Neuron, Vol.
15, 607-618, September,1995,Copyright© 1995by Ceil Press
Activation of Light-Dependent K + Channels
in Ciliary Invertebrate Photoreceptors Involves
cGMP but Not the IP3/Ca2÷ Cascade
Maria del Pilar Gomez and Enrico Nasi
Department of Physiology
Boston University School of Medicine
Boston, Massachusetts 02118
Marine Biological Laboratory
Woods Hole, Massachusetts 02543
Summary
The activation of light-dependent K* channels in ciliary
photoreceptors from Pecten was investigated using
intracellular dialysis of putative messengers and mod-
ulators. Neither elevated [Ca 2+] nor BAPTA changed
the membrane current in the dark or the light response.
IP3 and the antagonists heparin and decavanadate
were similarly ineffective, indicating that in these cells
the IPz/Ca 2÷ signaling pathway is not crucial for photo-
transduction. By contrast, 8-Br-cGMP and cGMP in-
duced an outward current accompanied by an increase
in membrane conductance; 8-Br-cAMP was ineffec-
tive. The identity between the cGMP-induced and the
light-induced currents is suggested by the following:
both are carried by K ÷ and blocked by 4-AP, and both
show outward rectification. In addition, guanine cyclic
nucleotides depressed the photoresponse and in-
duced single-channel currents in excised patches of
light-sensitive membrane. These light-dependent chan-
nels therefore appear to represent a link between the
families of cyclic nucleotide-gated channels and volt.
age-dependent K ÷ channels.
Introduction
The retina of some molluscan eyes is composed of a dou-
ble layer of photoreceptors: in addition to depolarizing
rhabdomeric cells similar to those found in most other in-
vertebrates, there is a distal layer of ciliary photoreceptors
that hyperpotarize in response to light stimulation (Gorman
and McReynolds, 1969). The properties of the light re-
sponse in hyperpolarizing photoreceptors have been char-
acterized in two species, Pecten (McReynolds and Gor-
man, 1970; Cornwall and Gorman, 1983a; Gomez and
Nasi, 1994a) and Lima (Cornwall and Gorman, 1983b; Go-
mez and Nasi, 1994a). Light stimulation under whole-cell
clamp elicits an outward current accompanied by a de-
crease in the input resistance of the cell. The reversal
potential of the light-activated current lies near EK and has
a Nernstian dependency on manipulations of [K+]o (Corn-
wall and Gorman, 1983a; Gomez and Nasi, 1994a), indi-
cating that the underlying conductance is K÷ selective.
There is presently no information about the transduction
cascade that couples photopigment stimulation to the acti-
vation of these light-dependent K÷ channels. Because
Ca2÷-activated K ÷ channels are commonly found in many
cell types (for a concise review, see Blatz and Magleby,
1987), and because light is known to evoke a rise in cyto-
solic Ca2÷in other invertebrate photoreceptors (Brown and
Blinks, 1974), it had previously been suggested that Ca2÷
could function as the internal messenger (Gorman and
McReynolds, 1974). On the other hand, there are some
remarkable similarities between invertebrate hyperpolar-
izing photoreceptors and vertebrate rods and cones, not
only on structural grounds (e.g., in both cases the trans-
ducing structure derives from modified cilia; Miller, 1958;
Tokuyasu and Yamada, 1959; Barber et al., 1967), but
also functionally: the photocurrent shows an outward recti-
fication that is due to voltage-dependent block by divalent
cations (Gomez and Nasi, 1994a) and is inhibited by I-cis-
-diltiazem (M. G. and E. N., unpublished data). This raises
the possibility that cGMP may also control the light-
dependent channels of hyperpolarizing invertebrate pho-
toreceptors, as it does in rods and cones (for review, see
Yau and Baylor, 1989). In the present report, we used
intracellular dialysis via a patch pipette to introduce a num-
ber of test substances implicated either in the inositol tri-
phosphate (IP3)/Ca2÷or the cGMP signaling pathways, and
we examined the effects of such treatments on the light
response, as well as directly on the conductance of the
membrane.
Results
Because in most of the experiments described below puta-
tive messengers or modulators of the transduction cas-
cade were applied internally to photoreceptor cells via the
patch electrode used for whole-cell voltage clamp, it was
critical to verify that the electrode solution effectively ex-
changes with the cytosol and reaches the site where light-
dependent channels are located. A simple test entails
replacing internal K÷ (which normally carries the photo-
current) with some impermeant ion and examining the re-
sulting changes in the photocurrent and their time course.
A control experiment was conducted (Figure 1A) in which
a photoreceptor was whole-cell clamped a t - 3 0 mV, using
the standard intracellular solution, and stimulated with
flashes of moderate intensity applied every 30 s for 6 min.
The first response was elicited a few seconds after ruptur-
ing the membrane patch to gain access to the cell interior.
The superimposed photocurrent traces have a virtually
identical shape, and their peak amplitude, plotted on the
right, shows no systematic variation over time. A similar
experiment was conducted using Cs ÷ in the electrode solu.-
tion (Figure 1B), instead of K÷. In this case, the photocur.-
rent was initially outward, but rapidly reversed direction
after the first flash, finally stabilizing as an inward pho-
tocurrent. This is expected from the fact that Cs÷ is almost
impermeant through the light-activated conductance (Cor-
nwall and Gorman, 1983a), and perfusion of the cytosol
eventually leads to an inversion of the gradient for the
permeant K+ (10 mM extracellular, - 0 mM intracellular).
The time course of the changes in peak amplitude of the
photocurrent is illustrated (Figure 1B, right); the exponen-
tial curve least-squares fitted to the data points has a time
constant ('~) of - 30 s. This result was confirmed in 4 addi-
Neuron
6O8
400pAI ~ ~ . ~
Light
b' i J i i
1000
i L i J i
2000
Time(ms)
400 pAI ~"~'=-
L t ~"'~4.5
tgh ~ L/ i i , i = i i J
0 1000 2000
Time (ms)
tional cells (average ~ = 28 _+ 12 s) and also in a photore-
ceptor dialyzed with Tris instead of Cs ÷. These observa-
tions confirm that dialysis of the cytosol is effective and
rapid and reaches the site of phototransduction.
Manipulations of Cellular Ca 2+
Ca 2+has long been considered an important putative intra-
cellular messenger for visual excitation in depolarizing in-
vertebrate photoreceptors (Bolsover and Brown, 1985;
Payne et al., 1986a, 1986b; Shin et al., 1993), although
a potential direct role in the gating of light-dependent chan-
nels remains controversial. A simple test of the possible
involvement of Ca 2÷in the generation of the light response
entails repetitively stimulating a photoreceptor during pro-
longed superfusion with Ca2÷-free artificial sea water
(ASW). In rhabdomeric cells from a variety of species, this
treatment leads to a progressive decline of the light re-
sponse (Limulus: Bolsover and Brown, 1985; Balanus:
Werner et al., !992; Lima: Nasi, 1991) presumably re-
flecting the depletion of internal Ca 2+ stores. Consistent
with this notion, these effects can in part be prevented
or rescued by intracellular ionophoretic injection of small
amounts of Ca 2+ (Bolsover and Brown, 1985; Werner et
al., 1992). In hyperpolarizing cells, however, a similar treat-
ment failed to extinguish the light response: the amplitude
of the photocurrent elicited by flashes delivered every 30
s (spanning an interval of 14 min) did not decline over time
(Figure 2A). Furthermore, the time course did not change
noticeably (n = 10; Figure 2A, inset). Sensitivity and maxi-
mal photocurrent amplitude tested after 20-40 min of su-
perfusion with Ca2+-free ASW also were not adversely af-
fected (data not shown). Because hyperpolarizing cells
may lack the Ca 2+ transport mechanisms that are promi-
nent in depolarizing photoreceptors, they may also be rela-
Figure 1. Effectiveness of Intracellular Dial-
1000 ysis
(A) Left panelshowssuperimposedrecordsof
800 photocurrents elicited by a repetitive 100 ms
E
600
light stimulus of fixed intensity (10.8 x 1013
photons x s-~ x cm-2) deliveredevery 30 s.
400 The cell was voltageclampedat -30 mV, and
the standard K+-based intracellular solution
200 was usedto fill the patchpipette.Graphat right
shows peak amplitude of the photocurrent
o';'~';'4'~'~ plotted as a function of time.
(B) Effect of intracellulardialysis with Cs+. Left
Time(rain)
panel shows light-inducedcurrent elicited by
repetitiveflashes (10.8 x 1013photons x s '
x cm 2) in a photoreceptorinternallydialyzed
with Cs+ (replacingK÷). The current inverted
1200
in polarity less than 1 min after the patch of
membranewas ruptured.Graphat rightshows
time course of the changes in peak amplitude
600 of the photocurrents.The smoothcurve repre-
sents a single-exponentialfunction fitted to the
"~ 0 data points by a Simplexalgorithmand has a
time constant ('~)of 29 s.
-600 \\
-1200 . . . . . . . . . . . . ;lr, ~ " ,
0 1 2 3
T i m e (rain)
tively insensitive to Ca 2+ changes in the external medium.
We therefore examined the effect of direct manipulations
of cytosolic Ca 2+. First, we used internal perfusion with a
high concentration of a rapid Ca 2+ buffer to attenuate any
transient increase in cytosolic free Ca 2+ that may accom-
pany the photoresponse. In Limulus ventral photorecep-
tors, injection of various Ca 2÷ chelators, such as EGTA,
Quin 2, and BAPTA, reduces the size and sensitivity of
the photocurrent and substantially slows down its kinetics
(Payne et al., 1986b; Shin et al., 1993). The records shown
in Figure 2B were obtained from a hyperpolarizing Pecten
photoreceptor voltage-clamped with a patch pipette con-
taining 10 mM BAPTA (no Ca 2÷added). The cell was repeti-
tively stimulated with flashes of constant intensity deliv-
ered every 30 s (Figure 2B, left) Photoresponsiveness was
clearly preserved (n = 3), and the amplitude and time
course of the response were comparable to those mea-
sured in cells dialyzed with standard intracellular solution
(e.g., see Figure 1A). Subsequently, a light intensity series
was administered (Figure 2B, right); the saturating pho-
tocurrent amplitude and the light evoking half-maximal re-
sponse were comparable to those of control cells (see
Figure 3D). Similar results were also obtained in 2 addi-
tional cells dialyzed with 10 mM BAPTA and 0.2 I~M free
Ca 2+ (data not shown). By contrast, in depolarizing photo-
receptors from the same organism, this treatment severely
depressed the light response and slowed down its kinetics.
(M. G. and E. N., unpublished data). We also conducted
the complementary test of the proposition that Ca 2+ may
be an activator of the transduction cascade in ciliary cells,
by introducing into the cytosol solutions containing a high
concentration of buffered Ca 2÷. Excitatory effects of Ca 2÷
should manifest themselves as an outward current devel-
oping during dialysis in the absence of photostimulation,
cGMP and Light-Activated K÷ Channels
609

