A: Retinal pathway underlying L-IPSCs in axotomized Mb terminals. In the absence of direct visual input from the soma, a light signal can reach axotomized Mb terminals via the lateral inhibitory pathway. R: rod; C: cone; BC: bipolar cell; AC: amacrine cell; Glu: glutamate; Gly: glycine. Bi: Representative light evoked signals recorded from an axotomized Mb terminal. Averaged traces are shown (n=3). Light pulses (500 ms, λ=505 nm) were delivered once every 30 sec. Intensity of light pulses is given as number of photons (cm⁻²s⁻¹), indicated at the top of each trace. Note that “ON” L-IPSCs first develop at much lower intensities than “OFF” L-IPSCs. Calibration: 10 pA (vertical), 100 ms (horizontal). Bii: Quantification of L-IPSCs with response/intensity curve (Q/ I) for the same experiment shown in Bi. Note that our protocol consisted of flashes at 10 light intensities but only 6 are shown on Bi for clarity. Charge transfer was calculated by integrating the area under the current trace for 500 ms during the illumination (“ON”), as well as for 500 ms immediately following the offset of illumination (“OFF”). Charge values for each trace were plotted against the light intensity on a logarithmic scale. Biii: Summary Response/Intensity diagram (n=5) obtained with green light flashes (500 ms, λ=505 nm). R/R_max % calculation was performed for each cell before averaging values across cells at any given intensity. Error bars represent +/− SE. Note that sizeable ON responses were present at the lowest intensities applied in this study (in the high end of the rod sensitivity range), at intensities where OFF L-IPSCs were barely present. Also, ON L-IPSCs were nearly saturated at rod saturating intensities. Both of these observations support the notion that ON L-IPSCs are rod-dominant, whereas OFF L-IPSCs are cone-dominant. The difference between the ON and OFF data set was statistically significant (ANOVA, Holm-Sidak method, p<0.001). Ci: Representative light evoked signals recorded from an axotomized Mb terminal. Averaged traces are shown (n=2–4). Light pulses (500 ms, λ=660 nm) were delivered once every 30 sec, superimposed on a steady background (λ=505 nm). Intensity of light pulses is given as number of photons (cm⁻²s⁻¹), indicated at the top of each trace. The intensity of the green background light was 10¹⁰ photons*cm⁻²*s⁻¹. Note that both “ON” and “OFF” L-IPSCs were present, and that they developed at the same bright intensities. Calibration: 10 pA (vertical), 100 ms (horizontal). Cii: Quantification of L-IPSCs with response/intensity curve (Q/ I) for the experiment shown in Ci. Charge transfer was calculated as in Bii. Ciii: Summary Response/Intensity diagram (n=5) obtained with red light flashes (500 ms, λ=660 nm) superimposed on a steady green background (λ=505 nm). Data is presented as in Biii. Note that in the presence of (rod-saturating) background light the “ON” and “OFF” Q/I curves overlap, and there was no statistical difference between the ON and OFF data set (p=0.209, ANOVA, Holm-Sidak method,).

Ai: Both the ON and the OFF portions of the light-evoked (λ=505 nm, I=1.95×10¹² photons cm⁻²s⁻¹, 500 ms) L-IPSCs were inhibited by the GABA_CR antagonist TPMPA (150 µM, red trace). Average traces are shown for each condition (n=5). Aii: Quantification of TPMPA’s effect on ON and OFF charge transfer of L-IPSCs. Means are shown as filled circles connected with solid black lines. Error bars represent ± SE. TPMPA significantly reduced the charge of both ON and OFF L-IPSCs (*p<0.05, **p<0.02, paired Student’s t-test, two-tailed). Bi: SR95531 (25 µM, red trace) a GABA_AR antagonist markedly increased L-IPSCs, indicating that ACs providing GABAergic feedback to Mb terminals receive serial inhibition via GABA_A receptors. Bii: Quantification of SR95531’s effect on ON and OFF charge transfer of L-IPSCs. SR95531 significantly increased the charge of both ON and OFF L-IPSCs (*p<0.05, **p<0.02, paired Student’s t-test, two-tailed). Biii: Same cell as in Bi; traces obtained by consecutive treatments are divided into two panels for increased visibility. The SR95531-elevated L-IPSCs (red trace) were reduced by TPMPA (here 300 µM shown, blue trace), but not eliminated completely. Complete LIPSC block of SR95531-elevated L-IPSCs could be achieved upon addition of picrotoxin (100 µM, PTX, green trace). Biv: Quantification of PTX effect on the ON and OFF charge transfer of L-IPSCs. PTX significantly decreased the charge of both ON and OFF L-IPSCs (*p<0.05, **p<0.02, paired Student’s t-test, two-tailed). C: Summary diagram of normalized GABAergic drug effects on light-evoked L-IPSCs. Data is presented as mean± SE.

