Targeting the gatekeeper residue in phosphoinositide 3-kinases

Article abstract of DOI:10.1016/j.bmc.2005.02.021

A single residue in the ATP binding pocket of protein kinases-termed the gatekeeper-has been shown to control sensitivity to a wide range of small molecule inhibitors (Chem. Biol. 2004, 11, 691; Chem. Biol. 1999, 6, 671). Kinases that possess a small side chain at this position (Thr, Ala, or Gly) are readily targeted by structurally diverse classes of inhibitors, whereas kinases that possess a larger residue at this position are broadly resistant. Recently, lipid kinases of the phosphoinositide 3-kinase (PI3-K) family have become the focus of intense research interest as potential drug targets (Chem. Biol. 2003, 10, 207; Curr. Opin. Pharmacol. 2003, 3, 426). In this study, we identify the residue that corresponds structurally to the gatekeeper in PI3-Ks, and explore its importance in controlling enzyme activity and small molecule sensitivity. Isoleucine 848 of p110α was mutated to alanine and glycine, but the mutated kinase was found to have severely impaired enzymatic activity. A structural bioinformatic comparison of this kinase with its yeast orthologs identified second site mutations that rescued the enzymatic activity of the I848A kinase. To probe the dimensions of the gatekeeper pocket, a focused panel of analogs of the PI3-K inhibitor LY294002 was synthesized and its activity against gatekeeper mutated and wild-type p110α was assessed.

Full text of DOI:10.1016/j.bmc.2005.02.021

Bioorganic & Medicinal Chemistry 13 (2005) 2825–2836
Targeting the gatekeeper residue in phosphoinositide 3-kinases
Peter J. Alaimo,a, Zachary A. Knightb, and Kevan M. Shokata,c,
*
aDepartment of Cellular and Molecular Pharmacology, University of California, San Francisco, CA 94143-2280, USA
bProgram in Chemistry and Chemical Biology, University of California, San Francisco, CA 94143, USA
cDepartment of Chemistry, University of California, Berkeley, CA 94720, USA
Received 1 December 2004; accepted 12 February 2005
Abstract—A single residue in the ATP binding pocket of protein kinases—termed the gatekeeper—has been shown to control sen-
sitivity to a wide range of small molecule inhibitors (Chem. Biol. 2004, 11, 691; Chem. Biol. 1999, 6, 671). Kinases that possess a
small side chain at this position (Thr, Ala, or Gly) are readily targeted by structurally diverse classes of inhibitors, whereas kinases
that possess a larger residue at this position are broadly resistant. Recently, lipid kinases of the phosphoinositide 3-kinase (PI3-K)
family have become the focus of intense research interest as potential drug targets (Chem. Biol. 2003, 10, 207; Curr. Opin. Pharmacol.
2003, 3, 426). In this study, we identify the residue that corresponds structurally to the gatekeeper in PI3-Ks, and explore its impor-
tance in controlling enzyme activity and small molecule sensitivity. Isoleucine 848 of p110a was mutated to alanine and glycine, but
the mutated kinase was found to have severely impaired enzymatic activity. A structural bioinformatic comparison of this kinase
with its yeast orthologs identified second site mutations that rescued the enzymatic activity of the I848A kinase. To probe the dimen-
sions of the gatekeeper pocket, a focused panel of analogs of the PI3-K inhibitor LY294002 was synthesized and its activity against
gatekeeper mutated and wild-type p110a was assessed.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
pocket (corresponding to threonine 338 in Hck) has
been shown to control kinase sensitivity to a wide
range of structurally unrelated compounds,1 including
pyridinylimidazoles,8 pyrazolopyrimidines,2 purines,9
quinazolines,10 phenylaminopyrimidines,11 and stau-
rosporines.12 This residue is conserved as a threonine
or larger amino acid in the human kinome (no wild-type
protein kinases contain an alanine or glycine at this po-
sition), and structural analysis has shown that the size of
this gatekeeper residue restricts access to a pre-existing
cavity within the ATP binding pocket.13 Kinases that
possess a threonine at this position are readily targeted
by diverse classes of small molecule inhibitors that can
access this natural pocket. Moreover, mutation of the
gatekeeper residue to a smaller amino acid, such as ala-
nine or glycine, has been shown to induce sensitivity to
pyrazolopyrimidine inhibitors at low nanomolar con-
centrations in over 30 protein kinases—even though in
many cases the wild-type kinase is completely insensitive
to compounds of this class.12
In the past decade protein kinases have emerged as one
of the most important new classes of drug targets. Pro-
tein kinases play a central role in many signaling path-
ways disregulated in disease, and these enzymes can be
readily targeted with cell permeable, small molecule
inhibitors.5 These facts have led to the hope that inhib-
itors of individual protein kinases might be tailored to
specific diseases based on an understanding of their
molecular etiology.6 Recently, this concept has been
dramatically validated by the clinical success of Gleevec,
an inhibitor of the Abl tyrosine kinase,7 in the treatment
of chronic mylogenous leukemia, a disease driven by the
activity of the BCR-ABL oncogene.
The search for protein kinase inhibitors has led to the
realization that not all kinases are equally amenable to
targeting with potent, ATP-competitive small molecules.
In this regard, a single residue in the ATP binding
The importance of the gatekeeper in controlling inhibi-
tor sensitivity is underscored by the fact that most ki-
nase inhibitors currently in clinical use target kinases
that contain a threonine at this position, even though
threonine is found in only ꢀ20% of the human kinome
Keywords: PI3-kinase; Inhibition; Gatekeeper.
*
Corresponding author. Tel.: +1 415 514 0472; fax: +1 415 514
0822; e-mail: shokat@cmp.ucsf.edu
These authors contributed equally to this work.
0968-0896/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.02.021

2826
P. J. Alaimo et al. / Bioorg. Med. Chem. 13 (2005) 2825–2836
(e.g., Iressa: EGFR; Gleevec: Abl, PDGFR, and c-Kit;
BAY43-9006: Raf). Indeed, analysis of mutations in
BCR-Abl that confer drug resistance has shown that
mutation of the gatekeeper to a larger amino acid
(T315I) is one of the most common mechanisms of resis-
tance to Gleevec.11 Remarkably, second generation
BCR-Abl inhibitors designed to target resistant alleles
have been shown to effectively inhibit every naturally
occurring mutant of this kinase except those mutated
at the gatekeeper residue.14,15
the phosphoacceptor. In protein kinases, conserved
hydrogen bonds to the N6 and N1 of positions of the
adenine ring of ATP are made by the backbone amides
of two conserved residues (corresponding to Glu 339
and Met 341 of Hck, respectively). The gatekeeper resi-
due immediately precedes Glu 339, and forms the hydro-
phobic interior face of the ATP binding pocket, with Cb
˚
of the gatekeeper typically positioned 4–7 A from the
N6 and N7 residues of adenine. Inspection of the ATP
binding pocket of p110c reveals a similar set of contacts.
Hydrogen bonds to ATP are made by backbone amides
from two residues—Glu 880 and Val 882—and these
residues are immediately preceded by a large hydropho-
bic residue (Ile 879 in p110c) that is positioned approx-
Recently, lipid kinases of the phosphoinositide 3-kinase
(PI3-K) family have attracted considerable interest as a
new class of drug targets.3,4 These enzymes act by gener-
ating the lipid second messengers phosphatidylinositol-
3,4-bisphosphate (PI(3,4)P2) and phosphatidylinositol-
3,4,5-trisphosphate (PI(3,4,5)P3), which in turn activate
downstream enzymes in a wide-range of signaling path-
ways involved in cell growth, survival, differentiation,
and motility.16 Activating mutations in the PI3-K
isoform p110a have recently been identified at high
frequency in several types of cancer,17 and PTEN, the
lipid phosphatase that reverses the phosphorylation
reaction, has been identified as one of the most com-
monly inactivated tumor suppressors in the human gen-
ome.18 Moreover, the clinical efficacy of recently
approved agents that target the epidermal growth factor
receptor in breast19 and lung20,21 cancers has been dem-
onstrated to correlate with the dependence of those can-
cers on aberrant PI3-K signaling and the ability of these
agents to suppress that pathway. For these reasons, con-
siderable effort has been directed toward the develop-
ment of selective inhibitors of these enzymes as
potential cancer therapeutics.22
˚
imately 5 A from N6 and N7 of adenine (Fig. 1A). To
directly compare the orientation of these two residues
(Ile 879 in p110c and Thr 338 in Hck), we superimposed
the crystal structures of Hck and p110c with respect to
the adenine ring of ATP using the WebLab Viewer
4.0. This structural alignment indicates that these two
residues occupy a largely overlapping, but not identical,
space within the interior of the ATP binding pocket,
forming the innermost face that contacts the adenine
ring of ATP (Fig. 1B). The most important difference
between the p110c and Hck structures in this region is
that Ca of Ile 879 in p110c is shifted slightly perpendic-
˚
ular to the plane of the adenine ring (ꢀ2 A) relative to
Thr 338 in Hck, such that Ile 879 also participates in
forming the interior roof of the ATP binding pocket,
whereas Thr 338 is more completely fixed within the
plane of the adenine ring.
To compare this region of primary sequence between
PI3-Ks and protein kinases, a structure-based sequence
alignment was generated. This analysis reveals that the
Although PI3-Ks possess very low overall sequence
homology to protein kinases, they possess the same
overall fold, share several consensus sequences (e.g.,
the DFG motif that is responsible for coordination of
Mg2+), and are sensitive to two pan-specific protein ki-
nase inhibitors (staurosporine23 and quercetin24). Due
to the importance of these enzymes as an emerging class
of drug targets, we sought to identify the residue in lipid
kinases analogous to the gatekeeper residue, and to ex-
plore how the size of this residue affects the active site
structure and inhibitor sensitivity of this class of
enzymes.
2. Results and discussion
2.1. The gatekeeper residue is conserved in lipid kinases
The crystal structure of the PI3-K p110c bound to ATP
has been solved,25 and the catalytic domain was found
to share several features with reported protein kinase
structures. These similarities include a two-lobed struc-
ture consisting of a N-terminal lobe containing a 5–7-
stranded b-sheet, a loop connecting two of these strands
that interacts with the phosphate groups of ATP, a con-
served lysine residue that positions the a and b phos-
phate groups of ATP for the phosphotransfer reaction,
and a primarily a-helical C-terminal lobe that binds
Figure 1. Conservation in the ATP binding pocket of PI3Ks. (A)
Schematic illustrating the hydrogen bonding contacts to ATP. (B)
Surface representation of the interior face of the ATP binding pocket
of Hck (gray) and p110c (orange), from published crystal structure
data. (C) Structure-based sequence analysis of several residues lining
the ATP binding pockets of Hck and several PI3Ks.

