Targeting Alzheimer's disease by investigating previously unexplored chemical space surrounding the cholinesterase inhibitor donepezil

Article abstract of DOI:10.1016/j.ejmech.2016.10.036

A series of twenty seven acetylcholinesterase inhibitors, as potential agents for the treatment of Alzheimer's disease, were designed and synthesised based upon previously unexplored chemical space surrounding the molecular skeleton of the drug donepezil, which is currently used for the management of mild to severe Alzheimer's disease. Two series of analogues were prepared, the first looking at the replacement of the piperidine ring in donepezil with different sized saturated N-containing ring systems and the second looking at the introduction of different linkers between the indanone and piperidine rings in donepezil. The most active analogue 5,6-dimethoxy-1-oxo-2,3-dihydro-1H-inden-2-yl 1-benzylpiperidine-4-carboxylate (67) afforded an in vitro IC₅₀value of 0.03 ± 0.07 μM against acetylcholinesterase with no cytotoxicity observed (IC₅₀of >100 μM, SH-SY5Y cell line). In comparison donepezil had an IC₅₀of 0.05 ± 0.06 μM and an observed cytotoxicity IC₅₀of 15.54 ± 1.12 μM. Molecular modelling showed a strong correlation between activity and in silico binding in the active site of acetylcholinesterase.

Full text of DOI:10.1016/j.ejmech.2016.10.036

Accepted Manuscript
Targeting Alzheimer's disease by investigating previously unexplored chemical space
surrounding the cholinesterase inhibitor donepezil
Divan G. van Greunen, Werner Cordier, Margo Nell, Chris van der Westhuyzen,
Vanessa Steenkamp, Jenny-Lee Panayides, Darren L. Riley
PII:
S0223-5234(16)30898-4
10.1016/j.ejmech.2016.10.036
EJMECH 9001
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 23 June 2016
Revised Date: 30 September 2016
Accepted Date: 16 October 2016
Please cite this article as: D.G. van Greunen, W. Cordier, M. Nell, C. van der Westhuyzen, V.
Steenkamp, J.-L. Panayides, D.L. Riley, Targeting Alzheimer's disease by investigating previously
unexplored chemical space surrounding the cholinesterase inhibitor donepezil, European Journal of
Medicinal Chemistry (2016), doi: 10.1016/j.ejmech.2016.10.036.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical abstract
IC (AChE) 0.03 ± 0.07 μM
50
Cytotoxicity IC > 100 μM
50
1

ACCEPTED MANUSCRIPT
Targeting Alzheimer's disease by Investigating Previously Unexplored
Chemical Space Surrounding the Cholinesterase Inhibitor Donepezil
†
§
§
‡
Divan G. van Greunen, Werner Cordier, Margo Nell, Chris van der Westhuyzen, Vanessa
§
‡
,†
Steenkamp, Jenny-Lee Panayides, and Darren L. Riley,*
†
Department of Chemistry, Faculty of Natural- and Agricultural Sciences, University of Pretoria,
Lynnwood Road, Pretoria, South Africa
Pioneering Health Sciences, CSIR Biosciences, Meiring Naudé Road, Pretoria, South Africa
‡
§
Department of Pharmacology, Faculty of Health Sciences, University of Pretoria, Bophelo Road,
Pretoria, South Africa
ABSTRACT
A series of twenty seven acetylcholinesterase inhibitors, as potential agents for the treatment of
Alzheimer’s disease, were designed and synthesised based upon previously unexplored chemical
space surrounding the molecular skeleton of the drug donepezil, which is currently used for the
management of mild to severe Alzheimer’s disease. Two series of analogues were prepared; the first
looking at the replacement of the piperidine ring in donepezil with different sized saturated N-
containing ring systems and the second looking at the introduction of different linkers between the
indanone and piperidine rings in donepezil. The most active analogue 5,6-dimethoxy-1-oxo-2,3-
dihydro-1H-inden-2-yl 1-benzylpiperidine-4-carboxylate (67) afforded an in vitro IC value of 0.03 ±
50
0
.07 μM against acetylcholinesterase with no cytotoxicity observed (IC of >100 μM, SH-SY5Y cell
50
line). In comparison donepezil has an IC of 0.05 ± 0.06 μM and an observed cytotoxicity IC of
50
50
1
5.54 ± 1.12 μM. Molecular modelling showed a strong correlation between activity and in silico
binding in the active site of acetylcholinesterase.
Key Words
Acetylcholinesterase, Alzheimer’s disease, cytotoxicity, donepezil, piperidine
Introduction
Dementia is a general term that may be characterized as a loss of brain function that occurs with or
due to certain diseases. Usually the first symptom appears as forgetfulness. Later symptoms include
difficulty with areas of mental function such as language, memory dysfunction, perception,
[
1, 2]
emotional behaviour or personality and cognitive skills.
One of the most extensively studied
causes of dementia is Alzheimer’s disease (AD). AD is a multifaceted neurodegenerative disorder,
characterized by protein misfolding and aggregation, oxidative stress, mitochondrial abnormalities
[
3, 4]
and neuro-inflammatory processes at a molecular level.

ACCEPTED MANUSCRIPT
Unfortunately to date there is still no definitive consensus on the cause of AD, however several
hypotheses have been proposed. The most significant is the cholinergic hypothesis[5-7] which
postulates that cognitive decline can be linked to a decrease in the amount of the neurotransmitter
acetylcholine (ACh). The amyloid hypothesis links AD to the deposition of β-amyloid (Aβ) plaques in
the brain.[8, 9] The regulation of ACh levels in the brain is mediated by the enzyme
acetylcholinesterase (AChE), which catalyzes the breakdown of the neurotransmitter ACh. To a
lesser extent and at a slower rate ACh is also broken down by butyrylcholinesterase (BuChE).[10] In
addition, AChE has also been shown to bind to Aβ and play a role in the formation of Aβ plaques.[11]
Inhibitors of AChE could potentially improve cholinergic neurotransmission by controlling the
clearance of ACh and reducing the formation of Aβ plaques. To date the US Food and Drug
Administration (US FDA) has three approved AChE inhibitors for the treatment of AD: donepezil 1,
rivastigmine 2 and galantamine 3 (Figure 1). [12] tacrine 4 was formally FDA approved but has since
been discontinued in the US due to safety concerns over hepatotoxicity. [13]
Figure 1. Current and former FDA approved AChE inhibitors for the treatment of AD
Arguably the most successful of these has been the reversible inhibitor donepezil 1, identified by
Sugimoto and co-workers as a new class of AChE inhibitor.[14-17] donepezil 1 does not suffer from
dose-limiting hepatotoxicity,[18-20] has a relative bioavailability of 100% following oral
administration,[21] a considerably longer half-life (70-80 h) relative to other cholinesterase

ACCEPTED MANUSCRIPT
[
17]
inhibitors (0.3 to 12 h)[22] and has less side effects than other drugs on the market. donepezil 1
was developed as a racemic mixture with the two enantiomers readily interconverting in aqueous
solution, via a keto-enol intermediate.[23] Pharmacokinetic studies have shown that S-donepezil has
higher plasma levels than R-donepezil and that the enantiomers declined almost biphasically in
parallel with each other.[24]
Extensive structure-activity relationship (SAR) studies have been conducted on Donepezil 1 and
related piperidine-based AChE inhibitors,[15-17, 25, 26] including the synthesis and biological
evaluation of hybrid inhibitors.[27-29] The chemical space surrounding the N-benzylpiperidine ring
and the methyl linker between the indanone has also received some attention,[25, 26, 29] however
we felt the space still afforded opportunities for investigation. As such in our endeavours to develop
new AChE inhibitors we designed and synthesized a range of donepezil 1 analogues, replacing the
piperidine ring system with a range of different saturated N-benzylated ring systems and replacing
the methyl linker with several different functionalities. Reported in this article is the synthesis of
these analogues, their in vitro biological evaluation for AChE inhibition and cytotoxicity, and in silico
assessment of compound binding to the active site of AChE.
2
2
. Chemistry
.1 Synthesis of different ring sized analogues of donepezil 1
The preparation of analogues of donepezil 1 in which the piperidine ring system was replaced with
various saturated N-benzyl containing ring systems was envisaged through an aldol coupling
between 5,6-dimethoxy-1-indanone 5 with aldehydes (6-13). The aldehydes were prepared
individually using several different synthetic approaches.
Aziridine aldehyde 6 was prepared in four steps,[30, 31] firstly bromination of methyl acrylate 14
with liquid bromine afforded dibromo methyl ester 15 in 99% yield (Scheme 1). Subsequent
condensation with benzylamine in acetonitrile resulted in the preferential formation of unwanted
amide 16, replacement of the solvent with methanol negated the issue and gave aziridine 17 in 97%
yield. Red-Al mediated reduction afforded alcohol 18 in 84% yield followed by Swern oxidation to
the required aldehyde 6 in 93% yield.

