Neuron,

Vol. 8, 643-651,

April,

1992, Copyright

0

1992 by Cell Press

The frp Gene Is Essential for a Light-Activated Ca*+ Channel in Drosophila Photoreceptors Roger C. Hardie* and Baruch Minke+ *Department of Zoology Cambridge University Cambridge CB2 3EJ England +Department of Physiology and the Minerva Center for Studies of Visual Transduction The Hebrew University Hadassah Medical Jerusalem 91010 Israel

School

Summary Invertebrate phototransduction is an important model system for studying the ubiquitous inositol-lipid signaling system. In the transient receptor potential (trp) mutant, one of the most intensively studied transduction mutants of Drosophila, the light response quickly declines to baseline during prolonged intense light. Using whole-cell recordings from Drosophila photoreceptors, we show that the wild-type response is mediated by at least two functionally distinct classes of light-sensitive channels and that both the frp mutation and a Ca*+ channel blocker (La3+) selectively abolish one class of channel with high Caz+ permeability. Evidence is also presented that Ca*+ is necessary for excitation and that Caz+ depletion mimics the trp phenotype. We conclude that the recently sequenced frp protein represents a class of light-sensitive channel required for inositide-mediated Ca*+ entry and suggest that this process is necessary for maintained excitation during intense illumination in fly photoreceptors. Introduction A large body of evidence indicates that excitation in invertebrate rhabdomeric photoreceptors is at least partially mediated by the widely used inositol-lipid signaling cascade (Fein et al., 1984; Brown et al., 1984; Payne et al., 1988; Yoshioka et al., 1985; Devary et al., 1987; Nagy, 1991). The Drosophila compound eye has been an important model for studies of this system because of its molecular genetic potential and, in particular, the ability to analyze the phototransduction machinery using mutations at a variety of key stages in the process (reviews, Pak, 1979,199l; Selinger and Minke, 1988; Minkeand Selinger, 1992a; Ranganathan et al., 1991b). For example, intracellular recordings show that severe mutations of the norpA gene, which codes for a light-activated phospholipase C (Bloomquist et al., 1988; Toyoshima et al., 1990), lead to a complete loss of all responsiveness to light except for the early receptor potential (Minke and Selinger, 1992b), strongly suggestingthatexcitation is mediated exclusively via the inositol signaling pathway.

One of the most intensively investigated “transduction mutants” in Drosophila is the transient receptor potential mutant, trp (Cosens and Manning, 1969; Minke et al., 1975; Minke, 1982). The trp mutant and its counterpart nss (no steady state) in Lucilia(Howard, 1984) are so called because the light response decays to baseline during prolonged bright illumination. In addition, many manifestations of light adaptation, including the speeding up of response kinetics, reduction in bumpamplitude,and migration of intracellular pigment granules, also appear to be blocked by this mutation (Lo and Pak, 1981; Minke, 1982; Kirschfeld and Vogt, 1980; Howard, 1984; Suss et al., 1989; Minke and Selinger, 1992a). Since Caz+ is known to be an intracellular messenger of adaptation (Lisman and Brown, 1972) and probably also excitation (Bolsover and Brown, 1985; Payne et al., 1986; Hardie, 1991a) in arthropod photoreceptors, it has been suggested that the mutation interferes with the normal rise in intracellular Ca*+ during light (Suss-Toby et al., 1991; Minke and Selinger, 1992a). A major source of Ca2+ is found in inositol 1,4,5-trisphosphate (IP&sensitive pools of the endoplasmic reticulum, represented in arthropod photoreceptors by the so-called submicrovillar cisternae (Walz, 1982; Baumann et al., 1991). To account for the above results Minke and Selinger (1992a) have suggested that the trp gene may code for a plasma membrane “Ca *+ transporter” essential for replenishing these intracellular stores and that in its absence, the stores become rapidly depleted, resulting in both excitation and adaptation becoming blocked. The trp gene, which has recently been cloned and sequenced (Monte11 and Rubin, 1989; Wong et al., 1989), codes for a novel membrane protein. Minke and Selinger’s model, which also proposes that the trp protein is activated via direct coupling to the IP3 receptor, implicates this novel protein in the process of inositide-mediated Ca*+ entry, presently one of the most controversial areas in cellular Ca*+ homeostasis (Takemura et al., 1989; Berridge, 1990; Irvine, 1990, 1991; Bird et al., 1991). Recently it has proved possible to make voltageclamp measurements of the light-induced current (LIC) in Drosophila photoreceptors using the wholecell recording technique on dissociated ommatidia (Hardie, 1991a, 1991b; Ranganathan et al., ,1991a, 1991b). Unexpectedly the results indicated that the light-sensitive channels are primarily permeable to Ca*+. It might seem that these themselves should be a sufficient source of fresh Ca2+ to replenish the stores. To reconcile this finding with the trp phenotype and Minke and Selinger’s hypothesis, it was therefore suggested (Hardie, 1991a) that the trp gene product might in fact represent the light-sensitive channels. To test this suggestion directly, the LIC has now been investigated in the trp mutant and in wildtype flies in the presence of La3+, which mimics many

NWJKXl 644

al wild -

type

wild

tw -

-

r cl

b1 -

type

(0

-

Ca*+)

-

-2.0 -1.0 0.0

Ils 100

Figure sponse cated),

1. Whole-Cell to Prolonged at a Holding

pA

Voltage-Clamped Inward Currents (3 s) Stimuli (Relative Log Intensity Potential of -40 mV

in ReIndi-

(a) Wild type and (b) trp both recorded in normal Ringer’s solution with 2 mM extracellular Cal+. At higher intensities adaptation reduces the initial transient to a steady-state value in wild type, whilst in trp photoreceptors, there is a simple exponential decay of the response to baseline (same intensities as wild type). (c) The trp phenotype is closely mimicked in wild-type photoreceptors following prolonged exposure (>I hr) to nominally zero Ca2+ Ringer’s solution (2.5 uM extracellular Cal+), in this casewith intracellular Ca*+ buffered to 50 nM with 5 mM EGTA.

of the effects of the trp mutation (Hochstrate, 1989; Suss-Toby et al., 1991) and has been hypothesized to interfere with the trp gene product directly. The results show clearly that both external La3+ and the trp mutation affect the basic properties of the lightsensitive conductance itself and in particular demonstrate that the trp gene product is essential for its high Ca2+ permeability. Together with a recent study showing homologies of the trp protein with voltagesensitive Ca2+ channels (Phillips et al., 1992), we conclude that trp is a structural gene for one class of light-sensitive channel.