P. J. Alaimo et al. / Bioorg. Med. Chem. 13 (2005) 2825–2836
2827
gatekeeper residue is situated within a small region of
sequence homology between lipid and protein kinases
(Fig. 1C). As observed for the gatekeeper residue in pro-
tein kinases, Ile 879 is conserved as a large hydrophobic
residue in all PI3-Ks (isoleucine, leucine, methionine, or
valine, although no PI3-Ks contain threonine at this po-
sition). The residue immediately following the Ile 879 is
conserved in PI3-Ks as a glutamate (83%), and this is
also the most common residue at that position in protein
kinases (75%). The residues at the ꢁ1 and +1/+2 posi-
tions relative to Ile 879 in PI3-Ks are conserved as large
hydrophobic residues, and the corresponding positions
in protein kinases show a similar preference (Fig. 1C).
For example, position 337 is either isoleucine, leucine,
valine, or methionine in 86% of protein kinases and in
92% of human PI3-Ks. On this basis, we conclude that
p110c residue Ile 879 structurally corresponds to the
gatekeeper residue in protein kinases.
alleles encode active kinases. The activity of VPS34 is
required26 for growth of yeast at elevated temperature
(37 ꢁC) or high salt (1.0 M NaCl), whereas MEC1 activ-
ity is required27,28 for growth in the presence of agents
that alkylate DNA (0.02% methylmethanesulfonate)
or inhibit ribonucleotide reductase (60 mM hydroxyl-
urea). Growth of the mutant strains under these condi-
tions showed that the alanine gatekeeper mutant of
VPS34 (I670A) as well as the alanine and glycine gate-
keeper mutants of MEC1 (L2129A and L2129G) are
able to functionally complement for the knockout at a
level comparable to the wild-type enzyme (Fig. 2).
Importantly, transformation with the empty vector or
a vector containing a catalytically inactive, kinase-dead
(kd) mutant of each protein did not to rescue the
knockout phenotype, confirming that the observed
complementation is due to the catalytic activity of these
proteins (Fig. 2).
As the gatekeeper residue is structurally conserved be-
tween PI3-Ks and protein kinases, we asked whether it
is also functionally conserved, and in particular whether
mutagenesis of this position to a smaller residue might
induce inhibitor sensitivity as has been observed for pro-
tein kinases.12 Three representative members of the PI3-
K family were selected for this analysis: the yeast PI3-K
VPS34, the yeast PI3-K-related protein kinase MEC1,
and the prototypical mammalian PI3-K p110a. The
gatekeeper residue in each of these kinases was mutated
to alanine and glycine, and the effect of this mutation on
enzyme activity and inhibitor sensitivity was assessed.
For p110a, the gatekeeper residue was mutated to ala-
nine and glycine (I848A and I848G) and the mutant ki-
nases were expressed by transient transfection in cos-1
cells. The myc-tagged mutated kinases were purified by
immunoprecipitation and their lipid kinase activity
was assessed using a kinase assay in vitro. The enzy-
matic activity of the I848A and I848G mutants was sig-
nificantly impaired relative to wild-type p110a, with the
gatekeeper mutants displaying approximately one-
hundredth (I848A) to one-thousandth (I848G) of the
wild-type activity (Fig. 3A). The expression level of the
wild-type and I848A proteins was similar by western
blotting (Fig. 3B), suggesting that the gatekeeper muta-
tion had disrupted the integrity of the enzyme active site
in this mutant without globally destabilizing the protein.
2.2. Effects of gatekeeper mutation on enzymatic activity
Plasmids encoding mutated VPS34 and MEC1 kinases
were transformed into S. cerevisiae knockout strains,
such that the mutant allele functionally replaces the
wild-type enzyme. Although yeast are viable in the ab-
sence of these proteins, specific growth conditions can
induce a requirement for their catalytic activity, allow-
ing us to examine whether the gatekeeper mutated
As VPS34 and MEC1 appear to tolerate the gatekeeper
mutation, the impaired activity of I848A p110a was sur-
prising, and led us to search for differences in primary
sequence between these related kinases that might
account for the observed differences in biochemical
activity. Based on the crystal structure of p110c, the
Figure 2. Growth of yeast expressing various alleles of VPS34 and MEC1 under nonselective and selective growth conditions. S. cerevisiae expressing
VPS34 alleles (wild type, empty vector, kinase dead (kd; D749E), I670G, or I670A) were plated on nonselective media (SD(URA)) or selective media
(SD(URA), 1 M NaCl). S. cerevisiae expressing MEC1 alleles (wild type, empty vector, kinase dead (kd; D2243E), L2129G, or L2129A) were plated
on media (SD(URA)) or selective media (SD(URA), 0.02% (MMS); or SD(URA), 60 mM hydroxyurea (HU)).

2828
P. J. Alaimo et al. / Bioorg. Med. Chem. 13 (2005) 2825–2836
the context of second site mutations C838V, C838I,
and C838L. The double mutant enzymes were expressed
in cos-1 cells, and their expression levels and enzymatic
activity was assessed. The I848A/C838V and I848A/
C838I double mutants both showed ꢀ10-fold increased
enzymatic activity relative to the I848A single mutant
enzyme (Fig. 3A). This enhancement in activity occurred
without significantly changing the expression level of the
protein (Fig. 3B), and confirms that mutation of this
position to a b-branched residue can restore enzymatic
activity in the background of the gatekeeper mutation.
Strikingly, the I848A/C838L double mutant enzyme
showed virtually undetectable catalytic activity and
expression of this protein could not be detected by wes-
tern blotting (Fig. 3), indicating that the isomeric I848A/
C838I and I848A/C838L p110a proteins have dramati-
cally different stability. This suggests that the enhanced
activity of the isoleucine and valine mutants is likely a
specific consequence of the b-branched residue at posi-
tion 838 stabilizing b-strands 6 and 7, rather than a
nonspecific consequence of additional hydrophobic sur-
face in the protein core.
2.3. Effects of gatekeeper mutation on inhibitor sensitivity
We next sought to assess how mutation of the gate-
keeper residue to a smaller amino acid would affect
inhibitor sensitivity. In protein kinases, mutation of
the gatekeeper to alanine or glycine has been shown to
induce sensitivity to both pyrazolopyrimidine inhibitors
based on the Src-family kinase inhibitor PP1 and ana-
logs of the natural product K252a.12 p110c has been
shown to be sensitive to staurosporine (Kd = 0.29 lM),23
which is structurally related to K252a, suggesting that
K252a analogs may also target engineered lipid kinases.
Although wild-type PI3-Ks are not sensitive to the pro-
tein kinase inhibitor PP1, mutation of the gatekeeper
residue in protein kinases can induce sensitivity to this
class of compounds in kinases that otherwise show no
affinity for this scaffold.12 Following this reasoning, we
screened a small panel of these compounds in vitro
against the wild-type and I848A p110a and in vivo
against the wild-type and I670A mutant VPS34 (Supple-
mentary Fig. 1). None of these compounds showed
selective inhibition of the I848A allele of p110a at a con-
centration of 50 lM in vitro or I670A VPS34 at a con-
centration of 1 mM in a yeast halo assay. These results
suggest that the structural differences between protein
and lipid kinases may be too significant to bridge with
a single inhibitor scaffold, and that new inhibitor ana-
logs may be necessary to explore the gatekeeper pocket
in lipid kinases.
Figure 3. Relative expression levels and enzymatic activities of Myc-
tagged p110a alleles. (A) Equal amounts of soluble protein (29 lg)
from transiently transfected cos-1 cells were subjected to Western
analysis (anti-Myc antibody, 9E10). (B) Relative enzymatic activities
of p110a alleles. Cos-1 cells were co-transfected with equal amounts of
expression vectors containing the indicated genes and p85 (a p110a
adaptor protein). Protein was immunoprecipitated (anti-Myc anti-
body) from 950 lg of total soluble protein, and the activity assay was
performed. Error bars represent the standard deviation from the mean
from three independent measurements.
residues that form the core of the ATP binding pocket
were identified, and these residues were aligned among
15 members of the PI3-K family (Fig. 4). We then iden-
tified residues that are common to VPS34 and MEC1,
but different in p110a, that might account for the ob-
served difference in tolerating the gatekeeper mutation.
This analysis focused our attention on cysteine 838 in
p110a (Position 13, Fig. 4). Among the class I PI3-Ks
such as p110a, this residue is conserved as a cysteine,
whereas among the class III PI3Ks (e.g., VPS34) and
PI3-K-related protein kinases (e.g., MEC1), this residue
is conserved as a b-branched residue (valine or isoleu-
cine). Importantly, inspection of the crystal structure
of p110c reveals that C838 is located in b-strand 6, di-
rectly adjacent to I848 in b-strand 7 (Fig. 5), suggesting
that it may cooperate with the gatekeeper residue to sta-
bilize this region of the protein. As b-branched amino
acids have been shown to promote b-sheet formation,29
we reasoned that the presence of an isoleucine or valine
residue at this position might stabilize that region of the
protein when the gatekeeper is mutated, and thereby ac-
count for the differences in activity between MEC1/
VPS34 and p110a.
To probe more directly the engineered gatekeeper pock-
et, we prepared a panel of analogs of the PI3-K inhibitor
LY294002.30 LY294002 reversibly inhibits PI3-Ks at
IC50 values in the low micromolar range, but shows little
selectivity among individual family members.31 The
crystal structure of LY294002 bound to p110c has been
solved,23 and reveals that the C3 position of LY294002
is located adjacent to the gatekeeper residue at a dis-
˚
To test this hypothesis, we prepared double mutants of
p110a containing the I848A gatekeeper mutation in
tance of approximately 4 A. Molecular modeling based
on this structure suggests that analogs of LY294002

