To identify new chemical series with enhanced binding affinity to the BTB domain of B-cell lymphoma 6 protein, we targeted a subpocket adjacent to Val18. With no opportunities for strong polar interactions, we focused on attaining close shape complementarity by ring fusion onto our quinolinone lead series. Following exploration of different sized rings, we identified a conformationally restricted core which optimally filled the available space, leading to potent BCL6 inhibitors. Through X-ray structure-guided design, combined with efficient synthetic chemistry to make the resulting novel core structures, a >300-fold improvement in activity was obtained by the addition of seven heavy atoms.
To identify new chemical series with enhanced binding affinity to the BTB domain of B-cell lymphoma 6 protein, we targeted a subpocket adjacent to Val18. With no opportunities for strong polar interactions, we focused on attaining close shape complementarity by ring fusion onto our quinolinone lead series. Following exploration of different sized rings, we identified a conformationally restricted core which optimally filled the available space, leading to potent BCL6 inhibitors. Through X-ray structure-guided design, combined with efficient synthetic chemistry to make the resulting novel core structures, a >300-fold improvement in activity was obtained by the addition of seven heavy atoms.
BCL6 (B-cell lymphoma
6 protein) is a master regulator of the germinal
center (GC) B-cell phenotype and is required for GC formation.[1−3] By binding to and repressing target genes, BCL6 attenuates cell
differentiation and the DNA-damage response, thereby facilitating
the process of somatic hypermutation required for antibody maturation.[4,5] This phenotype broadly resembles that of malignant cells, so perhaps
unsurprisingly, most B-cell lymphomas, including diffuse large B-cell
lymphoma (DLBCL), derive from GC B-cells. Many of these tumors remain
dependent on BCL6 expression for their survival.[6]Transcriptional repression by BCL6 requires binding
of corepressors
including BCOR and NCOR to the BTB domain of BCL6.[7] Disruption of this protein–protein interaction hence
relieves BCL6-mediated gene repression. Both inhibitors and degraders
of BCL6 have been shown to cause selective growth inhibition in BCL6-driven
lymphoma cell lines,[8,9] but to date, the therapeutic potential
of inhibition or degradation of BCL6 in vivo in lymphomas has not
been thoroughly tested. Improved tool compounds with suitable pharmacokinetic
properties along with strong binding affinity are required to support
this objective.Previously, both we and others have described
compounds which bind
to the BCL6 BTB domain and displace corepressors.[8−14] Although the corepressors bind along a long, extended groove formed
by dimerization of the BTB domain,[5,7] reported hit
compounds all bind in the same small area and form common interactions
in the binding site: aromatic/hydrophobic contacts with Tyr58 and
hydrogen bonding to the backbone C=O of Met51 and the backbone
NH of Glu115. From these starting points, various approaches have
been used to reach the high binding affinity required, including conformational
restriction via macrocyclization[13] and
adding a further backbone H-bond interaction.[9] In this study, we focus on further filling a largely hydrophobic
subpocket defined by residues including His14, Asp17, Val18, and Cys53
(the “HDCH site”[15]) by fusing
an additional ring onto a previously identified series of quinolinone
inhibitors.[16] Our design strategy focused
on filling space and obtaining a close shape complementary to the
protein surface. Progress was enabled by the development of short,
efficient synthetic routes to access a variety of novel core structures
designed using X-ray structural information. Our resulting lead compound 1 (CCT372064) shows a >300-fold improvement in activity
compared
with starting point 2 (Figure ), gained entirely through improved binding
in the HDCH site without modification to the 3-chloro-4-cyanopyridine
substituent. In our subsequent paper, we report how the novel tricyclic
core of 1 provided the basis for the design of a potent
degrader of BCL6 suitable for testing the therapeutic hypothesis in
vivo.
Figure 1
Potency improvements gained in the HDCH site through growing and
cyclization strategies. The methyl or cyclopropyl group shown in bold
occupies a hydrophobic region close to Val18.
Potency improvements gained in the HDCH site through growing and
cyclization strategies. The methyl or cyclopropyl group shown in bold
occupies a hydrophobic region close to Val18.
Results
and Discussion
We previously reported the identification
of the quinolinone hit
compound 2, in which the terminal methyl of the N-ethyl group occupies a hydrophobic pocket close to Val18
in the HDCH site. Maintaining this methyl group in place, we further
optimized affinity by growing the molecule into this broad binding
pocket, displacing or interacting with bound water molecules.[16] Although our resulting lead compound 3 (CCT369347) showed good potency of 26 nM in the biochemical time-resolved
fluorescence energy transfer (TR-FRET) assay (Figure ), we saw a substantial drop-off in the cellular
nanoBRET assay (IC50 570 nM), despite high passive permeability.[16] Our aim was therefore to discover a chemical
scaffold with improved binding affinity for the BTB domain of BCL6,
which we expected in turn would lead to more potent activity in cellular
assays.We hence inspected the X-ray structure of 2 (Figure A) for opportunities
to design in additional interactions. We noticed that adding a substituent
in the 3-position of the quinolinone ring may enable access to an
unfilled region of the pocket. However, introducing such a substituent
would be expected to cause a change in conformation for the 4-amino
substituent to avoid steric clash. This was confirmed by synthesis
of 4 (Table ) and subsequent solving of its BCL6-bound crystal structure.
This structure showed that the 3-methyl group induces rotation of
the N-ethyl group (Figure B). As this alternate conformation cannot
be accommodated by the protein, the knock-on effect is a “flip”
of the whole quinolinone ring system which points the 4-amino group
out to the solvent, resulting in a 4-fold reduction in biochemical
potency.
Figure 2
(A) X-ray structure of the BCL6 BTB domain with bound ligand 2 (PDB: 7OKH, cyan) showing the shape of the pocket (protein surface shown in
gray). The unfilled pocket space potentially accessible from the quinolinone
3-position is highlighted in orange. (B) X-ray structure of the BCL6
BTB domain with bound ligand 4 (PDB: 7Q7S, magenta). The addition
of a 3-quinolinone methyl group causes the 4-ethylamino group to twist,
becoming orthogonal to the quinolinone ring. As a result, the quinolinone
ring is forced to adopt a different binding mode, pointing the 4-substituent
out to the solvent.
Table 1
Biochemical
and Cellular Data for
4-Aminoquinolinones and Fused-Tricyclic Derivativesa[18,19]
Individual
replicates and errors
are shown in Table S1. NT = not tested.
Quoted values represent the
geometric
mean of at least three replicates. n = 2 where indicated
by *.
Quoted values represent
the geometric
mean of two replicates. n = 1 where indicated by
*.
Kinetic solubility measured
by HPLC
in 10 mM phosphate-buffered saline buffer at pH 7.4 containing 1%
DMSO. Where + shown, solubility was measured by NMR in HEPES buffer
at pH 8, containing 4% DMSO.
(A) X-ray structure of the BCL6 BTB domain with bound ligand 2 (PDB: 7OKH, cyan) showing the shape of the pocket (protein surface shown in
gray). The unfilled pocket space potentially accessible from the quinolinone
3-position is highlighted in orange. (B) X-ray structure of the BCL6
BTB domain with bound ligand 4 (PDB: 7Q7S, magenta). The addition
of a 3-quinolinone methyl group causes the 4-ethylamino group to twist,
becoming orthogonal to the quinolinone ring. As a result, the quinolinone
ring is forced to adopt a different binding mode, pointing the 4-substituent
out to the solvent.Individual
replicates and errors
are shown in Table S1. NT = not tested.Quoted values represent the
geometric
mean of at least three replicates. n = 2 where indicated
by *.Quoted values represent
the geometric
mean of two replicates. n = 1 where indicated by
*.Kinetic solubility measured
by HPLC
in 10 mM phosphate-buffered saline buffer at pH 7.4 containing 1%
DMSO. Where + shown, solubility was measured by NMR in HEPES buffer
at pH 8, containing 4% DMSO.To overcome this problem, we constrained 3- and 4-quinolinone substitution
into a fused ring. We targeted the synthesis of a set of six- and
seven-membered fused quinolinones to control the conformation of the
quinolinone 3- and 4-substituents, aiming to more fully occupy the
space in this region of the pocket (Table ). Each of these novel cores required development
of new synthetic routes, as described in more detail in the “synthesis”
section below.The IC50 values for these compounds
are reported in Table . Compound 5a shows no improvement over 4, and we thus speculate
that this may be adopting the same “flipped” binding
mode. A methyl group was added to mimic the terminal methyl of 2 and hence maintain the interaction with the Val18 hydrophobic
pocket. The resulting compound 5b and ether analogue 6 showed >3-fold potency improvement over 5a,
but their activity was still no better than acyclic compound 2. In contrast, our first seven-membered ring derivative 7 showed a >10-fold potency increase over 2 despite
a reduction in lipophilicity. The large increase in lipophilic ligand
efficiency (LLE) from 3.5 to 5.9 suggested that new interactions were
likely being formed.[17]We solved
an X-ray structure of 7 bound to BCL6 (Figure A), which showed
the same key interactions seen previously in the 4-aminoquinolinone
series;[16] the cyano-chloro-pyridine ring
is clamped between Tyr58 and Asn21, forming a π–π
interaction with Tyr58; there are H-bonding interactions to the backbone
amides of Met51, Ala52, and Glu115; and the methyl group of the seven-membered
ring sits in the lipophilic region of the pocket near Val18. In addition,
a new interaction was observed between the lactone carbonyl and a
conserved water molecule deep in the HDCH pocket. This water forms
part of a network with two other water molecules, which in turn contact
the main chain carbonyl of Cys53 and the His14 side chain.
Figure 3
X-ray structures
of the BCL6 BTB domain with bound ligands 7 [(A,C); yellow;
PDB: 7Q7T] and 9a [(B,D,E); green;
PDB: 7Q7U].
