mGluR6 deletion renders the TRPM1 channel in retina inactive

Abstract

In darkness, glutamate released from photoreceptors activates the metabotropic glutamate receptor 6 (mGluR6) on retinal ON bipolar cells. This activates the G protein G_o, which then closes transient receptor potential melastatin 1 (TRPM1) channels, leading to cells' hyperpolarization. It has been generally assumed that deleting mGluR6 would render the cascade inactive and the ON bipolar cells constitutively depolarized. Here we show that the rod bipolar cells in mGluR6-null mice were hyperpolarized. The slope conductance of the current-voltage curves and the current noise were smaller than in wild type. Furthermore, while in wild-type rod bipolar cells, TRPM1 could be activated by local application of capsaicin; in null cells, it did not. These results suggest that the TRPM1 channel in mGluR6-null rod bipolar cells is inactive. To explore the reason for this lack of activity, we tested if mGluR6 deletion affected expression of cascade components. Immunostaining for G protein subunit candidates Gα_o, Gβ₃, and Gγ₁₃ showed no significant changes in their expression or distribution. Immunostaining for TRPM1 in the dendritic tips was greatly reduced, but the channel was still present in the soma and primary dendrites of mGluR6-null bipolar cells, where a certain fraction of TRPM1 appears to localize to the plasma membrane. Consequently, the lack of TRPM1 activity in the null retina is unlikely to be due to failure of the channels to localize to the plasma membrane. We speculate that, to be constitutively active, TRPM1 channels in ON bipolar cells have to be in a complex, or perhaps require an unidentified factor.

in retina, an increment of light intensity hyperpolarizes the photoreceptor and sets off two opposing signals in the OFF and ON bipolar cells. In darkness, glutamate released by photoreceptors depolarizes OFF bipolar cells and hyperpolarizes ON bipolar cells. The key steps in this ON bipolar cell “sign inverting” cascade are as follows: glutamate activates the metabotropic glutamate receptor 6 (mGluR6) (Akazawa et al. 1994; Kikkawa et al. 1993; Masu et al. 1995; Vardi and Morigiwa 1997), this activates the G protein G_o (Dhingra et al. 2002; Dhingra et al. 2000; Nawy 1999; Okawa et al. 2010; Vardi et al. 1993), and this in turn leads to closure of a presumably constitutively active nonselective cation channel, recently identified as transient receptor potential melastatin 1 (TRPM1) (Koike et al. 2010b; Morgans et al. 2009). Mutations of this channel in humans lead to night blindness (Audo et al. 2009; Li et al. 2009; Nakamura et al. 2010; van Genderen et al. 2009).

Several studies investigated the effect of mGluR6 deletion on the visual path, animal behavior, and retinal development. The general structure of the retina and the ON bipolar cells appears normal (Masu et al. 1995; Tagawa et al. 1999), although ectopic ribbons and mislocalization of mGluR7 have been seen in ON cone bipolar cells (Ishii et al. 2009; Tsukamoto et al. 2007). The visual ON pathway, as revealed by electroretinogram, certain behavioral tasks (Takao et al. 2000), pupillary responses, and optokinetic nystagmus (Iwakabe et al. 1997), is impaired. However, a late ON response is still observed in superior colliculus (Sugihara et al. 1997) and ganglion cells (Renteria et al. 2006), a feature that may explain why both mGluR6-null and wild-type (WT) mice respond similarly to light increments in an avoidance test (Masu et al. 1995). Although direct recordings from ON bipolar cells have not been performed, it has been generally assumed that deleting mGluR6 would render the cascade inactive and the ON bipolar cells constitutively depolarized (Ishii et al. 2009; O'Connor et al. 2006; Tagawa et al. 1999). Whether or not this actually occurs, is important when analyzing the effect that blocking the ON response will have on developmental and signaling processes and also in night blindness associated with mutations in mGluR6.

In this study, we measured the resting membrane potential of ON bipolar cells in mGluR6-null mice and found that the rod bipolar cells were constitutively hyperpolarized. We further found that mGluR6 deletion reduced TRPM1 channel expression at the ON bipolar's dendritic tips, but it did not affect its overall expression on the cell surface. We speculate that the activity of the TRPM1 channel depends on it being part of a complex.

MATERIALS AND METHODS

C57BL/6J WT mice came from Charles River Laboratories; the mGluR6-null mouse was a gift from Drs. Nakanishi and Copenhagen and was described in Masu et al. (1995), and the mGluR6/PCP2 double knockout mouse was generated by crossing the PCP2-null mouse (Xu et al. 2008) with the mGluR6-null mouse. Mice were deeply anesthetized by intraperitoneal injection of a mixture of 85 μg/g ketamine and 13 μg/g xylazine; the eyes were enucleated; and the mouse was killed by anesthetic overdose. Mice were treated in compliance with Federal regulations and University of Pennsylvania policy; protocol 803174 was approved by University IACUC. For Western blotting, retinas were detached quickly and frozen in liquid nitrogen.

Whole cell recording.

Retinal slices were prepared as described previously (Xu et al. 2008). Briefly, retinas were isolated and cut into 200-μm-thick slices with a tissue slicer (Narishige). The slices were transferred to a recording chamber, secured with vacuum grease, and then moved to the microscope stage of an Olympus microscope equipped with a ×60 water immersion objective. The chamber was perfused at a rate of 0.5–1 ml/min with oxygenated (95% O₂, 5% CO₂) Ames medium (Sigma, St. Louis, MO) containing sodium bicarbonate (1.9 g/l) at 32–34°C.

