- Email: [email protected]
Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters
Report
Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters Highlights d
Cerebellar unipolar brush cells mediate slow synaptic responses using AMPAR
d
After exocytosis, the slow signals arise as AMPARs recover from desensitization
Authors Hsin-Wei Lu, Timothy S. Balmer, Gabriel E. Romero, Laurence O. Trussell
Correspondence
d
The slow responses require the stargazin accessory subunit for AMPAR
[email protected]
d
Exposure of AMPAR to glutamate is controlled by plasma membrane transporters
Glutamate receptors that normally participate in rapid synaptic signaling can also create ultra-slow signals. Lu et al. show that slow signals require TARP accessory subunits to glutamate receptors as well as tight control of glutamate time course mediated by transporters.
Lu et al., 2017, Neuron 96, 1–8 September 27, 2017 ª 2017 Elsevier Inc. http://dx.doi.org/10.1016/j.neuron.2017.08.043
In Brief
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
Neuron
Report Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters Hsin-Wei Lu,1,4 Timothy S. Balmer,2,4 Gabriel E. Romero,3 and Laurence O. Trussell2,5,* 1Neuroscience
Graduate Program Hearing Research Center and Vollum Institute 3Department of Physiology and Pharmacology Oregon Health and Science University, Portland, OR, USA 4These authors contributed equally 5Lead Contact *Correspondence: [email protected] http://dx.doi.org/10.1016/j.neuron.2017.08.043 2Oregon
SUMMARY
AMPARs mediate the briefest synaptic currents in the brain by virtue of their rapid gating kinetics. However, at the mossy fiber-to-unipolar brush cell synapse in the cerebellum, AMPAR-mediated EPSCs last for hundreds of milliseconds, and it has been proposed that this time course reflects slow diffusion from a complex synaptic space. We show that upon release of glutamate, synaptic AMPARs were desensitized by transmitter by >90%. As glutamate levels subsequently fell, recovery of transmission occurred due to the presence of the AMPAR accessory protein stargazin that enhances the AMPAR response to low levels of transmitter. This gradual increase in receptor activity following desensitization accounted for the majority of synaptic transmission at this synapse. Moreover, the amplitude, duration, and shape of the synaptic response was tightly controlled by plasma membrane glutamate transporters, indicating that clearance of synaptic glutamate during the slow EPSC is dictated by an uptake process. INTRODUCTION AMPA receptors (AMPARs) mediate the majority of fast excitatory postsynaptic currents (EPSCs) in the brain (Jonas, 2000). The brevity of EPSCs and rapid deactivation of AMPARs depend upon a short lifetime of synaptically released glutamate, estimated to be about 1 ms (Clements et al., 1992). Another key factor contributing to fast AMPAR-mediated EPSCs is rapid desensitization, which decreases response amplitudes by >90% within 10 ms upon prolonged exposure to glutamate (Colquhoun et al., 1992; Raman and Trussell, 1992; Silver et al., 1996; Trussell and Fischbach, 1989). Thus, even when cleft glutamate clearance is slow, desensitization still forces AMPARmediated EPSCs to decay quickly (Trussell et al., 1993). In stark contrast to this picture of typical AMPAR synapses is the large mossy fiber-unipolar brush cell (UBC) synapse in the granular layer of cerebellar cortex and cochlear nucleus (Floris
et al., 1994; Rossi et al., 1995). Stimulation of this synapse evokes typical fast EPSCs, but these are followed by a slow, AMPAR-mediated EPSC lasting hundreds of milliseconds (Borges-Merjane and Trussell, 2015; Kinney et al., 1997; Rossi et al., 1995). The mossy fiber-UBC synapse features an extensive, convoluted synaptic cleft between the presynaptic terminal and postsynaptic brush-like dendrite (Rossi et al., 1995). Kinney et al. (1997) proposed that the slow current is the combined result of delayed clearance from this large synaptic cleft and the biophysical properties of AMPARs. Upon such prolonged glutamate exposure, synaptic AMPARs would enter steadystate desensitization and occasionally reopen, generating the slow EPSC. However, direct evidence for ‘‘glutamate entrapment’’ (Rossi et al., 1995) requires information about the kinetic state of receptors during synaptic transmission, their molecular properties, and the forces that determine the glutamate lifetime in the cleft. We tested this hypothesis in UBCs of the vestibular cerebellum. Fast UV uncaging of glutamate after synaptic stimuli revealed that after exocytosis of glutamate, >90% of AMPARs become desensitized. Thereafter, receptors slowly recover from desensitization concurrent with an increase in the EPSC amplitude. Doseresponse relations show that AMPARs produce smaller equilibrium responses to millimolar levels of glutamate than to micromolar levels, suggesting that the slow EPSC tracks recovery from desensitization as glutamate is removed. This decrease in response to high glutamate levels was absent in UBCs from g2 transmembrane AMPAR regulatory protein (TARP) mutant stargazer (stg) mice, and slow EPSCs in these mice were reduced in amplitude. Finally, the slow time course of glutamate was dictated by glutamate transporters, as block of transport profoundly distorted synaptic responses. Thus, at the mossy-fiber-UBC synapse, exocytosis initiates a process of transmission controlled by a balance between glutamate uptake and the heightened sensitivity to transmitter bestowed by TARPs. RESULTS A Slow EPSC Mediated by Synaptic AMPA Receptors The fast-then-slow EPSC sequence that is characteristic of UBCs was evoked by a 100-Hz, 10-stimulus train in mouse cerebellar brain slices (STAR Methods; Figures 1A and 1B) (BorgesNeuron 96, 1–8, September 27, 2017 ª 2017 Elsevier Inc. 1
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
Figure 1. AMPARs Mediate a Slow EPSC at Mossy Fiber-UBC Synapse (A) Top: UBC was filled with Alexa 488 to visualize the location of the brush. Bottom: recording diagram: electrode placed nearby the brush to stimulate a mossy fiber. (B) Left: AMPAR antagonist GYKI-53655 (red) blocked both fast and slow EPSCs evoked by 10 stimuli at 100 Hz. Black, control. Right: this stimulus regime increased spike rate during the slow EPSC. (C) Expanded view of (B). Left: fast EPSCs evoked during train stimulation (triangles) depress profoundly. Right: a slow EPSC occurred after train stimulation stops. (D) Short-term depression kinetics of the fast EPSCs can be fit by a two-exponential decay (red). The peak of slow EPSC occurred 150 ms after end of stimulation. (E) The slow EPSC begins after stimuli stop, regardless of number of stimuli (103–403). Bars indicate duration of 100-Hz stimulation applied to the same cell. (F) Summary of the normalized time to peak from last stimulus of the slow EPSC versus number of stimuli from nine cells. Each symbol represents a different cell. Linear regression fit (red) showed no significant correlation. (G) Application of 5% dextran solution slowed the rise and decay of slow EPSC. Traces are normalized to the peak slow EPSC. (H) Summary of dextran’s effect on decay time constant and rise time of the slow current. Error bars given as ±SEM.
Merjane and Trussell, 2015; van Dorp and De Zeeuw, 2014; Rossi et al., 1995; Zampini et al., 2016). A complete block of the EPSC by GYKI-53655 (GYKI) (75 mM) confirmed that the slow currents were AMPAR mediated (Figure 1B). As previously described, the slow current increases the firing rate of UBCs (Figure 1B; Borges-Merjane and Trussell, 2015; Rossi et al., 1995). In 17 cells tested, amplitudes of fast EPSCs depressed rapidly and nearly completely during train stimulation (fit to decay of peak amplitudes: fast decay t: 9.7 ± 7.7 ms, slow decay t: 18.5 ± 2.0 ms, fast component 55% of total; average of EPSC5–10/EPSC1: 4.3% ± 0.5%; Figures 1C and 1D). Following strong synaptic depression, a slow EPSC emerged after stimuli were terminated, reaching a peak of 13.6 ± 1.5 pA (19.1% ± 5.0% of EPSC1) 130.1 ± 9.0 ms after the train stimulation and decaying with a t of 453.5 ± 53.5 ms back to baseline (Figures 1C and 1E). The slow EPSC generated a charge transfer seven 2 Neuron 96, 1–8, September 27, 2017
times that of all prior fast EPSCs plus the tonic current between each fast EPSC (charge transferred during fast EPSCs: 1.1 ± 0.1 pC, slow EPSC: 7.6 ± 1.1 pC, n = 17). Thus, the slow EPSC was the dominant synaptic signal. The slow EPSC may be a spillover current mediated by extrasynaptic receptors (Nielsen et al., 2004). If this is the case, one would expect the slow EPSC to begin during a long stimulus train as accumulated glutamate spilled out to distant receptors. However, when we varied the number of stimuli, the slowly rising EPSC was observed to begin only after the termination of stimulation (Figure 1E). Furthermore, the current’s rise time after stimuli were halted was constant, regardless of the number of stimuli (range 5–100 stimuli; p = 0.24, R2 = 0.06, linear regression fit; Figure 1F). This result argues against the spillover hypothesis, as such diffusion should not depend upon the cessation of transmitter release. Rather, the results indicate that synaptic AMPARs were partially inhibited by glutamate accumulated in the synaptic cleft, and this inhibition is relieved as transmitter levels gradually decline (Kinney et al., 1997). If indeed the slow EPSC time course arises as glutamate gradually falls, its decay should be prolonged by restricting the diffusion time of glutamate. We thus applied 5% dextran in the bath
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
solution to increase the viscosity of extracellular space and retard transmitter diffusion (Min et al., 1998). After 15 min in dextran, the time to peak and decay time constant of the slow EPSC were prolonged (time to peak: from 135.0 ± 12.0 to 214.3 ± 38.5 ms, p = 0.03, n = 8; decay time constant: from 464.5 ± 74.2 to 688.9 ± 125.7 ms, p = 0.007, n = 8; paired t tests), confirming that the mossy fiber-UBC cleft limits the clearance of transmitter (Figures 1G and 1H). Role of Glutamate Transporters The prolongation of the slow EPSC by dextran indicates a role for diffusion in controlling the clearance of glutamate but does not exclude that glial or neuronal glutamate transporters might also affect glutamate decay. We thus tested the role of transporters using the selective blocker TBOA and found a striking dosedependent effect on the EPSC. At the highest TBOA dose examined (50 mM), the rise time of the slow EPSC was lengthened over 3-fold and the decay time lengthened over 2-fold over control (Figures 2A, 2C, and 2D). By contrast, the fast EPSC elicited by the first stimulus in the train was only slightly slower at the same concentration (Figures 2G–2J). Coincident with these actions, however, was a significant increase in inward holding current even with 10 mM TBOA (Figures 2A and 2E); this current was abolished by subsequent application of GYKI and therefore must represent tonic activation of AMPAR by ambient glutamate (Figure 2A). Given the sensitivity of AMPARs, ambient glutamate may cause tonic AMPAR desensitization. Indeed, we observed an 40% reduction in the peak amplitude of the first fast EPSC in 50 mM TBOA (Figures 2G and 2I). Moreover, there was an apparent interaction between the tonic current and the steady-state current during synaptic activation: the steady-state current was slightly less inward (relative to the baseline current level before stimulation) in 10 mM TBOA compared to control and became clearly outward (relative to baseline) in 50 mM TBOA (Figures 2B and 2F; 6 of 8 cells tested). Following the stimuli, the slow EPSC then emerged, becoming inward again before finally settling back to the baseline defined by the tonic current (Figures 2A and 2B). Subsequent application of GYKI blocked this complex synaptic waveform (Figure 2A). Thus, loss of transporter activity slowed the synaptic response and led to accumulation of ambient glutamate, generating a tonic current. The ongoing exocytosis associated with the train stimuli further desensitized the AMPARs, leading to the apparent outward current. Therefore, a function of the transporters is not only to narrow the EPSC time course, but to limit both glutamate accumulation at rest and receptor desensitization during stimuli. Slow EPSC Mediated by Recovery of AMPARs from Desensitization The delayed onset of the slow EPSC, and the appearance of an outward current when transporters are blocked, strongly suggests that synaptic transmission at the UBC synapse is accompanied by significant desensitization of AMPARs. We used glutamate uncaging to directly observe the magnitude of desensitization and its recovery during the EPSC. MNI-glutamate (1 mM) was bath applied, and a 2-ms UV flash was delivered over the UBC brush to locally uncage glutamate (Figure 3A); this method was combined with synaptic stimulation at varying inter-
vals. As the flash area contains the compact region of the brush, we expect it to activate all synaptic receptors. Control experiments confirmed that AMPARs were largely restricted to the brush region and that uncaged glutamate accessed synaptic receptors (STAR Methods; Figure S1). We found that the peak of the fast transient uncaging response (uEPSC) was greatly reduced when evoked closer to the end of train stimulation (Figures 3B and 3C), with maximal decrease of the test pulse of over 90%. Furthermore, the uEPSC gradually recovered as the synaptic-uncaging stimulus intervals increased, with a t of 267.1 ± 24.4 ms (Figure 3C; n = 6). A similar phenomenon was also observed using a paireduncaging pulse protocol, in which the peak uEPSC evoked by the second uncaging pulse recovered with a t of 413.9 ± 68.1 ms (Figures 3E and 3F; n = 8). Assuming the concentration of uncaged glutamate is similar in each trial (see STAR Methods), the reduction in peak uEPSC is most easily explained by desensitization of AMPARs. As uncaging activates the same set of receptors utilized by synaptic transmission, we conclude that synaptically released glutamate desensitizes AMPARs, and this desensitization is maximal at the end of the last stimulus. The initial rise in the slow EPSC after stimuli cease must be generated by a decline in cleft glutamate occurring simultaneously with recovery from desensitization. This explains two observations. First, an ‘‘undershoot’’ of current was seen right after uncaging, most obvious near the peak of the slow EPSC (Figures 3B and 3E). This undershoot current always reached a level similar to the current level just after the last synaptic stimulus and was followed by its own rise and fall in slow current (Figures 3B and 3E). Thus, cleft glutamate increased upon uncaging, further desensitized the receptors, and then evoked a slow EPSC as glutamate levels subsequently declined. Second, when uncaging pulses were applied at the beginning of the rise of the slow EPSC or at a later point when current had fallen back to the same level, the uncaging responses were distinctly different (Figure 3B, inset). Plotting baseline current just before an uncaging pulse against the fractional recovery of the uncaging response from desensitization revealed a bell-shaped relation, with a maximum at 40% recovery (Figures 3D and 3G). As the slow EPSC peaked when 60% of AMPARs were desensitized, this result suggests that an intermediate glutamate concentration was retained in the synaptic cleft during the slow EPSC. Thus, the slow EPSC was mediated by AMPARs that reopened as they recovered from desensitization. Glutamate Sensitivity of AMPAR and Role of TARPs The ability of AMPARs to produce larger currents in response to falling levels of transmitter implies a molecular specialization for non-monotonic sensitivity to glutamate. TARP co-expression is associated with reduced desensitization and enhanced steadystate AMPAR currents (Jackson and Nicoll, 2011; Tomita, 2010). To test for a role of TARPs in the slow EPSC, we recorded EPSCs in UBCs from stg mice, which lack g2 TARP (Letts et al., 1998), comparing their amplitudes to those of wild-type (WT) mice. EPSCs in stg UBCs showed both fast and slow components upon train stimulation, as in WT UBCs (Figures 4A and 4B). Scatterplots of peak amplitude of fast EPSCs versus slow EPSCs from different UBCs were positively correlated (R2 = 0.163; p = 0.0027, n = 53), presumably as synapses varied in their Neuron 96, 1–8, September 27, 2017 3