Ca-free ASW Figure 2~ Insensitivity of the Photocurrent to

Ca2+ Manipulations
(A) Survival of the photocurrent elicited by re-
petitive flashes (12.6 x 1013photons x s-~ x
cm 2, 100 ms, 2 flashes/min) during prolonged
2o0 superfusion with Ca2÷-freeASW (-30 mV hold-
ing potential). The peak amplitude of the light-
0 L evoked current is plotted as a function of time.
0 3 6 9 12 15 Inset bars, 200 ms and 200 pA.
~me(min) (B) Graph at left shows repetitive stimulation
(100 ms, 10.8 x 1013photons x s 1 x cm-2,
2 flashes/min) of a photoreceptor internallydia-
B Intemal BAPTA lyzed with 10 mM BAPTA and voltage clamped
30OO at -30 mV. Inset shows superimposed light re-
sponses (bars, 200 ms and 200 pA). Graph at
right shows intensity series measured in the
same cell; peak response amplitude is plotted
,oo as a function of log attenuation of the stimulat-
O 1000 ing light (unattenuated beam, 16.4 x 10TM pho-
tons x s ~ x cm-2).Inset shows superimposed
0 family of records (bars, 200 ms and 500 pA).
0 1 2 3 4 5 -5 -4 -3 -2 -1 0
(C) Effect of internal perfusion with 10 pM Ca2÷.
Time(min) rog(I/lo)
Left panel shows holding current measured in
High internal Ca2+
the dark at -30 mV as a function of time after
accessing the cell interior. No systematic drift
400(3 was observed. Graph at right shows light ~nten-
sity series in the same photoreceptor; the satu-
m m D m m
3000 rating photocurrent amplitude and the light
100 pA [ B
intensity required to elicit a half-maximal re-
c== sponse were comparable to those measured
with 10 mM BAPTA (see [B]) and to control
100C measurements using standard internal solu-
tion. Inset bars, 200 ms and 1000 pA.
0 I 2 3 -5 -4 ~3 -2 -1 0
Time (rain) Iog(I/I o)

much as intracellular Ca 2÷ injections have been shown to recent results with optical Ca 2+ indicators have revealed
elicit an inward current in Limulus photoreceptors (Payne that the measured increase in cyto$olic Ca 2÷ that accom-
et al., 1986a; Shin et al., 1993) and in Balanus (Brown et panies the light response is due to influx from the extracel-
al., 1988). The holding current was sampled every 30 s lutar c o m p a r t m e n t (Peretz et al., 1994; Ranganathan et
beginning just after the patch was broken in a cell main- al., 1994) rather than to release from internal stores. A
tained in darkness (Figure 2C, left); the current did not similar p h e n o m e n o n was previously reported in Balanus
c h a n g e in any systematic w a y during the 3 min of re- photoreceptors (Brown and Blinks, 1974). This raises the
cording, and variations were < 2 0 pA. Responsiveness to possibility that inositol polyphosphates may be coupled
light flashes of increasing intensity was subsequently ex- to light-dependent m e m b r a n e conductance changes via
amined (6-10 min after accessing the cell interior), and mechanisms other than the release of Ca 2÷ (Hardie and
the sensitivity and m a x i m u m amplitude of the responses Minke, 1992; Minke and Selinger, 1992). For example, a
were normal (Figure 2C, right). None of the cells that were direct action of IP3 on m e m b r a n e channel gating has been
dialyzed with high concentrations of free Ca 2+ (10 ~M, demonstrated in other systems, including T-lymphocytes
n = 7; 100 ~M, n = 1) exhibited any noticeable change (Kuno and Gardner, 1987), X e n o p u s oocytes (Snyder et
in m e m b r a n e conductance or light-evoked currents that al., 1988; DeLisle et al., 1992), and olfactory neurons (Fa-
differed from those of control cells dialyzed with nanomolar dool and Ache, 1992; Restrepo et al., 1992). We therefore
levels of free Ca 2+. Together, this set of observations indi- also e x a m i n e d the effects of treatments that target the
cates that, in these photoreceptors, Ca 2~ is not crucially phosphoinositoid pathway upstream of the Ca 2" release
involved in visual excitation. process. The first approach was to introduce IP3 into the
photoreceptors directly, at 4 I~M (n = 2) and 10 ~M (n =
3); these concentrations c o m p a r e favorably with the doses
Manipulations of the Phosphoinositoid Cascade that were found effective in Limulus ventral photorecep-
In Drosophila, the pivotal role of the phosphoinositoid sig- tors (Brown et al., 1984; P a y n e et al., 1986b). Both the
naling p a t h w a y in phototransduction is supported by the holding current and the m e m b r a n e conductance remained
excitatory effects of IP3 and the demonstration of a light- unchanged during dialysis with tP3 (Figure 3A). Further-
activated phospholipase C (Devary et al., 1987) that is more, testing with photostimulation - 6 min later resulted
lacking in a blind mutant, norpA (Deland and Pak, 1973; in photocurrents that did not differ from those measured
Inoue et al., 1985; Selinger and Minke, 1988). However, under control conditions. For comparison, Figure 3D
 Neuron
610

iP3 Figure 3. Lack of Effects of IP3 and Antago-

nists
(A) Lack of effect of 4 p.M IP3 internally applied
~JL_ to a photoreceptor voltage clamped at -30 mV.
J~
200 pA[~~ Left panel shows membrane current continu-
ously recorded in the dark, starting immedi-
ately after rupturing the patch of membrane;
rectangular command steps (4 mV amplitude,
200 pA
4s
Lgiht ~
,

0
, ,
........
, .