Patch pipettes with resistances of 5–7 MΩ were fabricated from borosilicate glass using an electrode puller (Sutter, Novato, CA). Pipettes were filled with the following solutions (in mM): 110 potassium-gluconate, 10 EGTA, 10 HEPES, 10 KCl, 5 NaCl, 4 MgATP, and 1 LiGTP. pH was adjusted to 7.4 with KOH, and osmolality was 290 mosmol/kgH₂O. The solution was aliquoted, stored at −20°C, and defrosted before each experiment. In experiments testing the activity of the TRPM1 channel, transient receptor potential vanilloid 1 (TRPV1) channel agonist capsaicin was delivered to the bipolar soma with a puffing pipette that used positive pressure (2–4 psi) via a computer-controlled solenoid valve (Picrosprizer; General Valve). To construct a current-voltage (I-V) curve, we increased the holding voltage from −90 mV to +60 mV in 30-mV increments. All chemicals were obtained from Sigma.

Whole cell recordings were obtained with an Axopatch 1D amplifier (Molecular Devices). Membrane potentials were corrected for the liquid junction potential that was calculated to be ∼15 mV. Cells were discarded if the baseline current exceeded −100 pA at a holding potential of −60 mV. Voltage command generation and data acquisition were accomplished with Clampex (Molecular Devices). Cells were either current clamped with zero current injection to measure the resting membrane potential, or voltage clamped to measure the I–V curve and current noise.

Light stimulus and data analysis.

The retina was stimulated using a green full-field light generated by a light-emitting diode with a peak wavelength of 565 nm. The light illuminance ranged from 1.7 to 2.9 log photons/μm² as described before (Xu et al. 2008). For each cell, a sequence of increasing light intensities or puffing durations was repeated three times, and the recordings were averaged using Igor (WaveMetrics). The averaged responses are displayed in the results as recording traces. Waveform analysis of the response was done offline with Clampfit (Molecular Devices).

Resting membrane potential was determined by averaging the voltage values during the first 1-s period of recorded spontaneous activity. Current noise was defined as the standard deviation of the spontaneous activity during the first second of recording while voltage clamping the cell at −60 mV. Results of WT and null mice were compared with Student's t-test. Differences were considered significant when P ≤ 0.05. All data are reported as means ± SE (standard error of the mean).

Western blotting.

Retinas were homogenized using a polytron homogenizer in a lysis buffer containing 5 mM Tris·HCl (pH 7.5), 2 mM EDTA, and 350 mM sucrose. Homogenate was centrifuged at 6,000 g for 10 min, and supernatant was collected. Total membrane protein was prepared by centrifuging the homogenate at 20,000 g for 30 min at 4°C. To check the expression of TRPM1 in the plasma membrane fraction, Plasma Membrane Protein Extraction Kit (Abcam, Cambridge, MA) was used as per the manufacturer's protocol. Briefly, the supernatant obtained from a 10,000 g spin was resuspended in the Upper Phase Solution and extracted using the Lower Phase Solution (both solutions from the manufacturer). This was followed by centrifugation to pellet the plasma membrane. Protein assay was carried out using BSA protein reagent (Bio-Rad, Hercules, CA). The proteins were run on 7.5%, 10%, or 4–15% SDS-PAGE gel and transferred to a nitrocellulose membrane using semiwet transfer apparatus (Bio-Rad). After a brief rinse in PBS, the blots were incubated sequentially in the following: Odyssey blocking buffer diluted with PBS (1:1, blocking buffer) at room temperature for 1 h; primary antibody diluted in the blocking buffer containing 0.1% Tween 20 at 4°C overnight; a wash in PBS + 0.1% Tween 20 (PBST); IRDye-conjugated secondary antibodies (from LI-COR Biosciences or Rockland, 1:15,000 dilution in the blocking buffer containing 0.1% Tween 20) for 1 h at room temperature; a wash in PBST; and a final rinse in PBS. The blots were then scanned using the Odyssey Infrared Imaging system (LI-COR Biosciences), according to manufacturer's instructions. To control for equal loading and to check for plasma membrane markers, we incubated blots simultaneously with antibodies against the test protein and against a marker protein (e.g., β-actin or Na-K-ATPase).

Immunocytochemistry.

Eyes were enucleated from an anesthetized mouse, and a small cut was made through the lens. The eyeball was immersion fixed in 4% paraformaldehyde for 10 min or 1 h; rinsed in phosphate buffer; soaked overnight in 30% buffered sucrose; and embedded in a mixture of two parts 20% sucrose in phosphate buffer and one part tissue freezing medium. The eye was cryosectioned radially at a 10- to 15-μm thickness. Sections were soaked in diluent containing 10% normal goat serum, 5% sucrose, and 0.5% Triton X-100 in phosphate buffer. They were then incubated in primary antibodies at 4°C overnight; washed; incubated for 3 h in secondary antibodies conjugated to a fluorescent marker; rinsed; and mounted in Vectashield (Vector Laboratories, Burlingame, CA).

Cell dissociation.

The eye was hemisected, the anterior segments and the vitreous body were removed, and the retina was detached from the pigment epithelium. The retina was incubated for 5 min in warm buffer containing the following (in mM): 135 NaCl, 5 KCl, 1 MgCl₂, 2 CaCl₂, 10 HEPES, and 10 glucose (pH 7, 4), plus 3 U/ml papain (Sigma, P4762) and 0.15 mg/ml l-cysteine (Sigma, C7352) at 37°C. Retinal fragments were rinsed three times with the buffer containing 1 mg/ml bovine serum albumin (Sigma A7906) and then three times with buffer alone. The retina was then triturated using gradually smaller pipette tips until a cloudy solution was achieved. Drops of the cell suspension were dispersed onto a slide coated with 1 mg/ml concanavalin A (Sigma C2010). The cells were allowed to settle for 1 h and then were rinsed three times with buffer. The settled cells were fixed with 4% paraformaldehyde for 30 min and then processed for immunocytochemistry.