P. J. Alaimo et al. / Bioorg. Med. Chem. 13 (2005) 2825–2836
2829
Figure 4. Structure-based sequence alignment of 15 members of the PI3K family. Residue coloring: hydrophobic aliphatic (green), hydrophobic
aromatic (light blue), small (yellow), polar charged (dark blue), basic (purple), and acidic (orange).
containing extended substituents at C3 would access the
nascent pocket created by the gatekeeper mutation (Fig.
5). A series of C3-substituted analogs of LY294002 were
designed (lacking the 8-phenyl group of LY294002, as
this moiety has been shown to be dispensable for bind-
ing to PI3-Ks30) and a panel of such analogs was pre-
pared. This series of compounds was designed to
include analogs that possess C3 substituents similar in
size to the space created directly by the isoleucine to ala-
nine mutation (e.g., Et, n-Pr), as well as those that con-
tain much larger substituents (Ph, Bn). These latter
compounds were included to probe whether the gate-
keeper mutation would allow access to a deeper cavity
within the kinase active site; in protein kinases, mutation
of the gatekeeper has been shown to facilitate binding of
substrate and inhibitor analogs containing bulky sub-
stituents much larger than the space created directly
by the amino acid change.12,13
was modestly reduced for the double mutant. As the size
of the C3-moiety was increased, IC50 values against both
the mutant and wild-type p110a increased by ꢀ10 to
>100-fold, and this increase tracked with the size of
the C3 substituent, with the weakest inhibition by com-
pounds 3g (Ph) and 3h (Bn). Surprisingly, only com-
pound 3c (Et) showed enhanced binding to mutant
p110a relative to wild-type p110a, with modest selectiv-
ity for the engineered kinase (ꢀ6-fold). This compound
contains an ethyl moiety at the C3 position, which is the
most similar in size to the space created directly by the
amino acid change (Ile to Ala) at the gatekeeper posi-
tion. This suggests that mutation of this residue does
create a corresponding cavity in the kinase active site,
but that this cavity does not expose a preexisting pocket
or allow for the binding of significantly larger inhibitor
analogs with enhanced affinity, as has been observed in
the protein kinase family.
Synthesis was accomplished by addition of an excess of
the appropriate Grignard reagent to salicylaldehyde to
afford benzylic alcohols 1c–g (Scheme 1). The resulting
alcohols were oxidized to the corresponding ketones
2c–g using MnO2, and the ring closing was performed
using a morpholine phosgeniminium salt30,34 to afford
the desired 3-substituted analogs of LY294002 in high
purity as white solids.
3. Conclusions
We have explored the role of the gatekeeper residue in
PI3-Ks by a convergent engineering approach that com-
bines mutagenesis of the target residue with design of
inhibitor analogs to complement this mutation. This
study suggests that mutation of the gatekeeper residue
in lipid kinases can create a nonnatural pocket, but that
this mutation does not provide access to a deeper pocket
such as that found in protein kinases. This difference
may reflect the different way that protein and lipid
kinases utilize their primary sequence to construct the
innermost wall of the ATP binding pocket. Crystal
structures of protein kinases reveal that residues from
This panel of LY294002 analogs was tested for inhibi-
tion of the wild-type and C838V/I848A mutant of
p110a in a PI3-K assay in vitro (Table 1). Both
LY294002 and the desphenyl analog LY292223 inhib-
ited wild-type and C838V/I848A p110a at low micromo-
lar concentrations, although inhibition by LY292223

2830
P. J. Alaimo et al. / Bioorg. Med. Chem. 13 (2005) 2825–2836
Table 1. IC50 values of LY292223 analogues against wild type and
mutant p110a enzymes
Analogue
IC50 (lM)
Wild type
C838V I848A
LY294002
LY292223 (3a)
3b (Me)
3c (Et)
1.1
2.6
1.1
15
33
28
40
4.4
3d (Pr)
31
51
68
>200
48
3e (i-Pr)
3f (Bu)
3g (Ph)
25
>200
>200
>200
>200
3h (Bn)
from b-strand 6 (which is otherwise positioned in a sec-
ond sphere of residues that do not contact ATP directly)
infiltrates the ATP binding pocket to make a direct con-
tact with adenine near N6. The analogous residue in
protein kinases (corresponding to Leu 325 in Hck) is
positioned within a second sphere that is obstructed
from accessing ATP by the gatekeeper. Thus, it appears
that in lipid kinases two residues (Tyr 867 and Ile 879)
collaborate to fill the space that is otherwise occupied
by a single residue in protein kinases. Unfortunately,
we have found that mutation of Tyr 836 in p110a (which
corresponds to Tyr 867 in p110c) to any other amino
acid tested (glycine, alanine, threonine, aspartic acid,
leucine, methionine, or histidine) results in a complete
loss of catalytic activity, indicating that this region of
the protein is not amenable to structural modification
(Z.A.K. and K.M.S., unpublished observations). These
results, as well as the nearly perfect conservation of
Tyr 867 within the PI3-K family (Fig. 4), suggest that
a different set of residues in lipid kinases are likely
responsible for controlling sensitivity to small molecule
inhibitors. Ultimately, the identification of rules guiding
inhibitor sensitivity for this important family of enzymes
will require the discovery of new structural classes of
PI3-K inhibitors and the broad characterization of their
specificity against individual PI3-K isoforms, a process
that is currently underway in our laboratory31 and
others.32
Figure 5. Co-crystal structure of p110c-LY294002. (A) Surface repre-
sentation and (B) ribbon representation of the protein illustrating the
important hydrogen-bonding contacts and residues mutated in this
study.
4. Experimental
4.1. Protein expression
Mutations were introduced by Quikchange (Stratagene),
and confirmed using standard dideoxy-based sequenc-
ing. Myc-tagged p110a was expressed by transient trans-
fection of cos-1 cells. Cells were lysed in lysis buffer
(50 mM Tris (pH 7.4), 300 mM NaCl, 5 mM EDTA,
0.02% NaN3, 1% Triton X-100), protease inhibitors
(protease inhibitor cocktail tablets (Roche); sodium
orthovanadate (8 mM); PMSF (83 lM)), and 8 mM
DTT. The kinase was immunoprecipitated using a Pro-
tein-G-9E10 antibody complex, and washed twice with
buffer A (PBS, 1 mM EDTA, 1% Triton X-100), twice
with buffer B (100 mM Tris (pH 7.4), 500 mM LiCl,
1 mM EDTA), twice with buffer C (50 mM Tris
(pH 7.4), 100 mM NaCl), and once with PBS. In control
Scheme 1. Reagents and conditions: (i) Grignard reagent, THF;
(ii) MnO2, CH2Cl2; (iii) TiCl4, i-Pr2EtN, CH2Cl2; (iv) 4-dichloro-
methylenemorpholin-4-ium chloride.
b-strand 5, which include the gatekeeper, form the deep-
est face of the ATP binding pocket, and that most of the
contacts that this face makes with adenine involve the
side chain of the gatekeeper residue directly. By compar-
ison, in the crystal structure of p110c, the gatekeeper is
shifted upward and away from the N6 of adenine by
˚
ꢀ2 A. To help fill this space, the side chain of Tyr 867

