Catalytic Activity Is Required for Efficient Recycling of Presynaptic ...

Report

DGKq Catalytic Activity Is Required for Efficient Recycling of Presynaptic Vesicles at Excitatory Synapses Graphical Abstract

Authors Hana L. Goldschmidt, Becky Tu-Sekine, Lenora Volk, Victor Anggono, Richard L. Huganir, Daniel M. Raben

Correspondence [email protected] (R.L.H.), [email protected] (D.M.R.)

In Brief Goldschmidt et al. identified a cellular role for the lipid kinase DGKq in the mammalian brain. DGKq was found to localize to excitatory synapses where its activity is required cell autonomously to promote efficient retrieval of synaptic vesicles following sustained neuronal activity.

Highlights d

DGKq functions at excitatory synapses of the mammalian CNS

d

Loss of DGKq slows presynaptic vesicle retrieval following neuronal stimulation

d

DGKq catalytic activity promotes efficient synaptic vesicle recycling in neurons

Goldschmidt et al., 2016, Cell Reports 14, 200–207 January 12, 2016 ª2016 The Authors http://dx.doi.org/10.1016/j.celrep.2015.12.022

Cell Reports

Report DGKq Catalytic Activity Is Required for Efficient Recycling of Presynaptic Vesicles at Excitatory Synapses Hana L. Goldschmidt,1 Becky Tu-Sekine,1 Lenora Volk,2 Victor Anggono,3 Richard L. Huganir,2,* and Daniel M. Raben1,* 1Department of Biological Chemistry, Johns Hopkins University School of Medicine, Hunterian 503, 725 N. Wolfe Street, Baltimore, MD 21205, USA 2Department of Neuroscience, Johns Hopkins University School of Medicine, Hunterian 1001, 725 N. Wolfe Street, Baltimore, MD 21205, USA 3Clem Jones Centre for Ageing Dementia Research, Queensland Brain Institute, The University of Queensland, Brisbane, QLD 4072, Australia *Correspondence: [email protected] (R.L.H.), [email protected] (D.M.R.) http://dx.doi.org/10.1016/j.celrep.2015.12.022 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

SUMMARY

Synaptic transmission relies on coordinated coupling of synaptic vesicle (SV) exocytosis and endocytosis. While much attention has focused on characterizing proteins involved in SV recycling, the roles of membrane lipids and their metabolism remain poorly understood. Diacylglycerol, a major signaling lipid produced at synapses during synaptic transmission, is regulated by diacylglycerol kinase (DGK). Here, we report a role for DGKq in the mammalian CNS in facilitating recycling of presynaptic vesicles at excitatory synapses. Using synaptophysin- and vGlut1pHluorin optical reporters, we found that acute and chronic deletion of DGKq attenuated the recovery of SVs following neuronal stimulation. Rescue of recycling kinetics required DGKq kinase activity. Our data establish a role for DGK catalytic activity at the presynaptic nerve terminal in SV recycling. Altogether, these data suggest that DGKq supports synaptic transmission during periods of elevated neuronal activity. INTRODUCTION Efficient communication between neurons is essential for proper brain function. This process is triggered by Ca2+-influx into presynaptic nerve terminals, resulting in fusion of synaptic vesicles (SVs) with the plasma membrane (exocytosis) and release of neurotransmitters into the synaptic cleft. A typical nerve terminal contains a relatively small number of vesicles, enough to maintain about 5–10 s of neurotransmission. Thus after exocytosis, SVs must be retrieved and recycled by endocy€dhof, 2004). tosis in order to maintain synaptic transmission (Su This becomes particularly critical during periods of elevated neuronal activity, where multiple SVs undergo exocytosis over a short period of time (Cheung et al., 2010). SV recycling is therefore essential for neuronal function, and its dysregulation may contribute to several neurological and psychiatric disorders (Kavalali, 2006). 200 Cell Reports 14, 200–207, January 12, 2016 ª2016 The Authors

Despite its being a well-studied cellular process, the mechanisms that mediate the steps of the SV cycle, particularly those involved in endocytosis, remain a matter of debate. To date, four mechanisms of SV endocytosis have been described: (1) clathrin-mediated endocytosis (CME), (2) activity-dependent bulk endocytosis (ADBE) (Cheung et al., 2010), (3) kiss-and-run €dhof, 2004), and (4) ultra-fast-endocytosis (Watanabe et al., (Su 2013). These pathways are differentially utilized depending on the strength and duration of neuronal activity, as well as differ in their molecular machinery, speed, and capacity for membrane retrieval (Clayton and Cousin, 2009; Kononenko and Haucke, €dhof, 2004; Watanabe et al., 2013; Wu et al., 2014). 2015; Su Numerous proteins regulate SV endocytosis in mammalian central neurons (Haucke et al., 2011). Equally important, the lipid composition of the presynaptic membrane plays an active role in this process. Of the membrane lipids studied so far, phosphoinositides have the most well established role in SV endocytosis (Puchkov and Haucke, 2013; Rohrbough and Broadie, 2005). Phosphatidylinositol-4,5-bisphosphate (PtdIns(4,5)P2) modulates SV recycling by recruiting and activating key molecules, such as synaptotagmin I (Chapman, 2008), clathrin adaptor protein AP2 and dynamin-1 (Burger et al., 2000; Di Paolo et al., 2004) to the presynaptic membrane. Genetic deletions of the lipid kinase (phosphatidylinositol phosphate kinase type Ig, PIPK1g) (Di Paolo et al., 2004), or the lipid phosphatase (synaptojanin 1) (Cremona et al., 1999; Mani et al., 2007) that mediate the generation and metabolism of PtdIns(4,5)P2, respectively, result in multiple synaptic defects, including impaired SV recycling. PtdIns(4,5)P2 is also a substrate for phospholipase C, which produces the signaling lipid, diacylglycerol (DAG). DAG has been implicated in synaptic function and may play at least three roles in the SV cycle (Tu-Sekine and Raben, 2011). First, DAG enhances the activity of Munc13-1, which mediates the priming of SVs, a crucial step in SV exocytosis during spontaneous and evoked synaptic transmission (Augustin et al., 1999; Bauer et al., 2007). Second, DAG activates protein kinase C (PKC), which phosphorylates and thereby regulates the activities of presynaptic SNARE complex proteins, including Munc-18 and SNAP-25 (Di Paolo et al., 2004; Rhee et al., 2002). Finally, termination of DAG signaling through its phosphorylation by DAG kinases (DGKs) results in the production of phosphatidic acid (PtdOH), an acidic phospholipid which can serve as a signaling

