Supplementary Information D-Serine regulates cerebellar LTD and motor coordination through the 2 glutamate receptor Wataru Kakegawa, Yurika Miyoshi, Kenji Hamase, Shinji Matsuda, Keiko Matsuda, Kazuhisa Kohda, Kyoichi Emi, Junko Motohashi, Ryuichi Konno, Kiyoshi Zaitsu & Michisuke Yuzaki Supplementary Figures Supplementary Figure 1 Little change in the PPF value and the EPSC kinetics of PF-EPSCs by application of D-Ser. (a) Representative EPSC responses to the paired-pulse stimulation of PFs before (Pre) and 30 min after (Post) the application of D-Ser (200 M for 10 min) in immature WT cerebellar slices. Both traces were normalized using the amplitude of the first PF-EPSCs (Normalized). (b d) Averaged data showing the PPF value (b), the 10 90% rise time (c), and the decay time constant (d) of the PF-EPSCs before and 30 min after D-Ser application. P values were obtained using the paired Student s t test. Data represent the means SEM. 1
Supplementary Figure 2 PF-EPSC consists of the AMPA component. (a,c,e,g) Representative PF-EPSC traces recorded from immature wild-type Purkinje cell by application of single PF stimulation (a) or burst PF stimulation identical to the LTD induction protocol (30 [10 PF stimuli at 50 Hz] at 1Hz; c,e,g). Recordings were performed at 80 mv in a,c,e and +60 mv in g. To detect the NMDA receptor component at 80 mv, Mg 2+ was removed from extracellular and patch-pipette solutions (Mg 2+ -free) in e. (b,d,f,h) Quantitative data of the PF-EPSC amplitude and charge transfer before (control, open columns) and after the application of NMDA receptor blockers (100- M D-AP5/25 M MK801, red columns) or an AMPA receptor blocker (50- M GYKI52466, green columns), and NMDA receptor and AMPA receptor 2
blockers (100- M D-AP5/25- M MK801/50- M GYKI52466, blue columns). PF-EPSCs were almost completely blocked by an AMPA receptor blocker, GYKI52466. Data represent the means SEM. 3
Supplementary Figure 3 D-Ser causes reduction of PF-EPSCs in immature GluD2-null/Tg but not GluD2-null mice in the absence of NMDA receptor blockers. (a c) Averaged data showing the effect of D-Ser on PF-EPSC in immature GluD2-null (a), GluD2-null/TgWT (b) or GluD2-null/TgR/K (c) cerebellar slices. The insets show PF-EPSCs observed just before (black traces) and 30 min after (gray traces) the application of D-Ser (200 M; 10 min during time 0 10 min). The data in the presence of NMDA receptor blockers were taken from Fig. 1. P values were obtained using the Mann Whitney s U test. Data represent the means SEM. 4
Supplementary Figure 4 Immunoblot analysis showing GluD2 Tg protein expression level in each immature transgenic mouse. (a) Band intensities of the GluD2 protein in 10 g of immature GluD2-null/TgWT (orange circle), GluD2-null/TgR/K (blue circle) and GluD2-null/Tg CT7 (green circle) cerebellar cell lysates. To compare with the WT GluD2 protein, various amounts of immature WT cerebellar lysates (10, 4, 2, 1, 0.4 g) were simultaneously loaded (open circles; slope = 25,824 using the linear least-squares fitting method). The X-axis shows the protein amount of the WT cerebellar cell lysates. (b) Quantitative data for the GluD2 Tg protein expression level in each transgenic mouse (compared with WT GluD2 protein expression). ** P < 0.01, * P < 0.05; ns, Not significant (ANOVA/Bonferroni correction for multiple comparison). Data represent the means SEM. 5
Supplementary Figure 5 Input output relationship of PF-EPSC amplitudes from each mouse. (a,b) The data of PF-EPSC amplitudes recorded from coronal cerebellar slices in immature (a) or mature (b) mice. The insets show representative PF-EPSC traces from GluD2-null/TgWT (orange traces) and GluD2-null/TgR/K (blue traces) mice. Both GluD2 Tg proteins completely rescued impaired PF synaptogenesis in immature and mature GluD2-null mice (GluD2-null/TgWT vs. GluD2-null/TgR/K; P = 0.980 and P = 0.697 at immature and mature stages, respectively). Stimulus duration, 10 s. Data represent the means SEM. 6