800
. . . ~

1600
, , , ,

2400
100 ms duration) were superimposed on the
holding potential to monitor the membrane con-
ductance. No changes in either the holding cur-
Time (ms) rent or the conductance were detected. Right
panel shows currents evoked by flashes of in-
creasing intensity (1 per min; unattenuated
B C
beam, 16.4 x 1016photons x s ~ x cm 2) in
Heparin Decavanadate the same cell, demonstrating that respon-

2 opA[
siveness is comparable to that of a control cell
(see [D]).
(B) Intensity series in a cell dialyzed with 1 mg/
If/\ \ ml low molecular weight heparin.
(C) A similar test conducted in a different photo-
receptor cell, with an internal solution con-
taining 20 I~M decavanadate.
(D) Responses to light stimuli of increasing in-
Ught ~ Light
tensity measured in a control cell dialyzed with
' ' 'B--~ . . . . 1600
. . . . 2400 ~) ' ' '8;o' ' %;0' ' '2~ standard intracellular solution.
Time(ms) Time(ms) (E) Normalized peak photocurrent amplitude
plotted as a function of log light attenuation for
the cells shown in (A) (triangles), (B) (squares),
D (C) (inverted triangles), and (D) (circles), illus-
Control trating the similarity in the intensity-response
1,0 relation after the different treatments.
0.8
P
o
o 0.6
"o

0.4
E
Z 0.2
Light ~ ' /
0 y ,e L I I I
b ' ' '8;0' ' '16'00' ' '24'o0 -5 -4 -2 -1 0
Time (ms) log(I/lo)

shows light responses to similar stimuli in a cell dialyzed a concentration of 1 mg/ml, which compares favorably with
with standard internal solution. the upper bound estimates of doses that have been found
In spite of the negative results above, the possibility effective in Limulus (0.2-2 mg/ml). All 4 cells treated in
exists that the IP3 may be rapidly dephosphorylated to this way exhibited normal responsiveness (Figure 3B). We
inert IP2 (Berridge and Irvine, 1989) as it diffuses into the also tested decavanadate, another agent that has recently
cell, thus precluding significant physiological effects. For been shown to inhibit IP3 binding competitively in endo-
example, in the retina of the crab, a high level of inositol crine cells with a K,/2 ~ 5 I~M (F6hr et al., 1991). Inclusion
trisphosphatase activity has been demonstrated (Trowell, of 20 ~M d e c a v a n a d a t e in the patch electrode did not
1988). As an alternative strategy, we therefore also investi- adversely affect the light response (n = 1; Figure 3C). In
gated the effects of antagonists of the phosphoinositoid Figure 3E, the normalized peak current amplitudes for
pathway. The best known inhibitor of a variety of effects each of the conditions (Figures 3A-3D) are plotted as a
controlled by IP3 is heparin, which has been shown to function of relative light intensity, underscoring the similar-
compete with IP3 in binding studies (Worley et al., 1987; ity in the resulting intensity-response relations. The pres-
Supattapone et al., 1988) and to antagonize IP3-induced ent results therefore indicate that in these cells tP3 is not
Ca 2+ release both in vitro (Cutlen et al., 1988; Tones et acting as a messenger via some process independent of
al., 1989) and in vivo (Gosh et al., 1988; Nilsson et al., its ability to mobilize internal Ca 2+.
1988), as well as the open;ng of IP3-sensitive channels
reconstituted in lipid bilayers (Ehrlich and Watras, 1988). Effects of Cyclic Nucleotides
In Limulus ventral photoreceptors, injections of heparin A dramatically different o u t c o m e was obtained when the
have been shown to attenuate the photoresponse (Frank poorly hydrolyzable cGMP analog, 8-bromo-cGMP (8-Br-
and Fein, 1991; Faddis and Brown, 1993). We examined cGMP), was included in the internal solution. Figure 4B
the effects of heparin dissolved in the internal solution at shows the current recorded in a cell dialyzed with 50 t~M
cGMP and Light-Activated K÷ Channels
611

Control D 250 p.M 8-Br-cAMP

200 pAI
4s

~11~1~iiHih~H~i~i~H~H~i~[~[~i~[1i~1~H~i~[~i~H~H~iH~i~[~i~1~1~H~]~i~i~i~H]~i~iH~1~i~L~[~]~iH~H~iH~H~iiH
IJlJIIIJIIl~lJlfflll]rrlJ]l~llftlllJl[llll~llfNill,ill,r[ ir ~rr~[r~j~iF~J~r~i~J~r~j~H~j#]q[~JJ~r~u~]~r~r~JJ~i~i~N~J~fr~r~i~r~H~;~i~

50 IJ,M 8-Br-cGMP E 500 p,M cGMP

200 pA I
4s

---h--- ---h----
!!!!!!!!!!~!!!!!!!!~J!!!!!!!!!!!!!!!!!!!!~!!!!!!!!!!!!!!!"~!!!!!!~!!!!!~J!!~!!!!~!!!!!!~!!"J!!!~"!!

250 IIM 8-Br-cGMP F 8 mM cGMP

Figure 4. Effects of Dialysis with Cyclic Nucleotides on the Holding Current and the Membrane Conductance
All recordings were started immediately after accessing the cell interior and were performed in the dark at a holding voltage of -30 mY, while
the cells were superfused with normal ASW. The repetitive voltage steps administered to measure merebrane conductance were 4 mV in amplitude,
100 ms duration; representative current responses to command steps are shown above the continuous records, on a time scale expanded 10-fold.
(A) Control recording in a cell dialyzed with the standard intracellutar solution; no changes were observed in either the current level or the cell
input resistance throughout the recording period (40 s).
(B) Effect of including 50 p.M 8-Br-cGMP in the patch pipette. A distinct outward current developed after - 1 5 s of recording; the membrane
conductance concomitantly increased by - 250%.
(C) Recording obtained with 250 I~M 8-Br-cGMP in the pipette. A large conductance change is observed (490%), and a small inward dip precedes
the onset of the outward current.
(D) Lack of effect of internal dialysis with 250 I~M 8-Br-cAMP.
(E and F) The effects of cGMP at a concentration of 500 ~M Were insignificant (E). When cGMP was increased to 8 mM (F), a clear outward
current and an increase in membrane conductance were observed.

