Ruthenium half-sandwich complexes as protein kinase inhibitors: Derivatization of the pyridocarbazole pharmacophore ligand

Article abstract of DOI:10.1039/b700433h

A general route to ruthenium pyridocarbazole half-sandwich complexes is presented and applied to the synthesis of sixteen new compounds, many of which have modulated protein kinase inhibition properties. For example, the incorporation of a fluorine into t

Full text of DOI:10.1039/b700433h

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Ruthenium half-sandwich complexes as protein kinase inhibitors:
derivatization of the pyridocarbazole pharmacophore ligand†
Nicholas Pagano,a Jasna Maksimoska,a Howard Bregman,a Douglas S. Williams,a Richard D. Webster,b
Feng Xuec and Eric Meggers*a
Received 3rd January 2007, Accepted 5th March 2007
First published as an Advance Article on the web 20th March 2007
DOI: 10.1039/b700433h
A general route to ruthenium pyridocarbazole half-sandwich complexes is presented and applied to the
synthesis of sixteen new compounds, many of which have modulated protein kinase inhibition
properties. For example, the incorporation of a fluorine into the pyridine moiety increases the binding
affinity for glycogen synthase kinase 3 by almost one order of magnitude. These data are supplemented
with cyclic voltammetry experiments and a protein co-crystallographic study.
Introduction
An important objective for the discovery of compounds with new
and unique biological activities is the development of methods
for the straightforward synthesis of molecular scaffolds with
defined three-dimensional shapes. Towards this goal, we recently
started a research program that explores the scope of using metal
complexes.1–8 In our concept, structures are built from a metal
center with the main function of the metal being to organize the
organic ligands in three-dimensional space.
As a proof-of-principle study, we have designed over the
last two years organometallic inhibitors for protein kinases by
using the class of ATP-competitive indolocarbazole alkaloids
(e.g. staurosporine,9 Fig. 1) as a lead structure.2–8 For this, we
replaced the indolocarbazole alkaloid scaffold with simple metal
complexes such as 1 and 2 in which the main features of
the indolocarbazole aglycon are retained in the metal-chelating
pyridocarbazole ligand, whereas the carbohydrate is replaced
by a ruthenium fragment. Although structurally related, these
new scaffolds turned out to possess properties which are distinct
from their parent indolocarbazole alkaloids. For example, whereas
staurosporine is a highly unselective protein kinase inhibitor, the
half-sandwich compounds 1 and 2 show a remarkable preference
for just a few kinases, especially GSK-3 and Pim-1.
In order to modulate the potency and selectivity profiles
of organometallics 1 and 2, as well as their pharmacokinetic
properties for future in vivo studies, we were interested in func-
tionalizing this scaffold. The pyridocarbazole heterocycle is a main
pharmacophore which is designed to bind in the adenine pocket of
Fig. 1 Staurosporine as a lead structure for functionalized ruthenium
pyridocarbazole half-sandwich complexes 1–3. All complexes are racemic
but only the (S)-isomer is shown.
the ATP binding site and thus constitutes an important target for
modifications. Previous synthetic approaches from our group to
aDepartment of Chemistry, University of Pennsylvania, 231 South 34th
Street, Philadelphia, PA, 19104, USA. E-mail: meggers@sas.upenn.edu;
derivatize this heterocycle required cumbersome readjustments of
the reaction conditions for every new compound.4,7 We here now
report a more general and efficient synthetic strategy for the syn-
thesis of functionalized pyridocarbazole ruthenium compounds,
apply it to the synthesis of sixteen new derivatives 3a–p (Fig. 1),
and present preliminary kinase inhibition results and structural
data.
Fax: +1 215 746 0348; Tel: +1 215 573 1953
bDivision of Chemistry and Biological Chemistry, Nanyang Technological
University, Singapore 637616
cDepartment of Chemistry, National University of Singapore, 3 Science Drive
3, Singapore 117543
† Electronic supplementary information (ESI) available: Synthetic proto-
cols, analytical data of new compounds and NMR spectra. See DOI:
10.1039/b700433h
1218 | Org. Biomol. Chem., 2007, 5, 1218–1227
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
ones from our group is the ability to access a large variety
of pyridoindoles through a reliable Suzuki coupling instead of
Fischer indole synthesis.
Results and discussion
Synthesis
Derivatives of indole 8 were prepared in two or three synthetic
steps from the appropriate commercially available indoles 10a and
10b by Boc-protection to 11a and 11b, followed by introducing the
boronic acid to 8a and 8b (Scheme 2).11 These boronic acids can
then optionally be transformed into the crystalline trifluoroborates
8c and 8d by reaction with KHF2.12 The trifluoroborates 8c and
8d are much more stable compared to 8a and 8b and can simply
be stored on the bench. In addition, as discussed below, they
afford significantly higher yields in the Suzuki coupling with a-
halogenated pyridines.
A retrosynthetic analysis of ruthenium complexes 1–3 is shown
in Scheme 1. The ruthenium fragment is introduced by reacting
a tert-butyldimethylsilyl (TBS)-protected pyridocarbazole 4 with
the ruthenium half-sandwich complex 5 in the presence of a weak
base, followed by TBAF deprotection to afford the ruthenium
pyridocarbazoles 1–3. Thus, the key aspect of a general synthesis is
the formation of the TBS-protected pyridocarbazole 4. Heterocy-
cle 4 can be constructed from the reaction of pyridoindoles 6 with
TBS-protected dibromomaleimide 7 by a sequence of substitution
and photocyclization.4 Pyridoindoles are accessible by Suzuki
coupling of indole boronic acids 8 or their trifluoroborates with
a-halogenated pyridines 9.10 Thus, overall, in our new synthetic
scheme, ruthenium pyridocarbazoles 1–3 are assembled from four
easily accessible building blocks (5, 7–9) in three consecutive C–
C bond formation steps followed by a coordination chemistry
step. A major improvement of this synthetic scheme over previous
Scheme 2 Synthesis of the indole building block 8.
Subsequent Suzuki cross-coupling of indoles 8a–d with a-
halogenated pyridines 9a–j furnished pyridoindoles with the Boc-
protection group partially removed. Unable to easily separate the
two compounds in most cases, the purified mixtures in each case
were adsorbed onto silica gel and heated under high vacuum to
afford pyridoindoles 6a–m (Table 1).13 Reaction yields were in the
range of 47–91% over the two steps. In general, good to excellent
Scheme 1 Retrosynthetic analysis of 1–3.
Table 1 Pyridoindole synthesis by Suzuki couplinga
Entry
Pyridine (R2, R3, Y)
Indole
Pyridoindole (R1, R2, R3)
Yield (%)
1
2
3
4
5
6
7
8
9
10
11
12
13c
9a
9b
9c
9d
9e
9f
9g
9h
9i
9a
9b
9c
9j
F, H, Br
CF3, H, Cl
CH2OTBS, H, Br
CH3, H, Br
N(CH3)2, H, Br
OCH3, H, Br
H, Cl, I
H, Br, I
H, CH2OTBS, Br
F, H, Br
CF3, H, Cl
CH2OTBS, H, Br
CO2Z, H, Br
8a
6a
6b
6c
6d
6e
6f
6g
6h
6i
H, F, H
H, CF3, H
H, CH2OTBS, H
H, CH3, H
H, N(CH3)2, H
H, OCH3, H
H, H, Cl
H, H, Br
H, H, CH2OTBS
OBn, F, H
OBn, CF3, H
OBn, CH2OTBS, H
OBn, CO2Z, H
62
57
84
91
86
85
66
52
8a
8a
8c
8c
8c
8a
8a
8a
67
8b (8d)
8b (8d)
8b
6j
6k
6l
57 (87)b
52 (73)b
68
8b
6m
47
a Reaction conditions: first, 1.1 equiv of 8a–d, 1.0 equiv 9a–j, 0.1 equiv of Pd(0) or Pd(II) catalyst, 2.75 equiv Na2CO3, DME–H2O, reflux, overnight, then
10 : 1 mass ratio silica gel : indole, high vacuum, 80 ◦C, overnight. b Yields with trifluoroborate 8d. c Z = CH2CH2Si(CH3)3.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1218–1227 | 1219
©

View Article Online
yields were observed when using the potassium trifluoroborates
8c and 8d as the indole coupling partner. This difference can be
attributed to a suppressed formation of a bis-indole homocoupling
product which is present when using boronic acids 8a and 8b. From
these studies we can conclude that the synthesis of pyridoindoles
through Suzuki coupling with trifluoroborates is superior to the
more traditional Fischer indole synthesis with respect to generality
and accessibility of a wide variety of derivatives.
