https://doi.org/10.

1038/s41586-024-07341-z

Accelerated Article Preview

W
Copper-catalyzed dehydrogenation or
E
I
lactonization of C(sp3)−H bonds
EV
PR
Received: 27 September 2023

Accepted: 21 March 2024

Accelerated Article Preview

Shupeng Zhou, Zi-Jun Zhang & Jin-Quan Yu

This is a PDF file of a peer-reviewed paper that has been accepted for publication.

E
Published online xx xx xxxx Although unedited, the content has been subjected to preliminary formatting. Nature

L
Cite this article as: Zhou, S. et al. is providing this early version of the typeset paper as a service to our authors and
Copper-catalyzed dehydrogenation or readers. The text and figures will undergo copyediting and a proof review before the

C
lactonization of C(sp3)−H bonds. Nature paper is published in its final form. Please note that during the production process
https://doi.org/10.1038/s41586-024-07341-z

I
errors may be discovered which could affect the content, and all legal disclaimers
(2024)

T
apply.

AR
ED
AT
R
E
EL
C
C
A

Nature | www.nature.com
 1 Copper-catalyzed dehydrogenation or lactonization of C(sp3)−H bonds

W
2 Authors: Shupeng Zhou,1,† Zi-Jun Zhang,1,† Jin-Quan Yu1,*

IE
3 Affiliations: 1Department of Chemistry, The Scripps Research Institute, 10550 North Torrey Pines Road, La

EV
4 Jolla, California 92037, United States.

PR
5 *Correspondence to: [email protected]

6 †These authors contributed equally to this work.

LE
7

8
C
Cytochrome P450 enzymes are known to catalyze bimodal oxidation of aliphatic acids via radical
TI
9 intermediates, which partition between pathways of hydroxylation and desaturation5,6. Developing
AR

10 analogous catalytic systems for remote C−H functionalization remains a significant challenge14,15,16.

11 Here we report the development of Cu(I)-catalyzed bimodal dehydrogenation/lactonization reactions

E D

12 of synthetically common N-methoxyamides via radical abstractions of the γ-aliphatic C−H bonds. The
AT

13 feasibility of switching from dehydrogenation to lactonization has also been demonstrated by altering

14 reaction conditions. The use of a readily available amide as both radical precursor and internal oxidant
ER

15 allowed for the development of a redox-neutral C−H functionalization reactions with methanol as the
EL

16 sole side product. These C−H functionalization reactions using Cu(I) catalyst of loading as low as 0.5
C

17 mol% have been applied to the diversification of a wide range of aliphatic acids including drug
AC

18 molecules and natural products. The exceptional compatibility of this catalytic system with a wide

1
19 range of oxidatively sensitive functionality demonstrates the unique advantage of using simple amide

20 substrate as the mild internal oxidant.

W
Main Text: Olefins and carbon-oxygen bonds are ubiquitous and highly important in organic synthesis1. A

IE
21

EV
22 particularly attractive strategy for constructing such moieties is the direct oxidation of inert C−H bonds. The

23 generation of a C−O bond at the expense of a C−H bond leads to increased complexities and enables the

PR
24 direct synthesis of the target of interest from a simple hydrocarbon fragment2,3. From the perspective of

25 downstream diversification, the desaturation of aliphatic chains by C−H dehydrogenation can be even more

LE
26 versatile than the direct conversion of C−H to C−O bonds. Nature has evolved various enzymes to catalyze

27
C
the direct oxidation of hydrocarbon skeletons with great precision4. In some cases, a single enzyme was found
TI
28 to simultaneously catalyze different types of reactions5,6. In 1987, a landmark discovery by Baillie and
AR

29 coworkers showcased that a type of hepatic cytochrome P450, normally considered a hydroxylase, could also

30 act as a desaturase5. It is believed that a common carbon-centered radical intermediate is responsible for this
E D

31 mixed hydroxylase/desaturase activity (Fig. 1a). Inspired by this remarkable enzymatic chemistry, we
AT

32 envisioned the possibility of accessing remote bimodal C−H dehydrogenation/oxygenation reaction with

33 metal catalysts through radical abstraction. Such a reaction could be useful for direct oxidation state elevation
ER

34 of hydrocarbon frameworks2,3, and it would be a valuable tool for late-stage modifications and
EL

35 diversifications of natural products and drug molecules7-9. Although biomimetic dehydrogenation and
C

36 oxygenation reactions based on the hydrogen atom abstraction strategy have been reported starting with the
AC

37 pioneering work of Breslow10,11 and Groves12,13 et al., dual desaturation/oxygenation reaction remains a

38 grand challenge. The few early examples suffered from limited scope (only for benzylic C−H)14,15 and/or

2
39 overoxidation15,16. Moreover, these methods require exogenous stoichiometric oxidants, most of which give

40 mixed dehydrogenated/hydroxylated products and cannot be tuned to produce a single major product.

W
41 Therefore, the development of a controllable bimodal dehydrogenation/oxygenation reaction would be highly

IE
42 desirable.

