The trp and trpl genes are thought to encode two classes of light-activated ion channels inDrosophila. A previous report indicated that a nulltrpl mutant does not display any mutant phenotype. This lack of detectable mutant phenotypes made it difficult to suggest functions for the transient receptor potential-like (TRPL) channel in photoreceptor responses. Here, the properties of trplphotoreceptor responses were studied by using electroretinogram (ERG) and intracellular recording techniques in combination with light stimuli of relatively long durations. Distinct mutant phenotypes were detectable under these conditions. These consisted of a reduced sustained component, oscillations superimposed on the response, a poststimulus hyperpolarization, and altered adaptation properties to dim background light. Comparison of photoreceptor responses obtained from wild type, trp, and trpl showed that the responses obtained from the trp andtrpl null mutants did not sum up to that of the wild-type response. To explain the nonlinear summation at the peak of the response, Reuss et al. (1997) proposed that Ca2+ions entering through the TRP channel modulate TRP and TRPL channel activities differentially. However, nonlinear summation was present not only at the peak but throughout the duration of response. Two lines of evidence are presented to suggest that, in addition to the interaction proposed by Reuss et al. (1997), there are other forms of interactions between TRP and TRPL channels, probably involving the channel proteins themselves.
Drosophila photoreceptors respond to light with a depolarization mediated by a phospholipase C (PLC)-dependent signaling pathway (Bloomquist et al., 1988), resulting in the opening of two types of cation channels, transient receptor potential (TRP) and TRP-like (TRPL). The mechanism of activation of either channel is not known. The TRP and TRPL channel subunits are encoded by the transient receptor potential (trp) andtrp-like (trpl) genes, respectively (Montell and Rubin, 1989; Wong et al., 1989; Phillips et al., 1992). The protein products of trp and trpl share 39% amino acid identity and some homology to neuronal voltage-gated channel subunits, although neither TRP nor TRPL is voltage-gated (Stühmer et al., 1989; Phillips et al., 1992). The TRP channel is highly calcium-permeable, whereas the TRPL channel is nonspecifically cation-selective (Hardie and Minke, 1992; Niemeyer et al., 1996; Reuss et al., 1997).
Previous workers have demonstrated that trp mutants exhibit severely impaired photoreceptor responses (Cosens and Manning, 1969;Minke et al., 1975; Pak, 1979; Minke, 1982). However, a nulltrpl mutant,trpl302 , has been reported not to show any mutant phenotype (Niemeyer et al., 1996), raising questions about the role of the TRPL channel in phototransduction. It was suggested that TRP and TRPL might play overlapping roles. Subsequently,Reuss et al. (1997) showed differences in ionic permeabilities between wild type and trpl302 . However, the differences were subtle, and the question still remained whether trpl302 has a clear-cut phenotype. Previous studies of photoreceptor responses oftrpl302 were performed on dissociated photoreceptors (Hardie, 1991), using patch-clamp techniques (Hamill et al., 1981) and short-duration light stimuli. In an attempt to keep the preparations as physiologically intact as possible, we used living flies to perform extracellular (electroretinograms, ERG) and intracellular recordings of photoreceptor responses. Moreover, we used light stimuli of relatively long durations to explore properties of the sustained component of the receptor potential. We show that, under these recording conditions,trpl302 photoreceptor responses do display distinct mutant phenotypes.
TRP and TRPL channel activities appear to influence each other. For example, the similarity in the current amplitudes obtained fromtrpl302 and wild type was explained in terms of differential regulation of the two channels by Ca2+ entering through the TRP channel (Reuss et al., 1997). These authors found that an increase in the internal Ca2+ concentration first facilitated and then suppressed the TRP channel while suppressing the TRPL channel. Thus, during the initial part of the photoreceptor response in wild-type flies the calcium ions coming in through the TRP channels would facilitate the activities of TRP channels but suppress those of TRPL channels. Because the TRP channels are the only major contributors to the early part of the response in both wild type andtrpl302 , their peak response amplitudes would be similar. Such interactions are important because they modulate and shape the responses of photoreceptors. We, therefore, sought to determine what other forms of interactions might be present between TRP and TRPL channels.
MATERIALS AND METHODS
Materials. The wild-type strain used was Oregon R. All flies were marked with the mutations w(white) or cn bw (cinnabar brown) to remove the screening pigments in the eye. Thetrpl302 mutant was generated by Niemeyer et al. (1996). All other mutants were generated on an Oregon R background by ethylmethane sulfonate mutagenesis in this laboratory. The light stimuli, originating from a tungsten halogen lamp (Bausch & Lomb, Rochester, NY), were delivered to the preparation with a fiber optics light guide. The unattenuated intensity at the level of the fly was ∼800 μW/cm2.
