(9Z)- and (11Z)-8-Methylretinals for Artificial Visual Pigment Studies: Stereoselective Synthesis, Structure, and Binding Models

Article abstract of DOI:10.1002/chem.200304847

Artificial visual pigment formation was studied by using 8-methyl-substituted retinals in an effort to understand the effect that alkyl substitution of the chromophore side chain has on the visual cycle. The stereoselective synthesis of the 9-cis and 11-c

Full text of DOI:10.1002/chem.200304847

FULL PAPER
(9Z)- and (11Z)-8-Methylretinals for Artificial Visual Pigment Studies:
Stereoselective Synthesis, Structure, and Binding Models
Rosana Alvarez,[a] Marta DomÌnguez,[a] Yolanda Pazos,[b] Fredy Sussman,[b] and
Angel R. de Lera*[a]
Abstract: Artificial visual pigment for-
mation was studied by using 8-methyl-
substituted retinals in an effort to un-
derstand the effect that alkyl substitu-
tion of the chromophore side chain has
on the visual cycle. The stereoselective
synthesis of the 9-cis and 11-cis isomers
of 8-methylretinal, as well as the 5-de-
methylated analogues is also described.
The key bond formations consist of a
thallium-accelerated Suzuki cross-cou-
pling reaction between cyclohexenyl-
boronic acids and dienyliodides (C6ꢀ
hexenylboronic acid was prepared by
trapping the precursor cyclohexenyl-
lithium species with B(OiPr)3 or
B(OMe)3. The cyclohexenyllithium
species is itself obtained by nBuLi-in-
duced elimination of a trisylhydrazone
(Shapiro reaction), or depending upon
the steric hindrance of the ring, by
iodine metal exchange. In binding ex-
periments with the apoprotein opsin,
only 9-cis-5-demethyl-8-methylretinal
yielded an artificial pigment; 9-cis-8-
methylretinal simply provided residual
binding, while evidence of artificial pig-
ment formation was not found for the
11-cis analogues. Molecular-mechanics-
based docking simulations with the
crystal structure of rhodopsin have al-
lowed us to rationalize the lack of
binding displayed by the 11-cis ana-
logues. Our results indicate that these
isomers are highly strained, especially
when bound, due to steric clashes with
the receptor, and that these interac-
tions are undoubtedly alleviated when
9-cis-5-demethyl-8-methylretinal binds
opsin.
Keywords: cross-coupling ¥ docking
C7), and
a highly stereocontrolled
models
¥ membrane proteins ¥
Horner Wadsworth Emmons or Wittig
condensation (C11ꢀC12). The cyclo-
receptors
Introduction
ces, which span the lipid bilayer of the receptor cell mem-
brane. In contrast to other GPCR subfamilies, the ligand in
Rhs is covalently attached to the protein as an inverse ago-
nist. In the binding pocket of bovine Rhs (a 40 kDa protein
with 348 aminoacids) the light-sensitive chromophore 11-cis-
retinal 1 (Figure 1) is attached to Lys296 in helix VII
through a protonated Schiff base, while Glu113 in helix III
acts as a counterion.[2] Upon light absorption, the 11-cis
bond of the protonated Schiff base photoisomerizes to the
trans geometry; this is one of the fastest (less than 200 fs)[3]
and most efficient (quantum yield of 0.67) chemical reac-
tions known. As a result, the cyclohexenyl ring and several
of the polyene sidechain substituents are displaced to other
regions within the binding pocket. These perturbations are
linked to the protein structure alterations (such as displace-
ment of helices and reorganization of citoplasmic loops), the
Schiff base deprotonation (with concomitant proton uptake
by Glu134), and detachment of trans-retinal 2 (Figure 1)
from the apoprotein.[4,5] The isolation of intermediates,
which were spectrometrically detected at low temperature,
and their carefully monitored interconversion provides evi-
dence for these changes.[2] The Meta II intermediate, which
contains a deprotonated trans-retinal Schiff base, is the
The light-sensitive visual rhodopsin (Rhs) pigments are
members of the G-protein coupled receptor (GPCR) super-
family, which also consists of proteins sensitive to cell stimu-
li as diverse as calcium ions, neurotransmitters, hormones,
and even other proteins.[1] Common to this large family of
receptors is a protein backbone architecture of seven a-heli-
[a] Prof. Dr. A. R. de Lera, Dr. R. Alvarez, M. DomÌnguez
Departamento de QuÌmica Orgµnica
Facultad de Ciencias, Universidade de Vigo
36200 Vigo (Spain)
Fax : (+34)986-812-556
E-mail: qolera@uvigo.es
[b] Dr. Y. Pazos, Dr. F. Sussman
Departamento de QuÌmica Orgµnica
Facultad de QuÌmicas
Universidad de Santiago de Compostela
15700 Santiago de Compostela (Spain)
Fax : (+34)981-591-014
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
5821
Chem. Eur. J. 2003, 9, 5821 5831
DOI: 10.1002/chem.200304847
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

FULL PAPER
form visual pigments (rhodop-
sin and isorhodopsin, respec-
tively),[15] we herein describe
the stereoselective synthesis of
each of these isomers of 8-
methylretinal and 5-demethyl-
8-methylretinal (compounds 4
7, Figure 1). Opsin-binding
studies have revealed the pres-
ence of polyene distortions, es-
pecially in the 11-cis derivatives
4 and 5, which fail to bind the
apoprotein. These are a result
of the structural perturbation
introduced by the methyl group
Figure 1. 9-cis- and 11-cis-Retinal analogues with a methyl substituent at C8 of the parent retinal isomer.
active state of Rhs that interacts with a heterotrimeric G-
protein, transducin, through the extracellular loops that con-
nect the seven transmembrane helices. The protein protein
interaction then triggers a biochemical cascade which leads
to neural signals that provide the sensation of vision.[2,4,5]
Advances in protein extraction protocols have facilitated
the growth of Rhs crystals suitable for X-ray analysis. Al-
though the ground-state (inactive) structure of Rhs[6a] re-
veals detailed information about the chromophore opsin in-
teraction, the resolution (2.8 ä) is still insufficient for the
11-cis-retinal conformation to be unambiguously assigned.[6]
In particular, some discrepancies of the C6ꢀC7 bond[7] have
been noted with respect to the values reported for the 6-s-
cis (by X-ray diffraction)[6a] or 6-s-trans conformer (by solid-
state deuterium NMR spectroscopy using a labeled chromo-
phore).[8] Therefore, the use of synthetic retinals[9] that have
structural modifications in the cyclohexenyl ring or in the
adjacent side chain might help clarify the role that confor-
mational changes of the chromophore play in the visual
cycle.[10] Interestingly, recent results suggest that the 6-s-cis
to 6-s-trans transition, which occurs by a ring flip following
at C8. Structural and molecular mechanic docking simulation
studies of the apoprotein ligand interaction based on the
crystal structure of rhodopsin are also presented. The bind-
ing rankings of compounds 4 and 5 were calculated relative
to 1, and thus allowed us to put forth a rationale for the
binding experiment results observed for the 11-cis- and 9-
cis-retinal analogues. Our results indicate that the 11-cis de-
rivatives do not bind because of the strain induced in these
molecules from steric clashes between the putative ligands
and the Trp265 residue, which is located in the binding
pocket of the receptor. Therefore, the superior binding be-
havior of one of the 9-cis-retinal analogues (compound 7)
arises because of the smaller strain energy it encounters
when bound.
Results and Discussion
Synthesis: Our previously described protocol for the prepa-
ration of 9-cis-retinoic acid was considered to be the most
convenient approach to the desired analogues.[16] This
straightforward stereocontrolled synthesis involves the con-
struction of the C6ꢀC7 bond by convergent Suzuki cou-
pling[17] of a cyclohexenylboronic acid 8 and a stereodefined
w-iodotetraenyl ester 9 (path A, Scheme 1).
the 11-cis to trans photoisomerization, is involved in the ac-
13]
tivation step of the visual process.[11
Indeed, the X-ray
structure of Rhs shows C5ꢀCH3 and C8ꢀH to be in close
proximity.[6] We reasoned that by substituting the native hy-
drogen atom at C8 with
a
methyl group (8-methylretinal),
steric interactions of greater se-
verity would occur between
C8ꢀCH3 and C5ꢀCH3, and that
this might lead to changes in
the C6ꢀC7 conformation.[14]
Furthermore, it was considered
that positional exchange of C5ꢀ
CH3 and C8ꢀH (5-demethyl-8-
methylretinal) might result in a
small shift of steric bulk within
the chromophore,[6] and would
afford
opsin-bound
a
derivative with an
conformation
close to that of the native chro-
mophore.
Since both 11-cis-retinal 1
and 9-cis-retinal 3 are known to Scheme 1. Retrosynthesis of compounds 4 7.
5822
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 5821 5831

(9Z)- and (11Z)-8-Methylretinals
5821 5831
The appropriate choice of isomer of the commercially
available 3-methylpent-4-yn-1-ol [(E)-11 or (Z)-11] starting
material should control the C9ꢀC10 geometry of the retinal.
Functionalization of these enynols by methylalumination io-
dination,[18] which proceeds by syn-addition, should afford
(E)-10 or (Z)-10, and would provide the C7ꢀC8 trans geom-
etry required in the desired analogues 4 7. Moreover, ster-
eoselective Wittig or related condensations can be utilized
to attach the terminal five-carbon fragment of the polyenic
side chain, and thus, depending upon the order of steps in
the sequence, set the C11ꢀC12 configuration in either 4 7
(path B) or 9 (path A).
change prior to their subsequent cross-coupling with organo-
metallic reagents.[16]
With the shorter iododienols (Z)-10 and (E)-10 in hand,
attention was directed to the sequence indicated in path B
of Scheme 1. Here, the Suzuki reaction to afford (Z)-12 and
(E)-12 precedes the condensation step (C11ꢀC12 bond in
retinoid numbering) that completes the polyenic side
chain.[16] Cyclohexenylboronic acids such as 8 are routinely
acquired from the corresponding organolithium species by a
lithium boron exchange process. The desired alkenyllithium
compound can be obtained in two ways from commercially
available ketones; the choice depends on the steric hin-
drance of the precursor carbonyl group. Since the Shapiro
reaction of unhindered cyclohexanone hydrazones is
known,[24] trienols (Z)-12a and (E)-12a were prepared in
this manner, as shown in Scheme 3. 2,2-Dimethylcyclohexa-
The attempted stereocontrolled route to w-iodotetraenyl-
ester (Z)-9 is shown in Scheme 2. Zirconium-promoted
methylalumination[18] of cis-enynol (Z)-11 and subsequent
treatment of the resultant dimethylalkenylallane intermedi-
Scheme 2. a) i) [ZrCp2Cl2], Me3Al, CH2Cl2, 258C; ii) I2, THF, ꢀ508C
[(Z)-10, 52%; (E)-10, 60%]; b) MnO2, K2CO3, CH2Cl2, 258C, 1.5 h [(Z)-
15, 84%]; c) phosphonate 14, nBuLi, DMPU, THF, ꢀ78!ꢀ358C (9,
75%; (Z)-16, 61%).
