Fe/Cu-Mediated One-Pot Ketone Synthesis

Article abstract of DOI:10.1021/acs.orglett.7b01128

An Fe/Cu-mediated one-pot ketone synthesis was reported. Unlike Ni- and Pd-mediated one-pot ketone syntheses, the reported Fe/Cu-mediated method allowed selective activation and coupling of alkyl iodides over vinyl iodides. The newly developed one-pot ket

Full text of DOI:10.1021/acs.orglett.7b01128

Letter
pubs.acs.org/OrgLett
Fe/Cu-Mediated One-Pot Ketone Synthesis
Vemula Praveen Kumar, Vaddela Sudheer Babu, Kenzo Yahata, and Yoshito Kishi*
Department of Chemistry and Chemical Biology, Harvard University, 12 Oxford Street, Cambridge, Massachusetts 02138, United
States
S
* Supporting Information
ABSTRACT: An Fe/Cu-mediated one-pot ketone synthesis
was reported. Unlike Ni- and Pd-mediated one-pot ketone
syntheses, the reported Fe/Cu-mediated method allowed
selective activation and coupling of alkyl iodides over vinyl
iodides. The newly developed one-pot ketone synthesis was
applied to a synthesis of vinyl iodide/ketone 13, the left half of
halichondrin B, as well as vinyl iodide/ketone 8a, the C20−
C26 building block of halichondrins.
he addition of organometallic reagents onto carboxylic
acids or derivatives gives a simple solution for ketone
(3) LiCl effect,9 (4) 1a:2a molar ratio,10 (5) reducing metal,11
(6) solvent and concentration,12 and (7) additives.13 Through
this study, the conditions that include “1a (1.0 equiv), 2a (3.0
equiv), Fe(TMHD)3 (10 mol %; 4 in Scheme 2), CuCl2 (1.0
equiv), Mn (2 equiv), LiCl (3 equiv), DME (C 0.4 M), 0 °C,
15 h” were found to be effective for the (1a + 2a)-coupling
(76% isolated yield).14
T
synthesis. One drawback associated with this method is the fact
that desired ketones often react further with organometallic
reagents. Weinreb ketone synthesis offers an ingenious solution
to overcome this drawback.1 In recent years, it has been
demonstrated that Ni- or Pd-mediated coupling of an activated
form of carboxylic acid with an organometallic offers an
alternative solution.2
Scheme 2. Fe-Catalysts Used in the Study
We recently reported a one-pot ketone synthesis with
alkylzinc halides, prepared in situ from alkyl halides via a single
electron transfer (SET) process, and were curious to extend
this concept to development of Cu-mediated one-pot ketone
synthesis for two reasons.3 First, Cu-mediated one-pot ketone
synthesis might exhibit a reactivity profile different from Ni-
and/or Pd-mediated one-pot ketone syntheses; for example, the
latter methods do not match well with nucleophiles bearing
vinyl and/or aryl halides.4 Second, it is well recognized that
overaddition of organometallic reagents is not the issue for
cuprate-based ketone synthesis.5 In this letter, we report the
development and application of Fe/Cu-mediated one-pot
ketone synthesis.
The scope and limitation of this method was then studied
under the conditions of Method A in Scheme 3. Overall, the
efficiency of Cu-mediated one-pot ketone synthesis was
comparable with that of Pd- and Ni-mediated one-pot ketone
syntheses. As expected, however, the Cu-mediated method
exhibited an appealing reactivity profile; that is, unlike the Pd-
and Ni-mediated method, the Cu-mediated method allowed
selective activation and coupling of an alkyl iodide over a vinyl
or aryl iodide, cf., 1j−m. Related to the ongoing project, we
were interested in this selectivity. In particular, we recognized
the possibility of synthesizing 8a, the C20−C26 building block
of halichondrins, via the coupling of 6 with 7.15
For the feasibility study of the proposed reductive coupling,
we chose the substrates shown in Scheme 1, because this pair of
Scheme 1. Substrates Used for Feasibility Test
Under the condition of Method A (Scheme 3), the one-pot
ketone coupling gave the desired product 8a. However, fine-
tuning was necessary to efficiently perform the coupling. The
activation rate of alkyl iodide in 7 was slower than that in the
model 1j. With a higher loading of Fe(TMHD)3, the coupling
rate was accelerated as expected, but the product was
substrates was used for the study on Pd- and Ni-mediated one-
pot ketone synthesis in this laboratory.3,6 The first attempt
under the arbitrarily chosen condition [2a (5 equiv), 1a (1
equiv), MnPc (10 mol %), CuCN (1 equiv), LiCl (3 equiv),
THF (C = 0.2 M), rt, 6 h] gave 3a in 35% isolated yield. Being
encouraged by this result, we optimized the coupling condition,
including (1) radical initiator and loading,7 (2) copper source,8
Received: April 13, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.7b01128
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
might be difficult to achieve by the Pd- or Ni-mediated one-pot
ketone synthesis. Nevertheless, we were anxious to improve
one particular aspect of this coupling, i.e., a 3:1 molar ratio of
electrophiles and nucleophiles. For the case of (6 + 7)-
coupling, the unbalanced molar ratio of electrophile and
nucleophile was acceptable, because 6 was significantly less
expensive than 7. However, it would not be appealing to apply
this coupling for the inverse case, i.e., a more expensive
electrophile compared to the nucleophile.
