Slow self-activation enhances the potency of viridin prodrugs

Article abstract of DOI:10.1021/jm800374f

When the viridin wortmannin (Wm) is modified by reaction with certain nucleophiles at the C20 position, the compounds obtained exhibit an improved antiproliferative activity even though a covalent reaction between C20 and a lysine in the active site of PI3 kinase is essential to Wm's ability to inhibit this enzyme. Here we show that this improved potency results from an intramolecular attack by the C6 hydroxyl group that slowly converts these inactive prodrugs to the active species Wm over the 48 h duration of the antiproliferative assay. Our results provide a guide for selecting Wm-like compounds to maximize kinase inhibition with the variety of protocols used to assess the role of PI3 kinase in biological systems, or for achieving optimal therapeutic effects in vivo. In addition, the slow self-activation of WmC20 derivatives provides a mechanism that can be exploited to obtain kinase inhibitors endowed with physical and pharmacokinetic properties far different from man-made kinase inhibitors because they do not bind to kinase active sites.

Full text of DOI:10.1021/jm800374f

J. Med. Chem. 2008, 51, 4699–4707
4699
Slow Self-Activation Enhances The Potency of Viridin Prodrugs
Joseph Blois,†,# Hushan Yuan,†,# Adam Smith,† Michael E. Pacold,†,‡ Ralph Weissleder,†,§,| Lewis C. Cantley,|, and
Lee Josephson*,†
Center for Molecular Imaging Research, Massachusetts General Hospital and HarVard Medical School, 149 13th Street, Charlestown,
Massachusetts 02129, Center for Cancer Research, Massachusetts Institute of Technology, 40 Ames Street, Cambridge, Massachusetts 02142,
Center for Systems Biology, Massachusetts General Hospital, Boston, Massachusetts 02114, Department of Systems Biology, HarVard Medical
School, 200 Longwood AVenue, Boston Massachusetts 02115, Department of Systems Biology, HarVard Medical School and DiVision of Signal
Transduction, Beth Israel Deaconess Medical Center, 77 AVenue Louis Pasteur, Boston Massachusetts 02115
ReceiVed April 2, 2008
When the viridin wortmannin (Wm) is modified by reaction with certain nucleophiles at the C20 position,
the compounds obtained exhibit an improved antiproliferative activity even though a covalent reaction between
C20 and a lysine in the active site of PI3 kinase is essential to Wm’s ability to inhibit this enzyme. Here we
show that this improved potency results from an intramolecular attack by the C6 hydroxyl group that slowly
converts these inactive prodrugs to the active species Wm over the 48 h duration of the antiproliferative
assay. Our results provide a guide for selecting Wm-like compounds to maximize kinase inhibition with the
variety of protocols used to assess the role of PI3 kinase in biological systems, or for achieving optimal
therapeutic effects in vivo. In addition, the slow self-activation of WmC20 derivatives provides a mechanism
that can be exploited to obtain kinase inhibitors endowed with physical and pharmacokinetic properties far
different from man-made kinase inhibitors because they do not bind to kinase active sites.
Introduction
work found that a variety of Wm derivatives made by modifica-
tion of the electrophilic C20 (see Figure 1 for numbering) with
a variety of constituent groups were inactive as PI3 kinase
inhibitors and in cell-based assays.14,23–25 These studies indicated
an essential role for an intact C20 in the bioactivity of Wm and
suggested C20 might covalently react with an amino acid in
the active site of a target enzyme. Indeed, a covalent modifica-
tion of the catalytic subunit of PI3 kinase by Wm was then
reported, followed by crystallographic studies showing a modi-
fied Wm in the ATP site.12,13 Despite these findings on the
crucial role played by the intact C20, more recent studies have
found Wm derivatives modified at C20 were markedly superior
to the starting Wm, both as PI3 kinase inhibitors and in their
antiproliferative activity.22,26
Viridins, a class of compounds isolated from fungi of which
wortmannin (Wm) is a member, have long intrigued pharma-
cologists because of their potent antiproliferative and immune
suppressive properties.1 Viridins are steroid-like molecules,
which have a ”extra” furan ring as a common, key structural
feature.1 Wm, the most commonly used viridin, has been widely
employed to define the role of PI3 kinase and the PI3 kinase
pathway in many biological systems. Wm is a nanomolar PI3
kinase inhibitor,2,3 but at higher concentrations inhibits polo-
like kinase,4 mTOR,5 DNA-dependent protein kinase,6 and
ATM.7 Unlike man-made PI3 kinase inhibitors,8–11 a specific
carbon on Wm, C20, reacts with a lysine in the ATP site of
PI3 kinase.1,12,13 Wm’s antiproliferative and immune suppressive
activities have been explored over the years beginning in
1974,14–20 and it continues to be used as a scaffold for drug
development.21,22 Although viridins have been widely used to
define the role of the PI3 kinase pathway and for drug
development, relationships between viridin chemistry (structure,
stability, chemical reactivity) on the one hand, and their
bioactivity on the other hand, are sometimes not easily
reconciled.
We hypothesized that the bioactivities of WmC20 compounds
might be related to the ability of WmC20 derivatives to form
Wm;27 that is the bioactivity of a WmC20 derivative might
reflect a combination of the rate of Wm self-activation and the
duration of the assay involved. In this view, cell-based bioassays
using long incubation times, (e.g., a 48 h antiproliferative assay),
would be more indicative of antiproliferative activity of WmC20
modified compounds in vivo, where sustained, long-term effects
are generally defined as efficacy. To investigate this possibility,
we modified a recently developed fluorescent Wm29 with various
constituent groups at C20, to obtain a panel of fluorescent
WmC20 derivatives that formed the original fluorescent Wm
at widely different rates. We then examined the activity of the
panel with a short duration PI3 kinase assay and a long duration
antiproliferative assay and employed these compounds’ fluo-
rescence to determine their fates in cultured cells. We show
that certain Wm derivatives modified at C20 (WmC20 deriva-
tives) are poor PI3 kinase inhibitors yet exhibit enhanced
potency with a 48 h antiproliferative assay due to a self-
activation mechanism that yields the active Wm. As the
chemistry of Wm and its derivatives becomes fully understood,
these unique natural products can be better exploited to define
A seeming inconsistency involves the bioactivity of Wm
derivatives made by reaction of the electrophilic C20 carbon
with nucleophiles: because a covalent reaction between the C20
of Wm and PI3 kinase is an essential part of its mechanism of
kinase inhibition, can a reaction between C20 and nucleophiles
yield WmC20 derivatives that are more potent than Wm? Early
* To whom correspondence should be addressed. Phone: (617) 726-6478.
Fax: (617) 726-5708. E-mail: josephso@helix.mgh.harvard.edu.
†
Center for Molecular Imaging Research, Massachusetts General
Hospital and Harvard Medical School.
‡
Center for Cancer Research, Massachusetts Institute of Technology.
Center for Systems Biology, Massachusetts General Hospital.
Department of Systems Biology, Harvard Medical School.
§
|
Department of Systems Biology, Harvard Medical School and Division
of Signal Transduction, Beth Israel Deaconess Medical Center.
#
Contributed equally to this work.
10.1021/jm800374f CCC: $40.75
2008 American Chemical Society
Published on Web 07/17/2008

4700 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 15
Blois et al.
Figure 1. Synthesis of WmC20 compounds. The positions of the C3, C6, C11, and C20 carbons are shown for Wm, 1. 1 was modified at C11 by
the attachment of NBD and a 6-carbon linker to obtain NBD-Wm, 3. Then 3 was then modified at C20 by the attachments of certain linkers (L’s),
to yield self-activating fluorescent WmC20 derivatives. An intramolecular attack of the hydroxyl at C6 provides a mechanism of self-activation and
Wm generation as in Figure 1 of ref 27. Structures of Linkers attached to C20 (a-e) are provided. The C20 designation indicates the point of
attachment to NBD-Wm or Wm. The physical properties of low-molecular-weight fluorescent derivatives (4a, 4b) were transformed by reaction
with a carrier amino dextran to form 6a and 6b. Oxygens at C3 and C6 flanking C20 form hydrogen bonds when Wm (or NBD-Wm) has reacted
with a lysine in the ATP binding site of PI3 kinase gamma, see discussion regarding Figure 4. Synthetic Methods: (1) see ref 29; (2) for 2a and
4a, 6-methylaminohexanoic acid (N(Me)HA), for 2b and 4b, 6-aminohexanoic acid (N(H)HA), for 2c and 4c, 4-methylaminopyridine (N(Me)Pyr),
for 2d, L-proline (Pro), for 2e, N-acetyl lysine (Lys)), triethylamine, DMSO, room temperature; (3) (a) for 5a, 2a, (for 6a, 4a; for 5b, 2b; for 6b,
4b), N-hydroxysuccinimide, 1,3-diisopropylcarbodiimide; (b) aminodextran, DMSO, PBS, pH 7.0, 37 °C.