EV
PR
LE
C
TI
AR
E D
AT
ER
EL

43
C

44 N-methoxyamides have become a major class of practically useful substrates in Pd(II)- and Rh(III)-
AC

45 catalyzed C−H activation reactions17 since their first introduction18. They can be readily prepared in large

46 quantities from carboxylic acids in a single step and are bench-stable for long-term storage. Despite advances

3
47 in the field of amidyl radical generation via N−O cleavage19,20, specifically from activated O-acyl and O-aryl

48 hydroxamides with photocatalysis21-24, the simple N-methoxyamide substrates have not been shown to be

W
49 compatible with these chemistry, presumably due to its relatively high reduction potential25. Notably,

IE
50 developing C−H functionalization reactions based on such radical abstraction require external traps20 or

EV
51 photocatalysis of highly redox-active groups26 to close the redox catalytic cycle. Inspired by previously

PR
52 reported Cu(I)-mediated formation of iminyl radicals from active oxime esters and subsequent cyclization

53 with tethered olefins27, we wondered whether Cu(I) could reduce N-methoxyamides to form amidyl radicals

LE
54 (the efforts to overcome the relatively high BDE of N−O bond of N-methoxyamide are summarized in the

C
55 SI). If successful, this reductive method would essentially use the methoxy group as the green internal oxidant,
TI
56 thus omitting the photoredox catalysts, and external oxidants which often lead to overoxidation. This amidyl
AR

57 radical I can then perform 1,5-H atom abstraction to form the γ-carbon-centered radical IV (Fig. 1b). We

58 envisioned that the combination of this alkyl radical with Cu(II) species could either favor oxidative
D

59 elimination to afford dehydrogenation products (V to II, path A) or undergo oxidative substitution leading
E

to the formation of lactone III via intramolecular trapping of carbocationic intermediate VI (path B)28-30. In
AT

60

61 both pathways, Cu(I) would be regenerated, closing this redox-neutral catalytic cycle. The only by-product
ER

62 of the dehydrogenation reaction is MeOH (or MeOH and ammonium salt for the lactonization). Here, we
EL

63 report a redox-neutral Cu-catalyzed bimodal dehydrogenation/lactonization reaction of N-methoxyamides

64 without using highly reactive redox-active groups which require additional installation steps. This platform
C
AC

65 would allow the rapid diversification of widely available carboxylic acids into valuable lactones and

4
66 dehydrogenated primary amides (Fig. 1c), where the amides could be further diversified by transformations

67 of the amide, functionalization of the alkene and a cyclization between the amide and the double bond.

W
68 To our delight, the realization of the dehydrogenation was enabled by using CuF2 (10 mol%) as the

IE
69 catalyst, 8-methoxyquinoline (20 mol%) as the ligand, and AcOH as the additive in 1,4-dioxane or DCE (Fig.

EV
70 2; the details of the reaction optimization are discussed in the SI, see Figure S1-S6). Under these reaction

PR
71 conditions, N-methoxyamide A27 gave dehydrogenated product B27 in 70% yield.

72 With the optimized reaction conditions in hand, we began to explore the substrate scope of the

LE
73 dehydrogenation (Fig. 2 and Extended Data Fig. 1). Both unactivated (A1-A8) and benzylic (A9-A25) γ-

C
74 C−H bonds reacted to form the corresponding olefins in good yields with both acyclic and cyclic substrates.
TI
75 β,γ-olefins can also be obtained as the major product when the δ-position was blocked (B4 and B5). A
AR

76 particularly striking example is the formation of γ,δ-unsaturated skipped diene B2 rather than the conjugated

77 β,γ-olefin. We were delighted to find that N-methoxyamides bearing γ-methines were also competent
D

78 substrates, affording the desired olefins (B26-B38). Although small amounts of β,γ-unsaturated isomers were
E
AT

79 formed in some cases (B33-B36), the reaction generally exhibited a strong preference for the formation of

80 γ,δ- rather than β,γ-olefins. The reaction was also found to tolerate a wide range of functionality, including
ER

81 alkyl acetates (B3, B6, and B28), pre-existing olefins (B2, B20, and B26), alkynes (B21), (thio)ethers (B10,
EL

82 B22, B24, and B25), carbamates (B19), heterocycles (B11, B22, and B23), and potential cross coupling

83 partners for downstream elaborations such as aryl halides (B13, B14, and B18) and boronic esters (B15). The
C
AC

84 exceptional ability of this oxidative catalytic system to tolerate a wide range of oxidatively sensitive

5
85 functionality demonstrates the unique advantage of using amide substrate as the mild internal oxidant rather

86 than using harsh exogenous oxidants.

W
IE
EV
PR
LE
C
TI
AR
E D
AT
ER
EL

87
C

88 We next examined the reaction in a range of more complex settings. A variety of α-, β-, and γ-amino
AC

89 acid derivatives were subjected to the reaction, affording the corresponding dehydrogenated products in

90 synthetically useful yields (B39-B50, Extended Data Fig. 1). N-Methoxyamide derivatives of natural
6
 91 products and pharmaceuticals also proved amenable to dehydrogenation (B51-B65, Fig. 2). The exclusive

92 formation of mono-desaturation product B52 from valproic acid derivative A52 offers a particularly notable

W
93 example, as our internal oxidant strategy, unlike the use of external oxidant, prevents further dehydrogenation.

IE
94 Site-selectivity became somewhat more complicated with polycyclic substrates. Betulinic acid derivative

EV
95 B62 offers a particularly informative example wherein high yields of a single product were obtained despite

PR
96 the presence of four competing γ-C−H bonds. Notably, the major product was formed by HAT at an

97 unactivated methine rather than the allylic γ-methine, indicating that geometric factors can outweigh the

LE
98 intrinsic reactivity of the competing C−H bonds. As with the examples discussed above, the reaction

C
99 displayed a remarkable tolerance for sensitive functionality such as the conjugated triene system formed in
TI
100 B59, though migration of the pre-existing double bond was observed in B64 and B65 (see the SI for details).
AR

101 Considering that abietane-type structures are widely found in nature and are also precursors of many complex

102 terpenes, we decided to test a large-scale reaction on the derivative of dehydroabietic acid. Treatment of 1.6
D

103 gram of A58 with CuF2 (5 mol%) and CSA (0.5 equiv.) gave B58 in 72% yield after refluxing for 20 hours.
E