molecule itself as well as a precursor for the generation of PtdIns(4,5)P2 (Antonescu et al., 2010; Luo et al., 2004). Despite the importance of DAG and PtdOH in SV recycling, not much is known regarding the roles of DGKs in regulating SV recycling and presynaptic function. Understanding these functions is complicated by the fact there are ten mammalian DGK isoforms (a, b, g, d, ε, z, h, q, i, k), all of which possess the same catalytic activity, are expressed in the CNS, and nine of the ten isoforms are found in neurons (Me´rida et al., 2008; Tu-Sekine and Raben, 2011). Several functional studies have implicated individual DGK isoforms (b, z, ε, h) in modulating spine dynamics, neuronal plasticity, and neurological disorders (Kakefuda et al., 2010; Kim et al., 2010; Musto and Bazan, 2006; Shirai et al., 2010). However, the roles of DGK isoforms localized to the presynaptic terminal are less well understood. Indeed, DGKi is the only isoform that has been implicated in presynaptic release (Yang et al., 2011). The notion that this family of enzymes might play a role in regulating SV recycling is supported by studies in C. elegans that found that DGK-1 activity suppressed acetylcholine release from the neuromuscular junction (McMullan et al., 2006; Nurrish et al., 1999). The ortholog of DGK-1 in mammals is DGKq, the only type V DGK isoform and whose function has never been described in the nervous system. Here, we report that DGKq plays an important role in modulating SV recycling in mammalian central neurons. Both shRNA-mediated knockdown of DGKq and neurons derived from DGKq knockout (KO) mice exhibit a decreased rate of SV endocytosis compared to wild-type neurons. Importantly, this defect can only be rescued by ectopic expression of enzymatically active DGKq but not a kinase-dead mutant of the enzyme. Our data establish a role for DGKq kinase activity in the regulation of SV recycling and suggest that DGKq supports synaptic transmission during periods of sustained neuronal activity. RESULTS DGKq Localizes to Excitatory Synapses in the Mouse Forebrain DGKq expression has been detected in multiple regions of the late embryonic (Ueda et al., 2013) and adult mouse brain (Houssa et al., 1997; Tu-Sekine and Raben, 2011), albeit at low cellular and temporal resolution. In order to study the role of DGKq in brain function, we first examined DGKq expression across multiple brain regions in adult mice. Western blot analysis using an isoform-specific antibody that recognizes the C-terminal region of DGKq revealed wide-spread expression of DGKq protein in all brain regions examined, including the cortex and hippocampus (Figure 1A), which is consistent with the previously reported mRNA expression pattern (Houssa et al., 1997). DGKq was detected in the forebrain but not in cultured glial cells prepared from the same brain region (Figure 1B), indicating a specific expression of DGKq in neuronal cells. The expression of DGKq was upregulated during development, which peaked at postnatal day 14 (P14), coincided with the expression of synaptophysin, an integral SV protein with an established function at synapses (Figure 1C). A similar increase in DGKq protein during synapse forma-

tion and development was also observed in cultured neurons (DIV 7–14, data not shown). Subcellular fractionation analysis of an adult mouse brain showed a general distribution of DGKq in the cytosolic, microsomal, and synaptosomal fractions (Figure 1D). DGKq was also detected in the post-synaptic density fraction, albeit at a lower level, suggesting its presence at both pre- and postsynaptic sites. To verify the biochemical results we examined endogenous DGKq localization by immunofluorescence microscopy. Consistent with fractionation experiments, DGKq was detected throughout the neuron, including the soma, MAP2-positive dendrites, and MAP2-negative axons (Figures 1E and 1F). Strikingly, we found that DGKq had a punctate distribution along dendrites that significantly overlapped with the excitatory presynaptic protein vGlut1 (Figures 1E and 1F). Quantification of the overlap between these two signals showed that 62.6% ± 1.4% of DGKq overlapped with vGlut1 and 56.6% ± 1.5% of vGlut1 overlapped with DGKq per micrometer of dendrite (Figure 1G). In contrast, only 10.0% ± 0.7% of DGKq overlapped with the inhibitory presynaptic protein VGAT per micrometer of dendrite (Figures 1F and 1G), suggesting that DGKq preferentially localizes to excitatory synapses. Knockdown of DGKq Slows the Rate of Synaptic Vesicle Recycling DGKq localization to excitatory synapses and the onset of its expression during synaptogenesis suggested a potential role for this enzyme in excitatory synaptic transmission. Given the role of the DGKq ortholog in C. elegans (McMullan et al., 2006; Nurrish et al., 1999), we predicted that DGKq might possess a similar synaptic function at the presynaptic terminal. To test this, we generated a short-hairpin RNA construct directed against endogenous DGKq mRNA (Figure 2A) to suppress DGKq protein expression. Western blot analysis showed that lentiviral-mediated expression of DGKq-shRNA in cortical neurons resulted in >90% knockdown of DGKq protein compared to control shRNA-infected cells (Figure 2B). To determine the effect of reduced DGKq protein levels on presynaptic function, we used the pH-sensitive optical reporter synaptophysin-pHluorin (sypHy; Granseth et al., 2006) to monitor SV recycling dynamics. Cortical neurons were co-transfected with sypHy and DGKq-shRNA or a control shRNA and allowed to express for 48 hr prior to measuring SV recycling kinetics. Knockdown of DGKq significantly slowed the rate of SV endocytosis following KCl-induced neuronal depolarization compared to control neurons (Figures 2C and 2D; t = 59.2 ± 4.2 s and 32.4 ± 1.9 s, respectively). Moreover, expression of shRNAresistant human DGKq (mycDGKq) was sufficient to rescue the observed defect in endocytosis in neurons expressing DGKqshRNA, thus verifying the specificity of DGKq-shRNA in our study (Figure 2D; t = 34.9 ± 4 s). Importantly, similar results were obtained when neurons were stimulated with a train of 300 action potentials (APs, 10Hz), supporting the role of DGKq in regulating SV recycling in central neurons (Figures 2E, 2F, and S1A). To confirm that the effect of DGKq knockdown was not due to a specific defect in the sorting of a particular SV protein (synaptophysin in this case), we measured SV recycling kinetics using a different pHluorin reporter, vGlut1-pHluorin Cell Reports 14, 200–207, January 12, 2016 ª2016 The Authors 201

