Type II phosphatidylinositol 4-kinase regulates nerve terminal growth and synaptic vesicle recycling

Abstract

Type II phosphatidylinositol 4-kinase (PI4KII) is thought to be associated with synaptic vesicles (SVs) and to be responsible for the majority of PI4K activity in the nervous system. However, the function of PI4KII at the synapse is unknown. We characterized the synaptic phenotypes of a Drosophila melanogaster PI4KII null mutant. We found increased nerve terminal growth in PI4KII null mutants indicating that PI4KII restrains nerve terminal growth. Evoked neurotransmitter release elicited in response to low frequency stimulation and spontaneous neurotransmitter release were not altered in PI4KII null mutants. However, PI4KII null mutants displayed reduced FM1-43 uptake in response to stimulation by high K⁺ saline, indicating impaired SV endocytosis. PI4KII null mutants did not display any defects in FM1-43 unloading, consistent with normal SV exocytosis. Thus, PI4KII is required for SV endocytosis but dispensable for SV exocytosis. Overall, our data show that PI4KII regulates both nerve terminal growth and SV recycling.

Keywords:

• Presynaptic
• Drosophila
• endocytosis
• phosphoinositides

Introduction

Phosphoinositides affect a vast array of processes at synapses (reviewed in Frere, Chang-Ileto, & Di Paolo, 2012; Lauwers, Goodchild, & Verstreken, 2016). Phosphatidylinositol (4,5)-bisphosphate (PI(4,5)P₂) regulates axonal growth (Khuong, Habets, Slabbaert, & Verstreken, 2010), ion channel function (Suh, Leal, & Hille, 2010; Suh & Hille, 2002) and several steps of the synaptic vesicle (SV) cycle (Di Paolo et al., 2004; Verstreken et al., 2009; Walter et al., 2017). The role of the PI(4,5)P₂ precursor phosphatidylinositol 4-phosphate (PI4P) at the synapse is less clear. Mounting evidence suggests that PI4P is functionally important and not simply a precursor for PI(4,5)P₂ ( reviewed in D'Angelo, Vicinanza, Di Campli, & De Matteis, 2008; De Matteis, Wilson, & D'Angelo, 2013; Tan & Brill, 2014 ). Phosphatidylinositol 4-kinase (PI4K) is responsible for the synthesis of PI4P, which is subsequently converted to PI(4,5)P₂ by phosphatidylinositol 4-phosphate 5-kinase (PI4P5K).

PI4K activity is associated with SVs and pharmacological inhibition of PI4K activity in intact synaptosomes attenuates evoked release of glutamate (Wiedemann, Schäfer, Burger, & Sihra, 1998). PI4KIIα accounts for the majority of PI4K activity in brain extracts (Guo et al., 2003) and is associated with SVs (Takamori et al., 2006). Mutant mice lacking the PI4KIIα catalytic domain are viable but display age-dependent axonal degeneration in the spinal cord (Simons et al., 2009). It is not entirely clear if these mice display defects in axonal growth earlier in development. Furthermore, the role of PI4KIIα in SV exocytosis or endocytosis was not characterized.

Drosophila encodes one PI4KII and two PI4KIIIs (Brill, Hime, Scharer-Schuksz, & Fuller, 2000; Burgess et al., 2012; Tan, Oh, Burgess, Hipfner, & Brill, 2014). The C-terminal kinase domain of Drosophila PI4KII is homologous to both human PI4KIIα and PI4KIIβ, whereas N-terminal regions are not conserved (Burgess et al., 2012). The Drosophila larval neuromuscular junction (nmj) is a well-established model system for studying nerve terminal growth and SV exocytosis and endocytosis (reviewed in Harris and Littleton, 2015). Phosphoinositides have been shown to regulate these processes at this nmj (Forrest et al., 2013; Khuong et al., 2010; Verstreken et al., 2009). Here, we used a Drosophila PI4KII null mutant (Burgess et al., 2012) to characterize the role of PI4KII in nerve terminal growth and SV exocytosis and endocytosis at the Drosophila larval nmj.

Material and methods

Fly stocks

Fly stocks were grown in uncrowded conditions at 22 °C on standard molasses based fly medium. Wandering third-instar larvae were used for all experiments. All flies used in this study were in a w¹¹¹⁸ background and were previously described (Burgess et al., 2012). w¹¹¹⁸ flies were used as a control strain. PI4KII null mutants contain a deletion that removes the coding regions of PI4KII (CG2929) and CG14671 (Burgess et al., 2012). They also contain a P{w⁺, CG14671} genomic rescue. mCherry-PI4KII^wt is a rescue transgene that expresses wild-type PI4KII under control of the alpha1-tubulin promoter in a PI4KII null background (Burgess et al., 2012). mCherry-PI4KII^ATP is a rescue transgene that expresses a catalytically inactive PI4KII under control of the alpha1-tubulin promoter in a PI4KII null background (Burgess et al., 2012).

Immunohistochemistry

Immunohistochemistry was performed as previously described (Romero-Pozuelo, Dason, Atwood, & Ferrús, 2007). Briefly, fixed preparations were incubated overnight at 4 °C with FITC-conjugated anti-horseradish peroxidase (HRP, West Grove, PA) antibody (1:800 dilution; Jackson ImmunoResearch) to visualize neurons and the mouse monoclonal bruchpilot (brp) antibody (1:100 dilution; Iowa Hybridoma Bank, Iowa City, IA; Wagh et al., 2006) to visualize active zones. All primary antibodies were diluted in blocking solution. Preparations were mounted in Permafluor (Immunon, Pittsburgh, PA) on a glass slide with a cover slip and viewed under an Olympus FV1000 laser-scanning microscope (Heidelberg, Germany) with a 60× oil-immersion objective (1.42 NA).

Electrophysiology

Intracellular recordings were performed as previously described (Romero-Pozuelo et al., 2007) in HL6 saline (Macleod, Hegstrom-Wojtowicz, Charlton, & Atwood, 2002) supplemented with 1 mM CaCl₂. Briefly, the ventral longitudinal muscle fiber 6 (abdominal segment 3) of dissected larvae was impaled with a sharp glass electrode filled with 3 M KCl (∼40 MΩ) to measure spontaneously occurring miniature excitatory junction potentials (mEJPs) and stimulus-evoked excitatory junction potentials (EJPs). Cut segmental nerves were stimulated at 0.05 Hz using a suction electrode. Electrical signals were recorded using the MacLab/4S data acquisition system (ADInstruments, Colorado Springs, CO).

FM1-43 imaging

FM1-43 experiments were performed using a Leica TCS SP5 confocal laser-scanning microscope with a 63× water dipping objective (0.9 NA). FM1-43 experiments were performed as previously described (Dason, Smith, Marin, & Charlton, 2010). Briefly, FM1-43 loading and unloading was induced by high K⁺ depolarization using the following high K⁺ saline: 25 mM NaCl, 90 mM KCl, 10 mM NaHCO₃, 5 mM HEPES, 30 mM sucrose, 5 mM trehalose, 10 mM MgCl₂, 2 mM CaCl₂, pH 7.2 (Verstreken, Ohyama, & Bellen, 2008). Preparations were loaded with FM1-43 (Invitrogen, Carlsbad, CA) by high K⁺ depolarization for 2 min and were extensively washed for 5 min in Ca²⁺ free HL6. To reduce background fluorescence from extracellular FM1-43, 75 μM Advasep-7 was included for the first 1 min of the wash (Kay et al., 1999). After an image of FM1-43 uptake was taken, high K⁺ depolarization for 2 min was used to induce exocytosis. Another image was then taken to document FM1-43 unloading. The released fraction was calculated using the following formula: (fluorescence of load − fluorescence of unload)/fluorescence of load.

Statistical analysis

SigmaPlot (version 11.0; Systat Software, San Jose, CA) was used for statistical analysis. Statistical analyses were performed using unpaired t-tests. Error bars in all figures represent standard error of the mean (SEM).

Results

PI4KII restrains nerve terminal growth

We assessed whether PI4KII regulated nerve terminal growth by staining nmjs for HRP (a neuronal membrane marker) and brp (an active zone marker). We counted the number of synaptic boutons on segment 3 of muscle fibers 6 and 7 in PI4KII null mutants and controls. These muscle fibers are innervated by two types of axons that result in type 1b and type 1s boutons, which differ in both their morphological and physiological properties (Atwood, Govind, & Wu, 1993; Kurdyak, Atwood, Stewart, & Wu, 1994). These boutons can be distinguished anatomically: 1b boutons are 3–5 μm in diameter, whereas 1s boutons are 1–3 μm in diameter (Atwood et al., 1993). PI4KII null mutants had significantly more 1b and 1s boutons than controls (Figure 1(A, B)). In addition, the number of brp spots per nmj in PI4KII null mutants was significantly enhanced compared to controls (w¹¹¹⁸ – 235.60 ± 28.00, PIKII null mutants – 303.29 ± 15.41; p < .05). Thus, PI4KII plays a role in restricting nerve terminal growth.

Figure 1. PI4KII restrains nerve terminal growth. (A, C) Representative images of fixed third instar larval nmjs stained with an FITC-conjugated anti-HRP antibody and anti-brp antibody. (B) There was a significant increase (p < .001) in the number of 1b and 1s boutons in PI4KII null mutants (n = 14) in comparison to controls (w¹¹¹⁸; n = 15). (D) There was a significant increase (p < .01) in the number of 1b and 1s boutons in nmjs of larvae expressing PI4KII^ATP in a PI4KII null mutant background (n = 7) in comparison to larvae expressing PI4KII^wt in a PI4KII null mutant background (n = 9). Error bars represent SEM.

To determine whether PI4KII kinase activity is required for nerve terminal growth, we expressed wild-type or catalytically inactive PI4KII in a PI4KII null mutant background and examined their ability to rescue the nerve terminal overgrowth seen in the PI4KII null mutant. We found that wild-type PI4KII rescued the nerve terminal overgrowth seen in PI4KII null mutants, whereas catalytically inactive PI4KII was unable to rescue this phenotype (Figure 1(C, D)). Thus, PI4KII kinase activity is required to restrain nerve terminal growth.

Basal levels of synaptic transmission were not altered in PI4KII mutants

We next examined the role of PI4KII in neurotransmitter release by recording the compound excitatory junction potential (EJP) generated by tonic-like type 1b and phasic-like type 1s boutons from segment 3 of muscle fiber 6 in third-instar larvae in response to low frequency stimulation (0.05 Hz). We found no significant differences in EJP amplitude between PI4KII null mutants and controls (Figure 2(A, B)). Similarly, there were no significant differences in the amplitude (Figure 2(A, C)) or frequency (Figure 2(A, D)) of spontaneously occurring miniature excitatory junction potentials (mEJP) between PI4KII null mutants and controls. Taken together, these data suggest that PI4KII is dispensable for basal levels of synaptic transmission.

Figure 2. Evoked and spontaneous neurotransmitter release are not affected by the absence of PI4KII. (A) Representative traces of EJPs and mEJPs in controls (w¹¹¹⁸) and PI4KII null mutants. (B) There was no significant difference (p > .05) in EJP amplitude in response to 0.05 Hz stimulation between controls (w¹¹¹⁸; n = 7) and PI4KII null mutants (n = 5). (C, D) There were no significant differences in the amplitude or frequency of mEJPs between controls (w¹¹¹⁸; n = 7) and PI4KII null mutants (n = 5). Error bars represent SEM.

PI4KII mutants display impaired synaptic vesicle recycling

To determine if PI4KII is required for SV exocytosis or endocytosis, we used the lipophilic dye FM1-43 (Betz & Bewick, 1992) to monitor SV cycling. We assessed SV cycling in control and PI4KII mutants by measuring FM1-43 uptake by stimulating preparations with high K⁺ saline for 2 min in the presence of FM1-43 (Figure 3(A)). FM1-43 uptake was significantly reduced by 33% in PI4KII null mutants in comparison to the controls (Figure 3(B, C)). We next measured FM1-43 unloading by applying high K⁺ saline for 2 min to determine if this impairment was due to a defect in SV exocytosis and calculated the released fraction. We found that PI4KII null mutants released a similar fraction in comparison to the controls (Figure 3(B, D)) demonstrating that SV exocytosis was not impaired. Thus, the defects in FM1-43 uptake can be attributed to an impairment of SV endocytosis.

Figure 3. SV recycling is impaired in PI4KII null mutants. (A) Experimental protocol indicating time at which images were captured (indicated by arrows). (B, E) Representative images of presynaptic terminals loaded with FM1-43 during high K⁺ for 2 min and subsequently unloaded with high K⁺ stimulation. Preparations were then washed in 0 mM Ca²⁺ HL6 (with 75 μM Advasep-7 for first minute) for 5 min to remove extracellular FM1-43 and fluorescence was measured. Then high K⁺ saline was reapplied for 2 min to cause unloading and fluorescence measured again (unload). (C) PI4KII null mutants took up significantly less FM1-43 than controls (n = 7; p < .05), demonstrating impaired vesicle cycling. (D) A similar fraction of FM1-43 was released in controls and PI4KII null mutants, demonstrating that exocytosis was not impaired (p > .05). (F) Larvae expressing PI4KII^ATP in a PI4KII null mutant background took up less FM1-43 than larvae expressing PI4KII^wt in a PI4KII null mutant background (n = 6; p < .05). (G) A similar fraction of FM1-43 was released by larvae expressing either PI4KII^ATP or PI4KII^wt in a PI4KII null mutant background, demonstrating that exocytosis was not impaired (p > .05). Fluorescence (F) was reported with background F subtracted. Error bars represent SEM.

To determine whether PI4KII kinase activity is required for SV recycling, we expressed wild-type or catalytically inactive PI4KII in a PI4KII null mutant background and examined their ability to rescue the impaired SV recycling seen in the PI4KII null mutant. We found that wild-type PI4KII rescued the impaired FM1-43 uptake seen in PI4KII null mutants, whereas catalytically inactive PI4KII was unable to rescue this phenotype (Figure 3(E, F)). The released fraction of FM1-43 following high K⁺ stimulation was similar between genotypes (Figure 3(E, G)). Thus, PI4KII kinase activity is required for SV recycling.

Discussion

Our study uses a PI4KII null mutant to characterize the effects of loss of PI4KII on nerve terminal growth and SV cycling. We show that PI4KII functions to restrain nerve terminal growth and is required for SV recycling.

PI4KII and nerve terminal growth

Three PI4Ks are found in Drosophila, PI4KII (Burgess et al., 2012), PI4KIIIα (Khuong et al., 2010; Tan et al., 2014; Yan, Denef, Tang, & Schüpbach, 2011), and PI4KIIIβ (Brill et al., 2000). Detailed expression data for all three PI4Ks are lacking and it is not known whether these three PI4Ks are expressed at the Drosophila larval nmj. However, the effects of all three PI4Ks on nerve terminal growth at the larval nmj have now been characterized. We found that nerve terminal growth is increased in the absence of PI4KII and that PI4KII kinase activity is required for this effect (Figure 1). Similarly, presynaptic knock down PI4KIIIα increased nmj length (Khuong et al., 2010). The phosphoinositide phosphatase Sac1 predominantly dephosphorylates PI4P and presynaptic knock down of Sac1 reduces the number of synaptic boutons (Forrest et al., 2013). Collectively, these data suggest that PI4P levels negatively correlate with the number of synaptic boutons. Interestingly, nerve terminal growth was not altered in fwd (Drosophila PI4KIIIβ) null mutants (Dason et al., 2009). It is possible that fwd is not expressed at the larval nmj or that PI4KII or PI4KIIIα compensate for the absence of fwd. It appears that PI4KII or PI4KIIIα are not able to compensate for the loss of each other as both result in increased nerve terminal growth.

PI(4,5)P₂ has previously been shown to restrict nerve terminal growth at the Drosophila larval nmj. Knocking down PI4P5K or reducing PI(4,5)P₂ levels by expressing a pleckstrin homology domain of PLC_δ1 results in nmj overgrowth (Khuong et al., 2010). These effects were due to activation of presynaptic Wiscott–Aldrich syndrome protein/WASP. It is possible that PI4KII acts to restrain nerve terminal growth through this pathway. Alternatively, PI4KII may be acting through a PI4P-dependent mechanism that does not involve PI(4,5)P₂.

