A 19F NMR study of rhodopsin analogs: use of vinylfluororetinal chromophores

Article abstract of DOI:10.1021/jp952850k

¹⁹F NMR spectra of 11-cis and 9-cis isomers of six fluorinated rhodopsin analogs with the label(s) located at the vinylic positions of the polyene chain (8-F, 10-F, 12-F, 14-F, 8,12-F2, 10,14-F2) are reported along with their UV-vis and CD spec

Full text of DOI:10.1021/jp952850k

J. Phys. Chem. 1996, 100, 9175-9180
9175
19
A F NMR Study of Rhodopsin Analogs: Use of Vinylfluororetinal Chromophores
Leticia U. Colmenares, Walter P. Niemczura, Alfred E. Asato, and Robert S. H. Liu*
Department of Chemistry, 2545 The Mall, UniVersity of Hawaii, Honolulu, Hawaii 96822
ReceiVed: September 26, 1995; In Final Form: March 13, 1996X
19F NMR spectra of 11-cis and 9-cis isomers of six fluorinated rhodopsin analogs with the label(s) located at
the vinylic positions of the polyene chain (8-F, 10-F, 12-F, 14-F, 8,12-F2, 10,14-F2) are reported along with
their UV-vis and CD spectra. The regiospecific F chemical shift data are analyzed in terms of chromophore
changes and local perturbation resulting from specific interactions with protein. Two analogs (11-cis-12-F
and 11-cis-8-F) and also 9,11-di-cis-12-F possess FOS (fluorine opsin shift) values uniquely different from
others. Ab initio F NMR chemical shielding calculations of model structures provide support that a strong
protein perturbation to the 12-F position prevails in the binding cavity and that F-8 shift is sensitive to variation
of the nearby dihedral angle(s). Possible causes for the broad line width of the F signals of these membrane
proteins are discussed.
Introduction
were obtained using the GE QE-300 or GN-Omega-500
spectrometer. Assignment of chemical shifts was based on
coupling constants and decoupling experiments.
The application of 19F NMR spectroscopy in studying the
structure and function of proteins is widely known.1 In most
cases, fluorine is introduced into the protein by using fluorinated
amino acid. In a recent study of the visual pigment rhodopsin,2
we reported preliminarily the use of fluorine-labeled chro-
mophores (substrate) instead. By introducing fluorine as a
reporting group at different parts of the chromophore, the local
environment around the chromophore within the binding pocket
of rhodopsin is probed. And since fluorine replaces hydrogen,
it probes into regions beyond those occupied by the carbon
frame work. Thus the method does not duplicate the well-
established 13C NMR work.3,4 We now present a complete
report on chain-labeled retinal chromophores, including difluoro
as well as monofluoro analogs.
Rhodopsin is the visual pigment in the rod photoreceptor cell
that is responsible for scotopic (dim light) vision in vertebrates.
It consists of a 348-residue polypeptide chain that shapes into
seven R-helices interconnected by water-soluble loops, its
interior being a hydrophobic cavity where the chromophore (11-
cis-retinal) resides with a protonated Schiff base (PSB) linkage
attached to Lys-296. Upon light absorption, the chromophore
isomerizes to the all-trans geometry, thereby triggering its
expulsion from the cavity and an eventual neural signal. Much
useful information on structural properties related to protein-
chromophore interactions in rhodopsin have been obtained from
electron diffraction,5 X-ray diffraction,6 and electron cryomi-
croscopy studies.7 Abbreviations used in this paper are as
follows: FOS, 19F NMR opsin shift; OS, opsin shift; PSB,
protonated Schiff base; SB, Schiff base.
The 19F NMR spectra were taken using either the Nicolet
NT-300 (operating at 283 MHz) or the GE GN-Omega-500
(operating at 470 MHz). All 19F NMR chemical shifts were
referenced to an external standard, 20% (v/v) CF3CCl3 in CDCl3
(-82.2 ppm). Spectra were recorded with 16K data points over
a variable width ranging from 6 to 20 kHz. These were obtained
with a 20 µs pulse width and usually a 1 s delay between pulses.
For protein samples, a total of less than 10 000 free induction
decays were accumulated (block averaged), zero-filled, and then
Fourier transformed with a line broadening of 80 Hz. Line
widths were obtained by fitting the observed spectra to a
Lorentzian curve using the LF (line fitting) routine. Unless
otherwise stated, the probe temperature was 20 °C.
UV-vis spectra were taken using the Perkin-Elmer λ-19
spectrophotometer. Circular dichroic spectra were obtained
using a Jasco J-600 spectropolarimeter by taking the difference
between those before and after photobleaching. UV-vis and
CD experiments were carried out using samples of approxi-
mately 1.0 absorbance and at 20 °C. The scan speed for
acquisition of CD spectra is 50 nm/min. For photobleaching,
a 500-W tungsten light source was used with an appropriate
glass cutoff filter.