8-Br-cGMP. Repetitive voltage steps (4 mV amplitude, 10 tended to become shorter, the outward current was often
Hz) were superimposed on the steady holding potential preceded by a transient inward dip, and its amplitude
o f - 3 0 mV, in order to monitor changes in m e m b r a n e resis- started to decline slowly after several seconds. All of the
tance. Approximately 20 s after the patch of m e m b r a n e cells tested with 250 I~M 8-Br-cGMP at - 3 0 mV responded
was ruptured, an outward current began to develop, even- with an outward current (n -- 8). The peak amplitude of
tually reaching an amplitude of several hundred picoamp- the current averaged 471 _ 283 pA, and the associated
eres and remaining stable thereafter. The current was mean increase in m e m b r a n e conductance was 3 5 7 % -+
accompanied by a substantial increase in m e m b r a n e con- 312%. In 6 of 8 cells, the response was preceded by an
ductance from 16 to 83 nS, as revealed by the increased inward dip and displayed a gradual decline, but in all cases
size of the responses to the voltage steps. Three of four the conductance increase remained unabated as the out-
cells tested in this way displayed the effect (mean peak ward current decreased.
current amplitude, 280 _ 261 pA; average conductance A similar pattern of results was observed with hydrolyz-
increase, 141% -4- 163%). By contrast, control recordings able cGMP. In this case, however, higher concentrations
performed with the standard intracellular solution (Figure were required (>2 raM). Figure 4F shows an e x a m p l e ob-
4A) never showed any drift larger than 20 pA nor any de- tained in a cell perfused internally with 8 mM cGMP and
tectable change in m e m b r a n e conductance ,over time (n held at - 3 0 mV. No initial inward transient was observed,
= 6). When the concentration of 8-Br-cGMP in the pipette and the outward current was sustained with little or no
was increased to 250 pM (Figure 4C), the response delay indication of decay, similar to the effect of the low dose
Neuron
612
[~
A Control 501~M8-Br-cGMP 250p,M 8-Br-cGMP 250y,M 8-Br-cAMP
500 pA
Light ~ _ _ _
200 ms
(n=6)
4000 F = level. Stars above histogram bars indicate a
~ 3000 (n 3)
i,ooFi I T
c6
(50 I~M) of the slowly hydrolyzable analog. Comparable
results were obtained in 3 other cells (mean amplitude,
533 ___94pA; mean conductance change, 223% _+ 91%).
A first question to be addressed concerns the specificity
of the response to cGMP and its analog. To that effect,
we employed 250 I~M 8-bromo-cAMP (8-Br-cAMP; Figure
4E). In 3 cells examined, no outward current was elicited
(in 1 cell a barely perceptible slow drift reaching only - 3 5
pA was detected over a period of 4 min) nor any detectable
changes in the cell input resistance. These observations
indicate a substantial selectivity for guanosine over adeno-
sine. cyclic nucleotides.
The above results demonstrate the induction by cGMP
of a membrane conductance and the activation of an out-
ward current at physiological membrane potentials. Clues
were then sought to establish whether such a response
may be identified with the light-activated, K+-selective con-
ductance of these cells. An observation that may critically
bear on this issue is the effect of cGMP stimulation on the
photocurrent: if light and cGMP activate the same conduc-
tance, one would predict that, when a fraction of the chan-
nels has already been opened by cGMP, the number of
channels available for activation by a light stimulus will
necessary be reduced. To this end, many of the cells inter-
nally perfused with the compounds described above were
tested with a standard light intensity series protocol. It was
found that photoreceptors dialyzed with guanine cyclic nu-
cleotides exhibited dramatically decreased light respon-
siveness as compared with control photoreceptors and
photoreceptors dialyzed with 8-Br-cAMP (Figure 5A). To
quantitate the effect, the amplitude of the saturating pho-
tocurrent was determined by taking the asymptote of a
sigmoidal function least-squares fitted to the data points
from the intensity series and averaging the obtained val-
ues for each of the conditions (Figure 5B). A Kruskal-
Wallis nonparametric analysis of variance (Hollander and
Figure 5. Depressionofthe Light-EvokedCur-
8 mM oGMP
rent after InternalAdministrationof Cyclic Nu-
cleotides
(A) Superimposed photocurrents elicited by
flashesof increasingintensityin representative
cells treatedas indicatedaboveeach family of
curves.
(B) Mean value of the maximal photocurrent
measuredundereach of the variousconditions
of intracellulardialysis. The overall intergroup
differences were significant at the p = .02
significant difference(p < .05) with respect to
the control group. Only the cells treated with
8-Br-cAMP failed to show a reliable reduction
of the photoresponse.Error bars indicate SD.
E
Wolfe, 1973) confirmed that the overall differences across
conditions were statistically highly significant (p < .001),
and that each of the groups treated with 8-Br-cGMP or
with cGMP differed from the control group (p < .05). In
the case of 8-Br-cAMP, the apparent reduction in photore-
sponse amplitude failed to attain statistical significance.
Finally, there was a significantly different degree of pho-
tocurrent suppression induced by the two concentrations
of 8-Br-cGMP (50 and 250 I~M), demonstrating dose de-
pendency of the effect (mean residual saturating response
amplitudes of 401 and 157 pA, respectively). We also ex-
amined the correlation between the size of the dialysis-
induced outward current and that of the light response by
pooling all the photoreceptors treated with guanine nucle-
otides. This approach partly overcomes the cell to cell
variability that may stem from differences in the efficacy
of internal perfusion. The obtained value for Spearman's
correlation coefficient was -0.53; Kendall's rank analysis
confirmed that the codependency between the two vari-
ables was statistically significant (K = 64, n = 19, p <
.02). An exact quantitative relation is difficult to establish
because an additional conductance is sometimes ob-
served during dialysis, as previously discussed. However,
the suggestion that light and guanine cyclic nucleotides
tap a common effector mechanism is strengthened by the
observation that the average sum of the conductance in-
duced by 250 pM 8-Br-cGMP (41.3 __. 18 nS; n = 7) and
the residual conductance induced by light (5.1 -+ 4.4 nS)
did not differ significantly from the saturating light-
dependent conductance in untreated cells (56.7 _+ 25 nS;
n = 6).
We also examined whether the cGMP-evoked current
shares any salient properties with the light-induced cur-
rent. The first aspect scrutinized was the current-voltage
relation. This type of measurement presents significant
problems because, as soon as the membrane patch is
cGMP and Light-Activated K÷ Channels
613

A Figure 6. Outward Rectification and K÷ De-

2nd pendency of the Current Induced by Dialysis
1.0 with Guanine Nucleotides
(A) A photoreceptor was whole-cell voltage
~o 0.8 clamped with an electrode containing 8 mM
cGMP. A few seconds after rupturing the mem-
0.6 brane patch to gain access to the intracellular
compartment, a ramp was applied, sweeping
o.4
.E from -85 to +40 mV (1st trace). A second ramp
o 0.2 was applied - 1 rain later, when an outward
Time (ms) membrane current of -170 pA (at -30 mV
0,.~0' , 3,00-40' , , 6~, ,0 ' 9~140
o ;,II
-80
i
-60
tl
-~o
i
-20
i holding potential) had developed (2nd trace).
The difference between the two traces, repre-
Voltage (mV) Voltage(mV) senting the cGMP-activated current, rectifies
markedly in the outward direction and be-
comes vanishingly small as the command volt-
age is swept near the negative end of the
c
range. The dual-labeled horizontal axis indi-
cates both time and applied voltage.
10 mM [K]o
600 (B) Comparison of the current-voltage relation
for the current induced by 8 mM cGMP (open
squares) and the light-induced current (closed
~ 400 squares). Values of the cGMP-activated cur-
rent during stimulation with a voltage ramp
== were normalizedwith respect of the current at 0
5 0 m M [K]o 200
0 mV and averaged for 3 cells. Error bars indicate
SD. Peak photocurrent amplitudes in response
to a half-saturating flash were obtained in 5
cells, normalized, and averaged.
Time (ms) (C) K÷ dependency of the cGMP-induced cur-
100 pA
0 500 1000 rent. The two recordings were obtained in 2
20 s
i ,i
-90
i 150 t h, i
-10
i 30L' cells dialyzed with cGMP and voltage clamped
Voltage(mV) at the same holding potential (-60 mV). The
top trace was obtained in normal ASW con-
taining 10 mM K÷ and shows an outward cur-
rent developing during dialysis. The bottom trace was obtained during superfusion with high K÷ (50 raM) ASW; in this case, the cGMP-induced
current was inwardly directed.
(D) Reversal of the cGMP-activated current in the presence of 50 mM [K+]o,measured with voltage ramps. The trace represents the subtraction
of the record obtained immediately after patch rupturing from that measured 1 rain later, near the peak of the cyclic nucleotide effect. The nutl
point occurred at -43 mV (arrow), close to the calculated value of EK (-45 mV).