A

CTX HIP

CB

OB

E

MB

vGlut1

DGKθ

DGKθ

Tubulin

B

Merge

MAP2

Whole brain

Glia DGKθ

GFAP Synaptophysin

C P14

P25

Synaptophysin

DGKθ

Synaptophysin

PSD-1

SYN

P2

LM

S3

S2

S1

D

PSD-1 sol

Tubulin

Merge MAP2 VGAT DGKθ

P5

DGKθ

G

80 % overlap on dendrites

P0

Merge MAP2 vGlut1 DGKθ

F

60

40

20

DGKθ / vGlut1 vGlut1 / DGKθ DGKθ / VGAT VGAT / DGKθ

Figure 1. DGKq Is an Excitatory Synaptic Protein (A) Protein extracts prepared from adult mouse cortex (CTX), hippocampus (HIP), cerebellum (CB), olfactory bulb (OB), and midbrain (MB) were assayed for DGKq protein expression by western blot using specific antibodies against DGKq and tubulin (loading control). (B) Whole-cell extracts from primary glial cultures and rat forebrain were blotted with specific antibodies against DGKq, GFAP (glial marker), and synaptophysin (neuronal marker). (C) Whole-brain lysates from mice between postnatal day 0 (P0) and 25 (P25) were assayed for DGKq protein expression by western blot using antibodies against DGKq, synaptophysin, and tubulin. (D) Biochemical fractionation of adult mouse brain reveals a wide distribution of DGKq in various subcellular compartments. Synaptophysin and PSD-95 (pre- and postsynaptic protein markers, respectively) were used as controls for successful isolation of synaptic fractions. Post-nuclear supernatant (S1), cytosol after P2 precipitation (S2), cytosol after LM precipitation (S3), light membranes (LM), crude synaptosomes/membranes (P2), synaptosomes (SYN), postsynaptic density (PSD-I), remaining soluble fraction after PSD-I precipitation (PSD-I sol). 20 mg of protein was loaded for each fraction. (E and F) Cultured hippocampal neurons (DIV 28) were immunostained with specific antibodies against DGKq, vGlut1 (excitatory presynaptic protein) or VGAT (inhibitory synaptic marker) and MAP2 (dendritic protein). Scale bars represent 20 mm and 5 mm (cropped region). (G) Quantification of the average overlap between DGKq/vGlut1 and DGKq/VGAT staining on MAP2-positive secondary dendrites. Data represent mean ± SEM (n = 3 independent coverslips).

PSD-95 0

(vGlut1-pH, (Balaji and Ryan, 2007)). Consistent with the sypHy data, the kinetics of vGlut1-pH recycling following a train of 300 APs (10Hz) were also significantly slower in DGKqshRNA expressing neurons compared to controls (Figure S1B). Together, these data indicate that DGKq is involved in the general retrieval mechanism of SVs, rather than regulating the sorting of a specific SV protein. Because the strength of the stimulus can activate distinct recycling mechanisms (Clayton and Cousin, 2009; Kononenko and Haucke, 2015), we speculated that elevated neuronal activity may exaggerate the defect in SV recycling observed in DGKqknockdown neurons. As shown in Figures S1B and S1C, the delay in endocytosis kinetics of sypHy or vGlut1-pH in DGKq knockdown neurons was larger when stimulated with trains of 600 APs (50 Hz) compared to 300 APs (10 Hz). Furthermore, a relatively smaller, but significant, delay in the rate of endocytosis was also observed when neurons were stimulated with a train of 40 APs (20Hz), which is known to mobilize primarily the readily releasable 202 Cell Reports 14, 200–207, January 12, 2016 ª2016 The Authors

pool of SVs (Burrone et al., 2006) (Figure S1D). Taken together, these data demonstrate that DGKq is crucial for maintaining efficient recycling of SVs via distinct pathways of membrane retrieval in a use-dependent manner. DGKq KO Mice Have Reduced DGK Activity in the Brain and Exhibited an Impairment in SV Recycling Efficiency Next, we investigated the role of DGKq with the use of a DGKq homozygous KO mouse (see methods, Figure 3A). The genotypes and the levels of DGKq protein expression in wild-type (WT), heterozygous (Het) and KO mice were confirmed by PCR and western blotting analyses, respectively (Figures 3A and 3B). DGKq KO mice appeared overtly healthy and did not display any obvious gross morphological differences in body size, mating, and lifespan compared to WT mice (data not shown). In addition, we did not observe any compensatory increase in other neuronal DGK isoforms, all of which are capable of catalyzing the production of PtdOH, in brain extracts from DGKq KO mice

N

A RA

PH

D

Ctrl DGKθ-shRNA

0.6 0.4 0.2 0.0 -0.2

0

50

100 150 Time (s)

200

KCl Average post-stimulus endocytic τ (s)

KCl

1.0

250

E Normalized Fluorescence

GAPDH

80 60

1.2 10 Hz, 300 AP 1.0 0.8 0.6

Ctrl DGKθ-shRNA

0.4 0.2 0.0 -0.2

*** ###

Ctrl DGKθ-shRNA myc DGKθ myc DGKθ-rescue

###

40 20 0

F

10 Hz, 300 AP Average post-stimulus endocytic τ (s)

Normalized Fluorescence

C

0.8

G

DGKθ

PR CRD Catalytic domain extCRD

1.2

D

C

trl

DGKθ-shRNA

Kθ

-s

hR

B

A

60

*** 40

***

### ###

Ctrl DGKθ-shRNA myc DGKθ myc DGKθ-rescue

Figure 2. DGKq Regulates the Kinetics of SV Recycling in Cortical Neurons (A) Schematic of the domain structure of DGKq and the relative location of the DGKq-shRNA target. (B) Lysates from cultured cortical neurons infected with either a control (Ctrl) or DGKq-shRNA lentivirus were subjected to western blot analysis with antibodies against DGKq and GAPDH (loading control). (C) Normalized average traces from neurons expressing sypHy with either control (black) or DGKq-shRNA (red) in response to 60 s stimulation with high K+ buffer. (D) Comparison of average t values between control and DGKq-shRNA from (C), mycDGKq (+ctrl shRNA, gray), and mycDGKq-rescue (green) neurons. The decay phases of the traces were fitted with single exponential functions and t values were calculated from the fits. (E) Normalized average traces from neurons expressing sypHy with control or DGKq-shRNA in response to a train of 300 APs (10 Hz). (F) Comparison of average t values between control and DGKq-shRNA from (E), mycDGKq and myc DGKq-rescue neurons in response to the 10 Hz stimulus. For all experiments shown in Figure 2, data represent mean ± SEM from R100 boutons; ***p < 0.001 against Ctrl; ### p < 0.001 against DGKqshRNA, a one-way ANOVA with Tukey’s post hoc test.