4b): buffer A:buffer B (80:20) isocratic for 5 min, linear gradient
to buffer A:buffer B (20:80) over 30 min, then isocratic for 5 min,
gradient back to 80:20 (buffer A:buffer B) for 5 min and isocratic
for 5 min; flow: 6.0 mL/min; λmax: 408 nm. System 5 (for 4a):
buffer A:buffer B (70:30) linear gradient to buffer A:buffer B (0:
100) over 25 min, then isocratic for 5 min and gradient back to
70:30 (buffer A:buffer B) in 5 min and isocratic for 5 min; flow:
1.0 mL/min; λmax: 280 and 490 nm. System 6 (for 4b): buffer
A:buffer B (50:50) linear gradient to buffer A:buffer B (23:77) over
20 min, then isocratic for 5 min and gradient back to 50:50 (buffer
A:buffer B) in 5 min and isocratic for 5 min; flow: 1.0 mL/min;
the role of the PI3 kinase in biological systems or to obtain
compounds with acceptable efficacy and toxicity for use in vivo.
Materials and Methods
Compound Synthesis. The strategy employed and compounds
synthesized are shown in Figure 1. Compounds 3, 2a, 2b, 2d, and
2e were prepared as described.27–29 Wortmannin (Wm) was a gift
of the natural products branch of the National Cancer Institute.
Amino-dextran was a 70 kDa species from Invitrogen, Carlsbad,
CA. All reagents and solvents were of standard quality. NMR was
performed on a Varian 400 MHz and Varian Unity Inova 500 MHz
instruments with CDCl3. High resolution mass spectra were obtained
on a Micromass LCT instrument by using the time-of-flight ESI
technique. Low resolution mass spectra were collected on a Waters
Micromass ZQ MAA288 instrument. Compounds were purified by
HPLC (Varian Prostar 210 with a variable wavelength PDA 330
detector), which employed reverse phase C18 columns (VYDAC,
catalogue no. 218TP1022 for preparative work; Varian, catalogue
no. R0086200C5 for analytical work) with water (Millipore,
containing 0.1% trifluoroacetic acid) (buffer A), and acetonitrile
(containing 20% buffer A) (buffer B) as the elution buffers. System
1 (for 2c and 4c): buffer A:buffer B (70:30) linear gradient to buffer
A:buffer B (0:100) over 15 min, then gradient back to 70:30 (buffer
A:buffer B) for 3 min and isocratic for 5 min; flow: 1.0 mL/min;
λ
max: 280 and 490 nm.
Synthesis of 2c. A mixture of Wm 1 (12 mg, 0.028 mmol) and
4-(N-methylamino) pyridine (60 mg, 0.55 mmol) in anhydrous
CH2Cl2 (1 mL) was stirred at room temperature for 15 h. The
mixture was purified by preparative silica gel TLC developed by
CH2Cl2:MeOH (10:1). A dark-yellow solid was obtained. Crude
yield: 40% with purity 96% (system 1). A further HPLC purification
(system 2) was needed for both the PI3 kinase enzyme and
antiproliferative cell-based assays. A bright-yellow powder (2c)
would be obtained after lyophilization. HRMS: C29H32N2O8, expt
537.2237 (M + H+), obsd. 537.2220. LRMS: 537.0. 1H NMR
(CDCl3, ppm, 500 MHz): δ 0.86 (3H, s, C13-CH3), 1.24 (1H, br,
OH), 1.64 (3H, s, C10-CH3), 1.79-1.84 (1H, q, J1 ) 5.2 Hz, J2 )
14.3 Hz, H-12), 2.04-2.12 (4H, m, H-15, OCCH3), 2.24-2.29 (1H,
m, H-16), 2.39-2.44 (1H, q, J1 ) 7.4 Hz, J2 ) 14.3 Hz, H-12),
2.53-2.62 (1H, m, H-16), 2.89-3.02 (2H, m, H-14, OCH2),
3.11-3.17 (4H, m, H-15, N-CH3), 3.25 (3H, s, OCH3), 3.27-3.32
(1H, q, J1 ) 2.19 Hz, J2 ) 11.0 Hz, OCH2), 4.61-4.63 (1H, q, J1
) 2.5 Hz, J2 ) 6.6 Hz, H-1), 6.07-6.10 (1H, m, H-11), 7.00 (2H,
d, J ) 6.4 Hz, Py-H), 8.39 (1H, s, H-20), 8.35 (2H, d, J ) 6.4 Hz,
Py-H).
λmax: 410 and 495 nm. System 2 (for 2c and 4c): buffer A:buffer B
(70:30) linear gradient to buffer A:buffer B (0:100) over 20 min,
isocratic for 5 min, then gradient back to 70:30 (buffer A:buffer
B) for 5 min and isocratic for 5 min; flow: 4.9 mL/min; λmax: 410
and 495 nm. System 3 (for 4a): buffer A:buffer B (80:20) isocratic
for 5 min, linear gradient to buffer A:buffer B (20:80) over 30 min,
then gradient back to 80:20 (buffer A:buffer B) for 5 min and
isocratic for 5 min; flow: 6.0 mL/min; λmax: 410 nm. System 4 (for

Self-ActiVation Enhances Potency of Viridin Prodrugs
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 15 4701
Figure 2. Time course for NBD-Wm cell protein labeling when cells were incubated with C20 reactive NBD-Wm or NBD-WmC20 compounds
with various rates of self-activation. Cells were incubated with 10 µM of NBD-Wm, either as NBD-Wm or a WmC20 derivative with various rates
of self-activation for specified times and lysed. After SDS electrophoresis, NBD-modified proteins were visualized with an anti-NBD Western blot
method as described.29 Wortmannylated (NBDylated) proteins were seen with the principal bands at 35 kDa (possibly a doublet), 55 kDa, and 75
kDa. With all compounds the 55 kDa band appeared first. Time course of cell protein labeling (total area of 35, 55, and 75 kDa bands from
densitometry) was further analyzed and is shown in Figure 3A.
Synthesis of 4a. NBD-Wm 3 (43.8 mg, 0.065 mmol), 6-(N-
methylamino)-hexanoic acid hydrogen chloride (53 mg, 0.3 mmol),
and triethylamine (40 µL) were mixed in anhydrous DMSO (2 mL).
The mixture was stirred at room temperature and completed
immediately. After dilution with 50% acetonitrile in water (1:1)
before the injection, the mixture was purified by HPLC (system 3)
and gave a red powder after lyophilization. Analysis of 4a by HPLC
(system 5) showed less than 3% of NBDWm: 38.1 mg, 67.6%.
HRMS: C41H51N5O13, calcd 822.3561 (M + H+), found 822.3594.