Synthesis of Fluorine-Labeled Retinals, SB and PSB. The
retinal analogs used in this study are the following: 8-F, 10-F,
12-F, 14-F, 8,12-F2-, and 10,14-F2. The reaction schemes for
the syntheses of the individual fluorine-labeled retinals are
essentially the same as the ones outlined.8 The 11-cis and 9-cis
isomers present in the synthetic mixture or generated by
photoisomerization were individually purified by HPLC.
Possible formation of fluorinated visual pigment analogs with
the retinyl chromophore replaced by F-labeled retinals was
demonstrated by this group.8 Stereoselectivity of opsin (i.e.,
its preference for 11-cis and 9-cis isomers, a lesser degree for
the 7-cis and unreactive toward 13-cis)9 has been retained by
the fluorinated retinal analogs,8 thus giving some assurance that
the same binding site is involved in the labeled analogs as well
as the parent retinal. As a result, we are able to use fluorine as
an alternate reporting group for information related to the
binding site of rhodopsin.
The n-butylretinylidene SB’s were prepared by adding 3 µL
of n-butylamine to about 5 mg of the freshly purified retinal in
2 mL of dry hexane. The solution was kept at 0 °C until the
reaction was complete (monitored by UV-vis). Excess n-
butylamine was removed by evaporation under vacuum. After
thorough drying, the SB was taken up in CD2Cl2 for NMR.
Protonation of the Schiff bases was done by adding excess
trichloroacetic acid (in chloroform) to the SB at <10 °C.
Preparation of Fluorinated Rhodopsin Analogs. Dark-
adapted frozen cattle retinas were purchased from Excel Corp.,
KS. Procedures for opsin extraction and purification as well
as the preparation of fluorinated rhodopsin analogs for NMR
Materials and Methods
NMR and Spectroscopic Methods. 1H NMR spectra of the
retinals, Schiff bases (SB), and protonated Schiff bases (PSB)
X Abstract published in AdVance ACS Abstracts, May 1, 1996.
S0022-3654(95)02850-4 CCC: $12.00 © 1996 American Chemical Society

9176 J. Phys. Chem., Vol. 100, No. 21, 1996
Colmenares et al.
TABLE 1: UV-Vis Data and Pigment Yield Data of Fluorinated Retinals, SB, PSB,a and Rhodopsin Analogsb
11-cis
CHO, nmc
SB, nm
PSB, nm
Rh, nm
Rh yield
OS cm-1
R-CD
â-CD
retinal
8-F
10-F
12-F
380
354
350
382
378
356
348
350
340
346
442
413
428
446
459
423
432
500
463
499
507
527
476
526
+++
+++
+++
+++
+++
++
2600
2600
3300
2700
2800
2600
4100
485 (3.22)
444 (1.73)
495 (4.08)
508 (0.79)
529 (1.95)
480 (-0.99)
NDd
335 (5.82)
340 (4.53)
337 (9.00)
346 (4.30)
351 (3.30)
346 (4.31)
14-F
357
341
341
8,12-F2
10,14-F2
9-cis
+++
retinal
8-F
10-F
12-F
14-F
9-CF3
10,14-F2
9,11-dic-12F
373
354
370
375
377
336
373
374
433
410
420
436
451
386
446
443
484
459
486
493
510
454
512
484
+++
+++
+++
+++
+++
++
2400
2600
3200
2650
2600
3900
2900
1900
normal
341
450 (4.28)
478 (1.88)
484 (1.46)
508 (2.79)
448 (5.17)
low R
324 (4.27)
332 (3.81)
334 (3.06)
343 (2.18)
335 (8.11)
357
360
334
361
350
++
++
no â
a In ethanol. b in 2% CHAPS. c In hexane. d ND, not determined.
The UV opsin shift (OS, in cm-1) values,12 i.e., the differences
between the absorption maxima of the pigment analog and its
PSB, are also tabulated in Table 1.
Fluorine NMR Spectroscopy. Freshly reconstituted and
concentrated pigment analogs were used for 19F NMR studies
for recording the “before photoirradiation” and “after irradiation”
spectra. Spectra of the monofluoro analogs are shown as
representative examples in Figures 2-5 and the dicis and
difluoro analogs in the supporting information. The signal that
disappears upon photobleaching is identified as the pigment
analog, the new signal that appears upon irradiation of the
photobleaching product, stereochemistry identified by their F
shifts and the HPLC retention time and UV data of the extracted
retinal analog. Chemical shifts are listed in Table 2.
Figure 1. CD spectra of visual pigment analogs: (a) 11-cis-rhodopsin,
(b) 9,11-di-cis-12-F-, (c) 11-cis-12-F-, and (d) 11-cis-8,12-F2-rhodopsin
in 2% CHAPS.