P. J. Alaimo et al. / Bioorg. Med. Chem. 13 (2005) 2825–2836
2831
experiments, no differences in inhibitor sensitivity were
observed between wild type p110a protein that was
obtained from transfected cos-1 cells, Sf9 cells using a
baculovirus system, or commercially available recombi-
nant protein (Jena Bioscience).
selected on SD-URA media. Cells were streaked onto
plates containing SD-URA or the same media supple-
mented with either 1 M NaCl (VPS34) or 0.02% MMS
or 60 mM HU (MEC1), grown at 30 or 37 ꢁC for 3 days,
and photographed.
For SDS-PAGE and Western blot analyses, protein
concentrations in cell lysates were determined using a
Bradford assay. Proteins were loaded onto a Tris-gly-
cine gel (8–16% gradient; Gradipore) and separated by
SDS-PAGE before being transferred onto a nitrocellu-
lose membrane. The blot was treated with blocking
reagent (5% dry milk in TBST) for 1 h, then primary
antibody (9E10 (anti-Myc; Santa Cruz Biotech), 1:500
in TBST) overnight at 4 ꢁC. The blot was then rinsed
(5 min each) with deionized water (1 · 20 mL) and
TBST (3 · 20 mL) before treatment with secondary
anti-mouse-HRP antibody (1:1000 in TBST) for
30 min at rt.
4.4. Chemical synthesis
4.4.1. General methods. All reactions were performed
under argon in oven- or flame-dried glassware fitted
with rubber septa, and were stirred magnetically. Thin
layer chromatography was performed on Merck pre-
coated silica gel F-254 plates (0.25 mm). Flash column
chromatography was performed using Merck silica gel
60 (230–400 mesh). Proton NMR spectra were recorded
at 400 MHz and are reported in d (ppm) as s (singlet), d
(doublet), t (triplet), q (quartet), m (multiplet), or br
(broad), and are referenced to the residual solvent sig-
nal: CDCl3 (7.26) or C6D6 (7.15). Carbon NMR spectra
were recorded at 100 MHz and are reported in d (ppm),
and are referenced to the solvent signal: CDCl3 (77.0),
C6D6 (128.0). Infrared spectra were recorded on a Nic-
olet Impact 400 spectrometer using thin films of sample
and resonances are reported in wavenumbers (cmꢁ1).
Mass spectra were recorded on a VG-70S mass spec-
trometer and are reported in units of mass/charge
(m/z). Unless otherwise noted, all materials were
obtained from commercial sources and used without
further purification. In experiments involving air- or
moisture-sensitive compounds, solvents and reagents
were either distilled or purchased as anhydrous grade
material. Dichloromethylenemorpholin-4-ium chloride
was prepared according to literature procedures.33,34
4.2. PI3-Kinase enzymatic assay
The PI3-K assay was performed essentially as de-
scribed.31 Briefly, a mixture of kinase, inhibitor, buffer
(25 mM HEPES (pH 7.4), 10 mM MgCl2), and freshly
sonicated phosphatidylinositol (200 lg/mL) was pre-
pared at 4 ꢁC, and aliquotted into eppendorf tubes.
The tubes were allowed to warm to rt over 5 min, and
the enzymatic reaction was initiated with the addition
of ATP (10 lCi c-32P-ATP; final [ATP] = 20 nM). Reac-
tions were incubated for 20 min at rt, and quenched by
addition of 1 M HCl (105 lL) followed by 1:1 MeOH–
CHCl3 (160 lL). The resulting biphasic mixture was
vortexed (ca. 5 s), briefly centrifuged (ca. 5 s), and the
organic phase (ꢀ100 lL) was transferred to a new tube
using a gel-loading tip pre-coated with CHCl3. This ex-
tract was spotted on TCL plates (silica gel 60 F254,
250 lm) and developed for 3–4 h using 65:35 1-propa-
nol–2 M AcOH as eluent. The TLC plates were then
air-dried (ca. 45 min), and the reaction products quanti-
tated using a phosphorimager (Molecular Dynamics).
For each compound, kinase activity was measured at
15 inhibitor concentrations representing twofold dilu-
tions from the highest concentration tested (usually
400 lM). For compounds showing significant activity,
IC50 value determinations were repeated (2–5 times)
and the value reported is the average of these indepen-
dent measurements. Preincubation of the enzyme with
cold ATP (1 lM, 10 min) was sometimes used to sup-
press the observation of radiolabeled impurities that
were kinase-independent (data not shown). In control
experiments, no difference was observed in inhibitor sen-
sitivity between performing the assay in the absence or
presence (1 lM) of cold ATP.
4.4.2. General procedure for the preparation of alcohols
1c–g. A round bottom flask charged with salicylalde-
hyde and THF was cooled to ꢁ78 ꢁC. To this mixture
was added the appropriate Grignard reagent (5–
10 equiv) by syringe over 10 min, and the resulting
mixture was allowed to stir at ꢁ78 ꢁC for 10 min. The
stirred reaction mixture was allowed to warm to rt
overnight. In air, this mixture was slowly added to a
0 ꢁC saturated NH4Cl solution. The aqueous phase
was extracted with ether and the combined organic
phases were dried over Na2SO4, filtered, and concen-
trated to give a crude oil. Further purification and char-
acterization of each derivative is described below.
4.4.2.1. 2-(1-Hydroxybutyl)phenol (1c). Synthesis was
performed using the general procedure described above
using 1.15 g salicylaldehyde (9.4 mmol), 94 mL THF,
94 mmol propylmagnesium chloride (47 mL of 2.0 M
solution in ether), saturated NH4Cl solution (75 mL),
and ether (3 · 40 mL). The crude oil was purified by
flash column chromatography (SiO2, 10% EtOAc–hex-
anes) to give a clear, pale yellow oil (1.226 g, 7.4 mmol,
79%). 1H NMR: (CDCl3) d 8.16 (s, 1H), 7.12 (t,
J = 8 Hz, 1H), 6.91 (d, J = 8 Hz, 1H), 6.82 (m, 2H),
4.76 (t, J = 7 Hz, 1H), 3.35 (s, 1H), 1.83 (m, 1H), 1.74
(m, 1H), 1.43 (m, 1H), 1.30 (m, 1H), 0.91 (t, J = 8 Hz,
3H). 13C{1H} NMR: (CDCl3) d 155.2, 128.7, 127.8,
127.2, 119.7, 116.9, 75.5, 39.3, 18.9, 13.8. IR: 3321.
HRMS: (EI) calcd for C10H14O2 (M+) 166.0994; found
4.3. Analysis of yeast expressing PI3-K mutant alleles
The VPS34 S. cerevisiae knockout strain 15149 was ob-
tained from the ATCC. The MEC1/SML1 S. cerevisiae
knockout strain yBRT7-4b was a gift from David Toc-
zyski (UCSF). Knockout strains were transformed with
plasmids (pRS416 CEN URA3) encoding wild-type and
mutant VPS34 and MEC1 and transformants were

