Structure based design of novel 6,5 heterobicyclic mitogen-activated protein kinase kinase (MEK) inhibitors leading to the discovery of imidazo[1,5-a] pyrazine G-479

Article abstract of DOI:10.1016/j.bmcl.2014.08.008

Use of the tools of SBDD including crystallography led to the discovery of novel and potent 6,5 heterobicyclic MEKi's [J. Med. Chem. 2012, 55, 4594]. The core change from a 5,6 heterobicycle to a 6,5 heterobicycle was driven by the desire for increased structural diversity and aided by the co-crystal structure of G-925 [J. Med. Chem. 2012, 55, 4594]. The key design feature was the shift of the attachment of the five-membered heterocyclic ring towards the B ring while maintaining the key hydroxamate and anilino pharamcophoric elements in a remarkably similar position as in G-925. From modelling, changing the connection point of the five membered ring heterocycle placed the H-bond accepting nitrogen within a good distance and angle to the Ser212 [J. Med. Chem. 2012, 55, 4594]. The resulting novel 6,5 benzoisothiazole MEKi G-155 exhibited improved potency versus aza-benzofurans G-925 and G-963 but was a potent inhibitor of cytochrome P450's 2C9 and 2C19. Lowering the log D by switching to the more polar imidazo[1,5-a] pyridine core significantly diminished 2C9/2C19 inhibition while retaining potency. The imidazo[1,5-a] pyridine G-868 exhibited increased potency versus the starting point for this work (aza-benzofuran G-925) leading to deprioritization of the azabenzofurans. The 6,5-imidazo[1,5-a] pyridine scaffold was further diversified by incorporating a nitrogen at the 7 position to give the imidazo[1,5-a] pyrazine scaffold. The introduction of the C7 nitrogen was driven by the desire to improve metabolic stability by blocking metabolism at the C7 and C8 positions (particularly the HLM stability). It was found that improving on G-868 (later renamed GDC-0623) required combining C7 nitrogen with a diol hydroxamate to give G-479. G-479 with polarity distributed throughout the molecule was improved over G-868 in many aspects.

Full text of DOI:10.1016/j.bmcl.2014.08.008

Bioorganic & Medicinal Chemistry Letters 24 (2014) 4714–4723
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure based design of novel 6,5 heterobicyclic mitogen-activated
protein kinase kinase (MEK) inhibitors leading to the discovery of
imidazo[1,5-a] pyrazine G-479
a
b
b
b
c
Kirk D. Robarge a, , Wendy Lee , Charles Eigenbrot , Mark Ultsch , Christian Wiesmann , Robert Heald ,
Steve Price c, Joanne Hewitt c, Philip Jackson c, Pascal Savy c, Brenda Burton c, Edna F. Choo d, Jodie Pang d,
Jason Boggs d, April Yang d, Xioaye Yang d, Matthew Baumgardner d
⇑
a Discovery Chemistry, Genentech Inc., 1 DNA Way, South San Francisco, CA, USA
b Structural Biology, Genentech Inc., 1 DNA Way, South San Francisco, CA, USA
c Argenta, Discovery Services Charles River, 8-9 Spire Greene Centre, Harlow, Essex CM19 5TR, United Kingdom
d Drug Metabolism and Pharmacokinetics, Genentech Inc., 1 DNA Way, South San Francisco, CA, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Use of the tools of SBDD including crystallography led to the discovery of novel and potent 6,5 heterobi-
cyclic MEKi’s [J. Med. Chem. 2012, 55, 4594]. The core change from a 5,6 heterobicycle to a 6,5 heterobi-
cycle was driven by the desire for increased structural diversity and aided by the co-crystal structure of
G-925 [J. Med. Chem. 2012, 55, 4594]. The key design feature was the shift of the attachment of the
five-membered heterocyclic ring towards the B ring while maintaining the key hydroxamate and anilino
pharamcophoric elements in a remarkably similar position as in G-925. From modelling, changing the
connection point of the five membered ring heterocycle placed the H-bond accepting nitrogen within a good
distance and angle to the Ser212 [J. Med. Chem. 2012, 55, 4594]. The resulting novel 6,5 benzoisothiazole
MEKi G-155 exhibited improved potency versus aza-benzofurans G-925 and G-963 but was a potent
inhibitor of cytochrome P450’s 2C9 and 2C19. Lowering the logD by switching to the more polar
imidazo[1,5-a] pyridine core significantly diminished 2C9/2C19 inhibition while retaining potency. The
Received 27 June 2014
Revised 1 August 2014
Accepted 5 August 2014
Available online 15 August 2014
Keywords:
Mitogen-activated protein kinase kinase
inhibitors
MEK inhibitors
RAS signaling pathway
Imidazo[1,5-a] pyrazine
Abbreviations: MAPK/ERK kinase (MEK) signaling cascade, The pathway includes many proteins, including MAPK (mitogen-activated protein kinases, originally called ERK,
extracellular signal-regulated kinases), which communicate by adding phosphate groups to a neighboring protein, which acts as an ‘on’ or ‘off’ switch. Also known as the RAS-
RAF-MEK-ERK mitogen-activated protein kinase signaling pathway; KRAS, GTPase KRas also known as V-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog and KRAS, is a
protein that in humans is encoded by the KRAS gene; NRAS, The N-ras oncogene is a member of the Ras gene family and an enzyme that in humans is encoded by the NRAS
gene. It was named NRAS for its initial identification in human neuroblastoma cells; HRAS, GTPase HRas also known as transforming protein p21 is an enzyme that in humans
is encoded by the HRAS gene. Once bound to guanosine triphosphate, activates a Raf kinase like c-Raf, the next step in the MAPK/ERK signaling cascade; BRAF, BRAF is a
human gene that makes a protein called B-Raf. The gene is also referred to as proto-oncogene B-Raf and v-Raf murine sarcoma viral oncogene homolog B, while the protein is
more formally known as serine/threonine-protein kinase B-Raf; RAF, proto-oncogene serine/threonine-protein kinase also known as c-RAF; AKT, AKT is also known as protein
kinase B(PKB), is a serine/threonine-specific protein kinase that plays a key role in multiple cellular processes such as glucose metabolism, apoptosis, cell proliferation,
transcription and cell migration; PI3K, phosphatidylinositol-4,5-bisphosphate 3-kinase are a family of enzymes involved in cellular functions such as cell growth,
proliferation, differentiation, motility, survival and intracellular trafficking, which in turn are involved in cancer; MEKi’s, MAPK/ERK kinase (MEK) signaling cascade inhibitors.
