Discovery of a Potent Nonpeptidomimetic, Small-Molecule Antagonist of Cellular Inhibitor of Apoptosis Protein 1 (cIAP1) and X-Linked Inhibitor of Apoptosis Protein (XIAP)

Article abstract of DOI:10.1021/acs.jmedchem.6b01877

XIAP and cIAP1 are members of the inhibitor of apoptosis protein (IAP) family and are key regulators of anti-apoptotic and pro-survival signaling pathways. Overexpression of IAPs occurs in various cancers and has been associated with tumor progression and resistance to treatment. Structure-based drug design (SBDD) guided by structural information from X-ray crystallography, computational studies, and NMR solution conformational analysis was successfully applied to a fragment-derived lead resulting in AT-IAP, a potent, orally bioavailable, dual antagonist of XIAP and cIAP1 and a structurally novel chemical probe for IAP biology.

Full text of DOI:10.1021/acs.jmedchem.6b01877

Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)
Article
Discovery of a Potent Non-peptidomimetic, Small-Molecule
Antagonist of Cellular Inhibitor of Apoptosis Protein 1
(cIAP1) and X-linked Inhibitor of Apoptosis Protein (XIAP)
Emiliano Tamanini, Ildiko Maria Buck, Gianni Chessari, Elisabetta Chiarparin, James E. H. Day,
Martyn Frederickson, Charlotte M. Griffiths-Jones, Keisha Hearn, Tom D. Heightman, Aman
Iqbal, Christopher Norbert Johnson, Edward J. Lewis, Vanessa Martins, Torren Peakman,
Michael Reader, Sharna J. Rich, George A. Ward, Pamela A. Williams, and Nicola E. Wilsher
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 11 May 2017
Downloaded from http://pubs.acs.org on May 12, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 47
Journal of Medicinal Chemistry
1
2
3
4
Discovery of a Potent Nonꢀpeptidomimetic, SmallꢀMolecule Antagonist of Cellular
Inhibitor of Apoptosis Protein 1 (cIAP1) and Xꢀlinked Inhibitor of Apoptosis Protein
(XIAP)
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Emiliano Tamanini*, Ildiko M. Buck, Gianni Chessari, Elisabetta Chiarparin, James E. H.
Day, Martyn Frederickson, Charlotte M. GriffithsꢀJones, Keisha Hearn, Tom D. Heightman,
Aman Iqbal, Christopher N. Johnson, Edward J. Lewis, Vanessa Martins, Torren Peakman,
Michael Reader,. Sharna J. Rich, George A. Ward, Pamela A. Williams, Nicola E. Wilsher.
Astex Pharmaceuticals, 436 Cambridge Science Park, Milton Road, Cambridge CB4 0QA,
U.K.
Page 1 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 47
1
2
3
4
5
6
7
8
9
ABSTRACT: XIAP and cIAP1 are members of the inhibitor of apoptosis protein (IAP)
family and are key regulators of antiꢀapoptotic and proꢀsurvival signaling pathways.
Overexpression of IAPs occurs in various cancers and has been associated with tumor
progression and resistance to treatment. Structureꢀbased drug design (SBDD) guided by
structural information from Xꢀray crystallography, computational studies and NMR solution
conformational analysis was successfully applied to a fragmentꢀderived lead resulting in ATꢀ
IAP, a potent, orally bioavailable, dual antagonist of XIAP and cIAP1 and a structurally
novel chemical probe for IAP biology.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 2 of 47
ACS Paragon Plus Environment