Scheme 3. a) i) nBuLi, THF, ꢀ788C; ii) B(OiPr)3, 08C; iii) [Pd(PPh3)4],
iodide (Z)-10 or (E)-10, 10% aq. TlOH [(Z)-12a, 60%; (E)-12a, 98%];
b) i) tBuLi, THF, ꢀ788C; ii) B(OMe)3, 08C; iii) H2O (77%);
c) [Pd(PPh3)4], iodide (Z)-10 or (E)-10, 10% aq. TlOH [(Z)-12b, 60%;
(E)-12b, 65%].
ate with iodine in THF at ꢀ408C afforded iododienol (Z)-10
(52% yield). Owing to its degradation after rapid valence
isomerization to the a-pyran,[19] aldehyde (Z)-13, which was
obtained by MnO2 oxidation of (Z)-10 in the presence of
K2CO3,[20] was used immediately without purification. Hor-
ner Wadsworth Emmons (HWE) condensation of (Z)-13
with the anion derived from phosphonate 14 in the presence
of DMPU[21] provided tetraenyliodide 9. The C4ꢀC5 bond
none trisylhydrazone 17[25] was treated with nBuLi, and the
resultant alkenyllithium species was trapped with B(OiPr)3
to afford boronate 8a. Sequential addition of [Pd(PPh3)4],
iodide (Z)-10 or (E)-10, and a 10% aqueous TlOH solu-
tion[26] provided, after stirring for 4 h at 258C, alcohols (Z)-
12a or (E)-12a in 60 and 98% combined yields, respectively
(Scheme 3).
Synthesis of the 8-methylretinal analogues illustrates the
alternative method available for the generation of the alke-
nyllithium precursor.[25] The halogen lithium exchange reac-
tion is suitable for hydrazones that are derived from hin-
dered ketones in which deprotonation of the Ca tertiary
carbon with nBuLi, as in the Shapiro reaction, is inefficient.
Oxidation of the hydrazone derived from 2,2,6-trimethylcy-
clohexanone with iodine using Barton×s procedure[27]
(Scheme 3) afforded the cycloalkenylboronic acid 8b[25] pre-
1
formation was highly trans stereoselective, but the H NMR
spectrum of 9 displayed signals for both geometric isomers
at the C8ꢀC9 bond (1:2 ratio) (Scheme 2); this resulted
from the reversible valence isomerization indicated above.
Although reports describing the successful carboalumination
of polyene ynes[18,22] exist, iodide (Z)-9 could not be ob-
tained from (Z)-16 (itself prepared as indicated in Scheme 2
by HWE condensation of enynal (Z)-15,[23] and the phos-
phonate anion obtained from 14) by this procedure. These
and related observations highlight the instability of w-iodo-
tetraenyl esters, which should be able to be prepared from
the corresponding tetraenylstannanes by tin iodine ex-
5823
Chem. Eur. J. 2003, 9, 5821 5831
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

A. R. de Lera et al.
FULL PAPER
cursor, cycloalkenyliodide 18 (77% yield). Compound 8b
was immediately coupled to iodide (Z)-10 or its geometric
isomer in the same manner as that described for the de-
methylated analogues (Scheme 3) to afford, after stirring for
8 h, trienols (Z)-12b and (E)-12b in 60 and 65% yield, re-
spectively.
It was envisaged that the geometry of the C11ꢀC12
double bond, which is required to complete the polyene side
chain, could be controlled by the use of stereoselective var-
iations of the Wittig and HWE reactions. Trienols (Z)-12a
and (Z)-12b were oxidized with catalytic quantities of tetra-
n-propylammonium perruthenate (TPAP) in CH2Cl2 in the
presence of N-methylmorpholine N-oxide (NMO) as co-oxi-
dant (Scheme 4).[28] The resultant aldehydes (Z)-19a and
(Z)-19b were then treated with the anion of phosphonate 14
ment of aldehydes (E)-21a and (E)-21b at ꢀ788C with the
phosphorane derived from 22 provided the unstable 11-cis-
retinols. These were immediately oxidized under basic
MnO2 conditions to afford retinals 4 and 5 in 60 and 51%
combined yields, respectively, after purification by HPLC
(Scheme 5). Examination of the peak intensities in the
HPLC trace allowed the Z/E stereoselectivity for the Wittig
reaction to be estimated at a remarkable 20 22:1 ratio.
Structural studies: It was anticipated that the increase in
steric bulk at C8 relative to the native retinal would alter
the conformation of the bonds that are proximal to the hy-
drophobic ring, particularly for the 8-methyl retinal isomers
4 and 6. On the other hand, we considered that exchange of
the C5 and C8 substituents in the 5-demethyl-8-methylreti-
nal isomers 5 and 7 would mimic the steric interactions pres-
ent in the same region of the parent retinal (Figure 1). Dis-
tortions as a result of the C8 methyl substituent in the reti-
nal polyene conformation are supported by spectroscopic
and molecular mechanic studies. For example, broad signals
for the C1ꢀ(CH3)2 groups appeared in the room tempera-
1
ture H NMR spectra of 6, as well as in the simpler systems
(E)-12b and (E)-21b; this indicates a low-exchange regime.
When a solution of 6 in [D8]toluene was cooled to ꢀ358C,
individual methyl resonance signals were observed, and a
dynamic NMR study revealed that these signals coalesced at
ꢀ158C. From these experiments, the free energy of activa-
tion (DG°) at the coalescence temperature was estimated to
be about 14.2Æ0.5 kcalmolꢀ1.[30] For the shorter trienal
model (E)-21b, coalescence occurs at room temperature,
while at ꢀ208C, the two C1 methyl groups appear as sharp
singlets (DG° is about 14.9Æ0.5 kcalmolꢀ1). In principle,
the barrier to interconversion could be ascribed either to cy-
clohexene ring inversion or to C6ꢀC7 bond rotation. The
retinal cyclohexene has a half-chair conformation,[31] but
there are indications that a dynamic process occurs even in
the solid state, in which ring inversion through a cyclohex-
ene boat-form transition state with an energy barrier of
about 6.3 kcalmolꢀ1 interconverts these half-chair conforma-
tions. The conformation of the C6ꢀC7 bond has been found
Scheme 4. a) TPAP, NMO, 4 ä MS, CH2Cl2, 258C [ Z( )-19a, 77%; (Z)-
19b, 91%]; b) i) hosphonate 14, nBuLi, DMPU, THF, 08C; ii) aldehyde
(Z)-19a or (Z)-19b, ꢀ78!ꢀ408C [ Z()-20a, 95%; (Z)-20b, 94%];
c) i) DIBAL-H, THF, ꢀ788C; ii) MnO2, Na2CO3, CH2Cl2, 258C (7, 79%;
6, 90%).
as indicated[21] to afford the (11E)-pentaenyl esters (Z)-20a
and (Z)-20b in high yields. Subsequent functional-group ma-
nipulation, which included a DIBAL-H reduction and MnO2
oxidation, afforded the desired retinals 6 and 7.
For the stereoselective preparation of the (11Z)-reti-
noids, we turned our attention to the procedure described
by Kobayashi;[29] this uses the (E)-oxidoallylic phosphorane
reagent derived from phosphonium salt 22 (Scheme 5). Al-
though this (Z)-selective Wittig condensation is rarely used,
we found it to be a potentially useful and reliable method
when a freshly opened bottle of KHMDS was used. Treat-
to be close to s-cis (i.e., in the crystal structure of 11-cis-reti-
34]
nal).[32
Ab initio studies have also addressed the confor-
mational equilibria of retinal models derived from b-
ionone.[35] In the latter, diastereomeric minima were found
in which the C6ꢀC7 s-cis conformation exhibited dihedral
angles of 60.7 and ꢀ64.98, and barrier heights of about
5 kcalmolꢀ1 were estimated for bond rotation about the C6ꢀ
C7 bond.
The retinal derivatives (4 and 6) synthesized in this
study contain an additional methyl group at C8. It is reason-
able to assume that this sterically-demanding methyl group
has less influence on the cyclohexene ring inversion (i.e., se-
quential movement of C2 and C3 through the plane of the
double bond) than on the rotation of the ring sidechain
bond, as in the latter there is a severe steric interaction with
the C5ꢀCH3 group. Inspection of Dreiding molecular
models clearly shows that upon rotation about the C6ꢀC7
bond, steric clashes occur between C8ꢀCH3 and C5ꢀCH3 on
one side, and C8ꢀCH3 and C1ꢀCH3 on the other side.
Scheme 5. a) TPAP, NMO, 4 ä MS, CH2Cl2, 258C [ E( )-21a, 92%; (E)-
21b, 67%]; b) i) phosphonium salt 22, KHMDS, THF, ꢀ78!258C; ii) al-
dehyde (E)-21a or (E)-21b, THF, ꢀ78!258C; c) MnO2, Na2CO3,
CH2Cl2, 258C (5, 51%; 4, 60%).
5824
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 5821 5831

(9Z)- and (11Z)-8-Methylretinals
5821 5831
Table 1. Absorption maxima of the parent retinal isomers and their ana-
logues.
Therefore, we consider that the dynamic process observed
by H NMR spectroscopy corresponds to interconversion of
1
Retinal Absorption maxima [nm] Retinal Absorption maxima [nm]
the most stable conformers. Since the energy barrier is con-
siderably higher than that calculated for the model
system,[36] we computed the energy profile that corresponds
to the C6ꢀC7 bond rotation of compound 6 by using molec-
ular mechanic calculations on a CHARMM force field (see
Experimental Section).[37] Two energy minima, which
showed a high departure from planarity, were found at ꢀ40
and 1208. The barrier for interconversion was calculated to
be 13.8 kcalmolꢀ1; this value is similar to that found for the
experimental free energy of activation at the coalescence
temperature discussed above.
To further document the existence of steric clashes in
the C8ꢀCH3 C11ꢀH region, nOe difference spectra were re-
corded. The nuclear Overhauser effect study for 9-cis-reti-
nals 6 and 7 in CD3OD solution at ambient temperature re-
veals a highly distorted polyene, in particular, around the
C6ꢀC7 and C8ꢀC9 bonds. The most relevant data are de-
picted in Figure 2. The through-space interactions of C8ꢀ
1[a]
3[a]
2[a]
4
365[b], 380[d]
363[b], 373[d]
368[c], 383[d]
368[c]
5
6
7
371[c]
342[c], 344[d]
345[c], 343[d]
[a] Taken from ref. [38]. [b] In hexane. [c] In methanol. [d] In ethanol.
tion of a partial conjugative interaction of all the polyene
double bonds.
