In expression of our respect to Professor Barry M. Trost on the occasion of his 80th birthday.
Ring expansion of small-sized carbo- and heterocycles is among the most widely used transformations toward medium (or larger) carbo- and heterocycles.[1] Due to the presence of both a highly strained three-membered heterocycle and an activated double bond, vinyl aziridines are an extraordinary source of reactivity, and are therefore valuable building blocks in organic synthesis.[2]
[3]
[4] Consequently, we have been interested in exploring possibilities for performing ring-expansion processes with vinyl aziridines under mild and practical conditions toward novel useful building blocks. Moreover, N-heterocycles are highly valued and widely used in medicinal chemistry, revealing much promise for industrial purposes.[5]
Building on our previous work in the area of palladium-catalyzed [3+2]-cycloadditions of vinyl aziridines 1 with cyclic sulfoxyimines,[4] we were drawn to the potential use of silyl enol ethers 2 in the presence of Lewis acids to form the corresponding five-membered lactams 3 (Scheme [1]) bearing a quaternary center.
Scheme 1 Envisaged reactions of vinyl aziridines 1 with silyl enol ethers 2
When the reaction was attempted with model compound 1a and the silyl enol ether 2a in the presence of 5 mol% of Cu(OTf)2 at room temperature in dichloromethane, the expected compound 3a could not be observed in the crude material. Instead, along with the rather predictable elimination compounds 4a and 5a, pyrroline 6a was isolated from the reaction mixture in low yields (10–20%) (Scheme [2]).
Scheme 2 Attempted reaction of vinyl aziridine 1a with silyl enol ether 2a under copper catalysis
The copper-catalyzed ring expansion of vinyl aziridines to 3-pyrrolines is a known reaction, originally described by Njardarson and co-workers in 2008 (Scheme [3]).[6] However these reactions required the presence of an unusual copper complex, bis(hexafluoroacetylacetonato)copper [Cu(hfacac)2], under relatively harsh conditions (toluene, 150 °C).[7] As shown in Scheme [3], various substituted 3-pyrrolines could be efficiently obtained under these reaction conditions; the reaction mechanism was also unveiled.[6c]
Scheme 3 Some examples of vinyl aziridine rearrangements under Njardarson’s conditions[6a]
The very mild conditions (Scheme [2]) in which 6a was formed in our reaction prompted us to reinvestigate the reaction conditions (Table [1]) towards the exclusive, or at least, main formation of pyrroline 6a. Interestingly, in the absence of silyl enolate 2a, the pyrroline 6a was not observed (Table [1], entry 2). Suspecting an in situ Cu(II) → Cu(I) reduction in the reaction process,[8] we investigated the use of Cu(I) salts in the absence of silyl enolate 2a. Whereas various Cu(I) salts proved unsuccessful in catalyzing the reaction (entries 3–5), we were delighted to observe the clean formation of 6a in the presence of 10 mol% of (CuOTf)2·toluene (entry 6) at room temperature in THF.[9] To exclude the possibility of an organic acid-catalyzed transformation, we ran a trial in the presence of 5 mol% of TFA; in this case, no reaction was observed (entry 7).
Scheme 4 Synthesis of vinyl aziridines 1. For additional details, see Supporting Information.
Table 1 Pyrroline 6a Synthesis from Vinyl Aziridine 1a: Optimization Studya
|
Entry
|
Solvent
|
Cat. (5 mol%)
|
(4a + 5a)/6a
b
|
Yield (%) of 6a
|
|
1c
|
CH2Cl2
|
Cu(OTf)2
|
70:30
|
10
|
|
2
|
CH2Cl2
d
|
Cu(OTf)2
|
100:0
|
–
|
|
3
|
THF
|
[(i-Pr)CuCl]
|
NRe
|
–
|
|
4
|
THF
|
[(Ph3P)3CuF]
|
NR
|
–
|
|
5
|
THF
|
CuTCf
|
NR
|
–
|
|
6
|
THF
|
(CuOTf)2·toluene
|
<5: >95
|
60
|
|
7
|
THF
|
TFA
|
NR
|
–
|
a Reactions conditions: 1a (0.05 mmol), catalyst (5 mol%), solvent (0.5 mL), rt, argon atmosphere, 12 h.
b Determined by 1H NMR of the crude product.
c With silyl enolate 2a (1.5 equiv).
d Or THF.
e NR = no reaction.
f TC = thiophene-2-carboxylate.
After optimizing of the basic rearrangement reaction with vinyl aziridine 1a, we examined its general scope. For this purpose, a small collection of vinyl aziridines 1 were prepared by following a sequence involving an organocatalyzed aziridination/Wittig reaction (Scheme [4]).[6]
[10] The expected substrates needed for the ring-expansion transformation were obtained in moderate-to-good yields.