Imaging and quantification.

Sections or dissociated cells were photographed with an Olympus FV-1000 confocal microscope under ×40 or ×60 oil immersion objective. To compare the intensity of immunostaining between WT and null retina, we simultaneously fixed and processed a WT eye and one or two age-matched mGluR6-null eyes using exactly the same conditions and imaging parameters. Retinas that were processed in parallel were considered a set. Intensity measurements were taken from z-stacks (1 μm apart; same number of sections for WT and null mice) using Volocity software (Improvision, Coventry, UK). Regions of interest were drawn around the outer plexiform layer (OPL), the outer region of the inner nuclear layer (INL), the inner region of the INL, along with the OFF sublamina of the inner plexiform layer (IPL-OFF) and the ON sublamina (IPL-ON) (Fig. 1A). Mean background intensity level (measured from the outer nuclear layer) was subtracted from mean intensities of the regions of interest. The ratio of the mean intensities between paired WT and null mice was calculated for each region and averaged across three to five pairs. A large number of the same data set was analyzed independently by two authors; the second author used ImageJ (National Institutes of Health software). Paired t-test was used to compare the intensities between WT and null mice. Results obtained by the two authors were similar (3–10% difference with same significance), but here we report only those obtained by Volocity because this program analyzed the data in the given volume (rather than collapsing the confocal planes), and it retained the depth bit information.

Fig. 1.

Quantification methods. A: a representative confocal image of transient receptor potential melastatin 1 (TRPM1) staining shows the regions of interest used to find the average intensities for the different layers. For this and the following figures: ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer. The region of interest in the INL includes only the top tiers where ON bipolar cells reside, and the region of interest in the IPL is divided into IPL-OFF (which also includes the lower tiers of the INL) and IPL-ON. B and C: a representative confocal image of TRPM1 staining shows the region of interest for counting puncta in the OPL (B), and the puncta selected by Volocity (C; magenta rectangles). D: a dissociated ON bipolar cell immunostained for Na⁺-K⁺-ATPase (cyan; outlines the plasma membrane) and for TRPM1 (red). The cell was also stained with 4,6-diamidino-2-phenylindole (DAPI; blue) to locate the nucleus. Five lines perpendicular to the plasma membrane were drawn, and the intensity profiles for two of them are shown in E. E: intensity profiles (along lines 8 and 9 shown in D) plot the intensity of each pixel proceeding from the extracellular to the intracellular space. The profiles are aligned by the peak of the Na⁺-K⁺-ATPase stain (vertical line). The dotted boxed region in each profile overlaps with the nucleus and, therefore, shows the background level of TRPM1 staining in the cell.

Count of TRPM1 puncta in the OPL was performed with the aid of Volocity software. A region of interest that encompassed the OPL throughout the z-stack was drawn, and only puncta within this region were counted. Care was taken to have equal numbers of optical sections in the WT and null mice. The Volocity program counted every puncta that had a volume of 0.1–7 μm³ and was more intense than a predefined threshold value (Fig. 1, B and C). This volume range was chosen by considering the dimensions of the rod bipolar dendrite. The minimum volume assumes that the stained dendrite under a rod spherule should be a cylinder ∼0.5 μm in diameter and length. The maximum volume considers that, due to blurring, intense staining appears larger than the object's actual size. The value was chosen to exclude primary dendrites and the long rows of puncta under cone terminals. For the threshold intensities, we tried several intensity values and chose those that labeled the puncta judged to be dendritic tips by eye (intensity value of 841 for one set and 700 for the second). Most values tended to overcount puncta because the program often fragmented primary dendrites into several puncta. However, although this counting method may not provide the actual number of puncta in a certain volume, it does provide a relative comparison between the WT and null mice.

To determine whether TRPM1 in dissociated cells localizes to the plasma membrane of the soma, we adopted the following approach. We selected a round cell stained for Na⁺-K⁺-ATPase and TRPM1; turned off the channel that showed the TRPM1-staining and selected a single confocal image in which the membrane delineation by Na⁺-K⁺-ATPase staining appeared sharp; and drew about three to eight lines across the membrane and had FV1000 return intensity profiles for each line (Fig. 1, D and E). We then assigned each profile to one of four categories that describe qualitatively the relationships between the location of TRPM1 and the membrane stain (see results). To determine the average intensities across all lines, we used a program in Matlab to align profiles according to the peak of the membrane staining, and we then averaged the corresponding pixels (Fig. 1E).

Antibodies used.

The following antibodies were used: rabbit anti-Gα_o, 1:100 (MAB 3073, Millipore, Billerica, MA); rabbit anti-Gβ₃, 1:300 (HPA005645, Sigma, St. Louis, MO); rabbit anti-Gγ₁₃, 1:500 (gift of Dr. Robert Margolskee); rabbit anti-TRPM1, 1:100 (prepared by Dr. Furukawa's group, for immunocytochemistry); sheep anti-TRPM1, 1:1,1000 (Gift of Dr. Kirill Martemayanov, for Western blots); rabbit anti-Gβ₅, 1:500 (gift of Theodore Wensel); rabbit anti-RGS11, 1:1000 (gift of Theodore Wensel); mouse anti-Na⁺-K⁺-ATPase, 1:100 (MA3–915, Thermo Scientific, Rockford, IL); and rabbit anti-ERGIC-53/58 (p58), 1:1000 (Sigma-Aldrich).

RESULTS

Rod bipolar cells in mGluR6-null mice are hyperpolarized.

