Triple molecular target approach to selective melanoma cytotoxicity

Article abstract of DOI:10.1039/b920260a

Phenylalanine-linked pyrrolo[1,2-a]benzimidazoles were successfully designed to target melanoma cells in vitro. Our design utilised three molecular targets: a phenylalanine pump, the reducing enzyme DT-diaphorase, and IMP dehydrogenase. We describe the synthesis of these compounds as well as the results of in vitro, in vivo, and QSAR studies.

Full text of DOI:10.1039/b920260a

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Triple molecular target approach to selective melanoma cytotoxicity†
Edward B. Skibo,* Akmal Jamil, Brittany Austin, Douglas Hansen and Armand Ghodousi
Received 29th September 2009, Accepted 21st December 2009
First published as an Advance Article on the web 28th January 2010
DOI: 10.1039/b920260a
Phenylalanine-linked pyrrolo[1,2-a]benzimidazoles were successfully designed to target melanoma cells
in vitro. Our design utilised three molecular targets: a phenylalanine pump, the reducing enzyme
DT-diaphorase, and IMP dehydrogenase. We describe the synthesis of these compounds as well as the
results of in vitro, in vivo, and QSAR studies.
Introduction
A molecular target is a protein or enzyme that is highly expressed
in human tumour cell lines. The levels of over one thousand
biologically relevant molecular targets have been determined in
human tumour cell lines from measurements of mRNA and
enzyme activity levels.1 A therapeutic agent directed toward one
or more of these molecular targets has the potential of exhibiting
cancer-cell selectivity. A classic molecular target is the quinone
two-electron reducing enzyme DT-diaphorase. The unusually high
expression of DT-diaphorase in some histological cancer types2–6
contributes to the tumour’s tendency to reduce quinones. The
Fig. 1 Design features of the structure utilising three molecular targets.
IUPAC numbering of the pyrrolo[1,2-a]benzimidazole ring is shown.
clinically used quinone-based antitumour agent mitomycin C
is a well known example of an alkylating agent activated by
DT-diaphorase (i.e. a bioreductive-alkylating agent).7–11 Selective
bioreductive alkylation in high DT-diaphorase cancer types
Inosine monophosphate dehydrogenase (IMPDH) is respon-
(melanoma, renal and non-small-cell lung cancers)1 would exhibit
sible for nicotinamide adenine dinucleotide (NAD)-dependent
maximal antitumour activity with minimal side effects.
Our hypothesis is that a structure utilising three molecular
oxidation of IMP to xanthine monophosphate (XMP), which is
the rate-limiting step in the guanosine monophosphate (GMP)
biosynthetic pathway.23,24 Human inosine monophosphate dehy-
drogenase exists in two isoforms type I and II, but type II is
considered a more important target in chemotherapeutic design
since this isoform is up-regulated in neoplastic cells and is the
predominant form.25 IMPDH inhibitors possess a broad spectrum
of biological activity including antitumour, immunosuppression,
antiviral, and antibacterial activity.23,25,26 This broad spectrum of
activity reflects the importance of GMP in diverse biological
pathways. Therefore, the proposed IMPDH inhibitor shown in
Fig. 1 was expected to exert cytostatic and cytotoxic activity on
melanoma cells.
Known IMPDH inhibitors usually mimic either the enzyme’s
IMP or NAD substrates.27 Mycophenolic acid, which binds to
the NAD cofactor site, represents the earliest inhibitors of IMP
dehydrogenase. Bredinin and ribavirin28,29 are two examples of
nucleoside-based inhibitors that bind in the IMP site of the enzyme
upon their metabolic conversion to the 5¢-monophosphate form.
Selenazofurin and benzamide riboside30 are potent inhibitors of
inosine monophosphate dehydrogenase that binds in the NAD site
of the enzyme upon metabolic conversion to the corresponding
NAD analogue. With crystal structures of IMPDH in hand,
it has been possible to design non-nucleotide inhibitors.31–34,35,36
Similarly, we used reported crystal structures of human IMPDH
type II (Protein Data Bank # 1B30)37 to aid in our IMPDH
inhibitor design.
targets, rather than just one, would possess a high degree of histo-
logic specificity. We chose to design a melanoma-specific structure
utilising the L-phenylalanine pump found in melanocytes12 as
well as the elevated levels of both DT-diaphorase and inosine
monophosphate dehydrogenase (IMPDH) found in melanoma
cells.1 Upon uptake by the L-phenylalanine, reductive activation
by DT-diaphorase would then be followed by the inactivation
of IMPDH. Fig. 1 shows the structure designed to utilise
three molecular targets. This structure possesses a phenylalanine
dipeptide portion that is designed to play a role in uptake of the
compound into the melanoma cell as well as to promote IMPDH
binding.
The pyrrolo[1,2-a]benzimidazole (PBI) ring system shown in
Fig. 1 possesses the benzimidazole ring system designed to mimic
the hypoxanthine ring of the inosine monophosphate (IMP)
substrate of IMPDH. The PBI ring system is functionalised as an
aziridinylquinone to provide bioreductive alkylation capability.
The synthesis and biological properties of PBI derivatives have
been extensively documented by this laboratory.13–22
Department of Chemistry and Biochemistry, Arizona State University,
Tempe, Arizona 85287-1604. E-mail: eskibo@asu.edu; Fax: +1 480 965
2747; Tel: +1 480 965 3581
† Electronic supplementary information (ESI) available: Spectroscopic
data for all compounds and in vitro assay results. See DOI:
10.1039/b920260a
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 1577–1587 | 1577
©

View Online
the quinone or the hydroquinone forms of the PBI resulting
in irreversible inactivation of IMPDH, Scheme 1. However, the
aziridinyl hydroquinone form is expected to be a more reactive
alkylating agent based on literature precedents.14,42
Results and discussion
IMPDH inhibitor design
In 1980, one of us (EBS) designed phenyl-substituted IMP
analogues that exhibited enhanced binding to the IMP binding
site by taking advantage of a hydrophobic pocket located at the 8-
position of the purine ring.38 The actual presence of the pocket
was verified by crystal structures obtained many years later.39
In 2004, COMPARE analysis of 60-cancer cell line mean graph
data for a substituted PBI analogue revealed that IMPDH was
its molecular target.22 The COMPARE analysis correlates mean
graph data with known molecular target levels in cell lines and
thereby generate hypotheses concerning the agent’s mechanism of
action.40,41 Considering the aforementioned results, we postulated
that substituted PBIs could exploit the hydrophobic pocket located
at the 8-position of the purine ring
The model used to design these inhibitors was developed by
superimposing the benzimidazole ring of the L-phenylalanine-
linked pyrrolobenzimidazole system unto the purine ring of active
site bound 6-ChloroIMP. After the superimposition, the active
site IMP structure is subtracted from the model and the active
site bound pyrrolobenzimidazole and enzyme complex minimised
to afford the structure shown in Fig. 2. This structure shows that
the amino acid residue of the PBI occupies a cleft different than
that occupied by the ribofuranosyl and phosphate residues of the
IMP substrate. The pyrrolo ring of the pyrrolobenzimidazole ring
system, as well as the S-stereochemistry of the 3-position, directs
the amino acid residues to this cleft. Modeling studies revealed
that the L-phenylalanine (the S-enantiomer) binds more tightly
than the D-form (the R-enantiomer). Modelling the hydroquinone
derivative of the PBI into the IMPDH active site afforded a
structure similar to that shown in Fig. 2. These modelling studies
showed that a PBI-based inhibitor, either as the quinone or
hydroquinone, could bind to the IMPDH active site.
Scheme 1 Aziridinyl alkylation mechanisms.
PBI–phenylalanine analogue synthesis
The PBI-linked phenylalanine analogues prepared for this study
are shown in Fig. 3 and 4. The diphenylalanine analogues 1 shown
in Fig. 3 possess all four D- and L-stereochemical possibilities.
The goal was to investigate the influence of stereochemistry on
melanoma histological specificity and in vivo activity. According
to our hypothesis, the D-enantiomers should possess diminished
melanoma specificity. We also prepared the analogues 2 shown in
Fig. 4 to investigate the role of aromatic substituents on histolog-
ical specificity. The goal was to discover a PBI analogue linked
to a substituted L-phenylalanine that still possesses melanoma
specificity. A quantitative structure–activity relationship (QSAR)
derived from in vitro data for analogues 2 would guide future
analogue development.