The interaction surface [line, calculated using MOE (Chemical Computing
Group)] is a zero-potential contour of the van der Waals potential
of a probe atom—atoms sitting on the surface are at optimal
distance for vdW interactions with the protein. Hydrogen atoms on
the seven-membered ring of 9a sit on this surface, showing
that the pocket is well filled in this area (D), whereas the lactone
ring of 7 (C) does not map closely to the surface, although
it does form an additional strong H-bond to a water molecule (A).
The carbonyl of Cys53 sits close to the ring carbon alpha to the ether
oxygen of 9a (E). Replacing this carbon with an oxygen
(as in 8) may be disfavored due to clashing lone pairs,
preventing close surface contact.
X-ray structures
of the BCL6 BTB domain with bound ligands 7 [(A,C); yellow;
PDB: 7Q7T] and 9a [(B,D,E); green;
PDB: 7Q7U].
The interaction surface [line, calculated using MOE (Chemical Computing
Group)] is a zero-potential contour of the van der Waals potential
of a probe atom—atoms sitting on the surface are at optimal
distance for vdW interactions with the protein. Hydrogen atoms on
the seven-membered ring of 9a sit on this surface, showing
that the pocket is well filled in this area (D), whereas the lactone
ring of 7 (C) does not map closely to the surface, although
it does form an additional strong H-bond to a water molecule (A).
The carbonyl of Cys53 sits close to the ring carbon alpha to the ether
oxygen of 9a (E). Replacing this carbon with an oxygen
(as in 8) may be disfavored due to clashing lone pairs,
preventing close surface contact.Despite the high biochemical activity, no activity was observed
in cells using the NanoBRET assay. We believed that this was due to
the very low permeability (PAMPA Pe <0.2
× 10–6 cm s–1 at pH 7.4),
driven in turn by the low hydrophobicity of 7 [clogD7.4 1.1, topological polar surface area (TPSA)
109 Å2]. We decided to investigate the removal of
the carbonyl in the seven-membered ring and prepared compound 8 (clogD7.4 1.7, TPSA 92 Å2). Disappointingly, this change also removed most of the activity
of the compound (Table ). To understand the reason for this loss of activity, we reinspected
the crystal structure of 7 and noted that the ring appeared
to sit across the top of the pocket rather than align closely to the
surface (Figure C).
We hypothesized that this could be due to the backbone carbonyl of
Cys53 at the back of the pocket, which could repel the ring oxygen
of both lactone 7 and ether 8. Moreover,
we reasoned that moving the oxygen by one position—as in compound 9a—would remove that repulsion and allow the seven-membered
ring to form closer contacts with the protein surface, resulting in
improved activity. Gratifyingly, 9a showed a significantly
increased TR-FRET activity of ∼100 nM, while having a much
lower TPSA (92 Å2) and increased lipophilicity (Mo
Kα log D7.4 2.5) compared to 7. This profile translated into a measurable, albeit modest,
cellular activity (2 μM) in NanoBRET for 9a.X-ray crystallography confirmed our hypothesis that ether tricycle 9a adopts a significantly different conformation to lactone
tricycle 7 (Figure ). While both positioned the terminal methyl group
in the same location near Val18, the ether tricycle more closely follows
the surface of the pocket, with the methylene group alpha to oxygen
sitting 3.2 Å away from the carbonyl of Cys53 (Figure E). We hypothesized that the
resulting displacement of an additional water molecule by 9a and the hydrophobic contacts formed were able to compensate for
the loss of the polar water-mediated interaction made by the carbonyl
of 7. Switching to methyl epimer 9b or des-methyl 10 led to weaker activity in the TR-FRET assay, again reinforcing
the importance of positioning the terminal methyl group in the Val18
pocket.Having identified a scaffold that effectively shape-matches
the
available space, we looked for opportunities to further optimize the
potency. From the crystal structure of 9a, it appeared
that there was additional space around the methyl group close to Val18
that could be exploited (Figure A). This space appeared to be “triangular”
in shape, and we speculated that a small alkyl or cycloalkyl group
might fit into this pocket; we prepared a range of such groups to
test this hypothesis. Racemic analogues were prepared initially due
to building block availability, and the results are shown in Table .
Figure 4
X-ray structures of the
BCL6 BTB domain with bound ligands 9a [(A); green; PDB: 7Q7U] and 12a [(B); cyan; PDB: 7Q7V]. Molecular surface
shown in gray. The interaction surface around the methyl group (orange
grid, see Figure legend
for details) shows a triangular shape (A), which the cyclopropyl group
of 12a fills effectively (B).
Table 2
Biochemical and Cellular Data for
Fused-7-Ring Tricyclic Quinolinones with Differing Groups in the Pocket
Adjacent to Val18a
Individual replicates
and errors
are shown in Table S1. NT = not tested.
Quoted values represent the
geometric
mean of at least three replicates.
Quoted values represent the geometric
mean of at least two replicates.
Kinetic solubility measured by NMR
in HEPES buffer at pH 8, containing 4% DMSO.
X-ray structures of the
BCL6 BTB domain with bound ligands 9a [(A); green; PDB: 7Q7U] and 12a [(B); cyan; PDB: 7Q7V]. Molecular surface
shown in gray. The interaction surface around the methyl group (orange
grid, see Figure legend
for details) shows a triangular shape (A), which the cyclopropyl group
of 12a fills effectively (B).Individual replicates
and errors
are shown in Table S1. NT = not tested.Quoted values represent the
geometric
mean of at least three replicates.Quoted values represent the geometric
mean of at least two replicates.Kinetic solubility measured by NMR
in HEPES buffer at pH 8, containing 4% DMSO.Consistent with the pocket shape, cyclopropyl 11a provided
the optimal solution, with both smaller (ethyl, 11b)
and larger (cyclobutyl 11c and isopropyl 11d) substituents showing 8- to 55-fold reduction in activity in comparison.
This highlights the importance of matching the geometric requirements
of the pocket. Cyclobutyl and isopropyl substituents were not well
tolerated as they do not fit into the triangular pocket—the
larger bond angles (90–109.5°) between the methyl or methylene
groups would lead to a clash with the protein. Reducing the size of
this substituent by replacing each of the methyl groups of the isopropyl
group with fluorine (11e) did not significantly recover
this lost activity.Single enantiomers of the cyclopropyl derivative
were prepared,
and as expected, the preferred enantiomer was (R)-cyclopropyl 12a, in which the cyclopropyl points in the same direction
as the (S)-methyl in 9a. Although the
potency gain from methyl to cyclopropyl is modest (∼3-fold),
this improvement combined with a lipophilicity-driven increase in
permeability led to a 10-fold increase in cellular assay activity
for 12a compared to 9a.The crystal
structure of 12a was solved and showed
the same binding conformation as 9a (Figures , 6). This supports our finding that the cyclopropyl group is the optimal
solution in this region; it efficiently fills the additional “triangular”
space, closely mapping to the surface.
Figure 6
X-ray
structures of the BCL6 BTB domain with bound ligand 1 [(D); orange] compared to 2 [(A); yellow], 9a [(B); green], and 12a [(C), cyan]. The protein
surface is represented as a gray grid, and the ligand surface is shown
in colors as above. The space in the pocket is more completely filled
as potency is optimized from (A) through to (D).
Having identified the
cyclopropyl ring as the optimal substituent
adjacent to NH, we explored other positions of the seven-membered
ring where the crystal structures showed potential for increasing
the hydrophobic surface contact area (Table ). First, methyl groups were incorporated
alpha to the ether oxygen of 9a, aiming to fill a small
hydrophobic area. However, no potency improvement was observed for
monomethyl 13a, and dimethyl analogue 13b showed a dramatic drop-off in potency, indicating that the space
in this area was limited. Dimethyl or cyclopropyl beta to the oxygen (13c and 13d, respectively)
also reduced TR-FRET activity to micromolar levels, suggesting that
these groups were too large. We therefore targeted difluoro substitution
as a more subtle change. Adding difluoro substitution to 12a led to a further potency breakthrough, with 1 (CCT372064)
showing 4 nM activity in TR-FRET. This represents a 10-fold improvement
for the addition of 2 fluorine atoms. This is a larger increase than
would be expected from lipophilicity alone, as demonstrated by the
increase in LLE from 4.5 (for 12a) to 5.4 (for 1). This can be explained by the improved shape matching between
the ligand and the protein observed in the X-ray structure of 1 (Figures , 6). The electronegative
fluorine atoms map closely to the protein surface, pointing toward
a C–H bond in the Cys53 side chain and the aryl C–H
of Phe89. In addition, the short distance (3.1 Å) between the
backbone carbonyl of Cys53 and the ring CH2 suggests a
possible Sutor bond,[20−22] favored by the electron-withdrawing nature of the
adjacent CF2 group.
Table 3
Biochemical and Cellular Data for
Fused-7-ring Tricyclic Quinolinones with Additional Pocket-Facing
Substitutiona
Individual replicates
and errors
are shown in Table S1. NT = not tested.
Quoted values represent the
geometric
mean of at least three replicates. † indicates n = 2.
Quoted
values represent the geometric
mean of two replicates. n = 1 where indicated by
*.
Kinetic solubility measured
by NMR
in HEPES buffer at pH 8, containing 4% DMSO.