Amidyl radicals can in principle be utilized to synthesize γ-lactones via derivations of the classical
AT

104

105 Hofmann-Löffler-Freytag reaction31-35, but both the multiple-step reaction conditions involving external
ER

106 oxidants and narrowed substrate scope remain to be substantially improved (see Figure S7 of the SI for a
EL

107 detailed comparison). Nevertheless, these early studies prompted us to investigate if our catalytic system

108 involving a similar radical intermediate can be engineered to switch from dehydrogenation to γ-C−H
C
AC

109 lactonization, thereby realizing bimodal catalysis using a copper catalyst. When performing dehydrogenation

110 of substrate A29 under our standard reaction conditions, we observed the formation of a trace amount of

7
111 lactone C29. Hence, a series of experiments were conducted to favor the lactonization reaction pathway

112 (Figure S8 of the SI). In particular, DCE suppressed the formation of lactone, while replacing the AcOH

W
113 additive with a stronger acid such as TFA increased the yield of lactone. To our delight, when dioxane was

IE
114 used as the solvent and TFA was included as an additive, C29 was formed in 78% yield with no B29 detected.

EV
115 For substrates with unactivated γ-methylene C−H bonds, a combination of dioxane/nitromethane as solvent

PR
116 with [(CH3CN)4Cu]BF4 was found to be optimal for lactonization (Figure S9 of the SI). Representative

117 examples of this conditions-based switch in reactivity from dehydrogenation to lactonization are given in Fig.

LE
118 3a (see Figure S10 of the SI for details). Selectivity for lactonization over dehydrogenation can be perfectly

C
119 controlled for substrates bearing tertiary or benzylic γ-C−H bonds (e.g., C27-C29, C35, C37, C38, C48).
TI
120 Lactonization of long-chain fatty acids and amino acids with unactivated γ-methylene were also feasible via
AR

121 this method (e.g., C43 and C51), though competing dehydrogenation was observed with these starting

122 materials. Owing to the large differences between these two products, the rapid diversification of carboxylic
D

123 acids could be envisioned because of this dual dehydrogenation/lactonization reactivity, which is valuable
E
AT

124 from the perspective of chemical space expansion.

ER
EL
C
AC

8
 W
IE
EV
PR
LE
C
TI
AR
E D
AT
ER

125

126 A variety of additional substrates were prepared to test the generality of the lactonization reaction
EL

127 (Fig. 3b and Extended Data Fig. 2). Activated benzylic, allylic, and propargylic γ-C–H bonds (Extended Data
C

128 Fig. 2) generally underwent efficient lactonization regardless of whether the position was primary (C68-C73),
AC

129 secondary (C74-C83, C87-C90) or tertiary (C91). As noted above, substrates with unactivated secondary

130 (B84-C86) and tertiary (C92) γ-C−H bonds (Fig. 3b) also performed well in the reaction. In the case of C84,
9
131 although a more reactive δ-benzylic site is present, γ-lactone is still exclusively formed via 1,5-H atom

132 abstraction. As expected, the method displayed favorable functional group compatibility with synthetically

W
133 valuable aryl halides (C78) and oxidatively sensitive groups such as olefins (C81, C83, and C88), alkynes

IE
134 (C87 and C88), and highly electron rich arenes (C74), which are unlikely to be tolerated by conventional

EV
135 approaches that rely on strong external oxidants.

PR
136 We next sought to examine the reliability and synthetic utility of the reaction by examining the

137 lactonization of several derivatives of natural products and pharmaceuticals (C93-C102, Fig. 3b). As with

LE
138 the dehydrogenation, the γ-lactonization tolerated structural and functional complexity of these compounds

139
C
well. Notably, when the γ position is blocked, the amidyl radical can instead abstract a δ-hydrogen, for
TI
140 example, affording rearranged product C95 in 41% yield (see the SI for details). Interestingly, treatment of
AR

141 the derivative of estrone with the standard conditions gave the lactam as the main product (C102) with lactone

142 C101 as the minor product. This suggests that the reaction may proceed through a alkyl Cu(III) intermediate
D

that undergoes C−N and C−O bond formation to avoid the unfavorable formation of an unstabilized primary
E

143
AT

144 carbocation. We are currently exploiting this pathway to construct lactams from N-methoxyamides.
ER

145 Remarkably, simultaneous bimodal oxidation reactions at two different sites are also realized for

146 diacid substrates (Fig. 4a), allowing for either dehydrogenation-lactonization (BC102), or dehydrogenation-
EL

147 lactamization (BC103) sequentially. It is noteworthy that our reactions were also efficient using low loading
C

148 of catalyst with prolonged reaction time (Fig. 4b). Both dehydrogenation and lactonization proceeded
AC

149 smoothly with as low as 0.5 mol% of copper under an argon atmosphere, and the products were obtained in

150 almost the same yields as when 10 mol% catalyst was used. In addition, our reactions can also proceed at

10
151 lower temperature (100 ˚C) with extended reaction time (see Figure S12 for details). A series of experiments

152 were conducted to investigate the mechanism of the dehydrogenation and lactonization reactions (Extended

W
153 Data Fig. 3). While screening the reaction conditions, we found that when 2,2’-biquinoline was added as a

IE
154 ligand, the reaction mixture turned an intense purple color rather than the usual blue or green (see Figure S14

EV
155 of the SI for details). This is a characteristic of the formation of Cu(I)-biquinoline complex36,27 and suggests

PR
156 that Cu(I) produced by disproportionation of the CuF2 precatalyst is present in solution as a potential active

157 catalyst37. A radical clock experiment was performed to investigate the radical intermediacy. When

LE
158 compound A66 was subjected to the standard conditions, diene B66 was obtained in 45% yield (Extended

C
159 Data Fig. 3b). Presumably, the carbon-centered radical Int-1 formed by 1,5-H-abstraction rearranged via an
TI
160 opening of the adjacent cyclopropane, forming the primary allylic radical Int-2, which can subsequently
AR

161 undergo oxidative elimination to afford diene B66. Taken together, these results support a mechanism

162 initiated by Cu(I)-catalyzed oxidation of the N-methoxyamide to afford an amidyl radical, which then
D

163 rearranges to a carbon-centered radical at the γ-position via 1,5-HAT.

