The discovery and preclinical evaluation of BMS-707035, a potent HIV-1 integrase strand transfer inhibitor

Article abstract of DOI:10.1016/j.bmcl.2018.05.027

BMS-707035 is an HIV-1 integrase strand transfer inhibitor (INSTI) discovered by systematic optimization of N-methylpyrimidinone carboxamides guided by structure-activity relationships (SARs) and the single crystal X-ray structure of compound 10. It was r

Full text of DOI:10.1016/j.bmcl.2018.05.027

Accepted Manuscript
The discovery and preclinical evaluation of BMS-707035, a potent HIV-1 in-
tegrase strand transfer inhibitor
B. Narasimhulu Naidu, Michael A. Walker, Margaret E. Sorenson, Yasutsugu
Ueda, John D. Matiskella, Timothy P. Connolly, Ira B. Dicker, Zeyu Lin,
Sagarika Bollini, Brian J. Terry, Helen Higley, Ming Zheng, Dawn D. Parker,
Dedong Wu, Stephen Adams, Mark R. Krystal, Nicholas A. Meanwell
PII:
S0960-894X(18)30429-3
https://doi.org/10.1016/j.bmcl.2018.05.027
BMCL 25844
DOI:
Reference:
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date:
Revised Date:
Accepted Date:
20 February 2018
10 May 2018
12 May 2018
Please cite this article as: Naidu, B.N., Walker, M.A., Sorenson, M.E., Ueda, Y., Matiskella, J.D., Connolly, T.P.,
Dicker, I.B., Lin, Z., Bollini, S., Terry, B.J., Higley, H., Zheng, M., Parker, D.D., Wu, D., Adams, S., Krystal, M.R.,
Meanwell, N.A., The discovery and preclinical evaluation of BMS-707035, a potent HIV-1 integrase strand transfer
inhibitor, Bioorganic & Medicinal Chemistry Letters (2018), doi: https://doi.org/10.1016/j.bmcl.2018.05.027
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Graphical Abstract
The discovery and preclinical evaluation of BMS-707035, a potent HIV-1 integrase strand transfer
inhibitor
B. Narasimhulu Naidu, Michael A. Walker, Margaret E. Sorenson, Yasutsugu Ueda, John D. Matiskella,
Timothy P. Connolly, Ira B. Dicker, Zeyu Lin, Sagarika Bollini, Brian J. Terry, Helen Higley, Ming
Zheng, Dawn D. Parker, Dedong Wu, Stephen Adams, Mark R. Krystal, and Nicholas A. Meanwell

The discovery and preclinical evaluation of BMS-707035, a potent HIV-1 integrase
strand transfer inhibitor
1
†
1
1
1
B. Narasimhulu Naidu, * Michael A. Walker , Margaret E. Sorenson, Yasutsugu Ueda, John D.
1
1
2
2
2
2
Matiskella, Timothy P. Connolly , Ira B. Dicker, Zeyu Lin, Sagarika Bollini, Brian J. Terry, Helen
2
3
3
4
3
2
Higley, Ming Zheng, Dawn D. Parker, Dedong Wu, Stephen Adams, Mark R. Krystal, and
1
Nicholas A. Meanwell
1
2
3
4
Departments of Discovery Chemistry, Virology, Pharmaceutical Candidate Optimization, Materials
Chemistry and Crystallography, Bristol-Myers Squibb Research and Development, 5 Research Parkway,
Wallingford, Connecticut 06492.
Keywords:
HIV-1 Integrase
Strand transfer inhibitors
INSTI
Pyrimidinone carboxamides
Antiviral activity
Phamacokinetics
ABSTRACT: BMS-707035 is an HIV-1 integrase strand transfer inhibitor (INSTI) discovered by
systematic optimization of N-methylpyrimidinone carboxamides guided by structure-activity
relationships (SARs) and the single crystal X-ray structure of compound 10. It was rationalized that the

unexpectedly advantageous profiles of N-methylpyrimidinone carboxamides with a saturated C2-
substitutent may be due, in part, to the geometric relationship between the C2-substituent and the
pyrimidinone core. The single crystal X-ray structure of 10 provided support for this reasoning and
guided the design of a spirocyclic series 12 which led to discovery of the morpholino-fused
pyrimidinone series 13. Several carboxamides derived from this bicyclic scaffold displayed improved
antiviral activity and pharmacokinetic profiles when compared with corresponding spirocyclic analogs.
Based on the excellent antiviral activity, preclinical profiles and acceptable in vitro and in vivo toxicity
profiles, 13a (BMS-707035) was selected for advancement into phase I clinical trials.
------------------------------------------------------------------------------------------------------------------------------
Human immunodeficiency virus-1 (HIV-1) is the causative agent of acquired immunodeficiency
syndrome (AIDS), and without treatment results in a debilitating disease that cripples the immune
system of the patient, increasing susceptibility to opportunistic diseases, ultimately leading to death.
Since the approval of zidovudine (AZT) as the first HIV-1 inhibitor in 1987, significant advances have
been made in the treatment of HIV- infection. The current preferred treatment, referred to as highly
active antiretroviral therapy (HAART) or combination antiretroviral therapy (cART), consists of
combinations of two or more drugs that target specific viral processes. Since the advent of HAART, the
quality of life and overall life span of the HIV-1 infected patients has significantly improved. Recent
UNAIDS estimates indicate that approximately 36.7 million people are living with HIV-1 globally and
that there were 1.8 million new infections and 1.0 million AIDS deaths in 2016, a significant decline
when compared with the 3.4 million new infections (47%) estimated to have occurred in 2001 and 2.3
1
million AIDS deaths (56%) in 2005. HIV-1 therapy requires life-long adherence to a treatment regimen
that if missed can lead to the rapid development of resistance to the one or more drugs used in the
combination. The development of resistance consequently limits the number of treatment options
available to an infected individual. Although current drug regimens are highly efficacious and generally