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
Figure 2. Transporters Determine Glutamate Time Course (Ai and Aii) Examples of the effect of TBOA on the response to synaptic stimulation. Dashed line indicates holding current in control (0 mM TBOA). TBOA increased the inward holding current, measured immediately before stimulation train (circle). TBOA decreased the steady-state inward current, measured at the end of the train (square) relative to the holding current of the same trace (circle) (i). TBOA increased the time to peak and the decay of the slow EPSC that occurs at the end of the stimulation train (triangle). In 50 mM TBOA (ii), the steady-state current became outward relative to the standing glutamate current. Addition of 50 mM GYKI blocked both standing current and the outward current, indicating that they were due to tonic AMPAR activation and desensitization, respectively. (B) Same traces as in (A), except that they are overlaid with baselines subtracted to show that steady-state current became outward in 50 mM TBOA. (C) TBOA increased the time to peak of the slow EPSC that begins at the offset of the synaptic stimuli. Paired t tests: 0 mM versus 10 mM: p = 7E–5, n = 11; 10 mM versus 50 mM: p = 0.0006, n = 8; 0 mM versus 50 mM: p = 9E–5, n = 8. (D) TBOA increased the decay of the slow EPSC that begins at the offset of the synaptic stimuli (measured from the inward peak to 10% of the baseline. Paired t tests: 0 mM versus 10 mM: p = 0.009, n = 11; 0 mM versus 50 mM: p = 0.018, n = 8. (E) The inward holding current was increased by TBOA and blocked by GYKI. Paired t tests: 0 mM versus 10 mM, p = 0.0003, n = 11; 10 mM versus 50 mM, p = 0.006, n = 8. Illustrated holding current values are relative to the holding current in the previous condition, such that the 10 mM value is relative to the 0 mM, 50 mM is relative to 10 mM, and GYKI is relative to 50 mM TBOA. Paired t test: 50 mM versus GYKI, p = 0.012, n = 6. (F) Change in steady-state current, measured between the 9th and 10th synaptic stimulus, relative to holding current. A positive change in steady state indicates that the current went outward during the stimulus train, as shown in (A) and (B). Paired t tests: 10 mM to 50 mM, p = 0.005, n = 8; 0 mM to 50 mM, p = 0.005, n = 8. Addition of 50 mM GYKI blocked the steady-state current, indicating that the outward current was an AMPAR-mediated current. Paired t test including cells that had an outward steady-state current: 50 mM to GYKI, p = 0.017, n = 6. (G) The first EPSC in the train in (A). (H) The first EPSC in the train peak normalized to illustrate small change in decay rate. (I) TBOA reduces amplitude of the first EPSC in the train. Paired t tests: 10 mM versus 50 mM, p = 0.006, n = 8; 0 mM versus 50 mM, p = 0.008, n = 8. (J) TBOA increases the decay of the first EPSC in the train. Paired t test: 0 mM versus 50 mM: p = 0.037, n = 8. Error bars given as ±SEM.
size, quantal content, and numbers of AMPARs (Zampini et al., 2016). However, when plotted for each mouse line, stg slow currents appeared to be smaller than in WT over a similar range of 4 Neuron 96, 1–8, September 27, 2017
fast EPSCs (Figures 4A and 4B, red versus black traces and points). Given that some of the variance in the population was accounted for by a correlation between EPSC amplitudes, we
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
Figure 3. AMPARs Are Desensitized and Recover during Slow EPSC (A) Experimental configuration. UV-laser spot was targeted to the brush to uncage MNI-glutamate after synaptic stimulation. (B) Uncaging at different times (arrowheads) during slow EPSC. Gray trace shows the control uEPSC. Inset: expanded view showing diminished size of uncaging response (*; black trace) delivered 15 ms after final synaptic stimulation. Blue trace delivered 85 ms later shows recovery even though baseline current is the same as in black trace. (C) Summary of recovery time course shown in (B). Each symbol represents a different cell. The recovery was fit by a single exponential (black line). (D) Steady-state current of the slow EPSC evoked just before the test uncaging pulse plotted against the proportional recovery of test pulse amplitude. Slow EPSC was evoked by synaptic stimulation as in (B). Each color represents a different cell (n = 6). The rising and then falling curves show that the same steady-state current corresponds to two different recovery states, indicating that they were produced by different glutamate levels. (E) Paired-pulse uncaging also showed a desensitized second uEPSC (triangle). Note the undershoot current (arrow) in the second uEPSC. (F) Recovery time course for the second uEPSC in paired uncaging paradigm. Each symbol represents a different cell. Recovery was fit with a single exponential (black line). (G) Steady-state current of the slow current evoked by an uncaging pulse just before a second test uncaging pulse as in (E), plotted against the proportional recovery of test pulse amplitude.
used permutation tests to determine whether WT and stg synapses differed in the amplitudes of slow EPSCs and fast EPSCs and in their slow/fast ratios (see STAR Methods). Slow EPSC amplitudes were 50% smaller in stg UBCs than in WT UBCs (stg median: 5.74 pA, inter-quartile range (IQR): 3.66–6.47 pA, n = 21; WT median: 11.62 pA, IQR: 8.42–17.79 pA, n = 32; permutation test, p < 0.0001). Fast EPSCs were also significantly smaller in stg UBCs, but the difference between medians was smaller than for slow EPSCs (29%), (stg median: 48.33 pA, IQR: 43.5–66.45 pA, n = 21; WT median: 68.47 pA, IQR: 54.51– 90.11 pA, n = 32; permutation test, p = 0.0099). The ratio of the slow/fast EPSCs were used to determine whether the slow EPSCs were generally more affected by the stg mutation than the fast EPSC. Indeed, the slow/fast EPSC ratios were significantly smaller in stg UBCs, indicating that AMPARs in the stg mutant mice have a reduced ability to generate the slow EPSC (stg median: 0.09, IQR: 0.06–0.14, n = 21; WT median: 0.15,
IQR: 0.10–0.23, n = 32; permutation test, p = 0.0133), and suggested a key role for stargazin in boosting sensitivity of the receptors to the slow synaptic glutamate transients at the UBC synapse. A striking effect of heterologous TARP expression on AMPAR function is a bellshaped glutamate dose-response relation, termed ‘‘auto-inactivation’’ (Morimoto-Tomita et al., 2009; Semenov et al., 2012), a feature described in some native AMPARs (Geoffroy et al., 1991; Raman and Trussell, 1992). Given that the slow EPSC rises as cleft glutamate decays, and that TARPs support this slow current, we expected that WT, but not stg, UBCs might feature a non-monotonic sensitivity to glutamate. UBCs enzymatically isolated with intact brush dendrites were used for constructing dose-response relations (Figure 4C), as glutamate transporters in slices may limit the access of exogenous glutamate to receptors. As shown in Figures 4D–4F, when 1 mM glutamate was applied to the cell, a steady current was generated without apparent desensitization. A fast transient current could be observed at higher glutamate levels (32–1,000 mM), which quickly desensitized and reached steady state. This steady-state response is presumed to originate from bound AMPARs at microscopic equilibrium between open, closed, and desensitized states. Absolute and normalized Neuron 96, 1–8, September 27, 2017 5
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
Figure 4. Stargazer UBCs Showed Reduced Slow EPSC and Monotonic Dose-Response Relation (A) Overlay of EPSCs from a WT (black) and a stg UBC (red). The peak of the fast EPSC is scaled to the same level and each trace represents the median slow/fast ratio of each group. Added filtering was applied to the slow EPSC. Inset: expanded view of slow EPSC. (B) Plot of slow versus fast EPSC amplitudes from 32 WT and 21 stg UBCs. The peak amplitudes of slow, but not fast, EPSC in stg UBCs were significantly smaller compared to WT. (C) A GFP-labeled dissociated UBC with intact dendritic brush. (D) AMPAR-mediated currents evoked by glutamate (gray bar) at the indicated concentrations applied to dissociated WT and stg UBCs. (E and F) Comparison of absolute (E) and normalized (F) amplitudes of steady-state AMPAR currents in WT and stg UBCs. Error bars given as ±SEM.
reduced slow EPSC reflects an inability of the AMPARs to give a rebound response as synaptic glutamate decays. DISCUSSION
steady-state responses at different glutamate concentrations are shown in Figures 4E and 4F. The steady-state response increased with glutamate concentration up to 32 mM (156.1% ± 0.1% of the 1 mM response, n = 9), and mean absolute values for currents at 32 mM were significantly larger than at 1 mM (n = 9, p = 0.004, paired t test). These experiments were then repeated on stg UBCs and revealed clear differences with WT UBCs. Maximal steady-state current was about half that of WT (Figure 4E; e.g., at 32 mM, WT: 11.5 ± 1.6 pA, n = 9; stg: 4.8 ± 0.7 pA, n = 7, p = 0.004, unpaired t test). These values at 32 mM were comparable to the peak amplitude of the slow EPSC within the same genotype (WT: p = 0.95; stg: p = 0.30, unpaired t tests). Notably, the shape of the stg dose-response relation was monotonic (Figure 4F), such that responses at 32 mM, normalized to responses for 1 mM, were not significantly different from 1 (p = 0.83, one-sample t test). Overall, in comparing the normalized curves between WT and stg, significant differences were observed for responses to 1, 10, 32, and 100 mM glutamate (p = 0.04–0.004). Together, these data indicate that stg mice have reduced sensitivity to glutamate, particularly in the range of 1–100 mM that would normally give rise to the non-monotonic dose-response curve, suggesting that the 6 Neuron 96, 1–8, September 27, 2017
This study showed that the slow EPSC of UBCs is determined by two factors. Continual transporter activity controls the time course of the EPSC, the depth of desensitization, and the baseline level of glutamate in the synaptic cleft. Moreover, a TARP-containing AMPAR renders the cell sensitive to the slowly changing levels of glutamate established by the transporters. In the absence of transport, glutamate levels fall extremely slowly and never reach a level below the sensitivity of the AMPAR, highlighting the ‘‘entrapment’’ hypothesis originally proposed by Rossi et al. (1995). This situation is dramatically different from that of mossy fiber synapses made onto granule cells or of fenestrated calyceal synapses, both of which have exclusively fast phasic AMPAR-mediated transmission even after transporter blockade (Renden et al., 2005; Sylantyev et al., 2013); thus, entrapment, transporter action, and the AMPAR-TARP complex allow the UBC to transmit in a nearly tonic mode despite sensing phasic exocytosis. Desensitized AMPARs during Synaptic Transmission The remarkable speed of AMPAR desensitization (Kiskin et al., 1986; Tang et al., 1989) raised the possibility that desensitization could determine the decay time of the EPSC (Trussell and Fischbach, 1989; Trussell et al., 1993), or impact the amplitude of EPSCs during repetitive activity (DiGregorio et al., 2007; Trussell et al., 1993), depending upon ease of clearance of glutamate and the rate of recovery from desensitized states. This study demonstrates a very different role for desensitization. In UBCs, recovery
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
from desensitization of TARP-containing AMPARs during slowly changing levels of transmitter leads to sizeable currents and accounts for a large fraction of transmitter action. Moreover, the balance between open and desensitized states is regulated by glutamate transporters, such that loss of transport reduces inward excitatory current during synaptic stimuli by enhancing desensitization. In this way, the UBC synapse can more effectively integrate presynaptic activity over longer periods of time than would be obtained with fast EPSCs. Role of Stargazin in Synaptic Transmission The 50% reduction of slow EPSC in stg UBCs suggests that stargazin is present in WT UBCs and regulates glutamate sensitivity of AMPARs. Interestingly, cerebellar granule cells also express stargazin (Hashimoto et al., 1999; Yamazaki et al., 2010), and stg granule cells lack functional AMPARs at mossy fiber synapses because of impaired AMPAR trafficking (Chen et al., 2000; Hashimoto et al., 1999). Stargazin regulates AMPAR gating, slowing down deactivation and desensitization rates (Priel et al., 2005; Tomita et al., 2005; Turetsky et al., 2005), and miniature EPSCs (Tomita et al., 2005). Stargazin also enhanced the equilibrium AMPAR current and generated non-monotonic dose-response curves (Morimoto-Tomita et al., 2009; Semenov et al., 2012). The physiological significance of the non-monotonic dose-response relation was previously obscure because slow glutamate transients typically do not occur at fast synapses. Our observations at UBC synapses reveal that stargazin also regulates synaptic transmission by allowing the synapse to respond to slowly changing, low levels of glutamate. STAR+METHODS Detailed methods are provided in the online version of this paper and include the following: d d d d
d
KEY RESOURCES TABLE CONTACT FOR REAGENT AND RESOURCE SHARING EXPERIMENTAL MODEL AND SUBJECT DETAILS B Animals and Genotyping METHOD DETAILS B Brain Slice Preparation B Electrophysiology B MNI-Glutamate Uncaging B Dissociated Cell Preparation B Immunohistochemistry B Fast Application of Glutamate B Pharmacology QUANTIFICATION AND STATISTICAL ANALYSIS
SUPPLEMENTAL INFORMATION Supplemental Information includes one figure and can be found with this article online at http://dx.doi.org/10.1016/j.neuron.2017.08.043.
