(Received for publication, July 17, 1996, and in revised form, September 25, 1996)
From the Department of Biochemistry and Molecular Biology and Ophthalmology, State University of New York Health Science Center, Syracuse, New York 13210
Phototransduction in cones differs significantly
from that in rods in sensitivity, kinetics, and recovery following
exposure to light. The contribution that the visual pigment makes in
determining the cone response was investigated biochemically by
expressing a Xenopus violet cone opsin (VCOP) cDNA in
COS1 cells and assaying the light-dependent activation of
transducin. Light-exposed VCOP stimulated [35S]guanosine
5-(
-thio)triphosphate nucleotide exchange on bovine rod transducin
in a time-dependent manner with a half-time for activation
of 0.75 min, similar to that of bovine rhodopsin. In exhaustive binding
assays, VCOP and rhodopsin activity showed similar concentration
dependence with half-maximal activation occurring at 0.02 mol of
pigment/mol of transducin. Although VCOP was able to activate as many
as 12 transducins per photoisomerization, rhodopsin catalyzed
significantly more. When assays were performed with
> 420 nm
illumination, VCOP exhibited rapid regeneration and high affinity for
the photoregenerated 11-cis-retinal. Recycling of the
chromophore and reactivation of the pigment resulted in multiple
activations of transducin, whereas a maximum of 1 transducin per VCOP
was activated under brief illumination. The decay of the active species
formed following photobleaching was complete in <5 min, ~10-fold
faster than that of rhodopsin. In vitro, VCOP activated rod
transducin with kinetics and affinity similar to those of rhodopsin,
but the active conformation decayed more rapidly and the apoprotein
regenerated more efficiently with VCOP than with rhodopsin. These
properties of the violet pigment may account for much of the difference
in response kinetics between rods and cones.
Photopic vision is mediated by specialized photoreceptor cone
cells that function at high levels of illumination, respond to rapid
changes in light, and permit color discrimination (1, 2). Each cone
cell expresses a cone opsin. The cone opsins are members of a larger
family of visual pigments (3, 4, 5) and share significant amino acid
sequence homology with the rod pigment rhodopsin (6). Among the cone
pigments, the short wavelength pigments (Group S, with wavelengths of
peak absorbance (max)1 ~ 415-440 nm) permit vision in the violet/blue region of the spectrum
and are represented by the mammalian blue, chicken violet, and
Xenopus violet pigments.2
Although cone cells expressing Group S pigments are in the minority in
the vertebrate retina, they are an integral part of vision. For
example, in blue monochromats, i.e. humans with red-green pigment mutations, the blue opsin is the sole mediator of photopic vision (7, 8).
The phototransduction mechanisms have been extensively investigated in rods (for reviews see Refs. 2 and 9). Following absorption of light and isomerization of 11-cis-retinal to all-trans-retinal, rhodopsin undergoes a series of conformational changes that eventually leads to a transient state, coincident with metarhodopsin(II) (MetaII), that activates the second messenger cascade (9). MetaII interacts with the heterotrimeric rod G protein, transducin, and thereby catalyzes the exchange of guanyl nucleotide, which leads to dissociation of the transducin subunits. Rods and cones express a very similar set of transduction proteins (10); however, in general, the electrical response of a cone to light is faster and less sensitive than that of a rod (11, 12). Moreover, there are differences in the properties of responses even between the various cone subtypes. In salamanders, responses of violet cones to light exhibit more rod-like behavior, with slower kinetics and increased sensitivity than long wavelength (red) cones (11). The molecular basis for the differences between cone and rod responses and among the cone subtypes has not been established.
In order to investigate the phototransduction properties of the short wavelength visual pigments, a Xenopus violet cone opsin (VCOP) was cloned and characterized.2 The protein, which was efficiently expressed in COS1 cells, formed a Schiff base with 11-cis-retinal and absorbed maximally at 425 nm. Characterization of the light-dependent interaction of VCOP with transducin was performed to identify the contribution of the pigment to the physiological responses of cones. Using a guanyl nucleotide exchange assay and purified bovine rod transducin, we found that VCOP activated rod transducin with kinetics and affinity similar to those of rhodopsin but that the active conformation decayed more rapidly and the apoprotein regenerated more efficiently than rhodopsin.