Protein binding–molecular modelling studies: Preliminary
opsin-binding experiments revealed that the added steric
bulk at C8 in compounds 4 and 5 with respect to native 11-
cis-retinal 1 was detrimental to apoprotein binding, since
pigment formation was not detected for 4, while the yield
for 5 was very low (~5%); this discouraged further stud-
ies.[39] At first it was speculated that the added steric bulk at
C8 had induced a conformational change around the C6ꢀC7
bond, particularly in the more substituted derivative 4, and
that this had prevented its entry into the binding pocket.
However, a conformational search of 4 and 5 using molecu-
lar mechanic calculations (see Figure 3 below) revealed that
the polyene side chain is folded in a manner similar to that
adopted in the crystal structure of free and opsin-bound 11-
cis-retinal.[6,32] As expected, the absorption spectra of ana-
logues 4 and 5 (Table 1) do not significantly differ from that
of 1; a small blue-shift (around 12 nm) is observed for the
more substituted 4. Therefore, on the assumption that the
conjugated polyene structure is close to that of 1, the reti-
nals 4 and 5 fail to form pigments because of enhanced
ligand protein steric interactions. Given the recent availabil-
ity of a rhodopsin crystal structure,[6] we performed molecu-
lar modelling calculations on the free and bound ligand (see
Experimental Section), and attempted to find a rationale for
the results based on energy criteria.
Figure 2. Relevant through-space interactions (nOe difference values) for
compounds 6 and 7.
CH3 with C11ꢀH, and C9ꢀCH3 with C7ꢀH are negligible
(less than 1% nOe); this indicates that the C8ꢀC9 bond of
the unbound chromophore has a twisted conformation.
Through-space interaction is not apparent between C8ꢀCH3
and the ring methyl substituents, as is expected in a predom-
inantly twisted conformation.
The UV spectra of analogues 6 and 7 (listed in Table 1)
also suggest that the C8 methyl group causes a loss of pla-
narity since the absorption maxima displays a 30 nm blue-
shift with respect to 9-cis-retinal 3. The 11-cis-analogues, 4
and 5, show a moderate shift of around 10 nm relative to
the parent system 1. Therefore, we surmised that the hypso-
chromic shift in 6 and 7 origi-
The energy difference for Schiff base formation with
opsin between the cognate retinal and its analogues can
be calculated and analyzed by using the thermodynamic
+
cycle shown in Scheme 6,[40] in which Opꢀ NH3 is the opsin
molecule, R1ꢀCHO and R4ꢀCHO/R5ꢀCHO are the cognate
11-cis-retinal 1 and analogues 4 and 5, respectively, and
nates from the C8ꢀCH3¥¥¥C11ꢀ
H steric interaction, and that
this is lessened in the C8ꢀ
CH3¥¥¥C10ꢀH region of the 11-
cis-retinal analogues. Taken to-
gether, the spectroscopic data
for the 8-methylretinal series 4
7 indicates that the ring-chain
steric crowding is somehow al-
leviated by both a twist of the
adjacent C8ꢀC9 single bond
Scheme 6. Thermodynamic cycle for analysis of the energy difference for Schiff base formation with opsin be-
and by the simultaneous reten- tween the cognate retinal and its analogues.
5825
Chem. Eur. J. 2003, 9, 5821 5831
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

A. R. de Lera et al.
FULL PAPER
+
Rhoꢀ NH=CHR are the native and artificial rhodopsins,
our calculations (see values in parenthesis, Table 2). The
values for DG4 are rather small relative to DG3. As a result,
they do not affect the outcome, that is, the condensation re-
action of analogues 4 and 5 is highly unfavorable in compar-
ison to Schiff base formation with 11-cis-retinal.
The rationale behind the preferential binding can be
found by analyzing the internal energy of the bound ligands.
Table 3 lists the internal energy differences for the opsin-
that is, the putative Schiff base adducts. DG1 and DG2 are
the reaction energies for Schiff base formation, DG3 is the
energy when the Schiff base adduct of 11-cis-retinal 1 trans-
forms into the Schiff base adducts of analogues 4 and 5, and
DG4 gives the energy difference between native retinal and
an analogue in the unbound state.
The closure property of this cycle yields the differential
binding energy between any two ligands [Eq. (1)]:
Table 3. Differences in internal energy between 11-cis-retinal and its ana-
DDGbin ¼ DG2ꢀDG1 ¼ DG3ꢀDG4
ð1Þ
logues [Kcalmolꢀ1
]
4
5
In other words, as Karplus has pointed out,[40a] the bind-
ing free-energy change can be calculated either from the
™alchemical∫ vertical legs or from the ™chemical∫ horizontal
legs of the thermodynamic cycle.
including Me group (bound)
excluding Me group (bound)
unbound
52.5
29.6
3.0
55.6
27.4
ꢀ3.3
The ranking binding free-energy function is made up of
an internal energy [DDGbin(int)] term, which can be calculat-
ed by molecular mechanics, and a solvation term [DDGbin(-
solv)]. Entropy contributions have been neglected, because
we expect that the analogue modification would have negli-
gible effects. Hence, the binding ranking can be written as a
sum of two terms [Eq. (2)]:
bound analogues (4 and 5) relative to the native ligand. This
quantity contributes to DDGbin(int) [see Eq. (2)], and corre-
sponds to the change in strain energy between the native
ligand and its analogues when bound. As can be seen, when
the Schiff bases are formed the internal energy of the ana-
logues is much greater than that of 11-cis-retinal; this indi-
cates that the binding of these analogues is burdened by a
higher amount of strain than the original ligand. To estimate
the direct effect the additional methyl group has on the
strain energy of the molecule, we evaluated the strain ener-
gies of the unoptimized bound retinal analogue structure
without the C8 methyl group. The difference in strain
energy with respect to the native ligand was still found to be
large. This indicates that the strain resides not only around
the modified region (at the C8 position), but is also trans-
mitted to the rest of the molecule, such as the ring moiety
and the polyenic system. To visualize the distortions induced
by binding we have superimposed the resultant models for
the free and bound structures for every analogue. The re-
sults are displayed in Figure 3A C and allow the extent of
the distortion that the ligands undergo upon binding to be
compared. In agreement with the energy results, the smallest
changes in ligand conformation upon binding arise in the
native ligand 11-cis-retinal 1. The largest changes brought
about by binding can be found in the ring conformation and
in the orientation of the methyl groups attached to the poly-
enic system of analogues 4 and 5, both of which are critical
structural determinants for efficient binding.[6,8]
The strain energy difference is considerably smaller for
the unbound analogues (see Tables 2 and 3). Hence, our re-
sults indicate that 4 and 5 fail to form artificial visual pig-
ments because of the strain they experience upon being
bound to the apoprotein opsin. The source of this strain can
be explained upon inspection of the unoptimized artificial
rhodopsin models. Figure 3D shows a close up of the bind-
ing region for the structure on which the rhodopsin adducts
were modeled, and includes the bound analogue 4 and resi-
due Trp265, which is the closest fragment to the additional
methyl group. As can be seen, the additional methyl group
is located well below van der Waals contact distances from
the methyl group located off the ligand ring, as well as from
the indole ring of residue Trp265. Structure optimization of
DDGbin ¼ DDGbinðintÞ þ DDGbinðsolvÞ
ð2Þ
The solvation term contains a polar component, which
penalizes the desolvation of polar groups, as well as a term
for hydrophobic desolvation, which favors binding.[41] The li-
gands studied here are highly apolar. Hence, we have ap-
proximated the desolvation energy in a hydrophobic term
that is directly proportional (the proportionality constant
has been suggested by Sharp et al.[42]) to the solvent accessi-
bility.
Table 2 lists the differential binding energy values for an-
alogues 4, 5, and native 11-cis-retinal 1, as given by Equa-
tions (1) and (2). As seen from Table 2, the internal energy
differences for the rhodopsin adducts (DG3) in the thermo-
Table 2. Differential binding free energies for the schiff base formation
of ligands 4 and 5 with opsin relative to 1.
Energy difference[a]
4
5
DG3(int)[b]
DG4(int)
DDGbind(int)
DDGbind(solv)[c]
DDGbind
110.2 (53.0)[d]
3.0
118.3 (51.3)
ꢀ3.3
107.2 (50.0)
ꢀ1.3
121.6 (54.6)
ꢀ2.0
105.9 (48.7)
119.6 (52.6)
[a] Energies in Kcalmolꢀ1. [b] The first two rows list the difference in the
internal energy component for the terms indicated in Equation (1).
[c] Differential in solvation energy. [d] The numbers in parenthesis were
obtained taking the full molecule into account.
dynamic cycle are positive for both 11-cis-retinal analogues;
this indicates that the Schiff base formation will be adverse-
ly effected when the cognate 11-cis-retinal 1 is replaced with
analogues 4 or 5. The general trend in DG3(int) did not
change when interactions were restricted only to those resi-
dues that were used in the energy minimization (10 ä), or
when the whole molecule was taken into consideration for
5826
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 5821 5831

(9Z)- and (11Z)-8-Methylretinals
5821 5831
isomer 7, but not the 11-cis
isomer 6, forms the corre-
sponding artificial visual pig-
ment, albeit in low yield.
Experimental Section
General: All reactions were carried
out under an atmosphere of argon,
and those that did not involve aque-
ous reagents were carried out in
oven-dried glassware. Tetrahydro-
furan (THF) was distilled over
sodium benzophenone ketyl, and di-
chloromethane was distilled over
calcium hydride. Flash column chro-
matography was carried out under
pressure using Merck Kieselgel 60
(230 400 mesh). High-performance
liquid chromatography was per-
formed on a Waters machine with a
dual-wave detector (254 and
390 nm). Analytical thin-layer chro-
matography (TLC) was performed
on aluminum plates with Merck
Kieselgel 60 F254, and were visual-
ised by UV irradiation (254 nm) or
by staining with a solution of phos-
phomolybdic acid. UV/Vis spectra
were recorded on a HP5989A spec-
trophotometer. Infrared spectra of
thin films deposited onto NaCl glass
Figure 3. A C) Modeled structure of the 11-cis-retinals (1, 4, and 5) as they form a Schiff base with Lys296
(magenta) in rhodopsin; this has been superimposed over the modelled free ligands (coloured by their atom
types). For clarity, hydrogens have been left out of the displayed structures. The upper left (A), upper right
(B), and lower left (C) panels correspond to native 11-cis-retinal 1 and its analogues 4 and 5, respectively. D)
The lower right panel depicts a close up of the unoptimized rhodopsin complex binding site in the presence of
analogue 4. The retinal adduct and the closest residue (Trp265) to the additional methyl substituent are includ-
ed.
were obtained on a MIDAC Prospect FTIR spectrophotometer. Electron
ionization mass spectra were obtained on a Hewlett-Packard HP59970 in-
strument operating at 70 eV. High-resolution mass spectra (HRMS) were
taken on a VG Autospec instrument. 1H NMR spectra were recorded in
CDCl3 or C6D6 at ambient temperature on a Bruker AMX-400 spectrom-
eter at 400 MHz using residual protic solvent as the internal reference
(CHCl3, dH =7.23; C6D6, dH =7.26); chemical shifts (d) are given in ppm,
and coupling constants (J) are given in Hz. The 1H NMR spectra are re-
ported as follows: d (multiplicity, coupling constant J, number of protons;
assignment). 13C NMR spectra were recorded in CDCl3 at ambient tem-
perature on the same spectrometer at 100 MHz using the central peak of
CHCl3 (d=77.0) or C6H6 (d=128.0) as the internal reference. DEPT135
spectra aided the assignment of signals in the 13C NMR spectra; differ-
ence nOe experiments were also performed in a number of cases.