20

0 0 40 80 120 160 200 240 280 Time (s)

(Figure 3C). Surprisingly, we detected a significant decrease in the total DGK activity measured in protein extracts from adult KO forebrain compared to WT (Figure 3D; Tu-Sekine and Raben, 2012). These data highlight an important role for DGKq in the production of PtdOH in the brain. To confirm the role of DGKq in regulating the kinetics of SV recycling, we measured the efficiency of sypHy retrieval in neurons derived from WT and DGKq KO mice. Consistent with our previous results, the rate of endocytosis was significantly slower in KO neurons following a train of 600 APs compared to WT controls (Figures 4A and 4B; tWT = 27.9 ± 0.8 s versus tKO = 60.2 ± 1.1 s). The endocytic defect in KO neurons was accompanied by a significant increase in sypHy fluorescence intensity 200 s post-stimulation (Figure S2B), indicating the accumulation of uninternalized SV proteins on the cell surface. We also ruled out the role of DGKq in regulating SV exocytosis in neurons (Figures S2C– S2E). Importantly, expression of mycDGKq completely rescued the endocytic defect in KO neurons (Figure 4B; tmycDGKq = 37.0 ± 1.1 s). Due to the significant reduction of total DGK activity in DGKq KO brain tissue, we predicted that the catalytic activity of DGKq might be essential for promoting efficient recycling

of SVs. To test this hypothesis, we transiently transfected KO neurons with myc DGKq harboring the G648A point mutation that renders the enzyme catalytically inactive (Los et al., 2004) . Consistent with our hypothesis, expression of the mycDGKq kinasedead mutant (mycDGKq-kd) failed to rescue the delay in SV recycling kinetics in KO neurons (Figure 4B; tmycDGKq-kd = 62.1 ± 1.5 s). This was not due to the low expression or mis-targeting of the mutant as mycDGKq-kd colocalized with sypHy along neuronal processes (Figure 4C). Taken together, these data demonstrate that DGKq catalytic activity is necessary for efficient recycling of SVs following neuronal activity. Finally, we evaluated the effects of frequency and duration of stimuli on the rate of endocytosis. Whereas WT neurons displayed comparable endocytic time constants across various neuronal stimuli, DGKq KO neurons displayed a significant augmentation of the SV recycling defect with increases in the strength and frequency of neuronal stimulation (Figures 4D– 4F). Thus, we conclude that during periods of sustained neuronal activity, when more APs are fired, DGKq plays a more critical role in promoting efficient retrieval of SVs. DISCUSSION Despite speculations regarding the roles of mammalian DGKs in synaptic transmission, the function of DGKq in the brain has Cell Reports 14, 200–207, January 12, 2016 ª2016 The Authors 203

A

loxP

FLP

1

lacZ SA

FLP loxP

2

neo

pA

loxP

3

pA

WT HET KO Wild-type 629 bp Floxed 341 bp

WT HET KO

B DGKθ

Tubulin Synaptophysin

KO

WT

C DGKθ DGKδ DGKγ DGKι DGKζ

PSD-95

Synaptophysin

6.0x10-04 Total DGK activity (nmol PtdOH*μg-1*min-1)

D

4.0x10-04

2.0x10-04 ***

0

WT KO

Figure 3. Total DGK Activity Is Reduced in DGKq KO Mice (A) Top: schematic of DGKq KO allele (Dgkqtm1a(KOMP)Wtsi). Exon 1 is spliced to the artificial splice acceptor (SA) in front of lacZ instead of another exon in the DGKq gene. The poly-adenylation site (pA) terminates transcription after lacZ, preventing the transcription of the DGKq RNA. Bottom: typical result of PCR for

204 Cell Reports 14, 200–207, January 12, 2016 ª2016 The Authors

remained unknown. In this study, we tested the hypothesis that DGKq modulates neurotransmitter release at central synapses. We found that DGKq protein expression is elevated during synaptogenesis and localizes specifically to excitatory synapses. Both acute and chronic loss of DGKq slowed SV retrieval following neuronal stimulation. The endocytic defect could be rescued by ectopic expression of WT, but not catalytically inactive DGKq, thus implicating DGKq enzymatic activity in promoting the efficient recycling of SVs following neuronal activity. A potential consequence of reduced synaptic DGK activity observed in DGKq KO mice could be elevated levels of DAG in the plasma membrane. Since functional analogs of DAG are known to potentiate synaptic transmission (Rhee et al., 2002), we hypothesized that the slowed recycling kinetics measured in DGKq KO neurons could be the result of augmented SV exocytosis. However, the rate of SV exocytosis reported by sypHy, assayed in the presence of the vesicular ATPase inhibitor, bafilomycin A1, was essentially identical in WT and KO neurons (Figure S2). These findings suggest that the SV recycling defect observed in DGKq KO neurons is not secondary to altered exocytosis, and argues that DGKq directly regulates the rate of SV endocytosis. If DGKq is regulating a distinct pool of DAG, not relevant for SV exocytosis, it raises the intriguing possibility that it is the PtdOH produced by DGKq, rather than its consumption of DAG, that is crucial for maintaining efficient SV recycling. This notion is corroborated by the fact that DGKq is responsible for generating a significant amount of PtdOH and, potentially PtdIns, in the brain. Interestingly, the role of DGKq becomes much more prominent during intense neuronal stimulation, presumably due to an increase demand of PtdOH production at synapses. Consistent with this notion, previous studies in non-neuronal cells have shown that PtdOH production by DGKs as well as phospholipase D is important for clathrinmediated endocytosis (Antonescu et al., 2010; Kawasaki et al., 2008; Los et al., 2004). Although our data implicate a role for DGK-mediated PtdOH production in SV endocytosis at central synapses, the involvement of a PLD cannot be ruled out. A deeper understanding of the distinct mechanisms used by individual lipid-metabolizing enzymes within the presynaptic terminal and how these pathways are integrated to ensure lipid homeostasis and efficient neuronal function will be critical for understanding the molecular basis of synaptic transmission as well as neurological diseases.

genotyping. Bands at 629bp and 341bp are indicative of DGKq WT and KO alleles, respectively. (B) Western blot analysis of protein extracts prepared from DGKq WT, HET, and KO mouse brain tissue confirms the loss of DGKq protein in KO lysates. Tubulin and synaptophysin blot showed equal loading of protein lysates. (C) Western blot analysis of brain tissue isolated from two pairs of WT and KO mice run on the same gel. Samples were immunoblotted with antibodies against DGKq, -g, -ii, -z, representing four classes of DGKs. Synaptophysin and PSD-95 blots showed equal loading of protein lysates. (D) Average total DGK activity measured in vitro in 5 mg of S1 fractions from five pairs of age-matched WT and KO forebrain tissues. Averages include three technical replicates per sample. Error bars represent SEM. Student’s t test, ***p < 0.0001 against WT.