An
expression construct, pMT-VCOP, containing the first 328 codons of the
Xenopus violet cone opsin coding region and the last 14 codons of bovine rod opsin was expressed in mammalian COS1 cells by
transient transfection (13). The epitope-tagged violet cone pigment
(VCOP) was generated by incubation with 11-cis-retinal and
purified in buffer W (50 mM HEPES, pH 6.6, 140 mM NaCl, 3 mM MgCl2, 20% (w/v)
glycerol, and 0.1% dodecyl maltoside (DM)), as described
elsewhere.2 Xenopus rhodopsin, in the construct
pMT-XOP1 (14), was similarly expressed and purified. Bovine rhodopsin
from rod outer segments (ROS) was purified in parallel with the
expressed pigments. UV-visible absorption spectra of purified and
solubilized visual pigments were recorded in the indicated buffers at
20 °C with a Beckman Instruments DU 640 single beam
spectrophotometer using an extinction coefficient of 39,400 M1 cm
1 at 425 nm for VCOP.
Bleaching illumination was provided with light from a 300-watt
projector (Eastman Kodak Co.) at a distance of 50-60 cm. Unfiltered
light from the projector was termed white light, and high-pass colored
glass cutoff filters (>420 nm, >440 nm, and >455 nm, Edmund
Scientific, Inc., Barrington, NJ) were used as indicated. Total radiant
energy of the various filtered light stimuli were measured using an
optical power meter (3M, St. Paul, MN, Photodyne) equipped with a
calibrated silicon photodiode (EGG, Inc., Princeton, NJ). Unfiltered
light had 4.0 milliwatts, whereas the various filters had 3.4-3.6
milliwatts. Bleaching of visual pigments was monitored by recording
spectra from a single sample before and several times after
illumination under conditions identical to transducin assay conditions
(see below). All spectra were analyzed using SigmaPlot Software
(Jandel, San Rafael, CA).
Bovine transducin was purified
and the stimulation of the binding of [35S]GTPS to
bovine transducin by visual pigments was measured with the use of a
nitrocellulose filter binding assay (15, 16). Unless otherwise stated,
reaction volumes were 200 µl and assays were performed at
22-25 °C in 0.007-0.01% DM (Anatrace, Maumee, OH). The reaction
mixtures contained 10-62.5 nM VCOP or 2.5-62.5 nM bovine rhodopsin, 2 µM GTP
S (with
2000-5000 cpm [35S]GTP
S/pmol, DuPont NEN), and 300 nM bovine transducin in 10 mM Tris acetate, pH
7.0, 100 mM NaCl, 5 mM MgCl2, 5 mM 2-mercaptoethanol. Both initial rate and extent
reactions were initiated by illumination. The binding of
[35S]GTP
S to transducin in the absence of visual
pigment was subtracted from all data and was typically 5-8% of the
total transducin for a 30-min reaction. Samples that were exposed to
light prior to transducin assay were illuminated for 1 min, incubated
for various times, and then added into the transducin assay for a
30-min reaction. Assays were performed in disposable cuvettes that were
transparent in the UV region.
In order
to examine the interaction between short wavelength visual pigments and
transducin, the violet opsin was expressed in COS1 cells containing a
carboxyl-terminal 1D4 epitope tag from bovine rhodopsin (VCOP). Visual
pigment was generated in COS1 membranes by incubation with
11-cis-retinal, solubilized in DM, and purified by
immunoaffinity chromatography on 1D4-Sepharose.2 The
presence of the altered carboxyl terminus in the expressed VCOP was not
expected to perturb the transducin activity significantly since the
carboxyl domain does not appear to play an important role in transducin
activation by rhodopsin (17, 18). The low abundance of transducin from
cone cells, in either Xenopus or bovine retina, precluded
the use of cone transducin in the biochemical assays. However, the high
degree of homology between VCOP and bovine rhodopsin in the regions
shown to be involved in the interaction of light-activated rhodopsin
with rod transducin (17, 19) and the similarity of rod and cone
transducin subunits in mammals (10) and in
Xenopus3 suggested that VCOP
would interact with rod transducin.
VCOP activated bovine rod transducin in a time-dependent
manner (Fig. 1A) with a half-time of
activation (t1/2) of 0.75 min. The reaction was
complete in 5 min and resulted in only 24% of the available transducin
being activated. Both COS1-expressed Xenopus rhodopsin and
bovine ROS rhodopsin were able to completely activate transducin (Fig.
