Optimization of 5-substituted thiazolyl ureas and 6-substituted imidazopyridines as potential HIV-1 latency reversing agents

Article abstract of DOI:10.1016/j.ejmech.2020.112254

A persistent latent reservoir of virus in CD4⁺ T cells is a major barrier to cure HIV. Activating viral transcription in latently infected cells using small molecules is one strategy being explored to eliminate latency. We previously described the use of a FlpIn.FM HEK293 cellular assay to identify and then optimize the 2-acylaminothiazole class to exhibit modest activation of HIV gene expression. Here, we implement two strategies to further improve the activation of viral gene expression and physicochemical properties of this class. Firstly, we explored rigidification of the central oxy-carbon linker with a variety of saturated heterocycles, and secondly, investigated bioisosteric replacement of the 2-acylaminothiazole moiety. The optimization process afforded lead compounds (74 and 91) from the 2-piperazinyl thiazolyl urea and the imidazopyridine class. The lead compounds from each class demonstrate potent activation of HIV gene expression in the FlpIn.FM HEK293 cellular assay (both with LTR EC₅₀s of 80 nM) and in the Jurkat Latency 10.6 cell model (LTR EC₅₀ 220 and 320 nM respectively), but consequently activate gene expression non-specifically in the FlpIn.FM HEK293 cellular assay (CMV EC₅₀ 70 and 270 nM respectively) manifesting in cellular cytotoxicity. The lead compounds have potential for further development as novel latency reversing agents.

Full text of DOI:10.1016/j.ejmech.2020.112254

Journal Pre-proof
Optimization of 5-substituted thiazolyl ureas and 6-substituted imidazopyridines as
potential HIV-1 latency reversing agents
William Nguyen, Jonathan Jacobson, Kate E. Jarman, Timothy R. Blackmore, Helene
Jousset Sabroux, Sharon R. Lewin, Damian F. Purcell, Brad E. Sleebs
PII:
S0223-5234(20)30221-X
DOI:
https://doi.org/10.1016/j.ejmech.2020.112254
EJMECH 112254
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 30 December 2019
Revised Date: 13 February 2020
Accepted Date: 17 March 2020
Please cite this article as: W. Nguyen, J. Jacobson, K.E. Jarman, T.R. Blackmore, H.J. Sabroux,
S.R. Lewin, D.F. Purcell, B.E. Sleebs, Optimization of 5-substituted thiazolyl ureas and 6-substituted
imidazopyridines as potential HIV-1 latency reversing agents, European Journal of Medicinal Chemistry
(2020), doi: https://doi.org/10.1016/j.ejmech.2020.112254.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. 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.
©
2020 Published by Elsevier Masson SAS.

Optimization of 5-Substituted Thiazolyl Ureas and
6
-Substituted Imidazopyridines as Potential HIV-1
Latency Reversing Agents
†
,‡
€
†,‡
†,‡
William Nguyen, Jonathan Jacobson, Kate E. Jarman, Timothy R. Blackmore, Helene
†
,‡
€, ¥, ǁ
€
†,‡,*
Jousset Sabroux, Sharon R. Lewin,
Damian, F. Purcell, Brad E. Sleebs.
†
The Walter and Eliza Hall Institute for Medical Research, Parkville, Victoria, 3052,
Australia.
‡
Department of Medical Biology, University of Melbourne, Parkville, Victoria, 3052,
Australia.
€
Department of Microbiology and Immunology, University of Melbourne at The Peter
Doherty Institute of Infection and Immunity, University of Melbourne, Parkville, Victoria,
000, Australia.
3
¥
The Peter Doherty Institute for Infection and Immunity, The University of Melbourne and
Royal Melbourne Hospital, Parkville, Victoria 3000, Australia.
ǁ
Department of Infectious Diseases, Alfred Health and Monash University, Melbourne,
Victoria, 3004 Australia.
*
Correspondence addressed to this author.
KEYWORDS
HIV-1, AIDS, latency, thiazole, imidazopyridine, transcription.

ABSTRACT
A persistent latent reservoir of virus in CD4+ T cells is a major barrier to cure HIV.
Activating viral transcription in latently infected cells using small molecules is one strategy
being explored to eliminate latency. We previously described the use of a FlpIn.FM HEK293
cellular assay to identify and then optimize the 2-acylaminothiazole class to exhibit modest
activation of HIV gene expression. Here, we implement two strategies to further improve the
activation of viral gene expression and physicochemical properties of this class. Firstly, we
explored rigidification of the central oxy-carbon linker with a variety of saturated
heterocycles, and secondly, investigated bioisosteric replacement of the 2-acylaminothiazole
moiety. The optimization process afforded lead compounds (74 and 91) from the 2-
piperazinyl thiazolyl urea and the imidazopyridine class. The lead compounds from each
class demonstrate potent activation of HIV gene expression in the FlpIn.FM HEK293 cellular
assay (both with LTR EC s of 80 nM) and in the Jurkat Latency 10.6 cell model (LTR EC
50
50
2
20 and 320 nM respectively), but consequently activate gene expression non-specifically in
the FlpIn.FM HEK293 cellular assay (CMV EC 70 and 270 nM respectively) manifesting in
5
0
cellular cytotoxicity. The lead compounds have potential for further development as novel
latency reversing agents.

INTRODUCTION
The emergence of antiretroviral therapies (ART) 25 years ago has allowed for the control of
1
, 2
HIV-1 by suppressing plasma viral loads to undetectable levels. However, discontinuation
of treatment results in rebound of the virus, constraining patients to a lifelong drug treatment
in order to subdue HIV replication and consequently decrease the severity of opportunistic
3
, 4
pathogenic and viral infections. This in itself poses many challenges including the need for
long term compliance, the emergence of multidrug resistant viruses and drug toxicities
5
resulting from life-long therapy. The virus can rebound after cessation of ART due to long
lived and proliferating latently infected CD4+ T cells that harbour replication-competent
6
, 7
virus.
These latently infected cells are established early after infection, prior to the
8
, 9
detection of viremia, and persist on ART, presenting a major barrier to the eradication of
1
0, 11
the virus.
One strategy being investigated to eliminate latently infected CD4+ T cells is known as
shock and kill”. The aim of this approach is the pharmacological activation of proviral
“
transcription and subsequent virion production. Infected cells would then be susceptible to
immune clearance and viral cytopathic effects, allowing for their elimination. Used in
combination with ART, this approach could potentially eradicate or reduce the number of
12, 13
latently infected cells allowing for enhanced virological control when ART is stopped.
A diverse range of epigenetic proteins have been investigated as latency reversing agents
LRAs) for the “shock and kill” approach and include histone deacetylase (HDACs),
(

bromodomain and extra-terminal motif (BET) proteins, histone methyltransferases (HMT)
1
4, 15
and protein kinase C (PKC) activators.
Several inhibitors of these proteins have
progressed to human clinical trials. One example is the HDAC inhibitor, Vorinostat. In this
clinical trial, people living with HIV (PLWH) on suppressive ART were treated with
Vorinostat (Figure 1) for 14 days and showed increased levels of CD4+ T cell associated
1
6
unspliced (CA-US) HIV RNA, however, plasma HIV RNA, concentration of HIV DNA,
integrated DNA and inducible virus in CD4+ T-cells remained unchanged. Significant
changes in host gene expression were also observed, questioning the safety and long-term use
of Vorinostat. Similar findings with preclinical and clinical trials on other HDAC inhibitors,
1
7, 18
19, 20
21,22
including panobinostat
and romidepsin
and the BET inhibitor, JQ1
(Figure 1). In
general, epigenetic modulators demonstrate poor efficacy in ex vivo models and do not
2
2
induce substantial increases in intracellular HIV-1 mRNA or free virions in patient cells.
Therefore, new LRAs are needed with a novel mechanism of action.
We recently reported on the design of a FlpIn.FM luciferase reporter cellular assay
2
3
capable of detecting small molecules that activate HIV-1 gene expression. The FlpIn.FM
reporter cell line will detect both transcriptional and post-transcriptional steps of gene
expression, including effects on RNA capping, RNA splicing, RNA transport, RNA-
modifications, RNA stability and translation of protein. The HEK293 derived reporter cell
line expresses a HIV-1 Nef – click beetle red (CBR) fusion protein from a spliced mRNA
expressed under direction from the long terminal repeat (LTR) HIV promoter, allowing for
detection of early viral gene expression. The cell line also contains a click beetle green
(CBG) luciferase reporter driven from a cytomegalovirus (CMV) promoter allowing for the
detection of non-specific host cell protein expression.
We used the HEK293 FlpIn.FM cell line in a high throughput screen and the identified the
2
3
2
-acylaminothiazole class (Figure 2). Our initial optimization efforts concentrated on

