Facile synthesis of 1,2-dione-containing abietane analogues for the generation of human carboxylesterase inhibitors

Article abstract of DOI:10.1016/j.ejmech.2018.02.052

Recently, a series of selective human carboxylesterase inhibitors have been identified based upon the tanshinones, with biologically active molecules containing a 1,2-dione group as part of a naphthoquinone core. Unfortunately, the synthesis of such compounds is complex. Here we describe a novel method for the generation of 1,2-dione containing diterpenoids using a unified approach, by which boronic acids are joined to vinyl bromo-cyclohexene derivatives via Suzuki coupling, followed by electrocyclization and oxidation to the o-phenanthroquinones. This has allowed the construction of a panel of miltirone analogues containing an array of substituents (methyl, isopropyl, fluorine, methoxy) which have been used to develop preliminary SAR with the two human carboxylesterase isoforms. As a consequence, we have synthesized highly potent inhibitors of these enzymes (K_i < 15 nM), that maintain the core tanshinone scaffold. Hence, we have developed a facile and reproducible method for the synthesis of abietane analogues that have resulted in a panel of miltirone derivatives that will be useful tool compounds to assess carboxylesterase biology.

Full text of DOI:10.1016/j.ejmech.2018.02.052

Accepted Manuscript
Facile synthesis of 1,2-dione-containing abietane analogues for the generation of
human carboxylesterase inhibitors
Randall J. Binder, M. Jason Hatfield, Liying Chi, Philip M. Potter
PII:
S0223-5234(18)30188-0
10.1016/j.ejmech.2018.02.052
EJMECH 10229
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 20 December 2017
Revised Date: 13 February 2018
Accepted Date: 14 February 2018
Please cite this article as: R.J. Binder, M.J. Hatfield, L. Chi, P.M. Potter, Facile synthesis of 1,2-dione-
containing abietane analogues for the generation of human carboxylesterase inhibitors, European
Journal of Medicinal Chemistry (2018), doi: 10.1016/j.ejmech.2018.02.052.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Facile synthesis of 1,2-dione-containing abietane analogues for the generation of
human carboxylesterase inhibitors
Randall J. Binder, M. Jason Hatfield, Liying Chi and Philip M. Potter
Department of Chemical Biology and Therapeutics,
262 Danny Thomas Place,
St. Jude Children's Research Hospital,
Memphis, TN 38105,
United States
Email addresses:
Randall.binder@stjude.org
Jason.hatfield@stjude.org
Liying.chi@stjude.org
Phil.potter@stjude.org
Corresponding author.
Philip M. Potter
Department of Chemical Biology and Therapeutics,
St. Jude Children's Research Hospital,
262 Danny Thomas Place,
Memphis,
TN 38105, United States
Tel: 901-595-6045
Fax: 901-595-4293
1

ACCEPTED MANUSCRIPT
E-mail: phil.potter@stjude.org.
2

ACCEPTED MANUSCRIPT
Keywords
Abietane analogues, carboxylesterase; enzyme inhibition; synthesis.
3

ACCEPTED MANUSCRIPT
ABSTRACT
Recently, a series of selective human carboxylesterase inhibitors have been identified
based upon the tanshinones, with biologically active molecules containing a 1,2-dione
group as part of a naphthoquinone core. Unfortunately, the synthesis of such
compounds is complex. Here we describe a novel method for the generation of 1,2-
dione containing diterpenoids using a unified approach, by which boronic acids are
joined to vinyl bromo-cyclohexene derivatives via Suzuki coupling, followed by
electrocyclization and oxidation to the o-phenanthroquinones. This has allowed the
construction of a panel of miltirone analogues containing an array of substituents
(methyl, isopropyl, fluorine, methoxy) which have been used to develop preliminary SAR
with the two human carboxylesterase isoforms. As a consequence, we have
synthesized highly potent inhibitors of these enzymes (Ki <15nM), that maintain the core
tanshinone scaffold. Hence, we have developed a facile and reproducible method for
the synthesis of abietane analogues that have resulted in a panel of miltirone
derivatives that will be useful tool compounds to assess carboxylesterase biology.
4

ACCEPTED MANUSCRIPT
INTRODUCTION
Carboxylesterases (CEs1) are ubiquitous enzymes that are responsible for the
hydrolysis of ester containing molecules, typically xenobiotics.[1-3] This includes natural
products such as cocaine and heroin[4, 5], many clinically used drugs (oseltamivir,
meperidine, lidocaine, procaine, etc)[6, 7] as well as agents used in the agricultural
industry (e.g., malathion, aldicarb).[8, 9] Since CEs are responsible for the activation
and inactivation of such compounds, modulating their enzymatic activity may have a
significant effect on the biodistribution and toxicity of the small molecules. Indeed, we
have demonstrated that compounds such as benzil, a potent inhibitor of human CEs,
can significantly modulate the cytotoxicity of CPT-11, an anticancer prodrug that this
activated by this class of enzymes.[10, 11] In efforts to try and identify CE inhibitors in
natural products, we recently determined that the Chinese herbal medicine Danshen
contains significant amounts of such compounds. These molecules (principally,
tanshinone IIA and miltirone), share structural features with benzil (1,2-dione moiety),
are cell permeable, and can modulate ester-containing prodrug hydrolysis mediated by
CEs.[12]
The abietane class of terpenoids have a unique scaffold composed of polycyclic
fused rings, and may contain an ortho-quinone moiety (e.g., miltirone and the related
tanshinones). While several reports describe the generation of these compounds to
produce libraries of analogues[13, 14], until now, these synthetic efforts required
1 Abbreviations used in this manuscript: CE – carboxylesterase; CPT-11 – irinotecan, 7-
ethyl-10-[4-(1-piperidino)-1-piperidino]carbonyloxycamptothecin; hAChE – human
acetylcholinesterase; hBChE – human butyrylcholinesterase; hCE1 – human liver CE;
hiCE – human intestinal CE; o-NPA – o-nitrophenyl acetate.
5