We prepared LD1, DL1, and DD1 using the synthetic method-
ology previously described for LL1.22 Scheme 2 outlines the
preparation PBI-linked phenylalanines 2 starting with 3-bromo-4-
nitrotoluene 3. The conversion of 3 to the (S)-amino pyrrolo[1,2-
a]benzimidazole derivative 4 is a multistep process previously
reported from this laboratory.14,17,20,43,44 The substituted phenylala-
nine was coupled to the amine group of 4 employing dicyclohexyl
carbodiimide (DCC). The coupling product 6 was subject to
catalytic reduction to both remove the bromo group and reduce
the nitro group to the amine. The amine was then subjected to
Fremy oxidation to afford the quinone analogues 7. Removal of
the t-Boc followed by aziridination afforded the final products 2.
Fig. 2 A PBI analogue modelled into the purine active site of human type
II IMPDH.
Melanoma histological specificity of the diphenylalanine PBIs 1
The compound series 1 shown in Fig. 3 was screened against the
National Cancer Institute’s 60 cell line human cancer panel.40,45–47
This panel consists of the nine major histological types of human
cancers. We used the average -log LC50 values for each cancer
Inspection of Fig. 2 also reveals that the aziridinyl group
is proximal to an active site cysteine residue (sulfur shown in
yellow). Reaction between these centres could occur with either
1578 | Org. Biomol. Chem., 2010, 8, 1577–1587
This journal is
The Royal Society of Chemistry 2010
©

View Online
Fig. 3 Stereoisomers of diphenylalanine-linked PBIs 1 prepared for this
Scheme 2 Synthesis of the PBI-linked phenylalanines 2.
study.
Fig. 4 Substituted L-phenylalanine-linked PBIs prepared for this study.
type to determine the sensitivity toward the stereoisomers of 1.
The LC50 value, the cytotoxicity parameter, is the concentration
of agent at which there is 50% lethality toward the cancer cell line.
The lower the LC50 value, the more potent the agent. Similarly,
the larger the -log LC50 value, the more potent the agent. The
bar graph in Fig. 5 shows the relative activity (-log LC50) of
the stereoisomers of 1 towards the NCI nine-cancer panel. The
raw data used to obtain Fig. 5 is found in the Supplementary
Information Section.
The stereoisomer LL-1 possesses the expected high cytotoxic
specificity toward the melanoma panel. Actually, many of the LC50
values for melanoma cell lines are < 10-8 M, so the melanoma bar
for LL-1 in Fig. 5 is even higher. In contrast, the stereoisomer LD-
1 possesses diminished melanoma specificity. This observation is
consistent with our hypothesis that the terminal L-phenylalanine
Fig. 5 Bar graph of the average -logLC50 values for series 1 plotted against
nine cancer panels. Higher bars represent higher activity.
of 1 utilises an L-phenylalanine pump found in melanocytes12 to
gain entry into the melanoma cell.
The stereoisomers DL-1 and DD-1 possess a low level of
cytotoxicity against all of the cancer panels. There is no apparent
melanoma specific cytotoxicity for either of these stereoisomers.
This result is consistent with our modelling studies that indicated
the first phenylalanine must be L to be accommodated by the
IMPDH active site.
The conclusion is the both phenylalanines of 1 must have the L
stereochemistry for there to be melanoma specificity.
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 1577–1587 | 1579
©

View Online
Table 1 Hollow fibre scores of LD-1 and those of clinically used
antitumour agents
In vivo activity of LL-1 and LD-1
The in vivo activity of the most cytotoxic analogues LL-1 and LD-
1 were evaluated in a xenograft human tumour model and in the
National Cancer Institute’s hollow fibre48,49 assay respectively.
We assayed LL-1 in the xenograft tumour model consisting
of human lung tumour cells (NCI-H460) implanted into SCID
mice. This readily available lung tumour model would primarily
provide toxicity information as well as evidence of antitumour
activity (LL-1 exhibits cytostatic activity at nanomolar levels
against a broad range of human cancers). Compound LL-1 was
administered intraperitoneally (IP) to SCID mice at doses of 1,
3, and 5 mg kg-1 at 4 day intervals for a total of six doses.
We used a vehicle control and a mitomycin C treated control
as references. Tumour growth inhibition was calculated at T/C
(treated/control) ¥ 100. T/C Values ≤ 42% are considered active,
indicated with an * in Fig. 6 and 7. Inspection of Fig. 6 reveals that
the 1 mg kg-1 dose did not reduce tumour mass compared to the
control, but resulted in 3/6 animal deaths. In contrast, mitomycin
C was active at all doses without animal deaths.
Intraperitoneal (IP)
score
Subcutaneous (SC)
score
Compound
Total score
LD-1
Taxol
Cisplatin
42
24
14
2
8
8
44
32
22
assay of LD-1 was carried out as previously described.48,49 The
hollow fibre score is broken down into a intraperitoneal (IP) and
a subcutaneous (SC) score that is based on the percent of cancer
left in the fiber (<50% = 2 and >50% = 0) The highest total score
(IP + SC) is 96. The typical antitumor agent only has a hollow
fiber score of ~5 while the highest score being 64 for the natural
product cyclin-dependent kinase inhibitor flavopiridol.50 A good
SC score (≥ 8) indicates that the drug is able to get to the tumor site
(subcutaneous) from a distant site (intraperitoneal) of injection.
Compound LD-1 had an IP score of 42/48 and an SC score of
2/48. These results indicate that compound LD-1 has excellent
antitumor activity in this assay. The hollow fibre scores shown in
Table 1 reveals that LD-1 even has higher activity than clinically
used antitumor agents.
The conclusion of the in vivo studies is that L-phenylalanine
linked analogs are toxic and lack antitumour activity unless
followed by D-phenylalanine as in LD-1. Enzymatic removal of
one or both phenylalanines would afford the amino-substituted
pyrrolobenzimidazole 8 that is known to be highly toxic due to
its cleavage of DNA,20 Scheme 3. Likewise, the L-phenylalanine-
linked PBI 2f22 is highly toxic in vivo showing much the same results
as LL-1.
Fig. 6 The tumour growth curves in SCID mice for LL1 (NSC # 723006),
the vehicle control and the mitomycin C control.
Fig. 7 The tumour growth curves in SCID mice are shown for LL1 (NSC
# 723006) at 2 and 5 mg kg-1. The vehicle control and the mitomycin C
control are also shown.
Scheme 3 Reactions that could be responsible for the high toxicity of
LL-1 and 2f as well as the low toxicity of LD-1.
Fig. 7 shows that the 2 and 5 mg kg-1 doses for LL-1 reduced
tumour mass somewhat. The 5 mg kg-1 dose showed an active
level of tumor reduction (37%) by the 5th dose. Animal survival
continued to be a problem at these dosages with a 3/6 and 6/6
death rate for 2 and 5 mg kg-1 respectively. In conclusion LL-1
possesses very little antitumor activity with toxicity even at low
doses.
In contrast, the unnatural D-phenylalanine of LD-1 is not likely
subject to in vivo hydrolysis and the toxic metabolite 8 is never
formed. However, the L-enantiomer of phenylalanine is preferred
as far as melanoma specificity is concerned. The substituted
analogues shown in Fig. 4 may circumvent hydrolytic removal
of the L-enantiomer. Perhaps a substituent would render the L-
phenylalanine “unnatural” and not susceptible to hydrolysis. Or
perhaps the substituent will enhance melanoma specificity and
cytotoxicity.
Compound LD-1 was far less toxic than LL-1 with a measured
maximum tolerated dose (MTD) of 6.5 mg kg-1. The hollow fiber
1580 | Org. Biomol. Chem., 2010, 8, 1577–1587
This journal is
The Royal Society of Chemistry 2010
©

View Online
Table 2 Results of matrix COMPARE for compound series 2. Pearson
correlation coefficients are shown for each compound pair
analogues will likewise consist of PBI linked to p-methoxy or p-
fluoro substituted diphenylalanine.
Inspection of Fig. 8 reveals that the substituent influences
the level of -log LC50 in the melanoma panel. We carried out
a QSAR analysis52 (quantitative structure activity relationship)
to find the substituent type required for optimal activity against
the melanoma panel. Substituent parameters used in the analysis
include p (lipophilicity), s (electron withdrawing/releasing), F
(field, a measure of inductive electronic effects), and MR (molar
refractivity, a measure of size and polarizability). Multivariable
regression analysis provided the following relationship (R squared
is 98%):
2f
2a
2b
2c
2d
2a
2b
2c
2d
2e
0.71
0.766
0.657
0.612
0.607
0.876
0.77
0.858
0.733
0.67
0.832
0.716
0.78
0.735
0.913
Matrix COMPARE of the monophenylalanines PBIs 2, melanoma
histological specificity, and QSAR
The goal of these studies was to determine the L-phenylalanine
substituent type that would optimise melanoma specificity and
cytotoxicity. Compound series 2, shown in Fig. 4, were screened
against the National Cancer Institute’s 60 cell line human cancer
panel.40,45–47 The raw data used to generate the results described
below are found in the Supplementary Information Section.