ACCEPTED MANUSCRIPT
Scheme 1: (i) 1.1 eq. Br , MeOH, 0°C, 1 h, 99 %, (ii) 1 eq. BnNH , 3 eq. Et N, MeOH, 0 °C to rt, 12 h, 97
2
2
3
%
, (iii) 2 eq. Red-Al, PhMe, 15-20°C, 2 h, 84 %, (iv) 1.2 eq. OxCl , 2.5 eq. DMSO, 5 eq. Et N, DCM, -
2 3
7
8°C, 2 h, 93 %
The 1,2-disubstituted azetidine aldehyde 7 was prepared in four steps from γ-butyrolactone 19
Scheme 4). The lactone 19 underwent a one pot ring-opening, dibromination and esterification on
treatment with bromine and catalytic phosphorus tribromide in methanol, affording dibromo ester
0 in 79% yield. Condensation with benzylamine afforded azetidine 21 in 26% yield,[32] subsequent
(
2
Red-Al reduction gave alcohol 22 in a 97% yield and finally Swern oxidation afforded the desired
aldehyde 6 in 96% yield.
Scheme 2: (i) 0.02 eq. PBr , 1.1 eq. Br , MeOH, 100 °C, 12 h, 79 %, (ii) 1 eq. BnNH , 3 eq. Et N, MeOH,
3
2
2
3
0
5
°C to rt, 12 h, 26 %, (iii) 1.2 eq. Red-Al, PhMe, 15 – 20 °C, 2 h, 97 %, (iv) 1.2 eq. OxCl , 2.5 eq. DMSO,
2
eq. Et N, DCM, -78 °C, 2 h, 96 %
3

ACCEPTED MANUSCRIPT
Preparation of the more challenging 1,3-substituted azetidine aldehyde 8 was achieved in 6 steps
from commercially available diethylmalonate 23 through a modification of the method reported by
Miller and co-workers (Scheme 3).[33] Diethyl malonate 23 was subjected to a double condensation
with formaldehyde to afford diol 24. Conversion of the diols to bromides at this stage followed by
condensation with benzylamine unfortunately results in the mono-displacement of one of the
[
34]
bromine atoms and elimination of the second bromine in the form of HBr. This issue is overcome
by activating the diols as triflates followed by condensation with benzylamine to afford diester 25 in
6
0% crude yield. To necessitate easy purification, the diester 25 was hydrolyzed in methanolic
sodium hydroxide and isolated as the di-sodium salt 26 in 84% yield, after which β-decarboxylation
afforded acid 27 in 48% yield.[35] Lithium aluminium hydride mediated reduction of 27 gave alcohol
2
8 in 85% yield; with a final Swern oxidation affording the aldehyde 8 in 89% yield.
Scheme 3: (i) 3 eq. CH O (20 % soln.), 0.08 eq. KHCO , rt, 18.5 h, 100 % crude, (ii) 2.1 eq. Tf O, 2.5 eq.
2
3
2
Hünig’s base, 1.5 eq. BnNH , AcCN, -20 °C to reflux, 2 h, 60 % crude, (iii) 5 M NaOH-MeOH, MeOH,
2
reflux, 23 h, 84 %, (iv) 6 M HCl, H O, reflux, 7 h, 48 %, (v) 1.5 eq. LiAlH , THF, 0 °C to reflux, 2 h, 85 %,
2
4
(
vi) 1.2 eq. OxCl , 2.5 eq. DMSO, 5 eq. Et N, DCM, -78 °C, 2 h, 89 %
2
3
The 1,2-disubstituted pyrrolidine aldehyde 9 was prepared in 3 steps from L-proline 29 (Scheme
)[36] Benzylation of L-proline 29 with benzylchloride afforded acid 30, subsequent borane
4

ACCEPTED MANUSCRIPT
reduction gave alcohol 31 in 73% yield [37] and finally Swern oxidation yielded aldehyde 9 in 80%
yield.
i
Scheme 4: (i) 4 eq. KOH, 1.5 eq. BnCl, PrOH, 40 °C, 6 h, 60 %, (ii) 3 eq. BH -DMS, THF, rt, 73 %, (iii)
3
1
.2 eq. OxCl , 2.5 eq. DMSO, 5 eq. Et N, DCM, -78 °C, 2 h, 91 %
2 3
The 1,3-disubstituted pyrrolidine aldehyde 10 was accessed in two steps from commercially
available methyl 1-benzyl-5-oxo-3-pyrrolidine carboxylate 32 (Scheme 5). A Red-Al mediated
reduction allowed the reduction of the ester functionality as well as the simultaneous
defunctionalisation of the amide carbonyl group affording alcohol 33 in 57% yield,[38] subsequent
Swern oxidation afforded the required aldehyde 10 in 94% yield.
Scheme 5: (i) 5 eq. Red-Al, PhMe, 15 – 20 °C, 2 h, 57 %, (ii) 1.2 eq. OxCl , 2.5 eq. DMSO, 5 eq. Et N,
2
3
DCM, -78 °C, 2 h, 94 %
Finally the three piperidine based aldehydes 11-13 were prepared using a common approach
Scheme 6). Benzylated nipecotate’s 34-36 were subjected to Red-Al reduction giving alcohols 37-39
(
in moderate yields (64-75%) and subsequent Swern oxidation afforded the desired aldehydes 11-13
in excellent yields (94-95%).

ACCEPTED MANUSCRIPT
Scheme 6: i) 2.5 eq. Red-Al, PhMe, 15 – 20 °C, 2 h, (ii) 1.2 eq. OxCl , 2.5 eq. DMSO, 5 eq. Et N, DCM, -
2
3
7
8 °C, 2 h.
The eight ring systems prepared were coupled to 5,6-dimethoxy-1-indanone 5 under aldol
condensation conditions affording compounds 40-47 with a methylene linker in moderate to good
yields (37-80%) with the exception of aziridine 48 the formation of which although observed by TLC
and crude NMR decomposed rapidly on standing. Subsequent catalytic hydrogenation with 10%
palladium on carbon afforded methyl coupled analogues 1, 49-54 in yields ranging from 16-99%
(
Scheme 7). Aziridine 48 unfortunately decomposed rapidly on standing and could not be isolated as
a pure derivative.
Scheme 7: (i) Aldehyde, 7 eq. K CO , MeOH, 2.5h, (ii) H (g), 10% Pd/C, MeOH, rt, 12 h
2
3
2
2
.2 Synthesis of donepezil 1 analogues with an ester-bridging linker
Ester linked derivatives 55-63 were obtained in two steps from 5,6-dimethoxy-1-indanone 5 (Scheme
). The indanone 5 was carboxylated by treatment with dimethyl carbonate to afford 64 in 81% yield,
8
subsequent Amberlyst-15 catalysed transesterification[39] with alcohols 17, 22, 28, 31, 33 and 37-39
afforded esters 53-60 in low to good yields (18%-82%). In the case of ester 61 the required alcohol
was prepared by reduction of commercially available 1-benzylpiperidine-4-one with Red-Al in 96%
yield (not shown).

ACCEPTED MANUSCRIPT
Scheme 8: (i) CO(CH ) , THF, reflux, 2 h, 81%, (ii) R-OH, 10 % (w/w) Amberlyst-15, toluene, reflux, 3 h
3
2
2
.3 Synthesis of donepezil 1 analogue with a reverse ester-bridging linker
The synthesis of a donepezil 1 analogue with an ester linker reversed with the carbonyl bound to the
piperidine ring was achieved in two steps (Scheme 9). 5,6-Dimethoxy-1-indanone 5 was brominated
affording the α-brominated indanone 64 in 81% yield, subsequent displacement of the halogen with
1
-benzylpiperidine-4-carboxylic acid 66 catalyzed by silver oxide afforded the reverse ester 67 in 85%
yield.
Scheme 9: (i) 1.05 eq. Br , MeOH, rt, 10 min, 81%, (ii) 66, Ag O, CH CN, 60 °C, 12 h, 85%
2
2
3
2
.4 Synthesis donepezil 1 analogue with a ketone-bridging linker
The preparation of a donepezil 1 analogue containing a ketone linker was achieved in two steps from
acid 66 (Scheme 10). Carboxylic acid 66 was treated with carbonyl diimidazole affording the
imidazole derivative 68, subsequent treatment with 5,6-dimethoxy-1-indanone 5 in the presence of
the methylsulfinyl carbanion[40] afforded diketone 69 in an 8% yield.
Scheme 10: (i) 1.2 eq. carbonyl diimidazole, CH Cl , 0 °C, 2 h, 93%, (ii) 4.5 eq. NaH, 1.3 eq. 5, DMSO,
2
2
rt, 5 h, 8%

