Complex thermal behavior of 11-cis-retinal, the ligand of the visual pigments

Article abstract of DOI:10.1021/jo801899k

Upon heating to 80 °C, 11-cis-retinal yields a mixture of all-trans-retinal and 13-cis-retinal. This isomerization has been studied by means of density functional theory methods, and the computational results suggest a close competition between two mechanisms of very different nature. A classical internal rotation around the C11-C12 cis double bond, via a diradical transition state, accounts for the formation of the all-trans isomer. An intricate sequence of pericyclic reactions, namely a reversible [1,7]-H sigmatropic shift and a reversible 6-π-oxa-electrocyclic reaction, is responsible for the formation of 13-cis-retinal. Experiments using 11-cis-retinal labeled with deuterium at C19 confirmed the mechanistic proposal and also revealed an unprecedented outcome on the product composition of isotopologues.

Full text of DOI:10.1021/jo801899k

Complex Thermal Behavior of 11-cis-Retinal, the Ligand of the
Visual Pigments
´
Carlos Silva Lo´pez, Rosana Alvarez, Marta Dom´ınguez, Olalla Nieto Faza, and
´
Angel R. de Lera*
Departamento de Qu´ımica Orga´nica, Facultade de Qu´ımica, UniVersidade de Vigo, Lagoas Marcosende,
E-36310, Vigo, Spain
qolera@uVigo.es
ReceiVed August 31, 2008
Upon heating to 80 °C, 11-cis-retinal yields a mixture of all-trans-retinal and 13-cis-retinal. This
isomerization has been studied by means of density functional theory methods, and the computational
results suggest a close competition between two mechanisms of very different nature. A classical internal
rotation around the C11-C12 cis double bond, via a diradical transition state, accounts for the formation
of the all-trans isomer. An intricate sequence of pericyclic reactions, namely a reversible [1,7]-H
sigmatropic shift and a reversible 6-π-oxa-electrocyclic reaction, is responsible for the formation of 13-
cis-retinal. Experiments using 11-cis-retinal labeled with deuterium at C19 confirmed the mechanistic
proposal and also revealed an unprecedented outcome on the product composition of isotopologues.
Introduction
absorbed light energy), the distorted conformation of the
chromophore relaxes and generates the metarhodopsin II state
through a series of intermediates. The active Meta II binds and
transiently activates many copies of the G-protein transducin,
which modulates the transmembrane potential within the cell
and induces an optical nerve signal that culminates in visual
perception in the brain.3,4
The chromophore in rhodopsin and cone opsins acts in fact
as an inverse agonist since in the dark it suppresses the activity
of the receptor to an undetectable level. Crucial to the efficient
functioning of 11-cis-retinal 1 as an inverse agonist is its thermal
Rhodopsin, the representative member of the G-protein
coupled receptor superfamily (GPCR) in rods, and the cone
visual pigments in cone photoreceptors are transmembrane
protein complexes formed by the apoprotein (opsin and cone
opsins, respectively) and the ligand 11-cis-retinal 1, which is
covalently attached via a protonated Schiff base.1 Light absorp-
tion by the bound chromophore 11-cis-retinal initiates the visual
cycle in vertebrates. Nature has optimized this light-capturing
protein complex through evolution to the extent that the
photochemically induced isomerization of 11-cis-retinal is one
of the fastest and most efficient reactions recorded.1,2 Upon
photoisomerization to all-trans-retinal 3 in intermediate bathor-
hodopsin (a highly energetic species that stores 2/3 of the
(2) (a) Yan, E. C. Y.; Ganim, Z.; Kazmi, M. A.; Chang, B. S. W.; Sakmar,
T. P.; Mathies, R. A. Biochemistry 2004, 43, 10867–10876. (b) Kukura, P.;
McCamant, D. W.; Yoon, S.; Wandschneider, D. B.; Mathies, R. A. Science
2005, 310, 1006–1009.
(3) (a) Rando, R. R. Chem. ReV. 2001, 101, 1881–1896. (b) Fishkin, N.;
Berova, N.; Nakanishi, K. Chem. Rec. 2004, 4, 120–135.