We carried out a matrix COMPARE40,46,51 by comparing pair-
wise the patterns of cytotoxicity against a 60 cell line human cancer
panel for 2a–f, Table 2. The substituted phenylalanine analogues
2a and 2b show the highest correlations with the analogue linked to
unsubstituted L-phenylalanine 2f. The data in Table 2 also indicate
that the substituted phenylalanine analogues 2a–e correlate well
with each other (coefficients 0.7 to 0.9). These high correlations
are expected given their structural similarity. The anomaly is the
pair 2d and 2e that correlated well with each other (0.913), but
not with the parent compound 2f (~0.6). In fact, both compounds
possess little melanoma specificity.
-logLC50 = 0.48p + 0.61F - 1.52s + 6.77
(1)
The constant 6.77 is the -logLC50 of the unsubstituted analogue,
the actual value is 6.81. This QSAR relationship shows that the
preferred substituent is electron releasing (s is negative) with an
electron withdrawing inductive component (F is positive) and
lipophilic character (p is positive). The para-methoxy substituent
possesses most of these preferred characteristics: s = -0.27, F =
0.26, and p = -0.02.53 Other suitable substituents include the
halogens and the N,N-dimethylamino group. At this time, we
are screening these phenyl-substituted derivatives of LL-1. The
limited number and type of substituents used to obtain this QSAR
relationship prohibits its extension to very large and very lipophilic
substituents.
COMPARE analysis
The bar graph in Fig. 8 shows the relative activity (-log
LC50) of 2a,b and 2f22 towards the NCI nine-cancer panel.
These results suggest that the p-methoxy and p-fluoro substituted
L-phenylalanines of 2a,b would be suitable mimics of natural
L-phenylalanine. Inspection of the Fig. 8 reveals that 2a,b and
2f possess specificity toward the melanoma panel, although not
as high as the diphenylalanine analogue LL-1. Fig. 5 shows
that melanoma histological specificity requires the presence of
two linked L-phenylalanines. The proposed second generation
COMPARE analysis correlates mean graph data with known
molecular target levels in cell lines and thereby generates hypothe-
ses concerning the agent’s mechanism of action. The levels of over
one thousand biologically relevant molecular targets have been
determined in these tumor cell lines from measurements of mRNA
and enzyme activity levels.1
The observed specificity of LL-1 and 2f for the melanoma panel
is likely related to specific cellular uptake. The events leading to cell
death, reductive activation and IMPDH inhibition, occur down
stream to the uptake process. The COMPARE analysis described
below provides insights into what molecular targets might be
involved in uptake.
COMPARE analysis of 2f showed a good correlation (0.59) with
ABC-B5, a gene for an ATP-binding membrane protein involved
with substrate transport. This protein is expressed in melanoma
cells as well as melanocytes54 and is implicated in the resistance
of malignant melanomas to doxorubicin.55 Perhaps the ABC-B5
transporter is involved in the actual transport of compound 2f
into melanoma cells. However, COMPARE analyses of the D-
stereoisomer of 2f and the substituted analogues 2a–e did not
correlate with ABC-B5.
COMPARE analysis of LL-1 showed a very good correlation
with ECM1 (0.66), a gene for the Extracellular Matrix Protein 1.
The level of this protein is associated with metastatic potential,56
and shows elevated values only in the melanoma cell lines of the
60-cell line panel. We do not know if LL-1 directly interacts with
the ECM1 protein or if the correlation happens to be coincidental.
COMPARE analysis of the D-stereoisomer (DL-1) showed loss of
the ECM1 correlation as well as loss of melanoma specificity.
Fig. 8 Bar graph of the average -logLC50 values for 2a,b and 2f plotted
against nine cancer panels. Higher bars represent higher activity.
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 1577–1587 | 1581
©

View Online
The conclusion of this section is that a L-phenylalanine is very
likely required for melanoma specificity. A second generation
derivative of LL-1, with a substituted L-phenylalanine and a
terminal L-phenylalanine, could display melanoma specificity
without toxicity.
Conclusions
We report the successful design of a melanoma-specific cytotoxic
agent LL-1. All four possible stereoisomers of 1 were prepared
and screened; only the natural L, L-diphenylalanine stereoisomer
showed melanoma specificity. Unfortunately, LL-1 is highly toxic
in vivo with mice deaths even at 1 mg kg-1 doses. The toxicity
problem was solved by utilising the unnatural D-phenylalanine in
the terminal position (LD-1). However, the melanoma specificity
was lost, apparently because of the terminal D-phenylalanine. We
postulate that the L-phenylalanines are removed enzymatically
to afford a toxic PBI. The unnatural D-phenylalanine of LD-1
prevents phenylalanine removal resulting in a maximum tolerated
dose over 6 mg kg-1.
The next step was to investigate substituted L-phenylalanines
as alternatives to L-phenylalanine. The goal was to determine the
L-phenylalanine substituent type that would optimise melanoma
specificity and cytotoxicity. Matrix COMPARE and QSAR anal-
yses revealed that the para-methoxy and the para-fluoro sub-
stituents (2a and 2b respectively) resemble natural L-phenylalanine
and possess enhanced potency against the melanoma cell
lines. Currently we are evaluating second-generation substituted
diphenylalanine PBI analogues in vivo.
IMPDH inhibition
These studies validate the modeling studies (Fig. 2) that predicted
2f could interact with the IMPDH active site. In addition, these
studies suggest that IMPDH could be a molecular target for
compound series 2.
The activity of human recombinant type II IMP dehydrogenase
was determined by monitoring the absorbance of NADH that
was formed during oxidation of inosine monophosphate (IMP)
to xanthosine monophosphate (XMP). To carry out these as-
says, the concentration of IMP was varied while keeping the
concentration of NAD constant. Following the formation of
NADH at 340 nm allowed the progress of the dehydrogenase
reaction to be monitored. The Lineweaver–Burk plot shown in
Fig. 9, obtained in the absence of inhibitor, afforded KM and kcat
values consistent with literature values of type II human IMP
dehydrogenase.57
COMPARE analysis of the unsubstituted phenylalanine ana-
logues 2f and LL-1 revealed correlations with the respective
molecular targets ABC-B5, a gene for a membrane transport
protein, and ECM1, a gene for the Extracellular Matrix Protein
1. Correlations with these molecular targets are lost upon a
change to the D-stereochemistry or the addition of a phenylalanine
substituent.
The role of DT-diaphorase and IMP dehydrogenase were
not obvious from these COMPARE analyses. Circumstantial
evidence of the role of these molecular targets include the
inhibition IMP dehydrogenase by 2f reported herein and the well
known reduction of simple PBIs by DT-diaphorase.19,58 The
involvement of either molecular target will require a monopheny-
lalanine analogue (2) rather than the more sizable diphenylalanine
analogue (1).
Fig. 9 Lineweaver–Burk plots for IMPDH-mediated oxidation of IMP
(mM) in the presence of 2f.
Therefore, another second-generation PBI analogue will possess
a substituted phenylalanine and a terminal L-phenylalanine. We
anticipate that the terminal phenylalanine will confer melanoma
specificity and that its enzymatic removal will permit interaction
with DT-diaphorase and IMP dehydrogenase.
Also shown in Fig. 9 are Lineweaver–Burk plots obtained with
type II IMP dehydrogenase in the presence of two concentrations
of pyrrolobenzimidazole 2f. These plots suggest that competitive
inhibition of IMP dehydrogenase is occurring. However, these
data are highly scattered perhaps due to irreversible inhibition
(see reactions in Scheme 1). Recall that Lineweaver–Burk analysis
assumes reversible inhibition. The Lineweaver–Burk Ki value is
~2 mM, which is the apparent dissociation of 2f from the IMP
dehydrogenase active site.
Experimental
General
The conclusion of the IMPDH assays is that 2f can interact
with the IMPDH active site. Modelling studies suggest the 2a,e
could be accommodated by the IMPDH active site as well.