ruptured in order to voltage-clamp a cell, dialysis of the
intracellular c o m p a r t m e n t with the activator begins. As
such, it is difficult to provide an uncontaminated temporal
w i n d o w to estimate the basal conductance to be sub-
tracted from the conductance measured subsequently, at
the peak of the nucleotide effect. Because voltage steps
are too time consuming, we resorted to ramps of voltage,
which can rapidly sweep the range of interest. Even then,
however, a margin of uncertainty remains regarding the
"purity" of the m e a s u r e m e n t of baseline conductance be-
cause, on several occasions, the time delay between
patch rupturing and onset of the physiological effects was
only on the order of a few seconds (e.g., see Figures 4C
and 4F). The two records in Figure 6A illustrate the mem-
brane current measured during application of a voltage
ramp in a cell dialyzed with 8 mM cGMP, shortly after
patch rupturing (i.e., within 10 s) and - 1 min later; the
difference b e t w e e n them therefore represents the compo-
nent activated by cGMP. In Figure 6B the normalized
mean current induced by 8 mM cGMP (average of 3 cells)
and the light-induced current elicited by half-saturating
stimuli (average of 5 cells) are plotted as a function of
m e m b r a n e voltage. Both exhibit a similar rectification in
the outward direction and approach zero near the negative
end of the voltage range. It is difficult to determine the
reversal potential with accuracy, as the current becomes
vanishingly small over an e x t e n d e d voltage range. In fact,
we had previously reported a dramatic reduction of the
light-sensitive conductance at very negative potentials,
which makes it difficult to record sizable inwardly directed
photocurrents under normal ionic conditions; by contrast,
if the cells were exposed to a high [K÷]o, large inward pho-
tocurrents could be readily obtained at modest hyperpolar-
izations with respect to EK (Gomez and Nasi, 1994a). Fig-
ure 6C illustrates the effect of manipulating [K+]o on the
cGMP-induced current in 2 different cells voltage-clamped
at the same holding potential, - 6 0 mV, a voltage at which
the light-dependent conductance is not drastically re-
duced. The extracellular solution contained either 10 mM
(normal ASW) or 50 mM [K÷]o; the calculated values for EK
are - 8 5 and - 4 5 mV, respectively, and the corresponding
driving force on K ÷ in the two cases was therefore approxi-
mately +25 versus - 2 0 inV. In normal [K÷]o, the nucleofide-
evoked current was outwardly directed (Figure 6C, top
trace), whereas in the presence of high K ÷ solution, it was
inward. Similar inward currents induced by cGMP in the
presence of elevated [K+]o were observed in 8 cells. Figure
6D shows the difference between a ramp administered
immediately after establishing whole-cell clamp and one
applied after 1 min of dialysis with cGMP, in a cell super-
Neuron
614
25 pA
Y
4-AP
5 pA
2s
8-Br-cGMP
in the outwarddirection.Channel activity subsidedupon returningto the control solution.Activation of the channelscould be obtainedrepetitively
on subsequentperiods of 8-Br-cGMP administration.Pipette potential,+40 mV.
fused with 50 mM [K+]o. Under these conditions a distinct
reversal occurred near the calculated equilibrium potential
for K+ (41 __. 2 mV; n -- 2). Together, the results indicate
that cGMP activates an outwardly rectifying current pri-
marily carried by K÷, similar to the light-dependent current.
In previous work (Gomez and Nasi, 1994b) it was demon-
strated that the light-sensitive channels of hyperpolarizing
scallop photoreceptors are susceptible to blockage by
4-aminopyridine (4-AP), with a submicromolar effective
K,/2. This feature provided an additional criterion to help
identify the nature of the conductance increase evoked
by the application of guanine cyclic nucleotides. We tested
the effect of external superfusion with 4-AP on the mem-
brane current that develops as a result of internal dialysis
with 8-Br-cGMP. Figure 7A illustrates the result of an ex-
periment in which a pipette filled with ASW containing 50
I~M 4-AP was positioned - 6 0 ~m from a cell voltage-
clamped with an electrode containing 8 mM cGMP. Shortly
after rupturing the patch of membrane, the usual outward
current is observed reaching an amplitude of - 9 0 pA.
At that point, 4-AP was pressure-ejected from the puffer
pipette for - 12 s, causing an immediate decrease in the
current, approximately back to the baseline level (i.e., the
holding current immediately after patch rupture). After in-
terrupting delivery of 4-AP, the current began slowly to
recover, and a second application of the blocker sup-
pressed it again. A similar, though less complete, blockage
was observed in another photoreceptor dialyzed with 100
p,M 8-Br-cGMP and bathed in 50 mM [K+]o while the mem-
brane potential was held at - 6 0 mV; in this case, an in-
wardly directed current developed and was subsequently
depressed by 4-AP. A control cell internally perfused with
normal intracellular solution (i.e., without cGMP) proved
insensitive to application of 4-AP.
Figure 7. cGMP-Dependent Conductance:
Pharmacological Blockage and Activation in
Cell-Free Patches
(A) Blockageof the cGMP-evoked current by
4-AP. A cell was dialyzedwith 8 mM cGMP at
-30 mV. An outward current developedwith a
latencyof a fewsecondsand reacheda plateau
in little morethan 1 min. At that point, a puffer
pipette was approached,and 50 #,M 4-AP dis-
solved in ASW was pressureejected,causing
rapid returnof the currentto baseline.Interrup-
4-AP tion of the local perfusion gave rise to a slow
recoveryof the outwardcurrent, and a second
applicationof 4-AP again suppressedit. Short
bars below trace indicate 4-AP applications;
vertical lines labeled"R" indicate the delivery
of a voltage ramp.
(B) Multichannel currents induced by 8-Br-
cGMP in an excisedpatch.The recordingwas
obtained from an inside-out patch excised
while locallyperfusingwith standardintracellu-
lar solution by meansof a double-barrelpuffer
pipette positioned -50 #m from the recording
electrodetip. During the period markedby the
long horizontalbar, the solutionwas switched
to one containing 100 p,M 8-Br-cGMP, which
resultedin the appearanceof channelcurrents
Additional insights on the effects of cGMP on the gating
of the light-sensitive channels can be gained by examining
currents in excised patches. Cell-attached recordings in
the ciliary appendages, presumably the light-sensitive
structures of the cell, are rather difficult to obtain, not only
because of the small size of the target but also because
the seal resistance rarely exceeded 2 G~; therefore, the
background noise was often too high for high resolution
measurements. In a previous study, however, we suc-
ceeded in characterizing the basic properties of light-
dependent single-channel currents (Gomez and Nasi,
1994a). Excision of such patches posed additional difficul-
ties, primarily owing to the fragility of the cilium that con-
nects the appendages to the soma; this breaks easily,
causing the whole appendage to detach and reseal. In
other instances, the entire cell tended to break loose from
the chamber, because of the precarious attachment of
these cells to the chamber (Gomez and Nasi, 1994a), and
preclude excision altogether. Only three patches that were
screened for on-cell light-dependent channel activity sur-
vived excision and could be stimulated with 8-Br-cGMP.
In 2 of these, ejection of the agonist from a puffer pipette
evoked outwardly directed channel currents. Figure 7B
shows an example in which the electrode potential was
+40 mV, the concentration of 8-Br-cGMP was 100 I~M,
and the puffer pipette was positioned - 5 0 #m from the
recording electrode tip. Vigorous channel activity ap-
peared with a short latency ( - 2 0 0 ms) and persisted
throughout the period of agonist presentation ( - 1 0 . 5 s).
When ejection of 8-Br-cGMP was interrupted, the channel
activity rapidly faded. This effect could be repeated numer-
ous times with little indication of desensitization. Because
nucleoside trisphosphates were not present in the control
and test solutions, the effects appear not to be mediated
cGMP and Light-Activated K ~ Channels
615
by a phosphorylation mechanism. The presence of several
channels in the patch, however, precluded determination
of the unitary amplitude and, hence, of the single-channel
conductance, in order to positively identify these channels
with those activated by photostimulation.
Discussion
The present observations demonstrate that in hyperpolar-
izing scallop photoreceptors guanine cyclic nucleotides
selectively induce a membrane conductance that shares
important similarities with the light-dependent conduc-
tance, including the outward rectification, the dependency
on K÷, and the susceptibility to blockage by 4-AP. Further-
more, direct chemical stimulation with cGMP was accom-
panied by a dramatic reduction in the responsiveness to
light stimuli. By contrast, neither the treatments designed
to stimulate or inhibit the IP3 pathway nor the manipula-
tions of cellular Ca 2÷and Ca2+buffering had any detectable
effect either directly on membrane currents or on the light
response. Together, the present results strongly suggest
that these photoreceptors rely on cGMP as an internal
messenger for the transduction cascade. The respon-
siveness of excised patches to 8-Br-cGMP stimulation in
the absence of ATP and GTP is consistent with the notion
that guanine cyclic nucleotides may interact directly with
the light-dependent K÷-selective ion channels, in a manner
similar to vertebrate rods and cones (Fesenko et al., 1985).
A precedent for cGMP controlling a light-dependent K÷
conductance has recently been found in a photosensitive
neuron in the abdominal ganglion of the opisthobranch
mollusc Onchidium verrunculatum. This cell generates
slow, depolarizing photoresponses due to the closing of
K÷ channels (Gotow, 1989) that are kept open in the dark
by cGMP (Gotow and Nishi, 1991; Gotow et al., 1994).
The lower doses of 8-Br-cGMP that were tested and
found effective (50 ~tM) compare favorably with those that
have been proved effective in vertebrate sensory cells.
The pronounced discrepancy in potency between cGMP
and 8-Br-cGMP is reminiscent of similar findings in am-
phibian rods, which have been accounted for by at least