ACKNOWLEDGMENTS This study is funded by National Institutes of Health (NIH) Grants NS028901 and DC004450 (to L.O.T.); N.L. Tartar Trust Fellowship (to H.-W.L.); DC014878 to T.S.B. We thank members of the Trussell lab for helpful discussions, Dr. Stephen David for assistance with statistical analysis, and Dr. Craig Jahr and Dr. Brett Carter for comments on the manuscript. We thank Michael Bateschell and Ruby Larisch for help with mouse colony management. Stargazer breeder mice were kindly provided by the Puthussery lab at the Casey Eye Institute. Received: August 9, 2016 Revised: July 25, 2017 Accepted: August 28, 2017 Published: September 14, 2017 REFERENCES Borges-Merjane, C., and Trussell, L.O. (2015). ON and OFF unipolar brush cells transform multisensory inputs to the auditory system. Neuron 85, 1029–1042. Chen, L., Chetkovich, D.M., Petralia, R.S., Sweeney, N.T., Kawasaki, Y., Wenthold, R.J., Bredt, D.S., and Nicoll, R.A. (2000). Stargazin regulates synaptic targeting of AMPA receptors by two distinct mechanisms. Nature 408, 936–943. Clements, J.D., Lester, R.A., Tong, G., Jahr, C.E., and Westbrook, G.L. (1992). The time course of glutamate in the synaptic cleft. Science 258, 1498–1501. Colquhoun, D., Jonas, P., and Sakmann, B. (1992). Action of brief pulses of glutamate on AMPA/kainate receptors in patches from different neurones of rat hippocampal slices. J. Physiol. 458, 261–287. DiGregorio, D.A., Rothman, J.S., Nielsen, T.A., and Silver, R.A. (2007). Desensitization properties of AMPA receptors at the cerebellar mossy fiber granule cell synapse. J. Neurosci. 27, 8344–8357. Floris, A., Din˜o, M., Jacobowitz, D.M., and Mugnaini, E. (1994). The unipolar brush cells of the rat cerebellar cortex and cochlear nucleus are calretinin-positive: a study by light and electron microscopic immunocytochemistry. Anat. Embryol. (Berl.) 189, 495–520. Geoffroy, M., Lambolez, B., Audinat, E., Hamon, B., Crepel, F., Rossier, J., and Kado, R.T. (1991). Reduction of desensitization of a glutamate ionotropic receptor by antagonists. Mol. Pharmacol. 39, 587–591. Hashimoto, K., Fukaya, M., Qiao, X., Sakimura, K., Watanabe, M., and Kano, M. (1999). Impairment of AMPA receptor function in cerebellar granule cells of ataxic mutant mouse stargazer. J. Neurosci. 19, 6027–6036. Jackson, A.C., and Nicoll, R.A. (2011). Stargazin (TARP g-2) is required for compartment-specific AMPA receptor trafficking and synaptic plasticity in cerebellar stellate cells. J. Neurosci. 31, 3939–3952. Jonas, P. (2000). The time course of signaling at central glutamatergic synapses. News Physiol. Sci. 15, 83–89. Kinney, G.A., Overstreet, L.S., and Slater, N.T. (1997). Prolonged physiological entrapment of glutamate in the synaptic cleft of cerebellar unipolar brush cells. J. Neurophysiol. 78, 1320–1333. Kiskin, N.I., Krishtal, O.A., and Tsyndrenko, A.Y.A. (1986). Excitatory amino acid receptors in hippocampal neurons: kainate fails to desensitize them. Neurosci. Lett. 63, 225–230. Letts, V.A., Felix, R., Biddlecome, G.H., Arikkath, J., Mahaffey, C.L., Valenzuela, A., Bartlett, F.S., 2nd, Mori, Y., Campbell, K.P., and Frankel, W.N. (1998). The mouse stargazer gene encodes a neuronal Ca2+-channel g subunit. Nat. Genet. 19, 340–347.
AUTHOR CONTRIBUTIONS
Min, M.-Y., Rusakov, D.A., and Kullmann, D.M. (1998). Activation of AMPA, kainate, and metabotropic receptors at hippocampal mossy fiber synapses: role of glutamate diffusion. Neuron 21, 561–570.
H.-W.L., T.S.B., and G.E.R. performed the experiments and analyzed data. H.-W.L., T.S.B., G.E.R., and L.O.T. designed experiments and wrote the manuscript.
Morimoto-Tomita, M., Zhang, W., Straub, C., Cho, C.-H., Kim, K.S., Howe, J.R., and Tomita, S. (2009). Autoinactivation of neuronal AMPA receptors via glutamate-regulated TARP interaction. Neuron 61, 101–112.
Neuron 96, 1–8, September 27, 2017 7
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
Morin, F., Din˜o, M.R., and Mugnaini, E. (2001). Postnatal differentiation of unipolar brush cells and mossy fiber-unipolar brush cell synapses in rat cerebellum. Neuroscience 104, 1127–1139.
Tang, C.M., Dichter, M., and Morad, M. (1989). Quisqualate activates a rapidly inactivating high conductance ionic channel in hippocampal neurons. Science 243, 1474–1477.
Nielsen, T.A., DiGregorio, D.A., and Silver, R.A. (2004). Modulation of glutamate mobility reveals the mechanism underlying slow-rising AMPAR EPSCs and the diffusion coefficient in the synaptic cleft. Neuron 42, 757–771.
Tomita, S. (2010). Regulation of ionotropic glutamate receptors by their auxiliary subunits. Physiology (Bethesda) 25, 41–49.
Priel, A., Kolleker, A., Ayalon, G., Gillor, M., Osten, P., and Stern-Bach, Y. (2005). Stargazin reduces desensitization and slows deactivation of the AMPA-type glutamate receptors. J. Neurosci. 25, 2682–2686.
Tomita, S., Adesnik, H., Sekiguchi, M., Zhang, W., Wada, K., Howe, J.R., Nicoll, R.A., and Bredt, D.S. (2005). Stargazin modulates AMPA receptor gating and trafficking by distinct domains. Nature 435, 1052–1058.
Raman, I.M., and Trussell, L.O. (1992). The kinetics of the response to glutamate and kainate in neurons of the avian cochlear nucleus. Neuron 9, 173–186.
Trussell, L.O., and Fischbach, G.D. (1989). Glutamate receptor desensitization and its role in synaptic transmission. Neuron 3, 209–218.
Renden, R., Taschenberger, H., Puente, N., Rusakov, D.A., Duvoisin, R., Wang, L.-Y., Lehre, K.P., and von Gersdorff, H. (2005). Glutamate transporter studies reveal the pruning of metabotropic glutamate receptors and absence of AMPA receptor desensitization at mature calyx of held synapses. J. Neurosci. 25, 8482–8497.
Trussell, L.O., Zhang, S., and Raman, I.M. (1993). Desensitization of AMPA receptors upon multiquantal neurotransmitter release. Neuron 10, 1185–1196.
Rossi, D.J., Alford, S., Mugnaini, E., and Slater, N.T. (1995). Properties of transmission at a giant glutamatergic synapse in cerebellum: the mossy fiber-unipolar brush cell synapse. J. Neurophysiol. 74, 24–42. €nen, K. Semenov, A., Mo¨ykkynen, T., Coleman, S.K., Korpi, E.R., and Keina (2012). Autoinactivation of the stargazin-AMPA receptor complex: subunit-dependency and independence from physical dissociation. PLoS ONE 7, e49282. Silver, R.A., Colquhoun, D., Cull-Candy, S.G., and Edmonds, B. (1996). Deactivation and desensitization of non-NMDA receptors in patches and the time course of EPSCs in rat cerebellar granule cells. J. Physiol. 493, 167–173.
Turetsky, D., Garringer, E., and Patneau, D.K. (2005). Stargazin modulates native AMPA receptor functional properties by two distinct mechanisms. J. Neurosci. 25, 7438–7448. van Dorp, S., and De Zeeuw, C.I. (2014). Variable timing of synaptic transmission in cerebellar unipolar brush cells. Proc. Natl. Acad. Sci. USA 111, 5403–5408. Watanabe, D., Inokawa, H., Hashimoto, K., Suzuki, N., Kano, M., Shigemoto, R., Hirano, T., Toyama, K., Kaneko, S., Yokoi, M., et al. (1998). Ablation of cerebellar Golgi cells disrupts synaptic integration involving GABA inhibition and NMDA receptor activation in motor coordination. Cell 95, 17–27.
Suter, B.A., O’Connor, T., Iyer, V., Petreanu, L.T., Hooks, B.M., Kiritani, T., Svoboda, K., and Shepherd, G.M.G. (2010). Ephus: multipurpose data acquisition software for neuroscience experiments. Front. Neural Circuits 4, 100.
Yamazaki, M., Fukaya, M., Hashimoto, K., Yamasaki, M., Tsujita, M., Itakura, M., Abe, M., Natsume, R., Takahashi, M., Kano, M., et al. (2010). TARPs g-2 and g-7 are essential for AMPA receptor expression in the cerebellum. Eur. J. Neurosci. 31, 2204–2220.
Sylantyev, S., Savtchenko, L.P., Ermolyuk, Y., Michaluk, P., and Rusakov, D.A. (2013). Spike-driven glutamate electrodiffusion triggers synaptic potentiation via a Homer-dependent mGluR-NMDAR link. Neuron 77, 528–541.
Zampini, V., Liu, J.K., Diana, M.A., Maldonado, P.P., Brunel, N., and Dieudonne´, S. (2016). Mechanisms and functional roles of glutamatergic synapse diversity in a cerebellar circuit. eLife 5, e15872.
