Structure-Kinetic Profiling of Haloperidol Analogues at the Human Dopamine D2 Receptor

Article abstract of DOI:10.1021/acs.jmedchem.9b00864

Haloperidol is a typical antipsychotic drug (APD) associated with an increased risk of extrapyramidal side effects (EPSs) and hyperprolactinemia relative to atypical APDs such as clozapine. Both drugs are dopamine D₂ receptor (D₂R) antagonists, with contrasting kinetic profiles. Haloperidol displays fast association/slow dissociation at the D₂R, whereas clozapine exhibits relatively slow association/fast dissociation. Recently, we have provided evidence that slow dissociation from the D₂R predicts hyperprolactinemia, whereas fast association predicts EPS. Unfortunately, clozapine can cause severe side effects independent of its D₂R action. Our results suggest an optimal kinetic profile for D₂R antagonist APDs that avoids EPS. To begin exploring this hypothesis, we conducted a structure-kinetic relationship study of haloperidol and revealed that subtle structural modifications dramatically change binding kinetic rate constants, affording compounds with a clozapine-like kinetic profile. Thus, optimization of these kinetic parameters may allow development of novel APDs based on the haloperidol scaffold with improved side-effect profiles.

Full text of DOI:10.1021/acs.jmedchem.9b00864

Subscriber access provided by UNIV OF GEORGIA
Article
Structure-Kinetic Profiling of Haloperidol
2
Analogues at the Human Dopamine D Receptor
Tim John Fyfe, Barrie Kellam, David A. Sykes, Ben Capuano, Peter J.
Scammells, J Robert Lane, Steven J. Charlton, and Shailesh N. Mistry
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b00864 • Publication Date (Web): 03 Oct 2019
Downloaded from pubs.acs.org on October 7, 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 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Structure-Kinetic Profiling of Haloperidol Analogues at the Human
Dopamine D Receptor
2
Tim J. Fyfe,†,‡,§ Barrie Kellam, David Sykes, Ben Capuano, Peter J. Scammells, J.
‡
≠
†
†
Robert Lane,≠, #,* Steven J. Charlton,≠,#, ^, * and Shailesh N. Mistry‡,*
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
†
§
Medicinal Chemistry, and Drug Discovery Biology, Monash Institute of Pharmaceutical Sciences,
Monash University, Parkville, Victoria 3052, Australia.
‡School of Pharmacy, Centre for Biomolecular Sciences, University of Nottingham, Nottingham
NG7 2RD, U.K.
≠School of Life Sciences, Queen’s Medical Centre, University of Nottingham, Nottingham NG7
2
UH, U.K.
#Centre of Membrane Protein and Receptors, Universities of Birmingham and Nottingham, U.K.
^ Excellerate Bioscience Ltd, BioCity, Nottingham, NG1 1GF, U.K.
.
ABSTRACT
Haloperidol is a typical antipsychotic drug (APD) associated with an increased risk of extrapyramidal
side-effects (EPS) and hyperprolactinemia relative to atypical APDs such as clozapine. Both drugs
are dopamine D receptor (D R) antagonists, with contrasting kinetic profiles. Haloperidol displays
2
2
fast association/slow dissociation at the D R whereas clozapine exhibits relatively slow
2
association/fast dissociation. Recently, we have provided evidence that slow dissociation from the
D R predicts hyperprolactinemia, whereas fast association predicts EPS. Unfortunately, clozapine
2
can cause severe side-effects independent of its D R action. Our results suggest an optimal kinetic
2
profile for D R antagonist APDs that avoids EPS. To begin exploring this hypothesis, we conducted
2
a structure-kinetic relationship study of haloperidol and reveal that subtle structural modifications
dramatically change binding kinetic rate constants, affording compounds with a clozapine-like kinetic
profile. Thus, optimisation of these kinetic parameters may allow development of novel APDs based
on the haloperidol scaffold with improved side-effect profiles.
.
INTRODUCTION
Haloperidol (1, Figure 1) is an effective, typical antipsychotic drug (APD) used in the treatment of
schizophrenia (SCZ). As for all current APDs, its mechanism of action is primarily through
antagonism of dopamine (DA) D receptors (D R) in the mesolimbic pathway, where excessive DA
2
2
activity is thought to underlie the positive symptoms of schizophrenia.1-3 Unfortunately, 1 along with
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 78
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
other typical APDs, are associated with severe on-target side effects including EPS (e.g., Parkinsonian
3,4
symptoms such as bradykinesia and tremor) and hyperprolactinemia. These symptoms are mediated
by blockade of D R signalling in the nigrostriatal and tuberoinfundibular DA pathways,
2
respectively.2-7 Tardive dyskinesia is also associated with long-term exposure to typical APDs such
as 1.8
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
Atypical APDs display a diminished incidence of EPS and hyperprolactinemia relative to typical
APSs.9-11 While the primary distinction between typicality and atypicality is based on such clinical
observations, the mechanism(s) that might drive this distinction remain unclear. Clozapine (2, Figure
1
) is a prototypical atypical antipsychotic. It has a complex pharmacological profile with high affinity
for other members of the biogenic amine receptor family and, in particular, a relatively high affinity
1
2
for the serotonin 2A receptor (5-HT R). Many atypical APDs have similar pharmacology leading
2
A
to the hypothesis that a relatively high affinity for the 5HT R as compared to the D R confers
2
A
2
atypicality.13-16 However, not all atypical APDs share this profile suggesting that this theory cannot
account for all examples of atypicality.13,17 Unfortunately, this lack of selectivity across aminergic
receptors is associated with off-target side-effects, including sedation, metabolic disorders, weight
1
8
gain, urinary incontinence and constipation. 2 can also cause acute agranulocytosis, a potentially
life-threatening white blood cell disorder.
Cl
N
OH
O
N
N
N
Cl
F
N
H
Haloperidol (1)
Clozapine (2)
Figure 1. Typical APD haloperidol (1) and atypical APD clozapine (2).
The relatively fast rate at which 2, and other related APDs, dissociate from the D R has also been
2
1
7
suggested to be the basis for an atypical profile. Rapid dissociation of an antagonist might allow a
fraction of D Rs to be occupied by transiently high concentrations of DA released into the synapse
2
whereas an antagonist with a slow dissociation rate would cause insurmountable antagonism. Central
to this hypothesis was the consensus that APDs exhibit similar association rates (k ) for the D R
on
2
1
9
meaning that affinity is largely mediated by differences in dissociation rate (k ). Olanzapine,
off
however, which has a similar high affinity for the D R as many typical antipsychotics, displays an
2
atypical profile.4
The incorporation of drug-receptor kinetic binding parameters into drug discovery programs is seen
as increasingly important for the development of next generation therapeutics.20-28 Previous efforts to
derive estimates of APD kinetic rate constants have used radiometric detection methods with limited
2
ACS Paragon Plus Environment

Page 3 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
assay throughput.19,29,30 We have recently developed a competition association assay using TR-FRET
31,32
to determine ligand kinetic parameters of unlabelled D R agonists,
and profiled an extensive
2
3
3
series of APDs in order to explore the kinetic basis for on-target side effects. We found that the
association rates of the APDs varied over three orders of magnitude and that association rates, rather
than dissociation rates, correlated with EPS. These observations led us to propose a revised kinetic
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
hypothesis whereby rapid association rate leads to drug rebinding at the D R, maintaining a higher
2
concentration of APD in the synaptic compartment. This causes increased competition with DA
leading to EPS. In contrast, hyperprolactinaemia was correlated with APD dissociation rate.33
Optimising D R binding kinetics may permit the design of novel tools to test this kinetic hypothesis,
2
as well as facilitate the generation of new APDs with an improved therapeutic profile.
Although clozapine (2) appears to possess the desired slow on/fast off kinetic profile for reduced on-
7
-1
-1
-1 33
target side effects (k = 8.23 ± 1.42 × 10 M min , k = 1.67 ± 0.25 min ), it displays affinity
on
off
for many aminergic GPCRs, contributing to its off-target side effects. Haloperidol (1), in contrast,
has a better off-target selectivity profile, but an undesirable fast on/slow off kinetic profile at the D R
2
9
-1
-1
-1
(
k = 1.29 ± 0.21 × 10 M min , k = 0.61 ± 0.04 min ) that contributes to its on-target side effects.
on off
The aim of the current study was to optimise the kinetic binding parameters of the more selective
scaffold of haloperidol towards a slow on/fast off profile. To this end, we herein describe the design
and synthesis of 50 analogues of 1, focusing on structural modification of four key moieties (figure
2
) and use competition association kinetic binding methodology to determine their association and
dissociation rates, and equilibrium affinities at the D R. We reveal that both the association and
2
dissociation kinetics of this scaffold can vary considerably with subtle structural modification.
Interestingly, we have identified previous analogues of 1, among others, that may have been
overlooked on the basis of affinity-driven scaffold optimisation, that possess favourable kinetic
profiles. These data reveal the structure-kinetic relationships (SKR) of 1, as well identify novel tool
compounds with which to interrogate the relationship between APD kinetic binding parameters and
on-target side-effect profiles. Although the structure-activity relationships (SAR) surrounding the
butyrophenone scaffold of APDs have been extensively studied in previous years,34-39 to our
knowledge, these data represent the first reported SKR relating to analogues of 1.
.
RESULTS AND DISCUSSION
Chemistry. To begin our structure-kinetic study, we focused on modifying four distinct regions of
, namely the para-fluorophenyl (red box), ketone and alkyl linker (green box), piperidinol (orange
1
box), and para-chlorophenyl (blue box) moieties, as depicted in figure 2. For completeness, we
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 78
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
included both established and novel analogues of 1 in our approach, covering 50 compounds in
totality.
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 2. Structural regions of haloperidol (1) investigated as part of the SKR study.
Variation of para-chlorophenyl moiety of 1. In evaluating the positional effects of halogen
substitution on the p-chlorophenyl moiety, we initially synthesised analogues bearing the chloro-
substituent in the ortho (8n) and meta (8a) positions, as well as incorporation of all possible dichloro
40
substitution patterns (2,3-diCl (8b); 2,4-diCl (8c); 2,5-diCl (8d); 2,6-diCl (8e); 3,4-diCl (8f); and
3,5-diCl (8g)). In addition, we wanted to assess the effects of halogen removal through the proteo
4
1
41
analogue (8h), as well as alternative para-substituents, including methyl (8i), trifluoromethyl (8j),
4
1
N,N-dimethylamino (8k) and fluoro (8l). The synthesis of these compounds is summarised in
scheme 1. Firstly, the appropriately substituted bromobenzene (3a-l) underwent lithiation using n-
BuLi, followed by treatment with commercially available tert-butyl 4-oxopiperidine-1-carboxylate
(4) to afford the corresponding N-Boc-protected phenylpiperidinols (5a-l). HCl-mediated N-Boc-
deprotection afforded the corresponding hydrochloride salts or free amines following basic work-up
(6a-l). Finally, nucleophilic displacement of key intermediate 4-chloro-1-(4-fluorophenyl)butan-1-
one (7a) with the appropriate substituted phenylpiperidinol (6a-l) was achieved by refluxing in
toluene in the presence of KI and NaHCO , to afford the desired final analogues (8a-l).
3
4
ACS Paragon Plus Environment