Page 3 of 47
Journal of Medicinal Chemistry
1
2
3
4
1. INTRODUCTION
5
6
7
8
Apoptosis, a closely regulated programmed cell death mechanism, is an essential process to
maintain tissue homeostasis. Resistance to apoptotic stimuli is one of the hallmarks of
cancer1, and can be achieved by overexpression of antiꢀapoptotic proteins. Inhibitor of
apoptosis proteins (IAP), such as cellular IAP1 and 2 (cIAP) and Xꢀlinked IAP (XIAP) play a
critical role in the antiꢀapoptotic and proꢀsurvival signaling pathways; XIAP directly inhibits
caspases, whilst cIAPs prevent the formation of proꢀapoptotic signaling complexes. This
leads to suppression of apoptosis through both the extrinsic and intrinsic apoptosis pathways.2
Deregulation of IAPs, through amplification, overexpression or loss of endogenous
antagonists, has been observed in various cancers and can contribute to tumor growth, disease
progression and poor prognosis. Therefore IAPs are attractive therapeutic targets for antiꢀ
cancer drug discovery.3 In addition, IAPs have been shown to play a role in resistance to
treatment; XIAP is upregulated in response to ionizing radiation,4 suggesting a role for IAP
antagonists in combination therapy.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
IAPs are distinguished by their baculovirus IAP repeat (BIR) domains which are mediators of
proteinꢀprotein interactions; family members such as cIAP and XIAP additionally possesses
really interesting new gene (RING) domains with ubiquitin ligase activity.5 The antiꢀ
apoptotic effect of XIAP are mediated by direct binding of its BIR domains to caspases 3, 7
and 9 resulting in caspase inactivation; caspases 3 and 7 bind to the linker between domains
BIR1 and BIR2 whilst caspase 9 binds at the peptide binding groove of the BIR3 domain.6
IAP antagonists such as the endogenous second mitochondria derived activator of caspases
(SMAC) bind to the IAP BIR domains and can disrupt interactions such as those between
caspase 9 and XIAP.7 On binding to cIAP1, SMAC induces a conformational change
resulting in activation of its E3 ligase activity with subsequent rapid autoubiquitination and
proteasomal degradation.8 This combined loss of cIAP1 and release of the XIAPꢀmediated
Page 3 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 47
1
2
3
4
5
6
7
8
9
block on caspases results in a sustained proꢀapoptotic effect via the extrinsic apoptosis
pathway in the presence of TNFꢀα. Tumors with sufficient levels of TNFꢀα in the tumor
microenvironment may, therefore, be particularly sensitive to IAP antagonism.9 In addition,
antagonism of XIAPꢀmediated caspase inhibition is important for promoting apoptosis via the
intrinsic apoptosis pathway, in response to stimulation by agent such as chemotherapeutics
and DNA damaging agents.10 This suggests that dual cIAP1/XIAP antagonists can be used to
promote apoptosis through both the extrinsic and intrinsic pathways.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
IAPs have been therapeutically targeted by antisense oligonucleotides and small molecule
antagonists. XIAP antisense oligonucleotide (AEG35156)2b sensitized cells to
chemotherapeutic agents and TRAIL receptor agonists and showed evidence of clinical
activity.2b The first generation of IAP antagonists is based on the SMAC Nꢀterminal peptide
sequence AVPI. SMACꢀmimetics, such as 1 – 411aꢀd (Figure 1), all contain a terminal alanine
group and have shown promising preclinical activity allowing them to progress into clinical
trials.2a, 2b
Page 4 of 47
ACS Paragon Plus Environment

Page 5 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Examples of SMAC mimetics in clinical trials
We previously reported the application of a fragmentꢀbased approach to the IAP family12
where weakly binding fragment 5 was optimized into early lead compound 6 (Figure 2). This
is the first published example of a nonꢀpeptidic, subꢀmicromolar affinity dual antagonist of
cIAP1 and XIAP, lacking the alanine group present in previously reported SMAC mimetics.
When dosed intraperitoneally, compound 6 exhibited in vivo biomarker modulation and
concomitant tumor growth inhibition in a mouse xenograft model based on a breast cancer
cell line (MDAꢀMBꢀ231). Here we describe further optimization of 6 into 26, an orally
bioavailable and efficacious low nM antagonist of both cIAP1 and XIAP.
Page 5 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 47
1
2
3
4
5
6
O
N
X
N
Y
N
HN
O
N
7
HN
O
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
X = CH, Y = N 6
XIAP IC50 38% @ 5 mM LE < 0.20
cIAP1 IC50 16% @ 5 mM LE < 0.20
XIAP IC50 = 160 nM LE = 0.30
cIAP1 IC50 = 10 nM LE = 0.35
X = N, Y = CH 7
XIAP IC50 = 220 nM LE = 0.29
cIAP1 IC50
= 14 nM LE = 0.35
Figure 2. Structures of previously published piperazine fragment 5 and early lead compound
612 are shown alongside close analogue 7, the starting point of the work presented in this
paper.
2. OPTIMIZATION OF THE AZAINDOLINE SCAFFOLD
Figure 3A shows the Xꢀray crystal structure of previously published compound 6 bound to
XIAPꢀBIR3. The piperazine ring occupies the P1 pocket with the protonated nitrogen
engaged in a bidentate hydrogen bond interaction with the side chain of Glu314 and the
backbone carbonyl of Asp309. The methyl substituent efficiently fills the small lipophilic
subꢀpocket in P1 forming a van der Waals contact with the side chain of Trp310, strongly
contributing to binding affinity for XIAP. The methoxymethyl group lies in the P2 pocket
and contributes 4 fold to binding affinity,12 despite making no obvious direct contact with the
protein. When we first synthesized compound 6, we lacked a good explanation for this
affinity increase but comparison with Xꢀray crystal structures of bound peptidic IAP
antagonists (e.g. as shown in Figure 3B) suggested that variation of the P2 substituent would
provide good scope for further optimization. Compound 6 extends into P3 with a 5ꢀ
azaindoline bicycle connected to the piperazine moiety via an amide linker. The carbonyl of
Page 6 of 47
ACS Paragon Plus Environment

