The mechanism of photoinduced acylation of amines by N-acyl-5,7- dinitroindoline as determined by time-resolved infrared spectroscopy

Article abstract of DOI:10.1021/ol050670f

(Chemical Equation Presented) The photochemistry of N-acyl-5,7- dinitroindoline (1) was studied in acetonitrile using nanosecond time-resolved infrared (TRIR) spectroscopy. Upon photolysis, two nearly but not completely overlapping sets of transient IR ba

Full text of DOI:10.1021/ol050670f

ORGANIC
LETTERS
2
005
Vol. 7, No. 14
845-2848
The Mechanism of Photoinduced
Acylation of Amines by
2
N-Acyl-5,7-dinitroindoline as Determined
by Time-Resolved Infrared Spectroscopy
†
‡
,‡
Andrew D. Cohen, C e´ line Helgen, Christian G. Bochet,* and
John P. Toscano*,
†
Department of Chemistry, Johns Hopkins UniVersity, 3400 North Charles Street,
Baltimore, Maryland 21218, and Department of Chemistry, UniVersity of Fribourg,
CH-1700 Fribourg, Switzerland
christian.bochet@unifr.ch; jtoscano@jhu.edu
Received March 29, 2005
ABSTRACT
The photochemistry of N-acyl-5,7-dinitroindoline (1) was studied in acetonitrile using nanosecond time-resolved infrared (TRIR) spectroscopy.
Upon photolysis, two nearly but not completely overlapping sets of transient IR bands are observed that are assigned to two non-interconvertible
conformers of mixed acetic nitronic anhydride 7. While syn-7 reverts rapidly to 1, anti-7 is long-lived and is able to acylate amines. Results
of density functional theory calculations support conclusions based on experimental TRIR data.
Photolabile protecting groups and caged compounds have
been widely used in organic synthesis for nearly three
decades. Their attractiveness comes from the possibility of
controlling both the location and the time of release by
exposure to light alone. A large variety of such protecting
groups have been developed, each of them tailored for
specific uses. Among them, the N-acyl-7-nitroindolines are
of particular interest because their photolysis in the presence
of a nucleophile under both neutral and relatively mild
conditions leads to acylated products. This reaction was first
explored by Amit et al., who first used the 5-bromo-7-
indoline moiety as a photoactivatible leaving group to form
1
2
both acids and esters and later extended this methodology
3
to form amides. Helgen and Bochet have expanded the
synthetic utility of this reaction by photolyzing N-acyl-5,7-
dinitroindoline (1) in dry acetonitrile in the presence of a
variety of primary and secondary amines to form, along with
†
Johns Hopkins University
University of Fribourg
‡
(2) Amit, B.; Ben-Efraim, D. A.; Patchornik, A. J. Am. Chem. Soc. 1976,
(
1) (a) Bochet, C. G. J. Chem. Soc., Perkin Trans. 1 2002, 125-142.
b) Pelliccioli, A. P.; Wirz, J. Photochem. Photobiol. Sci. 2002, 1, 441-
58.
98, 843-844.
(
4
(3) Pass, S.; Amit, B.; Patchornik, A. J. Am. Chem. Soc. 1981, 103,
7674-7675.
1
0.1021/ol050670f CCC: $30.25
© 2005 American Chemical Society
Published on Web 06/10/2005