808.4 (M + H+), 825.4 (M + NH4+), 830.4 (M + Na+). 1H NMR
(CDCl3, ppm, 400 MHz): δ 0.82 (3H, s, C13-CH3), 1.42-1.50 (4H,
m, NCH2CH2CH2CH2CH2CO), 1.52 (3H, s, C10-CH3), 1.67-1.82
(9H, m, NCH2CH2CH2CH2CH2CO, H-12), 2.22-2.42 (7H, m,
NCH2CH2CH2CH2CH2CO, H-12, H-15, H-16), 2.53-2.61 (1H, m,
H-16), 2.73-2.95 (2H, H-14, OCH2), 3.14-3.18 (1H, m, H-15),
3.23 (3H, s, OCH3), 3.39-3.68 (8H, m, br, NCH3, NCH2, OCH2,
OH), 4.08 (2H, br, ArNCH2), 4.29 (1H, d, J ) 7.6 Hz, H-1),
5.99-6.02 (1H, m, H-11), 6.11 (1H, d, J ) 9.0 Hz, ArH), 8.46
(1H, d, J ) 9.0 Hz, ArH), 8.53 (1H, d, J ) 13.9 Hz, H-20),
9.82-9.88 (1H, m, NH). 13C NMR (ppm): 16.65, 22.62, 24.28,
26.02, 26.26, 26.53, 30.42, 33.51, 33.80, 36.62, 38.63, 42.43, 42.50,
43.88, 49.97, 49.98, 55.93, 59.43, 67.51, 73.31, 77.43, 81.31, 88.41,
101.31, 129.05, 135.70, 137.09, 137.56, 145.10, 145.57, 150.83,
159.61, 166.15, 172.62, 177.22, 178.65, 218.10.
Synthesis of 4c. A mixture of NBD-Wm 3 (21 mg, 0.031 mmol)
and 4-(N-methylamino) pyridine (0.268 g, 0.6 mmol) in anhydrous
CH2Cl2 (2 mL) was stirred at room temperature for 2 days. The
mixture was purified by preparative silica gel TLC developed by
CH2Cl2:MeOH (10:1). A dark-red waxy solid was obtained. Crude
yield: 16.7% with purity 95% (system 1). A further HPLC
purification (system 2) was needed for both the PI3 kinase enzyme
and antiproliferative cell-based assays. A bright-red powder (4c)
would be obtained after lyophilization. HRMS: C40H44N6O11, expt
785.3141 (M + H+), obsd. 785.3145. LRMS: 807.4 [M + Na+].
1H NMR (CDCl3, ppm, 400 MHz): δ 0.90 (3H, s, C13-CH3),
1.44-1.49 (2H, m, J ) 7.3 Hz, NCH2CH2CH2CH2CH2CO), 1.62
1
LRMS: 822.4. H NMR (CDCl3, ppm, 400 MHz): δ 0.86 (3H, s,
C13-CH3), 1.42-1.48 (4H, m, NCH2CH2CH2CH2CH2CO), 1.56
(3H, s, C10-CH3), 1.61-1.78 (9H, m, NCH2CH2CH2CH2CH2CO,
H-12), 2.05-2.16 (1H, m, H-16), 2.22-2.47 (4H,
m
NCH2CH2CH2CH2CH2CO), 2.54-2.64 (1H, m, H-15), 2.73-2.87
(3H, m, H-12, H-14, H-16), 2.92-2.99 (1H, m, OCH2), 3.09-3.17
(1H, m, H-15), 3.22 (3H, s, OCH3), 3.30-3.60 (11H, br, NCH3,
OCH2, NCH2, OH), 4.08 (2H, br, ArNCH2), 4.48-4.49 (1H, d, J
) 6.25 Hz, H-1), 6.07-6.13 (2H, m, H-11, ArH), 8.26 (1H, s,
H-20), 8.46 (1H, d, J ) 9.0 Hz, ArH).
Synthesis of 4b. NBD-Wm 3 (13 mg, 0.02 mmol) and 6-amino-
hexanoic acid (13 mg, 0.1 mmol) were mixed in anhydrous DMSO
(1 mL). The mixture was stirred at room temperature for 1.5 h.
After dilution with 50% acetonitrile in water (1:1) before the
injection, the mixture was purified by HPLC (system 4) and gave
a red powder after lyophilization. HPLC analysis (System 6) showed
a single peak to prove its high purity; 13.5 mg, 83.9%. HRMS:
C40H49N5O13, calcd 808.3405 (M + H+), found 808.3423. LRMS:

4702 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 15
Blois et al.
Figure 3. Time courses for the behavior of fluorescent WmC20
derivatives in cells. (A) Time course for the formation of wortman-
nylated cell protein with the C20 reactive NBD-Wm 3 or NBD-WmC20
compounds. (B) and (C): Time course for cell fluorescence, (uptake of
NBD). Relative cell fluorescence, the RCF on the y-axis, was the same
for the short (B) and long time scales (C). Lines in (A) or (B) were
obtained by fitting data to the equations given in Methods, with the
relative initial rates given in Table 1. Equivalent concentrations of 10
µM Wm were used throughout.
(3H, s, C10-CH3), 1.71-1.83 (5H, m, NCH2CH2CH2CH2CH2CO,
H-12), 2.02-2.12 (1H, m, H-15), 2.24-2.34 (1H, m, H-16), 2.37
(2H, t, J ) 7.32, NCH2CH2CH2CH2CH2CO), 2.45-2.49 (1H, q,
J1 ) 7.32 Hz, J2 ) 13.7 Hz, H-12), 2.57-2.63 (1H, q, J1 ) 8.8
Hz, J2 ) 19.5 Hz, H-16), 2.66 (1H, s, OH), 2.92-2.98 (2H, m,
H-14, OCH2), 3.17-3.19 (1H, m, H-15), 3.20 (3H, s, OCH3), 3.21
(3H, s, N-CH3), 3.34-3.37 (1H, q, J1 ) 2.9 Hz, J2 ) 10.7 Hz,
OCH2), 3.46 (3H, br, NCH3), 4.12 (2H, br, NCH2), 4.73-4.74 (1H,
q, J1 ) 2.9 Hz, J2 ) 5.9 Hz, H-1), 6.10-6.13 (2H, m, H-11, NBD-
H), 7.22 (2H, d, J ) 6.8 Hz, Py-H), 8.26 (1H, s, H-20), 8.46 (1H,
d, J ) 8.8 Hz, NBD-H), 8.63 (2H, d, J ) 6.8 Hz, Py-H).
Synthesis of 5a. To an amino-dextran (300 mg, 0.00429mmol)
in 4 mL PBS, pH 7 was added to the NHS ester of 2a in 4 mL
DMSO (28.7 mg, 0.0429 mmol), and the mixture incubated for
1.5 h at 37 °C. Low-molecular-weight impurities were removed
with Sephadex G-50 chromatography in 1 mM phosphate buffer
at pH 7.0. A yellow powder was obtained after lyophilization. The
Wm attached was determined from Wm absorbance by using UV
standard curves at 418 nm. The ratio of 2a to dextran was 5.9.
Synthesis of 5b. Amino-dextran (202 mg, 0.00429mmol) in 3
mL PBS, pH 7 was added to the NHS ester of 2b in 5 mL DMSO
(56 mg, 0.0858 mmol), and the mixture incubated for 3 h at 37 °C.
Low-molecular-weight impurities were removed as above, and a
yellow powder was obtained after lyophilization. The Wm attached
was determined from Wm absorbance by using UV standard curves
at 408 nm. The ratio of 2b to dextran was 7.6.
Synthesis of 6a. To a 70 kDa amino-dextran (35 mg, 0.0005mmol)
in 1 mL PBS, pH 7 was added to the NHS ester of 4a in 4 mL
DMSO (6.6 mg, 0.00718 mmol) and the mixture incubated for 2 h
at 37 °C. Low-molecularweight impurities were removed with
lyophilization as above. The NBD-Wm 3 attached was determined
from NBD absorbance by using UV standard curves at 490 nm.
The ratio of 4a to dextran was 3.7.
Synthesis of 6b. Amino-dextran (36 mg, 0.0005 mmol, 70 kDa)
was reacted with the NHS ester of 4b (5.5 mg, 0.005 mmol) and
purified as above. The ratio of 4b to dextran was 6.3.