Since inhibition is downstream to cellular uptake, a correlation
with the IMPDH molecular target is not observed. Interactions
of LL-1 with the IMPDH active site cannot occur due to its
larger size. However, enzymatic removal of a phenylalanine to
afford 2f will lead to IMPDH inhibition. Our next generation
of analogues will possess a substituted phenylalanine and a
terminal L-phenylalanine to facilitate uptake as well as enzymatic
removal.
All analytically pure compounds were dried under high
vacuum in a drying pistol over refluxing methanol. High
resolution mass spectra and MALDI were run at Arizona
State University. Melting points and decomposition points were
determined with a Mel-Temp apparatus. All TLCs were per-
formed on silica gel plates using a variety of solvent sys-
tems and a fluorescent indicator for visualization. IR spectra
were taken as KBr pellets and only the strongest absorbances
1
were reported. H NMR spectra were obtained with a 300 or
500 MHz spectrometer. All chemical shifts are reported relative to
TMS.
1582 | Org. Biomol. Chem., 2010, 8, 1577–1587
This journal is
The Royal Society of Chemistry 2010
©

View Online
Synthesis
(3S)-6-Aziridinyl-2,3-dihydro-7-methyl-1H -3-(N -L-phenyl-
under reduced pressure to yield a white residue. The white residue
contained a mixture of 6 and dicyclohexylurea side product. The
dicyclohexylurea impurity was readily removed upon conversion
to 7 in the next step. Recrystallisation of 6 from dichloromethane–
hexane afforded pure material suitable for characterisation.
alanyl-D-phenylalanylamino)-pyrrolo[1,2-a]benzimidazole-5,8-
dione (LD-1). Yield: 17 mg (5.13%); mp: 132–135 ◦C (dec); TLC:
1
[9 : 1 dichloromethane–methanol] Rf 0.26; H NMR (CDCl3) d
(3S)-6-Bromo-2,3-dihydro-7-methyl-5-nitro-3-[Boc-L-4-metho-
xyphenylalanyl]-1H-pyrrolo[1,2-a]benzimidazole (6a). 409 mg
(75% yield); mp: 150–153 ◦C; TLC: [9 : 1 dichloromethane–
methanol] Rf 0.60; 1H NMR: (DMSO-d6) 8.17 (1H, d, J ~ 8 Hz,
urethane NH), 7.02 (2H, d, J = 8.5 Hz aromatic protons), ~7.0
(1H, s, 8-H), 6.87 (1H, d, J ~ 8 Hz, 9-amide proton), 6.8 (2H, d,
J = 8.5 Hz aromatic protons & 1H, s, 8-H), 4.4 (1H, m, 3-methine
proton), 3.9 & 4.2 (2H, 2 m, 1-methylene protons), 3.66 (3H, s,
methoxy), 3.47 (1H, m, phenylalanine methine), 2.9 (1H, m, 2-
methylene proton), 2.73 (1H, m, 2-methylene proton), 2.44 (3H, s,
6-Me), 2.5 & 2.9 (2H, 2 m, phenylalanine methylene protons),
1.28 (9H, s, t-Boc protons); IR (KBr pellet): 3332, 2932, 2855,
1676, 1631, 1537, 1447, 1368, 1307, 1225, 1165, 1052, 827 cm-1;
HRMS: Calculated 588.1458 for C26H3179BrN5O6, found 588.1391,
calculated 590.1437 for C26H3181BrN5O6, found 590.1440.
7.92 (2H, br, amide NH), 7.30–7.0 (10H, 2 m, aromatic protons),
5.0 (1H, m, 3-methine proton), 4.65 (1H, br s, phenylalanine
methine proton), 4.28 (1H, m, 1-methylene proton), 4.09 (1H, m, 1-
methylene proton), 3.62 (1H, br s, phenylalanine methine proton),
2.91–3.18 (4H, m, phenylalanine methylene protons), 3.09 (1H, m,
C(2) methylene proton), 2.61 (1H, m, C(2) methylene proton), 2.28
(4H, s, aziridine protons), 1.98 (3H, s, methyl protons); IR (thin
film) 3357, 2929, 1623, 1462, 1404, 1381, 1234, 1125, 1082 cm-1;
MALDI: calculated for C31H32N6O4 + H+ (M + H)+ 553.256, found
553.251.
(3S)-6-Aziridinyl-2,3-dihydro-7-methyl-1H -3-(N -D-phenyl-
alanyl-L-phenylalanylamino)-pyrrolo[1,2-a]benzimidazole-5,8-
dione (DL-1). Yield: 16 mg (4.82%); mp: 110–112 ◦C (dec); TLC:
1
[9 : 1 dichloromethane–methanol] Rf 0.51; H NMR (CDCl3) d
8.2 (2H, br s, amide NH), 7.11–7.21 (10H, m, aromatic protons),
5.2 (1H, m, 3-methine proton), d 4.83 (1H, br d, phenylalanine
methine proton), 4.53 (1H, m, 1-methylene proton), 4.25 (1H, m, 1-
methylene proton), 4.02 (1H, br s, phenylalanine methine proton),
2.82–3.16 (4H, m, phenylalanine methylene protons), 2.58 (1H,
m, 2-methylene proton), 2.42 (1H, m, 2-methylene proton), 2.26
(4H, s, aziridine protons), 1.99 (3H, s, methyl protons); IR (thin
film) 3260, 2931, 1644, 1518,1250 cm-1; MALDI: calculated for
C31H32N6O4 + H+ (M + H)+ 553.256, found 553.251.
(3S)-6-Bromo-2,3-dihydro-7-methyl-5-nitro-3-[Boc-L-4-fluoro-
phenylalanyl]-1H-pyrrolo[1,2-a]benzimidazole (6b). 348 mg (65%
yield); mp: 145–155 ◦C; TLC: [9 : 1 dichloromethane–methanol] Rf
1
0.60; H NMR: (DMSO-d6): 8.6 (1H, d, J ~ 8 Hz, amide NH),
7.8 (1H, s, 8-H), 6.9 (1H, d, J ~ 8 Hz, urethane NH), 7.0 & 7.38
(4H, 2d, J = 8.5 Hz aromatic protons), 5.3 (1H, m, 3-methine
proton), 4.0 & 4.3 (2H, 2 m, 1-methylene protons), 3.95 (1H, m,
phenylalanine methine proton), 2.7 & 3.0 (2H, 2 m, 2-methylene
protons) 3.0 & 2.4 (2H, 2 m, phenylalanine methylene protons)
2.5 (3H, s, methyl), 1.28 (9H, s, t-Boc protons); IR: (KBr pellet)
3332, 2932, 2855, 1676, 1631, 1537, 1447, 1368, 1307, 1225, 1165,
1052, 827 cm-1; HRMS Calculated 576.1258 for C25H2879BrFN5O5,
found 576.1266; Calculated 578.1237 for C25H2881BrFN5O5, found
578.1289.
(3S)-6-Aziridinyl-2,3-dihydro-7-methyl-1H -3-(N -D-phenyl-
alanyl-D-phenylalanylamino)-pyrrolo[1,2-a]benzimidazole-5,8-
dione (DD-1). Yield 15 mg (4.52%); mp: 148–150 ◦C (dec); TLC:
1
[9 : 1 dichloromethane–methanol] Rf 0.40; H NMR (CDCl3): d
8.64 (1H, d, J = 6.5 Hz, amide NH), 7.97 (1H, d, J = 6.5 Hz,
amide NH), 7.14–7.32 (10H, m, aromatic protons), 4.87 (1H, q,
J = 7.5 Hz, 3-methine proton), 4.11 (1H, m, 1-methylene proton),
3.97 (1H, m, 1-methylene proton), 3.69 (1H, doublet of doublets,
J = 3 Hz, J = 6 Hz, phenylalanine methine proton), 3.03–3.14 (4H,
m, phenylalanine methylene protons), 2.89 (1H, m, 2-methylene
proton), 2.40 (1H, doublet of doublets, J = 11 Hz, J = 12.5 Hz,
phenylalanine methine proton), 2.33 (2H, d, J = 6.5 Hz, aziridine
protons), 2.30 (2H, d, J = 6 Hz, aziridine protons), 2.16 (1H,
m, 2-methylene protons), 1.93 (3H, s, methyl protons); IR: (thin
film) 3302, 2923, 1659, 1519, 1230 cm-1; MALDI: calculated for
C31H32N6O4 + H+ (M + H)+ 553.256, found 553.258.
(3S)-6-Bromo-2,3-dihydro-7-methyl-5-nitro-3-[Boc-L-4-trifluo-
romethylphenylalanyl]-1H-pyrrolo[1,2-a]benzimidazole
(6c).