well tolerated, there remain concerns about long-term toxicities. Development of cross-resistance within
a mechanistic class and poor tolerability has necessitated a search for new drugs that are active against
resistant viruses and safer for long-term use.
Reverse transcriptase (RT), integrase (IN) and protease (PR) are three key enzymes involved in
the HIV-1 life cycle. HIV-1 integrase irreversibly inserts the nascent reverse-transcribed, double
2
stranded viral DNA into the host genome by a multistep biochemical process. There are two well-
characterized catalytic steps involved in the integration process, the first of which is denoted as 3’-
processing, in which the integrase enzyme cleaves a GT dinucleotide pair from each 3’-end of the viral
cDNA long-terminal repeats, a process that occurs in the cytoplasm. The second catalytic step, referred
to as strand transfer (ST), occurs in the nucleus, where integrase catalyzes the covalent insertion of
reverse-transcribed, double-stranded viral DNA into the host chromosome. Inhibitors specifically
targeting the inhibition of strand transfer reaction are designated as integrase strand transfer inhibitors
(
INSTIs).
3
Previously, starting from the triketoacid lead 1 (Figure 1), optimization arrived at the amide
4
ketoacid series 2, which displayed an improved antiviral activity in the cell culture HIV inhibition assay
when compared with tri- or di-ketoacid derivatives. Further optimization studies focused on embedding
the crucial structural elements of 2 into heterocyclic ring systems, a survey that includes pyrrolinedione
5
6
carboxamides 3, the pyrimidinone carboxamides 4 and pyridinone carboxamides 5. Raltegravir, the
7
first integrase inhibitor approved for treating HIV/AIDS, belongs to the pyrimidinone class. In this
article, we describe efforts to optimize the antiviral and pharmacokinetic properties of pyrimidinone
carboxamides that led to the discovery of BMS-707035 (13a), a candidate that was advanced into
clinical trials.

Figure 1. Evolution of INSTI chemotypes
Initial efforts were focused on the identification of an optimal substituent at the C2-position of
the pyrimidinone core, and to this end a variety of C2-substituents were surveyed, including alkyl and
8
9
10
cycloalkyl moieties, saturated heterocycles, acyclic and cyclic amines, and aryl and heteroaryl
groups. The strand transfer and cell culture inhibition data for selected compounds which form the
foundation for the discovery of BMS-707035 (13a) are compiled in Table 1. Compounds 6–10
displayed good inhibitory activity in the strand transfer assay, which translated into potent antiviral
8
activity in cell culture. However, there are subtle differences in the effects of the C2-substituents based
on their structural attributes. The unsaturated cyclopentenyl- and 2,5-dihydrofuranyl-substituted analogs
6
and 8 are less potent than their corresponding saturated homologues 7 and 9, respectively. In addition,
the C2-carbocyclic carboxamides 6 and 7 exhibit reduced activity in the presence of human serum
albumin (HSA), attributed to their lipophilic nature, leading to a lower free fraction. Not surprisingly,
the more polar C2-heterocyclic compounds 8 and 9 exhibit a more moderate loss of activity in the

presence of HSA. Remarkably, however, the C2-methylthiotetrahydrofuranyl carboxamide 10 not only
exhibited potent cellular inhibitory activity but also displayed a lower serum shift than the
tetrahydrofuranyl analog 9, despite the fact that it is more lipophilic (cLogP = 0.98) than 9, (cLogP =
0
.43).
Table 1. Structure-activity relationships (SAR) for C2-substituted pyrimidinone carboxamide-based
HIV-1 integrase inhibitor
a
EC50 (nM)
ST
IC (nM)
Serum-
shift
Compd.
R
cLogP
1.79
5
0
b
1
0% FBS
HSA
6
8
2
84
1,534
214
18.3
23.8
8.7
7
8
9
9
206
9
1.96
1
9
1,783
0.42
0.43
0.98
4
4
60
26
6.7
3.7
1
0
7
a
b
ST = Strand transfer assay HSA-EC was determined in the presence of 10% FBS and 15 mg/mL
5
0
human serum albumin
Compounds 9 and 10 were evaluated for their pharmacokinetic properties in rats and the results
8
are summarized in Table 2. Following intravenous (IV) administration, both compounds displayed low

