Stereoselective Synthesis and Evaluation of C6-Substituted 5a-Carbasugar Analogues of SL0101 as Inhibitors of RSK1/2

Article abstract of DOI:10.1021/acs.orglett.7b00945

A convergent synthesis of 5a-carbasugar analogues of the n-Pr-variant of SL0101 is described. The analogues were synthesized in an effort to find compounds with potent in vivo efficacy in the inhibition of p90 ribosomal s6 kinase (RSK1/2). The synthesis derived the desired C-4 L-rhamnose stereochemistry from quinic acid and used a highly selective cuprate addition, NaBH₄ reduction, Mitsunobu inversion, and alkene dihydroxylation to install the remaining stereochemistry. A Pd-catalyzed cyclitolization stereoselectively installed the aglycon at the anomeric position. The analogues were evaluated as RSK1/2 inhibitors and found to have 3- to 6-fold improved activity.

Full text of DOI:10.1021/acs.orglett.7b00945

Letter
pubs.acs.org/OrgLett
Stereoselective Synthesis and Evaluation of C6″-Substituted
5a‑Carbasugar Analogues of SL0101 as Inhibitors of RSK1/2
Mingzong Li,‡,∥ Yu Li,‡,∥ Katarzyna A. Ludwik,†,∥ Zachary M. Sandusky,‡,∥ Deborah A. Lannigan,*,‡,†
and George A. O’Doherty*,§
†Departments of Pathology, Microbiology & Immunology, Nashville, Tennessee 37232, United States
‡Cancer Biology, Vanderbilt University School of Medicine, Nashville, Tennessee 37232, United States
§Department of Chemistry and Chemical Biology, Northeastern University, Boston, Massachusetts 02115, United States
S
* Supporting Information
ABSTRACT: A convergent synthesis of 5a-carbasugar analogues of the n-
Pr-variant of SL0101 is described. The analogues were synthesized in an
effort to find compounds with potent in vivo efficacy in the inhibition of
p90 ribosomal s6 kinase (RSK1/2). The synthesis derived the desired C-4
L-rhamnose stereochemistry from quinic acid and used a highly selective
cuprate addition, NaBH4 reduction, Mitsunobu inversion, and alkene
dihydroxylation to install the remaining stereochemistry. A Pd-catalyzed
cyclitolization stereoselectively installed the aglycon at the anomeric
position. The analogues were evaluated as RSK1/2 inhibitors and found to have 3- to 6-fold improved activity.
he p90 ribosomal s6 kinases (RSK) are a family of Ser/Thr
protein kinases.1 Two isoforms (RSK1/2) from this family
Scheme 1. SAR for SL0101 RSK Inhibition8
T
are involved in the etiology of a number of different cancers.2 In
an effort aimed at identifying RSK1/2 inhibitors, the flavonoid
glycoside natural product SL0101 (1) was discovered as a
relatively selective inhibitor of the N-terminal kinase domain
(NTKD) of RSK.3 RSK has two kinase domains where the N-
terminal domain (NTKD) is responsible for phosphorylation of
target substrates.4 Based on the crystal structure of the RSK2
NTKD complexed with SL0101, a major conformational
rearrangement of the N-lobe of the kinase domain generates
the inhibitor-binding pocket.5
Inspired by its unique activity and selectivity, we have been
exploring structure−activity relationship (SAR) requirements
associated with SL0101 (1).3,6 As part of these studies, we have
developed a de novo asymmetric synthesis7 of SL0101, its
enantiomer, and several congeners.8 Our studies have identified
several analogues with improved activity and have emphasized
the importance of the rhamnose sugar and its C-3 and C-4
acetates. In addition, we have found that substitution at the C-6
position8,9 and the ring oxygen (2 and 3) of the sugar gives
improved efficacy in the in vitro kinase assays and cell-based
studies.8,10
SL0101 (1) has a short biological half-life in vivo,5 which is
presumably due to the hydrolyzable C-3/C-4-acetates on the
sugar, as well as an O-glycosidic bond. To identify less labile
groups that could replace the ester without loss of affinity, we
have investigated replacing the rhamnose C-4-acetate (e.g., 5a−f
with a C-4 acetamide), the C-3/C-4-acetates (e.g., 4h with a C-3/
C-4-n-Pr-carbamates), and the ring oxygen with a methylene
group (i.e., 3a−d carbasugars).8 As part of our ongoing effort to
identify RSK1/2-inhibitors as potential therapeutics, we decided
to test the effects of combining these three substitutions and
targeted two analogues 2a−b (Scheme 1). Herein we disclose the
synthesis of cyclitol analogues 2a and 2b as well as the relative
RSK2 inhibitory activity.11
Received: March 29, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.7b00945