two factors. First, there appears to be a genuine difference
in the affinity of the two compounds for the channels: in
perfused, excised patches of outer segment membrane,
the half-maximal concentration for 8-Br-cGMP is near
2-3 ~M, which is approximately an order of magnitude
lower than that for cGMP (Zimmerman et al., 1985). In
addition, factors normally present in the intracellular envi-
ronment may also exert a powerful influence: in truncated
outer segments, Nakatani and Yau (1988) demonstrated
that 8-Br-cGM P is similarly effective as in excised patches
(i.e., Ky= = 4 ~M) but that the lower potency of the hydrolyz-
able form is even more evident (-100-fold). Because in-
clusion of isobutylmethylxanthine (IBMX) in the dialysis
solution lowered the K,~,for cGMP without affecting that
of 8-Br-cGMP, effectively reducing their relative potency
to a factor of - 2 0 , these authors concluded that in the
outer segments the phosphodiesterase activity is suffi-
ciently high to hydrolyze a substantial fraction of the cGMP
present in the intracellular solution. In preliminary tests in
Pecten photoreceptors, however, we failed to boost the
efficacy of moderate concentrations of cGMP (2 mM) by
codialyzing 1 mM IBMX. With 250 ~M 8-Br-cGMP, the
amplitude of the outward current declined over time. How-
ever, because this phenomenon was correlated with the
occurrence of an early, inwardly directed transient and
was never accompanied by a reduction in membrane con-
ductance, it appears not to be due to desensitization of
the response. Rather, high concentrations of the agonist
may also activate an additional conductance with a differ-
ent ionic selectivity. In amphibian rods, high doses of
cGMP have also been reported to induce the activation
of more than one conductance mechanism (Cameron and
Pugh, 1990; Koutalos et al., 1995).
A role for cGMP in the visual response of depolarizing
photoreceptors of Limulus had previously been suggested
(Johnson et al., 1986; Bacigalupo et al., 1991). In those
cells, however, a rise in cytosolic Ca2+appears to be both
necessary (Bolsover and Brown, 1985; Shin et al., 1993)
and sufficient to produce photoreceptor excitation (Payne
et al., 1986a; Shin et al., 1993). As a consequence, it was
proposed that cGMP operates downstream in the trans-
duction cascade and that phosphoinositoids retain a criti-
cal role in the generation of the visual response, as in
other depolarizing photoreceptors (Devary et at., 1987).
By contrast, in the present case, IP3 and Ca2+do not appear
to be implicated as precursors to cGMP mobilization. The
disparate organization of the transduction pathway in pho-
toreceptor cells from the same or from related organisms
may reflect the separate evolutionary lines of ciliary versus
rhabdomeric visual cells postulated by Eakin (1965). Alter-
natively, one can speculate that, because the transduction
gain in depolarizing photoreceptors is orders of magnitude
greater than that of hyperpolarizing cells, the IPJCa 2+
pathway may have evolved as an added gain stage.
The notion that some invertebrate photoreceptors utilize
an internal messenger similar to that used by rods and
cones raises the issue of the similarity between their re-
spective light-dependent ion channels. An obvious dis-
crepancy exists concerning the properties of ionic selectiv-
ity. The cyclic nucleotide-gated channels of vertebrate
sensory cells are characteristically Cationic nonselective,
in sharp contrast with the high selectivity to K÷ displayed
by the light-sensitive conductance of the ciliary photore-
ceptors in Pecten and Lima (Cornwall and Gorman, 1983a;
Gomez and Nasi, 1994b). Such a difference in conduction
properties, however, may stem from small structural alter-
ations: recent m utagenesis experiments in Shaker demon-
strate that a switch from high K÷ selectivity to cationic
nonselectivity may entail only minor amino acid substitu-
tions (Heginbotham et al., 1992). Furthermore, the cloning
of the cyclic nucleotide-activated channel from vertebrate
photoreceptors (Kaupp et al., 1989) and olfactory neurons
(Dhallan et al., 1990) has revealed a structural kinship to
voltage-gated K÷ channels, suggesting that the two fami-
lies are related (Kaupp, 1991 ; Guy et al., 1991 ; Yau, 1994).
The light-dependent channels of Pecten photoreceptors
might bridge the gap between these two classes of ion
channels.
Neuron
616
Experimental Procedures
Specimens of Pecten irradians were obtained from the Marine Re-
sources Center at the Marine Biological Laboratory (Woods Hole, MA)
and used immediately. Isolated retinas were enzymatically and me-
chanically dispersed (Gomez and Nasi, 1994a), and the resulting cell
suspension was transferred to a recording flow chamber pretreated
with concanavatin A to promote cell adhesion, as previously described
(Nasi, 1991). After plating, the chamber was continuously superfused
with ASW containing 480 mM NaCI, 10 mM KCI, 49 mM MgCI2, 10
mM CaCI2, 10 mM HEPES, and 5 mM glucose (pH 7.8, adjusted with
NaOH). All experiments were conducted at room temperature (22°C -
24°C). Whole-cell tight-seal recording was performed in isolated hy-
perpolarizing photoreceptors, as previously described (Gomez and
Nasi, 1994a, 1994b). Patch electrodes, fabricated with thin-wall boro-
silicate capillary tubing (Garner Glass, Claremont, CA), were filled with
an intracellular solution containing 100 mM KCI, 200 mM K-aspartate
or K-glutamate, 6 mM Na2ATP, 6 mM MgCI2, 10 mM HEPES, 1 mM
EGTA, 100 p.M GTP, and 300 mM sucrose (pH 7.3). In control experi-
ments designed to verify effectiveness of perfusion, intracellular K+
was replaced with either Cs÷ or Tris. Test substances were simply
added to the standard internal solution from stock solutions prepared
daily, which required dilution by either 1:50 or 1:100 to achieve the
desired final concentration, cGMP and its slowly hydrolyzable analog
8-Br-cGMP were purchased from either Pluka (Ronkonkoma, NY) or
Sigma (St. Louis, MO); 8-Br-cAMP, IP3, and low molecular weight
(4000-6000) heparin were obtained from Sigma. Decavanadate was
prepared from sodium orthovanadate (Sigma) by titration with HCI
to pH 3.7 (as described in FShr et al, 1989) and used immediately.
Intracellular solutions designed to buffer transients of cytosolic free
Ca2÷contained 10 mM BAPTA (Molecular Probes, Eugene, OR), either
Ca2÷-freeor with an added amount of Ca2+to yield a free concentration
of 0.2 ~M, as calculated using the program Chelator (kindly provided
by Dr. T. Schoenmakers). Intracellular solutions with elevated levels
of Ca2÷ were prepared with a mixture of Ca 2÷ buffer (1 mM) and an
added amount of CaCI2 to result in 10 or 100 ~,M Ca 2÷. For 10 ~M free
Ca2÷,we used EGTA. HEDTA would be the buffer of choice; however, in
several pilot tests, it proved toxic to the transduction process, severely
depressing the light response, irrespective of the presence of Ca2+.
In the case of 100 t~M free Ca2÷, the buffer was nitrilotriacetic acid
(NTA). The resistance of the patch electrodes, measured in ASW,
ranged between 2 and 6 M~. Series resistance errors were corrected
via a positive feedback circuit in the amplifier (maximal residual error
typically < 2 mV). Macroscopic currents were low-pass filtered at 500
Hz (Bessel 4-pole filter) and digitized on-line at a 2 kHz sampling rate;
single-channel currents were filtered at 2-3 kHz and sampled at 7.5-
12 kHz. Voltage and light stimuli were applied by a microprocessor-
controlled programmable stimulator (Stim 6, Ionoptix, Milton, MA), ex-
cept for the voltage ramps, which were produced by an arbitrary wave-
form generator (HP-33120A, Hewlett Packard, Palo Alto, CA).
Solution changes were performed either by superfusing the whole
recording chamber via a system of reservoirs and manifolds, or by
local perfusion using puffer pipettes; the latter was preferred when
rapidity of solution change was critical. Puffer pipettes were pulled
from either single- or double-barrel capillary glass (Glass Company
of America, Millville, NJ) to a tip with an outside diameter of 3-4 I~m,
filled, and connected to a solenoid-operated valve that delivered pres-
surized nitrogen (2-5 Ib/in2).
Light steps were provided by two optical stimulators. The beam
of one was combined with that of the microscope illuminator via a
beam-splitter prism, whereas the other was brought by a fiber optics
light guide to the epifluorescence port of the inverted microscope, as
previously described (Gomez and Nasi, 1994a). Light intensity was
measured with a calibrated radiometer (Mod. 370, UDT, Hawthorne,
CA). An in vivo calibration (n = 3) was performed, as previously de-
scribed (Gomez and Nasi, 1994a, 1994b), to compare the effect of
monochromatic stimulation (500 nm peak, 10 nm half-width; Ditric
Optics, Hudson, MA) with that of white or broadband (515-650 nm) light
used in most experiments, in order to provide a quantitative reference.
Light intensity is expressed either in terms of equivalent photon flux
at 500 nm or, when light attenuation was varied in a given protocol,
as log,0(l/lo) (where Io is the intensity of the unattenuated beam light).
Calibrated neutral-density filters (Melles Griot, Irvine, CA) provided
controlled attenuation. During experimental manipulations, the cells
were viewed with a Newvicon TV camera (Mod. WV-1550, Panasonic,
Secaucus, NJ) using a near-infrared long-pass filter for illumination
(;~ > 780 nm; Andover Corporation, Salem, NH). The infrared illumina-
tor was turned off for several minutes before testing light responses.
Acknowledgments
This study was supported by National Institutes of Health grant RO1-
EY07559.
The costs of publication of this article were defrayed in part by
the payment of page charges. This article must therefore be hereby
marked "advertisement" in accordance with 18 USC Section 1734
solely to indicate this fact.
Received March 10, 1995; revised June 17, 1995.
References
Bacigalupo, J., Johnson, E. C., Vergara, C., and Lisman, J. E. (1991).
Light-dependent channels from excised patches of Limulus ventral
photoreceptors are opened by cGMP. Proc. Natl. Acad. Sci. USA 88,
7938-7942.
Barber, V. C., Evans, E. M, and Land, M. F. (1967). The fine structure
of the eye of the mollusk Pecten maximus. Z. Zellfors. Mikro. Anat.
76, 295-312.
Berridge, M. J., and Irvine, R. F. (1989). Inositol phosphates and cell
signalling. Nature 341, 197-205.