Results Voltage-Clamped Measurements and trp Photoreceptors

of Adult Wild-Type

Previous studies (Hardie, 1991a, 1991b; Ranganathan et al., 199la) of the Drosophila LIC used photoreceptors from late stage pupae, since their ommatidia are easier to dissociate and have better electrical propertiesforwhole-cell recording(higher input resistances, lower cell capacitance). In an early stage of the present study itwas found that photoreceptors prepared from late stage trp pupae usually have no or very little response to light. Reliable responses to light were first found only after the flies had emerged, and thus experiments were performed on adult flies within 24 hr of emergence. All control experiments were performed on wild-type adult flies at the same stage to ensure that the basic properties of the LIC described previously (Hardie, 1991a) were essentially the same in adults as in pupae.

Figure 1 shows responses to prolonged lights of varying intensities in cells clamped near resting potential (-50 mV). In wild type, dim light produces a maintained, basically square wave inward current. At higher intensities there is a rapid transient that declines to a lower steady-state value due to light adaptation. At the brightest intensity this is seen to develop a complex waveform with a notch. In previous measurements made from membrane voltage, this might have been attributed, for example, to voltage-dependent K’ channels, which are a dominant feature of these cells (Hardie, 1991b); however, under the present conditions (voltage clamp and intracellular substitution of K+ for Cs’) it must be attributed to the complex dynamics of light adaptation mediated, at least in part, via influx of Ca2+ ions from the extracellular space (Hardie, 1991a; Ranganathan et al., 1991a). In the trp mutant the response to dim light looks nearly normal as previously reported; however, at higher intensities the response decays completely to baseline, only recovering after about 1 min in the dark. Furthermore at all intensities, the decay follows a single exponential time course, the time constant simplydecreasing with intensity. This is the typical trp phenotype (Cosens and Manning, 1969; Minke et al., 1975), which has previously been shown to be closely mimicked by micromolar concentrations of external La3+ in wildtype photoreceptors (Hochstrate, 1989; Suss-Toby et al., 1991). Figure Ic shows that prolonged exposure to nominally zero extracellular Ca?+ (2.5 PM) combined with tion nM)

5 mM EGTA buffering of the internal (free intracellular Ca*+ concentration, can also mimic this behavior.

Ionic Selectivity

of the Light-Sensitive

pipette solu[Ca’+],, 50

Channels

In wild-type Drosophila photoreceptors the reversal potential (E,,,) of the LIC shows a strong dependence on extracellular Ca’+. Assuming constant field theory the Ca*+-dependent shift in reversal potential indicates that the channels are about 40x more permeable to Ca*+ than monovalent cations such as Na’and Cs’(see Figure 3a; Hardie, 1991a). Measurements performed in trp mutants (Figure 2; Figure 3) reveal a striking difference: E,,, is significantly more negative, and its dependence upon extracellular Ca’+ is much reduced. This clearly demonstrates that the ionic selectivity of the channels is altered in the mutant and specifically that their permeability to CaL+ is greatly reduced. Identical measurements were also made in wild-type photoreceptors in the presence of 10 kM La-l+. The resultswere indistinguishable from those for trp photoreceptors (Figure 3a). The data in Figure 3a are for the peaks of flash responses, and the question arises whether the plateau phase of the response has the same ionic selectivity. Figure 3b shows that this is the case in both wild-type and trp flies, since the peaks reverse at the same potential as the plateau (see also, Hardie, 1991a). To obtain measurable plateau potentials in trp photoreceptors, we used trp flies raised at 19%, which, as shown

trp: A Light-Sensitive 645

Ca’+ Channel

in Drosophila

al

al wild

type

A

DA

200

ms

L

La3+

trP

100

30 5: -5

Relative

20 -

5

Ca Mg cs Na

permeability w.t frP 40 3.5 8 1.9 1 1 1 0.7

‘;

-10

1 ,

,

,

, ,,,,

.005 .Ol

, ,,,, .l

external

bl

b) holding -40

potential

(mV)

0

[CaZ+l

10

mM

10 mM Ca++ wild

w 40

, , , ,,I 1

type

80

500 400

2 E :: ?! x : a

200 wild

100

type

-

0 -100

Figure

-200

Figure 2. Flash mM Extracellular

Responses Cal’)

at Different

Holding

Potentials

(0.5

(a) Voltage-clamped responses to identical 20 ms flashes (log -2.0) delivered at times indicated by arrows at holding potentials between -60 and +40 mV in 20 mV steps. Cells bathed in normal Ringer’s solution with 0.5 mM extracellular Ca’+. The wild-type response kinetics show a strong voltage dependence, which is greatly reduced in trp flies or wild-type flies in the presence of external LaJc (IO PM). (b) Current-voltage relationship for the peak responses in (a): notice that the reversal potential in both trp and La’+-treated photoreceptors is shifted by about -10 mV with respect to wild type. In addition, the marked inward rectification of the wildtype current-voltage curve is absent in trp and La’+-treated flies. All curves become strongly outwardly rectifying at depolarized holding potentials.

by Minke (1983), maintain a significant plateau at higher intensities. Although the permeability to Ca2+ is drastically reduced in trp, Figure3 shows that there is in fact a small but significant (p < .OOl) Ca2+-dependent shift in E,,, consistent with a relative permeability (Pca:Pcs) of about 3.5:1. To demonstrate permeability to divalent cations more directly we replaced all monovalent cations with the impermeant cation N-methyl-o-glucamine, leaving Mg*+ as the only possible charge carrier for inward currents (Figure 4). As expected, in the absence of external Mg2’, only outward currents could be detected in the N-methyl+glucamine Ring-

