Neuron, Vol. 32, 89–98, October 11, 2001, Copyright 2001 by Cell Press

T-Type Ca2� Channels Mediate NeurotransmitterRelease in Retinal Bipolar Cells

Bipolar cells in a variety of species have been reportedto express both LVA and L-type HVA Ca2� currents (La-sater, 1988; Maguire et al., 1989; Kaneko et al., 1989;

Zhuo-Hua Pan,1,3 Hui-Juan Hu,1

Paul Perring,2 and Rodrigo Andrade2

1Department of Anatomy and Cell Biology2Department of Psychiatry and Behavioral Pan and Lipton, 1995; de la Villa et al., 1998; Pan, 2000).

Previous studies on Mb1 bipolar cells in goldfish haveNeurosciencesWayne State University School of Medicine demonstrated that activation of L-type Ca2� currents

can trigger transmitter release (Tachibana et al. 1993;Detroit, Michigan 48201von Gersdorff and Matthews, 1994). This has led to theidea that transmitter release from bipolar cells is medi-ated exclusively by L-type Ca2� channels. However,Summarypatch-clamp recording of isolated presynaptic terminalsof mammalian bipolar cells indicates that LVA Ca2� cur-Transmitter release in neurons is thought to be medi-

ated exclusively by high-voltage-activated (HVA) Ca2� rents are also present at the presynaptic terminals (Pan,2001). This raises the possibility that LVA Ca2� currentschannels. However, we now report that, in retinal bipo-

lar cells, low-voltage-activated (LVA) Ca2� channels may mediate transmitter release in bipolar cells. Herewe show that T-type Ca2� channels underlie the LVAalso mediate neurotransmitter release. Bipolar cells

are specialized neurons that release neurotransmitter Ca2� currents in retinal bipolar cells and that activationof these channels directly triggers transmitter release.in response to graded depolarizations. Here we show

that these cells express T-type Ca2� channel subunitsand functional LVA Ca2� currents sensitive to mibef- Resultsradil. Activation of these currents results in Ca2� influxinto presynaptic terminals and exocytosis, which we Molecular Composition of LVA Ca2� Currentsdetected as a capacitance increase in isolated termi- in Bipolar Cellsnals and the appearance of reciprocal currents in reti- To determine the molecular composition of LVA Ca2�

nal slices. The involvement of T-type Ca2� channels current in bipolar cells, we first tested for the presencein bipolar cell transmitter release may contribute to of T-type Ca2� channel subunit mRNAs in whole retinaretinal information processing. mRNA. Amplification using degenerate primers for all

three T-type Ca2� channel subunits (�1G, �1H, and �1I)Introduction revealed two bands of about 520 bp and 470 bp (Figure

1A, Cmn lane). Restriction analysis suggested that theOne of the primary functions of voltage-dependent Ca2� smaller band contained fragments amplified from �1H

channels in the CNS is triggering transmitter release and �1I (lanes H and I). The larger band was cut by XhoI,at presynaptic terminals (Dunlap et al., 1995). Voltage- as expected for a fragment amplified from �1G (line G),activated Ca2� channels have been divided into low- except that it was �30 bp longer than predicted by thevoltage-activated (LVA) and high-voltage-activated rat neuronal �1G sequence (492 bp) (Perez-Reyes et al.,(HVA) Ca2� channels that include L-, N-, P/Q-, and 1998). To confirm the identification and to resolve thisR-types (Tsien et al., 1995). LVA Ca2� channels are also discrepancy, we subcloned and sequenced the prod-commonly referred to as T-type and are characterized ucts. The sequence confirmed that the two short frag-by activation at hyperpolarized potentials near rest and ments corresponded to �1H and �1I. The larger fragmentfast inactivation (Huguenard, 1996). Three members of corresponded to a recently described, alternativelythe T-type Ca2� channel family (�1G, �1H, and �1I) have spliced form of �1G (525 bp) that is expressed in pancre-recently been cloned (Perez-Reyes et al., 1998; Cribbs atic � cells (Zhuang et al., 2000).et al., 1998; Lee et al., 1999). Interestingly, to date, fast We next examined the expression of T-type Ca2�

transmitter release at presynaptic terminals in the CNS channel subunits in rod and cone bipolar cells usinghas been reported only in response to Ca2� influx single-cell RT-PCR. The intracellular contents of iso-through HVA Ca2� channels (Dunlap et al., 1995; Wu et lated single bipolar cells were harvested into whole-al., 1998). cell recording pipettes, and mRNA amplification was

Retinal bipolar cells, second-order neurons in the ret- conducted using seminested primer pairs in which theina, relay the visual signal from photoreceptors to third- second pair was subunit specific. Figure 1B shows aorder neurons, amacrine and ganglion cells. Bipolar cells cone bipolar cell from which all three fragments couldare classified into rod and cone types on the basis of be successfully amplified (�1G, line G; �1H, line H; andtheir synaptic inputs, as well as ON and OFF types on �1I, line I). No product was amplified when the cellularthe basis of their light-response polarity (Werblin and template was replaced by water (line W, �1G-specificDowling, 1969). Bipolar cells respond to light with a primers). Overall, we were able to detect at least onesustained response and belong to the category of cen- subunit in 59% of the bipolar cells tested (13 out of 22),tral neurons that primarily uses a graded potential for and in 5 of these cells all three subunits were detected.signal transmission. This detection rate is in line with previous studies (Plant

et al., 1998). These results suggest that mRNAs for T-typeCa2� channel subunits are expressed in bipolar cells.3 Correspondence: zhpan@med.wayne.edu

Neuron90

the differential voltage sensitivity of the Ca2� currents,provide a viable approach to examining the roles of LVAand HVA Ca2� channels in mediating neurotransmitterrelease in these cells.

Calcium Influx into Axon Terminalsthrough T-Type Ca2� ChannelsCa2� influx into the presynaptic terminals has been re-ported to be directly correlated with transmitter release(Augustine and Neher, 1992; Llinas et al., 1992; BorstFigure 1. Expression of T-Type Ca2� Channel Subunit mRNAs in

Whole Retina and in Single Retinal Bipolar Cells and Sakmann, 1996). Therefore, we next examined(A) RT-PCR analysis of whole-retina T-type Ca2� channel mRNAs. whether T-type Ca2� channels could contribute to Ca2�

Amplification of mRNAs for all three T-type Ca2� channel �1 subunits transients in axon terminals of isolated rod and conewas obtained using a set of common primers (Cmn lane). The prod- bipolar cells. To accomplish this, Ca2� sensitive dyesuct of this reaction was then subjected to restriction digest using

were introduced into cells through the recording elec-enzymes targeting unique sites in each sequence on the basis oftrode. The cell membrane potential was controlled byavailable rat sequences (XhoI for the �1G subunit, lane G; Eco0109Ivoltage-clamp, and the Ca2� response at the axon termi-for �1H subunit, lane H; and AflII for the �1I subunit, lane I). The sizes

of the resulting fragments corresponded to those predicted from nal was recorded (Figure 3A). Typical results from a rodthe �1G, �1H, and �1I subunit sequences as described in the text (G, bipolar cell are shown in Figure 3B. We elicited Ca2�

