A re(V)-catalyzed C-N bond-forming route to human lipoxygenase inhibitors

Article abstract of DOI:10.1021/ol050897a

(Chemical Equation Presented) A regioselective synthesis of propargylamines by the coupling of propargyl alcohols with tosylamines and carbamates catalyzed by an air-and moisture-tolerant rhenium-oxo complex is described. The ability to couple functionali

Full text of DOI:10.1021/ol050897a

ORGANIC
LETTERS
2005
Vol. 7, No. 12
2501-2504
A Re(V)-Catalyzed C−N Bond-Forming
Route to Human Lipoxygenase
Inhibitors
Rachana V. Ohri,† Alexander T. Radosevich,‡ K. James Hrovat,‡
Christine Musich,‡ David Huang,‡ Theodore R. Holman,*,† and F. Dean Toste*,‡
Department of Chemistry and Biochemistry, UniVersity of California-Santa Cruz,
Santa Cruz, California 956064, and Department of Chemistry, UniVersity of
California-Berkeley, Berkeley, California 94720
tholman@chemistry.ucsc.edu; fdtoste@uclink.berkeley.edu
Received April 22, 2005
ABSTRACT
A regioselective synthesis of propargylamines by the coupling of propargyl alcohols with tosylamines and carbamates catalyzed by an air-
and moisture-tolerant rhenium oxo complex is described. The ability to couple functionalized components allows for convergent approaches
−
to nitrogen-containing heterocyclic compounds such as the marine antibiotic pentabromopseudilin. These compounds were assayed against
human lipoxygenase and found to be both potent and selective.
The biosynthesis of linear eicosanoids such as leukotrienes
and lipoxins is initiated by the regio- and stereoselective
hydroperoxidation of arachidonic acid.1 This pivotal inflam-
mation response is performed by the lipoxygense family of
enzymes, which contain a catalytically active non-heme iron
cofactor.2 In addition to their native regulatory function, the
lipoxygenases have been implicated in a number of human
diseases such as asthma and cancer,3 making them important
targets for potential therapeutic inhibitors. Previously, we
have shown that marine-derived polybrominated phenols are
potent inhibitors against human lipoxygenases, which dem-
onstrate selectivity for human reticulocyte 15-lipoxygenase-1
(15-hLO) over human platelet 12-lipoxygenase (12-hLO).4
These initial results suggested that the degree of bromination
of these inhibitors might be related to their potency; however,
a modifiable chemical scaffold was required in order to
perform systematic structure/function analyses. With this goal
in mind, we developed a novel synthetic methodology
utilizing our established oxorhenium(V)-catalyzed propar-
gylation reaction5 to synthesize pentabromopseudilin6 (1), a
marine natural product that has structural similarities to our
previously discovered brominated phenol lipoxygenase in-
hibitors. In doing so, we hoped to extend the scope of our
† University of California, Santa Cruz.
‡ University of California, Berkeley.
(1) (a) Funk, C. D. Science 2001, 294, 1871. (b) Funk, C. D.; Chen, X.
S.; Johnson, E. N.; Zhao, L. Prost. Oth. Lipid M. 2002, 68-69, 303.
(2) Solomon, E. I.; Zhou, J.; Neese, F.; Pavel, E. G. Chem. Biol. 1997,
4, 795.
(3) (a) Poff, C. D.; Balazy, M. Curr. Drug Targets: Inflammation Allergy
2004, 3, 19. (b) Navab, M.; Ananthramaiah, G. M.; Reddy, S. T.; Van
Lenten, B. J.; Ansell, B. J.; Fonarow, G. C.; Vahabzadeh, K.; Hama, S.;
Hough, G.; Kamranpour, N.; Berliner, J. A.; Lusis, A. J.; Fogelman, A. M.
J. Lipid Res. 2004, 45, 993. (c) Hinz, B.; Brune, K. Curr. Opin. Rheumatol.
2004, 16, 628. (d) Xiong, H. Q. Cancer Chemother. Pharmacol. 2004, 54,
S69. (e) Steele, V. E.; Holmes, C. A.; Hawk, E. T.; Kopelvich, L.; Lubet,
R. A.; Crowell, J. A.; Sigman, C. C.; Kelloff, G. J. Cancer Epidemiol.
Biomarkers ReV. 1999, 8, 467.
(4) Segraves, E.; Shah, R.; Segraves, N.; Johnson, T.; Whitman, S.; Sui,
J.; Kenyon, V.; Cichewicz, R.; Crews, P.; Holman, T. R. J. Med. Chem.