Scheme 3. Scope and Limitation
It was speculated that some of 6 was wasted during the
activation process of the Fe-catalyst; in order for Fe(TMHD)3
to function as a radical initiator, Fe(III) should be reduced to
Fe(II) by Mn metal. The reduction released one molecule of β-
diketone, which consumed some of 6 in a nonproductive
manner. If so, this side reaction could be avoided with use of an
Fe(II)-initiator. For this reason, various radical initiators were
screened for the Cu-mediated ketone coupling. Among them,
FeBr2·(dppb) was found to promote the (1a + 2a)-coupling
well.17 Through optimization, it was found that the coupling
was effectively achieved under the following conditions: FeBr2·
(dppb) (5 mol %), acid chloride (1.0 equiv), iodide (1.2 equiv),
CuCl2 (1.0 equiv), LiCl (3 equiv), Mn (2 equiv), DME (C =
0.4 M), 0 °C, 15 h.13 Interestingly, FeBr2·(dppb) gave a better
coupling yield than the corresponding FeCl2·(dppb).18 Under
the optimized condition, the coupling was tested with the molar
ratio of 1a/2a being 1.2/1.0 and 1.0/1.2, to give 3a in 90% and
87% isolated yield, respectively. With the 1.2/1.0 molar ratio of
nucleophile and electrophile, the coupling efficiency was
studied for the substrates listed in Method B in Scheme 3.
The FeBr2·(dppb) condition was applied for the (6 + 7)-
coupling, to give 8a in 72% isolated yield. The coupling yield
was further improved up to an 80% yield, by replacing FeBr2·
(dppb) with the FeBr2-complex prepared from the SciOPP-
ligand, recently reported by Nakamura and co-workers.19
Phosphine-based FeBr2-catalysts allowed us to efficiently
achieve the one-pot ketone synthesis, even with a near 1:1
molar ratio of nucleophiles and electrophiles. Encouraged by
this result, we attempted to apply this approach to a synthesis
of vinyl iodide 13, the left half in the halichondrin-B series, as
well as closely related vinyl iodide 11.20 In the latter series, we
were able to prepare the acid chloride and demonstrated that
the coupling gave the desired product 11 in 20−25% overall
yield from the carboxylic acid.21 However, in the former series,
we could not prepare the required acid chloride. Overall,
because of their instability during preparation and/or coupling,
these acid chlorides were not useful for our purpose. Under
these circumstances, we focused on a 2-thiopyridine ester as an
alternative electrophile, because it proved to be an excellent
electrophile in the Zr/Ni-mediated one-pot ketone syn-
thesis.6,22
accompanied by a small amount of side-product 8b (X = H).16
Tuning the Fe(TMHD)3 loading provided a practical solution
to overcome this problem; with 13−15 mol % catalyst, the (6 +
7)-coupling gave the desired ketone 8a (75% isolated yield),
along with a trace amount of 8b (<1% yield), in 15 g scale
experiments (Scheme 4).
Scheme 4. Synthesis of the C20−C26 Building Block of
Halichondrins via Fe/Cu-Mediated One-Pot Coupling
With this background, we tested the (1a + 2b → 3a)
coupling under the conditions of Method B and obtained the
desired product in ∼15% yield. Being encouraged by this result,
we optimized the coupling, thereby revealing that the 2-
thiopyridine ester was not stable under the conditions of
Method B; namely, 2b rapidly decomposed in the presence of
CuCl2, as anticipated from the literature precedent.23 However,
the 2-thiopyridine ester was found to be stable in the presence
of CuI, immediately suggesting the coupling in the presence of
CuI, instead of CuCl2 (Method C). The coupling efficiency
under this condition was studied for each substrate listed in
Scheme 3.
Similarly, the Fe/Cu-mediated coupling with the vinyl
bromide corresponding to 7 gave the desired product in a
comparable yield.