Page 5 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 1. Synthesis of haloperidol (1) analogues with modification to para-chlorophenyl
moietya
O
O
Br
N
O
(i)
HO
+
N
O
R1
O
R1
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
3a-l
4
5a-l
ii)
(
OH
NH
HO
O
R1
(iii)
O
N
Cl
R1
F
8a-l
6a, 6c, 6e-f, 6h-l (free base)
6b, 6d, 6g (HCl salt)
a R1 = 3-Cl
1
g R1 = 3,5-Cl
F
7a
8
1
b R = 2,3-Cl h R = H
1
1
1
1
c R = 2,4-Cl
i
j
R
R
= 4-Me
= 4-CF3
1
d R = 2,5-Cl
1
e R = 2,6-Cl
k R = 4-N(CH3)2
1
f
R
= 3,4-Cl
R1 = 4-F
l
aReagents and conditions: (i) n-BuLi, THF, -78 ºC, 3-8 h, 50-88% (5a-l); (ii) HCl (4 M), 1,4-dioxane,
1
2
-3 h, 75-98% (6a, 6c, 6e-f, 6h-l (free base) 6b, 6d, 6g (HCl salt)); (iii) NaHCO , KI, toluene, reflux,
3
4 h, 45-70% (8a-l).
Though established, the employed lithiation chemistry proved to be problematic towards the synthesis
of the o-chloro analogue (8n, scheme 2). Standard conditions failed to deliver the desired N-Boc-
protected piperidinol intermediate from 3m, instead producing the biphenyl piperidinol (5m). Whilst
unintended, this molecule would still provide additional information to our study and was N-Boc
deprotected to give 6m, followed by N-alkylation with 7a using conditions outlined previously, to
furnish biphenyl analogue (8m). In contrast, the desired o-chloro analogue was accessed in three steps
using an alternative approach (scheme 2). Grignard addition of 3m to 4 yielded o-chlorophenyl
piperidinol intermediate 5n, which underwent N-Boc deprotection to give 6n. Final N-alkylation with
7
a, furnished 8n.
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 78
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
Scheme 2. Synthesis of biphenyl side-product and ortho-Cl analogues of 1a
OH
R2
Cl
O
OH
(i) or (ii)
(iii)
Br
R2
+
O
N
O
N
HN
6m, 6n
O
O
3m
4
5m, 5n
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
OH
R2
O
(
iv)
N
R2
=
R2
=
Cl
7
a
F
Cl
5n, 6n, 8n
8m, 8n
5m, 6m, 8m
aReagents and conditions: (i) n-BuLi, THF, -78 ºC, 6 h, 55% (5m); (ii) Mg, I2(cat.), Et O, 0 ºC –
2
reflux, 3 h, 54% (5n); (iii) HCl, 1,4-dioxane, rt, 2 h, 72-96% (6m, 6n); (iii) NaHCO , KI, toluene,
3
reflux, 24 h, 62-76% (8m, 8n).
Variation of the p-fluorophenyl moiety of 1. To investigate positional effects of fluorine substitution
on the butyrophenone phenyl ring on the kinetics of 1, we generated analogues with all possible mono
4
2
42
and di-fluoro substituents (2-F (14a); 3-F (14b); 2,3-diF (14c); 2,4-diF (14d); 2,5-diF (14e); 2,6-
4
2
diF (14f); 3,4-diF (14g); 3,5-diF (14h), as well as two ortho-substituted analogues (2-Cl (14i) and
43
41
2
-Me (14j)), a para-substituted analogue (4-Cl (16k)) and a des-fluoro variant (14l). As detailed
in scheme 3, commercially available 3-butynol (9) was treated with SOCl and catalytic pyridine at
2
reflux temperature, followed by distillation to afford 4-chlorobut-1-yne (10).42 The appropriate
42
iodobenzene (11a-j) was then employed in a Pd-catalysed Sonogashira cross-coupling reaction with
10, affording the corresponding internal aryl alkynes (12a-j). Next, we utilised a TfOH-catalysed
metal-free regioselective Markovnikov-type hydration protocol44 with 2,2,2-trifluoroethanol as
solvent in the presence of H O, furnishing the corresponding aryl ketones (13a-j). Finally, N-
2
alkylation of commercially available key intermediate 7b with each synthesised alkyl chloride (13a-
b, 13d-e, 13g-j) furnished final analogues 14a-b, 14d-e, and 14g-j. Alternatively, 4-chloro-1-(4-
chlorophenyl)butan-1-one (13l) was accessed via Freidel-Crafts acylation, followed by N-alkylation
of 7b to afford 14k. Finally, commercially available 4-chloro-1-phenylbutan-1-one (13l) was
aminated with 7b to afford the des-fluoro analogue 14l.
6
ACS Paragon Plus Environment

Page 7 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 3. Synthesis of analogues of 1 with modification to p-fluorophenyl moietya
Cl
(i)
(ii)
OH
Cl
I
R3
9
10
12a-j
R3
11a-j
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
Cl
O
OH
(iii)
Cl
(iv)
OH
O
N
R3
HN
Cl
1
3a-j
R3
7b
4 a R3 = 2-F
g R3 = 3,4-F
1
3
3
b R = 3-F
c R = 2,3-F i R = 2-Cl
d R = 2,4-F j R = 2-Me
e R = 2,5-F k R = 4-Cl
f R = 2,6-F
h R = 3,5-F
3
3
3
3
3
3
O
3
3
Cl
l R = H
R3
3
1
3k R = 4-Cl
3
13l R = H
a
Reagents and conditions: (i) SOCl , pyridine, 0 ºC – reflux, 30 min, 82%; (ii) PdCl (PPh ) , CuI,
2
2
3 2
Et N, 1,4-dioxane, 50 ºC, 1-3 h, 45-85% (12a-j); (iii) TfOH, H O, CF CH OH, 60 ºC, 3-8 h, 50-90%
3
2
3
2
(13a-j); (iv) NaHCO , KI, toluene, reflux, 24 h, 50-82% (14a-b, d-e, g-l); 14c and 14f were detected
3
but unable to be isolated in appreciable yield.
The synthesis of analogues containing 2,3-difluorophenyl (14c) and 2,6-difluorophenyl (14f)
substituents were problematic. When attempting to N-alkylate key intermediate 7b with the
corresponding alkyl halides (13c, 13f) major side-products due to a competing S Ar reaction were
N
observed, making purification of the target compounds by FCC and preparative HPLC extremely
challenging. These side-products are believed to arise due to activation of the position ortho to the
ketone moiety, when a fluoro-substituent is present. To circumvent the S Ar reaction, syntheses of
N
the affected analogues were modified to incorporate ketal protection/deprotection of the ketone,
permitting nucleophilic displacement of the alkyl halide only (scheme 4). Beginning with ketal
protection of 13c, we employed a pTsOH-catalysed reaction with trimethyl orthoformate in MeOH
at room temperature, to afford the corresponding dimethyl ketal (15c). Alternatively, 13f was reacted
with 1,2-ethanediol in the presence of catalytic pTsOH in toluene under Dean-Stark conditions, to
afford the corresponding 1,3-dioxolane (15f). These compounds were then subjected to nucleophilic
displacement using 7b to furnish 16c and 16f, followed by acid-catalysed hydrolysis in acetone at
reflux, affording final compounds 14c and 14f.
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 78
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
Scheme 4. Synthesis of 2,3- and 2,6-difluoro analogues of 1 using various protection strategiesa
R4 R5
O
(i) or (ii)
(iii)
O
O
Cl
Cl
R3
7b
R3
3
3
4
5
1
1
3c R = 2,3-diF
15c R = 2,3-F, R = Me, R = Me
3
3
4
5
3f R = 2,6-diF
15f R = 2,6-F, R = R = (CH2)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
Cl
Cl
R4 R5
OH
OH
(iv)
O
O
O
N
N
R3
R3
3
3
4
5
14c R = 2,3-F
1
6c R = 2,3-F, R = Me, R = Me
3
3
4
5
14f R = 2,6-F
16f R = 2,6-F, R = R = (CH2)2
a
Reagents and conditions: (i) trimethylorthoformate, p-TsOH.H O, MeOH, rt, 12 h, 77% (15c); (ii)
2
ethylene glycol, p-TsOH.H O, toluene, reflux (Dean-Stark), 16 h, 83% (15f); (iii) NaHCO , KI,
2
3
toluene, reflux, 24 h, 71-77% (16c, 16f); (iv) p-TsOH.H O, 15:1 acetone/H O, reflux, 48 h, 76-82%
2
2
(14c, 14f).
Variation of ketone and linker moiety of 1. We focused on replacement of the ketone group of 1 with
a range of moieties, including ether, thioether and the corresponding carbinol (racemic). The ether-
45
and thioether-variants of 1 were accessed using a literature procedure in three steps (17a-b, Figure
3, Supplementary Scheme 1), whilst the corresponding secondary alcohol was afforded in two-steps
4
6
also through literature procedure (18, Figure 3) (Supplementary Scheme 2).
Cl
Cl
Cl
OH
OH
OH
OH
O
N
X
N
N
F
1
1
7a X = S
7b X = O
F
18
F
42
Cl
Cl
O
O
O
N
N
N
N
F
4
3
3
47
48
F
F
OH
OH
O
O
N
N
Cl
5
54
Figure 3. Literature analogues of 1 synthesised using various methodologies. Ether- and thioether
analogues45 (17a-b, respectively); racemic alcohol analogue46 (18) ; tropanyl analogue34 (42);
8
ACS Paragon Plus Environment