3.

trp and

La’+ Reduce

Ca’+ Permeability

of the

LIC

(a) E,,, as a function of [Ca2’],, in wild type (closed squares), trp (open squares), and wild-type photoreceptors in the presence of 10 PM external LaJ+ (closed triangles). E,., in trp and La’+-treated flies is significantly more negative and shows a reduced Ca’+ dependency. Data points are means (f SD) from flash responses in at least eight cells for each point, corrected for junction potential (-3 mV). The theoretical curves were calculated assuming constant field theory using the permeability ratios indicated (see Experimental Procedures). (b) Responses to 1 s steps of light (log -2.25) in 10 mM extracellular Cal’ at 10 mV intervals between voltages indicated. Peak and plateau reverse at the same potential: E,,, trp = f5 mV, wild type = +24 mV (after correction for junction potential). Notice also the greater noise in the plateau of the trp response. This is because of the lack of adaptation in trp (Minke et al., 1975).

er’s solution; however, the addition of 8 mM Mg*’ allowed inward currents to be measured in both wildtype and trp photoreceptors. The very different reversal potentials (trp, -28 mV; wild type, -3 mV) again directly confirm the much reduced permeability for divalent cations in the mutant (Figure 4). Although Ca2+ can also be shown directly to permeate the channels using this strategy (Hardie, 1991a) quantitative measurements of this nature using Ca2+ proved problematic, since in the presence of external Ca2+ and absence of external Na’ there is rapid and irreversible loss of all responsiveness to light, probably because of Na+-Ca2+ exchange (Minke and Tsacopoulos, 1986; Ziegler and Walz, 1989) running in reverse.

N~UUXl h46

0 Mg2+

0.5 mM Ca2+

6 mM Mg”

10 mM Caz+

wild type

wild type

100 pA 100 ms

L

rrp Figure ions

4. Direct

0 mV -60 Demonstration

of Permeation

by Divalent

mV Cat-

Responses to identical 20 ms flashes (log -1.5) delivered at times indicated by arrows at a variety of holding potentials in 20 mV steps starting at -60 mV in ranges indicated. (Left) With N-methylo-glucamine (130 mM) as the only external cation (0 Mg2’), only outward currents are detected. (Right] With 8 mM Mg*+ added to the bath (different cell from the same fly), inward currents carried by Mg2+ are observed in both wild type and trp. The reversal potentials (-3 f 2 mV, n = 5 in wild type; -28 f 3 mV, n = 8 in trp) demonstrate that the permeability for divalent cations in trp, whilst finite, is greatly reduced compared with wild type.

Evidence for Two Classes of Light-Sensitive Channels Figures 2-4 demonstrate that both the trp mutation and micromolar concentrations of external Las+ alter the ionic selectivity of the light-activated conductance.The most direct interpretation of these findings is that the trp gene codes for a subset of the lightsensitive channels. This in turn implies the existence of a second class of light-sensitive channel with lower Caz+ permeability. Unless the kinetics of these two channels are identical the LIC should thus not show a unique reversal potential as previously reported (Hardie, 1991a). A detailed reexamination now reveals that under certain conditions the LIC does indeed exhibitabiphasicreversal potential indicativeofchannels with different ionic selectivities (Figure 5) as has also been reported in Limulus (Nagy, 1991) and the mollusc Lima (Nasi, 1991). Surprisingly perhaps, such biphasic reversal potentials are most clearly seen with rather low extracellular Ca*+ concentrations (0.5 mM), at which the reversal potentials of the two components are rather close (Figure 3a). Under these conditions, at a holding potential of around +I0 mV, the LIC is initially outward, before becoming inward during the later phase of the response. This indicates that the early phase of the response is carried by channels with relatively low Ca2+ permeability, whilst channels highly permeable to Ca2+ dominate the later phase and the plateau (Figure 3b). At higher Caz+concentrations, the reversal potential appears to be unique, as previously reported (Hardie, 1991a), and indicative of high Ca*+ permeability (Figure 3b; Figure 5). Apparently, either the kinetics of the two classes of channel are now identical or the second class of channel (with lower Ca2+ permeability) no

Figure

5. Biphasic

Reversal

Potential

Flash responses (IO ms flashes, log -1.5, at time indicated by artifact), recorded at 5 mV intervals within the ranges indicated. (Left) In 0.5 mM extracellular Caz+, wild-type photoreceptors show a biphasic reversal potential (arrows) between +I0 and +I5 mV. This biphasic behavior was not observed at IO mM extracellular Ca*+ or in trp photoreceptors at either Ca*+ concentration. Notice also that raising the [Caz+10 shifts E,,, by 15-20 mV in wildtype photoreceptors but by only about 5 mV in trp. The inset shows three flash responses (wild type) to increasing intensity (-2.0, -1.0, and 0.0) all at +I0 mV: at the highest intensity there is apparently yet a third (early inward) component to the currents.

longer operates under these conditions. In trp photoreceptors, the reversal potential appears to be unique with both low and high extracellular Caz+ concentration ([Ca*‘],) (Figure 5). With very bright flashes, there were indications of yet a third component to the responses (inset Figure 5). This was not studied further in the present investigation, since the very large conductances activated at these intensities result in unreliable space-clamp conditions. Voltage Dependence of Response Kinetics In wild-type flies, hyperpolarization greatly accelerates the kinetics of the light response in a Ca*+dependent manner (Figure 2; Figure 6; Hardie, 1991a). Previously evidence was presented indicating that this effect is due to sequential positive and negative feedback by Ca2+ permeating the light-sensitive channels-with more Ca2+ entering as the cell is hyperpolarized (Hardie, 1991a). Since the light-sensitive channels in trp appear to be some 10x less permeable to Ca*+ (Figure 3), one might predict that this effect would be attenuated in the mutant. This indeed is the case, and in trp there are only very slight voltagedependent shifts in the time to peak (Figure 6 [right], closed symbols). In wild-type photoreceptors, external La3+ (IO PM) has essentially the same effect (Figure 6 [right], open symbols). The control experiments in adult wild-type flies basically confirm previous measurements (Hardie, 1991a); however, at low (0.5 mM) [Ca2+10 the voltage dependence of time to peak shows a marked discontinuity, which is a reflection of the biphasic reversal potential (Figure 5), and with nominally zero extracellular Ca2+ the time to peak actually shows a small, but significant decrease with depolar-

trp: A Light-Sensitive

Cal+ Channel

in Drosophila

647

wild

type

trp

Figure 6. Voltage sponse Kinetics

/ La’+

VI

0 ..

