HIV-1 Tat protein inhibits neurosecretion by binding to phosphatidylinositol 4,5-bisphosphate

The human immunodeficiency virus type I (HIV-1) transcriptional activator (Tat) is essential for viral gene expression and replication (Berkhout and Jeang, 1989; Gaynor, 1995). Tat is an ∼11 kDa basic protein that is actively released by infected T-cells into the extracellular medium (Ensoli et al., 1990; Rayne et al., 2010a). Tat does not have a signal sequence and is unconventionally secreted (Rayne et al., 2010b). Circulating Tat in the serum of HIV patients can reach nanomolar concentrations (Westendorp et al., 1995; Xiao et al., 2000). Tat is also present within the central nervous system (CNS) since it can cross the blood brain barrier and be produced in situ by HIV-1-infected cells such as macrophages and microglia. Thus, Tat may be involved in HIV-1 neuropathogenesis (Banks et al., 2005). HIV-1-associated neuronal symptoms include motor or cognitive dysfunctions and behavioral changes affecting nearly 30% of patients undergoing intensive chemotherapy. Moreover, 10% of patients develop HIV-1-associated dementia at the late stage of AIDS (Ghafouri et al., 2006). The molecular details underlying these CNS perturbations during HIV-1 infection are scarce.

It is generally believed that extracellular Tat operates as a viral toxin (Huigen et al., 2004; Rubartelli et al., 1998) that can enter several types of uninfected cells and affect their physiology (Li et al., 2009). Tat internalization seems to involve different pathways depending on cell type. Indeed, Tat can enter HeLa cells using caveolar endocytosis (Tyagi et al., 2001), and T-cells using clathrin/AP-2-dependent uptake (Vendeville et al., 2004). In neurons, extracellular Tat interacts with the low-density lipoprotein receptor-related protein (LRP) and is then internalized into endosomal structures (Liu et al., 2000), where Tat concentration can reach 2 pg/ml (Deshmane et al., 2011). Upon delivery to endosomes, the amino acid tryptophan 11 initiates Tat insertion into the late endosome membrane (Yezid et al., 2009). Tat transmembrane transport is then catalyzed by the cytosolic chaperone Hsp90 that helps Tat to refold on the trans side of the endosomal membrane (Vendeville et al., 2004).

Phosphoinositides control the activity of many membrane protein complexes implicated in a variety of signaling and trafficking pathways including neurosecretion (Balla and Varnái, 2009). PtdIns(4,5)P₂ is a key phospholipid concentrated on the cytoplasmic leaflet of the plasma membrane where it functions as a recruiting platform for proteins that are essential components of the exocytotic or endocytotic machineries (Di Paolo and De Camilli, 2006; van den Bout and Divecha, 2009). The PH_PLCδ–GFP fusion protein has been widely used to visualize PtdIns(4,5)P₂ at exocytotic sites in neuroendocrine cells and PH_PLCδ–GFP overexpression inhibits exocytosis in chromaffin cells (Holz et al., 2000). Since Tat was recently shown to bind to PtdIns(4,5)P₂ with a high affinity (Rayne et al., 2010b), we examined whether Tat might affect neurosecretion by altering PtdIns(4,5)P₂-dependent processes. We show here that extracellular Tat is able to enter chromaffin and PC12 cells. Once internalized, Tat strongly inhibits hormone secretion. We observed that Tat modifies the localization of annexin A2, a major organizer of the exocytotic platform in neuroendocrine cells (Umbrecht-Jenck et al., 2010). Tat also affects the reorganization of the cortical actin cytoskeleton required for exocytosis in neuroendocrine cells. Thus, through its capacity to bind PtdIns(4,5)P₂, Tat profoundly disorganizes the exocytotic machinery leading to the inhibition of neurosecretion.

We first examined whether extracellular Tat could enter neuroendocrine cells. After a 30 min incubation on ice with recombinant Tat, PC12 cells were chased at 37°C for different periods of time. At time zero, Tat was mostly observed at the cell periphery, showing continuous staining along the cell membrane (Fig. 1A). This Tat labeling at the cell periphery was also detected in the absence of cell permeabilization (data not shown), suggesting that at this early time point, Tat is mostly associated with the extracellular face of the plasma membrane. After a 20 min chase at 37°C, Tat was not detected anymore at the cell periphery but on early endosomal structures labeled with EEA1 (Fig. 1A). Quantification revealed colocalization of Tat with EEA1, being maximal (50±1.8%) after a 20 min chase and thereafter decreasing (Fig. 1B). At later chase times, Tat was observed in vesicles labeled with the late endo-lysosomal marker Lamp-1: Tat/Lamp-1 colocalization peaked at 2 hours and then decreased to be barely significant after 4–6 hours (supplementary material Fig. S1). Interestingly, after 4–6 hours of chase Tat could be detected again at the plasma membrane (Fig. 1A, arrowheads).