Blatz, A. M., and Magleby, K. L. (1987). Calcium-activated potassium
channels. Trends Neurosci. 11,463-467.
Bolsover, S. R., and Brown, J. E. (1985). Calcium ion, an intracellular
messenger of light adaptation, also participates in excitation of Limulus
photoreceptors. J. Physiol. 364, 381-393.
Brown, J. E., and Blinks, J. R. (1974). Changes in intracellular free
calcium concentration during illumination of invertebrate photorecep-
tors. J. Gen. Physiol. 64, 643-665.
Brown, J. E., Rubin, L. J., Ghalayni, A. J., Tarver, A. P., Irvine, R. F.,
Berridge, M. J., and Anderson, R. E. (1984). myo-inositol polyphos-
phate may be a messenger for visual excitation in Limulus photorecep-
tors. Nature 311, 160-163.
Brown, H. M., Rydqvist, B., and Moser, H. (1988). Intracellular calcium
changes in Balanus photoreceptor. A study with calcium ion-selective
electrodes and Arsenazo II1. Cell Calcium 9, 105-119.
Cameron, D. A., and Pugh, E. N. (1990). The magnitude, time course
and spatial distribution of current induced in salamander rods by cyclic
guanine nucleotides. J. Physiol. 430, 419-430.
Cornwall, M. C., and Gorman, A. L. F. (1983a). The cation selectivity
and voltage dependence of the light-activated potassium conductance
in scallop distal photoreceptor. J. Physiol. 340, 287-305.
Cornwall, M. C., and Gorman, A. L. F. (1983b). Ionic and spectral
mechanisms of the off response to light in hyperpolarizing photorecep-
tors of the clam, Lima scabra. Cell. Mol. Neurobiol. 3, 311-328.
Cullen, P. G., Comerford, G. G., and Dawson, A. P. (1988). Heparin
inhibits the inositol 1,4,5-trisphosphate-induced Ca2+-release from rat
liver microsomes. FEBS Lett. 228, 57-59.
Deland, M. C., and Pak, W. L. (1973). Reversible temperature sensitive
phototransduction mutant of Drosophila melanogaster. Nature 244,
184-186.
DeLisle, S., Pittet, D., Potter, B. V. L., Lew, P. D., and Welsh, M J.
(1992). InsP3 and Ins(1,3,4,5)P4 act in synergy to stimulate influx of
extracellular Ca2÷ in Xenopus oocytes. Am. J. Physiol. 262, C1456-
C1463.
Devary, O., Heichal, O., Blumenfeld, A., Cassel, D., Suss, E., Barash,
S., Rubinstein, C. T., Minke, B., and Selinger, Z. (1987). Coupling of
photoexcited rhodopsin to inositol phospholipid hydrolysis in fly photo-
receptors. Proc. Natl. Acad. Sci. USA 84, 6939-6943.
Dhallan, R. S., Yau, K.-W., Schrader, K., and Reed, R. R. (1990).
Primary structure and functional expression of a cyclic nucleotide-
activated channel from olfactory neurons. Nature 347, 184-189.
cGMP and Light-Activated K÷ Channels
617
Eakin, R. M. (1965). Evolution of photoreceptors. Cold Spring Harb.
Symp. Quant. Biol. 30, 363-370.
Ehrlich, B. E., and Watras, J. (1988). Inositol 1,4,5-trisphosphate acti-
vates a channel from smooth muscle sarcoplasmic reticulum. Nature
336, 583-586.
Faddis, M. N., and Brown, J. E. (1993). Intracellular injection of heparin
and polyamines. Effects on phototransduction in Limulus ventral photo-
receptors. J. Gen. Physiol. 101,909-931.
Fadool, D. A., and Ache, B. W. (1992). Plasma membrane inositol
1,4,5-trisphosphate-activated channels mediate signal transduction in
lobster olfactory receptor neurons. Neuron 9, 907-918.
Fesenko, E. E., Kolesnikov, S. S., and Lyubarsky, A. L. (1985). Induc-
tion by cyclic GMP of cationic conductance in plasma membrane of
retinal rod outer segment. Nature 313, 310-313.
FShr, K. J., Scott, J., Ahnert-Hilger, G., and Gratzl, M. (t989). Charac-
terization of the inositol 1,4,5-trisphosphate-induced calcium release
from permeabilized endocrine cells and its inhibition by decavanadate
and p-hydroxymercurybenzoate. Biochem. J. 262, 83-89.
FShr, K. J., Wahl, Y., Engling, R., Kemmer, T. P., and Gratzl, M. (1991).
Decavanadate displaces inositol 1,4,5-trisphosphate (IP3) from its re-
ceptor and inhibits IP3 induced Ca~÷release in permeabilized pancre-
atic acinar cells. Cell Calcium 12, 735-742.
Frank, T. M., and Fein, A. (1991). The role of inositol phosphate cas-
cade in visual excitation of invertebrate microvillar photoreceptors. J.
Gen. Physiol. 97, 697-723.
Gomez, M., and Nasi, E. (1994a). The light-sensitive conductance of
hyperpolarizing invertebrate photoreceptors: a patch-clamp study. J.
Gen. Physiol. 103, 939-956.
Gomez, M., and Nasi, E. (1994b). Blockage of the light-sensitive con-
ductance in hyperpolarizing photoreceptors of the scallop. Effects of
tetraethylammonium and 4-aminopyridine. J. Gen. Physiol. 104,487-
505.
Gorman, A. L. F., and McReynolds, J. S. (1969). Hyperpolarizing and
depolarizing receptor potentials in the scallop eye. Science 165, 309-
310.
German, A. L. F., and McReynolds, J. S. (1974). Control of membrane
K÷ permeability in a hyperpolarizing photoreceptor: similar effects of
light and metabolic inhibitors. Science 185, 620-621.
Gosh, T. K., Eis, P. S., Mullaney, J. M., Ebert, C. L., and Gill, D. L.
(1988). Competitive, reversible and potent antagonism of inositol 1,4,5
trisphosphate-activated calcium release by heparin. J. Biol. Chem.
263, 11075-1 t079.
Gotow, T. (1989). Photoresponse of an extraocular photoreceptor as-
sociated with a decrease in membrane conductance in an opistho-
branch mollusc. Brain Res. 479, 120-129.
Gotow, T., and Nishi, T. (1991). Roles of cyclic GMP and inositol tris-
phosphate in phototransduction of the molluscan extraocular photore-
ceptor. Brain Res. 557, 121-128.
Gotow, T., Nishi, T., and Kijima, H. (1994). Single K÷ channels closed
by light and opened by cyclic GMP in molluscan extraocular photore-
ceptor cells. Brain Res. 662, 268-272.
Guy, H. R., Durelt, S. R., Warmke, J., Drysdale, R., and Ganetzky,
B. (1991). Similarities in amino acid sequences of Drosophila eag and
cyclic nucleotide-gated channels Science 254, 730.
Hardie, R. C., and Minke, B. (1992). The trp gene is essential for a
light-activated Ca 2+ channel in Drosophila photoreceptors. Neuron 8,
643-651.
Heginbotham, L., Abramson, T., and MacKinnon, R. (1992). A func-
tional connection between the pores of distantly related ion channels
as revealed by mutant K÷ channels. Nature 258, 1152-1155.
Hollander, M., and Wolfe, D. A. (1973). Nonparametric Statistical Meth-
ods (New York: John Wiley & Sons).
Inoue, H., Yoshioka, T., and Hotta, Y. (1985). A genetic study of inositol
trisphosphate involvement in phototransduction using Drosophila mu-
tants. Biochem. Biophys. Res. Commun. 132, 513-519.
Johnson, E. C., Robinson, P. R., and Lisman, J. E. (1986). Cyclic GMP
is involved in the excitation of invertebrate photoreceptors. Nature
324, 468-470.
Kaupp, U. B. (1991). The cyclic nucleotide-gated channels of verte-
brate photoreceptors and olfactory epithelium. Trends Neurosci. 14,
150-157.
Kaupp, U. B., Niidome, T., Tanabe, T., Terada, S., B6nigk, W.,
St~hmer, W., Cook, N. J., Kangawa, K., Matsuo, H., Hirose, T., et al.
(1989). Primary structure and functional expression from complemen-
tary DNA of the rod photoreceptor cyclic GMP-gated channel. Nature
342, 762-766.
Koutalos, Y., Nakatani, K., and Yau, K.-W. (1995). Cyclic GMP diffusion
coefficient in rod photoreceptor outer segments. Biophys. J. 68, 373-
382.
Kuno, M., and Gardner, P. (1987). Ion channels activated by inositol
1,4,5-trisphosphate in plasma membrane of human T-lymphocytes.
Nature 326, 301-304.
McReynolds, J. S., and German, A. L. F. (1970). Membrane conduc-
tances and spectral sensitivities of Pecten photoreceptors. J. Gen.
Physiol. 56, 392-406.
Miller, W. H. (1958). Derivatives of cilia in the distal sense cells of the
retina of Pecten. J. Biophys. Biochem. Cytol. 4, 227-228.
Minke, B., and Selinger, Z. (1992). Inositol lipid pathway in fly photore-
ceptors. Excitation, calcium mobilization and retinal degeneration. In
Progress in Retinal Research, Vol. 11, N. N. Osborne and G. J. Chader,
eds. (Oxford: Pergamon Press), pp. 99-124.
Nakatani, K., and Yau, K.-W. (1988). Guanosine 3',5'-cyclic monophos-
phate activated conductance studied in a truncated rod outer segment
of the toad. J. Physiol. 395, 731-753.
Nasi, E. (1991). Electrophysiotogical properties of isolated photorecep-
tors from the eye of Lima scabra. J. Gen. Physiol. 97, 17-34.
Nilsson, T. J., Zwiller, J., Boynton, A. L., and Berggren, P. O. (1988).
Heparin inhibits IP3-induced Ca2÷release in permeabilized pancreatic
beta-cells. FEBS Lett. 229, 211-214.
Payne, R., Corson, D. W., and Fein, A. (1986a). Pressure injection of
calcium both excites and adapts Limulus ventral photoreceptors. J.
Gen. Physiol. 88, 107-126.
Payne, R., Corson, D. W., Fein, A., and Berridge, M. J. (1986b). Excita-
tion and adaptation of Limulus ventral photoreceptors by inositol 1,4,5
trisphosphate result from a rise in intracellular calcium. J. Gen. Physiol.
88, 127-142.
Peretz, A., Suss-Toby, E., Rom-Glas, A., Arnon, A., Payne, R., and
Minke, B. (1994). The light response of Drosophila photoreceptors is
accompanied by an increase in cellular calcium: effects of specific
mutations. Neuron 12, 1257-1267.
Ranganathan, R., Bacskai, B. J., Tsien, R. Y., and Zuker, C. S. (1994).
Cytosolic calcium transients: spatial localization and role in Drosophila
photorecepter cell function. Neuron 13, 837-848.
Restrepo, D., Teeter, J. H., Honda, E., Boyle, A. G., Marecek, J. F.,
Prestwich, G. D., and Kalonoski, D. L. (1992). Evidence for an InsP~-
gated channel protein in isolated rat olfactory cilia. Am. J. Physiol.
263, C667-C673.
Selinger, Z., and Minke, B. (1988). Inositol lipid cascade of vision stud-
ied in mutant flies. Cold Spring Harb. Symp. Quant. Biol. 53, 333-
341.
Shin, J., Richard, E. A., and Lisman, J. E. (1993). Ca2÷is an obligatory
intermediate in the excitation cascade of Limulus photoreceptors. Neu-
ron 11,845-855.
Snyder, P. M., Krause, K.-H., and Welsh, M. J. (1988). Inositol trisphos-
phate isomers but not inositol 1,3,4,5-tetrakisphosphate induce cal-
cium influx in Xenopus laevis oocytes. J. Biol. Chem. 263, 11048-
11051.
Supattapone, S., Worley, P. F., Baraban, J. M., and Snyder, S. H.
(1988). Solubilization, purification and characterization of an inositol
trisphosphate receptor. J. Biol. Chem. 263, 12132-12136.
Tokuyasu, K., and Yamada, E. (1959). The fine structure of the retina
studied with the electron microscope. IV. Morphogenesis of outer seg-
ments of retinal rods. J. Biophys. Biochem. Cytol. 6, 225-230.
Tones, M. A., Bootman, M. D., Higgins, B. F., Lane, D. A., Pay,
G. F., and Lindhal, U. (1989). The effect of heparin on the inositol
Neuron
618