333 bp and 192 bp; H, 270 bp and 201 bp; I, 274 bp and 197 bp). influx by briefly depolarizing the cell to �40 mV (leftThe L lane corresponds to a 100 bp ladder.

traces) or �10 mV (right traces) from a holding potential(B) RT-PCR analysis of T-Type Ca2� channel subunit mRNA expres-of �80 mV. Depolarization to �40 mV would be expectedsion in a single retinal bipolar cell. Amplification of mRNAs for T-typeto activate T-type Ca2� channels, whereas depolariza-Ca2� channel �1 subunits from a single bipolar cell was obtained

using seminested PCR reactions. This cell was found to express tion to �10 mV would be expected to activate bothmRNAs for all three subunits (predicted sizes: G, 432 bp; H, 343 bp; types because HVA Ca2� currents are significantly acti-I, 376 bp). The L lane corresponds to a 100 bp ladder. The lane vated only at potentials more positive than �30 mVlabeled as W illustrates a water control in which no template was

under our recording conditions (Pan, 2000). In controladded to the RT-PCR reaction aimed at detecting the �1G subunit.conditions, depolarizations to �40 mV and �10 mV bothWater controls for each of the three subunits were consistentlyproduced increases in intraterminal Ca2� concentration.negative in all these experiments.A Ca2� increase was still evoked at both potentials dur-ing the application of nimodipine (5 �M) (traces in thesecond row), although the recovery of the signal didPharmacological Properties of LVAbecome faster at �10 mV. This suggests some contribu-and HVA Ca2� Currentstion by L-type Ca2� channels to the Ca2� transientsTo examine the functional role of LVA Ca2� channels inevoked by stepping to �10 mV but not at �40 mV. Inbipolar cells, we needed to develop experimental proto-contrast, application of mibefradil (10 �M) eliminatedcols that could reliably separate effects mediated bythe Ca2� response at �40 mV and reduced that evokedLVA and HVA Ca2� channels. A low concentration ofby stepping to �10 mV (traces in the third row). Finally,mibefradil has been reported to selectively block T-typethe Ca2� response evoked at both potentials was com-Ca2� channels (Mishra and Hermsmeyer, 1994). Con-pletely blocked by the application of nimodipine andversely, L-type Ca2� currents are sensitive to a low con-mibefradil. These results indicate that both types of Ca2�centration of dihydropyridine compounds such as nimodi-channels contribute to Ca2� influx by membrane depo-pine. Therefore, we tested whether the combined uselarization to �10 mV but that Ca2� influx in responseof these two compounds could reliably isolate currentsto smaller depolarizations (e.g., �40 mV) is mediatedcarried by LVA or HVA Ca2� currents in bipolar cells.predominantly by T-type Ca2� channels. Similar resultsAs expected for a T-type current, the LVA Ca2� cur-were observed from all recorded rod bipolar cells (n �rents in bipolar cells were found to be highly sensitive24) and from the majority of cone bipolar cells (17 outto mibefradil. An example of this inhibition is illustratedof 24). In 7 of the cone bipolar cells tested, Ca2� influxin Figure 2A. No differences in this effect were observedat both �40 mV and �10 mV was found to be mediatedbetween rod and cone bipolar cells. The ability of mibe-mainly, if not exclusively, by LVA Ca2� channels becausefradil to inhibit LVA Ca2� currents was concentrationthe Ca2� increases evoked at both potentials were resis-dependent and exhibited an IC50 of 0.81 �M (n � 15)tant to nimodipine but were completely blocked by mi-(Figure 2C). This contrasted with the much lower po-befradil (data not shown). Overall, these results indicatetency of this compound for inhibiting HVA Ca2� currentsthat the T-type Ca2� channel contributes the Ca2� influx(Figures 2B and 3C; IC50 � 8.5 �M, n � 15). In contrastto presynaptic terminals of all bipolar cells.to mibefradil, nimodipine showed marked selectivity for

the HVA Ca2� currents in rod and cone bipolar cells(Figures 2D–2F). Thus, nimodipine inhibited HVA Ca2� Exocytosis Evoked by T-Type Ca2� Channels

Capacitance measurements have been used to monitorcurrents in these cells with an IC50 of 0.90 �M (n �15) but was more than 20-fold less potent at LVA Ca2� exocytosis and transmitter release in a number of prepa-

rations including the giant Mb1 bipolar cells in goldfishchannels (IC50 � 21.9 �M, n � 27). These results supportthe idea that two Ca2� currents in bipolar cells can be (Neher and Marty, 1982; Parsons et al., 1994; von Gers-

dorff and Matthews, 1994; Rieke and Schwartz, 1996).distinguished and isolated using mibefradil and nimodi-pine. These pharmacological differences, coupled with To determine whether T-type Ca2� currents could trigger

T-Type Ca2� Channels Mediate Transmitter Release91

Figure 2. Effects of Mibefradil and Nimodipine on LVA and HVA Ca2� Currents of Bipolar Cells

(A–C) Effects of mibefradil on LVA and HVA Ca2� currents. (A) Sample recordings illustrating the inhibition of LVA Ca2� currents by mibefradil(mib) in a cone bipolar cell. (B) Sample recordings illustrating the inhibition of HVA Ca2� currents by mibefradil in a rod bipolar cell. (C)Concentration-response curves for mibefradil’s inhibition of LVA (circle) and HVA (triangle) Ca2� currents. The IC50 and Hill coefficient are 0.81�m and 1.0 (n � 15) for LVA Ca2� currents and 8.5 �m and 1.4 (n � 15) for HVA Ca2� currents.(D–F) Effects of nimodipine on LVA and HVA Ca2� currents. (D) Sample recordings illustrating the concentration-dependent inhibition of LVACa2� currents by nimodipine (nim) on a rod bipolar cell. (E) Sample recordings illustrating the concentration-dependent inhibition of HVA Ca2�

currents by nimodipine in a cone bipolar cell. (F) Concentration-response curves for nimodipine on LVA (circle) and HVA (triangle) Ca2� currents.The IC50 and Hill coefficient are 21.9 �m and 1.5 (n � 27) for LVA Ca2� currents and 0.90 �m and 1.0 (n � 15) for HVA Ca2� currents.The LVA Ca2� currents in (A) and (D) were evoked by stepping to �40 mV from the holding potential of �80 mV. HVA Ca2� currents in (B) and(D) were evoked by stepping to �10 mV from the holding potential of �45 mV. Data points in (C) and (F) are means, and error bars representstandard deviation.

transmitter release, we went on to determine whether depends on Ca2� influx and is not simply due to gatingcharge movements (Horrigan and Bookman, 1994).capacitance measurements of exocytosis evoked by

Ca2� currents could be resolved in mammalian bipolar Consistent with the interpretation that the recordedcapacitance increase reflected vesicle fusion, when acells. To this effect we took advantage of a recently

developed isolated axon terminal preparation of rat rod lower concentration of EGTA (0.5 mM) was used in theelectrode solution, the average capacitance increasebipolar cells (Figure 4A) (Pan, 2001).