A MEK inhibitor is a chemical or drug that inhibits the mitogen-activated protein kinase kinase enzymes MEK1 and/or MEK2. Allosteric MEK inhibitors described herein
inhibit both MEK1 and MEK 2; MEK 1,2 inhibitors, MEK 1,2 inhibitors are synonymous with MEKi’s; ATP, adenosine triphosphate is a nucleoside triphosphate used in cells as a
coenzyme; H-bond, hydrogen bond; BID, twice daily; QD, once daily; CYP, cytochrome P450; logD (measured), distribution-coefficient is the ratio of concentrations of a
compound in a mixture of two immiscible phases at equilibrium. These coefficients are a measure of the difference in solubility of the compound in these two phases. LogD’s
have a pH dependence. LogD measurements for MEKi’s disclosed herein were performed at pH 7.4; clogD, calculated logD @ pH 7.4; clogP, calculated partition-coefficient
(clogP) is the calculated ratio of concentrations of a compound in a mixture of two immiscible phases at equilibrium. These coefficients are a measure of the difference in
solubility of the compound in these two phases. The clogP value reflects the overall lipophilicity of a molecule; IV, delivered directly into the vein; PO, by mouth; Mic(H/R/M/
D/C), Microsomes H(human)/R(rat)/M(mouse)/D(dog)/C(cyno). Incubation of the MEKi’s described herein was conducted with liver microsomes from human(HLM) and pre-
clinical species rat(RLM), mouse(MLM), dog(DLM), and cyno(CLM) to assess comparative metabolic stability; PEG 400, polyethylene glycol 400 is a low-molecular-weight
grade of polyethylene glycol miscible with water and used as vehicle for dosing rats IV and PO; CLp, In vivo total plasma clearance is defined as the volume of blood or plasma
cleared of drug in a unit time. It is related to the volume in which the drug is dissolved and the rate at which it is eliminated; T1/2 (h), half life of MEK inhibitor when dosed
intravenously; ppb, plasma protein binding; Oral exposure, area under the plasma concentration versus time curve (AUC); %F, oral bioavailability which is the dose-corrected
area under curve (AUC) after oral dosing divided by AUC intravenous; IV/IV, in vitro/in vivo clearance correlation (CL predicted from rat microsome incubation/versus CL
observed after intravenous administration in rat).
⇑
Corresponding author. Tel.: +1 650 225 2320; fax: +1 650 742 4943.
E-mail address: kir@gene.com (K.D. Robarge).
http://dx.doi.org/10.1016/j.bmcl.2014.08.008
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

K. D. Robarge et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4714–4723
4715
Imidazo[1,5-a] pyridine
Structure based drug design (SBDD)
Oncology
imidazo[1,5-a] pyridine G-868 exhibited increased potency versus the starting point for this work (aza-
benzofuran G-925) leading to deprioritization of the azabenzofurans. The 6,5-imidazo[1,5-a] pyridine
scaffold was further diversified by incorporating a nitrogen at the 7 position to give the imidazo[1,5-a]
pyrazine scaffold. The introduction of the C7 nitrogen was driven by the desire to improve metabolic sta-
bility by blocking metabolism at the C7 and C8 positions (particularly the HLM stability). It was found
that improving on G-868 (later renamed GDC-0623) required combining C7 nitrogen with a diol hydroxa-
mate to give G-479. G-479 with polarity distributed throughout the molecule was improved over G-868
in many aspects.
Ó 2014 Elsevier Ltd. All rights reserved.
The MAPK/ERK kinase (MEK) signaling cascade is
a
key
group is typically a hydroxamate, but other polar functionalities
such as an amide, a reverse sulfonamide, or a sulfonyl urea can
be utilized. A third binding feature is an H-bond acceptor to
Ser212. It has recently been disclosed that certain MEKi’s with
superior efficacy against KRAS tumors form a strong H-bond inter-
action with the Ser212 in MEK that is critical for blocking MEK
feedback phosphorylation by wild-type RAF.19 The strength of
the Ser212 H-bond interaction and position relative to the aniline
and hydroxamate is determined by the ‘A-ring’ to which these moi-
eties are attached (Fig. 1). It has been proposed that increasing the
strength of the Ser212/H-bond acceptor interaction results in
greater tolerance in the other pharmacophoric elements. A potent
MEK inhibitor requires having all three of these binding interac-
tions in place.
regulator of cellular proliferation, differentiation and survival
downstream of RAS activation.2 Upregulation of this pathway
occurs in a large fraction of tumors, frequently owing to oncogenic
activating mutations in KRAS, NRAS, HRAS and BRAF.3 Constitutive
activation of the extracellular regulated kinase (ERK) cascade,
through oncogenic forms of RAS and mutations in BRAF, has been
observed in lung, colon, pancreas, kidney, and ovary primary
human tumor samples.4 Whereas BRAF inhibitors have shown
remarkable efficacy against melanomas with BRAF (V600E) muta-
tions, these compounds are not effective against KRAS-mutant
tumors owing to inhibitor-mediated priming of wild-type RAF sig-
naling.5–7 As MEK is a common effector downstream of wild-type
and mutant RAF,8,9 MEK inhibitors have the potential to target all
tumors dependent on MAPK pathway signaling.
In our search for novel MEK inhibitors, we leveraged key infor-
mation from the X-ray crystal structure of our previously disclosed
5,6 heterobicyclic azabenzofuran G-925 to scaffold hop to a new
class of 6,5 heterobicyclic MEKi’s1 which, like the 5,6 azabenzofu-
ran, engaged the Ser212 in a monodentate interaction. The scaffold
hop from the 6,5 to the 5,6 heterobicyclic scaffold was driven by
two factors: (1) increasing the structural diversity and (2) to fur-
ther our understanding of the influence of the A ring on the place-
ment of the Ser212 H-bond accepting nitrogen and how that
impacts potency. Modelling indicated the 6,5 isobenzothiazole G-
155 would be tolerated in the MEK binding pocket with the H-
bond accepting nitrogen capable of making a good interaction with
Ser212. As shown in Figure 2, the 6,5 heterobicyclic isobenzothiaz-
ole G-155, while equipotent in the biochemical assay showed
increased potency versus its azabenzofuran comparator G-963 in
both cell proliferation assays. Typically for our potent MEKi’s, the
biochemical assay was not as useful as the 72 h cell proliferation
assays for differentiation. The cell proliferation (particularly the
HCT116 cell proliferation assay) potency was the decision making
potency readout used to differentiate our potent MEKi’s and deter-
mine their progression.