4 and 5 may relieve the steric clashes that preclude them
from binding by substantially increasing the internal strain
energy of these analogues.
Interestingly, when analogue 7, in which the C8ꢀH and
C5ꢀCH3 groups from the parent 9-cis-retinal 3 have been
exchanged, was incubated with apoprotein, the artificial pig-
ment 5-demethyl-8-methylisorhodopsin was formed, albeit
only in 20% relative yield with respect to the native 9-cis-
retinal. The pigment showed roughly similar UV maxima
(525 nm) to that displayed by native isorhodopsin; this indi-
cates that the structural changes were adequately compen-
sated by the protein binding-pocket flexibility. The proper-
ties of the pigment so generated have been described else-
where.[39]
General procedure for methylalumination iodination
(2Z,4E)-3,4-Dimethyl-5-iodopenta-2,4-dien-1-ol ((Z)-10): Trimethylalu-
minum (3.1 mL, 31.26 mmol) and (Z)-11 (1.0 g, 10.42 mmol) were added
Since a crystal structure is not yet available for isorho-
dopsin, we can only hypothesize that the bound conforma-
tion of compound 7 has reduced steric clashes with the re-
ceptor, and is therefore, not highly strained.
Since only residual pigment formation was detected
upon incubating 6 with opsin, the additional methyl group
probably causes a greater alteration of the chromophore
structure, which in turn makes it unable to fit in the binding
pocket.
In summary, an additional methyl substituent at the C8
sidechain position abrogates the ability of the retinal to bind
to the apoprotein opsin. Molecular modeling studies indi-
cate that the strain penalty energy that analogues 4 and 5
must incur upon binding precludes them from forming a
stable artificial pigment. If the structural perturbation is
compensated by removal of the C5 methyl group, the 9-cis
to
a cooled (08C) suspension of [ZrCp2Cl2] (3.03 g, 10.42 mmol) in
CH2Cl2 (20 mL). After stirring for 12 h, the mixture was cooled (ꢀ508C),
and a solution of iodine (7.93 g, 31.26 mmol) in THF (27 mL) was added.
The reaction was carefully poured over THF/H2O (50:50, v/v) and was
then extracted with Et2O (3î). The combined organic layers were
washed with saturated Na2S2O3 solution (3î) and H2O (3î), dried over
Na2SO4, and the solvent was removed. The residue was purified by
column chromatography (silica gel, 80:20 hexane/EtOAc) to afford a
1
yellow oil (1.29 g, 52%). H NMR (400.13 MHz, CDCl3, 258C, TMS): d=
1.54 (s, 3H; CH3), 1.74 (d, J(H,H)=1.2 Hz, 3H; CH3), 3.90 (d, J(H,H)=
6.8 Hz, 2H; CH2), 5.30 (t, J(H,H)=6.8 Hz, 1H; CH), 5.91 ppm (q,
J(H,H)=1.2 Hz, 1H; CH); 13C NMR (100.62 MHz, CDCl3): d=22.1 (q),
23.3 (q), 59.7 (t), 78.9 (d), 127.4 (d), 140.0 (s), 147.3 ppm (s); IR (NaCl):
n˜ =3600 3100, 2966, 2917, 2872, 1595, 1437, 1252, 998 cmꢀ1; MS: m/z
(%): 221 (18) [M+ꢀ17], 189 (3), 173 (6), 149 (12), 127 (7), 111 (20), 94
(100), 79 (64), 77 (28); HMRS m/z: calcd for C7H11O: 111.0810; found:
111.0809 [M+ꢀI].
(2E,4E)-3,4-Dimethyl-5-iodopenta-2,4-dien-1-ol ((E)-10): (E)-11 (0.5 g,
5.21 mmol) was treated with [ZrCp2Cl2] (1.51 g, 5.21 mmol), Me3Al
5827
Chem. Eur. J. 2003, 9, 5821 5831
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

A. R. de Lera et al.
FULL PAPER
(1.6 mL, 15.6 mmol), and iodine (3.96 g, 15.6 mmol) following the general
procedure for methylalumination iodination to give, after purification by
column chromatography (silica gel, 80:20 hexane/EtOAc), a yellow oil
(0.74 g, 60%). 1H NMR (400.13 MHz, CDCl3): d=1.83 (s, 3H; CH3),
2.05 (d, J(H,H)=0.8 Hz, 3H; CH3), 4.27 (d, J(H,H)=6.4 Hz, 2H; CH2),
5.81 (dd, J(H,H)=6.4, 0.8 Hz, 1H; CH), 6.42 ppm (s, 1H; CH); 13C
NMR (100.62 MHz, (CD3)2CO): d=14.6 (q), 22.8 (q), 59.8 (t), 80.1 (d),
130.6 (d), 135.7 (s), 148.8 ppm (s); IR (NaCl): n˜ =3500 3200, 2923, 2864,
1570, 1442, 1376, 1003 cmꢀ1; MS: m/z (%): 238 (36) [M+], 236 (7),
219 (13), 149 (22), 148 (17), 131 (42), 119 (25), 111 (63), 94 (31), 93 (33),
78 (28), 69 (100); HMRS m/z: calcd for C7H11IO: 237.9855; found:
237.9849.
(100), 81 (15); HMRS m/z: calcd for C9H17BO2: 168.1322; found:
168.1322.
General procedure for Suzuki cross coupling
(2Z,4E)-3,4-Dimethyl-5-(2,6,6-trimethylcyclohex-1-en-1-yl)penta-2,4-
dien-1-ol ((Z)-12b):
A solution of alkenylboronic acid 8b (0.23 g,
1.36 mmol), [Pd(PPh3)4] (0.125 g, 0.108 mmol), and 10% aqueous TlOH
(9.2 mL, 4.16 mmol) in THF (7 mL) was added to a solution of (Z)-10
(0.26 g, 1.08 mmol) in THF (7.0 mL). After stirring for 8 h at 258C, the
mixture was diluted with Et2O (6 mL) and washed with aqueous
NaHCO3 (3î). The organic layer was dried over Na2SO4 and the solvent
was evaporated. The residue was purified by column chromatography
(silica gel, 80:15:2 hexane/EtOAc/Et3N) to afford a yellow oil (0.15 g,
60%). 1H NMR (400.13 MHz, CDCl3): d=0.95 (s, 6H; 2CH3), 1.40 1.70
(m, 4H; 2CH2), 1.53 (s, 3H; CH3), 1.56 (d, J(H,H)=1.3 Hz, 3H; CH3),
1.85 (s, 3H; CH3), 1.90 2.00 (m, 2H; CH2), 4.20 (d, J(H,H)=6.8 Hz, 2H;
CH2), 5.40 (tq, J(H,H)=6.8, 1.3 Hz, 1H; CH), 5.61 ppm (s, 1H; CH); 13C
NMR (100.61 MHz, CDCl3): d=16.9 (q), 19.3 (t), 21.2 (q), 23.4 (q), 28.4
(2q), 31.9 (t), 34.7 (s), 39.1 (t), 60.6 (t), 123.8 (d), 126.2 (d), 128.3 (s),
135.3 (s), 137.6 (s), 144.3 ppm (s); IR (NaCl): n˜ =3600 3100, 2927, 2864,
1436, 1371, 1001 cmꢀ1; UV (MeOH): lmax =226 nm; MS: m/z (%): 234 (2)
[M+ꢀ17], 217 (64), 203 (41), 161 (18), 147 (31), 133 (100), 119 (26), 109
(15), 105 (16), 95 (16), 83 (20), 69 (42); HMRS m/z: calcd for C16H26O:
234.1984; found: 234.1979.
General procedure for Shapiro Suzuki cross coupling
(2Z,4E)-5-(6,6-Dimethylcyclohex-1-en-1-yl)-3,4-dimethylpenta-2,4-dien-1-
ol ((Z)-12a): A cooled (ꢀ788C) solution of 17[25] (0.87 g, 2.14 mmol) in
THF (6 mL) was treated with nBuLi (2.6 mL, 2.47m in hexane,
6.43 mmol). After stirring for 30 min, B(OiPr)3 (0.98 mL, 4.250 mmol)
was added, and the temperature was raised to 08C. The mixture was stir-
red for 1 h at 08C and for 10 min at 258C. The solution was then added,
by cannula, to a separate flask that contained a solution of [Pd(PPh3)4]
(0.2 g, 0.17 mmol) and (Z)-10 (0.41 g, 1.70 mmol) in THF (4 mL). After
addition of a 10% aqueous TlOH solution (14.7 mL, 6.57 mmol), the
mixture was stirred for 4 h at 258C. The reaction mixture was diluted
with Et2O (8 mL) and was then washed with aqueous NaHCO3 (3î).
The organic layer was dried over Na2SO4 and the solvent was evaporated.
The residue was purified by column chromatography (silica gel, 80:15:2
hexane/EtOAc/Et3N) to afford a yellow oil (0.22 g, 60%). 1H NMR
(250.13 MHz, CDCl3): d=0.99 (s, 3H; CH3), 1.00 (s, 3H; CH3), 1.40 1.70
(m, 4H; 2CH2), 1.76 (d, J(H,H)=1.4 Hz, 3H; CH3), 1.83 (d, J(H,H)=
1.1 Hz, 3H; CH3), 2.00 2.20 (m, 2H; CH2), 4.15 (d, J(H,H)=6.4 Hz, 2H;
CH2), 5.20 5.50 (m, 2H; CH2), 5.67 ppm (s, 1H; CH); 13C NMR
(100.62 MHz, CDCl3): d=16.9 (q), 19.1 (t), 22.9 (q), 25.8 (t), 28.2 (q, 2î
), 33.8 (s), 38.9 (t), 60.4 (t), 123.9 (d), 125.2 (d), 127.6 (d), 136.5 (s), 142.2
(s), 144.6 ppm (s); IR (NaCl): n˜ =3600 3100, 2957, 2926, 1434, 1374,
999 cmꢀ1; UV (MeOH): lmax =234 nm; MS: m/z (%): 203 (6) [M+ꢀ17],
189 (11), 167 (18), 149 (88), 119 (100), 97 (82), 71 (86); HMRS m/z: calcd
for C15H24O: 220.1827; found: 220.1824.