431 mg (74% yield); mp: 185–195 ◦C TLC: [9 : 1 dichloromethane–
1
methanol] Rf 0.60; H NMR (DMSO) 8.7 (1H, d, J = 8 Hz,
amide NH), 8.2 (1H, s, C8-H), 7.45 & 7.6 (4H, 2d, J = 8.5 Hz
aromatic protons), 6.95 (1H, d, J = 8 Hz, urethane NH), 5.35
(1H, m, 3-methine proton), 4.0 (1H, m, phenylalanine methine
proton), 4.0 & 4.3 (2H, 2 m, 1-methylene protons), 2.8–3.1 (2H,
2 m, 2-methylene protons) 2.4 & 2.95 (2H, 2 m, phenylalanine
methylene protons), 2.5 (3H, s, methyl), 1.35 (9H, m, t-Boc
protons); IR: (KBr pellet) 3420, 2933, 1676, 1527, 1367, 1327,
1166, 1124, 1067, 1020, 826; HRMS: Calculated 626.1226 for
C26H2879BrF3N5O5, found 626.1260; Calculated 628.1206 for
C26H2881BrF3N5O5, found 628.1200.
General synthesis of intermediate 6
A
solution of 1 mmol of t-Boc protected L-substituted-
phenylalanine 5 in 15 mL of dry DCM was cooled to 0 ◦C.
To this solution, 1.1 mL (1.1 mmol) of a 1 M solution of
1,3-dicyclohexylcarbodiimide (DCC) was added. The resulting
mixture was stirred at 0 ◦C until the solution became cloudy.
To this mixture was added 290 mg (0.93 mmol) of 4 and
the mixture was removed from the ice bath and allowed to
warm to room temperature and stirred for 2 h. The mixture
was filtered, and the filtrate’s purity was tested via TLC (9 : 1
dichloromethane–methanol). The dichloromethane was removed
(3S)-6-Bromo-2,3-dihydro-7-methyl-5-nitro-3-[Boc-L-4-methyl-
phenylalanyl]-1H-pyrrolo[1,2-a]benzimidazole (6d). 329 mg (62%
yield); mp: 148–155 ◦C: TLC [9 : 1 dichloromethane–methanol] Rf
0.62; 1H NMR (DMSO-d6): 8.6 (1H, d, J = 8 Hz, amide NH), 7.8
(1H, s, C8-H), 7.05 & 7.18 (4H, 2d, J = 8.5 Hz aromatic protons),
6.8 (1H, d, J = 8 Hz, urethane NH), 5.3 (1H, m, 3-methine proton),
4.1 (1H, m, phenylalanine methine proton), 4.1 & 4.3 (2H, 2 m,
1-methylene protons), 2.95–2.75 (2H, 2 m, 2-methylene protons),
2.95 & 2.4 (2H, 2 m, phenylalanine methylene protons), 2.5 (3H, s,
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 1577–1587 | 1583
©

View Online
methyl), 2.2 (3H, s, phenylalanine methyl), 1.28 (9H, m, t-Boc
protons); IR: (KBr pellet) 3329, 2929, 2853, 1681, 1628, 1538,
1445, 1369, 1304, 1246, 1166, 1051, 877 cm-1; HRMS: Calculated
572.1508 for C26H3179BrN5O5, found 572.1535, calculated 574.1488
for C26H3181BrN5O5, found 574.1489.
phenylalanine methine proton), 4.1 & 4.3 (2H, 2 m, 1-methylene
protons), 2.95 & 2.7 (2H, 2 m, phenylalanine methylene protons),
2.9 & 2.4 (2H, 2 m, 2-methylene protons), 2.0 (3H, s, methyl
protons), 1.35 (9H, m, t-Boc protons); IR (KBr pellet) 3425, 2930,
1665, 1512, 1369, 1160 cm-1; HRMS Calculated: 483.2044 for
C25H28FN4O5, found 483.2031.
(3S)-6-Bromo-2,3-dihydro-7-methyl-5-nitro-3-[Boc-L-3-fluoro-
phenylalanyl]-1H-pyrrolo[1,2-a]benzimidazole (6e). 387 mg (72%
yield); mp: 183–190 ◦C; TLC: [9 : 1 dichloromethane–methanol] Rf
0.60; 1H NMR (DMSO-d6) 8.7 (1H, d, J ~ 8 Hz, amide NH), 8.2
(1H, s, 8-H), 7.8 (1H, d, J ~ 1 Hz, aromatic CH ortho to F), 7.1 &
7.3 (3H, 2 m, aromatic protons), 6.95 (1H, d, J ~ 8 Hz, urethane
NH), 5.4 (1H, m, 3-methine proton), 4.1 (1H, m, phenylalanine
methine proton), 4.0 & 4.3 (2H, 2 m, 1-methylene protons), 3.0 &
2.4 (2H, 2 m, phenylalanine methylene protons), 3.0 & 2.75 (2H, 2
m, 2-methylene protons), 2.51 (3H, m, methyl protons), 1.35 (9H,
m, t-Boc protons); IR: (KBr pellet) 3330, 2932, 1676, 1629, 1527,
1450, 1367, 1250, 1166, 783 cm-1; HRMS: Calculated 576.1258 for
C25H28BrFN5O5, found 576.1237.
(3S)-2,3-Dihydro-7-methyl-3-[Boc-L-4-trifluoromethylphenyl-
alanyl]-1H-pyrrolo[1,2-a]benzimidazole-5,8-dione (7c). 50 mg
(14% yield); mp 144–155 ◦C; TLC: [9 : 1 dichloromethane–
methanol] Rf 0.55; H NMR (DMSO-d6): 8.6 (1H, d, J = 8 Hz,
1
amide NH), 7.6 & 7.4 (4H, 2d, J = 8 Hz, aromatic protons), 7
(1H, d, J = 8 Hz, urethane NH), 6.55 (1H, s, 6-H), 5.2 (1H, m, 3-
methine proton), 4.05 (1H, m, phenylalanine methine proton),
4.3 & 4.1 (2H, 2 m, 1-methylene protons), 3.05 & 2.8 (2H, 2
m, phenylalanine methylene protons), 3.0 & 2.35 (2H, 2 m, 2-
methylene protons), 2.0 (3H, s, methyl protons), 1.35 (9H, m, t-
Boc protons); IR (KBr pellet): 3426, 2931, 1658, 1525, 1327, 1248,
1164, 1124 cm-1; HRMS Calculated: 533.2012 for C26H28F3N4O5,
found 533.2061.
General synthesis of intermediate 7
(3S)-2,3-Dihydro-7-methyl-3-[Boc-L-4-methylphenylalanyl]-1H-
pyrrolo[1,2-a]benzimidazole-5,8-dione (7d). 29 mg (9% yield);
mp: 164–172 ◦C; TLC: [9 : 1 dichloromethane–methanol] Rf 0.62;
1H NMR (DMSO-d6): 8.6 (1H, d, J = 8 Hz, amide proton),
7.15 & 7.05 (4H, 2d, J = 8 Hz, aromatic protons), 6.8 (1H,
d, J = 8 Hz, urethane NH), 6.57 (1H, s, 6-H), 5.2 (1H, m,
3-methine proton), 4.3 & 4.1 (2H, 2 m, 1-methylene protons),
4.1 (1H, m, phenylalanine methine proton), 2.95 & 2.75 (2H, 2
m, phenylalanine methylene protons), 2.95 & 2.30 (2H, 2 m, 2-
methylene protons), 2.2 & 2.0 (6H, 2 s, methyl protons), 1.35 (9H,
m, t-Boc protons); IR (KBr pellet) 3403, 2928, 1677, 1523, 1384,
1161 cm-1; HRMS Calculated: 479.2294 for C26H31N4O5, found
479.2295.
To a degassed solution of 387 mg (0.67 mmol) of 6 in 50 mL of
methanol was added 70 mg of 5% activated Pd/C. The resulting
mixture was hydrogenated at 50 psi for 6 h. The catalyst was
filtered off through Celite, and the methanol was removed from
the filtrate under reduced pressure. A solution consisting of 80 mL
of double distilled H2O containing 1.2 g of potassium phosphate
monobasic was added to the residue of the filtrate. To this solution
was added 1 g of freshly prepared Fremy’s salt, and the reaction
stirred at room temperature for 1 h. The completed reaction
was extracted with ethyl acetate until the organic extracts were
colorless. The ethyl acetate was removed under reduced pressure
to yield a brown residue of 7 along with trace deblocked (t-Boc)
side product. This residue was dissolved in a minimum amount of
chloroform and purified via silica gel chromatography employing
ethyl acetate as the eluant. The yellow product was recrystallised
from chloroform–hexane to afford analytically pure 7 as a yellow
solid.