1A). Thus, although a single rhodopsin was able to catalyze
the activation of multiple transducins, a single VCOP maximally
stimulated only one transducin. The kinetics of activation of the
rhodopsins were similar to those of VCOP (t1/2 = 0.58 and 1.0 min for Xenopus and bovine rhodopsin,
respectively), suggesting that the rate of interaction between
transducin and the active state of VCOP is similar to that of the
rhodopsins. However, the initial rate of activation was slower for VCOP
(0.46 mol of GTPS bound/mol of opsin/min) than for
Xenopus (1.44 mol of GTP
S bound/mol of opsin/min) and
bovine rhodopsin (2.1 mol of GTP
S bound/mol of opsin/min).
In order to determine the total number of interactions that occur
between VCOP and transducin, a second set of experiments was performed
with a fixed concentration of rod transducin. VCOP caused a
concentration-dependent stimulation of nucleotide exchange on transducin (Fig. 1B) with a sigmoidal behavior and
half-maximal activity occurring at 1.2 pmol of pigment. Bovine ROS
rhodopsin similarly stimulated transducin with a half-maximal activity
also occurring at 1.2 pmol of bovine rhodopsin (Fig. 1B).
Thus, the apparent affinity between VCOP and rod transducin was similar to that observed with rhodopsin. However, the maximum level of activity
of VCOP occurred at 21 pmol of GTPS bound, indicating that only 35%
of available transducin was activated, in contrast to bovine rhodopsin,
which activated all available transducin. The incomplete activation of
transducin by VCOP was not due to the inability of a single VCOP
molecule to activate more than one transducin, since multiple turnovers
were observed at lower pigment concentrations (e.g. 1.2 pmol
of pigment catalyzed the exchange of 14 pmol of GTP
S). By contrast,
bovine rhodopsin stimulated the turnover of up to 36 transducins at its
half-maximal concentration (Fig. 1B) in agreement with
previous studies of the activity of bovine rhodopsin in detergent (20).
The similarity of the initial rate t1/2 values of
VCOP and rhodopsin and the identical concentration dependence for
activation of transducin suggest that the interaction of the active
state (V-MetaII or V*) of VCOP with rod transducin has a rate and an
affinity comparable to those of rhodopsin.
The failure of VCOP to completely activate transducin was not due to a
decay of the pigment during the incubation period of the transducin
assay since the activity of VCOP preincubated for 30 min at room
temperature before addition of transducin was equal to the activity
without preincubation (data not shown). The hydrophobic media,
i.e. lipid and detergent, in which transducin reactions occur play an important role in stabilizing activity. The interaction between bovine rod transducin and VCOP was measured in 0.007-0.01% DM
since previous work has demonstrated that this detergent concentration range permits optimal enzymatic activity of solubilized bovine rhodopsin (20). The optimal concentration of DM for VCOP activity sharply peaked at 0.008% (Fig. 2), which is near the
critical micelle concentration of the detergent (21). The detergent
profiles of bovine and expressed Xenopus rhodopsin
activities were the same as that of VCOP (Ref. 20 and data not shown).
Thus, the incomplete activation of bovine rod transducin by VCOP at
saturation cannot be attributed to nonoptimal detergent conditions.
To directly demonstrate that the light intensity ( > 420 nm) was
not limiting in transducin binding assays, bleaching of VCOP was
monitored spectroscopically under conditions identical to assay
conditions except that transducin was omitted (Fig. 3, left panel). Photoactivation of VCOP caused a decay of the
pigment (
max 425 nm) and a concurrent formation of a
product with
max ~ 375 nm. This product could be free
all-trans-retinal, which is released from the pigment upon
bleaching, or V-MetaII, which is expected to be spectrally
indistinguishable from free retinal. Acid denaturation experiments have
shown that the product is most likely free all-trans-retinal
(data not shown). A dark-light difference spectra (Fig. 3B)
more clearly shows the progress of pigment bleaching, since the
absorbances of VCOP and the photoproducts significantly overlap. After
10 s of illumination, >85% of the 425 nm absorbance has been
converted. In addition, there was an increase in light scattering at
shorter wavelengths, indicating that VCOP aggregates on bleaching.
After 20 s of illumination, there was no additional change in the
425 nm peak, indicating that 20 s is enough time to bleach VCOP
completely. The spectroscopic analysis of VCOP bleaching provides
direct evidence that, under the transducin assay conditions, the VCOP
was sufficiently photostimulated and cannot account for the failure of
VCOP to activate all of the available transducin.