ACCEPTED MANUSCRIPT
upwards of 10 linear steps in order to furnish each compound.[15-17] In an effort to use
these molecules to act as probes to understand the architecture of the CE active sites,
and potentially design prototypic molecules that could be used to modulate drug
disposition in vivo, we sought to generate a panel of tanshinone analogues. However,
due to the complexity of the synthesis of such compounds, a novel route was required.
The scheme that we have devised takes advantage of a modular approach, suitable for
rapid library generation and evaluation of new compounds.
Initially, we attempted to replicate the work by Snyder and coworkers, who
employed sonication-enhanced Diels-Alder annulations.[15, 16] However, these
intermolecular cyclizations were relatively inefficient, and were compounded by the fact
that each of the various dienes were laborious to synthesize. Furthermore, a review of
the literature indicated that no truly efficient approach existed by which miltirone-like
compounds could be rapidly synthesized and that were amenable to producing defined
analogue libraries.[17-24] Therefore, drawing upon inspiration from Harvey and
coworkers, we devised a method by which decorated phenanthrols could be converted
to o-quinones.[25-27] The methodology that we have developed is relatively facile, can
be used to rapidly generate different regioisomers, and requires mild reagents and
protection strategies. We describe here, the chemistry associated with this approach
and the development of a small library of miltirones that demonstrate potent and
selective inhibition of human CEs.
6

ACCEPTED MANUSCRIPT
Materials and Methods
Reagents and instrumentation
Unless otherwise noted, all reactions were carried out under nitrogen using anhydrous
solvents. All reagents were used as-is from commercial sources and were of ACS
grade or better. A detailed description of the synthesis and characterization of all
compounds used in this article is reported in the Supporting Information. Compounds
were purified using silica cartridge and high pressure chromatography. NMR spectra
(1H, 13C and 19F) of purified molecules dissolved in DMSO-d6, CDCl3 or CHCl3, were
taken on either 400 MHz Bruker Avance or 500 MHz Bruker Ascend spectrometers with
TMS or fluorobenzene as standards. A Waters Acquity UPLC – Xevo G2 QTOF
spectrometer was used to determine purity and collect HRMS data.
Chromatographically pure human CEs, hCE1 and hiCE, were generated as
previously described[5, 28] and human acetyl- (hAChE) and butyrylcholinesterase
(hBChE) were obtained from Sigma Biochemicals (St., Louis, MO).
Enzyme inhibition
Inhibition of CEs was determined in 96 well plates using 3mM o-nitrophenyl acetate (o-
NPA) as a substrate.[10] Briefly, inhibitors (dissolved in DMSO) were aliquoted into
individual wells containing 50mM Hepes, pH7.4 and o-NPA, and the reaction was
initiated by the addition of enzymes. All data points assays were run in triplicate, at least
8 concentrations of inhibitor were used and samples were compared to DMSO-alone
treated controls. Results were fitted to a multifactorial equation[29] and then analyzed
using Akaike’s information criteria[30, 31] to assess the best model for CE inhibition.
Once established, data were plotted using Prism (GraphPad, La Jolla, CA) and Ki
values were determined from derived curve fits.
7

ACCEPTED MANUSCRIPT
A determination of the inhibition of hAChE or hBChE was undertaken in a similar
fashion except that acetylthiocholine or butyrylthiocholine were used as a substrates,
respectively.[32, 33] Inhibitors were assayed at a single concentration (5µM).
Computer-assisted docking
Compounds were docked into the active site of hCE1 using ICMPro software (Molsoft,
San Diego, CA) and the x-ray coordinates derived from PDB: 1MX1.[34] Small
molecules were minimized and docked within the active site using standard program
descriptors. To maximize the accuracy of the finals models, the ‘Thoroughness’
parameter was set to 10. Distances from the serine Oγ atom to the carbonyl carbon
atoms present within the 1,2-dione were then measured. Statistical correlations
between these datasets and the Ki values for the human CEs were evaluated using
Prism software.
8

ACCEPTED MANUSCRIPT
RESULTS AND DISCUSSION
Retrosynthetic strategy
Previous analysis of biologically active 1,2-diones and the tanshinone structures
indicated that this chemotype was essential for enzyme inhibition, presumably to act as
a target for the active site serine residue in CEs.[12] In addition, the potency and
substrate specificity of such compounds was significantly modulated by the groups
appended to the core cyclic structure. Using this as a guide, we developed a
retrosynthetic approach towards the synthesis of such molecules by joining two
fragments that would allow for the introduction of chemical diversity and generate the
1,2-dione chemotype (Scheme 1). This was accomplished by the use of Suzuki
reactions to couple vinyl bromo aldehydes (4) with phenylboronic acids (5) to yield enol
ethers (3). Subsequent acid-mediated electrocyclization of 3 would then generate the
tricyclic compounds (2), which following subsequent deprotection and oxidation would
yield the corresponding naphthols and o-naphthoquinones (1). This segmented
approach would be amenable to library generation since a myriad of substituted
phenylboronic acids and vinyl bromo-aldehydes can be readily obtained and/or
synthesized.
Therefore, to accomplish the above scheme, the synthesis of two classes of
reactants was required: a panel of bromovinyl aldehydes that allowed for the formation
of the central aromatic rings and the introduction of diversity on the left hand side of the
molecule; and a series of boronic acids or esters, that allow for the construction of the
1,2-dione and provides sufficient flexibility to permit inclusion of a variety of chemotypes
for analogue synthesis.
Development of required reagents
9