Page 9 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
piperazinyl analogue34 (43); phenyl- and p-chlorophenylpiperidine analogues38,47 (47-48,
48
41
respectively); reverse substitution analogue (53) ; des-halo analogue (54).
Our subsequent focus was to further understand the effect and importance of geometry on the kinetics
of conformationally restricted analogues of 1, via synthesis of both olefin geometric isomers. The
trans-olefin 23 was accessed through a five step chemical synthesis as outlined in scheme 5,
beginning with a one-pot base-mediated intramolecular enolate alkylation of key intermediate 7a, to
furnish cyclopropyl(4-fluorophenyl)methanone (19) in quantitative yield. Subsequent reduction with
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
NaBH afforded secondary alcohol 20, followed by a vanadyl acetylacetonate-catalysed
4
stereoselective isomerisation in chlorobenzene to yield (E)-4-(4-fluorophenyl)but-3-en-1-ol (21) as
the exclusive geometric isomer. Compound 21 was subsequently activated with methanesulfonyl
chloride to give mesylate 22. This was followed by N-alkylation of 7b using standard conditions,
affording final olefin analogue 23.
Scheme 5. Synthesis of trans-olefin analogue of 1a
O
O
OH
(i)
(ii)
(iii)
Cl
F
F
F
7a
19
20
Cl
(E)
(E)
OH
OMs
(iv)
(iii)
(E)
OH
N
7b
F
F
21
F
22
23
a
Reagents and conditions: (i) NaOH, MeOH, 60 ºC, 5 h, quantitative; (ii) NaBH , MeOH, 0 ºC – rt,
4
3
h, 99%; (iii) VO(acac) , BHT, PhCl, 80 ºC, 48 h, 35%; (iv) MsCl, DCM, Et N, rt, 3h, 88%; (v)
2 3
NaHCO , KI, toluene, reflux, 24 h, 82%.
3
The cis-isomer 28, was accessed through a three-step synthesis as outlined in scheme 6. Initially, Ni-
catalysed stereoselective arylation of 2,3-dihydrofuran 24 with (4-fluorophenyl)magnesium bromide
4
9
(
25) at -30 ºC, successfully afforded the cis-olefin 26 as the exclusive isomer. This compound was
then mesylated using standard conditions to afford 27, followed by N-alkylation of 7b using
conditions outlined previously, furnishing 28. As outlined in scheme 7, both trans and cis-isomers
(21 and 26, respectively) were treated with diethylzinc and diiodomethane using Simmons-Smith50
conditions to access the corresponding racemic trans- and cis-cyclopropanes (29 and 30,
respectively). This was followed by mesylation to give 29a and 30a, and subsequent N-alkylation of
7
b to afford racemic 29b and 30b.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 78
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
Scheme 6. Synthesis of cis-olefin analogue of 1a
Cl
N
OH
OMs
OH
(i)
F
(ii)
F
(
iii)
O
F
7
b
(Z)
(Z)
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
MgBr
F
24
26
27
25
(Z)
2
8
a
Reagents and conditions: (i) Ni[COD] , 1,3-bis-(2,6-diisopropylphenyl)imidazolinium chloride,
2
LiCl, THF, -30 ºC, 8 h, 31%; (ii) MsCl, Et N, DCM, rt, 24 h, 90%; (iii) NaHCO , KI, toluene, reflux,
3
3
2
4 h, 75%.
a
Scheme 7. Synthesis of both trans- and cis-cyclopropane enantiomers of 1
(E)
OH
F
21 = trans
(i)
(ii)
(iii)
N
F
(Z)
7b
F
OH
F
OMs
OH
OH
29 = trans
0 = cis
29a = trans
30a = cis
29b = trans
30b = cis
Cl
F
3
26 = cis
a
Reagents and conditions: (i) Et Zn, CH I , DCM, 0 ºC – rt, 24 h, 95-98% (29, 30); (ii) MsCl, Et N,
2
2 2
3
DCM, rt, 90-95% (29a, 30a); (iii) NaHCO , KI, toluene, reflux, 24 h, 67-70% (29b, 30b).
3
Next, we focused on the synthesis of both propiophenone and valerophenone analogues of 1 that
maintained the ketone functionality (scheme 8). Beginning with Friedel-Crafts acylation chemistry,
the appropriate commercially available acyl chloride (31a, 31c) was reacted with fluorobenzene 32
51
52
in the presence of stoichiometric AlCl to afford the corresponding phenones (33a, 33c ). This was
3
5
3
followed by N-alkylation of 7b to afford final analogues (34a, 34c ). In addition, we wanted to access
the 1,3-propylene, 1,4-butylene and 1,5-pentylene analogues of 1 (scheme 8). To achieve this,
phenones 33a, 7a, and 33c were treated with triethylsilane in TFA, followed by evaporation and direct
chromatographic purification to furnish the corresponding reduced intermediates 35a-c. Lastly, N-
alkylation of key intermediate 7b furnished alkane analogues 36a-c.
1
0
ACS Paragon Plus Environment

Page 11 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 8. Synthesis of ketone and alkane analogues of 1a
O
O
O
(i)
(ii)
N
n
Cl
Cl
Cl
n
n
7b
F
OH
F
3
3
1a n = 1
1c n = 3
33a n = 1
33c n = 3
34a n = 1
34c n = 3
F
Cl
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
32
(iii)
O
N
(vi)
n
(iii)
Cl
Cl
n
F
7b
OH
F
F
3
3
5a n = 1
5b n = 2
36a n = 1
36b n = 2
36c n = 3
Cl
7a
35c n = 3
a
Reagents and conditions: (i) AlCl , DCM, 0 ºC – rt, 6 h, 85% (33a, 33c); (ii) NaHCO , KI, toluene,
3
3
reflux, 24 h, 75-90% (34a, 34c); (iii) triethylsilane, trifluoroacetic acid, 0 ºC, 2-3 h, 75-85% (35a-c);
iv) NaHCO , KI, toluene, reflux, 24 h, 72-91% (36a-c).
(
3
To assess analogues of 1 incorporating internal aromatic alkynes (scheme 9), 1-fluoro-4-iodobenzene
37) was initially subjected to modified Sonagashira conditions54 using commercially available
(
5
5
alcohols (38a, 38c), affording aryl alkynes (39a, 39c). Next, the alcohols were converted to their
5
6
corresponding mesylates (40a, 40c), followed by N-alkylation of 7b with the appropriate mesylate
to afford the corresponding final propynyl and pentynyl analogues (41a and 41c, respectively). The
butynyl analogue 41b was accessed via the cross-coupling reaction between key intermediate 10 and
1-fluoro-4-iodobenzene (37), providing aryl alkyne intermediate 40b, which underwent amination
with 7b.
Scheme 9. Synthesis of internal alkyne analogues of 1a
F
F
(i)
(ii)
OH
n
+
OH
n
I
37
38a n = 1
8c n = 3
39a n = 1
39c n = 3
3
F
OH
Cl
F
HN
Cl
OH
7b
OMs
N
n
(iv)
n
4
4
0a n = 1
0c n = 3
41a n = 1
4
4
1b n = 2
1c n = 3
Cl
(iii)
Cl
F
10
40b
F
I
37
1
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 78
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
Reagents and conditions: (i) Pd(PPh ) Cl , CuI, Et N, MeCN, rt 5 h, 94% (39a, 39c); (ii)
3
2
2
3
,
methanesulfonyl chloride, Et N, DCM, 24 h, 93% (40a, 40c); (iii) PdCl (PPh ) , Et N, 1,4-dioxane,
3
2
3 2
3
5
0 ºC, 3 h, 75% (40b); (iii) NaHCO , KI, toluene, reflux, 24 h, 68-74% (41a-c).
3
Variation of the piperidinol moiety of 1. Modification to the piperidinol moiety of 1 was another key
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
interest in our SKR investigation. To observe the kinetic effect of introducing an ethylene bridge on
3
6
the piperidinol, we synthesised tropanyl analogue 42 according to a literature procedure (Figure 3,
Supplementary Scheme 3), utilising n-BuLi in place of Grignard chemistry. We then sought to modify
the tertiary alcohol group, beginning with synthesis of piperazinyl analogue 43 (Figure 3). This
compound was accessed in two steps via the construction of the piperazine ring and subsequent N-
3
4
alkylation with 7a (Supplementary Scheme 4).
Removing the tertiary alcohol within 1 to generate the corresponding 3,6-dihydropyridine (45) was
our next focus, as well as further elaboration of the olefin to yield the corresponding cycloalkane
derivative (46) (scheme 10). Key piperidinol intermediate 7b was firstly dehydrated using neat
concentrated HCl followed by an alkaline work-up to afford the 1,2,3,6-tetrahydropyridine (44).
5
7
Displacement of 7a with 44 furnished olefin 45. This molecule was subsequently treated using
Simmons-Smith50 conditions, as outlined previously, to afford the corresponding cyclopropane
analogue 46.
Scheme 10. Synthesis of dihydropyridyl and fused cyclopropane analogues of 1a
Cl
Cl
Cl
OH
(i)
(ii)
O
N
7a
HN
HN
7b
44
F
45
O
Cl
(iii)
N
F
Rac
46
Reagents and conditions: (i) HCl (conc.), reflux, 5 h, quantitative; (ii) NaHCO , KI, toluene, reflux,
3
2
4 h, 65% (45); (iii) Et Zn, CH I , DCM, 0 ºC – rt, 24 h, 75% (46).
2 2 2
47
In addition, we synthesised two analogues of 1 containing modified phenyl piperidine (47) and p-
3
8
chlorophenyl piperidine (48) cores (Figure 3), of which their synthesis is detailed in Supplementary
Scheme 5.
1
2
ACS Paragon Plus Environment