(3S)-2,3-Dihydro-7-methyl-3-[Boc-L-3-fluorophenylalanyl]-1H-
pyrrolo[1,2-a]benzimidazole-5,8-dione (7e). 71 mg (22% yield);
mp: 143–152 ◦C; TLC [9 : 1 dichloromethane–methanol]: Rf 0.58;
1H NMR (DMSO-d6): 8.6 (1H, d, J = 8 Hz, amide proton), 7.25 &
7.05 (4H, 2 m, aromatic protons), 6.95 (1H, d, J = 8 Hz, urethane
NH), 6.57 (1H, s, 6-H), 5.2 (1H, m, 3-methine proton), 4.05 (1H,
m, phenylalanine methine proton), 4.3 & 4.1 (2H, 2 m, 1-methylene
protons), 2.95 & 2.75 (2H, 2 m, phenylalanine methylene protons),
2.9 & 2.30 (2H, 2 m, 2-methylene protons), 2.0 (3H, s, 7 methyl
protons), 1.35 (9H, m, t-Boc protons); IR (KBr pellet): 3331, 2931,
2854, 1663, 1528, 1520, 1368, 1249, 1160 cm-1; HRMS Calculated:
483.2044 for C25H28FN4O5, found 483.2048.
(3S)-2,3-Dihydro-7-methyl-3-[Boc-L-4-methoxyphenylalanyl]-
1H-pyrrolo[1,2-a]benzimidazole-5,8-dione (7a). 45 mg (13%
yield); mp: 118–123 ◦C; TLC: [9 : 1 dichloromethane–methanol]
Rf 0.58; 1H NMR (DMSO-d6) 8.6 (1H, d, J = 8 Hz, amide NH),
7.18 & 6.8 (4H, 2d, J = 8.5 Hz aromatic protons), 6.8 (1H, d, J =
8 Hz, urethane NH), 6.6 (1H, s, 6-proton), 5.2 (1H, m, 3-methine
proton), 4.05 (1H, m, phenylalanine methine proton), 4.3 & 4.1
(2H, 2 m, 1-methylene protons), 3.7 (3H, s, methoxy protons),
2.95 & 2.75 (2H, m & t respectively, J = 8 Hz, phenylalanine
methylene protons), 2.9 & 2.3 (2H, 2 m, 2-methylene protons), 2.0
(3H, s, methyl protons), 1.30 (9H, m, t-Boc protons); IR, (KBr
pellet) 3337, 2933, 2857, 1704, 1662, 1514, 1450, 1369, 1249, 1172,
1039, 827 cm-1; HRMS Calculated: 495.2244 for C25H31N4O6,
found 495.2234.
General synthesis of 2 from intermediate 7
This transformation was carried out in two steps: removal of the
t-Boc protecting group with trifluoroacetic acid (TFA) followed
by aziridination. A solution of 0.1 mmol of 7 in 10 mL and
TFA (0.25 mL) was stirred at room temperature. The reaction
was monitored by TLC (9 : 1 dichloromethane–methanol). Upon
completion, the reaction solvent was removed under reduced
pressure. The residue was dissolved in 15 mL of anhydrous
methanol and cooled to 0 ◦C. To this mixture was added 0.4 mL
of aziridine and stirring continued at 0 ◦C for 30 min. The reaction
mixture was then allowed to warm to room temperature and stirred
for 3 h. The solvent was removed under reduced pressure and
(3S)-2,3-Dihydro-7-methyl-3-[Boc-L-4-fluorophenylalanyl]-1H-
pyrrolo[1,2-a]benzimidazole-5,8-dione (7b). 66 mg (23% yield);
mp: 142–149 ◦C; TLC [9 : 1 dichloromethane–methanol] Rf 0.60;
1H NMR (DMSO-d6) 8.6 (1H, d, J = 8 Hz, amide NH), 7.25 & 7.05
(4H, 2 m, aromatic protons), 6.9 (1H, d, J = 8 Hz, urethane NH),
6.5 (1H, s, 6-proton), 5.2 (1H, m, 3-methine proton), 4.05 (1H, m,
1584 | Org. Biomol. Chem., 2010, 8, 1577–1587
This journal is
The Royal Society of Chemistry 2010
©

View Online
the residue was dissolved in a minimum amount of chloroform
which was spotted onto a preparatory TLC plate employing 9 : 1
dichloromethane and methanol. The bright red band was scraped
off the plate and extracted with ethyl acetate. The ethyl acetate
was removed under reduced pressure, and 2 was recrystallized
from chloroform–hexane. The yield of purified product ranged
from 2 to 7%.
(3S)-6-Aziridinyl-2,3-dihydro-7-methyl-3-(L-4-methylphenyl-
alanylamino)-1H-pyrrolo[1,2-a]benzimidazole-5,8-dione
(2d).
Yield: 22 mg (7.00%); mp: 149–150 ◦C (dec); TLC [9 : 1
1
dichloromethane–methanol]: Rf 0.48; H NMR (CDCl3) d 8.24
(1H, br s, amide NH), 6.99 (4H, m, aromatic protons), 5.12 (1H,
q, J = 7.20 Hz, 3-methine proton), 4.20–4.31 (1H, m, 1-methylene
proton), 3.99–4.08 (1H, m, 1-methylene proton), 3.60–3.70 (1H,
dd, J = 3.9 Hz, J = 3.6 Hz, phenylalanine methine proton),
3.05–3.12 (1H, dd, J = 3.9 Hz, J = 3.6 Hz, phenylalanine
methylene proton), 2.973–3.03 (1H, dd, J = 4.80 Hz, J = 3.3 Hz,
phenylalanine methylene proton), 2.62–2.70 (1H, m, 2-methylene
proton), 2.4–2.5 (1H, m, 2-proton), 2.24 (4H, s, aziridine), 2.20
(3H, s, methyl protons), 1.94 (3H, s, methyl protons). IR (thin
film) 3324, 2924, 1675, 1583, 1518, 1486, 1379, 1313, 1250, 1140,
1098, 1032, 991, 783 cm-1; MALDI: calculated for C23H25N5O3 +
H+ (M + H)+ 420.203, found 420.208; calculated for C23H25N5O3 +
Na+ (M + Na)+ 442.186, found 442.186.
(3S)-6-Aziridinyl-2,3-dihydro-7-methyl-3-(L-p-methoxyphenyl-
alanylamino)-1H-pyrrolo[1,2-a]benzimidazole-5,8-dione
(2a).
Yield: 15 mg (4.59%); mp 145–146 ◦C (dec); TLC
[dichloromethane–methanol (90 : 10)]: Rf 0.39; 1H NMR
(CDCl3) d 8.29 (1H, br s, amide NH), 7.11 (2H, d, J = 8.5 Hz
aromatic protons), 6.82 (2H, d, J = 8.5 Hz aromatic proton), 5.21
(1H, m, 3-methine proton), 4.34 (1H, m, 1-methylene proton),
4.12 (1H, m, 1-methylene proton), 3.77 (3H, s, methoxy protons),
3.68 (1H, br s, phenylalanine methine proton), 3.12 (2H, dd, J =
10.5 Hz, J = 3.5 Hz, phenylalanine methylene proton), 2.74 (1H,
m, 2-methylene proton), 2.51 (1H, m, 2-methylene proton), 2.33
(4H, s, aziridine), 2.02 (3H, s, methyl protons); IR (thin film)
3311, 2924, 2854, 1672, 1581, 1512, 1461, 1378, 1302, 1247, 1179,
1139, 1033, 812 cm-1; MALDI: calculated for C23H25N5O4 + H+
(M + H)+ 436.198, found 436.195; calculated for C23H25N5O4 +
Na+ (M + Na)+ 458.180, found 458.173.
(3S)-6-Aziridinyl-2,3-dihydro-7-methyl-3-(L-m-fluorophenyl-
alanylamino)-1H-pyrrolo[1,2-a]benzimidazole-5,8-dione
(2e).