As illustrated in Figure 4B, Ca2� influx into isolated became larger (19.8 � 7.1 fF; n � 8) than that recordedin our normal Ca2� buffer condition (5 mM EGTA) (Figureterminals resulted in a detectable capacitance increase.

The membrane capacitance was monitored before and 4D). Under this condition, a continued slight upwarddrift in the capacitance after the termination of the stimu-after a depolarizing pulse to �10 mV from the holding

potentials of �85 mV. A capacitance jump (arrowhead) lus was usually observed (see Figure 4D). Although re-cordings with low EGTA were associated with a rela-was observed right after the termination of the depolar-

ization pulse, whereas no detectable change was ob- tively large tail current, probably indicating the presenceof Ca2�-dependent conductances, this was not foundserved in the series resistance (GS, lower trace; left col-

umn). Overall, we detected capacitance increases (�5 to affect the capacitance measurements. A reanalysisof this data, skipping the measurements obtained forfF) in response to a depolarization pulse to �10 mV in

about 70% of recorded isolated terminals. The average 100 ms after the termination of the stimulus to removeany contribution of the tail current, did not alter thevalue for the increase was 13.0 � 5.0 fF (mean � SD;

n � 32). In contrast, no significant increase in capaci- estimate of the change in capacitance (20.1 � 7.4 fFversus 19.8 � 7.1 fF). Taken together, these results indi-tance was observed when recordings were made in the

Co2�-containing solution (0.21 � 2.9 fF; n � 14) (Figure cate that the capacitance increase is sensitive to intra-cellular Ca2� buffer as would be expected if it resulted4B, right column) or when recordings were obtained

using an intracellular solution containing a high concen- from Ca2�-triggered vesicle exocytosis.Having established that we could record capacitancetration of Ca2� buffer BAPTA (10 mM) (1.6 � 3.3 fF; n �

4) (Figure 4C). This suggests that the capacitance jump changes associated with exocytosis in isolated bipolar

Neuron92

cytosis at these terminals, a capacitance increase couldbe detected in about 30% (7 out of 24) of the terminalsfollowing depolarization to �40 mV. In addition, capaci-tance increases were also detected following depolar-ization to both �30 mV and �10 mV in the presence ofnimodipine. These results strongly suggest a role forT-type Ca2� channels in neurotransmitter release at thisterminal.

To further test the role for T-type Ca2� channels, wenext compared capacitance changes elicited by stepdepolarizations to �30 mV and �10 mV in the presenceor absence of mibefradil (10 �M). As illustrated in Figures5B and 5C, the capacitance increase evoked by depolar-ization to �30 mV was essentially abolished by mibef-radil. In contrast, the capacitance increase elicited bystepping to �10 mV was only partially inhibited. Thesefindings are consistent with the notion that mainlyT-type, but not L-type, Ca2� currents mediate releasefollowing depolarizations to �30 mV, whereas bothtypes of Ca2� currents mediate release following depo-larizations to �10 mV. In agreement with this idea, nosignificant increase in capacitance (�1.9 � 3.6 fF; n �7) was observed when recordings were made in thecombined presence of nimodipine (10 �M) and mibef-radil (10 �M) (Figure 5C). Overall, these results indicatethat capacitance increase at the axon terminals of rodbipolar cells could be evoked by Ca2� influx throughT- as well as L-type Ca2� channels.

Reciprocal Currents Evoked by T-TypeCa2� ChannelsRetinal bipolar cells are known to receive reciprocalinhibitory GABAergic feedback from amacrine cells(Dowling and Boycott, 1966), and the reciprocal inhibi-tory currents have been used to monitor bipolar celltransmitter release in retinal slices (Dong and Werblin,1998; Hartveit, 1999). This provided an additional avenueto examine the role of LVA Ca2� currents in bipolar cell

Figure 3. Fluorescence Measurement of Ca2� Influx into Axon Ter-transmitter release.minals of Bipolar Cells

Bipolar cells were recorded in rat retinal slices, and(A) Illustration of the recording set up. Ca2�-sensitive dye was loadedtheir identity was confirmed by fluorescence dye fillinginto the cell through the recording electrode, and the cell was voltagethrough the recording electrode (Figure 6A). We lookedclamped. Fluorescence signals from the axon terminals were re-

stricted by the use of a small window and recorded by a PMT-based for LVA Ca2� channel triggered synaptic events by in-photometric system. Scale bar is 10 �m. cluding nimodipine (10–20 �M) in all recording solutions(B) Intracellular Ca2� transients recorded from the axon terminal of to block L-type Ca2� channels. A sample recording froma rod bipolar cell. The left column depicts responses evoked by

a cone bipolar cell is shown in Figures 6B and 6C. Whenstepping to �40 mV, and the right column responses evoked bythe membrane potential was depolarized from �80 mVstepping to �10 mV. In all cases the holding potential was �80 mV.to ��50 mV, an inactivating inward current with voltageIn this cell, administration of nimodipine alone had little effect on

the Ca2� transients. In contrast, mibefradil blocked the response dependence consistent with a T-type Ca2� channel waselicited by stepping to �40 mV and reduced that elicited by stepping observed. This inward current was associated with theto �10 mV. Finally, the coadministration of mibefradil and nimodi- appearance of fluctuating currents in both rod (n � 17)pine blocked the Ca2� transients evoked by stepping to both �40

and cone (n � 4) bipolar cells. In addition, large tailmV and �10 mV. The black traces represent the raw data, and thecurrents were observed immediately after the termina-red traces are averaged data through a time window of 100 ms.tion of the test pulse. The transient fluctuating currentsevoked during the test pulse and the tail current arethought to represent reciprocal inhibitory currents, me-cell terminals, we next asked whether Ca2� influx

through LVA Ca2� channels could play a role in this diated mainly by GABAA and/or GABAC receptors (Dongand Werblin, 1998; Protti and Llano, 1998; Hartveit,process. To this effect we first compared capacitance

changes elicited by step depolarizations to �40 mV, 1999). Consistent with this idea, application of bicucul-line (200 �M) and 3-amino-propyl(methyl)phosphinic�30 mV, and �10 mV in the presence or absence of

nimodipine (10 �M). A sample recording is illustrated in acid (3-APMPA) (200 �M), antagonists of GABAA andGABAC receptors, respectively, blocked the fluctuatingFigure 5A, and the overall results are plotted in Figure

5C. Consistent with a role for LVA Ca2� currents in exo- currents during the test pulse as well as the tail currents