To confirm that Tat transits to the plasma membrane, we performed double staining experiments with syntaxin 1A, a specific plasma membrane marker in neurosecretory cells (Fig. 1C,D). At 20 min, there was almost no colocalization between Tat and syntaxin 1A (9±1.4%). However, after 4 hours of chase, the presence of Tat at the plasma membrane was revealed by a significant increase in Tat/syntaxin colocalization (45±2%; Fig. 1C,D). Additional experiments were designed to analyze the inner face of PC12 native plasma membrane sheets. Tat clusters were detected on these membrane lawns only when cells were incubated with Tat and chased for 4 hours (Fig. 1E), suggesting that Tat associates with the inner leaflet of the plasma membrane. In agreement, permeabilization with saponin was required to detect Tat staining on intact fixed cells after 20 min or longer chase periods. Together, these data suggest that in neuroendocrine cells (similar observations were obtained with bovine chromaffin cells; data not shown), exogenous Tat can be internalized by endocytosis into early endosomal structures, reach late endosomes and subsequently escape to bind to the plasma membrane inner leaflet, with kinetics that are similar to those described in neurons (Liu et al., 2000).

Because Tat binds with high affinity to PtdIns(4,5)P₂ (Rayne et al., 2010b), a key lipid for exocytosis, we examined the effect of Tat on chromaffin cell catecholamine secretion using carbon-fiber amperometry. Bovine chromaffin cells in primary culture were incubated in the presence of Tat for 4 hours and then stimulated by a local application of nicotine. We found that the number of amperometric spikes reflecting exocytotic fusion events was strongly decreased in Tat-exposed cells compared to control cells (Fig. 2A). Tat decreased by more than 50% the number of spikes per cell evoked by stimulation with either nicotine or a depolarizing concentration of K⁺ (Fig. 2B). The kinetic parameters (rising and decay time, half-width) of the remaining spikes were, however, unaffected (Fig. 2C), indicating that Tat inhibited catecholamine secretion by decreasing the frequency of fusion events. We verified using TRM labeling, that Tat did not trigger apoptosis under these experimental conditions (data not shown).

The effect of Tat on PC12 cell exocytosis was assessed in cells expressing human growth hormone (GH) (Vitale et al., 2001). Because GH is stored in secretory granules and released by exocytosis, it can be used as a secretory activity reporter. Extracellular Tat inhibited ATP-evoked GH release from PC12 cells in a time- and dose-dependent manner (Fig. 2D,E). At 10 nM, Tat produced a significant inhibitory effect on GH secretion whereas the boiled Tat protein was without effect (Fig. 2E). Similar observations were obtained with PC12 cells stimulated with a depolarizing K⁺ concentration (data not shown). Note that the GH expression level (∼400 pg/10⁵ cells) was not affected by the Tat concentrations used in these studies (supplementary material Fig. S2).

We then examined the effect of transient Tat expression on ATP-evoked GH release from PC12 cells. Transfections enabled us to use various Tat mutants that were unable to penetrate neurosecretory cells. When transiently expressed, Tat inhibited ATP-evoked GH release from PC12 cells (see Fig. 5), suggesting that expressed Tat reproduces the effect of exogenous Tat incubation on regulated exocytosis. Transfected Tat also strongly inhibited GH release triggered by 20 µM free calcium in PC12 cells permeabilized with digitonin to bypass the secretagogue-evoked calcium rise (Fig. 2F). Thus, Tat inhibits secretion at a step located downstream of the rise in cytosolic calcium, probably through a direct perturbation of the exocytotic machinery. Accordingly, using Fluo-4 dye to measure cytosolic calcium levels, we found that the transient calcium rise in response to cell stimulation was similar in control and Tat-expressing PC12 cells (supplementary material Fig. S3).