5
,7-dinitroindoline (2), mono- and disubstituted amides in
acetonitrile the major product is nitroindoline 5 (R ) CH
2
-
4
high yield and excellent purity (Scheme 1). Nicolaou et al.
CO CH ). The ratio of these products varies smoothly upon
2
3
varying the water content of acetonitrile. These and other
data suggest that both products could potentially derive from
one common acetic nitronic anhydride intermediate, the
product of acyl transfer to a nitro-oxygen (i.e., 6). Interest-
ingly, this initial step qualitatively resembles the hydrogen
transfer that initiates release of alcohols from nitrobenzyl
ethers and esters. Two different solvent-dependent reactions
are potentially available to an intermediate of structure 6; in
organic media, standard nucleophilic attack on the activated
carbonyl and protonation yields a nitroindoline 5, while in
aqueous media loss of acetate and deprotonation yields a
Scheme 1. Synthesis of Amides Using 1
10
have applied a similar synthetic strategy using a resin-bound
analogue to synthesize various amides with ease of purifica-
3
H-nitrosoindole, which tautomerizes to an isolable ni-
trosoindole (i.e., 4) (Scheme 2). Although the data are
compelling, nevertheless no direct structural evidence is
available concerning the nature of purported intermediate 6.
In this letter, we present a time-resolved infrared (TRIR)
spectroscopic study of N-acyl-5,7-dinitroindoline (1) in
acetonitrile that confirms the intermediacy of 6 (R ) NO ),
2
as well as computational data that also support this conclu-
sion.
The synthesis of 1 has been previously reported.11 Laser
photolysis (355 nm, 90 ns, 0.4 mJ) of 1 in both acetonitrile
3
and acetonitrile-d produces the TRIR difference spectrum
5
tion. More recently, this methodology has been applied to
6
the syntheses of N-glycosylated amino acids and various
7
carbamates. In aqueous media, though the aromatic byprod-
uct differs, carboxylic acids can be photochemically gener-
ated. Specifically, 7-nitroindoline derivatives have been used
as hydrolytically stable “caged” glutamate, GABA, and
glycine, releasing these neuroactive amino acids rapidly upon
8
photolysis.
Although the course of this reaction differs in aqueous
and organic media as evidenced by the different aromatic
products (Scheme 2), it nevertheless seems plausible that both
shown in Figure 1a. Instantaneous bleaching of the starting
material (negative bands) is accompanied by instantaneous
-
1
growth of transient IR bands at 1800, 1580, and 1160 cm .
The bands at 1800 and 1160 cm decay incompletely over
Scheme 2. Solvent-Dependent Reactivity of Purported
Intermediate 6
-
1
2
0 µs with concomitant and incomplete recovery of the bands
5
-1
due to the ground state (kobs ) 2.0 × 10 s ). Absorbance
-1
near 1800 cm is usually indicative of a carbonyl substituted
with a highly electron-withdrawing group. Upon closer
-1
examination of the absorption bands at 1800 and 1160 cm ,
it appears that two different intermediates (with distinct
reactivity) having similar IR spectra (i.e., similar structure)
are both formed within the laser pulse. Because IR bands
are typically broadened in solution relative to the gas phase,
we cannot discern two clear absorption maxima. We can,
however, distinguish between the two species by examining
their kinetic behavior at the two extremes of the observed
absorption bands. Specifically, the higher-energy edge of the
reactions proceed via a common intermediate. Wan, Corrie,
and co-workers have used a variety of steady-state and laser
flash photolytic techniques to study the mechanism of this
-
1
band at 1800 cm and the lower-energy edge of the 1160
-1
reaction in both water and wet acetonitrile using water-
cm band both exhibit a first-order decay over 20 µs, while
9
-1
soluble indoline 3 (R ) CH
2
CO
2
CH
3
) (Scheme 2). Along
the lower-energy edge of the band at 1800 cm and the
-
1
with acetic acid in both cases, in 100% water the major
higher-energy edge of the 1160 cm band are persistent for
more than 180 µs (Figure 1a). If a band is due to only one
species, then each edge of the band would display identical
kinetic behavior, such as is observed for the depletion band
product is nitrosoindole 4 (R ) CH CO CH ), while in wet
2
2
3
(
(
4) Helgen, C.; Bochet, C. G. Synlett 2001, 1968-1970.
5) Nicolaou, K. C.; Safina, B. S.; Winssinger, N. Synlett 2001, 900-
-
1
near 1300 cm . Interestingly, as the short-lived species
decays, the depletion bands due to 1 recover to a similar
extent (ca. 30%) and at the same rate, implying that the short-
lived intermediate does indeed revert to 1. The positive band
9
03.
6) (a)Vizvardi, K.; Kreutz, C.; Davis, A. S.; Lee, V. P.; Philmus, B. J.;
(
Simo, O.; Michael, K. Chem. Lett. 2003, 32, 348-349. (b) Simo, O.; Lee,
V. P.; Davis, A. S.; Kreutz, C.; Gross, P. H.; Jones, P. R.; Michael, K.
Carbohydr. Res. 2005, 340, 557-566.
(
7) Helgen, C.; Bochet, C. G. J. Org. Chem. 2003, 68, 2483-2486.
(8) (a) Canepari, M.; Papageorgiou, G.; Corrie, J. E. T.; Watkins, C.;
Ogden, D. J. Physiol. 2001, 533, 765-772. (b) Canepari, M.; Nelson, L.;
(10) (a) Corrie, J. E. T.; Barth, A.; Munasinghe, V. R. N.; Trentham, D.
R.; Hutter, M. C. J. Am. Chem. Soc. 2003, 125, 8546-8554. (b) Il’ichev,
Y. V.; Schworer, M. A.; Wirz, J. J. Am. Chem. Soc. 2004, 126, 4581.
(11) Mortensen, M. B.; Kamounah, F. S.; Christensen, J. B. Org. Prep.
Proc. Int. 1996, 28, 123-125.
Papageorgiou, G.; Corrie, J. E. T.; Ogden, D. J. Neurosci. Methods 2001,
1
12, 29-42. (c) Lowe, G. J. Neurophysiol. 2003, 90, 1737-1746.
9) Morrison, J.; Wan, P.; Corrie, J. E. T.; Papageorgiou, G. Photochem.
Photobiol. Sci. 2002, 1, 960-969.
(
2846
Org. Lett., Vol. 7, No. 14, 2005