ACCEPTED MANUSCRIPT
3
. Structure Activity Relationship Study
The inhibitory activity of all new synthetic compounds was assessed against AChE (type VI-S from
[
41]
[42]
electric eel) using Ellman’s spectrophotometric method with some minor modifications. Table 4
shows the anticholinesterase activities for the synthesised compounds expressed as half-maximal
inhibitory concentration (IC ) values. In addition selected compounds were assessed for cytotoxicity
50
in the SH-SY5Y neuroblastoma cell line using the sulforhodamine B (SRB) staining as described by
Vichai and Kirtikarahai[43] It should be noted that all analogues were prepared and screened as
racemic mixtures and that the biological results described are representative of the additive effect of
all the diastereomers for each analogue.
All compounds prepared showed better AChE activity than galantamine 3 (3.57 μM), with one
compound 67 showing activity (0.03 μM) comparable to that of donepezil (0.05 μM)(Table 1). The
results show good AChE inhibition for most of the compounds designed to explore the chemical
space around the piperidine ring system of donepezil 1 (40-54) in the presence of either a methylene
(
40-47) or a methyl linker (48-54). Exceptions were the aziridine ring system (40) and the 1,2-
disubstituted azetidine systems (41 and 49). In general, analogues with a methyl linker displayed
better activity in comparison to those with a methylene linker. The results suggest that the 1,4-
disubstituted piperidine ring system of donepezil 1 clearly affords the best inhibition. The
introduction of an ester linker (55-63) resulted in a complete loss of activity in all cases except 63, in
which the ester was directly linked to the 1,4-disubstituted piperidine ring system as opposed to via a
methyl linker. Interestingly, when the orientation of the ester linker in compound 62 was reversed
(
67) the activity increased to 0.03 μM which is comparable to that of donepezil 1 (0.05 μM).
Introduction of a carbonyl linker (69) also afforded good activity (0.66 μM). The results indicate that
the benzyl piperidine ring system in general affords good activity and the linker between the
indanone and piperidine rings plays an important role in AChE inhibition with subtle change in the
linkers resulting in dramatic changes in the ability to inhibit AChE in vitro.
Cytotoxic evaluation showed that in all cases IC50 values were above 1 μM, with compound 42
displaying the most potent cytotoxicity of the compounds screened (IC 1.26 μM). Compounds with
50
methylene linkers (42, 46 and 47) displayed higher cytotoxicity than the analogous compounds with
a simple methyl linker (50, 54 and 1). In the reduced derivatives, cytotoxicity seemed to increase as
the molecule became more linear in nature around the nitrogen containing ring system, shown by
azetidines 49 and 50 (IC values of 58.5 and 8.01 μM respectively) and piperidines 53, 54 and 1
50

ACCEPTED MANUSCRIPT
(
cytotoxicity IC values of >100, 66.74 and 15.54 μM respectively). Overall four compounds (53, 54,
50
6
1
7 and 69) showed AChE inhibition below 1 μM and less cytotoxicity than that of donepezil 1 (IC50
5.54 μM). Compound 67, with the reverse ester linker, displayed the highest AChE inhibition (0.03
μM) and low cytotoxicity with an IC value > 100 μM.
50
Table 1. In vitro AChE inhibitory activity, cytotoxicity and selectivity index of compounds 1 and 40-47
and 49-69.
O
MeO
A
B
MeO
a
Compound
Linker A
Ring B
AChE
IC ± SD (µM)
Cytotoxicity
IC50 ± SEM (µM)
a
c
50
[44]
4
4
4
0
1
2
=CH-
=CH-
=CH-
>
>
10
10
-
-
0
>
0
.70 ± 0.12
10
1.26 ± 1.08
18.98 ± 1.02
6.59 ± 1.14
4
4
3
4
=CH-
=CH-
.52 ± 0.15
N
Bn
4
4
5
6
=CH-
=CH-
0
1
.75 ± 0.14
.18 ± 0.08
-
4.72 ± 1.26
4
4
7
9
=CH-
-CH2-
0
>
.42 ± 0.11
10
3.99 ± 1.13
58.5 ± 1.18
5
0
-CH2-
1
.13 ± 0.05
8.01 ± 1.09

ACCEPTED MANUSCRIPT
5
5
1
2
-CH2-
-CH2-
3
1
.16 ± 0.07
.74 ± 0.06
-
-
N
Bn
5
5
3
4
-CH2-
-CH2-
0
0
.46 ± 0.08
.51 ± 0.16
>100
66.74 ± 1.13
5
5
>
>
10
10
-
-
5
5
6
7
>
10
>100
5
5
8
9
>
>
10
10
-
-
6
6
0
1
>
>
10
10
-
-
6
6
6
6
2
3
7
9
>
3
0
0
10
-
.03 ± 0.05
.03 ± 0.07
.66 ± 0.07
>100
>100
63.89 ± 1.13
donepezil 1
galantamine 3
0.05 ± 0.06
3.57 ± 1.42
15.54 ± 1.12
-
a
Compounds 1, 40-54, 67 and 69 were screened as their respective HCl salts[45]

ACCEPTED MANUSCRIPT
4
. Molecular Modelling
In an attempt to rationalise the SAR results, compounds 1, 40 – 63, 67 and 69 were docked into the
binding site of both human and torpedo eel acetylcholine esterase (pdb accession codes 4EY7 and
1
EVE, respectively) using CDOCKER (BIOVIA Discovery Studio 2016). The former was included for
comparison purposes only, with the latter being identical to the protein used in the assays performed
in this work. The crystal structures both contain the anti-AD compound donepezil 1 co-crystallised
with the protein, both in the R-configuration at the carbon atom alpha to the indanone carbonyl
group. In both instances, donepezil 1 was docked back into the structures to gauge the robustness of
the modelling performed (Section 6.3). It was found that a number of water molecules had to be
included in the protein binding site to obtain the correctly docked pose of donepezil 1, and as a
result the hydrogen-bonded water networks already existing within the binding sites of the crystal
structures were left intact in subsequent modelling processes.
From the docking results obtained using the torpedo eel AChE (1EVE), the strongest binding
contributions for the donepezil analogues stemmed from a series of strong π-π interactions between
the benzyl groups in the inhibitors with Trp84, Trp279 and His440, as well as a series of π-alkyl
interactions between the phenyl moieties of Phe330 and Tyr334 and the methylene groups of the
nitrogen heterocycle. In addition, the top binding materials all showed clear interactions with the
following water molecules in the binding site: HOH1160 (π-donor π-interaction with the benzylic
phenyl system), HOH1159 (hydrogen bonding to the nitrogen of the heterocycles), HOH1254
(
hydrogen bonding to the indanone carbonyl group) and HOH1234 (hydrogen bonding to the
methylene groups of the heterocyclic system). A stabilising interaction between the carbonyl group
of Ser286 and a methoxy group of the indanone was also present in many poses. These interactions
are summarised in Figure 2a for the strongest inhibitor tested, ester 67.

ACCEPTED MANUSCRIPT
Figure 2a: 2D interaction map of the docked pose of ester 67 in torpedo eel AChE (1EVE) showing
primary interactions detected with both adjacent amino acid residues and hydrogen-bonded water
molecules in the active site.
A three dimensional overlay of the best pose of the docked ester 67 with donepezil 1 (Figure 2b)
clearly indicates that ester 67 prefers an R-conformation alpha to the indanone carbonyl group
identical to that found in the bound form of donepezil 1 in the crystal structure. All the required
interactions are satisfied in this pose. The ester carbonyl projects into an empty pocket, allowing the
piperidine ring to occupy a chair conformation similar to that for donepezil 1. This pose, while nearly
identical to that of 1, seems to allow better interactions of the indanone moiety with the adjacent
residues due to a slight lengthening of the linker chain – measured from the alpha carbon of the
indanone moiety to C1 of the benzylic phenyl ring, from 7.52Å for 1 to 8.15Å for 67 – thus improving
the overall binding interaction. The S-conformation of 67 adopted a pose in which the indanone
system ‘flipped’ to form a weaker hydrogen bond to a water molecule in the active site (not shown),
along with correspondingly poorer π-interactions and poorer apparent binding overall. Noteworthy
for later comparison is the clear CH/π-interaction between the piperidine ring and the Phe330
phenyl moiety observed in both the R- and S-conformers of docked ester 67, along with the
hydrogen bond between the piperidine nitrogen and HOH1159 “above” the inhibitor as shown.