When expressed in PC12 cells using a bicistronic plasmid that encodes both EGFP and Tat, Tat was found to colocalize with the plasma membrane marker SNAP-25 (Fig. 3A) and with PtdIns(4,5)P₂ at the cell periphery (Fig. 3B). A similar localization was observed in chromaffin cells (supplementary material Fig. S4). Interestingly, Tat was efficiently displaced from the plasma membrane to the cytosol by neomycin, a PtdIns(4,5)P₂ masking drug (Fig. 3C). A semi-quantitative analysis revealed that neomycin reduced by nearly 50% the peripheral Tat (Fig. 3C), indicating that the association of Tat with the plasma membrane depends on PtdIns(4,5)P₂ binding.

To further support the notion that Tat localization at the cell periphery is dependent on PtdIns(4,5)P₂ level, we co-expressed in PC12 cells Tat and a drug-inducible type IV 5-phosphatase (Béglé et al., 2009) or the PtdIns(4,5)P₂ 5-phosphatase synaptojanin to reduce PtdIns(4,5)P₂ levels. In both type of experiments, reduction of PtdIns(4,5)P₂ levels strongly affected the peripheral distribution of Tat, which became largely cytosolic (Fig. 4). Altogether, these observations indicate that, in neuroendocrine cells, Tat either expressed or applied exogenously and internalized by endocytosis, tends to associate in a PtdIns(4,5)P₂-dependent manner to the plasma membrane.

Next, we investigated whether the inhibitory effect of Tat on neurosecretion might be linked to its binding to PtdIns(4,5)P₂. To this end, PC12 cells were transfected with plasmids encoding different Tat mutants and probed for GH secretion. The expression levels of Tat mutants were similar to that of the wild-type Tat protein (supplementary material Fig. S5). As illustrated in Fig. 5, expression of Tat(49–51)A that does not bind to PtdIns(4,5)P₂ (Rayne et al., 2010b) did not affect secretagogue-evoked GH secretion, whereas expression of Tat(55–57)A, also mutated in the basic region but still able to bind to PtdIns(4,5)P₂, inhibited GH release like wild-type Tat (Fig. 5A). Moreover, secretory activity was unaffected by the expression of the TatW11Y mutant that preserves the native structure and transactivation activity of Tat (Yezid et al., 2009), but does not bind PtdIns(4,5)P₂ (Rayne et al., 2010b) (Fig. 5A). Note that Tat(49–51)A accumulated mostly in the cytoplasm, whereas Tat(55–57)A was found at the cell periphery like the wild-type Tat protein (supplementary material Fig. S4). Thus, the interaction of Tat with PtdIns(4,5)P₂ at the plasma membrane seems to be necessary for Tat to block exocytosis.

To confirm these results, we examined whether increasing the intracellular PtdIns(4,5)P₂ concentration by overexpressing the phosphatidylinositol 4 phosphate 5-kinase Iγ (PIP5KIγ) isoform could counteract the inhibitory effect of Tat on secretion. PIP5KIγ is the major PtdIns(4,5)P₂ synthesizing enzyme in neurosecreting cells, and its overexpression has previously been shown to increase the amount of PtdIns(4,5)P₂ at the plasma membrane by ∼5-fold (Béglé et al., 2009; van den Bout and Divecha, 2009). As illustrated in Fig. 5B, overexpression of PIP5KIγ did not significantly modify GH release from PC12 cells, indicating that PtdIns(4,5)P₂ concentration is not a limiting factor for secretion. However, overexpression of PIP5KIγ did significantly reduce the inhibitory effect of Tat on exocytosis (Fig. 5B). Note that this rescuing effect on Tat-mediated inhibition of secretion was not due to a downregulation of Tat expression (Fig. 5C,D). Altogether, these data strongly suggest that impairment of secretion by Tat is linked to its binding capacity to PtdIns(4,5)P₂ at the plasma membrane.