1, I -60-40-20

I,

I, 0

I 20

I, 40

I 60

holding

I,, -60-40-20

potential

I

/. 0

I, 20

I. 40

0.5

mM

q w

2 mM

O+

10 mM

I 60

(mV1

ization (Figure 6 [left]). Hardie (1991a) also attributed the marked inward rectification of the LIC currentvoltage relationship at negative holding potentials to positive feedback by Ca2+ permeating the light-sensitive channels. Figure 2b shows that both the trp mutation and La3+ abolish this inward rectification, and the current-voltage relationship is now approximately linear at negative holding potentials.

Ca2- Dependence of Excitation Minke and Selinger (1992b) explained the decay of the light response in trp by assuming that a rise in [Ca2+]i is necessary for excitation and that the limited amount of Ca2+ in the intracellular pools can only support excitation for a short time during intense light. The apparent positive feedback of Ca*+ in accelerating response kinetics (Hardie, 1991a; Figure 2; Figure 6) suggests that Ca2+ does play an excitatory role in fly phototransduction; however, direct evidence, such as exists in Limulus (Bolsover and Brown, 1985; Payne et al., 1986), barnacle (Brown et al., 1988), and bee photoreceptors (Walz, 1992), has been lacking. To provide direct evidence for an excitatory role of Ca2+ in Drosophila we bathed cells in nominally zero extracellular Ca2+ (2.5 vM) for at least 30 min, presenting adapting flashes (3 s, log -1.0) every 5 min to encourage depletion of the intracellular pools. Recordings were then made alternately using intracellular Ca*+ buffered to either 50 nM or 700 nM with 5 mM EGTA in the recording pipette. With 50 nM intracellular Ca2’, sensitivity to light was greatly reduced in both wild type and the trp mutant under these conditions (Figure 7). However, the inclusion of 700 nM intracellular Ca*+ in the pipette rapidly restored sensitivity (within a few seconds). No exception to this behavior was found in five wild-type flies (39 cells) and six trp flies (44 cells), clearly indicating an excitatory role for intracellular Ca2+ in both wild type and the trp mutant. The only obvious difference between wild type and trp in this experiment was that sensitivity in trp was consistently about 0.5 log units lower than in wild type.

Dependence

of

Re-

Time to peak of flash responses plotted against holding potential at various extracellular Ca:+ concentrations as indicated. Each curve is from a different cell. (Left) Wild type; thetime to peak shows a marked dependence on both voltage and [CaZ+],; the discontinuity in the curve at 0.5 mM Ca” is a reflection of the biphasic reversal potential (see Figure 5). (Right) In trp photoreceptors (closed symbols) and wild-type photoreceptors in the presence of 10 FM external La’+ (open symbols) both voltage and Ca2+ dependence are much reduced. With nominally zero extracellular Ca” (2.5 PM), voltage dependence of the time to peak is also greatly reduced in wild type (left open circles).

Discussion The main purpose of this study was to test the hypothesis that the trp gene encodes a class of light-activated Ca*+ channel induced via the phosphoinositide pathway, and in the following we will first present arguments that do indeed lead to this important conclusion. We will then consider mechanisms by which a defect in such a channel might lead to the trp phenotype. Although the resolution of this issue does not influence our main conclusion, it may have important implications forthe mechanism of phototransduction in invertebrate photoreceptors and more generally the mechanism of receptor-mediated Ca*+ entry. frp Encodes a light-Activated Ca2+ Channel These results demonstrate that the trp mutation alters fundamental properties of the light-sensitive channels. Where tested, these effects are closely mimicked by external application of micromolar concentrations of Lb+, thus complementing detailed quantitative studies on the effects of La3+ in both Drosophila and larger flies (Hochstrate, 1989; Suss-Toby et al., 1991). In trpfliesthe relativeCa*+ permeability(Pc,:P&ofthe light-sensitive conductance is reduced about IO-fold (Figure 3). In addition the voltage dependence of the response kinetics and inward rectification, previously attributed to feedback by Ca2+ permeating the channels (Hardie, 1991a), are almost abolished (Figure 2; Figure 6). The most obvious interpretation of these findings is that there are at least two classes of lightactivated channel and that the trp gene codes for a subset of channels with high Ca*+ permeability. This conclusion is supported by the demonstration of a biphasic reversal potential in wild-type (but not trp) photoreceptors, similar to the behavior reported in photoreceptors of Limulus (Nagy, 1991) and the mollust Lima (Nasi, 1991). The suggestion that frp encodes a plasma membrane Ca2+ channel is strongly supported by a recent reevaluation of the trp protein sequence (Phillips et

a)