0.8

80

WT KO

0.6 0.2

*** ***

### ###

###

###

WT WT+ mycDGKθ WT+ mycDGKθ-kd KO KO+ mycDGKθ KO+ mycDGKθ-kd

20

0 40 80 120 160 200240 280 Time (s)

E

10 Hz stimulation WT KO

75

DGKθ KO neurons

60

F

50 Hz stimulation

Average post-stimulus endocytic τ (s)

D

0

WT KO

75 60

45 *

***

15

***

30

***

15

0 0

100 200 300 400 Number of Stimuli

WT KO

75 60

45

30

300 AP at 10 Hz vs 50 Hz


0.0 -0.2


***

60

40

0.4

DGKθ-kd

Merge mCherry sypHy

50 Hz, 600 AP

1.0


Normalized Fluorescence

1.2

myc

C Myc

B

A

45 ***

30

***

15

0 0

200 400 600 800 Number of Stimuli

0 0

20 40 Frequency (Hz)

60

Figure 4. DGKq Enzymatic Activity Is Required for Efficient SV Recycling (A) Normalized average traces from WT (black) and KO (red) neurons expressing sypHy in response to a train of 600 APs (50 Hz). (B) Comparison of average tendo values between WT and KO neurons expressing empty vector from (A), mycDGKq, or kinase-dead DGKq (mycDGKq-kd) in response the 50 Hz stimulus. (C) DGKq KO neurons (DIV20) expressing sypHy, mCherry, mycDGKq-kd were stained with anti-myc (red), anti-GFP (green), and anti-mCherry (blue) antibodies. Merge image (bottom) shows mycDGKq-kd colocalizes with sypHy reporter. Scale bar represents 5 mm. (D–F) Defect in SV recycling is exaggerated with increasing number of stimuli (D and E) and higher frequency stimulation (F). (D) Average t values measured in WT and KO neurons following trains of 100 or 300 APs delivered at 10 Hz. (E) Average t values measured in WT and KO neurons following trains of 300 or 600 APs delivered at 50 Hz. (F) Comparison of the average t values from (D) and (E) following a train of 300 APs delivered at 10 Hz and 50 Hz. For all experiments shown in Figure 4, data represent ± SEM from R200 boutons per condition; ***p < 0.001, against WT, ### p < 0.001 against KO, ANOVA with Tukey’s post hoc test. EXPERIMENTAL PROCEDURES Animals The DGKq KO mouse (Dgkqtm1a(KOMP)Wtsi) was obtained from the KOMP Repository. Sprague Dawley rats were used to generate embryos for neuronal cultures. All animals were treated in accordance with the Johns Hopkins University Animal Care and Use Committee guidelines. Neuronal Culture and Transfection Cortical neurons from E18 rat or E17 mouse pups were plated onto polyL-lysine coated dishes or 18 mm coverslips in Neurobasal growth medium supplemented with 2%B27, 2 mM Glutamax, 50 U/ml penicillin, 50 mg/ml streptomycin, and 5% Horse serum. FDU was added at days in vitro (DIV) 4 and neurons were maintained in glial-conditioned growth medium (1% serum) and fed twice a week. Neurons were transfected at DIV 11-12 using lipofectamine 2000 (Invitrogen) according to manufacturer’s instruction. SypHy and vGlut1-pH live-cell imaging was performed DIV 14–17. Live-Cell Imaging Coverslips containing neurons were mounted into a custom-built perfusion chamber and held at 37 C on the heated microscope stage. Cells were contin-

uously perfused at with pre-warmed ACSF (in millimoles: 122.5, NaCl, 2.5 KCl, 2 CaCl2, 2 MgCl2, 30 D-glucose, 25 HEPES [pH 7.4]) and imaged at 0.5 Hz through a 403 (1.6 NA) oil objective using a Zeiss spinning-disk confocal microscope. SypHy and vGlut1-pH fluorescence was imaged at 488 nm excitation and collected through a 505–550 nm filter, whereas mCherry signal was imaged at 561 nm excitation and 575–615 nm emission. For KCl stimulation (ACSF with 50 mM KCl, 75 mM NaCl), 1 mM tetrodotoxin (TTX) was added to all buffers. For field stimulation, 10 mM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and 50 mM D,L-2 amino-5-phosphonovaleric acid (AP5) were added to the ACSF instead of TTX. APs were delivered using platinum wires embedded in the imaging chamber at 100 mA and 1 ms pulse width. Quantitative imaging analyses were performed with ImageJ using the time-series plugin (Granseth et al., 2006), and the data were fitted using Prism 5 software (GraphPad Software). See Supplemental Experimental Procedures for details.

SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures and two figures and can be found with this article online at http://dx.doi.org/ 10.1016/j.celrep.2015.12.022.

Cell Reports 14, 200–207, January 12, 2016 ª2016 The Authors 205

AUTHOR CONTRIBUTIONS Conceptualization, H.L.G, B.T. and D.M.R, Methodology, H.L.G, V.A., R.L.H. and D.M.R., Investigation, H.L.G, V.A., B.T. and L.V., Writing - Original Draft, H.L.G, R.L.H. and D.M.R., Writing - Review & Editing draft, H.L.G, V.A., B.T., L.V., R.L.H. and D.M.R., Funding Acquisition, H.L.G, V.A., R.L.H. and D.M.R., Supervision, R.L.H. and D.M.R. ACKNOWLEDGMENTS The authors thank Brian Loeper and Rebecca White for assistance in generating DGKq KO mice; Seth Margolis, Michael Wolfgang, and members of the R.L.H. lab for critical discussions. This work was supported by RO1NS077923 (to D.M.R.), T32GM007445 (to H.L.G., BCMB Training Grant), RO1NS036715 (to R.L.H.), and the John T. Reid Charitable Trusts (to V.A.).

Houssa, B., Schaap, D., van der Wal, J., Goto, K., Kondo, H., Yamakawa, A., Shibata, M., Takenawa, T., and van Blitterswijk, W.J. (1997). Cloning of a novel human diacylglycerol kinase (DGKtheta) containing three cysteinerich domains, a proline-rich region, and a pleckstrin homology domain with an overlapping Ras-associating domain. J. Biol. Chem. 272, 10422– 10428. Kakefuda, K., Oyagi, A., Ishisaka, M., Tsuruma, K., Shimazawa, M., Yokota, K., Shirai, Y., Horie, K., Saito, N., Takeda, J., and Hara, H. (2010). Diacylglycerol kinase b knockout mice exhibit lithium-sensitive behavioral abnormalities. PLoS ONE 5, e13447. Kavalali, E.T. (2006). Synaptic vesicle reuse and its implications. Neuroscientist 12, 57–66. Kawasaki, T., Kobayashi, T., Ueyama, T., Shirai, Y., and Saito, N. (2008). Regulation of clathrin-dependent endocytosis by diacylglycerol kinase delta: importance of kinase activity and binding to AP2alpha. Biochem. J. 409, 471–479.

Received: January 14, 2015 Revised: October 27, 2015 Accepted: November 25, 2015 Date: December 31, 2015

Kim, K., Yang, J., and Kim, E. (2010). Diacylglycerol kinases in the regulation of dendritic spines. J. Neurochem. 112, 577–587.

REFERENCES

Los, A.P., van Baal, J., de Widt, J., Divecha, N., and van Blitterswijk, W.J. (2004). Structure-activity relationship of diacylglycerol kinase theta. Biochim. Biophys. Acta 1636, 169–174.