ACCEPTED MANUSCRIPT
a) Bromovinyl aldehydes
The appropriate vinylbromo-aldehydes 4 were synthesized by subjecting
cyclohexanone derivatives to Vilsmeier-Haack conditions[35]. Using this approach, we
have generated 18 compounds with yields ranging from 40 – 70%. In general, these
reactions proceeded well, and due to their versatility, we have been able to adapt them
to a panel of substituted cyclohexanes and benzene derivatives. As indicated in
Scheme 2, this allows for significant diversity to be introduced into these molecules, and
allows for generation of novel miltirone analogues.
b) Boronic acids/esters
In general, the boronic acids (5) used to generate the right hand side of the miltirone
analogues were either purchased or synthesized using standard approaches[36, 37].
Typically boronic acids were generated from bromobenzene derivatives with the desired
substituents. Treatment of the bromobenzenes with butyllithium followed immediately by
triisopropyl borate and quenching with aqueous acid provided the desired boronic acids
in good yields. Using this methodology, we acquired or generated 8 building blocks
(Scheme 3) that allowed for the inclusion of a variety of substituents at the 2- and 3-
positions. This included fluorine, simple alkyl substituents and methoxy groups, and
suggests that other chemotypes could be readily incorporated into these molecules.
Synthesis of miltirone analogues
Scheme 4 outlines the synthetic route used to generate miltirone analogues. The
vinylbromo-aldehydes 4 were converted into methyl enol ethers 7 in good yields (75-
80%) using Wittig reagent 6. This was then joined to boronic acids (or boronic ester) 5
in high yields by palladium tetrakis(triphenylphosphine)-catalyzed Suzuki coupling.[38]
Subsequent acid-catalyzed electrocyclization/aromatization furnished the tricyclic
10

ACCEPTED MANUSCRIPT
intermediate 2 and the minor regioisomer 8, which, following deprotection, gave rise to
phenols 9 and 10 over two steps. Subsequent treatment of these phenols with 2-
iodoxybenzoic acid (11) produces the o-quinones 1 and 12.[26] The route has been
optimized such that overall yields are as high as 38%, starting from the vinyl bromo-
aldehyde.
Synthetic efforts to generate and use derivatives of enol ether 7 were taken with
care as these dienes were relatively unstable and appeared to self-condense
(dependent upon the concentration of the solution). Due to their apparent limited shelf
life, these compounds were synthesized and typically used within 24 hours of
purification. If necessary, they were stored at 4C as dilute solutions in diethyl ether
before use. However, the overall yields of miltirone analogues generated from this
synthetic route from derivatives of enol ether 7 were up to 40%.
In some instances, the harsh conditions required to unmask the naphthol from a
methyl ether were not amenable to analogues with oxygen-containing functionalities,
requiring protecting groups that could be removed under milder deprotection conditions,
specifically as tert-butyldimethyl silyl ethers. In addition, to eliminate potential losses
from purification and chromatography, the crude tricyclic intermediates were directly
deprotected to yield the resulting naphthols. This step typically increased overall yield
by 10-15%.
Introduction of diversity
As shown, the above route allows for a fair degree of chemical diversity on either end of
the molecule. On the quinone portion, decorations are nearly limitless, depending upon
the commercial availability of boronic acids/esters or their ease of synthesis. Suzuki
couplings were conducted in tetrahydrofuran to maximize solubility of both coupling
11

ACCEPTED MANUSCRIPT
partners and were adapted from a similar coupling by Frontier and co-workers.[38] In
some cases, when yields of the Suzuki coupling were poor, a “ligandless” catalyst
method was employed. In a modified approach, vinylbromo aldehydes 4 were directly
attached to the boronic acid using a palladium acetate-mediated Suzuki coupling
(Scheme 5). The resulting vinyl benzene cyclohexenal products 14 were then converted
to the enol ether using the same Wittig reagent as per the normal route. This approach
was less desirable from a library-generation perspective as the branch point for adding
diversity is added earlier in the reaction sequence. However, this approach also
illustrates the broader applicability and overall flexibility of the strategy.
This approach allows for a fair degree of chemical diversity on either end of the
molecule by taking advantage of the modular approach (Table 1). Inclusion of a pyran
within the scaffold in the left-hand ring demonstrates the utility of this approach towards
generating heterocyclic analogues of these compounds. Significant modification also
can be introduced into the quinone moiety to provide numerous analogues, dependent
upon the commercial availability of boronic acids/esters and/or their ease of synthesis.
Occasionally the product quinones were relatively unstable, however this tended to
occur only with compounds containing electronegative substituents on the C ring. We
believe that compounds bearing electron-donating groups possess longer shelf-lives as
they are less prone to the aromatization. Additionally, LCMS data strongly suggested
that dimerization via an inverse-demand hetero Diels-Alder cycloaddition could account
for decomposition.
Inhibition of human CEs by miltirone analogues
Having generated a panel of miltirone analogues in good yield, we assessed their ability
to inhibit both hCE1 and hiCE, using o-NPA as a substrate. As indicated in Table 2, the
12