Supplementary Figure 6 A single CJ-stim is sufficient to induce a maximal cerebellar LTD. CJ-stim (30 [PF stimulation plus Purkinje cell depolarization from 60 mv to 20 mv] at 1 Hz, which was used in Fig. 2) was continuously applied at the 0- and 30-min time points (arrows). The insets show representative PF-EPSCs observed just before the first CJ-stim (1) and just before (2) and 30 min after (3) the second CJ-stim. Data represent the means SEM. 7
Supplementary Figure 7 Micro 2D-HPLC analysis with fluorescence detection. (a) Structures of the NBD-amino acids to be analyzed. (b) Flow diagram of the 2D-HPLC system for determining the Ser enantiomers. CS, column-switching valve; D, fluorescence detector; P, pump. Using a reversed-phase micro-ods column (red), a fraction of NBD-Ser (a mixture of D and L forms) was isolated and automatically introduced to an enantioselective column (blue) to separate and determine D- and L-Ser. 8
Supplementary Figure 8 DAAO selectively degraded D-Ser. Representative 2D-HPLC spectrograms measured from extracellular solutions superfusing immature cerebellar slices with PF stimulation in the absence [DAAO ( ); black trace] or presence [DAAO (+); red trace] of DAAO treatment (0.125 U/mL). Before the application of PF burst stimulation, slices were preincubated with DAAO for at least 60 min, which is a similar experimental condition with Fig. 4b,c. Filled and open circles represent the signal peaks for D- and L-Ser, respectively. 9
Supplementary Figure 9 PF burst stimulation increased the [Gly] in the extracellular solution in WT cerebellar slices. PF burst stimulation (PF burst, 30 [10 application of PF stimuli at 50 Hz] at 1 Hz at 10 different points) was applied to immature (top) and mature (bottom) WT cerebellar slices. Extracellular [Gly] was calculated from the fluorescence intensity of the signal peak for Gly in 2D-HPLC spectrograms. *** P < 0.001, ** P < 0.01 (Mann Whitney s U test). Data represent the means SEM. 10
Supplementary Figure 10 Expression pattern of serine racemase in the cerebellum of a wild-type mouse. Immnunostaining using antibodies against serine racemase (red), Calbindin (a Purkinje cell-positive marker; green in upper panels) or 3-PGDH (a glial cell marker; green in lower panels) was performed in wild-type cerebellar slices. The asterisks represent the soma of Purkinje cells. 11
Supplementary Figure 11 Characterization of LTD in immature cerebellar slices. (a d) Minimal changes in presynaptic functions during cerebellar LTD. Representative continuous plots of PF-EPSC amplitude (a) and PPF value (b) during cerebellar LTD recording from a WT cerebellar slice. To estimate the PPF value, EPSCs were evoked by stimulating PFs using paired-pulse stimulation (50-ms inter-stimulus interval). The inset traces show representative PF-EPSCs observed just before (1), 2 min (2) and 30 min after (3) CJ-stim. Averaged data of PF-EPSC 12
amplitudes (c) and PPF values (d) obtained during recordings of cerebellar LTD. *** P < 0.001, ** P < 0.01, * P < 0.05 (Mann Whitney s U test). (e) Cerebellar LTD is dependent on intracellular Ca 2+. LTD recorded from immature WT Purkinje cells loaded with the Ca 2+ chelator BAPTA (20 mm) in the patch pipettes. The control data was taken from Fig. 4d. (f) Cerebellar LTD is caused by endocytosis of postsynaptic AMPA receptors. LTD recorded from immature WT Purkinje cells loaded with AP2-specific binding peptide (pep- A849 Q853; 500 M; filled circles) or the control peptide (pep-k844a; 500 M; open circles) in the patch pipettes. CJ-stim was applied at the 0-min time point (arrow). The inset traces show representative PF-EPSCs obtained just before (black traces) and 30 min after (gray traces) CJ-stim for each condition. The P value was obtained using the Mann Whitney s U test. Data represent the means SEM. 13