Dextran interfered with the PI3 kinase assay (data not shown),
so the ability of 5a, 5b, 6a, and 6b to inhibit PI3 kinase was
obtained with compounds where an amino glucose carrier replaced
the amino dextran. The NHS ester of 2a (or 2b, 4a, 4b) (0.04 mmol)
was mixed with 2-amino glucose hydrogen chloride (0.4 mmol) in
anhydrous DMSO (1 mL). After the addition of triethylamine (60
Figure 4. Binding of NBD-Wm 3 into the active site of PI3 kinase
gamma. Model is based on the published structure of Wm in the ATP
site of PI3 kinase gamma (PDB: 1E7U) with NBD adduct added. (A)
Desacyl Wm (Wm lacking the acetyl group at C11) is orange. The
NBD adduct (NBD and six carbon linker) is colored cyan, with nitrogen
in blue and oxygen in red. The NBD and the six-carbon linker protrude
out of the ATP-binding pocket into the solvent while Wm is buried
deeply in the ATP-binding pocket. (B) Another view of NBD-Wm
showing the relationship between the surface of the enzyme and NBD-
Wm. Lysine 833, covalently reacted with the C20 of Wm, is in magenta
and extends through a blue, positively charged surface. The proximity
of the surface prevents Wm derivatives modified at C20 from binding
in the ATP-binding pocket.
µL), the mixtures were incubated under 37 °C for 2.5 h. The solution
was diluted with water (1 mL) and applied to HPLC using system
5: buffer A:buffer B (90:10) linear gradient to buffer A:buffer B
(40:60) over 25 min, then isocratic for 5 min and gradient back to
90:10 (buffer A:buffer B) in 5 min and isocratic for 5 min; flow:
4.9 mL/min; λmax: 280 and 410 nm. Yield: 65-80%. HRMS: 7a:
C36H50N2O14 expt 735.3335 [M + H+], obs 735.3325; LRMS 757.4
[M + Na+]. 7b: C35H48N2O14 expt 721.3178 [M + H+], obs
721.3176; LRMS 743.4 [M + Na+]. 8a: C47H62N6O17 expt 983.4244
[M + H+], obs 983.4256; LRMS 1005.4 [M + Na+]. 8b:
C46H60N6O17 expt 991.3907 [M + Na+], obs 991.3914; LRMS
991.7 [M + Na+].
Generation of Wm or NBD-Wm 3. The half-times for Wm
formation were determined by incubating WmC20 compounds at
1.5 mM Wm equivalents in PBS, pH 6.8, and 37 °C. The mixture
was analyzed by HPLC using MHB as an internal standard (1 µg
of methyl-4-hyrdroxybenzoate). With 4a, 4b, and 4c, 10% aceto-
nitrile and 5% DMSO were added to ensure solubility of the NBD-
Wm that was formed. System 1 (for 2c and 4c), system 5 (for 4a),
and system 6 (for 4b) were used for HPLC analysis. With the
WmC20 dextran conjugates of Wm and NBD-Wm insolubility was
utilized as a separation mechanism. A solution of 5a or 6a (1.5
mM Wm equivalents) in 300 µL in PBS (pH 7.0) was incubated at
37 °C. For 5a, release was monitored by a decrease in absorbance
at 322 nm that occurs with Wm formation. For 6a, released NBD-

Self-ActiVation Enhances Potency of Viridin Prodrugs
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 15 4703
Table 1. Wm Formation and Activities of WmC20 Compoundsa
half-time of
cell protein
cell uptake,
linker
Wm or NBD-Wm
formation (h)
PI3K IC50
(µM)
antiprolif.
labeling, relative relative initial
compound R groupb@ C11
(L) b@ C20
carrierb
cIC50 (µM)
initial rate
rate
1.00
1
3
acetyl,R1
NBD,R2
R1
none
none
no
no
not applicable
not applicable
2.30 (2.13-2.42)
1.87 (1.69-2.73)
8.7 (7.83-9.75)
8.0 (6.96-9.53)
9.9 (7.29-15.53)
9.1 (7.42-11.65)
9.2 (8.54-10.08)
>100
0.0176 ( 0.017 11.4 ( 0.5
0.0242 ( 0.016 12.2 ( 0.8
1.00
2c
4c
2a
4a
2d
5a
6a
2b
4b
5b
6b
2e
N(Me)-Pyr, (c) no
(c) no
N(Me)-HA,(a) no
0.119 ( 0.011
0.262 ( 0.020
0.128 ( 0.016
0.174 ( 0.012
0.105 ( 0.011
0.458 ( 0.021
0.453 ( 0.017
>10
9.86 ( 2.52
9.99 ( 1.25
0.96 ( 0.33
3.85 ( 0.36
1.79 ( 0.18
1.05 ( 0.42
1.04 ( 0.10
27.9 ( 3.3
48.7 ( 13.3
not detected
not detected
29.8 ( 0.9
R2
R1
R2
R1
0.206
0.246
1.199
(a)
no
no
yes
yes
no
0.187
Pro, (d)
R1
R2
R1
R2
(a)
(a)
(b)
(b)
(b)
(b)
0.0656
0.00613
none
0.0661
0.0271
0.0064
no
>100
>10
R1
R2
R2
yes
yes
no
>100
>10
>100
>10
NAclys, (e)
>100
>10
a Non-C20 protected compounds(1 and 3) are given first, followed by WmC20 derivatives grouped from top to bottom by half-times of Wm formation.
b Structures of compounds, R groups at C11 and linkers L at C20, as well abbreviations used for linkers, are provided in Figure 1 and its legend. c Antiprolif.
is the antiproliferative assay.
Wm was removed by centrifugation (5 min, 13000 rpm, Eppendorf
benchtop microfuge). The absorbance of the supernatant at 480 nm
was used to obtain NBD-Wm released. Absorbances (322 or 480
nm) were fit to a first-order decay process (log absorbance versus
time), with coefficients of correlation of greater than 0.95 in all
cases. First-order decay constants were converted to half-times.
Activity-Based Bioassays. The antiproliferative assay employed
sulforhodamine B binding to determine cell mass as described.29
The PI3 kinase assay employed the gamma catalytic subunit of
PI3 kinase and was the fluorescence energy transfer assay.30 The
human, nonsmall cell lung carcinoma A549 cell line was used in
the cell proliferation assay, cell pharmacokinetic assays, and as a
tumor model, that is, for all studies reported here. All concentrations
are expressed as moles of Wm.
Cellular Pharmacokinetics. NBD-Wm cell protein labeling
assay: Cells were treated with 10 µM of NBD-Wm 3 or an NBD-
WmC20 derivative for the indicated time and whole cell lysates
generated. Cell protein was normalized between samples and
wortmannylated (NBDylated) protein was visualized using an anti-
NBD rabbit polyclonal antibody (Biogenesis, Munich, Germany)
as described29 and quantified using densitometry. Labeled cell
protein was taken as the sum of the areas of three prominent bands
(35, 55, and 75 kDa). Values of labeled cell protein as a function
of time (L(t)) were fit to a two step sequential model where k1 was
the rate of increase of labeled protein, k2 was the rate of decrease
of labeled protein, and L0 was the maximum using GraphPad Prism
Software (Irvine, CA): L(t) ) L0 × k1/(k2 - k1) exp((-k1t)
-exp(-k2t)). Initial rates were the increase in L(t) over the initial
5 min of reaction as shown in Figure 3A normalized to fastest
compound.
tives were prepared by reacting the C20 of NBD-Wm 3 or Wm
1, which behave similarly as PI3 kinase inhibitors and in the
antiproliferative assay,29 with nucleophiles. This yielded low-
molecular-weight WmC20 derivates 2a-e (from 1) and 4a-c
(from 3). The modification of Wm at C11 for NBD attachment
and the modification at C20 occur independently. Low-molec-
ular-weight WmC20 derivatives featuring a carboxylic acid (2a,
4a, 2b, and 4b) were then reacted with the amino group of a
70 kDa amino-dextran, which increases their molecular weight
and increases their hydrophilicity through an association with
many hydroxyls on the polyglucose-like dextran polymer.