defining structure activity relationship (SAR) of the 2-acylaminothiazole class and resulted in
compounds with enhanced HIV gene expression (LTR) in both HEK293 FlpIn.FM and the
Jurkat 10.6 latent HIV-GFP reporter cell line. The optimized compounds also increased CA-
US HIV RNA in resting CD4+ T cells isolated from PLWH on ART. While a modest
enhancement in HIV gene expression was demonstrated with optimized analogues, the
selectivity window between activation of HIV and non-specific (CMV) gene expression was
limited. This is comparable to the level of selectivity observed with epigenetic modulators
2
3
used as LRAs in clinical trials.
In this report we describe further optimization of the 5-substituted 2-acylaminothiazole
class focused on rigidifying the central oxy-carbon chain and bioisosteric substitution of the
2
-acylaminothiazole moiety to improve LTR specific activity and physicochemical
properties. We show the 5-substituted thiazolyl urea and 6-substituted imidazopyridine
analogues derived from this study have enhanced HIV-1 LTR activity in both the FlpIn.FM
cell line and J.Lat10.6 cell lines and improved in vitro metabolic stability.
RESULTS AND DISCUSSION
The initial focus of the research reported here was to rigidify the central aliphatic chain,
linking the thiazole and the pendant aryl group (Figure 2), by replacement with a variety of 4-
,
5-, 6- and 7- membered unsaturated heterocyclic systems. It was proposed that this change
could improve HIV-1 LTR activity by reducing conformational energy required for the
substituted thiazolyl and aryl motifs to engage with pockets of the unknown cellular target. In
addition, it was hypothesized, the incorporation of heterocyclic systems may improve
aqueous solubility and in vitro metabolism. We previously reported the 2-acylaminothiazole
class possessed poor aqueous solubility (<1.6 μM at pH 6.5) and was rapidly degraded by
2
3
both human and mouse liver microsomes. In addition, the thiazole moiety is a known

toxicophore because of cytochrome P450-mediated oxidative conversion of the 4,5-
2
4
unsaturated bond to an epoxide which can act as a nucleophile trap. While 4- or 5-
substitution of the thiazole ring is known to suppress this metabolic event, we endeavoured to
eliminate this liability by introducing a 5,6-bicyclic heterocycle that is a bioisoteric
replacement for the 2-acylaminothiazole moiety. Herein, we describe efforts to improve HIV-
1
LTR specific potency and physiochemical properties by concentrating on substituting the
central aliphatic linker and the 2-acylaminothiazole moiety with 4-, 5-, 6- and 7- membered
heterocyclic and 5,6-bicyclic systems respectively.
Chemistry
The synthesis of 2-thiazoloyl urea analogues (Tables 1-5) started with the preparation of
phenyl 2-thiazolecarbamates (4) by reaction of 5-substituted 2-aminothiazoles (3) with
phenyl chloroformate under basic conditions (Scheme 1). The heterocyclic substituted aryl
building block (6) was generated employing a Buchwald–Hartwig amination between an aryl
halide (5) and a N-Boc protected heterocycle (L in Tables 1-5) followed by N-Boc
deprotection. The phenyl 2-thiazolecarbamate (4) and the heterocyclic substituted aryl
building block (6) were then used to afford the 2-thiazoloyl urea analogues (7) in Tables 1-5.
The synthesis of the imidazopyridine analogues (Table 6) began with the alkylation of a
substituted phenol or pyridinol (8) with a protected halogenated butanoate (Scheme 2).
Deprotection of the ester followed and the carboxylic acid (9) was then transformed into a
acid chloride using POCl The acid chloride was on-reacted with a substituted pyridine-
3.
diamine yielding the imidazopyridine analogues (10) in Table 6.
Access to the 2-piperazine substituted imidazopyridines (Table 7) started with the
condensation of a substituted pyridine-diamine (11) with disuccinimidyl carbonate under

aprotic conditions to yield the imidazopyridinone (12, Scheme 3). The imidazopyridinone
was then reacted with POCl to yield the 2-chloroimidazopyridine (13). Finally, the
3
nucleophilic substitution of
a
substituted aryl piperazine (14) with the 2-
chloroimidazopyridine (13) afforded the 2-piperazine substituted imidazopyridines (15) in
Table 7.
Structure and Activity Relationship
Our previous SAR studies established the 5-position of the thiazole ring and 3,4-
disubstitution on the aryl ring of the 2-acylaminothiazole class were important for HIV-1
2
3
LTR activity. In particular, combinations of 5-methyl or 5-chloro substitution on the
thiazole ring and 3-methyl and 4-chloro disubstitution on the phenoxy ring (Figure 2), were
optimal for HIV-1 LTR activity and selectivity against the CMV non-specific reporter, and
thus this substitution pattern would be used as a benchmark and remain consistent for the
analogues produced in the present study.
We first investigated the suitability of 5- and 6- membered saturated heterocycles as
surrogates for an oxy-carbon linker between the carbonyl and the aryl ring. The 5- and 6-
membered heterocycles were proposed because they have similar steric spacing compared to
the 4-atom oxy-carbon linker. The thiazolyl urea analogues generated were evaluated against
the HEK293 FlpIn.FM cell line and their LTR and CMV activities are shown in Tables 1 and
2
. The results show that introduction of the piperazine, seen in compounds 16 and 17, gave a
- and 20- fold improvement in LTR activity (EC 0.6 and 0.4 µM) compared to 1 and 2
4
50
(Figure 2), respectively. It was also noted that the 3-chloro and 4-methyl substitution on the
aryl ring seen in 34 and 35 (LTR EC of 0.47 and 0.25 µM) (Table 2) did not affect activity
5
0
compared to 16 and 17, indicating that the configuration of the 3- and 4- chloro and methyl

substitution was interchangeable with the thiazolyl urea and consistent with previous SAR
2
3
with the 2-acylaminothiazole class.
The selectivity window between the non-specific CMV reporter and the LTR reporter for
all analogues shown in Tables 1 and 2, was approximately 2- to 3- fold, except for analogues,
3
6 and 37. The observed selectivity window was generally consistent for all analogues
produced throughout this report, and therefore our efforts herein concentrated on improving
the LTR activity. While CMV activity was monitored, achieving selectivity between the LTR
and CMV reporter was not the ongoing primary focus.
Analogues with heterocycles replacing the oxy-carbon linker such as 4-hydroxy, 4-amino
and 4-methylamino piperidine (18-22) and 3-hydroxy and 3-amino piperidine (23-27) were
all inactive (LTR EC >20 µM) (Table 1). The corresponding 5-membered systems, the 3-
5
0
amino and 3-hydroxy pyrrolidine analogues, 28-31, displayed LTR activity (EC 0.79 – 6.2
5
0
µM) comparable to 1 and 2, but N-methylation of the 3-amino pyrrolidine functionality (32
and 33) was detrimental to LTR activity. Analogues 36 and 37 with the 4-pipecolic amide
scaffold retained LTR activity (EC50 2.5 and 2.0 µM) (Table 2) compared to 2-
acylaminothiazoles 1 and 2.
Smaller 4-membered heterocycles were all investigated as replacements for the oxy-
carbon linker (Table 2). The analogues, 38 and 39, with 3-aminoazetidine functionality
maintained comparable LTR activity (EC 0.67 and 0.23 µM) to the piperazine analogues 34
5
0
and 35. N-Methylation of the 3-amino group of the azetidine moiety (40 and 41) was
detrimental to LTR potency, mirroring the SAR observed with the N-methylation of the 3-
amino pyrrolidine analogues, 32 and 33. The 3-oxy azetidine analogues, 42 and 43, exhibited
comparable LTR activity (EC 0.22 and 0.53 µM) to 38 and 39, suggesting that the nitrogen
5
0
and oxygen atoms are interchangeable at this position, consistent with the SAR previously
23
reported with the 2-acylaminothiazoles series.