(4) (a) McBee, J. K.; Palczewski, K.; Baehr, W.; Pepperberg, D. R. Prog.
Ret. Eye Res. 2001, 20, 469–529. (b) Ridge, K. D.; Palczewski, K. J. Biol. Chem.
2007, 282, 9297–9301.
* To whom correspondence should be addressed. Fax: 34986811940.
(1) De Grip, W. J.; Rothschild, K. J. In Molecular Mechanisms of Visual
Transduction; Stavenga, D. G., De Grip, W. J., Pugh, E. N., Jr., Eds.; Elsevier:
Amsterdam, 2000; p 1.
10.1021/jo801899k CCC: $40.75
Published on Web 12/24/2008
2009 American Chemical Society
J. Org. Chem. 2009, 74, 1007–1013 1007

Lo´pez et al.
FIGURE 1. 1H NMR spectra recorded over time for the thermal isomerization of 11-cis-retinal 1 at 80 °C (C6D6). Characteristic 1H NMR chemical
shifts of the aldehyde group were used to track the reactant and products over time.
stability. The protein-ligand complex shows a very low level
of spontaneous thermal isomerization, a process that would
produce the agonist-activated protein complex. Previous reports
had indicated photoreceptor noise production due to thermal
isomerization of the unprotonated Schiff base.5 More recently,
a thermal denaturation study of rhodopsin at 55 °C showed that
the protein-ligand complex is more thermolabile than the free
chromophore.6
The first report on the thermal isomerization of 11-cis-retinal
1 in solution was due to Hubbard in 1966.7 A rate constant of
k ) 1.02 × 10-5 s-1 was determined at 80 °C, and Arrhenius
parameters of A ) 1 × 1011 s-1 and ∆Gq ) 26.2 kcal/mol were
measured in n-heptane solutions. UV monitoring led to the
proposal that the product was the all-trans-retinal isomer. This
reaction has not been revisited during the 40 year period of
intense and stimulating research in the field of vision in
vertebrates and contrasts with the comprehensive experimental8,1
and theoretical9 studies on the alternative photoisomerization
event.
We hereby provide a combined experimental-theoretical
(density functional theory, DFT) study of this thermal isomer-
ization which proves that all-trans-retinal 3 is not the unique
product of the reaction but is instead the minor component
(∼35%) of a mixture with the predominant 13-cis-retinal 2
(∼65%). Moreover, we show that the mechanistic pathway
yielding the latter product most intriguingly involves a cascade
of pericyclic reactions completely avoiding any diradical species
along the course of the formal double-bond isomerization steps.
Results and Discussion
Upon heating to 80 °C in C6D6, 11-cis-retinal 1 experienced
an unimolecular isomerization process affording signals in the
1H NMR spectra corresponding to both all-trans-retinal 3 and
13-cis-retinal 2 (see Figure 1). These isomers were identified
by comparison of their 1H NMR data with those of commercial
samples and by HPLC coelution with standards. Reproducible
(5) Barlow, R.; Birge, R.; Kaplan, E.; Tallent, J. Nature 1993, 366, 64–66.
(6) Valle, L. J. D.; Ramon, E.; Bosch, L.; Manyosa, J.; Garriga, P. Cell.
Mol. Life Sci. 2003, 60, 2532–2537, The authors pointed out that the thermal
instability of rhodopsin at 55 °C may suggest a connection between some retinal
diseases and mutations in the apoprotein that mimic the destabilizing effect of
the temperature.
(7) Hubbard, R. J. Biol. Chem. 1966, 241, 1814–1818.
(8) Liu, R. S. H.; Asato, A. E. Tetrahedron 1984, 40, 1931–1969.
(9) Gasco´n, J. A.; Sproviero, E. M.; Batista, V. S. Acc. Chem. Res. 2006,
39, 184–193.