(2E,4E)-3,4-Dimethyl-5-(2,6,6-trimethylcyclohex-1-en-1-yl)penta-2,4-
dien-1-ol ((E)-12b): Following the general procedure for the Suzuki cross
coupling, (E)-12b was obtained in 65% yield after purification by
column chromatography (silica gel, 80:15:2 hexane/EtOAc/Et3N). 1H
NMR (400.13 MHz, CDCl3): d=0.70 1.20 (m, 6H; 2CH3), 1.38 (s, 3H;
CH3), 1.40 1.60 (m, 4H; 2CH2), 1.60 (s, 3H; CH3), 1.82 (s, 3H; CH3),
1.90 2.00 (m, 2H; CH2), 4.26 (d, J(H,H)=6.6 Hz, 2H; CH2), 5.69 (t,
J(H,H)=6.6 Hz, 1H; CH), 6.04 ppm (s, 1H; CH); 13C NMR
(100.62 MHz, CDCl3): d=14.2 (q), 15.3 (q), 19.4 (t), 21.1 (q), 28.3 (q, 2î
), 31.9 (t), 34.9 (s), 39.2 (t), 60.1 (t), 124.4 (d), 126.2 (d), 128.1 (s), 136.3
(s), 137.7 (s), 139.1 ppm (s); IR (NaCl): n˜ =3600 3200, 2926, 2863, 1448,
1373, 1011 cmꢀ1; MS: m/z (%): 234 (11) [M+], 203 (17), 147 (17), 134
(12), 133 (100), 119 (21), 84 (15), 83 (11), 73 (60); HMRS m/z: calcd for
C16H26O: 234.1984; found: 234.1976.
(2E,4E)-5-(6,6-Dimethylcyclohex-1-en-1-yl)-3,4-dimethylpenta-2,4-dien-1-
ol ((E)-12a): Following the general procedure for the Shapiro Suzuki
cross coupling reaction, (E)-12a was obtained in 98% yield after purifi-
cation by column chromatography (silica gel, 80:15:2 hexane/EtOAc/
Et3N). 1H NMR (400.13 MHz, CDCl3): d=0.92 (s, 6H; 2CH3), 1.40 1.50
(m, 2H; CH2), 1.50 1.60 (m, 2H; CH2), 1.83 (s, 3H; CH3), 1.84 (s, 3H;
CH3), 1.90 2.00 (m, 2H; CH2), 4.30 (d, J(H,H)=6.7 Hz, 2H; CH2), 5.28
(td, J(H,H)=3.8, 1.4 Hz, 1H; CH), 5.73 (t, J(H,H)=6.7 Hz, 1H; CH),
6.13 ppm (s, 1H; CH); 13C NMR (100.62 MHz, CDCl3): d=14.6 (q), 15.8
(q), 19.6 (t), 26.3 (t), 28.7 (2q), 34.6 (s), 39.4 (t), 60.5 (t), 125.4 (d), 125.7
(d), 127.7 (d), 137.2 (s), 139.9 (s), 143.5 ppm (s); IR (NaCl): n˜ =3600
3200, 2930, 2865, 1455, 1377, 1019 cmꢀ1; UV (MeOH): lmax =247 nm.
MS: m/z (%): 221 (28) [M+], 1203 (7), 193 (13), 181 (31), 177 (21), 175
(26), 165 (31), 137 (27), 133 (21), 123 (18), 119 (10bb0), 109 (26), 108
(28), 105 (20), 95 (24); HMRS m/z: calcd for C15H24O: 220.1827; found:
220.1822.
General procedure for alcohol oxidation using TPAP/NMO
(2Z,4E)-5-(6,6-Dimethylcyclohex-1-en-1-yl)-3,4-dimethylpenta-2,4-dienal
((Z)-19a): A solution of (Z)-12a (0.22 g, 1.01 mmol) in CH2Cl2 (3 mL)
was added to a cooled (08C) and stirred suspension of N-methylmorpho-
line N-oxide (0.18 g, 1.51 mmol) and 4 ä molecular sieves in CH2Cl2
(6 mL). After stirring for 10 min, TPAP (0.017 g, 0.05 mmol) was added
and the mixture was stirred at 258C for 4 h. The mixture was diluted with
CH2Cl2 (10 mL) and was washed with aqueous Na2SO3 (3î). The organic
layer was dried over Na2SO4 and the solvent was evaporated. The residue
was purified by column chromatography (silica gel, 93:5:2 hexane/
EtOAc/Et3N) to afford
a
yellow oil (0.17 g, 77%). 1H NMR
(400.13 MHz, C6D6): d=0.93 (s, 6H; 2CH3), 1.40 1.60 (m, 4H; CH2),
1.60 (d, J(H,H)=1.3 Hz, 3H; CH3), 1.62 (d, J(H,H)=1.3 Hz, 3H; CH3),
1.80 2.00 (m, 2H; CH2), 5.30 (dt, J(H,H)=3.8, 1.3 Hz, 1H; CH), 5.92
(dq, J(H,H)=7.8, 1.3 Hz, 1H; CH), 5.95 (s, 1H; CH3), 10.04 ppm (d, J=
7.8 Hz, 1H; CH); 13C NMR (100.61 MHz, C6D6): d=16.6 (q), 19.4 (t),
23.1 (q), 26.1 (t), 28.2 (q, 2î), 33.9 (s), 39.0 (t), 126.6 (d), 129.4 (d), 132.3
(d), 135.2 (s), 141.9 (s), 164.6 (s), 191.1 ppm (d); IR (NaCl): n˜ =2958,
2929, 2865, 2833, 1678, 1611, 1384, 1140 cmꢀ1; UV (MeOH): lmax =234,
282, 338 nm; MS: m/z (%): 218 (18) [M+], 189 (34), 133 (19), 119 (100),
109 (28), 105 (18), 91 (17), 79 (15), 77 (16); HMRS m/z: calcd for
C15H22O: 218.1671; found: 218.1676.
2,6,6-Trimethylcyclohex-1-en-1-ylboronic acid (8b): tBuLi (2.34 mL, 1.7m
in pentane, 3.99 mmol) was added to a cooled (ꢀ788C) solution of alken-
yliodide 18[25] (0.48 g, 1.90 mmol) in THF (22 mL) at ꢀ788C, and the re-
sultant mixture was stirred for 30 min. Trimethylborate (2.16 mL,
18.9 mmol) was then added, and the reaction was stirred for 2 h at 258C.
Water (5 mL) was subsequently added, and the mixture was stirred for
2 h. The solution was diluted with Et2O, the layers were separated, and
the aqueous layers were extracted with tert-butyl methyl ether (TBDME;
3î). The combined organic layers were washed with 1n HCl (2î), dried
over Na2SO4, and the solvent was evaporated. The residue was purified
by column chromatography (silica gel, 80:20 hexane/EtOAc) to afford a
white solid (0.24 g, 77%) (m.p. 108 1098C, hexane/EtOAc). 1H NMR
(400.13 MHz, CD3OD): d=1.25 (s, 6H; 2CH3), 1.50 1.60 (m, 2H; CH2),
1.80 1.90 (m, 2H; CH2), 1.87 (s, 3H; CH3), 2.08 (t, J(H,H)=6.1 Hz, 2H;
CH2), 5.05 ppm (s, 2H; CH2); 13C NMR (100.62 MHz, CD3OD): d=21.0
(d), 25.1 (q), 31.1 (q, 2î), 34.5 (d), 35.7 (s), 39.7 (d), 135.1 ppm (s, 2î);
IR (NaCl): n˜ =3500 3300, 2925, 1321 cmꢀ1; MS: m/z (%): 169 (5)[M+
+1], 168 (33) [M+], 154 (15), 153 (99), 152 (23), 123 (13), 110 (12), 109
(2Z,4E)-3,4-Dimethyl-5-(2,6,6-trimethylcyclohex-1-en-1-yl)penta-2,4-
dienal ((Z)-19b): Following the general procedure for TPAP/NMO oxi-
dation, reaction of (Z)-12b (0.17 g, 0.71 mmol) with TPAP (0.011 g,
0.034 mmol) and NMO (0.13 g, 1.06 mmol) in CH2Cl2 (18 mL) afforded,
after purification by column chromatography (silica gel, 93:5:2 hexane/
EtOAc/Et3N), a yellow oil (0.15 g, 91%). 1H NMR (400.13 MHz, C6D6):
d=0.89 (s, 6H; 2CH3), 1.37 (s, 3H; CH3), 1.39 (d, J=1.2 Hz, 3H; CH3),
1.40 1.60 (m, 4H; 2CH2), 1.58 (s, 3H; CH3), 1.70 1.90 (m, 2H; CH2),
5.85 (t, J(H,H)=1.0 Hz, 1H; CH), 5.90 (dq, J(H,H)=7.9, 1.2 Hz, 1H;
CH), 10.12 ppm (d, J(H,H)=7.9 Hz, 1H; CH); 13C NMR (100.62 MHz,
C6D6): d=16.6 (q), 19.5 (t), 21.3 (q), 23.5 (q), 28.4 (q, 2î), 32.1 (t), 34.8
5828
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 5821 5831

(9Z)- and (11Z)-8-Methylretinals
5821 5831
(s), 39.3 (t), 129.4 (s), 129.5 (d), 131.2 (d), 135.2 (s), 136.5 (s), 164.1 (s),
190.9 ppm (d); IR (NaCl): n˜ =2927, 2864, 2830, 1679, 1612, 1438, 1383,
1139 cmꢀ1; UV (MeOH): lmax =226, 272, 334 nm; MS: m/z (%): 232 (4)
[M+], 203 (40), 161 (17), 147 (17), 134 (15), 133 (100), 119 (22), 109 (29),
105 (12), 91 (16), 83 (19), 77 (12); HMRS m/z: calcd for C16H24O:
232.1827; found: 232.1833.
1H; CH), 6.02 (d, J(H,H)=15.4 Hz, 1H; CH), 7.03 (dd, J(H,H)=15.4,
10.9 Hz, 1H; CH), 9.98 ppm (d, J(H,H)=7.9 Hz, 1H; CH); 13C NMR
(100.62 MHz, (CD3)2CO): d=14.1 (q), 18.2 (q), 20.9 (t), 24.7 (q),
27.5 (t), 29.7 (q, 2î), 35.7 (s), 40.6 (t), 127.2 (d), 127.3 (d), 130.6 (d),
130.9 (d), 134.8 (d), 135.8 (d), 138.9 (s), 143.9 (s), 151.4 (s), 156.2 (s),
192.2 ppm (d); IR (NaCl): n˜ =2927, 2850, 1663, 1594, 1449, 1196 cmꢀ1
;
UV (MeOH): lmax =345 nm; MS: m/z (%): 285 (24) [M++1], 284 (84)
[M+], 269 (37), 202 (25), 199 (21), 187 (54), 185 (28), 176 (44), 173 (68),
171 (35), 161 (69), 159 (57), 157 (37), 145 (47), 133 (39), 119 (100),
105 (38), 95 (64); HMRS m/z: calcd for C20H28O: 284.2140; found:
284.2135.