2832
P. J. Alaimo et al. / Bioorg. Med. Chem. 13 (2005) 2825–2836
166.0995. Anal. Calcd for C10H14O2: C, 72.26; H, 8.49.
Found: C, 72.19; H, 8.41.
of 2.0 M solution in THF), saturated NH4Cl (75 mL),
and ether (3 · 50 mL). The crude oil was purified by
flash column chromatography (SiO2, 50% EtOAc–hex-
anes) to give a colorless solid (1.42 g, 6.6 mmol, 75%).
1H NMR: (CDCl3) d 8.01 (s, 1H), 7.34–7.14 (m, 5H),
6.93–6.79 (m, 4H), 4.98 (dt, J = 3 Hz, J = 7 Hz, 1H),
3.10 (d, J = 7 Hz, 2H), 2.81 (d, J = 3 Hz, 1H). Lit.37
(partial) d 5.02 (t, J = 7 Hz, 1H), 3.11 (d, J = 7 Hz,
2H). 13C{1H} NMR: (CDCl3) d 151.6, 133.6, 125.8,
125.3, 125.0, 123.5, 123.3, 122.8, 116.1, 113.5, 73.0,
40.4. IR: 3327. HRMS: (EI) 214.0993 calcd for
C14H14O2 (M+); found 214.1003.
4.4.2.2. 2-(1-Hydroxypentyl)phenol (1d).35,36 Synthe-
sis was performed using the general procedure described
above using 2.29 g salicylaldehyde, 100 mL THF,
94 mmol of butylmagnesium chloride (47 mL of 2.0 M
solution in THF), saturated NH4Cl solution (75 mL),
and ether (4 · 50 mL). The crude oil was purified by
flash column chromatography (SiO2, 10% EtOAc–hex-
anes) to give a clear colorless oil (3.37 g, 18.7 mmol,
1
99%). H NMR: (CDCl3) d 8.32 (s, 1H), 7.12 (m, 1H),
6.93 (m, 1H), 6.82 (m, 2H), 4.74 (t, J = 7 Hz, 1H), 3.84
(br s, 1H), 1.90–1.71 (m, 2H), 1.44–1.22 (m, 4H), 0.90
(t, J = 7 Hz, 3H). Lit.35 d 7.93 (s, 1H), 7.4–6.7 (m,
4H), 4.82 (t, J = 5.6 Hz, 1H), 2.53 (br, 1H), 2.0–1.1
(m, 6H), 0.90 (t, J = 5.9 Hz, 3H). 13C{1H} NMR:
(CDCl3) d 155.2, 128.7, 128.2, 127.0, 119.8, 117.0,
73.8, 46.0, 24.5, 23.1, 22.0. IR: 3357. HRMS: (EI)
180.1150 calcd for C11H16O2 (M+); found 180.1144.
Lit.35 180.1150 calcd for C11H16O2 (M+); found
180.1119.
4.4.3. General procedure for the preparation of ketones
2c–g. A round bottom flask was charged with the appro-
priate diol 1, MnO2, and CH2Cl2, and stirred for 7 h at
rt. The heterogeneous reaction mixture was filtered
through a pad of Celite, and the solids were washed with
500 mL CH2Cl2. The filtrate was concentrated in vacuo
to give a crude oil. The crude oil was purified by flash
column chromatography (SiO2, 10% EtOAc–hexanes)
to give a colorless oil. Characterization of each deriva-
tive is described below.
4.4.2.3. 2-(1-Hydroxy-3-methylbutyl)phenol (1e). Syn-
thesis was performed using the general procedure de-
scribed above using 1.15 g salicylaldehyde, 95 mL
THF, 94 mmol of isobutylmagnesium bromide (47 mL
of 2.0 M solution), saturated NH4Cl (75 mL), and ether
(3 · 50 mL). The crude oil was purified by flash column
chromatography (SiO2, 10% EtOAc–hexanes) to give a
clear, pale yellow oil (1.66 g, 9.2 mmol, 98%). 1H
NMR: (CDCl3) d 8.09 (s, 1H), 7.12 (m, 1H), 6.91 (m,
1H), 6.81 (m, 2H), 4.84 (br d, J = 3 Hz, 1H), 3.22 (s,
1H), 1.81 (m, 1H), 1.70 (m, 1H), 1.56 (m, 1H), 0.93
(m, 6H). Lit.38 (CCl4–C6D6 7:1) d 7.76 (s, 1H), 6.50–
7.20 (m, 4H), 4.50–4.90 (m, 1H), 2.55 (s, 1H), 1.00–
2.00 (m 3H), 0.75–1.00 (m, 6H). 13C{1H} NMR:
(CDCl3) d 155.2, 128.6, 128.1, 126.9, 119.8, 116.9,
73.7, 45.9, 24.5, 23.0, 22.0. IR: 3341. Lit.38 3384 (s, b),
1596 (m), 1384 (m), 1372 (m), 760 (s). HRMS: (EI)
180.1150 calcd for C11H16O2 (M+); found 180.1152.
4.4.3.1. 1-(2-Hydroxyphenyl)butan-1-one (2c). Synthe-
sis was performed using the general procedure described
above using 1c (2.5 g, 15.1 mmol), MnO2 (11.5 g,
132 mmol), and CH2Cl2 (140 mL). Compound 2c was
1
isolated as a colorless oil (1.13 g, 6.9 mmol, 46%). H
NMR: (CDCl3) d 12.38 (s, 1H), 7.68 (dd, J = 1 Hz,
J = 7 Hz, 1H), 7.36 (dt, J = 2 Hz, J = 7 Hz, 1H), 6.90
(dd, J = 1 Hz, J = 7 Hz, 1H), 6.80 (dt, J = 2 Hz,
J = 7 Hz, 1H), 2.87 (t, dt, J = 8 Hz, 2H), 1.70 (sextet,
J = 8 Hz, 2H), 0.96 (t, J = 8 Hz, 3H). Lit.40 d 12.40 (s,
1H), 7.82–6.80 (m, 4H), 2.96 (t, J = 7.4 Hz, 2H), 1.78
(sextet, J = 7.4 Hz, 2H), 1.02 (t, J = 7.4 Hz, 3H).
13C{1H} NMR: (CDCl3) d 206.7, 162.5, 136.1, 129.9,
119.3, 118.8, 118.4, 40.1, 17.8, 13.7. Lit.40 d 206.79,
162.49, 136.17, 130.00, 119.36, 118.82, 118.48, 40.17,
17.90, 13.38. IR: 1641, 1614, 1581, 1488, 1447, 1265,
1202, 1158, 754. Lit.40 1640, 1448, 1226, 1203. HRMS:
(EI) 164.0837 calcd for C10H12O2 (M+); found
164.0840. Lit.40 164. Anal. Calcd for C10H12O2: C,
73.15; H, 7.37. Found: C, 73.22; H, 7.45.
4.4.2.4. 2-(1-Hydroxyhexyl)phenol (1f). Synthesis was
performed using the general procedure described above
using 2.29 g salicylaldehyde, 100 mL THF, 80 mmol of
pentylmagnesium bromide (40 mL of 2.0 M solution),
saturated NH4Cl (75 mL), and ether (3 · 50 mL). The
crude oil was purified by flash column chromatography
(SiO2, 10% EtOAc–hexanes) to give a clear, pale yellow
oil (3.6 g, 18.7 mmol, 99%). 1H NMR: (CDCl3) d 8.44 (s,
1H), 7.12 (m, 1H), 6.96 (m, 1H), 6.83 (m, 2H), 4.76 (br
m, 1H), 4.16 (br s, 1H), 1.91–1.72 (m, 2H), 1.48–1.38 (m,
1H), 1.30 (m, 5H), 0.91 (t, J = 7 Hz, 3H). 13C{1H}
NMR: (CDCl3) d 155.5, 128.5, 127.9, 127.1, 119.9,
116.8, 75.5, 37.1, 31.5, 25.3, 22.5, 13.9. IR: 3348.
HRMS: (EI) 194.1306 calcd for C12H18O2 (M+); found
194.1302. Combustion analysis was not obtained for this
compound due to its similarity to 1c–g.
4.4.3.2. 1-(2-Hydroxyphenyl)pentan-1-one (2d).36
Synthesis was performed using the general procedure de-
scribed above using 1d (2.0 g, 11.1 mmol), MnO2 (9.7 g,
110 mmol), and CH2Cl2 (100 mL). Compound 2d was
isolated as a colorless oil (796 mg, 4.47 mmol, 40%).
1H NMR: (CDCl3) d 12.38 (s, 1H), 7.71 (dd, J = 1 Hz,
J = 8 Hz, 1H), 7.41 (m, 1H), 6.92 (dd, J = 1 Hz,
J = 8 Hz, 1H), 6.84 (m, 1H), 2.93 (t, J = 7 Hz, 2H),
1.68 (m, 2H), 1.38 (m, 2H), 0.93 (t, J = 7 Hz, 3H).
13C{1H} NMR: (CDCl3) d 208.7, 162.3, 135.9, 129.8,
119.0, 118.5, 118.2, 37.8, 26.3, 22.1, 13.6. IR: 1640,
1582, 1487, 1446, 1353, 1249, 119, 1157, 753. HRMS:
(EI) 178.0994 calcd for C11H14O2 (M+); found
178.0996. Combustion analysis was not obtained for this
compound due to its similarity to 2c–g.
4.4.2.5. 2-(1-Hydroxy-2-phenylethyl)phenol (1g). Syn-
thesis was performed using the general procedure de-
scribed above using 1.15 g salicylaldehyde, 95 mL
THF, 94 mmol of benzylmagnesium bromide (47 mL
4.4.3.3. 1-(2-Hydroxyphenyl)-3-methylbutan-1-one
(2e). Synthesis was performed using the general proce-