As shown in Scheme [5], various mono- and disubstituted pyrrolines 6 were isolated in good yields. In the presence of both R1 and R2 groups in the starting vinyl aziridine, pyrrolines 6a and 6b were obtained in yields of 60 and 33%, respectively. Notably, no reaction was observed when an electron-withdrawing group (R2 = CO2Et; compound 6c) was present, and for which Njardarson’s conditions are efficient.[6] From monosubstituted aziridines 1d–g, the corresponding pyrrolines 6d–g were obtained in yields of 60–94%. Of note, the presence of functionalized side-chains was well tolerated (6g and 6j). The rearrangement of vinyl aziridine 1d was also performed at a 1 mmol scale (as opposed to 0.14 mmol) to give the expected pyrroline 6d in a comparable isolated yield of 83%. The ring expansion of vinyl aziridines 1h–j with substituents at the R3 position was next investigated. In this case, pyrrolines 6h–j were formed in moderate to good yields of 40–57%. From these results, we can deduce that hindrance around the NTs functional group (the R2 and R3 groups) has a deleterious effect on the ring-expansion process, as the expected compounds were isolated in lowers yields (<60%). In these cases, the reactions were complete according to TLC monitoring, but were accompanied by the formation of unidentified byproducts. This trend was reinforced when substrate 1k was submitted to the ring-expansion conditions, as none of the trisubstituted pyrroline 6k was formed. In contrast, pyrrolines monosubstituted at R1 were obtained in excellent yields (≤94%). Interestingly, Cu(MeCN)4·PF6 could also be used as catalyst for the synthesis of 6b, 6d, and 6i, affording similar yields (32, 90, and 40%, respectively).
Scheme 5 Scope of the rearrangement of vinyl aziridines 1. Reagents and conditions: 1 (0.14 mmol), (CuOTf)2·toluene (5 mol%), THF (0.7 mL), rt, 2 h. Isolated yields after purification by silica gel column chromatography are reported. NR = no reaction. a Pyrroline 6b was, for an unknown reason, very difficult to purify.
Thanks to an easy access to 2-substituted vinyl cyclopropanes,[11] a related reaction of the vinylcyclopropane 7 was next attempted under the same reaction conditions, but with no success (Scheme [6]). The rearrangement of vinyl epoxides had been previously reported by Njardarson [Cu(hfacac)2 (5 mol%), toluene, 150 °C].[6d] We therefore attempted a rearrangement of vinyl epoxide 9 in the presence of (CuOTf)2 at room temperature, but the reaction led to an intractable mixture of products, and dihydropyran 10 could not be isolated.
Scheme 6 Attempted copper-catalyzed rearrangements of vinylcyclopropane 7 and epoxide 9
The mild conditions used in this transformation (THF, rt) prompted us to check the reaction stereospecificity. Starting from the enantiomerically enriched vinyl aziridine (Z)-1a (80% ee),[10] the corresponding pyrroline 6a was obtained in 62% yield in a racemic form (Scheme [7]).
Scheme 7 Rearrangement of enantiomerically enriched vinyl aziridine (Z)-1a
These results, including the racemization of enantiomerically enriched vinyl aziridines, are in full accordance with Njardarson’s observations. Mechanistic studies performed by Njardarson and co-workers highlighted a copper(I) insertion into the C–N bond to give an allylcopper intermediate, ultimately leading, to the pyrroline 6 after reductive elimination (Scheme [8]; left-hand pathway).[6c] An alternative mechanistic scenario (Scheme [8]; right-hand pathway) involving copper coordination and subsequent heterolytic rupture of the C–N bond, followed by the attack of the N-tosyl anion, cannot be totally excluded under our reaction conditions.
Scheme 8 Proposed mechanistic scenarios
These mechanistic insights prompted us to test the use of palladium salts in these ring-opening aziridine reactions. In a model reaction with (Z)-1a, we eventually found that, under aza-Wacker conditions,[12] the corresponding pyrrole 11 could be obtained in an unoptimized 50% yield (Scheme [9]).[13]
Scheme 9 Aza-Wacker-type rearrangement of (Z)-1a
In conclusion, we have developed a practical protocol[14] for the synthesis of 3-substituted pyrrolines from accessible vinyl aziridines. The ring-expansion process takes place under Cu(I) catalysis. In particular, mild conditions (THF, rt) are sufficient to accomplish the rearrangements. In the presence of R2 and R3 groups (steric hindrance around the NTs functionality), yields ranged from 33 to 60%, whereas 3-pyrrolines (R2 = R3 = H) were isolated in higher yields (>90%). Further investigations will be carried in an attempt to understand the correlation (if any) between conversion and steric encumbrance, and also to broaden the actual scope. The present procedure might be an efficient method for the synthesis of 3-pyrrolines, which are useful building blocks in both organic and medicinal chemistry. Moreover, the Wacker-type reactions disclosed are currently under investigation.