1000 1

wild

type

a a

-3.5

bl

IOOO,

-3.0

-2.5

-2.0

-2.5

-2.0 log

7. Calcium

-0.5

0

w

-3.0

Figure

-1.0

-1.5

Dependence

-1.5

-1.0

-0.5

0

I

of Excitation

Response (PA) versus intensity (log I) functions for (a) wild-type and (b) trp photoreceptors recorded after at least 30 min exposure to Ca’+-free Ringer’s solution (2.5 FM). Each curve is generated from a series of flash responses at a holding potential of -40 mV in a different photoreceptor from the same preparation (e.g., insets in [b], with log attenuation range indicated), 1-2 min after establishing the whole-cell configuration. With intracellular Ca2+ in the recording pipette buffered to 50 nM, sensitivity is very low, some cells only responding weakly to flashes of the brightest intensity available, e.g., single data points in (b). With 700 nM intracellular Ca*+, however, sensitivity was quickly restored (compare dotted lines, which are from responses determined in 0.5 mM extracellular Ca2+ and no EGTA in the pipette). The inset in (a) shows responses to identical flashes (log -2.5) repeated at 8 s intervals starting 4 s after establishing the whole-cell configuration. Responses are rapidly facilitated and, in this case, superimposed upon a gradually developing noisy inward current that may represent stimulus-independent excitation of channels by Ca”. In these experiments the whole-cell pipette solution contained 140 mM KCI, 10 mM TES, 4 mM MgS04, and 5 mM EGTA.

al., 1992). Original reports of the primary structure (Monte11 and Rubin, 1989; Wong et al., 1989) revealed no homologies with known proteins, although it was concluded that it represented an integral membrane protein associated with the rhabdomeres. Phillips et al. (1992), however, have recently sequenced a trp homolog gene (t/p/) from the Drosophila eye and now report that both trp and trpl have significant homologies with vertebrate brain voltage-sensitive Ca*+ channels (dihydropyridine receptors), particularly in the putative membrane-spanning regions S3, S4, S5, and S6. Interestingly in the S4 region, the positively charged residues believed to represent the voltage

sensors (Noda et al., 1984) are missing. Both trp and h-p/ encode only one membrane domain compared with the four found in the dihydropyridine receptor and are thus likely to represent single subunits of a tetrameric channel (Phillips et al., 1992). Whether they are subunits of the same or a different channel remains to be seen. Taken together, these results provide compelling evidence that the trp gene is a structural gene for a light-activated channel with high Ca2+ permeability. Presumably the pathway leading to the activation of this channel must be the light-activated inositide cascade, since, with the exception of the early receptor potential, a// responsiveness to light in Drosophila is eliminated by null mutations of the norpA gene (Minke and Selinger, 1992133, which codes for a lightactivated phospholipase C (Bloomquist et al., 1988; Toyoshima et al., 1990). Phospholipase C produces two intracellular messengers, IP3 and diacylglycerol, the latter exerting its effects via protein kinase C. The trp-dependent channel is apparently activated via the IP3 arm of the pathway (rather than protein kinase C), since in the inaC mutant, which lacks eye-specific protein kinase C (Smith et al., 1991), the Ca*+ permeability of the light-induced conductance is not affected (Ranganathan et al., 1991a). These considerations lead to the conclusion that the trp gene product is a membrane channel responsible for phosphoinositide-mediated Ca2+ entry. It will be interesting to see whether homologous proteinsare responsibleforthis process in other cells as well.

Receptor-Mediated Ca*+ Entry and the Origin of the trp Phenotype The trp phenotype was originally defined by the inability of the photoreceptors to sustain a maintained response during intense illumination (Cosens and Manning, 1969; Minke et al., 1975). One possibile explanation for this phenotype is that the trp-dependent channel alone underlies the steady-state response and that the other class of channel inactivates during prolonged intense stimulation. This simple model seems unlikely to be a sufficient explanation; however, since under certain conditions (in flies raised at 19”C), there is a significant steady-state response in trp photoreceptors even during bright lights (Minke, 1983). Here we use reversal potential measurements to confirm that it is the non-trp-dependent channels that sustain the plateau response under these conditions (Figure 3b). A different class of explanation has been proposed by Minke and Selinger (1992a) based on more general models of receptor-mediated Ca*+ entry in a variety of cells where the release of Ca*+ from the IP3-sensitive Ca2+ pools is followed by Ca*+ influx across the plasma membrane (Takemura et al., 1989; Berridge, 1990; Irvine, 1990). The mechanism underlying this coordinated process, which is important for efficient refilling of the pools, remains controversial (Berridge, 1990; Irvine, 1991) and represents one of the major unan-

trp: A Light-Sensitive I149

swered

questions

Ca”

Channel

in

cellular

in Drosophila

Ca*+

homeostasis.

By

analogy with the interaction between the ryanodine receptor and voltage-sensitive Ca*+ channels (dihydropyridine receptors) in muscle (Block et al., 1988), the models that attempt to explain this process of inositide-mediated Ca*+ entry postulate that a surface membrane Ca*+ channel (of unknown molecular identity) is linked to theendoplasmic reticulum viaadirect interaction with the IP3 receptor, which has extensive sequence homologies with the ryanodine receptor (Furichi et al.. 1989). How this hypothetical membrane channel is activated is a matter of intense speculation, with depletion of the luminal Ca2+ pool (Takemura et al., 1989; Berridge, 1990), binding of IP4 (Irvine, 1990), or both having been suggested (Irvine, 1991). Based on a large body of completely independent evidence, Minke and Selinger (1992a) suggested that the trp gene product represented such a membrane channel or transporter. Our direct demonstration of the role of trp in inositide-mediated Ca’J+ entry and the observation of Phillips et al. (1992) of homology between trp and the dihydropyridine receptor greatly strengthen this hypothesis. To explain the decay of the response in trp mutants, Minke and Selinger (1992a) assumed that a rise in intracellular Ca*+ is required for excitation, which cannot therefore be sustained without influx of Ca*+ to replenish the stores. This assumption is supported by our finding that Ca2’ restores sensitivity in Caz+-depleted cells (Figure 7), whilst, as also reported in the blowfly Calliphora (M. Mojet, personal communication) and the barnacle (Werner et al., 1992), Ca2+ depletion can induce a trplike decay in the response (Figure Ic). Finally, a strong prediction of the model, namely that the trp protein is located close to the submicrovillar cisternae, thus allowing direct interaction with the IP3 receptor, has recently been confirmed by immunolocalization of the trp protein to the base of the microvilli (J. A. Pollock, personal communication). Whilst our results are consistent with and indeed provide further support for Minke and Selinger’s hypothesis, the final resolution of the origin of the trp phenotype will probably require a complete understanding of the mechanism of phototransduction, which in invertebrates, in comparison with the situation in vertebrates, still leaves many questions to be answered.