Figure 5
X-ray structure of the BCL6 BTB domain
with bound ligand 1 [(A); orange; PDB 7Q7R], highlighting key
distances between the protein and
the ligand. Possible interactions observed include H-bonds with Met51
(backbone C=O), Glu155 (backbone NH), and Ala52 (backbone C=O);
C–H···O interactions (Sutor bonds) with Cys53
(backbone C=O); vdW or dipole–dipole contacts between
“δ−” fluorine atoms and “δ+”
hydrogen atoms on Phe89 and Cys53; and vdW contacts with Val18 and
π–π interactions between the electron-deficient
pyridine ring and the electron-rich Tyr58 side chain. Close-up views
of the CF2 group of 1 (C) show improved surface
contact compared with the analogous CH2 group of 12a (B). The cyclopropyl group has been truncated to methyl
in (B,C) to enable a clear view of this region.
X-ray structure of the BCL6 BTB domain
with bound ligand 1 [(A); orange; PDB 7Q7R], highlighting key
distances between the protein and
the ligand. Possible interactions observed include H-bonds with Met51
(backbone C=O), Glu155 (backbone NH), and Ala52 (backbone C=O);
C–H···O interactions (Sutor bonds) with Cys53
(backbone C=O); vdW or dipole–dipole contacts between
“δ−” fluorine atoms and “δ+”
hydrogen atoms on Phe89 and Cys53; and vdW contacts with Val18 and
π–π interactions between the electron-deficient
pyridine ring and the electron-rich Tyr58 side chain. Close-up views
of the CF2 group of 1 (C) show improved surface
contact compared with the analogous CH2 group of 12a (B). The cyclopropyl group has been truncated to methyl
in (B,C) to enable a clear view of this region.X-ray
structures of the BCL6 BTB domain with bound ligand 1 [(D); orange] compared to 2 [(A); yellow], 9a [(B); green], and 12a [(C), cyan]. The protein
surface is represented as a gray grid, and the ligand surface is shown
in colors as above. The space in the pocket is more completely filled
as potency is optimized from (A) through to (D).Individual replicates
and errors
are shown in Table S1. NT = not tested.Quoted values represent the
geometric
mean of at least three replicates. † indicates n = 2.Quoted
values represent the geometric
mean of two replicates. n = 1 where indicated by
*.Kinetic solubility measured
by NMR
in HEPES buffer at pH 8, containing 4% DMSO.Pleasingly, 1 showed potent cellular
activity, with
an IC50 value of 12 nM in the NanoBRET assay. This compound
therefore shows not only a >300-fold improvement in biochemical
potency
over the ethylamine derivative 2 but also a 50-fold improvement
in cellular potency (NanoBRET) over our previous best BCL6 inhibitor 3 (Figure ). This was achieved through effective shape matching and space filling,
without adding any new strong polar interactions (Figure ). With this area of the pocket
now filled, we had achieved our goal of discovering a potent, ligand
efficient core. The optimization of this core into a potent degrader
suitable for sustained depletion of BCL6 in vivo is described in a
subsequent paper.
Chemistry
To enable our strategy
of cyclization, we needed to rapidly develop
and execute a range of varied synthetic routes to access largely unprecedented
cores. We therefore targeted synthesis of novel nitro-substituted
tricyclic cores 15–19, which could be converted
to final compounds by standard reduction and SNAr reactions,
and aimed to make these from building blocks we had previously prepared
(Scheme ).[16]
Scheme 1
Synthetic Approach to Tricyclic Quinolinones
Previously reported synthetic routes to “piperidine-fused”
quinolinones (2,3,4,6-tetrahydrobenzo[h][1,6]naphthyridin-5(1H)-ones, Table ) were lengthy,[23,24] and our desired substitution
pattern was not precedented. However, while exploring azetidine substituents,
a surprising solution presented itself: we observed a ring expansion
of the azetidine of 20e under the high-temperature SNAr reaction conditions forming the desired fused compound 5a (Scheme ). We speculated that azetidine ring opening may be taking place,
possibly by the chloride ion generated during the SNAr.
Intramolecular attack from the presumably nucleophilic quinolinone
3-position would then form the observed product. To investigate this
hypothesis, we replaced the azetidine with alkyl tosylate 21 and were pleased to observe the desired cyclization to form target
compound 5b.
Scheme 2
Synthesis of 2,3,4,6-tetrahydrobenzo[h][1,6]naphthyridin-5(1H)-ones
(a) Azetidine, NMP, 160 °C
μW, 1 h, 56%; (b) SnCl2, EtOH, 120 °C μW,
1 h, 97%; (c) 2,4-dichloropyridine-3-carbonitrile, TEA, NMP, 160 °C
μW, 1 h, 8–51%; (d) NaH, DMF, 0 °C, 10 m, then MeI,
rt, 1.25 h, 81%; (e) 3-aminobutan-1-ol, DIPEA, NMP, 160 °C, 20
h, 66%; (f) TsCl, py/DCM, 0 °C to rt, 20 h, 44%; (g) DIPEA, NMP,
160 °C μW, 1 h, 67%; (h) Pd/C, ammonium formate, EtOH/NMP,
60 °C, 0.5 h, 100%.
Synthesis of 2,3,4,6-tetrahydrobenzo[h][1,6]naphthyridin-5(1H)-ones
(a) Azetidine, NMP, 160 °C
μW, 1 h, 56%; (b) SnCl2, EtOH, 120 °C μW,
1 h, 97%; (c) 2,4-dichloropyridine-3-carbonitrile, TEA, NMP, 160 °C
μW, 1 h, 8–51%; (d) NaH, DMF, 0 °C, 10 m, then MeI,
rt, 1.25 h, 81%; (e) 3-aminobutan-1-ol, DIPEA, NMP, 160 °C, 20
h, 66%; (f) TsCl, py/DCM, 0 °C to rt, 20 h, 44%; (g) DIPEA, NMP,
160 °C μW, 1 h, 67%; (h) Pd/C, ammonium formate, EtOH/NMP,
60 °C, 0.5 h, 100%.The “fused
morpholine” (2,3-dihydro-1H-[1,4]oxazino[2,3-c]quinolin-5(6H)-one, Scheme ) core
had not previously been reported in the literature. To enable rapid
synthesis, we aimed to use existing 4-aminoquinolinone intermediates
and therefore targeted making the key C–O bond via the Ullmann
ether synthesis. Iodination of 22a occurred selectively
on the quinolinone 3-position, and copper-catalyzed C–O bond
formation provided access to the desired target.
Scheme 3
Synthesis of the
2,3-Dihydro-1H-[1,4]oxazino[2,3-c]quinolin-5(6H)-one Core
(a)
2-Aminopropan-1-ol, DIPEA,
NMP, 160 °C, 24 h, 87%; (b) iodine, methanol/water, 60 °C,
2 h, 38%; (c) 1,10-phen, CuI, cesium carbonate, NMP, 120 °C,
1 h, 41%; (d) Pd/C, ammonium formate, methanol, 80 °C, 1.5 h,
81–100%; (e) 2,4-dichloropyridine-3-carbonitrile, TEA, NMP,
160 °C, μW, 1.5 h, 60%.
Synthesis of the
2,3-Dihydro-1H-[1,4]oxazino[2,3-c]quinolin-5(6H)-one Core
(a)
2-Aminopropan-1-ol, DIPEA,
NMP, 160 °C, 24 h, 87%; (b) iodine, methanol/water, 60 °C,
2 h, 38%; (c) 1,10-phen, CuI, cesium carbonate, NMP, 120 °C,
1 h, 41%; (d) Pd/C, ammonium formate, methanol, 80 °C, 1.5 h,
81–100%; (e) 2,4-dichloropyridine-3-carbonitrile, TEA, NMP,
160 °C, μW, 1.5 h, 60%.As the
potency of 6 showed no improvement over the
acyclic compounds, we moved on to explore seven-membered rings, aiming
to more completely fill the available space in the pocket. Although
our targeted cyclic lactone structures (2,3-dihydro-[1,4]oxazepino[6,5-c]quinoline-5,6(1H,7H)-diones, Scheme ) were previously
unreported, we had observed the formation of 17 as a
byproduct during SNAr reactions of ester 20b with aminoalcohols.[16] Following nitro
reduction to 24a, the resulting intermediate could be
used to synthesise 7. Access to a further target, 8—containing the 2,3,5,7-tetrahydro-[1,4]oxazepino[6,5-c]quinolin-6(1H)-one core—required
synthesis of ether 24b (Scheme ). This was prepared by reduction of the
cyclic lactone group of 24a using sodium borohydride
in the presence of a Lewis acid.
Scheme 4
Synthesis of 2,3-Dihydro-[1,4]oxazepino[6,5-c]quinoline-5,6(1H,7H)-dione
and 2,3,5,7-Tetrahydro-[1,4]oxazepino[6,5-c]quinolin-6(1H)-one Cores
(a) Diethyl malonate, NaH, 0
°C–rt, 3 h, 94%; (b) POCl3, 80 °C, 2.5
h, 57%; (c) (S)-2-aminopropan-1-ol, DIPEA, NMP, 160
°C, 1 h, then LiCl 160 °C, 1 h, 37–70%; (d) H2, Pd/C, ethanol, rt, 16 h, 92%; (e) 2,4-dichloropyridine-3-carbonitrile,
DIPEA, NMP, 160 °C, 1 h, 10–19%; (f) BF3·OEt2, THF, NaBH4, 0 °C, 2 h, 32%.