In addition to the 3-amino (38 and 39) and 3-oxy azetidine (42 and 43), the piperazine
heterocycle (34 and 35) was the most suitable replacement for the oxy-carbon linker,
therefore we further explored iterations on the piperazine (Table 2). Expanding the piperazine
ring to homo-piperazine (44 and 45) resulted in 20- and 100-fold losses in activity when
compared to the corresponding piperazine compounds (34 and 35). 2-Methyl substituted
piperazine analogues (46-49) were consistently less potent (EC 1.3 to 8.9 µM) than the
50
unsubstituted piperazine analogues (34 and 35). The oxo-piperazine analogue, 50 (EC 0.6
5
0
µM), was 2-fold less active than 35, although it had enhanced LipE (3.0). The 2,6-
diazaspiro[3.3]heptane heterocycle was also trialled because Reilly et al. demonstrated it was
2
5
a suitable piperazine isostere with reduced cytotoxicity. It was found that 2,6-
diazaspiro[3.3]heptane scaffold was not a suitable replacement for piperazine in this instance,
with the analogues 51 and 52 displaying modest LTR activity (EC 40 and 16 µM). Overall,
5
0
the unsubstituted piperazine was established as the most appropriate group for replacement of
the oxy-carbon linker.
Another goal of this work was to improve the in vitro metabolism and solubility, as the 2-
acylaminothiazole analogues (Figure 2) were shown to possess poor aqueous solubility and
2
3
modest in vitro metabolic stability (Table 8). Although the LipE was increased with the
piperazine analogue 35 compared to 1, the aqueous solubility was still limited (<1.6 μM at
pH 2.0 and pH 6.5), however, the metabolic stability of 35 in mouse liver microsomes (Clint
5
3 μL/min/mg protein) was enhanced compared to the 2-acylaminothiazole 1 (Table 8). A
similar outcome was also observed for the 3-amino azetidine analogue 39 with modest
metabolic stability in mouse liver microsomes (Cl 49 μL/min/mg protein) and low aqueous
int
solubility. Analogue 29 with 3-amino pyrrolidine functionality possessed poor metabolic
stability in mouse microsomes (Cl 151 μL/min/mg protein), while the inclusion of the oxo-
int
piperazine group (50) enhanced stability in mouse liver microsomes (Cl 29 μL/min/mg
int

protein) and improved LipE, but did not improve aqueous solubility. Overall, this data
demonstrated further improvements in aqueous solubility and metabolism were required.
To further enhance LTR mediated gene-expression and potentially improve
physicochemical properties, a variety of substitutions on the aryl group were explored while
retaining the 5-methyl thiazolyl piperazine pharmacophore (Table 3). Consistent with
previous observations (Figure 2) that the 3-mono and 3,4-disubstitution are important for
2
3
LTR potency, the thiazolyl piperazine analogues with either no substitution, 4-chloro or 3-
chloro substitution (53 and 54) were significantly less potent than the disubstituted 3-, 4-
chloro methyl analogues, 16 and 34. The 3,4-dichloro substituted analogue (55) was
equipotent to the two chloro-methyl substituted analogues (16 and 34) again suggesting that
the chloro and methyl groups are interchangeable in the 3- and 4- positions. Introduction of a
nitrile substituent on the 3-position of the aryl group (56) modestly improved LTR activity
(
EC LTR of 0.23 µM) and LipE (3.7).
50
We previously determined a chloro, nitrile or trifluoromethyl substituent in the 5-position
of the thiazole was important for LTR activity and improving metabolic stability. In
2
3
particular, the nitrile substitution was shown to decrease thiazole oxidation. In an effort to
further improve LTR activity and metabolic stability, we next explored the impact of
substitution in the 5-position of the thiazole ring on LTR activity while maintaining either a
4
4
-substitution or a 3,4-disubstition on the aryl ring (surveyed in Table 3). The results in Table
show that 5-chloro and 5-nitrile thiazole substitution with a 4-chloro aryl substitution (57
and 58) gives a 2- and 7-fold improvement in LTR activity (EC 0.35 and 1.5 µM) compared
5
0
to the 5-methyl substituted analogue (54). Similarly, 5-chloro and 5-nitrile thiazole
substitution with a 3-chloro-4-methyl aryl group (35 and 60) resulted in a 2-fold
improvement in LTR activity (EC 0.24 and 0.24 µM) compared to the 5-methyl substituted
5
0
comparator analogue 34, while the 5-trifluoromethyl substitution (61) was 2-fold less active

(
EC 1.3 µM). The 5-chloro substitution on the thiazole with a 4-nitrile aryl substitution (65)
50
was less active than its 5-methyl substituted counterpart (56). An undecorated thiazole
analogue (59) possessed moderate LTR activity (EC 3.4 µM), but in contrast, we previously
50
2
3
showed that the unsubstituted 2-acylaminothiazole was inactive, highlighting the influence
of the piperazinyl scaffold on LTR potency. Overall, analogues with 5-chloro and 5-nitrile
thiazole substitution and a 3,4-disubstituted aryl group, 17 and 62-64, consistently displayed
improved LTR activity compared to its 5-methyl thiazole counterpart (Table 3). In addition,
the inclusion of 5-nitrile substitution on the thiazole enhanced LipE values (>2.6).
The changes made to the 5-thiazole and aryl substituents enhanced the metabolic stability
of 63 in the presence of mouse liver microsomes (Clint 15 μL/min/mg protein) when
compared to 35 (Clint 53 μL/min/mg protein) (Table 8). Previous analysis of our 2-
acylaminothiazoles, identified several potential metabolites including those predicted to
originate from the mono-oxidation of both the phenyl ring and the 3- or 4- methyl substituent
2
3
on the aryl ring. Mass spectrometry analysis of 63 only detected a single mono-oxidation
site (Table S1) demonstrating that installation of 3,4-dichloro substitution on the aryl ring
significant improves intrinsic clearance. The in vitro metabolism data on 65 (Clint 38
μL/min/mg protein) (Table 8) provides further evidence of improved metabolic stability
when a methyl substituent is not present on the thiazole or the pendant aryl ring.
To further improve metabolic stability and enhance aqueous solubility of the thiazolyl urea
series, we explored whether differentially substituted pyridines would be a suitable
replacement for the 3,4-disubstituted aryl ring. Analogues were designed to incorporate an
endocyclic nitrogen in the 5- or 6- position while maintaining 3,4-disubstitution on the aryl
ring in combination with a 5-substituted thiazole (Table 5). The results of the analogues
evaluated in FlpIn.FM HEK293 assay show the 5-methyl thiazole analogue 66 with 6-(3-
cyano-4-methylpyridine) substitution exhibited modest LTR activity (EC50 4.2 µM).