For the following reaction with TBS-protected dibromoma-
leimide 7, the benzyl-protection group located on pyridoindoles
6j–m was first replaced by a TBS group in order to simplify
a later deprotection (Scheme 3). This was accomplished by
hydrogenolysis with H2, Pd/C furnishing the phenol derivatives
12j–m, followed by the treatment with TBS-triflate in presence
of the base di(isopropyl)ethylamine to afford the TBS-protected
pyridoindoles 13j–m.
Scheme 4 Assignments of R1–R4: see Fig. 1 for compounds 3a–p. For all
intermediates, see Table 1 (a–i and n–p) and Scheme 3 (j–m).
Scheme 3 Protection group exchange of pyridoindoles 6j–m to 13j–m.
For the next two C–C bond formations, pyridoindoles 6a-i
and 13j–m were condensed with dibromomaleimide 7 as shown
in Scheme 4. This method was successfully used by Murase
et al. for the synthesis of granulatimides and their analogs.4,14
Accordingly, pyridoindoles 6a–i and 13j–m were treated with
at least two equivalents of LiHMDS and quenched with TBS-
protected dibromomaleimide 7, furnishing the relatively unstable
monobromides 14a–m in yields of 34–78%. As a trend, reactions
involving pyridoindoles bearing only electron-withdrawing groups
on the pyridine were sluggish (14a, 14b, 14g, and 14h), likely as a
result of the decreased nucleophilicity of the joined heterocycles.
Cyclization of the monobromide intermediates 14a–m by an
anaerobic photocyclization with a medium pressure Hg lamp
using an uranium filter (50% transmission at 350 nm) yielded
the corresponding pyridocarbazoles 4a–m. In the course of the
photocyclization, the solubility decreases significantly and the
compounds generally precipitate out. The major side product
of this reaction was found in the competing cyclization by
substituting the bromide with the pyridine nitrogen, leading to
pyridinium salts. This competing salt formation can be influenced
by steric effects and the polarity of the solvent. For example,
the monobromides 14g, 14h and 14i undergo photocyclization in
outstanding yields (86–96%) and it is likely that substitution at the
2-position of the pyridine heterocycle precludes the formation of
a pyridinium bromide salt due to steric hindrance of the pyridine
nitrogen. Yields also improve significantly by using a less polar
solvent such as toluene which appears to disfavor the transition
state of the salt side product.
The pyridocarbazoles 4 can now further become functionalized
after their assembly. For example, our group has had some
success in directly functionalizing the pyridocarbazole ligands
1 and 2 utilizing electrophilic aromatic substitution reactions.7
In the absence of previously existing functionality at the indole
moiety, both chlorination product 4n and bromination product
4o were isolated with exclusive addition to the 5-position of the
indole heterocycle, by reaction with NCS and NBS, respectively
(Scheme 4). In the case where this position at the indole is already
occupied, addition was directed towards the 7-position of the
indole. We have previously reported the bromination of 2 to
afford the disubstituted product 4p by reaction with PhMe3NBr3
(Scheme 4).7
Finally, the reaction of each pyridocarbazole ligand 4 with
the common ruthenium precursor 5 furnishes TBS-protected
organometallic complexes 15a–p. Despite the uncommon involve-
ment of an indole, the heterocycles 4 are reliable coordinating
ligands which lose a proton upon introduction into the coordina-
tion sphere of a metal. However, reactions with pyridocarbazoles
bearing substitutions at the 2-position of the pyridine (4g–i) are
sluggish, likely because they sterically interfere with complex
formation. It is also noteworthy that the TBS group in the
heterocycles 4 is necessary in order to increase the solubility of
the ligands in organic solvents. Ten of the sixteen complexes
were isolated and desilylated in a separate synthetic step, while
the remaining six were desilylated in situ providing the final
organometallic inhibitors 3a–p.
1220 | Org. Biomol. Chem., 2007, 5, 1218–1227
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Protein kinase inhibition and cyclic voltammetry
Screening data for these new compounds at a single concentration
against the protein kinase GSK-3 and Pim-1 are displayed in Fig. 2.
These data demonstrate that derivatizations of the indole and/or
pyridine moiety lead to significant modulations in inhibition
potencies and selectivities. Although compounds 3g, 3h, 3n, and 3o
display similar or reduced potencies compared to 1, all remaining
compounds are modestly or significantly improved inhibitors for
GSK-3 and Pim-1. Importantly, substituents also influence the
kinase selectivities. Whereas compound 3m strongly prefers Pim-
1, 3p instead possesses a switched selectivity towards GSK-3 as
has been demonstrated previously.7
Fig. 3 IC50 curves of compounds 1 and 3a against GSK-3 (a-isoform)
demonstrating the effect of substituting one hydrogen against fluorine.
The ATP concentration was 100 lM.
chemical robustness for the design of GSK-3 inhibitors to be
used as molecular probes for signaling pathways involving GSK-
3. Improving selectivities, especially with respect to Pim-1, can be
accomplished in future work by additional modifications at the
cyclopentadienyl moiety as has been demonstrated recently.8
Cocrystal structure of (S)-3a with Pim-1
We succeeded in cocrystallizing full-length human Pim-1 with
Fig. 2 Screening of ruthenium complexes 1–3 against GSK-3 and Pim-1
at a concentration of 10 nM. Staurosporine (S) was used as a reference.
the enantiomer (S)-3a and solved the structure to a resolution
15
˚
of 2.2 A. This structure is related to previous Pim-1 cocrystal
Fig. 2 also reveals that all synthesized modifications at the meta-
position of the pyridine moiety (R2) are beneficial for the inhibition
of GSK-3 and Pim-1 (3a–f). Especially interestingly, adding a
single fluorine (3a) increases the binding to GSK-3 significantly.
The concentration at which 50% of the kinase is inhibited (IC50)
was determined to be 2.5 nM for 3a at 100 lM ATP, which is a
decrease by almost one order of magnitude just by exchanging a
single hydrogen for a fluorine (Fig. 3).
In addition, as has been observed before and can also be
extracted from the screening data in Fig. 2, an OH group at the
indole R1 position generally increases the affinity of compounds
significantly (see 2, 3j–m, and 3p).3 However, it is noteworthy that
the OH group poses a potential risk in terms of enzymatic or
chemical oxidation in a biological environment. For example, a
comparison of cyclic voltammetry data for compounds 1 and 2
reveals that the addition of an OH group into the indole moiety
(2) results in an oxidative peak potential which is ca. 0.18 V lower
as compared to 1. Furthermore, whereas the oxidation of 1 is
chemically reversible at a scan rate of 1 V s−1 at 293 K, compound
2 displays complicated oxidative electrochemistry and appears
to decompose/react after initial oxidation (data not shown). In
contrast to the OH group in 2, the fluorine moiety in 3a increases
the resistance to (chemically reversible) oxidation, yielding an
oxidative peak potential that is 0.06 V higher compared to 1.