Antonescu, C.N., Danuser, G., and Schmid, S.L. (2010). Phosphatidic acid plays a regulatory role in clathrin-mediated endocytosis. Mol. Biol. Cell 21, 2944–2952. €dhof, T.C., and Brose, N. (1999). Munc13-1 is Augustin, I., Rosenmund, C., Su essential for fusion competence of glutamatergic synaptic vesicles. Nature 400, 457–461. Balaji, J., and Ryan, T.A. (2007). Single-vesicle imaging reveals that synaptic vesicle exocytosis and endocytosis are coupled by a single stochastic mode. Proc. Natl. Acad. Sci. USA 104, 20576–20581. Bauer, C.S., Woolley, R.J., Teschemacher, A.G., and Seward, E.P. (2007). Potentiation of exocytosis by phospholipase C-coupled G-protein-coupled receptors requires the priming protein Munc13-1. J. Neurosci. 27, 212–219. Burger, K.N., Demel, R.A., Schmid, S.L., and de Kruijff, B. (2000). Dynamin is membrane-active: lipid insertion is induced by phosphoinositides and phosphatidic acid. Biochemistry 39, 12485–12493.

Kononenko, N.L., and Haucke, V. (2015). Molecular mechanisms of presynaptic membrane retrieval and synaptic vesicle reformation. Neuron 85, 484–496.

Luo, B., Prescott, S.M., and Topham, M.K. (2004). Diacylglycerol kinase zeta regulates phosphatidylinositol 4-phosphate 5-kinase Ialpha by a novel mechanism. Cell. Signal. 16, 891–897. Mani, M., Lee, S.Y., Lucast, L., Cremona, O., Di Paolo, G., De Camilli, P., and Ryan, T.A. (2007). The dual phosphatase activity of synaptojanin1 is required for both efficient synaptic vesicle endocytosis and reavailability at nerve terminals. Neuron 56, 1004–1018. McMullan, R., Hiley, E., Morrison, P., and Nurrish, S.J. (2006). Rho is a presynaptic activator of neurotransmitter release at pre-existing synapses in C. elegans. Genes Dev. 20, 65–76. Me´rida, I., Avila-Flores, A., and Merino, E. (2008). Diacylglycerol kinases: at the hub of cell signalling. Biochem. J. 409, 1–18. Musto, A., and Bazan, N.G. (2006). Diacylglycerol kinase epsilon modulates rapid kindling epileptogenesis. Epilepsia 47, 267–276.

Burrone, J., Li, Z., and Murthy, V.N. (2006). Studying vesicle cycling in presynaptic terminals using the genetically encoded probe synaptopHluorin. Nat. Protoc. 1, 2970–2978.

Nurrish, S., Se´galat, L., and Kaplan, J.M. (1999). Serotonin inhibition of synaptic transmission: Galpha(0) decreases the abundance of UNC-13 at release sites. Neuron 24, 231–242.

Chapman, E.R. (2008). How does synaptotagmin trigger neurotransmitter release? Annu. Rev. Biochem. 77, 615–641.

Puchkov, D., and Haucke, V. (2013). Greasing the synaptic vesicle cycle by membrane lipids. Trends Cell Biol. 23, 493–503.

Cheung, G., Jupp, O.J., and Cousin, M.A. (2010). Activity-dependent bulk endocytosis and clathrin-dependent endocytosis replenish specific synaptic vesicle pools in central nerve terminals. J. Neurosci. 30, 8151–8161.

Rhee, J.S., Betz, A., Pyott, S., Reim, K., Varoqueaux, F., Augustin, I., Hesse, €dhof, T.C., Takahashi, M., Rosenmund, C., and Brose, N. (2002). Beta D., Su phorbol ester- and diacylglycerol-induced augmentation of transmitter release is mediated by Munc13s and not by PKCs. Cell 108, 121–133.

Clayton, E.L., and Cousin, M.A. (2009). The molecular physiology of activitydependent bulk endocytosis of synaptic vesicles. J. Neurochem. 111, 901–914. €thi, A., Kim, W.T., Takei, K., Daniell, Cremona, O., Di Paolo, G., Wenk, M.R., Lu L., Nemoto, Y., Shears, S.B., Flavell, R.A., et al. (1999). Essential role of phosphoinositide metabolism in synaptic vesicle recycling. Cell 99, 179–188.

Rohrbough, J., and Broadie, K. (2005). Lipid regulation of the synaptic vesicle cycle. Nat. Rev. Neurosci. 6, 139–150. Shirai, Y., Kouzuki, T., Kakefuda, K., Moriguchi, S., Oyagi, A., Horie, K., Morita, S.-Y., Shimazawa, M., Fukunaga, K., Takeda, J., et al. (2010). Essential role of neuron-enriched diacylglycerol kinase (DGK), DGKbeta in neurite spine formation, contributing to cognitive function. PLoS ONE 5, e11602.

Di Paolo, G., Moskowitz, H.S., Gipson, K., Wenk, M.R., Voronov, S., Obayashi, M., Flavell, R., Fitzsimonds, R.M., Ryan, T.A., and De Camilli, P. (2004). Impaired PtdIns(4,5)P2 synthesis in nerve terminals produces defects in synaptic vesicle trafficking. Nature 431, 415–422.

€dhof, T.C. (2004). The synaptic vesicle cycle. Annu. Rev. Neurosci. 27, Su 509–547.

Granseth, B., Odermatt, B., Royle, S.J., and Lagnado, L. (2006). Clathrin-mediated endocytosis is the dominant mechanism of vesicle retrieval at hippocampal synapses. Neuron 51, 773–786.

Tu-Sekine, B., and Raben, D.M. (2011). Regulation and roles of neuronal diacylglycerol kinases: a lipid perspective. Crit. Rev. Biochem. Mol. Biol. 46, 353–364.

Haucke, V., Neher, E., and Sigrist, S.J. (2011). Protein scaffolds in the coupling of synaptic exocytosis and endocytosis. Nat. Rev. Neurosci. 12, 127–138.

Tu-Sekine, B., and Raben, D.M. (2012). Dual regulation of diacylglycerol kinase (DGK)-q: polybasic proteins promote activation by phospholipids and increase substrate affinity. J. Biol. Chem. 287, 41619–41627.

206 Cell Reports 14, 200–207, January 12, 2016 ª2016 The Authors

Ueda, S., Tu-Sekine, B., Yamanoue, M., Raben, D.M., and Shirai, Y. (2013). The expression of diacylglycerol kinase theta during the organogenesis of mouse embryos. BMC Dev. Biol. 13, 35.

Wu, Y., O’Toole, E.T., Girard, M., Ritter, B., Messa, M., Liu, X., McPherson, P.S., Ferguson, S.M., and De Camilli, P. (2014). A dynamin 1-, dynamin 3- and clathrin-independent pathway of synaptic vesicle recycling mediated by bulk endocytosis. eLife 3, e01621.

Watanabe, S., Rost, B.R., Camacho-Pe´rez, M., Davis, M.W., So¨hl-Kielczynski, B., Rosenmund, C., and Jorgensen, E.M. (2013). Ultrafast endocytosis at mouse hippocampal synapses. Nature 504, 242–247.

