2-carbonyl)azetidin-3-
yloxy)benzaldehyde, 4-((1-(5-phenyl-1,3,4-oxadiazole-2-carbonyl)azetidin-3-
yl)oxy)benzaldehyde, 1-(5-(4-methoxyphenyl)-1,3,4-oxadiazole-2-carbonyl)azetidin-3-yl
methanesulfonate, 1-(5-phenyl-1,3,4-oxadiazole-2-carbonyl)azetidin-3-yl methanesulfonate and
(3-hydroxyazetidin-1-yl)(5-phenyl-1,3,4-oxadiazol-2-yl)methanone were all synthesized as
2
1
previously described. All 5-phenyl-1,3,4-oxazole intermediates were synthesized from the
2
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
42
corresponding benzohydrazines. Compounds 29, 30, 31, 32, 33 and 34 have previously been
2
7
described. Compounds 14 and 35 were synthesized according to a previously described
procedure.25
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(3-(4-(2-oxa-6-azaspiro[3.3]heptan-6-ylmethyl)phenoxy)azetidin-1-yl)(5-(4-
(
difluoromethyl)phenyl)-1,3,4-oxadiazol-2-yl)methanone (3).
(
5-(4-(difluoromethyl)phenyl)-1,3,4-oxadiazol-2-yl)-(3-hydroxyazetidin-1-yl)methanone (37).
Ethyl 5-(4-(difluoromethyl)phenyl)-1,3,4-oxadiazole-2-carboxylate (1.0 g, 3.73 mmol) and
azetidin-3-ol hydrochloride (490 mg, 4.47 mmol) were mixed in MeOH (20 mL). TEA (0.62 mL,
4
.47 mmol) was added followed by NaCN (0.18 g, 3.73 mmol) and the reaction was stirred at rt
for 2 days. Most of the MeOH was evaporated off and water (20 mL) was added into the residue.
The mixture was stirred for 5 min and the formed precipitate was filtered off, washed with water
1
(
10 mL) and dried under vacuum to afford 37 (603 mg, 55%). H NMR (400 MHz, CDCl
3
) δ
8
.26 (d, J = 7.3 Hz, 2H), 7.69 (d, J = 7.3 Hz, 2H), 6.72 (t, J = 56 Hz, 1H), 4.75 – 5.11 (m, 2H),
4.43 – 4.69 (m, 2H), 4.16 (d, J = 10.6 Hz, 1H), 2.39 (s, 1H).
(
1-(5-(4-(difluoromethyl)phenyl)-1,3,4-oxadiazole-2-carbonyl)azetidin-3-yl) methanesulfonate
(
38). 37 (603 mg, 2.04 mmol) was dissolved in DCM (10 mL) and cooled in an ice-bath. TEA
(0.311 mL, 2.25 mmol) was added and then MsCl (0.167 mL, 2.14 mmol) was added dropwise
over 5 min. The ice-bath was removed and the reaction stirred to rt for 3h. More DCM and aq.
NaHCO
3
were added, the mixture filtered through a phase separator and evaporated to give 38
1
(807 mg, 106%). H NMR (400 MHz, CDCl
3
) δ 8.26 (d, J = 8.2 Hz, 2H), 7.70 (d, J = 8.2 Hz,
2
H), 6.72 (t, J = 56 Hz, 1H), 5.34 – 5.48 (m, 1H), 5.06 – 5.20 (m, 1H), 4.83 – 4.95 (m, 1H), 4.59
4.72 (m, 1H), 4.37 – 4.50 (m, 1H), 3.13 (s, 3H).
–
2
6
ACS Paragon Plus Environment

Page 27 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
4
-(1-(5-(4-(difluoromethyl)phenyl)-1,3,4-oxadiazole-2-carbonyl)azetidin-3-yl)oxybenzaldehyde
(39). 38 (800 mg, 2.14 mmol), 4-hydroxybenzaldehyde (275 mg, 2.25 mmol) and cesium
carbonate (733 mg, 2.25 mmol) were mixed in DMA (10 mL) and stirred at 100°C over night.