In earlier experiments when we used a 40 µs pulse and a 50
ms interscan delay time, the intensity of the F signal was lower
than that determined by UV absorption. This initiated a
relaxation study in an attempt to determine the optimal delay
between scans for data acquisition. It was also hoped that such
a study could shed light on the origin of the observed broad
lines (∼1 ppm) for the protein-bound chromophores.
studies have already been reported.2 The reconstituted protein
samples were stored at -70 °C.
Chemical Shift Calculations. The 19F chemical shielding
values for AM1 (MOPAC 6.0)-optimized model structures were
calculated using the Turbomole and TurboNMR programs
(GIAO method10) of Biosym Technologies of San Diego using
the standard DZ + P basis set. These computations were
performed on an IBM RISC/6000 520 computer.
Upon determination of the relaxation parameters (T1 and T2)
of the 9-cis-9-CF3- and 11-cis-12-F-rhodopsin analogs (Table
3), subsequent experiments (Figures 2a, 3, 4a, 6, and 7) were
carried out using >1 s interscan delay time. The data of the
11-cis of 12-F, 10-F, and 10,14-F2 and 9-cis of 10-F and 10,-
14-F2 are in agreement with the ones obtained previously
(Colmenares et al., 1991). However, the result of the new 11-
cis-8,12-F2-rhodopsin analog showed a discrepancy with the 11-
cis-8-F pigment taken previously. A repeat experiment of the
latter was undertaken, using different delay times (see supporting
information). It is apparent that a 50 ms interscan delay time
used in the earlier experiment is too short for the 11-cis-8-F-
pigment signal to recover (return to equilibrium).
Results
Characterization of Fluorinated Retinals, SB and PSB.
1H NMR (see tables of chemical shift and coupling constants
in the supporting information) and UV-vis absorption data
(Table 1) of retinal analogs, their SB and PSB were collected.
The 11-cis and 9-cis isomers of both retinals and SB of the
fluorinated samples used in this study exhibit the characteristic
1
UV and H NMR features for their ready assignment.8,11
Characterization of Bovine Rhodopsin Analogs. The
absorption maximum wavelength data and pigment yields of
the reconstituted pigment analogs are listed in Table 1. The
pigment yields (+++, ++, and + corresponding to 70-100%,
30-70%, and 3-30%, respectively) were determined from the
ratio of absorbances at maximum wavelength of the analog and
the reference 11-cis-rhodopsin. Also measured were their CD
spectra by taking the difference curves of the pigment before
and after bleaching. The ratios of the R/â CD band ellipticities
(in mdeg) are tabulated in Table 1. All the measured R/â CD
ellipticity ratios are comparable to those of 11-cis- and 9-cis-
rhodopsins. The spectra of those that deviate greatly (11-cis-
12-F-rhodopsin, 9,11-di-cis-12-F-rhodopsin, and 11-cis-8,12-
F2-rhodopsin) are shown in Figure 1.
The updated F-chemical shifts of the rhodopsin analogs and
the corresponding PSB including the difference in these two
values (FOS) are listed in Table 2.
More recent PSB chemical shift data were acquired using
methylene chloride instead of chloroform, the solvent used
earlier. The change in solvent was a precautionary measure to
prevent isomerization.13 We found that such a solvent change
can cause 19F NMR chemical shift variation of about 0.5 ppm,
in agreement with the reported small solvent shift of 0.3-0.7
ppm.14 Medium effects are said to be dependent on the relative
polarizabilities of the solvents, and since the two mentioned
solvents are similar, our PSB F NMR data in either of the two
solvents should not significantly affect the relative FOS values.

19F NMR Study of Rhodopsin
J. Phys. Chem., Vol. 100, No. 21, 1996 9177
Figure 2. 283 MHz 19F NMR spectra of isomers of 8-F-rhodopsin in CHAPS before (lower) and after photoirradiation (upper): (a) 11-cis (pulse
delay, D5, ) 5.0 s, number of acquisitions, NA ) 5200, line broadening, LB ) 80 Hz); (b) 9-cis (D5 ) 50 ms, NA ) 160 000, LB ) 80 MHz).
Disappearance of the excess 9-cis aldehyde was due to repeated formation and bleaching of pigment during the irradiation process.
Figure 3. 283 MHz 19F NMR spectra of 10-F-rhodopsin in CHAPS before (lower) and after photoirradiation (upper): (a) 11-cis (D5 ) 1 s, NA
) 7200, LB ) 80 Hz); 9-cis (D5 ) 1 s, NA ) 4220, LB ) 80 Hz).