Decay of the Active State
The ability of VCOP to activate rod
transducin suggests that illumination creates a metastable active state
analogous to MetaII. To further characterize the interaction between
VCOP and transducin, the lifetime of the photoexcited state (V*) was
measured. V* was formed by brief illumination of VCOP and its lifetime
was monitored by addition of transducin and [35S]GTPS
for nucleotide exchange assay at various times after illumination. The
light-activated VCOP exhibited a rapid decay of its ability to
stimulate GTP
S exchange. The half-time of decay was approximately 2 min and a complete loss of activity occurred in 5 min (Fig. 4). In contrast, light-activated rhodopsin exhibited
more prolonged activity with a half-time of decay of 20 min and
detectable activity even after 2 h. Thus, the reduced activity of
VCOP when compared to that of rhodopsin is attributable to the
relatively reduced lifetime of its photoactive state and not to a
reduced rate of interaction with or affinity for rod transducin.
Efficient Photoregeneration of Violet Cone Opsin
Unlike
rhodopsin (data not shown), transducin activation by VCOP under these
in vitro conditions depended on the illumination conditions
during the assay (Fig. 5). Using > 420 nm light, there was an approximately linear increase in the activity observed with the illumination time, showing an almost 10-fold increased activity with 30 min of illumination compared to 1 min. The activity of
VCOP was also measured as a function of the spectral composition of the
saturating illumination used during the transducin assay in reactions
continuously exposed to one of the following illumination conditions:
> 455 nm,
> 440 nm,
> 420 nm, or unfiltered projector
light (white light) (Fig. 6). Maximal and equivalent activities were observed when VCOP was illuminated for 30 min with
white light or with
> 420 nm light. Illumination of the pigment
for 30 min with
> 440 nm or
> 455 nm filtered light resulted
in increasingly reduced activities even though VCOP was bleached by
these lighting conditions.
The photoisomerization of all-trans-retinal to an
equilibrium mixture of isomers, including 11-cis-retinal, is
produced with high intensity near-UV light in aqueous solutions (22,
23). The recycling of released and photoisomerized
all-trans-retinal chromophore by subsequent regeneration of
VCOP apoprotein is consistent with the observed increase of activity
with illumination (Fig. 5) and with the spectral profile of activity
(Fig. 6). In order to determine whether pigment recycling was
occurring, activation of transducin by VCOP was carried out in the
presence and absence of horse liver alcohol dehydrogenase (ADH) and its
cofactor, NADH (Fig. 7). In the presence of excess NADH
at pH 7.0, ADH reduces retinal to retinol (24), thereby preventing the
regeneration of visual pigment by recycling of the chromophore. The
increase of VCOP activity with the time of illumination was eliminated when excess ADH and NADH were added, whereas the activity for the brief
illumination was not changed. The sensitivity to ADH and NADH provides
evidence that the illumination time-dependent VCOP activity
is due to the photoisomerization and recycling of released
chromophore.
In another set of experiments, the regeneration of VCOP with recycled
chromophore was explicitly monitored by measuring the ADH-sensitive
transducin activity of bleached VCOP. Before addition of rod
transducin, VCOP was illuminated ( > 420 nm) for 30 s and
preincubated for 30 min in the dark at room temperature to allow
complete decay of the resulting active state. In addition, the dark
incubation was carried out in the presence and absence of ADH and NADH
and/or 11-cis-retinal. Upon addition of transducin to the
bleached VCOP, the reaction proceeded for 30 min in the dark or under
constant illumination (
> 420 nm) to assay
illumination-dependent activity (Fig. 8).
The bleached VCOP showed light-sensitive activity. There was no
activity when the assay was performed in the dark. The light-sensitive
activity of the bleached VCOP was severely inhibited by the addition of
excess ADH and NADH. This result shows that the Xenopus
violet apoprotein rebinds 11-cis-retinal, formed by
photoisomerization of the released chromophore, to regenerate an active
pigment. The addition of 1-2 molar excess over pigment of exogenous
11-cis-retinal during the dark preincubation, immediately after preillumination, enhanced the light-sensitive activity of the
bleached pigment by 1.7-fold, and this activity was also sensitive to
ADH and NADH. The increase of activity caused by illumination, despite
substoichiometric formation of 11-cis-retinal by
photoisomerization, suggests that the affinity of Xenopus
violet opsin for 11-cis-retinal is quite high.