E
AT
ER
EL
C
AC

11
 W
IE
EV
PR
LE
C
TI
AR
D

164
E

165 Based on Kochi’s mechanistic studies on radical pathways with copper30, we envision two plausible
AT

166 pathways of the elimination: 1) a Cu(II)-induced oxidation of the alkyl radical to a carbocation followed by
ER

167 a subsequent β-deprotonation; 2) a recombination of the alkyl radical and Cu(II) to form an alkylcopper(III)

168 intermediate followed by a β-oxidative elimination. If the elimination proceeds through β-deprotonation, the
EL

169 more acidic β-H should be preferentially eliminated, which is contrary to the regioselectivity we observed.
C

170 Furthermore, we treated alkyl chloride A67 with AgBF4 for in-situ carbocation generation (Extended Data
AC

171 Fig. 3c). Contrary to the regioselectivity of our reaction, this reaction afforded the β,γ- and γ,δ-olefins in a

172 10:1 mixture (Figure S16 of the SI). Thus, our copper-catalyzed dehydrogenation unlikely proceeds via a
12
173 cation intermediate, and the second pathway involving organocopper(III) species is more plausible38. In this

174 scenario, the more hydridic H is preferentially eliminated (Figure S19 of the SI). In addition, the coordination

W
175 of the amide to the copper(III) to form a metallocycle, favoring the elimination of an exocyclic δ-H rather

IE
176 than an endocyclic β-H39, could also contribute to the γ,δ-selectivity. Kochi and coworkers revealed that the

EV
177 ratio of elimination to substitution of organocopper(III) intermediates was found to be largely controlled by

PR
178 nature of the substrate (with substrates that can form more stable carbocations typically favoring substitution),

179 but also exhibited sensitivity to the reaction conditions30, particularly copper sources and solvents (more

LE
180 polar solvents favor oxidative substitution). These findings are in line with our experimental data. Of note,

C
181 C68-C73, C85, and C101 rule out the possibility that the lactone is formed via a cyclization of the
TI
182 dehydrogenated amide under acidic conditions (Figure S17 of the SI).
AR

183 Based on our experimental results and literature precedents, a proposed mechanism for this bimodal

184 dehydrogenation/lactonization reaction is outlined in Fig. 4c. An in-situ generation of Cu(I) initiates the
D

reaction by promoting the reductive cleavage of the N−O bond of N-methoxyamide I to form amidyl radical
E

185
AT

186 II and a Cu(II) species. Once formed, II undergoes 1,5-H atom abstraction to afford alkyl radical III, which

187 could recombine with Cu(II) to form alkylcopper(III) intermediate IV, analogous to the Kharasch allylic
ER

188 oxidation40. From this point forward, the reaction diverges into two paths. On the one hand, alkylcopper(III)
EL

189 species have significant carbocationic character. Hence, they can undergo oxidative elimination to generate
C

190 alkene V. On the other hand, in a more polar environment (with more polar solvent and a more acidic
AC

191 additive), the elimination may be suppressed, and instead, carbocation intermediate VI can form.

192 Intramolecular trapping of cation VI and subsequent iminium hydrolysis provides lactone VII. Remarkably,

13
193 for substrates where the carbocationic intermediate VI is relatively stabilized (e.g., when the γ position is

194 benzylic, tertiary, or allylic), the selectivity between the two pathways could be perfectly controlled.

W
IE
195 In conclusion, we have developed a bimodal Cu-catalyzed dehydrogenation/lactonization of

EV
196 synthetically common N-methoxyamides. This redox-neutral process led to two controllable reaction

197 pathways for synthesizing γ,δ-unsaturated primary amides and γ-lactones from various carboxylic acids (as

PR
198 the precursors of N-methoxyamides). Both the dehydrogenation and the lactonization could serve as

199 strategies for the diversifications of a variety of drug molecules and natural products. Further development

LE
200 of the current methodology for the synthesis of lactams is underway in our laboratory.

201
C
Acknowledgements: We acknowledge The Scripps Research Institute, NIH (NIGMS, 2R01GM084019) for
TI
202 financial support. The content is solely our responsibility and does not necessarily represent the official views
AR

203 of the National Institutes of Health. We thank D. Strassfeld for proofreading and providing helpful
D

204 suggestions in preparing the manuscript. We thank Z. Li and Y.-K. Lin for proofreading supplementary
E

205 information and repeating reactions. We thank M. Gembicky and J. Bailey of the UCSD Crystallography
AT

206 Facility for X-ray crystallographic analysis. We thank D.-H. Huang, L. Pasternack, and G. Kroon of the
ER

207 Nuclear Magnetic Resonance Facility of the Scripps Researcher Services for their assistance with NMR

208 analysis. We thank B. Webb and E. Billings of the Scripps Center for Metabolomics and Mass Spectrometry
EL

209 and Q. N. Wong of the Scripps Automated Synthesis Facility for assistance with mass spectrometry. This
C

210 paper is dedicated to the memory of the late Dr. Dong-Hui, Wang.
AC

14
211 Author contributions: J.-Q. Y. conceived the concept. S. Z. and Z.-J. Z. discovered and developed the

212 dehydrogenation/lactonization reaction. S. Z. and Z.-J. Z. conducted the mechanistic studies. S. Z., Z.-J. Z.