ACCEPTED MANUSCRIPT
Figure 2b: Overlay of the docked poses of ester 67 (white) and donepezil 1 (green) in torpedo eel
AChE (1EVE). The π-π stacking interactions with Trp84, His440 and Trp279, and the π-alkyl interaction
with Tyr334 are indicated in dark and light pink, respectively. Hydrogen bonds between water and CH
groups (dull blue) and C=O moieties (pale blue), as well as carbonyl/CH bonding interactions (pale
green) are indicated for 67 only. Selected water molecules (red) are represented by the oxygen atoms
only, and protein structures not directly involved in the binding processes have been hidden for
clarity.
In the subset of molecules having an alkenyl moiety (40 – 47), the increased rigidity of the
dihydroindanone system limits the ability of the linking cyclic system to accommodate the residues of
the active site. In the case of unsaturated donepezil precursor 47, this resulted in a shortening of the
accommodated linker length to 7.17 Å (compared to 7.52 Å for 1). In this instance, only the Z-isomer
could be docked (Figure 3). Similar consequences are observable with the remainder of the series (40
–
46), although the larger rings and better positioning of the benzyl group towards the opposite side

ACCEPTED MANUSCRIPT
of the ring relative to the indanone allows for substantially improved interactions with the protein
residues as is seen with 42 and 44, although these materials still displayed poorer activity relative to
donepezil 1. In all these latter cases, the linker length fell in the 5.9 Å – 7.1Å range.
Figure 3: Overlay of the docked poses of the Z-isomer of 47 (white) and donepezil 1 (green) in
torpedo eel AChE (1EVE). The π-π stacking interactions with His440 and Trp279, and the π-alkyl
interaction with Phe330 and Tyr334 are indicated in dark and light pink, respectively. Hydrogen
bonds between water and CH groups (dull blue) as well as carbonyl/CH bonding interactions (pale
green) are indicated for 47 only. Selected water molecules (red) are represented by the oxygen atoms
only, and protein structures not directly involved in the binding processes have been hidden for
clarity.

ACCEPTED MANUSCRIPT
In a similar manner, β-ketoesters 55 – 63 and β-diketone 69 are likely to suffer a similar penalty due
to enolisation of the ketone, imparting additional rigidity to the entire ketoester moiety by the
formation of a six-membered hydrogen-bonded ring between the two carbonyl groups. This
hydrogen bonded ring set was not modelled, but the activity trend points to this conclusion, with the
activity of 53 and 54 being similar to 47 for the reasons described above, for example.
The modelling also reinforced the preference of the compounds to adopt an orientation
corresponding to the R-isomer of donepezil 1 at α-C of the indanone skeleton. While all the docking
as well as biological evaluations were done on an ensemble of isomers, invariably the best docked
pose of each inhibitor having this stereocentre adopted an orientation in the active site
corresponding to donepezil 1.
While the biological tests were not evaluated on human AChE, docking studies performed using the
same protocols on the human protein (4EY7) indicated the same general trends were followed as
seen for the torpedo eel results, with ester 67 affording the most energetically favoured poses of the
data set. However, one notable difference between the poses obtained from the docking into human
AChE and torpedo eel AChE for each inhibitor ensemble was evident: the orientation of the central
piperidine ring. An overlay of the best docking pose of 67 in either enzyme indicates that the
observed “ring flip” of the piperidine chair conformation, observed in the original crystal structure
with donepezil 1, [46] is conserved here (Figure 4a). The only observable difference in the binding
site stems from the presence of Tyr337 in the human protein that replaces Phe330 in the torpedo eel
site. The residue forms a hydrogen-bonded network with HOH951 “beneath” the docked inhibitor in
Figure 4a, enabling a hydrogen bond to form with the piperidinyl nitrogen – thus enforcing the
flipped chair conformation. (Note that the same interaction for the torpedo eel system was satisfied
by HOH1159 in a hydrogen-bonded water network “above” the docked inhibitor, with a CH/π-
interaction to Phe330 stabilising the piperidinyl system – see Figure 2b). Interestingly, this bonding
interaction enforces an axial conformation of the ester substituent attached to the piperidinyl moiety
in the human protein. The result is hydrogen-bonding to the site’s water network by both the ester
and indanone carbonyl groups, as well as a “flipped” pose of the indanone system relative to
donepezil 1 in the native crystal structure (Figure 4b).

ACCEPTED MANUSCRIPT
Figure 4a: Overlay of the poses obtained for the dockings of ester 67 in human AChE (white) and
torpedo eel AChE (green) with the human AChE (4EY7) binding site shown. Binding interactions for
the docking with human AChE are shown (key as for Figure 2b).

ACCEPTED MANUSCRIPT
Figure 4b: Overlay of the docked poses of 67 (white) and donepezil 1 (green) in human AChE (4EY7).
The π-π stacking interactions with Try86, Trp286 and Tyr341, and the π-alkyl interaction with Tyr337,
Phe338 and Tyr341 are indicated in dark and light pink, respectively. Hydrogen bonds (blue) are
indicated for 67 only. Selected water molecules (red) are represented by the oxygen atoms only, and
protein structures not directly involved in the binding processes have been hidden for clarity.
5
. Conclusion
Two libraries of potential AChE inhibitors were designed and prepared based upon the molecular
skeleton of the AChE inhibitor donepezil 1. Analogues were prepared by either changing the size of
the piperidine ring and/or altering the methyl linker between the indanone and piperidine ring
system found in donepezil 1. Structure activity relationships revealed that the piperidine ring system
afforded the best activities. Introduction of an alkene linker reduced activity and replacement of the

ACCEPTED MANUSCRIPT
linker with an ester moiety bound to the indanone via the carbonyl group resulted in a complete loss
of activity in all but one case. Interestingly reversal of the ester to afford binding to the indanone via
the oxygen atom improved activity to the range of that of donepezil 1. Cytotoxic assessment
indicated that the more flexible systems with a methyl or ester linker were less cytotoxic than those
with a methylene linker. The most active molecule in the series, reverse ester 67, displayed
nanomolar activity and no observable cytotoxicity. It should be noted, however, that the ester
functionality could pose a metabolic liability and current work is focused on addressing this issue.
Molecular modelling correlated well with the observed biological activities and indicated that the
primary interactions stem from a series of strong π-π interactions between the benzyl groups in the
inhibitors with Trp84, Trp279 and His440, as well as a series of π-alkyl interactions between the
phenyl moieties of Phe330 and Tyr334 and the methylene groups of the nitrogen heterocycle. The
distance between the indanone α-C and the phenyl ring was shown to be of critical importance and
could be used to rationalise why the more rigid systems, including those with the methylene and
enolisable carbonyls which imparted a “shorter” distance, showed less activity than the
corresponding less rigid systems with methyl linkers. Finally, molecular modelling also suggested that
the (R)-isomer at the α-carbon of the indanone ring shows more favourable docking interactions, and
although in biological systems this position is readily epimerised, development of non-epimerisable
analogues with the same relative orientation could allow for the preparation of more potent
inhibitors.
6
6
6
. Experimental
.1 Chemistry
.1.1 General methods
All solvents, chemicals, and reagents were obtained commercially and used without further
1
13
purification. H NMR (300 MHz) and C NMR (75 MHz) spectra were recorded on Bruker AVANCE-III-
00 instrument using CDCl , DMSO-d or D O as solvent. CDCl contained tetramethylsilane as an
3
3
6
2
3
internal standard. Chemical shifts, δ, are reported in parts per million (ppm), and splitting patterns
are given as singlet (s), doublet (d), triplet (t), quartet (q), or multiplet (m). Coupling constants, J, are
expressed in hertz (Hz). Mass spectra were recorded in ESI mode on a Waters Synapt G2 Mass
Spectrometer at 70 eV and 200 mA. Samples were made up in acetonitrile (containing 0.1 % formic
acid) to an approximate concentration of 10 μg/ml. Infrared spectra were run on a Bruker ALPHA