Yang, J., Seo, J., Nair, R., Han, S., Jang, S., Kim, K., Han, K., Paik, S.K., Choi, J., Lee, S., et al. (2011). DGKi regulates presynaptic release during mGluRdependent LTD. EMBO J. 30, 165–180.

Cell Reports 14, 200–207, January 12, 2016 ª2016 The Authors 207

Cell Reports Supplemental Information

DGK Catalytic Activity Is Required for Efficient Recycling of Presynaptic Vesicles at Excitatory Synapses Hana L. Goldschmidt, Becky Tu-Sekine, Lenora Volk, Victor Anggono, Richard L. Huganir, and Daniel M. Raben

A

B DGKθ

myc


Myc

vGlut1-pH

Merge dsRed sypHy

C

10 Hz, 300 AP

D

50 Hz, 600 AP

60

20 0

20 Hz, 40 AP sypHy

***

Ctrl DGKθ-shRNA

20 0

Figure S1 (related to Figure 2)



40

***

Ctrl DGKθ-shRNA

40

vGlut1-pH sypHy 60

***

30 20 10 0

*

Ctrl DGKθ-shRNA

A

B Average post-stimulus fluorescence (F200-220s)

90 60

ns

30 0

WT KO

0.4 0.3 0.2

***

0.1 0.0

WT KO

D

C

0.8 0.6

10 Hz, 900 AP

0.75 WT KO

0.4 0.2 0.0 -0.2

0

50

100 Time (s)

150

Figure S2 (related to Figure 4)

Maximum Amplitude

Normalized Fluorescence (ΔF/Fo)

1 μM Bafilomycin

E Exocytic time constant (s)

Average baseline fluorescence (Fo)

120

0.60 0.45 0.30 0.15 0.00

WT KO

20 15 10 5 0

WT KO

Supplementary Figure Legends

Figure S1, Related to Figure 2. SV recycling kinetics measured with vGlut1-pHluorin are slower in DGKθ-knockdown neurons. (A) Cultured cortical neurons transfected with mycDGKθ, sypHy, and mCherry were and immunostained with specific antibodies against Myc (mycDGKθ, blue), GFP (sypHy, green) and dsRED (mCherry). The overlap between the Myc and GFP signals along neuronal processes indicates mycDGKθ rescues SV recycling kinetics (Figure 2) in DGKθ-shRNA expressing neurons neurons at synaptic sites. Scale bar, 5µm. (B) Comparison of average post-stimulus endocytic time constants (τ) values measured using the vGlut1-pHluorin reporter (vGlut1-pH) between control (Ctrl, black) and DGKθshRNA (red) in response to a train of 300 APs (10Hz). Data represent mean ± SEM from >300 boutons for each condition. Student’s t test, *** P300 boutons (vGlut1-pH) and >100 boutons (sypHy) for each condition; *** P500 boutons per genotype. (B) Average sypHy fluorescence intensity is significantly higher in DGKθ KO neurons >2 minutes following termination of a train of 600APs (50Hz). Values were first normalized to Fo and the average fluorescence intensities were calculated from 10 frames (20 s) 160 s following the end of the 50 Hz stimulus (F200-220s). Data represent mean ± SEM from >160 boutons per genotype; Student’s t test. (C) DGKθ KO neurons show no significant difference in the kinetics of SV exocytosis compared to WT. Average traces of sypHy fluorescence in WT and KO neurons in response to a train of 900 APs (10Hz) delivered in the presence of bafilomycin A1 (baf). Values were normalized to the average baseline before stimulation (Fo, as in described in (A)). Representative traces from 1 coverslip of each genotype are shown; data represent mean from ± SEM from >100 boutons per genotype. (D) Comparison of the average maximum sypHy amplitudes between WT and KO neurons in response to a train of 900 APs (10Hz) delivered in the presence of bafilomycin. Values were calculated from ΔF/F curves (example traces shown in (C)). There was no significant difference between sypHy amplitudes measured in WT and KO neurons. Data represent mean ± SEM from ≥300 boutons per genotype, Student’s t test.

(E) Exocytic rate constants for WT and KO neurons were determined by first normalizing ΔF/F traces for each ROI to the maximum sypHy amplitude (0-1 scale). Tau values were calculated by fitting the rise-portion of the response during stimulation to a one-phase association curve. Tau values measured in WT and KO neurons were not significantly different. Data represent mean ± SEM from ≥300 boutons per genotype, Student’s t test.

Detailed Supplementary Experimental Procedures

Reagents Imaging reagents: TTX, DL-AP5, and CNQX (Tocris), were resuspended with water to make 1000x stock solutions (1 mM, 50 mM, 10 mM, respectively), aliquots were stored at -20°C until use. Bafilomycin A1 (Calbiochem) was resuspended in DMSO to make 1mM stock solution, aliquots were stored at -20°C until use. Antibodies: mouse antiDGKθ, DGKι, DGKγ from BD biosciences, chicken anti-MAP2 (Novus), mouse anti-PSD95 (NeuroMab), guinea pig anti-vGlut1 (Synaptic systems), mouse β-Tubulin (Sigma), rabbit anti-GluR1 (RLH lab), rabbit anti-dsRed (Clontech), chicken anti-GFP (Abcam). IRDye-conjugated secondary antibodies were used for infrared imaging (Odyssey; LICOR).

DNA constructs (Molecular Biology) shRNA-knockdown and rescue plasmids. Complementary primers containing DGKθ or control shRNA (Ctrl) sequences were annealed and cloned into SacI and XhoI sites of pSuper vector and driven by H1 RNA polymerase III promoter. Ctrl shRNA sequence was specific for luciferase. shRNA targeting sequences: DGKθ-shRNA (Baldanzi et al., 2010), 5'- GTGTACATTTGGACGTCTA -3'; luciferase shRNA, 5'CGCTGAGTACTTCGAAATGTC -3'. For rescue experiments, human DGKθ or DGKθkinase dead (DGKθ-kd, point-mutation G648A in kinase domain (Los et al., 2004)) was subcloned into pRK5-Myc between SalI and NotI restriction sites and driven by CMV promoter (MycDGKθ and MycDGKθ-kd, respectively). vGlut1-pHluorin was subcloned into pRK5 from a lentiviral vector (Beijing University). All constructs were verified by restriction enzyme digest and DNA sequencing. Other DNA plasmids used: synaptophysin-pHluorin (CMV::SypHy A4) was obtained from Addgene (plasmid 24478).

Lentivirus Control and DGKθ-shRNA were subcloned into FuGW lentiviral vector and lentivirus was produced as described previously(Anggono et al., 2011). Briefly, HEK 293 cells were transfected with FuGW, Δ8.9, VSV-G, using lipofectamine 2000. Cells were treated with sodium butyrate 18hrs post-transfection, Virus was collected 8hrs, and 24hrs post treatment. Virus was concentrated, and resuspended in NB, aliquots were frozen and stored at -80C. For DGKθ knockdown experiments, rat cortical neurons were infected with control or DGKθ-shRNA lentivirus DIV 8, and lysates were assayed by western blot DIV 14-16.