2004, 47, 4060.
(5) (a) Sherry, B. D.; Radosevich, A. T.; Toste, F. D. J. Am. Chem. Soc.
2003, 125, 6076. (b) Luzung, M. R.; Toste, F. D. J. Am. Chem. Soc. 2003,
125, 15760. (c) Kennedy-Smith, J. J.; Young, L. A.; Toste, F. D. Org. Lett.
2004, 6 (8), 1325.
(6) (a) Burkholder, P. R.; Pfister, R. M.; Leitz, F. H. Appl. Microbiol.
1966, 14, 649. (b) Anderson, R. J.; Wolfe, M. S.; Faulkner, D. Mar. Biol.
1974, 27, 281. (c) Laatsch, H.; Pudleiner, H. Liebigs Ann. Chem. 1989,
863. (d) Xu, Z.; Lu, X. J. Org. Chem. 1998, 63, 5031.
10.1021/ol050897a CCC: $30.25
© 2005 American Chemical Society
Published on Web 05/12/2005

new synthetic methodology to meet the demands of the
current biological study.
or products derived from reaction of 4 with acetonitrile.12
Furthermore, under these conditions, the rearrangement of
the propargyl alcohol to an R,â-unsaturated ketone13 was
completely suppressed.
Encouraged by these results, we considered employing
other protected amines as nucleophiles in the reaction. In
particular, the diverse protocols for the cleavage of carbamate
protecting groups made them synthetically attractive for our
purposes; however, the use of carbamates as nucleophiles
typically requires the stoichiometric deprotonation of the
carbamate N-H.14 As a result, we were pleased to find that
a wide range of common carbamates participated effectively
in the (dppm)ReOCl3-catalyzed substitution without the need
for a stoichiometric base (Scheme 2).
The need to assess the effect of halogenation on the
potency of lipoxygenase inhibition precluded the synthesis
of pentabromopseudilin analogues by transition metal-
catalyzed biaryl coupling. In view of this, we envisioned the
preparation of the required analogues by the strategy outlined
in Scheme 1, in which diene 2 can be obtained from
Scheme 1
Scheme 2
propargylic amine 3. In turn, propargylic amines 3 could be
readily prepared from simple amines and aryl propargyl
alcohols7 using our rhenium-catalyzed propargylic substitu-
tion reaction.5 Avoiding the requirement for prior activation
of either the alcohol8 or the amine, we hoped the mild
conditions for the substitution reaction would allow for the
rapid synthesis of 1a, as well as analogues with diverse
substitution patterns.
With these preliminary results in hand, the substrate and
nucleophile scope of this reaction was investigated (Table
1). Substitution occurred with a range of propargylic alcohol
substrates, including phenyl (entries 1-5) and aryl rings
substituted with electron-withdrawing (entry 10) and electron-
donating (entries 11-13) groups. Additionally, both sterically
encumbered ortho disubstituted (entries 8 and 9) and het-
eroaromatic (entry 15) substrates cleanly participated in good
yield. Aryl-halide bonds are unaffected by the transforma-
tion (entries 6, 7, and 10). Notably, amides are unreactive
toward propargylation (entry 14), allowing for orthogonal
protection of nitrogen functionality in the propargylic adduct.
A variety of substituents on the alkynyl moiety are equally
well tolerated. Simple methyl and primary alkyl substituents
are exemplary, and although slightly increased reaction
temperatures were required, the more sterically demanding
phenyl (entry 4) and trimethylsilyl (entries 6, 7, and 12-
14) groups also perform well. Substrates containing a
terminal alkynyl unit also participate in the reaction, although
in somewhat diminished yields (compare entries 8 and 9).
Other potentially reactive groups, including terminal olefins
(entry 5) and conjugated esters (entry 11), prove tractable.