General procedure for Horner Wadsworth Emmons olefination
Ethyl (2E,4E,6Z,8E)-9-(6,6-dimethylcyclohex-1-en-1-yl)-3,7,8-trimethyl-
nona-2,4,6,8-tetraenoate ((Z)-20a): A cooled (08C) solution of 14 (0.22 g,
0.82 mmol) and DMPU (0.19 g, 1.61 mmol) in THF (2 mL) was treated
with nBuLi (2.03 mL, 2.35m in hexane, 0.87 mmol) and stirred for
20 min. The mixture was cooled to ꢀ788C and a solution of (Z)-19a
(0.13 g, 0.60 mmol) in THF (2 mL) was added. The resultant mixture was
allowed to warm to ꢀ408C, at which temperature H2O (4 mL) was
added. The reaction was extracted with Et2O (3î) and the organic layers
were washed with brine (3î), dried over Na2SO4, and the solvent was
evaporated. The residue was purified by column chromatography (silica
gel, 95:5 hexane/EtOAc) to afford a yellow oil (0.19 g, 95%). 1H NMR
(400.13 MHz, CDCl3): d=1.02 (s, 6H; 2CH3), 1.27 (t, J(H,H)=7.1 Hz,
3H; CH3), 1.40 1.60 (m, 4H; 2CH2), 1.81 (d, J(H,H)=1.4 Hz, 3H; CH3),
1.93 (s, 3H; CH3), 2.00 2.20 (m, 2H; CH2), 2.27 (d, J(H,H)=0.8 Hz, 3H;
CH3), 4.16 (q, J(H,H)=7.1 Hz, 2H; CH2), 5.30 5.50 (td, J(H,H)=3.8,
1.3 Hz, 1H; CH), 5.73 (s, 1H; CH), 5.82 (d, J(H,H)=1.3 Hz, 1H; CH),
5.95 (d, J(H,H)=11.0 Hz, 1H; CH), 6.19 (d, J(H,H)=15.3 Hz, 1H; CH),
6.92 ppm (dd, J(H,H)=15.3, 11.0 Hz, 1H; CH); 13C NMR (100.62 MHz,
CDCl3): d=13.8 (q), 14.3 (q), 16.9 (q), 19.2 (t), 23.6 (q), 25.9 (t), 28.3 (q,
2î), 34.0 (s), 38.9 (t), 59.5 (t), 117.9 (d), 125.2 (d), 125.3 (d), 129.1 (d),
132.9 (d), 133.2 (d), 136.7 (s), 142.2 (s), 148.4 (s), 153.2 (s), 167.2 ppm (s);
IR (NaCl): n˜ =2959, 2931, 2866, 1711, 1603, 1444, 1350, 1239, 1150,
967 cmꢀ1; UV (MeOH): lmax =320 nm; MS: m/z (%): 328 (100) [M+],
313 (47), 267 (19), 255 (25), 201 (65), 185 (33), 175 (24), 173 (51), 171
(38), 161 (36), 139 (57); HMRS m/z: calcd for C22H32O2: 328.2402;
found: 328.2387.
(2E,4E,6Z,8E)-3,7,8-Trimethyl-9-(2,6,6-trimethylcyclohex-1-en-1-yl)nona-
2,4,6,8-tetraenal (6) [(9Z)-8-methylretinal]: Following the general com-
bined procedure for DIBAL-H reduction/MnO2 oxidation, compound 6
was obtained in 75% yield after purification by column chromatography
(silica gel, 95:5 hexane/EtOAc). 1H NMR (400.13 MHz, C6D6): d=1.01
(s, 6H; CH3), 1.40 1.50 (m, 2H; CH2), 1.54 (s, 3H; CH3), 1.58 (s, 3H;
CH3), 1.60 1.70 (m, 2H; CH2), 1.78 (s, 3H; CH3), 1.80 (s, 3H; CH3),
1.90 2.00 (m, 2H; CH2), 5.81 (d, J(H,H)=11.0 Hz, 1H; CH), 5.85 (s, 1H;
CH), 6.00 (d, J(H,H)=7.9 Hz, 1H; CH), 6.02 (d, J(H,H)=15.4 Hz, 1H;
CH), 7.08 (dd, J(H,H)=15.4, 11.0 Hz, 1H; CH), 9.99 ppm (d, J(H,H)=
7.9 Hz, 1H; CH); 13C NMR (100.62 MHz, (CD3)2CO): d=14.2 (q), 18.2
(q), 21.0 (t), 22.5 (q), 25.0 (q), 29.8 (q, 2î), 33.6 (t), 36.5 (s), 40.9 (t),
127.3 (d), 129.4 (d), 130.4 (s), 130.6 (d), 134.9 (d), 135.9 (d), 137.0 (s),
140.3 (s), 151.1 (s), 156.2 (s), 192.2 ppm (d); UV (MeOH): lmax =344 nm;
MS: m/z (%): 299 (23) [M++1], 298 (100) [M+], 216 (20), 213 (20), 201
(37), 187 (24), 173 (20), 171 (23), 159 (37), 147 (27), 145 (28), 91 (28);
HMRS m/z: calcd for C21H30O: 298.2297; found: 298.2299.
(2E,4E)-5-(6,6-Dimethylcyclohex-1-en-1-yl)-3,4-dimethylpenta-2,4-dienal
((E)-21a): Following the general procedure for oxidation of alcohols with
TPAP/NMO, reaction of (E)-12a (0.15 g, 0.68 mmol) with TPAP (0.012 g,
0.034 mmol) and NMO (0.12 g, 1.02 mmol) in CH2Cl2 (7 mL) afforded,
after purification by column chromatography (silica gel, 95:5 hexane/
EtOAc), (E)-21a as a yellow oil (0.14 g, 92%). 1H NMR (400.13 MHz,
CD2Cl2): d=1.04 (s, 6H; 2CH3), 1.50 1.60 (m, 2H; CH2), 1.60 1.70 (m,
2H; CH2), 1.95 (d, J(H,H)=1.1 Hz, 3H; CH3), 2.12 (m, 2H; CH2), 2.35
(d, J(H,H)=0.9 Hz, 3H; CH3), 5.45 (td, J(H,H)=3.9, 1.5 Hz, 1H; CH),
6.08 (d, J(H,H)=7.9 Hz, 1H; CH), 6.67 (s, 1H; CH), 10.16 ppm (d,
J(H,H)=7.9 Hz, 1H; CH); 13C NMR (100.62 MHz, CD2Cl2): d=14.5 (q),
15.5 (q), 19.4 (t), 26.3 (t), 28.4 (q, 2î), 34.5 (s), 39.2 (t), 126.5 (d), 127.4
(d), 134.1 (d), 137.0 (s), 143.1 (s), 158.7 (s), 192.2 ppm (d); UV (MeOH):
lmax =305 nm; MS: m/z (%): 218 (9) [M+], 203 (11), 189 (23), 179 (15),
148 (12), 133 (16), 119 (100), 105 (11), 91 (12); HMRS m/z: calcd for
C15H22O: 218.1671; found: 218.1674.
Ethyl
(2E,4E,6Z,8E)-3,7,8-trimethyl-9-(2,6,6-trimethylcyclohex-1-en-1-
yl)nona-2,4,6,8-tetraenoate ((Z)-20b): Following the general procedure
for HWE olefination, (Z)-20b was obtained in 94% yield after purifica-
tion by column chromatography (silica gel, 95:5 hexane/EtOAc). 1H
NMR (400.13 MHz, CDCl3): d=1.01 (s, 6H; 2CH3), 1.28 (t, J(H,H)=
7.1 Hz, 3H; CH3), 1.40 1.70 (m, 4H; 2CH2), 1.60 (s, 3H; CH3), 1.62 (d,
J(H,H)=1.2 Hz, 3H; CH3), 1.96 (s, 3H; CH3), 2.00 2.20 (m, 2H; CH2),
2.27 (d, J(H,H)=0.8 Hz, 3H; CH3), 4.16 (t, J(H,H)=7.1 Hz, 2H; CH2),
5.74 (s, 1H; CH), 5.76 (s, 1H; CH), 5.98 (d, J(H,H)=10.9 Hz, 1H; CH),
6.20 (d, J(H,H)=15.4 Hz, 1H; CH), 7.00 ppm (dd, J(H,H)=15.4,
10.9 Hz, 1H; CH); 13C NMR (100.62 MHz, CDCl3): d=13.8 (q), 14.3 (q),
16.8 (q), 19.3 (t), 21.1 (q), 23.8 (q), 28.4 (q, 2î), 32.0 (t), 34.8 (s), 39.1 (t),
59.5 (t), 117.9 (d), 125.2 (d), 127.7 (d), 128.4 (s), 133.0 (d), 133.1 (d),
135.3 (s), 138.0 (s), 148.0 (s), 153.1 (s), 167.2 ppm (s); IR (NaCl): n˜ =
2960, 2928, 2864, 1710, 1603, 1442, 1238, 1151, 969 cmꢀ1; UV (MeOH):
lmax =322 nm; MS: m/z (%): 342 (100) [M+], 327 (20), 281 (21), 269 (23),
215 (22), 199 (39), 187 (27), 185 (33), 159 (26), 157 (19), 139 (61), 133
(25), 119 (19), 105 (14), 91 (16); HMRS m/z: calcd for C23H34O2:
342.2559; found: 342.2559.
(2E,4E)-3,4-Dimethyl-5-(2,6,6-trimethylcyclohex-1-en-1-yl)penta-2,4-
dienal ((E)-21b): Following the general procedure for TPAP/NMO oxi-
dation, reaction of (E)-12b (0.061 g, 0.26 mmol) with TPAP (0.004 g,
0.011 mmol) and NMO (0.05 g, 0.39 mmol) in CH2Cl2 (3 mL) afforded,
after purification by column chromatography (silica gel, 95:5 hexane/
EtOAc), (E)-21b as a yellow oil (0.04 g, 67%). 1H NMR (400.13 MHz,
CDCl3): d=0.80 1.10 (brm, 6H; 2CH3), 1.45 (s, 3H; CH3), 1.40 1.50
(brt, J(H,H)=5.7 Hz, 2H; CH2), 1.60 1.70 (m, 2H; CH2), 1.72 (s, 3H;
CH3), 1.90 2.00 (m, 2H; CH2), 2.36 (s, 3H; CH3), 6.13 (d, J(H,H)=
7.8 Hz, 1H; CH), 6.62 (s, 1H; CH), 10.17 ppm (d, J(H,H)=7.8 Hz, 1H;
CH); 13C NMR (100.62 MHz, CDCl3): d=14.3 (q), 15.2 (q), 19.2 (t), 21.1
(q), 28.4 (q, 2î), 31.9 (t), 35.0 (s), 39.0 (t), 125.6 (d), 129.5 (s), 133.6 (d),
135.7 (s), 137.7 (s), 157.6 (s), 192.1 ppm (d); UV (MeOH): lmax =289 nm;
MS: m/z (%): 232 (3) [M+], 205 (12), 203 (20), 191 (16), 177 (15), 162
(17), 161 (15), 147 (21), 134 (100), 119 (32), 69 (84); HMRS m/z: calcd
for C16H24O: 232.1827; found: 232.1824.