A Novel

Possible Mechanism

of Excitation

Our results reinforce findings in Limulus (Nagy, 1991) and Lima (Nasi, 1991) that there appear to be two or more distinct classes of light-activated channels, although in Limulus none appear to be significantly permeable to Ca*+ (Brown and Blinks, 1974; Deckert and Stieve, 1991). The possibility thus exists that these may require different messengers for activation, such as Ca*+ (Payne et al., 1988) and cGMP (Johnson et al., 1986; Feng et al., 1991). The lack of specific pharmacological agents blocking one or the other channel in Limulus has proved an obstacle to analysis of these

potentially divergent pathways (but see Frank and Fein, 1991). The demonstration of both a mutation (trp) and a chemical agent (La3+) affecting only one class of channel in Drosophila can be expected to be of significant value in elucidation of the final pathways of excitation, which, at least in the fly, can presumably only diverge downstream from the activation of phospholipasec. Finally, it should be noticed that ouranalogy with models for receptor-mediated Ca2+ entry in other systems (Berridge, 1990; Irvine, 1991; Minke and Selinger, 1992a) suggests a mechanism of excitation not previously explicitly considered for invertebrate phototransduction, namely, that the trp-dependent light-sensitive channel may be gated by the reduction in luminal CaZ+ resulting from IP3-induced Ca*+ release from the submicrovillar cisternae. An interesting feature of this mechanism is that, at least in the short term, it should be resistant to strong Ca2+ buffering in the cytosol. Experimental

Procedures

Flies The wild-type strain was Oregon R. Both red and white eyed (w) flies were used with no obvious differences; the mutant allele was w; trpCM (Cosens and Manning, 1969), which is a genuine null (protein negative) strain (Monte11 and Rubin, 1989; Wong et al., 1989). Except for the experiment of Figure 3b flies were raised at 25T. Preparation, Recording, and Solutions Dissociated ommatidia were prepared and recorded from under dim red light as previously described (Hardie et al., 1991; Hardie, 1991a, 1991b) except that adult flies (less than 24 hr post emergence) were used rather than pupae. Unless otherwise stated bath Ringer’s solution contained 130 mM NaCI, 5 mM KCI, IO mM TES, 8 mM MgCI, (pH 7.15), with CaClz added as indicated on individual experiments, and whole-cell recording pipettes contained 120 mM CsCI, 15 mM tetraethylammonium chloride, 2mMMgS04,and10mMTES(pH 7.15);ECTAwaseitherincluded at low concentrations (0.1 mM) or omitted altogether, since higher concentrations result in profound changes in response kinetics. Data were recorded via an Axopatch IB amplifier (Axon Instruments Inc.) and sampled on line using pClamp 5.5 software. Voltage steps were applied 50-100 ms prior to light flashes from a holding potential of -40 or -50 mV. All recordings were made at room temperature (20°C k 1.5OC). Illumination, delivered by a light guide placed 2 cm over the bath, was from a 12 V QI lamp filtered via Wratten ND filters and a yellow (Schott OC530) cutoff filter to ensure that only responses from one class of photoreceptor (RI-R6) were recorded (Hardie, 1985, 1991a). Relative intensities were calibrated using a pin diode. Free Ca?’ concentrations in Ca-ECTA solutions or nominally zero Ca” solutions were calibrated in a fluorescence spectrometer using fura(below 1 PM) or flue-3 (above 1 PM). Clamp Quality Unsylgarded electrodes with resistances of 7-12 MO gave series resistances typically in the range 15-25 Mn. Although some of the largest responses measured involved series resistance errors (series resistance x total current) of up to 15 mV, for all critical quantitative data (e.g., reversal potentials) the calculated errors (which were corrected for) were less than 5 mV. Whole-cell capacitances were in the range 35-70 pF, generating clamp time constants (series resistance x whole-cell capacitance) usually less than 1 ms. Resting input resistances of 400-1000 Ma ensure excellent space clamp for the moderate currents analyzed in this study (see also Hardie, 1991a, 1991b).

Nf2UKNl 650

Determination of Permeability Ratios Reversal potentials were determined from families of responses at different holding potentials (e.g., Figure 2a), corrected for series resistance and junction potential errors. Relative permeability ratios were fitted to the data by eye using the GoldmanHodgkin-Katz constant current equation (e.g., Hille, 1984). PNs: PC, was estimated from data where Mg>+ and Cs’ were the only permeant ions (e.g., Figure 4), then Pr,:PN, was estimated from values determined in the absence of external Cal+, and finally PC. was varied to fit the data as extracellular Ca’+ was raised. Unity activity coefficients were assumed (leading, if anything, to an underestimate of the true Caj+ permeability). Acknowledgments We thank Drs. S. Laughlin, T. Cheek, and 2. Selinger for critical reading of an earlier version of this manuscript. This research was supported by grants from the Science and Engineering Research Council and the Royal Society (R. C. H.) and the National Institutes of Health, The US-Israel Binational Science Foundation, and the German Israeli Foundation (B. M.). 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

November

22, 1991;

revised

January

23, 1992.

Baumann, O., Walz, B., Somlyo, A. V., and Somlyo, A. P. (1991). Electron probe microanalysis of calcium releaseand magnesium uptake by endoplasmic reticulum in bee photoreceptors. Proc. Natl. Acad. Sci USA 88, 741-744. M. J. (1990).

Calcium

oscillations.

J. Biol.

Chem.