Neuronal exocytosis requires the presence of PtdIns(4,5)P₂ at the plasma membrane (Aoyagi et al., 2005; Milosevic et al., 2005; James et al., 2008; Wen et al., 2011). The amount of plasma membrane located PtdIns(4,5)P₂ determines various parameters of exocytosis, most likely through the interaction with proteins involved in docking, priming and/or fusion (Umbrecht-Jenck et al., 2010; Di Paolo and De Camilli, 2006; James et al., 2008; Rescher et al., 2004; van den Bout and Divecha, 2009). Thus, one hypothesis might be that Tat interferes with the PtdIns(4,5)P₂-mediated recruitment of constituent(s) of the exocytotic machinery to the plasma membrane. Annexin A2 is a well known PtdIns(4,5)P₂-binding protein recruited to the plasma membrane at early stages of exocytosis in chromaffin and PC12 cells (Umbrecht-Jenck et al., 2010). Annexin A2 is cytosolic in resting cells. Secretagogue-evoked stimulation triggers its rapid translocation to the plasma membrane (Fig. 6A), where it stabilizes lipid micro-domains required for exocytosis (Chasserot-Golaz et al., 2005). In Tat-expressing PC12 cells, neither ATP (Fig. 6A) nor Ba²⁺ or high K⁺ (data not shown) were able to induce the translocation of annexin A2 to the cell periphery. Semi-quantitative analysis (Fig. 6C) confirmed that annexin A2 recruitment to the plasma membrane was severely impaired in cells expressing Tat whereas expression of the Tat(49–51)A mutant which is unable to bind PtdIns(4,5)P₂ had no effect (Fig. 6B,C). These experiments suggest that, through its ability to bind PtdIns(4,5)P₂, Tat may also affect the lipid organizing activity of annexin A2 required for exocytosis.

Because PtdIns(4,5)P₂ is also able to bind and modulate several actin regulatory proteins known to mediate the actin redistribution required for exocytosis in neuroendocrine cells (Trifaró et al., 2008), we monitored actin depolymerization at the cell periphery using fluorescent phalloidin. As illustrated in Fig. 7, PC12 cell stimulation reduced peripheral phalloidin staining in line with the transient actin reorganization occurring in cells undergoing exocytosis. Expression of Tat impaired peripheral actin depolymerization (Fig. 7A,C), whereas Tat(49–51)A had no effect (Fig. 7B,C). In other words, Tat through its binding to PtdIns(4,5)P₂ interfered with the subplasmalemmal actin reorganization in stimulated cells, an effect that may also be involved in its capacity to impair the exocytotic response.

Tat protein plays a pivotal role in the control of transcription and replication of HIV-1 (Coiras et al., 2010). Although Tat function has been thought for a long time to be restricted to the control of viral RNA production (Chiu et al., 2002), recent findings indicated that Tat binds to PtdIns(4,5)P₂ with a very high affinity at the plasma membrane. This binding is required for Tat release to extracellular space (Rayne et al., 2010b). Strikingly, the capacity of Tat to be endocytosed and reach the cytosol in many cell types (Debaisieux et al., 2012) suggests that circulating Tat might interfere with key cellular processes relying on PtdIns(4,5)P₂ such as endocytosis, phagocytosis, exocytosis, cell migration and division (Di Paolo and De Camilli, 2006).

We describe here that Tat taken up by an endocytotic pathway or expressed by transient transfection inhibits secretion from neuroendocrine chromaffin and PC12 cells, probably through its binding to PtdIns(4,5)P₂ at the plasma membrane. We found that Tat affected the PtdIns(4,5)P₂-mediated recruitment of annexin A2, a key component of the exocytotic machinery. Tat also interferes with the peripheral actin cytoskeleton reorganization upon neuroexocytosis, thereby compromising the secretory response at a step following the rise in cytosolic calcium.

HIV-1-infected cells massively export Tat (Ensoli et al., 1990; Rayne et al., 2010a; Rayne et al., 2010b) and extracellular Tat can enter into various cell types via multiple endocytic routes (Liu et al., 2000; Tyagi et al., 2001). We report here that like in neurons (Liu et al., 2000), Tat is efficiently internalized in neurosecretory chromaffin and PC12 cells, making these two models suitable to study the effect of Tat on neurosecretion. Consistent with the Tat uptake pathway observed in other cell types (Liu et al., 2000; Vendeville et al., 2004), our results show that internalized Tat is first found in early endosomal structures as revealed by colocalization with EEA1, then in Lamp-1-positive structures. After 4 hours, Tat-containing vesicular structures were barely detectable suggesting that endocytosed Tat progressively dissipates into the cytosol, due to translocation across the endosome membrane (Vendeville et al., 2004; Yezid et al., 2009). Concomitantly, Tat signal became evident on the inner leaflet of the plasma membrane as revealed on plasma membrane sheets prepared from Tat-exposed cells. When transfected and expressed in PC12 cells, Tat localized essentially at the plasma membrane. The binding of Tat to PtdIns(4,5)P₂ is most likely responsible for this recruitment of Tat to the plasma membrane because both neomycin, a PtdIns(4,5)P₂ masking drug, and PtdIns(4,5)P₂ 5-phosphatases, which reduce cellular PtdIns(4,5)P₂ levels when overexpressed, efficiently displaced Tat from the plasma membrane to the cytosol.