P. J. Alaimo et al. / Bioorg. Med. Chem. 13 (2005) 2825–2836
2833
dure described above using 1e (2.0 g, 11.1 mmol), MnO2
(10.0 g, 115 mmol), and CH2Cl2 (120 mL). Compound
2e was isolated as a colorless oil (650 mg, 3.65 mmol,
33%). 1H NMR: (CDCl3) d 12.45 (s, 1H), 7.69 (dd,
J = 2 Hz, J = 8 Hz, 1H), 7.38 (dt, J = 2 Hz, J = 8 Hz,
1H), 6.91 (dd, J = 2 Hz, J = 8 Hz, 1H), 6.82 (dt,
J = 2 Hz, J = 8 Hz, 1H), 2.77 (d, J = 7 Hz, 2H), 2.24
Et3N and MeOH were added. The stirred reaction mix-
ture was allowed to warm to rt over 3 h. The volatile
materials were removed in vacuo to give a crude red-col-
ored residue that was washed with NaHCO3 (satd,
30 mL), then the solids were extracted with CH2Cl2
(3 · 15 mL). The combined organics were dried over
MgSO4, filtered, and concentrated in vacuo to give a
crude oil, which was purified by flash column chroma-
tography, and lyophilized from C6H6 to give a flocculent
white powder. Characterization of each derivative is de-
scribed below.
(nonet, J = 6 Hz, 1H), 0.96 (d, J = 6 Hz, 6H). Lit.40
d
12.47 (s, 1H), 7.81–6.82 (m, 4H), 2.85 (d, J = 6.9 Hz,
2H), 1.78 (nonet, J = 6.7 Hz, 1H), 1.02 (d, J = 6.6 Hz,
6H). 13C{1H} NMR: (CDCl3) d 206.4, 162.5, 136.0,
130.0, 119.5, 118.6, 118.3, 46.9, 25.3, 22.5. Lit.40
d
207.2, 163.1, 136.7, 130.6, 120.1, 119.3, 119.0, 47.6,
26.0, 23.2. IR: 1638, 1488, 1447, 1306, 1202, 1158, 753.
Lit.40 1639, 1488, 1447, 1158. HRMS: (EI) 178.0994
calcd for C11H14O2 (M+); found 178.0999. Lit.40 178.
4.4.4.1. 2-Morpholin-4-yl-chromen-4-one (LY292223)
(3a). Synthesis was performed using the general proce-
dure described above using 20-hydroxyacetophenone
(565 mg, 4.15 mmol), CH2Cl2 (11 mL), TiCl4 (6.5 mL
of a 1.0 M CH2Cl2 solution, 6.5 mmol), diisopropyleth-
ylamine (2.7 mL, 15.5 mmol), 4-dichloromethyl-
enemorpholin-4-ium chloride (1.2 g, 5.0 mmol), Et3N
(1.0 mL), and MeOH (5 mL). Column chromatography:
SiO2, 40% EtOAc–hexanes + 1% Et3N to 90% EtOAc–
MeOH + 1% Et3N gradient (344 mg, 1.5 mmol, 36%).
1H NMR: (CDCl3) d 7.95, 7.35, 7.14, 7.08, 5.27 (s,
1H), 3.63 (t, J = 5 Hz, 4H), 3.30 (t, J = 5 Hz, 4H).
Lit.40 d 8.15 (m, 1H), 7.56 (m, 1H), 7.31 (m, 2H), 5.49
(s, 1H), 3.83 (t, J = 4.7 Hz, 4H), 3.50 (t, J = 4.7 Hz,
4H). 13C{1H} NMR: (CDCl3) d 176.8, 162.5, 153.5,
132.2, 125.2, 124.6, 122.8, 116.3, 87.0, 65.8, 44.5. Lit.40
d 176.9, 162.4, 153.5, 132.1, 125.3, 124.6, 122.7, 116.1,
87.2, 65.8, 44.4. IR: 1616, 1555, 1418, 1300, 1251,
1117, 985, 766. Lit.40 1622, 1559. HRMS: (EI)
231.0895 calcd for C13H13NO3 (M+); found 231.0887.
Anal. Calcd for C13H13NO3: C, 67.52; H, 5.66; N,
6.00. Found: C, 67.59; H, 5.69; N, 5.97.
4.4.3.4. 1-(2-Hydroxyphenyl)hexan-1-one (2f). Synthe-
sis was performed using the general procedure described
above using 1f (2.0 g, 10.3 mmol), MnO2 (8.0 g,
92 mmol), and CH2Cl2 (110 mL). Compound 2f was
isolated as a colorless oil (900 mg, 4.68 mmol, 46%).
1H NMR: (CDCl3) d 12.38 (s, 1H), 7.70 (dd, J = 2 Hz,
J = 8 Hz, 1H), 7.40 (m, 1H), 6.92 (m, 1H), 6.83 (m,
1H), 2.92 (t, J = 7 Hz, 2H), 1.70 (m, 2H), 1.33 (m,
4H), 0.88 (t, J = 7 Hz, 3H). Lit.40 d 12.40 (s, 1H),
7.82–6.80 (m, 4H), 2.98 (t, J = 7.6 Hz, 2H), 1.84–1.25
(m, 6H), 0.91 (t, J = 6.7 Hz, 3H). 13C{1H} NMR:
(CDCl3) d 206.8, 162.4, 136.0, 129.9, 119.2, 118.7,
118.3, 38.1, 31.3, 24.0, 22.4, 13.8. Lit.40 d 207.18,
162.70, 136.36, 130.18, 119.53, 119.01, 118.70, 38.49,
31.64, 24.41, 22.67, 14.11. IR: 3426, 1639, 1487, 1447,
1197, 1156, 752. HRMS: (EI) 192.1150 calcd for
C12H16O2 (M+); found 192.1145. Lit.40 192.
4.4.3.5. 1-(2-Hydroxyphenyl)-2-phenylethanone (2g).
Synthesis was performed using the general procedure de-
scribed above using 1g (1.5 g, 7.0 mmol), MnO2 (6.0 g,
69 mmol), and CH2Cl2 (80 mL). Compound 2g was iso-
4.4.4.2.
3-Methyl-2-morpholin-4-yl-chromen-4-one
(3b). Synthesis was performed using the general proce-
dure described above using 200 mg 20-hydroxypropio-
phenone, 3.5 mL CH2Cl2, 2.0 mmol TiCl4, 600 mg
i-Pr2EtN, 308 mg dichloromethylenemorpholin-4-ium
chloride, 0.4 mL Et3N, and 2 mL MeOH. Column chro-
matography: SiO2, 50% EtOAc–hexanes + 1% Et3N
1
lated as a colorless oil (624 mg, 2.94 mmol, 42%). H
NMR: (CDCl3) d 12.27 (s, 1H), 7.86 (dd, J = 2 Hz,
J = 8 Hz, 1H), 7.45 (m, 1H), 7.36 (m, 2H), 7.31 (m,
1
3H), 6.99 (m, 1H), 6.88 (m, 1H), 4.29 (s, 2H). Lit.39
d
(77 mg, 0.31 mmol, 24%). H NMR: (C6D6) d 8.52 (d,
12.1 (s, 1H), 7.73–6.91 (m, 9H), 4.2 (s, 2H). 13C{1H}
J = 8 Hz, 1H), 7.07 (m, 1H), 6.93 (m, 2H), 3.31 (t,
J = 5 Hz, 4H), 2.70 (t, J = 5 Hz, 4H), 2.01 (s, 3H).
13C{1H} NMR: (C6D6) d 177.9, 161.5, 154.2, 132.0,
126.4, 124.6, 123.4, 116.7, 103.6, 66.5, 48.3, 11.0. IR:
1612, 1556, 1410, 1114. HRMS: (EI) 245.1051 calcd
for C14H15NO3 (M+); found 245.1050. Combustion
analysis was not obtained for this compound due to its
similarity to 3a–i.
NMR: (CDCl3) d 203.7, 162.7, 136.4, 133.8, 130.3,
129.3, 128.6, 127.0, 118.9, 118.9, 118.5, 44.9. Lit.39
d
203.7, 162.7, 136.3–118.4 (11C), 44.9. IR: 1638, 1487,
1446, 1344, 1275, 1156, 754. Lit.39 1680. HRMS: (EI)
212.0837 calcd for C14H12O2 (M+); found 212.0835.
Lit.39 212.
4.4.4. General procedure for the preparation of com-
pounds 3a–i. A round bottom flask was charged with the
appropriate 20-hydroxyketophenol(2), CH2Cl2, and
cooled to ꢁ78 ꢁC. TiCl4 (1.0 M CH2Cl2 solution) was
added dropwise by syringe and was stirred for 1 h at
ꢁ78 ꢁC to produce a reddish-brown-colored solution.
Diisopropylethylamine was added and the reaction
was stirred at ꢁ78 ꢁC for 1 h. 4-Dichloro-
methylenemorpholin-4-ium chloride was added and the
reaction mixture was stirred at ꢁ78 ꢁC then warmed to
0 ꢁC for 2 h, then allowed to warm to rt and stirred over-
night. The reaction mixture was cooled to 0 ꢁC, and
4.4.4.3. 3-Ethyl-2-morpholin-4-yl-chromen-4-one (3c).
Synthesis was performed using the general procedure de-
scribed above using 407 mg 2c, 30 mL CH2Cl2,
3.71 mmol TiCl4, 3.2 g i-Pr2EtN, 759 mg dichloro-
methylenemorpholin-4-ium chloride, 0.6 mL Et3N, and
3 mL MeOH. Column chromatography: SiO2, 30%
EtOAc–hexanes + 1% Et3N (364 mg, 1.40 mmol, 57%).
1H NMR: (C6D6) d 8.29 (d, J = 8 Hz, 1H), 7.12 (m,
1H), 6.95 (m, 2H), 3.40 (t, J = 5 Hz, 4H), 2.83 (t,
J = 5 Hz, 4H), 2.48 (q, J = 7 Hz, 2H), 1.19 (t,
J = 7 Hz, 3H). 13C{1H} NMR: (C6D6) d 177.8, 161.8,