After cooling to rt, EtOAc was added to the reaction mixture which was washed with water (4x),
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
dried (MgSO ) and evaporated. The residue was purified by automated flash chromatography on
4
a 50g column. A gradient from 30% to 100% of EtOAc in heptane over 15CV was used as
mobile phase. The product was collected using the wavelength 272 nm. Relevant fractions were
1
pooled and evaporated to afford 39 (333 mg, 39%). H NMR (400 MHz, CDCl
3
) δ 9.93 (s, 1H),
8
.27 (d, J = 8.4 Hz, 2H), 7.85 – 7.93 (m, 2H), 7.70 (d, J = 8.2 Hz, 2H), 6.91 (d, J = 8.7 Hz, 2H),
6.79 (t, J = 56 Hz, 1H), 5.14 – 5.25 (m, 2H), 4.77 – 4.86 (m, 1H), 4.72 (dd, J = 6.2, 12 Hz, 1H),
.38 (dd, J = 2.1, 11.7 Hz, 1H).
3-(4-(2-oxa-6-azaspiro[3.3]heptan-6-ylmethyl)phenoxy)azetidin-1-yl)(5-(4-
4
(
(difluoromethyl)phenyl)-1,3,4-oxadiazol-2-yl)methanone (3). 39 (55 mg, 0.14 mml), 2-oxa-6-
azaspiro[3.3]heptane oxalate (32 mg, 0.17 mmol) and TEA (0.057 mL, 0.41 mmol) were mixed
in DCM (2 mL) and stirred at rt for 30 min. Sodium triacetoxyhydroborate (58 mg, 0.28 mmol)
was added and the reaction was stirred at rt over weekend. NaHCO (sat. 2 mL) and DCM (2
3
mL) were added, the mixture stirred, filtered through a phase separator and evaporated. The
product was purified by preparative HPLC eluting with a gradient of ACN (5-95%) in 0.2%
1
ammonia buffer to afford 3 (40 mg, 60%). H NMR (400 MHz, CDCl
3
) δ 8.26 (d, J = 8.3 Hz,
2H), 7.69 (d, J = 8.2 Hz, 2H), 7.19 (d, J = 8.6 Hz, 2H), 6.88 – 6.55 (m, 3H), 5.18 – 5.03 (m, 2H),
4
4
.79 – 4.75 (m, 1H), 4.74 (s, 4H), 4.69 – 4.61 (m, 1H), 4.37 – 4.30 (m, 1H), 3.50 (s, 2H), 3.38 (s,
+
+
H). HRMS calcd for C25
H
25
F
2
N
4
O , [M+H] , 483.1844; found 483.1884.
4
2
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 59
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
3-(4-(2-oxa-6-azaspiro[3.3]heptan-6-ylmethyl)phenoxy)azetidin-1-yl)(5-(4-
(difluoromethoxy)phenyl)-1,3,4-oxadiazol-2-yl)methanone (4). Ethyl 5-(4-
(difluoromethoxy)phenyl)-1,3,4-oxadiazole-2-carboxylate (60 mg, 0.21 mmol) and 6-(4-
azetidin-3-yloxy)benzyl)-2-oxa-6-azaspiro[3.3]heptane (50 mg, 0.19 mmol) were mixed in
MeOH (6 mL). NaCN (3.8 mg, 0.08 mmol) was added and the reaction was stirred at rt for 3 h.
The mixture was diluted with DCM (50 mL), the organic layer washed with Na CO (aq), dried
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
2
3
by filtering through a phase separator and evaporated. The crude product was purified by Si flash
column chromatography (first eluting with EtOAc (250 mL) and then eluting the product with
1
DCM : MeOH (2M NH
3
) 20:1) to afford 4 (65 mg, 68%). H NMR (500 MHz, CDCl
3
) δ 8.18 (d,
J = 8.7 Hz, 2H), 7.35 – 7.18 (m, 4H), 6.84 – 6.44 (m, 3H), 5.17 – 5.04 (m, 2H), 4.76 (bs, 4H),
4
.70 – 4.60 (m, 2H), 4.43 – 4.29 (m, 2H), 4.13 – 3.94 (m, 1H), 3.61 (bs, 4H). HRMS calcd for
+
+
C
25
H
25
F
2
N
4
O , [M + H] , 499.1793; found 499.1767.