The half-times of Wm formation of the 12 WmC20 com-
pounds shown in Figure 1 were then determined. Compounds
featuring a N(Me)pyridine group at C20 (2c and 4c) had half-
times for Wm formation of 2.30 and 1.87 h, respectively. These
compounds are termed “fast Wm forming” WmC20 derivatives
and are shown in italics in Table 1. Five WmC20 derivatives
featured an alkyl-based tertiary enamine at C20 (2a, 2d, 4a,
5a, and 6a) and formed Wm with half-times between 8.0 and
9.9 h. These compounds are termed ”slow Wm forming”
WmC20 compounds and are shown in bold lettering in Table
1. Five WmC20 compounds featuring a secondary enamine at
C20 (2b, 2e, 4b, 5b, and 6b) were termed “non-Wm-forming”
WmC20 derivatives. With these compounds the formation of
Wm could not be detected (less than 2% of the WmC20
compound) after 10 h in PBS, our standard assay, or when the
incubation period was extended to 100 h. On the basis of their
behavior under these conditions, they are termed “non-Wm
formers,” although Wm formation might occur below our
detection limits or in biological systems. They are listed as
having half-times of greater than 100 h in Table 1. The
designation of WmC20 derivatives as “fast,” “slow”, or “non”
Wm reflected their rate of formation, which in turn was
determined by the nature and type of atoms near C20 (leaving
group) and which was independent of the group present at C11
(NBD adduct vs acetyl group) or the attachment of dextran
attached to the carboxyl group of 2a and 2b, or 4a and 4b.
Cellular Pharmacokinetics, NBD-Wm Cell Uptake Assay
by Fluorescence. Cells were incubated with 10 µM of 3 or an NBD-
WmC20 derivative for the indicated time. Cells were detached with
trypsin and fluorescence determined by FACS.29 Cell fluorescence
as a function of time (F(t)) was measured as the relative cellular
fluorescence (RCF), which is the geometric mean of cells exposed
to an NBD compound divided by geometric mean of unexposed
cells.29,31,32 All data for the first hour were fit to a single exponential
curve (F(t) ) F0(1 -exp -kt). Initial rates were the increase in
F(t) over the initial 5 min of reaction as shown in Figure 3B
normalized to fastest compound.
The X-ray crystal structure of Wortmannin bound to PI3K
gamma (1E7U12) was loaded into PyMOL (DeLano, W.L. The
PyMOL Molecular Graphics System (2002), DeLano Scientific,
Palo Alto, CA) and the 6-carbon linker and NBD were attached
using the PyMOL Build function.
We next measured the behavior of fast, slow and non-Wm
forming WmC20 compounds using Wm derivatives that lacked
modifications at C20, that is Wm 1 and NBD-Wm 3 as controls,
in our four bioassays. Data obtained is provided in Table 1.
Activity-Based Antiproliferative and PI3 Kinase Assays. With
the antiproliferative assay, fast Wm releasing compounds had
IC50s similar to Wm, slow Wm releasing compounds had IC50s
3-12 times lower than the Wm IC50 (better than Wm), and
Results
Our strategy for the synthesis of Wm-based viridin derivatives
is shown in Figure 1. Low-molecular-weight WmC20 deriva-

4704 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 15
Blois et al.
non-Wm releasers had IC50s at least 2.6 times higher than that
of Wm (worse than Wm). With the in vitro PI3 kinase assay
on the other hand, all WmC20 compounds tested were notably
less effective inhibitors than Wm. Fast or slow Wm releasing
compounds had IC50s 5-30 times higher than Wm. Non-Wm
formers did not inhibit PI3 kinase even at the highest concentra-
tion employed (10 µM). All WmC20 derivatives tested were
weaker PI3 kinase inhibitors than Wm, reflecting the incomplete
generation of Wm during this 0.5 h assay, with the rate of Wm
formation governed by the nature of the leaving group at C20.
To ascertain whether the variable rates of Wm formation
achieved with our panel of WmC20 derivatives in PBS
determined the pharmacokinetics with which compounds inter-
acted with cells, we exposed intact cells to fluorescent WmC20
compounds and determined their fate.
attached to dextran (6a > 6b), although as with the labeled cell
protein assay, dextran attachment resulted in a decreased rate
of compound uptake (4a > 6a and 4b > 6b).
Discussion
WmC20 Compounds Have Variable Rates of Wm
Formation. WmC20 compounds formed Wm at highly variable
rates that depended on the nature of the leaving group at C20.
With tertiary aromatic enamines such as N-methyaminopyridine
at C20, the delocalization of electrons by the pyridine ring,
together with a tertiary enamine at C20, resulted in half-times
for Wm formation of 1.87 and 2.30 h for 2c and 4c, respectively
(Table 1). With tertiary aliphatic enamines at C20, a slow Wm
formation was obtained: half-times were between 8.0 and 9.9 h.
Slow Wm forming compounds included N(Me) hexanoic acid
(abbreviated N(Me)HA) based WmC20 derivatives (2a, 4a, 5a,
6a) and 2d, which featured a proline leaving group. With
secondary enamines at C20, Wm formation was not observed:
the half-time for Wm formation was greater than 100 h.
Compounds in this class included N(H) hexanoic acid based
compounds (2b, 4b, 5b, 6b) or 2e, which featured a lysine
leaving group. The rates of Wm release were independent of
the presence of the group at C11 (acetyl or NBD-adduct) and
the attachment of the low molecular carboxylic acid WmC20
derivatives (2a, 2b, 4a, 4b) to a dextran carrier (5a, 5b, 6a,
6b). The attachment of low-molecular-weight WmC20 deriva-
tives to dextran increased the size and water solubility of the
Wm (data not shown) without altering Wm release kinetics. For
example, the half-time for Wm formation with 2a was 8.7 h,
while that for 2a with dextran attached (5a) was 9.1 h.
WmC20 Compounds as Inhibitors of PI3 Kinase. All
WmC20 compounds were substantially weaker PI3 kinase
inhibitors than Wm, though they fell into two classes. Com-
pounds which did not form Wm (or NBD-Wm), that is 2b, 4b,
5b, 6b, or 2e, featured a secondary amine at C20 and failed to
inhibit PI3 kinase at the highest concentration employed (10000
nM). Fast Wm forming (2c, 4c) or slow Wm forming com-
pounds (2a, 4a, 5a, 6a, 2d) inhibited the enzyme but with higher
IC50s than Wm due to the incomplete production of Wm in
the 0.5 h time of the assay. These results were consistent with
our earlier work on the inhibition of PI3 kinase by 2a and 2b,
which used a radioactive lipid phosphorylation assay,27 while
the current study employed a fluorescence energy transfer
method.30
WmC20 Compounds as Inhibitors of Cell Proliferation.
The conclusion that the kinetics of Wm formation controlled
the antiproliferative activity of WmC20 compounds was sup-
ported by the correlation between the rates of Wm formation
and the antiproliferative activities of the WmC20 compounds
in culture (Table 1). Fast Wm releasers with half-times for Wm
formation of 1.87-2.30 h had IC50s similar to Wm, which were
short compared to the 48 h time for the antiproliferative assay
and reflects the negligible effect of the Wm formation reaction
under these conditions. Slow Wm releasers (half-time for Wm
formation about 9 h) had lower IC50s than Wm (more potent
than Wm), a behavior we related to their slow generation of
Wm over the 48 h time of the assay. On the other hand, non-
Wm formers had IC50s higher than Wm. Because their half-
times for Wm formation were greater than 100 h, about twice
the 48 h assay period, these compounds would be expected to
largely remain as inactive C20 modified Wm derivatives for
the duration of the assay. Although the rate of Wm formation
with WmC20 compounds appeared to explain the activity of
these compounds in PI3 kinase and antiproliferative assays, we
NBD-Wm Cell Protein Labeling. The time course for the
formation of wortmannylated cell protein provided a measure
of a cell’s exposure to the reactive species (NBD-Wm 3), when
an NBD-WmC20 compound was applied to cells. Cells were
exposed to 10 µM of the non-C20 protected NBD-Wm or 3,
the fast Wm releasing 4c, the slow releasing 4a and 6a, and
the nonreleasing 4b and 6b for various times. After lysis, cell
protein was subjected to SDS gel electrophoresis and the
presence of wortmannylated proteins (NBDylated proteins)
determined by reacting the gel with anti-NBD as shown in
Figure 2. With all compounds, the 55 kDa band was detected
before other bands and was the only band seen with the non-
Wm-releasing 4b. The ability of 4b to label the 55 kDa protein
indicates some NBD-Wm formation was occurring under these
conditions (48 h incubation with cells), although 4b was
designated as a “non” Wm forming compound based on its
behavior in PBS. This small difference may reflect slightly
different activation mechanisms in cells versus PBS or reflect
a higher sensitivity of the enzyme amplified Western blot
method for detecting wortmannylated protein in Figure 2.