Although isobenzothiazole G-155 was potent, progression was
halted due to potent 2C9/2C19 CYP inhibition (Table 1). It was rea-
soned that the CYP 2C9/2C19 inhibition could be reduced by low-
ering lipophilicity and logD. A decision was made to introduce
polarity into the core and scaffold hop to 6,5 imidazo[1,5-a] pyri-
dines. In addition to lowering the lipophilicity (reflected in the
measured logD @ pH 7.4) versus the isobenzothiazole, the imi-
dazo[1,5-a] pyridine core increased the structural diversity of the
6,5 heterobicyclic MEKi’s and differentiated the pKa’s (cpKa’s using
MOKA1.1.0 for G-155/G-868 = 1.83/4.78) which can effect the
strength of the H-bond interaction with Ser212. Both imi-
dazo[1,5-a] pyridines G-86820,21 and G-60622 retained cell prolifer-
ation potency and exhibited greatly reduced 2C9(warfarin)/2C19
(mephenytoin) inhibition (Table 1). Imidazo[1,5-a] pyridines G-
868 and G-606 did not inhibit cytochrome P450’s 3A4 (testoster-
one/midazolam), 2D6 (dextromethorphan) and 1A2 (phenacetin).
With three examples of 6,5 heterobicyclic MEKi’s with increased
The RAS-RAF-MEK-ERK mitogen-activated protein kinase sig-
naling pathway has emerged as one of the most promising targets
for molecularly targeted antitumor agents. Currently, at least
twelve MEK inhibitors are at some stage of clinical evaluation for
the treatment of cancer (three others have been discontinued)
and one has been approved by the FDA for use as a single agent
for melanoma (Mekinist™, aka trametinib or GSK 1120212). Many
in the field of targeted therapies believe there is greater potential
for MEK inhibitors if they are used in combination with other tar-
geted agents such as BRAF,10 AKT,11 PI3K inhibitors12,13 and immu-
notherapies.14 Combination therapy has been suggested to better
inhibit the MAPK/ERK kinase pathway and therefore prolong treat-
ment effect. The results of phase 1 and 2 clinical studies utilizing a
selective BRAF inhibitor, TafinlarTM (aka dabrafenib), and the selec-
tive MEK inhibitor, trametinib, demonstrated the benefit of a com-
bination approach. It was found that dabrafenib and trametinib
could be safely combined at 150 mg BID/2 mg QD with rare
dose-limiting side effects. The combination therapy significantly
improved progression free survival and patients showed a higher
rate of response in comparison to dabrafenib monotherapy.10
However, despite the benefit demonstrated by the MEK/BRAF com-
bination, resistance still developed in most patients after an aver-
age of 9.4 months. The mechanisms of resistance to combined RAF/
MEK inhibition remain poorly understood.15
MEK 1,2 inhibitors16 bind to an allosteric binding site proximal
to the ATP binding site and as a result are exquisitely selective.
These compounds form a tertiary complex with the MEK enzyme
and ATP and are ATP noncompetitive.17,18 The binding site for allo-
steric MEK inhibitors has been previously described.1 There are
seven structurally related MEK inhibitors in the clinic that all pos-
sess an anilino ‘B-ring’ which occupies a lipophilic pocket formed
by Leu118, Ile126, Val127, Ile141, Met143, Phe129, Phe209, and
Val211 (Fig. 1). These MEKi’s all incorporate a halogen at the para
position which is lipophilic and polarizable and capable of engag-
ing the carbonyl of Val127 in a halogen bond. Additionally these
ATP noncompetitive MEKi’s all have a polar group capable of
engaging the terminal phosphate of ATP and Lys97. This polar

4716
K. D. Robarge et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4714–4723
H
O
H
O
O
HN
O
F
H
N
A
B
Br
F
I
F
PD318088 Chemical Structure
Figure 1. The key interactions of biaryl aniline MEK inhibitors: crystal structure of MEK1 in complex with ATP (green) and PD318088 (orange) (PDB code 1S9J). Figure 1
adapted with permission from Ref. 1. Copyright 2012 American Chemical Society.
Figure 2. As predicted based on ‘fit’ to the MEK pharmacophore model, 6,5 heterobicyclic isobenzothiazole G-155 exhibits increased cell proliferation potency versus its
azabenzofuran comparator. HCT116 cell line was derived from malignant human colon cancer cells. A375 cell line was derived from malignant human melanoma cells.
potency versus their direct aza-benzofuran comparator (G-963),
the 5,6 aza-benzofuran core was deprioritized.
Our qualitative met ID studies indicated G-868 was stable in rat
(82%). Some metabolism occurred on the hydroxamate (10% acid,
5% amide, see Supplemental material; in mouse there is some
turnover-NADPH, which appears to be hydrolysis to the acid and
accounts for 1.2% of total). Even though our met ID data did not
indicate C8 as a site of metabolism, MetasiteTM23,24 did (Fig. 4).
The potency of the 6,5 heterobicyclic MEKi’s indicate the
positioning of the pharmacophore elements, including the Ser212
H-bond acceptor nitrogen, are well tolerated in the MEK binding
pocket. The polar functionality (A-ring dependent) are positioned
optimally as shown by the X-ray co-crystal structures in Figure 3.
The imidazopyridine class of MEKi’s replaced the benzoiso-
thiazoles with the 5-imidazo[1,5-a] pyridines preferred versus
the 8-imidazo[1,5-a] pyridines. G-868 was chosen as our new lead.
G-868 exhibited improved predicted metabolic stability in dog and
cyno microsomes, low total CL in vivo, longer IV half life, higher
AUC and %F versus the its direct comparator 8-imidazo[1,5-a]
pyridine G-606 (Table 2). The G-868/G-606 comparison and
conclusion is representative of the bigger data set of 5-imidazo
versus 8-imidazo[1,5-a] pyridines. Note that both the 5 and
8-imidazo[1,5-a] pyridines inhibit the hERG channel (Table 2).
For the 5-imidazo[1,5-a] pyridines we observed poor in vitro/
in vivo CL correlation (IV/IV, see Supplemental material) in rat
and G-868 was no exception. Despite efforts to understand the
IV/IV disconnect, a satisfactory explanation was elusive and a
decision was made to put more weight on in vivo studies in rat.
Thus, G-879 (HCT116 IC50 = 0.14 lM), containing a C8-fluorine
was synthesized. From a cell proliferation potency standpoint,
the C8 fluorine was not detrimental to potency. However, consis-
tent with the Metasite analysis, C8 fluorine incorporation lowered
in vivo CLp, increased half life and oral exposure in rat, consistent
with blocking a site of metabolism. Unfortunately CYP2C9 inhibi-
tion activity for G-879 increased sevenfold.