8 Neuron 96, 1–8, September 27, 2017
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
STAR+METHODS KEY RESOURCES TABLE
REAGENT or RESOURCE
SOURCE
IDENTIFIER
Chicken IgY Anti-GFP
Aves
Cat # GFP-1020 RRID: AB_10000240
Alexa 488-conjugated AffiniPure Donkey Anti-Chicken IgY
Jackson ImmunoResearch
Cat # 703-545-155 RRID: AB_2340375
GYKI-53655
Tocris
Cat # 2555
JNJ-16259685
Tocris
Cat # 2333
LY-341495
Tocris
Cat # 1209
MNI-caged-L-glutamate
Tocris
Cat # 1490
(+)-MK-801 hydrogen maleate
Sigma
Cat # M107
Picrotoxin
Sigma
Cat # P1675
Strychnine hydrochloride
Sigma
Cat # S8753
Dextran (MW 35,000–45,000)
Sigma
Cat # D1662
Alexa 488 hydrazide sodium salt
Thermo Fisher Scientific
Cat # A10436
DL-TBOA
Tocris
Cat # 1223
Papain, Suspension
Worthington
Cat # LS003124
Minimum Essential Medium Eagle, no L-glut
Sigma
Cat # M0275
Trypsin Inhibitor, Ovomucoid
Worthington
Cat # LS003085
Mouse: C57BL/6J
Jackson Laboratory
RRID: IMSR_JAX:000664
Mouse: C57BL/6J-TgN(grm2-IL2RA/GFP)1kyo (mGluR2-GFP)
Gift from Robert Duvoisin
Watanabe et al., 1998
Mouse: B6C3Fe a/a-Cacng2stg/J (Stargazer)
Gift from Teresa Puthussery
RRID: IMSR_JAX:001756
Wavemetrics
RRID: SCR_000325
Antibodies
Chemicals, Peptides, and Recombinant Proteins
Experimental Models: Organisms/Strains
Software and Algorithms Igor Pro MATLAB
MathWorks
RRID: SCR_001622
Ephus
Suter et al., 2010
http://www.ephus.org
Excel
Microsoft
N/A
pClamp
Molecular Devices
RRID: SCR_011323
Prism 7
GraphPad
RRID: SCR_002798
Neuromatic
Jason Rothman
RRID: SCR_004186
Axograph
Axograph
RRID: SCR_014284
CONTACT FOR REAGENT AND RESOURCE SHARING Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact Laurence Trussell ([email protected]). EXPERIMENTAL MODEL AND SUBJECT DETAILS Animals and Genotyping Three mouse lines were used in this study. Wild-type C57BL/6J or transgenic C57BL/6J-TgN(grm2-IL2RA/GFP)1kyo (referred to as mGluR2-GFP) mouse line were used as control animals. In the latter mouse line, GFP is tagged to human interleukin-2 receptor a subunit with expression driven by the metabotropic glutamate receptor 2 (mGluR2) promoter (Watanabe et al., 1998). Since
Neuron 96, 1–8.e1–e4, September 27, 2017 e1
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
UBCs express mGluR2 (Borges-Merjane and Trussell, 2015), this mouse line was used for identifying UBCs in the brain slice or dissociated cell preparations, although in later experiments UBCs could be recognized with transmitted light, and confirmed by dye fills. For simplicity, these two lines were referred to as ‘‘wildtype’’ (wt) in this study. A third mouse line, B6C3Fe a/a-Cacng2stg/J (stargazer or stg) (Hashimoto et al., 1999), was used for experiments to measure the effect of TARP g2 (stargazin) mutant. This mouse line has a retroviral-like, early transposon insertion in the second intron of stargazin gene, resulting in a significant reduction of protein expression (Letts et al., 1998). Stg pups were produced by crossing heterozygous males with either heterozygous or homozygous females. The pups were genotyped according to the protocol described on Jackson Lab website (https://www.jax.org/strain/001756). Stg pups also expressed ataxic phenotype as described in previous studies (Hashimoto et al., 1999). Breeding of all three mouse lines were maintained in the animal facility managed by the Department of Comparative Medicine and all procedures were approved by the Oregon Health and Science University’s Institutional Animal Care and Use Committee. Pups from both sexes were used. Because synapse formation between mossy fiber and UBC is mature in animals older than postnatal day 21 (P21) (Morin et al., 2001), we only used pups older than this age (P21 – P39) for experiments. METHOD DETAILS Brain Slice Preparation Animals were anesthetized with isoflurane, decapitated, and the cerebellum was dissected from the skull under ice-cold high-sucrose artificial cerebral spinal fluid solution (ACSF) containing (in mM): 87 NaCl, 75 sucrose, 25 NaHCO3, 25 glucose, 2.5 KCl, 1.25 NaH2PO4, 0.5 CaCl2, 7 MgCl2, bubbled with 5% CO2/95% O2. Sagittal cerebellum sections containing lobe X were cut at 300 mm with a vibratome (VT1200S, Leica) in ice-cold high-sucrose ACSF. After cutting, slices were allowed to recover at 35 C for 30-40 min in ACSF containing (in mM): 130 NaCl, 2.1 KCl, 1.2 KH2PO4, 3 Na-HEPES, 10 glucose, 20 NaHCO3, 2 Na-pyruvate, 2 CaCl2, 1 MgSO4, and 3 myo-inositol, bubbled with 5% CO2/95% O2 (Borges-Merjane and Trussell, 2015). After recovery, slices were kept at room temperature (22 C) in ACSF until use. Electrophysiology For whole-cell recording in brain slices, slices containing cerebellar lobe X were transferred to a recording chamber perfused with warm ACSF (34 C) controlled by a peristaltic pump. Cells were visualized with a 40X magnification objective on the stage of an upright microscope equipped with an infrared Dodt contrast mask and a blue LED for fluorescence optics. In slices from mGluR2-GFP mouse line, UBCs were identified as GFP-positive neurons with brushy dendritic structure. In brain slices from wildtype or stargazer mouse line, UBCs were first identified as round cells with soma diameter 10 mm in the cerebellar granular layer. Only data from UBCs in the lobe X were used for analysis in this study. All cells recorded were filled with 10 mM Alexa Fluor 488 hydrazide sodium salt (Molecular Probes) via the recording pipette in order to visualize the dendritic morphology. Only cells with a brushy dendrite and rebound burst firing were identified as UBCs (Borges-Merjane and Trussell, 2015). Recording electrodes (6 – 8 MU) were pulled from borosilicate glass (WPI 1B150F-4) by a horizontal puller (Sutter Instrument P97). Series resistances were usually < 30 MU and were compensated by 50%–60%. For data acquisition we used a Multiclamp 700B amplifier and pClamp 10 (Molecular Device) or Ephus software (Suter et al., 2010) (http://www.ephus.org) running in MATLAB 2007b (Mathworks). Signals were sampled at 40 KHz, low-pass filtered at 10 KHz using a Digidata (1440A, Molecular Devices) analog-digital converter board (for pClamp) or a NI-6229 board (National Instruments). The recording pipette solution contained (in mM): 113 K-gluconate, 9 HEPES, 4.5 MgCl2, 0.1 EGTA, 14 Tris-phosphocreatine, 4 Na2-ATP, 0.3 Tris-GFP, with osmolality adjusted to 290 – 300 mOsm with sucrose and pH adjusted to pH 7.3 with KOH. In some cases, 5 mM QX-314 was added to prevent escaping spikes during electrical stimulation. To isolate AMPAR-mediated currents, NMDAR blocker MK-801 (3 – 15 mM), inhibitory synaptic blockers picrotoxin (100 mM) and strychnine (0.5 mM) were added to the bath solution. Recent studies in the dorsal cochlear nucleus and cerebellum showed that, depending on its response to glutamate, UBCs can be categorized into two groups: an ON cell type which shows an slow inward current mediated by mGluR1 receptors, and an OFF cell type which shows an outward current mediated by mGluR2 receptors (Borges-Merjane and Trussell, 2015). Because both types of UBCs have an AMPAR-mediated slow EPSC and both types of mGluRs may contribute to the synaptic response, we always applied mGluR1 and mGluR2 antagonists to the bath solution (1 mM JNJ16259685 and 1 mM LY-341495, respectively) to isolate AMPAR-mediated responses. Data collected in this study therefore were presumably from both populations of UBCs. In brain slice recordings, cells were voltage-clamped at –70 mV to –80 mV with junction potential correction. Extracellular stimulation of mossy fibers was achieved by applying voltage pulses (10 – 50 V, 100 ms) using a stimulus isolation unit (Iso Flex, A.M.P.I.) via an ACSF-filled monopolar glass electrode (same as the recording electrode) placed near the brush. In some cases a concentric electrode placing in the white matter was used. For dissociated cell recording, the same internal solution was used; however, the recording bath solution was oxygenated Tyrode’s solution (see below) instead of ACSF. Recordings were made at room temperature and the cells were voltage-clamped at –70 mV with junction potential correction. MNI-Glutamate Uncaging The UV-laser uncaging setup was modified from the Laser Scanning Photostimulation System (LSPS) and was controlled by Ephus software (Suter et al., 2010). Voltage-controlled mirror galvanometers (Model 6210; Cambridge Technology) was used to target a 355 nm UV laser beam (3500-SMPS, DPSS Lasers) at the dendritic brush of the recorded UBC, which was filled with Alexa 488 e2 Neuron 96, 1–8.e1–e4, September 27, 2017
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
and could be visualized under the fluorescence microscope through a CCD camera (Ret 2000R, QImaging). Laser power was controlled by Ephus via a Pockels cell (Conoptics) and a neutral density filter wheel (Edmund Optics). The online laser power was measured by a UV-sensitive photodiode (Edmund Optics) detecting a deflected beam in the light path and calibrated for the power measured offline under the objective (40X, NA 0.8, Olympus) by a UV-sensor (PM3Q, Coherent) and power meter (FieldMate, Coherent). Unless otherwise stated, the UV-uncaging power was estimated to be 5 mW under the objective. The duration of uncaging pulse (2 ms) was controlled by Ephus via a mechanical shutter (VMM-D1, Uniblitz). MNI-Glutamate (1 mM; Tocris) was used as caged glutamate compound and bath applied in the recirculating recording ACSF (total volume 30 mL). We confirmed that uncaged glutamate activates the same set of receptors as those reached by synaptic stimulation by testing the spatial resolution of uncaging. A fast-then-slow uncaging-evoked EPSC (uEPSC) was evoked by the flash only when the UV uncaging beam targeted the brush but not the soma or dendrite of the recorded neuron or an area adjacent to the brush (Figure S1, n = 3), indicating that AMPARs are largely restricted to the brush. Second, we tested whether uncaging restricted to the brush activates the same AMPARs as those activated by synaptic stimulation. We applied synaptic stimulation immediately after uncaging, and found that the uncaging pulse markedly desensitized subsequent fast synaptic EPSCs (not shown). The peak of the fast EPSC decreased to 3.9% ± 3.9% (n = 2) or 32.6% ± 9.6% (n = 5) of its original size 15 ms and 200 ms after uncaging, respectively. The particularly profound desensitization 15 ms after uncaging suggests that uncaging activates the same AMPARs as those evoked by synaptic stimulation. These data confirm that MNI-glutamate uncaging can selectively activate synaptic AMPARs when targeted to the brush. Each uncaging pulse led to a peak inward current followed by a trough, where desensitization was maximal, and then a gradually rising and falling slow current. When pulses were applied at different intervals following a conditioning pulse or synaptic stimulus train, the trough was always of similar amplitude, suggesting the amount of glutamate uncaged was also similar. Dissociated Cell Preparation Acute cerebellar slices were collected and then placed into a 31 C oxygenated (100% O2) chamber containing Eagle’s minimal essential medium, 10 mM HEPES (pH 7.4), 40 U of papain (Worthington), 0.5 mM EDTA and 1 mM cysteine (Raman and Trussell, 1992). Slices were incubated in this solution for 20 min, and then washed twice with oxygenated (100% O2) Eagle’s minimal essential medium containing 10 mM HEPES (pH 7.4), 1 mM EDTA, 2 mM cysteine, 1 mg/ml bovine serum albumin, and 1 mg/ml trypsin inhibitor. Lobes IX and X of the cerebellum were dissected using black-enameled insect pins, then washed with oxygenated (100% O2) Tyrode’s solution (150 mM NaCl, 4 mM KCl, 2 mM CaCl2, 2 mM MgCl2,10 mM HEPES, 10 mM glucose, pH 7.4), and cut into smaller chunks by eye (300 mm3). The cerebellar tissue was transferred to poly-D-lysine-coated glass-coverslips and triturated 10-15 times with fine-tipped 8’’-long Pasteur pipettes. The dissociated neurons were allowed to settle undisturbed for 1 hour. Isolated UBCs were recognized by their morphology, and confirmed by expression of GFP when using mGluR2-GFP mice. Immunohistochemistry Dissociated UBC cells were fixed to poly-D-lysine coated glass coverslips with 4% paraformaldehyde for 20 min. The cells were then washed in 0.1M phosphate-buffered saline (PBS) three times, 10 min each wash. The cells were then permeablized and blocked with 0.2% Triton X-100, 2% BSA, and 2% fish gelatin in PBS for 45 min, then washed in PBS three times 10 min each wash. The dissociated cells were then incubated in primary antibody for 1 hr (1:2000 dilution of Chicken IgY Anti-GFP and 1% fish gelatin in PBS) then washed again in PBS three times, 10 min each wash. The dissociated cells were then incubated in secondary antibody for 1 hr (1:500 dilution of Alexa Fluor 488-conjugated AffiniPure Donkey Anti-Chicken IgY and 1% fish gelatin in PBS) then washed again in PBS three times, 10 min each wash. The slides were then mounted to a glass coverslip with Fluoromount-G (Southern Biotech). Fast Application of Glutamate The recording chamber consisted of two-compartments: one for the dissociated cells and a second for rapid application of glutamate. Once a whole-cell patch was achieved both chambers were flooded with Tyrode’s solution and the cell was transported to the application-chamber. Glutamate was applied through a Perfusion Fast-Step System (SF-77B, Warner Instruments) with a glass three square-barrel piping system. Multiple solutions were fed into each square-barrel (700 mm width and height), totaling 10 solutions: one control solution (oxygenated Tyrode’s solution containing synaptic blockers) per barrel and seven concentrations of glutamate (control solution with 0.1 to 1000 mM glutamate). Each solution was gated by a two-way valve, and the control solution was used to wash each barrel between differing glutamate concentrations. During experiments, the dissociated cell was placed about 100-200 mm from the barrel interface and centered in front of a barrel primed with control solution. Flow for all three barrels was started, and the barrel adjacent to the control barrel was switched to a glutamate concentration. Exchange of solution was accomplished by rapid movement of the barrels. The flow from each barrel was approximately 450 ml/min, and solutions were exchanged within 50 ms over the surface of the cell. Pharmacology All drugs unless otherwise noted were purchased from Sigma. All drugs in the slice experiments were bath applied. In dextran-treated experiments, the solution was applied after the rise and decay of the slow EPSC became stable for 7 min. Receptor antagonists
Neuron 96, 1–8.e1–e4, September 27, 2017 e3
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
used this study contained: GYKI-53655 (AMPAR; Tocris), JNJ-16259685 (mGluR1; Tocris), LY-341495 (group II mGluR; Tocris), MK-801 (NMDAR; Sigma), picrotoxin (GABAAR; Sigma), strychnine (GlyR; Sigma). MNI-glutamate was purchased from Tocris. Dextran (MW 35000 – 45000) was purchased from Sigma. QUANTIFICATION AND STATISTICAL ANALYSIS All waveforms were analyzed in IgorPro, Axograph, pClamp or MATLAB with custom written scripts. For analysis in IgorPro, a Neuromatic package (http://www.neuromatic.thinkrandom.com/) was also used. Graphs were made with IgorPro or Prism. For slow EPSC measurements, rise times are defined as the interval between the onset and the peak. Decay time constant was measured by fitting a single-exponential decay function from the peak to 10% of the EPSC amplitude in the falling phase. The change in steady-state current during synaptic stimulation was measured between the 9th and 10th stimulus, relative to the holding current. Statistics were done in IgorPro, MATLAB, R (http://www.R-project.org), or Excel (Microsoft). To test the effect of TBOA on the slow EPSC, measurements were made from Hodgkin-Huxley type fits made to the waveforms in Axograph: A13eðxx0Þ=tau1 3A23ð1 eðxx0Þ=tau2 Þ + constant: The sample size (n) provided in this study refers to the number of neurons recorded. Unless otherwise stated, all data are displayed as mean ± SEM. A two-tailed, one-sample t test was conducted for comparison between the normalized current to a single value. For two-group comparisons, a two-tailed paired or unpaired Student’s t test was conducted if the datasets were normally distributed. The normality test was conducted using Shapiro-Wilk test or P-P plot. For comparison between EPSCs from WT and stg UBCs, a nonparametric permutation test was conducted. Permutation distributions were constructed using randomly shuffled datasets to directly measure the probability that the observed datasets were produced by chance. Data points from both genotypes were pooled and randomly assigned to one of two groups, proportioned as in the actual dataset. The p value was determined using a one-tailed test, as the fraction of 10,000 such randomly shuffled datasets that produced a difference between randomly labeled X and Y groups that was larger than observed in the actual data. In all tests, statistical significance was considered if p < 0.05. In figures, *, **, *** indicates p < 0.05, p < 0.01, and p < 0.001, respectively.
e4 Neuron 96, 1–8.e1–e4, September 27, 2017
Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters Highlights d
Cerebellar unipolar brush cells mediate slow synaptic responses using AMPAR
d
After exocytosis, the slow signals arise as AMPARs recover from desensitization
Authors Hsin-Wei Lu, Timothy S. Balmer, Gabriel E. Romero, Laurence O. Trussell
Correspondence
d
The slow responses require the stargazin accessory subunit for AMPAR
[email protected]
d
Exposure of AMPAR to glutamate is controlled by plasma membrane transporters
Glutamate receptors that normally participate in rapid synaptic signaling can also create ultra-slow signals. Lu et al. show that slow signals require TARP accessory subunits to glutamate receptors as well as tight control of glutamate time course mediated by transporters.