Figure 1. (a) TRIR difference spectra averaged over the time scales indicated following laser photolysis (355 nm, 90 ns, 0.4 mJ) of a
solution of 1 (2 mM) in either acetonitrile (1850-1650 cm-1 and 1250-1100 cm ) or acetonitrile-d
-1
(1650-1250 cm ) and (b) TRIR
-1
3
12
difference spectra from panel a overlaid with bars representing calculated IR frequencies (scaled by 0.96 ) and relative intensities. Negative
signals are due to depletion of reactants, and positive signals are due to new transients or products.
at 1580 cm- does not display this behavior due to
1
are only 2.8 Å apart (well within the sum of the van der
Waals radii), consistent with this species easily reverting to
1. Interestingly, the relative position of analogous vibrational
modes calculated for syn-7 and anti-7 match well with the
expectation that the syn conformer will be short-lived and
the anti conformer will be persistent. This is most clearly
seen for the carbonyl stretch near 1800 cm , where the
calculations predict the syn conformer to have a carbonyl
stretching frequency that is higher energy than the anti
complicating overlap with depletion bands.
Based on these observations, we assign the positive
absorption bands to two non-interconvertible conformers of
the acetic nitronic anhydride (syn-7 and anti-7). These
kinetically distinct yet structurally similar intermediates
would be expected to display different reactivity, based on
the differing proximity of the acyl group to the indoline
nitrogen. To corroborate the assignment of the transient IR
bands, we performed geometry optimizations and frequency
calculations at the B3LYP/6-31G* level of theory on both
syn-7 and anti-7 (see Supporting Information for computa-
tional details). As an internal check of calculation accuracy,
we also performed geometry optimizations and frequency
calculations on indoline 1. All three structures optimized to
stable minima. The difference in energy between syn-7 and
anti-7 was ca1culated to be 1.7 kcal/mol favoring the former.
The calculated IR bands are overlayed with the TRIR
spectrum in Figure 1b. The calculated frequencies for both
conformers of 7 are similar to each other (as would be
expected based on their structural similarity) and match well
with the experimentally observed bands. In the calculated
geometry of syn-7, the acyl carbon and indoline nitrogen
-
1
-
1
conformer. Indeed, when we monitor the decay at 1810 cm
-
1
ca. 50% of the signal decays, whereas at 1785 cm , less
than 10% decays (Figure 2). Similar behavior is observed
for the band near 1160 cm . Again, this is in contrast to
the depletion bands between 1300 and 1400 cm , which
are completely symmetric.
-
1
-
1
We also examined the behavior of 7 in the presence of
added n-butylamine, to confirm that this species is the active
electrophile in the photochemical acylation reaction. Indeed,
the persistent band due to anti-7 decays only in the presence
of this added nucleophile and at an observed rate proportional
to the concentration of added amine. Additionally, we
observe the growth of the product (N-butylacetamide) at 1670
-1
cm at the same rate as the decay of anti-7 (data not shown).
Org. Lett., Vol. 7, No. 14, 2005
2847