T-Type Ca2� Channels Mediate Transmitter Release93

Figure 4. Capacitance Measurements in Iso-lated Bipolar Cell Axon Terminals

(A) Illustration of an isolated presynaptic ter-minal of a rod bipolar cell. Scale bar is 10 �m.(B) Sample recordings of capacitance froman isolated axon terminal under control con-ditions (left column) and in the presence of 4mM Co2� (right column). The top traces showthe stimulation protocol used for capacitancemeasurements. The traces labeled Cm depictcapacitance measurements before and aftera 400 ms depolarization pulse to �10 mV fromthe holding potential of �85 mV. A capaci-tance jump was observed after the stimula-tion in control conditions (arrowhead) but notin the presence of Co2� (right column). Thetraces labeled ICa show the Ca2� elicited bythe depolarizing step. Ca2� currents wereblocked by Co2� (right column). The traceslabeled Gs show the series resistance. No de-tectable changes were observed in either case.(C) A sample capacitance measurement whenthe recording electrode contained the fastCa2� buffer BAPTA (10 mM). No jump in ca-pacitance was evoked despite the presenceof a large Ca2� current. The Ca2� current re-cordings in BAPTA displayed less inactiva-tion, suggesting that Ca2�-dependent regula-tion may be involved.(D) A sample capacitance measurement whenthe recording electrode contained a low con-centration of EGTA (0.5 mM). A large jump incapacitance was evoked. The black tracesrepresent the raw data, and the red tracesare averaged data through a time window of40 ms. Dotted lines represent the levels be-fore stimulation.

(Figure 6C; n � 6). These results indicate that bipolar tected in the intracellular contents harvested from singlebipolar cells. In addition, LVA Ca2� currents in bipolarcell synaptic transmission persists even after the block-

ade of HVA Ca2� channels. To test for the involvement cells were shown to be highly sensitive to mibefradil, aT-type Ca2� channel antagonist. The IC50 for this effectof T-type Ca2� channels in supporting synaptic activity,

we again used mibefradil (10 �M). As illustrated in Figure was close to that previously reported for native T-typeCa2� channels in other neuron preparations or for re-6C, administration of this compound completely elimi-

nated the inactivating inward current and associated combinant T-type Ca2� channels (McDonough andBean, 1998; Cribbs et al., 1998; Martin et al., 2000).reciprocal inhibitory currents (n � 4). Mibefradil (10 �M)

was not found to reduce the currents evoked by bath- Furthermore, the partial sensitivity of the LVA Ca2� cur-rents of bipolar cells to nimodipine is also consistentapplied GABA (100 �M) in either rod or cone bipolar

cells (data not shown). Taken together, these results with the pharmacological properties of T-type Ca2� cur-rents reported in other preparations (Cohen and McCar-indicate that activation of T-type Ca2� currents evokes

reciprocal inhibitory currents in bipolar cells in retinal thy, 1987; Richard et al., 1991; Randall and Tsien, 1997).Taken together, these results lead us to conclude thatslices.the LVA Ca2� currents in retinal bipolar cells are encodedby T-type Ca2� channel subunits.Discussion

The main finding of this study is the demonstration that T-Type Ca2� Currents Trigger Transmitter ReleaseT-type Ca2� currents can mediate transmitter release In this study, we also demonstrate that activation offrom synaptic terminals of retinal bipolar cells. Several T-type Ca2� channels in bipolar cells could directly trig-lines of evidence support this conclusion. ger transmitter release. First, we used Ca2� imaging to

show that activation of T-type Ca2� channels contributesto the Ca2� transients at the presynaptic terminals of allT-Type Ca2� Channels Underlie LVA Ca2� Currents

in Bipolar Cells recorded bipolar cells. In most cells, including all rodbipolar cells, both T- and L-type Ca2� channels contrib-In this study, we show that LVA Ca2� currents in mamma-

lian retinal bipolar cell synaptic terminals are of the ute to presynaptic Ca2� transients. These findings areconsistent with the presence of both LVA and L-typeT-type. By using RT-PCR, we detected the mRNAs for

all three T-type Ca2� channel subunits in the retina. The HVA Ca2� currents in isolated axon terminals of rodbipolar cells (Pan, 2001). Interestingly, in about 30% ofmRNAs of these T-type Ca2� channels were also de-

Neuron94

Figure 5. Voltage Dependency and Effects ofNimodipine and Mibefradil on CapacitanceChanges

(A) Sample recordings illustrating the capaci-tance changes from an isolated axon terminalfollowing step depolarizations to �40 mV,�30 mV, and �10 mV. The capacitance in-crease at �30 mV was persistent in the pres-ence of nimodipine (10 �M), whereas that at�10 mV was partially inhibited by nimodipine.(B) Sample recordings illustrating the effectof mibefradil (10 �M) on capacitance changesfrom an isolated axon terminal following stepdepolarizations to �30 mV and �10 mV. Thecapacitance increase at �30 mV was com-pletely blocked by mibefrail, whereas that at�10 mV was partially inhibited by mibefradil.(C) Plot summarizing the capacitance changesat different stimulation potentials observed incontrol conditions (square) or in the presenceof 10 �M nimodipine (circle; red), 10 �M mibef-radil (up triangle; blue), or 10 �M nimodipineplus 10 �M mibefradil (down triangle). Datapoints are mean, and error bars are standarddeviation from the indicated number of re-corded terminals in each condition.Some recordings in (C) for testing the effect ofmibefradil, including the sample recordingsshown in (B), were performed using 0.5 mMEGTA in the intracellular solution, and all oth-ers were made using 5 mM EGTA.

cone bipolar cells, these responses were evoked mainly Two lines of evidence support the idea that activationof these T-type Ca2� channels can trigger neurotransmit-through T-type Ca2� channels, suggesting that Ca2�

channels localized at the axon terminals of certain sub- ter release. The first line of evidence is based on capaci-tance measurement in presynaptic terminals. Studiessets of cone bipolar cells could be predominantly T-type.

Figure 6. LVA Ca2� Currents Evoke Recipro-cal Inhibitory Currents in Bipolar Cells in Reti-nal Slices

(A) A bipolar cell (rod bipolar cell) in a retinalslice visualized by injection of Alexa 488.Scale bar is 10 �m.(B) An inactivating inward current superim-posed with transient fluctuating currents isobserved in a cone bipolar cell following de-polarizing steps to voltages positive to �50mV from a holding potential of �80 mV.(C) Application of bicuculline (200 �M) plus3-APMPA (200 �M) blocks the transient fluc-tuating currents (thick trace). The applicationof mibefradil (10 �M) blocks all inward cur-rents.Recordings of (B) and (C) were made fromthe same cone bipolar cell with a series resis-tance of 27 M. Nimodipine (10 �M) was in-cluded in bath and all recording solutions.