Lu et al., 2017, Neuron 96, 1–8 September 27, 2017 ª 2017 Elsevier Inc. http://dx.doi.org/10.1016/j.neuron.2017.08.043
In Brief
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
Neuron
Report Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters Hsin-Wei Lu,1,4 Timothy S. Balmer,2,4 Gabriel E. Romero,3 and Laurence O. Trussell2,5,* 1Neuroscience
Graduate Program Hearing Research Center and Vollum Institute 3Department of Physiology and Pharmacology Oregon Health and Science University, Portland, OR, USA 4These authors contributed equally 5Lead Contact *Correspondence: [email protected] http://dx.doi.org/10.1016/j.neuron.2017.08.043 2Oregon
SUMMARY
AMPARs mediate the briefest synaptic currents in the brain by virtue of their rapid gating kinetics. However, at the mossy fiber-to-unipolar brush cell synapse in the cerebellum, AMPAR-mediated EPSCs last for hundreds of milliseconds, and it has been proposed that this time course reflects slow diffusion from a complex synaptic space. We show that upon release of glutamate, synaptic AMPARs were desensitized by transmitter by >90%. As glutamate levels subsequently fell, recovery of transmission occurred due to the presence of the AMPAR accessory protein stargazin that enhances the AMPAR response to low levels of transmitter. This gradual increase in receptor activity following desensitization accounted for the majority of synaptic transmission at this synapse. Moreover, the amplitude, duration, and shape of the synaptic response was tightly controlled by plasma membrane glutamate transporters, indicating that clearance of synaptic glutamate during the slow EPSC is dictated by an uptake process. INTRODUCTION AMPA receptors (AMPARs) mediate the majority of fast excitatory postsynaptic currents (EPSCs) in the brain (Jonas, 2000). The brevity of EPSCs and rapid deactivation of AMPARs depend upon a short lifetime of synaptically released glutamate, estimated to be about 1 ms (Clements et al., 1992). Another key factor contributing to fast AMPAR-mediated EPSCs is rapid desensitization, which decreases response amplitudes by >90% within 10 ms upon prolonged exposure to glutamate (Colquhoun et al., 1992; Raman and Trussell, 1992; Silver et al., 1996; Trussell and Fischbach, 1989). Thus, even when cleft glutamate clearance is slow, desensitization still forces AMPARmediated EPSCs to decay quickly (Trussell et al., 1993). In stark contrast to this picture of typical AMPAR synapses is the large mossy fiber-unipolar brush cell (UBC) synapse in the granular layer of cerebellar cortex and cochlear nucleus (Floris
et al., 1994; Rossi et al., 1995). Stimulation of this synapse evokes typical fast EPSCs, but these are followed by a slow, AMPAR-mediated EPSC lasting hundreds of milliseconds (Borges-Merjane and Trussell, 2015; Kinney et al., 1997; Rossi et al., 1995). The mossy fiber-UBC synapse features an extensive, convoluted synaptic cleft between the presynaptic terminal and postsynaptic brush-like dendrite (Rossi et al., 1995). Kinney et al. (1997) proposed that the slow current is the combined result of delayed clearance from this large synaptic cleft and the biophysical properties of AMPARs. Upon such prolonged glutamate exposure, synaptic AMPARs would enter steadystate desensitization and occasionally reopen, generating the slow EPSC. However, direct evidence for ‘‘glutamate entrapment’’ (Rossi et al., 1995) requires information about the kinetic state of receptors during synaptic transmission, their molecular properties, and the forces that determine the glutamate lifetime in the cleft. We tested this hypothesis in UBCs of the vestibular cerebellum. Fast UV uncaging of glutamate after synaptic stimuli revealed that after exocytosis of glutamate, >90% of AMPARs become desensitized. Thereafter, receptors slowly recover from desensitization concurrent with an increase in the EPSC amplitude. Doseresponse relations show that AMPARs produce smaller equilibrium responses to millimolar levels of glutamate than to micromolar levels, suggesting that the slow EPSC tracks recovery from desensitization as glutamate is removed. This decrease in response to high glutamate levels was absent in UBCs from g2 transmembrane AMPAR regulatory protein (TARP) mutant stargazer (stg) mice, and slow EPSCs in these mice were reduced in amplitude. Finally, the slow time course of glutamate was dictated by glutamate transporters, as block of transport profoundly distorted synaptic responses. Thus, at the mossy-fiber-UBC synapse, exocytosis initiates a process of transmission controlled by a balance between glutamate uptake and the heightened sensitivity to transmitter bestowed by TARPs. RESULTS A Slow EPSC Mediated by Synaptic AMPA Receptors The fast-then-slow EPSC sequence that is characteristic of UBCs was evoked by a 100-Hz, 10-stimulus train in mouse cerebellar brain slices (STAR Methods; Figures 1A and 1B) (BorgesNeuron 96, 1–8, September 27, 2017 ª 2017 Elsevier Inc. 1
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
Figure 1. AMPARs Mediate a Slow EPSC at Mossy Fiber-UBC Synapse (A) Top: UBC was filled with Alexa 488 to visualize the location of the brush. Bottom: recording diagram: electrode placed nearby the brush to stimulate a mossy fiber. (B) Left: AMPAR antagonist GYKI-53655 (red) blocked both fast and slow EPSCs evoked by 10 stimuli at 100 Hz. Black, control. Right: this stimulus regime increased spike rate during the slow EPSC. (C) Expanded view of (B). Left: fast EPSCs evoked during train stimulation (triangles) depress profoundly. Right: a slow EPSC occurred after train stimulation stops. (D) Short-term depression kinetics of the fast EPSCs can be fit by a two-exponential decay (red). The peak of slow EPSC occurred 150 ms after end of stimulation. (E) The slow EPSC begins after stimuli stop, regardless of number of stimuli (103–403). Bars indicate duration of 100-Hz stimulation applied to the same cell. (F) Summary of the normalized time to peak from last stimulus of the slow EPSC versus number of stimuli from nine cells. Each symbol represents a different cell. Linear regression fit (red) showed no significant correlation. (G) Application of 5% dextran solution slowed the rise and decay of slow EPSC. Traces are normalized to the peak slow EPSC. (H) Summary of dextran’s effect on decay time constant and rise time of the slow current. Error bars given as ±SEM.
Merjane and Trussell, 2015; van Dorp and De Zeeuw, 2014; Rossi et al., 1995; Zampini et al., 2016). A complete block of the EPSC by GYKI-53655 (GYKI) (75 mM) confirmed that the slow currents were AMPAR mediated (Figure 1B). As previously described, the slow current increases the firing rate of UBCs (Figure 1B; Borges-Merjane and Trussell, 2015; Rossi et al., 1995). In 17 cells tested, amplitudes of fast EPSCs depressed rapidly and nearly completely during train stimulation (fit to decay of peak amplitudes: fast decay t: 9.7 ± 7.7 ms, slow decay t: 18.5 ± 2.0 ms, fast component 55% of total; average of EPSC5–10/EPSC1: 4.3% ± 0.5%; Figures 1C and 1D). Following strong synaptic depression, a slow EPSC emerged after stimuli were terminated, reaching a peak of 13.6 ± 1.5 pA (19.1% ± 5.0% of EPSC1) 130.1 ± 9.0 ms after the train stimulation and decaying with a t of 453.5 ± 53.5 ms back to baseline (Figures 1C and 1E). The slow EPSC generated a charge transfer seven 2 Neuron 96, 1–8, September 27, 2017
times that of all prior fast EPSCs plus the tonic current between each fast EPSC (charge transferred during fast EPSCs: 1.1 ± 0.1 pC, slow EPSC: 7.6 ± 1.1 pC, n = 17). Thus, the slow EPSC was the dominant synaptic signal. The slow EPSC may be a spillover current mediated by extrasynaptic receptors (Nielsen et al., 2004). If this is the case, one would expect the slow EPSC to begin during a long stimulus train as accumulated glutamate spilled out to distant receptors. However, when we varied the number of stimuli, the slowly rising EPSC was observed to begin only after the termination of stimulation (Figure 1E). Furthermore, the current’s rise time after stimuli were halted was constant, regardless of the number of stimuli (range 5–100 stimuli; p = 0.24, R2 = 0.06, linear regression fit; Figure 1F). This result argues against the spillover hypothesis, as such diffusion should not depend upon the cessation of transmitter release. Rather, the results indicate that synaptic AMPARs were partially inhibited by glutamate accumulated in the synaptic cleft, and this inhibition is relieved as transmitter levels gradually decline (Kinney et al., 1997). If indeed the slow EPSC time course arises as glutamate gradually falls, its decay should be prolonged by restricting the diffusion time of glutamate. We thus applied 5% dextran in the bath
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
solution to increase the viscosity of extracellular space and retard transmitter diffusion (Min et al., 1998). After 15 min in dextran, the time to peak and decay time constant of the slow EPSC were prolonged (time to peak: from 135.0 ± 12.0 to 214.3 ± 38.5 ms, p = 0.03, n = 8; decay time constant: from 464.5 ± 74.2 to 688.9 ± 125.7 ms, p = 0.007, n = 8; paired t tests), confirming that the mossy fiber-UBC cleft limits the clearance of transmitter (Figures 1G and 1H). Role of Glutamate Transporters The prolongation of the slow EPSC by dextran indicates a role for diffusion in controlling the clearance of glutamate but does not exclude that glial or neuronal glutamate transporters might also affect glutamate decay. We thus tested the role of transporters using the selective blocker TBOA and found a striking dosedependent effect on the EPSC. At the highest TBOA dose examined (50 mM), the rise time of the slow EPSC was lengthened over 3-fold and the decay time lengthened over 2-fold over control (Figures 2A, 2C, and 2D). By contrast, the fast EPSC elicited by the first stimulus in the train was only slightly slower at the same concentration (Figures 2G–2J). Coincident with these actions, however, was a significant increase in inward holding current even with 10 mM TBOA (Figures 2A and 2E); this current was abolished by subsequent application of GYKI and therefore must represent tonic activation of AMPAR by ambient glutamate (Figure 2A). Given the sensitivity of AMPARs, ambient glutamate may cause tonic AMPAR desensitization. Indeed, we observed an 40% reduction in the peak amplitude of the first fast EPSC in 50 mM TBOA (Figures 2G and 2I). Moreover, there was an apparent interaction between the tonic current and the steady-state current during synaptic activation: the steady-state current was slightly less inward (relative to the baseline current level before stimulation) in 10 mM TBOA compared to control and became clearly outward (relative to baseline) in 50 mM TBOA (Figures 2B and 2F; 6 of 8 cells tested). Following the stimuli, the slow EPSC then emerged, becoming inward again before finally settling back to the baseline defined by the tonic current (Figures 2A and 2B). Subsequent application of GYKI blocked this complex synaptic waveform (Figure 2A). Thus, loss of transporter activity slowed the synaptic response and led to accumulation of ambient glutamate, generating a tonic current. The ongoing exocytosis associated with the train stimuli further desensitized the AMPARs, leading to the apparent outward current. Therefore, a function of the transporters is not only to narrow the EPSC time course, but to limit both glutamate accumulation at rest and receptor desensitization during stimuli. Slow EPSC Mediated by Recovery of AMPARs from Desensitization The delayed onset of the slow EPSC, and the appearance of an outward current when transporters are blocked, strongly suggests that synaptic transmission at the UBC synapse is accompanied by significant desensitization of AMPARs. We used glutamate uncaging to directly observe the magnitude of desensitization and its recovery during the EPSC. MNI-glutamate (1 mM) was bath applied, and a 2-ms UV flash was delivered over the UBC brush to locally uncage glutamate (Figure 3A); this method was combined with synaptic stimulation at varying inter-
vals. As the flash area contains the compact region of the brush, we expect it to activate all synaptic receptors. Control experiments confirmed that AMPARs were largely restricted to the brush region and that uncaged glutamate accessed synaptic receptors (STAR Methods; Figure S1). We found that the peak of the fast transient uncaging response (uEPSC) was greatly reduced when evoked closer to the end of train stimulation (Figures 3B and 3C), with maximal decrease of the test pulse of over 90%. Furthermore, the uEPSC gradually recovered as the synaptic-uncaging stimulus intervals increased, with a t of 267.1 ± 24.4 ms (Figure 3C; n = 6). A similar phenomenon was also observed using a paireduncaging pulse protocol, in which the peak uEPSC evoked by the second uncaging pulse recovered with a t of 413.9 ± 68.1 ms (Figures 3E and 3F; n = 8). Assuming the concentration of uncaged glutamate is similar in each trial (see STAR Methods), the reduction in peak uEPSC is most easily explained by desensitization of AMPARs. As uncaging activates the same set of receptors utilized by synaptic transmission, we conclude that synaptically released glutamate desensitizes AMPARs, and this desensitization is maximal at the end of the last stimulus. The initial rise in the slow EPSC after stimuli cease must be generated by a decline in cleft glutamate occurring simultaneously with recovery from desensitization. This explains two observations. First, an ‘‘undershoot’’ of current was seen right after uncaging, most obvious near the peak of the slow EPSC (Figures 3B and 3E). This undershoot current always reached a level similar to the current level just after the last synaptic stimulus and was followed by its own rise and fall in slow current (Figures 3B and 3E). Thus, cleft glutamate increased upon uncaging, further desensitized the receptors, and then evoked a slow EPSC as glutamate levels subsequently declined. Second, when uncaging pulses were applied at the beginning of the rise of the slow EPSC or at a later point when current had fallen back to the same level, the uncaging responses were distinctly different (Figure 3B, inset). Plotting baseline current just before an uncaging pulse against the fractional recovery of the uncaging response from desensitization revealed a bell-shaped relation, with a maximum at 40% recovery (Figures 3D and 3G). As the slow EPSC peaked when 60% of AMPARs were desensitized, this result suggests that an intermediate glutamate concentration was retained in the synaptic cleft during the slow EPSC. Thus, the slow EPSC was mediated by AMPARs that reopened as they recovered from desensitization. Glutamate Sensitivity of AMPAR and Role of TARPs The ability of AMPARs to produce larger currents in response to falling levels of transmitter implies a molecular specialization for non-monotonic sensitivity to glutamate. TARP co-expression is associated with reduced desensitization and enhanced steadystate AMPAR currents (Jackson and Nicoll, 2011; Tomita, 2010). To test for a role of TARPs in the slow EPSC, we recorded EPSCs in UBCs from stg mice, which lack g2 TARP (Letts et al., 1998), comparing their amplitudes to those of wild-type (WT) mice. EPSCs in stg UBCs showed both fast and slow components upon train stimulation, as in WT UBCs (Figures 4A and 4B). Scatterplots of peak amplitude of fast EPSCs versus slow EPSCs from different UBCs were positively correlated (R2 = 0.163; p = 0.0027, n = 53), presumably as synapses varied in their Neuron 96, 1–8, September 27, 2017 3