W
213 and J.-Q. Y. wrote the manuscript. J.-Q. Y. directed the project.

IE
214 Competing interests: J.-Q.Y., S. Z., and Z.-J. Z. are inventors on a patent application related to this work

EV
215 (US Patent application 63/605,065) filed by The Scripps Research Institute. The authors declare no other

PR
216 competing interests.

217 Supplementary Information is available in the online version of the paper.

LE
218 Data availability: Crystallographic data for compounds B63, B64, C27, C86, C101, and C102, as well as

C
219 for derivatives of B62 (labeled as B62-ketone) and B65 (labeled as B65-Ac) are available in the
TI
220 Supplementary Information files and from the Cambridge Crystallographic Data Center under reference
AR

221 numbers CCDC 2279927, CCDC 2271734, CCDC 2271733, CCDC 2271730, CCDC 2271732, CCDC

222 2271731, CCDC 2296322, and CCDC 2296327, respectively. All other data supporting the findings of this
D

223 study are available in the Article and its Supplementary Information files.
E
AT

224 References:

225 1. Larock, R. C. Comprehensive Organic Transformation (Wiley, 2018).

ER

226 2. Ishihara, Y. & Baran, P. S. Two-phase terpene total synthesis: historical perspective and application
227 to the Taxol® problem. Synlett 12, 1733−1745 (2010).
Qiu, Y. & Gao, S. Trends in applying C−H oxidation to the total synthesis of natural products. Nat.
EL

228 3.
229 Prod. Rep. 33, 562−581 (2016).
C

230 4. Buist, P. H. Fatty acid desaturases: selecting the dehydrogenation channel. Nat. Prod. Rep. 21,
231 249−262 (2004).
AC

232 5. Rettie, A. E., Rettenmeier, A. W., Howald, W. N. & Baillie, T. A. Cytochrome P-450-catalyzed
233 formation of Δ4-VPA, a toxic metabolite of valproic acid. Science 235, 890−893 (1987).

15
234 6. Zhou, J. et al. Spectroscopic studies of substrate interactions with clavaminate synthase 2, a
235 multifunctional a-KG-dependent non-heme iron enzyme: correlation with mechanisms and
236 reactivities. J. Am. Chem. Soc. 123, 7388−7398 (2001).

W
237 7. Cernak, T.; Dykstra, K. D.; Tyagarajan, S.; Vachal, P. & Krska, S. W. The medicinal chemist’s
238 toolbox for late stage functionalization of drug-like molecules. Chem. Soc. Rev. 45, 546−576 (2016).

IE
239 8. Blakemore, D. C. et al. Organic synthesis provides opportunities to transform drug discovery. Nat.
240 Chem. 10, 383−394 (2018).

EV
241 9. Hong, B.; Luo, T. & Lei, X. Late-stage diversification of natural products. ACS Cent. Sci. 6, 622−635
242 (2020).

PR
243 10. Breslow, R. & Baldwin, S. W. Conversion of cholestanol to 12-oxocholestanol and to cholest-14-enol
244 and-8(14)-enol by remote oxidation. J. Am. Chem. Soc. 92, 732−734 (1970).
245 11. Breslow, R. Biomimetic chemistry and artificial enzymes: catalysis by design. Acc. Chem. Res. 28,

LE
246 146−153 (1995).
247 12. Groves, J. T.; Nemo, T. E. & Myers, R. S. Hydroxylation and epoxidation catalyzed by iron-porphine
248
C
complexes. Oxygen transfer from iodosylbenzene. J. Am. Chem. Soc. 101, 1032−1033 (1979).
TI
249 13. Huang, X. & Groves, J. T. Beyond ferryl-mediated hydroxylation: 40 years of the rebound mechanism
250 and C−H activation. J. Biol. Inorg. Chem. 22, 185−207 (2017).
AR

251 14. Kim, C.; Dong, Y. & Que, L. Jr. Modeling nonheme diiron enzymes: hydrocarbon hydroxylation and
252 desaturation by a high-valent Fe2O2 diamond core. J. Am. Chem. Soc. 119, 3635−3636 (1997).
D

253 15. Hull, J. F. et al. Manganese catalysts for C−H activation: an experimental/ theoretical study identifies
254 the stereoelectronic factor that controls the switch between hydroxylation and desaturation pathways.
E

255 J. Am. Chem. Soc. 132, 7605−7616 (2010).

AT

256 16. Bigi, M. A.; Reed, S. A. & White, M. C. Diverting non-haem iron catalyzed aliphatic C−H
257 hydroxylations towards desaturations. Nat. Chem. 3, 216−222 (2011).
ER

258 17. Zhu, R.-Y.; Farmer, M. E., Chen, Y.-Q. & Yu, J.-Q. A simple and versatile amide directing group for
259 C−H functionalizations. Angew. Chem. Int. Ed. 55, 10578−10599 (2016).
260 18. Wang, D.-H.; Wasa, M. Giri, R. & Yu, J.-Q. Pd(II)-catalyzed cross-coupling of sp3 C−H bonds with
EL

261 sp2 and sp3 boronic acids using air as the oxidant. J. Am. Chem. Soc. 130, 7190−7191 (2008).
262 19. Zard, S. Z. Recent progress in the generation and use of nitrogen-centred radicals. Chem. Soc. Rev.
C

263 37, 1603 (2008).