Yield: 23 mg (7.24%); mp: 294–295 ◦C (dec); TLC [9 : 1
1
dichloromethane–methanol]: Rf 0.35; H NMR (CDCl3): d 8.36
(1H, br s, amide NH), 7.17 (1H, d, J = 7.50 Hz, aromatic protons),
6.93 (1H, d, J = 7.50 Hz, aromatic proton), 6.82–6.88 (2H, m, J =
9.3 Hz, J = 8.7 Hz aromatic protons), 5.14 (1H, br s, 3-methine
proton), 4.23–4.32 (1H, m, 1-methylene proton), 4.00–4.10 (1H,
m, 1-methylene proton), 3.71 (1H, s, phenylalanine methine
proton), 3.01–3.17 (2H, m, phenylalanine methylene proton), 2.81
(1H, m, 2-methylene proton), 2.47 (1H, m, 2-methylene proton),
2.28 (4H, s, aziridine), 1.95 (3H, s, methyl protons); IR (thin film)
3329, 2926, 1675, 1639, 1582, 1518, 1378, 1312, 1250, 1139, 1072,
1032 cm-1; MALDI: calculated for C22H22N5O3F + H+ (M + H)+
424.178, found 424.180; calculated for C22H22N5O3F + Na+ (M +
Na)+ 446.161, found 446.163.
(3S)-6-Aziridinyl-2,3-dihydro-7-methyl-3-(L-p-fluorophenyl-
alanylamino)-1H-pyrrolo[1,2-a]benzimidazole-5,8-dione
(2b).
Yield: 21 mg (6.62%); mp: 275–276 ◦C (dec); TLC: [9 : 1
1
dichloromethane–methanol] Rf 0.33; H NMR (CDCl3): d 8.16
(1H, br s, amide NH), 7.06 (2H, dd, J = 7.8 Hz, J = 5.4 Hz
aromatic protons), 6.88 (2H, dd, J = 8.7 Hz, J = 8.1 Hz
aromatic protons), 5.11 (1H, m, 3-methine proton), 4.23 (1H, m,
1-methylene proton), 4.05 (1H, m, 1-methylene proton), 3.59 (1H,
dd, J = 5.40 Hz, J = 5.10 Hz, phenylalanine methine proton),
3.05 (2H, m, phenylalanine methylene proton), 2.71 (1H, m,
2-methylene proton), 2.40 (1H, m, 2-methylene proton), 2.24
(4H, s, aziridine), 1.93 (3H, s, methyl protons); IR (thin film):
3328, 2924, 2854, 1675, 1639, 1577, 1511, 1378, 1339, 1313, 1221,
1139, 1099 cm-1; MALDI: calculated for C22H22N5O3F + H+ (M +
H)+ 424.178, found 424.181; calculated for C22H22N5O3F + Na+
(M + Na)+ 446.161, found 446.161.
In vivo studies
Either male or female SCID mice were implanted with the
xenograft tumour model for human lung (NCI-H460). Compound
LL-1 was administrated intraperitoneally at doses of 1, 3, and
5 mg kg-1 at 4 day intervals for a total of 6 doses. Each mL of
the solvent for LL-1 consisted of 2 mg citric acid, 30 mg benzyl
alcohol, 80 mg polyoxyethylene sorbitan monooleate (tween 80),
650 mg polyethylene glycol 300 and 30.5% ethanol (w/w). The
tumour size, body weight and signs of overt animal toxicity after
LL-1 injection were monitored for 25 days.
(3S)-6-Aziridinyl-2,3-dihydro-7-methyl-3-(L-p-trifluoromethyl-
phenylalanylamino)-1H -pyrrolo[1,2-a]benzimidazole-5,8-dione
(2c). Yield: 8 mg (2.25%); mp: 182–184 ◦C (dec); TLC [9 : 1
1
dichloromethane–methanol]: Rf 0.28; H NMR (CDCl3): d 8.45
(1H, br s, amide NH), 7.55 (2H, d, J = 6.5 Hz aromatic protons),
7.36 (2H, d, J = 7.5 Hz aromatic protons), 5.20 (1H, br s, 3-
methine proton), 4.34 (1H, m, 1-methylene proton), 4.12 (1H, m, 1-
methylene proton), 3.84 (1H, br s, p-trifluoromethyl phenylalanine
methine proton), 3.28 (1H, d, J = 12.5 Hz, trifluoromethyl
phenylalanine methylene proton), 3.09 (1H, d, J = 7.5 Hz,
trifluoromethyl phenylalanine methylene proton), 2.93 (1H, m, 2-
methylene proton), 2.52 (1H, m, 2-methylene proton), 2.34 (4H, s,
aziridine), 2.02 (3H, s, methyl protons); IR (thin film): 3325,
2922, 2853, 1724, 1553, 1500, 1461, 1374, 1319, 1191, 1038 cm-1;
MALDI: calculated for C23H22N5O3F3 + H+ (M + H)+ 474.175,
found 474.170; calculated for C23H22N5O3F3 + Na+ (M + Na)+
496.156, found 496.151.
Acute toxicity assays
A concentration of LD-1 was prepared, consistent with adminis-
tration of 400 mg kg-1 intraperitoneally, in a volume of 1 mL. Treat
each SCID mouse with a single dose of 400, 200, and 100 mg kg-1
(1 mL, 0.5 mL, and 0.25 mL/mouse). The body weight and strain
of mouse used should be consistent with that to be used in the
subsequent antitumor evaluation protocols. Hold the animals for
a period of 14 days observing for morbidity (body weight loss)
and mortality. If the mice do not survive for 14 days, administer
a 50% lower dose to a single mouse and continue this process
(50 mg kg-1, 25 mg kg-1, 12.5 mg kg-1, etc.) until a tolerated dosage
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 1577–1587 | 1585
©

View Online
is determined. Use the maximum tolerated dosage as the basis for
selecting doses for antitumor drug evaluations.
8 D. Siegel, H. Beall, C. Senekowitsch, M. Kasai, H. Arai, N. W. Gibson
and D. Ross, Biochemistry, 1992, 31, 7879.
9 P. Schiltz and H. Kohn, J. Am. Chem. Soc., 1993, 115, 10510.
10 J. Cummings, Drug Resist. Updates, 2000, 3, 143.
11 W. Ping, S. Yang, Z. Lixia, H. Hanping and Z. Xiang, Curr. Med.
Chem., 2005, 12, 2893.
Hollow fibre assays48,49
12 K. U. Schallreuter and J. M. Wood, Biochem. Biophys. Res. Commun.,
1999, 262, 423.
13 I. Islam and E. B. Skibo, J. Org. Chem., 1990, 55, 3195.
14 E. B. Skibo and W. G. Schulz, J. Med. Chem., 1993, 36, 3050.
15 W. G. Schulz, E. Islam and E. B. Skibo, J. Med. Chem., 1995, 38, 109.
16 E. B. Skibo, Curr. Med. Chem., 1996, 2, 900.
17 E. B. Skibo, I. Islam, W. G. Schulz, R. Zhou, L. Bess and R. Boruah,
Synlett, 1996, 297.
18 R. Zhou and E. B. Skibo, J. Med. Chem., 1996, 39, 4321.
19 E. S. Skibo, S. Gordon, L. Bess, R. Boruah and J. Heileman, J. Med.
Chem., 1997, 40, 1327.
20 W. A. Craigo, B. W. LeSueur and E. B. Skibo, J. Med. Chem., 1999, 42,
3324.
21 X. Huang, A. Suleman and E. B. Skibo, Bioorg. Chem., 2000, 28, 324.
22 A. Ghodousi, X. Huang, Z. Cheng and E. B. Skibo, J. Med. Chem.,
2004, 47, 90.
23 R. I. Christopherson, S. D. Lyons and P. K. Wilson, Acc. Chem. Res.,
2002, 35, 961.
24 A. G. Zimmermann, J. J. Gu, J. Laliberte and B. S. Mitchell, Inosine-
5¢-monophosphate dehydrogenase: Regulation of expression and role
in cellular proliferation and T lymphocyte activation, in Progress in
Nucleic Acid Research and Molecular Biology, Vol. 61, ed. K. Moldave,
Academic Press Inc., San Diego, 1998, 181209.
Mice are treated with either a high or a low dose using a QD ¥
4 schedule (four daily treatments) administered intraperitoneally.
Altogether, twelve cell lines are studied resulting in 48 possible test
combinations (12 cell lines ¥ 2 sites ¥ 2 doses). A score of two is
given to each test in which there is a % T/C of 50 or less (tumor
mass 50% or less than the control). Thus the highest possible score
is a 96, but the typical score is only 5 and the highest score achieved
so is a 64. The score is broken down into a intraperitoneal (IP)
and a subcutaneous (SC) score. A good SC score (≥ 8) indicates
that the drug is able to get to the tumour site (subcutaneous) from
a distant site (intraperitoneal) of injection.