We conducted whole cell patch recordings from rod bipolar cells in slices of dark-adapted mouse retina. As expected, the cells of WT mice depolarized to light stimuli, while those of mGluR6-null mice failed to respond to light (Fig. 2A). However, contrary to expectations, the dark resting membrane potential of the null cells was significantly more hyperpolarized. While in WT rod bipolar cells, the potential averaged at −46 ± 3 mV (41 cells, 6 mice); in mGluR6-null cells, it averaged at −62 ± 2 mV (57 cells, 3 mice; P < 0.001), about a 16-mV difference. In accordance, the average holding current in the WT rod bipolar cell was −22.5 pA when the cells were clamped at −60 mV, and that in the null cell was 0.21 pA.

Fig. 2.

Rod bipolar cells in metabotropic glutamate receptor 6 (mGluR6)-null mice are hyperpolarized. A: representative voltage responses of rod bipolar cells from wild type (WT; dark line) and mGluR6-null [knockout (KO), gray line] retinas in current clamp mode with no current injection. Dotted vertical line represents the time of stimulus presentation, a 10-ms saturating light of 2.7 log photons/μm². B: averaged current-voltage (I-V) curves show that the slope conductance of null rod bipolar cells (gray, n = 52) is smaller than that of WT (black, n = 19). The voltage at which the graph crosses the x-axis represents the dark resting membrane potential. Error bars were calculated at every 20-mV increment.

The extra hyperpolarization of mGluR6-null rod bipolar cells may be due to either the closure of TRPM1 channels or the opening of potassium channels. To distinguish between these two possibilities, we measured the slope conductance of the I-V curves, as well as the current noise at −60 mV holding potential in the dark. Lower slope conductance in mGluR6-null rod bipolar cells would indicate the closure of TRPM1 channels, whereas higher values would indicate the opening of potassium channels. In the linear range of the I-V curve between −75 and −35 mV, the slope conductance for WT cells was 1.84 ± 0.26 nS (n = 19), and for null cells was significantly lower at 1.09 ± 0.08 nS (n = 51; P < 0.001) (Fig. 2B). Moreover, the current channel noise for WT cells was 7.0 ± 0.6 pA (n = 41), and for null cells was significantly lower at 2.6 ± 0.2 pA (n = 27, P < 0.001). Thus our data suggest that the greater hyperpolarization of mGluR6-null rod bipolar cells can be attributed to the closure of TRPM1 channels.

TRPM1 channels are inactive in mGluR6-null rod bipolar cells.

We next investigated whether mGluR6-null rod bipolar cells are totally devoid of channel activity or still exhibit residual activity. If activity persists, we should be able to open the channels with capsaicin, the prototypical TRPV1 agonist that elicits an inward current in rod bipolar cells (Morgans et al. 2010; Shen et al. 2009). In nearly one-half of the WT rod bipolar cells we tested (6 out of 15), local application of 20 μM capsaicin elicited an inward current (Fig. 3A, black line), and this response reversed direction when the cell was depolarized (Fig. 3B). In contrast, none of the 32 mGluR6-null rod bipolar cells we tested gave a response to capsaicin (Fig. 3A, gray line). Thus it is most likely that the TRPM1 channel in mGluR6-null mice is inactive.

Fig. 3.

TRPM1 channels are inactive in mGluR6-null rod bipolar cells. A: sample current responses of rod bipolar cells from WT and mGluR6-null (KO) mice to a puff of 20 μM capsaicin [holding voltage (V_hold) = −60 mV]. B: I–V relationship of the WT cell shown in A; the response to capsaicin reverses direction at around 0 mV.

In a different experiment, we addressed the question of residual activity by testing PCP2-mGluR6 double knockout mice. We previously reported that the PCP2-null rod bipolar cell shows a more depolarized membrane potential than the WT cell due to an increase in activity of the nonselective cation channel when PCP2 is absent (Xu et al. 2008). We reason that if there is residual TRPM1 activity in the mGluR6-null rod bipolar cell, the resting membrane potential of the mGluR6/PCP2 double-knockout cell should be more depolarized than that of the mGluR6-null cell. We found that the resting potential of the double knockout rod bipolar cell was −64 ± 2 mV (n = 34, 3 mice), almost identical to that of the mGluR6-null cell (−62 mV). Thus there is probably no residual TRPM1 channel activity.

The expression patterns of the G protein subunit candidates appear normal in mGluR6-null mice.

To investigate the reason for TRPM1 channel inactivity in the mGluR6-null rod bipolar cell, we tested if mGluR6 deletion affects the expression of mGluR6 cascade elements. We first tested expression of the G protein subunit Gα_o because this subunit is known to be required for the ON response (Dhingra et al. 2000, 2002). Western blotting of null and WT retinal proteins displayed similar bands at 37 kDa, with a null-to-WT intensity ratio of 1.24 (Fig. 4, A and G). Note that Gα_o staining in the IPL does not necessarily reflect staining of the ON bipolar cells (Vardi 1998), so Western blotting may not accurately reflect Gα_o expression in ON bipolar cells. Thus we used immunostaining. In the null mouse, immunostaining confirmed lack of mGluR6-stained puncta in the OPL (Fig. 4C) and showed normal Gα_o staining of the ON bipolar cells (Fig. 4D). In these mice, quantification of staining intensity in the OPL and INL showed some elevation of staining intensity compared with WT; however, the difference was not statistically significant (Fig. 4G).

Fig. 4.