ACCEPTED MANUSCRIPT
majority of the novel compounds were more potent than the natural product towards the
former enzyme. For example, molecules 1Ad and 1Be yielded Ki values that were
~210- and ~175-fold lower, respectively, than that seen for miltirone. While potency
towards hiCE was comparable to that seen for the natural product, several analogues
demonstrated increased potency (e.g., 1Bd and 1Bf). Indeed, 1Bd was the most potent
non-selective analogue, inhibiting both hCE1 (Ki = 42nM) and hiCE (Ki = 15nM) with
similar efficacy. However, selectivity was also observed, and in some instances the
relative potencies towards the different CE isozymes differed by ~60-fold, comparable to
miltirone. For example, 1Bf demonstrated Ki values of 31.2nM and 1,871nM for hiCE
and hCE1, respectively with a comparable ratio being observed for 1Cf (Ki values of
4,204nM with hCE1 and 70.7nM for hiCE; Table 2). Interestingly, while the selectively
between the different enzymes was highly variable, it was not readily apparent which
parts of the molecules contributed to the specificity of inhibition. Overall however, even
in this small series of analogues, we have increased the potency of molecules that
contain the miltirone scaffold towards hCE1 by 200-fold (see inhibition data for 1Ad),
and to hiCE by ~3-fold (compound 1Bd).
Gratifyingly, none of the compounds demonstrated any appreciable inhibition of
hAChE, with an observed maximum of 30% by 1Cf at 5µM. Two molecules did inhibit
hBChE (54% and 62% for 1Aa and 1Ad, respectively at 5µM), however these studies
confirm that the synthesized analogs are not promiscuous inhibitors of human
esterases, and demonstrate considerable selectivity for CEs.
The compounds synthesized were chosen to evaluate three main properties
within the quinone portion: the relative steric bulk; the overall hydrophobicity of the
molecule; and the biological effect of modulating the electronegativity of groups
13

ACCEPTED MANUSCRIPT
appended adjacent to the 1,2-dione chemotype. Initially, we focused on the quinone
portion of the molecule, hypothesizing that altering the electrophilicity of the carbonyl
carbon atoms would allow more facile attack by the serine Oγ atom present within the
CE active site. Therefore analogues were generated containing both electron-
withdrawing and donating groups as substituents, expecting that the more electron-
deficient quinones would be more electrophilic, and thus more potent inhibitors. As
anticipated, the electron-withdrawing nature of the fluorine on analogues 1Ad, 1Ae, and
1Be results in compounds with significantly greater potency, particularly with regard to
hCE1 (Table 2). Indeed, these represent some of the most effective inhibitors of this
isozyme. However, with hiCE, the most potent compounds possess either larger alkyl or
fluoro substituents at the α-position relative to the dione (e.g., 1Bf, 1Cf and 1Bd). While
the former group can be considered to be electronically neutral, they add more steric
bulk to the molecules and as such may play a crucial role in the positioning of the
compounds within the active site. In addition, the alkyl substituents significantly increase
the logP of these molecules, however in this limited series of analogues, the inclusion of
these groups did not dramatically increase the biological activity of the 1,2-diones as
compared to miltirone.
Addition of fluorine, either in the α or β position with respect to the 1,2-dione
group for the less complex molecules, resulted in significant increases in inhibitory
potency for both human CEs. For example, 1Ad and 1Ae are 10- and 5-fold more
potent than the unsubstituted analogue (1Aa) with regard to hCE1 (Table 2). This is
likely due to an increase in electropositivity of the carbonyl carbon atom allowing for
more facile attack by the serine Oγ atom, the initial step in the interaction between the
enzyme and the small molecule inhibitors. However, this was not obviously apparent for
14

ACCEPTED MANUSCRIPT
the gem dimethyl substituted compounds, as exemplified by molecules 1Ba, 1Bd, 1Be
and 1Bg. Essentially, no large changes in the Ki values were observed by inclusion of
the fluorine atoms in these compounds when assayed against hCE1. Interestingly, the
α-fluoro-substituted analogue (1Bd) was ~23-fold more potent towards hiCE than 1Ba,
and the disubstituted molecule 1Bg also demonstrated increased potency (~8-fold).
These results argue, that for this enzyme, the group appended immediately adjacent to
the carbonyl carbon atom significantly impacts the activity of the compound. This was
also observed in the α-methyl-substituted derivative (1Bb) which is ~4-fold more potent
towards hiCE, but ~75-fold less active against hCE1 (Table 2). Our data indicate that
the substitutions in this region play a significant role both in inhibitor potency and
isozyme selectivity.
Similarly, effects of substituent on the A ring have been considered from three
perspectives: hydrophobicity; the steric bulk associated with this domain; and the
planarity of the overall molecule. Specifically, we compared the biological activity of
compounds containing a either a benzene or cyclohexyl ring, or those containing the
geminal-dimethyl groups at various positions within the latter. A direct comparison of the
effect of aromaticity was demonstrated by compounds 1Ad and 1Dd. The latter
compound (A ring = benzene) demonstrated reduced potency towards hCE1 (4-fold)
than the former molecule (A ring = cyclohexyl), but was ~2-fold more potent towards
hiCE (Table 2). Due to the similarity in the clogP values of these two small molecules, it
would appear that a change in conformation (1Dd would be more planar, whereas 1Ad
possesses more bulk) likely influences enzyme inhibition. Interestingly, an analogue
containing geminal-dimethyl groups within the cyclohexyl ring (1Bd), was considerably
more potent than the parent compound (1Ad) towards hiCE. It should be noted
15