Supplementary Figure 12 LTD from immature WT cerebellar slices treated with heat-inactivated DAAO. Cerebellar slices were treated with heat-inactivated DAAO (100 C for 10 min; preincubation for at least 60 min and perfusion during recordings), and cerebellar LTD was induced by adding CJ-stim (arrow). The inset shows representative PF-EPSCs observed just before (black trace) and 30 min after (gray trace) CJ-stim. Data represent the means SEM. 14
Supplementary Figure 13 Gly-mediated PF-EPSC rundown is unaffected by DAAO treatment. Exogenous Gly (1 mm, during 0 10 min) was applied to immature WT cerebellar slices with (filled circles) or without (open circles) DAAO treatment (0.125 U/mL; preincubation for at least 60 min and perfusion during recordings). In this experiment, strychnine (1 M) was always added to the extracellular solution to block glycinergic inhibitory input. The insets show PF-EPSC traces just before (black trace) and 30 min after (gray trace) Gly application. The P value was obtained using the Mann Whitney s U test. Data represent the means SEM. The products of D-Ser degradation by DAAO did not inhibit PF-EPSC rundown. 15
Supplementary Figure 14 D-Ser-mediated PF-EPSC rundown in immature WT cerebellar slices treated with NaFAC or NASP. Exogenous D-Ser (200 M, during 0 10 min) induced a reduction in PF-EPSC in the absence (control; open circles) or presence of NaFAC (gray circles) or NASP (filled circles) treatment in immature WT cerebellar slices. The insets show PF-EPSC traces just before (black traces) and 30 min after (gray traces) D-Ser application in each condition. P values were obtained using the ANOVA. Data represent the means SEM. 16
Supplementary Figure 15 Cerebellar LTD recorded from mature DAAO +/+ and DAAO / mice in the absence of NMDA receptor blockers. CJ-stim was applied to mature DAAO +/+ (open circles) and DAAO / (filled circles) cerebellar slices at the 0-min time point (arrow). The inset shows representative PF-EPSCs observed just before (black traces) and 30 min after (gray traces) CJ-stim. The P value was obtained using the Mann Whitney s U test. Data represent the means SEM. 17
Supplementary Figure 16 7-ClKA blocks LTD in immature WT cerebellar slices. 7-ClKA (200 M) was included in the extracellular solution during recordings in the absence (control, open circles) or presence (NMDAR blockers, filled circles) of NMDA receptor blockers. CJ-stim was applied at the 0-min time point (arrow). The inset shows representative PF-EPSCs observed just before (black traces) and 30 min after (gray traces) CJ-stim. Data represent the means SEM. 18
Supplementary Figure 17 D-Ser GluD2 interaction that triggers cerebellar LTD in developing mice. In the basal state, glutamate (Glu) released from PF terminus activates AMPA receptors expressed on Purkinje cell spines to elicit PF-EPSC (Pre; a). Following neuronal activities, endogenous D-Ser is released mainly from Bergmann glia and interacts with GluD2 to convey signals for the endocytosis of AMPA receptors via its cytoplasmic C-terminus (LTD induction; b). After that, PF-EPSC is reduced by the decreased number of postsynaptic AMPA receptors, which accompanies cerebellar LTD (Post; c). To simplify the scheme, the contribution of NMDA receptors is omitted. 19