clearance (CL), short half-lives (t ) and low volumes of distribution (V ). When dosed orally as a
1
/2
ss
solution, exposure was high and bioavailability (F) was >65% for both compounds.
Table 2. Pharmacokinetic profiles of 9 and 10 in male Sprague-Dawley rats
9
1
10
1
a
IV dose (mg/kg)
CL (mL/min/kg)
t (h)
4.9
3.6
1.63
0.14
5
1.08
0.15
5
1
/2
V (L/kg)
ss
a
PO dose (mg/kg)
Cmax (M)
10.45
0.67
33.89
68
8.04
1.67
39.45
67
tmax (h)
AUC (M*h)
F %
a
Vehicle: PEG-400/Ethanol (90:10)
It was reasoned that the subtle differences in the antiviral activity of the saturated and
unsaturated matched pairs and the lower serum shift for 10 may be, in part, a function of the geometrical
relationship between the pyrimidinone core and the C2-substituent. In the case of the unsaturated
2
substituent, it was hypothesized that the sp hybridization of the connecting carbon would favor a
coplanar arrangement between the pyrimidinone and the C2 substituent. In contrast, the tetrahedral
3
geometry (sp hybridization) associated with the connecting carbon atom in the saturated analogues
would theoretically favor a topography in which the pyrimidinone ring is perpendicular to the plane of
the C2 substituent. In support of this hypothesis, the single crystal X-ray structure of 10 revealed the
pyrimidinone core to be perpendicular to the plane of the tetrahydrofuranyl ring (Figure 2). With this

observation in mind, the next phase of the SAR survey focused upon designing compounds that would
lock the geometric relationship between the pyrimidinone core and the C2-substituent.
Figure 2. Single crystal X-ray structure of 10.
To that end, it was envisioned that this could be accomplished by linking the pyrimidinone N1 to
the 3’-carbon atom of the tetrahydrofuran ring to establish a rigid framework on which to deploy a
spirocycle, as depicted conceptually by 11 in Figure 3. In this arrangement, the pyrimidinone core is
perpendicular to the plane of the spirocyclic ring, thereby mimicking the topography observed in the
solid state structure of 10. This design concept was examined initially in the context of the
spirocyclopentane derivative 12a for which the single crystal X-ray confirmed the design hypothesis
(Figure 4).

Figure 3. Design principle leading to the spirocyclic fused pyrimidinone chemotype 11.
Figure 4. Molecular and single crystal X-ray structure 12a
The in vitro, antiviral activity and rat clearance of the series of spirocyclic benzyl carboxamides
1
2a-f are compiled in Table 3. Compound 12a displayed superior antiviral potency compared to the
progenitor 7 and, interestingly, displayed a lower serum shift despite being more lipophilic than 7.
Benzylamide 12b, containing a lipophilic ortho-methyl substituent, maintained both the intrinsic
enzyme inhibition and antiviral activity of 12a, but suffered from a substantial (30-fold) loss of activity
in the presence of HSA. In an effort to reduce the effect of serum, the benzylamides series 12c-f
incorporating hydrophilic substituents were surveyed. The methylsulfonyl-substituted analog 12c
displayed potent antiviral activity with a modest 8-fold shift in the EC value in the presence of HSA,
50
whereas the dimethylsulfamoyl analog 12d exhibited slightly reduced activity in both the strand transfer
and cell culture inhibition assays compared to 12a with or without added human serum albumen. On the
other hand, the morpholinosulfonamide analog 12e exhibited reduced potency in the biochemical assay
but demonstrated a similar antiviral profile to 12a in both cell culture assays. Lastly, the cyclic
sulfonamide 12f exhibited a comparable antiviral profile to 12a in all three assays. Of the spirocyclic

carboxamide derivatives that were evaluated in a rat IV pharmacokinetic screen, 12a and 12b displayed
intermediate clearance values, while 12c exhibited high clearance.
Table 3. Structure activity relationships for spirocycle pyrimidinone carboxamides 12a-f
a
EC50 (nM)
c
1
2
ST IC50
Rat CL
Compd.
R
R
cLogP
2.49
(nM)
b
(mL/min/kg)
1
0% FBS HSA
1
2a
H
H
H
Me
H
3
2
5
1
3
3
19
149
8
18
1
2b
2
4
3.00
1.30
1.49
1.30
17
46
1
2c
S(O) Me
2
d
1
2d
S(O)
2
NMe
2
H
6
27
14
nd
1
2e
H
11
nd
nd
O
O
S
1
2f
N
H
5
2
21
1.55

a
b
ST = Strand transfer, HSA-EC values were determined in the presence of 10% FBS and 15 mg/mL
5
0
c
human serum albumin, Male Sprague-Dawley rats were administered a 1 mpk IV dose of the compound
in a vehicle of PEG400/Ethanol (90:10) and blood samples were collected at 5 and 30 min and 1, 2 and
d
4
h following drug administration. nd = not done