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
Figure 2. Transporters Determine Glutamate Time Course (Ai and Aii) Examples of the effect of TBOA on the response to synaptic stimulation. Dashed line indicates holding current in control (0 mM TBOA). TBOA increased the inward holding current, measured immediately before stimulation train (circle). TBOA decreased the steady-state inward current, measured at the end of the train (square) relative to the holding current of the same trace (circle) (i). TBOA increased the time to peak and the decay of the slow EPSC that occurs at the end of the stimulation train (triangle). In 50 mM TBOA (ii), the steady-state current became outward relative to the standing glutamate current. Addition of 50 mM GYKI blocked both standing current and the outward current, indicating that they were due to tonic AMPAR activation and desensitization, respectively. (B) Same traces as in (A), except that they are overlaid with baselines subtracted to show that steady-state current became outward in 50 mM TBOA. (C) TBOA increased the time to peak of the slow EPSC that begins at the offset of the synaptic stimuli. Paired t tests: 0 mM versus 10 mM: p = 7E–5, n = 11; 10 mM versus 50 mM: p = 0.0006, n = 8; 0 mM versus 50 mM: p = 9E–5, n = 8. (D) TBOA increased the decay of the slow EPSC that begins at the offset of the synaptic stimuli (measured from the inward peak to 10% of the baseline. Paired t tests: 0 mM versus 10 mM: p = 0.009, n = 11; 0 mM versus 50 mM: p = 0.018, n = 8. (E) The inward holding current was increased by TBOA and blocked by GYKI. Paired t tests: 0 mM versus 10 mM, p = 0.0003, n = 11; 10 mM versus 50 mM, p = 0.006, n = 8. Illustrated holding current values are relative to the holding current in the previous condition, such that the 10 mM value is relative to the 0 mM, 50 mM is relative to 10 mM, and GYKI is relative to 50 mM TBOA. Paired t test: 50 mM versus GYKI, p = 0.012, n = 6. (F) Change in steady-state current, measured between the 9th and 10th synaptic stimulus, relative to holding current. A positive change in steady state indicates that the current went outward during the stimulus train, as shown in (A) and (B). Paired t tests: 10 mM to 50 mM, p = 0.005, n = 8; 0 mM to 50 mM, p = 0.005, n = 8. Addition of 50 mM GYKI blocked the steady-state current, indicating that the outward current was an AMPAR-mediated current. Paired t test including cells that had an outward steady-state current: 50 mM to GYKI, p = 0.017, n = 6. (G) The first EPSC in the train in (A). (H) The first EPSC in the train peak normalized to illustrate small change in decay rate. (I) TBOA reduces amplitude of the first EPSC in the train. Paired t tests: 10 mM versus 50 mM, p = 0.006, n = 8; 0 mM versus 50 mM, p = 0.008, n = 8. (J) TBOA increases the decay of the first EPSC in the train. Paired t test: 0 mM versus 50 mM: p = 0.037, n = 8. Error bars given as ±SEM.
size, quantal content, and numbers of AMPARs (Zampini et al., 2016). However, when plotted for each mouse line, stg slow currents appeared to be smaller than in WT over a similar range of 4 Neuron 96, 1–8, September 27, 2017
fast EPSCs (Figures 4A and 4B, red versus black traces and points). Given that some of the variance in the population was accounted for by a correlation between EPSC amplitudes, we
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
Figure 3. AMPARs Are Desensitized and Recover during Slow EPSC (A) Experimental configuration. UV-laser spot was targeted to the brush to uncage MNI-glutamate after synaptic stimulation. (B) Uncaging at different times (arrowheads) during slow EPSC. Gray trace shows the control uEPSC. Inset: expanded view showing diminished size of uncaging response (*; black trace) delivered 15 ms after final synaptic stimulation. Blue trace delivered 85 ms later shows recovery even though baseline current is the same as in black trace. (C) Summary of recovery time course shown in (B). Each symbol represents a different cell. The recovery was fit by a single exponential (black line). (D) Steady-state current of the slow EPSC evoked just before the test uncaging pulse plotted against the proportional recovery of test pulse amplitude. Slow EPSC was evoked by synaptic stimulation as in (B). Each color represents a different cell (n = 6). The rising and then falling curves show that the same steady-state current corresponds to two different recovery states, indicating that they were produced by different glutamate levels. (E) Paired-pulse uncaging also showed a desensitized second uEPSC (triangle). Note the undershoot current (arrow) in the second uEPSC. (F) Recovery time course for the second uEPSC in paired uncaging paradigm. Each symbol represents a different cell. Recovery was fit with a single exponential (black line). (G) Steady-state current of the slow current evoked by an uncaging pulse just before a second test uncaging pulse as in (E), plotted against the proportional recovery of test pulse amplitude.
used permutation tests to determine whether WT and stg synapses differed in the amplitudes of slow EPSCs and fast EPSCs and in their slow/fast ratios (see STAR Methods). Slow EPSC amplitudes were 50% smaller in stg UBCs than in WT UBCs (stg median: 5.74 pA, inter-quartile range (IQR): 3.66–6.47 pA, n = 21; WT median: 11.62 pA, IQR: 8.42–17.79 pA, n = 32; permutation test, p < 0.0001). Fast EPSCs were also significantly smaller in stg UBCs, but the difference between medians was smaller than for slow EPSCs (29%), (stg median: 48.33 pA, IQR: 43.5–66.45 pA, n = 21; WT median: 68.47 pA, IQR: 54.51– 90.11 pA, n = 32; permutation test, p = 0.0099). The ratio of the slow/fast EPSCs were used to determine whether the slow EPSCs were generally more affected by the stg mutation than the fast EPSC. Indeed, the slow/fast EPSC ratios were significantly smaller in stg UBCs, indicating that AMPARs in the stg mutant mice have a reduced ability to generate the slow EPSC (stg median: 0.09, IQR: 0.06–0.14, n = 21; WT median: 0.15,
IQR: 0.10–0.23, n = 32; permutation test, p = 0.0133), and suggested a key role for stargazin in boosting sensitivity of the receptors to the slow synaptic glutamate transients at the UBC synapse. A striking effect of heterologous TARP expression on AMPAR function is a bellshaped glutamate dose-response relation, termed ‘‘auto-inactivation’’ (Morimoto-Tomita et al., 2009; Semenov et al., 2012), a feature described in some native AMPARs (Geoffroy et al., 1991; Raman and Trussell, 1992). Given that the slow EPSC rises as cleft glutamate decays, and that TARPs support this slow current, we expected that WT, but not stg, UBCs might feature a non-monotonic sensitivity to glutamate. UBCs enzymatically isolated with intact brush dendrites were used for constructing dose-response relations (Figure 4C), as glutamate transporters in slices may limit the access of exogenous glutamate to receptors. As shown in Figures 4D–4F, when 1 mM glutamate was applied to the cell, a steady current was generated without apparent desensitization. A fast transient current could be observed at higher glutamate levels (32–1,000 mM), which quickly desensitized and reached steady state. This steady-state response is presumed to originate from bound AMPARs at microscopic equilibrium between open, closed, and desensitized states. Absolute and normalized Neuron 96, 1–8, September 27, 2017 5
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
Figure 4. Stargazer UBCs Showed Reduced Slow EPSC and Monotonic Dose-Response Relation (A) Overlay of EPSCs from a WT (black) and a stg UBC (red). The peak of the fast EPSC is scaled to the same level and each trace represents the median slow/fast ratio of each group. Added filtering was applied to the slow EPSC. Inset: expanded view of slow EPSC. (B) Plot of slow versus fast EPSC amplitudes from 32 WT and 21 stg UBCs. The peak amplitudes of slow, but not fast, EPSC in stg UBCs were significantly smaller compared to WT. (C) A GFP-labeled dissociated UBC with intact dendritic brush. (D) AMPAR-mediated currents evoked by glutamate (gray bar) at the indicated concentrations applied to dissociated WT and stg UBCs. (E and F) Comparison of absolute (E) and normalized (F) amplitudes of steady-state AMPAR currents in WT and stg UBCs. Error bars given as ±SEM.
reduced slow EPSC reflects an inability of the AMPARs to give a rebound response as synaptic glutamate decays. DISCUSSION
steady-state responses at different glutamate concentrations are shown in Figures 4E and 4F. The steady-state response increased with glutamate concentration up to 32 mM (156.1% ± 0.1% of the 1 mM response, n = 9), and mean absolute values for currents at 32 mM were significantly larger than at 1 mM (n = 9, p = 0.004, paired t test). These experiments were then repeated on stg UBCs and revealed clear differences with WT UBCs. Maximal steady-state current was about half that of WT (Figure 4E; e.g., at 32 mM, WT: 11.5 ± 1.6 pA, n = 9; stg: 4.8 ± 0.7 pA, n = 7, p = 0.004, unpaired t test). These values at 32 mM were comparable to the peak amplitude of the slow EPSC within the same genotype (WT: p = 0.95; stg: p = 0.30, unpaired t tests). Notably, the shape of the stg dose-response relation was monotonic (Figure 4F), such that responses at 32 mM, normalized to responses for 1 mM, were not significantly different from 1 (p = 0.83, one-sample t test). Overall, in comparing the normalized curves between WT and stg, significant differences were observed for responses to 1, 10, 32, and 100 mM glutamate (p = 0.04–0.004). Together, these data indicate that stg mice have reduced sensitivity to glutamate, particularly in the range of 1–100 mM that would normally give rise to the non-monotonic dose-response curve, suggesting that the 6 Neuron 96, 1–8, September 27, 2017
This study showed that the slow EPSC of UBCs is determined by two factors. Continual transporter activity controls the time course of the EPSC, the depth of desensitization, and the baseline level of glutamate in the synaptic cleft. Moreover, a TARP-containing AMPAR renders the cell sensitive to the slowly changing levels of glutamate established by the transporters. In the absence of transport, glutamate levels fall extremely slowly and never reach a level below the sensitivity of the AMPAR, highlighting the ‘‘entrapment’’ hypothesis originally proposed by Rossi et al. (1995). This situation is dramatically different from that of mossy fiber synapses made onto granule cells or of fenestrated calyceal synapses, both of which have exclusively fast phasic AMPAR-mediated transmission even after transporter blockade (Renden et al., 2005; Sylantyev et al., 2013); thus, entrapment, transporter action, and the AMPAR-TARP complex allow the UBC to transmit in a nearly tonic mode despite sensing phasic exocytosis. Desensitized AMPARs during Synaptic Transmission The remarkable speed of AMPAR desensitization (Kiskin et al., 1986; Tang et al., 1989) raised the possibility that desensitization could determine the decay time of the EPSC (Trussell and Fischbach, 1989; Trussell et al., 1993), or impact the amplitude of EPSCs during repetitive activity (DiGregorio et al., 2007; Trussell et al., 1993), depending upon ease of clearance of glutamate and the rate of recovery from desensitized states. This study demonstrates a very different role for desensitization. In UBCs, recovery
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
from desensitization of TARP-containing AMPARs during slowly changing levels of transmitter leads to sizeable currents and accounts for a large fraction of transmitter action. Moreover, the balance between open and desensitized states is regulated by glutamate transporters, such that loss of transport reduces inward excitatory current during synaptic stimuli by enhancing desensitization. In this way, the UBC synapse can more effectively integrate presynaptic activity over longer periods of time than would be obtained with fast EPSCs. Role of Stargazin in Synaptic Transmission The 50% reduction of slow EPSC in stg UBCs suggests that stargazin is present in WT UBCs and regulates glutamate sensitivity of AMPARs. Interestingly, cerebellar granule cells also express stargazin (Hashimoto et al., 1999; Yamazaki et al., 2010), and stg granule cells lack functional AMPARs at mossy fiber synapses because of impaired AMPAR trafficking (Chen et al., 2000; Hashimoto et al., 1999). Stargazin regulates AMPAR gating, slowing down deactivation and desensitization rates (Priel et al., 2005; Tomita et al., 2005; Turetsky et al., 2005), and miniature EPSCs (Tomita et al., 2005). Stargazin also enhanced the equilibrium AMPAR current and generated non-monotonic dose-response curves (Morimoto-Tomita et al., 2009; Semenov et al., 2012). The physiological significance of the non-monotonic dose-response relation was previously obscure because slow glutamate transients typically do not occur at fast synapses. Our observations at UBC synapses reveal that stargazin also regulates synaptic transmission by allowing the synapse to respond to slowly changing, low levels of glutamate. STAR+METHODS Detailed methods are provided in the online version of this paper and include the following: d d d d
d
KEY RESOURCES TABLE CONTACT FOR REAGENT AND RESOURCE SHARING EXPERIMENTAL MODEL AND SUBJECT DETAILS B Animals and Genotyping METHOD DETAILS B Brain Slice Preparation B Electrophysiology B MNI-Glutamate Uncaging B Dissociated Cell Preparation B Immunohistochemistry B Fast Application of Glutamate B Pharmacology QUANTIFICATION AND STATISTICAL ANALYSIS
SUPPLEMENTAL INFORMATION Supplemental Information includes one figure and can be found with this article online at http://dx.doi.org/10.1016/j.neuron.2017.08.043.