QSAR calculations
Multivarible regression analysis was carried out using GraphPad
InStat. Substituent constants were obtained from Hansch and
Leo.53
IMP dehydrogenase inhibition assays
25 Y. Natsumeda and S. F. Carr, Ann. N. Y. Acad. Sci., 1993, 696, 88.
26 P. Franchetti, L. Cappellacci and M. Grifantini, Farmaco, 1996, 51,
457.
These assays were carried out as previously described by this
laboratory.59 The IMP dehydrogenase activity was measured at
340 nm in 1.0 cm path length cuvette maintained at 30 ◦C in
a CARY-14 UV/VIS spectrophotometer. The final assay volume
was set to 1.0 mL and contained 100 mM Tris-HCl (pH 8.0),
100 mM KCl, 5 mM mercaptoethanol, 3 mM EDTA, 100 mL
type II IMP dehydrogenase, 1 mM DTT, 0.40 mM IMP, and
0.30 mM NAD+. The solution mixture, including everything
except the enzyme, was placed in the spectrophotometer and was
allowed to temperature equilibrate for 5 min. The dehydrogenase
reaction was initiated upon addition of type II IMP dehydrogenase
and formation of NADH was monitored for 30 min. Next,
to determine the Km value concentration of IMP was varied
from 0.25 mM to 25 mM and corresponding rates in terms of
optical density were obtained. To determine activity of type II
IMP dehydrogenase in the presence of potential inhibitors, the
experiment was repeated in the presence of 0.01 and 0.005 mM 2f.
27 K. W. Pankiewicz, Expert Opin. Ther. Pat., 2001, 11, 1161.
28 N. Minakawa and A. Matsuda, Curr. Med. Chem., 1999, 6, 615.
29 K. W. Pankiewicz, S. E. Patterson, P. L. Black, H. N. Jayaram, D. Risal,
B. M. Goldstein, L. J. Stuyver and R. F. Schinazi, Curr. Med. Chem.,
2004, 11, 887.
30 K. Gharehbaghi, W. Grunberger and H. N. Jayaram, Curr. Med. Chem.,
2002, 9, 743.
31 J. Jain, S. J. Almquist, D. Shlyakhter and M. W. Harding, J. Pharm.
Sci., 2001, 90, 625.
32 W. Markland, T. J. McQuaid, J. Jain and A. D. Kwong, Antimicrob.
Agents Chemother., 2000, 44, 859.
33 M. D. Sintchak and E. Nimmesgern, Immunopharmacology, 2000, 47,
163.
34 L. A. Sorbera, J. S. Silvestre, J. Castaner and L. Martin, Drugs Future,
2000, 25, 809.
35 Y. Jin, H.-Y. Li, L.-P. Lin, J. Tan, J. Ding, X. Luo and Y.-Q. Long,
Bioorg. Med. Chem., 2005, 13, 5613.
36 S. H. Watterson, T. G. M. Dhar, S. K. Ballentine, Z. Q. Shen, J. C.
Barrish, D. Cheney, C. A. Fleener, K. A. Rouleau, R. Townsend, D. L.
Hollenbaugh and E. J. Iwanowicz, Bioorg. Med. Chem. Lett., 2003, 13,
1273.
37 T. D. Colby, K. Vanderveen, M. D. Strickler, G. D. Markham and B. M.
Goldstein, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, 3531.
38 E. B. Skibo and R. B. Meyer, J. Med. Chem., 1981, 24, 1155.
39 T. D. Colby, K. Vanderveen, M. D. Strickler, G. D. Markham and B. M.
Goldstein, Proc. Natl. Acad. Sci. U. S. A., 1999, 96, 3531.
40 D. K. Paull, R. H. Shoemaker, L. Hodes, A. Monks, D. A. Scudiero,
L. Rubinstein, J. Plowman and M. R. Boyd, J. Natl. Cancer Inst., 1989,
81, 1088.
Acknowledgements
We wish to thank the Arizona Biomedical Research Commission
for their generous support.
41 D. S. Swaffar, C. M. Ireland and L. R. Barrows, Anti-Cancer Drugs,
1994, 5, 15.
42 E. B. Skibo and C. Xing, Biochemistry, 1998, 37, 15199.
43 I. Islam, E. B. Skibo, R. T. Dorr and D. S. Alberts, J. Med. Chem.,
1991, 34, 2954.
44 E. B. Skibo, I. Islam, M. J. Heileman and W. G. Schulz, J. Med. Chem.,
1994, 37, 78.
References
1 D. T. Ross, U. Scherf, M. B. Eisen, C. M. Perou, C. Rees, P. Spellman, V.
Iyer, S. S. Jeffrey, M. Van De Rijn, M. Waltham, A. Pergamenschikov,
J. C. E. Lee, D. Lashkari, D. Shalon, T. G. Myers, J. N. Weinstein, D.
Botstein and P. O. Brown, Nat. Genet., 2000, 24, 227.
2 P. L. Gutierrez, Free Radical Biol. Med., 2000, 29, 263.
3 A. M. Rauth, Z. Goldberg and V. Misra, Oncol. Res., 1997, 9, 339.
4 M. D. Garrett and P. Workman, Eur. J. Cancer, 1999, 35, 2010.
5 I. J. Stratford and P. Workman, Anti-Cancer Drug Des., 1998, 13, 519.
6 P. Workman, Oncol. Res., 1994, 6, 461.
7 R. W. Franck and M. Tomasz, The Chemistry of Mitomycins, in The
Chemistry of Antitumor Agents, ed. D. E. Wilman, Blackie & Sons,
Ltd., Glasgow, Scotland, 1990, 379394.
45 M. R. Boyd, Status of the NCI Preclinical Antitumor Drug Discovery
Screen, in Principles and Practices of Oncology (PPO), 3 ed., ed.
V. T. DeVita, S. Hellman and S. A. Rosenberg, J. B. Lippincott,
Phildelphia, Pa., USA, 1989, Vol. 3, 112.
46 M. R. Boyd and K. D. Paull, Drug Dev. Res., 1995, 34, 91.
47 A. Monks, D. A. Scudiero, G. S. Johnson, K. D. Paull and E. A.
Sausville, Anti-Cancer Drug Des., 1997, 12, 533.
1586 | Org. Biomol. Chem., 2010, 8, 1577–1587
This journal is
The Royal Society of Chemistry 2010
©

View Online
48 M. G. Hollingshead, M. C. Alley, R. F. Camalier, B. J. Abbott, J. G.
Mayo, L. Malspeis and M. R. Grever, Life Sci., 1995, 57, 131.
49 M. G. Hollingshead, J. Plowman, M. Alley, J. Mayo and E. Sausville,
Relevance of Tumor Models in Anticancer Drug Development, in
Contrib. Oncol., Krager Verlag, Germany, 1999, 109120.
50 H. H. Sedlacek, J. Czech, R. Naik, G. Kaur, P. Worland, M. Losiewicz,
B. Parker, B. Carlson, A. Smith, A. Senderowicz and E. Sausville,
Int. J. Oncol., 1996, 9, 1143.
51 L. Hodes, K. Paull, A. Koutsoukos and L. Rubinstein, J. Biopharm.
Stat., 1992, 2, 31.
52 C. Hansch, Acc. Chem. Res., 2002, 2, 232.
53 C. Hansch and A. Leo, ‘Substituent Constants For Correlation Analysis
in Chemistry and Biology’, Wiley-Interscience, 1979.
54 S. Heimerl, A. K. Bosserhoff, T. Langmann, J. Ecker and G. Schmitz,
Melanoma Res., 2007, 17, 265.
55 N. Y. Frank, A. Margaryan, Y. Huang, T. Schatton, A. M. Waaga-
Gasser, M. Gasser, M. H. Sayegh, W. Sadee and M. H. Frank, Cancer
Res., 2005, 65, 4320.
56 L. Wang, J. Yu, J. Ni, X.-M. Xu, J. Wang, H. Ning, X.-F. Pei, J. Chen, S.
Yang, C. B. Underhill, L. Liu, J. Liekens, J. Merregaert and L. Zhang,
Cancer Lett., 2003, 200, 57.
57 J. E. McLean, N. Hamaguchi, P. Belenky, S. E. Mortimer, M. Stanton
and L. Hedstrom, Biochem. J., 2004, 379, 243.
58 A. Suleman and E. B. Skibo, J. Med. Chem., 2002, 45, 1211.
59 M. E. Pugh and E. B. Skibo, Comp. Biochem. Physiol., 1993, 105B,
381.
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 1577–1587 | 1587
©