Electroretinograms (ERGs). The ERGs were obtained as described in Larrivee et al. (1981), using glass microelectrodes filled with Hoyle's saline. White light stimuli attenuated by two log units were used. Signals were filtered below 100 Hz and sampled at 5 kHz with an analog-to-digital converter (Digidata 1200A), and the data were acquired and analyzed in a computer with Axoscope (Axon Instruments, Foster City, CA).
Intracellular recordings. Intracellular recordings were performed as described in Johnson and Pak (1986) on 1-d-old flies. The electrodes were introduced through a cut that covered small parts of both the cornea and the head. Vacuum grease was applied on the cut to prevent fluid loss. The recording electrodes had 30–50 MΩ resistance when filled with 2 m KCl. Signals were filtered out below 50 Hz and sampled at 2 kHz with an analog-to-digital converter (Digidata 1200A), and the data were acquired and analyzed in a computer with Axoscope (Axon Instruments).
Western blot analysis. Flies were raised at 25°C in a 12 hr light/dark illumination cycle. Heads were collected from six female and six male flies at 1 d after eclosion and homogenized in 60 μl of SDS-PAGE sample buffer containing 50 mmdithiothreitol (DTT). The homogenate was boiled for 5 min and centrifuged (12,000 × g for 3 min), and 10 μl of the supernatant was loaded onto SDS/8% acrylamide gels. To confirm the uniformity of total protein loading in each lane, we stained one of the several identically loaded gels with Coomassie blue. Western blots were generated by a standard protocol. The anti-TRP monoclonal antibody (Pollock et al., 1995) was used at 1:3000 dilution.
Confocal microscopy. After dissection, fly eyes were fixed in 4% formaldehyde (in PBS with 0.3% Triton X-100) for 1 hr and incubated in PBS containing 4% normal goat serum for 2 hr. Filamentous actin of the rhabdomeres was stained with phalloidin-tetramethylrhodamine B isothiocyanate (Sigma, St. Louis, MO). Optical sections of ∼1 μm thickness were viewed by confocal microscopy.
Phenotypes of trpl302
Among the parameters of photoreceptor responses that were examined by using ERG and intracellular recording techniques were (1) the time course of decay of the response during stimulus; (2)V–log I relations, which relate response amplitudes with the log of stimulus intensity, determined at two time points: the peak of the response and 2 sec after the onset of response; (3) oscillations superimposed on the response; (4) poststimulus hyperpolarizations; (5) adaptation to background illuminations of different intensities; (6) speed of recovery from a previous stimulation (refractory period); and (7) response latency.
As reported by others (Niemeyer et al., 1996; Reuss et al., 1997), the peak response amplitudes obtained fromtrpl302 and wild type were similar (Fig. 1 A,B). However, the sustained components of the two responses had very different amplitudes and time courses of decay. During a bright 20 sec light stimulus the responses of both wild type andtrpl302 declined from the peak amplitude to a lower sustained level (Fig. 1 A). However, the trpl302 response declined to a much lower level so that the response remaining 20 sec after the onset was significantly smaller intrpl302 than in wild type. The smaller sustained amplitudes of the trpl responses could be detected even 2 sec after the stimulus onset. As before, the peak amplitudes of the wild-type andtrpl302 responses were similar, but the amplitudes measured at 2 sec after the onset of response were significantly smaller intrpl302 than in wild type (Fig. 1 B). Tables 1 and2 summarize the results of amplitude measurements at the response peak and 2 sec after the response onset obtained at several stimulus intensities. Although the peak amplitudes were similar between wild type andtrpl302 (Table 1), the normalized amplitudes measured at 2 sec after the response onset were all significantly smaller intrpl302 than in wild type (Table 2).
In addition to the smaller sustained component, thetrpl302 response often was accompanied by oscillations superimposed on the response during stimulus and a hyperpolarization immediately after the stimulus termination (Fig. 1 B-b). To determine the frequency of occurrence of these two response phenotypes, we obtained ERGs from 30 trpl302 and 30 wild-type flies. As summarized in Table 3, 80% (24 of 30) and 37% (11 of 30) oftrpl302 flies exhibited oscillations and poststimulus hyperpolarizations, respectively, whereas none of the wild-type flies showed either property. Response oscillations and poststimulus hyperpolarizations also were examined in intracellularly recorded responses, allowing the average amplitude of hyperpolarizations to be determined. Of the 12 cells that were analyzed, 10 showed oscillations (83%) and seven (58%) showed hyperpolarizations, yielding similar frequencies of occurrence of these properties as in ERG recordings. The average amplitude of hyperpolarization in the seven cells was 2.2 ± 0.6 mV when light intensities were attenuated by two log units, i.e., logI/I 0 = −2, whereI and I 0 are as defined in the Figure 1 legend.