Synthesis of 2,3-Dihydro-[1,4]oxazepino[6,5-c]quinoline-5,6(1H,7H)-dione
and 2,3,5,7-Tetrahydro-[1,4]oxazepino[6,5-c]quinolin-6(1H)-one Cores
(a) Diethyl malonate, NaH, 0
°C–rt, 3 h, 94%; (b) POCl3, 80 °C, 2.5
h, 57%; (c) (S)-2-aminopropan-1-ol, DIPEA, NMP, 160
°C, 1 h, then LiCl 160 °C, 1 h, 37–70%; (d) H2, Pd/C, ethanol, rt, 16 h, 92%; (e) 2,4-dichloropyridine-3-carbonitrile,
DIPEA, NMP, 160 °C, 1 h, 10–19%; (f) BF3·OEt2, THF, NaBH4, 0 °C, 2 h, 32%.Our initial attempts to prepare the isomeric cyclic ethers
(1,2,3,4-tetrahydro-[1,4]oxazepino[2,3-c]quinolin-6(7H)-ones, Scheme ) from iodo-quinolinone 25b by the Ullmann cyclization
conditions used previously
(Scheme ) were unsuccessful,
leading only to dehalogenation (Scheme , step c). With their conformation proposed to enable
closer surface contact as described above, their synthesis remained
a high priority. However, no syntheses of this core were previously
reported, so we sought other methods for forming the desired C–O
bond. We found an intriguing report suggesting that simple aryl ethers
could be formed by microwave-heating alcohols with aryl halides in
dimethyl sulfoxide (DMSO) in the presence of potassium tert-butoxide.[25] Although the reported scope
was both limited and quite different to our desired reaction, we applied
these conditions to iodo-intermediate 25b. We were delighted
to observe the formation of desired intermediate 19,
albeit in a 1:1 ratio with dehalogenation product 25a. We thought that the weak C–I bond may be favoring dehalogenation
over the desired cyclisation and so instead prepared bromo-intermediate 25c by an acid-catalyzed electrophilic bromination. Treatment
of this intermediate with the cyclization conditions enabled improved
conversion to products, and the resulting conditions were sufficiently
effective to make final products 9a, 9b, 9r, 10, 11a–d, 12a–b, 13a–d, and 14 (Scheme ).
Scheme 5
Synthetic Approaches
to 1,2,3,4-Tetrahydro-[1,4]oxazepino[2,3-c]quinolin-6(7H)-ones
(a) 3-Aminobutan-1-ol, DIPEA,
NMP, 160 °C, 20 h, 66%; (b) iodine, methanol/water, 60 °C,
2 h, 38%; (c) 1,10-phen, CuI, cesium carbonate, NMP, 120 °C,
1 h, product not observed; (d) aminoalcohol, DIPEA or TEA, NMP, 160
°C, 1–4 d, 11–71%; (e) KOBu, THF/DMSO, μw, 100 °C, 10 min, 20%; (f) NBS,
TFA, DCM, 0 °C–rt, 10–40 min, 40–77%; (g)
KOBu, THF/DMSO, μw, 60 °C,
50–200 min, 7–51%; (h) ammonium formate, Pd/C, methanol,
80 °C, 20–30 min, 67–100%; (i) 2,4-dichloropyridine-3-carbonitrile,
DIPEA or TEA, NMP, 160 °C, 60–90 min, 10–80%; (j)
3-amino-4,4-difluoro-butan-1-ol, DIPEA, NMP, 160 °C, 2 h then
LiCl, 160 °C, 2 h, 70%; (k) 3-amino-3-cyclopropylpropan-1-ol,
DIPEA, MeCN, 85 °C, 22 h then NaOH, THF/water, 85 °C, 6
h, 96%.
Synthetic Approaches
to 1,2,3,4-Tetrahydro-[1,4]oxazepino[2,3-c]quinolin-6(7H)-ones
(a) 3-Aminobutan-1-ol, DIPEA,
NMP, 160 °C, 20 h, 66%; (b) iodine, methanol/water, 60 °C,
2 h, 38%; (c) 1,10-phen, CuI, cesium carbonate, NMP, 120 °C,
1 h, product not observed; (d) aminoalcohol, DIPEA or TEA, NMP, 160
°C, 1–4 d, 11–71%; (e) KOBu, THF/DMSO, μw, 100 °C, 10 min, 20%; (f) NBS,
TFA, DCM, 0 °C–rt, 10–40 min, 40–77%; (g)
KOBu, THF/DMSO, μw, 60 °C,
50–200 min, 7–51%; (h) ammonium formate, Pd/C, methanol,
80 °C, 20–30 min, 67–100%; (i) 2,4-dichloropyridine-3-carbonitrile,
DIPEA or TEA, NMP, 160 °C, 60–90 min, 10–80%; (j)
3-amino-4,4-difluoro-butan-1-ol, DIPEA, NMP, 160 °C, 2 h then
LiCl, 160 °C, 2 h, 70%; (k) 3-amino-3-cyclopropylpropan-1-ol,
DIPEA, MeCN, 85 °C, 22 h then NaOH, THF/water, 85 °C, 6
h, 96%.We previously reported that replacing
4-chloroquinolinone 20a with the more reactive chloro-ester 20b allowed
SNAr reactions on the quinolinone core to proceed more
readily.[16] The activating ester group could
then be removed by base-mediated hydrolysis/decarboxylation or treatment
with lithium chloride. This increased reactivity enabled us to access
electron-withdrawing substituents such as difluoromethyl (11e). We also applied this method to improve the yield for the synthesis
of cyclopropyl compound 12a following promising activity
data. Replacing 20a with ester 20b allowed
reduced temperature and higher yields, and desired intermediate 26i was obtained in 96% yield over two steps, compared with
45% from the original route using 20a (Scheme ).
Scheme 6
Synthesis of 2-Cyclopropyl-1,2,3,4-tetrahydro-[1,4]oxazepino[2,3-c]quinolin-6(7H)-ones
(a)
DIPEA (2.5 equiv), MeCN,
85 °C, 22 h, then 2 M NaOH (6 equiv), THF, 85 °C, 6 h, 96%;
(b) NBS (1.5 equiv), TFA (5 equiv), DCM, 0 °C, 20 min, 58–79%;
(c) LiOtBu (1 M in THF; 1.6 equiv), THF [0.1 M], 60 °C, 15 min,
99%; (d) H2, Pd/C, ethanol, 1–2 h, 99%; (e) 2,4-dichloropyridine-3-carbonitrile,
TEA, NMP, 160 °C μW, 90 min, 51–53%.
Synthesis of 2-Cyclopropyl-1,2,3,4-tetrahydro-[1,4]oxazepino[2,3-c]quinolin-6(7H)-ones
(a)
DIPEA (2.5 equiv), MeCN,
85 °C, 22 h, then 2 M NaOH (6 equiv), THF, 85 °C, 6 h, 96%;
(b) NBS (1.5 equiv), TFA (5 equiv), DCM, 0 °C, 20 min, 58–79%;
(c) LiOtBu (1 M in THF; 1.6 equiv), THF [0.1 M], 60 °C, 15 min,
99%; (d) H2, Pd/C, ethanol, 1–2 h, 99%; (e) 2,4-dichloropyridine-3-carbonitrile,
TEA, NMP, 160 °C μW, 90 min, 51–53%.The microwave-assisted conditions for the key cyclization
step
to form the seven-membered rings gave modest and variable yields due
to incomplete conversion, difficulty in separating the desired product
from the unreacted starting material, and debromination to 26i under the reaction conditions. Hence, we looked to optimize this
reaction to make the synthesis of the core compatible with future
scale-up. Optimization was carried out using the racemic methyl derivative 25c (Table S4).The standard
reaction conditions under microwave irradiation afforded
an approximately 3:1 ratio of cyclized product 19/debromination
product 25a, reaching 91% conversion. Intermediate 19 was isolated in 41% yield. We wanted to switch to conventional
heating to simplify scale-up, and this gave a similar ratio of ∼4:1
of 19/25a. A range of bases and other additives were
tried instead of the tert-butoxides, but product
formation was not observed. Switching the counterion to sodium was
tolerated, and the use of lithium tert-butoxide maintained
a similar ratio of 19/25a but gave improved conversion
of the starting material.The proposed mechanism for the published
work is via a benzyne
intermediate, and good yield was only obtained using DMSO as a solvent.[25] However, this mechanism is not possible for
our substrates, suggesting that a different reaction pathway must
be occurring. We therefore undertook a short solvent screen. In our
case, DMSO was not essential, and use of ether solvents or DCE minimized
debromination, with tetrahydrofuran (THF) proving optimal. We could
lower the equivalents of lithium tert-butoxide to
1.6 and increase the concentration of the reaction mixture to 0.1
M. Finally, we were able to significantly reduce the reaction time
with complete conversion seen after only 15 min. An isolated yield
of 81% of 19 was achieved after aqueous work-up with
no further purification required.Reoptimised conditions were
used for the synthesis of the more
potent cyclopropyl-containing tricyclic compounds (Scheme ). Gratifyingly, under the
new reaction conditions, cyclization occurred cleanly and a quantitative
isolated yield was obtained for intermediate 28i after
aqueous work-up with no chromatography required. For the difluoro
analogue, nitro intermediates 32, 33, and 34a were obtained in ∼85% ee due to the presence of the minor
(R)-isomer as an impurity in the starting material
(3S)-3-amino-3-cyclopropyl-2,2-difluoro-propan-1-ol
hydrochloride, which we obtained commercially. These isomers were
separated at the nitro stage by chiral supercritical fluid chromatography
(SFC) and used to prepare lead compound 1 and its enantiomer 14.
Conclusions
Obtaining sufficient binding affinity to
inhibit protein–protein
interactions can be challenging, and frequently, the resulting inhibitors
are large and complex molecules, posing development challenges. To
meet our objective of identifying chemical scaffolds with improved
BCL6 binding affinity and a good ligand efficiency, we therefore focused
on gaining as much binding affinity as possible from each area of
the binding site. In this study, we particularly explored optimizing
the shape complementarity in the BCL6 substrate binding region defined
by
residues Asp17, Val18, Cys53, and His14 (the HDCH site). Initial attempts
to achieve this through introducing an additional substituent failed,
likely because of the conformation effect on the adjacent substituent.