Encouraged it was possible to reduce total and free drug clear-
ance (G-879 ppb rat = 97%, free drug CL = 55 mL/min/kg vs
1320 mL/min/kg for G-868, ppb rat = 99%), increase T1/2 and oral
exposure by blocking C8, a second Metasite analysis (Fig. 5) indi-
cated that incorporation of a nitrogen at C7 would block metabo-
lism at both C7 and C8 (Fig. 5). Incorporation of a nitrogen at C7
of the imidazo[1,5-a] pyridine core was also anticipated to have
additional benefits by lowering the clogP/clogD and decreasing
the pKa for the imidazo[1,5-a] pyrazines. Decreasing lipophilicity

K. D. Robarge et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4714–4723
4717
Table 1
Increasing the polarity of the 6,5 heterobicyclic core lowers lipophilicity as reflected in measured logD @ pH 7.4
2C9/2C19 cytochrome P450 inhibition is decreased and cell proliferation potency maintained. Target EC50 for inhibition of cell proliferation was <0.050
lM (HCT116) and
<0.005 lM (A375).
Figure 3. Imidazo[1,5-a] pyridine binding mode. Left panel-co-crystal structure of
a 5,6 azabenzofuran (2.8 Å resolution; PDB code 4U80) overlayed with a 6,5
imidazopyridine (2.7 Å resolution; PDB code 4U81). The 6,5 imidazopyridine N is suitably placed for an H-bond with Ser212 while maintaining the key hydroxamate and
anilino pharmacophoric elements in a remarkably similar position as in G-925. Right panel-co-crystal structure of a 6,5 benzimidazole (AZD6244, 2.8 Å resolution; PDB code
4U7Z) overlayed with the same 6,5 imidazopyridine (PDB code 4U81). It is noteworthy that in the 6,5 imidazopyridine co-crystal structure the attachment point of the five-
membered ring heterocycle is shifted towards the B-ring versus its attachment point in AZD6244. Despite the shift, the imidazopyridine nitrogen is positioned to make a
robust H-bond interaction to the Ser212 while maintaining the key hydroxamate and anilino pharmacophoric elements in a remarkably similar position as in AZD6244. These
X-ray co-crystal structures support the hypothesis from modelling that novel 6,5 heterobicycles would result in MEKi’s that would fit well in the MEK binding pocket and
have a high probability for equivalent or improved potency versus the 5,6 azabenzofuran core.

4718
K. D. Robarge et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4714–4723
Table 2
Comparative potency, liver microsome stability (H = human, R = rat, M = mouse, D = dog, C = cyno), rat PK and hERG patch clamp IC50’s for G-868 (representative of the
5-imidazo[1,5-a] pyridine class of MEKi’) and G-606 (representative of the 8-imidazo[1,5-a] pyridine class of MEKi’s)
IV/PO dose = 1/5 mpk. AUC for G-868 (IV/PO dose = 1/10 mpk) was normalized to a 5 mpk dose. IV/PO formulations = solution (60%PEG400/40%H2O)/solution(60%PEG400/
40%H2O) for all compounds dosed. Dose volume = 5 mL/kg for all compounds dosed.
Liver microsomal stability ranges utilized for stable/moderate/labile designations; (CLhep) mL/min/kg; stable (H/R/M) = <6.2:<17:27; moderate (H/R/M) = 6.2–15:17–39:27–
63; labile (H/R/M) = >15:>39:>63.
Figure 4. C8 Fluorine substitution improved CLp, T1/2, and oral exposure (PO dose 5 mpk, rat) but increased CYP2C9 inhibition activity 7-fold. The more intense the red circle,
the higher probability of metabolism (color gradient shown on bar on right and indicates low probability to high probability of metabolism taking place on circled carbon).
The major metabolic site is indicated by a blue circle. For G-868, the MetaSite software identified C8 as the major metabolic site.
along with decreasing basicity would increase the probability of
improving the hERG IC50 versus G-868. Based on met ID data, the
hydroxamate was the major metabolic site of G-868 and one way
to modulate clearance would be to make the hydroxamate less
prone to hydrolysis. However, it was difficult to predict a priori
how the N7 nitrogen and lower lipophilicity would effect the
metabolism of the hydroxamate. G-593 was synthesized to
determine if an imidazo[1,5-a] pyrazine could further improve on
the imidazo[1,5-a] pyridine G-868 (Fig. 6).
As shown in Table 3, the incorporation of a nitrogen at C7 (G-
593) lowered the clogP/clogD versus G-868. Both MOKA and the
ACD software predicted a lower pKa, though the magnitude of
decrease is different. The measured pKa’s validated the predictions
as the pKa for G-593 is two units lower versus G-868. However,
despite the lower clogP/clogD and decreased pKa, there is no
change in the hERG IC50. The measured logD (@ pH 7.4) shows iden-
tical values for G-868 and G-593, a possible reason for lack of
improvement in the hERG IC50. That the overall hydrophobicity,
as reflected in logD measurements, is identical may reflect
‘‘buried’’ polar surface area in G-593 if the hydroxamate NH is
engaged in an H-bonding interaction with N7. Such an interaction
may not be reflected in the logD. The working hypothesis that
came out of this result was that for the imidazo[1,5-a] pyridine
MEKi’s (Table 3) any improvement in hERG IC50 is likely to be dri-
ven by decreased lipophilicity (i.e., lowering measured logD). The
imidazole nitrogen in G-868 is not very basic to begin with (pKa
4.30), so lowering its pKa does not impact the hERG IC50
.