Still another characteristic oftrpl302 was its inability to adapt to dim background illuminations (Fig.2). Light adaptation refers to the ability of photoreceptors to adjust their sensitivity in response to changing background illuminations, allowing them to operate over a wide range of illuminations. Figure 2 A compares theV–log I curves obtained from wild type andtrpl302 normalized to the maximal response obtained from each cell, using the brightest stimulus available in the absence of background illumination.V–log I curves obtained from wild-type photoreceptors shifted to increasingly higher intensities and saturated at progressively lower amplitudes as background illuminations increased in intensity (Fig. 2 A). For a test stimulus of given intensity the largest response was obtained with no background illumination, and progressively smaller responses were obtained with background illuminations of increasing intensity (Fig.2 A-a,B-a). Thetrpl302 photoreceptors responded similarly at higher background illuminations, but theV–log I curve obtained at the lowest background that was tested was indistinguishable from that generated in the absence of background illumination [compare open diamonds (−5 BG) with open squares (Dark BG) in Figure2 A-b,B-b], suggesting that trplphotoreceptors failed to adapted to the dim background light.
Two other response properties that were tested were unaltered intrpl302 , as was the case of peak amplitudes that were described previously. They were the refractory period and response latency. The refractory period was defined, in terms of the two-stimulus protocol illustrated in Figure3, as the time required for the response to the second stimulus, R2, to attain nearly the same amplitude as the response to the first stimulus, R1. The refractory periods of wild-type and trpl302 responses were indistinguishable in this protocol. In both cases R2 had an amplitude similar to that of R1 when two strong 2 sec stimuli, S1 and S2, were presented only 2 sec apart. The response latency was defined as the time interval between the beginning of a stimulus and the onset of response. The latency of R2 was shorter than that of R1 in bothtrpl302 and wild-type flies. Moreover, the magnitudes of latency were also very similar (Fig.3 C).
Phenotypes of trp
The phenotypes of trp have been described extensively by previous workers (Cosens and Manning, 1969; Minke et al., 1975; Pak, 1979; Minke, 1982). The present study of trp was performed (1) to serve as controls for some of the experiments to be described and (2) to assess the relative contributions of TRP and TRPL activities to the photoreceptor response. As has been shown by others, a functionally null trp mutant,trpP343 , had no sustained response during a strong prolonged light stimulation (Fig.3 D), a reduced peak amplitude (Fig. 3 A-c,D), a longer-than-normal refractory period (Fig. 3 A-c), and a longer-than-normal response latency (Fig. 3 B-c,C). A near-null trp mutant,trpP301 , gave similar results (see Fig. 5 A, B-b,C-b,D). As shown in Figure 3 D, the peak amplitude oftrpP343 , elicited by a bright stimulus, was ∼ that of the wild type, whereas that of trpl302 was approximately the same as wild type. There was no sustained component in atrpP343 response, and there was a significantly reduced sustained component intrpl302 . The summed response of trpP343 andtrpl302 had a peak amplitude larger than that of wild type and a sustained component smaller than that of wild type (Fig. 3 D). This observation suggested that there might be functional interactions between TRP and TRPL channel activities. We therefore performed the following experiments to see whether there is any other evidence of possible interactions between TRP and TRPL and to obtain clues to the nature of the interactions, if any are present.