Page 7 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the central amide forms a hydrogen bond with the backbone NH of Thr308. The saturated
ring of the azaindoline stacks against the side chain of Trp323 in the P3 pocket while the
electron deficient aromatic portion of the bicycle contacts a negative patch in the electrostatic
potential surface of the protein formed by the backbone carbonyl of Gly306 and the phenolic
oxygen in the side chain of Tyr 324.12 A benzylic substituent grows from Cꢀ6 of the
azaindoline into P4, partially filling the lipophilic pocket.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. (A) Xꢀray crystal structures of 6 bound to XIAPꢀBIR3 (PDB 5C83). Hydrogen
bonds between ligand and protein are shown as dashed red lines. The Connolly surface of the
protein is colored by electrostatic potential (red = negative, blue = positive). (B) Comparison
of Xꢀray crystal structures of 7 (orange) and the peptide AVPF (light green) bound to XIAPꢀ
BIR3. The Connolly surface of the protein is colored gray.
We started our SAR exploration by synthesizing a small number of analogues of 6 replacing
the methoxy group in the P2 pocket, with the aim of improving interactions in that region of
the protein. It became evident that some of these compounds showed different degrees of
chemical instability. In a 24 hour NMRꢀbased stability assay test compounds were dissolved
Page 7 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 47
1
2
3
4
5
6
7
8
9
in pH 1, 4 and 7 aqueous buffers (at a nominal concentration of 0.5 mM) and the compounds
showed different levels of cleavage of the central amide bond. This instability was
particularly notable when methoxy was replaced by a more nucleophilic substituent such as
an amine. We ascribed this observation to the electronic effect of the aromatic nitrogen (Nꢀ5)
of the indoline scaffold and its ability to stabilize an incipient negative charge on Nꢀ1 of the
cleavage product. Hence the 5ꢀazaindoline moiety is a relatively good leaving group and the
entire structure susceptible to amide bond cleavage. We reasoned that repositioning of the
nitrogen to Nꢀ4 (as in 7, Figure 2) would lessen the degree of anion stabilization and,
gratifyingly, this change allowed us to improve the chemical stability, whilst still maintaining
almost identical levels of affinity against the two target proteins as measured by XIAP and
cIAP1 fluorescence polarization assay (7 XIAP IC50 220 nM, LE 0.29; cIAP1 IC50 14 nM, LE
0.35).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We report herein the results of the lead optimization campaign that started with compound 7
with a particular emphasis on the structure activity relationships (SAR) for two of the subꢀ
pockets within the protein binding site: P2 and P4 (see Figure 3B). Data presented will
primarily focus on the 4ꢀazaindolines, though some 5ꢀazaindolines are shown in order to
illustrate key SAR points and learnings that were subsequently applied to the 4ꢀazaindoline
series.
3. OPTIMIZATION OF THE P2 AND P4 SIDE CHAINS
A number of features suggested that the P2 pocket was a promising region to target. Firstly,
the crystal structure of 7 bound to XIAPꢀBIR3 domain (Figure 3B) shows the ether side chain
on the piperazine ring occupies the P2 pocket of the protein mimicking the isoꢀpropyl group
of the valine amino acid of the AVPF sequence. Previous work on peptidomimetics13
suggests the potency of the ligand can be modulated by altering the size and polarity of the
Page 8 of 47
ACS Paragon Plus Environment

Page 9 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
substituent in this pocket. Secondly, the electron density of the ether side chain is not well
resolved in the crystal structure which suggests it can assume a number of energetically
similar conformations in the crystal structure of 7 and may not be optimal. Finally, we
consistently observed a water molecule which mediates an Hꢀbond interaction between the
tertiary amine of the piperazine ring and the NH of Trp323 (Figure 4A, labeled W1).
Modeling studies suggested that a carefully positioned substituent on the ligand scaffold,
such as the amide shown in Figure 4A, could displace the water molecule and engage in a
direct interaction with the protein resulting in an increase in binding affinity.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 9 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 47
1
2
3
4
Table 1. Initial exploration of structure activity relationships (SAR) in the P2 pocket
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
XIAP LE b
cIAP1 LE b
XIAP IC50
(nM)a
cIAP1 IC50
(nM)a
Compound
X
Y
R
(kcal molꢀ1per
nonꢀH atom)
(kcal molꢀ1per
nonꢀH atom)
7
8
N
N
CH
CH
220
250
0.29
0.27
14
0.35
47% @ 12
>0.33
9
N
CH
N
CH
N
150
49% @ 40
110
0.26
>0.29
0.27
62% @ 12
80% @ 12
64% @ 12
>0.30
>0.31
>0.31
10
11
CH
12
N
CH
44
0.29
82% @ 12
>0.31
aValues were determined by fluorescence polarization assay (see Experimental Section).
b
Values are the geometrical mean of at least two separate measurements. Ligand efficiency
(LE) values calculated according to the Hopkins formula.14
Page 10 of 47
ACS Paragon Plus Environment