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Retrosynthetically, we envisioned preparing 2a and 2b in a
route analogous to our previously established routes to SL0101
analogues 1−5 (e.g., 6 + 7, Scheme 2). Specifically, we expected
Our redesigned retrosynthesis turned the original design 180°
(Scheme 4), with the quinic acid tertiary alcohol becoming the
Scheme 4. Synthesis of Pd-Cyclitolization Donor
Scheme 2. Retrosynthesis of n-Pr Carbasugar Analogues
carbasugar C-1 anomeric position and the central alcohol of the
triol becoming the C-4 enone 8. Because of complications with
its conversion to enone 8 this effort turned to the synthesis of
allylic acetate 17. Specifically, the allylic benzoate 17 already had
a Pd-π-allyl leaving group at the pseudoanomeric position for the
cyclitolization reaction. In addition, benzoate 17 also had the
desired carbasugar C-4 acetate in the correct rhamno-stereo-
chemistry. The question that remained was could we find
conditions to selectively ionize the axial C-1 allylic p-NO2Bz
group without touching the allylic C-4 acetate. In addition, we
were concerned that the C-4 acetate may not control the
regiochemistry of nucleophilic attack to the π-allyl intermediate
(i.e., TS-1) as well as the C-4 ketone (i.e., TS-2).15 In general, we
found a C-4 ketone both improved the electrophilicity of the Pd-
π-allyl intermediate and helped direct nucleophilic addition to
the C-1 position. In this regard, we were hopeful that the
regiocontrol issues could also be controlled by the C-5 n-propyl
substituent.16 Altenatively, chiral ligands on the Pd-π-allyl could
be used if the C-4 acetate is not sufficient for controlling the
regiochemistry.
that 2a/b would arise from the Pd-catalyzed cyclitolization of 7
with enone 8.8,12 Base on our previous success, we viewed the
cyclitol donor 8 would arise from quinic acid 10 via enone 9,13
where the quinic acid tertiary alcohol would become the C-4
ketone and the central alcohol of the triol would become the C-1
anomeric position.
At the outset, our efforts to extend our previous cyclitol
synthesis (R = H) encountered difficulty, associated primarily
with achieving high stereocontrol in the conversion of 9 into 12b.
Specifically, we explored the possibility of installing the sugar L-
stereochemistry at C-5 by means of a selective hydrogenation of
the enone 11 from the exoface to selectively provide 12b over
12a. To our surprise, when we exposed 11 to typical
hydrogenation conditions (1 atm of H2 with Pd/C), we found
12a was the major isomer, but with poor diastereoselectivity
(3:1) (Scheme 3). This problem was exacerbated by the fact that
isomerization of the position α to the ketone occurred when we
tried to deprotect the acetonide.12 These two compounding
factors prompted a search for an alternative approach.14
Our redesigned synthesis returned to enone 9, which
underwent a highly stereoselective cuprate-promoted addition
of n-propyl anion to furnish ketone 13 in an excellent yield and as
a single diastereomer (89%).17 Exposure of 13 to DBU in
benzene gave allylic alcohol 14 in good yield (80%). The
required C-4 acetate was then installed by an acylation of 14 with
acetic anhydride/DMAP with Et3N to deliver allylic acetate 15
(74%). A stereoselective 1,2-reduction of enone 15 under Luche
conditions18 gave the allylic alcohol 16 in good yield (89%) and
diastereoselectivity (>10:1). Analysis of the allylic coupling
Scheme 3. Initial Attempt to α-L-Cyclitol Donors
1
constants in the H NMR for 16 indicated that the Luche
reduction occurred via axial attack to install an equatorial
alcohol,19 which corresponds to β-anomeric stereochemistry in
the resulting carbasugar. Thus, a Mitsunobu reaction was
preformed on 16 with DIAD/PPh3 and p-nitrobenzoic acid to
yield the desired cyclitol donor 17 in excellent yield (90%).20
Unfortunately, the conversion of the bis-allylic ester 17 to
enone 8 proved to be untenable, as conditions for the selective
hydrolysis of either of the two esters were not found. In an effort
to find a viable alternative, we decided to explore the use of allylic
p-NO2-benzoate 17 as a cyclitol donor, in the cyclitolization
reactions. The allylic p-nitrobenzoate 17 had some potential
B
DOI: 10.1021/acs.orglett.7b00945
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
advantages to 8 in our planned synthesis of SL0101 analogue 2a.