From all these data it can be concluded that a fluorine
substituent at the meta-position of the pyridine moiety (R2)
as in 3a offers a promising compromise between affinity and
structures with (R)-1 and (S)-2.6 Fig. 4 demonstrates that the
ruthenium complex occupies the ATP binding site of the protein
kinase.16 The pyridocarbazole moiety of (S)-3a is nicely placed
inside of the hydrophobic pocket formed by Leu44, Phe49, Val52,
as well as Ala65 of the N-terminal domain, and Ile104, Val126,
Leu174, and Ile185 of the C-terminal domain. In agreement with
the initial design, (S)-3a retains the hydrogen bonding pattern
of ATP by forming one characteristic hydrogen bond between
the imide NH group and the backbone carbonyl oxygen atom of
Glu121 within the hinge region. An additional water mediated
hydrogen bond is established between the maleimide group of
(S)-3a and Asp186 (Fig. 4b). The CO group points towards the
glycine rich loop and the cyclopentadienyl moiety packs against
Phe49 and Ile185, while the fluorine substituent points towards
the solvent without having any notable contact. The improved
binding affinity by introducing a fluorine (1 → 3a) may be due to
a subtle dipole effect.
Intriguingly, superposition of the binding positions of cocrys-
tallized (S)-3a and staurosporine (PDB code 1YHS) with Pim-
1 demonstrates how closely the ruthenium complex mimics the
binding mode of staurosporine.17 The pyridocarbazole ligand
perfectly mimics the position of the indolocarbazole moiety
of staurosporine while the cyclopentadienyl ring and the CO
group occupy the binding position of the carbohydrate moiety
of staurosporine (Fig. 4c). Importantly, the ruthenium center is
not involved in any direct interaction and has solely a structural
role.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1218–1227 | 1221
©

View Article Online
˚
Fig. 4 Cocrystal structure of (S)-3a with Pim-1 at 2.2 A. a) The ruthenium complex occupies the ATP binding site of Pim-1. b) The most important
interactions between (S)-3a and Pim-1. c) Superimposed relative binding positions of (S)-3a staurosporine to Pim-1 (PDB code 1YHS).
was carefully added. After stirring at room temperature for 5
Conclusions
minutes, the organic layer was separated and the aqueous layer
We have developed an efficient and general strategy for the syn-
was extracted with EtOAc. The combined organic layers were
thesis of modified pyridocarbazoles in ruthenium half-sandwich
kinase inhibitors. Preliminary kinase inhibition results demon-
strate that these modifications have a significant effect on the
kinase inhibition profiles. Future work will combine functional-
izations at the pyridocarbazole moiety with a derivatization at the
cyclopentadienyl moiety in order to generate GSK-3 and Pim-
1 inhibitors with improved selectivities and potencies, as well as
hopefully favorable in vivo properties.
washed with brine, dried using Na2SO4, filtered and concentrated
to dryness in vacuo. The crude material was adsorbed onto silica gel
and subjected to silica gel chromatography with hexanes–EtOAc
(15 : 1). The combined product eluents were dried in vacuo to
1
provide 11b (6.6 g, 93%) as a white solid. H NMR (360 MHz,
DMSO-d6): d (ppm) 7.92 (d, J = 9.0 Hz, 1H), 7.61 (d, J = 3.7 Hz,
1H), 7.46–7.31 (m, 5H), 7.21 (d, J = 2.5 Hz, 1H), 7.00 (dd, J =
9.0, 2.5 Hz, 1H), 6.61 (d, J = 3.7 Hz, 1H), 5.12 (s, 2H), 1.60 (s,
9H). 13C NMR (90 MHz, DMSO-d6): d (ppm) 154.9, 149.5, 137.7,
131.5, 129.7, 128.8, 128.2, 128.1, 127.1, 115.8, 114.1, 107.8, 105.4,
84.0, 70.0, 28.1. IR (film): m (cm−1) 2981, 2928, 1730, 1611, 1585,
Experimental
Synthesis
1470, 1453, 1374, 1347, 1334, 1286, 1268, 1163, 1150, 1119, 1083,
1018, 846, 807. HRMS calcd for C20H21NO3 (M)+ 323.1521, found
(M)+ 323.1533.
The exemplary complete syntheses of the fluorinated compounds
3a and 3j are described. See ESI† for all other compounds.
Reactions were carried out using oven-dried glassware and were
conducted under a positive pressure of argon unless otherwise
specified. NMR spectra were recorded on a DMX-360 (360 MHz)
or Bruker AM-500 (500 MHz) spectrometer. Infrared spectra were
recorded on a Perkin-Elmer 1600 series FTIR spectrometer. Low-
resolution mass spectra were obtained on an LC platform from Mi-
cromass using the ESI technique. ES-TOF spectra were measured
by Waters Micromass MS Technologies. High-resolution mass
spectra were obtained with a Micromass AutoSpec instrument
using either CI or ES ionization. Reagents and solvents were used
as received from standard suppliers. Chromatographic separations
were performed using Silia-P Flash silica gel from Silicycle Inc.
Compound 8b. A solution of 11b (6.0 g, 18.6 mmol) in
anhydrous THF (24 mL) was prepared after drying 11b overnight
under high vacuum. Following the addition of triisopropyl bo-
ra◦te (6.5 mL, 28.0 mmol), the reaction mixture was cooled to
0 C and lithium diisopropylamide (2.0 M solution in heptane–
tetrahydrofuran–ethylbenzene) (14 mL, 28.0 mmol) was added
dropwise over 1 hour. After stirring for an additional 30 minutes
at 0 ◦C, 2 M HCl (50 mL) was carefully added and allowed to
stir at room temperature for 10 minutes. The resulting layers were
separated and the aqueous layer was extracted using EtOAc. The
combined organic layers were dried using Na2SO4, filtered and
concentrated to dryness in vacuo to provide 8b (6.7 g, 99%) as a
tan solid. Attempts to further purify the crude material by column
chromatography resulted in significant decomposition of 8b.
Compound 11b. A solution of 10b (5.0 g, 22.4 mmol) in THF
(16.5 mL) was purged with argon while cooling to 0 C. To the
◦
solution was added di-tert-butyl dicarbonate (5.4 mL, 23.5 mmol)
followed by the careful addition of solid (dimethylamino)pyridine
(DMAP) (4.1 g, 33.6 mmol). Upon complete addition of DMAP,
the reaction mixture quickly solidified and the resulting white
solid was gently warmed until it began to bubble. The resulting
white suspension was stirred at room temperature overnight. The
reaction mixture was cooled to 0 ◦C and 1 M HCl (12.5 mL)
Compound 8c. A solution of 8a (24.0 g, 92.2 mmol) in MeOH
(300 mL) was purged with argon while cooling to 0 ◦C. A
saturated solution of potassium hydrogen fluoride (82.0 mL) was
slowly added and the resulting thick suspension was stirred at
room temperature for 4 hours. The light orange suspension was
evaporated to dryness in vacuo and dried under high vacuum
overnight to remove all residual water. To the resulting light
1222 | Org. Biomol. Chem., 2007, 5, 1218–1227
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
orange solid was added acetone (750 mL), and the suspension was
stirred at 40 ◦C for approximately 30 minutes. Any undissolved
material (inorganic salts) was removed by filtration and the filtrate
was concentrated to dryness in vacuo. The crude material was
redissolved in a minimal amount of acetone and the product was
precipitated with addition of ether. Collection of the precipitate
by filtration provided 8c (19.1 g, 64%) as white crystals. 1H NMR
(500 MHz, acetone-d6): d (ppm) 8.01 (d, J = 8.0 Hz, 1H), 7.39 (m,
1H), 7.09 (ddd, J = 8.4, 7.0, 1.4, 1H), 7.05 (td, J = 7.3, 1.2 Hz,
1H), 6.64 (s, 1H), 1.67 (s, 9H). 13C NMR (125 MHz, acetone-d6): d
(ppm) 152.5, 138.7, 132.6, 122.8, 122.3, 120.4, 116.3, 113.9, 83.0,
28.5 (C–B not observed). IR (film): m (cm−1) 3414, 2974, 1730,
1637, 1450, 1369, 1330, 1251, 1214, 1153, 1121, 983, 892, 847,
746. HRMS calcd for C13H14BF3NO2 (M)− 284.1070, found (M)−
284.1073.
solid. 1H NMR (360 MHz, CDCl3): d (ppm) 9.45 (s, 1H), 8.44 (d,
J = 2.8 Hz, 1H), 7.79 (ddd, J = 8.8, 4.3, 0.6, 1H), 7.66 (dd, J =
7.9, 0.9 Hz, 1H), 7.43 (m, 2H), 7.23 (ddd, J = 8.2, 7.0, 1.2 Hz, 1H),
7.13 (ddd, J = 8.0, 7.0, 1.0 Hz, 1H), 6.96 (dd, J = 2.0, 0.8 Hz, 1H).