General procedure for reduction of esters oxidation of alcohols
(2E,4E,6Z,8E)-9-(6,6-Dimethylcyclohex-1-en-1-yl)-3,7,8-trimethylnona-
2,4,6,8-tetraenal (7) [(9Z)-5-demethyl-8-methylretinal]: DIBAL-H
(0.7 mL, 1m in hexane, 0.7 mmol) was added to a solution of (Z)-20a
(0.06 g, 0.18 mmol) in THF (2 mL) at ꢀ788C, and the resultant suspen-
sion was stirred for 2 h. After careful addition of H2O, the mixture was
extracted with Et2O (3î), and the organic layers were dried over Na2SO4
and concentrated. The residue was oxidized without further purification.
General procedure for Wittig olefination and oxidation
Manganese dioxide (0.27 g, 3.08 mmol) and Na2CO3 (0.33 g, 3.08 mmol)
were added to a solution of the above compound in CH2Cl2 (4 mL), and
the suspension was stirred for 4 h. The mixture was filtered throught
Celite and the solvent was removed. The residue was purified by column
chromatography (silica gel, 95:5 hexane/EtOAc) to afford a yellow oil
(0.04 g, 79%). 1H NMR (400.13 MHz, C6D6): d=1.01 (s, 6H; 2CH3),
1.40 1.50 (m, 2H; CH2), 1.50 1.60 (m, 2H; CH2), 1.75 (d, J(H,H)=
1.3 Hz, 3H; CH3), 1.76 (s, 3H; CH3), 1.80 (d, J(H,H)=1.0 Hz, 3H; CH3),
1.90 2.00 (m, 2H; CH2), 5.46 (td, J(H,H)=3.9, 1.4 Hz, 1H; CH), 5.80 (d,
J(H,H)=10.9 Hz, 1H; CH), 5.94 (s, 1H; CH), 5.98 (d, J(H,H)=7.9 Hz,
(2E,4Z,6E,8E)-9-(6,6-Dimethylcyclohex-1-en-1-yl)-3,7,8-trimethylnona-
2,4,6,8-tetraenal (5) [(11Z)-5-demethyl-8-methylretinal]: KHMDS
(2.52 mL, 0.5m in toluene, 1.23 mmol) was added to a cooled (ꢀ788C)
suspension of phosphonium salt 22 (0.25 g, 0.56 mmol) in THF (4 mL).
After stirring for 10 min at ꢀ788C and 1 h at 258C, the mixture was
cooled to ꢀ788C and a solution of (E)-21a (0.14 g, 0.62 mmol) in THF
(4 mL) was added. The resultant mixture was stirred for 30 min at ꢀ788C
and for 30 min at 258C. The mixture was poured over H2O and extracted
with Et2O (4î). The combined organic layers were dried over Na2SO4
5829
Chem. Eur. J. 2003, 9, 5821 5831
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

A. R. de Lera et al.
FULL PAPER
and the solvent was evaporated. The residue was purified by column
chromatography (80:18:2 hexane/EtOAc/Et3N) to afford a yellow oil,
which was used immediately in the next reaction.
Acknowledgements
We thank the Spanish Ministerio de Ciencia y TecnologÌa (Grant SAF01-
3288: RamÛn y Cajal Research Contract to R.A.; FPU Fellowship to
M.D.) and Xunta de Galicia (Grant PGIDIT02PXIC30108PN) for finan-
cial support. We are indebted to Prof. Dave Kliger and Dr. Jim Lewis
(UC Santa Cruz) for opsin-binding experiments, and Prof. K. Palczewski
(University of Washington) for allowing us access to his lab facilities
during a summer stay by M.D.
Manganese dioxide (0.49 g, 5.66 mmol) and Na2CO3 (0.60 g, 5.66 mmol)
were added to a solution of the above compound in CH2Cl2 (15 mL), and
the suspension was stirred at room temperature for 3 h. The mixture was
filtered through Celite and the solvent was removed. The residue was pu-
rified by column chromatography (silica gel, 94:3:3 hexane/EtOAc/Et3N)
to afford a yellow oil (0.09 g, 51%). 1H NMR (400.13 MHz, C6D6): d=
1.03 (s, 6H; 2CH3), 1.40 1.70 (m, 4H; 2CH2), 1.81 (d, J(H,H)=0.8 Hz,
3H; CH3), 1.82 (s, 3H; CH3), 1.93 (d, J(H,H)=0.8 Hz, 3H; CH3), 1.90
2.00 (m, 2H; CH2), 5.46 (td, J(H,H)=3.8, 1.5 Hz, 1H; CH), 5.62 (d,
J(H,H)=11.7 Hz, 1H; CH), 6.13 (d, J(H,H)=7.7 Hz, 1H; CH), 6.40 6.50
(m, 1H; CH), 6.45 (t, J(H,H)=11.8 Hz, 1H; CH), 6.82 (d, J(H,H)=
11.8 Hz, 1H; CH), 9.98 ppm (d, J(H,H)=7.7 Hz, 1H; CH); 13C NMR
(100.62 MHz, C6D6): d=14.0 (q), 15.6 (q), 17.5 (q), 19.5 (t), 26.2 (t), 28.5
(q, 2î), 34.5 (s), 39.2 (t), 122.9 (d), 126.2 (d), 129.7 (d), 130.6 (d), 131.2
(d), 131.6 (d), 137.7 (s), 143.2 (s), 143.3 (s), 154.1 (s), 189.8 ppm (d); IR
(NaCl): n˜ =2928, 1662, 1594, 1445, 1116 cmꢀ1; UV (MeOH): lmax =255,
371 nm; MS: m/z (%): 284 (54) [M+], 269 (23), 202 (21), 187 (47), 176
(39), 173 (58), 171 (28), 161 (60), 159 (51), 145 (42), 133 (37), 119 (100),
105 (34), 95 (67), 91 (43); HMRS m/z: calcd for C20H28O: 284.2140;
found: 284.2150.
[1] T. Gudemann, T. Schoneberg, G. Schultz, Annu. Rev. Neurosci.
1997, 20, 399 427.
[2] For recent reviews, see: a) R. R. Rando, Chem. Rev. 2001, 101,
1881 1896; b) K. J. McBee, K. Palczewski, W. Baehr, D. R. Pepper-
berg, Prog. Retinal. Eye Res. 2001, 20, 469 529.
[3] W. R. Schoenlein, A. L. Peteann, A. R. Mathies, C. V. Shank,
Science 1991, 254, 412 415.
[4] L. Stryer, J. Biol. Chem. 1990, 265 266, 10711 10714.
[5] P. T. Sakmar, R. R. Franke, H. G. Khorana, Proc. Natl. Acad. Sci.
USA 1989, 86, 8309 8313.
[6] a) K. Palczewski, T. Kumasaka, T. Hori, C. A. Behnke, H. Motoshi-
ma, B. A. Fox, I. Le Trong, D. C. Teller, T. Okada, R. E. Stenkamp,
M. Yamamoto, M. Miyano, Science 2000, 289, 739 745; b) T.
Okada, K. Palczewski, Curr. Opin. Struct. Biol. 2001, 11, 420 426;
c) D. C. Teller, T. Okada, C. A. Behnke, K. Palczewski, R. Sten-
kamp, Biochemistry 2001, 40, 7761 7762.
[7] a) W. G‰rtner, Angew. Chem. 2001, 113, 3065 3069; Angew. Chem.
Int. Ed. 2001, 40, 2977 2981; b) L.-O. Essen, ChemBioChem 2001,
2, 513 516.
[8] G. Grˆbner, I. J. Burnett, C. Glaubitz, G. Choi, A. J. Mason, A.
Watts, Nature 2000, 405, 810 813; a recent 1H and 13C MAS NMR
study using 99% enriched uniformily 13C-labeled 11-cis-retinal con-
firmed the 6-s-cis conformation of the chromophore: A. F. L.
Creemers, S. Kiihne, P. H. M. Bovee-Geurts, W. J. DeGrip, J. Lug-
tenburg, H. J. M. Groot, Proc. Natl. Acad. Sci. USA 2002, 99, 9101
9106.
[9] K. Nakanishi, R. Crouch, Isr. J. Chem. 1995, 35, 253 272.
[10] The twisted polyene chain of 11-cis-retinal 1 within the binding
pocket is intrinsically chiral. The absolute sense of twist around the
6-s-cis bond has been determined as negative with the use of locked
synthetic retinal analogues; see: Y. Fujimoto, J. Ishihara, S. Maki, N.
Fujioka, T. Wang, T. Furuta, N. Fishkin, B. Borham, N. Berova, K.
Nakanishi, Chem. Eur. J. 2001, 7, 4198 4204. See also the discussion
in the same paper on the positive helicity about the 12-s-bond.
[11] A. Albeck, N. Friedman, M. Ottolenghi, M. Sheves, C. M. Einterz,
S. J. Hug, J. W. Lewis, D. S. Kliger, Biophys. J. 1989, 55, 233 241.
[12] G.-F. Jang, V. Kuksa, S. Filipek, F. Barti, E. Ritter, M. Gelb, P. K.
Hofmann, K. Palczewski, J. Biol. Chem. 2001, 276, 26148 26153.
[13] Recent photoaffinity labeling studies using the photolysis of diazo
groups attached to the hydrophobic ring appear to be consistent
with this proposal: B. Borhan, M. L. Souto, H. Imai, Y. Shichida, K.
Nakanishi, Science 2000, 288, 2209 2212.
[14] Substitution of the retinal side chain at C8 is uncommon. For previ-
ous studies on the interaction of a trans-retinal that was modified at
C8 with bacteriorhodopsin, which is the protein of the purple mem-
brane of Halobacterium salinarum; see: a) A. R. de Lera, B. Iglesias,
J. RodrÌguez, R. Alvarez, S. LÛpez, J. Villanueva, E. PadrÛs, J. Am.
Chem. Soc. 1995, 117, 8220 8231; b) 6-s-cis-locked retinoids with a
ring that spans the C18 and C8 positions have been described: Y.
Katsuta, Y. Aoyama, H. Ozone, A. Wada, M. Ito, J. Chem. Soc.
Perkin Trans. 1 1997, 1405 1410; c) Liu has also described retinals
fluorinated at C8, see: C. J. Simmons, L. U. Colmenares, R. S. H.