ACCEPTED MANUSCRIPT
-
1
Platinum ATR spectrometer. The absorptions are reported on the wavenumber (cm ) scale, in the
-
1
range 400-4000 cm . Melting points were measured on a Stuart Melting Point SMP10 microscope
and are uncorrected. The retention factor (Rf) values quoted are for thin layer chromatography (TLC)
on aluminium-backed Macherey-Nagel ALUGRAM Sil G/UV254 plates pre-coated with 0.25 mm silica
gel 60, spots were visualised with UV light and basic KMnO4 spray reagent. Chromatographic
separations were performed on Macherey-Nagel Silica gel 60 (particle size 0.063 – 0.200 mm). Yields
refer to isolated pure products unless stated otherwise.
6
.1.2 Methyl 2,3-dibromopropanoate 15
To a solution of methyl acrylate 14 (25.00 g, 290.4 mmol, 1.0 eq.) in dichloromethane (80 ml) was
added bromine (16.50 ml, 319.4 mmol, 1.1 eq.) at 0 °C. The resulting solution was allowed to stir at 0
°
C for 1 h. The reaction was quenched with a saturated sodium thiosulphate solution (50 ml). The
resulting suspension was allowed to stir for 1 h at rt. The mixture was diluted with dichloromethane
50 ml) and the organic layer separated. The aqueous layer was again extracted with
dichloromethane (3 x 50 ml). The organic layers were combined, washed with brine, dried (Na SO ),
(
2
4
filtered and evaporated in vacuo. The crude product was purified by column chromatography (25%
ethyl acetate: hexane) on silica gel gave the product, methyl 2,3-dibromopropanoate 15, as yellow
-
1
1
oil (71.01 g, 99%). Rf 0.24 (hexane); νmax (neat)/cm 1742, 1355, 1247, 1148, 835; H NMR (300
MHz, CDCl ); δ 4.44 (dd, 1H, J = 4.4 & 11.3 Hz, CHBr), 3.92 (dd, 1H, J = 9.9 & 11.3 Hz, CH Br), 3.84 (s,
3
2a
1
3
3
2
H, OCH ), 3.67 (dd, 1H, J = 4.4 & 9.9 Hz, CH Br); C NMR (75 MHz, CDCl ); δ 168.20, 53.54, 40.86,
3 2b 3
9.74.
6
.1.3 Methyl 1-benzylaziridine-2-carboxylate 17
A solution of methyl 2,3-dibromopropanoate 15 at 0°C (5.01 g, 20.4 mmol, 1.0 eq.) in methanol (10
ml) was added drop-wise to a stirred solution of triethylamine (6.18 g, 8.47 ml, 61.1 mmol, 3.0 eq.)
in methanol (40 ml) and stirred for 15 min. Benzylamine (2.18 g, 2.22 ml, 20.4 mmol, 1.0 eq.) in
methanol (40 ml) was then added drop-wise and the reaction mixture was stirred for a further 30
min after which time the solution was left to warm to room temperature and stirred overnight. The
reaction mixture was evaporated in vacuo, diluted with diethyl ether (100 ml), washed with water
(
200 ml). The aqueous phase was extracted with diethyl ether (3 x 50 ml). The combined organic
phases were dried (Na SO ), filtered, and evaporated in vacuo. The resulting material was purified by
2
4
column chromatography (25 % ethyl acetate: hexane) to afford methyl 1-benzylaziridine-2-

ACCEPTED MANUSCRIPT
-
1
carboxylate 17 as a yellow oil (3.79 g, 97 %). Rf 0.27 (25% ethyl acetate: hexane); νmax (neat)/cm
1
1
3
737, 1442, 1180, 1028; H NMR (300 MHz, CDCl ); δ 7.39–7.13 (m, 5H, ArH’s), 3.70 (s, 3H, OCH ),
3
3
.53 (s, 2H, CH Ph), 2.25 (dd, 1H, J = 1.1 & 3.1 Hz, CHN), 2.20 (dd, 1H, J = 3.1 & 6.4 Hz, CHCH N), 1.74
2
2a
1
3
(
dd, 1H, J = 1.0 & 6.4 Hz, CHCH N); C NMR (75 MHz, CDCl ); δ 174.40, 138.04, 128.49, 128.12,
2b 3
1
27.42, 63.98, 52.22, 37.40, 34.57; HRMS m/z (ESI) 192.1058 ([M + H]+ requires 192.1026).
6
.1.4 Methyl 2,4-dibromobutanoate 20
Phosphorous tribromide (0.61 g, 0.21 ml, 2.17 mmol, 0.02 eq.) was added to a stirring solution of γ-
butyrolactone 19 (10.1 g, 9.00 ml, 118 mmol, 1.0 eq.) and heated to 100 °C. To solution was added
bromine (20.5 g, 6.60 ml, 130 mmol, 1.1 eq.) in a drop-wise manner. The reaction was cooled to rt
and then to 0 °C. Methanol (48 ml) was added, followed by concentrated hydrochloric acid until pH
1
. The reaction was stirred overnight at rt. A solution of saturated sodium thiosulfate was added
until the red solution turned clear. The aqueous layer was extracted with diethyl ether (3 x 50 ml).
The combined organic layers were combined, dried (Na SO ), filtered and evaporated in vacuo to
2
4
afford the crude product as a transparent orange oil. The product was purified by column
chromatography (25% ethyl acetate: hexane) to afford methyl 2,4-dibromobutanoate 20 as a yellow
-
1
1
oil (24.31 g, 79 %). Rf 0.86 (25% ethyl acetate: hexane); νmax (neat)/cm 2955 , 1737, 1274, 1157; H
NMR (300 MHz, CDCl ); δ 4.52 (dd, 1H, J = 6.1 & 8.1Hz, CHBr), 3.80 (s, 3H, OCH ), 3.56–3.50 (m, 2H,
3
3
1
3
CH Br), 2.66–2.40 (m, 2H, -CH -); C NMR (75 MHz, CDCl ); δ 169.69, 53.33, 43.51, 37.08, 29.83.
2
2
3
6
.1.5 Methyl 1-benzylazetidine-2-carboxylate 21
A solution of methyl 2,4-dibromobutanoate 20 (10.01 g, 38.49 mmol, 1.0 eq.) in methanol (15 ml)
was added drop-wise to a stirred solution of triethylamine (11.68 g, 16.01 ml, 115.5 mmol, 3.0 eq.) in
methanol (40 ml) at 0°C. The solution was stirred for 15 min after which time benzylamine (4.33 g,
4
.41 ml, 40.4 mmol, 1.05 eq.) in methanol (25 ml) was added in a drop-wise manner and stirring was
continued for a further 30 min after which time the solution was left to warm to rt and finally left to
stir overnight. The reaction mixture was concentrated in vacuo and then diluted with diethyl ether
(
(
100 ml). The resulting solids were removed by filtration. The organic phase was washed with water
200 ml) and the aqueous phase was extracted with diethyl ether (3 x 50 ml). The combined organic
phases were dried (Na SO ), filtered and evaporated in vacuo. The crude material was purified by
2
4
column chromatography (25% ethyl acetate: hexane) to afford methyl 1-benzylazetidine-2-
-
1
carboxylate 21 as an orange oil (2.02 g, 26 %). Rf 0.35 (1:3 ethyl acetate: hexane); νmax (neat)/cm

ACCEPTED MANUSCRIPT
1
2
3
2
953, 2837, 1737, 1442 , 1201 , 1172 ; H NMR (300 MHz, CDCl ); δ 7.35–7.17 (m, 5H, ArH’s), 3.84–
3
.70 (m, 2H, CH Ph), 3.62 (s, 3H, OCH ), 3.58 (d, 1H, J = 12.5 Hz, COCH), 3.35–3.27 (m, 1H, NCH ),
2
3
2a
13
.98–2.88 (m, 1H, NCH ), 2.36 (dt, 1H, J = 8.7 & 17.6 Hz, CHCH ), 2.25–2.13 (m, 1H, CHCH ); C
2b
2a
2b
NMR (75 MHz, CDCl ); δ 173.10, 137.08, 129.17, 128.36, 127.32, 64.40, 62.45, 51.83, 50.85, 21.72;
3
HRMS m/z (ESI) 206.1198 ([M + H]+ requires 206.1175).
6
.1.6 Diethyl 2,2-bis(hydroxymethyl)malonate 24
A 35 % aqueous formaldehyde solution (16.07 ml, 187.3 mmol, 3.0 eq.) and potassium bicarbonate
0.53 g, 5.27 mmol, 0.08 eq.) were placed in a water-bath at 20°C. Diethylmalonate 23 (10.00 g,
2.43 mmol, 1.0 eq.) was added drop-wise and left to stir overnight. A saturated solution of sodium
sulfate was added and the product was extracted with diethyl ether. The organic layer was dried
Na SO ), filtered and evaporated in vacuo to afford the crude product, diethyl 2,2-
(
6
(
2
4
bis(hydroxymethyl)malonate 24 as a transparent colourless oil (13.79 g, 100%). The crude material
-
1
was used as is. Rf 0.31 (50% ethyl acetate: hexane); νmax (neat)/cm 3433, 2982, 1790, 1209, 1028;
1
H NMR (300 MHz, CDCl ); δ 4.22 (2 x q, 4H, J = 6.7 Hz, 2 x OCH CH ), 4.09 – 4.04 (m, 4 H, 2 x CH OH),
3
2
3
2
1
3
2
.99 (s, 2H, 2 x OH), 1.25 (t, 3H, J = 7.1 Hz, OCH CH ) 1.24 (t, 3H, J = 7.1 Hz, OCH CH ); C NMR (75
2 3 2 3
MHz, CDCl ); δ 169.58, 63.54, 61.97, 61.21, 14.04.
3
6
.1.7 Diethyl 1-benzylazetidine-3,3-dicarboxylate 25
A solution of diethyl 2,2-bis(hydroxymethyl)malonate 24 (7.01 g, 31.83 mmol, 1.0 eq.) in anhydrous
acetonitrile (105 ml) was cooled to -15 °C. Triflic anhydride (11.25 ml, 66.84 mmol, 2.1 eq.) was
added drop-wise while keeping the temperature below -10 °C. N,N-Diisopropylethylamine (27.72 ml,
1
59.2 mmol, 5.0 eq.) was then added drop-wise while maintaining the temperature below -10 °C.
After the addition the temperature was raised to rt. Benzylamine (5.22 ml, 47.8 mmol, 1.5 eq.) was
added in a single portion. The resulting mixture was heated to 70 °C and aged for 3 h. The reaction
mixture was cooled to ambient temperature, diluted with toluene (40 ml) and washed with brine (2
x 500 ml). The separated organic layer was dried over anhydrous sodium sulphate. The solvent was
evaporated in vacuo to afford the crude product, diethyl 1-benzylazetidine-3,3-dicarboxylate 25 as a
dark brown oil (5.59 g, 60 %) which was used directly in the next step. Rf 0.80 (25% hexane:ethyl
-
1
1
acetate); νmax (neat)/cm 2981, 1729, 1270, 1154, 1013, 857, 700; H NMR (300 MHz, CDCl ); δ
3
7
6
.34–7.20 (m, 5H, ArH’s), 4.25 (q, 2H, J = 7.1 Hz, OCH CH ), 4.23 (q, 2H, J = 7.1 Hz, OCH CH ), 3.66 (s,
2 3 2 3
13
H, CH Ph 2 x CH N), 1.28 (t, 3H, J = 7.1 Hz, OCH CH ), 1.26 (t, 3H, J = 7.1 Hz, OCH CH ); C NMR (75
2
2
2
3
2
3