Beginning our investigation of this new catalytic reaction,
the addition of allylamine (5a) to propargylic alcohol 4 in
the presence of a rhenium(V)-oxo catalyst, (dppm)ReOCl3,9
and ammonium hexafluorophosphate in acetonitrile failed
to yield the desired propargylic amine (eq 2). Presumably,
competitive binding of the Lewis basic amine to the rhenium
center precluded the complexation of the propargylic alcohol
necessary for catalytic activity. Thus, we reasoned that
attenuation of the Lewis basicity of the nitrogen nucleophile
would allow for a more viable catalytic process. Gratifyingly,
reaction of p-nitroaniline (5b) and p-toluenesulfonamide (5c)
with 4 under the catalytic reaction conditions gave the desired
propargylic amine adducts in good yields and without
contamination from bispropargylated products.10 Addition-
ally, N-alkyl (5d) and N-aryl (5e) derivatives participated
equally well in the propargylic amination without the need
for stoichiometric activation of either the sulfonamide or the
alcohol.11 In all cases, the propargylic amines were produced
without contamination by regioisomeric allenic sulfonamides
(7) For a Ru-catalyzed amination of propargyl alcohols bearing a terminal
alkyne, see: (a) Nishibayashi, Y.; Wakiji, I.; Hidai, M. J. Am. Chem. Soc.
2000, 122, 11019. (b) Nishibayashi, Y.; Milton, M. D.; Inada, Y.;
Yoshikawa, M.; Wakiji, I.; Hidai, M.; Uemura, S. Chem. Eur. J. 2005, 11,
1433.
(8) Propargyl phosphonates and esters have also been employed as
substrates. Cu-catalyzed: (a) Imada, Y.; Yuasa, M.; Nakamura, I.; Mura-
hashi, S.-I. J. Org. Chem. 1994, 59, 2282. Ti-catalyzed (b) Mahrwald, R.;
Quint, S. Tetrahedron Lett. 2001, 42, 1655.
(9) (a) Chatt, J.; Rowe, G. A. J. Chem. Soc. 1962, 4019. (b) Rossi, R.;
Marchi, A.; Marvelli, L.; Magon, L.; Peruzzini, M.; Casellato, U.; Graziani,
R. Inorg. Chim. Acta 1993, 204, 63.
(10) Amination of benzylic alcohols catalyzed by methyltrioxorhenium-
(VII) (MTO) has been described: Zhu, Z.; Espenson, J. H. J. Org. Chem.
1996, 61, 324. However, the MTO-catalyzed reaction of 1 and N-allyl ethyl
carbamate in acetonitrile (12 h, 65 °C) resulted in rearrangement to the
R,â-unsaturated ketone.
(11) For coupling of sulfonamides and propargyl alcohols using the
Mitsunobu reaction, see: Bell, K. E.; Knight, D. W.; Gravestock, M. B.
Tetrahedron Lett. 1995, 36, 8681.
(12) Nilsson, B. M.; Hacksell, U. J. Heterocycl. Chem. 1989, 26, 269.
(13) Various metal-oxo complexes have been reported as catalysts for
conversion of propargyl alcohols to enones (Meyer-Schuster rearrange-
ment). (a) Re: Narasaka, K.; Kusama, H.; Hayashi, Y. Tetrahedron 1992,
48, 2059. (b) V: Erman, M. B.; Gulyi, S. E.; Aulchenko I. S. MendeleeV
Commun. 1994, 89. (c) Mo: Lorber, C. Y.; Osborn J. A. Tetrahedron Lett.
1996, 37, 853.
(14) (a) Boger, D. L.; Ishizaki, T.; Wysocki, R. J., Jr.; Munk, S. A.;
Kitos, P. A.; Suntornwat, O. J. Am. Chem. Soc. 1989, 111, 6461. (b) Le
Strat, F.; Maddaluno, J. Org. Lett. 2002, 4, 2791. (c) Osipov, S. N.; Burger,
K. Tetrahedron Lett. 2000, 41, 5659.
2502
Org. Lett., Vol. 7, No. 12, 2005

Scheme 3. Synthesis of Pentabromopseudilin
Table 1. Re(V)-Catalyzed Propargyl Substitution
catalyzed ring-closing metathesis15 to afford dihydropyrrole
9. Oxidative aromatization followed by bromination then
completed the synthesis of the natural product (1a). The
modular design of this synthetic scheme allows for facile
analogue synthesis by appropriate choice of the initial
substituted benzaldehyde. Compounds synthesized in this
way were then advanced to biological testing.
Pentabromopseudilin (1a) has a structure similar to that
of our previously discovered brominated phenol LO inhibi-
tors;4 however, its structure is more compact than our
previously investigated brominated phenol ethers. Our inhibi-
tor data of 1a indicated that it is a potent inhibitor to 12-
hLO (IC50 ) 13 ( 3.6 µM), 15-hLO (IC50 ) 3.5 ( 0.5 µM),
and 15-sLO (IC50 ) 8.5 ( 0.7 µM) (Table 2, entry 1), with
Table 2. SAR of Lipoxygenase Inhibitors
a Reaction conditions: 1 M propargyl alcohol in MeCN, 3.0 equiv of
b
amine nucleophile. Isolated yield after chromatography. c Obtained as a
mixture of diastereomers. d (dppm)ReOCl3 (10%) and 10% NH4PF6 were
employed.