The expression of G protein subunit candidates Gα_o, Gβ₃, and Gγ₁₃ appears normal in mGluR6-null mice. A and B: Western blots show that the levels of Gα_o (A) and Gβ₃ (B) are the same in WT and mGluR6-null (KO) retinas. Actin is used to control for loading. C: mGluR6 is expressed in the OPL of WT retina, but is absent in mGluR6-null retina. For C–F, fixation and immunostaining of WT and mGluR6-null retinas were performed in parallel, and images were taken with the same settings. D, E, and F: immunostaining for Gα_o (D), Gβ₃ (E), and Gγ₁₃ (F) shows similar expression patterns in WT and mGluR6-null ON bipolar cells. G: average intensity ratios (for immunostaining or Western band intensity) of null to WT retinas for Gα_o, Gβ₃, and Gγ₁₃ in OPL, INL, ON layer in the IPL (IPL-ON), and the sum of these layers + the OFF sublamina (all layers).

Next we tested for Gβ₃, the Gβ subunit that is the best candidate because it localizes to the ON bipolar cell (Ritchey et al. 2010) and because its deletion, as is the case with Gα_o deletion, eliminates the electroretinogram b-wave (Dhingra et al. 2011). Evaluating protein level by Western blots showed that band intensities were similar, and the null-to-WT ratio was 1.06 (Fig. 4, B and G); however, these measurements included expression in cones (Peng et al. 1992). Thus we evaluated the protein levels by immunocytochemistry. We found that, consistent with Ritchey et al. (2010) and as shown by us (Dhingra et al. 2011), WT retina exhibited Gβ₃ staining in cone outer segments (not shown) and in ON bipolar cells. mGluR6-null retina displayed similar staining with somewhat higher intensity (P < 0.05, Fig. 4, E and G).

Last, we tested the expression of the G protein subunit Gγ₁₃. In retina, this subunit has been seen exclusively in ON bipolar cells, thus making it a good candidate for mediating the cascade (Huang et al. 2003). Since attempts to obtain this small molecule band in Western blots failed, we limited our analysis to immunocytochemistry. As has been previously reported, in WT retina, Gγ₁₃ antibodies selectively labeled ON-bipolar cells throughout the cells, including their dendrites, somas, and axon terminals that stratified in the inner half of the IPL (the ON sublamina). In the mGluR6-null retina, the staining pattern was identical with similar staining intensities in all layers (Fig. 4, F and G). We conclude that, although mGluR6 deletion may lead to slight upregulation of Gα_o and Gβ₃ subunits, this cannot explain the lack of channel activity.

TRPM1 channels are diminished at the dendritic tips of mGluR6-null ON bipolar cells.

To determine whether the lack of TRPM1 channel activity in mGluR6-null cells is due to downregulation of the TRPM1 channel itself, we immunostained for the channel. In WT retina, we found the expected staining pattern (Fig. 5A, left). The OPL was decorated with bright puncta corresponding to rod bipolar dendritic tips and with short rows composed of smaller puncta corresponding to cone bipolar dendritic tips (Koike et al. 2010b) (Fig. 5A, left, inset); the INL displayed strong staining of ON bipolar cell somas; and the IPL showed occasional faint staining of their axon terminals in the ON sublamina. In the mGluR6-null retina, the TRPM1 staining of somas and axon terminals was still present, but the brightly stained puncta in the OPL were greatly diminished (Fig. 5A, right, inset). Comparing relative intensity levels across the different layers showed that the intensity in the OPL of the null retina was significantly lower than that of the WT by ∼30%, but the intensity in the ON sublamina of the IPL was slightly higher (Fig. 5D). Counting the number of puncta in the OPL showed that this number for null retina was only 40% of the number for WT (an average of 430 puncta in 15,000 μm³ for null retina vs. 1,067 for WT; 3 mGluR6-null and 2 WT mice). Moreover, in mGluR6-null retina, these counted puncta were dimmer and smaller (Fig. 5, E and F). The total intensity in all layers was slightly lower in the null mouse (85% of WT by immunostaining and 77% by Western blots; Fig. 5D). These findings indicate that, although TRPM1 may be slightly downregulated in mGluR6-null retina, the main effect of eliminating mGluR6 is the change in TRPM1 localization in the dendritic tips of the ON bipolar cells.

Fig. 5.

The expression patterns of TRPM1, Gβ₅, and regulator of G protein signaling (RGS) 11 are abnormal in mGluR6-null mice. A: TRPM1 staining in sections of WT and mGluR6-null (KO) retinas. The brightly stained puncta that decorate the dendritic tips of ON bipolar cells disappear in null retina. Insets show magnified areas of the OPL. Cyan brackets indicate the row of puncta under a cone pedicle. For A–C, fixation and immunostaining of WT and mGluR6-null retinas were performed in parallel, and images were taken with the same settings (scale bars = 10 μm). B and C: immunostaining for Gβ₅ (B) and RGS11 (C) shows highly diminished expression at the dendritic tips of null ON bipolar cells. D: immunostaining intensity ratios of null to WT retina for TRPM1, Gβ₅, and RGS11 in OPL, INL, IPL-ON, and the sum of these layers + the OFF sublamina (all layers). Each average comes from 3–5 images. The ratio of TRPM1 band intensities (normalized to actin) of KO/WT in Western blots is also shown. E: intensity distribution of puncta that meet the criteria for the automated count by Volocity. The average puncta intensity for null retina (dotted line) was lower than that for WT (solid line) (data are from one set). F: volume distribution of the same puncta as in E. Puncta in the null retina were smaller than those in WT.

Gβ₅ and RGS11 are diminished at the dendritic tips of mGluR6-null ON bipolar cells.