Despite the similarities in the components of the cone and rod visual transduction cascade, the physiological responses of cone photoreceptor cells differ from those of rods in a number of ways (11, 12). First, cones are less sensitive than rods; 100 photons are required to elicit an electrical response of a cone cell, as opposed to 1 photon for a rod cell. Second, the kinetics of the electrical response of a cone cell have a rate 3-10-fold faster than that of the rod cell. Finally, the rate of regeneration of a cone is 500 times faster than that of a rod. However, the biochemical and molecular mechanisms that underlie these phenomena are unclear, largely due to the low abundance and heterogeneity of cone phototransduction components in most retinae. Using methods for mammalian cell expression of opsins, the purified violet cone visual pigment was generated in quantities amenable to a detailed study of its interaction with rod transducin, which can be purified in abundance and, is homologous to cone transducin (10). Although the rates and extents of activity measured using rod transducin may not be the same for the activation of cone transducin, the observed activity is proportional to the amount of V* formed. The characterization of the interaction of the violet pigment and rod transducin permitted the investigation of the role of the visual pigment itself in the response properties of the cone cell.
The Xenopus violet pigment was able to activate rod transducin in both a concentration- and a time-dependent manner. The amount of violet pigment required to stimulate half-maximal activity was the same as the half-maximal value for bovine rhodopsin, suggesting that the affinity of the violet pigment for rod transducin is similar to that of rhodopsin. The maximal rate of interaction between the violet pigment and transducin was also approximately equivalent to that of bovine rhodopsin. These results are consistent with both the conservation of the amino acid regions in rhodopsin and the Xenopus violet pigment that are involved in transducin binding2 and with the high homology (76%) between bovine rod and cone transducin (10) and suggest that the conformations of the violet pigment and its enzymatically active state are very similar to those of rhodopsin. These results are consistent with the similarities found between the MetaII photointermediates of chicken red cone opsin and rhodopsin using circular dichroism (25).
The transducin activation catalyzed by the Xenopus violet
pigment differed markedly from that of bovine rhodopsin in the number of molecules activated by transducin. The lifetime of the enzymatically active state, V*, was about 10 times shorter than the active state of
rhodopsin and thus may account for the fewer turnovers and hence the
reduced sensitivity to light of cones compared to rods. It was also
demonstrated that photolysis of the violet pigment was accompanied by a
light-driven isomerization of all-trans-retinal, which was
able to recombine with the violet opsin to regenerate a photosensitive
pigment and further stimulate transducin. This system thus exhibited
efficient pigment recycling and directly implicated the effective
regeneration of the cone pigment in mediating the enhanced dark
adaptability of the cones. A summary of these findings is illustrated
in Fig. 9. Our experiments have lent direct support to
the hypothesis that the rates of pigment regeneration and V* decay play
an important role in determining cone responsiveness (4, 26). In
addition, the rapid decay of the active species raises the question of
whether a rhodopsin kinase/arrestin pathway (9) is required in cone
cells to inactivate responses to light.
These experiments demonstrate the effectiveness of using expressed Xenopus violet cone opsin as a model system for exploring cone visual pigment biochemistry and structure-function relationships. Previously, several studies on the chicken red cone pigment have demonstrated that it activates transducin (4, 25, 27) and the cGMP phosphodiesterase (28) and is a substrate for rhodopsin kinase (29). Although these studies have demonstrated the basic interactions, kinetic comparisons relevant to cone-rod physiological differences were not carried out. The experiments reported here have extended the biochemistry of phototransduction to the cone pigments and are novel for the short wavelength group of cone pigments. Recent spectroscopic work on chicken green opsin and its photobleaching intermediates (26, 30) have postulated differences in signal transduction properties between rod and cone pigments arising from differences in the MetaII lifetimes. The results presented in this report have confirmed this idea. Future work will be directed toward identifying the molecular basis for the different phototransduction properties exhibited by the short wavelength visual pigment and rhodopsin.
We gratefully acknowledge Dr. Arjun Surya for the transducin used in this work and for many helpful suggestions, Drs. Marianne Max and Ed Pugh for critical reading of the manuscript, and Dr. Jack Sullivan for help measuring light intensities.