Although the spirocyclic derivatives 12a-f demonstrated targeted antiviral potency, these
compounds still suffered from poor pharmacokinetic and pharmaceutical/physicochemical properties.
These compounds displayed aqueous solubility of less ≤5 g/ml, providing a focus for further
optimization that was also anticipated to offer the potential for enhanced pharmacokinetic properties.
To reduce the lipophilicity of 12a, the effect of replacing the cyclopentyl ring with a gem-dimethyl
moiety was explored, while the introduction of a heteroatom into the fused B ring was examined as an
avenue toward further increasing polarity (Figure 5). The initial efforts were directed at scaffold 13a
which incorporates an embedded morpholine heterocycle, a moiety that has found extensive application
as a means of improving physicochemical properties. Analysis of developability data for compounds
containing a variety of heterocyclic rings has found that those containing a morpholine heterocycle fared
well when compared to other heterocycles, generally displaying improved aqueous solubility, lower
1
1
protein binding and a lower potential for CYP450 inhibition. This scaffold was anticipated to exhibit
reduced lipophilic character as reflected in the LogP value of 1.02 calculated for 13a which is
substantially lower than the LogP of 2.49 calculated for 12a.
Figure 5. Design of morpholine embedded pyrimidinone scaffold

The biochemical, cell culture activities and in vivo properties of bicyclic carboxamides 13a-j are
compiled in Table 4, with the prototype of the series, 4-fluorobenzyl carboxamide 13a, displaying an
antiviral profile comparable to the spirocyclic carboxamide 12a. The introduction of a small
hydrophobic methyl (13b) and chloride (13c) substituents at the meta-position of the benzyl ring did not
significantly affect the intrinsic antiviral activity, although the HSA-derived potency shift was
exacerbated. In contrast, in the presence of HAS, the ortho-methyl and ortho-chloro analogs 13d and
1
1
3e, respectively, exhibited substantially reduced activities. In contrast to the observations made with
3d-e, analogs 13f-j, containing polar ortho-substituents, exhibited potent antiviral activities and
displayed lower potency shift in the presence of HSA. The morpholino analog 13f showed potent
antiviral activity that was reduced about 10-fold in the presence of HSA. The ortho-methylcarboxamide
1
3g displayed an interesting profile, expressing potent antiviral activity that was negligibly affected by
HSA in spite of reduced potency in the biochemical assay. Finally, the methylsulfonyl (13h), the
dimethylsulfamoyl (13i) and the cyclic-sulfonamide (13j) analogs exhibited potent antiviral activities
that were subject to modest ≤4-fold shift in the EC value in the presence of HSA. Several of these
50
compounds were evaluated in rat pharmacokinetic studies and the clearance data following IV dosing is
presented in the Table 4. Compounds 13a, 13b and 13h displayed low clearance whereas 13f, 13g, 13i
and 13j exhibited moderate to high clearance in the rat.
Table 4. Enzyme inhibitory and antiviral data for pyrimidinone carboxamides 13a-j
a
EC50 (nM)
c
1
2
ST IC50
Rat CL
(mL/min/kg)
Compd.
R
R
cLogP
(nM)
b
1
0% FBS HSA

1
3a
H
H
H
Me
Cl
H
3
3
2
1
17
90
84
557
857
19
3
1.02
1.53
1.63
1.53
1.63
0.90
0.09
-0.17
0.02
0.08
9.6
11
1
3b
d
1
3c
H
3
2
nd
1
3d
Me
Cl
2
21
53
2
nd
nd
29
22
6.9
62
38
1
3e
H
10
5
1
3f
H
1
3g
CO
2
NHMe
H
11
6
2
1
3h
S(O)
2
Me
H
3
6
1
3i
S(O)
2
NMe
2
F
10
9
3
12
7
1
3j
H
2
a
b
ST = Strand transfer, HSA-EC were determined in the presence of 10% FBS and 15 mg/mL human
5
0
c
serum albumin, Male Sprague-Dawley rats were administered a 1 mpk IV dose of the compound in a
vehicle of PEG400/Ethanol (90:10) and blood samples were collected at 5 and 30 min and 1, 2 and 4 h
following drug administration. nd = not done.
d
Compound 13a with potent antiviral activity and low potency-shift (10x) in the presence of
either 15 mg/mL HSA or 40% human serum (HS) was further evaluated in additional in vitro studies,
the results of which are summarized in Table 5. Plasma protein binding of 13a was high in all species
(
93.7%) and the compound was not overtly cytotoxic to several cell lines, with CC values of 45 M.
50
The crystalline solubility of 13a ranged from 9 to 63 g/mL in the physiological pH range at room