2834
P. J. Alaimo et al. / Bioorg. Med. Chem. 13 (2005) 2825–2836
154.2, 132.3, 126.0, 124.5, 123.5, 116.9, 110.4, 66.6, 49.3,
18.6, 13.3. IR: 1615, 1556, 1467, 1401, 1361, 1264, 1223,
1116, 759. HRMS: (EI) 259.1208 calcd for C15H17NO3
(M+); found 259.1199. Anal. Calcd for C15H17NO3: C,
69.48; H, 6.61; N, 5.40. Found: C, 69.43; H, 6.60; N,
5.36.
1.01 mmol TiCl4, 870 mg i-Pr2EtN, 206 mg di-
chloromethylenemorpholin-4-ium chloride, 0.1 mL
Et3N, and 0.5 mL MeOH. Column chromatography:
SiO2, 20% EtOAc–hexanes + 1% Et3N (144 mg,
0.47 mmol, 70%). 1H NMR: (C6D6)
d 8.39 (d,
J = 8 Hz, 1H), 7.47 (m, 2H), 7.20 (m, 2H), 7.11–7.02
(m, 2H), 6.96–6.90 (m, 2H), 3.11 (t, J = 5 Hz, 4H),
2.68 (t, J = 5 Hz, 4H). 13C{1H} NMR: (C6D6) d 175.5,
160.9, 153.6, 134.5, 132.2, 131.3, 128.2, 126.9, 126.6,
124.7, 124.0, 116.6, 105.2, 66.1, 47.8. IR: 1614, 1550,
1417, 1340, 1232, 1116, 760. HRMS: (EI) 307.1208 calcd
for C19H17NO3 (M+); found 307.1206. Combustion
analysis was not obtained for this compound due to its
similarity to 3a–i.
4.4.4.4. 2-Morpholin-4-yl-3-propylchromen-4-one (3d).
Synthesis was performed using the general procedure
described above using 68 mg 2d, 5 mL CH2Cl2,
0.57 mmol TiCl4, 490 mg i-Pr2EtN, 117 mg dichloro-
methylenemorpholin-4-ium chloride, 0.1 mL Et3N, and
0.5 mL MeOH. Column chromatography: SiO2, 20%
EtOAc–hexanes + 1% Et3N (58 mg, 0.21 mmol, 56%).
1H NMR: (C6D6) d 8.46 (d, J = 6 Hz, 1H), 7.08 (m,
1H), 6.94 (m, 2H), 3.38 (t, J = 5 Hz, 4H), 2.77 (t,
J = 5 Hz, 4H), 2.57 (t, J = 8 Hz, 2H), 1.77 (tq,
J = 8 Hz, 2H), 0.99 (t, J = 8 Hz, 3H). 13C{1H} NMR:
(C6D6) d 178.2, 162.0, 154.5, 132.2, 126.5, 124.6, 123.6,
116.8, 110.2, 66.6, 49.4, 27.3, 22.0, 14.7. IR: 1616,
1556, 1467, 1400, 1115, 759. HRMS: (EI) 273.1365 calcd
for C16H19NO3 (M+); found 273.1369. Combustion
analysis was not obtained for this compound due to its
similarity to 3a–i.
4.4.4.8.
3-Benzyl-2-morpholin-4-yl-chromen-4-one
(3h). Synthesis was performed using the general proce-
dure described above using 1.0 g 20-hydroxy-3-phenyl-
propiophenone, 12 mL CH2Cl2, 6.63 mmol TiCl4,
2.7 mL i-Pr2EtN, 1.2 g dichloromethylenemorpholin-4-
ium chloride, 0.2 mL Et3N, and 1.2 mL MeOH. Column
chromatography: SiO2, 35% EtOAc–hexanes + 1% Et3N
1
(552 mg, 1.80 mmol, 41%). H NMR: (C6D6) d 8.18 (d,
J = 8 Hz, 1H), 7.57 (t, J = 8 Hz, 1H), 7.34 (m, 2H), 7.19
(m, 5H), 3.94 (s, 2H), 3.67 (t, J = 5 Hz, 4H), 3.28 (t,
J = 5 Hz, 4H). 13C{1H} NMR: (C6D6) d 178.5, 162.8,
153.9, 140.0, 132.6, 128.4, 127.9, 126.1, 125.9, 124.7,
122.6, 116.7, 106.5, 66.6, 48.8, 30.5. IR: 1615, 1555,
1467, 1402, 1236, 1114, 759. HRMS: (EI) 321.1365 calcd
for C20H19NO3 (M+); found 321.1368. Combustion
analysis was not obtained for this compound due to its
similarity to 3a–i.
4.4.4.5. 3-Isopropyl-2-morpholin-4-yl-chromen-4-one
(3e). Synthesis was performed using the general pro-
cedure described above using 150 mg 2e, 10 mL
CH2Cl2, 1.24 mmol TiCl4, 1.06 g i-Pr2EtN, 253 mg
dichloromethylenemorpholin-4-ium chloride, 0.2 mL
Et3N, and 1.0 mL MeOH. Column chromatography:
SiO2, 20% EtOAc–hexanes + 1% Et3N (138 mg,
0.50 mmol, 60%). 1H NMR: (C6D6)
d 8.36 (d,
J = 8 Hz, 1H), 7.01 (m, 1H), 6.89 (m, 2H), 3.47 (t,
J = 5 Hz, 4H), 2.96 (m, 1H), 2.72 (t, J = 5 Hz, 4H),
1.54 (d, J = 7, 6H). 13C{1H} NMR: (C6D6) d 178.2,
162.1, 154.4, 132.2, 126.3, 124.5, 124.4, 116.7, 115.2,
66.4, 50.1, 27.7, 20.5. IR: 1616, 1557, 1466, 1390, 1217,
1115, 759. HRMS: (EI) 273.1364 calcd for C16H19NO3
(M+); found 273.1364. Combustion analysis was not ob-
tained for this compound due to its similarity to 3a–i.
4.4.4.9. 3-Decyl-2-morpholin-4-yl-chromen-4-one (3i).
Synthesis was performed using the general procedure de-
scribed above for 102 mg 20-hydroxydodecanophenone,
3.0 mL CH2Cl2, 0.551 mmol TiCl4, 473 mg i-Pr2EtN,
113 mg dichloromethylenemorpholin-4-ium chloride,
0.06 mL Et3N, and 0.3 mL MeOH. Column chromato-
graphy: SiO2, 20% EtOAc–hexanes + 1% Et3N.
(37 mg, 0.098 mmol, 27%). 1H NMR: (C6D6) d 8.50
(d, J = 8 Hz, 1H), 7.07 (m, 1H), 6.93 (m, 2H), 3.42 (t,
J = 5 Hz, 4H), 2.80 (t, J = 5 Hz, 4H), 2.66 (t, J = 8 Hz,
2H), 1.82 (m, 2H), 1.48–1.20 (m, 14H), 0.90
(t, J = 7 Hz, 3H). 13C{1H} NMR: (C6D6) d 178.2,
162.0, 154.5, 132.2, 126.5, 124.6, 123.7, 116.8, 110.5,
66.6, 49.4, 32.3, 30.4, 30.11, 30.09, 29.9, 29.8, 28.7,
25.2, 23.1, 14.3. IR: 1615, 1561, 1466, 1395, 1117.
HRMS: (EI) 371.2460 calcd for C23H33NO3 (M+);
found 371.2458. Combustion analysis was not obtained
for this compound due to its similarity to 3a–i.
4.4.4.6. 3-Butyl-2-morpholin-4-yl-chromen-4-one (3f).
Synthesis was performed using the general procedure de-
scribed above using 375 mg 2f, 20 mL CH2Cl2,
2.92 mmol TiCl4, 2.51 g i-Pr2EtN, 597 mg dichloro-
methylenemorpholin-4-ium chloride, 0.4 mL Et3N, and
2.0 mL MeOH. Column chromatography: SiO2, 20%
EtOAc–hexanes + 1% Et3N (296 mg, 1.03 mmol, 53%).
1H NMR: (C6D6) d 8.48 (d, J = 8 Hz, 1H), 7.07 (m,
1H), 6.93 (m, 2H), 3.39 (t, J = 5 Hz, 4H), 2.77 (t,
J = 5 Hz, 4H), 2.61 (t, J = 8 Hz, 2H), 1.75 (p,
J = 8 Hz, 2H), 1.41 (s, J = 8 Hz, 2H), 0.95 (t, J = 8 Hz,
3H). 13C{1H} NMR: (C6D6) d 178.1, 162.0, 154.5,
132.2, 126.5, 124.6, 123.7, 116.8, 110.4, 66.6, 49.4,
30.8, 24.8, 23.3, 14.2. IR: 1615, 1558, 1466, 1398, 1223,
1116, 759. HRMS: (EI) 287.1521 calcd for C17H21NO3
(M+); found 287.1520. Combustion analysis was not ob-
tained for this compound due to its similarity to 3a–i.
Acknowledgements
We thank the members of the Shokat research group,
particularly Quincey Justman, Scott Lazerwith, Jennifer
Paulson, and Daniel Rauh, as well as Professor James
Bobbitt (Univ. Connecticut) for helpful discussions.
Funding was provided by postdoctoral fellowships from
the Susan G. Komen Breast Cancer Foundation and the
California Section of the American Cancer Society
(P.J.A.), a predoctoral fellowship from the Howard
4.4.4.7. 2-Morpholin-4-yl-3-phenylchromen-4-one (3g).
Synthesis was performed using the general procedure
described above using 143 mg 2g, 5 mL CH2Cl2,