5
(
3-(4-(2-oxa-6-azaspiro[3.3]heptan-6-ylmethyl)phenoxy)azetidin-1-yl)(5-(4-fluorophenyl)-
1,3,4-oxadiazol-2-yl)methanone (5). Ethyl 5-(4-fluorophenyl)-1,3,4-oxadiazole-2-carboxylate
55 mg, 0.23 mmol) and 6-(4-(azetidin-3-yloxy)benzyl)-2-oxa-6-azaspiro[3.3]heptane (55 mg,
.21 mmol) were mixed in MeOH (6 mL). NaCN (4.1 mg, 0.08 mmol) was added and the
reaction was stirred at rt for 3 h. The mixture was diluted with DCM (50 mL), washed with
Na CO (aq), dried by filtering through a phase separator and evaporated. The crude product was
(
0
2
3
purified by Si flash column chromatography (first eluting with EtOAc and then eluting the
1
product with DCM:MeOH (2M NH
CDCl
H), 4.79 – 4.69 (m, 5H), 4.69 – 4.59 (m, 1H), 4.32 (dd, J = 11.5, 2.5 Hz, 1H), 3.49 (s, 2H), 3.37
3
) 20:1) to afford 5 (75 mg, 79%). H NMR (500 MHz,
3
) δ 8.23 – 8.12 (m, 2H), 7.31 – 7.11 (m, 4H), 6.73 (d, J = 8.5 Hz, 2H), 5.17 – 5.00 (m,
2
+
+
(s, 4H). HRMS calcd for C24
H24FN
4
O , [M+H] , 451.1781; found 451.1773.
4
2
8
ACS Paragon Plus Environment

Page 29 of 59
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
3-(4-(2-oxa-6-azaspiro[3.3]heptan-6-ylmethyl)-3-chlorophenoxy)azetidin-1-yl)(5-phenyl-
1
,3,4-oxadiazol-2-yl)methanone (6).
3-chloro-4-(2-oxa-6-azaspiro[3.3]heptan-6-ylmethyl)phenol (40). 2-Chloro-4-hydroxy-
benzaldehyde (0.5 g, 3.19 mmol) and 2-oxa-6-azaspiro[3.3]heptane (0.55 g, 3.83 mmol) were
mixed in DCM (35 mL). After stirring for 20 min, sodium triacetoxyborohydride (1.02 g, 4.79
mmol) was added and the reaction mixture stirred at rt overnight. The mixture was diluted with
DCM and transferred to a separatory funnel. Water was added and the organic phase was rinsed
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
out (the product is in the water layer). The water phase was saturated with K
2
CO and DCM was
3
added (organic phase is placed above the saturated water layer). The organic layers were
separated and the saturated water phase was extracted two more times with DCM. The organic
1
layers were combined, dried (phase separator) and concentrated to give 40 (0.653 g, 85%). H
NMR (500 MHz, CDCl
3
) δ 7.09 (d, J = 8.4 Hz, 1H), 6.71 (d, J = 2.5 Hz, 1H), 6.55 (dd, J = 2.5,
-
8.4 Hz, 1H), 4.74 (s, 4H), 3.62 (s, 2H), 3.50 (s, 4H). LCMS (ES-) m/z [M - H] : 238.1.
(
3-(4-(2-oxa-6-azaspiro[3.3]heptan-6-ylmethyl)-3-chlorophenoxy)azetidin-1-yl)(5-phenyl-1,3,4-
oxadiazol-2-yl)methanone (6). 40 (0.33 g, 1.38 mmol) was dissolved in dry DMF (10 mL) and
Cs CO (0.90 g, 2.75 mmol) was added. The reaction mixture was stirred for 10 min before 1-(5-
2
3
phenyl-1,3,4-oxadiazole-2-carbonyl)azetidin-3-yl methanesulfonate (0.67 g, 2.07 mol) was
o
added. The reaction mixture was stirred at 90 C for 24 h, then filtered and evaporated. The
residue was purified by preparative HPLC eluting with a gradient of ACN (15-55%) in 0.2%
1
ammonia buffer to afford the product. H-NMR revealed a minor impurity, so the product was
re-purified by silica gel chromatography eluting the product with DCM/MeOH(2M NH ) 20:1
3
1
followed by trituration with diethyl ether to afford 6 (213 mg, 33%). H NMR (500 MHz,
CDCl ) δ 8.21 – 8.10 (m, 2H), 7.65 – 7.49 (m, 3H), 7.28 (d, J = 8.5 Hz, 1H), 6.79 (d, J = 2.5 Hz,
3
2
9
ACS Paragon Plus Environment