Though the anticipated improvement in hERG IC50 was not real-
ized, further exploration of the imidazo[1,5-a] pyrazines was con-
ducted. As shown in Table 4, the cell proliferation potency lost by
the addition of a nitrogen at the 7 position could be regained with a
more hydrophobic side chain without increasing the logD (G-593
vs G-327). G-593 and G-327 exhibited higher total CL and lower
exposure than G-868. However, unlike the imidazo[1,5-a] pyridine
class of MEKi’s, the in vivo clearance was in line with that pre-
dicted from in vitro RLM stabilities. The lower total clearance
exhibited by G-868 may be due in part to its high plasma protein
binding .The most pronounced effect of the N7 modification was

K. D. Robarge et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4714–4723
4719
decreasing plasma protein binding. Plasma protein binding was
decreased across species (3–9%, 7% lower in rat). The nitrogen in
G-593 increased metabolic stability and so lowered the unbound
clearance (G-593 600 mL/min/kg versus G-868 1320 mL/min/kg;
ppb (rat) G-868 = 99%, G-593 = 92%). Permeability in the A–B direc-
tion increased for G-593 with reduced efflux (B–A/A–B = 0.7 vs 2.60
for GDC-868).The improved permeability may be a result of an
effective decrease in polar surface area due to the hydroxamate
NH engaging in an H-bonding interaction with N7. While such an
interaction may not be reflected in the logD it could result in
improved permeability. Lowering of the plasma protein binding
resulting in a lowering of the free drug clearance and improved
permeability demonstrated C7 nitrogen modification could influ-
ence drug like properties in a positive way.
Encouraged by the improved permeability and the decreased
free drug clearance we revisited our working hypothesis that we
could improve the hERG IC50 by lowering logD. It was reasoned that
adding a more polar hydroxamate side chain would lower the logD.
Additionally, spreading polarity throughout the molecule is one
strategy that has been utilized to improve hERG by destabilizing
the interaction with the lipophilic cavity of the hERG channel.25
Our microsome stability calculations indicated improvement in
Figure 5. Metasite analysis showing that a nitrogen incorporated at the C7 position
removes two potential sites of metabolism. Note the major site of metabolism
predicted by MetaSite is now the imidazo ring of the imidazo[1,5-a] pyridine core.
Figure 6. Imidazo[1,5-a] pyrazines G-593/G-327/G-479.
Table 3
Comparative cell proliferation potencies, clogD/clogP, logD @ pH 7.4, calculated and measured pKa’s and hERG patchclamp IC50’s for imidazo[1,5-a] pyridine G-868 and
imidazo[1,5-a] pyrazine G-593 Illustrates the value of obtaining measured log D’s on a routine basis (RSD for Log D = 0.30)
For the imidazo[1,5-a] pyrazine core, the log D may exert a bigger influence on the hERG IC50 than decreasing pKa.

4720
K. D. Robarge et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4714–4723
Table 4
Comparative potency, cLogP/LogD@ pH 7.4, thermodynamic solubility, liver microsome stability, rat PK, MDCK permeability and plasma protein binding (ppb, human, rat, mouse)
for novel 6,5-heterobicyclic MEKi’s G-868 and G-606 (imidazo[1,5-a] pyridine class), and G-593/G-327/G-479 (imidazo[1,5-a] pyrazine class). IV/PO dose = 1/5 mpk and 1/10 mpk
G-868
G#
structure
Cell prolif IC50
HCT116/A375
(l
M) clogP LogD TS
(pH 6.5,
g/mL)
Mic(H/R/M)
(mL/min/kg) (mL/min/kg),
PK(rat)CLp/CL(free) MDCK (A–B;B–A; ppb
hERG patch
clamp IC50 (lM)
B–A/A–B)
(H/R/M)
(%)
l
F(%), AUC (
l
M⁄h)
10À6 cm/s
2.3
2.05
14:38:64
M/M/L
13.2:1320
100,20*
0.049:0.004
0.035:0.002
0.120:0.008
41(cryst)
6.20:16.5:2.6
20.4:14:0.7
19.7:14:0.7
99.5:99:94
4.4
2.3
2.05
15:40:75
M/M/L
14:636
6,37
110 (amorph)
150 (amorph)
98.5:97.8:95.1 2.1
1.6
2.10
11:41:54
M/L/M
48:600
38,1.48
91:92:89
3.3
2.1
2.2
10:44:56
M/L/M
44:800
15,0.61
0.052:0.005
0.107:0.007
48 (amorph)
12 (amorph)
ND
90:94.5:91
86:87:87
3.3
14
0.95
1.1
8:19:29
S/M/M
12:92
65,9.2
7.9:8.7:1.3
Dose volume = 5 mL/kg for all compounds dosed.
*
AUC normalized to a 5 mpk dose for G-868. IV/PO formulations = solution (60%PEG400/40%H2O)/solution (60%PEG400/40%H2O) for all compounds dosed.
metabolic stability across species with the additional polarity.
Incorporation of the diol hydroxamate sidechain in G-479 lowered
the measured logD and an improvement in the hERG IC50 was real-
ized (Table 4). As predicted by in vitro RLM’s (the pyrazines exhib-
ited an excellent IV/IV correlation for rat), G-479 exhibited
improved metabolic stability in rat, lower plasma protein binding
across species, improved MDCK permeability (less efflux), lower
free drug CL (92 mL/min/kg vs 1300 mL/min/kg for G-868), and
increased LLE (Supplemental material) compared with G-868. The
in vivo total CLp in rat was equivalent to G-868, the exposure and
%F much improved versus G-593 and G-327. The oral exposure of
G-479 was only two fold lower than that of the lead imidazo
[1,5-a] pyridine G-868. Since the secondary hydroxyl of the diol
hydroxamate had no effect on cell proliferation potency (G-593
120 nM vs G-479 107 nM), the potency loss suffered by G-479 can
be rationalized by a weaker interaction with the Ser212 nitrogen
due to the less basic H-bond accepting nitrogen.
taken on directly. This result also indicated the imidazo[1,5-a] pyr-
azine core is stable under acidic but not basic reaction conditions.
(e)-(h) Crude material was carried through steps (e)-(h) and
purification performed on intermediate 9.
(i) It was necessary to hydrolyze the ester using tin mediated
hydrolysis conditions at 85 °C34 as the more standard hydrolysis
conditions of LiOH/THF/MeOH/H2O resulted in decarboxylation.
This result was consistent the previous observation that the
imidazo[1,5-a] pyrazine core was stable under acidic reaction
conditions (even with heating to 85 °C!!) but not basic reaction
conditions.
(j) Changing the base from DIPEA to 4-methylmorpholine
improved the yield of the coupling reaction 4.5 fold (11–50%).
In planning for success and scalability, there was precedent in
the MEK patent literature for hydroxamate formation directly from
the ester (74% yield on a 200 g scale).38 While the imidazo[1,5-a]
pyrazines were prepared on 50–200 mg scale with the final step
being an EDCI/HOBT coupling, we did also explore hydroxamate
formation directly from the methyl ester (Scheme 2). When pre-
paring the hydroxamate from the methyl ester, it was critical to
use three equivalents of the LHMDS base. Two representative
examples are shown below.
A plot of the biochemical potency versus logD and clogP (over-
all hydrophobicity) can be found in the Supplemental material.