Consistent with previous trends, transition from the 5-methyl to 5-chloro substituent on the
thiazole (67) resulted in a significant increase in potency (EC 0.46 µM). This trend was also
5
0
observed for the analogues 68 and 69 comprising a 5-(2-cyano-3-methylpyridine) system
EC of 10 µM vs. 0.21 µM). The similar LTR activity observed between 67 and 69 suggests
(
5
0
the location of the endocyclic nitrogen appears to be unimportant for LTR activity.
Substitution of a trifluoromethyl group into the 4-pyridyl position (70 and 71) was also
tolerated (LTR EC 0.73 and 0.92 µM), but the 5-chloro analogue 71 was 2-fold less active
5
0
than nitrile counterpart 67. Furthermore, the inclusion of the nitrile substituent resulted in
increased LipE values for analogues 67 (3.4) and 69 (4.0). Analogues 72-75 comprised of
3
,4-dichloropyridin-5-yl substitution in combination with substitution of methyl, nitrile,
chloro and trifluoromethyl in the 5-position of the thiazole, respectively, showed a 2- to 10-
fold increase in LTR activity (EC <1.0 µM) and improved LipE compared to their phenyl
5
0
counterparts (55, 62 and 63). Compound 74 exhibited the greatest LTR potency (EC 80 nM)
5
0
of the thiazolyl urea series. To evaluate whether a pyrimidine system could replace the
substituted pyridinyl system, the analogues 76 and 77 were generated, but pyrimidine
substitution was not tolerated (EC >40 and 8 μM).
5
0
The LipE values were improved for the pyridinyl analogues in Table 5 compared to the
aryl analogues, but the aqueous solubility was not improved (Table 8). The metabolic
stability of the pyridinyl analogue 74 in mouse liver microsomes (Cl 32 μL/min/mg protein)
int
was comparable to the corresponding aryl analogues 35 and 50. Similarly, the pyridinyl
analogue 73 possessed similar metabolic stability (Clint 22 μL/min/mg protein) than the
comparator phenyl analogue 63 (Clint 15 μL/min/mg protein). Overall, incorporating a
heterocyclic nitrogen in the aryl ring was beneficial for LTR activity and LipE, maintained in
vitro metabolism, but did not improve aqueous solubility.

Unsubstituted thiazoles are notorious toxicophores due to their susceptibility for
cytochrome P450 mediated oxidative metabolism of the 4,5-double bond producing an
epoxide, which is a nucleophilic trap. In vitro metabolism studies, including our own, have
2
3, 26
shown that 5-substitution of the thiazole ring can mitigate this oxidative event.
To
overcome the oxidative metabolism, we investigated the bioisoteric replacement of the 2-
aminoacylthiazole system. We hypothesized a low-lying σ* orbital of the sulfur atom in the
thiazole was acting as an electron acceptor for an electron rich oxygen atom on the
amide/urea carbonyl. This interaction is likely to result in a non-covalent intramolecular
interaction capable of fixing the rotational confirmation of carboxamide in relation to the
thiazole. Non-covalent sulfur interactions are a common phenomenon and have been recently
2
7
reviewed. We proposed the imidazopyridine scaffold could act as a suitable bioisosteric
replacement for the 2-acylaminothiazole moiety, in which the pyridine nitrogen would mimic
the thiazole nitrogen and the imidazole NH would emulate the 2-acylaminothiazole NH.
To determine if the imidazopyridine system was an appropriate replacement for the 2-
acylaminothiazole structure, a set of analogues were generated and evaluated in the FlpIn.FM
HEK293 cellular assay (Table 6). The results show that the 6-methyl and 6-chloro
imidazopyridine analogues 78 and 79 exhibited modest LTR activity (EC 12 and 9.4 µM)
50
comparable to the LTR activity of the corresponding 2-acylaminothiazoles 1 and 2 (Figure 2).
The imidazopyridine analogues 78, 80 and 82 possessed similar LTR activity suggesting 3,4-
disubstitution pattern on the phenoxy aryl ring was not as sensitive to change as the 2-
acylaminothiazole scaffold. The imidazopyridine analogues with the 5-chloro substituent, 79,
8
1 and 83 were consistently more active than the 6-methyl substituted imidazopyridine
analogues 78, 80 and 82, demonstrating the same trend seen with the 2-acylaminothiazole
class. The thiazopyridine analogues, 84 and 85 were inactive (LTR EC50 >40 µM),
demonstrating the requirement for an NH in the 3-position of the imidazopyridine. The in

vitro metabolism of 79 and 83 in the presence of mouse liver microsomes was high (Clint
866 and 122 μL/min/mg protein) (Table 8). It was discovered the high metabolism was a
>
result of a result of mono and bis-oxygenation of the aryl phenoxy ring (Table S1). The
introduction of the imidazopyridine system did not improve aqueous solubility, for example
7
9 and 83 has comparable solubility to the 2-acylaminothiazole analogues in Table 8.
Although no improvement in in vitro physiochemical properties was observed, the LTR
activity of the imidazopyridine analogues was comparable to the corresponding 2-
acylaminothiazole derivatives, reinforcing imidazopyridine as a suitable bioisosteric
replacement.
To improve LTR mediated gene-expression and physicochemical properties of the
imidazopyridine scaffold, we next replaced the oxy-carbon chain with piperazine; an
approach previously undertaken with the thiazolyl urea class (Tables 1-5). A focus set of 2-
piperazinyl imidazopyridine analogues was synthesized and evaluated in the FlpIn.FM
HEK293 cell assay (Table 7). The results show the imidazopyridine analogues 86, 88 and 90
(EC50 0.75 - 3.2 µM) are moderately less potent than the corresponding thiazolyl urea
analogues, 34, 62 and 72. The 6-halo imidazopyridine analogues 87 and 91 are 10-fold more
potent (EC 0.32 and 0.08 µM) than the 6-methyl derivatives 86 and 90 (EC 3.2 and 1.3
5
0
50
µM), consistent with the trend observed for the 5-chloro thiazolyl urea class. In addition,
imidazopyridine analogue 91 was equipotent (EC of 0.08 µM) to 74, suggesting the 2-
50
piperazinyl imidazopyridine derivatives were closely aligned with the trends in SAR and
LTR activity of the 2-piperazinyl thiazolyl urea series, further supporting the application of
the imidazopyridine bioisostere. The imidazopyridine analogues 86 and 91 exhibited low
metabolic stability in the presence of mouse liver microsomes (Cl 224 and 132 μL/min/mg
int
protein) (Table 8) compared to the corresponding thiazolyl urea derivatives 35 and 74. The
aqueous solubility of 86 and 91 at pH 6.5 was also low, but the solubility at pH 2 was

enhanced (25 - 100 μM), because of the presence of an ionizable nitrogen on the
imidazopyridine system.
In summary of the SAR (Figure 3), it was initially determined that the piperazine, 3-
aminoazetidine and 3-aminopyrrolidine were suitable replacements for the oxy-carbon chain
substituted producing the thiazolyl urea scaffold with improved LTR activity (Tables 1-5)
and physicochemical properties (Table 8) compared to the 2-acylaminothiazole class. The
SAR of the thiazolyl urea series was also closely aligned with the 2-aminoacylthiazole
2
3
class. In general, either 3,4-dichloro or 3,4-chloro methyl substitution on the aryl ring and
either 5-chloro or 5-nitrile substitution on the thiazole were optimal for LTR directed gene
expression. This activity and SAR trend was also observed for the imidazopyridine
bioisosteric replacements of the 2-acylaminothiazole and the thiazolyl urea series (Table 6
and 7). 5-Chloro and 3,4-dichloro substitution on both the 2-acylaminothiazole and the
thiazolyl urea series displayed the most robust in vitro metabolic stability in comparison to
other substitution patterns (Table 8). The imidazopyridine system exhibited slightly improved
aqueous solubility but possessed significantly lower metabolic stability compared to 2-
acylaminothiazole and the thiazolyl urea series suggesting an intrinsic metabolic liability
associated with the bicyclic scaffold.
In general, improvements in LTR activated reporter gene expression also resulted in
proportionate increases in CMV activity. In cases where selectivity was observed, LTR
activity reached a maximum before decreasing or plateauing with increasing compound
concentration (Figure S1), suggesting that either off-target cellular toxicity circumvents
transcriptional or modulates post-transcriptional gene expression, or an increase in
transcriptional activation and gene expression perturbs cellular homeostasis leading to cell
death.