13C NMR (90 MHz, CDCl3): d (ppm) 158.8 (dC–F, J = 255.5 Hz),
147.1 (dC–F, J = 4.0 Hz), 137.4, (dC–F, J = 24.4 Hz), 136.7, 136.1,
129.3, 124.0 (dC–F, J = 19.1 Hz), 123.5, 121.4, 120.9 (dC–F, J =
4.4 Hz), 120.5, 111.5, 100.5. IR (film): m (cm−1) 3424, 3056, 1597,
1549, 1467, 1449, 1425, 1376, 1350, 1302, 1263, 1230, 1116, 1046,
1010, 933, 909, 837, 798, 742. HRMS calcd for C13H9FN2 (M)+
212.0750, found (M)+ 212.0753.
Compound 6j using 8b.
A biphasic suspension of 8b
(2.87 g, 7.82 mmol), 9a (1.25 g, 7.10 mmol), bis(triphenyl-
phosphine)palladium(II) chloride (500 mg, 0.71 mmol) and 2 M
Na2CO3 (40 mL) in 1,2-dimethoxyethane (72 mL) and water
(8 mL) was purged with argon and refluxed overnight. The reaction
mixture was cooled to room temperature, diluted with water
and extracted with EtOAc. The combined organic layers were
washed with brine, dried using Na2SO4, filtered and concentrated
to dryness in vacuo. The crude material was adsorbed onto silica gel
and subjected to silica gel chromatography with hexanes–EtOAc
(10 : 1). The combined product eluents were isolated as a mixture
of Boc-protected pyridoindole and 6j. The resulting white solid
(1.82 g) was dissolved into CH2Cl2 and adsorbed onto silica gel
(18.2 g) using rotary evaporation. The white powder was heated
to 80 ◦C overnight under high vacuum. The silica gel was cooled
to room temperature, filtered through celite with EtOAc and the
filtrate dried in vacuo to provide 6j (1.28 g, 57% over two steps) as
a white solid. 1H NMR (360 MHz, CDCl3): d (ppm) 9.35 (s, 1H),
8.41 (d, J = 2.8 Hz, 1H), 7.75 (dd, J = 8.8, 4.3, 1H), 7.50–7.30 (m,
7H), 7.16 (d, J = 2.4 Hz, 1H), 6.98 (dd, J = 8.8, 2.4 Hz, 1H), 6.87
(d, J = 1.9 Hz, 1H), 5.12 (s, 2H). 13C NMR (90 MHz, CDCl3): d
(ppm) 158.7 (dC–F, J = 255.7 Hz) 153.9, 147.1 (dC–F, J = 3.8 Hz),
137.9, 137.4 (dC–F, J = 24.4 Hz), 136.7, 132.3, 129.7, 128.7, 128.0,
127.8, 124.0 (dC–F, J = 19.2 Hz), 120.8 (dC–F, J = 4.3 Hz), 114.8,
112.3, 104.4, 100.3, 71.1. IR (film): m (cm−1) 3438, 3095, 3051, 3025,
2884, 2849, 1598, 1549, 1462, 1448, 1422, 1374, 1347, 1290, 1264,
1228, 1207, 1158, 1114, 1035, 956, 934, 916, 837. HRMS calcd for
C20H15FN2O (M)+ 318.1168, found (M)+ 318.1183.
Compound 8d. A solution of 8b (16.0 g, 43.6 mmol) in MeOH
(240 mL) was purged with argon while cooling to 0 ◦C. A
saturated solution of potassium hydrogen fluoride (38.8 mL) was
slowly added and the resulting thick suspension was stirred at
room temperature for 6 hours. The light orange suspension was
evaporated to dryness in vacuo and dried under high vacuum
overnight to remove all residual water. To the resulting light
orange solid was added acetone (560 mL), and the suspension
was stirred at 40 ◦C for approximately 1 hour. Any undissolved
material (inorganic salts) was removed by filtration and the filtrate
was concentrated to dryness in vacuo. The crude material was
redissolved in a minimal amount of acetone and the product was
precipitated with addition of ether. Collection of the precipitate
by filtration provided 8d (9.16 g, 49%) as a white solid. 1H NMR
(500 MHz, acetone-d6): d (ppm) 7.91 (d, J = 9.0 Hz, 1H), 7.50–
7.29 (m, 5H), 7.03 (d, J = 2.5 Hz, 1H), 6.80 (dd, J = 9.0,
2.6 Hz, 1H), 6.56 (s, 1H), 5.12 (s, 2H), 1.66 (s, 9H). 13C NMR
(125 MHz, acetone-d6): d (ppm) 155.4, 152.4, 139.2, 133.7, 133.5,
129.2, 128.4, 116.8, 113.8, 111.9, 104.7, 104.6, 82.7, 70.9, 28.5 (C–
B not observed). IR (film): m (cm−1) 2975, 2870, 1721, 1615, 1544,
1450, 1372, 1327, 1253, 1187, 1158, 1128, 1096, 988, 869. HRMS
calcd for C20H20BF3NO3 (M)− 390.1488, found (M)− 390.1487.
Compound 6a.
A biphasic suspension of 8a (6.50 g,
25.0 mmol), 9a (4.00 g, 22.7 mmol), tetrakis(triphenylphosphine)-
palladium(0) (2.62 g, 2.27 mmol) and Na2CO3 (6.62 g, 62.4 mmol)
in 1,2-dimethoxyethane (227 mL) and water (25 mL) was purged
with argon and refluxed overnight. The resulting orange reaction
mixture was cooled to room temperature, diluted with water
and extracted with EtOAc. The combined organic layers were
washed with brine, dried using Na2SO4, filtered and concentrated
to dryness in vacuo. The crude material was adsorbed onto silica gel
and subjected to silica gel chromatography with hexanes–EtOAc
(15 : 1 to 10 : 1). The combined product eluents were isolated as a
mixture of Boc-protected pyridoindole and 6a. The resulting white
solid (4.5 g) was dissolved into CH2Cl2 and adsorbed onto silica
gel (45.0 g) using rotary evaporation. The white powder was heated
to 80 ◦C overnight under high vacuum. The silica gel was cooled
to room temperature, filtered through celite with EtOAc and the
filtrate concentrated to dryness in vacuo. The crude material was
adsorbed onto silica gel and subjected to silica gel chromatography
with hexanes–EtOAc (10 : 1). The combined product eluents were
dried in vacuo to provide 6a (3.00 g, 62% over two steps) as a white
Compound 6j using 8d. A biphasic suspension of 8d (268 mg,
0.625 mmol), 9a (100 mg, 0.568 mmol), tetrakis(triphenyl-
phosphine)palladium(0) (66 mg, 0.057 mmol) and Na2CO3
(165 mg, 1.56 mmol) in 1,2-dimethoxyethane (2.5 mL) and water
(0.73 mL) was purged with argon and refluxed overnight. The
resulting yellow reaction mixture was cooled to room temperature,
diluted with water and extracted with EtOAc. The combined
organic layers were washed with brine, dried using Na2SO4, filtered
and concentrated to dryness in vacuo. The crude material was
adsorbed onto silica gel and subjected to silica gel chromatography
with hexanes–EtOAc (8 : 1). The combined product eluents were
isolated as a mixture of Boc-protected pyridoindole and 6j. The
resulting white solid (209 mg) was dissolved into EtOAc and
adsorbed onto silica gel (2.09 g) using rotary evaporation. The
white powder was heated to 80 ◦C overnight under high vacuum.