Page 13 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Further emphasis was placed on the tertiary alcohol contained within 1, where we sought to assess
the impact of O-methylation (scheme 11). N-Boc-protection of key intermediate 7b gave 49, followed
by O-alkylation with methyl iodide to afford the corresponding methyl ether 50. This was followed
by N-Boc-deprotection to give the secondary amine hydrochloride 51.Final N-alkylation with key
intermediate 7a afforded compound 52.
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
Scheme 11. Synthesis of methyl ether analogue of 1a
Cl
Cl
O
OH
(i)
(ii)
OH
Cl
O
N
O
N
HN
O
O
7b
49
50
O
(iii)
O
(iv)
O
Cl
Cl
Cl
N
7a
H2N
F
51
52
a
Reagents and conditions: (i) Boc O, Et N, DCM, rt, 4 h, 85%; (ii) NaH, MeI, DMF, rt, 24 h, 80%;
2
3
(
iii) HCl, 1,4-dioxane, rt, 2 h, 95%; (iv) NaHCO , KI, toluene, reflux, 24 h, 70%.
3
Dual modification to halo-aryl moieties of 1. Recent molecular dynamics (MD) simulations by
5
8
Thomas et al. were used to understand the ligand binding pathways of 1 and 2 at the D R/D R. The
2
3
final stable pose of 1 was shown to occupy the same space as predicted in a number of molecular
docking studies;59-61 however, the molecular orientation was contradictory to these data by 180º, with
the butyrophenone moiety buried most deeply in the receptor. Therefore, and due to confounding
studies regarding the orientation of 1 at the D R, it was of interest to investigate the kinetic effects of
2
modifying both phenyl moieties of 1 simultaneously. Accordingly, we synthesised a further two
structural analogues of 1 (Figure 3). These modifications included swapping both aromatic termini
(53), as well as removal of these aromatic substituents (54). 53 was synthesised according to a
literature procedure following Friedel-Crafts acylation and N-alkylation (Figure 3, Supplementary
Scheme 6).43 Compound 54 was similarly accessed through literature methods (Figure 3,
Supplementary Scheme 7).41
Pharmacology. Characterisation of PPHT-red binding. Specific equilibrium binding of the agonist
PPHT-red (Cisbio Bioassays) to human D receptor (hD R) was saturable and best described by the
2
L
2L
interaction of PPHT-red with a single population of binding sites (Supplementary Figure 1A). From
these studies, the equilibrium dissociation constant (K ) of the fluorescent ligand was determined to
d
be 43.2 ± 0.37 nM. The binding kinetics of PPHT-red were characterised by monitoring the observed
1
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 78
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
association rates at six different ligand concentrations (Supplementary Figure 1B). The observed rate
of association was related to PPHT-red concentration in a linear fashion (Supplementary Figure 1C).
Kinetic rate parameters for PPHT-red were calculated by globally fitting the association time courses,
6
-1
-1
-1
resulting in a k of 9.21 ± 0.24 × 10 M min and k of 0.35 ± 0.01 min . The resulting K (k /k )
on
off
d
off on
of 46.3 ± 0.15 nM was comparable to that obtained from equilibrium studies.
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
Characterisation of kinetic binding parameters of unlabelled analogues of 1 at the D R. The
2
competition association binding method allows the characterisation of the kinetic rate parameters of
unlabelled compounds (k , k ) and the subsequent calculation of a kinetically derived (k /k )
on
off
on off
equilibrium dissociation constant (K ). The binding affinity of the various ligands for the hD R were
d
2L
measured at equilibrium at 37 ºC in a buffer containing 5′-guanylyl imidodiphosphate (GppNHp) (0.1
mM) to ensure that antagonist and tracer binding only occurred to the G protein-uncoupled form of
the receptor. K values for compound 1, and the 50 structural analogues studied are summarised in
i
Tables 1-5, and representative competition curves are presented in Figure 4A. In these tables we have
separated the analogues into five groups, those that have been modified at the para-chlorophenyl,
para-fluorophenyl, piperidinol, ketone/alkyl linker, and concurrent phenyl ring moiety modification,
as indicated in Figure 2. Representative kinetic competition curves for selected analogues are in
Figures 4B-D. Association curves for PPHT-red alone and in the presence of competitor were globally
fitted to Eq. 3 enabling the calculation of both k (k3) and k (k4) for each of the ligands, as reported
on
off
in Tables 1-5. To validate the rate constants, we compared the kinetically derived dissociation
constant (K ) values (k /k ) with the dissociation constant (K ) obtained from equilibrium
d
on off
i
competition binding experiments (Figure 5). There was a good correlation between these two values
2
for all compounds tested (two-tailed Pearson’s correlation r = 0.99, p < 0.0001), indicating that the
parameters determined in the kinetic assay were in agreement with those determined at equilibrium.
Characterisation of the kinetic profile of 1 at the hD R. The equilibrium affinities and kinetic rate
2
L
3
3
constants of 1 and 2 have recently been determined using the aforementioned TR-FRET assay. Prior
to initiating an investigation into 1, we also assessed its parameters and determined similar estimates
33
-1
9
-1
-1
in agreement with literature (k = 0.61 ± 0.04 min , k = 1.29 ± 0.21 × 10 M min , pK = 9.31
off
on
d
± 0.05, Table 1), validating our experimental conditions and further demonstrating that 1 is indeed a
high affinity, fast k /slow k compound at the hD R. Kinetic estimates for 1 are outlined in Tables
on
off
2L
1
-5 and all experimental structure-kinetic data will make specific reference to these data as a
-1
comparison. Furthermore, compounds with fast k values approaching >1.0 min were reassessed
off
using a modified injection protocol, whereby the hD R membrane homogenates were introduced
2
L
1
4
ACS Paragon Plus Environment

Page 15 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
using an online injector whilst simultaneously measuring TR-FRET binding. This is to avoid any
delay between membrane addition and initial TR-FRET measurement, improving the quality of the
non-linear fit for compounds with rapid equilibration kinetics and thus increasing our confidence in
the rate parameter estimate. Characterisation of 1 using this methodology returned comparable
estimates to the offline injection protocol. Additional data acquired for selected compounds using this
methodology are located in Tables 2 and 3.
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. Equilibrium and competition association binding. (A) Competition between PPHT-red
(12.5 nM) and increasing concentrations of 1 and representative analogues (8b, 8d, 8g, 8l, 14j, 14k,
4
5, 46, 48, 52, 53) at the hD R. PPHT-red competition association curves in the presence of (B) 30b;
2
L
(C) 34c; (D) 42. All binding reactions were performed at 37 ºC in the presence of GppNHp (100 µM)
with non-specific binding levels determined by inclusion of haloperidol (10 µM). Kinetic and
equilibrium data were fitted to the equations described in “Methods” section to calculate K , K , and
i
d
k and k values for the unlabelled ligands: these are summarised in tables 1-5. Data are presented
on
off
as singlet values from a representative of four.
1
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 78
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. Correlating equilibrium and kinetically derived parameters for haloperidol (1) and
1 structural analogues at the dopamine D receptor. Correlation between pK and kinetically
5
2
i
derived pK for the 51 test ligands including haloperidol. pK values were taken from PPHT-red
d
i
competition binding experiments at equilibrium as exemplified in Figure 4A. The values composing
the kinetically derived pK (k /k ) were taken from competition kinetic association experiments as
d
off on
exemplified in Figures 4B-D. All data used in these plots are detailed in Tables 1-5. Data are presented
as mean ± S.E.M. from four separate experiments.
Measurement of the functional activity of analogues of 1. 1 is a hD R antagonist. It is possible,
2
L
however, that modification of this structure may yield agonists. We measured the activity of all
analogues in an assay measuring inhibition of intracellular cAMP production stimulated by forskolin
using a bioluminescence resonance energy transfer (BRET) biosensor. This is a measurement of Gi/o
G protein activation by the hD R. We performed two types of measurement. In the first, we tested
2
L
the ability of a 10 µM concentration of each analogue alone to activate the hD R and in the second
2L
we measured the ability of each analogue to antagonise the action of an EC (30 nM) concentration
8
0
of the agonist dopamine. The results of these experiments revealed that none of the compounds
displayed agonist activity apart from 47 which displayed 20% of the maximal effect of dopamine at
a concentration of 10 µM (Supplementary Table 1). All compounds antagonised the effect of
dopamine to a basal (unstimulated) level except for 47 which reduced the effect of dopamine to a
level consistent with the intrinsic activity determined in the agonist assay protocol (Supplementary
Table 1).
Kinetic effects of variation of the para-chlorophenyl moiety of 1. Initially focusing on modification
of the para-chlorophenyl moiety of 1, we sought to assess the kinetic effect of all possible mono (8a,
1
6
ACS Paragon Plus Environment