Specifically, the C-1 allylic-benzoate 17 already had a Pd-π-allyl
leaving group at the anomeric position for the cyclitolization
reaction. In addition, the p-nitrobenzoate 17 also had the desired
C-4 acetate with the correct stereochemistry installed, and thus,
the aglycon portion would go through two less transformations
(Figure 1).
Scheme 6. Synthesis of Bis-carbamate Analogues
Figure 1. Competing Pd-π-allyl mechanisms.
To our delight, when a mixture of p-nitrobenzoate 17 and
flavonol 7 was exposed to our typical glycosylation conditions
(2.5 mol % Pd2(dba)3·CHCl3 and 10 mol % of PPh3 in CH2Cl2 at
0 °C) the reaction proceeded smoothly to provide the desired
product in good yield (80%), in excellent regio- and stereo-
selectivity (Scheme 5). Exposure of the allylic acetate 18 to the
Et3N in MeCN). Exposure of the allylic carbamate 23 to the
Upjohn conditions (OsO4/NMO; 73%) stereoselectively
converted it into the rhamno-diol 24. Re-exposure of the C-2/
C-3 diol 24 to the carbamate forming conditions (n-Pr-
isocyanate, 10% DBU/Et3N in MeCN) gave a mixture of C-3
and C-2 carbamates 20 and 21 (1:1). The desired C-3/C-4
biscarbamate 25 could be isolated from that mixture in a 40%
overall yield. Unfortunately, the undesired C-2/C-4 biscarbamate
26 could not be isomerized under basic conditions as was the
case with the diacetate. Finally, the desired C-3/C-4 regioisomer
25 was globally deprotected by an exhaustive hydrogenolysis (1
atm of hydrogen with Pd/C), to afford the target C-3/C-4
biscarbamate 2b in good overall yield (72%).
Scheme 5. Synthesis of n-Pr Carbasugar SL0101 Analogue
The efficacy of the two cyclitol analogues 2a and 2b to inhibit
RSK2 activity was determined in in vitro kinase assays using
purified recombinant RSK2 (Table 1).3 The data were fit using
nonlinear regression analysis. Both SL0101 analogues showed
improved RSK inhibition over the lead structure SL0101, the n-
Pr cyclitol diacetate 2a having a >6-fold decrease in IC50 (54
nM), whereas the bis-carbamate 2b had an ∼3-fold decrease in
a
Table 1. In Vitro Potency of SL0101 (1) and Analogues
Upjohn conditions21 (OsO4/NMO; 70%) stereoselectively
converted it into the rhamno-diol 19 which, when acylated
with acetyl chloride and Hunig’s base, gave a mixture of the C-3
and C-2 acetates 20 and 21 (1:1.6) in 38% yield. Fortunately, the
undesired C-2 acetate 21 could be isomerized to a mixture of
acetates which favored the desired C-3 acetate 20 (3:1). Finally,
the desired regioisomer 20 was globally deprotected by an
exhaustive hydrogenolysis (1 atm of hydrogen with Pd/C),
which produced the target C-3/C-4 diacetate 2a in good yield
(70%).
With access to the desired diacetate 2a having been
established, we turned our attention to the second target
compound biscarbamate 2b (Scheme 6). Our approach to
biscarbamate 2b returned us to allylic acetate 18, which could be
hydrolyzed into allylic alcohol 22 with K2CO3 in methanol
(84%). The allylic alcohol 22 was converted into n-Pr-carbamate
23 by exposure of it to n-Pr-isocyanate and base (10% DBU/
a
b
For procedure see Supporting Information. See ref 11.
C
DOI: 10.1021/acs.orglett.7b00945
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(3) Smith, J. A.; Poteet-Smith, C. E.; Xu, Y.; Errington, T. M.; Hecht, S.
M.; Lannigan, D. A. Cancer Res. 2005, 65, 1027.
(4) Anjum, R.; Blenis, J. Nat. Rev. Mol. Cell Biol. 2008, 9, 747.