ACCEPTED MANUSCRIPT
MHz, CDCl ); δ 169.93, 137.22, 128.66, 128.52, 127.38, 62.91, 62.00, 59.65, 49.61, 14.12; HRMS m/z
3
(
ESI) 292.1632 ([M + H]+ requires 292.1543).
6
.1.8 Sodium 1-benzylazetidine-3,3-dicarboxylate 26
To a solution of the diester 25 (3.99 g, 13.69 mmol, 1.0 eq.) in methanol (30 ml) was added 5 M
methanolic sodium hydroxide (6.7 ml) in one portion at rt. The mixture was heated to 50 °C and
aged over 45 min. The resultant slurry was cooled to room temperature and sampled for assay,
which showed incomplete hydrolysis, an additional portion of 5 M methanolic sodium hydroxide (2
ml) was added. The slurry was aged at 50 °C for an additional 30 min, cooled to rt and filtered. The
solid was washed with methanol (3 x 30 ml). The wet solid was diluted with methanol (30 ml) and
stirred overnight at rt. The slurry was then filtered and washed with methanol (2 x 20 ml). The
product was vacuum dried overnight to afford sodium 1-benzylazetidine-3,3-dicarboxylate 26 as an
off-white solid (3.22 g, 84 %). The salt was used without additional purification. mp > 300 °C; νmax
-
1
1
(
neat)/cm 1610, 1432, 1332; H NMR (300 MHz, D O); δ 7.49 – 7.20 (m, 5H, ArH’s), 3.68 (s, 2H,
2
1
3
CH Ph), 3.62 (s, 4H, 2 x CH N); C NMR (75 MHz, D O); δ 179.04, 136.79, 128.91, 128.63, 127.59,
2
2
2
6
1.89, 60.86, 53.46.
6
.1.9 1-Benzylazetidine-3-carboxylic acid 27
To a solution of sodium 1-benzylazetidine-3,3-dicarboxylate 26 (4.47 g, 16.01 mmol, 1.0 eq.) in water
17.6 ml), was added concentrated aqueous hydrochloric acid (4.00 ml, 48.03 mmol, 3.0 eq.) at rt.
(
The resulting suspension was heated to reflux for 7 h. The mixture was cooled to rt and solid sodium
bicarbonate was added to adjust the pH to 5-6. The water removed in vacuo until the point where
solid precipitation of the monoacid and the inorganic salts was observed. The mixture was filtered
through a sintered glass funnel and washed with ethanol. The solvent was filtered off, dried
(
Na SO ), filtered and evaporated in vacuo to afford 1-benzylazetidine-3-carboxylic acid 27, as an off-
2 4
white solid (1.46 g, 48 %) which was used as is without further purification. Rf 0.14 (1:1
-
1
1
dichloromethane: methanol); mp 132–133 °C; νmax (neat)/cm 3395, 1577, 1358, 1244, 962, 833; H
NMR (300 MHz, D O); δ 7.53–7.43 (m, 5H, Ar-H’s), 4.38 (d, 2H, J = 12.7 Hz, CH Ph), 4.21 (2 x dt, 4H, J
2
2
1
3
=
1
6.5 & 11.1, 2 x CH N), 3.52 (p, 1H, J = 9.1 Hz, COCH); C NMR (75 MHz, D O); δ 177.23, 129.96,
2 2
29.86, 129.36, 68.37, 56.39, 34.39; HRMS m/z (ESI) 192.1048 ([M + H]+ requires 192.1019).
.1.10 (1-Benzylazetidin-3-yl)methanol 28
6

ACCEPTED MANUSCRIPT
To a stirring slurry of lithium aluminium hydride (0.34 g, 8.42 mmol, 1.5 eq., 95 %) in dry
tetrahydrofuran (25 ml) at 0°C, was added 1-benzylazetidine-3-carboxylic acid 27 (1.05 g, 5.49 mmol,
1
.0 eq.) in a single portion. The mixture was stirred for 30 min at 0 °C after which time the solution
was heated to reflux for 2 h. The reaction mixture was then left to cool to rt and then quenched with
a 3 M aqueous sodium hydroxide solution. The solvent was evaporated in vacuo and
dichloromethane (30 ml) was added to the residue. The organic layer was washed with 10 %
aqueous sodium hydroxide (3 x 25 ml). The organic layer was separated, dried (Na SO ), filtered and
2
4
evaporated in vacuo to afford (1-benzylazetidin-3-yl)methanol 28, as a viscous clear oil (0.83 g, 85
). The product was used as is without any further purification. Rf 0.47 (20%
%
-
1
1
methanol:dichloromethane); νmax (neat)/cm 3323, 1450. 1031, 740, 697; H NMR (300 MHz, CDCl );
3
δ 7.35–7.22 (m, 5H, ArH’s), 3.76 (d, 2H, J = 5.7 Hz, CH OH), 3.60 (s, 2H, CH Ph), 3.35 (t, 2H, J = 7.6 Hz,
2
2
1
3
2
x CH N), 3.09 (dd, 2H, J = 5.5 & 7.4 Hz, 2 x CH N), 3.00 (s, 1H, OH), 2.70–2.57 (m, 1H, CH); C NMR
2a 2b
(
75 MHz, CDCl ); δ 137.96, 128.57, 128.44, 127.17, 64.59, 63.49, 56.99, 32.80; HRMS m/z (ESI)
3
3
55.2386 ([2M + Na]+ requires 355.2380).
6
.1.11 (S)-1-Benzylpyrrolidine-2-carboxylic acid 30
A solution of L-proline 29 (5.00 g, 43.44 mmol, 1.0 eq.) and potassium hydroxide (9.78 g, 174. 3
mmol, 4 eq.) in iso-propanol (50 ml) was stirred at 40 °C until the solution turned clear, after which
benzyl chloride (8.25 g, 7.50 ml, 65.2 mmol, 1.5 eq.) was added and the solution was stirred for 6 h
at 40 °C. The solution was neutralised with concentrated aqueous hydrochloric acid to pH 5-6.
Chloroform (30 ml) was added after neutralization and the mixture was stirred overnight. The
resulting precipitate was removed by filtration and washed with chloroform (30 ml). The organic
fractions were combined and evaporated in vacuo. The resulting residue was treated with acetone
(
30 ml) resulting in the precipitation of crude product which was filtered and again washed with
acetone to afford the product (S)-1-Benzylpyrrolidine-2-carboxylic acid 30, as a white solid (5.39 g,
6.26 mmol, 60 %). Rf 0.09 (9:1 dichloromethane: methanol); mp 175 °C (literature [47] 167 °C); νmax
2
-
1
1
(
neat)/cm 3041, 2992, 2969, 1634, 1450, 1375, 1311, 1190, 753, 704; H NMR (300 MHz, D O); δ
2
7
.53 (s, br, 5 H, Ar-H’s), 4.40 (s, 2H, CH Ph), 4.01 (dd, 1H, J = 6.7 & 9.3 Hz, CHCO H), 3.73-3.58 (m, 1H,
2 2
CH CH N), 3.38–3.21 (m, 1H, m, CH CH N), 2.62–2.41 (m, 1H, CH CH), 2.27–1.89 (m, 3H, CH CH &
2
2a
2
2b
2a
2b
1
3
CH CH N); C NMR (75 MHz, D O); δ 173.54, 130.55, 130.05, 129.96, 129.25, 68.24, 58.30, 54.61,
2
2
2
2
8.79, 22.78; HRMS m/z (ESI) 206.1215 ([M + H]+ requires 206.1165).