Figure 4. 283 MHz 19F NMR spectra of 12-F-rhodopsin in CHAPS before and after photoirradiation : (a) 11-cis (D5 ) 2 s, NA ) 1,624, LB )
80 Hz), (b) 9-cis (D5 ) 50 ms, NA ) 150 000, LB ) 80 Hz).
Sample concentration15 and temperature16 are also known to
have minimal effects on the 19F NMR chemical shift.
The FOS values for the 9-cis as well as the 11-cis pigments
are listed in Table 2. They have a mean value of 6.6 ppm with
95% of them distributed between 4.6 and 8.6 ppm. Only three
FOS values are exceptional, falling outside this range: those
of 9,11-di-cis-12-F (-2.1 ppm), 11-cis-8-F (13.1 ppm), and 11-
cis-12-F (13.2 ppm) and to a lesser degree also for the
disubstituted 11-cis-8,12-F2 (11.8, 11.7 ppm). The trends are
evident in the plot of FOS values versus the F position along
the chromophore chain shown in Figure 6.
Discussion
Fluorine Opsin Shift (FOS). The F chemical shift is very
sensitive to changes in the environment. In this study, the F
shift of each fluorinated pigment (a PSB) is compared to that
of the solution value of the corresponding free PSB. The
difference in their 19F NMR chemical shifts represents the
change imposed by the local environment (the protein binding
cavity) on the F probe. Parallel to the concept of opsin shift
(OS) in UV-vis spectroscopy (hence UV OS) introduced by
the Columbia group,12 we now call this value the 19F NMR
opsin shift (FOS).
There are a number of factors that could contribute to the
FOS value of each pigment. Their partial cancellation renders
interpretation of any small difference between these values (i.e.,
those within the 4.6-8.6 range) very difficult. However, the

9178 J. Phys. Chem., Vol. 100, No. 21, 1996
Colmenares et al.
Figure 5. 283 MHz 19F NMR spectra of 14-F-rhodopsin in CHAPS before and after photoirradiation: (a) 11-cis (D5 ) 50 ms, NA ) 196 136,
LB ) 80 Hz); (b) 9-cis (D5 ) 50 ms, NA ) 200 000, LB ) 80 Hz). The 13-cis isomer was from dark isomerization.
TABLE 2: F NMR Chemical Shifts of Retinylidene PSB’s,
Rhodopsin Pigments, and Corresponding F NMR Opsin
Shifts
the 12-s-cis and 12-s-trans conformations entail less than 1 ppm
shift for both the 10-F and 12-F analogs.
Thus, we believe that internal structural variation can only
exert a small effect of F shifts of labels at the 10- or 12-position.
The unusually large FOS value observed for 11-cis-12-F (13.2
ppm) must then be due to environmental changes of the
chromophore when protein bound. The deshielded value for
the 11-cis-12-F pigment analog is in agreement with the
presence of a nearby negative charge. Rhodopsin (11-cis)
models24-26 show that Glu-11327 is in the vicinity of C-12,
contributing to electrical interactions between the 12-F label
and the protein. Analog studies also showed crowdedness
analog
PSB (CD2Cl2) pigment (CHAPS) F NMR OSa
11-cis-8-F
-116.8
-112.2b
-107.8b
-125.5
-117.0
-104.6
-115.3
-122.7
-105.9
-119.7b
-120.9
-131.4
-119.1
-128.8
-108.7
-64.6
-103.7
-107.4
-94.6
-117.1
-105.2
-92.9
-111.8
-117.6
-99.3
-115.3
-114.3
-123.5
-115.4
-121.2
-110.8
-60.2
13.1
4.8
13.2
8.4
11.8
11.7
3.5
5.1
6.6
4.4
6.6
7.9
3.7
7.6
-2.1
4.4
11-cis-10-F
11-cis-12-F
11-cis-14-F
11-cis-8,12-F2
11-cis-10,14-F2
9-cis-8-F
9-cis-10-F
9-cis-12-F
9-cis-14-F
9-cis-10,14-F2
nearby the 12-substituent.24 Semiempirical calculations and 13
C
NMR data28 place one of the oxygens of the carboxylate
counterion (Glu-113) 2.6 Å from H-12. Orientation of F-12 in
the opposite side of the chromophore, as in the 9,11-di-cis
isomer, showed a negative FOS value (-2.1 ppm).29 The
upfield shift is consistent with the view that the fluorine probe
is directed away from the negative counterion into a less
crowded region as revealed in other analog studies.30
9,11-di-cis-12-F
9-cis-9-CF3
a F NMR opsin shift ) (column 3 - column 2) in δ, ppm. b in
CDCl3.
major FOS differences noted above undoubtedly reveal large
structural and environmental changes occurring at local sites.
We now consider a number of important factors known to affect
F shifts and interpret the unusual FOS data in light of these
factors and discuss in parallel with their CD spectra.