1008 J. Org. Chem. Vol. 74, No. 3, 2009

Complex Thermal BehaVior of 11-cis-Retinal
FIGURE 2. Isomerization pathways of 11-cis-retinal 1 to all-trans-retinal 3 and 13-cis-retinal 2 explored in this work. All the free energies (in
kcal/mol) are relative to 11-cis-retinal 1.
FIGURE 3. Sequential (and torquoselective) four-electron ring-closure/ring-opening process for the isomerization of both double bonds of the
terminal dienal fragment of 11-cis-retinal 1 (see Figure 2, path D).
first-order kinetics (k ) 1.02 × 10-5 s-1 and t1/2 of 22 h) were
experimentally determined in C6D6 at 80 °C. The mixture of
isomers obtained after subjecting 11-cis-retinal 1 to thermal
conditions (see Figure 1) contrasts with the equilibrium com-
position found for retinal isomers obtained with CF3COOH
(61.8:23.9:10.9:4.0:0.1 all-trans/13-cis/9-cis/9,13-di-cis/11-cis
from all-trans-retinal)10 and also with the photostationary
mixture of retinal in hexane solution (54:5:41 all-trans/9-cis/
13-cis from all-trans-retinal).8 It is, however, very similar to
that obtained upon treatment of 11-cis-retinal 1 with iodine at
40 °C (65.8:31.6:2.6 all-trans/13-cis/11-cis).10
experimental data.7 Given the feasibility of the direct olefin
isomerization process, we then envisioned that all-trans-retinal
3 could undergo further E/Z isomerization at C13-C14 to yield
13-cis-retinal 2. Control experiments, however, indicated no
significant interconversion between the final isomers upon
heating isolated all-trans-retinal 3 or 13-cis-retinal 2 at the same
temperature for extended reaction times.12 Therefore, the
pathways for formation of the single- (trans-) and double-
isomerization (13-cis) retinal isomers do not appear to converge
(10) (a) Sperling, W.; Carl, P.; Rafferty, Ch. N.; Dencher, N. A. Biophys.
Struct. Mech. 1977, 3, 79–94. (b) Rando, R. R.; Chang, A. J. Am. Chem. Soc.
1983, 105, 2879–2882.
(11) Bonacic-Koutecky`, V.; Koutecky`, J.; Michl, J. Angew. Chem., Int. Ed.
1987, 26, 170–189.
(12) Some thermal experiments showed that 13-cis-retinal 2 produced ca.
10% all-trans-retinal 3 upon prolonged heating even in carefully deactivated
glassware. But under these conditions, the 13-cis 2 and all-trans 3 ratio never
reach the ca. 65:35 ratio found in the thermal isomerization of 11-cis-retinal 1.
The proposed 11-cis 1 to all-trans-retinal 3 isomerization
might involve a simple internal rotation about the C11-C12
double bond via a diradical transition state (Figure 3, path A).11
Our DFT calculations provided an activation barrier (∆Gq )
27.3 kcal/mol) for this process in close agreement with the
J. Org. Chem. Vol. 74, No. 3, 2009 1009

Lo´pez et al.
at a common intermediate. This requirement could however be
met if the order of the double-bond isomerization steps was
reversed: first, a 13E to 13Z rotation yielding 11,13-di-cis-retinal
4 and then an 11Z to 11E rotation to obtain 2, thus avoiding
all-trans-retinal 3 in the reaction path. The computed activation
energy for the first E to Z isomerization (see reaction path B in
Figure 2) was found to be 35.11 kcal/mol, rendering this
possibility out of competition (compared with 27.3 kcal/mol).