Page 17 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
8
n) and di-chlorophenyl substituents (8b-g), as well as variation of the para- substituent (8h-m)
through the synthesis of 14 structural analogues (Table 1). These compounds exhibited a 17-fold
variation in affinity, which was driven by interesting changes in kinetic parameters, spanning a >10-
8
-1
-1
9
-1
-1
fold variation in association rate (k = 1.22 ± 0.20 × 10 M min to 2.95 ± 0.30 × 10 M min ),
on
-1
-1
and a ~4-fold variation in dissociation rate (k = 0.30 ± 0.01 min to k = 1.25 ± 0.09 min ).
off
off
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
The data show that analogues lacking an electron withdrawing group (EWG) (chloro) substituent at
the meta- and para-positions have reduced binding affinity, and this loss is mirrored by a decrease in
k and an increase in k relative to 1. For example, the ortho-Cl analogue (8n) displayed an ~8-fold
on
off
8
-1
reduction in affinity resulting from a decreased k and increased k (k = 3.54 ± 0.16 × 10 M
on
off
on
-1
-1
min , k = 1.16 ± 0.11 min ). This was also evident for the 2,6-diCl analogue (8e) losing ~6-fold
off
affinity, also mediated by a slowed association and increased dissociation rate (k = 5.07 ± 0.47 ×
on
8
-1
-1
-1
1
0 M min , k = 1.05 ± 0.05 min ). This trend continued with the des-Cl analogue 8h, as it also
off
8
-1
-1
-
revealed a similar change in rate constants (k = 1.22 ± 0.20 × 10 M min , k = 1.02 ± 0.10 min
on
off
1).
Addition of a strong electron donating group (EDG) (N,N-dimethylamino, 8k) results in a >10-fold
decrease in affinity (pK = 8.12 ± 0.04) and again appears to be driven by a decrease in k and an
d
on
8
-1
-1
-1
increase in koff (k = 1.64 ± 0.12 × 10 M min , k = 1.25 ± 0.09 min ). Furthermore, other
on
off
analogues bearing weakly electron donating substituents (e.g. para-tolyl analogue 8i) saw a smaller
decrease in affinity (~3-fold), similarly mediated by a change in both rate constants towards a slow
9
-1
-1
-1
on, fast off profile (k = 1.00 ± 0.06 × 10 M min , k = 0.98 ± 0.02 min ).
on
off
Conversely, insertion of a meta-Cl substituent (exemplified by 8a), despite decreasing affinity ~8-
8
-1
-1
fold, acts only to decrease the k whilst having no effect on k (k = 3.62 ± 0.94 × 10 M min ,
on
off
on
-1
k = 0.64 ± 0.10 min ), and this similarly applies to para-Cl substituents. The trend continued with
off
2
,4-dichloro (8c) and 2,5-dichloro (8d) analogues, losing ~7-fold and 3-fold affinity, respectively.
8
-
Again, this loss was largely mediated by a decreased association rate (8c: k = 2.87 ± 0.56 × 10 M
on
1
-1
-1
8
-1
-1
-1
min , k = 0.70 ± 0.16 min ; 8d: k = 5.29 ± 0.36 × 10 M min , k = 0.56 ± 0.06 min ), relative
off
on
off
to 1. Interestingly, when the ortho- and meta-chloro substituents are combined (2,3-diCl analogue
8b)), affinity increases ~5-fold, and this is now predominantly mediated by both a ~2-fold increase
in association rate and ~2-fold decrease in dissociation rate (pK = 9.84 ± 0.08, k = 2.20 ± 0.30 ×
(
d
on
9
-1
-1
-1
1
0 M min , k = 0.30 ± 0.01 min ). Substitution with a strongly electron withdrawing para-CF
off 3
substituent (8j) maintained affinity, with no effect on the kinetic profile of the analogue relative to 1
9
-1
-1
-1
(k = 1.36 ± 0.07 × 10 M min , k = 0.62 ± 0.02 min ). Furthermore, replacing the para-chloro
on off
substituent for a para-fluoro (8l) predominantly decreased kon.
1
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 78
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
All compounds bearing an ortho-substituent (8n, 8c, 8d, 8e, 8m), with the exception of 8b, displayed
a reduced on-rate, indicating potential sensitivity to steric bulk at this position through resulting
rotation of the phenyl group relative to the piperidinol. Interestingly, the 2,3-diCl analogue (8b),
contains the privileged 2,3-dichlorophenylpiperidine pharmacophore known to confer high affinity
in other molecules at both the D -like and 5HT receptors. This particular substitution pattern may
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
therefore support a different binding mode. Both increased lipophilicity and steric bulk are preferred
at the meta- and para–positions of the ring, with the 4-position being optimal, which is supported by
8
h (4-fluoro) and 8l (4-H) being less favoured. For the off rate, the substituent effect is reversed in
terms of increasing k (o>m>p). This parameter appears to be less impacted by steric factors, and
off
instead the electronics may play a greater role (8k, 8i, 8h). In summary, these initial data provide
insight into how structural modifications of haloperidol (1) impact upon individual kinetic
parameters, demonstrating the potential for differential modification of rate constants towards a slow
on, fast off profile, depending on the position and nature of the aryl substituents of the 4-
phenylpiperidin-4-ol moiety.
Table 1. Kinetic binding parameters for haloperidol (1) and unlabelled analogues of 1 for
human D receptors estimated using TR-FRET assay.
2
L
OH
R2
O
N
F
Haloperidol (1), 8a-8n
R2
k (M min )
-1
-1 a
k (min )
-1
a
t1/2 (min) a
pKd a
pKi b
on
off
1
Cl
1.29 ± 0.21 × 109 0.61 ± 0.04
6.32 ± 0.94 × 108 0.64 ± 0.10
1.15 ± 0.08 9.31 ± 0.05
1.19 ± 0.22 9.00 ± 0.05
9.33 ± 0.09
8a
9.00 ± 0.03
8.67 ± 0.09
9.92 ± 0.03
8.58 ± 0.01
Cl
8
8
n
b
3.54 ± 0.16 × 108 1.16 ± 0.11
2.20 ± 0.30 × 109 0.30 ± 0.01
2.87 ± 0.56 × 108 0.70 ± 0.16
0.61 ± 0.05 8.49 ± 0.03
2.28 ± 0.04 9.84 ± 0.08
1.15 ± 0.24 8.62 ± 0.02
Cl
Cl
Cl
Cl
8c
Cl
1
8
ACS Paragon Plus Environment