We solved this issue by constraining both substituents into a ring
and guided by X-ray crystallography designed a series of six- and
seven-membered fused rings. A broad range of synthetic methods were
used, taking advantage of available intermediates, to explore this
diverse set of ring fusions. Cyclic lactone 7 provided
the first validation of this approach, with high binding affinity,
but poor permeability prevented further progression of this series.
X-ray crystallography enabled us to prioritize possible single-point
changes to the core structure—including moving the ether oxygen
to form cyclic ether 9a, predicted and later confirmed
to have a 3D shape which more completely filled the corepressor binding
pocket on the BTB domain of BCL6. The synthesis of this potent core
required development of modified cyclization reaction conditions,
and the broad scope of this reaction enabled us to explore a range
of substituents, which we again prioritized based on X-ray structural
information. We further optimized shape complementarity and achieved
another breakthrough in potency by introducing a cyclopropyl group
and then difluoro substituents to give our lead molecule 1. The improvement in shape complementarity throughout this optimization
can be seen in the comparison of crystal structures shown in Figure . The increase in
LLE observed across this series (Figure ) also demonstrates the potential utility
of LLE plots in identifying compounds with improved shape complementarity.
Overall, this work illustrates how optimizing shape complementarity
leads to improved biological activity, while maintaining the overall
lipophilicity under control and without introducing additional polar
features.
Figure 7
LLE plot of compounds from this study, reporting pIC50 vs calculated
log D7.4 (Moka[18,19]), with iso-LLE lines. Hydrophobic substituents that improve shape
complementarity generally exhibit increased LLE (10, 9a/r, 11/12a, 1). In contrast, adding
hydrophobic substituents that do not produce optimal surface contacts
may maintain (9r, 11b) or increase (10, 11c/d) potency but show no improvement in
LLE. Other compounds from the tricyclic ether series (9b, 11e, 12b, 13a, 13c, 13d, 14) are shown as gray circles, and
those from other series presented in this paper (compounds 2–8) are shown as white circles.
LLE plot of compounds from this study, reporting pIC50 vs calculated
log D7.4 (Moka[18,19]), with iso-LLE lines. Hydrophobic substituents that improve shape
complementarity generally exhibit increased LLE (10, 9a/r, 11/12a, 1). In contrast, adding
hydrophobic substituents that do not produce optimal surface contacts
may maintain (9r, 11b) or increase (10, 11c/d) potency but show no improvement in
LLE. Other compounds from the tricyclic ether series (9b, 11e, 12b, 13a, 13c, 13d, 14) are shown as gray circles, and
those from other series presented in this paper (compounds 2–8) are shown as white circles.Lead compound 1 is a potent and ligand-efficient inhibitor
of BCL6 and shows excellent translation into the cellular NanoBRET
assay (Table ). Although
our previous lead inhibitor 3 showed a >20-fold drop-off, 1 shows ∼3-fold from 3.9 nM (TR-FRET) to 12 nM (NanoBRET).
We speculate that this discrepancy could result from differences in
binding to the TR-FRET assay construct (BTB domain) and full length
BCL6 used in the nanoBRET: this hypothesis has been previously proposed
to explain mismatches between cell-free and cellular assays for a
different series of BCL6 inhibitors.[26] The
improved cellular potency of 1 translates into increased
antiproliferative activity in BCL6-high OCI-Ly1 and Karpas-422 cell
lines, and selectivity over the BCL6-low OCI-Ly3 line is maintained
(Table ). With improved
potency, a reduced molecular weight and polar surface area, and hence
an increased ligand efficiency compared to 3 (Table ), compound 1 (CCT372064) represents a new core structure for further
elaboration toward the identification of candidate-quality BCL6 inhibitors;
a subsequent publication will describe how this new core enabled the
discovery of more potent, in vivo active degraders of BCL6.
Table 4
Comparison of Lead Compounds: Activity
in Biochemical, Cellular, and Antiproliferative Effects; Permeability;
Calculated Physicochemical Properties; and Ligand Efficiency Metrics
No
BCL6 TR-FRET IC50
(μM)
cellular
NanoBRET IC50 (μM)
OCI-Ly1 GI50 (μM)
Karpas 422 GI50 (μM)
OCI-Ly3 GI50 (μM)
PAMPA (×
10–6 cm s–1) Pe (7.4)
MW
Moka[18,19] log D7.4
TPSA
Å2
LE
LLE
2
1.2
>30
NT
NT
NT
<0.1
340
2.4
94
0.35
3.5
3 (CCT369347)
0.026
0.57
1.68
3.32
>10
35
486
3.0
109
0.30
4.6
1 (CCT372064)
0.0039
0.012
0.17
0.19
3.4
5.5
458
3.0
92
0.36
5.3
Experimental Section
General Synthetic Information
All anhydrous solvents
and reagents were obtained from commercial suppliers and used without
further purification. Evaporation of solvents was carried out using
a rotary evaporator under reduced pressure at a bath temperature of
up to 60 °C. Flash column chromatography was carried out using
a Biotage purification system using SNAP KP-Sil cartridges or on the
reverse-phase mode using SNAP Ultra C18 cartridges. Semi-preparative
separations were carried out using a 1200 series preparative high-performance
liquid chromatography (HPLC) over a 15 min gradient elution. Microwave-assisted
reactions were carried out using a Biotage initiator microwave system.
Final compounds were purified to ≥95% purity. NMR data were
collected on a Bruker AVANCE 500 spectrometer equipped with a 5 mm
BBO/QNP probe or on a Bruker AVANCE Neo 600 spectrometer equipped
with a 5 mm TCI Cryo-Probe. NMR data are presented in the form of
chemical shift δ (multiplicity, coupling constants, integration)
for major diagnostic protons, given in parts per million relative
to tetramethylsilane, referenced to the internal deuterated solvent.
High-resolution mass spectrometry (HRMS) was assessed using an Agilent
1200 series HPLC and a diode array detector coupled to a 6120 time
of a flight mass spectrometer with a dual multimode APCI/ESI source
or on a Waters Acquity UHPLC and a diode array detector coupled to
a Waters G2 QToF mass spectrometer fitted with a multimode ESI/APCI
source.
Preparation of Compounds
Compounds 2, 3, 20a–c, and 23 were prepared
as previously reported.[16]
A mixture
of 34a (1.08 g, 3.09 mmol) and Pd/C (10 wt %, 108 mg)
in ethanol (15 mL) was stirred at 60 °C under an atmosphere of
H2 for 1 h. The reaction mixture was allowed to cool to
rt and then filtered through Celite, and the solids were washed with
ethanol (60 mL). The filtrate was concentrated under reduced pressure,
affording the title compound (1.03 g, 100%) as an orange solid, which
was used without purification in step 2. LCMS (2 min; ESI) RT 0.89
min; m/z: calcd for C16H18F2N3O2+ [M + H]+, 322.1362; found, 322.1370.
To a stirred solution of 2,4-dichloro-3-methyl-6-nitro-quinoline[27] (480 mg, 1.87 mmol) in 1,4-dioxane (3.73 mL)
was added conc. HCl (0.46 mL, 14.9 mmol) dropwise. The reaction mixture
was refluxed for 18 h and then poured into ice water and stirred for
1 h. The resulting solid was filtered and dried under vacuum to afford
the title compound (363 mg, 81%) as a white solid. The material was
taken forward to the next step with no further purification. 1H NMR (500 MHz, DMSO-d6): δ
12.59 (s, 1 H), 8.57 (d, J = 2.5 Hz, 1 H), 8.37 (dd, J = 9.0, 2.5 Hz, 1 H), 7.47 (d, J = 9.0
Hz, 1 H), 2.24 (s, 3 H); LCMS (2 min; ToF) RT 1.44 min m/z: calcd for C10H8ClN2O3+ [M + H]+, 239.0218; found,
239.0200.
To a solution of the product of step 1 (100
mg, 0.42
mmol) in dimethylformamide (DMF) (0.84 mL) at rt was added sodium
hydride (60% in mineral oil, 33.5 mg, 0.84 mmol), followed by iodomethane
(39 μL, 0.63 mmol). The mixture was stirred at rt for 1 h; then,
water was added with care to the reaction mixture, forming an orange
precipitate which was filtered under suction. The orange solid was
washed several times with water. The residual solid was washed with
DCM; then, the filtrate was separated, and the organic phase was concentrated.
The resulting solid was combined with the solid isolated by filtration
to give the title compound (101 mg, 95%). 1H NMR (500 MHz,
DMSO-d6): δ 8.69 (d, J = 2.6 Hz, 1 H), 8.45 (dd, J = 9.3, 2.6 Hz, 1 H),
7.81 (d, J = 9.3 Hz, 1 H), 3.71 (s, 3 H), 2.31 (s,
3 H); LCMS (2 min; ToF) RT 1.46 min m/z: calcd for C11H10ClN2O3+ [M + H]+, 253.0374; found, 253.0350.
To a suspension of the product of step 2
in NMP (1.42 mL) was added ethylamine (2.0 M in THF, 0.71 mL, 1.42
mmol), and the resulting mixture was heated to 200 °C for 12
h. After purification by HPLC (10–100% methanol in water, 0.1%
formic acid), the title compound was isolated (7 mg, 19%, 91% purity
by LCMS) as a pale-yellow solid, which was used without further purification
in the next step. LCMS (2 min; ToF) RT 0.53 min m/z: calcd for C13H18N3O+ [M + H]+, 232.1444; found, 232.1422.