With respect to the nucleophile, the rhenium(V)-oxo-
catalyzed propargylic amination is quite general. As previ-
ously demonstrated, the choice of carbamate protecting group
is largely unrestricted. Methyl and ethyl carbamates, as well
as the more commonly employed Alloc, Boc, and Cbz
protecting groups, all reacted effectively. Pendant olefin
(entry 15) and ester (entries 3 and 7) groups and function-
alized heterocycles (entry 12) are carried through the
substitution event. Carbamates with branching at the R-posi-
tion also participated, but the increased steric bulk leads to
reduced yields of the desired substitution product (entry 4).
Notably, Re-catalyzed propargylic substitution using benzyl
carbamate as an ammonia equivalent also proceeded smoothly
(entry 13).
Having successfully developed the Re(V)-catalyzed C-N
bond-forming reaction, we turned to the application of this
methodology in the synthesis of pentabromopseudilin 1 and
analogues. To begin the synthetic sequence (Scheme 3),
propargylic alcohol 7, available by nucleophilic acetylide
addition to o-benzyloxybenzaldehyde, was catalytically ami-
nated with N-allyl methyl carbamate to give adduct 8 in 99%
yield. Subsequent basic desilylation and Lindlar reduction
of the trimethylsilyl alkyne set the stage for ruthenium-
a Reduction of 15-hLO was screened with the observation of HPOD
degradation. b n/a ) samples we not screened for 15-hLO reduction due to
either large IC50 values, which make them undetectable, or alkylation of
the redox hydroxide.
a slight selectivity against 15-hLO inhibition. The mechanism
of action for 1a against 15-sLO is through reduction, as seen
by the fluorescence screen,16 presumably through the free
hydroxyl on the phenyl ring. In accord with this hypothesis,
Org. Lett., Vol. 7, No. 12, 2005
2503

methylation of the phenolic hydroxy (1b, entry 2) abolished
the reduction of the active-site iron in 15-sLO and dramati-
cally increased the IC50 (IC50 > 1000 µM). In contrast, the
IC50 of 1b did not increase against the human enzymes 12-
hLO (IC50 ) 5.7 ( 1.6 µM) and 15-hLO (IC50 ) 5.5 ( 0.5
µM), indicating that 1b is not a redox inhibitor against the
human enzymes and, by inference, neither is 1a. This result
was confirmed with our pseudoperoxidase assay with 15-
hLO, which clearly indicates that 1a and 1b are not redox
inhibitors of 15-hLO.
hLO through more general hydrophobic interactions and does
not require the precise positioning needed to ligate and reduce
the iron.
Alkylation of the pyrrole nitrogen (12a) abolished all
activity against 12-hLO (IC50 > 1000 µM) and 15-hLO (IC50
> 1000 mM), indicating that the pyrrole nitrogen is important
for inhibition activity. Interestingly, the addition of the bulky
hydrophobic benzyl group in 12b increased the inhibition
potency relative to 12a for both 12-hLO (IC50 ) 7.2 ( 1.6
µM) and 15-hLO (IC50 ) 3.9 ( 1 µM). A similar increase
in activity was noted on benzylation of 13a to give the
isomeric benzyl ether 13d, which showed increased activity
The unbrominated pyrrole 10a showed decreased inhibi-
tion against 12-hLO (IC50 ) 21 ( 2 µM) but a slightly
increased potency against 15-hLO (IC50 ) 1.6 ( 0.4 µM)
relative to 1a, indicating that the larger size of the brominated
pyrrole ring is more important for binding to 12-hLO than
15-hLO. To probe if the bromines affect the redox potential
of the phenol or the steric bulk of the inhibitor, the bromines
were replaced by methyl groups (10b). For 12-hLO, the IC50
value for 10b increased (IC50 ) 73 ( 25 µM); however, the
IC50 for 15-hLO decreased (IC50 ) 0.43 ( 0.1 µM) relative
to 10a (entry 4). Considering that both 10a and 10b display
reduction inhibition against 15-hLO, as seen by the pseudop-
eroxidase assay, and are of comparable size, this opposite
effect on 12-hLO and 15-hLO inhibition may be due to their
difference in hydrophobicity (cLogP). For 11a, the IC50 for
12-hLO (IC50 ) 52 ( 17 µM) increased, but for 15-hLO
the IC50 (IC50 ) 3.1 ( 0.8 µM) did not change relative to