ACKNOWLEDGMENTS This study is funded by National Institutes of Health (NIH) Grants NS028901 and DC004450 (to L.O.T.); N.L. Tartar Trust Fellowship (to H.-W.L.); DC014878 to T.S.B. We thank members of the Trussell lab for helpful discussions, Dr. Stephen David for assistance with statistical analysis, and Dr. Craig Jahr and Dr. Brett Carter for comments on the manuscript. We thank Michael Bateschell and Ruby Larisch for help with mouse colony management. Stargazer breeder mice were kindly provided by the Puthussery lab at the Casey Eye Institute. Received: August 9, 2016 Revised: July 25, 2017 Accepted: August 28, 2017 Published: September 14, 2017 REFERENCES Borges-Merjane, C., and Trussell, L.O. (2015). ON and OFF unipolar brush cells transform multisensory inputs to the auditory system. Neuron 85, 1029–1042. Chen, L., Chetkovich, D.M., Petralia, R.S., Sweeney, N.T., Kawasaki, Y., Wenthold, R.J., Bredt, D.S., and Nicoll, R.A. (2000). Stargazin regulates synaptic targeting of AMPA receptors by two distinct mechanisms. Nature 408, 936–943. Clements, J.D., Lester, R.A., Tong, G., Jahr, C.E., and Westbrook, G.L. (1992). The time course of glutamate in the synaptic cleft. Science 258, 1498–1501. Colquhoun, D., Jonas, P., and Sakmann, B. (1992). Action of brief pulses of glutamate on AMPA/kainate receptors in patches from different neurones of rat hippocampal slices. J. Physiol. 458, 261–287. DiGregorio, D.A., Rothman, J.S., Nielsen, T.A., and Silver, R.A. (2007). Desensitization properties of AMPA receptors at the cerebellar mossy fiber granule cell synapse. J. Neurosci. 27, 8344–8357. Floris, A., Din˜o, M., Jacobowitz, D.M., and Mugnaini, E. (1994). The unipolar brush cells of the rat cerebellar cortex and cochlear nucleus are calretinin-positive: a study by light and electron microscopic immunocytochemistry. Anat. Embryol. (Berl.) 189, 495–520. Geoffroy, M., Lambolez, B., Audinat, E., Hamon, B., Crepel, F., Rossier, J., and Kado, R.T. (1991). Reduction of desensitization of a glutamate ionotropic receptor by antagonists. Mol. Pharmacol. 39, 587–591. Hashimoto, K., Fukaya, M., Qiao, X., Sakimura, K., Watanabe, M., and Kano, M. (1999). Impairment of AMPA receptor function in cerebellar granule cells of ataxic mutant mouse stargazer. J. Neurosci. 19, 6027–6036. Jackson, A.C., and Nicoll, R.A. (2011). Stargazin (TARP g-2) is required for compartment-specific AMPA receptor trafficking and synaptic plasticity in cerebellar stellate cells. J. Neurosci. 31, 3939–3952. Jonas, P. (2000). The time course of signaling at central glutamatergic synapses. News Physiol. Sci. 15, 83–89. Kinney, G.A., Overstreet, L.S., and Slater, N.T. (1997). Prolonged physiological entrapment of glutamate in the synaptic cleft of cerebellar unipolar brush cells. J. Neurophysiol. 78, 1320–1333. Kiskin, N.I., Krishtal, O.A., and Tsyndrenko, A.Y.A. (1986). Excitatory amino acid receptors in hippocampal neurons: kainate fails to desensitize them. Neurosci. Lett. 63, 225–230. Letts, V.A., Felix, R., Biddlecome, G.H., Arikkath, J., Mahaffey, C.L., Valenzuela, A., Bartlett, F.S., 2nd, Mori, Y., Campbell, K.P., and Frankel, W.N. (1998). The mouse stargazer gene encodes a neuronal Ca2+-channel g subunit. Nat. Genet. 19, 340–347.
AUTHOR CONTRIBUTIONS
Min, M.-Y., Rusakov, D.A., and Kullmann, D.M. (1998). Activation of AMPA, kainate, and metabotropic receptors at hippocampal mossy fiber synapses: role of glutamate diffusion. Neuron 21, 561–570.
H.-W.L., T.S.B., and G.E.R. performed the experiments and analyzed data. H.-W.L., T.S.B., G.E.R., and L.O.T. designed experiments and wrote the manuscript.
Morimoto-Tomita, M., Zhang, W., Straub, C., Cho, C.-H., Kim, K.S., Howe, J.R., and Tomita, S. (2009). Autoinactivation of neuronal AMPA receptors via glutamate-regulated TARP interaction. Neuron 61, 101–112.
Neuron 96, 1–8, September 27, 2017 7
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
Morin, F., Din˜o, M.R., and Mugnaini, E. (2001). Postnatal differentiation of unipolar brush cells and mossy fiber-unipolar brush cell synapses in rat cerebellum. Neuroscience 104, 1127–1139.
Tang, C.M., Dichter, M., and Morad, M. (1989). Quisqualate activates a rapidly inactivating high conductance ionic channel in hippocampal neurons. Science 243, 1474–1477.
Nielsen, T.A., DiGregorio, D.A., and Silver, R.A. (2004). Modulation of glutamate mobility reveals the mechanism underlying slow-rising AMPAR EPSCs and the diffusion coefficient in the synaptic cleft. Neuron 42, 757–771.
Tomita, S. (2010). Regulation of ionotropic glutamate receptors by their auxiliary subunits. Physiology (Bethesda) 25, 41–49.
Priel, A., Kolleker, A., Ayalon, G., Gillor, M., Osten, P., and Stern-Bach, Y. (2005). Stargazin reduces desensitization and slows deactivation of the AMPA-type glutamate receptors. J. Neurosci. 25, 2682–2686.
Tomita, S., Adesnik, H., Sekiguchi, M., Zhang, W., Wada, K., Howe, J.R., Nicoll, R.A., and Bredt, D.S. (2005). Stargazin modulates AMPA receptor gating and trafficking by distinct domains. Nature 435, 1052–1058.
Raman, I.M., and Trussell, L.O. (1992). The kinetics of the response to glutamate and kainate in neurons of the avian cochlear nucleus. Neuron 9, 173–186.
Trussell, L.O., and Fischbach, G.D. (1989). Glutamate receptor desensitization and its role in synaptic transmission. Neuron 3, 209–218.
Renden, R., Taschenberger, H., Puente, N., Rusakov, D.A., Duvoisin, R., Wang, L.-Y., Lehre, K.P., and von Gersdorff, H. (2005). Glutamate transporter studies reveal the pruning of metabotropic glutamate receptors and absence of AMPA receptor desensitization at mature calyx of held synapses. J. Neurosci. 25, 8482–8497.
Trussell, L.O., Zhang, S., and Raman, I.M. (1993). Desensitization of AMPA receptors upon multiquantal neurotransmitter release. Neuron 10, 1185–1196.
Rossi, D.J., Alford, S., Mugnaini, E., and Slater, N.T. (1995). Properties of transmission at a giant glutamatergic synapse in cerebellum: the mossy fiber-unipolar brush cell synapse. J. Neurophysiol. 74, 24–42. €nen, K. Semenov, A., Mo¨ykkynen, T., Coleman, S.K., Korpi, E.R., and Keina (2012). Autoinactivation of the stargazin-AMPA receptor complex: subunit-dependency and independence from physical dissociation. PLoS ONE 7, e49282. Silver, R.A., Colquhoun, D., Cull-Candy, S.G., and Edmonds, B. (1996). Deactivation and desensitization of non-NMDA receptors in patches and the time course of EPSCs in rat cerebellar granule cells. J. Physiol. 493, 167–173.
Turetsky, D., Garringer, E., and Patneau, D.K. (2005). Stargazin modulates native AMPA receptor functional properties by two distinct mechanisms. J. Neurosci. 25, 7438–7448. van Dorp, S., and De Zeeuw, C.I. (2014). Variable timing of synaptic transmission in cerebellar unipolar brush cells. Proc. Natl. Acad. Sci. USA 111, 5403–5408. Watanabe, D., Inokawa, H., Hashimoto, K., Suzuki, N., Kano, M., Shigemoto, R., Hirano, T., Toyama, K., Kaneko, S., Yokoi, M., et al. (1998). Ablation of cerebellar Golgi cells disrupts synaptic integration involving GABA inhibition and NMDA receptor activation in motor coordination. Cell 95, 17–27.
Suter, B.A., O’Connor, T., Iyer, V., Petreanu, L.T., Hooks, B.M., Kiritani, T., Svoboda, K., and Shepherd, G.M.G. (2010). Ephus: multipurpose data acquisition software for neuroscience experiments. Front. Neural Circuits 4, 100.
Yamazaki, M., Fukaya, M., Hashimoto, K., Yamasaki, M., Tsujita, M., Itakura, M., Abe, M., Natsume, R., Takahashi, M., Kano, M., et al. (2010). TARPs g-2 and g-7 are essential for AMPA receptor expression in the cerebellum. Eur. J. Neurosci. 31, 2204–2220.
Sylantyev, S., Savtchenko, L.P., Ermolyuk, Y., Michaluk, P., and Rusakov, D.A. (2013). Spike-driven glutamate electrodiffusion triggers synaptic potentiation via a Homer-dependent mGluR-NMDAR link. Neuron 77, 528–541.
Zampini, V., Liu, J.K., Diana, M.A., Maldonado, P.P., Brunel, N., and Dieudonne´, S. (2016). Mechanisms and functional roles of glutamatergic synapse diversity in a cerebellar circuit. eLife 5, e15872.