Both the LLE versus measured logD and calculated cLogP plots
show a progression of the 6,5 heterobicyclic MEKi’s towards
greater LLE. Imidazo[1,5-a] pyrazine G-479 gained 1.5 units in
LLE and is the most ligand efficient 6,5 heterobicycle described in
this manuscript.
The imidazo[1,5-a] pyrazines were synthesized by the following
10 step synthetic route (Scheme 1).
Some interesting features of the chemistry included the
following:
(a) The reaction did not work at all using LDA obtained from
commercial sources. It was necessary to freshly prepare the LDA
to obtain 2. (b) In the conversion of 2 to 3, if LDA was used as
the base it was prepared fresh since that was considered best
practice. However, commercial LHDMS (Aldrich) could be used as
the base with comparable yields.
(e) In the synthesis of G-868 the reduction of the nitrile to the
benzyl amine was accomplished in good yield by a CoCl2/NaBH4/
MeOH reduction.20,21 Applying the same procedure to 5 gave no
product. In the case of pyrazine 5, the major product isolated from
the CoCl2/NaBH4 mediated reduction of 5 was the 2-F 4-TMS ani-
line. This result necessitated the screening of several reduction
conditions to find one that cleanly transformed 5 into 6. It was
found that it was absolutely necessary to run the hydrogenation
reaction in glacial acetic acid to achieve a high yield of crude 6
(presumably the acetate salt) which was of suitable purity to be
Leveraging X-ray crystal structure data for G-925 gave rise to
novel 6,5-heterobicyclic cores.39 The starting point for this work
was the novel 6,5 heterobicyclic isobenzothiazole MEKi which
interacted with Ser212 in a monodentate fashion as did G-925
but exhibited increased cell proliferation potency.1 Unfortunately
isobenzothiazole G-155 was a potent 2C9/2C19 inhibitor which
prompted a core change to a more polar 6,5 heterobicyclic
core. The imidazo[1,5-a] pyridine core with lower lipophilicity

K. D. Robarge et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4714–4723
4721
F
H2N
Scheme 1. Reagents and conditions:26 (a) (i) 1.02 equiv freshly prepared LDA, THF, 1.0 equiv 1, À78 °C, 2 h; (ii) CO2 (xs), 41% yield. (b) (i) 2.0 equiv27
,
Si
3.0 equiv LHMDS or LDA, THF, À78 °C, 1 h; (ii) 1.0 equiv 2, THF, À78 °C, 0.5 h to rt, 63% yield. (c) 70% Toluene, 30% MeOH (0.06 M), 0 °C, 2.3 equiv TMSCHN2 (2.0 M in hexane),
warm to rt, 1 h, 71% yield, crude pdt. Compound 4 used directly. (d) DMF,28 1.0 equiv 4 (0.16 M), 1.1 equiv Zn(CN)2, 0.125 equiv Pd(PPh3), microwave, 300 W, 150 °C, 25 min,
78% yield. (e) Compound 5, glacial acetic acid (0.15 M),29 10% Pd–C (20% by wt.), H2 (1 atm, balloon), 1.5 h,30 98% yield, crude pdt. Compound 6 used directly. (f) Compound 6,
75% formic acid/25% acetic anhydride,31 rt 1.5 h,32 80% yield, crude product 7 used directly. (g) Compound 7, anhydrous DCM (0.1 M), 0 °C, 2.0 equiv iodine monochloride
(1.0 M solution in DCM), 0° to rt, 1 h,33 98% yield, crude pdt. Compound 8 used directly. (h) Compound 8, toluene (0.025 M, suspension), 4.2 equiv POCl3, 95 °C, 1 h, SiO2
purified (ISCO 35–100% EA/hex, followed by 0–30% MeOH/EA) 46% yield. (i) Compound 9, anhydr. 1,2 DCE (0.14 M), 3.5 equiv Sn(CH3)3OH, 85 °C, 1 h, 98% yield.34,35 (j)
O
O
Compound 10, anhydr. DMF (0.2 M), 1.04 equiv
or
, or
1.05 equiv HOBT, 4.8 equiv 4-methylmorpholine, 1.05 equiv EDCI, rt,
O
H2N
O
O
(S)
O
H2N
OH
H2N
17 h–3 d, 50% yield G-327.36 (k) Anhydrous DCM/MeOH (2:1), 1.5 equiv 4 M HCL in 1,4 dioxane, rt, 1 h; (ii) 6.0 equiv solid NaHCO3, rt, 15 min, then Na2SO4 filter and purify on
SiO2 (ISCO, 0–15% MeOH/DCM), 48%, 32% over coupling/deprotection steps G-593/G-479).37
Scheme 2. Preparation of hydroxamate directly from the methyl ester.
eliminated 2C9/2C19 inhibition and increased structural diversity.
A new 6,5 heterobicyclic lead emerged, the imidazo[1,5-a] pyridine
G-868.20 G-868 went on to become a clinical candidate and was
later renamed GDC-0623.19,21 The 6,5-imidazo[1,5-a] pyridine
scaffold was further optimized by incorporating a nitrogen at the
7 position to yield the 6,5-heterobicyclic imidazo[1,5-a] pyrazine
core. The introduction of the C7 nitrogen was driven by increasing
the structural diversity of the 6,5 heterobicycles while also adding
polarity in a position where potency would be retained and meta-
bolic stability improved. Even though in vivo met ID data did not
lead us down this road, testing a hypothesis based on a MetaSite
prediction resulted in the discovery of a fourth novel and potent
6,5 heterobicyclic scaffold, the imidazo[1,5-a] pyrazines. The nitro-
gen at the 7 position did not result in an improvement in hERG IC50
as expected. The learning here was that lowering the lipophilicity
(as reflected in lowering the logD) is more important for improving
hERG than pKa. The N7 modification did have an effect on physico-
chemical properties by reducing the plasma protein binding across
species. The nitrogen increased metabolic stability and so lowered
the unbound clearance. A nitrogen at C7 was not sufficient in itself
to improve upon the oral exposure of G-868 (GDC-0623).
Combining N7 with a diol hydroxamate which spread the polarity

4722
K. D. Robarge et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4714–4723
4. Hoshino, R.; Chantani, Y.; Yamori, T.; Tsuruo, T.; Oka, H.; Yoshida, O.; Shimada,
Y.; Ari-I, S.; Wada, H.; Fujimoto, J.; Kohno, M. Oncogene 1999, 18, 813.