The silica gel was cooled to room temperature, filtered through
celite with EtOAc and the filtrate dried in vacuo to provide 6j
(158 mg, 87% over two steps) as a white solid.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1218–1227 | 1223
©

View Article Online
Compound 12j. A suspension of 6j (1.22 g, 3.84 mmol) in
ethanol (80 mL) was purged with argon for 15 minutes. Following
thorough argon purge, palladium (10 wt% on activated carbon,
wet, Degussa type E101 NE/W) (1.22 g) was added and the
reaction mixture was carefully purged with hydrogen gas and
stirred at room temperature under an atmosphere of hydrogen
for 5 hours. The resulting dark blue reaction mixture was filtered
through celite with EtOAc and the filtrate concentrated to
dryness in vacuo. The crude material was subjected to silica gel
chromatography with hexanes–EtOAc (3 : 1 to 1 : 1). The combined
product eluents were dried in vacuo to provide 12j (720 mg, 82%)
as a yellow solid. 1H NMR (360 MHz, DMSO-d6): d (ppm) 11.31
(s, 1H), 8.69 (s, 1H), 8.57 (d, J = 2.8 Hz, 1H), 7.97 (dd, J = 8.9,
4.4 Hz, 1H), 7.77 (td, J = 8.8, 2.9 Hz, 1H), 7.24 (d, J = 8.7 Hz,
1H), 6.91 (s, 1H), 6.85 (d, J = 2.0 Hz, 1H), 6.65 (dd, J = 8.7,
room temperature overnight. The resulting dark purple reaction
mixture was carefully poured into stirring ice cold 10% HCl
(100 mL) and extracted with EtOAc. The combined organic layers
were washed with saturated NaHCO3 followed by brine, dried
using Na2SO4, filtered and concentrated to dryness in vacuo. The
crude material was adsorbed onto silica gel and subjected to silica
gel chromatography with hexanes–EtOAc (6 : 1). The combined
product eluents were dried in vacuo to provide 14a (880 mg, 47%) as
an orange solid. The unstable material was immediately subjected
to the photocyclization step.
Compound 14j. A solution of 13j (525 mg, 1.54 mmol) in THF
(4.0 mL) was prepared after drying 13j under high vacuum. After
cooling the solution to −15 ◦C, lithium bis(trimethylsilyl)amide
(1 M solution in hexanes) (4.62 mL, 4.62 mmol) was added
dropwise over 15 minutes and the resulting red suspension was
stirred at −15 ◦C for an additional 45 minutes. A solution of 7
(598 mg, 1.62 mmol) in THF (4.0 mL) cooled to 0 ◦C was added
and the reaction mixture was stirred at −15 ◦C for 20 minutes and
room temperature overnight. The resulting dark purple reaction
mixture was carefully poured into stirring ice cold 10% HCl
(50 mL) and extracted with EtOAc. The combined organic layers
were washed with saturated NaHCO3 followed by brine, dried
using Na2SO4, filtered and concentrated to dryness in vacuo. The
crude material was adsorbed onto silica gel and subjected to
silica gel chromatography with hexanes–EtOAc (6 : 1 to 4 : 1).
The combined product eluents were dried in vacuo to provide 14j
(620 mg, 64%) as an orange solid. 1H NMR (360 MHz, CDCl3):
d (ppm) 9.65 (s, 1H), 8.44 (s, 1H), 7.26 (m, 3H), 6.94 (d, J = 2.0,
1H), 6.84 (dd, J = 8.7, 2.1, 1H), 1.01 (s, 18H), 0.49 (s, 6H), 0.22
(s, 6H). 13C NMR (90 MHz, CDCl3): d (ppm) 173.5, 171.0, 158.8
(dC–F, J = 258.3 Hz) 150.7, 146.2 (dC–F, J = 4.0 Hz), 141.5, 138.0
(dC–F, J = 24.4 Hz), 136.4, 131.6, 128.2, 125.9, 123.5 (dC–F, J =
18.9 Hz), 123.2 (dC–F, J = 4.3 Hz), 119.0, 112.4, 110.9, 101.4, 26.5,
26.0, 19.1, 18.4, −4.2. IR (film): m (cm−1) 3292, 2951, 2929, 2884,
2857, 1763, 1703, 1625, 1579, 1530, 1477, 1376, 1312, 1266, 1231,
1214, 1178, 1086, 1056, 970, 894, 842, 826, 801, 780, 750. HRMS
calcd for C29H37BrFN3NaO3Si2 (M + Na)+ 652.1439, found (M +
Na)+ 652.1415.
2.2 Hz, 1H). 13C NMR (90 MHz, DMSO-d6): d (ppm) 158.0 (dC–F
,
J = 252.9 Hz), 151.1, 147.5 (dC–F, J = 3.6 Hz), 137.0 (dC–F, J =
23.9 Hz), 136.5, 132.0, 129.1, 124.2 (dC–F, J = 18.9 Hz), 121.0
(dC–F, J = 4.4 Hz), 113.2, 112.4, 104.0, 99.9. IR (film): m (cm−1)
3406, 3365, 2545, 1700, 1624, 1579, 1548, 1447, 1427, 1366, 1229,
1148, 1118, 1011, 951, 915, 839, 814, 794, 753, 733. HRMS calcd
for C13H9FN2O (M)+ 228.0699, found (M)+ 228.0698.
Compound 13j. A solution of 12j (475 mg, 2.08 mmol) in DMF
(7.5 mL) was cooled to 0 ◦C and N,N-diisopropylethylamine
(1.81 mL, 10.4 mmol) was added dropwise. After stirring for 10
minutes at 0 ◦C, tert-butyldimethylsilyl trifluoromethanesulfonate
(524 lL, 2.29 mmol) was added dropwise over 15 minutes and
the reaction mixture was stirred at room temperature for an
additional hour. The resulting yellow solution was cooled to 0 ◦C
and 1 M NH4OAc (10 mL) was carefully added. The mixture was
then diluted with water and extracted with CH2Cl2. The combined
organic layers were dried using Na2SO4, filtered and concentrated
to dryness in vacuo. The crude material was adsorbed onto silica gel
and subjected to silica gel chromatography with hexanes–EtOAc
(10 : 1). The combined product eluents were dried in vacuo to
provide 13j (640 mg, 90%) as a white solid. 1H NMR (360 MHz,
CDCl3): d (ppm) 9.29 (s, 1H), 8.41 (d, J = 2.9 Hz, 1H), 7.74 (ddd,
J = 8.8, 4.3, 0.6, 1H), 7.43 (ddd, J = 8.8, 8.1, 2.9, 1H), 7.24 (dt,
J = 8.7, 0.7, 1H), 7.06 (dd, J = 1.7, 0.6 Hz, 1H), 6.83 (dd, J =
2.1, 0.8 Hz, 1H), 6.80 (dd, J = 8.7, 2.3 Hz, 1H), 1.02 (s, 9H),
Compound 4a. An orange solution of 14a (300 mg,
0.601 mmol) in toluene (250 mL) was irradiated with a medium
pressure Hg lamp using a uranium filter (50% transmission at
350 nm) for 3 hours with constant argon flow through the
solution. The resulting orange suspension was concentrated to
dryness in vacuo, adsorbed onto silica gel and subjected to silica
gel chromatography with hexanes–EtOAc (8 : 1). The combined
product eluents were dried in vacuo to provide 4a (179 mg, 71%) as
a yellow solid. 1H NMR (360 MHz, CDCl3): d (ppm) 9.86 (s, 1H),
9.06 (m, 2H), 8.89 (d, J = 2.8 Hz, 1H), 7.65 (d, J = 8.0 Hz, 1H),
7.59 (t, J = 7.6 Hz, 1H), 7.45 (t, J = 7.4 Hz, 1H), 1.07 (s, 9H), 0.64
(s, 6H). 13C NMR (90 MHz, CDCl3): d (ppm) 175.4, 174.2, 158.1
(dC–F, J = 258.4 Hz), 141.8 (dC–F, J = 27.8 Hz), 140.3, 139.5, 135.4,
131.8, 127.7, 125.8, 122.5, 122.4, 122.2, 120.4, (dC–F, J = 5.1 Hz),
117.9 (dC–F, J = 19.3 Hz), 115.0, 111.6, 26.7, 19.4, −3.8. IR (film): m
(cm−1) 3454, 2926, 2856, 1748, 1692, 1615, 1561, 1524, 1464, 1418,
1401, 1363, 1339, 1303, 1257, 1234, 1216, 1183, 1062, 1043, 1006,
938, 898, 850, 828, 768. HRMS calcd for C23H21FN3O2Si (M −
H)− 418.1387, found (M − H)− 418.1397.