8 Neuron 96, 1–8, September 27, 2017
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
STAR+METHODS KEY RESOURCES TABLE
REAGENT or RESOURCE
SOURCE
IDENTIFIER
Chicken IgY Anti-GFP
Aves
Cat # GFP-1020 RRID: AB_10000240
Alexa 488-conjugated AffiniPure Donkey Anti-Chicken IgY
Jackson ImmunoResearch
Cat # 703-545-155 RRID: AB_2340375
GYKI-53655
Tocris
Cat # 2555
JNJ-16259685
Tocris
Cat # 2333
LY-341495
Tocris
Cat # 1209
MNI-caged-L-glutamate
Tocris
Cat # 1490
(+)-MK-801 hydrogen maleate
Sigma
Cat # M107
Picrotoxin
Sigma
Cat # P1675
Strychnine hydrochloride
Sigma
Cat # S8753
Dextran (MW 35,000–45,000)
Sigma
Cat # D1662
Alexa 488 hydrazide sodium salt
Thermo Fisher Scientific
Cat # A10436
DL-TBOA
Tocris
Cat # 1223
Papain, Suspension
Worthington
Cat # LS003124
Minimum Essential Medium Eagle, no L-glut
Sigma
Cat # M0275
Trypsin Inhibitor, Ovomucoid
Worthington
Cat # LS003085
Mouse: C57BL/6J
Jackson Laboratory
RRID: IMSR_JAX:000664
Mouse: C57BL/6J-TgN(grm2-IL2RA/GFP)1kyo (mGluR2-GFP)
Gift from Robert Duvoisin
Watanabe et al., 1998
Mouse: B6C3Fe a/a-Cacng2stg/J (Stargazer)
Gift from Teresa Puthussery
RRID: IMSR_JAX:001756
Wavemetrics
RRID: SCR_000325
Antibodies
Chemicals, Peptides, and Recombinant Proteins
Experimental Models: Organisms/Strains
Software and Algorithms Igor Pro MATLAB
MathWorks
RRID: SCR_001622
Ephus
Suter et al., 2010
http://www.ephus.org
Excel
Microsoft
N/A
pClamp
Molecular Devices
RRID: SCR_011323
Prism 7
GraphPad
RRID: SCR_002798
Neuromatic
Jason Rothman
RRID: SCR_004186
Axograph
Axograph
RRID: SCR_014284
CONTACT FOR REAGENT AND RESOURCE SHARING Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact Laurence Trussell ([email protected]). EXPERIMENTAL MODEL AND SUBJECT DETAILS Animals and Genotyping Three mouse lines were used in this study. Wild-type C57BL/6J or transgenic C57BL/6J-TgN(grm2-IL2RA/GFP)1kyo (referred to as mGluR2-GFP) mouse line were used as control animals. In the latter mouse line, GFP is tagged to human interleukin-2 receptor a subunit with expression driven by the metabotropic glutamate receptor 2 (mGluR2) promoter (Watanabe et al., 1998). Since
Neuron 96, 1–8.e1–e4, September 27, 2017 e1
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
UBCs express mGluR2 (Borges-Merjane and Trussell, 2015), this mouse line was used for identifying UBCs in the brain slice or dissociated cell preparations, although in later experiments UBCs could be recognized with transmitted light, and confirmed by dye fills. For simplicity, these two lines were referred to as ‘‘wildtype’’ (wt) in this study. A third mouse line, B6C3Fe a/a-Cacng2stg/J (stargazer or stg) (Hashimoto et al., 1999), was used for experiments to measure the effect of TARP g2 (stargazin) mutant. This mouse line has a retroviral-like, early transposon insertion in the second intron of stargazin gene, resulting in a significant reduction of protein expression (Letts et al., 1998). Stg pups were produced by crossing heterozygous males with either heterozygous or homozygous females. The pups were genotyped according to the protocol described on Jackson Lab website (https://www.jax.org/strain/001756). Stg pups also expressed ataxic phenotype as described in previous studies (Hashimoto et al., 1999). Breeding of all three mouse lines were maintained in the animal facility managed by the Department of Comparative Medicine and all procedures were approved by the Oregon Health and Science University’s Institutional Animal Care and Use Committee. Pups from both sexes were used. Because synapse formation between mossy fiber and UBC is mature in animals older than postnatal day 21 (P21) (Morin et al., 2001), we only used pups older than this age (P21 – P39) for experiments. METHOD DETAILS Brain Slice Preparation Animals were anesthetized with isoflurane, decapitated, and the cerebellum was dissected from the skull under ice-cold high-sucrose artificial cerebral spinal fluid solution (ACSF) containing (in mM): 87 NaCl, 75 sucrose, 25 NaHCO3, 25 glucose, 2.5 KCl, 1.25 NaH2PO4, 0.5 CaCl2, 7 MgCl2, bubbled with 5% CO2/95% O2. Sagittal cerebellum sections containing lobe X were cut at 300 mm with a vibratome (VT1200S, Leica) in ice-cold high-sucrose ACSF. After cutting, slices were allowed to recover at 35 C for 30-40 min in ACSF containing (in mM): 130 NaCl, 2.1 KCl, 1.2 KH2PO4, 3 Na-HEPES, 10 glucose, 20 NaHCO3, 2 Na-pyruvate, 2 CaCl2, 1 MgSO4, and 3 myo-inositol, bubbled with 5% CO2/95% O2 (Borges-Merjane and Trussell, 2015). After recovery, slices were kept at room temperature (22 C) in ACSF until use. Electrophysiology For whole-cell recording in brain slices, slices containing cerebellar lobe X were transferred to a recording chamber perfused with warm ACSF (34 C) controlled by a peristaltic pump. Cells were visualized with a 40X magnification objective on the stage of an upright microscope equipped with an infrared Dodt contrast mask and a blue LED for fluorescence optics. In slices from mGluR2-GFP mouse line, UBCs were identified as GFP-positive neurons with brushy dendritic structure. In brain slices from wildtype or stargazer mouse line, UBCs were first identified as round cells with soma diameter 10 mm in the cerebellar granular layer. Only data from UBCs in the lobe X were used for analysis in this study. All cells recorded were filled with 10 mM Alexa Fluor 488 hydrazide sodium salt (Molecular Probes) via the recording pipette in order to visualize the dendritic morphology. Only cells with a brushy dendrite and rebound burst firing were identified as UBCs (Borges-Merjane and Trussell, 2015). Recording electrodes (6 – 8 MU) were pulled from borosilicate glass (WPI 1B150F-4) by a horizontal puller (Sutter Instrument P97). Series resistances were usually < 30 MU and were compensated by 50%–60%. For data acquisition we used a Multiclamp 700B amplifier and pClamp 10 (Molecular Device) or Ephus software (Suter et al., 2010) (http://www.ephus.org) running in MATLAB 2007b (Mathworks). Signals were sampled at 40 KHz, low-pass filtered at 10 KHz using a Digidata (1440A, Molecular Devices) analog-digital converter board (for pClamp) or a NI-6229 board (National Instruments). The recording pipette solution contained (in mM): 113 K-gluconate, 9 HEPES, 4.5 MgCl2, 0.1 EGTA, 14 Tris-phosphocreatine, 4 Na2-ATP, 0.3 Tris-GFP, with osmolality adjusted to 290 – 300 mOsm with sucrose and pH adjusted to pH 7.3 with KOH. In some cases, 5 mM QX-314 was added to prevent escaping spikes during electrical stimulation. To isolate AMPAR-mediated currents, NMDAR blocker MK-801 (3 – 15 mM), inhibitory synaptic blockers picrotoxin (100 mM) and strychnine (0.5 mM) were added to the bath solution. Recent studies in the dorsal cochlear nucleus and cerebellum showed that, depending on its response to glutamate, UBCs can be categorized into two groups: an ON cell type which shows an slow inward current mediated by mGluR1 receptors, and an OFF cell type which shows an outward current mediated by mGluR2 receptors (Borges-Merjane and Trussell, 2015). Because both types of UBCs have an AMPAR-mediated slow EPSC and both types of mGluRs may contribute to the synaptic response, we always applied mGluR1 and mGluR2 antagonists to the bath solution (1 mM JNJ16259685 and 1 mM LY-341495, respectively) to isolate AMPAR-mediated responses. Data collected in this study therefore were presumably from both populations of UBCs. In brain slice recordings, cells were voltage-clamped at –70 mV to –80 mV with junction potential correction. Extracellular stimulation of mossy fibers was achieved by applying voltage pulses (10 – 50 V, 100 ms) using a stimulus isolation unit (Iso Flex, A.M.P.I.) via an ACSF-filled monopolar glass electrode (same as the recording electrode) placed near the brush. In some cases a concentric electrode placing in the white matter was used. For dissociated cell recording, the same internal solution was used; however, the recording bath solution was oxygenated Tyrode’s solution (see below) instead of ACSF. Recordings were made at room temperature and the cells were voltage-clamped at –70 mV with junction potential correction. MNI-Glutamate Uncaging The UV-laser uncaging setup was modified from the Laser Scanning Photostimulation System (LSPS) and was controlled by Ephus software (Suter et al., 2010). Voltage-controlled mirror galvanometers (Model 6210; Cambridge Technology) was used to target a 355 nm UV laser beam (3500-SMPS, DPSS Lasers) at the dendritic brush of the recorded UBC, which was filled with Alexa 488 e2 Neuron 96, 1–8.e1–e4, September 27, 2017
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
and could be visualized under the fluorescence microscope through a CCD camera (Ret 2000R, QImaging). Laser power was controlled by Ephus via a Pockels cell (Conoptics) and a neutral density filter wheel (Edmund Optics). The online laser power was measured by a UV-sensitive photodiode (Edmund Optics) detecting a deflected beam in the light path and calibrated for the power measured offline under the objective (40X, NA 0.8, Olympus) by a UV-sensor (PM3Q, Coherent) and power meter (FieldMate, Coherent). Unless otherwise stated, the UV-uncaging power was estimated to be 5 mW under the objective. The duration of uncaging pulse (2 ms) was controlled by Ephus via a mechanical shutter (VMM-D1, Uniblitz). MNI-Glutamate (1 mM; Tocris) was used as caged glutamate compound and bath applied in the recirculating recording ACSF (total volume 30 mL). We confirmed that uncaged glutamate activates the same set of receptors as those reached by synaptic stimulation by testing the spatial resolution of uncaging. A fast-then-slow uncaging-evoked EPSC (uEPSC) was evoked by the flash only when the UV uncaging beam targeted the brush but not the soma or dendrite of the recorded neuron or an area adjacent to the brush (Figure S1, n = 3), indicating that AMPARs are largely restricted to the brush. Second, we tested whether uncaging restricted to the brush activates the same AMPARs as those activated by synaptic stimulation. We applied synaptic stimulation immediately after uncaging, and found that the uncaging pulse markedly desensitized subsequent fast synaptic EPSCs (not shown). The peak of the fast EPSC decreased to 3.9% ± 3.9% (n = 2) or 32.6% ± 9.6% (n = 5) of its original size 15 ms and 200 ms after uncaging, respectively. The particularly profound desensitization 15 ms after uncaging suggests that uncaging activates the same AMPARs as those evoked by synaptic stimulation. These data confirm that MNI-glutamate uncaging can selectively activate synaptic AMPARs when targeted to the brush. Each uncaging pulse led to a peak inward current followed by a trough, where desensitization was maximal, and then a gradually rising and falling slow current. When pulses were applied at different intervals following a conditioning pulse or synaptic stimulus train, the trough was always of similar amplitude, suggesting the amount of glutamate uncaged was also similar. Dissociated Cell Preparation Acute cerebellar slices were collected and then placed into a 31 C oxygenated (100% O2) chamber containing Eagle’s minimal essential medium, 10 mM HEPES (pH 7.4), 40 U of papain (Worthington), 0.5 mM EDTA and 1 mM cysteine (Raman and Trussell, 1992). Slices were incubated in this solution for 20 min, and then washed twice with oxygenated (100% O2) Eagle’s minimal essential medium containing 10 mM HEPES (pH 7.4), 1 mM EDTA, 2 mM cysteine, 1 mg/ml bovine serum albumin, and 1 mg/ml trypsin inhibitor. Lobes IX and X of the cerebellum were dissected using black-enameled insect pins, then washed with oxygenated (100% O2) Tyrode’s solution (150 mM NaCl, 4 mM KCl, 2 mM CaCl2, 2 mM MgCl2,10 mM HEPES, 10 mM glucose, pH 7.4), and cut into smaller chunks by eye (300 mm3). The cerebellar tissue was transferred to poly-D-lysine-coated glass-coverslips and triturated 10-15 times with fine-tipped 8’’-long Pasteur pipettes. The dissociated neurons were allowed to settle undisturbed for 1 hour. Isolated UBCs were recognized by their morphology, and confirmed by expression of GFP when using mGluR2-GFP mice. Immunohistochemistry Dissociated UBC cells were fixed to poly-D-lysine coated glass coverslips with 4% paraformaldehyde for 20 min. The cells were then washed in 0.1M phosphate-buffered saline (PBS) three times, 10 min each wash. The cells were then permeablized and blocked with 0.2% Triton X-100, 2% BSA, and 2% fish gelatin in PBS for 45 min, then washed in PBS three times 10 min each wash. The dissociated cells were then incubated in primary antibody for 1 hr (1:2000 dilution of Chicken IgY Anti-GFP and 1% fish gelatin in PBS) then washed again in PBS three times, 10 min each wash. The dissociated cells were then incubated in secondary antibody for 1 hr (1:500 dilution of Alexa Fluor 488-conjugated AffiniPure Donkey Anti-Chicken IgY and 1% fish gelatin in PBS) then washed again in PBS three times, 10 min each wash. The slides were then mounted to a glass coverslip with Fluoromount-G (Southern Biotech). Fast Application of Glutamate The recording chamber consisted of two-compartments: one for the dissociated cells and a second for rapid application of glutamate. Once a whole-cell patch was achieved both chambers were flooded with Tyrode’s solution and the cell was transported to the application-chamber. Glutamate was applied through a Perfusion Fast-Step System (SF-77B, Warner Instruments) with a glass three square-barrel piping system. Multiple solutions were fed into each square-barrel (700 mm width and height), totaling 10 solutions: one control solution (oxygenated Tyrode’s solution containing synaptic blockers) per barrel and seven concentrations of glutamate (control solution with 0.1 to 1000 mM glutamate). Each solution was gated by a two-way valve, and the control solution was used to wash each barrel between differing glutamate concentrations. During experiments, the dissociated cell was placed about 100-200 mm from the barrel interface and centered in front of a barrel primed with control solution. Flow for all three barrels was started, and the barrel adjacent to the control barrel was switched to a glutamate concentration. Exchange of solution was accomplished by rapid movement of the barrels. The flow from each barrel was approximately 450 ml/min, and solutions were exchanged within 50 ms over the surface of the cell. Pharmacology All drugs unless otherwise noted were purchased from Sigma. All drugs in the slice experiments were bath applied. In dextran-treated experiments, the solution was applied after the rise and decay of the slow EPSC became stable for 7 min. Receptor antagonists
Neuron 96, 1–8.e1–e4, September 27, 2017 e3
Please cite this article in press as: Lu et al., Slow AMPAR Synaptic Transmission Is Determined by Stargazin and Glutamate Transporters, Neuron (2017), http://dx.doi.org/10.1016/j.neuron.2017.08.043
used this study contained: GYKI-53655 (AMPAR; Tocris), JNJ-16259685 (mGluR1; Tocris), LY-341495 (group II mGluR; Tocris), MK-801 (NMDAR; Sigma), picrotoxin (GABAAR; Sigma), strychnine (GlyR; Sigma). MNI-glutamate was purchased from Tocris. Dextran (MW 35000 – 45000) was purchased from Sigma. QUANTIFICATION AND STATISTICAL ANALYSIS All waveforms were analyzed in IgorPro, Axograph, pClamp or MATLAB with custom written scripts. For analysis in IgorPro, a Neuromatic package (http://www.neuromatic.thinkrandom.com/) was also used. Graphs were made with IgorPro or Prism. For slow EPSC measurements, rise times are defined as the interval between the onset and the peak. Decay time constant was measured by fitting a single-exponential decay function from the peak to 10% of the EPSC amplitude in the falling phase. The change in steady-state current during synaptic stimulation was measured between the 9th and 10th stimulus, relative to the holding current. Statistics were done in IgorPro, MATLAB, R (http://www.R-project.org), or Excel (Microsoft). To test the effect of TBOA on the slow EPSC, measurements were made from Hodgkin-Huxley type fits made to the waveforms in Axograph: A13eðxx0Þ=tau1 3A23ð1 eðxx0Þ=tau2 Þ + constant: The sample size (n) provided in this study refers to the number of neurons recorded. Unless otherwise stated, all data are displayed as mean ± SEM. A two-tailed, one-sample t test was conducted for comparison between the normalized current to a single value. For two-group comparisons, a two-tailed paired or unpaired Student’s t test was conducted if the datasets were normally distributed. The normality test was conducted using Shapiro-Wilk test or P-P plot. For comparison between EPSCs from WT and stg UBCs, a nonparametric permutation test was conducted. Permutation distributions were constructed using randomly shuffled datasets to directly measure the probability that the observed datasets were produced by chance. Data points from both genotypes were pooled and randomly assigned to one of two groups, proportioned as in the actual dataset. The p value was determined using a one-tailed test, as the fraction of 10,000 such randomly shuffled datasets that produced a difference between randomly labeled X and Y groups that was larger than observed in the actual data. In all tests, statistical significance was considered if p < 0.05. In figures, *, **, *** indicates p < 0.05, p < 0.01, and p < 0.001, respectively.
e4 Neuron 96, 1–8.e1–e4, September 27, 2017