Evaluation of Cellular Growth Inhibition
The CMV reporter was built into the FlpIn.FM HEK239 cell line to detect fluctuations in
non-specific Luc reporter protein expression. To determine whether increased LTR and CMV
reporter activity are related to host cytotoxicity, a selection of compounds was evaluated in a
Cell Titre Glo growth inhibition assay using the HepG2 cell line (Table 9). In this assay, all
the analogues tested inhibited cell growth more potently than the 2-acylamidothiazole
analogue 2. The HepG2 cell growth inhibition data robustly correlated with the LTR activity,
suggesting transcriptional activation was related to cell growth inhibition. This observation
provides a reason why the LTR activity reaches a maximum before gradually decreasing with
higher concentrations of compound (Figure S1). This is consistent with clinically used LRA’s
such as Romidepsin and Panobinostat, whereby the cell growth inhibition is linked to their
2
8
mechanism of action. Compounds with increasing LTR activity were also associated with
potent CMV activity, however, the window of selectivity between the CMV and LTR
reporter was not consistent for each analogue, and therefore it is difficult to ascertain whether
cell growth inhibition is directly associated with LTR or CMV activity. It is unknown or this
class of HIV LRAs whether it will be possible to improve cell cytotoxicity while maintaining
LTR potency, but the strong correlation with both the LTR and CMV activity suggests gene
expression from these viral promoters and cellular toxicity are intrinsically linked. The
association between activation of non-specific gene expression and cell toxicity is a common
2
8
trait in other established LRAs. Ongoing efforts to uncover the mechanism of action of the
-acylaminothiazole and thiazolyl urea series may assist in understanding this relationship
2
and improving the selectivity between cell growth inhibition and activation of LTR gene
expression.
Evaluation of Thiazolyl Ureas Against Targets Previously Linked to HIV Latency

Our efforts to establish the mechanism by which the thiazolyl urea analogues enhance
cellular LTR activity began with profiling a representative compound against a panel of
2
9
epigenetic enzymes known to be associated with HIV gene expression (Table S2 to S4). 74
displayed no inhibitory activity when profiled against all histone deacetylase (HDAC)
isoforms and SIRT1 (Table S2). Compound 74 also exhibited no detectable inhibition of nine
methyltransferases and four acetyltransferases previously linked to HIV-1 transcriptional
2
9, 30
modulation in patients with chronic HIV-1 infection and latency
(Tables S3 and S4). The
thiazolyl urea scaffold possess a kinase hinge-binding motif and is closely related to reported
3
1,32
cyclin dependent kinase inhibitors,
35) against 369 kinases at a concentration of 0.5 μM (Table 5). The results show analogue
5 did not show significant inhibition of any of the kinases in the panel, except for the modest
and therefore we screened a representative analogue
(
3
inhibition (40-50%) of three cyclin-dependent kinase 9 (CDK9) proteins, cyclin K, T1 and
T2. Subsequently, 74 was tested against 22 kinases of the CDK family (Table S6) at
concentration of 0.5 μM. Consistent with the kinase inhibition data on 35, 74 exhibited
approximately 60% inhibition the CDK9 proteins, cyclin K, T1 and T2. CDK9 proteins are
known to be recruited by Tat to initiate HIV-1 transcription by phosphorylating the COOH-
3
3
terminus of RNA-pol II and increasing processivity. While this is within the realm of
pharmacological relevance as a potential mechanism, direct inhibition of CDK-9 in HIV-1
3
4
replication assays has been shown to potently block HIV-1 transcription. In addition, the
cell growth inhibition and LTR activity for both 35 and 74 (EC <0.1 µM) is significantly
5
0
lower than the measured CDK9 inhibition values. Therefore, it is likely that the CDK9 kinase
activity is unrelated to observed cellular and LTR activity. In summary, the screening of
analogues against a panel of epigenetic proteins and kinases suggests the thiazolyl urea series
has a mechanism of action not previously associated with HIV gene expression. Target
identification of the thiazolyl urea class will be key aspect of future studies.

Evaluation of Analogues in the Jurkat Latency 10.6 Cell Line
The J.Lat10.6 clone is a latently infected T-cell line commonly employed to evaluate the
35, 36
activity of LRAs to activate the HIV LTR directed reporter gene expression.
This cell
line contains a full-length integrated HIV-1 genome with green fluorescent protein (GFP) in
place of the Nef gene. Treatment of J.Lat10.6 with TNF-α or compound activates viral gene
expression which is quantified by flow cytometry.
We evaluated a selection of the analogues using the J.Lat10.6 cell line and compared their
LTR activity to selection of known LRAs (Table 9). The results show compounds 29, 35, 39,
5
0, 63 and 73 enhanced the LTR GFP reporter with an EC range of 1 - 5 µM, comparable to
50
the activity of known LRAs, Vorinostat and JQ1 (EC 2.3 and 8.1 µM). The LTR activity
50
displayed by the thiazolyl urea analogues was a significant improvement on the LTR activity
23
of the 2-amainoacylthiazole 2 reported previously (EC 29 µM). The analogues, 65, 74 and
5
0
9
1, potently enhanced the LTR GFP reporter (EC <0.5 µM) comparable to the activity of
50
the known LRAs, Romidepsin (EC 0.22 µM) and Panobinostat (EC 0.13 µM). There is a
5
0
50
robust correlation between the FlpIn.FM HEK293 LTR activity and the J.Lat10.6 LTR
reporter, and this is most evident with the analogues 74 and 91, which display potent activity
in both cell lines. Overall, the LTR activated Luc-reporter gene expression from the FlpIn.FM
HEK293 cell assay robustly correlates with the LTR activated GFP expression in the
J.Lat10.6 cellular model, demonstrating the utility of the FlpIn.FM HEK293 in detecting
LRAs.
CONCLUSIONS

The FlpIn.FM HEK293 cellular assay was previously utilized to identify the 2-
acylaminothiazole hit class. Exploration of the structure activity relationship afforded
compounds with modest activity in HIV cellular models (Figure 2). Building on our previous
work, we applied two design strategies to further enhance physicochemical properties and
improve activity in HIV gene expression models. The first approach explored rigidification of
the central oxy-carbon linker with a variety of 4-, 5-, 6- and 7- membered saturated
heterocycles. Piperazine, 3-aminopyrrolidine and 3-azetidine were the most suitable
surrogates for the central oxy-carbon linker (thiazolyl urea analogues 74 and 91) which
exhibited a 20-fold enhancement in LTR activity (EC 80 and 80 nM) compared to the 2-
50
acylaminothiazole 2 (EC 1.9 µM). The thiazolyl urea analogues also possessed enhanced
5
0
metabolic stability in the presence of both human and mouse liver microsomes (for example
3, Cl 11 and 15 μL/min/mg protein), however aqueous solubility remains low and will
6
int
require further optimization.
The second approach investigated the bioisosteric replacement of the 2-acylaminothiazole
moiety with imidazopyridine framework. We determined the imidazopyridine scaffold was a
reasonable bioisostere and mirrored the structure activity relationship and the LTR activity
observed with the 2-acylaminothiazole and the thiazolyl urea scaffolds. The imidazopyridine
analogues possessed low metabolic stability in the presence of both human and mouse liver
microsomes (for example, 86 Cl 115 and 224 μL/min/mg protein), suggesting an intrinsic
int
liability with the scaffold that could not be determined through metabolism identification.
Future work is required to address both metabolism and aqueous solubility of the
imidazopyridine class to permit its utility as an LRA in in vivo studies.
The thiazolyl urea and imidazopyridine series potently activated LTR activity in the
FlpIN.FM HEK293 cellular assay, but also variably activated the non-specific transcription
via the CMV reporter. Only a select few analogues displayed a significant window of

selectivity (>10-fold for 36-37, 63, 66 and 70) and in general, most analogues displayed a 2-
to 3- fold difference between LTR and CMV activity (Tables 1-7). It is unknown whether the
same mechanism of action is responsible for modulating both the LTR and CMV promoter
directed gene expression, but the robust correlation in the HIV LTR directed reporter activity
and cellular toxicity (Table 9) would suggest the mechanisms are intrinsically linked.
Therefore, it is proposed the cellular toxicity of the analogues is the result of a significant
increase in HIV-promoter directed protein expression that disrupts cellular homeostasis and
consequently manifests in cellular toxicity. The non-selective transcriptional and/or post-
transcriptional cytotoxicity observed is a reoccurring trait in epigenetic modulators, such as
2
8
Romidepsin and Panobinostat (Table 9), commonly utilized in the shock and kill approach.
It is likely that identifying the cellular target(s) of the thiazolyl urea and imidazopyridine
series will help unravel the mechanisms imparting the cytotoxicity observed. Our efforts to
elucidate the cellular target involved screening thiazolyl urea compounds against a panel of
epigenetic proteins and kinases but did not reveal the genuine target (Tables S2-S6). The
discovery of the mechanism remains a key aspect of our future work.
In summary, we have used a 2-aminoacylthiazole scaffold as a template to prepare a set of
thiazolyl ureas and imidazopyridine analogues that potently activate HIV transcription in
FlpIn.FM HEK293 and J.Lat10.6 latency models. Furthermore, the activity observed with
lead analogues (74 and 91) is comparable to the clinical evaluated LRAs, Romidepsin and
Panobinostat, and highlights the potential of this class for application to modify HIV gene
expression as a strategy to eliminate latent provirus.
EXPERIMENTAL SECTION
Biology Experimental.