AC

264 20. Fazekas, T. J.; Alty, J. W.; Neidhart, E. K.; Miller, A. S.; Leibfarth, F. A. & Alexanian, E. J.
265 Diversification of aliphatic C−H bonds in small molecules and polyolefins through radical chain
266 transfer. Science 375, 545−550 (2022).

16
267 21. Davies, J.; Svejstrup, T. D.; Fernandez Reina, D. F.; Sheikh, N. S. & Leonori, D. Visible-light-
268 mediated synthesis of amidyl radicals: transition-metal-free hydroamination and N-arylation
269 reactions. J. Am. Chem. Soc. 138, 8092−8095 (2016).

W
270 22. Davies, J.; Morcillo, S. P.; Douglas, J. J. & Leonori, D. Hydroxylamine derivatives as nitrogen-radical
271 precursors in visible-light photochemistry. Chem. - Eur. J. 24, 12154−12163 (2018).

IE
272 23. Jin, W. & Yu, S. Photoinduced and palladium-catalyzed remote desaturation of amide derivatives.
273 Org. Lett. 23, 6931−6935 (2021).

EV
274 24. Kwon, K.; Simons, R. T.; Nandakumar, M. & Roizen, J. L. Strategies to generate nitrogen-centered
275 radicals that may rely on photoredox catalysis: development in reaction methodology and applications

PR
276 in organic synthesis. Chem. Rev. 122, 2353−2428 (2022).
277 25. Bach, R. D. & Schlegel, H. B. The bond dissociation energy of the N−O bond. J. Phys. Chem. A 125,
278 5014−5021 (2021).

LE
279 26. Stateman, L. M.; Dare, R. M.; Paneque, A. N. & Nagib, D. A. Aza-heterocycles via copper-catalyzed,
280 remote C−H desaturation of amines. Chem 8, 210−224 (2022).
281 27.
C
Faulkner, A.; Race, N. J.; Scottb, J. S. & Bower, J. F. Copper catalyzed Heck-like cyclizations of
TI
282 oxime esters. Chem. Sci. 5, 2416−2421 (2014).
283 28. Kochi, J. K. The decomposition of peroxides catalyzed by copper compounds and the oxidation of
AR

284 alkyl radicals by cupric salts. J. Am. Chem. Soc. 85, 1958−1968 (1963).
285 29. Kochi, J. K. Mechanisms of organic oxidation and reduction by metal complexes: electron and ligand
286 transfer processes form the basis for redox reactions of radicals and metal species. Science, 155,
D

287 415−424 (1967).

E

288 30. Kochi, J. K.; Bemis, A. & Jenkins, C. L. Mechanism of electron transfer oxidation of alkyl radicals
AT

289 by copper(II) complexes. J. Am. Chem. Soc. 90, 4616−4625 (1968).

290 31. Barton, D. H. R.; Beckwith, A. L. J. & Goosen, A. Photochemical transformations. Part XVI. A novel
291 synthesis of lactones. J. Chem. Soc. 181−190 (1965).
ER

292 32. Neale, R. S.; Marcus, N. L. & Schepers, R. G. The chemistry of nitrogen radicals. IV. The
293 rearrangement of N-halamides and the synthesis of iminolactones1. J. Am. Chem. Soc. 88, 3051−3058
EL

294 (1966).
295 33. Chen, K.; Richter, J. M. & Baran, P. S. 1,3-Diol synthesis via controlled, radical-mediated C−H
C

296 functionalization. J. Am. Chem. Soc. 130, 7247−7249 (2008).

AC

297 34. Chen, K. & Baran, P. S. Total synthesis of eudesmane terpenes by site-selective C−H oxidations.
298 Nature, 459, 824−828 (2009).
299 35. Richers, J.; Heilmann, M.; Drees, M. & Tiefenbacher, K. Synthesis of lactones via C−H
300 functionalization of nonactivated C(sp3)−H bonds. Org. Lett. 18, 6472−6475 (2016).
17
301 36. Hoste, J. On a new copper specific group. Anal. Chim. Acta, 4, 23−27 (1950).
302 37. Ribas, X. et al. Aryl C−H activation by CuII to form an organometallic aryl-CuIII species: a novel
303 twist on copper disproportionation, Angew. Chem. Int. Ed. 41, 2991−2994 (2002).

W
304 38. Xu, J. et al. Copper-catalyzed trifluoromethylation of terminal alkenes through allylic C−H bond
305 activation. J. Am. Chem. Soc. 133, 15300−15303 (2011).

IE
306 39. Wu, X.; Riedel, J. & Dong, V. M. Transforming olefins into γ,δ-unsaturated nitriles through copper

EV
307 catalysis. Angew. Chem. Int. Ed. 56, 11589−11593 (2017).
308 40. Beckwith, A. L. J. & Zavitsas, A. A. Allylic oxidations by peroxy esters catalyzed by copper salts.

PR
309 The potential for stereoselective syntheses. J. Am. Chem. Soc. 108, 8230−8234 (1986).
310

311

LE
312 Fig. 1: Bimodal oxidation via a common radical intermediate. (a) Nature’s approach: hepatic P450-
313 mediated bimodal oxidation of valproic acid. (b) Our design of Cu-catalyzed bimodal oxidation of N-

C
314 methoxyamides via a common alkyl radical intermediate. (c) Synthetic approach: dual
315 desaturation/lactonization reactivity enables rapid diversification of carboxylic acids.
TI
316
AR

317 Fig. 2: Substrate scope for the dehydrogenation reaction. Reaction conditions: A (0.1 mmol), CuF2 (10
318 mol%), AcOH (8 eq.) or CSA (0.5 eq.), dioxane (0.50 mL), at 100-125 °C for 1-20 h (see the SI for details).
319 Isolated yields are reported. aL (20 mol%) was added. bThe solvent is DCE. cThe acid is TsOH•H2O (0.5 eq.).
D