We investigated the effect of Tat on neurosecretion using chromaffin cells to monitor exocytosis by amperometry (Bader et al., 2002; Bader and Vitale, 2009). In this cell type, Tat strongly inhibited the number of exocytotic events induced by two different secretagogues (nicotine, high K⁺), as revealed by a reduction in amperometric spike frequency. Tat also inhibited hormone release from intact PC12 cells stimulated by various secretagogues and from permeabilized PC12 cells in which the intracellular calcium level was buffered at 20 µM. These data indicate that the impairment of secretion was not the consequence of an altered calcium signal, but most likely resulted from a direct effect of Tat on the exocytotic machinery. Accordingly, using a fluorescent calcium dye, we found that Tat did not modify the secretagogue-evoked rise in intracellular calcium.

Using Tat mutants unable to bind PtdIns(4,5)P₂, we demonstrated that inhibition of secretion was directly linked to the interaction of Tat with PtdIns(4,5)P₂ at the plasma membrane. TatW11Y was especially informative since this mutant shows native transcriptional activity (Yezid et al., 2009), but no significant binding to PtdIns(4,5)P₂ (Rayne et al., 2010b). The fact that it did not modify secretion indicated that Tat transcriptional activity was not involved in the inhibition of secretory activity. Additionally, PIP5K was able to rescue secretion from cells exposed to Tat, confirming the idea that Tat blocks secretion through its capacity to bind to the plasma membrane-associated PtdIns(4,5)P₂ with high affinity.

We attempted to identify the PtdIns(4,5)P₂-binding components of the exocytotic machinery that might be displaced by Tat when present at the plasma membrane. Translocation of the PtdIns(4,5)P₂-binding protein annexin A2 to the plasma membrane following cell stimulation is a hallmark of chromaffin cell exocytosis. Annexin A2 plays an essential role in calcium-regulated exocytosis by promoting ganglioside GM1/cholesterol-containing lipid regions in the plasma membrane (Chasserot-Golaz et al., 2005; Chasserot-Golaz et al., 2010; Umbrecht-Jenck et al., 2010). We recently described that, at the plasma membrane, annexin A2 is found in the vicinity of docked secretory granules, colocalizing with syntaxin-1A in PtdIns(4,5)P₂-enriched domains (Umbrecht-Jenck et al., 2010), and proposed that annexin A2 is a promoter or stabilizer of genuine exocytotic platforms in neuroendocrine cells. As shown here, Tat is able to displace annexin A2 from the cell periphery in stimulated cells, suggesting that the high affinity binding of Tat to PtdIns(4,5)P₂ interferes with the recruitment of annexin A2. Annexin A2 could in principle bind to other acidic lipids present at the inner surface of the plasma membrane such as phosphatidylserine (PS). However, compared to PtdIns(4,5)P₂, Tat does not significantly interact with PS (Rayne et al., 2010b). The fact that annexin A2 recruitment is severely impaired in Tat-expressing cells is also an indication that binding to PtdIns(4,5)P₂ is most likely the trigger for the recruitment of annexin A2 to the plasma membrane in cells stimulated for exocytosis. Impairment of annexin A2 translocation in cells exposed to Tat might led to a reduction in the number of functional exocytotic sites and, as observed here, to a concomitant decrease in the number of exocytotic events.

The cortical F-actin network constitutes a negative clamp on the movement of secretory vesicles to release sites, and it must be locally disassembled and reorganized to allow translocation of secretory vesicles in preparation for exocytosis (Vitale et al., 1995). During stimulation, the entry of Ca²⁺ activates scinderin and gelsolin, which in turn provoke the disassembly of the cortical F-actin network. The activity and the binding to actin of these two F-actin severing proteins that are constituents of the exocytotic machinery have been shown to be regulated by PtdIns(4,5)P₂ (Trifaró et al., 2008). Thus the observation that Tat hampers the actin cytoskeleton reorganization during neurosecretion may also be a direct consequence of Tat binding to PtdIns(4,5)P₂.