TRPL helps to maintain a higher amount of TRP inInaDP215
Western blot analyses performed on 1-d-oldtrpl302 showed that thetrpl302 mutation did not affect the amount of the TRP protein (Fig.4 A). TheInaD gene encodes a PDZ protein to which several proteins involved in phototransduction, including TRP, bind to form a signaling complex (for review, see Montell, 1999; Tsunoda et al., 1999). In theInaDP215 mutant the TRP protein reportedly cannot bind to the signaling complex because of a defect in the TRP-binding site on the INAD protein (Shieh and Zhu, 1996; Chevesich et al., 1997; Tsunoda et al., 1997). Tsunoda et al. (1997) have reported that the amount of TRP was normal in newly eclosedInaDP215 flies but decreased to an undetectable level by 10 d after eclosion. Consistent with this report, the amount of TRP in 1-d-oldInaDP215 fly heads was ∼70% of the wild-type amount (Fig. 4 A). If the TRPL protein were removed fromInaDP215 (i.e., in thetrpl302InaDP215 double mutant), however, the amount of TRP was reduced to <10% of the wild-type amount. To see whether degeneration of the rhabdomeres could be the cause of this large reduction in the amount of TRP in thetrpl302InaDP215 double mutant, we performed confocal microscopy on the 1-d-old double mutant. The results (Fig. 4 B) showed that the rhabdomeres of the double mutant were intact at this age, excluding degeneration of the rhabdomeres as a primary cause of the reduction in TRP in very youngtrpl302InaDP215 . The above data, thus, suggested that the amount of TRP in 1-d-oldInaDP215 fly heads depended strongly on the presence of TRPL.
TRP that is not bound to INAD contributes significantly to photoreceptor response in the presence, but not in the absence, of TRPL
To explore how the binding of TRP to the signaling complex through INAD might affect the photoreceptor responses, we compared the responses of InaDP215 and trpP301. trpP301 is a near-null mutant that has no immunodetectable TRP protein but has small yet detectable physiological activities of TRP when examined in the double mutanttrpl302;trpP301 . Thus, 1-d-old InaDP215 heads contained ∼70% of the wild-type amount of TRP (Fig.4 A), which presumably was not bound to the signaling complex, whereas trpP301 heads contained a small, undetectable amount of TRP that, however, was bound properly to the complex. If the TRP protein that was not bound to the signaling complex were nonfunctional, then one might expect the phenotype of InaDP215 to be similar to that oftrpP301 . The comparison between the photoreceptor responses ofInaDP215 andtrpP301 showed otherwise. The responses ofInaDP215 were more wild-type-like (less severe) in the peak amplitude (Fig.5 A), amount of decay during light illumination (Fig. 5 A), refractory period (Fig.5 B), and response latency (Fig. 5 C,D) than those of trpP301 . The results suggested that either some TRP remained bound to the signaling complex in InaDP215 mutants or the TRP channels that were not bound to the signaling complex contributed significantly to photoreceptor responses.
A very different picture emerged if the TRPL protein were removed from the above two mutants via the construction of the double mutantstrpl302InaDP215 andtrpl302;trpP301 . As may be seen in Figure6 A, the responses of bothtrpl302InaDP215 andtrpl302;trpP301 were small and transient, but the response amplitude oftrpl302InaD P215was significantly smaller than that oftrpl302;trp P301. The comparison of such other properties as the refractory period (Fig.6 B) and response latency (Fig. 6 C,D) further showed that these phenotypes oftrpl302InaD P215were also more severe than those of thetrpl302;trp P301. Thus, in the presence of TRPL, TRP that was not bound to INAD contributed significantly to photoreceptor responses (InaD P215 vstrp P301), but in the absence of TRPL, TRP that was not bound to INAD was unable to contribute significantly to responses (trpl302;trp P301vstrpl302InaD P215). Although the amount of detectable TRP intrpl302InaD P215was much lower than in wild type (see Fig. 4 A), this reduced TRP content could not account for its phenotypes, because the phenotypes were much more severe than those oftrpl302;trp P301, which had no detectable amount of TRP. The fact that the ability of non-INAD-bound TRP to contribute to photoreceptor responses depends on the presence of TRPL suggests that there may be interactions between TRPL and the non-INAD-bound TRP protein.
We have shown that thetrpl302 mutants have clearly detectable mutant phenotypes. The absence of phenotype reported by previous workers may be attributable to differences in recording procedures. Previous studies of thetrpl302 mutants were done by using the whole-cell patch-clamp technique on dissociated photoreceptors, whereas the present studies were performed by using the ERG and intracellular recording techniques on photoreceptors of intact, living flies. Perhaps an even more important difference might be the duration of stimuli used to elicit photoreceptor responses. Previous workers used light pulses of millisecond durations, whereas we used stimulus durations of seconds to tens of seconds. Properties of photoreceptor responses that are readily detectable with short-duration light flashes, such as the peak amplitude and response latency, indeed are essentially indistinguishable betweentrpl302 and wild type. However, a number of readily detectable phenotypes become apparent if longer stimulus durations are used and recordings are made from living flies. These include oscillations, poststimulus hyperpolarizations, reduced sustained components, and altered adaptation properties. These results suggest that the TRPL channels may have a role in sustaining the photoreceptor response during prolonged illuminations and in adaptation to dim light stimuli, because eliminating the TRPL channels (i.e., in trpl302 ) substantially reduces the sustained component (see Fig.1 A) and makes the photoreceptors insensitive to dim background illuminations (see Fig. 2). However, complicating the interpretation regarding its role in sustaining the response is the observation that, in the absence of TRP channels (i.e., intrpP343 ), the TRPL channels essentially make no contributions to the sustained component (see Fig. 1 A). It thus may be that both classes of channels contribute to the sustained component and that they both must be present to allow mutual interactions so that normal channel outputs might be generated.