Previous reports have indicated that four of the cascade's modulators, Gβ₅, regulator of G protein signaling (RGS) 7, RGS11, and R9AP, are either downregulated or redistributed in the nob4 mouse, an mGluR6-deficient model (Cao et al. 2009; Morgans et al. 2007). To test if this holds true for the mGluR6-null mouse, we looked at the expression patterns of Gβ₅ and RGS11. Consistent with previous results (Morgans et al. 2007), in WT ON bipolar cells, these proteins were restricted to the dendritic tips, but in the mGluR6-null cells, they appeared more diffusely distributed throughout the ON-bipolar cell with weaker staining of the tips than in WT. For Gβ₅, staining in the outer plexiform was lower for null retina than for WT, and, in the INL and IPL, the average intensity was higher for null than for WT (Fig. 5, B and D). This indicates redistribution or mislocalization of Gβ₅ rather than downregulation. For RGS11, staining in the OPL of the null retina was devoid of strong puncta, but the total intensity in this layer was similar to that in WT (Fig. 5, C and D). This means that RGS11 does not make it to the tip, but remains in the primary dendrites.

TRPM1 is likely present on the plasma membrane of the ON bipolar cell's soma in WT and mGluR6-null retinas.

While the expression of TRPM1 and the cascade modulators in the dendritic tips is greatly reduced, it is not totally eliminated. Therefore, the reduction on its own cannot explain the inactivity of the presumably constitutively active TRPM1 channel. Because the rod bipolar cell is very compact, the opening of only a few channels should produce a significant voltage change and noise. We, therefore, tested the possibility that TRPM1 fails to be targeted to the plasma membrane in the mGluR6-null cells. Toward this end, we used Western blots of enriched membrane fractions with antibodies against TRPM1. Antibodies against Na⁺-K⁺-ATPase or KCC2 were used as markers for the plasma membrane, and against P58 as a marker for intracellular membranes. For WT retinas we found that, as expected for a channel protein, a strong TRPM1 (∼180 kDa) band appeared in the fraction that was enriched in total cellular membrane. This band was missing from the supernatant that contained soluble proteins (Fig. 6A). The TRPM1 band also appeared when we prepared fractions enriched only in plasma membrane (Fig. 6B), and the band intensity in the mGluR6-null retinas was similar to that of the WT with a null/WT TRPM1 intensity ratio of 0.77 ± 0.21. The fraction of plasma membrane was not contaminated by intracellular membranes since P58, which was strong in the unpurified fraction that is typically discarded in this preparation, and was undetectable in the plasma membrane fraction (Fig. 6B, bottom blot).

Fig. 6.

TRPM1 is present on the plasma membrane of ON bipolar cell somas. A: Western blots of “crude” membrane fractions that include total cell membranes (tm) and of supernatant (cytosolic fraction; cyto). The strong band for KCC2 indicates that these are membrane fractions. The same blot was also stained for actin to show loading in the cytosolic fraction. Bands intensities of TRPM1 (180 kDa, green) are similar in WT and mGluR6-null (KO) retinas. B: Western blots of enriched plasma membrane (pm) and the unpurified (up) fractions stained for TRPM1 and the plasma membrane marker Na⁺-K⁺-ATPase (ATPase). Same blot was also probed for the ER-Golgi intermediate compartment marker. The plasma membrane fraction, which has a strong ATPase band, has similar TRPM1 bands in WT and mGluR6-null retinas. This fraction was not contaminated by ER-Golgi membranes since P58 was not detected in the “pm” fraction, although it was strong in the “up” fraction. C: a dissociated ON bipolar cell from WT shows staining for Na⁺-K⁺-ATPase (cyan), TRPM1 (red), or both. Arrows point to regions of concentrated TRPM1 located between regions of concentrated ATPase. D: line intensity profiles plotted as in Fig. 1E (x-axis is pixel number, y-axis is intensity) for the triple-labeled dissociated cells. These profiles demonstrate the four categories of relationships between the TRPM1-intensity profile (red) and the membrane stain (Na-K-ATPase, blue). E: a histogram shows the percentage of lines that fall into each intensity profile category; categories are numbered as in D. Total number of lines analyzed: 198 for WT and 219 for null mouse from two independent experiments. F: average intensity profiles for WT and mGluR6-null cells are very similar. The average staining intensity for Na⁺-K⁺-ATPase is plotted in blue (mem), and that for TRPM1 in red (TRPM). Solid traces indicate WT, and dashed traces indicate null cells. The blue line on top of the x-axis denotes the region where the nucleus resides, so TRPM1 staining in this region is mainly due to out-of-focus information and is considered background; the dashed arrow indicates the level of this background. The average intensity at the membrane is indicated by the short arrow.

To further test if some TRPM1 was present on the plasma membrane, we increased the resolution of immunostained images (beyond what can be achieved in a 10- to 15-μm-thick section) by marking the plasma membrane of bipolar cells with Na⁺-K⁺-ATPase (ATPase) (Molday et al. 2007) and observing TRPM1 localization in the somas and primary dendrites of dissociated cells. The staining patterns of the cells in WT and mGluR6-null mice were similar. TRPM1 staining was often punctate, suggesting that the channels are clustered (Fig. 6C). These clusters were found both in the cytosol and on (or close to) the plasma membrane. ATPase staining was also punctate in most images, and the ATPase puncta often interdigitated with the TRPM1 puncta (arrows in Fig. 6C). Intensity profiles along lines that cross the membrane perpendicularly showed the relationship between the TRPM1 stain (red arrows in Fig. 6D) and the membrane location (which we took to be the peak of the Na⁺-K⁺-ATPase stain; blue arrows in Fig. 6D). We have categorized these profiles into four types: 1) TRPM1 shift L profiles in which a peak in TRPM1 staining precedes the Na⁺-K⁺-ATPase peak (i.e., staining appears to be slightly outside the membrane, likely due to staining above or below the confocal plane where the membrane may be shifted relative to the confocal plane; 2) on membrane profiles in which TRPM1 staining has a clear peak that corresponds exactly to the ATPase peak (note that this peak could be accompanied by additional intracellular peaks); 3) shift R profiles in which the TRPM1 peak is shifted inward by one pixel; and 4) not on membrane profiles in which the TRPM1 peak is shifted inward by two or more pixels (Fig. 6D). This categorization showed that, in ∼45% of the profiles, TRPM1 did not appear on the plasma membrane but only in the cytosol (profile 4) (Fig. 6E). Profiles where the TRPM1 peak was shifted by only one pixel to the right (profile 3) were considered ambiguous because the peak may result by blur from intracellular TRPM1 staining. Profiles 1 and 2 were considered to indicate membrane staining because the peak in staining intensity is unlikely to result from intracellular staining. These profiles made up at least 30% of the profiles, and the percentage was similar for WT and null cells (Fig. 6E).

