Total synthesis of celogentin C by stereoselective C-H activation

Article abstract of DOI:10.1002/anie.200905134

(Figure Presented) A total gent: Inspired by the biosynthesis of celogentin C, a highly stereoselective and efficient palladium-catalyzed C-H functionalization strategy is employed to construct the key Leu-Trp linkage of this bicyclic compound. A streamlined synthesis is completed in 23 steps from simple amino acid building blocks.

Full text of DOI:10.1002/anie.200905134

Communications
DOI: 10.1002/anie.200905134
ꢀ
C H Activation
ꢀ
Total Synthesis of Celogentin C by Stereoselective C H Activation**
Yiqing Feng and Gong Chen*
Dedicated to Professor Samuel J. Danishefsky
Celogentin C (1) is a bicyclic nonribosomal peptide that was
isolated from the seeds of Celosia argentea (Figure 1).[1] It is
hydrogenation conditions to dehydroamino acid precursors.
Enantio- and/or diastereoselectivities ranged from 1:1 to 16:1
for these 10–14 step sequences. More recently, Castle and co-
workers developed a novel Knoevenagel condensation/radi-
cal conjugate addition approach to the Leu-Trp linkage. They
completed the first celogentin synthesis through an elegant
ꢀ
NCS-mediated Trp-His C N coupling by utilizing the major
diastereomer product, albeit in modest enantio- and diaste-
reoselectivity.[4] Herein, we report a highly stereoselective and
efficient synthesis of celogentin C using a novel palladium-
ꢀ
catalyzed C H functionalization strategy.
The highly constrained structure of 1 is probably assem-
bled in vivo from the much simpler linear peptide precursor
through
a
series of enzymatic oxidative cross-links
(Figure 1).[5] Inspired by these simple yet powerful trans-
formations found in nature, we envisioned developing a
synthetic equivalent of these processes in a direct approach to
celogentin. Our synthetic strategy relied on the direct regio-
ꢀ
and stereoselective activation of the b C H bond of a Leu
ꢀ
moiety and on the subsequent coupling of the derived C Pd
species with a suitable Trp partner.[6] The recent report by
Corey et al.[7] of the carboxamide-directed b C H function-
ꢀ
alization of amino acids served as the starting point for our
ꢀ
venture. Corey demonstrated that the b C H bond of the N-
Figure 1. Celogentin C and our synthetic strategy.
phthaloyl amino acid 8-aminoquinoline amide can be effi-
ciently activated and then arylated with simple aryl iodides
under Pd(OAc)2 catalysis, a procedure built on the seminal
discovery of Daugulis and co-workers for the functionaliza-
the most potent isolate (IC50 = 0.8 mm; IC = inhibitory con-
centration) from the celogentin/moroidin family, whose
members possess inhibitory activity against tubulin polymer-
ization. Its highly unusual architecture, which is characterized
by the direct linkages of Trp C6 to Leu Cb, and Trp C2 to His
N1 (Figure 1), and its biological activity have prompted a
number of synthesis studies.[2] Although N-linked His residues
are known to occur in other macrocyclic peptides, the Leu-Trp
linkage is extremely rare and poses a difficult synthetic
challenge.[3] To access the key Leu-Trp motif, Moody and
Bentley,[2c] and Campagne et al.,[2f] applied asymmetric
3
[8]
ꢀ
tion of inactivated sp C H bonds. The quinoline moiety
serves as a chelating auxiliary for palladium coordination, and
promotes the formation of trans-palladacycle intermediate 4.
This palladium(II) intermediate presumably undergoes cross-
coupling with an aryl iodide partner to provide the final
arylated product which has an erythro stereochemical prefer-
ence. To our delight, we were able to achieve the high-yielding
and highly stereoselective 6-indolylation of N-phthaloyl
leucine (Scheme 1a). Upon simple heating of precursors 2
(2.0 equiv) and 3 (1.0 equiv), with Pd(OAc)2 (0.2 equiv), and
AgOAc (1.5 equiv), at 1108C in tBuOH, the desired diaste-
reomer 5 was formed exclusively, and the slight excess of 2
could be largely recovered. About 3% of deiodinated side
product 6 was also generated.