The fluorinated 9-cis series does not show any unusual FOS
value (3.6-7.9 ppm, Figure 6). At the 12-F position, it is not
surprising for the 9-cis and 11-cis chromophores to have
different FOS values since the distances of the F-probe to the
counterion are different in the two cases.30a The unusual FOS
value for 8-F rhodopsin analog (11-cis), 13.1 ppm, was at first
puzzling. However, its difference from the 9-cis value (6.6 ppm)
is in agreement with other evidence in the literature indicating
that 9-cis and 11-cis chromophores reside differently in the
binding cavity (by solid state 13C NMR31 and resonance Raman
studies32).
The major components contributing to chemical shielding are
believed to be the short-range or electronic structural contribu-
tions and the long-range or relatively weak electrical interactions
with the surrounding protein and solvent molecules.17 The
former includes torsion angles, bond lengths, bond angles, and
strong hydrogen bonds, the latter the external electrical charge
field generated by vicinal multipoles and weak electrical
interactions. Ring currents and other local group anisotropies
are believed to exert a small effect, ∼2 ppm shift.18,19
To have a better understanding of the cause for the downfield
shift, the F chemical shielding ab initio calculations of model
structures for F-8 were carried out. Retinals are known to
assume 6-s-cis conformations in solution and in the protein.
From MOPAC AM1 calculations of model compound 3, two
minimized structures with almost identical energies have
dihedral angles of -55.4° and 87.9° were obtained. The
computed isotropic shieldings show that F shifts can be very
sensitive to the ring-chain dihedral angle (probably due to the
extent of π-delocalization), a chemical shift difference of 3.8
ppm between the two calculated low-energy structures with the
former being less shielded. Without a nearby charged amino
acid unit,24-26 we propose that the observed large downfield
shift of 8-F in the 11-cis analog (FOS ) 13.2 ppm) is primarily
due to exclusive adoption of the -55.4° conformer in the
binding site, a degree of twist in agreement with binding
The 11-cis chromophore,20,21 while existing in an equilibrium
mixture of 12-s-cis and 12-s-trans conformations in solution,
binds selectively with the protein in the latter conformation.22,23
This structural change could affect the FOS values of the 11-
cis analogs of 10-F and 12-F. The energy differences between
pairs of conformers, obtained from MOPAC AM1 calculations,
are 0.3 and 0.6 kcal/mol for 11-cis-10-F and 11-cis-12-F,
respectively, and larger for 14-F and 10,14-F2 analogs (2.7 and
2.3 kcal/mol, respectively, 12-s-trans being lower in energy).
To determine the extent of contribution of changes in confor-
mational population to the FOS value, ab initio calculations of
the F chemical shieldings of AM1 optimized model structures
for the two conformers of 10-F and 12-F (structures 1 and 2 in
Figure 7) were attempted. The calculated results indicate that

19F NMR Study of Rhodopsin
J. Phys. Chem., Vol. 100, No. 21, 1996 9179
TABLE 3: Relaxation Data for 9-cis-9-CF3-retinal (1) and 11-cis-12-F-retinal (2) and Pigment Analogs
283 MHz
470.5 MHz
283 MHz
nOe,d %
a
b
c
a
b
c
T1
T2
ν1/2
T1
T2
ν1/2
analog 1
CHCl3-d
detergent, CHAPS
rhodopsin bound
and 9,13-di-cis-1e
and -1e
1.44 ( 0.08
0.37 ( 0.03
0.52 ( 0.02
0.42 ( 0.06
0.28 ( 0.05
490 ( 10
73 ( 1
0.6
13
126
61
1.01 ( 0.01
0.27 ( 0.01
1.06 ( 0.08
0.32 ( 0.05
0.30 ( 0.05
270 ( 10
42 ( 1
0.7
12
185
61
∼+2
-89
3.0 ( 0.2
1.8 ( 0.3
-62
4.8 ( 0.5
6.0 ( 0.1
12.2 ( 0.5
10.8 ( 0.5
∼-100
51
46
∼-100
analog 2
CHCl3-d
2.2 ( 0.01
0.30 ( 0.05
0.95 ( 0.04
0.27 ( 0.02
714 ( 10
0.5
1.8 ( 0.02
282 ( 10
32 ( 0.2
0.5 ( 0.1
2.7 ( 0.3
0.8
+35
-89
detergent, CHAPS
rhodopsin bound
and -2e
43 ( 0.8
1.6 ( 0.2
5.4 ( 0.2
22
155
60
0.42 ( 0.05
1.71 ( 0.1
22
239
95
-83
0.057 ( 0.006
∼-100
a In seconds. b In milliseconds. c In hertz. Line widths were determined by fitting the experimental points to a theoretical Lorentzian line shape.
d In percent of the nonirradiated peak measured only at 283 MHz. Values noted ∼-100 indicate that no peak could be observed. e Excess chromophore
in rhodopsin sample.