Altogether, our findings suggest that the inhibitory effect of Tat on neurosecretion is linked to its high affinity binding to PtdIns(4,5)P₂, which interferes with important components of the exocytotic machinery relying on PtdIns(4,5)P₂. Inhibition of secretion was observed at nM concentrations of Tat, comparable with the concentrations reported for circulating Tat (Westendorp et al., 1995; Xiao et al., 2000), thereby arguing for the physiological significance of our findings. Yet, the ability of Tat to inhibit the exocytotic machinery, once present at the inner face of the plasma membrane, remains an intriguing observation. Intracellular PtdIns(4,5)P₂ concentrations have been estimated in the micromolar range (McLaughlin et al., 2002). It is therefore difficult to postulate that Tat may titrate the entire pool of cellular PtdIns(4,5)P₂ and impact its availability, especially when exogenously supplied. One hypothesis is that Tat efficiently binds and concentrates at specific location of the plasma membrane, like the exocytotic sites characterized by a specific protein and lipid composition. This would confer a preferential titration of the PtdIns(4,5)P₂ at these sites and directly impair the exocytotic machinery. Alternatively, Tat may efficiently activate PtdIns(4,5)P₂-catalyzing pathways or inhibit the synthesis of PtdIns(4,5)P₂. Either way, these combined effects at exocytotic sites during stimulation may efficiently starve the exocytotic machinery from the required PtdIns(4,5)P₂.

HIV-1 infection frequently leads to neurological disorders such as depression, motor dysfunction, and cognitive problems (Li et al., 2009). Because HIV-1 does not infect neurons, indirect mechanisms are probably responsible for neurological problems (King et al., 2006) and likely involve viral proteins like Tat (Ghafouri et al., 2006). Indeed, Tat can be toxic to neurons via glial cell activation (Ghafouri et al., 2006; King et al., 2006). Tat can also induce reversible synapse loss in primary neuron cultures (Kim et al., 2008) and completely capsize neuron/glia ratio in the hippocampus (Fitting et al., 2010). In addition, it is becoming increasingly apparent that Tat can directly alter neuronal functions (Neri et al., 2007). For instance, Tat is able to activate NMDA receptors by receptor phosphorylation (King et al., 2010; Li et al., 2008; Neri et al., 2007). Tat also appears to have an impact on dopamine (Ferris et al., 2009), norepinephrine (Longordo et al., 2006), glutamate and GABA release (Musante et al., 2010). Moreover, it was recently shown that Tat application increased dorsal root ganglion neuron excitability (Chi et al., 2011). Although these findings suggest that Tat is able to affect neurotransmission, direct evidence of a Tat effect on the molecular process of exocytosis has been missing.

We demonstrate here for the first time that Tat directly affects exocytosis of large dense core secretory granules in neuroendocrine cells as early as a few hours after application. Using the classical synaptic vesicle (SV) dye FM1-43, we observed that Tat prevented vesicle loading in primary neurons (supplementary material Fig. S6). Since SV loading requires efficient exocytosis and subsequent endocytosis, these findings suggest that Tat also affects the exo-endocytosis cycle in neurons. Therefore, it is likely that Tat perturbs SV trafficking and thus influences neurotransmitter release, as suggested earlier (Neri et al., 2007).

HIV-1-infected patients not only develop AIDS, but also several other serious related pathologies. These include lypohypertrophy due to decrease in growth hormone secretion (Leung and Glesby, 2011), pubertal delay caused by the dysregulation of hypothalamic-pituitary axis (Majaliwa et al., 2009) and impaired thyroid functions (Hoffmann and Brown, 2007). All of these aberrant functions are linked to impaired hormone secretion, in line with the results presented here.

In conclusion altered neurotransmission and hormone secretion may contribute to the development of neurological dysfunctions in HIV-infected patients. Tat-mediated inhibition of neurosecretion is likely involved in the development of HIV-1-associated disorders that arise even in patients treated by highly active antiretroviral therapy (Ghafouri et al., 2006).