Although the quinoline carboxamide serves as an effec-
tive auxiliary in arylation chemistry, its efficient removal
under mild conditions would be required for this process to
become a useful tool for natural product synthesis. However,
the cleavage of the amide linkage was particularly problem-
atic, owing to both steric hindrance and the lability of the N-
phthaloyl group.[9] This phthaloyl group, which provides both
bis-protection of the a-amino group and steric bias, is critical
[*] Y. Feng, Prof. G. Chen
Department of Chemistry, The Pennsylvania State University
104 Chemistry Building, University Park, PA 16802 (USA)
Fax: (+1)814-863-5319
E-mail: guc11@psu.edu
[**] We thank The Pennsylvania State University for support of this work.
We thank Prof. Ray Funk, Prof. Ken Feldman, and their groups for
helpful discussions and their generous sharing of chemicals, Dr.
Alan Benesi for NMR analysis, and Tyler Grove for HPLC separation.
We also thank Prof. Steven Castle from Brigham Young University
for providing synthetic celogentin C for HPLC comparison.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200905134.
958
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 958 –961

Angewandte
Chemie
Scheme 2. Synthesis of the key Leu-Trp intermediate. Reagents and
conditions: a) tBuBr, TEBAC, DMA, 558C, 89%; HNO3 (1.7 equiv),
HOAc (5.0 equiv), CH2Cl2, RT; NaH, TsCl, DMF, 08C to RT; 24% yield
over 2 steps; b) H2, 10% Pd/C, MeOH, RT, 78% yield; NOBF4, KI, I2,
CH3CN, ꢀ408C, 62% yield; c) 14 (1.0 equiv), 2 (2.0 equiv), Pd(OAc)2
(0.2 equiv), AgOAc (1.5 equiv), tBuOH, 1108C, 36 h, 85% yield;
d) Ethylenediamine, nBuOH, RT; TfN3, NEt3, CH2Cl2, RT; 82% yield
over 2 steps; e) Boc2O (15 equiv), DMAP (3.0 equiv), CH3CN, 708C,
89% yield; f) LiOH (1.2 equiv), H2O2 (5.0 equiv), THF/H2O, 08C to RT,
quantitative yield. tBuBr=tert-butyl bromide, TEBAC=triethyl benzyl
ammonium chloride, DMA=N,N’-dimethylacetamide,.
ꢀ
Scheme 1. C H activation–indolylation and auxiliary removal. Reagents
and conditions: a) Ethylenediamine, nBuOH, 908C; TfN3, NEt3,
CH2Cl2, RT; 81% yield over 2 steps; b) Boc2O, DMAP, CH3CN, RT;
c) LiOH, H2O2, THF/H2O, RT; 91% yield over 2 steps. Ts=toluenesul-
fonyl, TfN3 =triflic azide, Boc2O=di-tert-butyl dicarbonate, DMAP=4-
dimethylaminopyridine.
to the arylation reaction; replacement with either a mono-
carbamate or a dibenzyl amine shuts down the reaction
completely. Eventually, this dilemma was successfully
addressed by converting the phthaloyl unit into a much
smaller azido function. This transformation was conveniently
achieved through the initial ethylenediamine deprotection of
5, followed by a diazo transfer reaction with TfN3.[10] Gratify-
ingly, Boc activation of the reformulated amide 7 proceeded
smoothly with Boc2O and DMAP (Scheme 1b).[11] Upon
treatment of 8 with the Evans hydrolytic conditions (LiOH/
H2O2) at room temperature, the desired azido acid 9 was
formed in quantitative yield, and with complete chiral
integrity (Scheme 1b).[12] Interestingly, the azido substrate
11 completely failed in the indolylation reaction.