The increase in activation energy can be attributed to the drastic
reduction in conjugation for the functionalized radical moiety
in the transition state (see Figure 2, path B).13
Since double-bond isomerization via diradical pathways is
clearly out of the energy range for competition at the working
temperatures, nonradical alternatives for the rearrangement of
1 to 2 had to be surveyed. A considerable number of pericyclic
reactions are compatible with the high degree of unsaturation,
the cis geometry, the allylic methyl groups, and the conjugated
aldehyde of 11-cis-retinal 1.14 This pentaenal structure can
accommodate, in addition to the already explored diradical
structures, a wide variety of pericyclic transition states (for
example, electrocyclic, heteroelectrocyclic, sigmatropic reac-
tions, etc.) that could operate in domino sequences to afford
different isomers of 11-cis-retinal 1. Pericyclic processes
involving derivatives of other retinal isomers are known. An
electrocyclic reaction of a formal dieniminium ion takes place
at physiological temperature in the retina during the formation
of the A2E pigment (a fluorescent amphiphilic pyridinium bis-
retinoid involved in age-related macular degeneration) from two
molecules of trans-retinal.15,16 Furthermore, 13-cis-retinal Schiff
bases afford at ambient temperature the corresponding dihy-
dropyridines by heteroelectrocyclic ring closure.17
pentaenal structure of retinal, including six- and four-electron
electrocyclizations, six- and four-electron heteroelectrocyclic
reactions, and [1,7] sigmatropic hydrogen shifts. We envisioned
four different pericyclic cascades that could afford 13-cis-retinal
2 from 11-cis-retinal 1 (see Figure 2). They are labeled C, D,
E, and F and characterized by the increasing number of electrons
involved in the pericyclic reaction initiating the cascade (4 in
C and D, 6 in E, and 8 in F).
The reaction path C implies a four-electron heteroelectrocyclic
ring closure at the carbonyl end of the polyenal to afford 2H-
oxete derivative 5, followed by the four-electron ring opening
of this ring in the same rotatory direction yielding 11,13-di-
cis-retinal 4.18 11,13-Di-cis-retinal 4 would then undergo a six-
electron heteroelectrocyclic ring closure to form 2H-pyran 6,19
finally converted to 13-cis-retinal by ring opening.20
Reaction path D represents perhaps the shortest pericyclic
pathway since a combined four-electron ring-closure/ring-
opening process in the same rotatory directionality (see Figure
3) has the overall effect of isomerizing both double bonds of
the involved diene. The directionality of the conrotation would
moreover be enforced by the strong CHO-in torquoselectivity
inherent to the ring opening of 3-formylcyclobutenes.21
Pathway E implies two pairs of ring-closing/opening elec-
trocyclizations. A six-electron ring closure (1 to 8),14a involving
the C9-C14 triene, followed by the corresponding ring opening
of 8 in the same rotatory direction to furnish 9,11,13-tri-cis-
retinal 9, for which conversion to 13-cis-retinal can be attained
by double isomerization of the cis double bonds at C9 and C11
through a four-electron electrocyclic ring closing/opening
sequence as illustrated above.
These three different mechanistic options C-E show rate-
limiting steps with activation energies much higher than 27 kcal/
mol (see values in Figure 2), thus rendering them noncompetitive
with path A.
Based on these precedents, we systematically considered
typical low energy unimolecular pericyclic reactions14 for the
(13) In the restricted environment of the opsin protein (as has been found in
bacteriorhodopsin, the proton pumping device of halobacteria, which uses a
protonated Schiff base of all-trans-retinal as chromophore) or in the solid state,
retinoids have been reported to undergo double-bond isomerizations via concerted
motions (hula twist and bicycle pedal); see: (a) Liu, R. S. H.; Browne, D. T.
Acc. Chem. Res. 1986, 19, 42–48. (b) Baudry, J.; Crouzy, S.; Roux, B.; Smith,
J. C. Biophys. J. 1999, 76, 1909–1917. The same studies accept that, in solution,
the most favored isomerization mechanism is the sequential pathway, where each
double-bond rotation occurs as a single and independent step. We thoroughly
explored this possibility and reached similar conclusions: the concerted isomer-
ization about C11-C12 and C13-C14 lies very high in energy (ca. 48.6 kcal/
mol) and the stationary point associated with such a transformation was found
to be a second-order saddle point (see the Supporting Information). Other
theoretical studies rule out the bicycle pedal motion of the photoexcited
chromophore, which moreover appears to restrict its movement to the C10-
C11-C12-C13 fragment (a so-called “photochemical hot spot”), whereas the
rest of the molecule begins to rotate only upon relaxation to the ground state;
see: (c) Weingart, O. J. Am. Chem. Soc. 2007, 129, 10618–10619.