P. J. Alaimo et al. / Bioorg. Med. Chem. 13 (2005) 2825–2836
2835
recognition by the tyrosine kinase v-Src. Bioorg. Med.
Chem. 1998, 6, 1219–1226.
Hughes Medical Institute (Z.A.K.), NIH grant
AI440099 (K.M.S.), and an award from the Sandler
Program for Asthma Research (K.M.S.). Mass spectra
were provided by the Center for Mass Spectrometry at
the University of California, San Francisco, supported
by the NIH division of Research Resources.
14. Shah, N. P.; Tran, C.; Lee, F. Y.; Chen, P.; Norris, D.;
Sawyers, C. L. Overriding imatinib resistance with a novel
ABL kinase inhibitor. Science 2004, 305, 399–401.
15. La Rosee, P.; Corbin, A. S.; Stoffregen, E. P.; Deininger,
M. W.; Druker, B. J. Activity of the Bcr-Abl kinase
inhibitor PD180970 against clinically relevant Bcr-Abl
isoforms that cause resistance to imatinib mesylate (Glee-
vec, STI571). Cancer Res. 2002, 62, 7149–7153.
16. Katso, R.; Okkenhaug, K.; Ahmadi, K.; White, S.;
Timms, J.; Waterfield, M. D. Cellular function of phos-
phoinositide 3-kinases: implications for development,
homeostasis, and cancer. Annu. Rev. Cell Dev. Biol.
2001, 17, 615–675.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.bmc.
2005.02.021.
17. Samuels, Y.; Wang, Z.; Bardelli, A.; Silliman, N.; Ptak, J.;
Szabo, S.; Yan, H.; Gazdar, A.; Powell, S. M.; Riggins, G.
J.; Willson, J. K.; Markowitz, S.; Kinzler, K. W.;
Vogelstein, B.; Velculescu, V. E. High frequency of
mutations of the PIK3CA gene in human cancers. Science
2004, 304, 554.
18. Leslie, N. R.; Downes, C. P. PTEN function: how normal
cells control it and tumour cells lose it. Biochem. J. 2004,
382, 1–11.
References and notes
1. Blencke, S.; Zech, B.; Engkvist, O.; Greff, Z.; Orfi, L.;
Horvath, Z.; Keri, G.; Ullrich, A.; Daub, H. Character-
ization of a conserved structural determinant controlling
protein kinase sensitivity to selective inhibitors. Chem.
Biol. 2004, 11, 691–701.
2. Liu, Y.; Bishop, A.; Witucki, L.; Kraybill, B.; Shimizu, E.;
Tsien, J.; Ubersax, J.; Blethrow, J.; Morgan, D. O.;
Shokat, K. M. Structural basis for selective inhibition of
Src family kinases by PP1. Chem. Biol. 1999, 6, 671–678.
3. Ward, S.; Sotsios, Y.; Dowden, J.; Bruce, I.; Finan, P.
Therapeutic potential of phosphoinositide 3-kinase inhib-
itors. Chem. Biol. 2003, 10, 207–213.
4. Ward, S. G.; Finan, P. Isoform-specific phosphoinositide
3-kinase inhibitors as therapeutic agents. Curr. Opin.
Pharmacol. 2003, 3, 426–434.
19. Nagata, Y.; Lan, K. H.; Zhou, X.; Tan, M.; Esteva, F. J.;
Sahin, A. A.; Klos, K. S.; Li, P.; Monia, B. P.; Nguyen, N.
T.; Hortobagyi, G. N.; Hung, M. C.; Yu, D. PTEN
activation contributes to tumor inhibition by trastuzumab,
and loss of PTEN predicts trastuzumab resistance in
patients. Cancer Cell 2004, 6, 117–127.
20. Cappuzzo, F.; Magrini, E.; Ceresoli, G. L.; Bartolini, S.;
Rossi, E.; Ludovini, V.; Gregorc, V.; Ligorio, C.; Cancel-
lieri, A.; Damiani, S.; Spreafico, A.; Paties, C. T.;
Lombardo, L.; Calandri, C.; Bellezza, G.; Tonato, M.;
Crino, L. Akt phosphorylation and gefitinib efficacy in
patients with advanced non-small-cell lung cancer. J. Natl.
Cancer Inst. 2004, 96, 1133–1141.
5. Noble, M. E.; Endicott, J. A.; Johnson, L. N. Protein
kinase inhibitors: insights into drug design from structure.
Science 2004, 303, 1800–1805.
6. Druker, B. J. Molecularly targeted therapy: have the
floodgates opened?. Oncologist 2004, 9, 357–360.
7. Druker, B. J.; Tamura, S.; Buchdunger, E.; Ohno, S.;
Segal, G. M.; Fanning, S.; Zimmermann, J.; Lydon, N. B.
Effects of a selective inhibitor of the Abl tyrosine kinase on
the growth of Bcr-Abl positive cells. Nat. Med. 1996, 2,
561–566.
21. Sordella, R.; Bell, D. W.; Haber, D. A.; Settleman, J.
Gefitinib-sensitizing EGFR mutations in lung cancer
activate anti-apoptotic pathways. Science 2004, 305,
1163–1167.
22. Luo, J.; Manning, B. D.; Cantley, L. C. Targeting the
PI3K-Akt pathway in human cancer: rationale and
promise. Cancer Cell 2003, 4, 257–262.
8. Eyers, P. A.; Craxton, M.; Morrice, N.; Cohen, P.;
Goedert, M. Conversion of SB 203580-insensitive MAP
kinase family members to drug-sensitive forms by a single
amino-acid substitution. Chem. Biol. 1998, 5, 321–328.
9. Shah, K.; Liu, Y.; Deirmengian, C.; Shokat, K. M.
Engineering unnatural nucleotide specificity for Rous
sarcoma virus tyrosine kinase to uniquely label its direct
substrates. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 3565–
3570.
23. Walker, E. H.; Pacold, M. E.; Perisic, O.; Stephens, L.;
Hawkins, P. T.; Wymann, M. P.; Williams, R. L.
Structural determinants of phosphoinositide 3-kinase
inhibition by wortmannin, LY294002, quercetin, myrice-
tin, and staurosporine. Mol. Cell 2000, 6, 909–919.
24. Matter, W. F.; Brown, R. F.; Vlahos, C. J. The inhibition
of phosphatidylinositol 3-kinase by quercetin and analogs.
Biochem. Biophys. Res. Commun. 1992, 186, 624–
631.
10. Blencke, S.; Ullrich, A.; Daub, H. Mutation of threonine
766 in the epidermal growth factor receptor reveals a
hotspot for resistance formation against selective tyrosine
kinase inhibitors. J. Biol. Chem. 2003, 278, 15435–15440.
11. Gorre, M. E.; Mohammed, M.; Ellwood, K.; Hsu, N.;
Paquette, R.; Rao, P. N.; Sawyers, C. L. Clinical
resistance to STI-571 cancer therapy caused by BCR-
ABL gene mutation or amplification. Science 2001, 293,
876–880.
12. Bishop, A. C.; Ubersax, J. A.; Petsch, D. T.; Matheos, D.
P.; Gray, N. S.; Blethrow, J.; Shimizu, E.; Tsien, J. Z.;
Schultz, P. G.; Rose, M. D.; Wood, J. L.; Morgan, D. O.;
Shokat, K. M. A chemical switch for inhibitor-sensitive
alleles of any protein kinase. Nature 2000, 407, 395–401.
13. Liu, Y.; Shah, K.; Yang, F.; Witucki, L.; Shokat, K. M. A
molecular gate which controls unnatural ATP analogue
25. Walker, E. H.; Perisic, O.; Ried, C.; Stephens, L.;
Williams, R. L. Structural insights into phosphoinositide
3-kinase catalysis and signalling. Nature 1999, 402, 313–
320.
26. Herman, P. K.; Emr, S. D. Characterization of VPS34, a
gene required for vacuolar protein sorting and vacuole
segregation in Saccharomyces cerevisiae. Mol. Cell Biol.
1990, 10, 6742–6754.
27. Kato, R.; Ogawa, H. An essential gene, ESR1, is required
for mitotic cell growth, DNA repair and meiotic recom-
bination in Saccharomyces cerevisiae. Nucleic Acids Res.
1994, 22, 3104–3112.
28. Desany, B. A.; Alcasabas, A. A.; Bachant, J. B.; Elledge,
S. J. Recovery from DNA replicational stress is the
essential function of the S-phase checkpoint pathway.
Genes Dev. 1998, 12, 2956–2970.

2836
P. J. Alaimo et al. / Bioorg. Med. Chem. 13 (2005) 2825–2836
29. Pal, D.; Chakrabarti, P. Beta-sheet propensity and its
correlation with parameters based on conformation. Acta
Crystallogr. D Biol. Crystallogr. 2000, 56(Pt. 5), 589–594.
30. Vlahos, C. J.; Matter, W. F.; Hui, K. Y.; Brown, R. F. A
specific inhibitor of phosphatidylinositol 3-kinase, 2-(4-
morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002).
J. Biol. Chem. 1994, 269, 5241–5248.
31. Knight, Z. A.; Chiang, G. C.; Alaimo, P. J.; Kenski, D.
M.; Ho, C. B.; Coan, K.; Abraham, R. T.; Shokat, K. M.
Isoform-specific phosphoinositide 3-kinase inhibitors from
an arylmorpholine scaffold. Bioorg. Med. Chem. 2004, 12,
4749–4759.
32. Sadhu, C.; Dick, K.; Tino, W. T.; Staunton, D. E.
Selective role of PI3K delta in neutrophil inflammatory
responses. Biochem. Biophys. Res. Commun. 2003, 308,
764–769.
33. Morris, J.; Wishka, D. G.; Fang, Y. A novel synthesis of
2-aminochromones via phosgeniminium salts. J. Org.
Chem. 1992, 57, 6502–6508.
35. Tanque, Y.; Terada, A.; Seto, I.; Umezu, Y.; Tsuge, O.
One-pot ortho hydroxylations of 2-(1-hydroalkyl)naphth-
ylenes and (1-hydroxyalkyl)benzenes. Bull. Chem. Soc.
Jpn. 1988, 61, 1221–1224.
36. Aniol, M.; Wawrzenczyk, C. Chromenes and chroma-
nones. 4. The birch reduction of 2,2-dimethyl-4-chroma-
none and its 7-substituted analogues. Heterocycles 1997,
45, 1069–1079.
37. Foster, K. L.; Baker, S.; Brousmiche, D. W.; Wan, P.
o-Quinone methide formation from excited state intra-
molecular proton transfer (ESIPT) in an o-hydroxysty-
rene. J. Photochem. Photobiol. A—Chem. 1999, 129, 157–
163.
38. Cimarelli, C.; Palmieri, G. Alkylation of dianions derived
from 2-(1-iminoalkyl) phenols: synthesis of functionalized
2-acyl phenols. Tetrahedron 1998, 54, 15711–15720.
39. Dolhem, E.; Barhdadi, R.; Folest, J. C.; Nedelec, J. Y.;
Troupel, M. Nickel catalysed electrosynthesis of ketones
from organic halides and iron pentacarbonyl. Part 2:
unsymmetrical ketones. Tetrahedron 2001, 57, 525–529.
40. Morris, J.; Wishka, D. G.; Fang, Y. A cyclodehydration
route to 2-aminochromones. Synth. Commun. 1994, 24,
849–858.
34. Morris, J.; Luke, G. P.; Wishka, D. G. Reaction of
phosgeniminium salts with enolates derived from Lewis
acid complexes of 20-hydroxypropiophenones and related
beta-diketones. J. Org. Chem. 1996, 61, 3218–3220.