Page 11 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. (A) Xꢀray crystal structure of compound 9 (light green) bound to XIAPꢀBIR3
overlaid with modeled morpholine side chain (purple in ball and stick representation). (B)
and (C) Xꢀray crystal structure of XIAPꢀBIR3 bound to compound 10 (B) and 12 (C). (D)
Water clusters (red maps) identified from the analysis of on molecular dynamics simulation
performed on a truncated analog of 7 lacking the P2 substituent (light blue ligand). Hydrogen
bonds are shown as dashed lines and Connolly surface of the protein is colored gray.
PDB Coordinates for Computational Models reported in Figures 4A and 4D are available in
the Supporting Information.
The results of the assay (Table 1), however, indicate that the amide linker in P2 (compounds
8 and 9) is tolerated but did not lead to a significant increase in XIAP binding affinity. The
Page 11 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 47
1
2
3
4
5
6
7
8
9
crystal structure of 9 (Figure 4A) reveals that the morpholine amide side chain does not bind
on the top of Gln319 as predicted by modeling studies. Instead, the morpholine prefers to
stack on the top of the main scaffold above the tertiary amide, with the conserved water W1
still bound. A similar binding mode with folded conformation was observed with the 4ꢀ
methyl pyrazole 10 (see Figure 4B) and we were intrigued to discover whether this bound
conformation represented a low energy state for the ligand. Interestingly, NMR spectroscopy
(see Supporting Information, Figure S1) on 10 dissolved in phosphate buffer indicated that
the Xꢀray pose is consistent with the lowest energy conformation in solution.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In order to investigate the impact of the folded conformations on potency, we designed a
small array of compounds where the P2 substituents were selected so that their electrostatic
potential surfaces (EPS) were complementary to the EPS of the central amide region of the
piperazineꢀindoline section of the ligand (see dashed area in Figure 5A). This approach
allowed us to maximize the intramolecular stacking interaction between the P2 substituent
and the central amide moiety of the same molecule in order to favor collapsed folded
geometries. Figures 5B and 5C show two P2 sidechains (an aromatic and a saturated
heterocycle) which satisfy the ESP complementarity and that were selected for synthesis.
Page 12 of 47
ACS Paragon Plus Environment

Page 13 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Electrostatic potential surface (EPS) of the piperazineꢀazaindoline scaffold (A) and
a selection of P2 side chains which were selected for synthesis (B and C). Negative areas are
colored red and positive areas are colored blue. The key region surrounding the central amide
of A is enclosed by the dashed red line.
A 2ꢀfold increase in activity was observed with 4ꢀfluoro pyrazole 11, and the 2ꢀpyrrolidinone
12 was the most ligand efficient group we identified in this first round of SAR exploration,
with XIAP IC50 44 nM (LE = 0.29). The crystal structure of 12 confirmed that the
pyrrolidinone ring stacked on top of the central linker amide group, resulting in the folded
conformation (see Figure 4C).
Interestingly, the lactam ring gives a 5ꢀfold increase in potency compared to the simple ether
side chain 7 despite no direct involvement in any van der Waals interactions with the protein.
In order to understand this effect, molecular dynamic (MD) simulations were performed on
the complex between XIAPꢀBIR3 and a truncated analog of 7 lacking the P2 substituent. An
analysis of the MD trajectories revealed a wellꢀdefined water molecule (Figure 4D, labeled
Page 13 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 47
1
2
3
4
5
6
7
8
9
W2) on the top of the key hydrogen bond interaction between the linker carbonyl group of the
ligand and the backbone NH of Thr308. We hypothesized that the lactam side chain in
compound 12 displaces W2, shielding the hydrogen bond interaction from bulk water. This
may have a positive impact on the overall potency of the ligand.15,16
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
With lactam 12 in hand, we started exploration of the P4 pocket (see Table 2) and quickly
found that many compounds reached potency levels close to or below the lower limit of the
fluorescence assays, hence we increasingly relied on XIAP and cIAP1 cellꢀbased assays12 for
generation of primary SAR. For XIAP, we used a XIAPꢀcaspase 9 immunoprecipitation (IP)
assay in HEK293 cells overexpressing XIAP and caspase 9, which directly measures the
ability of compounds to disrupt the proteinꢀprotein interaction inside the cell. For cIAP1, we
measured intracellular cIAP1 degradation in MDAꢀMBꢀ231 cells as a surrogate for cIAP1
affinity.17 SAR were also generated in a MDAꢀMBꢀ231 cell proliferation assay (highly
sensitive to cIAP1 inhibition)12 and an IAPꢀinsensitive human colon cancer cell line (HCTꢀ
116) to control for offꢀtarget cytotoxicity. XIAP/cIAP1 selectivity was defined as the ratio
between the EC50 values in the HEK293 XIAPꢀcaspase 9 IP and cIAP1 degradation assays.
The crystal structure of compound 12 bound to XIAPꢀBIR3 (Figure 4C) showed that the
benzyl group is slightly smaller than the pocket suggesting that small substituents at Cꢀ4 or
Cꢀ2 and Cꢀ4 of the phenyl ring could maximize shape complementarity between the ligand
and the protein.
Page 14 of 47
ACS Paragon Plus Environment