ACCEPTED MANUSCRIPT
6
.1.12 (S)-(1-Benzylpyrrolidin-2-yl)methanol 31
(
S)-1-Benzylpyrrolidine-2-carboxylic acid 30 (1.50 g, 7.32 mmol, 1.0 eq.) was dissolved in dry
tetrahydrofuran (30 ml) at 0 °C. To this stirring solution was added borane dimethyl sulfide complex
4.16 ml, 43.9 mmol, 3.0 eq.). The resulting mixture was stirred at 25 °C for 4.5 h. The mixture was
(
then cooled to 0 °C and 3 M aqueous sodium hydroxide (15 ml) was added to destroy excess hydride
and to hydrolyse the amine-borane complex, which required 12 h at 25°C. The pH of the resulting
solution was adjusted to pH 11 by adding a few pellets of sodium hydroxide. The aqueous phase was
saturated with saturated potassium carbonate (30 ml), the organic phase was separated and the
aqueous phase was extracted with diethyl ether (5 x 30 ml). The combined organic extracts were
dried (Na SO ), filtered and evaporated in vacuo to afford (S)-(1-benzylpyrrolidin-2-yl)methanol 31
2
4
(
1.03 g, 5.37 mmol, 73 %) as a viscous yellow oil which was used without any further purification. Rf
-
1
1
0
.18 (2% methanol:dichloromethane); νmax (neat)/cm 3363, 2955, 2797, 1450, 1034; H NMR (300
MHz, CDCl ); δ 7.32–7.21 (m, 5H, ArH’s), 3.95 (d, 1H, J = 13.0 Hz, CH Ph), 3.64 (dd, 1H, J = 3.6 &
3
2a
1
1
1
1
1
0.7, CH OH), 3.42 (dd, 1H, J = 2.2 & 10.8, CH OH), 3.34 (d, 1H, J = 13.0 Hz, CH Ph), 3.00 – 2.90 (m,
2a 2b 2b
H, CHN), 2.72 (ddd, 2H, J = 3.3, 5.9 & 11.9 Hz, CH CH N), 2.28 (dd, 1H, J = 9.2 & 16.8 Hz, CH CH N),
2
2
2
2
1
3
.97–1.75 (m, 2H, CH ), 1.74–1.63 (m, 2H, CH ); C NMR (75 MHz, CDCl ); δ 139.43, 128.79, 128.41,
2
2
3
27.13, 64.36, 61.90, 58.65, 54.54, 27.87, 23.53; HRMS m/z (ESI) 192.1411 ([M + H]+ requires
92.1383).
6
.1.13 (1-Benzylpyrrolidin-3-yl)methanol 33
To a stirred solution of methyl-1-benzyl-5-oxo-3-pyrrolidine carboxylate 32 (3.00 g, 12.86 mmol, 1.0
eq.) in toluene (50 ml), was added Red-Al (20.06 g, 19.30 ml, 64.50 mmol, 65 wt % in toluene) at 15–
2
0 °C. The contents were stirred for 2-3 hours at rt. The reaction was quenched by the slow addition
of 10 % aqueous sodium hydroxide (1 ml) and then diluted with water (40 ml). The contents were
stirred for 30 min, and the toluene layer was separated, washed with a saturated solution of sodium
chloride (30 ml). The organic layer was separated, dried (Na SO ), filtered and evaporated in vacuo
2
4
to yield (1-benzylpyrrolidin-3-yl)methanol 33 as a clear oil (1.41 g, 57 %) which was used without any
-1
further purification. Rf 0.19 (10% methanol:dichloromethane); νmax (neat)/cm 3341, 2792, 1450,
1
1
050, 697; H NMR (300 MHz, CDCl ); δ 7.34–7.19 (m, 5H, ArH’s), 3.68–3.57 (m, 4H, -OH, CH OH &
3
2a
CH Ph), 3.49 (dd, 1H, J = 5.2 & 10.1, CH OH), 2.77 (dt, 1H, J = 4.5 & 8.8 Hz, CHCH N), 2.64–2.49 (m,
2
2b
2a
2
H, CH CH N), 2.44–2.27 (m, 2H, CHCH N & CHCH OH), 2.05–1.90 (m, 1H, CH CH N), 1.73–1.59 (m,
2 2 2b 2 2a 2

ACCEPTED MANUSCRIPT
1
3
1
5
H, CH CH N); C NMR (75 MHz, CDCl ); δ 138.44, 128.91, 128.43, 127.23, 67.20, 60.23, 58.04,
2b 2 3
3.78, 38.93, 27.03; HRMS m/z (ESI) 192.1422 ([M + H]+ requires 192.1383).
6
.1.14 General method for Red-Al reduction
To a stirred solution of the carboxylate (1.50 g, 7.85 mmol, 1.0 eq.) in toluene (25 ml), was added
Red-Al (65 wt % in toluene, 3.66 g, 3.54 ml, 11.8 mmol, 1.5 eq.) at 15-20°C. The contents were
stirred overnight at rt. The reaction was quenched by the slow addition of 10 % sodium hydroxide (~
1
ml) and then the addition of water. The contents were stirred for 30 min, and the toluene layer
was separated, washed with a saturated solution of sodium chloride. The organic layer was dried
anhydrous sodium sulphate), filtered and evaporated in vacuo to yield the alcohol which was used
without any further purification.
(
6
.1.14.1 (1-Benzylaziridin-2-yl)methanol 18
-
1
(
Yield, 84 %); Yellow oil; Rf 0.32 (2:98 methanol: dichloromethane); νmax (neat)/cm 3351, 1026, 734
1
and 696; H NMR (300 MHz, CDCl3); δ 7.38–7.20 (m, 5H, ArH’s), 3.75 (dd, 1H, J = 3.0 & 11.8 Hz,
CH OH), 3.46 (d, 2H, J = 5.2 Hz, CH Ph), 3.42–3.32 (m, 1H, CH OH), 2.85 (br s, 1H, OH), 1.87–1.79
2a
2
2b
1
3
(
m, 2H, CH N), 1.47 (d, 1H, J = 6.4 Hz, CHN); C NMR (75 MHz, CDCl3); δ 138.92, 128.58, 128.16,
2
1
27.36, 64.12, 62.72, 40.54, 31.26; HRMS m/z (ESI) 164.1077 ([M + H]+ requires 164.1070).
.1.14.2 (1-Benzylazetidin-2-yl)methanol 22
6
-
1
(
Yield, 97 %); Yellow oil; Rf 0.22 (ethyl acetate); νmax (neat)/cm 3374, 2924, 2839, 1450, 1101, 733,
1
6
3
97; H NMR (300 MHz, CDCl3); δ 7.40–7.19 (m, 5H, ArH’s), 3.65 (dd, 2H, J = 12.7 & 32.0 Hz, CH Ph),
2
.44 (dt, 1H, J = 2.8 & 10.8 Hz, CH OH), 3.38–3.27 (m, 3H, CH OH, OH & CH N), 2.98 – 2.89 (m, 2H,
2a
2b
2
1
3
CH N & CHN), 2.31–2.12 (m, 1H, CH CH N), 1.92 (dtd, 1H, J = 2.5, 8.0, 10.5 Hz, CH CH N); C NMR
2
2a
2
2b
2
(
75 MHz, CDCl ); δ 137.98, 128.77, 128.48, 127.34, 66.87, 62.33, 61.95, 51.39, 18.70; HRMS m/z (ESI)
3
1
78.1227 ([M + H]+ requires 178.1226).
6
.1.14.3 (1-Benzylpiperidin-2-yl)methanol 37
-
1
(
Yield, 64 %); Orange oil; Rf 0.38 (10 % methanol:dichloromethane); νmax (neat)/cm 3381, 2928,
1
1
447, 1055, 1033, 734, 697; H NMR (300 MHz, CDCl ); δ 7.35–7.21 (m, 5H, ArH’s), 4.07 (d, 1H, J =
3

