Angewandte
Communications
Chemie
Table 1: Epoxide-opening ether cyclizations to form expanded and contracted
rings.[a]
cane 5aA was also observed as an important side
product in HFB 6, but not with AcOH. Cyclization of
6,7-epoxy alcohols 5b–5g without a cis alkene did
not occur (entries 8–13, Figure 2e).
7[b]
S[c] t [h][d] ht [%][e] A/B/O[f]
6[b]
H+[b]
auto[g] t [h][d] ht [%][e] A/B/O[f] ht [%][h]
Access to large rings, from 4 fA to 5aA, was
consistent with the working hypothesis of anion–p
templated entropy-centered ground-state destabili-
zation (Figure 2a); anti-Baldwin selectivity with
contributions from SN1-type behavior. The latter
was supported by allyl alcohol 5aO as a side product
(entry 7, Figure 2d) and 5bO and 5cO as the only
products obtained with 7 (entry 8, 9, Figure 2e). Poor
conversion of other epoxides 5d–5g confirmed the
decrease in reactivity with larger rings (entries 10–
13). Pinacol rearrangement into ketone 5dO’ as well
as the allyl alcohol 5dO is often considered as
evidence for trapped carbocation intermediates
(Figures 2e, S93–S95).
Moving from large to small rings, preorganiza-
tion with gem-dimethyl groups[22] in 3,4-epoxy alco-
hol 2a provided clean access to 2aA with all
catalysts, although anion–p catalyst 7 performed
better than the controls (entry 14, Figures 3b, S97–
S101). The templated[8] anti-Baldwin cyclization of
2b into trans-fused 2-oxabicyclo[3.3.0]octane 2bA
occurred only with 7, neither in HFB 6 nor with
AcOH (entry 15). In 2c, a reversal from 5-endo-tet to
4-exo-tet Baldwin selectivity was possible with 7
(entry 16, Figures 3a, S107–S111). With time, oxe-
tane 2cB expanded into oxolane 2cO (entry 17,
Figures S112–S116). The substitution pattern of 2cO
suggested that alcoholate–p interactions catalyze
oxetane opening (2cO°) followed by methyl migra-
tion (2cO°’) to the tertiary carbocation and ring
closure (2cO°’’, Figure 3a). This interpretation was
1
2
3
4a
4b
4c
280 79
210 81
140 100
160 100
0
230 100
160 100
400 100
120 100
260 100
120 100
0:100:0
0:100:0
24:76:0
0:100:0
–
52:9:39
59:0:41
0:0:100
0:0:100
mxt
mxt
–
–
100:0:0
100:0:0
0:92:8
0:0:100
0:100:0
mxt
–
+
+
+
–
–
+
–
–
–
–
–
140 89
170 59
170 48
210 81
0:100:0
0:100:0
11:89:0
89
66
53[i]
4[j] 4d
0:100:0 100
5
4e >200
–
160
–
0
–
–
0
0
0
0
0
0
0
0
0
6
7
8
9
4 f
5a
5b
5c
160 100
40:0:60
>200
0
0
0
0
–
–
–
–
–
–
–
–
>200
>200
>200
–
10 5d
11 5e
12 5 f >200
13 5g >200
14 2a
15 2b
16 2c
17 2c
18 2d
19 2d
20 2e >200
21 2 f >200
22 2g
23 2h >200
24 2i >200
25 2j
0
0
–
–
40 100
270 75
90 100
160 100
310 50
570 100
+
+
+
–
+
–
–
–
–
–
310 81
100:0:0 100
>200
0
–
0
–
0
–
–
0
–
–
0
–
–
–
–
–
–
–
–
–
–
–
0
–
160
–
160
–
45[k]
25[k]
0
0
–
–
mxt
–
–
mxt
0
0
0
0
0
0
–
40 100
>200
–
–
160
0
0
–
–
140 100
[a] Conditions: 1.0m S with i) 5 mol% 7 in CD2Cl2, ii) 6 as a solvent, or
iii) 1.0 equivalent AcOH (H+) in CD2Cl2, 408C; followed by 1H NMR spectroscopy.
[b] Catalysts, see Figure 1. [c] Substrates, see Figures 2, 3. [d] Reaction time.
[e] Conversion at given reaction time. [f] Product distribution: A=anti-Baldwin,
B=Baldwin, O=other product; for NMR spectra, see Figures S51–S119;
mxt=product mixtures. [g] Autocatalytic behavior noted (not studied). [h] Con-
version by 1 equivalent AcOH (H+) after at least 200 h. [i] A/B=11:89 after 200 h.
[j] Reaction run at room temperature. [k] A/B/O=0:100:0 after 300 h.
Baldwin selectivity also with AcOH controls, supported that
epoxide opening with alcoholate–p interactions could cause
a shift of reactivity from formal SN2- towards SN1-type
behavior.
supported by epoxide 2d with a gem-isopropyl alcohol
nucleophile. Formation of 2dB was the same as with 2cC
(entry 18, Figure S117). The following ring expansion, how-
ever, resulted in a complex mixture (entry 19), thus support-
ing the importance of methyl migration and tertiary carbo-
cations for the cascade transformation of 2c into 2cO
(entry 17).
Other substrates for cyclization into small rings were not
converted (2e[9]–2j, entries 20–25). Autocatalytic behavior,
assumed also for entropy-centered ground-state destabiliza-
tion (Figure 2a–c), was noticed for most reactions catalyzed
by anion–p catalyst 7, but never with conventional AcOH
catalysis (Table 1, Figure S2). Anion–p autocatalysis was,
however, not investigated further because it has already been
explored in detail for cyclization into standard oxetanes, in
experiment and theory.[10]
In summary, anion–p catalysis of epoxide-opening ether
cyclizations into rings of different size, from oxetanes to
oxocanes, provides the access to new reactivities that was
expected from catalysis with new interactions. The trans-fused
2-oxabicyclo[3.3.0]octane 2bA, the rearrangement of the
small oxetane 2cB into oxolane 2cO, and the large dioxe-
panes 4 fA and oxocane 5aA are all obtained only with
anion–p catalysis and not with Brønsted acid controls under
The insertion of a gem-dimethyl group in 4d to preor-
ganize cyclization[6,22] caused rate enhancements compared to
4a without changing selectivity for 4dB (entry 4). Replace-
ment of the gem-dimethyl motif by an oxygen atom to
inactivate the endo carbon relative to the exo carbon removed
all reactivity for 4e (entry 5, Figure 2d).[7] Considering the
efficient conversion with increasing anti-Baldwin selectivity
of 4c with anion–p catalyst 7 (entry 3), the exo carbon in 4e
was equipped with two methyl groups. Conversion of the
resulting 4 f with NDI 7 afforded the formal anti-Baldwin 1,4-
dioxepane 4 fA, together with allyl alcohol 4 fO and traces of
4 fB (entry 6; Figures S75–S84). The cyclization into the large
ring 4 fA was unique for strong anion–p catalysts, it occurred
neither in HFB 6 nor with AcOH.
In brevetoxin B, the hypothetical anti-Baldwin cyclization
into an oxacane is preorganized by a cis alkene (Fig-
ure 1a).[1,3] 6,7-Epoxy alcohol 5a with a cis alkene in position
2 was obtained by epoxidation of the monoterpene nerol
(Scheme S6). Reaction under standard conditions with 7, 5a
afforded oxocane 5aA (entry 7, Figures 2d, S85–S89). Oxo-
Angew. Chem. Int. Ed. 2020, 59, 1 – 6
ꢀ 2020 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!