To more quantitatively determine the time course of cell
protein labeling, films of Western blots were scanned and areas
of the 35, 55, and 75 kDa bands were summed and plotted
versus time as shown in Figure 3A, with the relative initial rates
of cell protein labeling given in Table 1. The formation of
wortmannylated cell protein reflects a reaction between the C20
of Wm, either added as NBD-Wm 3 or released from a NBD-
WmC20 derivative (4a-c, 6a, 6b), and the epsilon amino groups
of protein lysines. Reactions between Wm and lysine (an N
acetylated lysine), including exchange reactions between WmC20
protected derivatives and lysine, occur under physiological
conditions in vitro.28 The decrease of wortmannylated protein
with long incubation times likely reflects protein degradation
and synthesis. All WmC20 protected compounds labeled protein
more slowly than the non-C20 protected NBD-Wm. The low-
molecular-weight 4c labeled protein similarly to 4a, which in
turn was faster than 4b (4c ∼ 4a > 4b). The difference between
4a and 4b (4a > 4b) was maintained when they were attached
to dextran (6a > 6b), although attachment to dextran resulted
in a deceased rate of protein labeling (4a > 6a and 4b >).
NBD-Wm Cell Uptake by Cell Fluorescence. To obtain
another measure of the cellular pharmacokinetics of NBD-
WmC20 compounds, cells were incubated with these com-
pounds, and the resulting NBD-derived cellular fluorescence as
a function of time determined as shown in Figure 3B. Relative
initial rates of fluorescence increase, which reflect compound
uptake, are provided in Table 1. Here 4c was taken up faster
than 4a, which was taken up faster than 4b (4c > 4a > 4b).
Again, this pattern (4a > 4b) maintained when 4a or 4b were

Self-ActiVation Enhances Potency of Viridin Prodrugs
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 15 4705
sought to confirm that Wm release controlled the pharmacoki-
netic behavior of these compounds in more complex situations
involving cells.
With two assays reflecting the pharmacokinetic interaction
of our compounds with cells (formation of NBD-Wm labeled
cell protein by anti-NBD-Western blot and uptake of NBD-
Wm by cell fluorescence), the widely varying rates of NBD-
Wm formation in vitro correlated with the time course obtained
(Table 1). Thus the fast Wm forming 4c entered cells faster
than the slowly Wm forming 4a, which in turn was faster in
these respects than the non-Wm forming 4b (4c > 4a > 4b).
The non-Wm forming compounds 4b and 6b exhibited ex-
tremely slow pharmacokinetics in cell protein labeling and were
poorly internalized by cells. Results with the labeled cell protein
assay generally paralleled those with the cell fluorescence assay,
the exception being that the rate of cell protein labeling for 4c
was slightly slower than 4a, although on the basis of Wm
formation rates, 4c should be faster.
The view that the energy for the high affinity between PI3
kinase and Wm is largely derived from the reaction between
the electrophilic C20 and a lysine in the ATP site of the catalytic
subunit of PI3 kinase stems from observations that Wm when
modified at C20 inhibits PI3 kinase either very poorly or not at
all,14,23 although the opposite result, that WmC20 derivatives
are better PI3 kinase inhibitors than Wm have also been
described.22,26 In light of the different reports on the inhibition
of PI3 kinase by WmC20 compounds, it is of interest to consider
the crystal structure of Wm in the ATP site of PI3 kinase, and
what it implies for the self-activating mechanism of action
proposed here.
Figure 5. Fate of NBD-WmC20 derivatives in the PI3 kinase and
antiproliferative assay. (A) PI3 kinase assay: With a 0.5 h incubation,
fast and slow activating WmC20 derivatives partially activate by
generating NBD-Wm 3 and have higher IC50s than non-C20-protected
forms of Wm like 3. 3 is stable and active under these conditions and
has a low IC50. (B) Antiproliferative assay: NBD-Wm in culture media
reacts with nucleophiles like amino acids, yielding a mixture of NBD-
WmC20, components of which slowly regenerate NBD-Wm, endowing
the culture media with a continuing antiproliferative activity though
NBD-Wm has disappeared. When a slowly NBD-Wm forming prodrug
(4a or 6a) is used, the viridin NBD-Wm forms over the assay period.
The slow formation of 4a or 6a an increased potency relative to 3,
although the slow formation of NBD-Wm makes these compounds poor
PI3 kinase inhibitors. NBD-Wm rapidly enters cells after release.
A model of NBD-Wm in the ATP site of PI3 kinase gamma
subunit based on earlier crystallographic studies12 is shown in
Figure 4 and provides important support for two key features
of the postulate that NBD-WmC20 derivatives are in effect
“fluorescent Wm prodrugs.” First, when NBD-Wm is formed,
it can bind in the ATP site of the kinase similarly to Wm. As
shown in Figure 4A, NBD-Wm binds deeply in the ATP pocket,
while NBD (cyan) protrudes far out into the solvent, consistent
with similar IC50s for PI3 kinase that result when NBD is
attached to Wm (Table 1). Second, WmC20 derivatives should
not be able to bind in the ATP site of the kinase, a possibility
examined in Figure 4B. An NBD-Wm modified at C20 cannot
bind to PI3 kinase because of the proximity of the surface of
the enzyme shown in blue. Other aspects of the Wm binding to
PI3 kinase not shown in Figure 4 indicate a deeply buried C20
is essential for formation of important stabilizing hydrogen
bonds between (i) the carbonyl oxygen at the C3 of Wm and
hydroxyl group of S806 and (ii) between the hydroxyl group at
C6 and an oxygen of D964.12 As shown in Figure 1, C3 and
C6 are positioned close to the reactive C20.
As summarized in Figure 5, the behavior of NBD-Wm C20
derivatives in the PI3 kinase assays (5A) and antiproliferative
assay (5B) depends on a combination of assay duration and the
rate of release of NBD-Wm. In the 0.5 h kinase assay, the fast
activating (4c) and slow activating (4a, 6a) NBD-WmC20
derivatives partially generate NBD-Wm 3 and have higher IC50s
(are weaker) than the non-C20-protected NBD-Wm. NBD-Wm
does not have to self-activate and therefore has a lower IC50.
Non-Wm forming compounds (4b, 6b) failed to inhibit PI3
kinase. If NBD-Wm is added to culture media (Figure 5B),
reactions occur at the nonprotected C20 position with nucleo-
philes like amino acids as previously described28 and as shown
in Figure 5B. An uncontrolled mixture of NBD-WmC20
compounds is formed, components of which slowly generate
Wm, endowing the culture media with a continuing antiprolif-
erative activity though NBD-Wm has disappeared. When a
slowly NBD-Wm forming compound (4a or 6a) is added, NBD-
Wm slowly forms over the assay period, which is about 5 half-
times (48/9). Cell fluorescence increases slowly (Figure 3B) and
wortmannylated proteins form slowly (Figure 2), indicating the
rate limiting step in culture media with these compounds is the
slow release of NBD-Wm. Thus, WmC20 derivatives that
behave as slow release, self-activating prodrugs exhibit an
improved antiproliferative activity, although they are relatively
poor PI3 kinase inhibitors.
Results with the slow Wm forming, dextran conjugate 6a are
of particular interest for two general reasons. First, with 6a the
formation of NBD-Wm 3 is associated with a drastic change in
physical properties. When present as the 6a prodrug, NBD-Wm
has a molecular weight of 72 kDa and is hydrophilic and water
soluble, properties conferred by its attachment to the 70 kDa
dextran carrier. Released from 6a, NBD-Wm (MW ) 677 Da)
is poorly soluble in water. Therefore, with 6a, predominant
release of NBD-Wm is extracellular and the released NBD-
Wm undergoes a rapid cellular uptake. Supporting this view is
the fact that when NBD-Wm is attached to dextran with a
nonreleasing linkage (6b), it exhibited no antiproliferative
activity and failed to enter cells, both by the cell fluorescence
and modified protein assays. A second reason for focusing on
6a is that when used in vivo, the dextran can be an inert
macromolecular carrier, extending blood half-time, slowing
hepatic uptake, and providing enhanced tumor targeting due to
the enhanced permeability and retention effect.33,34 It should

4706 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 15
Blois et al.
be noted that the A549 tumor cells used here internalize
fluorescent dextrans very poorly (unpublished observations), so
that the slow extracellular release of NBD-Wm, followed by
uptake of NBD-Wm, is the predominant pathway of NBD-Wm
internalization. However, if 6a is exposed to other types of cells
(e.g., monocytes or macrophages after IV injection), 6a might
be internalized intact with an intracellular NBD-Wm release).