ACCEPTED MANUSCRIPT
1
3
3.4 Hz, CH Ph), 3.86 (dd, 1H, J = 4.2 & 10.9 Hz, CH OH), 3.54 (dd, 1H, J = 4.0 & 10.9 Hz, CH OH),
2a
2a
2b
.35 (d, 1H, J = 13.4 Hz, CH Ph), 2.94 – 2.70 (m, 2H, CHCH OH & OH), 2.47 (dq, 1H, J = 4.0 & 8.1 Hz,
2b
2
1
3
CH CH N), 2.22 – 2.09 (m, 1H, CH CH N), 1.77 – 1.49 (m, 4H, CH ), 1.47 – 1.27 (m, 2H, CH ); C
2
2a
2
2b
2
2
NMR (75 MHz, CDCl ); δ 138.77, 129.02, 128.44, 127.16, 62.36, 61.14, 57.72, 50.92, 27.34, 24.08,
3
2
3.44; HRMS m/z (ESI) 206.1591 ([M + H]+ requires 206.1539).
.1.14.4 (1-Benzylpiperidin-3-yl)methanol 38
6
-
1
(
Yield, 75 %); Orange oil; Rf 0.36 (10% methanol:dichloromethane); νmax (neat)/cm 3341, 2926,
1
1
1
447, 1352, 1163, 1042, 740, 697; H NMR (300 MHz, CDCl ); δ 7.33–7.20 (m, 5H, ArH’s), 3.57 (dd,
3
H, J = 5.2 & 10.6 Hz, CH OH), 3.49 (s, 2H, CH Ph), 3.47 (dd, 1H, J = 5.8 & 10.5 Hz, CH OH), 3.27 (s,
2a
2
2b
br, 1H, OH), 2.82 (dd, 1H, J = 2.9 & 10.9 Hz, CH N), 2.66 – 2.57 (m, 1H, CH N), 2.14 (t, 1H, J = 9.1 Hz,
2
2
CH N), 2.02 (t, 1H, J = 9.7 Hz, CH N), 1.86 – 1.49 (m, 4H, CH, CH & CH ), 1.13 (dt, 1H, J = 6.2 & 18.1,
2
2
2a
2
13
CH ); C NMR (75 MHz, CDCl3); δ 137.88, 129.36, 128.31, 127.17, 67.00, 63.69, 57.46, 54.16, 37.94,
2b
2
7.58, 24.60; HRMS m/z (ESI) 206.1591 ([M + H]+ requires 206.1539).
.1.14.5 (1-Benzylpiperidin-4-yl)methanol 39
6
-
1
(
Yield, 65 %); Orange oil; Rf 0.31 (9:1 dichloromethane: methanol); νmax (neat)/cm 3331, 2915,
1
2
803, 1448, 1367, 1112, 1040, 737, 697; H NMR (300 MHz, CDCl ); δ 7.35–7.19 (m, 5H, ArH’s), 3.50
3
(
s, 2H, CH Ph), 3.46 (d, 2H, J = 6.4 Hz, CH OH), 2.91 (d broad, 2H, J = 11.6 Hz, 2 x CH CH N), 2.53 (s,
2 2 2 2a
1
1
H, OH), 1.96 (td, 2H, J = 2.3 & 11.7 Hz, 2 x CH CH N), 1.70 (d, 2H, J = 13.9 Hz, 2 x CH CH N), 1.57 –
2 2b 2a 2
13
.40 (m, 1H, m, CHCH OH), 1.28 (ddd, 2H, J = 3.7, 12.2 & 24.2 Hz, 2 x CH CH N); C NMR(75 MHz,
2
2b
2
CDCl ); δ 138.26, 129.43, 128.27, 127.10, 67.83, 63.54, 53.50, 38.61, 28.83; HRMS m/z (ESI)
3
2
06.1591 ([M + H]+ requires 206.1539).
6
.1.15 General method for the Swern oxidation
A solution of dimethylsulfoxide (0.77 ml, 10.8 mmol, 2.5 eq.) in dichloromethane (5 ml) was added
drop-wise to a pre-cooled solution of oxalyl chloride (0.45 ml, 5.20 mmol, 1.2 eq.) in
dichloromethane (6 ml) and left to stir for 30 min. A solution of the alcohol (4.33 mmol, 1.0 eq.) in
dichloromethane (5 ml) was added drop-wise and stirred for 30 min. Triethylamine (3.0 ml, 21.7
mmol, 5.0 eq.) was then added and the solution is stirred for an additional 30 min. Upon warming
the solution to room temperature, water was added and the organic layer separated and washed

ACCEPTED MANUSCRIPT
with brine. The organic layer was then dried over anhydrous sodium sulfate, filtered and the solvent
was evaporated in vacuo to afford the aldehyde. The aldehyde was used as is without any further
purification.
6
.1.15.1 1-Benzylaziridine-2-carbaldehyde 6
-
1
(
Yield, 93 %); Orange oil; Rf 0.77 (75% ethyl acetate: hexane); νmax (neat)/cm 2923, 1655, 1446,
1
6
97; H NMR (300 MHz, CDCl ); δ 8.93 (d, 1H, J = 6.4 Hz, CHO), 7.42–7.19 (m, 5H, ArH’s), 3.56 (s, 2H,
3
J = 13.3 Hz, CH Ph), 2.31 (d, 1H, J = 2.9 Hz, CHCHO), 2.22 (dt, 1H, J = 3.0, 6.5 Hz, CH N), 1.93 (d, 1H, J
2
2a
1
3
=
3
6.6 Hz, CH N); C NMR (75 MHz, CDCl ); δ 199.93, 137.73, 128.71, 128.14, 127.70, 63.50, 44.69,
2b 3
2.85; HRMS m/z (ESI) 162.0941 ([M + H]+ requires 162.0914).
.1.15.2 1-Benzylazetidine-2-carbaldehyde 7
6
-
1
1
(
Yield, 89 %); Yellow oil; Rf 0.22 (ethyl acetate); νmax (film)/cm 2923, 1667, 1070; H NMR (300
MHz, CDCl ); δ 9.41 (d, 1H, J = 2.9 Hz, CHO), 7.34–7.21 (m, 5H, ArH’s), 3.66 (2H, dd, J = 12.4 & 41.8
3
Hz, CH Ph), 3.63 (dt, 1H, J = 3.9 & 11.2 Hz, CHCHO), 3.47–3.38 (1H, m, CH N), 3.16–3.04 (1H, m,
2
2a
1
CH N), 2.34–2.10 (2H, m, CH CH N); C NMR (75 MHz, CDCl ); δ 202.59, 137.22, 129.08, 128.60,
2b
2
2
3
1
27.65, 70.87, 62.80, 52.18, 19.39; HRMS m/z (ESI) 176.1050 ([M + H]+ requires 176.1070).
.1.15.3 1-Benzylazetidine-3-carbaldehyde 8
6
-
1
(
Yield, 65 %); Orange oil; Rf 0.14 (5% methanol:dichloromethane); νmax (neat)/cm 2948, 2825,
1
1
7
724, 1450, 1360, 1075, 980, 697; H NMR (300 MHz, CDCl ); δ 9.87 (d, 1H, J = 2.6 Hz, CHO), 7.36–
3
.15 (m, 5H, ArH’s), 3.60 (s, 2H, CH Ph), 3.48–3.36 (m, 4H, 2 x CHCH N), 3.30–3.19 (m, 1H, CHCHO);
2
2
13
C NMR (300 MHz, CDCl ); δ 201.35, 137.49, 128.55, 128.53, 127.34, 63.42, 54.55, 41.54; HRMS m/z
3
(
ESI) 176.1108 ([M + H]+ requires 176.1070).
6
.1.15.4 (S)-1-Benzylpyrrolidine-2-carbaldehyde 9
-
1
(
Yield, 91 %); Dark orange oil; Rf 0.32 (5% methanol:dichloromethane); νmax (neat)/cm 2921, 1667,
1
1
3
447, 1073, 697; H NMR (300 MHz, CDCl ); δ 9.29 (d, 1H, J = 4.0 Hz, CHO), 7.33–7.20 (m, 5H, ArH’s),
3
.69 (q, 2H, J = 12.9 Hz, CH Ph), 3.13–3.06 (m, 1H, CHCHO), 2.97 (ddd, 1H, J = 4.1, 6.5 & 10.7Hz,
2
CH CH N), 2.38 (dd, 1H, J = 8.4 & 17.1 Hz, CH CH N), 2.05–1.91 (m, 1H, CH ), 1.90–1.75 (m, 3H, CH
2
2
2a
2
2b
2a