Page 15 of 47
Journal of Medicinal Chemistry
1
2
3
4
Table 2. Exploration of SAR in the P4 pocket
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
XIAP cIAP1
MDAꢀMB
231
HCTꢀ116
EC50
Selectivity
XIAP IP/
IP
deg.
EC50
(nM)b
Compound
R
EC50
(nM)a
EC50
(nM)c
(µM)d
cIAP1 deg.
12
13
14
15
16
17
18
ꢀCH2Ph
ꢀCH2ꢀ4FꢀPh
ꢀCH2ꢀ2,4,diFꢀPh
ꢀCF2Ph
120
7.5
2.3
2.4
20
16
100
ꢀ9% @ 10
2% @ 10
1% @ 10
10% @ 10
ꢀ10% @ 10
1% @ 10
ꢀ8% @ 10
73
32
28
20
69
19
17
0.8
0.5
1.0
1.3
140
2000
1800
210
ꢀCF2Me
130
150
32
240
150
24
ꢀCF2Et
ꢀCF2Pr
aValues were determined by immunoprecipitation assay. Intracellular cIAP1 degradation in
b
c
MDAꢀMBꢀ231 cells (see Experimental Section). MDAꢀMBꢀ231 cell proliferation assay
d
(sensitive to cIAP1 inhibition). Human colon cancer cell line (HCTꢀ116) cell proliferation
assay (IAP insensitive).
Page 15 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 47
1
2
3
4
5
6
7
8
Encouragingly, the fluorinated analogues 13 and 14 gave a modest XIAP potency increase in
the cellular XIAP IP assay (respectively EC50 73 nM and 69 nM) compared to 12 (EC50 120
nM) and were also 5ꢀfold more active in the MDAꢀMBꢀ231 cell proliferation assay (EC50 20
nM and 19 nM respectively) compared to 12 (EC50 100 nM).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We have previously described how XIAP affinity can be improved by increasing the electron
deficient character of the aromatic ring of the bicycle, due to close proximity of this moiety to
an electronegative patch on the protein surface.12 Therefore two fluorine atoms were
introduced at the benzylic position (compound 15). Gratifyingly, this change elicited a 7ꢀfold
increase in XIAP potency. However, a lack of increased cIAP1 potency is reflected in no
potency improvement in the MDAꢀMB231 cell proliferation assay. Encouraged by the XIAP
potency increase seen with 15 a series of 1,1ꢀdifluoroalkyl substituents (16ꢀ18) was also
investigated in order to mimic the P4 binding mode of the Ile side chain of the SMAC
tetrapeptide AVPI. Of these, 18 had exceptional potency in the IP assay for its size. However,
all these compounds had lower activity in the MDAꢀMBꢀ231 cell assay compared to the
fluorobenzyl analogues, likely reflecting only moderate affinity for cIAP1.
The contrasting selectivity profiles achieved with larger P4 groups (e.g. fluorobenzyl) versus
smaller P4 groups (e.g. 1,1ꢀdifluoropropyl) can be explained by differences in P4 pocket size
between XIAP and cIAP1. The base of the P4 pocket is formed by the side chain of Leu292
in XIAP, resulting in a smaller pocket than cIAP1 which contains Val292 in this position.
Hence for small P4 groups, cIAP1 P4 is improperly filled (Figure 7) and the compounds tend
toward weaker cIAP1 affinity. For a large P4 group, the act of filling P4 optimally for cIAP1
results in suboptimal contact in XIAP between the azaindoline bicycle and the protein
surface, leading to a tendency for greater cIAP1 selectivity.
Page 16 of 47
ACS Paragon Plus Environment

Page 17 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Whilst we were delighted by the potent XIAP activity obtained with small P4 substituents,
we were concerned with the relative lack of cIAP1 activity. cIAP1 lies upstream on the
apoptotic pathways and there is a risk that insufficient target engagement here would result in
poor caspase activation. Given that XIAP lies downstream of caspase activation, high
potency XIAP antagonism might be expected to be effective only in the presence of a strong
caspase activation signal. We therefore decided to prioritize the 2,4ꢀdifluorobenzyl or the 4ꢀ
fluorobenzyl as P4 substituents and focus our attention to further developing SAR in the P2
pocket. In particular, we wanted to continue with the identification of P2 groups capable of
adopting the folded conformation that shields the key hydrogen bond interaction with
Thr308.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As described above, potential chemical instability was observed with 5ꢀazaindolines bearing
a nucleophilic P2 substituent, suggesting an interaction between frontier orbitals of the amide
carbonyl and P2 substituents. Therefore we investigated orbital overlap between the Lowest
Unoccupied Molecular Orbital (LUMO) of the piperazineꢀindoline core (Figure 6A) and the
Highest Occupied Molecular Orbital (HOMO) of different P2 substituents (e.g. Figures 6B
and 6C). Although the orbitals have contributions from other atoms, the LUMO corresponds
to the π* orbital on the carbonyl of the piperazineꢀazaindoline core, while the HOMO on the
morpholine or fluoro substituted pyrrolidine corresponds to the nitrogen lone pair. Fluoro
substituted pyrrolidine and morpholine rings, for example, showed high degree of
complementarity with the piperazineꢀindoline scaffold and were selected for synthesis. Table
3 shows the activity value of P2 nucleophile containing compounds as measured in the XIAP
IP cell assay, with lactam 19 as comparison.
Page 17 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 47
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. LUMO of the piperazineꢀazaindoline scaffold (A) and HOMO of examples of P2
side chains (B and C) which were selected for synthesis. Red arrows indicate the positive
orbital overlap between the piperazineꢀazaindoline scaffold and the P2 side chains.
The pyrrolidine 20 (Table 3) had lower activity compared to lactam 19. The result could be
explained by considering the ionization state of the P2 amine at physiological pH. The
calculated pKa of this side chain was 9.4 suggesting appreciable protonation and a
consequent disruption of the favorable HOMOꢀLUMO interaction.
Page 18 of 47
ACS Paragon Plus Environment