320 d
[(MeCN)4Cu]BF4 (10 mol%) was used instead of CuF2. rr is the ratio of γ,δ-alkene/β,γ-alkene.
E

321
AT

322 Fig. 3: Substrate scope for the lactonization reaction. (a) Representative examples of bimodal
323 dehydrogenation/lactonization, see Figure S10 of the SI for details. (b) More substrates for the lactonization.
324 Reaction conditions: A (0.1 mmol), CuF2 (10 mol%) or [(CH3CN)4Cu]BF4 (10 mol%), acid (5 eq. TFA or
ER

325 0.5 eq. CSA), dioxane or AcOH (0.50 mL), at 125 °C for 1-20 h (see the SI for details). Isolated yields are
326 reported. aThe solvent is dioxane/MeNO2 (0.25 mL/0.25 mL). b0.5 eq. TsOH•H2O was used. d.r. =
327 diastereomer ratio, see the SI for details.
EL

328

329 Fig. 4: Reactions of diacid derivatives or using a low loading of copper and plausible mechanism. (a)
C

330 Simultaneous bimodal oxidations within complex systems. L: dehydrogenation-lactonization; R:

AC

331 dehydrogenation-lactamization. (b) Experiments using a low loading of copper. Reaction conditions for
332 dehydrogenation: A (0.1 mmol), catalyst (0.5-1 mol%), CSA (0.5 eq), dioxane (0.50 mL), 125 °C, 20-24 h.
333 Reaction conditions for lactonization: A (0.1 mmol), catalyst (0.5-1 mol%), TFA (5 eq), dioxane or AcOH

18
334 (0.50 mL), 125 °C, 24 h. See the SI for details. Isolated yields are reported. (c) Plausible mechanism based
335 on the mechanistic studies and literature precedents.

336

W
337 Extended Data Fig. 1: Continued Substrate scope for the dehydrogenation reaction. Reaction conditions:

IE
338 A (0.1 mmol), CuF2 (10 mol%), AcOH (8 eq.) or CSA (0.5 eq.), dioxane (0.50 mL), at 100-125 °C for 2-20
339 h (see the SI for details). Isolated yields are reported. aL (20 mol%) was added. bThe solvent is DCE. rr is the

EV
340 ratio of γ,δ-alkene/β,γ-alkene.
341

PR
342 Extended Data Fig. 2: Continued substrate scope for the lactonization reaction. Reaction conditions: A
343 (0.1 mmol), CuF2 (10 mol%) or [(CH3CN)4Cu]BF4 (10 mol%), TFA (5 eq.), dioxane (0.50 mL), at 125 °C
344 for 1-20 h (see the SI for details). Isolated yields are reported. aThe solvent is dioxane/MeNO2 (0.25 mL/0.25
345 mL). aThe acid is TsOH•H2O (1 eq.). b0.5 eq. CSA was used. c2.5 eq. TFA was used. d.r. = diastereomer

LE
346 ratio, see the SI for details.
347

348
C
Extended Data Fig. 3: Mechanistic studies. (a) Investigation of Cu(I) generated in-situ. (b) Radical clock
TI
349 experiment. (c) Inverted regioselectivity of elimination from a cationic intermediate. See the SI for details.
AR
E D
AT
ER
EL
C
AC

19
 a
dehydrogenation COOH

Me

hepatic P450
Me COOH Me 4 COOH VPA-4ene
O2

Me Me

W
valproic acid (VPA) common radical oxygenation Me COOH
intermediate
OH Me

IE
VPA-4ol
b
green internal oxidant

EV
R O no extra installation steps
O O
R' O
γ α OMe
H β N R NH2
H
H
III I II

PR
O-trapping Cu(I)-initiated then 1,5-H atom
path B amidyl radical abstraction -H elimination path A
of cation

Cu(II)
O recombination R O Cu(II) R O
& oxidation recombination
R'

LE
NH2 H NH2 H NH2
Cu(III)
VI common radical V
intermediate IV

C
c

O CuII CuI CuII CuI H2N O

O O
TI NHOMe Me
Me Me
Me
H
AR

Me H Me
MeOH+NH4+ Me MeOH

lactone N-methoxyamide desaturated primary amide

one step from carboxylic acid
open for further
diversification
√controllable bimodal oxidation √redox-neutral √inexpensive catalyst √high regioselectivity √good functional group tolerance
√readily available and stable substrates, >120 examples
E D
AT
ER
EL
C
AC
 R O CuF2 (10 mol%),
AcOH (8 equiv.) or CSA (0.5 eq.) O

H NHOMe N
DCE or dioxane (0.2 M), R NH2
H OMe L
A 100-125 °C, 1-20 h B

secondary
O Me NH2 O
Me Me O O
MeO NH2
NH2 O R
NH2 NH2 AcO NH2

W
OAc O
B4, R = tBu, 76%a,b O H2N O
B1, 80%a B2, 56%a B3,75%a,b B5, R = 1-admantyl, 72%a,b B6, 52%a,b B7, 78%a,b B8, 80%a,b

tertiary

IE
O H2N O
H2N O H2N O Me H2N O
Me Me O
NH2 O iPr

Me Me NH2

EV
AcO NH2
Me Me Me
iPr
B26, 60%c B27, 70%a B28, 72%a,b B29, 60%b B30, 66%b B31, 85%a B32, 62%b

natural products and drugs CONH2 Me

PR
Me Me
CONH2 CONH2 Me CONH2
MeO2C
Me
C12H25 Me Me O Me
CONH2 AcO
tBu
Me Me Me
Me Me CONH2
B51, 62%, 12:1 rr, 3:2 E/Z B52, 74%, 6:1 rr B53, 67% B54, 66%b B55, 53%b B56, 90%b
from stearic acid from valproic acid from camphoric acid from gemfibrozil from celestolide from santonin