T-Type Ca2� Channels Mediate Transmitter Release95

of bipolar cells in goldfish have demonstrated that a ated mainly by T-type Ca2� channels. This finding isconsistent with the report of the transient Ca2� responsecapacitance increase is directly correlated to bipolar

cell transmitter release (von Gersdorff et al., 1998; von in axon terminals of certain cone bipolar cells in ratretinal slices (Protti and Llano, 1998). Combined, theseGersdorff and Matthews, 1999). In this study, by using

an isolated axon-terminal preparation, we show that a observations suggest that T-type Ca2� channels mayplay a key role in mediating neurotransmitter release incapacitance change could also be detected in mamma-

lian bipolar cells. Although the capacitance increase in certain subsets of cone bipolar cells.One marked feature of information processing in ret-rat bipolar cells was found to be approximately one

order of magnitude smaller than that of the giant Mb1 ina is the conversion of the relatively sustained re-sponses of bipolar cells to both sustained and transientbipolar cells in goldfish (von Gersdorff and Matthews,

1994), it was clearly detectable in this preparation. This responses in retinal third-order neurons. The formationof the transient response in third-order neurons hascapacitance increase was dependent on Ca2� influx into

the terminal and sensitive to intracellular Ca2� buffer, been found to be due in part to transient transmitterrelease from bipolar cells (Dixon and Copenhagen, 1992;suggesting that it was secondary to exocytosis. Most

importantly, we show that the capacitance increase Bieda and Copenhagen, 2000). Several mechanismshave been proposed to account for the transient trans-could be elicited by activation of T-type Ca2� channels,

as demonstrated by the combination of membrane po- mitter release, including intrinsic release kinetics (vonGersdorff and Matthews, 1997; Neves and Lagnado,tentials and Ca2� channel antagonists. These results are

consistent with the finding that transmitter release from 1999), response waveforms (Awatramani and Slaughter,2000), or the properties of glutamate receptors (DeVries,bipolar cells can be triggered by T-type Ca2� currents.

Further evidence that T-type Ca2� channels can medi- 2000). The present findings suggest that the transienttransmitter release of bipolar cells could also result fromate neurotransmitter release in bipolar cells was ob-

tained from the study of reciprocal inhibitory currents the involvement of T-type Ca2� channels. If this is thecase, the different activation and inactivation propertiesin retinal slices. Our results show that GABA-mediated

reciprocal currents could be observed in bipolar cells of LVA and HVA Ca2� channels, and the possible hetero-geneous expression of these two Ca2� channel types atin the presence of nimodipine, an L-type Ca2� channel

antagonist, at concentrations sufficient for a complete the axon terminals of different subtypes of bipolar cells,could contribute to the diversity of synaptic transmis-blockade of L-type Ca2� currents, but they were inhib-

ited by administration of mibefradil. Again, these results sion between bipolar and third-order neurons in theretina.support the idea that T-type Ca2� channels can mediate

synaptic transmission in retinal bipolar cells. Previous studies have documented the involvementof LVA Ca2� currents in hormone secretion and gradedsynaptic transmission in invertebrates (Cohen et al.,Implications in Retinal Synaptic Transmission1988; Tsien et al., 1988; Ivanow and Calabrese, 2000).Transmitter release from bipolar cells mediated by L-typeHowever, previously, it has not been reported that acti-HVA Ca2� currents has been demonstrated in the Mb1vation of T-type Ca2� channels can directly trigger fastbipolar cells of goldfish (Tachibana et al., 1993; von Gers-transmitter release in the CNS. Because retinal bipolardorff and Matthews, 1994). The Mb1 bipolar cells, how-cells use graded potentials for signal transmission, itever, express only L-type Ca2� currents (Tachibana andmakes physiological sense for these cells to utilize LVAOkada, 1991; Heidelberger and Matthews, 1992). WeCa2� channels to control transmitter release. This pro-now report that neurotransmitter release in mammaliancess contrasts with traditional synaptic transmission inbipolar cells is mediated by both L- and T-type Ca2�

the CNS where fixed amplitude action potentials propa-channels. Specifically, neurotransmitter release in re-gate into the presynaptic terminal to activate HVA Ca2�sponse to modest depolarizations (�30 mV–�40 mV) ischannels and evoke quantal transmitter release (Dunlapmediated predominantly by T-type Ca2� currents,et al., 1995). Can T-type Ca2� channels mediate synapticwhereas release in response to stronger depolarizationstransmission elsewhere in the brain? Interestingly, therein most bipolar cells involves both T- and L-type.are sites in the brain where synaptic transmission isSeveral considerations suggest that these T-type Ca2�

resistant to antagonists for L-, N-, and P/Q-type Ca2�currents in bipolar cells can be activated under physio-channels (Mintz et al., 1995; Turner et al., 1995; Sabatinilogic conditions in vivo as first suggested by Kaneko etand Svoboda, 2000). Furthermore, dendrodendritic sig-al. (1989). The membrane potentials of bipolar cells werenaling based on decremental amplitude action poten-reported to range from �70 mV to �20 mV (Simon ettials (Margrie et al., 2001), or altogether independent ofal., 1975; Ashmore and Falk, 1980; Ashmore and Copen-action potentials (e.g., Isaacson and Strowbridge, 1998),hagen, 1983; Saito and Kaneko, 1983), which are in parthas also been reported. Given the abundant expressionwithin the operating range of T-type Ca2� channels. Also,of T-type Ca2� channel subunits in the brain (Talley eta recent study showed that the dark membrane poten-al., 1999), it will be interesting to see whether thesetials of some OFF-type cone bipolar cells in rat werechannels can mediate synaptic transmission at these or��40 mV–�50 mV (Euler and Masland, 2000). There-other sites in the CNS.fore, it is possible that the membrane potentials of some

these cells might not reach the level for activating HVAExperimental ProceduresCa2� channels because OFF bipolar cells hyperpolarize

in response to light. Furthermore, during the course of Cell Isolation and Retinal Slice Preparationthis study we observed a subset of cone bipolar cells Dissociated bipolar cells and isolated axon terminals from Long

Evans rats �4 weeks of age were prepared as previously describedin which Ca2� influx into presynaptic terminals was medi-

Neuron96

(Pan and Lipton, 1995; Pan, 2000, 2001). Retinas were removed and CaCl2, 2.5; MgSO4, 0.5; MgCl2, 0.5; HEPES-NaOH, 5; and glucose,22.2; with phenol red, 0.001% v/v (pH 7.2). All other recordings wereplaced in a Hanks’ solution (in mM): NaCl, 138; NaHCO3, 1; Na2HPO4,

0.3; KCl, 5; KH2PO4, 0.3; CaCl2, 1.25; MgSO4, 0.5; MgCl2, 0.5; HEPES- made in 10 mM Ca2� extracellular solutions that contained (in mM):NaCl, 95; KCl, 5; CsCl, 10; TEA-Cl, 20; MgCl2, 1; CaCl2 10; HEPES,NaOH, 5; and glucose, 22.2; with phenol red, 0.001% v/v (pH 7.2).