ACCEPTED MANUSCRIPT
however, that 1Bd is likely to be much more hydrophobic (clogP values for 1Ad, 1Dd,
and 1Bd were calculated to be 2.85, 2.45 and 3.89, respectively using ChemBioDraw
Ultra) than 1Ad. Overall these results indicate that it is a combination of factors (steric
bulk, hydrophobicity, etc) that impact inhibitor potency with this class of molecules.
This was further demonstrated by comparing compounds 1Ab, 1Bb, and 1Cb. As
can be seen from Table 2, the addition of the geminal dimethyl groups does not
drastically impact the potency of these molecules toward hCE1 with 1Cb being the most
effective (although still in the mid nM range). Additionally, with hiCE, the 3-substituted
compound (1Bb) was more effective than the molecule that maintains the 2-substitution
(1Cb). This is potentially due to steric clashes within the hiCE active site, since there
was no statistically significant correlation observed between the hiCE Ki values and the
clogP of the molecules. In past studies with other 1,2-dione-containing structures, we
have observed good correlations between these parameters[39].
Molecular docking of miltirone analogues
Having generated a panel of miltirone-derived CE inhibitors that demonstrated biological
activity, we sought to determine whether computer docking studies might allow for a
determination of the potency of such molecules. Therefore, we used ICM Pro to dock
the respective compounds into the active site of hCE1 using the crystal structure
coordinates of this protein (1MX1). Following docking, the distances from the serine Oγ
atom to the carbonyl carbon atoms (the point at which this amino acid would attack the
1,2-diones) was determined (Table 3). As indicated in Table 4, there were weak linear
correlations between the potency of hCE1 inhibition and the docking of the molecules
within the active site when assessing either the interatomic distances, or the clogP
values of the small molecules (r2 values of ~ 0.5). However, when correlation tests were
16

ACCEPTED MANUSCRIPT
performed using these same datasets, highly significant Pearson r values were obtained
(P<0.001; Table 4). Unfortunately, using this information to design more potent
inhibitors is unlikely to be fruitful. However, these analyses do allow a gross analysis of
inhibitor binding and, as can be seen in Figure 1, 1Be docks in a different conformation
to miltirone. Since the difference in potency of these two molecules towards hCE1 is
~180-fold, these results argue that the more efficacious analogues localize within the
active site in an alternative manner to the natural product. Presently, the crystal
structure of hiCE has not been determined, and hence we have been unable to perform
similar studies with this enzyme.
Advantages of the developed methodology
The described Suzuki coupling approach should work well with most β-bromovinyl α-β
unsaturated aldehydes and aromatic boronic acids. Due to the robustness of the
reaction, the building-block strategy we have developed towards generating polycyclic
compounds, allows for the installation of diverse functionality within either portion of the
molecule. Additionally, we have demonstrated the ability to generate heterocyclic
precursors for incorporation into this class of compounds as exemplified by compound
1Eb (Table 1). However, it is likely that the elaboration of more complex functional
groups may require forethought before developing syntheses adhering to this route.
While the methods originally described by Harvey and co-workers used exclusively
aromatic building blocks[26], we have demonstrated that more saturated starting
materials can be coupled with considerable success. In doing so, we have
demonstrated a facile route towards diversely-substituted polycyclic aromatic
compounds, a scaffold common to many terpenoid natural products and derivatives.
Furthermore, due to the commercial availability, and ease of synthesis of boronic acids,
17

ACCEPTED MANUSCRIPT
this approach also requires minimal synthetic manipulations to yield analogues that can
be used to develop structure-activity relationships.
Other approaches to this class of compounds have used Diels-Alder annulations
or involved cascade cyclizations to generate polycyclic molecules.[15-17, 21] The
major drawbacks to these strategies is the laborious synthetic work required to yield the
key intermediates, typically requiring multiple synthetic steps. As a consequence, this
results in generally lower yields of the final products and reduced cost effectiveness.
Additionally, these routes tend to be linear and do not offer many opportunities for the
introduction of diversity without significant modification of the chemistry. The approach
described here minimizes the material requirements to generate each analogue, results
in high overall yields, and is suitable for small scale library generation.
Synthetic efforts by Snieckus and coworkers have employed a similar strategy by
which key bonds are formed using Suzuki couplings, followed by annulations via ortho-
metalation/condensation onto pendant amides.[40, 41] These approaches have been
employed towards the synthesis of fully aromatic phenanthrenes and β-lapachone.
However, there are functional group incompatibilities that exist in the former, and the
generation of the natural product requires at least 8 steps. Hence there are potential
limitations of these methods. Our route provides access to compounds containing a
saturated A-ring in a far more efficient manner.
It is also worth noting the regioselectivity of the annulation, as two regioisomers
are typically generated using this route (compounds 9 and 10 in Scheme 4). Scheme 6
illustrates how the selectivity between these isomers is dictated by competing steric and
electronic forces within the benzene portion of intermediates 3a and 3b. Since free
rotation can occur around the carbon-carbon bond between the benzene and
18