A suspension of 20c(16) (43 mg, 0.19 mmol) and azetidine (77
μL, 1.15 mmol) in NMP (1.00 mL) was stirred at 160 °C under
microwave irradiation for 1 h. Water was added, and the precipitate
was collected by filtration. 20d (26.5 mg, 56%, 91% purity
by LCMS) was obtained as a yellow solid, which was used without further
purification in step 2. 1H NMR (500 MHz, DMSO-d6): δ 11.51 (s, 1 H), 8.53 (d, J = 2.5 Hz, 1 H), 8.31 (dd, J = 9.1, 2.5 Hz, 1 H),
7.39 (d, J = 9.1 Hz, 1 H), 5.17 (s, 1 H), 4.35 (t, J = 7.5 Hz, 4 H), 2.43 (p, J = 7.5 Hz,
2 H); LCMS (2 min; ToF) RT 1.24 min, m/z: calcd for C12H12N3O3+ [M + H]+, 246.0884; found, 246.0886.
A mixture of 21 (from step 2, 34 mg, 0.076
mmol) and DIPEA (40 μL, 0.23 mmol) in NMP (0.76 mL) was heated
at 160 °C under microwave irradiation for 1 h. The reaction mixture
was allowed to cool to rt and then diluted with water and extracted
with EtOAc. The organic extracts were combined, washed with water
and brine, dried (MgSO4), and concentrated under reduced
pressure. Purification by flash chromatography (0 to 10% methanol
in DCM) afforded 15 (14 mg, 67%) as an orange solid.
LCMS (2 min; ToF) RT 1.36 min, m/z: calcd for C14H16N3O3+ [M + H]+, 274.1186; found, 274.1161.
To a solution
of 15 (from step 3, 14 mg, 0.0026 mmol)
in ethanol (1 mL) and NMP (0.2 mL) was added Pd/C (10 wt %, 2.7 mg,
0.05 eq.), followed by ammonium formate (32 mg, 0.51 mmol). The resulting
mixture was heated under an argon atmosphere at 60 °C in a sealed
vial for 30 min and then filtered through Celite and purified using
an SCX-2 (2 g) column, eluting with methanol (20 mL), followed by
2 N methanolic ammonia (20 mL). The ammonia fraction was concentrated
under reduced pressure, affording the title compound (16 mg, 100%)
as a yellow oil. LCMS (2 min; ToF) RT 0.62 min, m/z: calcd for C14H18N3O+ [M + H]+, 244.1444; found, 244.1443.
To a mixture of 22a (from step 1, 52 mg, 0.19 mmol) and iodine (145 mg, 0.57
mmol) under an argon atmosphere was added anhydrous methanol (1.2
mL), and the reaction mixture was heated at 60 °C under microwave
irradiation for 30 min. Water (0.6 mL) was added, and the reaction
mixture was heated at 60 °C under microwave irradiation for a
further 90 min. The reaction mixture was allowed to cool to rt, diluted
with methanol, and loaded onto silica. Purification by flash chromatography
(0 to 15% methanol in DCM) afforded 22b (29 mg, 38%)
as a yellow solid. 1H NMR (500 MHz, CDCl3):
δ 8.92 (d, J = 2.6 Hz, 1 H), 8.40 (dd, J = 9.3, 2.6 Hz, 1 H), 7.46 (d, J = 9.3
Hz, 1 H), 4.62 (d, J = 10.6 Hz, 1 H), 3.99–3.90
(m, 1 H), 3.86–3.76 (m, 5 H), 3.71 (dd, J =
11.2, 5.9 Hz, 1 H), 1.37 (d, J = 6.6 Hz, 3 H); LCMS
(2 min; ToF) RT 1.31 min; m/z: calcd
for C13H15IN3O4+ [M + H]+, 404.0102; found, 404.0108.
To a mixture of 16 (from step
3, 11 mg,
0.0392 mmol), Pd/C (10 wt %, 2 mg), and ammonium formate (13 mg, 0.20
mmol) under an argon atmosphere was added anhydrous methanol (0.4
mL), and the reaction mixture was stirred in a sealed vial at 80 °C
for 90 min. The reaction mixture was allowed to cool to rt and filtered
through Celite, and the solids were washed with methanol (20 mL).
The filtrate was concentrated under reduced pressure, redissolved
in methanol, and passed through an SCX-2 (2 g) column, washing with
methanol (30 mL) and then eluting with 2 N methanolic ammonia (30
mL). The ammonia fraction was concentrated under reduced pressure,
affording the title compound (7.8 mg, 81%) as an off-white solid,
which was used without further purification. LCMS (2 min; ToF) RT
0.21 min; m/z: calcd for C13H16N3O2+ [M + H]+, 246.1237; found, 246.1253.
A suspension of 17 (from
step 1, 48 mg, 0.16 mmol) and Pd/C (10 wt %, 3.4 mg) in ethanol (4
mL) was stirred under an atmosphere of H2 for 16 h. The
reaction mixture was filtered through Celite, and the solids washed
with methanol. The filtrate was concentrated under reduced pressure,
affording 24a (40 mg, 92, 85% purity by LCMS) as a yellow
oil, which was used without further purification. LCMS (2 min; ToF)
RT 0.19 min; m/z: calcd for C14H16N3O3+ [M +
H]+, 274.1186; found, 274.1179.
Boron trifluoride diethyl etherate (∼50% BF3;
0.1 mL, 0.41 mmol) was added to a stirred suspension of 24a (from compound 7 synthesis, step 2, 12 mg, 0.042 mmol)
in THF (4 mL) at 0 °C. The reaction mixture was stirred at 0
°C for 15 min, after which sodium borohydride (5 mg, 0.127 mmol)
was added. The reaction mixture was stirred at 0 °C for a further
2 h. The reaction mixture was quenched with the addition of methanol
and then concentrated under reduced pressure. Brine was added to the
residue, and the aqueous mixture was extracted with EtOAc. The organic
layer was dried (Na2SO4) and concentrated. Purification
by flash chromatography (5% methanol in EtOAc) afforded 24b (4 mg, 32%) as a yellow solid. LCMS (4 min; ToF) RT 0.69 min; m/z: calcd for C14H18N3O2+ [M + H]+, 260.1394;
found, 260.1399.
To a mixture of 20a (800 mg, 3.4 mmol) and (S)-3-aminobutan-1-ol
(446 mg, 5.0 mmol) in a dry vial under argon was added anhydrous NMP
(10 mL), followed by DIPEA (1.2 mL, 6.9 mmol). The reaction mixture
was heated at 160 °C in a heating block for 20 h and then allowed
to cool to rt and diluted with water (100 mL), and the aqueous mixture
was extracted with EtOAc (100 mL). The organic extract was washed
with water (2 × 25 mL). The aqueous washings were combined and
further extracted with EtOAc (3 × 50 mL). The organic extracts
were combined, dried (Na2SO4), and concentrated
under reduced pressure. The crude reaction mixture was dry-loaded
onto silica and purified by flash chromatography (0 to 10% methanol
in DCM), affording 26a (547 mg, 56%) as a yellow solid. 1H NMR (500 MHz, DMSO-d6): δ
9.11 (d, J = 2.5 Hz, 1 H), 8.37 (dd, J = 9.4, 2.5 Hz, 1 H), 7.60 (d, J = 9.4 Hz, 1 H),
7.11 (d, J = 7.9 Hz, 1 H), 5.59 (s, 1 H), 4.57 (t, J = 5.0 Hz, 1 H), 3.80–3.71 (m, 1 H), 3.55 (s, 3
H), 3.53–3.48 (m, 2 H), 1.94–1.87 (m, 1 H), 1.67–1.60
(m, 1 H), 1.23 (d, J = 6.4 Hz, 3 H). LCMS (2 min;
ToF); RT 1.21 min; m/z: calcd for
C14H18N3O4+ [M + H]+, 292.1292; found, 292.1274.
Trifluoroacetic acid
(0.72 mL, 9.4 mmol) was added to a stirred mixture of N-bromosuccinimide (509 mg, 2.9 mmol) and 26a (547 mg,
1.9 mmol) in anhydrous DCM (10 mL) at 0 °C under Ar. The reaction
mixture was stirred at 0 °C for 10 min and at rt for 30 min and
then diluted with ethyl acetate (100 mL) and washed with water (30
mL) and saturated aq. NaHCO3 (3 × 30 mL). The aqueous
washings were combined and re-extracted with EtOAc (30 mL). The organic
extracts were combined, dried (Na2SO4), and
concentrated under reduced pressure and then dry-loaded onto silica
and purified by flash chromatography (0 to 10% methanol in DCM), affording 27a (532 mg, 77%) as a yellow solid. 1H NMR (500
MHz, DMSO-d6): δ 8.89 (d, J = 2.6 Hz, 1 H), 8.42 (dd, J = 9.4, 2.6
Hz, 1 H), 7.72 (d, J = 9.4 Hz, 1 H), 5.82 (d, J = 9.8 Hz, 1 H), 4.51 (t, J = 4.7 Hz,
1 H), 4.28–4.19 (m, 1 H), 3.69 (s, 3 H), 3.51–3.46 (m,
2 H), 1.90–1.82 (m, 1 H), 1.79–1.71 (m, 1 H), 1.29 (d, J = 6.5 Hz, 3 H). LCMS (2 min; ToF); RT 1.31 min; m/z: calcd for C14H17BrN3O4+ [M + H]+, 370.0397;
found, 372.0390.