5. Hatzivassiliou, G.; Song, K.; Yen, I.; Brandhuber, B. J.; Anderson, D. J.;
Alvarado, R.; Ludlam, M. J.; Stokoe, D.; Gloor, S. L.; Vigers, G.; Morales, T.;
Aliagas, I.; Liu, B.; Sideris, S.; Hoeflich, K. P.; Jaiswal, B. S.; Seshagiri, S.;
Koeppen, H.; Belvin, M.; Friedman, L. S.; Malek, S. Nature 2010, 464, 431.
6. Heidorn, S. J.; Milagre, C.; Whittaker, S.; Nourry, A.; Niculescu-Divas, I.;
Dhomen, N.; Hussain, J.; Reis-Filo, J. S.; Springer, C. J.; Pritchard, C.; Marais, R.
Cell 2010, 140, 209.
throughout the molecule gave imidazo[1,5-a] pyrazine G-479 which
was improved over G-868 (GDC-0623) in many aspects (e.g., hERG,
lower ppb across species, 14Â lower free drug CL, improved(H/R/
M) microsome stability and passive permeability, Table 4). The LLE
analyses done retrospectively (Supplemental material) demon-
strates the progression of novel 6,5-heterobicycles towards
increased LLE after moving away from the 5,6-aza-benzofuran
scaffold.
7. Poulikakos, P. I.; Zhang, C.; Bollag, G.; Shokat, K. M.; Rosen, N. Nature 2010, 464,
427.
The core change to the 6,5 heterobicyclic scaffold was potency
driven and once discovered further diversified to give additional
novel and potent 6,5 heterobicyclic MEKi’s. G-479 did improve
on G-868 (GDC-0623) in many aspects but due to the two fold
potency loss in HCT116 cell proliferation assay coupled with the
two fold lower oral exposure, G-479 was not able to displace
G-868 (GDC-0623) as the development candidate.
8. Solit, D. B.; Garraway, L. A.; Pratilas, C. A.; Sawai, A.; Getz, G.; Basso, A.; Ye, Q.;
Lobo, J. M.; She, Y.; Osman, I.; Golub, T. R.; Sebolt-Leopold, J.; Sellers, W. R.;
Rosen, N. Nature 2006, 439, 358.
9. Wellbrock, C.; Karasarides, M.; Marais, R. Nat. Rev. Mol. Cell Biol. 2004, 5, 875.
10. Flaherty, K. T.; Infante, J. R.; Daud, A.; Gonzalez, R.; Kefford, R. F.; Sosman, J.;
Hamid, O.; Schuchter, L.; Cebon, J.; Ibrahim, N.; Kaduchadkar, R.; Burris, H. A.,
III; Falchook, G.; Algazi, A.; Lewis, K.; Long, G. V.; Puzanov, I.; Lebowitz, P.;
Singh, A.; Little, S.; Sung, P.; Allred, A.; Ouellet, D.; Kim, K. B.; Patel, K.; Weber, J.
N. Eng. J. Med. 2012, 367, 1694.
This work is yet another example in the medicinal chemistry lit-
erature where small structural changes can have big consequences.
The 6,5 heterobicycles were enabled by the crystal structure of the
5,6 aza-benzofuran G-925.1 Once the discovery of the first 6,5-
heterobicyclic core was made, it offered an opportunity to utilize
a successful medicinal chemistry strategy (especially for kinase
inhibitors) of incorporating nitrogen atoms in specific places to
add polarity, increase structural diversity, improve potency, PK,
and properties and give rise to novel MEKi scaffolds. Use of the
predictive tool MetaSite was rewarded by the discovery of
imidazo[1,5-a] pyrazine G-479. The design strategy taken was a
successful one as demonstrated by the discovery of clinical
candidate GDC-0623 which has progressed to phase 1b.
11. Clinical Trials.gov NCT01138085 Safety, PK of AKT and MEK Combination
(GSK1120212 and GSK2141795) Currently recruiting patients.
12. Haagensen, E. J.; Kyle, S.; Beale, G. S.; Maxwell, R. J.; Newell, D. R. Br. J. Cancer
2012, 106, 1386.
13. Shimizu, T.; Tolcher, A. W.; Papadopoulos, K. P.; Beeram, M.; Rasco, D. W.;
Smith, L. S.; Gunn, S.; Smetzer, L.; Mays, T. A.; Kaiser, B.; Wick, M. J.; Alvarez, C.;
Cavazos, A.; Mangold, G. L.; Paitnaik, A. Clin. Cancer Res. 2012, 18, 2316.
14. Menzies, A. M.; Long, G. V. Ther. Adv. Med. Oncol. 2013, 5, 278.
15. Wagle, N.; Van Allen, E. M.; Treacy, D. J.; Frederick, D. T.; Cooper, Z. A.; Taylor-
Weiner, A.; Rosenberg, M.; Goetz, E. M.; Sullivan, R. J.; Farlow, D. N.; Friedrich,
D. C.; Anderka, K.; Perrin, D.; Johannessen, C. M.; McKenna, A.; Cibulskis, K.;
Kryukov, G.; Hodis, E.; Lawrence, D. P.; Fisher, S.; Getz, G.; Gabriel, S. B.; Carter,
S. L.; Flaherty, K. T.; Wargo, J. A.; Garraway, L. A. Cancer Discov. 2014, 4, 61.
16. Price, S. Expert Opin. Ther. Pat. 2008, 18, 603.
17. Ohren, J. F.; Chen, H.; Pavlovsky, A.; Whitehead, C.; Zhang, E.; Kuffa, P.; Yan, C.;
McConnell, P.; Spessard, C.; Banotai, C.; Mueller, W. T.; Delaney, A.; Omer, C.;
Sebolt-Leopold, J. S.; Dudley, D. T.; Leung, I. K.; Flamme, C.; Warmus, J.;
Kaufman, M.; Barrett, S.; Tecle, H.; Hasemann, C. A. Nat. Struct. Mol. Biol. 2004,
11, 1192.
Acknowledgments
18. Fischmann, T. O.; Smith, C. K.; Mayhood, T. W.; Myers, J. E., Jr.; Reichert, P.;
Mannarino, A.; Carr, D.; Zhu, H.; Wong, J.; Yang, R.; Le, H. V.; Madison, V. S.
Biochemistry 2009, 48, 2661.
The authors gratefully thank the purification and analytical
groups within Genentech Small Molecule Drug Discovery for their
support. We also thank Dr Joseph Lyssikatos for helpful discussions
and encouragement in making the hydroxamate directly from the
methyl ester as described in the patent literature, Dr. Mark Zak for
his suggestion to try the trimethyl tin mediated hydrolysis of 9 after
the more conventional hydrolysis procedure unexpectedly gave
exclusively the decarboxylation product. Dr. Mark Zak and Dr.