(14) (a) Okamura, W. H.; de Lera, A. R. In ComprehensiVe Organic
Synthesis; Trost, B. M., Fleming, I., Paquette, L. A.,Volume Ed.; Pergamon Press:
London, 1991; Vol. 5, Chapter 6.2, pp 699-750. (b) Marvell, E. N. Thermal
Electrocyclic Reactions; Academic Press: New York, 1980. (c) Burnier, J. S.;
Jorgensen, W. L. J. Org. Chem. 1984, 49, 3001–3020.
(15) (a) Sakai, N.; Decatur, J.; Nakanishi, K.; Eldred, G. E. J. Am. Chem.
Soc. 1996, 118, 1559–1560. (b) Parish, C. A.; Hashimoto, M.; Nakanishi, K.;
Dillon, J.; Sparrow, J. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 14609–14613. (c)
Ben-Shabat, S.; Parish, C. A.; Hashimoto, M.; Liu, J.; Nakanishi, K.; Sparrow,
J. R. Bioorg. Med. Chem. Lett. 2001, 11, 1533–1540. (d) Fishkin, N.; Jang, Y.-
P.; Itagaki, Y.; Sparrow, J. R.; Nakanishi, K. Org. Biomol. Chem. 2003, 1, 1101–
1105. For the synthesis of A2E model systems involving an aza-6π-
electrocyclization of Schiff bases, see: (e) Tanaka, K.; Katsumura, S. Org. Lett.
2000, 2, 373–375.
(16) A dimerization mode of retinal alternative to that leading to A2E involves
also a pericyclic reaction and affords a pigment epithelial cell fluorophore: Fishkin,
N.; Sparrow, J. R.; Allikmets, R.; Nakanishi, K. Proc. Natl. Acad. Sci.U.S.A.
2005, 102, 7091–7096.
A mechanistic alternative (labeled as F in Figure 2) for the
isomerization of the terminal double bond in conjugated polyene
1 would be a reversible antarafacial [1,7]-H sigmatropic shift22
(1 f 11 f 4) between the methyl group at C9 and H at C14.
The migration of diastereotopic hydrogens at C14 in 10-cis,12-
cis-19,14-retro-retinal 11 or the back migration of the same atom
after helix inversion would both effect the isomerization of the
C13-C14 double bond of 1. This step would be followed by
(18) The ring-closure activation energy computed (49.77 kcal/mol) is in good
agreement with previous computations of analogous systems at the MP4SDTQ/
6-31G(d) level; see: Yu, H.; Chan, W. T.; Goddard, J. D. J. Am. Chem. Soc.
1990, 112, 7529–7537.
(19) The mechanism of isomerization of di-cis-dienals and dienones to the
monocis isomers through formation of R-pyrans had been previously studied by
Kluge and Lillya: (a) Kluge, A. F.; Lillya, C. P. J. Org. Chem. 1971, 36, 1977–
1988. (b) Kluge, A. F.; Lillya, C. P. J. Org. Chem. 1971, 36, 1988–1995.
(20) (a) 11,13-Di-cis-retinal was first shown to convert to 13-cis-retinal by
Wald et al. in 1955: Wald, G.; Hubbard, R.; Brown, P. K.; Oroshnik, W. Proc.
Natl. Acad. Sci. U.S.A. 1955, 41, 438–451. (b) The isomerization of the 11-cis
double bond of 11,13-di-cis-retinal and 9,11,13-tri-cis-retinal was studied by
Okamura: Knudsen, C. G.; Chandraratna, R. A. S.; Walkeapaa, L. P.; Chauhan,
Y. S.; Carey, S. C.; Cooper, T. M.; Birge, R. R.; Okamura, W. H. J. Am. Chem.
Soc. 1983, 105, 1626–1631. (c) In addition, Liu described the thermal
rearrangement of the four labile isomers of retinal (11,13-di-cis-, 7,11,13-tri-
cis-, 9,11,13-tri-cis-, and all-cis-retinal) presumably occurring via consecutive
6πe--heteroelectrocyclization reactions and determined the kinetic parameters
of these processes: Zhu, Y.; Ganapathy, S.; Liu, R. S. H. J. Org. Chem. 1992,
57, 1110–1113.