Immunocytochemistry Neurons were rinsed with PBS and fixed for 15 minutes at room temperature in PBS containing 4% paraformaldehyde (PFA) and 4% sucrose. Cells were washed 3 times and permeabilized with 0.25%TX-100 in PBS for 10 minutes. Neurons were then blocked with 10% BSA serum in PBS for 1hr at 37oC and incubated in primary antibodies in PBS containing 3% BSA at room temperature. Neurons were washed 5 times before they were incubated with fluorescently labeled secondary antibodies (goat conjugated Alexafluor 647, 568, or 488). Following 5 final washes with PBS, coverslips and mounted onto glass slides using Fluoromount-G (Southern Biotech). Images were obtained using a 510-laser scanning confocal microscope (Zeiss).

DGK activity assay Assay was performed as described previously (Tu-Sekine and Raben, 2012). Large unilamellar vesicles were prepared on the day of assay as follows: stock lipids were

combined, dried under nitrogen and stored under vacuum at 4° C for 2-20 hours to remove residual CHCl3. Lipid films were rehydrated in DGK assay buffer (55 mM Hepes, 100 mM NaCl) for 30 min at 40°C with occasional vortexing and sonication (30 seconds, Branson Sonicator). Vesicles were formed at 37-40C by extrusion through a 0.1µm polycarbonate membrane using an Avanti mini extruder per manufacturer instructions. Lipid composition was POPC:POPE:POPS:DOG at mole fractions of 26:51:15:8, respectively. 5µg of S1 fractions were added to the vesicles and reactions were incubated 30 min at 37C without agitation. The final assay contained: 1.2 mM total lipid in 50 mM Hepes pH 7.5, 1 mM DTT, 1.5 mM MgCl2, 1 mM [γ32 P]ATP (specific activity = 2.5x105 cpm/nmol ATP). Reactions were terminated by addition of chloroform/methanol/1M NaCl (1:2:0.8) (v:v), and phases separated by addition of 1 ml each of CHCl3 and 1M NaCl. The organic phase was washed with 2 ml of 1M NaCl, dried under nitrogen gas, resuspended in CHCl3:MeOH (95:5) and spotted onto a silica gel 60 TLC plate. Phosphatidic acid (PtdOH) was separated from other lipids with chloroform:acetone:methanol:acetic acid:water (10:4:3:2:1) (v:v). The amount of [γ32P]PtdOH was measured by liquid scintillation spectrophotometry in a Wallac 1410 liquid scintillation counter. DGK activity was quantified as nmol PtdOH min-1 µg -1 of protein. Assay conditions used in these experiments are routinely used to measure DGK activity of purified forms of DGKθ, DGKζ, and DGKδ (personal communication from Becky Tu-Sekine). By excluding calcium from the DGK assay buffer, we minimized activity measured in cell fraction from the type I DGK isoforms (-α, -β, and -γ).

PSD preparation Forebrain or whole brain was isolated from adult mouse and frozen on dry ice. Brain was thawed in homogenization buffer (320 mM sucrose, 5 mM sodium pyrophosphate, 1 mM

EDTA, 10 mM HEPES pH 7.4, 200 nM okadaic acid, protease inhibitor cocktail (Roche)) and homogenized 30 times in glass homogenizer. The homogenate was centrifuged at 800xg for 10 minutes at 4oC to yield P1 and S1. S1 was further centrifuged at 10,000xg for 20 minutes at 4oC to yield P2 and S2. P2 was layered on top of a sucrose gradient composed of 0.8, 1, 1.2 mM sucrose buffers, respectively, containing protease and phosphatase inhibitors. The gradient was centrifuged 82,500 xg for 2hrs at 4oC . The layer containing synaptosomes (between 1 and 1.2mM sucrose) was isolated, diluted with 10mM Hepes buffer, pH 7.4, and synaptosomes were pelleted by centrifugation at 150,000 x g for 30 minutes at 4oC. Synaptosomal pellet (SYN) was resuspended in 50mM HEPES pH 7.4, mixed with an equal volume of 1% triton X-100, and incubated with agitation at 4oC for 15 minutes. The PSD was generated by centrifugation at 32,000xg for 20 minutes at 4oC. The final PSD pellet was resuspended in 50mM HEPES pH 7.4 followed by protein quantification and western blot.

Live cell imaging and analysis Neurons were transfected DIV 11-12 using lipofectamine 2000 (Invitrogen) and sypHy live-cell imaging was performed DIV 14-20. Concentration of DNAs used are as follows: 1 µg sypHy or vGlut1-pH, 0.4 µg mCherry, 0.5 µg empty vector (pRK5-Myc) or MycDGKθ, or MycDGKθ-kd; for shRNA-knockdown 1 µg of control or DGKθ-specific shRNA (shLuc and sh#1, respectively) was also included. Coverslips containing neurons were mounted into a custom-built perfusion and stimulation chamber which held at 37°C on the heated microscope stage. Healthy, transfected neurons were identified by mCherry expression. During image acquisition, cells were continuously perfused at 0.5-1.0 mL/min with pre-warmed ACSF. In all experiments, images were acquired for 30 s prior to stimulation to establish a stable baseline.

Images were analyzed in ImageJ (http://rsb.info.nih.gov/ij/) using the time-series plugin (http://rsb.info.nih.gov/ij/plugins/time-series.html). τ values were calculated from fluorescence intensity changes individual boutons (≥3 coverslips per genotype, from ≥3 batches of neurons). Values were normalized to the initial baseline, ΔF/Fo, (Fo = average F values 20 s prior to stimulation) and then normalized to their maximum amplitude (ΔF). τ values were calculated by fitting the decay phase for each ΔF trace (corresponding to 1 bouton) to a single exponential function in Prism 5.

Supplemental References Anggono, V., Clem, R.L., and Huganir, R.L. (2011). PICK1 loss of function occludes homeostatic synaptic scaling. J. Neurosci. 31, 2188–2196. Baldanzi, G., Alchera, E., Imarisio, C., Gaggianesi, M., Dal Ponte, C., Nitti, M., Domenicotti, C., van Blitterswijk, W.J., Albano, E., Graziani, A., et al. (2010). Negative regulation of diacylglycerol kinase theta mediates adenosine-dependent hepatocyte preconditioning. Cell Death and Differentiation 17, 1059–1068. Los, A.P., van Baal, J., de Widt, J., Divecha, N., and van Blitterswijk, W.J. (2004). Structure-activity relationship of diacylglycerol kinase theta. Biochim. Biophys. Acta 1636, 169–174. Tu-Sekine, B., and Raben, D.M. (2012). Dual regulation of diacylglycerol kinase (DGK)θ: polybasic proteins promote activation by phospholipids and increase substrate affinity. Journal of Biological Chemistry 287, 41619–41627.