Page 19 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Cl
8d
5.29 ± 0.36 × 108 0.56 ± 0.06
1.27 ± 0.11 8.98 ± 0.02
8.99 ± 0.01
Cl
Cl
8
8
e
f
5.07 ± 0.47 × 108 1.05 ± 0.05
2.95 ± 0.30 × 109 0.42 ± 0.02
6.58 ± 0.67 × 108 0.43 ± 0.02
0.67 ± 0.05 8.68 ± 0.04
1.67 ± 0.11 9.84 ± 0.03
1.94 ± 0.30 9.27 ± 0.06
8.83 ± 0.08
9.81 ± 0.04
9.35 ± 0.04
Cl
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
Cl
Cl
Cl
Cl
8
8
g
h
1.22 ± 0.20 × 108 1.02 ± 0.10
1.00 ± 0.06 × 109 0.98 ± 0.11
1.36 ± 0.07 × 109 0.62 ± 0.02
1.64 ± 0.12 × 108 1.25 ± 0.09
0.71 ± 0.08 8.07 ± 0.04
0.73 ± 0.07 9.02 ± 0.03
1.12 ± 0.04 9.34 ± 0.04
0.56 ± 0.03 8.12 ± 0.04
8.12 ± 0.02
9.14 ± 0.08
9.49 ± 0.10
8.28 ± 0.09
8i
8
8
j
CF3
k
N
F
2.61 ± 0.24 × 108 0.78 ± 0.08
3.25 ± 0.77 × 108 0.82 ± 0.11
0.92 ± 0.11 8.52 ± 0.02
0.89 ± 0.12 8.57 ± 0.05
8.48 ± 0.01
8.61 ± 0.03
8
8
l
m
Cl
a
The rate constants k , k , the half-life (t ), and the kinetically derived pK were obtained from competition
off on
1/2
d
b
kinetic association experiments using PPHT-red. pK values were taken from PPHT-red competition binding
i
experiments at equilibrium. Data are presented as mean ± S.E.M. from four experiments performed in singlet.
*
Completed using online injection protocol.
Kinetic effects of variation of the para-fluorophenyl moiety of 1. We examined the effect of fluoro
substituents at both ortho-(14a) and meta-(14b) positions of the phenone moiety, as well as all
possible di-fluorophenyl substituents (14c-h), together with three additional ortho-analogues (o-Cl
(
14i), o-CH (14j) o-Cl (14k)), and an unsubstituted analogue (14l). Modification to this moiety
3
caused large decreases in affinity relative to 1, spanning over 100-fold from pK = 6.67 ± 0.01 (14h)
d
to pK = 8.75 ± 0.02 (14l), and is associated with a wide range of association and dissociation rate
d
constants. These losses in affinity are mediated through concurrent changes in both k and k . This
on
off
applies to all but the para-Cl analogue (14k), as it lost affinity 10-fold relative to 1, but this was
8
-1
largely mediated by a decreased rate of association (k = 1.38 ± 0.05 × 10 , k = 0.70 ± 0.03 min ).
on
off
1
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 78
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
The des-fluoro analogue (14l) maintained the highest affinity, and similar to the previous series, this
8
-1
-1
was facilitated by a shift in both rate constants (pK = 8.75 ± 0.02, k = 6.33 ± 1.06 × 10 M min ,
d
on
-1
k = 1.12 ± 0.18 min ). Of the three ortho-substituted analogues (14a (m-F), 14i (m-Cl)), 14j (m-
off
CH )), the fluoro substituent was the least favourable in terms of affinity, decreasing ~13-fold relative
3
to 1, whereas the ortho-tolyl substituent only reduces affinity by 6-fold. However, these changes are
likewise mediated by a decreased association rate and increased rate of dissociation. Notably, the m-
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
Cl (14i) and m-CH (14j) substituents have similar Van der Waals radii, but very different electronic
3
effects, thus highlighting a steric factor as being important. The meta-fluoro substituted analogue
(
14b) also dramatically reduced the affinity and was similarly driven by a decreased k and increased
on
7
-1
-1
-1
k (k = 2.55 ± 0.27 × 10 M min k = 1.08 ± 0.21 min ).
off
on
, off
Di-fluoro substitution of the phenyl ring revealed no clear SKR and commonly caused substantial
losses in binding affinity. However, unlike the previous chloro series, greater increases in the rate of
dissociation were observed. Interestingly, using our online injection protocol, we identified
compounds with even slower k values relative to 2, coupled with equal to or faster k values, despite
on
off
their affinities being lower than 2. For example, the 2,3-(14c), 2,4-(14d) and 2,5-difluoro (14e)
analogues of 1 (pK = 7.28 ± 0.04, 6.92 ± 0.05 and 6.85 ± 0.04, respectively) showed dissociation
d
-1
rates faster than any compound identified in the previous series (k = 1.70 ± 0.09 min , k = 1.36 ±
off
off
-1
-1
0.21 min and k = 1.49 ± 0.36 min , respectively). In conclusion, these preliminary data suggest
off
that different fluorine substitution patterns dramatically reduce binding affinities, mediated through
changes in both kinetic parameters. However, the relationship between the nature of substituents, the
substitution pattern and the corresponding kinetic profile is unclear.
Table 2. Kinetic binding parameters of unlabelled analogues of 1 with modification to the
para-fluorophenyl moiety for human D receptors estimated using TR-FRET assay.
2
L
OH
O
Cl
N
R6
Haloperidol (1), 14a-l
R6
k (M min )
-1
-1
k (min )
-1
t1/2 (min)
pKd
pKi
on
off
1
1
1.29 ± 0.21 × 109
9.93 ± 1.44 × 107
0.61 ± 0.04
1.04 ± 0.11
1.15 ± 0.08
9.31 ± 0.05
9.33 ± 0.09
F
4a
0.69 ± 0.02
7.99 ± 0.05
8.10 ± 0.15
F
2
0
ACS Paragon Plus Environment

Page 21 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
14b
2.55 ± 0.27 × 107
0.96 ± 0.13
1.08 ± 0.21
1.29 ± 0.26
1.70 ± 0.09
0.76 ± 0.09
7.43 ± 0.03
7.39 ± 0.01
7.38 ± 0.02
7.28 ± 0.04
7.44 ± 0.03
7.42 ± 0.02
7.39 ± 0.04
7.33 ± 0.04
F
F
F
F
1
1
1
4b*
2.64 ± 0.49 × 107
3.01 ± 0.47 × 107
3.24 ± 0.34 × 107
0.70 ± 0.12
0.60 ± 0.10
0.41 ± 0.10
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
4c
F
F
4c*
1
1
4d
4e
1.07 ± 0.02 × 107
1.15 ± 0.20 × 107
1.13 ± 0.38 × 107
1.36 ± 0.21
1.14 ± 0.15
1.49 ± 0.36
0.51 ± 0.07
0.63 ± 0.08
0.58 ± 0.16
6.92 ± 0.05
7.00 ± 0.04
6.85 ± 0.04
7.08 ± 0.10
7.12 ± 0.12
6.95 ± 0.12
F
F
F
F
F
F
14e*
F
1
1
4f
1.84 ± 0.26 × 107
1.36 ± 0.21 × 108
4.25 ± 0.52 × 106
1.09 ± 0.12
1.03 ± 0.20
0.92 ± 0.13
0.66 ± 0.09
0.75 ± 0.14
0.79 ± 0.09
7.22 ± 0.05
8.13 ± 0.03
6.67 ± 0.01
7.20 ± 0.04
8.11 ± 0.03
6.77 ± 0.08
F
4g
F
F
F
F
14h
1
4i
3.16 ± 0.54 × 108
3.66 ± 0.91 × 108
1.38 ± 0.05 × 108
6.33 ± 1.06 × 108
1.11 ± 0.25
0.83 ± 0.21
0.70 ± 0.03
1.12 ± 0.18
0.76 ± 0.22
1.01 ± 0.25
1.00 ± 0.04
0.67 ± 0.11
8.47 ± 0.03
8.71 ± 0.25
8.30 ± 0.02
8.75 ± 0.02
8.45 ± 0.01
8.82 ± 0.02
8.30 ± 0.02
8.75 ± 0.05
Cl
14j
1
1
4k
4l
Cl
a
The rate constants k , k , the half-life (t ), and the kinetically derived pK were obtained from competition
off on
1/2
d
b
kinetic association experiments using PPHT-red. pK values were taken from PPHT-red competition binding
i
experiments at equilibrium. Data are presented as mean ± S.E.M. from four experiments performed in singlet.
*
Completed using online injection protocol.
2
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 78
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
Kinetic effects of variation of ketone and linker moieties of 1. We next examined the effect of
modification to the linker and ketone moieties of 1 through synthesis of a further 15 analogues.
Specific linker-modified compounds included propiophenone (34a) and valerophenone (34c)
analogues of 1, alongside 3-5 carbon alkyl (36a-c) and alkyne analogues (41a-c). In addition, a
thorough analysis of modification to the ketone moiety was undertaken via synthesis of geometric
olefin isomers (23, 26) and their corresponding cyclopropane derivatives (29b, 30b), through to
isosteric replacement with sulfur (17a) or oxygen (17b), as well as conversion of the ketone to the
corresponding secondary alcohol (18).
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
All compounds in this series lost binding affinity relative to 1 and, for the most part, this was mediated
through a decrease in k and an increase in k . Converting the ketone to its corresponding secondary
on
off
alcohol (18) (racemic), whilst engendering a 13-fold reduction in affinity compared to 1, was
6
-1
-1
exclusively caused by a slowed k (pK = 7.04 ± 0.01, k = 6.19 ± 0.41 × 10 M min ). Replacement
on
d
on
of the carbonyl moiety with sulfur (17a) or oxygen (17b) modulated both kinetic binding parameters,
8
-1
-1
though their respective association rates varied ~6-fold (k = 4.99 ± 0.59 × 10 M min and k =
on
on
9
-1
-1
1
.22 ± 0.35 × 10 M min , respectively). This difference may be due to a number factors, including
the electronegativity and size difference between the sulfur and oxygen atoms, the relatively longer
S-C bond length compared to that of the O-C bond and the orbital arrangement around each
heteroatom (resulting in considerably smaller bond angles in the thioether compared to the ether).
The trans alkene (23) lost ~10-fold affinity relative to 1, and this was again predominantly due to a
7
-1
-1
decreased k (k = 6.39 ± 0.88 × 10 M min ). Interestingly, the cis- isomer (26) saw a further 5-
on
on
fold reduction in affinity (pK = 7.49 ± 0.12); however, this was predominantly due to a change in
d
association rate, displaying a k almost 20-fold slower and a k 2-fold faster than 1 (k = 3.35 ±
on
off
on
7
-1
-1
-1
0
.72 × 10 M min , k = 1.25 ± 0.16 min ). Analysis of the racemic cycloalkane diastereomers was
off
also interesting; introduction of the trans-cyclopropane (29b) resulted in a ~10-fold increase in affinity
relative to the parent trans-olefin 23, which was predominantly due to a ~10-fold increase in
8
-1
-1
association rate (k = 6.07 ± 0.87 × 10 M min ). Conversely, introduction of the cis-cyclopropane
on
(30b) had no effects on affinity relative to the parent cis-olefin 28; however, this substituent
7
-1
-1
marginally decreased k whilst increasing the k (k = 4.47 ± 0.47 × 10 M min , k = 1.37 ±
on
off
on
off
-1
0
.09 min ). These data indicate that cis-geometry is preferred as opposed to trans- with respect to
this sub-set of compounds in reference to tuning the kinetic profile towards "slow on, fast off”
characteristics, and demonstrates the importance of geometry in the corresponding pharmacological
profile of APDs. Analysis of the propiophenone and valerophenone analogues of 1 returned further
intriguing results. Decreasing the linker length by just one carbon (34a) relative to 1 resulted in
2
2
ACS Paragon Plus Environment