The retinas were incubated for �40–50 min at 34C–37oC in an 5; and glucose, 22.2; with phenol red, 0.001% v/v (pH 7.2).enzymatic solution that consisted of the normal Hanks’ solutiondescribed above, supplemented with 0.2 mg/ml DL-cysteine, 0.2

Capacitance Measurementsmg/ml bovine serum albumin, and 1.6 U/ml papain and mechanicallyCapacitance measurements were performed in the “sine � DC”dissociated by gently triturating in a Ca2�-free Hanks’ solution. Themode of the LockIn extension of PULSE software with a frequencydissociated cells were plated onto culture dishes in normal or Ca2�-of 400–800 Hz and a peak-to-peak voltage of 30 mV added to thefree Hanks’ solution. The contents of Ca2�-free Hanks’ solution areholding potential of �85 mV or �80 mV. Recordings were performedthe same as the normal Hanks’ described above, except that Ca2�

using the conventional whole-cell configuration on isolated bipolarions are omitted. Bipolar cells and isolated axon terminals werecell axon terminals. The whole-cell seal was established in the Ca2�-identified by their morphological appearance.free solution to avoid overstimulation of the release. Terminals withRetinal slices were prepared according to procedures previouslydetectable changes in the series resistance during the capacitancereported (Werblin, 1978). Retinal slices �200 �m thick were mountedmeasurement were not included in the data analysis (Gillis, 1995).in a glass-bottomed recording chamber. The recording chamberThe capacitance values for the isolated axon terminals range fromwas mounted on the stage of an upright microscope equipped with a0.32–2.1 pF, with an average of 1.46 pF (n � 49). Capacitance40� water-immersion objective and DIC and epifluorescence optics.changes were evoked by a 400 ms stimulation pulse. The membraneThe recording chamber (�0.9 ml) was continuously superfused withcapacitance was not monitored during or 5–25 ms right before andoxygenated Hanks’ solution at the rate of �1.5 ml/min. All recordingsafter the pulse. The capacitance change was determined by thewere performed at room temperature (�22oC).difference in the averaged mean during 0.4 s before and 1 s afterthe pulse. The endocytosis process was not found to be significant

RT-PCR during the course of the capacitance measurement. Because a sig-For whole-retina RT-PCR, poly (A�) RNAs were harvested from nificant capacitance increase was not always observed in recordedwhole rat retinae. For single-cell RT-PCR, patch pipettes (�10 M terminals, only those terminals with a capacitance increase of �5tip resistance) were filled with an internal solution containing (in fF in control (without blockers) were included in the average analysis,mM): KCl, 140; CaCl2, 1; EGTA, 5; and HEPES, 5 (pH adjusted with with the exception of the measurements made in the Co2� where allKOH to 7.4). To collect bipolar cell cytoplasm content, the cells were the recorded terminals were included in the analysis. The averages inheld under a fast stream of control solution, and constant positive the application of nimodipine and/or mibefradil were obtained onlypressure was applied to the inside of the electrode. The cell cyto- from those terminals where the capacitance increase in control wasplasm content was harvested by applying negative pressure. Re- observed.verse transcription of the harvested materials was performed usingthe SuperScript kit (Life Technologies). For restriction analysis, sub-cloning, and sequencing, cDNA fragments were amplified using the Fluorometric Intracellular Ca2� Measurementsfollowing primers: forward primer 5�-GGCGT(G/C)GT(G/C)GT For measurement of intracellular Ca2� changes, patch-clamp re-(G/C)GAGAACTT-3�, and reverse primer 5�-GATGATGGTGGG(A/G) cording electrodes containing the membrane-impermeant, Ca2�-TTGAT-3�. These primers are common to all three T-type Ca2� chan- sensitive fluorescence dyes Fluo-4 or Oregon green 488 at concen-nel subunits: �1G, �1H, and �1I. These primers were adopted from a trations of 100–200 �M were used. The electrode solution containedrecently published study (Lambert et al., 1998). For single-cell RT- (in mM): CsCl, 120; TEA-Cl, 20; MgCl2, 4; HEPES, 30; Na-GTP, 0.5;PCR, cDNA fragments were amplified by two rounds of PCR using and Mg-ATP, 4 (pH 7.4). The membrane potential was controlledfirst the common primer pairs described above and then a sem- through the patch electrode. The fluorescence signal was detectedinested primer pair specific for each subunit. The seminested primer by a photomultiplier tube (PMT)-based photometric system coupledpairs were composed of the following specific forward primers: �1G, with the EPC-9 amplifier and PULSE software (TILL photonics). The5�-TACGGAGGCTGGAGAAAA-3�; �1H, 5�-CGCAGACTATTCACA excitation light (480 nm) was generated by a scanning monochroma-CAC-3�; and �1I, 5�-GGAAAAGAAGCGCCGTAA-3�, and the common tor and coupled to the microscope through an optical fiber. Thereverse primer described above. The amplification procedure was dichroic and emission filters had center wavelengths at 505 and 515as follows: 1 cycle at 95oC for 15 min; and 34 cycles composed of nm, respectively. The fluorescence signal in the axon terminal was40 s at 94oC, 40 s at 58oC, 1 min at 72oC, and 10 min at 72oC. The measured through an adjustable window.temperature was optimized by using 10 pg RNA from total retina.Possible contamination artifacts were checked by replacing the

Chemical Agent Application and Data Analysiscellular template with water. To exclude the possible contributionMibefradil was a kind gift from F. Hoffmann (La Roche, Basel, Swit-of the products from genomic DNA, control experiments where thezerland). All other chemicals were purchased from Sigma. In theRT was omitted in the reverse transcription for the harvest materialsisolated cell recordings, chemical agents were applied by local per-were also performed. Both of these controls were negative.fusion through multiple-barreled, gravity-driven superfusion pi-pettes placed �200–300 �m from the cell. In the slice recordings,

Patch-Clamp Recordingschemical agents were applied by bath perfusion. Nimodipine (10–20

Recordings with patch electrodes in the whole-cell configuration�M) was included in the bath and all recording solutions in the

were made using standard procedures with an EPC-9 amplifierslice experiments. Data were analyzed off-line using PULSE-FIT

(Heka Electronik, Lambrecht/Pfalz, Germany) and PULSE softwareand ORIGIN programs. All data are given as mean and standard

(Heka Electronik) as previously described (Pan, 2000). Electrodesdeviation.

were coated with SYLGARD (Dow Corning, Midland, MI) and pol-ished. The resistance of the electrode was 7–14 M for both somaand terminal recordings. Series resistance ranged from 12–40 M. AcknowledgmentsExcept for the intracellular Ca2� measurements (see below), theelectrode solution contained (in mM): CsCl, 120; TEA-Cl, 20; MgCl2, We thank Dr. E.A. Ertel, Dr. E. Gutknecht, and Dr. P. Weber of F.1; CaCl2, 0.5; EGTA, 5; HEPES, 10; Na-GTP, 0.5; and Na-ATP, 2 (pH Hoffmann-La Roche for the gift of mibefradil; Dr. M.M. Slaughter,adjusted with CsOH to 7.4). In some capacitance measurements, Dr. S.A. Lipton, Dr. R. Pourcho, Dr. R. Heidelberger, and Dr. E. ChapinBAPTA (10 mM) or a low concentration of EGTA (0.5 mM) was used for comments. This work was supported by NIH grants EY12180in the recording solution as indicated in the text. For slice recordings, (Z.-H.P.), MH 43985 (R.A.), and MH 49355 (R.A.).the electrode solution contained 100 �M Alexa 488 (MolecularProbes, Eugene, OR). The extracellular solution for slice recordingscontained (in mM): NaCl, 138; Na2HPO4, 0.3; KCl, 5; KH2PO4, 0.3; Received January 26, 2001; revised July 31, 2001.