Page 19 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
Table 3. SAR for tertiary amine P2 substituents
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
XIAP IPa
EC50
Compound
R
(nM)
19
20
67
180
21
24
22
23
17
14
aValues were determined by immunoprecipitation assay.
Basic side chains with reduced basicity were then targeted, such as fluoropyrrolidines (21 and
22) and morpholine 23. As anticipated, all had much improved potency compared to 20, with
Page 19 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 47
1
2
3
4
5
6
7
8
9
23 close to 10 nM in the XIAP cell IP assay. An Xꢀray crystal structure of 23 bound to XIAPꢀ
BIR3 (data not shown) also showed the morpholine ring binding in the folded conformation,
with the nitrogen lone pair close to linker carbonyl of the ligand. Comparison of the ligand
and protein surfaces suggested an opportunity to append methyl groups adjacent to the
morpholine nitrogen. These changes were tested in our 4ꢀazaindoline main series (Table 4) in
order to avoid the potential issue of chemical instability. We decided to prioritize the 4ꢀ
fluorobenzyl substituent in P4 for further investigation over 2,4ꢀdifluorobenzyl based on the
identical level of potency measured for compounds 13 and 14 (Table 2) and on grounds of
reduced molecular weight and lipophilicity. Compound 26 stood out by virtue of subꢀ10 nM
cell based XIAP activity and exceptional potency in the cIAP1 degradation assay. Addition of
the methyl group results in a gain of 2.6ꢀfold in XIAP potency and a significantly larger
increase in cIAP1 potency compared to compound 24. The high cIAP1 potency led to very
potent antiproliferative effect in the MDAꢀMBꢀ231 assay.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Given the high potency of 26 we wanted to know whether the folded conformation seen in
the Xꢀray crystal structure (Figure 7) was representative of the conformation in aqueous
solution as previously shown for 10.
Page 20 of 47
ACS Paragon Plus Environment

Page 21 of 47
Journal of Medicinal Chemistry
1
2
3
4
Table 4. SAR for substituted morpholine P2 substituents
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
cIAP1
Deg.
MDAꢀ
XIAP
IP
Selectivity
XIAP IP/
HCTꢀ116
EC50
MBꢀ231
Compd
R
R1
EC50
EC50
EC50
cIAP1 deg.
(µM)d
(nM)a
(nM)b
(nM)c
24
25
26
13
2.5
2.4
5.2
20
39
10% @ 10
ꢀ10% @ 10
24% @ 3
O
N
37
15.4
15.9
5.1
0.32
4.4
b
aValues were determined by immunoprecipitation assay. Intracellular cIAP1 degradation in
c
MDAꢀMBꢀ231 cells (see Experimental Section). MDAꢀMBꢀ231 cell proliferation assay
d
(sensitive to cIAP1 inhibition). Human colon cancer cell line (HCTꢀ116) cell proliferation
assay (IAP insensitive).
Page 21 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 47
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Xꢀray crystal structures of compound 26 bound to XIAPꢀBIR3 (A and B) and to
cIAP1ꢀBIR3 (C and D). Water molecules are represented as red spheres. Red dashed lines
represent key hydrogen bond interactions between the ligand and the protein residues.
Connolly surface are shown for XIAP (green) and cIAP1 (gray)
A full NMR structural assignment was performed on 26 and the solution conformation was
assessed in pH 7 phosphate buffer at 20 °C using a ROESY experiment (see Figure 8).
Correlations were seen between hydrogens 1 and 3 and between hydrogens 2 and 4, defining
the conformation of the linker relative to the azaindoline moiety. Crucially, strong ROESY
correlation was observed between hydrogens 5 and 6 (Figure 8) confirming the close
proximity of the morpholine to C7 of the azaindoline, consistent with the Xꢀray bound
Page 22 of 47
ACS Paragon Plus Environment