The summed amplitude oftrpP343 andtrpl302 mutant receptor potentials is larger at the peak but smaller throughout the sustained part of the response than the wild-type response amplitude (see Fig.3 D). As discussed previously, Reuss et al. (1997) proposed that the nonlinear summation at the peak of the response arises because Ca2+ ions entering through the TRP channel suppress the TRPL channel activity. Thus, in this form of TRP–TRPL interaction the current entering through one class of channel influences the activities of both the same and the other class of channels. For the purpose of present discussion, this form of interactions may be referred to as “functional interactions.” However, nonlinear summation of the two channel activities is present not just at the peak but throughout the response. Moreover, the difference between the summed mutant amplitude and the wild-type amplitude is positive at the peak of the response but reverses in sign to be negative throughout the sustained component (see Fig.3 D). It is unlikely that a single mechanism, e.g., that proposed by Reuss et al. (1997), is responsible for the observed nonlinear summation of mutant responses throughout their entire time course.
The idea that other forms of interactions may be present in addition to “functional interactions” is supported by the results obtained withInaDP215 andtrpl302InaDP215 mutants. The amount of TRP in 1-d-oldInaDP215 mutants is substantially larger than that in thetrpl302InaDP215 double mutants of the same age (see Fig. 4 A). Moreover, the photoreceptor potential recorded fromInaDP215 is large and robust (see Fig. 5 A), whereas that recorded from the double mutanttrpl302InaDP215 is very small—even smaller than that oftrpl302;trp P301(see Fig. 6 A), which has no immunodetectable TRP protein. Several groups have shown previously, via in vitrobinding assays, that the TRP protein is unable to bind to the INAD protein in InaDP215 mutants (Shieh and Zhu, 1996; Chevesich et al., 1997; Tsunoda et al., 1997). The above results thus suggest that the presence of the TRPL protein somehow stabilizes the TRP protein even when the latter is not bound to the INAD protein and allows TRP to contribute substantially to the photoreceptor response. If one supposes that the TRPL protein is associated with the signaling complex, as suggested by Xu et al. (1998), a possible interpretation of these results is that a significant amount of TRP is associated, however weakly, with the signaling complex via interaction with the TRPL protein, thus contributing to the photoreceptor response, even when TRP cannot bind directly to the signaling complex itself. In the absence of TRPL, however, the association is lost and much of the response is lost also.
The idea that TRP and TRPL proteins interact need not necessarily be wedded to the above specific model. For example, the extent to which the TRP protein fails to bind to INAD in the intact photoreceptors of the InaD P215 mutant is unclear because most of the binding studies were performed in vitro. Regardless of the extent of TRP binding to INAD in theInaD P215 mutant, however, the results of this study suggest that the presence of TRPL stabilizes the TRP channel and allows the latter to contribute to the photoreceptor response inInaD P215, implying physical interactions between these two proteins.
The molecular nature of interactions between TRP and TRPL proteins cannot be specified from the results of the present study. However, it is pertinent to note that previous workers showed interaction between TRP and TRPL in immunoprecipitation assays (Xu et al., 1997) and presented evidence for the formation of heteromultimeric channels when both TRP and TRPL are coexpressed in oocyte (Gillo et al., 1996) or cell culture (Xu et al., 1997) systems. Whatever the actual mechanism of interaction responsible for our results, the results suggest that, in addition to “functional interactions,” direct “physical interactions” between TRP and TRPL also may be a part of the photoreceptor process.
This work was supported by a grant from the National Eye Institute, National Institutes of Health (W.L.P.). The departmental shared confocal microscope facility was supported by National Science Foundation Grant BIR-9512962. We thank Dr. John A. Pollock of Carnegie Mellon University for providing anti-TRP antibodies and Dr. Charles Zuker of University California San Diego for making thetrpl 302 mutant available.
Correspondence should be addressed to Dr. William L. Pak, Purdue University, Department of Biological Sciences, 1392 Lilly Hall, West Lafayette, IN 47907-1392. E-mail:.