To get an idea of the average profile, we aligned single profiles according to membrane position and averaged the TRPM1 intensity values for each pixel. Consistent with the above interpretation, the average TRPM1 staining profile peaked intracellularly, but staining under the membrane was more intense than the background level inside the cell (the background inside the cell was higher than the intensity just outside the membrane because of emission collected from above and below the focal plan) (Fig. 6F). The average intensity of TRPM1 staining at the membrane (background subtracted) was 53% and 68% of the intracellular peak for WT and null cells, respectively (an average of ∼200 lines from 27 WT cells and 22 null cells, two experiments). Moreover, in many cells that retained their initial dendrite or axon, the TRPM1 staining localized mainly to the plasma membrane (Fig. 7). This was true for both WT and null retinas (3 experiments). Thus, using several techniques and approaches, we find that a significant fraction of TRPM1 protein is present on the plasma membrane of the bipolar soma, and that this expression level is very similar between WT and mGluR6-null retinas.

Fig. 7.

TRPM1 is present on the plasma membrane of mGluR6-null ON bipolar cell's processes. Shown are dissociated cells that retained their processes; staining for Na⁺-K⁺-ATPase is in cyan, for TRPM in red, and DAPI stain is blue. A1: low magnification of a WT bipolar cell. A2: TRPM1 staining of the process is shown at higher magnification; the stain forms two parallel lines that correspond to the plasma membrane (arrowheads). B and C: cells dissociated from mGluR6-null mouse. Arrows in B2 point to plasma membrane on a primary dendrite. In C, several somas (s) (that contain a DAPI-stained nucleus), and several dendrites (*) (smaller circles with no DAPI) are viewed in a cross section. TRPM1 staining clearly outlines the processes.

DISCUSSION

Deletion of cascade elements greatly reduces channel activity.

We found that deleting mGluR6 from rod bipolar cells greatly diminished TRPM1 channel activity. This was evident from the mGluR6-null rod bipolar cell's more hyperpolarized (compared with WT) dark resting membrane potential, lower conductance, lower current noise, and lack of activation by capsaicin. This level of reduced channel activity seems similar to that seen in the TRPM1-null or nyctalopin-null rod bipolar cell (Gregg et al. 2007; Koike et al. 2010b), although precise comparison is impossible due to the different recording conditions. Our finding of diminished TRPM1 channel activity was initially unexpected since it has been assumed that deletion of the mGluR6 receptor renders the G protein inactive and the channel open. Indeed this assumption was used to explain why application of L-AP4, which hyperpolarizes ON bipolar cells, affected ganglion cell stratification, while deletion of mGluR6, which may render them depolarized, did not (Bisti et al. 1998; Bodnarenko and Chalupa 1993; Bodnarenko et al. 1995; Gargini et al. 1998; Tagawa et al. 1999). The fact that mGluR6-null rod bipolar cells were even more hyperpolarized than the WT may be surprising in face of reports that most cation channels in the dark-adapted rod bipolar cell are closed (Sampath and Rieke 2004). However, the state of dark adaptation in different studies is likely different as may be indicated by a large range of holding currents at the same potential (e.g., Gregg et al. 2007; Okawa et al. 2010; Sampath and Rieke 2004).

Although our results were initially unexpected, as more genes involved in the mGluR6 cascade are deleted or mutated, it becomes clear that this reduction of constitutive channel activity is the rule and not the exception. Thus similar to mGluR6-null rod bipolar cells, rod bipolar cells of Gα_o-null or nyctalopin-null retinas show low channel noise and small holding currents at potentials close to −60 mV (Gregg et al. 2007; Okawa et al. 2010). Since active Gα_o closes the channel (Koike et al. 2010b), its deletion is expected to eliminate the closing agent and keep the channel open. However, this does not happen. As for nyctalopin, lack of this extracellular protein that is concentrated at the dendritic tips of ON bipolar cells did not affect expression of mGluR6 or Gα_o (Ball et al. 2003). However, similar to the effect of mGluR6 deletion, absence of nyctalopin did reduce TRPM1 localization at the dendritic tips (Pearing et al. 2011). In the nyctalopin study, the authors assumed that the TRPM1 channel was not localized to the plasma membrane on the soma, so they attributed the lack of TRPM1 activity to the reduced TRPM1 staining at the tips. We have shown here that, even after mGluR6 deletion, a fraction of TRPM1 is very likely present on the plasma membrane of the dendrites and the somas. Moreover, our quantification of staining in the OPL has shown that TRPM1 is reduced, but not totally eliminated. Therefore, we cannot presume that lack of channel activity is due to the diminished staining at the dendritic tips.

Possible modulation/gating of the TRPM1 channel.