(21) Dolbier, W. R., Jr.; Koroniak, H.; Houk, K. N.; Sheu, C. Acc. Chem.
Res. 1996, 29, 471–477.
(22) For the first demonstration of the antarafacial nature of a [1,7]-H
hydrogen shift in the previtamin D series, see: (a) Hoeger, C. A.; Okamura,
W. H. J. Am. Chem. Soc. 1985, 107, 268–269. (b) Hoeger, C. A.; Johnston,
A. D.; Okamura, W. H. J. Am. Chem. Soc. 1987, 109, 4690–4698.
(17) (a) Okamura, W. H.; Lera, A. R. D.; Reischl, W. J. Am. Chem. Soc.
1988, 110, 4462–4464. (b) de Lera, A.; Reischl, W.; Okamura, W. J. Am. Chem.
Soc. 1989, 111, 4051–4063.
1010 J. Org. Chem. Vol. 74, No. 3, 2009

Complex Thermal BehaVior of 11-cis-Retinal
FIGURE 4. 1H NMR spectra of the aldehyde region recorded over time for the thermal isomerization of 11-cis-retinal-19,19,19-d3 at 80 °C (C6D6).
consecutive six-electron heteroelectrocyclic reactions19 of 11,13-
di-cis-retinal 4 through the corresponding 2H-pyran 6 to the
final 13-cis-retinal 2 as indicated above. Examination of the
activation energies for the sequence reveals that this alternative
pathway to 13-cis-retinal from 11-cis-retinal is highly competi-
tive (a rate-limiting step of 28.70 kcal/mol for the second [1,7]-
H sigmatropic migration) with that to the all-trans isomer
described before (27.27 kcal/mol).
The computations led us to propose that the thermal isomer-
ization of 11-cis-retinal 1 yields all-trans-retinal 3 via internal
rotation around the double bond at C11-C12 involving a
diradical transition structure and 13-cis-retinal 2 via [1,7]-H
sigmatropic shifts between the methyl group at C9 and H at
C14, followed by a six-electron electrocyclic ring closure of
the resulting 11,13-di-cis-retinal 4 and the reverse ring opening
of the corresponding 2H-pyran 6.20
The implication of a [1,7]-H sigmatropic migration reaction
in the thermal isomerization of 11-cis-retinal was empirically
investigated through deuterium-labeling experiments. If, as
predicted by our calculations, such rearrangement is operating,
the introduction of deuterium at the C9-CH3 position should
provide experimental evidence of the hydrogen migration. 11-
cis-Retinal-19,19,19-d3 1-d3 was prepared and subjected to the
thermal conditions (80 °C in C6D6) and the reaction monitored
by 1H NMR. Two relevant observations could be derived from
this study: first, as expected, reaction times increased consider-
ably with respect to those of the protium isotopologue due to
the primary isotope effect involved in the sigmatropic shift step
(cf. Figures 1 and 4); second, the product distribution is inverted
in the thermal isomerization of 1-d3 when compared to that of
1 (see Figure 4). To our knowledge, such extreme reversal in
product distribution upon isotope labeling is unprecedented. This
observation could only be explained by cooperatively combining
our experimental and computational results.
In the parent system 1, a product distribution of 65/35 under
kinetic control at 80 °C implies23 that the activation energy of
the [1,7]-H sigmatropic shift step is just 0.45 kcal/mol lower
than the activation of the double bond for the isomerization
along the diradical route A. Additionally, we computed through
harmonic and thermochemical analysis a large value of kH/kD)
5.5 for the initial [1,7]-H sigmatropic shift. This KIE is in very
good agreement with the values found by Okamura et al. in the
vitamin D field.24 The computed KIE of 5.5 is due to a 1.2
kcal/mol increase in the activation energy of the [1,7]-H
(23) Using the Maxwell-Boltzmann distribution.