Page 23 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
7
-1
dramatic changes in both association and dissociation rate constants (k = 1.33 ± 0.17 × 10 M
on
-1
-1
min , k = 1.95 ± 0.32 min ), resulting in a loss of affinity at the D R by >20-fold (pK = 6.84 ±
off
2
d
0
.05).
Perhaps the most exciting compound to arise from our study was the valerophenone analogue (34c).
Despite losing affinity by >10-fold relative to 1, this compound displayed a ~10-fold slower k and
on
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
a >3.5-fold faster k than 1. Both the kinetic profile and affinity are similar to that of 2, which our
off
3
3
previous studies predict would confer a low propensity to cause extrapyramidal side effects. The
alkane analogues of 1 (36a-c) exhibited a 10-fold variation in affinity with respect to one another,
with the butylene analogue (36b) found to be optimal in terms of affinity conservation relative to 1
(
pK = 8.17 ± 0.03), despite all having >10-fold losses in affinity relative to 1. Despite having a >20-
d
fold lower affinity compared to 1, the propylene analogue (36a) was found to have a “slow on, fast
7
-1
-1
-1
off” kinetic profile (pK = 7.18 ± 0.06, k = 2.29 ± 0.15 × 10 M min , k = 1.54 ± 0.07 min ).
d
on
off
Finally, analysis of the 3-5-carbon alkyne analogues (41a-c) saw a 10-fold variation in affinity, with
the pentyne analogue (41c) being optimal (pK = 7.75 ± 0.03), as well as displaying the largest change
d
7
in both rate constants towards a slow on, fast off profile (k = 6.41 ± 0.80 × 10 , k = 1.15 ± 0.15
on
off
-
1
min ).
Table 3. Kinetic binding parameters of unlabelled analogues of 1 with modifications to the
ketone and linker moieties for human D receptors estimated using TR-FRET assay.
2L
OH
O
Cl
N
F
Haloperidol (1), 17a-b, 18, 23, 26,
29b, 30b, 34a,c, 36a-c, 41a-c
-1
-1
-1
Structure
k (M min )
k (min )
t1/2 (min)
pKd
pKi
on
off
O
1
1
1.29 ± 0.21 × 109
0.61 ± 0.04
1.15 ± 0.08
9.31 ± 0.05
9.33 ± 0.09
S
4.99 ± 0.59 × 108
1.22 ± 0.35 × 109
7a
0.99 ± 0.15
0.91 ± 0.12
0.74 ± 0.10
0.81 ± 0.11
8.71 ± 0.04
9.10 ± 0.08
8.68 ± 0.04
8.97 ± 0.04
O
17b
OH
1
2
8
3
6.19 ± 0.41 × 106
6.39 ± 0.88 × 107
0.57 ± 0.05
0.76 ± 0.13
1.25 ± 0.13
0.74 ± 0.10
7.04 ± 0.01
7.94 ± 0.03
7.04 ± 0.02
7.95 ± 0.03
(E)
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 78
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
2
6
3.35 ± 0.72 × 107
1.25 ± 0.16
0.92 ± 0.14
1.37 ± 0.09
1.27 ± 0.15
0.58 ± 0.09
7.49 ± 0.12
8.82 ± 0.04
7.51 ± 0.03
6.89 ± 0.05
7.53 ± 0.06
(
Z)
(
±)-trans
29b
6.07 ± 0.87 × 108
4.47 ± 0.47 × 107
1.03 ± 0.18 × 107
0.82 ± 0.14
0.51 ± 0.03
0.57 ± 0.07
8.82 ± 0.04
7.53 ± 0.03
6.92 ± 0.02
(±)-cis
3
3
3
3
3
0b
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
O
O
4a
4a*
4c
1.33 ± 0.17 × 107
1.42 ± 0.24 × 108
1.80 ± 0.15 × 108
1.95 ± 0.32
1.65 ± 0.30
2.35 ± 0.19
0.43 ± 0.04
0.48 ± 0.10
0.30 ± 0.03
6.84 ± 0.05
7.94 ± 0.04
7.89 ± 0.01
6.88 ± 0.02
7.97 ± 0.02
7.92 ± 0.02
O
O
4c*
3
3
3
3
6a
2.45 ± 0.26 × 107
2.29 ± 0.15 × 107
1.57 ± 0.13 × 108
5.51 ± 0.69 × 107
1.33 ± 0.22
1.54 ± 0.07
1.11 ± 0.05
1.22 ± 0.15
0.57 ± 0.10
0.46 ± 0.10
0.63 ± 0.02
0.60 ± 0.08
7.28 ± 0.06
7.18 ± 0.06
8.17 ± 0.03
7.66 ± 0.03
7.25 ± 0.03
7.21 ± 0.03
8.27 ± 0.09
7.66 ± 0.02
6a*
6b
6c
41a
41b
3.37 ± 0.62 × 106
1.61 ± 0.24 × 107
6.41 ± 0.80 × 107
0.96 ± 0.14
1.10 ± 0.21
1.15 ± 0.15
0.77 ± 0.11
0.69 ± 0.11
0.65 ± 0.11
6.54 ± 0.01
7.15 ± 0.05
7.75 ± 0.03
6.55 ± 0.01
7.13 ± 0.06
7.72 ± 0.04
41c
a
The rate constants k , k , the half-life (t ), and the kinetically derived pK were obtained from competition
off on
1/2
d
b
kinetic association experiments using PPHT-red. pK values were taken from PPHT-red competition binding
i
experiments at equilibrium. Data are presented as mean ± S.E.M. from four experiments performed in singlet.
*
Completed using online injection protocol.
Kinetic effects of variation of the piperidinol moiety of 1. The kinetic effect of structural modifications
to the 4-phenylpiperidin-4-ol moiety of 1 was explored through the synthesis of eight additional
analogues. We observed the effects of introducing an ethylene bridge (42), as well as modification
primarily to the tertiary alcohol through methyl ether formation (52) and its subsequent removal,
generating a variety of compounds containing piperazinyl (43), dihydropyridinyl (45), cyclopropyl
(
46), and piperidinyl (47, 48) functionalities. We observed a wide range of affinities that spanned a
~
30-fold difference, and unlike the previous chemical series, modification to the piperidinol moiety
2
4
ACS Paragon Plus Environment

Page 25 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
for the most part had relatively negligible effects on the k , with the majority maintaining similar
off
values to that of 1 (Table 3). Instead, a decrease in affinity relative to 1 was largely facilitated by a
decreased k . Notably, of the two analogues with higher affinities relative to 1, these were instead
on
largely mediated by an increase in k and decrease in k . For example, introducing the tropanyl
on
off
moiety (42) conferred a ~10-fold increase in affinity which was equally driven by an increase in kon
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
9
-1
-1
-1
and decrease in k (pK = 10.26 ± 0.06, k = 3.68 ± 0.64 × 10 M min , k = 0.19 ± 0.02 min ).
off
d
on
off
The cyclopropane variants (46) 5-fold improved affinity relative to 1 was also mediated by an
9
-1
-1
increased k and decreased k (pK = 9.84 ± 0.02, k = 2.03 ± 0.09 × 10 M min , k = 0.30 ±
on
off
d
on
off
-1
0
.01 min ). The improved affinities and decreased dissociation rates of 42 and 46 (tropanyl and
cyclopropane analogues, respectively) can perhaps be rationalised through a major conformational
difference induced by these substituents, resulting in a more entropically favourable binding event.
From these preliminary data, it appears that modification to the piperidinol moiety is not particularly
amenable to significant increases in the corresponding compounds rate of dissociation.
Table 4. Kinetic binding parameters of unlabelled analogues of 1 with modifications to the
piperidinol moiety for human D receptors estimated using TR-FRET assay.
2
L
O
R7
F
Haloperidol (1), 42-43, 45-48, 52
-1
-1
-1
#
R7
k (M min )
k (min )
off
t1/2 (min)
pKd
pKi
on
Cl
OH
1
1.29 ± 0.21 × 109 0.61 ± 0.04
3.68 ± 0.64 × 109 0.19 ± 0.02
2.86 ± 0.33 × 107 0.80 ± 0.07
6.52 ± 0.46 × 107 0.65 ± 0.05
1.15 ± 0.08
9.31 ± 0.05
9.33 ± 0.09
N
Cl
OH
10.26 ±
42
3.59 ± 0.39
0.89 ± 0.09
1.09 ± 0.09
10.28 ± 0.08
7.53 ± 0.04
8.00 ± 0.02
0
.06
N
Cl
4
4
3
5
N
7.55 ± 0.04
8.00 ± 0.02
N
N
Cl
2
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 78
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
Cl
4
4
4
6
7
8
2.03 ± 0.09 × 109 0.30 ± 0.01
8.99 ± 0.37 × 107 0.71 ± 0.05
4.31 ± 0.37 × 108 0.84 ± 0.09
3.60 ± 0.22 × 108 0.68 ± 0.03
2.35 ± 0.10
9.84 ± 0.02
8.10 ± 0.02
8.72 ± 0.03
8.72 ± 0.01
9.80 ± 0.02
N
0.99 ± 0.07
0.86 ± 0.09
1.03 ± 0.05
8.11 ± 0.03
8.76 ± 0.06
8.72 ± 0.01
N
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
Cl
Cl
N
N
O
52
a
The rate constants k , k , the half-life (t ), and the kinetically derived pK were obtained from competition
off on
1/2
d
b
kinetic association experiments using PPHT-red. pK values were taken from PPHT-red competition binding
i
experiments at equilibrium. Data are presented as mean ± S.E.M. from four experiments performed in singlet.
Dual modifications to both phenyl moieties of 1. Finally, we assessed the effect of swapping the
halogen substituents on each end of haloperidol (1) through compound (53), as well as their
simultaneous removal as exemplified by the des-halo analogue 54 (Table 5). These structural changes
all decreased affinity, which was reflected by decreases in the corresponding k , with only minor
on
effects on k relative to 1. Swapping the halogen atoms on each ring (53) caused a 16-fold loss in
off
affinity (pK = 7.73 ± 0.02), which was predominantly driven by a 16-fold decrease in k (k = 4.17
d
on
on
7
-1
-1
± 0.28 × 10 M min ). Finally, removal of both halogen atoms (54) simultaneously caused a ~18-
fold loss in affinity (pK = 7.51 ± 0.01), driven by a sole ~18-fold decrease in k relative to 1 (k =
d
on
on
7
-1
-1
2
.28 ± 0.13 × 10 M min ). This effect is unlike that of previous analogues bearing a para-halo
substituent on only one of the two phenyl rings (8h and 14l), whereby both k and k are altered
on
off
(tables 1 and 2, respectively).
Table 5. Kinetic binding parameters of unlabelled bi-functionalised analogues of 1 for human
D receptors estimated using TR-FRET assay.
2
L
OH
R2
O
N
R6
Haloperidol (1), 53, 54
2
6
ACS Paragon Plus Environment