We then set out to apply the indolylation chemistry and
auxiliary removal methodology to the total synthesis of
celogentin C. The key iodotryptophan precursor 14 was
rapidly prepared from commercially available compound 12
using a nitration–reduction–Sandmeyer reaction sequence
(Scheme 2).[13,2f] The Trp Ca was protected as the tert-butyl
ester for the anticipated construction of ring B (Figure 1). We
were delighted to find that the indolylation reaction of 2
(2.0 equiv) with iodotryptophan 14 (1.0 equiv) worked
extremely well under the Pd(OAc)2/AgOAc/tBuOH condi-
tions, affording an 85% yield of isolated 15 (along with 3% of
the de-iodinated side product), and complete diastereoselec-
tivity on a 4 gram scale. The N-Phth unit was then converted
into the azido group using the same sequence as described for
azide 9. Boc-activation of amide 16 proceeded smoothly with
concomitant di-Boc protection of Trp Na. The subsequent
amide cleavage of 17 proceeded quantitatively upon treat-
ment of LiOH and H2O2.
The resulting azido acid 18 was then converted into ester
19 using DCC and HOSu (Scheme 3). Upon coupling of 19
with the dipeptide NH2Val-LeuOH and then Boc deprotec-
tion with HCl in dioxane, the cyclization precursor 20 was
obtained in excellent yield. The macrolactamization of 20 was
effected cleanly by EDCI/HOOBt to give 21 in 82% yield
without detectable epimerization. The azido group of 21 was
then reduced and the resulting amine was coupled with
pyroglutamic acid to furnish 22. The N-toluenesulfonyl and
tert-butyl protecting groups were then removed to yield acid
23, which successfully intercepts the Castle synthesis. It is
1
worth noting that the H NMR spectroscopic chemical shifts
(especially the Trp protons) of the protons in [D4]methanol
can be affected by residual trifluoroacetic acid. The methyl
ester 24 was then prepared and matched the previously
reported spectrum perfectly (see the Supporting Informa-
tion).
Employing the elegant route developed by Castle et al.,
NH2ProOBn was installed at the Trp Ca, and the His N1 of
dipeptide CbzNHArg(Pbf)-HisOtBu was linked at the Trp C2
Angew. Chem. Int. Ed. 2010, 49, 958 –961
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
959

Communications
Scheme 4. Completion of the synthesis of 1. Reagents and conditions:
a) EDCI, HOBt, DIPEA, NH2ProOBn·HCl, CH2Cl2, RT, 75% yield;
b) NCS, DMP, NH2ProOBn, CH2Cl2, then CbzNHArg(Pbf)-HisOtBu,
RT; c) NH4HCO2, Pd/C, MeOH/H2O, RT; HBTU, HOBt, DIPEA, DMF,
RT; TFA/H2O, RT; approx. 30% over 4 steps. DMF=N,N-dimethyl-
formamide, EDCI=1-ethyl-3-(3-dimethylaminopropyl)carbodiimide,
HOBt=N-hydroxybenzotriazole, DIPEA=diisopropylethylamine,
NCS=N-chlorosuccinimide, DMP=1,4-dimethylpiperazine,
HBTU=O-benzotriazol-1-yl-N,N,N’,N’-tetramethyluronium hexafluoro-
phosphate„ Pbf=2,2,4,6,7-pentamethyldihydrobenzofuran-5-sulfonyl,
Bn=benzyl, Cbz=carbobenzoxy.