J. Org. Chem. Vol. 74, No. 3, 2009 1011

Lo´pez et al.
FIGURE 5. Mechanistic pathway for the formation of several isotopomers of 11-cis-retinal 1 and 13-cis-retinal 2 through [1,7]-H/D sigmatropic
shifts.
sigmatropic migration step. Remarkably, this computed energy
difference not only is sufficient to invert the balance of the
activation energies for the competing processes but it also
justifies the observed inversion in product distribution for the
thermal isomerization of the deuterated substrate 1-d3. Thus,
the 2/3 65/35 product distribution observed in the thermal
isomerization of 11-cis-retinal together with the 2-d3:3-d3 30/
FIGURE 6. Thermal rearrangement of 11-cis-11-fuororetinal 12.
70 distribution for the thermal isomerization of the 1-d3
isotopologue is also an indication of a ca. 1.2 kcal/mol energy
change of the corresponding barriers, approximately. This
difference is in very good agreement with our computed KIE
subjected to thermal isomerization (Figure 6). The reaction
proceeds very rapidly compared to the parent system, undergo-
ing complete conversion at only 40 °C in 4.5 h (see the
Supporting Information). This thermal rearrangement yielded
for the [1,7]-H migration reaction. Furthermore, we also
all-trans-11-fluororetinal 13 as exclusive product, which is in
computed the expected secondary KIE values for the diradical
good agreement with the hypothesis based on the mechanistic
transition state in path A and found this effect to be almost
routes proposed for the thermal isomerization of 11-cis-retinal
negligible (kH/kD ) 1.05). In conclusion, our experimental and
1.
computational data provide consistent evidence suggesting that
the dramatic changes in product distribution observed upon
Conclusions
deuterium labeling arise entirely from the primary KIE affecting
the [1,7]-H sigmatropic shift.
The thermal isomerization of 11-cis-retinal has been shown
to afford a mixture of double-bond isomers, namely 13-cis- and
all-trans-retinal in a 65:35 ratio, approximately. On the one
hand, the formation of all-trans-retinal occurs through direct
internal rotation around the double bond at C11 via a diradical
transition state. On the other hand, the formation of 13-cis-retinal
occurs through a complex pericyclic mechanism involving a
pair of [1,7]-H sigmatropic shifts followed by a six electrons
heteroelectrocyclic ring closure/opening process. Increase of the
activation energy for the latter and change of product distribution
was enforced by deuterium labeling of the CH3 group at C9,
which moreover proved the occurrence of a [1,7]-H sigmatropic
shift in the vitamin A field.
The reversibility of the [1,7]-H sigmatropic shifts together
with the presence of deuterium labeling promotes H/D ex-
changes that produce a complex mixture of isotopomers. The
1
identification through H NMR spectroscopy of 11-cis-retinal-
19,19,19-d3, 11-cis-retinal-14,19,19-d3, all-trans-retinal-19,19,19-
d3, all-trans-retinal-14,19,19-d3, 13-cis-retinal-19,19,19-d3, and
13-cis-retinal-14,19,19-d3 further confirms our mechanistic
proposal (see Figure 5).25 HPLC separation of the different
double-bond isomers revealed the extent of deuterium scram-
bling in the rearrangement manifold (see the Supporting
Information).
Another corollary of the drastically different nature of the
competing mechanisms detailed above is the possibility to tune
the product distribution stabilizing or destabilizing selectively
only one of the rate-limiting transition states (i.e., the diradical
structure leading to all-trans-retinal or the sigmatropic shifts
on route to 13-cis-retinal). The diradical transition state should
be favored by fluorine substitution at C11.26 11-cis-11-Fluo-
roretinal 12, prepared as described in previous studies,27 was
Computational and Experimental Methods
For the thermal studies, dilute solutions of 11-cis-retinal 1 in
C6D6 were heated to 80 °C in the NMR probe, and spectra were
collected at the appropriate times. The synthesis of 11-cis-retinal-
19,19,19-d3 is described in the Supporting Information.