0.22 (s, 6H). 13C NMR (90 MHz, CDCl3): d (ppm) 158.7 (dC–F
J = 255.6 Hz) 149.9, 147.1 (dC–F, J = 4.0 Hz), 137.4 (dC–F, J =
,
24.1 Hz), 136.7, 132.5, 130.0, 124.0 (dC–F, J = 19.2 Hz), 120.8 (dC–F
,
J = 4.4 Hz), 118.0, 111.9, 110.5, 100.1, 26.0, 18.5, −4.2. IR (film):
m (cm−1) 3415, 3050, 2947, 2928, 2894, 2859, 1621, 1594, 1575,
1550, 1462, 1448, 1427, 1382, 1362, 1317, 1288, 1255, 1239, 1151,
1117, 1009, 965, 938, 924, 890, 862, 837, 792, 779, 754, 738, 718.
HRMS calcd for C19H24FN2OSi (M + H)+ 343.1642, found (M +
H)+ 343.1653.
Compound 14a. A solution of 6a (800 mg, 3.77 mmol) in THF
(9.5 mL) was prepared after drying 6a under high vacuum. After
cooling the solution to −15 ◦C, lithium bis(trimethylsilyl)amide
(1 M solution in hexanes) (11.3 mL, 11.3 mmol) was added
dropwise over 15 minutes and the resulting yellow suspension was
stirred at −15 ◦C for an additional 45 minutes. A◦solution of 7
(1.46 g, 3.96 mmol) in THF (9.5 mL) cooled to 0 C was added
and the reaction mixture was stirred at −15 ◦C for 20 minutes and
1224 | Org. Biomol. Chem., 2007, 5, 1218–1227
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
Compound 4j. An orange solution of 14j (300 mg, 0.477 mmol)
in toluene (250 mL) was irradiated with a medium pressure Hg
lamp using a uranium filter (50% transmission at 350 nm) for
3 hours with constant argon flow through the solution. The
resulting red solution was concentrated to dryness in vacuo and
the experiment was repeated (200 mg, 0.318 mmol). The combined
crude material was adsorbed onto silica gel and subjected to silica
gel chromatography with hexanes–EtOAc (8 : 1). The combined
product eluents were dried in vacuo to provide 4j (320 mg, 73%) as
a yellow solid. 1H NMR (360 MHz, CDCl3): d (ppm) 9.86 (s, 1H),
9.03 (dd, J = 9.4, 2.9, 1H), 8.84 (d, J = 2.9, 1H), 8.58 (d, J = 2.5,
1H), 7.46 (dd, J = 8.2, 0.5, 1H), 7.14 (dd, J = 8.7, 2.5, 1H), 1.07
(s, 18H), 0.64 (s, 6H), 0.33 (s, 6H). 13C NMR (90 MHz, CDCl3):
d (ppm) 175.4, 173.8, 158.0 (dC–F, J = 258.1 Hz) 151.2, 141.6,
141.3, 140.8, 135.5, 134.7, 131.9, 123.2, 122.7 (dC–F, J = 7.1 Hz),
121.5, 119.8 (dC–F, J = 4.8 Hz), 117.9 (dC–F, J = 19.2 Hz), 114.9,
112.0, 26.7, 26.1, 19.3, 18.6, −3.8, −4.1. IR (film): m (cm−1) 3457,
3078, 2955, 2930, 2890, 2858, 1751, 1695, 1631, 1612, 1576, 1562,
1527, 1494, 1471, 1418, 1363, 1336, 1309, 1282, 1257, 1215, 1186,
1164, 1063, 1041, 1006, 966, 899, 884, 828, 814, 781, 769. HRMS
calcd for C29H35FN3O3Si2 (M − H)− 548.2201, found (M − H)−
548.2194.
3111, 2954, 2931, 2891, 2858, 1965, 1744, 1687, 1592, 1559, 1504,
1461, 1412, 1332, 1305, 1258, 1231, 1199, 1142, 1043, 1006, 973,
948, 908, 830, 778, 734. HRMS calcd for C35H41FN3O4RuSi2 (M +
H)+ 744.1662, found (M + H)+ 744.1658.
Compound 3a. A solution of 15a (19 mg, 0.031 mmol) in
CH2Cl2 (2 mL) was purged with argon while cooling to 0 ◦C. To the
solution was added tetrabutylammonium fluoride (1 M solution in
THF) (46 lL, 0.046 mmol) and the reaction mixture was allowed
to slowly warm to room temperature. The resulting dark red
reaction mixture was cooled to 0 ◦C and glacial acetic acid (2.6 lL,
0.046 mmol) was added, allowed to warm to room temperature
and concentrated to dryness in vacuo. The crude material was
subjected to silica gel chromatography with hexanes–EtOAc (3 :
1). The combined product eluents were dried in vacuo to provide
3a (10 mg, 67%) as a purple solid. 1H NMR (360 MHz, DMSO-
d6): d (ppm) 11.13 (s, 1H), 9.51 (t, J = 2.5 Hz, 1H), 8.76 (dd, J =
9.4, 2.3 Hz, 1H), 8.64 (d, J = 7.7 Hz, 1H), 7.61 (d, J = 8.1 Hz,
1H), 7.54 (ddd, J = 8.1, 6.9, 1.2 Hz, 1H), 7.35 (ddd, J = 7.9, 6.8,
1.0 Hz, 1H), 5.55 (s, 5H). 13C NMR (90 MHz, DMSO-d6): d (ppm)
200.5, 170.5, 170.2, 156.7 (dC–F, J = 248.9 Hz), 154.1, 152.8, 146.4
(dC–F, J = 34.2 Hz), 141.1, 131.4, 126.1, 123.9, 123.4, 120.5 (dC–F
,
J = 8.6 Hz), 119.7, 117.3 (dC–F, J = 20.2 Hz), 115.8, 114.3, 111.7
(dC–F, J = 4.9 Hz), 81.7. IR (film): m (cm−1) 1954, 1743, 1692, 1564,
1523, 1494, 1464, 1407, 1338, 1292, 1255, 1223, 1199, 1105, 1066,
987. HRMS calcd for C23H13FN3O3Ru (M + H)+ 499.9984, found
(M + H)+ 499.9978.
Compound 15a.