temperature and increased as the pH increased, all the while demonstrating satisfactory solution and
solid-state stability under various conditions. Caco-2 studies revealed high permeability with no
indication of efflux. In studies designed to assess the potential for drug-drug interactions, the IC50
values toward human CYP1A2, CYP2C9, CYP2C19, CYP2D6 and CYP3A4 enzymes were >40 M,
classifying 13a as a relatively weak CYP inhibitor, for which drug interactions involving CYP substrates
were not anticipated. Minimal transactivation of human PXR (EC20A and EC60A > 50 M) or induction
of CYP3A4 mRNA (tested up to 100 M of 13a in Fa2N-4 immortalized human hepatocytes) was
observed, and thus, significant induction of CYP3A4 was not expected. Compound 13a was not a
substrate of human P-glycoprotein (P-gp) and was a relatively weak inhibitor of this efflux pump (5%
3
inhibition of [ H]-digoxin transport at 10 M), and, therefore, drug interactions involving P-gp were not
expected. The half-life values of 13a in a phase I metabolic oxidation assay (CYP-NADPH) were  96
minutes in human, dog, monkey and rat, data that suggest low potential for oxidation by the CYPs.
Compound 13a was not mutagenic in a bacterial reverse mutation assay (Salmonella typhimurium
strains TA98 and TA100) at 5000 g/plate with and without metabolic activation. An in vitro
cytokinesis-blocked micronucleus study was performed to determine the potential of 13a to induce
micronuclei in cultured Chinese hamster ovary (CHO) cells exposed to 500 g/mL 13a (1.4 mM).
Compound 13a was not clastogenic in CHO cells exposed either to 1.4 mM for 3 hours in the presence
of metabolic activation or to 0.36 mM for 22.5 hours in the absence of metabolic activation.
Table 5. In vitro profiling data for 13a.
Molecular weight
347.34

ST-IC50
3 (n = 98) nM
1
1
4
0%FBS-EC50
2.5 (n = 23) nM
5 mg/mL HSA-EC50
0% HS-EC50
16.5 (n = 14) nM
25.2 (n = 18) nM
Plasma protein binding
CC50
93.7% (rat), 95.9% (cyno), 98.5% (dog),
9
7.6% (human)
>200 (n = 44) M
45 M
HEPG2-Cytotoxicity IC50
Melting point = 179-180 C
Experimental Log D = 2.02 at pH 6.5
Physicochemical properties
Experimental pK = 6.6  0.2
a
cLogP = 1.02
9
6
g/mL at pH 2.3, 21 g/mL at pH 6.1
3 g/mL at pH 7.3 and 234 g/mL at pH 8.5
Solubility
Caco-2, (3 M) AB/BA
CYP inhibition IC50
PXR-TA EC50
334/240 nm/sec
All >40 M
>50 M
>120 (rat), 96 (cyno), 107 (dog), 110 (human)
CYP-NADPH t1/2
Genotoxicity
min
Ames TA98: negative
In vitro micronucleus: negative

Compound 13a emerged as a lead compound based on its antiviral activity and in vitro lead
profiling properties and was advanced into full pharmacokinetic profiling studies in three preclinical
species. The results of pharmacokinetic studies conducted in the rat, monkey, and dog, performed using
PEG 400/ethanol (90:10) solutions, are summarized in Table 5. Compound 13a was characterized as a
low clearance compound in the rat, dog and monkey with moderate to long elimination half-lives in all
species. The volume of distribution was low across the species, indicating that compound distribution
outside of the plasma is minimal. When 13a was dosed in solution, the oral exposure was high and
absolute oral bioavailabilities ranged between 56 and 129%. The tmax of 0.25 h suggests rapid
absorption of the compound in all 3 species. In dog cardiovascular and two week rat toxicity studies,
1
3a (BMS-707035) demonstrated good safety margins data that collectively contributed to its selection
as a clinical candidate
Table 5. Pharmacokinetic profiling data for 13a in the rat, monkey and dog
a
b
c
Rat
Monkey
Dog
d
IV dose (mg/kg)
CL (ml/min/kg)
t (h)
0.87
9.7
1
1
6.8
2.0
6.0
4.0
6.5
1
/2
V (L/kg)
0.86
4.4
0.87
5
0.45
5.2
ss
d
PO dose (mg/kg)
Cmax (M)
4.51
0.25
19.1
86
6.12
0.25
19.2
72.8
0.25
162
129
tmax (h)
AUC (M*h)
F %
56
a
b
c
d
Male Sprague-Dawley rats, cynomolgus monkeys, beagle dogs, Vehicel: PEG-400/Ethanol (90/10,
v/v) solution

The synthesis of C2-substituted pyrimidinone carboxamides 6-10 is outlined in Schemes 1 and 2.
Treatment of readily accessible nitriles 14a-c with N-methylhydroxylamine hydrochloride in the
presence of base provided the adducts 15a-c, which were subsequently reacted with diethyl
acetylenedicarboxylate to afford inseparable mixtures of 16a and 17a with nitrile 14a or 2,5-dihydro-
1
2
1
,2,4-oxadiazole intermediates 17b-c with nitriles 14b-c. Heating a xylenes solution of 16a or 17a-c at
1
3
reflux provided the desired pyrimidinone esters 18a-c. Treatment of 18a and 18c with benzoic
anhydride in pyridine furnished benzoate esters 19a and 19c. Oxidation of intermediate 19a with
mCPBA yielded the sulfone 20, which was subsequently converted into cyclopentyl derivative 21 by
reacting with 1,4-diiodobutane in the presence of Cs CO . Finally, heating 21 with 4-fluorobenzylamine
2
3
and triethylamine provided the final compound 6, the methylsulfonyl moiety was expelled by retro-ene
reaction under thermal condition. The C2-cyclopentyl compound 7 was prepared by heating
intermediate 18b with 4-fluorobenzylamine.
Scheme 1. Synthesis of 6 and 7