All of the structures presented in this work were optimized by
means of density functional theory methods.28,29 B3LYP was used
in conjunction with a 6-31+G(d,p) basis set.30 This level of theory
is known to provide an excellent balance between cost and accuracy
(MAD of 3.9 kcal/mol).30 The nature of all the stationary points
was determined by harmonic analysis. Due to the potential diradical
character of some of the structures considered in this work, the
(24) Isotope effects (kH/kD of about 6) have been measured in the [1,7]-H
sigmatropic shift of the previtamin D system: (a) Okamura, W. H.; Hoeger, C. A.;
Miller, K. J.; Reischl, W. J. Am. Chem. Soc. 1988, 110, 973–974. (b) For simple
trienes, the values are lower; see: Okamura, W. H.; Elnagar, H. Y.; Ruther, M.;
Dobreff, S. J. Org. Chem. 1993, 58, 600–610.
(25) all-trans-Retinal-14,19,19-d3 is formed by direct rotation about the C11-
C12 double bond in 11-cis-retinal-14,19,19-d3 which, in turn, is formed via [1,7]-
H sigmatropic shift driven H/D scrambling from the parent 11-cis-retinal-
19,19,19-d3 (see Figure 5).
(28) (a) Becke, A. D. Phys. ReV. A 1988, 38, 3098–3100. (b) Becke, A. D.
J. Chem. Phys. 1993, 98, 5648–5652.
(29) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B. 1988, 37, 785–789.
(30) Foresman, J. B. Frisch, Æ. Exploring Chemistry with Electronic Structure
Methods, 2nd ed.; Gaussian, Inc.: Pittsburgh, PA, 1996.
(26) Sreeruttun, R. K.; Ramasami, P.; Wannere, C. S.; Paul, A.; von R.
Schleyer, P.; Schaeffer, H. F., III. J. Org. Chem. 2005, 70, 8676–8686.
´
(27) Francesch, A.; Alvarez, R.; Lo´pez, S.; de Lera, A. R. J. Org. Chem.
1997, 62, 310–319.
1012 J. Org. Chem. Vol. 74, No. 3, 2009

Complex Thermal BehaVior of 11-cis-Retinal
internal and external stability of the wave functions was computed
via the Hermitian stability matrices A and B in all cases.31 For all
structures exhibiting unstable restricted wave functions, the spin-
symmetry constraint of the wave function was released (i.e.,
expanding the SCF calculation to an unrestricted space, UB3LYP),
leading to stable unrestricted wave functions.
The methodology here described yielded very accurate activation
barriers for the 11-cis-retinal to all-trans-retinal isomerization (∆G #)
27.3 kcal/mol at the B3LYP level) as confirmed through comparison
with experimental data (∆Gq ) 26.2 kcal/mol).7
the Centro de Supercomputacio´n de Galicia (CESGA) for
generous allocation of computational resources. We dedicate
this paper to Prof. W. H. Okamura for his mentorship and his
many contributions in the field of polyenes.
Supporting Information Available: Cartesian coordinates,
SCF energies, and the number of imaginary frequencies for each
structure are available. A two-dimensional potential energy
surface of the direct double bond isomerization and procedures
for the preparation of 11-cis-retinal-19,19,19-d3, protocols
followed for the kinetic experiments and NMR spectra of the
isomerization producs are also provided. This material is
available free of charge via the Internet at http://pubs.acs.org.
Acknowledgment. We are grateful to the MEC-Spain
(SAF2004-07131, SAF2007-63880-FEDER) and Xunta de Gali-
cia (IPP Research Contract to C.S.) for financial support and to
(31) Bauernschmitt, R.; Ahlrichs, R. J. Chem. Phys. 1996, 104, 9047–9052.
JO801899K
J. Org. Chem. Vol. 74, No. 3, 2009 1013