ACCEPTED MANUSCRIPT
cyclohexenyl rings, such movement juxtaposes either the Y or the OP groups adjacent
to the diene. It can be postulated that an electron-withdrawing group would reduce the
nucleophilicity of the benzene at the 2’ position allowing the 6’ position to act as the
intramolecular nucleophile. As a result, those compounds bearing a large or electron-
withdrawing group at the 3’ position would overwhelm the steric interactions imposed by
the protective group (P) on the 5’ oxygen, thereby producing the “undesired” isomer
(10). Thus, in these cases, larger protective groups were used to favor an equilibrium
that ensured the formation of the desired regioisomer (9). For example, with 3Ae and
3Be, where X and Y represent H and F, respectively (Schemes 1 and 4, Table 2), the
yield of the desired isomer was relatively low (~35:65, 2:8) when the 5’ oxygen was
protected as a methyl ether. However, when these cyclization reactions were
undertaken with a larger protective group (e.g., tert-butyldimethylsilyl ether), the
selectivity was inverted and improved to ~65:35 (2:8).’
Our primary future interests lie in generating members of the tanshinone
subclass of terpenoids, based upon our successful implementation of this approach,
specifically towards miltirone analogues. Given their known biological activity, we have
generated several analogues of these compounds. Furthermore, the wide range of
biological activities exhibited by similar terpenoids, including those isolated from Salvia
species, may make them attractive synthetic targets.[42-47] We anticipate that such
compounds could be readily synthesized, using in the described approaches, with
minimal modification. Consequently, most of our synthetic efforts are focused on
generating the Suzuki coupling precursors i.e., building in substitutions and functionality
prior to the key carbon-carbon bond forming step. Following this methodology, we can
develop more complex boronic acids with relative ease and apply them towards the
19

ACCEPTED MANUSCRIPT
generation of more complex polycyclic compounds. Given the relative ease with which
the necessary bromovinyl aldehydes are generated, a wide array of precursor cyclic
ketones can be used, provided they tolerate of the conditions used to install the β-
bromovinyl α-β unsaturated aldehyde functionality. Due to the flexibility and efficiency of
this approach, such studies are currently underway.
CONCLUSIONS
A facile synthetic route towards abietane analogues has been developed which can be
applied to incorporate a variety of functional groups. These syntheses are the shortest
routes to date of this class of molecule. Furthermore, the flexibility inherent in the
modular approach makes it ideally suited towards generating libraries of this class of
compounds. Using this synthetic approach, we have synthesized a growing library of
abietane analogues which have demonstrated varying degrees of inhibition of human
CEs, some of which are highly potent, with Ki values in single-digit nanomolar
concentrations. Furthermore, these compounds are not promiscuous inhibitors of
human esterases, indicated by their selectivity for CEs. As more molecules are
generated, a more complete structure-activity relationship will become apparent and will
guide the development of biological tools that can be used to assess CE activity and
function, both in vitro and in vivo.
20

ACCEPTED MANUSCRIPT
21

ACCEPTED MANUSCRIPT
Acknowledgements
This work was supported in part by an NIH grant AT007531, a Cancer Center Core
grant CA21765, and by ALSAC.
22

ACCEPTED MANUSCRIPT
REFERENCES
[1] B. Testa, J.M. Mayer, The Hydrolysis of Carboxylic Acid Esters, in: Hydrolysis in
Drug and Prodrug Metabolism, Verlag Helvetica Chimica Acta, Zurich, Switzerland,
2003.
[2] J. Cashman, B. Perroti, C. Berkman, J. Lin, Pharmacokinetics and molecular
detoxification, Environ. Health Perspect., 104 (1996) 23-40.
[3] M.R. Redinbo, P.M. Potter, Mammalian Carboxylesterases: From drug targets to
protein therapeutics., Drug Discov. Today, 10 (2005) 313-325.
[4] M.R. Brzezinski, B.J. Spink, R.A. Dean, C.E. Berkman, J.R. Cashman, W.F. Bosron,
Human liver carboxylesterase hCE-1: binding specificity for cocaine, heroin, and their
metabolites and analogs, Drug Metab. Dispos., 25 (1997) 1089-1096.
[5] M.J. Hatfield, L. Tsurkan, J.L. Hyatt, X. Yu, C.C. Edwards, L.D. Hicks, R.M. Wadkins,
P.M. Potter, Biochemical and molecular analysis of carboxylesterase-mediated
hydrolysis of cocaine and heroin, Br. J. Pharmacol., 160 (2010) 1916-1928.
[6] J. Zhang, J.C. Burnell, N. Dumaual, W.F. Bosron, Binding and hydrolysis of
meperidine by human liver carboxylesterase hCE-1, J. Pharmacol. Exp. Ther., 290
(1999) 314-318.
[7] D. Shi, J. Yang, D. Yang, E.L. LeCluyse, C. Black, L. You, F. Akhlaghi, B. Yan, Anti-
influenza prodrug oseltamivir is activated by carboxylesterase human carboxylesterase
1, and the activation is inhibited by antiplatelet agent clopidogrel, J. Pharmacol. Exp.
Ther., 319 (2006) 1477-1484.
[8] S. Whyard, R.J. Russell, V.K. Walker, Insecticide resistance and malathion
carboxylesterase in the sheep blowfly, Lucilia cuprina, Biochem. Genet., 32 (1994) 9-24.
23