(5) Utepbergenov, D.; Derewenda, U.; Olekhnovich, N.; Szukalska, G.;
Banerjee, B.; Hilinski, M. K.; Lannigan, D. A.; Stukenberg, P. T.;
Derewenda, Z. S. Biochemistry 2012, 51, 6499.
(6) (a) Smith, J. A.; Maloney, D. J.; Clark, D. E.; Xu, Y.; Hecht, S. M.;
Lannigan, D. A. Bioorg. Med. Chem. 2006, 14, 6034. (b) Smith, J. A.;
Maloney, D. J.; Hecht, S. M.; Lannigan, D. A. Bioorg. Med. Chem. 2007,
15 (14), 5018.
(7) (a) Iyer, A.; Zhou, M.; Azad, N.; Elbaz, H.; Wang, L.; Rogalsky, D.
K.; Rojanasakul, Y.; O’Doherty, G. A.; Langenhan, J. M. ACS Med. Chem.
Lett. 2010, 1, 326. (b) Shi, P.; Silva, M.; Wu, B.; Wang, H. Y. L.;
Akhmedov, N. G.; Li, M.; Beuning, P.; O’Doherty, G. A. ACS Med.
Chem. Lett. 2012, 3, 1086. (c) Wu, B.; Li, M.; O’Doherty, G. A. Org. Lett.
2010, 12, 5466.
(8) (a) Li, M.; Li, Y.; Mrozowski, R. M.; Sandusky, Z. M.; Shan, M.;
Song, X.; Wu, B.; Zhang, Q.; Lannigan, D. A.; O’Doherty, G. A. ACS
Med. Chem. Lett. 2015, 6, 95. (b) Mrozowski, R. M.; Sandusky, Z. M.;
Vemula, R.; Wu, B.; Zhang, Q.; Lannigan, D. A.; O’Doherty, G. A. Org.
Lett. 2014, 16, 5996. (c) Mrozowski, R. M.; Vemula, R.; Wu, B.; Zhang,
Q.; Schroeder, B. R.; Hilinski, M. K.; Clark, D. E.; Hecht, S. M.;
O’Doherty, G. A.; Lannigan, D. A. ACS Med. Chem. Lett. 2013, 4, 175.
(d) Shan, M.; O’Doherty, G. A. Org. Lett. 2010, 12, 2986.
(9) Wang, H. Y. L.; Xin, W.; Zhou, M.; Stueckle, T. A.; Rojanasakul, Y.;
O’Doherty, G. A. ACS Med. Chem. Lett. 2011, 2, 73.
IC50 (137 nM). In the parent series, cyclitol substitution (i.e., 1 to
3a) led to a modest improvement in the inhibitory activity (345
to 270 nM). In the n-Pr-series, however, the effect of cyclitol
substitution (i.e., 4b to 2a) trended toward reduced efficacy (20
to 56 nM). This loss in activity should be easily compensated for
by the expected improved bioavailability that results from the
hydrolysis-resistant cyclitol substitution. There was also a loss in
inhibitory activity (345 to 870 nM) by the n-Pr-carbamate
substitution in the parent series (1 to 4h). This negative effect
was also observed with the cyclitol series (i.e., 2a to 2b: 56 to 137
nM).
In conclusion, using a highly stereoselective Pd-cyclitolization
reaction, two new cyclitol analogues of the natural product
SL0101 were synthesized and evaluated. These analogues 2a and
2b showed significant improvement in RSK inhibitory activity.
Improved bioavailability has already been shown for 2a.11 The
synthesis of these new SL0101 cyclitol analogues required the
discovery of a novel synthesis of a new cyclitol donor, which
demonstrated that the previously believed requirement of a C-4
ketone in the cyclitol donor is not necessary for the reaction with
phenol-like nucleophiles. Further studies aimed at defining the
requirements for a specific-RSK1/2 inhibition are ongoing and
will be reported in due course.
(10) Wang, H. Y.; Wu, B.; Zhang, Q.; Kang, S. W.; Rojanasakul, Y.;
O’Doherty, G. A. ACS Med. Chem. Lett. 2011, 2, 259.
(11) For detailed in vivo studies of 2a, see: Ludwik, K. A.; Campbell, J.
P.; Li, M.; Li, Y.; Sandusky, Z. M.; Pasic, L.; Sowder, M. E.; Brenin, D. R.;
Pietenpol, J. A.; O’Doherty, G. A.; Lannigan, D. A. Mol. Cancer Ther.