Treatment of methylthiofuranyl intermediate 19c with tert-butyl carbonazidate (Boc-N ) in the
3
presence of anhydrous ferrous chloride furnished the 2,5-dihydrofuranyl derivative 24. Reaction of
Boc-N with 19c generates a transient sulfanylidene intermediate which instantaneously undergoes
3
retro-ene reaction to afford 24. Heating 24 with 4-fluorobenzylamine and triethylamine yielded final
compound 8, which was subsequently converted to C2-tetrahydrofuran-3-yl derivative 9 by reducing the

double bond under hydrogenation condition. Methylthiotetrahydrofuranyl analog 10 was obtained by
heating 19c with excess 4-fluorobenzylamine and triethylamine in DMF.
Scheme 2. Synthesis of 9 and 10
The spirocyclic pyrimidinone carboxamides 12a-f were prepared following the synthesis
sequence shown in Scheme 3. Alkylation of cyclopentylcarbonitrile 25 with 3-chloro-4-iodopropane led
to the intermediate 26. Heating 26 with hydoxylamine hydrochloride in the presence of base gave
spirocyclic intermediate 13, which when treated with diethyl acetylenedicarboxylate yielded the tricyclic
spirocycle 28. Upon heating the solution of 28 in refluxing 2,4-dichlorotoluene afforded the desired
pyrimidinone ester 29. The final spirocyclic pyrimidinone carboxamides 12a-f were prepared by
heating 29 with appropriate benzylamine 22 in the presence of triethylamine.
Scheme 3. Synthesis of spirocyclic pyrimidinone carboxamides 12

Finally, the bicyclic pyrimidinone carboxamides 13a-j were prepared in five steps starting from
simple starting materials as illustrated in Scheme 4. The nitrile 31 was prepared from acetone
1
4
cyanohydrin and 2-chloroethanol by following the procedure reported in the literature. Treatment of
1 with 50% aqueous hydroxylamine in the presence of sodium iodide provided intermediate 32, which
upon exposure to diethyl acetylenedicarboxylate yielded intermediate 33. Pyrolysis of 33 in boiling
,2,4-trimethylbenzene furnished the pyrimidinone ester 34, which was converted into the desired
3
1
1
5
carboxamides 13a-j by reacting 34 with appropriate benzylamine 22 and triethylamine at 90 C.
Scheme 4. Synthesis of pyrimidinone carboxamides 13

Observations of differences in the biological activities between the unsaturated and saturated C2-
substituted pyrimidinone carboxamides led to the hypothesis that the geometric arrangement between
the C2-substituent and the pyrimidinone core was an important factor. This led to the concept of a
spirocycle-based pyrimidinone chemotype designed to mimic what was hypothesized to be the bound
conformation of 10. The synthesis and evaluation of the spirocyclic carboxamides 12a-f validated the
design proposal, but these molecules suffered from poor physicochemical and pharmacokinetic
properties. This was addressed by truncating the spirocyclic ring and introducing a heteroatom into the
fused saturated ring of 12 to provide compounds with reduced lipophilicity. The antiviral activity,
preclinical pharmacokinetics and safety profile of BMS-707035 (13a) supported continued development
into phase I clinical trials conducted in normal healthy human volunteers. Single oral doses of 200, 600,
8
00 and 1200 mg of 13a were well tolerated with no clinically significant adverse events noted. The
mean C24h following the 600 mg QD dose was 340 ng/mL which is approximately 1.9-fold the protein
binding-adjusted-EC value of 180 ng/mL, lending support to a 600 mg QD dose being sufficient to
9
0
significantly inhibit viral replication in HIV-1-infected subjects. Unfortunately, further development of
3a was halted due to the occurrence of convulsions in two dogs with high Cmax exposure ( 67
1
g/mL) in a 12 month toxicological study.
Corresponding Author
*
B. Narasimhulu Naidu, Email: b.n.naidu@viivhealthcare.com. Phone: +1-203-643-6455
Present Address
†
ViiV Healthcare, 36 East Industrial Road, Branford, CT 06405, USA.