ACCEPTED MANUSCRIPT
[9] P.J. O’Brien, Y.A. Abdel-Aal, J.A. Ottea, J.B. Graves, Relationship of insecticide
resistance to carboxylesterases In Aphis gossypii (Homoptera: Aphididae) from
Midsouth cotton, J. Econ. Entomol., 85 (1992) 651-657.
[10] R.M. Wadkins, J.L. Hyatt, X. Wei, K.J. Yoon, M. Wierdl, C.C. Edwards, C.L. Morton,
J.C. Obenauer, K. Damodaran, P. Beroza, M.K. Danks, P.M. Potter, Identification and
characterization of novel benzil (diphenylethane-1,2-dione) analogues as inhibitors of
mammalian carboxylesterases, J. Med. Chem., 48 (2005) 2905-2915.
[11] J.L. Hyatt, L. Tsurkan, M. Wierdl, C.C. Edwards, M.K. Danks, P.M. Potter,
Intracellular inhibition of carboxylesterases by benzil: Modulation of CPT-11 cytotoxicity,
Mol. Cancer Ther., 5 (2006) 2281-2288.
[12] M.J. Hatfield, L.G. Tsurkan, J.L. Hyatt, C.C. Edwards, A. Lemoff, C. Jeffries, B.
Yan, P.M. Potter, Modulation of Esterified Drug Metabolism by Tanshinones from Salvia
miltiorrhiza ("Danshen"), J. Nat. Prod., 76 (2013) 36-44.
[13] Y. Dong, S.L. Morris-Natschke, K.H. Lee, Biosynthesis, total syntheses, and
antitumor activity of tanshinones and their analogs as potential therapeutic agents, Nat.
Prod. Rep., 28 (2011) 529-542.
[14] M.A. Gonzalez, Aromatic abietane diterpenoids: total syntheses and synthetic
studies, Tetrahedron, 71 (2015) 1883-1908.
[15] J. Lee, Snyder, J. K., Ultrasound-promoted cycloadditions in the synthesis of Salvia
miltiorrhiza abietanoid o-quinones, J. Org. Chem., 55 (1990) 4995-5008.
[16] J. Lee, Mei, H. S., and Snyder, J. K., Synthesis of miltirone by an ultrasound-
promoted cyclization, J. Org. Chem., 55 (1990) 5013-5016.
[17] H.M. Chang, K.Y. Chui, F.W. Tan, Y. Yang, Z.P. Zhong, C.M. Lee, S.H. L., H.N.
Wong, Structure-activity relationship of miltirone, and active central benzodiazepine
24

ACCEPTED MANUSCRIPT
receptor ligand isolated from Salvia miltiorrhiza Bunge (Danshen), J. Med. Chem., 34
(1991) 1675-1692.
[18] D. Nasipuri, A.K. Mitra, Synthetic studies in the diterpene series. Part VIII.
Synthesis of miltirone, a diterpenoid quinone, J. Chem. Soc.: Perkin Trans. 1, (1973)
285-287.
[19] S. Knapp, Sharma, S., Synthesis of the naturally occuring antioxidant
Rosmariquinone, J. Org. Chem., 50 (1985) 4996-4998.
[20] J.G. Luis, L.S. Andres, W.Q. Fletcher, E.H. Lahlou, P. Aurea, Rearrangement of
methyl 11,12-di-O-methyl-6,7-didehydro-caronsate in basic medium. Easy
hemisynthesis of miltirone, J. Chem. Soc.: Perkin Trans. 1, m (1996) 2207-2211.
[21] G. Majetich, S. Liu, J. Fang, D. Siesel, Y. Zhang, Use of conjugated dienones in
cyclialkylations: Total syntheses of arucadiol, 1,2-didehydromiltirone, (+/-)-hinokione,
(+/-)-nimbdiol, sageone, and miltirone, J. Org. Chem., 62 (1997) 6928-6951.
[22] C. Hall, Cuppett, S., and Dussault, P., Synthesis and antioxidant activity of
Rosmariquinone and several analogues, J. Agricul. Food. Chem., 46 (1998) 1303-1310.
[23] A.K. Banerjee, J.A. Azocar, M.C. Sulbaran de Carrasco, M. Laya Mimo, Potential
intermediates for diterpenoid quinones: Cryptotanshinone, tanshinone IIA, and miltirone,
Synth. Comm., 31 (2001) 2471-2478.
[24] W. Huang, J. Li, W. Zhang, Y. Zhou, C. Xie, Y. Luo, Y. Li, J. Wang, J. Li, W. Lu,
Synthesis of miltirone analogues as inhibitors of Cdc25 phosphates, Bioorg. Med.
Chem. Lett., 16 (2006) 1905-1908.
[25] D. Xu, T.M. Penning, I.A. Blair, R.G. Harvey, Synthesis of phenol and quinone
metabolites of benzo[a]pyrene, a carcinogenic component of tobacco smoke implicated
in lung cancer, J. Org. Chem., 74 (2009) 597-604.
25