FlpIn.FM HEK293 Cellular Assay. The FlpIn.FM HEK293 cellular assay was performed
2
3
according to previously described protocol. Briefly, FlpIn.FM HEK293 cells were
maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FCS.
A total of 1500 cells in 25 μL of DME media with 5% FCS were transferred to each well of
the assay plate using a Multidrop reagent dispenser. Plates were left at room temperature in a
single layer for 1 h to allow adhesion to commence and were then incubated at 37 °C and 5%
CO for 48 h. Assay plates were cooled to 20 °C prior to addition of 10 μL of Chroma-Glo
2
reagent to each assay plate well. The plates were incubated at 20 °C for 2 min, and plates
were read on a plate reader (Em: 613 and 537 nm). Percent activity of each compound was
calculated for the ClickBeetle Red (CBR) luminescence and ClickBeetle Green luminescence
(
CBG) according to the following equation: % activity = 100 – (100 ×
luminescencecompound – mean luminescencenegative control/mean luminescencepositive control – mean
(
luminescencenegative control)). For calculating the EC values of LTR- and CMV-driven gene
50
expression, the raw data was first normalized between the DMSO-treated negative control
(set to 0%) and the PMA-treated positive control (set to 100%). Data points associated with
cytotoxicity were excluded where LTR (ClickBeetle Red) activity begins to decrease with
increasing compound concentration. EC values were then obtained using XY analysis,
5
0
nonlinear regression (curve fit), and [agonist] versus normalized response–variable slope
analysis using Domatics Software.
Jurkat Latency 10.6 Clone (J.Lat10.6) Cellular Assay. The J.Lat10.6 cellular assay was
3
5
performed according to previously described protocol. Briefly, cells were maintained in
9
0% RPMI 1640 and 10% FBS supplemented with penicillin G (100 U/mL), streptomycin
4
(
100 μg/mL), and l-glutamine (2 mM, 0.3 mg/mL). J.Lat10.6 cells (4 × 10 ) were seeded into
a 96-well format in 50 μL of complete RPMI (supplemented with 10% FCS and 1%
Pen/Strep). Compounds were then added in 50 μL of complete RPMI and plates incubated at

o
3
7 C for 48 h. Bioluminescence was measured using a FACSCalibur. Positive population
data was normalized using the positive control (TNF-α) and negative control (DMSO) using
GraphPad Prism. The EC was extrapolated in the same manner as reported above for the
50
HEK293 FlpIn.FM dual reporter cellular assay.
HepG2 Cell Growth Inhibition Assay. The HepG2 cellular assay was performed according
2
3
to previously described protocol. Briefly, HepG2 cells were cultured in Dulbecco’s
modified Eagle’s medium (DMEM) supplemented with 10% fetal calf serum (FCS) in a
humidified incubator at 37 °C and 5% CO2. Assay plates were created by seeding 1000 cells
in 50 μL of DMEM with 10% FCS into each well of 384-well tissue culture treated plates.
Ten-point dilution series of compounds were prepared in DMSO and transferred to the assay
plate. Plates were then incubated at 37 °C and 5% CO for 48 h. Cell Titer-Glo reagent was
2
added and bioluminescence measured using an Envision plate reader. Cytotoxicity was
calculated as a percentage using DMSO as the positive growth control and 10 μM bortezomib
as a negative growth control. EC values were calculated using a four-parameter log dose,
50
nonlinear regression analysis, with sigmoidal dose−response (variable slope) curve fit and
constraint parameters of 0 and 100% using GraphPad Prism (v6.05).
Aqueous Solubility Determination Using Nephelometry. The compound in DMSO was
spiked into either pH 6.5 phosphate buffer or 0.01 M HCl (pH 2.0) with the final DMSO
concentration being 1%. Samples were then analysed via nephelometry to determine a
solubility range.
In Vitro Metabolism Using Mouse and Human Liver Microsomes. In vitro metabolism
and metabolism identification studies were performed following the previously described
2
3
procedure.
Chemistry Experimental.

General Chemistry Procedures. Solvents were obtained commercially and used without
further purification. Chromatography was performed with silica gel 60 (particle size 0.040-
0
.063 µm) using an automated purification system. NMR spectra were recorded on a Bruker
Avance DRX 300. Chemical shifts are reported in ppm on the δ scale and referenced to the
appropriate solvent peak. MeOD, DMSO-d , Acetone-D and CDCl contain H O.
6
6
3
2
LCMS were recorded on either a Waters LCMS system composed of a Waters 3100 Mass
Detector, Waters 2996 Diode Array Detector, Waters 2545 Binary Pump, Waters SFO
System Fluidics Organizer and a Waters 2767 Sample Manager (Method A), or an Agilent
LCMS system composed of an Agilent G6120B Mass Detector, 1260 Infinity G1312B
Binary pump, 1260 Infinity G1367E HiPALS autosampler and 1260 Infinity G4212B Diode
Array Detector (Method B). Conditions for LCMS Method A were as follows, column:
Kinetex TM XB-C18 5µm 4.6 × 50mm, injection volume 10 µL, 5-100% B over 3 min
(solvent A: water 0.1% formic acid; solvent B: AcCN 0.1% formic acid), flow rate: 1.5
mL/min, detection: 100-600 nm, acquisition time: 6 min. Conditions for LCMS Method B
were as follows, column: Poroshell 120 EC-C18, 2.1 × 50 mm 2.7 Micron at 20 ºC, injection
volume 2 µL, gradient: 5-100% B over 3 min (solvent A: water 0.1% formic acid; solvent B:
AcCN 0.1% formic acid), flow rate: 0.8 mL/min, detection: 254 nm, acquisition time: 5 min.
LCMS conditions used to assess purity of final compounds were as follows, column:
Phenomenex Gemini C18, 2.0 × 50 mm; injection volume 20 µL; gradient: 0-100% Buffer B
over 6 min (buffer A: 0.1% formic acid in autoclaved MilliQ water; buffer B: 0.1% formic
acid in 100% acetonitrile), flow rate: 1.0 mL/min, detection: 214 or 224 nm. Unless otherwise
noted, all compounds were found to be >95% pure by this method. HRMS were acquired
through The Bio21 Mass Spectrometry and Proteomics Facility using a Thermo Scientific™
nano-LC Q Exactive™ Plus Mass spectrometer.