ACKNOWLEDGMENTS
The authors would like to thank members of the Department of Discovery Synthesis (DDS) and
Biocon-Bristol Myers Squibb Research Center (BBRC) for scaling up key intermediates, analytical team
for analytical experiments, lead profiling group for evaluating compounds in an in vitro ADME&T
assays and discovery toxicology team for in vivo safety assessment in dogs and rats.
REFERENCES
1
2
3
. http://www.unaids.org/en/resources/fact-sheet; accessed Sept 1, 2017. .
. Craigie, R. The molecular biology of HIV integrase. Future Virol. 2012;7:679-686
. Reference for compound 1; a) Sechi,M.; Sannia, L.; Carta, F.; Palomba, M.; Dallocchio, R.;
Dessi, A.; Derudas, M.; Zawahur, Z. Neamati, N. Design of novel bioisostere of beta-diketo acid
inhibitors of HIV-1 integrase. Antivir. Chem. Chemother. 2005;16:41-61. B) Walker, M. A.;
Johnson, T.; Ma, Z.; Banville, J.; Remillard, R.; Kim, O.; Zhang, Y.; Staab, A.; Wong, H.; Torri,
A.; Samanta, H.; Lin. Z.; Deminie, C.; Terry, B.; Krystal, M. Meanwell, N. Triketoacid
inhibitors of HIV-integrase: A new chemotype useful for probing the integrase pharmacophore.
Bioorg. Med. Chem. Lett. 2006;16:2920-2924.
4
. (a) Walker MA, Banville J, Johnson T, Ma Z, Remillard R, Plamondron S, Bouthillier G, Martel
A, Torri AF, Casperson M, Bollini S, Terry BJ, Lin Z, Samanta HK, Krystal M, Zheng M,
Meanwell NA. 2009. Optimization of compound efficiency in the context of discovering an
th
amide ketoacid based HIV-integrase inhibitor with oral antiviral activity. 237 ACS National
Meeting, Salt Lake City, UT, USA, March 22-26, 2009, poster # MEDI 204. (b) Walker, M. A.;
Johnson, T.; Ma, Z.; Zhang, Y.; Banville, J.; Remillard, R.; Plamondon, S.; Pendri, A.; Wong,

H.; Smith, D.; Torri, A.; Samanta, H.; Lin. Z.; Deminie, C.; Terry, B.; Krystal, M. Meanwell, N.
Exploration of the diketoacid integrase inhibitor chemotype leading to the discovery of the
anilide-ketoacids chemotype. Bioorg. Med. Chem. Lett. 2006;16:5818-5821.
5
. Naidu, B. N; Banville, J.; Bollini, S.; Bouthillier, G.; Gulgeze, H. G.; Johnson, T.; Krystal, M.;
Lin, Z.; Martel, A.; Meanwell, N. A.; Parker, D. D.; Plamondon, S.; Remillard, R.; Samanta, H.
K.; Sorenson, M. S.; Terry, B. J.; Torri, A. F.; Walker, M. A.; Zhenh, M.; Zupping, M. Design,
synthesis and HIV-1 integrase inhibitory activity of N-benzyl-4-hydroxy-5-oxo-2,5-dihydro-1H-
th
pyrrole-3-carboxamides. The 237 American Chemical Society National Meeting, Salt Lake
City, UT, March 22-26, 2009, poster # MEDI 205.
6
. Three groups independently discovered pyrimidinone carboxamides as INSTIs: (a) Di Francesco,
M. E.; Gardelli, C.; Harper, S.; Matassa, V. G.; Muraglia, E.; Nizi, E.; Pace, P.; Pacini, B.;
Petrocchi, A.; Pioma, M.; Summa, V. Dihydroxypyrimidine carboxamide inhibitors of HIV
st
integrase. World Patent Application WO 2003035076, May 1 , 2003.; (b) Walker, M.; Gulgeze,
H. B.; Naidu, B. N; Sorenson, M. E.; Ueda, Y.; Matiskella, J. HIV integrase inhibitors. World
th
Patent Application WO 2004062613, July 29 , 2004.; (c) Fuji, M.; Matsushita, S.; Mikamiyama,
H. Antiviral agents containing nitrogen-containing heteroaromatic compounds. Japan Kokai
nd
Tokkyo Koho JP 2004244320, September 2 , 2004.
7
. Summa, V.; Petrocchi, A.; Bonelli, F.; Crescenzi, B.; Donghi, M.; Ferrara, M.; Fiore, F.;
Gardelli, C.; Paz, O. G.; Hazuda, D. J.; Jones, P.; Kinzel, O.; Laufer, R.; Monteaguda, E.;
Muraglia, E.; Nizi, E.; Orvieto, F.; Pace, P.; Pescatore, G.; Scarpelli, R.; Stillmock, K.; Witmer,
M. V.; Rowley, M. Discovery of Raltegravir, a potent, selective orally bioavailable HIV-
integrase inhibitor for the treatment of HIV-AIDS infection. J. Med. Chem. 2008;51:5843-5855.