A: Diagram depicting exclusive triggering of reciprocal feedback by direct step depolarization of an axotomized Mb terminal. Bi: Depolarization of an axotomized Mb terminal from the holding potential (HP) of −60 to 0 mV for 100 ms activated calcium influx through voltage-gated calcium channels (I_Ca), which triggered glutamate release, as evidenced by a jump in ΔC_m. The protocol used is shown in the bottom trace. The fast sinewave used to measure C_m was not delivered during the depolarization. The inhibitory feedback to the presynaptic terminal is expressed as a flurry of outward IPSCs superimposed on I_Ca (arrow). Experiments were performed in the presence of LY367385 (100 µM) to block mGluR1-dependent potentiation of reciprocal feedback. Under these conditions, consecutive depolarizations of the presynaptic terminal 1 min apart produced reciprocal feedback with similar magnitudes. There was no evidence of short-term depression (red trace). The resting C_m for this terminal was 5.5 pF. Bii: Same cell as in Bi. A light stimulus was applied (λ=505 nm, I=1.95×10¹² photons cm⁻²s⁻¹, 100 ms) to evoke pure lateral feedback between two reciprocal feedback steps (1 min apart), such that the light-evoked L-IPSCs preceded the second presynaptic depolarizations by 500 ms. However, the amplitude of reciprocal feedback did not decrease. C: The reverse of the experiment depicted in Bii. Consecutive, bright white light-evoked (I=7.32*10¹³ photons*cm⁻²*s⁻¹) L-IPSCs were triggered 1 min apart, but a 100 ms depolarization from −60 to 0 mV was delivered to the Mb terminal 300 ms before the second light pulse (red trace). No differences in the light-evoked L-IPSCs were noted.

Ai: Both the ON and the OFF portions of light-evoked (λ=505 nm, I=1.95×10¹² photons cm⁻²s⁻¹, 500 ms) L-IPSCs were reduced by the group III mGluR agonist L-AP4 (10–20 µM, red trace). Average traces are shown for each condition (n=5). Aii: Quantification of the effect of L-AP4 on the ON and OFF L-IPSCs charge transfer. L-AP4 significantly reduced ON LIPSC charge (*p<0.02, paired Student’s t-test, two-tailed), but the reduction of OFF L-IPSCs was not significant (p< 0.13, paired Student’s t-test, two-tailed). Measurements, analysis and presentation of drug effects on L-IPSCs are performed this way throughout the paper. Bi: The AMPAR/KAR antagonist NBQX (10 µM, red trace) substantially reduced the OFF L-IPSCs evoked by bright white light (400 ms, I=7.32*10¹³ photons*cm⁻²*s⁻¹), but markedly increased the ON L-IPSCs. Bii: Quantification of the NBQX effect on ON and OFF L-IPSC charge transfers. NBQX significantly reduced the OFF L-IPSCs (**: p<0.0001, paired Student’s t-test, two-tailed), and significantly enhanced the ON L-IPSCs (*: p<0.03, paired Student’s t-test, two-tailed). Ci: L-IPSCs were essentially eliminated by the combination of ionotropic glutamate receptor antagonists NBQX (AMPAR/KAR, 10 µM) and (R)-CPP (NMDAR, 20 µM) (red trace). Cii: Quantification of the effect of combined ionotropic glutamate receptor antagonists (NBQX +(R)-CPP, “N+C”) on the ON and OFF L-IPSCs. (R)-CPP +NBQX significantly reduced both the ON and OFF components of the red light (λ=660 nm, 500 ms, I=2.02*10¹² photons*cm⁻²*s⁻¹) evoked L-IPSCs’ charge (*p<0.05, paired Student’s t-test). D: Summary diagram of normalized effects of pharmacological agents affecting retinal glutamatergic signaling on light-evoked L-IPSCs. Data is presented as mean± SE.

Ai: Both the ON and OFF portions of the light-evoked (λ=505 nm, I=1.95×10¹² photons cm⁻²s⁻¹, 500 ms) L-IPSCs are affected by TTX, albeit differently. The example shown here depicts a cell in which the ON L-IPSCs were increased, whereas the OFF L-IPSCs were suppressed by TTX (2 µM, red trace). Average traces are shown for each conditions (n=5). Aii: Quantification of the TTX effect on the ON and OFF L-IPSC charge transfer (n=6). TTX significantly reduced OFF L-IPSCs charge (*p<0.03, paired Student’s t-test), but the effect of TTX on ON L-IPSCs was not significant (p<0.72, paired Student’s t-test). Note that TTX reduced ON L-IPSCs in 3/6 cells tested, and increased ON L-IPSCs in the other 3/6.

Light-evoked L-IPSCs in Mb terminals are triggered by at least two distinct retinal circuits, providing inhibition of Mb glutamate release with very different temporal characteristics under scotopic and photopic light conditons. R: rod; C: cone; RED arrow: glutamatergic synapse; BLUE arrow: GABAergic synapse.