Page 23 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
conformation. On acidification of the NMR sample to pH 4, the key correlation between
hydrogens 5 and 6 disappeared, consistent with protonation of the morpholine nitrogen and
disruption of the folded over conformation. Hence there is a strong degree of structural
preorganization of 26 in solution at neutral pH, favoring a conformation consistent with that
seen in the Xꢀray crystal structure of the coꢀcomplex. This suggests a low conformational
entropic penalty would be incurred on ligand binding to the protein, thus contributing to the
high affinity observed.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8. Chemical structure and solution conformation of 26 as observed by NMR. Red
arrows indicate through space ROESY correlations in the NMR spectrum.
When 26 was tested in the XIAP and cIAP1 cellular assays we observed that the compound
had a more balanced profile compared to IAP inhibitors in the clinic such as 1ꢀ4, which were
between 40 and 200 fold selective for cIAP1 over XIAP (see Supporting Information, Table
S1). This is likely due to the fundamental difference in the binding interactions of the
piperazine moiety in the P1 region compared to those of the alanine terminus of the
peptidomimetics, as has previously been discussed.12
Page 23 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 47
1
2
3
4
5
6
7
8
9
In vivo target engagement of cIAP1 after a single 30 mg/kg dose PO of 26 was demonstrated
in a PK/PD study in Balb/c scid mice bearing MDAꢀMBꢀ231 xenografts (see Figure 9A).
There was a clear reduction in tumor cIAP1 levels measured by Western blotting at 24 hours
(3/3 mice) and 48 hours (2/3 mice) post dose; before returning to vehicleꢀtreated levels 72
hours post dose. Furthermore, effects on markers of apoptosis were demonstrated in the same
study with elevation of both cleaved PARP and cleaved caspaseꢀ3. The mean plasma and
tumor levels of 26 after the single 30 mg/kg dose from this PK/PD study demonstrated a
prolonged exposure in the tumor compared to the plasma (see Figure 9B).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To demonstrate efficacy in a mouse xenograft model, 26 was dosed daily for 24 days at 7.5 or
15 or 30 mg/kg in Balb/c scid mice bearing MDAꢀMBꢀ231 xenografts. Significant doseꢀ
dependent reduction in tumor growth was measured with daily dosing PO of 26 (see Figure
9C), comparable with compound 2, with no significant effect on body weight. Efficacy was
also demonstrated in an A375 melanoma xenograft model (see Supporting Information,
Figure S2).
In vivo target engagement of XIAP by 26 was demonstrated in an engineered HEK293 cell
line xenograft model in which the HEK293 cell line stably overꢀexpresses both FLAGꢀtagged
XIAP and caspaseꢀ9. This cell line, when grown as subꢀcutaneous xenografts in Bald/c nude
mice, allowed us to demonstrate 26 induced disruption of the interaction between XIAP and
caspaseꢀ9 by immunoprecipitation of FLAGꢀtagged XIAP (Figure 9D).
Lanes 1ꢀ5 in the left panel of Figure 9D shows the results of the immunoprecipitation
experiments performed with antiꢀFLAG antibody on tumor lysates from mice dosed with
vehicle control or with 25 mg/kg of compound 26. In the control mice, caspaseꢀ9 is pulled
down in complex with FLAGꢀtagged XIAP in the presence of antiꢀFLAG antibody but the
amount of caspaseꢀ9 pulled down in animals dosed with compound 26 is considerably
reduced compared to the control mice (and also compared to the total pull down control
Page 24 of 47
ACS Paragon Plus Environment

Page 25 of 47
Journal of Medicinal Chemistry
1
2
3
4
experiment shown in the right panel, lanes 6ꢀ10, of Figure 9D) indicating that compound 26
5
6
efficiently disrupts the XIAPꢀcaspaseꢀ9 complex.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 9. (A) Pharmacodynamic (PD) and (B) pharmacokinetic (PK) effect of a single
30mg/kg dose of 26 PO in Balb/c scid mice bearing MDAꢀMBꢀ231 xenografts. (C) In vivo
efficacy in Balb/c scid mice bearing MDAꢀMBꢀ231 xenografts dosed daily PO with 26 at 30
mg/kg (blue line), 15 mg/kg (red line) and 7.5 mg/kg (green line) or dosed PO with 2 at
30mg/kg (orange line). (D) Immunoprecipitation (using antiꢀFLAG antibody) of FLAGꢀ
tagged XIAP from tumor lysates taken 2 h after a single 25 mg/kg dose of 26 to HEK293
xenografted Balb/c nude mice (lanes 1ꢀ5) and total pull down control experiments (without
antiꢀFLAG antibody) (lanes 6ꢀ10).
Pharmacokinetic parameters of compound 26 in mouse are shown in Table 5. Moderate
clearance in vivo (>30% liver blood flow), volume of distribution (Vss), and 3 h half life were
measured when compound 26 was dosed IV at 1 mg/kg in mice. Following single oral dosing
Page 25 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 47
1
2
3
4
5
6
in mice at 5 mg/kg, a rapid rate of absorption was observed and oral bioavailability (F) was
22%.
7
8
9
Table 5. Mouse pharmacokinetics of 26 dose IV at 1 mg/kg, and PO at 5 and 30 mg/kg
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Dose
Tmax
Cmax
CLp
Vss
Half life
AUC
F
Route
(mg/kg)
(h) (µg/mL) (mL/min/kg) (L/kg)
40.9 6.4
(h)
(µg*h/mL) (%)
1
IV
PO
PO
3.0
0.41
5
2
1
0.12
1.7
1.8
0.45
12.1
22
30
4. CHEMISTRY
For compounds 7ꢀ26 we required carboxylic acid functionalized piperazine (Scheme 1) and
azaindoline (Scheme 2) building blocks. Functionalized piperazines 28ꢀ29 were prepared
from intermediates (27a and 27b) described in a previous publication.12 Hydroxymethyl
piperazine 27b was readily oxidized to the corresponding carboxylic acid intermediate
followed by amide coupling with dimethylamine or morpholine to generate compounds 28a
and 28b. Carboxylic acid functionalized piperazines 30aꢀc, were prepared by reaction of
intermediate 27a with methanesulfonyl chloride, subsequent reaction with a variety of amine
or amide nucleophiles to give 29aꢀe, then reaction with benzyl bromoacetate and final
hydrogenolysis in the presence of palladium on carbon. The fluoroꢀpyrazole and methylꢀ
pyrazole compounds 31aꢀb were prepared via displacement of the mesylate formed in situ
from the protected hydroxymethyl piperazine and further elaboration as for 29aꢀc.
Page 26 of 47
ACS Paragon Plus Environment