Liu, Tetrahedron Lett. 1996, 37, 4103 4104.
(2E,4Z,6E,8E)-3,7,8-Trimethyl-9-(2,6,6-trimethylcyclohex-1-en-1-yl)nona-
2,4,6,8-tetraenal (4) [(11Z)-8-methylretinal]: Following the general proce-
dure for Wittig olefination and oxidation, compound 4 was obtained in
60% yield after purification by column chromatography (silica gel,
94:3:3 hexane/EtOAc/Et3N). 1H NMR (400.13 MHz, CDCl3): d=0.80
1.10 (brm, 6H; 2CH3), 1.49 (s, 6H; 2CH3), 1.50 1.60 (m, 2H; CH2),
1.70 1.80 (m, 2H; CH2), 1.83 (s, 3H; CH3), 1.86 (s, 3H; CH3), 1.90 2.00
(m, 2H; CH2), 5.66 (d, J(H,H)=11.6 Hz, 1H; CH), 6.18 (d, J(H,H)=
7.5 Hz, 1H; CH), 6.35 (s, 1H; CH), 6.48 (t, J(H,H)=11.6 Hz, 1H; CH),
6.87 (d, J(H,H)=11.6 Hz, 1H; CH), 9.97 ppm (d, J(H,H)=7.5 Hz, 1H;
CH); 13C NMR (100.63 MHz, (CD3)2CO): d=14.2 (q), 15.7 (q), 18.1 (q),
20.1 (d), 21,5 (q), 28.9 (q, 2î), 32.5 (d), 35.7 (s), 39.9 (d), 123.0 (d), 129.2
(d), 130.9 (d), 131.9 (d), 132.4 (d), 137.2 (s), 139.4 (s, 2î), 143.5 (s), 156.2
(s), 191.4 ppm (d); IR (NaCl): n˜ =2923, 2852, 1661, 1594, 1463 cmꢀ1; UV
(MeOH): lmax =253, 348, 368 nm; MS: m/z (%): 298 (28) [M+], 284 (14),
201 (25), 175 (28), 173 (35), 171 (29), 159 (58), 147 (34), 145 (42), 133
(99), 119 (63), 109 (65), 97 (60), 95 (74), 83 (51), 73 (61), 71 (61),
69 (100), 67 (60); HMRS m/z: calcd for C21H30O: 298.2297; found:
298.2292.
Calculation methods: The structure of bovine rhodopsin (opsin bound to
11-cis-retinal), pdb file 1HXZ, was used as the starting point for our cal-
culations. Hydrogen atoms were added and a CHARMM potential was
used throughout our calculations.[37] The unbound structures were mod-
eled by excising the retinal fragment from the complex and converting
the polar end of the polyene side chain to an aldehyde using the Builder
module found in the InsightII suite of programs (from Accelrys, Inc.).[43]
The retinal analogues, in their free and bound forms, were modeled by
adding (4) and then deleting (5) methyl groups from the original struc-
ture. To avoid having to deal with the residue gaps observed in the pro-
tein surface of the original crystallographic structure, the rhodopsin com-
plexes with retinals 1, 4, and 5 were optimized only 10 ä around the re-
action center; this left the surface gaps out of the calculation. All the rho-
dopsin structures were energy minimized by using the Adopted Basis
Newton Raphson (ABNR) protocol.[44]
The final structure of the unbound retinals was obtained from a 120 ps
NVT molecular dynamic (MD) simulation. The MD production stage
lasted 100 ps and generated a set of 100 frames that were ordered by
their internal energy. The lowest energy conformer obtained in the pro-
duction stage was taken as an energy minimum that yielded the model
used in our thermodynamic-cycle calculations.
Rotation barrier calculation: The rotation barrier around the C6ꢀC7
bond in compound 6 was evaluated using the CHARMM molecular mod-
eling suite.[37] A set of rotamers in 108 increments around this bond
was set up for the unbound analogue, and an energy minimization, in
which the torsion angle around this bond was frozen, was carried out for
all conformers. In all cases, an energy tolerance of 10ꢀ4 to 10ꢀ5 was
reached.
[15] T. Yoshizawa, G. Wald, Nature 1963, 197, 1279 1286.
[16] Y. Pazos, A. R. de Lera, Tetrahedron Lett. 1999, 40, 8287 8290.
[17] a) N. Miyaura, A. Suzuki, Chem. Rev. 1995, 95, 2457 2483; b) A.
Suzuki, J. Organomet. Chem. 1999, 576, 147 168.
[18] a) N. Okukado, E. Negishi, Tetrahedron Lett. 1978, 19, 2357 2360;
b) E. Negishi, L. F. Valente, M. Kobayashi, J. Am. Chem. Soc. 1980,
102, 3298 3299; c) E. Negishi, A. O. King, W. L. Klima, W. Patter-
5830
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 5821 5831

(9Z)- and (11Z)-8-Methylretinals
5821 5831
son, A. Silveira, J. Org. Chem. 1980, 45, 2526 2528; d) C. L. Rand,
D. E. Van Harn, M. W. Moore, E. Negishi, J. Org. Chem. 1981, 46,
4093 4096; e) E. Negishi, Pure Appl. Chem. 1981, 53, 2333 2357.
[19] a) A. F. Kluge, C. P. Lillya, J. Am. Chem. Soc. 1971, 93, 4458 4463;
b) R. A. S. Chandraratna, W. H. Okamura, Tetrahedron Lett. 1984,
25, 1003 1006.
[34] a) H. Leucke, B. Schobert, H.-T. Richter, J.-P. Cartailler, J. K. Lanyi,
J. Mol. Biol. 1999, 291, 899 911; b) A. Royant, K. Edman, T. Ursby,
E. Pebay- Peyroula, E. M. Landau, R. Neutze, Nature 2000, 406,
645 648; c) H. J. Sass, G. B¸dlt, R. Gessenich, D. Hehn, D. Neff, R.
Schlessinger, J. Berendzen, P. Ormos, Nature 2000, 406, 649 652.
[35] a) F. Terstegen, V. Buss, Chem. Phys. 1997, 225, 163 171; for other
ab initio studies of the same group on 11-cis-retinal and its Schiff
base, see: b) F. Terstegen, V. Buss, THEOCHEM 1998, 430, 209
218; c) V. Buss, K. Kolster, F. Terstegen, R. Vahrenhorst, Angew.
Chem. 1998, 110, 1997 2000; Angew. Chem. Int. Ed. 1998, 37,
1893 1895; d) V. Buss, O. Weingart, M. Sugihara, Angew. Chem.
2000, 112, 2895 2898; Angew. Chem. Int. Ed. 2000, 39, 2784 2786.
[36] The dynamic behaviour of 8-methyl-b-ionone observed by 13C NMR
was also interpreted as arising from the conformational interconver-
sion of the ring side chain: K. M¸llen, H. Schmickler, B. Frei, H. R.
Wolf, Tetrahedron Lett. 1986, 27, 477 480.
[37] B. R. Brooks, E. R. Bruccoleri, B. D. Olafson, D. J. States, S. Swami-
nathan, M. Karplus, J. Comput. Chem. 1983, 4, 187 217.
[38] R. Hubbard, P. K. Brown, D. Bownds, Methods Enzymol. 1971, 18,
628 653.
[39] I. Szundi, A. R. de Lera, Y. Pazos, R. Alvarez, M. Oliana, M.
Sheves, J. Lewis, D. S. Kliger, Biochemistry 2002, 41, 2028 2035.
[40] a) T. Simonson, G. Archontis, M. Karplus, Acc. Chem. Res. 2002, 35,
430 437; b) J. äquist, V. B. Luzhkov, B. O. Brandsdal, Acc. Chem.
Res. 2002, 35, 358 365.
[20] A. J. Fatiadi, Synthesis 1976, 65 104.
[21] a) Y. L. Bennani, M. F. Boehm, J. Org. Chem. 1995, 60, 1195 1200;
b) Y. L. Bennani, J. Org. Chem. 1996, 61, 3542 3544; c) B. DomÌn-
guez, B. Iglesias, A. R. de Lera, J. Org. Chem. 1998, 63, 4135 4139.
[22] B. H. Lipshutz, C. Lindsley, J. Am. Chem. Soc. 1997, 119, 4555
4556.
[23] F. Derguini, V. Balog-Nair, K. Nakanishi, Tetrahedron Lett. 1979, 20,
4899 4902.
[24] R. H. Shapiro, Org. React. 1976, 23, 405 507.
[25] Y. Pazos, B. Iglesias, A. R. de Lera, J. Org. Chem. 2001, 66, 8483
8489.
[26] J. Uenishi, J.-M. Beau, R. W. Armstrong, Y. Kishi, J. Am. Chem.
Soc. 1987, 109, 4756 4768.
[27] D. H. R. Barton, M. Chen, J. C. Jaszberenyi, D. K. Taylor, Org.
Synth. 1997, 74, 101 107.
[28] W. P. Griffith, S. V. Ley, G. P. Withcombe, A. D. White, J. Chem.
Soc. Chem. Commun. 1987, 1625 1627.
[29] A. Hosoda, T. Taguchi, Y. Kobayashi, Tetrahedron Lett. 1987, 28,
65 68.
[30] H. G¸nther, NMR Spectroscopy-An Introduction. Wiley, New York,
1980 Chapter 8, pp. 234 280.
[41] W. C. Still, A. Tempczyk, R. C. Hawley, T. J. Hendrickson, J. Am.
Chem. Soc. 1990, 112, 6127 6129.
[31] C. J. Simmons, R. S. H. Liu, M. Denny, K. Seff, Acta Crystallogr.
Sect. B 1981, 37, 2197 2205; ) C. J. Simmons, A. E. Asato, R. S. H.
Liu, Acta Crystallogr. Sect. C 1986, 42, 711 715, 1558 1563.
[32] R. Gilardi, Karle, L. I. J. Karle, W. Sperling, Nature 1971, 232, 187
189.
[33] However, upon binding the apoproteins, a selection of 6-s-cis or 6-s-
trans conformations is made by the receptor depending upon the tol-
erance of its binding pocket (i.e., 6-s-cis for 11-cis-retinal in rhodop-
sin ref. [6], and 6-s-trans for trans-retinal in bacteriorhodopsin
ref. [34], which is the protein of the purple membrane of Halobacte-
rium salinarum).
[42] K. A. Sharp, A. Nicholls, R. F. Fine, B. Honig, Science 1991,
252,106 109.
[43] InsightII, Builder and Biopolymer are trademarks of Accelrys, Inc.
San Diego, CA (USA).
[44] J. A. McCammon, S. C. Harvey, Dynamics of Proteins and Nucleic
Acids, Cambridge University Press, Cambridge, 1988.
Received: February 13, 2003
Revised: July 23, 2003 [F4847]
5831
Chem. Eur. J. 2003, 9, 5821 5831
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim