Doc2 supports spontaneous synaptic transmission by a Ca²⁺-independent mechanism

Abstract

Two families of Ca²⁺-binding proteins have been proposed as Ca²⁺-sensors for spontaneous release: synaptotagmins and Doc2’s, with the intriguing possibility that Doc2’s may represent high-affinity Ca²⁺-sensors that are activated by deletion of synaptotagmins, thereby accounting for the increased spontaneous release in synaptotagmin-deficient synapses. Here, we use an shRNA-dependent quadruple knockdown of all four Ca²⁺-binding proteins of the Doc2 family to confirm that Doc2-deficient synapses exhibit a marked decrease in the frequency of spontaneous release events. Knockdown in synaptotagmin-1 deficient synapses, however, failed to reduce either the increased spontaneous release or the decreased evoked release of these synapses, suggesting that Doc2 does not constitute a Ca²⁺-sensor for asynchronous release. Moreover, rescue experiments revealed that the decrease in spontaneous release induced by the Doc2 knockdown in wild-type synapses is fully reversed by mutant Doc2B lacking Ca²⁺-binding sites. Thus, our data suggest that Doc2’s are modulators of spontaneous synaptic transmission that act by a Ca²⁺-independent mechanism.

INTRODUCTION

At a synapse, three forms of neurotransmitter release are observed: evoked synchronous, evoked asynchronous and spontaneous ‘mini’ release. Synchronous release is triggered by Ca²⁺-binding to synaptotagmins, and represents the dominant release mode, whereas asynchronous release is mediated by Ca²⁺-binding to an as yet unknown Ca²⁺-sensor, and becomes manifest only under certain conditions (Goda and Stevens, 1994; Maximov and Südhof, 2005; Kerr et al., 2008). Spontaneous release is also largely Ca²⁺-dependent (Li et al., 2009; Xu et al., 2009). Confusingly, two Ca²⁺-sensors were proposed to trigger spontaneous release in wild-type synapses: synaptotagmins, suggesting that spontaneous release is simply an extension of evoked synchronous release (Xu et al., 2009), and proteins of the Doc2 family, suggesting that spontaneous and evoked releases are governed by distinct Ca²⁺-sensors (Groffen et al., 2010).

Synaptotagmins and Doc2 proteins are similar in that both contain two homologous C₂-domains, but differ in that the former include an N-terminal transmembrane region, whereas the latter are cytosolic (Orita et al., 1995; Sakaguchi et al., 1995). Each protein family comprises Ca²⁺-binding and Ca²⁺-independent members (8 of 16 synaptotagmins bind Ca²⁺, in particular the paradigmatic syntaptotagmin-1 (Syt1), while four Doc2-like proteins potentially bind Ca²⁺, namely Doc2A, 2B, 2G, and rabphilin). The two protein families exhibit the same overall C₂-domain architecture, and display Ca²⁺-dependent phospholipid- and SNARE-binding activities (Brose et al., 1992; Davletov and Südhof, 1993; Kojima et al., 1996; Groffen et al., 2006 and 2010). Synaptotagmins perform a well-established function as Ca²⁺-sensors for exocytosis, and Doc2 proteins were also shown to activate exocytosis (Orita et al., 1996; Mochida et al., 1998; Hori et al., 1999; Friedrich et al., 2008; Higashio et al., 2008). Consistent with a role for the Doc2 protein family in synaptic exocytosis, knockout (KO) studies suggested that rabphilin (which is closely related to Doc2’s but includes an N-terminal zinc-finger domain absent from other members of this protein family (Fukuda, 2005)) regulates repriming of vesicles for exocytosis (Deak et al., 2006). Strikingly, a recent double KO of Doc2A and Doc2B in neurons uncovered a large decrease in spontaneous release, suggesting that Doc2’s might act as Ca²⁺-sensors for spontaneous release (Groffen et al., 2010; Martens, 2010). Doc2 proteins are also interesting because the Doc2A gene is deleted or duplicated in 16p11.2 copy number variations associated with autism (Shinawi et al., 2010).