Page 27 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
t1/2
R6
R2
k (M min )
-1
-1
k (min )
-1
pKd
pKi
#
on
off
(
min)
1.15 ±
0.08
1
Cl
F
1.29 ± 0.21 × 109 0.61 ± 0.04
4.17 ± 0.28 × 107 0.77 ± 0.05
2.28 ± 0.13 × 107 0.71 ± 0.05
9.31 ± 0.05 9.33 ± 0.09
7.73 ± 0.02 7.75 ± 0.01
7.51 ± 0.01 7.49 ± 0.01
F
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
0
.91 ±
53
Cl
0.06
0
.98 ±
5
4
0.07
a
The rate constants k , k , the half-life (t ), and the kinetically derived pK were obtained from competition
off on
1/2
d
b
kinetic association experiments using PPHT-red. pK values were taken from PPHT-red competition binding
i
experiments at equilibrium. Data are presented as mean ± S.E.M. from four experiments performed in singlet.
*
Completed using online injection protocol.
Our studies show that modifying the scaffold of 1 produces compounds with a wide range of both
6
-1
-1
association rates (spanning ~3 orders of magnitude, from k = 3.37 ± 0.62 × 10 M min to 3.68 ±
on
9
-1
-1
-1
0
.64 × 10 M min ) and dissociation rates (spanning >10-fold, from k = 0.19 ± 0.02 min to 2.35
off
-1
± 0.19 min ), which constituted large variations in hD R affinities (spanning over three orders of
2
L
magnitude from K = 288 nM to 0.0549 nM). To further understand the relationship between kinetic
d
rate constants and the affinity of D R ligands, we have correlated the kinetic binding data of these 50
2
compounds (k , k ) with the derived equilibrium affinity estimates (pK ) (Figure 6A). Our data
on
off
d
2
confirms that pK is robustly correlated with association rate (see Figure 6A, Spearman’s r = 0.96, p
d
>
0.0001), whereas pK is, to a much lesser extent, correlated with dissociation rate (Figure 6B).
d
These data are in contrast to previous studies claiming the differences in APD affinities are
1
9
determined entirely by how fast they dissociate from the D R. This is due to the fact that association
2
rates have widely been assumed to be diffusion limited. Indeed, studies conducted at other systems,
namely the M muscarinic acetylcholine and A adenosine receptors, have found correlations
3
2A
between koff values and affinity.62-64 However, the association rate constants of a series of
metabotropic glutamate receptor 2 positive allosteric modulators were found to be strongly correlated
2
2
to affinity, whereas dissociation rate constants were not. This correlation has also been observed at
6
5,66
the orexin OX receptor and β -adrenoreceptors for ligands with distinct chemotypes.
2
2
It is evident that modification to the scaffold of 1 and the corresponding changes in affinity are
principally mediated by a change in the rate of association (Figure 6A). Though, our study highlights
that particular structural moieties of 1 are more appropriate for the modification of both kinetic
parameters towards a “slow on, fast off” profile. For example, when modification to the piperidinol
moiety caused a loss in binding affinity relative to 1, this was predominantly k mediated, whilst
on
2
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 78
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
having negligible effects on k . However, modification of the p-fluorophenyl or linker moieties and
off
subsequent losses in affinity saw greater changes in both kinetic rate constants, highlighting these
areas as a focal point for future SKR investigations. In addition, we were able to derive preliminary
SKR for the p-chlorophenyl moiety of 1. From our kinetic data obtained from a limited amount of
compound structural/chemical diversity, we determined that both the electronic nature and position
of substituents on the aromatic ring dictate the corresponding kinetic profile. We found that meta-
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
and para-EWG groups (depending on compound affinity), can either slow the k whilst having no
on
effect on k (8a, 8c, 8d), or equally, slow the k whilst increasing k (8b, 8f, 8g). Conversely,
off
on
off
compounds bearing ortho-Cl substituents and that are not meta- or para-substituted, act to slow the
k but increase the k (8e, 8n). This is also true for para-EDG substituents at these positions (8h, 8i,
on
off
8k). It may be possible to use such molecules as templates in an attempt to further increase affinity
via decoration of the aromatic termini, whilst maintaining an ‘attractive’ or slow on, fast off kinetic
profile.
A
B
pK vs. log k
off
pK vs. log k
d
d
on
10.5
10.0
1
1
0.5
0.0
9.5
9.0
9
.5
r2 = 0.97
9.0
8
8
7
7
.5
.0
.5
.0
8
8
7
7
6
6
.5
.0
.5
.0
.5
.0
r2 = 0.34
6.5
6.0
5
.5
-
5.5
0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
6
7
8
9
10
log kon (M-1 min )
-1
-1
log koff (min )
C
D
cLogP vs. log kon
log koff vs. log kon
0
.1 nM
7
6
5
4
3
10
Chlorpromazine
1
1
nM
Haloperidol
r2 = 0.007
9
0 nM
34c
8
Clozapine
30b
100 nM
14c
6a
3
7
1
000 nM
6
1
0,000 nM
5
6
7
8
9
10
-0.75
-0.50
-0.25
0.00
0.25
0.50
log kon (M-1 min )
-1
-1
log koff (min )
Figure 6. Correlating kinetically derived equilibrium dissociation constants vs. kinetic rate
constants of haloperidol (1) and 51 structural analogues at the dopamine D receptor. (A) A
2
plot of log k vs. pK demonstrates a statistically significant correlation (two-tailed Pearson’s
on
d
2
8
ACS Paragon Plus Environment

Page 29 of 78
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
correlation r = 0.96, p = < 0.0001) between these two variables. (B) Conversely, a plot of pK vs. log
d
2
k demonstrates a much poorer correlation (two-tailed Pearson’s correlation r = 0.34, p = < 0.0001)
off
despite the traditional scientific consensus that APD affinity is solely driven by changes in k . (C)
off
The observed association rate (log k ) and calculated partition coefficient (cLogP) show no
on
2
correlation (two-tailed Pearson’s correlation r = 0.007, p = 0.562). The central line corresponds to
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
the linear regression of the data, the dotted lines represent the 95% confidence intervals for the
regression. (D) Representing the diversity in affinity and corresponding kinetic profiles for analogues
of 1. A plot of log k vs. log k represents a spectrum of compounds with various kinetic profiles
off
on
3
3
(
~10-fold difference in k , ~30-fold difference in k ) identified from this study. Clozapine (2) and
off on
33
typical APD chlorpromazine are also included as reference points. These data identify several
compounds with interesting ‘slow on, fast off’ kinetic profiles (14c, 30b, 34c, 36a). Combinations of
k and k that result in identical affinity (K ) values are represented by diagonal dotted lines. All
on
off
d
data used in these plots apart from chlorpromazine are detailed in Tables 1-5. Data are presented as
mean from at least four separate experiments.
The derived association rate of all compounds was further assessed for any potential correlation with
physicochemical parameters such as clogP (Figure 6C) and topological polar surface area (tPSA)
(Supplementary Figure 2), to which there was found to be no relationship. This is unsurprising as this
study places particular emphasis on the kinetics of not only positional isomers between subsets of
compounds, but close structural analogues which display very similar properties of size, lipophilicity
and polarity. This further provides evidence that the observed changes to affinity and kinetic profile
are not simply due to modification of physicochemical properties. These data are in contrast to
previous observations at the D R reporting that compounds with fast dissociation rates are less
2
lipophilic and have lower molecular weights.67 This is notable as additional micro-
pharmacokinetic/pharmacodynamic mechanisms, such as ligand binding to the cell membrane, are
6
8
known to play a role in target binding kinetics. Although it is widely accepted that increasing
lipophilicity results in increased affinity, this study shows that for this subset of compounds this is
not the case, highlighting that careful analysis of kinetic parameters is essential and also likely to be
context/target dependent.
Our recent proposal to expand the kinetic hypothesis for APD side effects considers not only the
dissociation rate (and therefore the propensity to display insurmountable antagonism), but the
3
3
association rate and subsequent potential for receptor rebinding. Based on this hypothesis, we
proposed three broad classes of APDs in an attempt to explain how different kinetic characteristics
have the potential to influence on-target side effects. Class 1: fast on/slow off compounds exemplified
2
9
ACS Paragon Plus Environment