The results obtained with 6a, termed the SAV prodrug (self-
activating viridin prodrug), in animal models of cancer and
inflammation will be presented shortly.
Our slow release mechanism shown in Figure 5A,B provides
a basis for two recommendations for selecting Wm or Wm
derivatives to analyze signal transduction pathways. For short-
duration assays (minutes to about 1 h), Wm should be used
because WmC20 derivatives are inactive and will fail to form
Wm during these time periods, as evident from results with our
PI3 kinase assay. In addition, Wm’s lack of charge and
hydrophobicity allows it to enter cells rapidly, as was the case
for the minimally charged NBD-Wm (Figure 3B). For assays
where long incubations are employed, greater than about 6 h, a
slowly Wm forming, C20 protected Wm should be used. For
animal studies or pharmaceutical applications, WmC20 deriva-
tives are also recommended. As indicated by Wm half-time of
less than 10 min in culture media,35 the use of unprotected,
C20 reactive Wm will result in the formation of an uncontrolled
mixture of WmC20 compounds, which can give rise to variable
results, depending not on the design of the compound but on
the nucleophiles present in the media it encounters.
chemically diverse NBD-Wm releasing libraries of compounds
whose physical properties or interaction with molecular targets
could be determined by replacing the amino-dextran of 6a with
polysaccharides, peptides, antibodies, etc. The ability to rapidly
determine the fate of such materials in vivo could then be based
on NBD’s fluorescence or immunoreactivity, providing a
preliminary assessment of a compound’s biodistribution and
pharmacokinetic properties. Because man-made kinase inhibitors
achieve selectivity by binding to the active sites of target kinases
with high affinity, the delivery-based design of self-activating
Wm natural product prodrugs can be a completely distinct and
therefore valuable approach to obtaining kinase inhibitors.
Acknowledgment. Work was supported in part by NIH
grants T32P50-CA86355, RO1-EB004472, and T32-CA079443.
There are no competing financial interests.
References
(1) Wipf, P.; Halter Robert, J. Chemistry and biology of wortmannin. Org.
Biomol. Chem. 2005, 3, 2053–2061.
(2) Powis, G.; Bonjouklian, R.; Berggren, M. M.; Gallegos, A.; Abraham,
R.; Ashendel, C.; Zalkow, L.; Matter, W. F.; Dodge, J.; Grindey, G.;
et al. Wortmannin, a potent and selective inhibitor of phosphatidyli-
nositol-3-kinase. Cancer Res. 1994, 54, 2419–2423.
(3) Creemer, L. C.; Kirst, H. A.; Vlahos, C. J.; Schultz, R. M. Synthesis
and in vitro evaluation of new wortmannin esters: potent inhibitors of
phosphatidylinositol 3-kinase. J. Med. Chem. 1996, 39, 5021–5024.
(4) Liu, Y.; Shreder, K. R.; Gai, W.; Corral, S.; Ferris, D. K.; Rosenblum,
J. S. Wortmannin, a widely used phosphoinositide 3-kinase inhibitor,
also potently inhibits mammalian polo-like kinase. Chem. Biol. 2005,
12, 99–107.
The improved potency as antiproliferative agents of com-
pounds like 6a is related to their mechanism of slow self-
activation and generation of active NBD-Wm 3, and neither
implies nor requires that PI3 kinase be the molecular target of
the released NBD-Wm. Our use of the PI3 kinase assay and
antiproliferative assay to assess the bioactivity was based on
the incubation times employed by these assays, and by their
use in many laboratories, rather than a mechanistic relationship
between them. In fact determining the molecular target respon-
sible for Wm’s inhibition of cell proliferation is complicated
by the micromolar concentrations necessary to achieve this
effect, see Table 1 and refs 15, 16, 22 These concentrations are
considerably above the IC50s of Wm not only for PI3 kinase
(IC50 ) 1-5 nM)2,3,8 but also for polo-like kinase (IC50 ) 24
nM4), mTOR (IC50 ) 200 nM5), DNA-dependent protein kinase
(IC50 ) 200 nM)6 and ATM (IC50 ) 200 nM7). A second
observation for the specificity of Wm for PI3 kinase in vivo
was its selective modification of the p110 subunit of PI3 kinase
from neutrophils.13,36 However, more recent studies with a
variety of fluorescent Wm derivatives and cancer cell lines have
shown Wm can label multiple cell proteins.4,29,37 Our assay for
wortmannylated cell protein (Figure 3A) does not assume that
the labeled bands visualized (bands at 35, 55, and 75 kDa) are
involved in generating the biological effects of Wm. Instead,
this assay assumes that the degree of modification of these
proteins provides a quantitative indication of the exposure of
the cell to the chemically reactive C20 of NBD-Wm, in much
the same way that the level of glucose modified proteins like
hemoglobin AIc can be used as an indication of the exposure
of cells to glucose.38
(5) Brunn, G. J.; Williams, J.; Sabers, C.; Wiederrecht, G.; Lawrence,
J. C., Jr.; Abraham, R. T. Direct inhibition of the signaling functions
of the mammalian target of rapamycin by the phosphoinositide 3-kinase
inhibitors, wortmannin and LY294002. EMBO J. 1996, 15, 5256–
5267.
(6) Hartley, K. O.; Gell, D.; Smith, G. C.; Zhang, H.; Divecha, N.;
Connelly, M. A.; Admon, A.; Lees-Miller, S. P.; Anderson, C. W.;
Jackson, S. P. DNA-dependent protein kinase catalytic subunit: a
relative of phosphatidylinositol 3-kinase and the ataxia telangiectasia
gene product. Cell 1995, 82, 849–856.
(7) Sarkaria, J. N.; Tibbetts, R. S.; Busby, E. C.; Kennedy, A. P.; Hill,
D. E.; Abraham, R. T. Inhibition of phosphoinositide 3-kinase related
kinases by the radiosensitizing agent wortmannin. Cancer Res. 1998,
58, 4375–4382.
(8) Knight, Z. A.; Gonzalez, B.; Feldman, M. E.; Zunder, E. R.;
Goldenberg, D. D.; Williams, O.; Loewith, R.; Stokoe, D.; Balla, A.;
Toth, B.; Balla, T.; Weiss, W. A.; Williams, R. L.; Shokat, K. M. A
pharmacological map of the PI3-K family defines a role for p110alpha
in insulin signaling. Cell 2006, 125, 733–747.
(9) Ward, S. G.; Finan, P. Isoform-specific phosphoinositide 3-kinase
inhibitors as therapeutic agents. Curr. Opin. Pharmacol. 2003, 3, 426–
34.
(10) Vlahos, C. J.; Stancato, L. F. Inhibitors of cellular signaling targets:
designs and limitations. Methods Mol. Biol. 2004, 273, 87–102.
(11) Ruckle, T.; Schwarz, M. K.; Rommel, C. PI3Kgamma inhibition:
towards an “aspirin of the 21st century”. Nat. ReV. Drug DiscoVery
2006, 5, 903–918.
(12) Walker, E. H.; Pacold, M. E.; Perisic, O.; Stephens, L.; Hawkins, P. T.;
Wymann, M. P.; Williams, R. L. Structural determinants of phos-
phoinositide 3-kinase inhibition by wortmannin, LY294002, quercetin,
myricetin, and staurosporine. Mol. Cell 2000, 6, 909–919.
(13) Wymann, M. P.; Bulgarelli-Leva, G.; Zvelebil, M. J.; Pirola, L.;
Vanhaesebroeck, B.; Waterfield, M. D.; Panayotou, G. Wortmannin
inactivates phosphoinositide 3-kinase by covalent modification of Lys-
802, a residue involved in the phosphate transfer reaction. Mol. Cell.