The notion that Doc2 proteins may act as Ca²⁺-sensors for spontaneous exocytosis was attractive given their biochemical properties, but surprising since synaptotagmins were previously shown to mediate most of the Ca²⁺-triggering of spontaneous release (Xu et al., 2009). Thus, the question arises how two Ca²⁺-sensors can mediate spontaneous release, and whether one Ca²⁺-sensor is dominant over the other. Moreover, the continued expression of other similar Ca²⁺-binding proteins (Doc2G and rabphilin) in the Doc2A/2B double KO neurons prompts the question whether Doc2 proteins have additional functions that were occluded by the continued presence of these other Ca²⁺-binding proteins.

To address these questions, we developed a lentiviral knockdown (KD) approach that allows quadruple RNAi experiments coupled with rescue controls. Using this approach, we examined synaptic transmission in neurons lacking all Ca²⁺-binding members of the Doc2 family (Doc2A, 2B, 2G, and rabphilin). Our results confirm that suppression of Doc2 expression reduces spontaneous release dramatically (Groffen et al., 2010). However, Ca²⁺-triggered asynchronous release is unimpaired in the KD neurons, and the Doc2 KD phenotype in spontaneous release was fully rescued by expression of a Ca²⁺-binding deficient mutant of Doc2B, suggesting that Doc2 functions in spontaneous release not as a Ca²⁺-sensor, but as a structural support element. Our data thus are consistent with the notion that for spontaneous release, synaptotagmins remain the primary Ca²⁺-sensors under normal conditions.

RESULTS

A lentiviral RNAi system targeting four different mRNAs

To overcome potential functional redundancy among Doc2 protein family members (Doc2A, 2B, 2G, and rabphilin; Fig. 1A (Pang and Südhof, 2010)), we designed a lentiviral RNAi strategy to suppress expression of all four Doc2-like proteins in cultured mouse cortical neurons. We first screened for effective shRNAs that suppress each mRNA by at least 75%, as measured by quantitative rt-PCR of mRNA levels and immunoblotting. We then generated a lentivirus capable of expressing all four effective shRNAs from pol III promoters (the human H1 and U6 promoters), and a rescue construct from a pol II promoter (the ubiquitin promoter; Fig. 1B). Expression of the four shRNAs against Doc2 family proteins yielded good suppression of all targets except for Doc2A, although the KD efficiency was not as high as with lentiviruses expressing only a single shRNA. Thus, to maximize the Doc2A KD, we generated a second lentivirus expressing another Doc2A shRNA (Fig. 1B), and superinfected the cultured cortical neurons with both viruses. This procedure produced ~75% KD of all four targets, allowing us to analyze the effects of such a loss-of-function manipulation (Figs. 1C and 1D).

Figure 1. Knockdown of Doc2 proteins reduces spontaneous ‘mini’ release.

A. Domain structures of Doc2 proteins and rabphilin-3A. Note that rabphilin resembles a Doc2 protein with the characteristic C₂-domains but an extra N-terminal Zinc-finger domain (black dots = predicted Ca²⁺-binding sites).

B. Lentiviral system for KD of all four members of the Doc2/rabphilin protein family (H1 and U6 = human H1 and U6 pol III promoters; Ub = ubiquitin pol II promoter; WRE = woodchuck hepatitis virus regulatory element).

C & D. Measurement of KD efficiency. Cortical neurons cultured from newborn mice were co-infected at 4 days in vitro (DIV4) with a control lentivirus expressing only EGFP (Control), with the two lentiviruses described in panel B either without a rescue cDNA (Doc/Rph KD), or with an shRNA-resistant rescue cDNA encoding Doc2B (KD + Doc2B). Cells were harvested at DIV14, and mRNA levels for the three Doc2 isoforms were measured by quantitative rt-PCR (C), whereas the protein levels for rabphilin were assessed by immunoblotting (D; Synt. = syntaxin; VCP = vasolin-containing protein (loading controls)).

E & F. Representative traces (E) and summary graphs of the frequency (F, left) and amplitude (F, right) of inhibitory mIPSCs monitored in control neurons (Control) and Doc2/ rabphilin KD neurons without (Doc/Rph KD) or with expression of Doc2A or Doc2B rescue cDNA (KD + Doc2A or + Doc2B).

G & H. Same as E & F, except that excitatory mEPSCs were recorded.

I. Representative traces of mIPSCs monitored at different external Ca²⁺-concentrations in cortical neurons infected with control and Doc2/rabphilin KD lentiviruses.

J. Plot of the mIPSC frequency as a function of the external Ca²⁺-concentration.

K. Apparent Ca²⁺-affinity (left, estimated as the EC₅₀ for the mIPSC frequency) and Ca²⁺-cooperativity (right) of spontaneous mIPSCs in control and Doc2/rabphilin KD neurons, calculated by Hill function fits of individual Ca²⁺-titration experiments.

Data shown are means ± SEMs (in C, n=3 culture experiments; in F, H, and K, number of cells/experiments analyzed are shown in bars; n for J = K; statistical analyses for F, H, and K are by Student’s t-test [*=p<0.05 and ***=p<0.001], and for J by two-way ANOVA). See also Figure S1.

Doc2/rabphilin KD reduces spontaneous mini release

Since Doc2B is a proposed Ca²⁺-sensor for spontaneous release (Groffen et al., 2010), we first tested the effect of the Doc2/rabphilin KD on spontaneous miniature inhibitory and excitatory postsynaptic currents (mIPSCs and mEPSCs, respectively). Consistent with observations in Doc2A/2B double KO mice (Groffen et al., 2010), we found that the Doc2 KD reduced spontaneous inhibitory and excitatory mini release by >60% (Figs. 1E–1H), without altering neuronal cell density or synapse numbers and sizes (Fig. S1A).

With any shRNA-mediated KD, off-target effects are a major concern (Alvarez et al., 2006), even if the KD reproduces the KO phenotype (Groffen et al., 2010). To exclude off-target effects, we performed rescue experiments by co-expression of shRNA-resistant Doc2A or Doc2B alongside the shRNAs. Surprisingly, we found that Doc2A expression rescued the impairment of spontaneous mini release in excitatory but not inhibitory synapses in Doc2/rabphilin KD neurons, whereas Doc2B conversely rescued the mIPSC but not the mEPSC phenotype (Figs. 1E–1H).

To determine whether the Doc2/rabphilin KD acts postsynaptically, we transfected the lentiviral vectors, resulting in the expression of the Doc2/rabphilin shRNAs and EGFP in only a few neurons. Electrophysiological recordings from transfected, fluorescent neurons detected no changes in mIPSC frequency (Figs. S1C), suggesting a presynaptic role for Doc2 proteins.

Most spontaneous release is suppressed by BAPTA-AM, suggesting it is largely Ca²⁺-dependent (Li et al., 2009; Xu et al., 2009). To test whether the Doc2/rabphilin KD changes the Ca²⁺-dependence of spontaneous release, we titrated the extracellular Ca²⁺-dependence of the mini frequency. The Doc2/rabphilin KD decreased mini release at all Ca²⁺-concentrations (Figs. 1I and 1J), and produced a small increase in apparent Ca²⁺-affinity, but had no effect on apparent Ca²⁺-cooperativity (Fig. 1K). Thus, the Doc2 KD does not cause a major change in the Ca²⁺-dependence of mini release, but primarily suppresses the amount of release.

Doc2/rabphilin KD does not alter evoked synchronous or asynchronous release

Measurements of synaptic transmission evoked by isolated action potentials showed that the Doc2 KD did not decrease evoked synchronous release (Figs. 2A–2C), consistent with studies in Doc2A/2B double KO mice (Groffen et al., 2010). Moreover, the Doc2 KD did not alter the size of the readily-releasable pool of vesicles as measured by application of hypertonic sucrose (Figs. S2A and S2B).

Figure 2. Doc2/rabphilin KD does not change evoked synchronous or asynchronous release.

A–C. Representative traces (A) and summary graphs of the amplitude (B) and charge transfer (C) of IPSCs evoked by isolated action potentials in control neurons and Doc2/rabphilin KD neurons without or with Doc2B rescue. Neuronal KDs were performed as described for Figure 1C.

D–G. Representative traces (D), total charge transfer (E), charge transfer during delayed release (F), and the ratio of delayed to total release measured by charge transfer (G) of IPSCs evoked by a 10 Hz stimulus train applied for 1 s.

H & I. Representative traces (H) and the frequency (I, left) or amplitude (I, right) of mIPSCs monitored in cortical neurons cultured from Syt1 KO mice and infected with control lentivirus or the Doc2/rabphilin KD lentiviruses without or with Doc2B rescue as described in Figure 1C. For mEPSCs, see Fig. S2.

J & K. Representative traces (J) and total charge transfer (K) of IPSCs evoked by a 1 sec 10 Hz stimulus train in Syt1 KO neurons infected with control or KD lentiviruses as above.

Data shown are means ± SEMs (number of cells/experiments analyzed are shown in the bars; Student’s t-test failed to detect significant differences). See also Figure S2.

Since Doc2 proteins may have a higher apparent Ca²⁺-affinity than synaptotagmins (Groffen et al., 2010; McMahon et al., 2010), it is possible that they act as Ca²⁺-sensors for asynchronous release. To explore this possibility, we first measured the effect of the Doc2/rabphilin KD on delayed release, a form of asynchronous release that can be assessed following a 10 Hz stimulus train (Maximov and Südhof, 2005). We observed a ‘trend’ for decreased delayed release (Figs. 2D–2G). This trend, however, was not significant, prompting us to study asynchronous release further using cortical neurons from Syt1 KO mice in which synchronous release is absent (Geppert et al., 1994). In these mice, spontaneous mini release exhibits a paradoxical increase with a dramatically altered apparent Ca²⁺-dependence (Xu et al., 2009), and delayed release is enhanced (Maximov and Südhof, 2005), suggesting that Syt1 not only functions as Ca²⁺-sensor for spontaneous and evoked release, but also as a clamp for secondary Ca²⁺-sensors that mediate different forms of spontaneous and evoked release. Thus, we investigated the possibility that Doc2’s represent secondary Ca²⁺-sensors that become activated in Syt1 KO neurons, and may mediate these different forms of Ca²⁺-triggered release.

We found that the Doc2/rabphilin KD had no significant effect on spontaneous mini release in Syt1 KO neurons, suggesting that the KD effect requires Syt1 and that Doc2’s do not operate as the secondary Ca²⁺-sensors for the enhanced spontaneous release activated by the Syt1 KO (Figs. 2H, 2I, and S2C–S2F). Since the high mini release rates in Syt1 KO neurons may saturate the response, we also measured the effect of the Doc2 KD on mini frequency at a lower Ca²⁺-concentration (0.5 mM), but again failed to observe a change (Figs. S2G and S2H). Moreover, we examined the effect of the Doc2/rabphilin KD on evoked asynchronous release in Syt1 KO neurons, but again did not detect an impairment (Figs. 2J, 2K, S2I, and S2J). Thus, Doc2 proteins are not required for the increased spontaneous or asynchronous release in Syt1 KO neurons; the selective effect of the Doc2 KD on spontaneous release in wild-type but not Syt1 KO synapses reinforces the notion that spontaneous release in these two preparations represents distinct processes.

Generation of Ca²⁺-binding site mutants of Doc2B

Based on sequence alignments (Fig. S3) and the well-characterized Ca²⁺-binding sites of the Syt1 and rabphilin C₂-domains (Chen et al., 2002; Ubach et al., 1998 and 1999), the Doc2B C₂A- and C₂B-domains are predicted to bind two Ca²⁺-ions each (Fig. 3A). To test the functional role of Ca²⁺-binding to Doc2, we produced mutants of the Doc2B C₂-domains in which three of the five aspartate residues that ligate the Ca²⁺-ions have been exchanged for alanines (Fig. S3), analogous to similar mutations that block Syt1 function (Shin et al., 2009). To ensure that the mutant C₂-domains still folded properly, we purified them as recombinant proteins, and measured their circular dichroism spectra (Figs. 3B and 3C). The wild-type and mutant C₂A- and C₂B-domains exhibited similar characteristic β-sheet spectra, indicating that they were well folded.

Figure 3. Ca²⁺-binding deficient mutant C₂-domains of Doc2B are folded.

A. Schema of the predicted Ca²⁺-binding sites of the Doc2B C₂A- (left) and C₂B-domain Ca²⁺-binding sites (right), based on the atomic structures of the rabphilin C₂B-domain and the Syt1 C₂A- and C₂B-domains (Fernandez-Chacon et al., 2001; Fernandez et al., 2001; Ubach et al., 1999). Aspartate and glutamate residues involved Ca²⁺-binding are boxed; residues substituted for alanines in the Ca²⁺-binding site mutants are shown on a black background.

B. Purified wild-type or mutant Doc2B C₂-domains. The mutant Doc2B C₂A- and C₂B-domains contain three alanine substitutions each in critical Ca²⁺-binding residues (C2A^3A: D163A, D218A and D220A; C2B^3A: D303A, D357A and D359A).

C. Circular dichroism spectra of wild type and mutant Doc2B C₂-domains.

D & E. Ca²⁺-titration of intrinsic tryptophan fluorescence of wild-type and mutant Doc2B C₂B-domains. Panel D depicts fluorescence spectra of the wild-type (left) and mutant Doc2 C₂B-domain (right) as a function of increasing concentrations of free Ca²⁺, followed by addition of excess EGTA (5 mM) to remove bound Ca²⁺. Panel F plots the fluorescence changes (ΔF/F_max) as a function of the free Ca²⁺-concentration on a semi-logarithmic scale (note that 0 Ca²⁺ is plotted here at ~10⁻⁵ μM for illustration purposes).

Data are representative experiments (C-E) or means ± SEMs (F, n=3 independent experiments). See also Figure S3.

Since Ca²⁺-binding to Doc2 C₂-domains has not been directly measured, and it is uncertain whether Ca²⁺-binding to these C₂-domains is blocked in the mutations we introduced, we examined Ca²⁺-binding to the wild-type and mutant C₂B-domain. In these measurements, we took advantage of a tryptophan residue adjacent to the predicted Ca²⁺-binding site (W356), and monitored the intrinsic tryptophan fluorescence of the recombinant wild-type and mutant C₂B-domain as a function of Ca²⁺ (Fig. 3D). Similar to the C₂B-domain of rabphilin (Ubach et al., 1999), addition of Ca²⁺ quenched the intrinsic tryptophan fluorescence of wild-type but not of mutant C₂B-domain protein, demonstrating that the former but not the latter bound Ca²⁺. Plots of the titrations suggested a low micromolar intrinsic Ca²⁺-affinity of the C₂B-domain (Fig. 3E), consistent with indirect biochemical measurements suggesting that Doc2 proteins exhibit a higher apparent Ca²⁺-affinity than Syt1 (Groffen et al., 2010). Note that we chose to target intrinsic Ca²⁺-binding here instead of a secondary Ca²⁺-dependent binding property of Doc2B, such as phospholipid binding, in order to ensure that the mutation would block all Ca²⁺-dependent functions of Doc2B, and not just one particular property.

Ca²⁺-binding to Doc2B is not required for rescuing mini release

In a final set of experiments, we tested whether rescue of the decrease in spontaneous release induced by the Doc2/rabphilin KD requires Ca²⁺-binding to Doc2B. Surprisingly, mutant Doc2B in which all Ca²⁺-binding sites were inactivated by mutations of the aspartate Ca²⁺-ligands in both C₂-domains fully reversed the >60% decrease in mini frequency induced by the Doc2 KD (Figs. 4A and 4B), suggesting that Doc2B acts in spontaneous release not as a Ca²⁺-sensor, but as a structural element supporting continued supply of vesicles for spontaneous exocytosis.

Figure 4. Ca²⁺-binding deficient Doc2B rescues the decrease in miniature IPSC frequency in Doc2 KD neurons.

A & B. Representative traces (A) and summary graphs of the frequency (B, left) and amplitude (B, right) of mIPSCs monitored in control neurons (Control) and Doc2/rabphilin KD neurons without (Doc/Rph KD) or with expression of mutant Doc2B (KD + Doc2B^6A) in which all Ca²⁺-binding sites were ablated (see Figure 3A).

C. Representative traces of mIPSCs monitored at different external Ca²⁺-concentrations in cortical neurons infected with control lentivirus and Doc2/rabphilin KD lentiviruses without or with expression of mutant Doc2B^6A rescue cDNA.

D. Plot of the mean absolute (left) and normalized mIPSC frequency (right) as a function of the external Ca²⁺-concentration. mIPSCs were monitored in control infected neurons and Doc2/rabphilin KD neurons without and with rescue with mutant Doc2B^6A as described in C.

E. Apparent Ca²⁺-affinity (left, estimated as the EC₅₀ for the mIPSC frequency) and Ca²⁺-cooperativity (right) of spontaneous mIPSCs in control and Doc2/rabphilin KD neurons without or with rescue by mutant Doc2B^6A, as calculated from Hill function fits of individual Ca²⁺-titration experiments.

Data shown are means ± SEMs (number of cells/experiments analyzed are shown in the bars; in D, n corresponds to E; statistical analyses for B and E are by Student’s t-test [*=p<0.05 and ***=p<0.001], and for D by two-way ANOVA).

The unexpected rescue of the reduced mini frequency by mutant Doc2B in Doc2/rabphilin KD neurons could potentially be due to a shift in the Ca²⁺-dependence of spontaneous release, i.e. by activation of the secondary Ca²⁺-sensor that mediates spontaneous release in Syt1 KO synapses (Xu et al., 2009). To address this possibility, we titrated the Ca²⁺-dependence of mini release in Doc2-deficient neurons without or with rescue with mutant Doc2B (Figs. 4C and 4D). Strikingly, mutant Doc2B not only rescued mini release at all Ca²⁺-concentrations, but even slightly enhanced it (Fig. 4D), and reversed the small increase in apparent Ca²⁺-affinity observed in the Doc2 KD neurons (Fig. 4E). Thus, mutant Doc2B is fully active in this functional assay.

DISCUSSION

Spontaneous ‘mini’ release likely mediates important information transfer, and may be mechanistically distinct from evoked release (Sara et al., 2005; Fredj and Burrone, 2009; Stacey and Durand, 2000; Sutton et al., 2006). Most spontaneous release is Ca²⁺-dependent, and controlled by at least two different Ca²⁺-sensors: a low-affinity, high-cooperativity Ca²⁺-sensor in wild-type synapses, and a high-affinity, low-cooperativity Ca²⁺-sensor in synaptotagmin- or complexin-deficient synapses (Xu et al., 2009; Yang et al., 2010). For wild-type synapses, two Ca²⁺-sensors for spontaneous release were proposed: synaptotagmins (Xu et al., 2009) and Doc2A and Doc2B (Groffen et al., 2010). No candidate Ca²⁺-sensor exists for mini release in synaptotagmin-deficient synapses, although this Ca²⁺-sensor may be the same as that for asynchronous release, analogous to the proposed role of synaptotagmin as a Ca²⁺-sensor for both spontaneous and synchronous release in wild-type synapses. Both synaptotagmin and Doc2 are attractive Ca²⁺-sensor candidates for spontaneous release based on their biochemical properties, but only for synaptotagmin is there evidence linking changes in Ca²⁺-binding affinity to changes in spontaneous release (Xu et al., 2009). Here, we have examined the potential role of Doc2 proteins as Ca²⁺-sensors in spontaneous release, and their relation to asynchronous release. In doing so, we strove to avoid potential problems caused by the expression of four closely related isoforms of Doc2 proteins, which could produce functional redundancy, and developed an approach that allowed simultaneous KD of four different targets with a rescue control (Figs. 1A and 1B).