PI4KII and SV cycling

PI(4,5)P₂ has been shown to regulate SV exocytosis (Walter et al., 2017). Using a combination of electrophysiology and FM1-43, we show PI4KII is dispensable for SV exocytosis (Figures 2 and 3). Similarly, fwd mutants did not display any defects in neurotransmitter release (Dason et al., 2009). Pharmacological inhibition of PI4K activity by phenylarsine oxide in intact synaptosomes attenuated evoked release of glutamate (Wiedemann et al., 1998). The discrepancy between these results could be due to the other two PI4Ks compensating for loss of PI4KII or fwd in Drosophila. Functional redundancy was postulated as a reason for the lack of an effect on neurotransmitter release seen in PI4P5K or skittles mutants (Hassan et al., 1998). Alternatively, the pharmacological inhibitor, phenylarsine oxide, may have had non-specific effects on enzymes other than PI4Ks. Examining double or triple mutants of PI4Ks may help resolve these discrepancies.

We show that SV recycling is impaired in PI4KII null mutants (Figure 3). It is likely that this reduction in SV recycling in PI4KII null mutants would result in decreased neurotransmission during periods of sustained synaptic activity. Mammalian PI4KIIα is associated with SVs (Guo et al., 2003; Takamori et al., 2006) and generates the PI(4,5)P₂ precursor PI4P. PI(4,5)P₂ is required to recruit adaptor proteins and clathrin to endocytic sites for SV endocytosis (Di Paolo et al., 2004; Verstreken et al., 2009). Thus, it is possible that the reduced SV recycling in PI4KII null mutants is due to reduced PI(4,5)P₂ levels. Interestingly, cholesterol has been shown to regulate PI4KIIα activity in vitro (Banerji et al., 2010). Extraction of cholesterol from SVs impairs SV endocytosis (Dason et al., 2010) and prevents increases in the levels of PI(4,5)P₂ during SV recycling (Dason, Smith, Marin, & Charlton, 2014). It is possible that extraction of cholesterol from SVs reduces the activity of PI4KII, and this leads to reduced levels of PI(4,5)P₂. Thus, PI4KIIα and cholesterol together may play a critical role in SV recycling.

Nerve terminal growth and neurotransmission

Increased nerve terminal growth is associated with increased neurotransmitter release (Budnik, Zhong, & Wu, 1990; Zhong, Budnik, & Wu, 1992). However, these processes are not always tightly linked (Dason et al., 2009; Romero-Pozuelo et al., 2007, 2014). Here, we find that loss of PI4KII results in increased nerve terminal growth without any changes in neurotransmitter release. Similarly, reduced PI(4,5)P₂ levels results in nmj overgrowth with no changes in neurotransmitter release (Khuong et al., 2010). It is likely that homeostatic mechanisms are occurring to compensate for the altered nerve terminal growth (Stewart, Schuster, Goodman, & Atwood, 1996).

Conclusions

Our findings clearly show a role for PI4KII in nerve terminal growth and SV recycling. However, it is not clear if these effects are due to changes in PI4P or PI(4,5)P₂ levels. PI(4,5)P₂ has been shown to regulate both nerve terminal growth and SV recycling (Di Paolo et al., 2004; Khuong et al., 2010; Verstreken et al., 2009). Thus, it is tempting to attribute the changes we see in these two phenotypes to changes in PI(4,5)P₂. However, we are unable to rule out the possibility that PI4KII regulates nerve terminal growth and SV recycling through PI(4)P-dependent mechanisms. Future studies will need to target PI4P and PI(4,5)P₂ more acutely. This can likely be achieved using optogenetic approaches to acutely reduce phosphoinositide levels (Idevall-Hagren, Dickson, Hille, Toomre, & De Camilli, 2012) and by acutely increasing them using optical uncaging (Walter et al., 2017).

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by startup funds from the University of Windsor (to JSD) and by NSERC [RGPIN #06582; to JSD], by NSERC and Ontario Graduate scholarships (to JB), and by grants from the Cancer Research Society [#9059 and #11202], NSERC [RGPIN #262166] and CIHR [IG1-115714 and MOP-119483] (to JAB).