Fluo-4 direct calcium assay kit and FM1-43FX styryl dye were purchased from Molecular Probes. hGH ELISA kits were from Roche Applied Science, Lipofectamine 2000 from Invitrogen, and basic Nucleofector Kit from Lonza/Amaxa Biosystem. Recombinant Tat (BH10 isolate, 86 amino acids) was produced and purified from E. coli as described (Vendeville et al., 2004). Expression vectors coding for HIV-1 Tat or Tat mutants (BH10 isolate, 86 residues) were as described (Rayne et al., 2010a).

Mouse monoclonal anti-HIV-1 Tat antibody (sc-65912) and rabbit polyclonal anti-syntaxin 1A were from Santa Cruz Biotechnology. Mouse monoclonal and rabbit polyclonal anti-SNAP-25, and rabbit polyclonal anti-human GH were obtained as indicated (Béglé et al., 2009). Monoclonal anti-PtdIns(4,5)P₂ antibody (IgM) was purchased from Abcam (2C11). Rabbit polyclonal antibody against early endosomal antigen (EEA1) was from Abcam. Alexa-Fluor-488, -555 or-647-labeled secondary anti-mouse and anti-rabbit antibodies were obtained from Molecular Probes. The GFP-synaptojanin and inducible 5-phosphatase plasmids were kindly provided by Drs P. De Camilli (Yale School of Medicine) and T. Balla (NICHD, NIH), respectively.

PC12 cells were cultured as described previously (Zeniou-Meyer et al., 2007). Freshly dissected primary bovine chromaffin cells were cultured in DMEM in the presence of 10% foetal calf serum, 10 µM cytosine arabinoside, 10 µM fluorodeoxyuridine and antibiotics as described previously (Vitale et al., 1993).

Treatment of PC12 or chromaffin cells with exogenous recombinant HIV-1 Tat protein or citrate buffer (vehicle, 50 mM sodium citrate, 100 mM NaCl, pH 7) was performed the day after cell seeding, or 72 hours after plating, respectively. Cells were incubated with 50 nM Tat for 30 min on ice to allow Tat binding to cells. Cells were then washed three times with Locke’s buffer (140 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl₂, 1.2 mM MgSO₄, 15 mM HEPES and 11 mM glucose, pH 7.4) and either immediately fixed (time 0 min or no chase) or put into the incubator at 37°C (chase) for various periods of time before fixation or cell stimulation. Using the in situ cell death detection kit, TMR red (Roche), we found that Tat treatment and overexpression did not induce apoptosis in PC12 cells.

pBi-Tat, phGH, and pPIP5KIγ expression vectors (0.8 to 1 µg total DNA, 15×10⁴ cells, 24-well plate) were introduced into PC12 cells using Lipofectamine 2000. Chromaffin cells were transfected by Amaxa Nucleofector (Lonza/Amaxa Biosystem) according to the manufacturer’s protocol (3 µg DNA, 5×10⁶ cells, 24-well plate). Under these conditions, transfection efficiency ranged from 60 to 75% for PC12 cells and from 20 to 40% for chromaffin cells. Co-transfection efficacy was systematically verified by immunostaining and was between 80 and 90%. Experiments were performed 24 hours after transfection.

Plasma membrane sheets were prepared according to the following protocol. After incubation with exogenous recombinant Tat for 30 min on ice, cells on coverslips were washed four times with Locke’s solution and, when indicated, incubated for 4 hours at 37°C to allow Tat internalization. Cells were then submitted to hypotonic shock with water to burst cells following the lysis-squirting procedure described previously (Ziegler et al., 1998). Plasma membrane sheets attached on the coverslip were fixed with 2% paraformaldehyde at 4°C during 2 hours and labeled with anti-Tat or anti-syntaxin 1A antibodies.

PC12 or chromaffin cells were washed twice with warm Locke’s solution and stimulated 5 min with Locke’s solution containing a depolarizing concentration of potassium (59 mM) or 300 µM ATP before fixation by 4% paraformaldehyde in 120 mM phosphate buffer and permeabilization with 0.025% saponin in phosphate-buffered saline containing 0.1% BSA at room temperature. Cells were further processed for anti-Tat (1∶200), anti-EEA1 (1∶500), anti-GH (1∶200), anti-SNAP-25 (1∶200), anti-syntaxin 1A (1∶150) or anti-annexin A2 (1∶200) immunofluorescence as described previously (Béglé et al., 2009). Fixation and anti-PtdIns(4,5)P₂ (1∶100) staining were performed on ice as described (Hammond et al., 2006; Hammond et al., 2009). Stained cells were visualized using a SP5II Leica or LSM 510 Zeiss confocal microscope equipped with a 63× objective. Median confocal sections were recorded and quantified as indicated below. Images of membrane sheets were obtained using SP5II Leica with 63× objective (NA1.4) and pinhole adjustment of 0.68 Airy units.