rotations about the C7-N bond. The maximum energy of
these unoptimized structures lies at approximately 90°
rotation, as intuition predicts (see Supporting Information).
We then performed a series of single-point energy calcula-
tions at various C7-N bond lengths with the maximum
energy rotation. The minimum energy of these unoptimized
structures lies at approximately 1.43 Å (see Supporting
Information). These calculations provide us with an estimate
of the rotational barrier that is on the order of 42 kcal/mol.
Thus, it seems reasonable that syn-7 and anti-7 are indeed
independent entities with unique chemistry.
We have shown by TRIR spectroscopy that photolysis of
nitroindoline 1 initially yields two non-interconvertible
conformers of mixed acetic nitronic anhydride 7. One of
these conformers, syn-7, reverts to 1 rapidly, while the other,
anti-7 is much longer-lived. The subtle difference between
these conformers was detected by vibrational spectroscopy.
Thus, our TRIR results indicate that anti-7 is the potent
photochemically generated electrophile responsible for the
acylation of amines and alcohols discussed above. Appar-
ently, in an aqueous environment, in the absence of nucleo-
philes, an elimination pathway becomes favorable, resulting
Figure 2. Kinetic traces observed following laser photolysis (355
nm, 90 ns, 0.4 mJ) of a solution of 1 (2 mM) in acetonitrile.
A second-order rate constant of 1.1 × 10 M s-1 was
derived (Supporting Information). We were unable to add
enough n-butylamine to observe the reaction with syn-7
without obscuring the IR window.13 In addition, we were
4
-1
3
also unable to add enough methanol-d , to observe trapping
of anti-7 without obscuring the IR window.
14
in nitrosoindoles. In organic media such as methylene
The data suggest that syn-7 and anti-7 are non-intercon-
vertible species at room temperature. To confirm this, we
attempted to locate computationally a transition state cor-
responding to rotation about the C7-N bond. Although we
were unsuccessful, and therefore do not have an exact
calculation of the energetic barrier, we nevertheless have
suggestive evidence that this barrier is substantial. First, from
the optimized geometries (discussed above), the C7-N bond
length in 1 is 1.47 Å, but in both syn-7 and anti-7 the
analogous bond length is 1.34 Å, a substantial shortening
implying increased double-bond character. Second, we
performed a number of single-point energy calculations,
using anti-7 as a starting point, corresponding to various
chloride/dioxane mixtures or acetonitrile, this pathway is
unfavorable, and standard carbonyl addition/elimination
chemistry predominates.
Acknowledgment. J.P.T. thanks the National Science
Foundation (Grant CHE-0209350) for generous support of
this research. C.G.B. thanks the Swiss National Science
Foundation (Grant 2000-065047). A.D.C. acknowledges
support from an ACS Organic Division graduate fellowship.
C.H. acknowledges support from the Janggen-Poehn Stiftung.
Supporting Information Available: Experimental de-
tails, derivation of second-order rate constant, LFP data, and
computational details. This material is available free of
charge via the Internet at http://pubs.acs.org.
(12) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502-16513.
13) We also examined the photochemistry of 1 by laser flash photolysis
(
OL050670F
(
LFP) using UV detection. By LFP, syn-7 and anti-7 are indistinguishable,
so the transient absorption band was symmetric. Additionally, the derived
second-order rate constant was the same in both the LFP and TRIR
experiments (Supporting Information).
(14) Subtle structural changes can have an effect on the mechanism
because 5-nitro-7-nitrosoindolines were observed as products in the pho-
tolysis of carbamate derivatives of 5.7
2848
Org. Lett., Vol. 7, No. 14, 2005