To design an alternative protocol that could prevent the
hydrolysis side reaction anhydrous TBAF was prepared. The
recently reported procedure by DiMagno et al. was applied to
produce anhydrous TBAF (noted TBAF*).12 Using this activator
and the optimized reaction conditions for amines, the addition
of i-PrOH 11a was found to be slightly less selective since 5%
of type II ester 13a was detected by NMR (Table 4, entry 7),
but the ester 12a could be isolated with a good yield (69%)
and the formation of 7 was avoided. The same protocol was
applied to benzyl alcohol 11b, dodecyl alcohol 11c and solketal
11d which showed comparable reactivity and furnished the
corresponding esters 12b-d in good yields (Table 4, entries
8-10). In each case, the use of anhydrous TBAF* prevented
the hydrolysis side reaction leading to 7.
In regard of the mechanism of the reaction with TBAF, we
first envisaged the formation of an acid fluoride intermediate
by the regioselective addition of a fluoride anion on 1. With
the goal to detect such an intermediate, 1 equiv of 1 was mixed
with 2 equiv of TBAF in THF-d8 at -35 °C, and the resulting
solution was analyzed by 19F NMR (200 MHz) at low
temperature (-20, -35, and -78 °C); unfortunately, no peak
was observed that could be related to an acid fluoride (charac-
teristic chemical shift in the range 20-50 ppm).13 Thus, if an
acid fluoride is formed, this intermediate is probably unstable,
has a short lifetime, and reverts to the dilactone 1 in the absence
of a nucleophile. Interestingly, variations of the anion or the
cation of the activator alter a lot the regioselectivity of the
reaction. The larger and more polarizable the anion is, the lower
is the regioselectivity for the addition of 8a on 1 (Table 2, entries
16-19). On the other hand, changing tetrabutylammonium for
a smaller and more coordinating cation alters as well the
nucleophile addition regioselectivity (Table 2, entries 20-22).
It is noteworthy that in all experiments the total conversion of
1 into pulvinamides 9a and 10a was observed. It appears that
a high reactivity of the activator system leads to a high
regioselectivity. This observation, as well as our NMR experi-
ment, does not provide a clear explanation for the striking jump
in regioselectivity observed. Nevertheless, two hypotheses seem
to prevail: the direct reaction of fluoride on 1 that would lead
preferentially to a very reactive type I acid fluoride or the
activation of the nucleophile by fluoride acting as a strong
H-bond acceptor and able to modulate the outcome of the
reaction. The Table 4 results showing a variation in the
selectivity according to the alcohol used (entries 6-10) would
support the second hypothesis, whereas the Table 3 results
showing small selectivity variation when diverse amines were
used support either the first hypothesis or second hypothesis if
the H-bond acceptor effect is very high and equivalent for all
substrates.
In conclusion, the preparation of a monoaromatic pulvinic
dilactone 1 was described. It was found that TBAF was an
efficient activator for the regioselective ring opening of 1 by
amines and alcohols. Our methodology allowed the preparation
of 13 unsymmetrical and monoaromatic pulvinic acid derivatives
that are currently tested for their antioxidant capacities. The
study of the addition of other nucleophiles on 1 and the
preparation of type II monoaromatic derivatives (Figure 2b) are
in progress and will be reported in due course.
Experimental Section
General Procedure for the Regioselective Ring Opening of
Dilactone 1 with Amines. To a solution of dilactone 1 (1 equiv)
in 1.5 mL of THF was slowly added a solution of TBAF (1 M
solution in THF, 2 equiv) at -35 °C. The mixture was stirred for
15 min and then cooled to -78 °C. A solution of amine (1.3-2.5
equiv) in 1 mL of THF was added. After 15 min of stirring at -78
°C, the reaction was completed and the mixture was allowed to
warm to room temperature. After concentration in vacuo, the crude
residue was purified by chromatography on silica gel. Data for
9a: prepared from 1 (19.7 mg, 0.08 mmol) and 8a (9.7 µL, 0.10
mmol) and purified by flash chromatography (CH2Cl2/MeOH: 95/
5) as a yellow solid (31.0 mg, 94%): mp 86-87 °C; 1H NMR (300
MHz, CDCl3) δ 11.3 (brs, 1H), 8.04 (d, J ) 9.0 Hz, 2H), 6.93 (d,
J ) 9.0 Hz, 2H), 3.82 (s, 3H), 3.38 (t, J ) 7.6 Hz, 4H), 2.13 (s,
3H), 1.68-1.53 (m, 4H), 1.38-1.24 (m, 4H), 0.94 (t, J ) 7.1 Hz,
6H); 13C NMR (75 MHz, CDCl3) δ 172.1, 167.2, 162.0, 159.2,
159.1, 148.6, 129.1; 122.1, 114.3, 113.9, 104.2, 55.3, 47.6, 29.9,
20.1, 16.2, 13.8; IR 2959, 2932, 2871, 2837, 1759, 1604, 1537,
1432, 1249, 1094, 1072, 827 cm-1; HRMS (ES, pos) calcd for
C22H30NO5 [M + H]+ 388.2118, found 388.2144. Data for 9c:
prepared from 1 (20.2 mg, 0.08 mmol) and 8c (70 mg, 0.19 mmol)
and purified by flash chromatography (CH2Cl2/Et2O: 90/10) as a
yellow solid (38.2 mg, 80%): mp 48-49 °C; 1H NMR (300 MHz,
CDCl3) δ 11.26 (s, 1H), 8.05 (d, J ) 8.9 Hz, 2H), 6.94 (d, J ) 8.9
Hz, 2H), 3.83 (s, 3H), 3.38 (t, J ) 9.0 Hz, 4H), 2.17 (s, 3H),
1.62-1.60 (m, 4H), 1.29-1.24 (m, 36H), 0.88 (t, J ) 6.0 Hz, 6H);
13C NMR (75 MHz, CDCl3) δ 172.2, 167.2, 159.2, 159.1, 148.8,
129.1, 122.2, 114.3, 113.9, 104.2, 55.4, 47.9, 32.0, 29.8, 29.7, 29.5,
29.4, 27.9, 26.9, 22.8, 16.3, 14.3; IR 2921, 2850, 1769, 1468, 1454,
1250, 1182, 1079, 828 cm-1; HRMS (ES, pos) calcd for C38H62NO5
[M + H]+ 612.4623, found 612.4631.
Acknowledgment. This work was supported by the Centre
National de la Recherche Scientifique, Laboratoires Pierre Fabre
Dermo-Cosme´tique, and De´le´gation Ge´ne´rale pour l’Armement.
We are grateful to Cyril Antheaume for helpful NMR elucidations.
Supporting Information Available: General information,
analytical data, and 1H and 13C NMR spectra of all new
compounds. This material is available free of charge via the
(12) Sun, H.; DiMagno, S. G. J. Am. Chem. Soc. 2005, 127, 2050–2051.
(13) (a) Lu, C.; DesMarteau, D. D. J. Fluor. Chem. 2007, 128, 832–838. (b)
Cohen, O.; Sasson, R.; Rozen, S. J. Fluor. Chem. 2006, 127, 433–43.
JO801817S
J. Org. Chem. Vol. 73, No. 23, 2008 9493