Page 27 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Reagents and conditions: a) NaIO4, RuCl3, water/dimethyl carbonate/MeCN,
RT, 78%; b) Carbonyldiimidazole, then dimethylamine or morpholine, DCM, RT,
quant.; c) Pd/C, EtOH, H2, RT, 71% ꢀ quant.; d) NaBH(OAc)3, benzaldehyde,
dichloroethane, RT, 76%; e) methanesulfonyl chloride, triethylamine, DCM, RT, 77%; f)
NaH, R3H, DMF, 80°C; g) R3H, KI, K2CO3, MeCN, 70°C, 59% ꢀ quant.; h) Pd/C, H2,
Page 27 of 47
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 47
1
2
3
4
5
6
7
8
AcOH/EtOH, RT, 73ꢀ95%; i) benzyl bromoacetate, K2CO3, MeCN, RT, 56ꢀ98%; j)
Pd/C, H2, EtOAc, RT, 80ꢀ94%; k) methanesulfonic anhydride, triethylamine, 4ꢀfluoroꢀ
pyrazole or 4ꢀmethylꢀpyrazole, 0ꢀ20°C, 30ꢀ59%.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme
2.
Reagents
and
conditions:
a)
ArZnBr,
[1,3ꢀbis(2,6ꢀ
diisopropylphenyl)imidazolꢀ2ꢀylidene](3ꢀchloropyridyl)palladium(II) dichloride, LiBr,
NMP/THF, RT, 70% ꢀ quant.; b) 4M HCl in dioxane, MeOH, RT, 78% ꢀ quant.; c)
nBuLi, RCON(Me)OMe, ꢀ78 °C, 45ꢀ61%; d) Bis(2ꢀmethoxyethyl)aminosulfur
trifluoride, THF, 70ꢀ90 °C, 34ꢀ47%.
All 4ꢀazaindoline intermediates 32aꢀc and 33aꢀd were prepared from a common starting
material, Boc protected 3,3ꢀdimethylꢀ6ꢀbromoꢀ4ꢀazaindoline18 as shown in Scheme 2.
Differently substituted benzylic groups were introduced at the Cꢀ6 position of the indoline via
Negishi coupling19 with the appropriate benzylzinc bromide in the presence of [1,3ꢀbis(2,6ꢀ
diisopropylphenyl)imidazolꢀ2ꢀylidene](3ꢀchloropyridyl)palladium(II) dichloride catalyst. The
Page 28 of 47
ACS Paragon Plus Environment

Page 29 of 47
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
same procedure was also used to synthesize the 5ꢀazaindoline intermediates 34aꢀb starting
from Boc protected 3,3ꢀdimethylꢀ6ꢀchloroꢀ5ꢀazaindoline.12,18 Gemꢀdifluoro substituted
compounds 33aꢀd were prepared by fluorination of the corresponding ketones with bis(2ꢀ
methoxyethyl)aminosulfur trifluoride and subsequent deprotection. The required 6ꢀketo
intermediates were synthesised in good yields via lithiation at Cꢀ6 using nBuLi followed by
quench with the appropriate Weinreb amide. Three different routes were employed to
assemble compounds 7ꢀ26 as shown in Schemes 3, 4 and 5. In the first method (Scheme 3)
the carboxylic acid functionalised piperazines were directly coupled to the appropriate
indoline in a HATU or PyBroP mediated amide coupling reaction. Compound 7 required late
stage Negishi coupling with benzylzinc bromide to elaborate Cꢀ6 of the indoline scaffold.
Scheme 4 shows a second approach: the “three component coupling”. In this twoꢀstepꢀoneꢀ
pot reaction the indoline was firstly reacted with chloroacetylchloride to establish the amide
bond followed by the addition of the appropriately decorated piperazine to produce the Bocꢀ
protected final compounds in moderate yields. Finally, Scheme 5 describes the third synthetic
strategy. Using methods similar to those described previously, chloromethyl intermediates
35aꢀb were prepared allowing introduction of the P2 group at a late stage. All final
compounds were isolated as hydrochloride salts following deprotection under acidic
conditions.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Page 29 of 47
ACS Paragon Plus Environment