A
suspension of ligand 4a (20 mg,
0.048 mmol), [RuCp(CO)(CH3CN)2]+PF6 (30 mg, 0.072 mmol)
and K2CO3 (7.3 mg, 0.053 mmol) in CH3CN–CH3OH (3 : 1)
(2 mL) was stirred at room temperature overnight. The resulting
dark purple reaction mixture was concentrated to dryness in vacuo,
adsorbed onto silica gel and subjected to silica gel chromatography
with hexanes–EtOAc (6 : 1). The combined product eluents were
dried in vacuo to provide 15a (19 mg, 66%) as a dark red film. 1H
NMR (360 MHz, CDCl3): d (ppm) 8.98 (dd, J = 9.3, 2.3 Hz, 1H),
8.89 (d, J = 8.0 Hz, 1H), 8.86 (br, 1H), 7.58 (t, J = 7.5 Hz, 1H),
7.46 (d, J = 8.1 Hz, 1H), 7.41 (t, J = 7.5 Hz, 1H), 5.23 (s, 5H),
1.06 (s, 9H), 0.63 (s, 6H). 13C NMR (90 MHz, CDCl3): d (ppm)
199.2, 175.4, 174.8, 157.2 (dC–F, J = 250.8 Hz), 155.3, 153.5, 144.5
−
Compound 3j. A solution of 15j (21 mg, 0.028 mmol) in CH2Cl2
(2 mL) was purged with argon while cooling to 0 ◦C. To the
solution was added tetrabutylammonium fluoride (1 M solution in
THF) (70 lL, 0.070 mmol) and the reaction mixture was allowed
to slowly warm to room temperature. The resulting dark blue
reaction mixture was cooled to 0 ◦C and glacial acetic acid (4.0 lL,
0.070 mmol) was added, allowed to warm to room temperature and
concentrated to dryness in vacuo. The crude material was subjected
to silica gel chromatography with CH2Cl2–CH3OH (50 : 1 to 35 :
1). The combined product eluents were dried in vacuo to provide 3j
(14 mg, 100%) as a blue-green solid. 1H NMR (500 MHz, DMSO-
d6): d (ppm) 11.01 (s, 1H), 9.42 (t, J = 2.6, 1H), 9.17 (s, 1H), 8.71
(dd, J = 9.4, 2.4, 1H), 8.08 (d, J = 2.5, 1H), 7.43 (d, J = 8.7,
1H), 7.07 (dd, J = 8.7, 2.5, 1H), 5.52 (s, 5H). 13C NMR (125 MHz,
DMSO-d6): d (ppm) 200.4, 170.4, 170.3, 156.5 (dC–F, J = 248.4 Hz),
153.9, 151.7, 147.0, 145.5 (dC–F, J = 34.3 Hz), 141.5, 131.7, 123.8,
120.5 (dC–F, J = 8.7 Hz), 117.2, 117.0, 116.2 (dC–F, J = 18.2 Hz),
114.8, 110.2 (dC–F, J = 4.6 Hz), 108.3, 81.6. IR (film): m (cm−1)
3445, 3278, 2922, 1953, 1750, 1710, 1592, 1556, 1502, 1472, 1408,
1336, 1260, 1207, 1056, 1022, 997, 936, 924, 872, 807, 773, 700.
HRMS calcd for C23H13FN3O4Ru (M + H)+ 515.9933, found (M +
H)+ 515.9927.
(dC–F, J = 33.9 Hz), 142.4, 133.9, 126.5, 125.8, 124.3, 121.7 (dC–F
,
J = 8.5 Hz), 120.4, 119.3 (dC–F, J = 20.0 Hz), 115.6, 115.3, 113.8
(dC–F, J = 4.9 Hz), 80.7, 26.7, 19.4, −3.7. IR (film): m (cm−1) 3107,
2955, 2929, 2856, 1961, 1744, 1686, 1559, 1519, 1502, 1469, 1411,
1342, 1306, 1257, 1225, 1204, 1046, 1007, 908, 828, 745. HRMS
calcd for C29H27FN3O3RuSi (M + H)+ 614.0849, found (M + H)+
614.0839.
Compound 15j. A suspension of ligand 4j (30 mg, 0.055 mmol),
[RuCp(CO)(CH3CN)2]+PF6 (26 mg, 0.062 mmol) and K2CO3
−
(8.6 mg, 0.062 mmol) in CH3CN–CH3OH (3 : 1) (3 mL) was
stirred at room temperature overnight. The resulting dark purple
reaction mixture was concentrated to dryness in vacuo, adsorbed
onto silica gel and subjected to silica gel chromatography with
hexanes–EtOAc (8 : 1). The combined product eluents were dried
1
in vacuo to provide 15j (33 mg, 80%) as a purple film. H NMR
Protein kinase assays
(360 MHz, CDCl3): d (ppm) 8.94 (dd, J = 9.4, 2.4, 1H), 8.80 (t,
J = 2.4, 1H), 8.37 (d, J = 2.3, 1H), 7.30 (d, J = 8.7, 1H), 7.14
(dd, J = 8.7, 2.5, 1H), 5.22 (s, 5H), 1.07 (s, 9H), 1.05 (s, 9H), 0.62
(s, 6H), 0.32 (s, 6H). 13C NMR (90 MHz, CDCl3): d (ppm) 199.2,
175.5, 174.4, 157.1 (dC–F, J = 250.3 Hz), 155.4, 150.3, 148.8, 144.0
(dC–F, J = 33.8 Hz), 142.8, 134.1, 124.8, 121.7 (dC–F, J = 8.4 Hz),
120.5, 119.2 (dC–F, J = 20.1 Hz), 115.4, 114.7, 113.0 (dC–F, J =
4.7 Hz), 80.7, 29.9, 26.8, 19.3, 18.6, −3.7, −4.0. IR (film): m (cm−1)
Protein kinases (human) and substrates were purchased from
Upstate Biotechnology USA. Ten nM concentrations of stu-
arosporine, 1, 2 and 3a–p were incubated at room temperature
in 20 nM MOPS, 30 mM MgCl2, 0.8 lg lL−1 BSA, 5% DMSO
(resulting from the inhibitor stock solution), pH 7.0, in the
presence of substrate (GSK-3a: 20 lM phosphoglycogen synthase
peptide-2; Pim-1: 50 lM S6 kinase/Rsk2 substrate peptide 2) and
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1218–1227 | 1225
©

View Article Online
Table 2 Crystallographic data and refinement statisticsa
Space group P65
kinase (experiments in Fig. 2: 0.9 nM GSK-3a, 0.8 nM Pim-1;
experiments in Fig. 3: 0.1 nM GSK-3a). After 15 minutes, the
reaction was initiated by adding ATP to a final concentration of
100 lM, including approximately 0.2 lCi lL−1 [c-32P]ATP. Reac-
tions were performed in a total volume of 25 lL. After 30 minutes,
the reaction was terminated by spotting 17.5 lL on a circular P81
phosphocellulose paper (2.1 cm diameter, Whatman), followed by
washing four times (5 minutes each wash) with 0.75% phosphoric
acid and once with acetone. The dried P81 papers were transferred
to a scintillation vial, and 5 mL of scintillation cocktail was
added, and the counts per minute (CPM) were determined with a
Beckmann 6000 scintillation counter. IC50 values were defined to
be the concentration of inhibitor at which the CPM was 50% of
the control sample, corrected by the background.
˚
Cell dimension/A
Resolution/A
a,b = 98.10, c = 80.67
˚
2.2
Total observations (unique, redundancy)
Completeness (outer shell) [%]
Rmerge (outer shell)
57076 (20939, 2.7)
93.1 (96.4)
0.104 (0.450)
6.2
0.189 (0.245)
0.019
I/r
Rwork (Rfree
)
˚
Rmsd bond length/A
Rmsd bond angle/◦
1.89
a Rmsd = root mean square deviation.
Cyclic voltammetry
Voltammetric experiments were conducted with a computer
controlled Eco Chemie lAutolab III potentiostat with 1 mm
diameter planar Pt and glassy carbon (GC) working electrodes,
a Pt wire auxiliary electrode and an Ag wire reference electrode
(isolated by a salt bridge containing 0.5 M Bu4NPF6 in CH3CN).
Potentials were referenced to the ferrocene/ferrocenium redox
couple. The electrochemical cell was thermostatted at 293 K using
and Eyela PSL-1000 variable temperature cooling bath. Data were
obtained at a scan rate of 1 V s−1 in CH2Cl2 with 0.25 M Bu4NPF6
as the supporting electrolyte. Oxidative peak potentials (Epox): 1 =
+0.54 V, 2 = ca. +0.36 V, 3a = +0.60 V.
Protein expression, purification and crystallization
The plasmid encoding full-length Pim-1 (gi33304198) was a
generous gift from Dr S. Knapp, Oxford University, Centre for
Structural Genomics Botnar Research Centre, UK. The protein
was expressed and purified as described previously.18 Briefly,
expression of protein in BL21(DE3)pLysS cells was induced with
2 mM IPTG for 5 hours at 18 ◦C. Cells were collected by
centrifugation, resuspended in 50 mM HEPES pH 7.5, 500 mM
NaCl, 5% glycerol, and lysed applying high pressure (French-
press). The lysate was purified with DEAE cellulose column (DE52
Whatmann) and Ni-NTA chromatography (Qiagen). The protein
was treated overnight with lambda phosphatase and TEV protease
to remove phosphate and His6-tag respectively. Further purifica-
tion was achieved with a mono-Q column (Amersham-Biosiences),
which separated dephosphorylated and phosphorylated fractions,
and an additional Ni-NTA affinity column to obtain high purity.