265,

Bird, C. S. J., Rossier, M. F., Hughes, A. R., Shears, S. B., Armstrong, D. L., and Putney, J. W. (1991). Activation of Ca@ entry into acinar cells by a non-phosphorylatable inositol trisphosphate. Nature 352, 162-165. Block, B. A., Imagawa, T., Campbell, K. P., and FranziniArmstrong, L. (1988). Structural evidence for direct interaction between the molecular components of the transverse tubule/ sarcoplasmic reticulum junction in skeletal muscle. J. Cell Biol. 107, 2587-2600. Bloomquist, B. T., Shortridge, R. D., Schneuwly, S., Pedrew, M., Montell, C., Steller, H., Rubin, C., and Pak, W. L. (1988). Isolation of a putative phospholipase C gene of Drosophila, norpA, and its role in phototransduction. Cell 54, 723-733. Bolsover, S. R., and Brown, J. E. (1985). Calcium an intracellular messenger of light adaptation also participates in excitation of Limulus ventral photoreceptors. J. Physiol. (Land.) 364, 381-393. Brown, H. M., Rydqvist, B., and Moser, H. (1988). Intracellular calcium changes in Balanus photoreceptors, a study with calcium ion-selective electrodes and Arsenazo III. Cell Calcium 9, 105-119. Brown, calcium tection

J. E., and Blinks, J. R. (1974). Changes in intracellular during illumination of invertebrate photoreceptors: with Aequorin. J. Gen. Physiol. 64, 643-665.

free de-

Brown, J. E., Rubin, L. J., Chalayini, A. J., Tarver, A. P., Irvine, R. F., Berridge, M. J., and Anderson, R. E. (1984). Myo-inositol polyphosphate may bea messenger for visual excitation in LimuIus photoreceptors. Nature 377, 160-163. Cosens, D. J., and Manning, A. (1969). Abnormal electroretinogram from a Drosophila mutant. Nature 224, 285-287. Deckert, changer, Limulus

A., and Stieve, H. (1991). Electrogenic Na+-Ca2+ exthe link between intra- and extracellular calcium in the ventral photoreceptor. J. Physiol. 433, 467-482.

Devary, 0. O., Heichal, Barash, C. T., Rubinstein, Coupling of photoexcited

A., Blumenfeld, C.T., Minke, rhodopsin

D., Cassel, E., Suss, S., B., and Selinger,Z. (1987). to inositol phospholipid

in fly photoreceptors.

Proc.

Natl.

Acad.

Sci. USA 84,

Fein, A., Payne, R., Corson, D. W., Berridge, M. J., and Irvine, R. F. (1984). Photoreceptor excitation and adaptation by inositol 1,4,5trisphosphate. Nature 377, 157-160. Feng, J. J., Frank, T. M., and Fein, A. (1991). Excitation of Limulus photoreceptors by hydrolysis-resistant analogs of cGMP and CAMP. Brain Res. 552, 291-294. Frank, T. M., and Fein, A. (1991). The role of the inositol phate cascade in visual excitation of invertebrate microvillar toreceptors. J. Cen. Physiol. 97, 697-723.

phospho-

Furichi, T., Yoshikawa, S., Miyawaki, A., Wada, K., Madea, and Mokoshiba, K. (1989). Primary structure and functional pression of the inositol 1,4,5-trisphosphate-binding protein Nature 342, 32-38. Hardie, R. C. (1985). Functional Prog. Sens. Physiol. 5, l-79.

organization

of the fly

N., exPu,.

retina.

Hardie, R. C. (1991a). Whole-cell recordings of the light induced current in dissociated Drosophila photoreceptors, evidence for feedback by calcium permeating the light-sensitive channels. Proc. R. Sot. Lond. (B) 245, 203-210. Hardie, R. C. (1991b). Voltage-sensitive Drosophila photoreceptors. J. Neurosci.

potassium channels 17, 3079-3095.

in

Hardie, R. C., Voss, D., Pongs,O., and Laughlin, S. B. (1991). Novel potassium channels encoded by the Shaker locus in Drosophila photoreceptors. Neuron 6, 477-486. Hille, land,

References

Berridge, 9583-9586.

hydrolysis 6939-6943.

B. (1984). Ionic Massachusetts:

Channels Sinauer

of Excitable Associates).

Membranes

(Sunder-

Hochstrate, P. (1989). Lanthanum mimicks the trp photoreceptor mutant of Drosophila in the blowfly Cal1iphora.J. Comp. Physiol. (A) 166, 179-188. Howard, gration

J. (1984). Calcium enables photoreceptor in a mutant fly. J. Exp. Biol. 713, 471-475.

Irvine, R. F. (1990). “Quantal” entry by inositol phosphates-a 263, 5-9.

Cal+ release possible

pigment

mi-

and the control of Cal’ mechanism. FEBS Lett.

Irvine, R. F. (1991). lnositol tetrakisphosphate as a second messenger, confusions, contradiction and a potential resolution. Bioessays 73, 419-427. Johnson, E. C., Robinson, P. R., and Lisman, GMP is involved in theexcitation of invertebrate Nature 324, 468-470. Kirschfeld, migration 517.

J. E. (1986). Cyclic photoreceptors.

K., and Vogt, K. (1980). Calcium ions in fly photoreceptors. Naturwissenschaften

and

pigment 67, 516-

Lisman, J. E., and Brown, J. E. (1972). The effect of intracellular iontophoretic injection of calcium and sodium ions on the light response of Limulus ventral photoreceptors. J. Cen. Physiol. 59, 701-719. Lo, M. V.-C., and Pak, W. L. (1981). Light-induced migration in the retinular cells of Drosophila Comparison of wild type with ERG-defective Physiol. 77, 155-175. Minke, in the

B. (1982). Light-induced reduction trp mutant of Drosophila. J. Gen.

pigment granule melanogaster. mutants. J. Gen.

in excitation efficiency Physiol. 79, 361-385.