T-Type Ca2� Channels Mediate Transmitter Release97

References low-threshold Ca2� currents and graded synaptic transmission. J.Neurosci. 20, 4930–4943.

Ashmore, J.F., and Copenhagen, D.R. (1983). An analysis of trans- Lambert, R.C., Mckenna, F., Maulet, Y., Talley, E.M., Bayliss, D.A.,mission from cones to hyperpolarizing bipolar cells in the retina of Cribbs, L.L., Lee, J.-H., Perez-Reyes, E., and Feltz, A. (1998). Low-the turtle. J. Physiol. 340, 569–597. voltage-activated Ca2� currents are generated by members of the

CaVT subunit family (�1G/H) in rat primary sensory neurons. J. Neu-Ashmore, J.F., and Falk, G. (1980). Response of rod bipolar cells inrosci. 18, 8605–8613.the dark-adapted retina of the dogfish, scyliorhinus canicula. J.

Physiol. 300, 115–150. Lasater, E.M. (1988). Membrane currents of retinal bipolar cells inculture. J. Neurophysiol. 60, 1460–1480.Augustine, G.J., and Neher, E. (1992). Calcium requirements for

secretion in bovine chromaffin cells. J. Physiol. 450, 247–271. Lee, J.-H., Daud, A.N., Cribbs, L.L., Lacerda, A.E., Pereverzev, A.,Klockner, U., Schneider, T., and Perez-Reyes, E. (1999). Cloning andAwatramani, G.B., and Slaughter, M.M. (2000). Origin of transientexpression of a novel member of the low voltage-activated T-typeand sustained responses in ganglion cells of the retina. J. Neurosci.calcium channel family. J. Neurosci. 19, 1912–1921.20, 7087–7095.Llinas, R., Sugimori, M., and Silver, R.B. (1992). Microdomains ofBieda, M.C., and Copenhagen, D.R. (2000). Inhibition is not requiredhigh calcium concentration in presynaptic terminal. Science 256,for the production of transient spiking responses from retinal gan-677–679.glion cells. Vis. Neurosci. 17, 243–254.Maguire, G., Maple, B., Lukasiewicz, P., and Werblin, F. (1989).Borst, J.G., and Sakmann, B. (1996). Calcium influx and transmitter -aminobutyrate type B receptor modulation of L-type calciumrelease in a fast CNS synapse. Nature 383, 431–434.channel current at bipolar cell terminals in the retina of the tiger

Cohen, C.J., and McCarthy, R.T. (1987). Nimodipine block of calcium salamander. Proc. Natl. Acad. Sci. USA 86, 10144–10147.channels in rat anterior pituitary cells. J. Physiol. 387, 195–225.

Margrie, T.W., Sakmann, B., and Urban, N.N. (2001). Action potentialCohen, C.J., McCarthy, R.T., Barrett, P.O., and Rasmussen, H. propagation in mitral cell lateral dendrites is decremental and con-(1988). Ca2� channels in adrenal glomerulosa cells: K� and angioten- trols recurrent and lateral inhibition in the mammalian olfactory bulb.sin II increase T-type Ca2� channel current. Proc. Natl. Acad. Sci. Proc. Natl. Acad. Sci. USA 98, 319–324.USA 85, 2412–2416.

Martin, R.L., Lee, J.-H., Cribbs, L.L., Perez-Reyes, E., and Hanck,Cribbs, L.L., Lee, J.H., Yang, J., Satin, J., Zhang, Y., Daud, A., Bar- D.A. (2000). Mibefradil block of cloned T-type calcium channels. J.clay, J., Williamson, M.P., Fox, M., Rees, M., et al. (1998). Cloning Pharmacol. Exp. Ther. 295, 302–308.and characterization of �1H from human heart, a member of the McDonough, S.I., and Bean, B.P. (1998). Mibefradil inhibition ofT-type calcium channel gene family. Circ. Res. 83, 103–109. T-type calcium channels in cerebellar Purkinje neurons. Mol. Phar-de la Villa, P., Vaquero, C.F., and Kaneko, A. (1998). Two types of macol. 54, 1080–1087.calcium currents of the mouse bipolar cells recorded in the retinal Mishra, S.K., and Hermsmeyer, K. (1994). Selective inhibition ofslice preparation. Eur. J. Neurosci. 10, 317–323. T-type Ca2� channels by Ro 40-5967. Circ. Res. 75, 144–148.DeVries, S.H. (2000). Bipolar cells use kainate and AMPA receptors Mintz, I., Sabatini, B.L., and Regehr, W.G. (1995). Calcium controlto filter visual information into separate channels. Neuron 28, of transmitter release at a cerebellar synapse. Neuron 15, 675–688.847–856.

Neher, E., and Marty, A. (1982). Discrete changes of cell membraneDixon, D.B., and Copenhagen, D.R. (1992). Two types of glutamate capacitance observed under conditions of enhanced secretion inreceptors differentially excite amacrine cells in the tiger salamander bovine adrenal chromaffin cells. Proc. Natl. Acad. Sci. USA 79, 6712–retina. J. Physiol. 449, 589–606. 6716.Dong, C.-J., and Werblin, F.S. (1998). Temporal contrast enhance- Neves, G., and Lagnado, L. (1999). The kinetics of exocytosis andment via GABAC feedback at bipolar terminals in the tiger salaman- endocytosis in the synaptic terminal of goldfish retinal bipolar cells.der retina. J. Neurophysiol. 79, 2171–2180. J. Physiol. 515, 181–202.Dowling, J.E, and Boycott, B.B. (1966). Organization of the primate Pan, Z.-H. (2000). Differential expression of high- and two types ofretina: electron microscopy. Proc. R. Soc. Lond. B Biol. Sci. 166, low-voltage-activated calcium currents in rod and cone bipolar cells80–111. of the rat retina. J. Neurophysiol. 83, 513–527.