1a (Table 1, entry 5). These data indicate that the degree of
bromination has little effect on 15-hLO inhibition; however,
for 12-hLO, bromination decreases the IC50 approximately
5-fold. When the hydroxyl group of the unbrominated 11a
was methylated to give 11b, the inhibition potency against
both 12-hLO (IC50 > 100 µM) and 15-hLO (IC50 > 1000
µM) was lost. This observation implicates 11a as a redox
inhibitor, while 11b is not, which was confirmed by the
pseudoperoxidase assay for 15-hLO. This is in contrast to
the methylation of pentabromopseudilin (1a), which did not
change its potency toward 15-hLO because it is not a redox
inhibitor against 15-hLO, as seen by the pseudoperoxidase
assay. A possible explanation for the difference in inhibition
mechanism between 1a (nonredox) and 11a (redox) against
15-hLO is that the larger steric bulk of the brominated 1a
could hinder the approach of the free hydroxyl moiety to
the active site iron, hence prohibiting a redox mechanism
but retaining its ability to bind to the active site; however,
11a is small enough to have its hydroxyl group bind directly
to the active-site iron, rendering it a redox inhibitor.
Interestingly, the inhibitor potency of 11a against 15-hLO
is stronger than that against 12-hLO, indicating that position-
ing of 11a in the active site of 15-hLO is more efficient for
the reduction of the iron than in 12-hLO. On the other hand,
1a presumably binds to the active site of both 12- and 15-
against 12-hLO (IC50 ) 10 ( 1 µM) and 15-hLO (IC50
)
7.2 ( 0.9 µM). This is an unexpected result considering that
methylation at this position has little effect on potency and
suggests that the large size of the benzyloxy is important,
possibly as a consequence of increased steric bulk and/or
hydrophobicity (cLogP). Changing the position of the
hydroxyl group (from ortho to meta) drastically lowered the
potency of 13a toward 12-hLO (IC50 > 1000 µM) and 15-
hLO (IC50 > 1000 µM) versus compound 11a. This result
substantiates our hypothesis that 11a has a specific binding
mode that allows it to bind and reduce the iron, while 13a
cannot access this binding mode due to the change in the
location of the hydroxy group.
In conclusion, we have developed an air- and moisture-
tolerant rhenium-catalyzed regioselective coupling of prop-
argyl alcohols with sulfonamides and carbamates. The scope,
mild reaction conditions, and experimental ease of this
transformation have made it a valuable method for construc-
tion of C-N bonds in the context of bioactive nitrogen-
containing heterocycles. This was demonstrated by applica-
tion of this method to the construction of pentabromopseudlin
and its analogues, which were examined as lipoxygenase
inhibitors. With respect to the inhibition studies, we have
shown that pentabromopseudilin (1a) is a novel and potent
inhibitor against both 12- and 15-human lipoxygenase.
Analogues of 1a indicate that the degree of bromination can
regulate the mode of inhibition (redox versus nonredox), with
the larger brominated compounds being nonredox inhibitors.
Finally, the best selectivity was observed against 15-hLO
by the nonbrominated redox inhibitors 10a, 10b, and 11a,
which inhibited 15-hLO 13-, 170-, and 17-fold more ef-
fectively than 12-hLO, respectively.
Acknowledgment. F.D.T. gratefully acknowledges the
UC Cancer Research Coordinating Committee, Merck Re-
search Laboratories, Amgen, Inc., DuPont, GlaxoSmithKline,
and Eli Lilly & Co. for financial support. T.R.H. gratefully
acknowledges the NIH (GM56062) and the American Cancer
Society (RPG-00-219-01-CDD) for financial support.
Supporting Information Available: Experimental pro-
cedures and compound characterization data. This material
is available free of charge via the Internet at http://pubs.acs.org.
(15) Fu, G. C.; Grubbs, R. H. J. Am. Chem. Soc. 1992, 114, 7324.
(16) Agro, A. F.; Avigliano, L.; Egmond, M. R.; Veldink, G. A.;
Vliegenthart,.J. F. G. FEBS Lett. 1975, 52, 73.
OL050897A
2504
Org. Lett., Vol. 7, No. 12, 2005