1,4,5-trisphosphate receptor in rat liver microsomes. FIEBS Lett. 252,

105-108.
Trowell, S. C. (1988). Inositol trisphosphatase and bisphosphatase
activities in the retina of crab. FEBS Lett. 238, 281-284.
Werner, U., Suss-Toby, E., Rom, A., and Minke, B. (1992). Calcium
is necessary for light excitation in barnacle photoreceptors. J. Comp.
Physiol. 170, 427-434.
Worley, P. F., Baraban, J. M., Supattapone, S., Wilson, V. S., and
Snyder, S. H. (1987). Characterization of inositol trisphosphate recep-
tor binding in brain. J. Biol. Chem. 262, 12132-12136.
Yau, K.-W. (1994). Cyclic-nucleotide-gated channels: an expanding
new family of ion channels. Proc. Natl. Acad. Sci. USA 91, 3481-3483.
Yau, K.-W., and Baylor, D. A. (1989). Cyclic GMP-activated conduc-
tance of retinal photoreceptor cells. Annu. Rev. Neurosci. 12, 289-
327.
Zimmerman, A. L., Yamanaka, G., Eckstein, F., Baylor, D. A., and
Stryer, L. (1985), Interaction of hydrolysis resistant analogs of cyclic
GMP with the phosphodiesterase and light-sensitive channel of retinal
rod outer segments. Proc. Natl. Acad. Sci. USA 82, 8813-8817.