Dunlap, K., Luebke, J.I., and Turner, T.J. (1995). Exocytotic Ca2� Pan, Z.-H. (2001). Voltage-gated Ca2� channels and ionotropic GABAchannels in mammalian central neurons. Trends Neurosci. 18, 89–98. receptors localized at axon terminals of mammalian retinal bipolar

cells. Visual Neurosci. 18, 279–288.Euler, T., and Masland, R.H. (2000). Light-evoked responses of bipo-Pan, Z.-H., and Lipton, S.A. (1995). Multiple GABA receptor subtypeslar cells in a mammalian retina. J. Neurophysiol. 83, 1817–1829.mediate inhibition of calcium influx at rat retinal bipolar cell termi-Gillis, K.D. (1995). Techniques for membrane capacitance measure-nals. J. Neurosci. 15, 2668–2679.ments. In Single-Channel Recording, Second Edition, B. SakmannParsons, T.D., Lenzi, D., Almers, W., and Roberts, W.M. (1994). Cal-and E. Neher, eds. (New York: Plenum), pp. 155–198.cium-triggered exocytosis and endocytosis in an isolated presynap-Hartveit, E. (1999). Reciprocal synaptic interactions between rodtic cell: capacitance measurements in saccular hair cells. Neuronbipolar cells and amacrine cells in the rat retina. J. Neurophysiol.13, 875–883.81, 2923–2936.Perez-Reyes, E., Cribbs, L.L., Daud, A., Lacerda, A.E., Barclay, J.,Heidelberger, R., and Matthews, G. (1992). Calcium influx and cal-Williamson, M.P., Fox, M., Rees, M., and Lee, J.-H. (1998). Molecularcium current in single synaptic terminals of goldfish retinal bipolarcharacterization of a neuronal low-voltage-activated T-type calciumneurons. J. Physiol. 447, 235–256.channel. Nature 391, 896–900.

Horrigan, F.T., and Bookman, R.J. (1994). Releasable pools andPlant, T.D., Schirra, G., Katz, C.S., Uchitel, O.D., and Konnerth, A.

the kinetics of exocytosis in adrenal chromaffin cells. Neuron 13,(1998). Single-cell RT-PCR and functional characterization of Ca2�

1119–1129.channels in motoneurons of the rat facial nucleus. J. Neurosci. 18,

Huguenard, J.R. (1996). Low-threshold calcium currents in central 9573–9584.nervous system neurons. Ann. Rev. Physiol. 58, 329–348. Protti, D.A., and Llano, I. (1998). Calcium currents and calcium sig-Kaneko, A., Pinto, L.H., and Tachibana, M. (1989). Transient calcium naling in rod bipolar cells of rat retinal slice. J. Neurosci. 18, 3715–current of retinal bipolar cells of the mouse. J. Physiol. 410, 613–629. 3724.

Isaacson, J.S., and Strowbridge, B.W. (1998). Olfactory reciprocal Randall, A.D., and Tsien, R.W. (1997). Contrasting biophysical andsynapses dendritic signaling in the CNS. Neuron 20, 749–761. pharmacological properties of T-type and R-type calcium channels.

Neuropharmacology 36, 879–893.Ivanov, A.I., and Calabrese, R.L. (2000). Intracellular Ca2� dynamicsduring spontaneous and evoked activity of leech heart interneurons: Richard, S., Diochot, S., Nargeot, J., Baldy-Moulinier, M., and Val-

Neuron98

mier, J. (1991). Inhibition of T-type calcium currents by dihydropyri-dines in mouse embryonic dorsal root ganglion neurons. Neurosci.Lett. 132, 229–234.

Rieke, F., and Schwartz, E.A. (1996). Asynchronous transmitter re-lease: control of exocytosis and endocytosis at the salamander rodsynapse. J. Physiol. 493, 1–8.

Sabatini, B.L., and Svoboda, K. (2000). Analysis of calcium channelsin single spines using optical fluctuation analysis. Nature 408,589–593.

Saito, T., and Kaneko, A. (1983). Ionic mechanisms underlying theresponses of OFF-center bipolar cells in the carp retina I: studiesof responses evoked by light. J. Gen. Physiol. 81, 589–601.

Simon, E.J., Lamb, T.D., and Hodgkin, A.L. (1975). Spontaneousvoltage fluctuations in retinal cones and bipolar cells. Nature 256,661–662.

Tachibana, M., and Okada, T. (1991). Release of endogenous aminoacids from ON-type bipolar cells isolated from the goldfish retina.J. Neurosci. 11, 2199–2208.

Tachibana, M., Okada, T., Arimura, T., Kobayashi, K., and Piccolino,M. (1993). Dihydropyridine-sensitive calcium current mediates neu-rotransmitter release from bipolar cells of the goldfish retina. J.Neurosci. 13, 2898–2909.

Talley, E.M., Cribbs, L.L., Lee, J.-H., Daud, A., Perez-Reyes, E., andBayliss, D.A. (1999). Differential distribution of three members of agene family encoding low voltage-activated (T-type) calcium chan-nels. J. Neurosci. 19, 1895–1911.

Tsien, R.W., Lipscombe, D., Madison, D., Bley, K., and Fox, A. (1988).Multiple types of neuronal calcium channels and their selective mod-ulation. Trends Neurosci. 11, 431–438.

Tsien, R.W., Lipscombe, D., Madison, D., Bley, K., and Fox, A. (1995).Reflections on Ca2�-channel diversity, 1988–1994. Trends Neurosci.18, 52–54.

Turner, T.J., Lampe, R.A., and Dunlap, K. (1995). Characterizationof presynaptic calcium channels with omega-conotoxin MVIIC andomega-grammotoxin SIA: role for a resistant calcium channel typein neurosecretion. Mol. Pharmacol. 47, 348–353.

von Gersdorff, H., and Matthews, G. (1994). Dynamics of synapticvesicle fusion and membrane retrieval in synaptic terminals. Nature367, 735–739.

von Gersdorff, H., and Matthews, G. (1997). Depletion and replen-ishment of vesicle pools at a ribbon-type synaptic terminal. J. Neu-rosci. 17, 1919–1927.

von Gersdorff, H., and Matthews, G. (1999). Electrophysiology ofsynaptic vesicle cycling. Annu. Rev. Physiol. 61, 725–752.

von Gersdorff, H., Sakaba, T., Berglund, K., and Tachibana, M.(1998). Submillisecond kinetics of glutamate release from a sensorysynapse. Neuron 21, 1177–1188.

Werblin, F.S. (1978). Transmission along and between rods in thetiger salamander retina. J. Physiol. 280, 449–470.

Werblin, F.S., and Dowling, J.E. (1969). Organization of the retina ofthe mudpuppy, Necturus maculosus II: intracellular recording. J.Neurophysiol. 32, 339–355.

Wu, L.G., Borst, J.G., and Sakmann, B. (1998). R-type Ca2� currentsevoke transmitter release at a rat central synapse. Proc. Natl. Acad.Sci. USA 95, 4720–4725.

Zhuang, H., Bhattacharjee, A., Hu, F., Zhang, M., Goswami, T., Wu,S., Berggren, P.O., and Li, M. (2000). Cloning of a T-type Ca2� chan-nel isoform in insulin-secreting cells. Diabetes 49, 59–64.