Anthony Estrada provided helpful discussions on the experimental
details of the methyl ester hydrolysis. We acknowledge the use of
synchrotron X-ray sources at APS (Advanced Photo Source-Argonne
National Laboratory, outside Chicago IL) and SSRL (Stanford Syn-
chrotron Radiation Lightsource, Palo Alto CA) supported by the
Department of Energy’s (DOE) Office of Science. The SSRL Structural
Molecular Biology Program is also supported by DOE’s Office of Bio-
logical and Environmental Research, and by the National Institutes
of Health (NIH). Refinement statistics for the structures in Figure 3
are summarized in the Supplementary material.
19. Hatzivassiliou, G.; Haling, J. R.; Chen, H.; Song, K.; Price, S.; Heald, R.; Hewitt, J.
F. M.; Zak, M.; Peck, A.; Orr, C.; Merchant, M.; Hoeflich, K. P.; Chan, J.; Luoh, S.-
M.; Anderson, D. J.; Ludlam, M. J. C.; Wiesmann, C.; Ultsch, M.; Friedman, L. S.;
Malek, S.; Belvin, M. Nature 2013, 501, 232.
20. G-868 appeared in the patent literature in 2009: Price, S.; Heald, R.; Lee, W.;
Zak, M. E.; Hewitt, J.; Frances M. WO 85983 A1, 2009.
21. For the first public disclosure of the structure of GDC-0623, see Supplemental
material in Ref. [19].
22. G-606 appeared in the patent literature in 2009: Price, S.; Heald, R.; Savy, P. P.
A. WO 85980 A1, 2009.
23. Caron, G.; Ermondi, G.; Testa, B. Pharm. Res. 2007, 24, 480.
24. See also http://www.moldiscovery.com/publications.php MetaSite.
25. Jamieson, C.; Moir, E. M.; Rankovic, Z.; Wishart, G. J. Med. Chem. 2006, 49, 5029.
26. All reactions except (e) were run in oven dried glassware cooled under N2,
under an atmosphere of N2. Reagents were used as received from commercial
sources without any additional purification.
27. The preparation of 2-fluoro-4-(trimethylsilyl)aniline, was first disclosed in:
Goutopolis, A.; Askew, Jr., B. C.; Chen, X.; Srinivasa, K.; Yu, H. WO123936 A1,
2007. See also Price, S.; Heald, R.; Lee, W.; Zak, M. E.; Hewitt, J.; Frances M. WO
85983 A1, 2009; WO3025A1, 2010; WO3022A1, 2010; WO93013A1, 2009.
28. A fresh bottle of anhydrous DMF was degassed by sparging with nitrogen for
2 h prior to the running of the cyanation reaction.
29. Prior to adding the 10% Pd–C catalyst, the round bottomed flask was evacuated
under house vacuum and filled with nitrogen 3Â.
Supplementary data
30. Reaction times longer than 1.5 h (i.e., 3 h) resulted in over-reduction products
being formed.
31. Concd = 0.16 M, and a suspension resulted. The reaction mixture to form (7)
never became homogeneous.
32. After extractive workup (EtOAc, satd NaHCO3/water/brine wash, Na2SO4
drying), crude pdt. (7) was azeotroped with DCM to give an orange foam of
suitable purity by LC/MS to be used directly without purification for the next
step.
33. After extractive workup (EtOAc, satd NaHCO3/water/brine wash, Na2SO4
drying), crude pdt. (8) was azeotroped with DCM to give a yellow sold of
suitable purity to be carried on directly to the next step.
34. Nicolaou, K. C.; Zak, M.; Safina, B. S.; Estrada, A. A.; Lee, S. H.; Nevalainen, M. J.
Am. Chem. Soc. 2005, 127, 11176.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2014.08.
008.
References and notes
1. Heald, R. A.; Jackson, P.; Savy, P.; Jones, M.; Gancia, E.; Burton, B.; Newman, R.;
Boggs, J.; Chan, E.; Chan, J.; Choo, E.; Merchant, M.; Rudewicz, P.; Ultsch, M.;
Weismann, C.; Yue, Q.; Belvin, M.; Price, S. J. Med. Chem. 2012, 55, 4594. A
reading of Ref. [1] is recommended for key background information, the
genesis of the 5,6 scaffolds and the origin and significance of isobenzothiazole
G-155 which is used as the starting point for the G-479 discovery..
2. Downward, J. Nat. Rev. Cancer 2003, 3, 11.
35. After extractive workup (EtOAc, 1 M HCl (3Â)/water/brine wash, Na2SO4
drying), the crude product (10) was treated with DCM/Et2O/hexane. Crystals
resulted which were collected by suction filtration and dried under house
vacuum and used in the hydroxamate coupling reactions.
36. Crude G-327 was purified by SiO2 chromatography (ISCO, 65–100% EA/hex,
followed by 0–50% MeOH/EA) and the resulting viscous oil treated with DCM
(1 drop)/ether/hexane to afford a yellow solid. The ether/hexane/DCM solution
3. Bamford, S.; Dawson, E.; Forbes, S.; Clements, J.; Pettett, R.; Dogan, A.; Flanagan,
A.; Teague, J.; Futreal, P. A.; Stratton, M. R.; Wooster, R. Br. J. Cancer 2004, 91,
355.

K. D. Robarge et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4714–4723
4723
from which the yellow solid precipitated was removed from the round
bottomed flask via pipette, transferred to a second round bottomed flask and
concentrated. The resulting oil was purified by prep HPLC to give additional G-
327 of >98% purity. The yellow solid that had precipitated after treatment with
DCM (1 drop)/ether/hexane precipitate was dried under high vacuum
overnight.
37. Crude hydroxamate was purified by SiO2 chromatography (ISCO, 0–10% MeOH/
DCM). G-573 and G-479 were crystallized from DCM/ether/hexane. Collected
yellow solid via filtration and dried under high vacuum overnight.
38. Marlow, A. L.; Wallace, E.; Seo, J.; Lyssikatos, J. P.; Yang, H. W.; Blake, J.; Storey, R.
A.; Booth, R. J.; Pittam, J. D.; Leonard, J. et al. WO044084 A2 20070419, 2007.
39. Heald, R. A.; Jackson, P.; Lyssikatos, J.; Price, Stephen; Savy, P. P. WO2010/
3025A1, 2010.