As illustrated in the transformation of 6 + 7 → 8a, the Fe/
Cu-mediated one-pot ketone synthesis, initiated with Fe-
(TMHD)3, exhibited a unique profile of reactivity, which
B
DOI: 10.1021/acs.orglett.7b01128
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
The Fe/Cu-mediated one-pot ketone synthesis under the
conditions of Method C furnished vinyl iodide 13, the left half
in the halichondrin-B series, as well as closely related vinyl
iodide 11, with a 1.0:1.2 molar ratio of electrophile and
nucleophile (Scheme 5). This new synthetic route had a few
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.7b01128.
■
S
Experimental procedures, characterization data, and
copies of spectra (PDF)
Scheme 5. Substrates Used for Feasibility Test of Fe/Cu- or
Co/Cu-Mediated Ketone Synthesis
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: kishi@chemistry.harvard.edu.
ORCID
Yoshito Kishi: 0000-0001-5612-350X
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Dr. Changfeng Huang of this laboratory for his
participation in scaling studies on the C20−C26 building block.
Financial support from the Eisai USA Foundation is gratefully
acknowledged. K.Y. thanks the Naito Foundation for a
fellowship.
REFERENCES
■
(1) Nahm, S.; Weinreb, S. M. Tetrahedron Lett. 1981, 22, 3815.
(2) For a general review on ketone syntheses with organometallics,
see: (a) Dieter, R. K. Tetrahedron 1999, 55, 4177. For selected
references for metal-catalyzed ketone syntheses, see: Ni/Zn or Mn:
(b) Onaka, M.; Matsuoka, Y.; Mukaiyama, T. Chem. Lett. 1981, 10,
531. (c) Wotal, A. C.; Weix, D. J. Org. Lett. 2012, 14, 1476. (d) Wotal,
A. C.; Ribson, R. D.; Weix, D. J. Organometallics 2014, 33, 5874.
(e) Wu, F.; Lu, W.; Qian, Q.; Ren, Q.; Gong, H. Org. Lett. 2012, 14,
3044. (f) Zhao, C.; Jia, X.; Wang, X.; Gong, H. J. Am. Chem. Soc. 2014,
136, 17645 and references cited therein. (g) Cherney, A. H.; Kadunce,
N. T.; Reisman, S. E. J. Am. Chem. Soc. 2013, 135, 7442. RMgX/Ni:
(h) Fiandanese, V.; Marchese, G.; Ronzini, L. Tetrahedron Lett. 1983,
24, 3677. RMgX/Ni and RMgX/Fe: (i) Cardellicchio, C.; Fiandanese,
V.; Marchese, G.; Ronzini, L. Tetrahedron Lett. 1985, 26, 3595 and
references cited therein. RZnX/Pd: (j) Negishi, E.-i.; Bagheri, V.;
Chatterjee, S.; Luo, F.-T.; Miller, J. A.; Stoll, A. T. Tetrahedron Lett.
1983, 24, 5181. RSnX3/Pd: (k) Wittenberg, R.; Srogl, J.; Egi, M.;
Liebeskind, L. S. Org. Lett. 2003, 5, 3033. RB(OH)2/Pd:
(l) Liebeskind, L. S.; Srogl, J. J. Am. Chem. Soc. 2000, 122, 11260.
RSnX3/Cu: (m) Li, H.; He, A.; Falck, J. R.; Liebeskind, L. S. Org. Lett.
2011, 13, 3682. R2Zn/Ni: (n) Zhang, Y.; Rovis, T. J. Am. Chem. Soc.
2004, 126, 15964.
appealing aspects, including (1) a higher degree of convergency
and (2) introduction of the C39 vinyl group before the ketone
coupling via a standard transformation of terminal acetylene to
trans-vinyl iodide.24
Finally, the behavior of two commonly used radical probes
was tested under the three coupling conditions (Scheme 6). As
expected, the observed results indicated that a radical
intermediate(s) was involved in all three sets of coupling
conditions.
Scheme 6. Substrates Used for Feasibility Test of Fe/Cu- or
Co/Cu-Mediated Ketone Synthesis
(3) Lee, J. H.; Kishi, Y. J. Am. Chem. Soc. 2016, 138, 7178.
(4) In this paper, an alkyl iodide is referred to as a nucleophile,
because of its conversion to a nucleophile in situ.
(5) For a review, see for example: (a) Knochel, P.; Betzemeier, B.
Modern Organocopper Chemistry; Wiley-VCH: 2002. (b) Normant, J. F.
Synthesis 1972, 1972, 63. (c) Lipshutz, B. H. Synthesis 1987, 1987, 325.
(6) Ai, Y.; Ye, N.; Yahata, K.; Kishi, Y. Submitted for publication.
(7) Relative reactivity of radical initiators tested was roughly in the
following order: Fe(TMHD)3 > Fe(DBM)3 > CoPc > Fe(acac)3 ≈
ZnPc > MnPc ≈ FePc.