To 27a (111 mg, 0.30 mmol) under argon was added dry
DMSO (4 mL) and then potassium tert-butoxide (1 M
in THF; 0.54 mL, 0.54 mmol). The reaction mixture was heated at 60
°C under microwave irradiation for 50 min and then allowed to
cool to rt. Water (10 mL) was added, followed by EtOAc (10 mL). The
layers were separated, and the aqueous layer was further extracted
with EtOAc (10 mL). The organic extracts were combined and concentrated
under reduced pressure. The crude product was dissolved in DMSO (1.2
mL) and purified by reverse-phase chromatography (45–75% methanol
in water, 0.1% formic acid), affording 28a (36 mg, 41%)
as a dark-yellow solid. 1H NMR (600 MHz, CDCl3): δ 8.93 (d, J = 2.1 Hz, 1 H), 8.33 (dd, J = 9.2, 2.1 Hz, 1 H), 7.40 (d, J = 9.2
Hz, 1 H), 4.49–4.38 (m, 2 H), 4.14–4.08 (m, 1 H), 4.01
(br s, 1 H), 3.76 (s, 3 H), 2.23–2.26 (m, 1 H), 1.91–1.84
(m, 1 H), 1.47 (d, J = 6.3 Hz, 3 H). LCMS (2 min;
ToF); RT 1.26 min; m/z: calcd for
C14H16N3O4+ [M + H]+, 290.1135; found, 290.1131.
A mixture of 28a (35.8 mg, 0.124 mmol), Pd/C (10 wt
%) (6.3 mg, 0.006 mmol), and ammonium formate (53.3 mg, 0.845 mmol)
under argon in methanol (1.2 mL) was heated in a sealed vial to 80
°C for 20 min. Further, Pd/C (2.6 mg) and ammonium formate (32.5
mg) were added and heated to 80 °C for 10 min. The mixture was
filtered through Celite, washing with methanol (40 mL). The filtrate
was concentrated under reduced pressure, redissolved in methanol,
and purified using an SCX-2 (2 g) column, washing with methanol (40
mL) and then eluting with 2 N methanolic ammonia (40 mL). The methanolic
ammonia fraction was concentrated under reduced pressure, affording 29a (21.5 mg, 67%, 0.083 mmol) as a dark-yellow solid, which
was used without further purification. LCMS (2 min; ToF) RT 0.41 min m/z: calcd for C14H18N3O2+ [M + H]+, 260.1394;
found, 260.1393.
Prepared as described for its enantiomer, 9a, starting
from (R)-3-aminobutan-1-ol. 1H NMR as
for 9a. LCMS (4 min; ToF) RT = 2.58 min m/z: calcd for C20H19ClN5O2+ [M + H]+, 396.1222; found,
396.1213.
Prepared as described for a single enantiomer, 9a,
starting from rac-3-aminobutan-1-ol. 1H NMR as for 9a. LCMS (4 min; ToF) RT = 2.58 min m/z: calcd for C20H19ClN5O2+ [M + H]+, 396.1222;
found, 396.1217.
Obtained as a yellow solid (13 mg, 85%), which was used without
further purification. LCMS (2 min; ToF) RT 0.91 min; m/z: calcd for C16H20N3O2+ [M + H]+, 286.1550; found,
286.1551.
Prepared by transfer hydrogenation of 28e (2.5 mg,
0.008 mmol) using the procedure described for 29a (9a step 4), affording 29e (2 mg, 89%) as a yellow
oil, used without purification in the next step. LCMS (2 min; ToF)
RT 0.84 min; m/z: calcd for C15H20N3O2+ [M +
H]+, 274.1550; found, 274.1542.
Prepared by transfer hydrogenation of 28f (36 mg,
0.11 mmol) according to the procedure used for 29a. The
reaction time was 20 min, affording 29f (32 mg, 99%)
as a dark-orange solid. LCMS (2 min; ESI) RT 0.95 min; m/z: calcd for C17H22N3O2+ [M + H]+, 300.1707; found,
300.1716.
Prepared from 28h (16 mg, 0.05 mmol), according to
the procedure used for 29a. The reaction time was 20
min, affording 29h (14 mg, 96%) as a dark-yellow solid.
LCMS (2 min; ESI) RT 0.61 min; m/z: calcd for C14H16F2N3O2+ [M + H]+, 296.1205; found, 296.1224.
Method A: According to
the same procedure as 26a, reaction time 4 d at 160 °C,
from 20a (500 mg, 2.1 mmol) and (R)-3-amino-3-cyclopropylpropan-1-ol
(541 mg, 3.6 mmol), 26i obtained as a yellow solid (298
mg, 45%).Method B: (R)-3-Amino-3-cyclopropylpropan-1-ol
(0.79 g, 6.8 mmol), 20b (1.51 g, 4.8 mmol), DIPEA (2.1
mL, 12.1 mmol), and anhydrous acetonitrile (24 mL) in a dry flask
under argon were heated at 85 °C for 22 h. The reaction mixture
was concentrated under reduced pressure to remove MeCN. THF (15 mL)
was added, followed by sodium hydroxide (2 M, 14 mL, 28 mmol). The
reaction mixture was heated at 85 °C for 6 h. The reaction mixture
was cooled to rt, diluted with water (30 mL), and acidified to pH
5 with 3 M HCl. The resulting precipitate was filtered, washed with
water (70 mL), and dried, affording 26i (1.48 g, 96%)
as a dark-yellow solid. 1H NMR (500 MHz, DMSO-d6): δ 9.15 (d, J = 2.5 Hz, 1 H),
8.38 (dd, J = 9.4, 2.5 Hz, 1 H), 7.60 (d, J = 9.4 Hz, 1 H), 7.17 (d, J = 8.3 Hz,
1 H), 5.59 (s, 1 H), 4.51 (t, J = 5.0 Hz, 1 H), 3.59–3.53
(m, 4 H), 3.51–3.45 (m, 1 H), 3.27–3.22 (m, 1 H), 1.95–1.87
(m, 1 H), 1.84–1.77 (m, 1 H), 1.14–1.07 (m, 1 H), 0.53–0.47
(m, 1 H), 0.43–0.37 (m, 1 H), 0.28–0.23 (m, 2 H); LCMS
(2 min; ESI) RT 1.17 min; m/z: calcd
for C16H20N3O4+ [M + H]+, 318.1448; found, 318.1452.
Trifluoroacetic acid
(1.8 mL, 23.5 mmol) was added to a stirred mixture of N-bromosuccinimide (1.25 g, 7 mmol) and 26i (1.48 g,
4.67 mmol) in DCM (30 mL) at 0 °C under Ar. The reaction mixture
was stirred at 0 °C for 15 min and then diluted with DCM (30
mL) and washed with saturated aq. NaHCO3 (2 × 50 mL).
The aqueous washings were further extracted with DCM (30 mL). The
organic extracts were combined, washed with brine (50 mL), dried (Na2SO4), and concentrated under reduced pressure.
The resulting product was recrystallized from ethanol to give 498
mg of the desired product; the filtrate was dry-loaded onto silica,
and purification by flash chromatography (0 to 10% methanol in DCM)
afforded a further 582 mg of the desired product. Product batches
were combined to give 27i (1.08 g, 58%) as a yellow solid. 1H NMR (500 MHz, DMSO-d6): δ
8.94 (d, J = 2.6 Hz, 1 H), 8.41 (dd, J = 9.4, 2.6 Hz, 1 H), 7.73 (d, J = 9.4 Hz, 1 H),
5.67 (d, J = 10.3 Hz, 1 H), 4.52 (t, J = 4.7 Hz, 1 H), 3.70 (s, 3 H), 3.65–3.60 (m, 2 H), 3.56–3.48
(m, 1 H), 2.01–1.87 (m, 2 H), 1.15–1.06 (m, 1 H), 0.48–0.40
(m, 1 H), 0.35–0.28 (m, 1 H), 0.23–0.17 (m, 1 H), 0.08–0.02
(m, 1 H); LCMS (2 min; ESI) RT 1.25 min; m/z: calcd for C16H19BrN3O4+ [M + H]+, 396.0553; found, 396.0584.
Method A: To 27i (374 mg, 0.94 mmol) under argon was
added DMSO (12.6 mL), followed by potassium tert-butoxide
(1 M in THF, 1.7 mL, 1.7 mmol), and the reaction mixture was heated
at 60 °C under microwave irradiation for 50 min. To a further
portion of 27i (290 mg, 0.73 mmol) under argon was added
DMSO (9.8 mL), followed by potassium tert-butoxide
(1 M in THF, 1.32 mL, 1.32 mmol), and the reaction mixture was heated
at 60 °C under microwave irradiation for 50 min. The reaction
mixtures were combined, and water (100 mL) was added, followed by
EtOAc (100 mL). The layers were separated, and the aqueous layer was
further extracted with EtOAc (2 × 100 mL). The organic extracts
were combined and concentrated under reduced pressure. The crude product
was dissolved in DMSO (1.5 mL) and purified by reverse-phase chromatography
(45–65% methanol in water, 0.1% formic acid) to give 28i (247 mg, 47%) as a yellow solid.Method B: Lithium tert-butoxide (1 M in THF; 2.35 mL, 2.35 mmol) was added
to a suspension of 27i (582 mg, 1.47 mmol) in anhydrous
THF (14.7 mL) under Ar. A reflux condenser was fitted, and the reaction
mixture was heated at 60 °C for 15 min. The reaction mixture
was cooled to rt. Water (10 mL) was added, and the aqueous mixture
was extracted with DCM (4 × 10 mL). The organic extracts were
combined, washed with brine (2 × 10 mL), dried (Na2SO4), and concentrated under reduced pressure, affording 28i (463 mg, 99%) as an orange solid. 1H NMR (500
MHz, DMSO-d6): δ 9.03 (d, J = 2.6 Hz, 1 H), 8.31 (dd, J = 9.4, 2.6
Hz, 1 H), 7.60 (d, J = 9.4 Hz, 1 H), 6.62 (d, J = 4.4 Hz, 1 H), 4.26–4.18 (m, 2 H), 3.60 (s, 3
H), 2.94–2.87 (m, 1 H), 2.25–2.18 (m, 1 H), 2.05–1.99
(m, 1 H), 1.29–1.23 (m, 1 H), 0.57–0.49 (m, 2 H), 0.41–0.36
(m, 1 H), 0.30–0.24 (m, 1 H); LCMS (2 min; ESI) RT 1.25 min; m/z: calcd for C16H18N3O4+ [M + H]+, 316.1292;
found, 316.1304.