ACCEPTED MANUSCRIPT
[26] A. Wu, Y. Duan, D. Xu, T.M. Penning, R.G. Harvey, Regiospecific oxidation of
polycyclic aromatic phenols to quinones by hypervalent iodine reagents, Tetrahedron,
66 (2010) 2111-2118.
[27] A. Wu, D. Xu, D. Lu, T.M. Penning, I.A. Blair, R.G. Harvey, Synthesis of 13C4-
labelled oxidized metabolites of the carcinogenic polycyclic aromatic hydrocarbon
benzo[a]pyrene, Tetrahedron, 68 (2012) 7217-7233.
[28] S. Bencharit, C.L. Morton, Y. Xue, P.M. Potter, M.R. Redinbo, Structural basis of
heroin and cocaine metabolism by a promiscuous human drug-processing enzyme.,
Nat. Struct. Biol., 10 (2003) 349-356.
[29] R.M. Wadkins, J.L. Hyatt, K.J. Yoon, C.L. Morton, R.E. Lee, K. Damodaran, P.
Beroza, M.K. Danks, P.M. Potter, Discovery of novel selective inhibitors of human
intestinal carboxylesterase for the amelioration of irinotecan-induced diarrhea:
synthesis, quantitative structure-activity relationship analysis, and biological activity,
Mol. Pharmacol., 65 (2004) 1336-1343.
[30] H. Akaike, Information theory and an extension of the maximum likelihood principle,
in: B.N. Petrov, F. Csaki (Eds.) Second International Symposium on Information Theory,
Akademiai Kiado, Budapest, 1973, pp. 267-281.
[31] H. Akaike, A new look at the statistical model identification, IEEE Trans. Automatic
Control, AC-19 (1974) 716-723.
[32] G.L. Ellman, K.D. Courtney, V. Anders, R.M. Featherstone, A new and rapid
colorimetric determination of acetylcholinesterase activity, Biochem. Pharmacol., 7
(1961) 88-95.
[33] B.P. Doctor, L. Toker, E. Roth, I. Silman, Microtiter assay for acetylcholinesterase,
Anal. Biochem., 166 (1987) 399-403.
26

ACCEPTED MANUSCRIPT
[34] S. Bencharit, C.L. Morton, J.L. Hyatt, P. Kuhn, M.K. Danks, P.M. Potter, M.R.
Redinbo, Crystal structure of human carboxylesterase 1 complexed with the Alzheimer's
drug tacrine. From binding promiscuity to selective inhibition., Chem. & Biol., 10 (2003)
341-349.
[35] J.J. Lian, P.C. Chen, Y.P. Lin, H.C. Ting, R.S. Liu, Gold-catalyzed intramolecular
[3+2]-cycloaddition of arenyne-yne functionalities, J. Am. Chem. Soc., 128 (2006)
11372-11373.
[36] M.H. Norman, H.D. Smith, C.W. Andresw, F.L. Tang, C.L. Cowan, R.P. Steffen, 4-
(Heteroarythio)-2-biphenylyltetrazoles as Nonpeptide Angiotensin II Antagonists, J.
Med. Chem., 38 (1995) 4670-4678.
[37] J.V. Mello, N.S. Finney, Convenient synthesis and transformation of 2,6-dichloro-4-
iodopyridine, Org. Lett., 3 (2001) 4263-4265.
[38] W.T. Spencer, Frontier, A. J., Cycloaromatization protocol for synthesis of
polysubstituted phenol derivatives: Method development and mechanistic studies, J.
Org. Chem., 77 (2012) 7730-7736.
[39] J.L. Hyatt, T. Moak, J.M. Hatfield, L. Tsurkan, C.C. Edwards, M. Wierdl, M.K.
Danks, R.M. Wadkins, P.M. Potter, Selective inhibition of carboxylesterases by isatins,
indole-2,3-diones., J. Med. Chem., 50 (2007) 1876-1885.
[40] M.A.F. Brandao, A.B. de Oliveira, V. Snieckus, Combined directed metalation-cross
coupling strategies. A regiospecific route to heteroring-annelated ortho-
naphthoquinones and a short synthesis of b-lapachone, Tet. Lett., 34 (1993) 2437-2440.
[41] K.B. Jorgensen, Rantanen, T., Dorfler, T., and Snieckus, V., Directed metalation-
Suzuki-Miyaura cross-coupling strategies: regioselective synthesis of hydroxylated 1-
methylphenanthrenes, J. Org. Chem., 80 (2015) 9410-9424.
27

ACCEPTED MANUSCRIPT
[42] Y.B. Wu, Z.Y. Ni, Q.W. Shiu, M. Dong, H. Kiyota, Y.C. Gu, B. Cong, Constituents
from Salvia species and their biological activities, Chem. Rev., 112 (2012) 5967-6026.
[43] M.A. Gonzalez, Aromatic abietane diterpenoids: their biological activity and
synthesis, Nat. Prod. Rep., 32 (2015) 684-704.
[44] M.A. Gonzalez, Synthetic derivatives of aromatic abietane diterpenoids and their
biological activities, Eur. J. Med. Chem., 87 (2014) 834-842.
[45] A.K.S. Wan, S.W.S. Leung, D.-Y. Zhu, R.Y.K. Man, Vascular effects of different
lipophilic components of "Danshen", a traditional Chinese medicine, in the isolated
porcine coronary Artery, J. Nat. Prod., 71 (2008) 1825-1828.
[46] Y.M. Cui, E. Yasutomi, Y. Otani, T. Yoshinaga, K. Ido, T. Ohwada, Design,
synthesis, and characterization of podocarpate derivatives as openers of BK channels,
Bioorg. Med. Chem. Lett., 18 (2008) 5197-5200.
[47] Z. Dang, K. Jung, L. Zhu, H. Xie, K.H. Lee, C.H. Chen, L. Huang, Phenolic
diterpenoid derivatives as anti-influenza A virus agents, ACS Med. Chem. Lett., 6
(2015) 355-358.
28