Figure 6. Plot of FOS (ppm) for vinyl F labels at several locations on
the chain of the retinyl chromophore in the (0) 11-cis, dashed line,
(4) 9-cis, solid line, and (O) 9,11-di-cis configurations. Where
appropriate, the averaged FOS values from mono- and dilabeled analogs
were used for connecting lines. Solid dots are for CF3.
selectivity of opsin as demonstrated with a series of phenyl
retinal analogs.33
Computer calculations suggest that the source for the different
FOS value of the 9-cis-8-F chromophore (6.6 ppm) could be
due to a twisted C7-C8-C9-C10 dihedral angle in solution.
Figure 7. AM1 optimized 12-S-trans (1), 12-s-cis (2), and 6-s-cis (3)
Regaining planarity of this portion of the chromophore when
protein bound could balance against the downfield shift associ-
ated with the ring-chain twist as in the 11-cis chromophore.
Consistent with this interpretation is reported perturbations of
the C10-C11 stretching mode32 and an enhanced HC7dC8H
HOOP mode34 for 9-cis but not for 11-cis rhodopsin. Also, C
NMR studies suggest that the 9-cis chromophore in the protein
has more steric interaction between H-8 and the gem-dimethyl
groups making it more twisted and blue-shifted.31
model structures for retinal and corresponding ab initio computed 19F
NMR chemical shielding and averages.
rhodopsin, and 9,11-di-cis-12-F-rhodopsin (shown in Figure 1)
exhibit features different from those of rhodopsin. That of 11-
cis-12-F-rhodopsin has a negligible R-band, surprisingly similar
to those of the 11-cis small ring-locked analogs.36,37 The 9,-
11-di-cis-12-F-rhodopsin analog, on the other hand, has a
negligible â-band, similar to the 6-s-cis-locked rhodopsin
analog38 and unlike that of 9,11-di-cis-7-membered 11-cis-
locked rhodopsin analog.39 Since the R- and â-CD curves are
believed to be associated with respective twists in the C12-
C13 and C6-C7 bonds, one may then conclude that 11-cis-
and 9,11-di-cis-12-F-rhodopsin are nearly planar at the respec-
tive regions of the chromophore, possibly due to the specific
protein substrate interactions as noted in the above discussion
of FOS values. Additionally, we noticed that 11-cis-8,12-F2-
rhodopsin has a negative R-CD band, a novelty among rhodop-
sin analogs, possibly due to the above-noted 8-F effect causing
a further ring chain twist away from planarity in the opposite
direction as in rhodopsin.
UV-Vis and CD Spectroscopic Properties of the Fluori-
nated Rhodopsins. The UV-vis and CD spectral properties
of the fluorinated rhodopsin analogs are parallel to those of the
parent rhodopsin. A few exceptions are noted below.
The UV opsin shifts of 10-F-rhodopsin analogs (3300 cm-1
)
are larger than those of rhodopsin (2600 cm-1). These pigments
are known to undergo regiospecific photoisomerization, believed
to be associated with a specific interaction of the label with an
amino acid residue in close proximity.35 The “normal” FOS
values for such chromophores are consistent with the knowledge
of a small perturbation of F shift by hydrogen bonding.15 The
UV opsin shift of 11-cis-10,14-F2-rhodopsin analog (4100 cm-1)
is larger than those of either monofluorinated analog, indicating
an additive enhancement effect.
Line Width of Pigment Peaks. For these membrane
proteins, the broad line width (∼1 ppm) was on the one hand
puzzling and on the other annoying in significantly reducing
The CD spectra of 11-cis-12-F-rhodopsin, 11-cis-8,12-F2-

9180 J. Phys. Chem., Vol. 100, No. 21, 1996
Colmenares et al.
signal sensitivity. We have therefore carried out some explor-
atory experiments to determine the cause of the observed line
width.
constant data of all isomers of the fluorinated retinal analogs
used in this study and four figures showing 19F NMR spectra
(Figures 8-10) of di-cis and difluoro pigment analogs and a
case (Figure 11) demonstrating the effect of interscan delays
on signal intensity (6 pages). Ordering information is given
on any current masthead page.
It is well-known that the chemical shift anisotropy (CSA) of
the fluorine nucleus contributes to the relaxation rates40
especially transverse relaxation (T2). Subsequent studies by Ho
et al.41 have demonstrated that internal motions of the labeled
site will result in a modulation of the expected values.