Electrochemical measurements of catecholamine secretion were performed 4–6 hours after treatment with recombinant Tat protein or vehicle (citrate buffer) as described previously (Chasserot-Golaz et al., 2005). Catecholamine secretion was evoked by applying K⁺ (100 mM) or nicotine (100 µM) in Locke’s solution without ascorbic acid for 5 seconds to single cells by mean of a glass micropipette positioned at a distance of 40–50 µm from the cell.

PC12 cells (15×10⁴ cells, 24-well plates, 80% confluence) were transfected with pBi-Tat (0.6 µg/well) and a plasmid encoding GH (0.3 µg/well) using Lipofectamine 2000. 24 hours after transfection, cells were washed four times for 5 min with Locke’s solution and then incubated for 5 min in Locke’s solution (basal release) or stimulated for 5 min with a depolarizing concentration of K⁺ or with 300 µM ATP. Supernatants were collected and cells were harvested by scrapping in 10 mM phosphate-buffered saline and lyzed by three freeze and thaw cycles. The amounts of GH secreted into the medium or retained within the cells were measured using an ELISA assay (Roche Applied Science). GH secretion is expressed as a percentage of total GH present in the cells before stimulation (Lam et al., 2008). GH release from permeabilized cells was processed as described previously (de Barry et al., 2006).

Confluent PC12 cells grown in 3.5 cm diameter dishes were transfected with pBi-Tat and either an empty vector or a PIP5K-expressing plasmid (4 µg total DNA). The day after transfection, cells were harvested on ice by scraping in lysis buffer (10 mM Tris pH 7.4, 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM NaF, 20 mM Na₄P₂O₇, 2 mM Na₃VO₄, 1% Triton X-100, 10% glycerol, 0.1% SDS, 0.5% deoxycholate and protease inhibitors). Total cellular protein (30–40 µg) extracts were analyzed on 4–20% polyacrylamide–SDS gels. Immunoblots were then revealed with anti-actin and anti-Tat antibodies. Blots were processed using the Super Signal detection system (Pierce) as described previously (Corrotte et al., 2006). Immunoreactive bands were quantified using ImageJ. Tat band density was normalized to that of actin from the same lane and expressed as means ± s.e.m. of three experiments that were performed in duplicates using three different cell cultures.

Mask images were created with the LasAF Software 2.3.5 (Leica) using the colocalization application provided. After background threshold determination to eliminate low intensity pixels, colocalized pixels were visualized on black and white overlay images (Masks). Immunofluorescence images showing Tat (Figs 3, 4), annexin A2 (Fig. 6) or phalloidin (Fig. 7) staining were quantified using ImageJ. The amount of Tat or annexin A2 was measured as the average fluorescence intensity normalized to the corresponding surface area (D_f) at the cell periphery (plasma membrane level) and in the cytoplasm. Results are expressed as ratios of D_f (cell periphery)/D_f (cytoplasm). The amount of phalloidin staining was quantified at the level of cell periphery and expressed as fluorescence intensities normalized to surface units. This allows a quantitative cell-to-cell comparison of the fluorescent signal. Colocalization analyses were performed with LasAF Software 2.3.5 (Leica) using the colocalization application provided. Precisely, the threshold of each channel was set individually and independently from the other channel. Cut-off settings of the background fluorescence were generally around 20%, except for the EEA1 antibody, where they were around 30%. The overlay image and the scatter plot is automatically generated by the software algorithm resulting in a table that included colocalization percentages and Pearson’s correlation coefficient (PCF) values.

Statistical significance was estimated using Student’s t-test and data were considered significantly different when P<0.05. Gaussian distribution of the data was verified.

We are grateful to N. J. Grant and S. Gasman for critical reading of the manuscript. We thank S. Debaisieux and H. Yezid for Tat mutants. We also acknowledge V. Calco and T. Thahouly for technical assistance.