(8) Several Cu(I) salts were studied: CuCl, CuBr, CuI, CuCN, CuTc
gave 62%, 20%, 58%, 12%, and 10% yields, respectively. Also,
replacement of Cu(I) salts with Cu(II) salts was tested: CuCl2 and
CuBr2 yielded 2a in 76% and 32%, respectively.
In summary, an Fe/Cu-mediated one-pot ketone synthesis
was reported. Unlike Ni- and Pd-mediated one-pot ketone
syntheses, the Fe/Cu-mediated method allows selective
activation and coupling of alkyl iodides over vinyl iodides for
one-pot ketone synthesis. The newly developed method was
applied to a synthesis of vinyl iodide/ketone 13, the left half of
halichondrin B, as well as vinyl iodide/ketone 8a, the C20−C26
building block of halichondrins.
(9) LiCl, LiBr, and LiI were tested, thereby showing that 3 equiv of
LiCl gave the best result; see Supporting Information for details.
(10) The molar ratios of 1a:2a = 1.0:1.5, 1.0:2.0, 1.0:3.0 were tested.
Among them, 1.0:3.0 was found to be the best.
C
DOI: 10.1021/acs.orglett.7b01128
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(11) Mn- and Zn-powders were tested; Mn gave better results.
(12) DME gave slightly better results. Other solvents such as DMI,
DMA, DMF, and 1,4-dioxane gave only poor-to-low yields.
Concentration effects were also studied with C = 0.2, 0.3, and 0.4
M, but no significant difference was noticed. Thus, C = 0.4 M was
chosen for the study.
(13) Addition of Cp2ZrCl2, MnCl2, ZnCl2, MgCl2, and CrCl2 resulted
in a lower coupling efficiency.
(14) See Supporting Information for details.
(15) Previously, this coupling was done in two steps, i.e., Co/Cr-
mediated coupling, followed by oxidation; (a) Kim, D.-S.; Dong, C.-
G.; Kim, J. T.; Guo, H.; Huang, J.; Tiseni, P. S.; Kishi, Y. J. Am. Chem.
Soc. 2009, 131, 15636. (b) Dong, C.-G.; Henderson, J. A.; Kaburagi,
Y.; Sasaki, T.; Kim, D.-S.; Kim, J. T.; Urabe, D.; Guo, H.; Kishi, Y. J.
Am. Chem. Soc. 2009, 131, 15642.
(16) On treatment with the reagent mixture of Method A for 12 h at
room temperature, 8a was completely deiodinated, to give 8a.
(17) Phosphine complexes FeBr2·(dppb), FeCl2·(dppb), FeBr2·
(dppe), FeCl2·(dppe), and FeBr2·(PPh3)2 gave product 3a in 90%,
79%, 54%, 46%, and 48% yield, respectively, under the coupling
condition Method B (Scheme 3).
(18) Britovsek, G. J. P.; England, J.; Spitzmesser, S. K.; White, A. J. P.;
Williams, D. J. Dalton Trans. 2005, 945 and references cited therein.
(19) FeCl2(SciOPP) is known as a superior catalyst for Kumada−
Tamao−Corriu coupling; Hatakeyama, T.; Fujiwara, Y.; Okada, Y.;
Itoh, T.; Hashimoto, T.; Kawamura, S.; Ogata, K.; Takaya, H.;
Nakamura, M. Chem. Lett. 2011, 40, 1030.
(20) Buszek, K. R.; Fang, F. G.; Forsyth, C. J.; Jung, S. H.; Kishi, Y.;
Scola, P. M.; Yoon, S. K. Tetrahedron Lett. 1992, 33, 1553.
(21) With FeBr2·(dppb) and FeBr2·(SciOPP), 11 was obtained in
20% and 25% yields, respectively.
(22) (a) Araki, M.; Sakata, S.; Takei, H.; Mukaiyama, T. Bull. Chem.
Soc. Jpn. 1974, 47, 1777. (b) Onaka, M.; Matsuoka, Y.; Mukaiyama, T.
Chem. Lett. 1981, 10, 531.
(23) 2-Thiopyridine esters are known to be activated in the presence
of CuBr2 and CuCl2: Kim, S.; Lee, J. I. J. Org. Chem. 1984, 49, 1712.
(24) In the previous synthesis, the trans-vinyl iodide at C39 was
introduced via Takai trans-iodoolefination; see: Takai, K.; Nitta, K.;
Utimoto, K. J. Am. Chem. Soc. 1986, 108, 7408. (b) Takai, K.;
Ichiguchi, T.; Hikasa, S. Synlett 1999, 1999, 1268.
D
DOI: 10.1021/acs.orglett.7b01128
Org. Lett. XXXX, XXX, XXX−XXX