28i (463 mg, 1.47 mmol), Pd/C (10 wt %, 46 mg), and
ethanol (10 mL) were stirred at 60 °C under an atmosphere of
H2 for 2 h. The reaction mixture was allowed to cool to
rt. The reaction mixture was filtered through Celite, and the solids
were washed with EtOH (50 mL). The filtrate was concentrated under
reduced pressure, affording 29i (414 mg, 99%) as an orange
solid. LCMS (2 min; ESI) RT 0.76 min; m/z: calcd for C16H20N3O2+ [M + H]+, 286.1550; found, 286.1563.
Prepared by transfer hydrogenation of 28j (20 mg,
0.06 mmol) using the procedure described for 29a. Obtained
a yellow solid (17 mg, 92%). LCMS (2 min; ToF) RT 0.92 min; m/z: calcd for C16H20N3O2+ [M + H]+, 286.1550;
found, 286.1554.
Prepared from 28k using the procedure described for 29a. 29k (2 mg, 100%) was obtained as dark-yellow
solid. LCMS (2 min; ToF); RT 0.79 min; m/z: calcd for C15H20N3O2+ [M + H]+, 274.1550; found, 274.1552.
Prepared from 28l using the procedure described for 29a. 29l (3 mg, 88%) was obtained and used without
further purification. LCMS (2 min; ToF) RT 0.98 min; m/z: calcd for C16H22N3O2+ [M + H]+, 288.1707; found,
288.1684.
Prepared from 28m using the procedure described for 29a. 29m (7 mg, 71%) was obtained as a yellow
solid. LCMS (2 min; ToF) RT 0.96 min; m/z: calcd for C16H22N3O2+ [M + H]+, 288.1707; found, 288.1704.
Prepared from 28n using the procedure
described for 29a. 29n (12 mg, 99%) was
obtained as a yellow solid. LCMS (2 min; ToF) RT 0.74 min; m/z: calcd for C16H20N3O2+ [M + H]+, 286.1550;
found, 286.1555.
To a mixture of (S)-3-amino-3-cyclopropyl-2,2-difluoropropan-1-ol hydrochloride 31 [1.99 g, 10.6 mmol (supplied as ∼90% ee by SIA Enamine,
Latvia)] and 20b (3.0 g, 9.7 mmol) under argon was added
anhydrous acetonitrile (15 mL), followed by DIPEA (4.2 mL, 24.1 mmol).
The reaction mixture was heated at 160 °C under microwave irradiation
for 15 h. 2 M sodium hydroxide (30 mL, 60 mmol) was added, and the
reaction mixture was heated at 85 °C for 2 h. The reaction mixture
was cooled to rt. Water (80 mL) was added, and the reaction mixture
was acidified to pH 4–5 with 3 M HCl. The resulting precipitate
was filtered, washed with water (200 mL), and dried under vacuum,
affording 32 (3.18 g, 93%) as a beige solid. 1H NMR (500 MHz, DMSO-d6): δ 9.31
(d, J = 2.5 Hz, 1 H), 8.40 (dd, J = 9.4, 2.5 Hz, 1 H), 7.62 (d, J = 9.4 Hz, 1 H),
7.48 (d, J = 8.7 Hz, 1 H), 5.72 (s, 1 H), 5.60 (t, J = 6.1 Hz, 1 H), 3.90–3.71 (m, 2 H), 3.57–3.45
(m, 4 H), 1.38–1.29 (m, 1 H), 0.71–0.64 (m, 1 H), 0.63–0.56
(m, 1 H), 0.53–0.46 (m, 1 H), 0.27–0.20 (m, 1 H); LCMS
(2 min; ESI) RT 1.15 min; m/z: calcd
for C16H18F2N3O4+ [M + H]+, 354.1260; found, 354.1270.
Trifluoroacetic acid (3.44
mL, 44.9 mmol) was added to a stirred mixture of 32 (3.18
g, 8.99 mmol) and freshly recrystallized N-bromosuccinimide
(1.60 g, 8.98 mmol) in anhydrous DCM (60 mL) at 0 °C under Ar.
The reaction mixture was stirred at 0 °C for 10 min. The reaction
mixture was diluted with DCM (100 mL) and washed with saturated aq.
NaHCO3 (3 × 80 mL). The aqueous washings were further
extracted with DCM (100 mL). The organic extracts were combined, washed
with brine (80 mL), dried (Na2SO4), and concentrated
under reduced pressure, affording 33 (3.77 g, 97%) as
a yellow solid. 1H NMR (500 MHz, DMSO-d6): δ 8.95 (d, J = 2.5 Hz, 1 H),
8.43 (dd, J = 9.4, 2.5 Hz, 1 H), 7.75 (d, J = 9.4 Hz, 1 H), 5.86 (d, J = 11.1 Hz,
1 H), 5.63 (t, J = 5.9 Hz, 1 H), 4.05–3.95
(m, 1 H), 3.89–3.74 (m, 2 H), 3.71 (s, 3 H), 1.29–1.21
(m, 1 H), 0.68–0.62 (m, 1 H), 0.62–0.51 (m, 2 H), 0.50–0.44
(m, 1 H); LCMS (2 min; ESI) RT 1.31 min; m/z: calcd for C16H17BrF2N3O4+ [M + H]+, 432.0365;
found, 432.0369.
(S)-2-Cyclopropyl-3,3-difluoro-7-methyl-10-nitro-1,2,3,4-tetrahydro-[1,4]oxazepino-[2,3-c]quinolin-6(7H)-one (34a)
and (R)-2-Cyclopropyl-3,3-difluoro-7-methyl-10-nitro-1,2,3,4-tetrahydro-[1,4]oxazepino-[2,3-c]quinolin-6(7H)-one (34b)
Lithium tert-butoxide (1 M in THF; 13.9 mL, 13.9
mmol) was added to a suspension of 33 (3.77 g, 8.71 mmol)
in THF (87 mL) under Ar. The reaction mixture was heated at 60 °C
for 15 min and then cooled to rt. Water (100 mL) was added, and the
aqueous mixture was extracted with DCM (3 × 100 mL). The organic
extracts were combined, washed with brine (2 × 100 mL), dried
(Na2SO4), and concentrated under reduced pressure,
affording 34a (3.01 g, 98%) as a yellow solid. 1H NMR (500 MHz, DMSO-d6): δ 9.12
(d, J = 2.5 Hz, 1 H), 8.35 (dd, J = 9.4, 2.5 Hz, 1 H), 7.66 (d, J = 9.4 Hz, 1 H),
7.01 (d, J = 4.4 Hz, 1 H), 4.54–4.37 (m, 2
H), 3.62 (s, 3 H, NCH3), 3.29–3.22 (m, 1 H), 1.39–1.31
(m, 1 H), 0.76–0.69 (m, 1 H), 0.58–0.49 (m, 2 H), 0.37–0.30
(m, 1 H); LCMS (2 min; ESI) RT 1.29 min; m/z: calcd for C16H16F2N3O4+ [M + H]+, 352.1103; found,
352.1105. Due to the presence of the minor enantiomer in starting
material 31, a solution of this product (est. 90% ee)
at 25 mg/mL in DCM/methanol (9:1) was purified by SFC by Reach Separations,
Nottingham, using Chiralpak SA, 30:70% methanol/CO2, 0.2%
v/v NH3. Combined fractions of the major product were dissolved
in DCM and heptane, evaporated, and dried in a vacuum oven to give 34a as a yellow solid (1.6 g, 67%). Chiral purity was assessed
by SFC (YMA Amylose-C, 30:70 MeOH/CO2; 0.2% v/v NH3), which showed 99% ee. Fractions containing the minor isomer 34b were repurified by the same method, affording this compound
as an orange solid (143 mg, 6%). Chiral purity was assessed by SFC
(as above), which showed 98% ee. Compounds 34a and 34b were used without further purification in the preparation
of 1 and 14.
Authors: Alexandru F Ghetu; Connie M Corcoran; Leandro Cerchietti; Vivian J Bardwell; Ari Melnick; Gilbert G Privé Journal: Mol Cell Date: 2008-02-15 Impact factor: 17.970
Authors: Matthew G Lloyd; Rosemary Huckvale; Kwai-Ming J Cheung; Matthew J Rodrigues; Gavin W Collie; Olivier A Pierrat; Mahad Gatti Iou; Michael Carter; Owen A Davis; P Craig McAndrew; Emma Gunnell; Yann-Vaï Le Bihan; Rachel Talbot; Alan T Henley; Louise D Johnson; Angela Hayes; Michael D Bright; Florence I Raynaud; Mirco Meniconi; Rosemary Burke; Rob L M van Montfort; Olivia W Rossanese; Benjamin R Bellenie; Swen Hoelder Journal: J Med Chem Date: 2021-11-30 Impact factor: 7.446
Authors: Rosemary Huckvale; Alice C Harnden; Kwai-Ming J Cheung; Olivier A Pierrat; Rachel Talbot; Gary M Box; Alan T Henley; Alexis K de Haven Brandon; Albert E Hallsworth; Michael D Bright; Hafize Aysin Akpinar; Daniel S J Miller; Dalia Tarantino; Sharon Gowan; Angela Hayes; Emma A Gunnell; Alfie Brennan; Owen A Davis; Louise D Johnson; Selby de Klerk; Craig McAndrew; Yann-Vaï Le Bihan; Mirco Meniconi; Rosemary Burke; Vladimir Kirkin; Rob L M van Montfort; Florence I Raynaud; Olivia W Rossanese; Benjamin R Bellenie; Swen Hoelder Journal: J Med Chem Date: 2022-06-02 Impact factor: 8.039