General Procedure A: 4-(4-Chloro-3-methylphenyl)-N-(5-methylthiazol-2-yl)piperazine-1-
carboxamide (16). 92 (30 mg, 0.14 mmol), 93 (37 mg, 0.16 mmol) and caesium carbonate
(93 mg, 0.29 mmol) were combined in 1,4-dioxane (1 mL) and stirred at reflux for 5 h under
N . The reaction was then cooled to room temperature and the reaction mixture diluted with
2
EtOAc (20 mL), and then washed with water (10 mL) and brine (10 mL). The organic layer
was then dried with Na SO and concentrated in vacuo. The crude residue was then purified
2
4
by column chromatography (100% CyHex to 50% EtOAc/CyHex) to obtain 16 as a solid (19
1
mg, 38%). H NMR (300 MHz, CDCl ): δ 7.23 (d, J 8.6 Hz, 1H), 6.94 (d, J 1.3 Hz, 1H), 6.80
3
(d, J 2.9 Hz, 1H), 6.71 (dd, J 8.7, 3.0 Hz, 1H), 3.78 - 3.64 (m, 4H), 3.24 - 3.11 (m, 4H), 2.41 -
+
2
.33 (m, 6H). MS, m/z = 351 (100) [M+H] , 353 (30). HRMS found: (M + H) 351.1038;
C H ClN OS requires (M +H), 351.1041.
16
19
4
4
-(4-Chloro-3-methylphenyl)-N-(5-chlorothiazol-2-yl)piperazine-1-carboxamide
(17).
General Procedure A was followed using 94 (40 mg, 0.16 mmol) and 92 (30 mg, 0.14 mmol)
1
to obtain 17 as a solid (41 mg, 78%). H NMR (300 MHz, CDCl ): δ 7.23 (d, J 8.8 Hz, 1H),
3
7
3
1
.15 (s, 1H), 6.80 (d, J 2.9 Hz, 1H), 6.70 (dd, J 8.7, 3.0 Hz, 1H), 3.75 - 3.63 (m, 4H), 3.23 -
1
3
.12 (m, 4H), 2.35 (s, 3H). C NMR (75 MHz, CDCl ) δ 159.9, 152.8, 148.9, 136.3, 132.3,
3
+
29.1, 125.8, 120.0, 118.8, 115.1, 48.9, 43.6, 20.0. MS, m/z = 371 (100) [M+H] , 373 (60).
HRMS found: (M + H) 371.0493; C H Cl N OS requires (M + H), 371.0495.
15
16
2
4
4
-(4-Chloro-3-methylphenoxy)-N-(5-methylthiazol-2-yl)piperidine-1-carboxamide
(18).
General Procedure A was followed using 93 (30 mg, 0.13 mmol) and 95 (26 mg, 0.12 mmol)
1
to obtain 18 as a solid (20 mg, 47%). H-NMR (300 MHz; CDCl3): δ 7.23 (d, J 8.6 Hz, 1H),
6
3
3
+
.92 (s, 1H), 6.80 (d, J 2.9 Hz, 1H), 6.69 (dd, J 3.0, 8.7 Hz, 1H), 4.49 (tt, J 3.4, 6.4 Hz, 1H),
.81 - 3.63 (m, 2H), 3.61 - 3.44 (m, 2H), 2.40 - 2.29 (m, 6H), 2.04 – 1.74 (m, 4H). MS, m/z =
+
65 (100) [M+H] , 367 (30). HRMS found: (M + H) 366.1037; C H ClN O S requires (M
17
20
3
2
H), 366.1038.

4
-(4-Chloro-3-methylphenoxy)-N-(5-chlorothiazol-2-yl)piperidine-1-carboxamide
(19).
General Procedure A was followed using 94 (32 mg, 0.13 mmol) and 95 (26 mg, 0.12 mmol)
1
to obtain 19 as a solid (38 mg, 47%). H-NMR (300 MHz; CDCl ): δ 7.24 (d, J 8.6 Hz, 1H),
3
7
.13 (s, 1H), 6.81 (d, J 2.9 Hz, 1H), 6.70 (dd, J 2.9, 8.8 Hz, 1H), 4.57 - 4.47 (m, 1H), 3.76 -
+
3
.51 (m, 4H), 2.35 (s, 3H), 2.02 – 1.81 (m, 4H). MS, m/z = 386 (100) [M+H] . HRMS found:
(
M + H) 386.0490; C H Cl N O S requires (M +H), 386.0492.
16 17 2 3 2
4
-((4-Chloro-3-methylphenyl)amino)-N-(5-methylthiazol-2-yl)piperidine-1-carboxamide (20).
General Procedure A was followed using 93 (16 mg, 0.069 mmol) and 96 (13 mg, 0.058
1
mmol) to obtain 20 as a solid (10 mg, 47%). H-NMR (300 MHz; CDCl ): δ 7.12 (d, J 8.4
3
Hz, 1H), 6.93 (br. s., 1H), 6.48 (d, J 2.4 Hz, 1H), 6.39 (dd, J 2.9, 8.6 Hz, 1H), 4.16 - 4.04 (m,
2
1
3
H), 3.56 - 3.37 (m, 2H), 3.17 - 3.04 (m, 2H), 2.36 (s, 3H), 2.30 (s, 3H), 2.16 - 2.06 (m, 2H),
+
.43 - 1.31 (m, 2H). MS, m/z = 365 (100), 367 (30) [M+H] . HRMS found: (M + H)
65.1197; C H ClN OS requires (M +H), 365.1198.
17
21
4
4
-((4-Chloro-3-methylphenyl)amino)-N-(5-chlorothiazol-2-yl)piperidine-1-carboxamide (21).
General Procedure A was followed using 94 (18 mg, 0.069 mmol) and 96 (13 mg, 0.058
mmol). The crude residue was then purified by reverse phase preparatory HPLC using a
1
gradient of 95% water/ACN to 100% ACN to obtain 21 as a solid (2.5 mg, 11%). H-NMR
(
300 MHz; CDCl ): δ 7.24 - 7.12 (m, 2H), 6.61 - 6.45 (m, 2H), 4.22 - 4.03 (m, 2H), 3.56 -
3
3
1
3
.43 (m, 1H), 3.14 (ddd, J 3.3, 11.3, 14.3 Hz, 2H), 2.38 - 2.27 (m, 3H), 2.25 - 2.11 (m, 2H),
.61 - 1.39 (m, 2H). MS, m/z = 385 (100), 387 (60) [M+H]+. HRMS found: (M + H)
85.0650; C H Cl N OS requires (M +H), 385.0651.
16
18
2
4
4
-((4-Chloro-3-methylphenyl)(methyl)amino)-N-(5-chlorothiazol-2-yl)piperidine-1-
carboxamide (22). General Procedure A was followed using 94 (17 mg, 0.065 mmol) and 98
1
(
13 mg, 0.055 mmol) to obtain 22 as a solid (10 mg, 46%). H-NMR (300 MHz; CDCl ): δ
3

7
3
.22 - 7.11 (m, 2H), 6.69 (d, J 2.9 Hz, 1H), 6.62 (dd, J 3.0, 8.7 Hz, 1H), 4.30 - 4.16 (m, 2H),
.81 - 3.66 (m, 1H), 3.09 – 2.91 (m, 2H), 2.74 (s, 3H), 2.35 (s, 3H), 1.90 - 1.62 (m, 4H). MS,
+
m/z = 399 (100), 401 (60) [M+H] . HRMS found: (M + H) 399.0808; C H Cl N OS
17
20
2
4
requires (M +H), 399.0808.
-(4-Chloro-3-methylphenoxy)-N-(5-chlorothiazol-2-yl)piperidine-1-carboxamide
3
(23).
General Procedure A was followed using 94 (15 mg, 0.059 mmol) and 99 (12 mg, 0.053
mmol). The crude residue was then purified by reverse phase preparatory HPLC using a
1
gradient of 95% water/ACN to 100% ACN to obtain 23 as a solid (3.4 mg, 17%). H-NMR
(
300 MHz; CDCl ): δ 7.22 (d, J 8.8 Hz, 1H), 7.16 (s, 1H), 6.78 (d, J 3.3 Hz, 1H), 6.68 (dd, J
3
2
.9, 8.6 Hz, 1H), 4.34 (td, J 3.4, 6.5 Hz, 1H), 3.87 - 3.79 (m, 1H), 3.66 - 3.42 (m, 3H), 2.32
+
(
s, 3H), 2.04 – 1.86 (m, 3H), 1.67 - 1.55 (m, 1H). MS, m/z = 386 (100) [M+H] , 388 (60).
HRMS found: (M + H) 386.0490; C H Cl N O S requires (M +H), 386.0492.
16
17
2
3
2
(S)-3-((4-Chloro-3-methylphenyl)amino)-N-(5-methylthiazol-2-yl)piperidine-1-carboxamide
(24). General Procedure A was followed using 93 (16 mg, 0.067 mmol) and 100 (15 mg,
0
.067 mmol). The crude residue was then purified by reverse phase preparatory HPLC using
1
a gradient of 95% water/ACN to 100% ACN to obtain 24 as a solid (5.0 mg, 21%). H-NMR
(
300 MHz; CDCl ): δ 7.11 (d, J 8.58 Hz, 1H), 6.94 (s, 1H), 6.50 (d, J 2.86 Hz, 1H), 6.42 (dd,
3
J 3.0, 8.7 Hz, 1H), 4.12 - 4.02 (m, 1H), 3.82 - 3.62 (m, 1H), 3.52 - 3.41 (m, 1H), 3.40 - 3.29
(m, 1H), 3.14 - 3.01 (m, 1H), 2.36 (s, 3H), 2.28 (s, 3H), 2.10 – 1.97 (m, 1H), 1.92 – 1.79 (m,
+
1
H), 1.70 – 1.53 (m, 2H). MS, m/z = 365 (100) [M+H] . HRMS found: (M + H) 365.1197;
C H ClN OS requires (M +H), 365.1198.
17
21
4
(S)-3-((4-Chloro-3-methylphenyl)amino)-N-(5-chlorothiazol-2-yl)piperidine-1-carboxamide
(25). General Procedure A was followed using 94 (17 mg, 0.067 mmol) and 100 (15 mg,
0
.067 mmol). The crude residue was then purified by reverse phase preparatory HPLC using