2016, 15, 2598.
(12) Shan, M.; O’Doherty, G. A. Synthesis 2008, 2008, 3171.
(13) (a) Federspiel, M.; Fischer, R.; Hennig, M.; Mair, H. J.;
Oberhauser, T.; Rimmler, G.; Albiez, T.; Bruhin, J.; Estermann, H.;
Gandert, C.; Gockel, V.; Gotzo, S.; Hoffmann, U.; Huber, G.; Janatsch,
G.; Lauper, S.; Rockel-Stabler, O.; Trussardi, R.; Zwahlen, A. G. Org.
Process Res. Dev. 1999, 3, 266. (b) Trost, B. M.; Romero, A. G. J. Org.
Chem. 1986, 51, 2332. (c) Audia, J. E.; Boisvert, L.; Patten, A. D.;
Villalobos, A.; Danishefsky, S. J. J. Org. Chem. 1989, 54, 3738.
(14) (a) Guo, H.; O’Doherty, G. A. Tetrahedron 2008, 64, 304.
(b) Abrams, J. N.; Babu, R. S.; Guo, H.; Le, D.; Le, J.; Osbourn, J. M.;
O’Doherty, G. A. J. Org. Chem. 2008, 73, 1935. (c) Coral, J. A.; Guo, H.;
Shan, M.; O’Doherty, G. A. Heterocycles 2009, 79, 521. (d) Borisova, S.
A.; Guppi, S. R.; Kim, H. J.; Wu, B.; Penn, J. H.; Liu, H. W.; O’Doherty,
G. A. Org. Lett. 2010, 12, 5150.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.orglett.7b00945.
■
S
Experimental procedures and spectral data (PDF)
AUTHOR INFORMATION
Corresponding Authors
■
*E-mail: deborah.lannigan@vanderbilt.edu (D.A.L.)
*E-mail: G.ODoherty@neu.edu (G.A.O.)
ORCID
George A. O’Doherty: 0000-0002-1699-2249
Author Contributions
∥M.L., Y.L., K.A.L., and Z.M.S. co-first authors; the order is
alphabetical.
(15) Bajaj, S. O.; Sharif, E. U.; Akhmedov, N. G.; O’Doherty, G. A.
Chem. Sci. 2014, 5, 2230.
Notes
The authors declare the following competing financial
interest(s): The coresponding authors Deborah A. Lannigan
and George A. O'Doherty have applied for patent protection for
this class of compounds.
(16) Previously, we have found that 1,4-regiocontrol can be addressed
with chiral ligands; see ref 8b.
(17) Lipshutz, B. H.; Ellsworth, E. L. Tetrahedron Lett. 1988, 29, 893.
(18) Luche, J. L. J. Am. Chem. Soc. 1978, 100, 2226.
(19) Eliel, E. L.; Senda, Y. Tetrahedron 1970, 26, 2411.
(20) Mitsunobu, O. Synthesis 1981, 1981, 1.
(21) VanRheenen, V.; Kelly, R. C.; Cha, D. Y. Tetrahedron Lett. 1976,
17, 1973.
ACKNOWLEDGMENTS
■
This work was supported by the Susan G. Komen
(#IIR12223770 to D.A.L.) and the NSF (CHE-1565788 to
G.A.O.).
REFERENCES
■
(1) Eisinger-Mathason, T. S.; Andrade, J.; Lannigan, D. A. Steroids
2010, 75, 191.
(2) (a) Lara, R.; Mauri, F. A.; Taylor, H.; Derua, R.; Shia, A.; Gray, C.;
Nicols, A.; Shiner, R. J.; Schofield, E.; Bates, P. A.; Waelkens, E.;
Dallman, M.; Lamb, J.; Zicha, D.; Downward, J.; Seckl, M. J.; Pardo, O.
E. Oncogene 2011, 30, 3513. (b) Doehn, U.; Hauge, C.; Frank, S. R.;
Jensen, C. J.; Duda, K.; Nielsen, J. V.; Cohen, M. S.; Johansen, J. V.;
Winther, B. R.; Lund, L. R.; Winther, O.; Taunton, J.; Hansen, S. H.;
Frodin, M. Mol. Cell 2009, 35, 511.
D
DOI: 10.1021/acs.orglett.7b00945
Org. Lett. XXXX, XXX, XXX−XXX