Minke,B.(1983).ThetrpisaDrosophilamutantsensitivetodevelopmental temperature. J. Comp. Physiol. 757, 483-486. Minke, B., and Selinger, Z. (1992a). lnositol lipid pathway in fly photoreceptors, 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. 99124. Minke, B., and Selinger, Z. (1992b). The inositol-lipid necessary for light excitation in fly photoreceptors. Transduction, D. Corey and S. D. Roper, eds. (New feller University Press), in press.

pathway is In Sensory York: Rocke-

Minke, B., and Tsacopoulos, M. (1986). Light induced sodium dependent accumulation of calcium and potassium in the extracellular space of the bee retina. Vision Res. 26, 679-690.

trp: A Light-Sensitive 651

Ca”

Channel

in Drosophila

Minke, B., Wu, C.-F., and Pak, W. L. (1975). Induction of photoreceptor voltage noise in the dark in Drosophila mutant. Nature 258, 84-87. Montell, C., and Rubin, C. M. (1989). Molecular characterization of the Drosophila trp locus: a putative integral membrane tein required for phototransduction. Neuron 2, 1313-1323. Nagy, K. (1991). Biophysical processes ceptors, Recent progress and a critical Ius photoreceptors. Q. Rev. Biophys. Nasi, E. (1991). Two light-dependent domeric photoreceptors. J. Cen.

in invertebrate overview based 24, 165-226.

the honey press.

pro-

photoreon Limu-

conductances in Lima rhabPhysiol. 97, 55-72.

Noda, M., Shimizu, S., Tanabe,T., Takai, T., Kayano, T., Ikeda,T., Takahashi, H., Nakayama, H., Kanaoka, Y., Minamino, N., Kangawa, K., Matsuo, H., Raftery, M. A., Hirose, T., Inayama, S., Hayashida, H., Miyata, T., and Numa, S. (1984). Primary structure of Elecrrophorus electricus sodium channels deduced from cDNA sequence. Nature 312, 121-127. Pak, W. L. (1979). Study of photoreceptor function using Drosophila mutants. In Neurogenetics, Genetic Approaches to the Nervous System, X. Breakfield, ed. (Amsterdam: Elsevier NorthHolland Publishing Co.), pp. 67-99. Pak, W. 1. (1991). The molecular biology of the retina, basic and clinically relevant studies. Molecular genetic studies of photoreceptor function using Drosophila mutants. Prog. Clin. Biol. Res. 362, l-32. Payne, R., Corson, D. W., and Fein, A. (1986). Pressure injection of calcium both excites and adapts Limulus ventral photoreceptors. J. Cen. Physiol. 88, 107-126. Payne, R., Walz, B., Levy, S., and Fein, A. (1988). The localization of calcium release by inositol trisphosphate in Limulus photoreceptors and its control by negative feedback. Phil. Trans. Roy. Sot. (Land.) B 320, 359-379. Phillips, A. M., Bull, A., and Kelly, L. E. (1992). Identification of a Drosophila gene encoding a calmodulin-binding protein with homology to the trp phototransduction gene. Neuron, this issue. Ranganathan, R., Harris, G. L., Stevens, (IVVla). A Drosophila mutant defective dependent photoreceptor deactivation ture 354, 230-232.

C. F., and Zuker, C. S. in extracellular calciumand desensitization. Na-

Ranganathan, R., Harris, W. A., and Zuker, C. S. (IVVlb). The molecular genetics of invertebrate phototransduction. Trends Neurosci. 74, 486-493. Selinger, Z., and Minke, studied in mutant flies. 53, 333-341.

B. (1988). lnositol lipid cascade of vision Cold Spring Harbor Symp. Quant. Biol.

Smith, D. P., Ranganathan, R., Hardy, R. W., Marx, T., and Zuker, C. S. (1991). Photoreceptor deactivation degeneration mediated by a photoreceptor-specific nase C. Science 254, 1478-1484.

J., Tsuchida, and retinal protein ki-

Suss. E., Barash, S., Stavenga, D. G., Stieve, H., Selinger, Z., and Minke, B. (1989). Chemical excitation and inactivation in photoreceptors of the fly mutants trp and nss. J. Gen. Physiol. 94, 465491. Suss-Toby, E., Selinger, Z., and Minke, duces the excitation efficiency in fly Physiol. 98, 848-868.

B. (1991). Lanthanum rephotoreceptors. J. Cen.

Takemura, H., Hughes, A. R., Thastrup, O., and Putney, 1. W., Jr. (1989). Activation of calcium entry by the tumor promoter, thapsigargin, in parotid acinar cells. Evidence that an intracellular pool, and not an inositol phosphate, regulates calcium fluxes at the plasma membrane. J. Biol. Chem. 264, 12266-12271. Toyoshima, N., Matsumoto, P., Wang, P., Inoue, H., Yoshioka, T., Hotta, Y., and Osawa, T. (1990). Purification and partial amino acid sequences of phosphoinositide-specific phospholipase C of Drosophila eye. J. Biol. Chem. 265, 14842-14848. Walz, ulum 248.

B. (1982). Calcium-sequestering in retinula cells of the blowfly.

Walz,

B. (1992).

Enhancement

smooth endoplasmic reticJ. Ultrastruct. Res. 87, 24C-

of sensitivity

in photoreceptors

of

bee drone

by light and

by Cal+. J. Comp.

Physiol.,

in

Werner, U., Suss-Toby, E., Rom,A., and Minke, B. (1992). Calcium is necessary for light excitation in barnacle photoreceptors. J. Comp. Physiol., in press. Wong, F., Schaefer, E. L., Roop, B. C., LaMendola, J. N., JohnsonSeaton, D., and Shao, D. (1989). Proper function of the Drosophila trp gene product during pupal development is important for normal visual transduction in the adult. Neuron 3, 81-94. Yoshioka, T., Inoue, phosphatidyl-inositol sophila visual mutant, them. 97, 1251-1254.

H., and Hotta, Y. (1985). Absence of phosphodiesterase in the head of a DronorpA (no receptor potential A). I. Bio-

Ziegler, A., and Walz, and volume changes drone, Apis mellifera.

B. (1989). Analysis of extracellular calcium in the compound eye of the honeybee J. Comp. Physiol. (A) 765, 697-709.