1
a gradient of 95% water/ACN to 100% ACN to obtain 25 as a solid (5.3 mg, 21%). H-NMR
(
300 MHz; CDCl ): δ 7.16 - 7.10 (m, 2H), 6.51 (d, J 2.4 Hz, 1H), 6.43 (dd, J 2.5, 8.7 Hz,
3
1
H), 4.05 – 3.98 (m, 1H), 3.73 - 3.63 (m, 1H), 3.53 - 3.44 (m, 1H), 3.39 - 3.29 (m, 1H), 3.10
(dd, J 7.9, 13.4 Hz, 1H), 2.29 (s, 3H), 2.09 - 2.01 (m, 1H), 1.91 - 1.82 (m, 1H), 1.72 - 1.57
+
(
m, 2H). MS, m/z = 385 (100) [M+H] , 387 (60). HRMS found: (M + H) 385.0649;
C H Cl N OS requires (M +H), 385.0651.
16
18
2
4
(R)-3-((4-Chloro-3-methylphenyl)amino)-N-(5-methylthiazol-2-yl)piperidine-1-carboxamide
(26). General Procedure A was followed using 93 (31 mg, 0.13 mmol) and 101 (30 mg, 0.13
mmol). The crude residue was then purified by reverse phase preparatory HPLC using a
1
gradient of 95% water/ACN to 100% ACN to obtain 26 as a solid (2.8 mg, 6%). H-NMR
(
300 MHz; CDCl ): δ 7.11 (d, J 8.6 Hz, 1H), 6.94 (s, 1H), 6.50 (d, J 2.4 Hz, 1H), 6.42 (dd, J
3
2
.9, 8.4 Hz, 1H), 4.14 - 4.05 (m, 1H), 3.82 - 3.62 (m, 1H), 3.52 - 3.41 (m, 1H), 3.41 - 3.30
(m, 1H), 3.10 (dd, J 7.4, 13.3 Hz, 1H), 2.36 (s, 3H), 2.29 (s, 3H), 2.10 – 1.98 (m, 1H), 1.93 -
+
1
.81 (m, 1H), 1.76 - 1.52 (m, 2H). MS, m/z = 365 (100) [M+H] . HRMS found: (M + H)
65.1195; C H ClN OS requires (M +H), 365.1198.
3
17
21
4
(R)-3-((4-Chloro-3-methylphenyl)amino)-N-(5-chlorothiazol-2-yl)piperidine-1-carboxamide
(27). General Procedure A was followed using 94 (34 mg, 0.13 mmol) and 101 (30 mg, 0.13
mmol). The crude residue was then purified by reverse phase preparatory HPLC using a
1
gradient of 95% water/ACN to 100% ACN to obtain 27 as a solid (3.1 mg, 6%). H-NMR
(
300 MHz; CDCl ): δ 7.16 - 7.10 (m, 2H), 6.70 – 6.48 (m, 2H), 4.25 – 4.12 (m, 1H), 3.90 -
3
3
.80 (m, 1H), 3.53 - 3.17 (m, 3H), 2.31 (s, 3H), 1.90 - 2.05 (m, 1H), 1.75 – 1.54 (m, 3H). MS,
+
m/z = 385 (100) [M+H] , 387 (60). HRMS found: (M + H) 385.0648; C H Cl N OS
16
18
2
4
requires (M +H), 385.0651.

3
-((4-Chloro-3-methylphenyl)amino)-N-(5-methylthiazol-2-yl)pyrrolidine-1-carboxamide
(28). General Procedure A was followed using 93 (18 mg, 0.076 mmol) and 103 (16 mg,
1
0
8
1
–
.076 mmol) to obtain 28 as a solid (13 mg, 49%). H NMR (300 MHz, CDCl ): δ 7.14 (d, J
3
.6 Hz, 1H), 6.92 (s, 1H), 6.47 (d, J 2.6 Hz, 1H), 6.38 (dd, J 8.6, 2.9 Hz, 1H), 4.12 - 4.05 (m,
H), 3.82 – 3.74 (m, 1H), 3.66 - 3.52 (m, 2H), 3.44 - 3.34 (m, 1H), 2.39 - 2.18 (m, 7H), 2.05
+
1.95 (m, 1H). MS, m/z = 351 (100) [M+H] 353 (30). HRMS found: (M + H) 351.1041;
,
C H ClN OS requires (M +H), 351.1041.
16
19
4
3
-((4-Chloro-3-methylphenyl)amino)-N-(5-chlorothiazol-2-yl)pyrrolidine-1-carboxamide
(29). General Procedure A was followed using 94 (20 mg, 0.076 mmol) and 103 (16 mg,
1
0
8
3
1
1
.076 mmol) to obtain 29 as an oil (12 mg, 43%). H NMR (300 MHz, CDCl ): δ 7.14 (d, J
3
.6 Hz, 2H), 6.46 (d, J 2.6 Hz, 1H), 6.37 (dd, J 8.5, 3.0 Hz, 1H), 4.14 - 4.06 (m, 1H), 3.83 -
.69 (m, 1H), 3.65 - 3.50 (m, 2H), 3.44 – 3.34 (m, 1H), 2.39 - 2.21 (m, 4H), 2.09 – 1.93 (m,
1
3
H). C NMR (75 MHz, CDCl ) δ 159.1, 151.8, 144.7, 136.9, 132.6, 129.7, 123.8, 115.8,
3
12.2, 52.8, 51.9, 44.2, 31.4, 20.3 (1 quaternary C not observed). MS, m/z = 371 (100)
+
,
[
M+H] 373 (30). HRMS found: (M + H) 371.0492; C H Cl N OS requires (M +H),
15 16
2
4
3
71.0495.
3
-(4-Chloro-3-methylphenoxy)-N-(5-methylthiazol-2-yl)pyrrolidine-1-carboxamide
(30).
General Procedure A was followed using 93 (19 mg, 0.080 mmol) and 104 (17 mg, 0.080
1
mmol) to obtain 30 as a solid (9.3 mg, 33%). H-NMR (300 MHz; CDCl ): δ 7.23 (d, J 8.6
3
Hz, 1H), 6.92 (d, J 1.1 Hz, 1H), 6.74 (d, J 2.9 Hz, 1H), 6.63 (dd, J 2.9, 8.6 Hz, 1H), 4.99 -
1
3
4
1
1
.81 (m, 1H), 3.86 - 3.49 (m, 4H), 2.45 - 2.09 (m, 8H). C NMR (75 MHz, CDCl ) δ 159.5,
3
55.2, 152.5, 137.4, 133.2, 129.8, 126.7, 126.6, 118.2, 114.1, 75.9, 51.6, 44.0, 31.2, 20.3,
+
1.5. MS, m/z = 352 (100) [M+H] , 354 (30). HRMS found: (M + H) 352.0881;
C H ClN O S requires (M +H), 352.0881.
16
18
3
2