Our data confirm KO studies showing that Doc2 proteins are essential for normal mini release – in fact, the degree of impairment in spontaneous release we observed with a 75% KD of all four isoforms (Figs. 1 and S1) is strikingly similar to that described for the Doc2A and 2B double KO (Groffen et al., 2010). We show that in Doc2 KD synapses, the apparent Ca²⁺-dependence of mini release exhibits a small but significant increase (Fig. 1), but that otherwise no change in Ca²⁺-triggering of either spontaneous or evoked release is detected (Fig. 2). Moreover, our results indicate that the Doc2 KD does not alter synchronous or asynchronous evoked release, and – importantly – does not impair the enhanced spontaneous release detected in Syt1 KO synapses (Fig. 2). This latter result confirms the notion that spontaneous release events in Syt1 KO and wild-type neurons are qualitatively different, consistent with their distinct Ca²⁺-dependence (Xu et al., 2009).

To test the role of Ca²⁺-binding to Doc2 in spontaneous release, we generated Doc2 mutants unable to bind Ca²⁺ (Fig. 3). Rescue experiments surprisingly revealed that mutant Doc2B lacking functional Ca²⁺-binding sites was fully capable of rescuing the decrease in mini frequency induced by the Doc2 KD, and also rescued the altered apparent Ca²⁺-affinity of mini release (Fig. 4). Thus, Doc2 is unlikely to function as a Ca²⁺-sensor for mini release, but rather acts in a structural, Ca²⁺-independent role to maintain spontaneous mini release, consistent with a special status of spontaneous release (Sara et al., 2005; Fredj and Burrone, 2009).

Our results appear to contradict those of Groffen et al. (2010) who did not use mutations blocking Ca²⁺-binding to Doc2B to test its role in mini release, but other point mutations that supported a Ca²⁺-sensor role for Doc2B in mini release. However, this apparent contradiction can be explained if one considers our current understanding of C₂-domains. Groffen et al. examined a gain-of-function mutation in the Ca²⁺-binding mutations of the Doc2B C₂A-domain that was modeled after a similar mutation in Syt1 (Pang et al., 2006; Stevens and Sullivan, 2003), and was also independently tested for Doc2B in chromaffin cells (Friedrich et al., 2008). The fact that this mutation increases mini release in synapses does not necessarily mean that Doc2B is a direct Ca²⁺-sensor for release, but could equally change its structural role in mini release, especially since no correlation of a change in Ca²⁺-affinity of Doc2B with that of mini release, as documented for Syt1 (Xu et al., 2009), was reported. Thus, it seems likely that Doc2 proteins are evolutionarily novel effectors for spontaneous mini release which may have additional, as yet uncharacterized Ca²⁺-dependent functions.

METHODS

Generation of RNAi vectors

All shRNA expression, with and without rescue, was performed with the same lentiviral vector system (Pang et al., 2010; see Figure 1B for the schematic diagram of vectors). Oligonucleotide sequences are described in Suppl. Materials.

Production of Recombinant Lentiviruses was achieved by transfection of HEK293T cells using FuGENE-6 (Roche) as described (Pang et al., 2010; see Suppl. Materials).

Neuronal Cultures and Immunocytochemistry

Cortical neurons were cultured from neonatal wild-type or Syt1 KO mice as described (Pang et al., 2010), infected at DIV5, and analyzed at DIV14-16 (see Suppl. Materials for detailed descriptions).

Electrophysiological recordings were performed using whole-cell recordings and concentric extracellular stimulation electrodes (Maximov et al., 2007; see Suppl. Materials).

Purification and biophysical analyses of recombinant proteins were performed as described in the Suppl. Materials.

Miscellaneous

Immunocytochemistry and immunoblotting were performed as described (Chubykin et al., 2007).