Biol. 1996, 16, 1722–1733.
(14) Norman, B. H.; Shih, C.; Toth, J. E.; Ray, J. E.; Dodge, J. A.; Johnson,
D. W.; Rutherford, P. G.; Schultz, R. M.; Worzalla, J. F.; Vlahos,
C. J. Studies on the mechanism of phosphatidylinositol 3-kinase
inhibition by wortmannin and related analogs. J. Med. Chem. 1996,
39, 1106–1111.
Wm’s unique chemistry, the ability to attach the NBD
fluorochrome to the C11 carbon and then further modify Wm
at C20, permits the design of fluorescent WmC20 derivatives
which self-activate to produce the active NBD-Wm 3 and whose
disposition in biological systems can be readily ascertained. The
chemistry outlined here can be expanded to obtain more
(15) Schultz, R. M.; Merriman, R. L.; Andis, S. L.; Bonjouklian, R.;
Grindey, G. B.; Rutherford, P. G.; Gallegos, A.; Massey, K.; Powis,
G. In vitro and in vivo antitumor activity of the phosphatidylinositol-
3-kinase inhibitor, wortmannin. Anticancer Res. 1995, 15, 1135–1139.

Self-ActiVation Enhances Potency of Viridin Prodrugs
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 15 4707
(16) Lemke, L. E.; Paine-Murrieta, G. D.; Taylor, C. W.; Powis, G.
Wortmannin inhibits the growth of mammary tumors despite the
existence of a novel wortmannin-insensitive phosphatidylinositol-3-
kinase. Cancer Chemother. Pharmacol. 1999, 44, 491–497.
(17) Wiesinger, D.; Gubler, H. U.; Haefliger, W.; Hauser, D. Antiinflam-
matory activity of the new mould metabolite 11-desacetoxy-wortman-
nin and of some of its derivatives. Experientia 1974, 30, 135–136.
(18) Wymann, M. P.; Bjorklof, K.; Calvez, R.; Finan, P.; Thomast, M.;
Trifilieff, A.; Barbier, M.; Altruda, F.; Hirsch, E.; Laffargue, M.
Phosphoinositide 3-kinase gamma: a key modulator in inflammation
and allergy. Biochem. Soc. Trans. 2003, 31, 275–280.
(19) Laffargue, M.; Calvez, R.; Finan, P.; Trifilieff, A.; Barbier, M.; Altruda,
F.; Hirsch, E.; Wymann, M. P. Phosphoinositide 3-kinase gamma is
an essential amplifier of mast cell function. Immunity 2002, 16, 441–
451.
(20) Finan, P. M.; Thomas, M. J. PI 3-kinase inhibition: a therapeutic target
for respiratory disease. Biochem. Soc. Trans. 2004, 32, 378–382.
(21) Zhu, T.; Gu, J.; Yu, K.; Lucas, J.; Cai, P.; Tsao, R.; Gong, Y.; Li, F.;
Chaudhary, I.; Desai, P.; Ruppen, M.; Fawzi, M.; Gibbons, J.; Ayral-
Kaloustian, S.; Skotnicki, J.; Mansour, T.; Zask, A. Pegylated
wortmannin and 17-hydroxywortmannin conjugates as phosphoi-
nositide 3-kinase inhibitors active in human tumor xenograft models.
J. Med. Chem. 2006, 49, 1373–1378.
(22) Ihle, N. T.; Williams, R.; Chow, S.; Chew, W.; Berggren, M. I.; Paine-
Murrieta, G.; Minion, D. J.; Halter, R. J.; Wipf, P.; Abraham, R.;
Kirkpatrick, L.; Powis, G. Molecular pharmacology and antitumor
activity of PX-866, a novel inhibitor of phosphoinositide-3-kinase
signaling. Mol. Cancer Ther. 2004, 3, 763–772.
(23) Norman, B. H.; Paschal, J.; Vlahos, C. J. Synthetic studies on the
furan ring of wortmannin. Bioorg. Med. Chem. Lett. 1995, 5, 1183–
1186.
(24) Baggiolini, M.; Dewald, B.; Schnyder, J.; Ruch, W.; Cooper, P. H.;
Payne, T. G. Inhibition of the phagocytosis-induced respiratory burst
by the fungal metabolite wortmannin and some analogues. Exp. Cell
Res. 1987, 169, 408–418.
(25) Yano, H.; Nakanishi, S.; Kimura, K.; Hanai, N.; Saitoh, Y.; Fukui,
Y.; Nonomura, Y.; Matsuda, Y. Inhibition of histamine secretion by
wortmannin through the blockade of phosphatidylinositol 3-kinase in
RBL-2H3 cells. J. Biol. Chem. 1993, 268, 25846–25856.
heteroalkylation of the steroidal PI-3-kinase inhibitor wortmannin. Org.
Biomol. Chem. 2004, 2, 1911–1920.
(27) Yuan, H.; Luo, J.; Weissleder, R.; Cantley, L.; Josephson, L.
Wortmannin-C20 conjugates generate wortmannin. J. Med. Chem.
2006, 49, 740–747.
(28) Yuan, H.; Barnes, K. R.; Weissleder, R.; Cantley, L.; Josephson, L.
Covalent reactions of wortmannin under physiological conditions.
Chem. Biol. 2007, 14, 321–328.
(29) Barnes, K. R.; Blois, J.; Smith, A.; Yuan, H.; Reynolds, F.; Weissleder,
R.; Cantley, L. C.; Josephson, L. Fate of a Bioactive Fluorescent
Wortmannin Derivative in Cells. Bioconjugate Chem. 2007, 19, 130–
137.
(30) Gray, A.; Olsson, H.; Batty, I. H.; Priganica, L.; Peter Downes, C.
Nonradioactive methods for the assay of phosphoinositide 3-kinases
and phosphoinositide phosphatases and selective detection of signaling
lipids in cell and tissue extracts. Anal. Biochem. 2003, 313, 234–245.
(31) Koch, A. M.; Reynolds, F.; Kircher, M. F.; Merkle, H. P.; Weissleder,
R.; Josephson, L. Uptake and metabolism of a dual fluorochrome Tat-
nanoparticle in HeLa cells. Bioconjugate Chem. 2003, 14, 1115–1121.
(32) Koch, A. M.; Reynolds, F.; Merkle, H. P.; Weissleder, R.; Josephson,
L. Transport of surface-modified nanoparticles through cell monolay-
ers. ChemBioChem 2005, 6, 337–345.
(33) Vicent, M. J.; Duncan, R. Polymer conjugates: nanosized medicines
for treating cancer. Trends Biotechnol 2006, 24, 39–47.
(34) Duncan, R. Polymer conjugates as anticancer nanomedicines. Nat. ReV.
Cancer 2006, 6, 688–701.
(35) Holleran, J. L.; Egorin, M. J.; Zuhowski, E. G.; Parise, R. A.; Musser,
S. M.; Pan, S. S. Use of high-performance liquid chromatography to
characterize the rapid decomposition of wortmannin in tissue culture
media. Anal. Biochem. 2003, 323, 19–25.
(36) Thelen, M.; Wymann, M. P.; Langen, H. Wortmannin binds specifically
to 1-phosphatidylinositol 3-kinase while inhibiting guanine nucleotide-
binding protein-coupled receptor signaling in neutrophil leukocytes.
Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 4960–4964.
(37) Yee, M. C.; Fas, S. C.; Stohlmeyer, M. M.; Wandless, T. J.; Cimprich,
K. A. A cell-permeable, activity-based probe for protein and lipid
kinases. J. Biol. Chem. 2005, 280, 29053–29059.
(38) Koenig, R. J.; Peterson, C. M.; Jones, R. L.; Saudek, C.; Lehrman,
M.; Cerami, A. Correlation of glucose regulation and hemoglobin AIc
in diabetes mellitus. N. Engl. J. Med. 1976, 295, 417–420.
(26) Wipf, P.; Minion, D. J.; Halter, R. J.; Berggren, M. I.; Ho, C. B.;
Chiang, G. G.; Kirkpatrick, L.; Abraham, R.; Powis, G. Synthesis and
biological evaluation of synthetic viridins derived from C(20)-
JM800374F