Scheme 3. Synthesis of ring A. Reagents and conditions: a) HOSu,
DCC, CH2Cl2, RT, 92% yield; b) NH2Leu-ValOH, NaHCO3, DMF/H2O,
RT, 84% yield; 4molLꢀ1 HCl/dioxane, RT, 85% yield; c) EDCI, HOOBt,
CH2Cl2, RT, 82% yield; d) H2, 10% Pd/C, HOAc, EtOAc, RT, 85%
yield; pyroglutamic acid, EDCI, HOOBt, CH2Cl2, RT, 92% yield;
e) Magnesium, MeOH, sonication, RT, 87% yield; TFA/TIPS/H20, RT,
96% yield; f) SOCl2, MeOH, RT, 82% yield. HOSu=N-hydroxysuccin-
imide, DCC=dicyclohexylcarbodiimide, DMF=N,N-dimethylform-
amide, EDCI=1-ethyl-3-(3-dimethylaminopropyl)carbodiimide,
HOOBt=3-hydroxy-1,2,3-benzotriazin-4(3H)-one, TFA=trifluoroacetic
acid, TIPS=triisopropylsilane.
argon flushing, covered with aluminum foil, and heated at 1108C for
16 h. The reaction mixture was then cooled to room temperature,
diluted with 10 mL of dichloromethane, filtered through celite,
concentrated under vacuum, and purified by flash chromatography
(20–50% ethyl acetate in hexanes) to give compound 5 (43 mg, 80%),
compound 6 (approx. 1 mg), and recovered compound 2 (28 mg);
½aꢁ2D3 = ꢀ49 (c = 1.4, CHCl3); 1H NMR (300 MHz, CDCl3): d = 9.72 (s,
1H), 8.54 (dd, J = 3.1, 5.6 Hz, 1H,), 8.16 (s, 1H), 7.96 (m, 2H), 7.92 (d,
J = 8.9 Hz, 1H,), 7.85 (d, J = 2.6 Hz, 1H), 7.78 (m, 2H), 7.63 (d, J =
8.6 Hz, 2H), 7.58 (d, J = 8.0 Hz, 2H), 7.40–7.30 (m, 3H), 7.12 (m, 1H),
6.72 (d, 1H, J = 3.4 Hz), 6.65 (d, J = 7.9 Hz, 2H), 5.76 (d, J = 12.3 Hz,
1H), 4.34 (dd, J = 2.8, 12.1 Hz, 1H), 2.08 (m, 1H), 1.89 (s, 3H), 0.84
(d, J = 6.7 Hz, 3H), 0.74 ppm (d, J = 6.7 Hz, 3H); 13C NMR (75 MHz,
CDCl3): d = 168.7, 166.2, 147.7, 144.6, 138.1, 135.3, 134.7, 134.3, 133.1,
132.0, 130.3, 129.3, 127.4, 126.9, 126.7, 126.6, 123.8, 121.6, 121.5, 121.2,
by NCS-mediated oxidative coupling (Scheme 4).[4,14] Subse-
quently, the benzyl and carbobenzoxy protecting groups were
removed from 26 by hydrogenolysis, and ring B was cyclized
using a HBTU/HOBt-mediated coupling. The final depro-
tection steps with trifluoroacetic acid completed the total
synthesis of 1 in a total of 23 steps from simple amino acid
building blocks.
116.8, 108.8, 57.8, 48.9, 29.6, 21.6, 21.1, 16.1 ppm; IR (thin film): nmax
=
3281.0, 2953.2, 1709.3, 1527.6, 1381.5, 1172.2 cmꢀ1; HRMS (ESI+)
In summary, we have demonstrated a powerful method-
ꢀ
ology for the stereoselective indolylation of the b C H of N-
calcd. for C38H33N4O5S [M+H]+ 657.2166; found 657.2171.
ꢀ
Phth Leu by Pd(OAc)2 catalyzed C H activation. The
Received: September 14, 2009
Revised: November 15, 2009
Published online: December 22, 2009
aminoquinoline auxiliary was effectively detached under
mild conditions. A concise synthesis of celogentin C was
accomplished using this methodology; the synthesis and
biological studies of the other members of the celogentin
family and analogues are currently under investigation.