Longitudinal relaxation rates were determined for the more
abundant 12-F- and the more sensitive 9-cis-9-CF3-rhodopsin
pigment analogs. Data for the 12-F sample show the expected
field dependence of a large molecule with an overall correlation
time on the order of 10-8 s (observed T1 of 0.95 s at 283 MHz
and 1.7 s at 470 MHz). These observations are confirmed by
observed 1H-19F nOe (nuclear Overhauser effect) of -83% for
the pigment at 283 MHz. Relaxation studies of 12-F-11-cis-
retinal solubilized in CHAPS in water indicate that the CSA
contribution to T2 is small for this molecule (T1 ) 0.3 ( 0.05
s, T2 ) 43 ( 0.8 ms, nOe ) -89% at 283 MHz, and T1 ) 0.42
( 0.05 s, T2 ) 32 ( 0.2 ms at 470 MHz). In the pigment, the
19F line widths increase from 155 Hz at 283 MHz to 239 Hz at
470 MHz.
The observed line widths in Table 3 are consistent with a
predominant dipole-dipole relaxation mechanism and a small
contribution from the CSA mechanism. However, there is still
the wide range of line widths observed for the different analogs
bound to opsin. There are two possible explanations for the
observed differences in 19F line widths. One explanation is that
the retinal is bound to the protein in such a way that the nucleus
is experiencing a slow exchange between two or more confor-
mational sites within the binding pocket. Alternately, the
method used in solubilizing the protein micelles is such that a
range of particle sizes (different size micelles) is present in the
sample. This would result in signals to be of different line
widths and, possibly, different chemical shifts. Considering the
large magnitude of the FOS value, it is conceivable that minor
conformation changes can result in significant chemical shift
differences for the 19F nucleus. To date, attempts at verifying
this through variable-temperature studies have been unsuccessful
due to the limited temperature range accessible for these
membrane protein samples. Also, it is interesting to note the
different line width of the peaks for the bis-labeled analogs (and
the different line widths for different pigment analogs although
sample concentrations were not standardized), suggesting that
inhomogeneous micelle size is unlikely to be the only contribut-
ing factor to the peak width.
References and Notes
(1) Gerig, J. T. Prog. NMR Spectrosc. 1994, 293-370.
(2) Colmenares, L. U.; Asato, A. E.; Denny, M.; Mead, D.; Zingoni,
J. P.; Liu, R. S. H. Biochem. Biophys. Res. Commun. 1991, 179, 1337-
1343.
(3) Mollevanger, L.; Kentgens, A.; Pardoen, J.; Courtin, J.; Veeman,
W.; Lugtenburg, J.; deGrip, W. J. FEBS Lett. 1987, 163, 9-14.
(4) Smith, S. O.; Palings, I.; Copie, V.; Raleigh, D. P.; Courtin, J.;
Pardoen, J. A.; Lugtenburg, J.; Mathies, R. A.; Griffin, R. G. Biochemistry
1987, 26, 1606-1611.
(5) Henderson, R.; Unwin, P. N. T. Nature (London) 1975, 237, 28-
32.
(6) Henderson, R. J. Mol. Biol. 1975, 93, 123-138.
(7) Schertler, G. F. X.; Villa, C.; Henderson, R. Nature 1993, 362,
770-772.
(8) Liu, R. S. H.; Asato, A. E. In Chemistry and Biology of Synthetic
Retinoids; Dawson, M., Okamura, W., Eds.; CRC Press: Boca Raton, FL,
1990; pp 51-75.
(9) Liu, R. S. H.; Matsumoto, H.; Kini, A.; Asato, A. E.; Denny, M.;
Kropf, A.; DeGrip, W. J. Tetrahedron 1984, 40, 473-482.
(10) Ditchfield, R. Mol. Phys. 1974, 27, 789-807.
(11) Colmenares, L. U.; Liu, R. S. H. Magn. Reson. Chem. 1992, 30,
490-496.
(12) Arnaboldi, M.; Motto, M. G.; Tsujimoto, K.; Balogh-Nair, V.;
Nakanishi, K. J. Am. Chem. Soc. 1979, 101, 7082-7083.
(13) Childs, R., Shaw, G. J. Am. Chem. Soc. 1988, 110, 3013-3018.
(14) Evans, D. J. Chem. Soc. 1960, 877-880.
(15) Emsley, J. W.; Phillips, L. Prog. NMR Spectrosc. 1971, 7, 1-35.
(16) Jameson, C.; Jameson, A.; Cohen, S. J. Chem. Phys. 1977, 67,
2771-2774.
(17) Pearson, J. G.; Oldfield, E.; Lee, F. S.; Warshel, A. J. Am. Chem.
Soc. 1993, 115, 6851-6862.
(18) Gregory, D. H.; Gerig, J. T. Biopolymers 1991, 31, 845-858.
(19) Chen, R. L.; Colmenares, L. U.; Thiel, J. R.; Liu, R. S. H.
Tetrahedron Lett. 1994, 35, 7177-7180.