LE
Me Me Me Me
Me
Me Me
Me Me CONH2 H Me
Me Me Me
O Me Me
H Me

C
Me H
H H
H H
CONH2 Me O O
Me CONH2 Me CONH2 Me CONH2
TI
B57, 80%a,b b
B58, 73% ; 1.6 g scale, 72% B59, 40% B60, 72% a,b B61, 84%
from isolongifolic acid from dehydroabietic acid from abietic acid from (+)-manoyl oxide from cholest-4-en-3-one
AR

Me CONH2 Me Me Me CONH2
Me
Me
O
Me Me H CONH2 Me Me Me Me Me H CONH2 Me Me Me

H Me H Me Me H Me
D

HO H HO H AcO H HO H
Me Me Me Me Me Me Me Me
E

B62, 79%d B63, 70% B64, 66% B65, 65%

from betulinic acid from 18 -glycyrrhetinic acid from oleanic acid from 18 -glycyrrhetinic acid
AT
ER
EL
C
AC
 [Cu(MeCN)4]BF4
or CuF2 (10 mol%),
O CSA (0.5 eq.) or TFA (5 equiv.) O O
R' O
R NHOMe + R NH2
dioxane or AcOH (0.2 M),
A H2O, 110-125 °C, 1-20 h C B

a switching from dehydrogenation to lactonization

O O
O O O O O O
Me O O Me O

W
Me O
Me
Me Me O Me
Me
OAc NPhth
Me O

IE
C27, 76% C28, 58% C29, 78% C35, 73% C37, 74% C38, 41% C48, 62%
B27, N.D. B28, trace B29, N.D. B35, N.D. B37, N.D. B38, N.D. B48, trace
Me
NPhth O O Me
H O Me

EV
O Me O
N O O O
O O
O Me Me Me Ph
O PhthN
O O
Me NPhth NPhth
S
O PhthN + O
Me OMe
O
C50, 69%b C10, 71% C11, 46% C39, 48%a, d.r. 4:5 C41, 25%a, d.r. 2:1 C41', 22%a, C42, 65%a, d.r. 3:2

PR
B50, N. D. B10, trace + B11, 29% + B41, 28% d.r. 12:1 B42, N. D.
+ B39, 28%

O O
O O
O O O
O + PhthN
Me PhthN Me
NPhth PhthN Me
O C14H29
O
O

LE
O
C43, 52%a, d.r. 1:1 C45, 46%a + B45, 23% C46, 27%a + B46, 23% C46', 19%a C51, 33%a + B51, 42% C52, 35%a, d.r. 1:1
+ B43, 29% + B52, 39%

b Me
O O O

C
O O O
H H
O O Cl N Me O Me
O O N Me H
Me
H C5H11
TI H
O Me
O
PhthN O Cl O Me Me O
Me
Ph
O
AR

C84, 50%a C85, 78%; C86, 47%a C92, 60% C93, 49% C94, 55% C95, 41%
only -lactone 1.3 g scale, 72% ixazomib skeleton from citronellal from (+)-3-carene
O
iPr
20 O O
OAc Me Me
O Me O X
O Me AcO
Me O Me
Me H O H
O O H
Me H
D

OMe O O H H
Me H H
Me H AcO H H
Me Me H MeO
O AcO
E

H from estrone
C96, 59% C97, 51% C98, 62%, d.r. 1:1 C99, 52%, d.r. 25:2 C100, 45%, d.r. 5:4 C101, X = O, 15%
from mycophenolic acid from dehydroabietic acid from manool from androsterone from deoxycholic acid + C102, X = NH, 56%
AT
ER
EL
C
AC
 a
O NHOMe O
O O H O
NHOMe N
Me Me
Me CuF2 (20 mol%), CuF2 (20 mol%),
Me Me Me
CSA (1.0 eq.) Me CSA (1.0 eq.)
Me H Me

dioxane, 125 °C H H DCE, 125 °C H H

46% 63%
H H O O O O
Me NHOMe Me

W
NH2
O NHOMe O NH2
O O
A102 BC102 A103 BC103

b Me c

IE
Me O
H Me O Cu(II)
Me O
Me Me

EV
Me disproportionation
H H H2N R O R
O O V MeO
CONH2 Cu(I) N H
C29, 76% H
B61, 82% H
0.5 mol% [(MeCN)4Cu]BF4 oxidative
1 mol% CuF2 I
elimination

PR
N–O cleavage
oxidative
Me
substitution
O
R'
Me Me H CONH2 H2N
O
O R VI
O Cu(II) R
H Me

LE
O H2N H
HO H iminium HN H
Me Me Cu(III) lactone
B62, 72% C37, 80% IV hydrolysis H
II
0.5 mol% [(MeCN)4Cu]BF4 1 mol% CuF2

C
Me Me R' O
O
Me O

O
TI VII
Me Me H CONH2 Cu(II)
recombination 1,5-H-abstraction

Me O Cu(II) R
AR

Me
AcO H H2N H
Me Me B64, 65% C73, 73%
1 mol% CuF2 1 mol% CuF2 III
E D
AT
ER
EL
C
AC
AC
C

Extended Data Fig. 1

EL
ER
AT
ED
AR
TI
C
LE
PR
EV
IE
W
AC
C

Extended Data Fig. 2

EL
ER
AT
ED
AR
TI
C
LE
PR
EV
IE
W
AC
C

Extended Data Fig. 3

EL
ER
AT
ED
AR
TI
C
LE
PR
EV
IE
W