ꢀ
Keywords: amino acids · C H activation · natural products ·
palladium · total synthesis
.
Experimental Section
[1] J. Kobayashi, H. Suzuki, K. Shimbo, K. Takeya, H. Morita, J.
Org. Chem. 2001, 66, 6626 – 6633.
ꢀ
A typical C H indolylation procedure (synthesis of compound 5): A
4 mL glass vial with a PTFE lined cap was charged with compound 2
(63 mg, 0.16 mmol, 2.0 equiv), compound
1.0 equiv), Pd(OAc)2 (Aldrich 98%, 3.7 mg, 0.016 mmol, 0.2 equiv),
AgOAc (Aldrich 99%, 21 mg, 0.12 mmol, 1.5 equiv), and 0.3 mL of
tBuOH (ACS grade). The reaction vial was capped with or without
[2] a) J. Harrison, C. J. Moody, Tetrahedron Lett. 2003, 44, 5189 –
5191; b) S. L. Castle, G. S. C. Srikanth, Org. Lett. 2003, 5, 3611 –
3614; c) D. J. Bentley, C. J. Moody, Org. Biomol. Chem. 2004, 2,
3545 – 3547; d) D. J. Bentley, A. M. Z. Slawin, C. J. Moody, Org.
Lett. 2006, 8, 1975 – 1978; e) L. He, L. Yang, S. L. Castle, Org.
3 (32 mg, 0.08 mmol,
960
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 958 –961

Angewandte
Chemie
Lett. 2006, 8, 1165 – 1168; f) J. Michaux, P. Retailleau, J. M.
Campagne, Synlett 2008, 1532 – 1536.
[3] L. Feliu, M. Planas, Int. J. Pept. Res. Ther. 2005, 11, 53 – 97.
[4] B. Ma, D. N. Litvinov, L. He, B. Banerjee, S. L. Castle, Angew.
Chem. 2009, 121, 6220 – 6223; Angew. Chem. Int. Ed. 2009, 48,
6104 – 6107.
[5] For NRP biosynthesis, see: a) C. T. Walsh, Science 2004, 303,
1805 – 1810; b) S. A. Sieber, M. A. Marahiel, J. Bacteriol. 2003,
185, 7036 – 7043.
[8] a) V. G. Zaitsev, D. Shabashov, O. Daugulis, J. Am. Chem. Soc.
2005, 127, 13154 – 13155; b) D. Shabashov, O. Daugulis, Org.
Lett. 2005, 7, 3657 – 3659.
[9] The removal of the aminoquinoline auxiliary has not been
demonstrated in previous studies.
[10] P. T. Nyffeler, C.-H. Liang, K. M. Koeller, C.-H. Wong, J. Am.
Chem. Soc. 2002, 124, 10773 – 10778.
[11] D. L. Flynn, R. E. Zelle, P. A. Grieco, J. Org. Chem. 1983, 48,
2424 – 2426.
[12] D. A. Evans, T. C. Britton, J. A. Ellman, Tetrahedron Lett. 1987,
28, 6141 – 6144.
[13] R. S. Phillips, L. A. Cohen, J. Heterocycl. Chem. 1988, 25, 191 –
192.
[14] An alternative intramolecular oxidative cyclization strategy was
tested with only minimal success; more-exhaustive studies of
that approach are underway.
ꢀ
[6] For C H activation in natural product synthesis, see: a) K.
Godula, D. Sames, Science 2006, 312, 67 – 72; b) H. M. Davies,
J. R. Manning, Nature 2008, 451, 417 – 424; c) A. Hinman, J.
DuBois, J. Am. Chem. Soc. 2003, 125, 11510 – 11511.
[7] B. V. S. Reddy, L. R. Reddy, E. J. Corey, Org. Lett. 2006, 8, 3391 –
3394.
Angew. Chem. Int. Ed. 2010, 49, 958 –961
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
961