(20) Rowan, R.; Warshel, A.; Sykes, B. D.; Karplus, M. Biochemistry
1974, 13, 970-981.
(21) Shriver, J.; Mateescu, G.; Abrahamson, E. J. Am. Chem. Soc. 1979,
18, 4785-4792.
(22) Callender, R.; Douka, A.; Crouch, R.; Nakanishi, K. Biochemistry
1976, 15, 1621-1629.
(23) Chan, W.; Nakanishi, K.; Ebrey, T.; Honig, B. J. Am. Chem. Soc.
1974, 96, 3642-3644.
(24) Mirzadegan, T.; Liu, R. S. H. Retinal Res. 1991, 11, 57-74.
(25) Baldwin, J. M. EMBO J. 1993, 12, 1693-1703.
(26) Zhang, H.; Lerro, K. A.; Yamamoto, T.; Lien, T. H.; Sastry, L.;
Gawinowicz, M. A.; Nakanishi, K. J. Am. Chem. Soc. 1994, 116, 10165-
10173.
(27) Sakmar, T.; Franke, R.; Khorana, G. Proc. Natl. Acad. Sci. U.S.A.
1989, 86, 8309-8313.
(28) Han, M.; Smith, S. O. Biochemistry 1995, 34, 1425-1432.
(29) Colmenares, L. U.; Liu, R. S. H. Photochem. Photobiol. 1992, 56,
101-105.
(30) (a) Liu, R. S. H.; Mirzadegan, T. J. Am. Chem. Soc. 1988, 110,
8617-8623. (b) Asato, A. E.; Li, X.-Y.; Denny, M.; Mirzadegan, T.; Seff,
K.; Liu, R. S. H. Bioorg. Chem. 1989, 17, 410-421.
(31) Smith, S.; Palings, I.; Miley, M.; Courtin, J.; DeGroot, H.;
Lugtenburg, J.; Mathies, R.; Griffin, R. Biochemistry 1990, 29, 8158-8164.
(32) Palings, I.; Pardoen, J. A.; van den Berg, E.; Winkel, C.;
Lugtenburg, J.; Mathies, R. A. Biochemistry 1987, 26, 2544-2556.
(33) Colmenares, L. U.; Liu, R. S. H. Tetrahedron 1996, 52, 109-118.
(34) Eyring, G.; Curry, B.; Broek, A.; Lugtenburg, J.; Mathies, R. A.
Biochemistry 1982, 21, 384-393.
(35) Liu, R. S. H.; Crescitelli, F.; Denny, M.; Matsumoto, H.; Asato,
A. E. Biochemistry 1986, 25, 7026-7030.
(36) Fukada, Y.; Shichida, Y.; Yoshizawa, T.; Ito, M.; Kodama, A.;
Tsukida, K. Biochemistry 1984, 23, 5826-5832.
(37) Ito, M.; Hiroshima, T.; Tsukida, K.; Shichida, Y.; Yoshizawa, T.
J. Chem. Soc., Chem. Commun. 1985, 1443-1444.
(38) Ito, M. In Chemistry and Biology of Synthetic Retinoids; Dawson,
M., Okamura, W., Eds.; CRC Press: Boca Raton, FL, 1990; pp 77-97.
(39) Akita, H.; Tanis, S. P.; Adams, M.; Balogh-Nair, V.; Nakanishi,
K. J. Am. Chem. Soc. 1980, 102, 6370-6372.
(40) Hull, W. E.; Sykes, B. D. J. Mol. Biol. 1975, 98, 121-153.
(41) Ho, C.; Pratt, E. A.; Rule, G. S. Biochem. Biophys. Acta 1989,
988, 173-184.
Conclusions. The exceptional FOS values of 11-cis-8-F, 11-
cis-12-F, and 9,11-di-cis-12-F analogs can only be attributed
to local protein perturbations. The large positive FOS of 11-
cis-12-F analog is a corroborating evidence for the unusual
perturbation nearby the C-12 region of the chromophore. In
the case of 8-F, the change in the ring-chain and chain-chain
dihedral angles from that in solution to that protein bound is
believed to be the principal source of the unique FOS values
for the two isomeric pigments.
Therefore, through the use of a series of vinylic labeled
retinals, we have demonstrated possible use of F shift data for
studies of visual proteins. It is clear that the method of 19F
NMR could become more valuable if the signal peaks can be
significantly reduced. Hopefully, methods in solid state NMR
(e.g., MAS), so successfully applied to other nuclei, will also
be applicable to F in studies of membrane proteins.
Acknowledgment. This project was supported by a grant
from the U.S. Department of Health and Human Services (DK-
17806).
Supporting Information Available: Two tables (Table 4,
1
5) containing H and 19F NMR chemical shift and coupling
JP952850K