8
. (a) Naidu, B.N.; Sorenson, M. E.; Patel, M.; Ueda, Y.; Banville, J.; Beaulieu, F.; Bollini, S.;
Dicker, I.; Higley, H.; Lin, Z.; Pajor, L.; D. Parker, D.D; Terry, B.; Zheng, M.; Martel, A.;
Meanwell, N. A.; Krystal, M.; Walker, M. A. The Synthesis and HIV-1 Integrase Inhibitory
Activity of C2-C-Linked Heterocyclic-5-hydroxy-6-oxo-dihydropyrimidine-4-carboxamides.
th
The 240 American Chemical Society National Meeting, Boston, MA, March 22-26, 2009,
poster # MEDI 130. (b) Naidu, B.N.; Sorenson, M. E.; Patel, M.; Ueda, Y.; Banville, J.;
Beaulieu, F.; Bollini, S.; Dicker, I.; Higley, H.; Lin, Z.; Pajor, L.; D. Parker, D.D; Terry, B.;
Zheng, M.; Martel, A.; Meanwell, N. A.; Krystal, M.; Walker, M. A. The Synthesis and
evaluation of C2-carbon-linked heterocyclic-5-hydroxy-6-oxo-dihydropyrimidine-4-
carboxamides as HIV-1 Integrase Inhibitors. Bioorg. Med. Chem. Lett. 2015;25:717-720.
9
. Sorenson, M.; Naidu, B. N.; Matiskella, J.; Connolly, T.; Ueda, Y.; Meanwell, N. A.; Bollini, S.;
Samanta, H. K.; Terry, B. J.; Lin, Z.; Dicker, I.; Parker, D.; Pajor, L.; Zheng, M; Krystal, M.;
Walker, M. A. Study of C 2-N-linked heterocyclic pyrimidin-6-oxa-5-hydroxy-4-carboxamide
th
derivatives as HIV integrase inhibitors. The 240 American Chemical Society National Meeting,
Boston, MA, March 22-26, 2009, poster # MEDI 133.
1
1
0. Sorenson, M.; Bollini, S.; Casperson, M.; Krystal, M.; Lin, Z.; Meanwell, N.; Naidu, B. N.;
Samanta, H.; Terry, B.; Torri, A.; Walker, M. 2-Aryl substituted pyrimidin-6-oxa-5-hydroxy-4-
th
carboxamides as HIV integrase inhibitors. The 36 ACS Northeast Regional Meeting 2009,
Hartford, CT, October 7-10, 2009, poster # 74284.
1. Ritchie, T. J.; Macdonald, S. J. F.; Peace, S.; Pickett, S. D.; Luscombe, C. N. The developability
of heteroaromatic and heteroaliphatic rings-do some have a beter pedigree as potential drug
molecules than others? Med. Chem Commun. 2012;3:1062-1069.

1
1
2. Naidu, B. N.; Sorenson, M. E. Facile One-pot synthesis of 2,3,5-substituted-1,2,4-oxadiazolines
from nitriles in aqueous solution. Org. Lett. 2005;7:1391-1393.
3. Naidu, B. N. Tandem retro-Michael addition/Claisen rearrangement/intramolecular cyclization:
One-pot synthesis of densely functionalized ethyl dihydropyrimidine-4-carboxylates from simple
building blocks. Synlett, 2008;547-550. b) Pye, P. J.; Zhong, Y. L.; Jones, G. O.; Reamer, R. A.;
Houk, K. N.; Askin, D. A polar radical pathway to assemble the pyrimidinone core of HIV
integrase inhibitor raltegravir potassium. Angew. Chem. Int. Ed. 2008;47:4134-4136.
1
4. a) Navalokina, R. A.; Zil’berman, E. N. Production of unsymmetrical ethers from2-
hydroxyisobutyronitrile and various alcohols. J. Org. Chem. USSR (Engl. Trans.), 1980;16:1382-
1
386. b) Ramalingam, K. Process for preparing 2-methoxyisobutylisonitrile. United States Patent
th
US 4864051, September 5 , 1989.
1
5. Compound 13a: To a solution of ethyl 3-hydroxy-9,9-dimethyl-4-oxo-4,6,7,9-
tetrahydropyrimido[2,1-c][1,4]oxazine-2-carboxylate, (3.0 g, 11.19 mmol) in DMF (20 mL) and
ethanol (10 mL) was added triethylamine (1.55 mL) followed by 4-fluorobenzylamine (3.82 mL,
3
3.57 mmol). The mixture was stirred at 90 C for 2 h and then concentrated. The resultant oil
was partitioned between ethyl acetate (50 mL) and 1N aqueous HCl (35 mL). The aqueous layer
was back-extracted with ethyl acetate (20 mL) and the organic layers were combined and washed
with H O (4 x 20 mL) and brine then dried (Na SO ) and concentrated. The brown residue was
2
2
4
triturated with ether and the solids filtered and washed with ether. The pale brown solids were
recrystallized from 95:5 MeOH/H O to give the title compound as a colorless needles (3.18 g,
2
1
8
5
2% yield). H NMR (500 MHz, CDCl ) : 11.96 (1H, s), 7.77 (1H, brs), 7.30 (2H, dd, J = 8.4,
3
.3 Hz), 7.04 (2H, t, J = 8.7 Hz), 4.57 (2H, d, J = 6.1 Hz), 4.01 (4H, s), 1.56 (6H, s). HRMS
(
M+H) calcd for C H FN O : 348.13597; found: 348.1365. Anal calcd for C H FN O : C,
17
19
3
4
17 18
3
4
5
8.78; H, 5.22; N, 12.09. Found: C, 58.38; H, 5.23; N, 11.80.

Highlights:
Synthesis and antiviral activities of series of pyrimidinone carboxamide described
BMS-707035 is a potent HIV-1 integrase strand transfer inhibitor
It displayed good preclinical profiles; advanced to clinical trials
Development halted due to the findings in 12-month dog safety assessment study