Our finding that mGluR6 deletion renders the rod bipolar cell hyperpolarized, together with previous reports from other genetically modified mice, may give an insight into the gating and channel activity of the TRPM1 channel in the rod bipolar cell. Performing excised patch recordings on transfected CHO cells, Koike et al. reported that applying activated Gα_o to the cytoplasmic face of the patch deactivated TRPM1 channels (Koike et al. 2010b), while applying Gβγ did not (Koike et al. 2010a), indicating that TRPM1 is constitutively active, and active Gα_o closes the channel. However, the fact that Gα_o deletion does not render TRPM1 in rod bipolar cells open (Okawa et al. 2010) may indicate that the transduction channel needs inactive Gα_o to keep the channel open. If this were all that the channel needed, then it should be active in mGluR6-null cells where expression of Gα_o is normal. Our finding that mGluR6 deletion renders the channel closed may mean that, to be open, the channel needs to be in a protein complex. This idea further gains support from studies of bipolar cells in the rd retina. In this retina, the genetic composition of the rod bipolar machinery is intact, yet the mGluR6-related cascade proteins are incorrectly localized due to lack of photoreceptor input, and the rod bipolar cell remains hyperpolarized (Borowska et al. 2011; Strettoi and Pignatelli 2000). Since perturbation of the rod-to-rod bipolar synapse leads to dispersion of several cascade proteins, including Gβ₅, RGS11, and RGS7 (Cao et al. 2008), the factor or protein that is crucial is not clear. This theory, however, does not exclude the possibility that during development the channels were initially open, but a constant influx of calcium desensitizes them, as has been suggested to occur during light adaptation (Shiells and Falk 1999).

mGluR6 is required for correct localization of cascade elements at the dendritic tips of the ON bipolar cell.

We find here that the G protein subunit candidates, Gα_o, Gβ₃, and Gγ₁₃, retain normal expression level and pattern in the mGluR6-null bipolar cells, but TRPM1, Gβ₅, and RGS11 fail to concentrate at the dendritic tips. Lack of concentration at the dendritic tips of bipolar cells in nob4 mouse, a mouse model with a genetic mutation in mGluR6, has been shown before for Gβ₅, RGS11, RGS7, and R9AP (Cao et al. 2009) and recently also for TRPM1 (Cao et al. 2011). In contrast, deletion of, or mutations in, TRPM1, RGS7, RGS11, nyctalopin, or Gα_o does not significantly affect localization of mGluR6 (Cao et al. 2009; Dhingra et al. 2000; Gregg et al. 2007; Koike et al. 2010b; Zhang et al. 2010). This suggests that most elements of the mGluR6 cascade require the presence of mGluR6 to localize at the tip (the trafficking hypothesis). If so, why are the G protein subunits that are expected to interact directly with the receptor apparently not affected? Two explanations may hold. First, the distributions of the G protein subunits within the rod bipolar cell are clearly different from the distributions of mGluR6 and the mentioned modulators. While mGluR6, RGS7, RGS11, and Gβ₅ are restricted to the tips, the G protein subunits are present throughout the dendrites and somas and are even seen in the initial axon (Gα_o) and the axon terminals (Gγ₁₃). This may suggest that the G protein subunits have to be in excess, so there may be an independent mechanism that determines their distribution. Second, it is possible that the invaginating dendritic tips are devoid of Gα_o staining, but this small change is impossible to detect when the rest of the primary dendrite is stained.

The staining pattern of TRPM1 lies between the two patterns described above (punctate vs. diffuse). Within the dendrites, TRPM1 is restricted to the tips, displaying a punctate appearance like that of mGluR6. However, within the cell, this channel distributes diffusely in the soma, with a small amount also spreading to the axon, a pattern reminiscent of that obtained for the G protein subunits. Consequently, the effect of mGluR6 deletion significantly affects the localization of TRPM1 in the tip, as it does for proteins that give punctate appearance, and it does not affect TRPM1 localization in the soma, consistent with its lack of effect on the distribution of the G protein subunits.

An alternative to the “trafficking” hypothesis is the “activity” hypothesis. This hypothesis states that the cascade elements that concentrate at the dendritic tips find their way to the tips just as mGluR6 does, but they are not retained there because of a lack of activity. This idea arises from two observations. First, it appears that nob4 retina has a reduced number of invaginating rod bipolar dendrites (Cao et al. 2009). Second, mutations in bassoon, a presynaptic protein present on synaptic ribbons, or deletion of a presynaptic calcium channel (Cacna1f) reduces immunostaining of several postsynaptic proteins, including mGluR6. Furthermore, in the mutated bassoon retina, the puncta immunostained for Cacna1s continue to decrease with age (Specht et al. 2009). Thus it is possible that the integrity and the precise localization of the synaptic proteins require existence of a functional synapse.

GRANTS

This study was supported by National Institutes of Health Grants EY11105 (N. Vardi) and NEI P30 EY01583 (University of Pennsylvania); the Fundamental Research Funds for the Central Universities of China 21609101, 11611604; the National Basic Research Program of China (973 Program) 2011CB707501 (Y. Xu); and Grant-in-Aid for Scientific Research (B) and CREST from Japan Science and Technology Agency (T. Furukawa).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: Y.X., A.D., M.F., C.K., and T.F. performed experiments; Y.X. and A.D. analyzed data; Y.X., A.D., and N.V. interpreted results of experiments; Y.X., A.D., and M.F. prepared figures; Y.X., A.D., and N.V. drafted manuscript; Y.X., A.D., C.K., T.F., and N.V. edited and revised manuscript; Y.X., A.D., and N.V. approved final version of manuscript; N.V. conception and design of research.