Separated dephosphorylated and phosphorylated fractions were
concentrated to 5 mg mL−1 in crystallography buffer, 50 mM
HEPES pH 7.5, 250 mM NaCl, 5% glycerol, 10 mM DTT. Ruthe-
nium complex was added to the protein from 10 mM DMSO
stock solution to the final concentration of 1 mM. Crystals of
nonphosphorylated Pim-1 with (S)-3a were grown at 4 ◦C in 4 lL
sitting drops where 2 lL of protein solution was mixed with 2 lL
of the precipitate stock containing 0.2 M magnesium chloride,
100 mM BisTrisPropane (pH 6.5), 20% PEG3350, 10% ethylene
glycol and 0.3% DMSO.
Acknowledgements
We thank the National Institutes of Health for support (1 R01
GM071695–01A1) and the University of Pennsylvania for an
interdisciplinary pilot grant from the Institute for Medicine and
Engineering and the Alzheimer’s Disease Core Center (AG 10124).
We thank the group of Prof. David W. Christianson (University
of Pennsylvania), especially Dr German A. Gomez and Dr
Luigi Di Costanzo, for their support with crystallographic data
collection and structure determination. We thank the group of
Prof. Gary A. Molander (University of Pennsylvania), especially
Daniel E. Petrillo, for providing a general protocol for the
preparation of potassium trifluoroborates from boronic acids. We
thank the University of Marburg, namely Dr Uwe Linne, for the
measurement of some HRMS.
References
Data collection and structure determination
1 For bioorganometallic and medicinal organometallic chemistry, see:
K. Severin, R. Bergs and W. Beck, Angew. Chem., Int. Ed., 1998, 37,
1634–1654; D. B. Grotjahn, Coord. Chem. Rev., 1999, 190, 1125–1141;
special issue of J. Organomet. Chem., ed. G. Jaouen, 1999, 589, 1–126;
N. Metzler-Nolte, Angew. Chem., Int. Ed., 2001, 40, 1040–1043; R. H.
Fish and G. Jaouen, Organometallics, 2003, 22, 2166–2177; R. Stodt, S.
Gencaslan, I. M. Mu¨ller and W. S. Sheldrick, Eur. J. Inorg. Chem., 2003,
1873–1882; D. Schlawe, A. Majdalani, J. Velcicky, E. Heßler, T. Wieder,
A. Prokop and H.-G. Schmalz, Angew. Chem., Int. Ed., 2004, 43, 1731–
1734; D. R. van Staveren and N. Metzler-Nolte, Chem. Rev., 2004, 104,
5931–5985; I. Ott, B. Kircher and R. Gust, J. Inorg. Biochem., 2004,
485–489; Bioorganometallics, ed. G. Jaouen, Wiley-VCH, Weinheim,
2005; Y. K. Yan, M. Melchart, A. Habtemariam and P. J. Sadler, Chem.
Commun., 2005, 4764–4776; P. J. Dyson and G. Sava, Dalton Trans.,
2006, 1929–1933; U. Schatzschneider and N. Metzler-Nolte, Angew.
Chem., Int. Ed., 2006, 45, 1504–1507.
˚
Cryoprotected crystals yielded X-ray diffraction to 2.2 A on a
Rigaku RU-200HB rotating anode X-ray generator equipped with
R-axis IV++ image plate detector. Diffraction data were indexed
and merged using the program Crystal Clear (d*Trek).19 The
structure was determined by difference Fourier technique using
a structure of Pim-1 (PDB code 2BZI; inhibitor, solvent, and
ions were removed) as an initial model for phasing.6 Iterative
cycles of refinement and manual rebuilding of the model were
performed using the program REFMAC5 and O, respectively.20,21
Data collection and refinement statistics are listed in Table 2.
Coordinates for the structure have been deposited in the Protein
Data Bank (PDB code 2OI4).
1226 | Org. Biomol. Chem., 2007, 5, 1218–1227
This journal is
The Royal Society of Chemistry 2007
©

View Article Online
2 H. Bregman, D. S. Williams, G. E. Atilla, P. J. Carroll and E. Meggers,
J. Am. Chem. Soc., 2004, 126, 13594–13595.
3 D. S. Williams, G. E. Atilla, H. Bregman, A. Arzoumanian, P. S. Klein
and E. Meggers, Angew. Chem., Int. Ed., 2005, 44, 1984–1987.
4 H. Bregman, D. S. Williams and E. Meggers, Synthesis, 2005, 1521–
1527.
11 E. Vazquez, I. W. Davies and J. F. Payack, J. Org. Chem., 2000, 67,
7551–7552.
12 E. Vedejs, R. W. Chapman, S. C. Fields, S. Lin and M. R. Schrimpf,
J. Org. Chem., 1995, 60, 3020–3027.
13 T. Apelqvist and D. Wensbo, Tetrahedron Lett., 1996, 37, 1471–1472.
14 T. Yoshida, M. Nishiyachi, N. Nakashima, M. Murase and E. Kotani,
Chem. Pharm. Bull., 2002, 50, 872–876; T. Yoshida, M. Nishiyachi, N.
Nakashima, M. Murase and E. Kotani, Chem. Pharm. Bull., 2003, 51,
209–214.
5 H. Bregman, P. J. Carroll and E. Meggers, J. Am. Chem. Soc., 2006,
128, 877–884.
´
6 J. E. Debreczeni, A. N. Bullock, G. E. Atilla, D. S. Williams, H.
Bregman, S. Knapp and E. Meggers, Angew. Chem., Int. Ed., 2006,
15 The absolute configuration at the ruthenium has been assigned
5
45, 1580–1585.
according to the priority order of the ligands being g -C5H5 > pyridine
=
=
7 G. E. Atilla-Gokcumen, D. S. Williams, H. Bregman, N. Pagano and
E. Meggers, ChemBioChem, 2006, 7, 1443–1450.
[N C, C, C)] > indole [N C, C, lone pair)] > CO.
16 S. S. Taylor and E. Radzio-Andzelem, Structure, 1994, 2, 345–355.
17 M. D. Jacobs, J. Black, O. Futer, L. Swenson, B. Hare, M. Fleming and
K. Saxena, J. Biol. Chem., 2005, 280, 13728–13734.
8 H. Bregman and E. Meggers, Org. Lett., 2006, 8, 5465–5468.
9 For co-crystal structures of staurosporine with protein kinases, see for
example: L. M. Toledo and N. B. Lydon, Structure, 1997, 5, 1551–1556;
A. M. Lawrie, M. E. M. Noble, P. Tunnah, N. R. Brown, L. N. Johnson
and J. A. Endicott, Nat. Struct. Biol., 1997, 4, 796–801; L. Prade, R. A.
Engh, A. Girod, V. Kinzel, R. Huber and D. Bossemeyer, Structure,
1997, 5, 1627–1637.
´
18 A. N. Bullock, J. E. Debreczeni, O. Y. Fedorov, A. Nelson, B. D.
Marsden and S. Knapp, J. Med. Chem., 2005, 48, 7604–7614.
19 J. W. Pflugrath, Acta Crystallogr., Sect. D, 1999, 55, 1718–1725.
20 G. N. Murshudov, A. A. Vagin and E. J. Dodson, Acta Crystallogr.,
Sect. D, 1997, 53, 240–255.
10 C. B. de Koning, J. P. Michael and A. L. Rousseau, J. Chem. Soc.,
Perkin Trans. 1, 2000, 1705–1713.
21 T. A. Jones, J. Y. Zou, S. W. Cowan and M. Kjeldgaard, Acta
Crystallogr., Sect. A, 1991, 47, 110–119.
This journal is
The Royal Society of Chemistry 2007
Org. Biomol. Chem., 2007, 5, 1218–1227 | 1227
©